Document RpB2YNYY1E59O1Mq5dQ6r9voB

ATTACHMENT VI ASSESSMENT OF ESTIMATED RISK RESULTING FROM FLATOXINS IN CONSUMER PEANUT PRODUCTS AND OTHER FOOD COMMODITIES BUREAU OF FOODS FOOD AND DRUG ADMINISTRATION January 19, 1978 CMA 049498 Aiihc-rr* dirsc- -reef .'X l -^ i V- : e::is: , -i:;r=:rry :=s:s .**r;s :e=:es::2:s: "urthar, epidez^elzi'/ studies in Thailand and pares :: Africa show a significant relationship between liver cancer incidence and estimated levels of ailac;:;ir. intake. Aflatoxins cannot be completely eliminated froa food produces without eliminating the products theasalves. T2A has been enforcing "V~Iinct--of -u parts-per t:___con ^ppol tor ac.aTdTcIhT'ih certain foods and is considering reducing that liner to 15 rob for peanuts and peanu: products (2? 7-t 12713) . Unfortunately, the direct risks, id any, to hunans cron ingestion of aflacouins cannot be aeasured. Therefore, risk assessnents nust rely on nathenatical treacnent of aninal roniooiegy and epideniology stucies. ouc.t ris.-; assessnents nave seen nade bv 7TA and are described herein. i- i can occur m a variety o: grams anc nuts anc can also occur in che near, nil/., and eggs of food aninals chat have consuned aflatcnin-ccntadning feeds. Com products and peanu: products only have been considered in the rush calculations since inclusion of Ml ocher products, because of contamination levels anc/or frequency o centaainaticn or consumption does not significantly affect che numerical results. emphasis has been placed on com products in che Southeastern Tr.ited States because both consumption and afiatonin contamination levels are relatively high escon. CMA. 049499 di-.i, i: is estinated that ir.rs.ica ranges :::r ir. average of*~2. 7 to a naxirua ci ?.0 nanograas/kilcgran "* body weisht/day *' (0.1 tc 0.3 parrs per 'fillion in the diet) ir. the Southeast. An extensive body of data exists on aiiatoxin toxicology in laboratory aninais. 3ecause it would be_ impractical to sake risk calcularicns for each such study, five studies, involving long-tern feeding of aflaroxins to three species of rats, were selected. (3a.ts in general, and rale rats of the ?i3her strain in particular, are considered the nost aflatrxin-susteptibia naznals.) 1'sir.g the mathematical procedure of Mantel and Bryan, estinates of lifetine liver car.cer incidence rates were derived for each study, and for the five studies combined. These estinates range free 30 to HCO per 100,000 for the individual rat studies and fren 2i0 to 1100 per 100,000 fer their ccnbinatier.. 3y way of comparison, risk estinates derived fron studies on rats and ocher species ccebined by Cornfield and co-workers yield a range fren 17 to 125 per 100,000. The actual lifetine liver cancer incidence rate froa all causes in the U.S. is epproximacely 151 per 100,000. Actual incidence races in Georgia and Alabana are approximately 103 and 100 per 100,000 respectively. (Bates for other Southeastern states are net readily available.) Tie fact that the risk estinates icr rats in general exceed the human ir.cider.ca rata fsr liver cancer fren all causes can be attributed in oart, tc the conset-ralive mathematical rrreedures enoloved. In CMA 049500 CMA. 049502 chat establishes a tolerance for aflaoonir.s ir. shelled peanuts and peanut t products used as human food. The regulation vould set a 15 parts per billicr. (poo) tolerance for total ailatoxins (3^ + 2^ ~ - C-j) ir. these products. Ailatonins are mold-produced :::n:.s v'nich are unavcicabla contaminants in peanuts and a number of other food commodities (1). Although the human health effects of chronic eirrosure to afiazonir-s are net ar.evn vith certain there is evidence -hich associates chronic dietary exposure to aflatonina vith the induction of primary liver cancer. In particular, several studies have indicted ailatcxins as animal carcinogens (2). A: present, i: is no: possible to determine a level of dietary enpesur to cartir.ogenic substances that vould not lead to a finite rish of human cancer. Therefore, assurance of absolute safety for the consumer (i.e., no risk.of cancer) is achievable only through complete elimination of carcinoge.nio substances from the diet and from ocher sources, acuever, um_iae chemicals intentionally added to food, unavoidable food contaminants cannot be prohibited from the diet, 'unless the feed susceptible to contamination i itself prohibited. This approach to the as3urar.ee of absolute safety for food unavoidably c0r.tar4ir.ated with af laconic vould mean the complete elimination of those food commodities chat have been 3hcvn to be susceptible to 3uch ccncminso These food coooocdities include com, peanuts, 3orrhum, rice, wheat, soybear vslcuts, eimor.es, pecans, pistachio nuts, 3ra2ii nuts, figs, and the edibl CMA 049503 -SSU* - 4. - -- -- - .*55= V C*.-3 ^a:-s3 r.o: e.. .::s :: :r,45 6 :rcc3 aflacoxir.s , a- alternative approach :: indiscriminately eliminating chs /j _ * diet ooth contaminated anc ncr.-ccncaminacec food is co prevent friz er.cerin; consular channels only these lots demonstrated co contain aflaccxir.. In the case of consumer peanut produces, aiiacoxins a: or above a -sve. o: : oarrs oar oillior. (rob) can ba measured with the degree of certainty recuirec :: *_A co initiate ra-- r ^ - V ^ ,a.j., seizure; :o pravanc cha marketing of aflatoxin-ccr.canir.aced paar.ucs. :ha goal "=----------* * ; T3.ns _c.z s co an enforcement laval of 5 rob, i.a., the lowest level ac which i: can be determined vich certainty whether aflaccxir. is cr is noc orasanc ir. paanucs. (I: should ba nocad cha; sflstcxir.s ac levels lass chan 5 ppb can ba decaccad in paanucs, buz cha validity of measurements at such levels is questionable.) Vhiia a level of 5 ppb nay represent the lowest enforceable toler ance consistent vich TZh' s consumer protection objectives ci minimizing dietary exposures to ailatoxins, 3uch a tolerance would be beyond the c bility of manufacturers to meat consistently. This is the case because aflatoxim levels within and among peanut lots, suer. variability racuires that oualitv control limits in manufacturing ba sa: lover than cha regulatory tolerance (i). "or example, consistent orocu; of lots conforming to the prepesad tclarar.ca cf 15 rob vould racuira cu control limits on the order of 3-7 pro. Invar tolerances vould require oro- pcrtior.atelv lovar oualitv control limits, k ttlaranca of 3 rrb vould thus CMA 049504 A -S * i - ~ --- 5u.: -r. ina-r.i_5, :r*a sense :v ;-/A. * '/' In brief, if o* eanu: o* roducrs are nor co be er.riralv' orchibirad. consumer exposure co aflaroxins from these produces can only be kapr below deraceabie levels (i.e., 5 ppb) by seizure (and a signiiiczr.c loss of produce), "or a higher rolarar.es (a.g., If or 10 ppb), the emphasis would be or. nar.ufacrurir.r qualiry cor.rrol (ar.d rhars would be lass loss o: producr). Because of che foregoing rharar rerisrirs of rha ror.rrol siruarior., mrormscion is neecac sass rr.a rumar. risus associs: o: ar^acomin exposure. Vr.rcrrunare_y , suer, informanor. is r.oz raaiily avail able. However, evaluacion of rr.a: infcmacicr. which is available from roxi- cology arc epidemiology scudias does help oo pur rhe risk quescion in'peri spacrive. In rha following, risk ascimares are derivad from animal rases and from epidemiology srudias. Suer, esrimares car.r.or be oonsrruad as represenrir. acruel human risk bur rhey do provide an indication cf rhe ralarive ocrer.rial roKiciry of aflaroxins. ::rro?i-iv:~cy ::zbo;o re?, risk assbssmz:" Iris ring cara indirara char the most likely pc ear. rial efface from chrome.: aflaroxin exposure is primary liver cancer (1,2). In all risk esrimares ro follow, i: is this disease char is being considered. An asrimara of rhe risk of primary liver cancer in humans due ro expesur; ro diecary aflazoxins requires two types of information, first, soma measure of the rouitirv of aflzicxir. ro humans muse be available, "or curooses o: CMA. 04 9505 thlS ajiSs s"e"` , r ~ * "5S- -a a s long-tern, lr.;-l, |rqc?. Iwe = i;v, v-a; is scujr.i is a cc ic-rasoor.sa cur* S -W ..3d*r^ ta..: relaticr.sr.cp bar-ear. the level of aflazonin exposure (the cose) and the iacider.ee of liver cancer (the response) in the exposed group. The nose relevant and least suspect cf such indorsation would be that obtained by direct axpariner.caricn in hut:an beings. Clearly no such data era available r.tr, for ethical reasons, could they ever be developed, a ** reliance rust be placed or. data resultir.i scucies ir. TCUC5 C 2 HU-aIIS s.7.0** - CC have been inadvertently s:cpcs*u :: r_'i:xir.s ; z: 2) dcrscc sitzsrizs"t in laboratory ar.inals. Epiderdclrjic*! studies, which should provide the nest relevant infomaticn obtainable, nonetheless have several disadvantages, remaps the rest inportar.t of which are rneip inherent Imitations in establishing clear causal relationships ar.d chair insensiciv-.ty in detecting sue: . s _ i ,, _ or.s,, a . ; tucies m e:cre re cent al ictnals are usually note sensitive mar. epideciological studies ar.d, when properly conducted, are capable cf escaolishir.g direct causation and dose-response relationships cf relatively high certainty. But they have bear, criticized on the grounds of being irrelevant to huaans. I: nust also be hep: in nine that epiceniclogical and aninal data obtained froa relatively snail, honoger.sus population groups requires, for estmatior. cf risits in an entire popu.atior., extrapo.ati co very large, heterogeneous groups of hurar.s. It is not known how to estinate AO O w of these tyres. CM* 004499550066 ... ------------------------ czncaminz the laval - s-rast -0- s;S3ju.s... <:w: -3 * tn v:::: :. = :z::t r.s occur, tr.a .ava,s c: az.atCMins m tr.csa trees, :ns * .crecuancy and arccuncs or huzan consurr.pcizn :f chose foods, and the length : * cine of e:rpcsura. As vith the toxicity information, the human exposure fa: in the risk assassner.r aquarian is also subject to such uncertainty. TTITAHY IMPOST?.! TO AJIATOMTTS :xins m .r.a ::oc c-asses r.-.a; * L,,, . C Scares are dr7 mlled corn oroducrs (a.:..j.jm neal and grits) and peanuts ar.i casn.uc products (3). exposure to ailaronins through ronrauinarad corn occurs mainly in cha sourness cam scacas (i. a.} North Carolina, South Carolina, Georgia, Aladana, Ilorida, Mississippi, and Louisiana;, vhara che contamination of cor* is prco; local oririn. On che oasis of ITA and YSIA survev daca, esrinar: cne average _avai oc coca_ ar_aro::ins m .eo tom duces das deer, in che rang* of 5-13 ppb (3,5,7). Com produced and consuned in ocher regions of che r.arion has a substantially lover aflatonin concent. Therefore, che I'm:ad Scares population groups rha: appear co be ac highesc risx, chose of che southeastern scares, have been salaocad for emphasis in :r.is ns/*, assessner.: Tc be conservative, che upper level of 1C ppb is used : n * h A r- s *: * -- d * u' " t> a s m - : z**t n cr.? coucr.cn 3ttsS , lte ttzcesscu r. w t ICCrt SUTV75 TDA ir.dicace ar. esci- -- 2A ^ :v5^3: 1 :a_ a: r.sucsr p$T*u. rrzcuzts (3,5,7). This esc mac ed level is equally applicable co population groups on a natctna. or casts. CMA. 049507 T.-e estimate: average .sviii or a:_tc::tr.s m peanut ortd.ots and cry milled corn products (Table C, =r; derived from su--.1'ays :: these :::- icar. as marketed es opposed to food items prepared : r r consumotion (i.e.. es eater.). Therefore, conversion rectors vara applied, where appropriate, to era amount of intake cr rood trams listed in Tibia II ro obrsir. rha amount >o: intake expressed in-r^ms of rha food items listed in Table I. Calculation of the average and maximum daily intakes of foods suscaptibl to ailatcxin contamination also requires knowledge of the frequency a: which these foods are consumed. Dirac: knowledge of the frequency of consumption of these foods is not available. To estimate daily consumption, it was necessary to make use of the information that the total population in the South surveyed by VSTA consisted of o,253 persons, 3~0 of whom had eaten peanut products and 1,550 of whom had eater, torn products during the 21-hour period preceding the questioning. Since it would be expected that these individuals wculc not consume these foods every cay, the ratio of "eaters" for one day versus "non-eaters" for or.e cay was used to establish the esti mated frequency at which these foods were eaten. This ratio, or frequency factor, was applied "to both the average amount imd to the amount consumed by the 90ch peroenoile of the "eaters" surveyed to derive an estimate of average and maximum dailv intakes. These estimates are oresented ir. Table III. A1LZZ I--: estimated Consumption Xares Tor The Scutheasoarr. Tr.ioad States Of Toed Products k>.ich Mac- Contain Atlatouir.s. One cay average intake (grams) acts uom ? 15 53 One day maximum (?C:h percer tile) intake (grams) 11A ioo :r rrecuar.cy or rtnsumrnon .090 estimate: ever -i- --. Jr.take (grams) .scimatac maximum cai_v mta.-te 1.grams) 10 CMA 049510 eor.cainae ir. Table car. ;e conbir.ee vi:h cr.a levels c: irlaooxir. zzr.zs.~L: bi.cn (Tibia 1'; :: esrinsie average arc naxinun human a xp c su r e :: aflarrxir.s ,in che souchaascarr. :e:t cf cha ccurery (che exposura ir. all ocher regions i being lower). The rasulcs o: chase calculations are presences in Table TV. Also shown are che escinarsd average and maximum human incakes o: diecary aflaccxin expressed in ng/kg body weighc/day and in pares per billion (??b) in che diec. CAnll IV: oscirarad Avaraze and Maximum Tailv Aflacoxnn Ir.cake In The ocur r.eas: CCD CLASS Pear.ucs and Paar.ur Produces DAILY ATLATOTIV ITT.Ill Averaze ,.<_!/ _a/ maximum 1/ r~ nz' .<2 2/ . 3/ odo 3 0.15 0.0C5 20 0.37 0.013 Com Produces HO 2.33 0.093 170 3.53 lie 2.73 C.093 190 9.03 0.323* 2 1/ ng * namegrans (1 1C grans) 2 / r%/ a2 Wclg-: ; V4 rc2 - 3seCc*'"* "v-5"" 3^.3 _3/ ??b ir. ci*:; iver&j- d = iiy rood ineaxa or 1500 g CMA 049511 a Ar. ina1 Toxicity Dm: "he toxic of ailacoxir. have hear, .examined in several studies in a variety c: species end strains of experi- mental animals. These include several strains of rats as veil as mice, marmosets, tree shrevs, trout, ducks, rhesus monkeys, and hamsters (2). The literature is much too extensive to cover exhaustively in this report. Therefore, selected studies are employed in the following to illustrate the range of results obtained in animal studies. Generally, the rat, and in particular the rale Tisher strain rat, is considered the mammalian species that is most sensitive to aflaroxm carcinogenesis (2). "or this assessment, five studies involving three strains of rats were examined ar.d the statisti cal procedure developed by Mantel and Bryan (?) as modified (10) was app.lied to the experimental results. This statistical procedure is one of several available which allow the estimation of carcinogenic risks associated with dose levels e: the carcinogen below those necessary to yield a positive carcinogenic response in the animal experimentation. In the present example, the levels of interest are the average ar.d maximum intakes, C.l and C.3 ppb atlaConors, respectively (see Table 17). Mo experiments have been conducted in rats ac these low dose levels, so no direct measure of risk cam be made; but, as noted, the Mantel-3ryan procedure can be applied to the available animal data to estimate the risks associated with the 0.1 and 0.3 ppb dose levels. It must be kept in mind tr.at the Mirttl-Sryan procedure estimates animal, r.c: human, risks. The relative sensitivities of humar.s and, in the present case, rats, remains to be considered. The results of the risk assess ment in rats are tresented in Table 7. CMA 049512 As ;s- be Ui:. :::* Table v, : - * is nna: it Uterine liver incidence ruts vary viiely ar.cr.z ::.t it.rus, ranging :::: a l:v : 0 to * hi gn o: 1130 per 130,ICC. An;tret indicaticn ci the inherent * , variability' in these e:ti*>e riner.ts is n4 rcvided by the u*r*rer 99.1 cnr.rider.: / liaits cr risk shewn in parentheses in Table V. TABLI V: Estimated Lifetime Cancer Aisles Ir. Laboratory Animals (Rats)-- Incidence Rate Per 100,0001/ DATA 5CVP.CI AJV.-.T0ZI" (0.1 orb) (0.3 oob) (11) (12) (13) (U) Co) 2/( SO)-7 30( 220) ?0( 500) H0( S20) 320(1100) CCM3IMI0 .UT 3TTDII3 CCMZIMIT ANIMAL STTTIIS^7 210( 170) 17( -- ) 2,/C 320)-7 ISO( 990) 360(2300) 550(2100) HOO(iSCC) 1130(1900) 125( -- ) 1/ Incidence rates are based on the procedure cf Mantel and Bryan (9), as modified by Mantel, tz_ al. (10). 2/ Cannot be estinatad. !/ Numbers in rarer.theses are uooer 992 confidence Units. / Includes rats, rainbow trout, ducks, and tree shrevs. See tent. Risks art gnthitally estinattd. CMA 049513 The '-"per 59" cpzfiiar.ee limits or. the lifetime rij<i derived from 'the studies in Table V are act only conservative risk estina:es I - * they represea: a "vorst case'^escinate) , but also are useful far comparing the several scuries since the confidence Units cake into accaur.t differences in sample sire and ocher variables in eyperisen,tal_des1gn-------- -- . - Although the results of these five rat studies are too variable tc justify statistically their pooling, they have beer, combined in Zable V ta indicate an overall level af response ir. rat studies involving different 3 *s, Another appreach to aflatoxin risk assessment based cr. animal studies is that taken by Cornfield, al. (15). Combining data from experiments on the rzmbov trout, duck, tree shrev, and slue different studies on five different strains of rats, these authors fitted various mathematical medals. A prtbi: model vlth a slope of 1.2 prabios per log^Q css* appeared to jive the best fit. estimates from graphs in this paper yield lifetime incidence rates of approximately 17 and 125 per 1CC.CCC far cases o: 0.1 and 0.3 opb, respectively. As shown in Table 17, the average and maximum dietary intakes of ifla te -/.in contributed by peanuts are approximately 0.005 and 0.01 ppb, respec tively. For the combined rat studies, the risk estimates far chasa aflatoxir. levels are 2 and 7 per 1CC.0OC, respectively. Ctilinir.g the Cornfield approach yields risk estimates on the order of less than 0.1 per 100,000 (graphically estimated) for the 0.005 anc 0.01 ppb levels. CMA 049514 Me !: asseS3 che reasonableness :: erase arena! derived risk iszzzi.is, it is of interest zz ;n-para z'r.srz zz the zvsrz.ll hunan liver cancer incident: ' races zr. c.ne United States. Tie total crude incidence per year of primary liver canter iz the total United Seizes population, estimated in 19 = 9, -z-s * w* 2.2/lO5 (17). Tor purposes of conparisor. virh che Ilfetize anisai risk data, i: is necessary to break down che above crude yearly incidence race in cams of age-specific race$i,_;p.aul--p-ay-each-ege--specific'race by a aezor which reflaczs zhe orooortion cf the total liver cancer rite ezrtribuzad by uenbers of that age group, and then co nulziply each such agespecific rata by che expected hunar. lifespan of 70 years. Xecoabining the age-specific rates yields a calculated lifatine risk for liver cancer occurrence of 151 for ever/ 100,000 in the total U.i. population. Prinary liver cancer lifatine incidence rates .in Alabama and Georgia, eszinaztd in che sane way, are 100/10^ and 103/10^, raspectively (13). The rates for these two states, which are the only ones readily available for the south eastern states, era nose appropriate for ccnparison with the risks estinated iron the ar.inal data since they represent the region with the highest antici pated erpesura to afiacctcins. The risks calculated fron the rat cat* in Table 7 not only differ greatly among chenselves, deuonscraclng an intraspecies difference in sus ceptibility, but, at the 0.3 poo dose level, individually exceed the lifatine risks of prinary liver cancer in che hunan peculation of che United States fron ail causes. Possible animations for these differences are: 1) the level of hunan exposure to aflatoxin has beer, over-estinattd; 2) the har.cel3ryar. extrapolation procedure is overly conservative in this case; ar.b/cr 3) rats nay not be an appropriate node! for predicting aflatoxin-ir.duced CMA 049515 -Id primary liver canter i- humans. Vhile cats to rest :h* validity si the firs: or second ecrplmaticn are -or available, the third receives iczt cor roboration from studies shoving differences ir. aflaconic metabolism among various animal species, yosc relevanc of which are the differences between rats and humans observed _in vitro. Available _in vitro data indicate that the ability of an animal species to metabolize aflatoxin 3^ to a metabolite called aflato-icol (AH) is related to the sensitivity of the species to acute eilatonicosis amd-rrac--:he~rati'3 Jt recuctove aca oniiativt active. - ties, measured as the ratio of Aul concentration, to that of an oxidized metabolite known as Q. , is an index o: species susceptibility to the carcino genic effects o: 3^. Although iz vitro data for humans are extremely limits: the data available Indicate that, with respect to these modes of metabolic behavior, hunans are closer :o relatively resis cant species such as mice'and monkeys than they are to the highly susceptible rat strains (19). b. loldeniolorv Data: Tara reflecting actual consumption .of aflarcnir. over defined periods of tine by particular individuals who develop liver cancer ere not available. The data most closely representing that situation are those derived from studies in Thailand (20), Xenya (21), Mozambique (22), and Swaziland (23). ~cr two or more districts in each of these countries, average afiatonin content in samples of meals actually consumed were ceter- mined. These estimates cf average tflatcnin content were then correlated with official annual liver cancer mortality statistics. As previcusly starec, there are limitations to these epidemiology data, not the least cf which is their inherent inability to establish direct cause-effect relation ships. example, any comparison based cn these epidemiological data CMA 049516 .0- :i::s :r. :hi usu;?:;::. ::.i; sj::U tills i:i :i::it;t:ivt, t:.i: conranr-atior. levels are indicative :: test experience (i.e., 1C-2C vasts ego), arc that crate is no significant e:eposure to other possible causes of liver cancer. Virh respect to other possible causes, there is evidence that pyrrolidri-SiT.t"Ii"aToacT censuses in herbal cedicines and hepacicis-3 virus nay play a rola in the etiology of ::-tirLlive--tanner--ir-scuc-of-tdt-artas studies (2-,22). Thus, the appropriateness of applying the human euicetiiclcgy data derived iron these studies to low-incidence western countries is question able. Several other interfering factors, veil-described by van Ransburz (2i), further cloud the epidemiological data. The calculations sad* below should be considered against this background of 'uncertainty. The results of the various epidesiology studies have been standardized by van Rer.sburg (2-) and these data are presented in "idle 71. There is a highly significant positive correlation between cancer rat* and estimated level of aflatoxir. 3, intake in the geographical areas studied, and a linear relationship batveer log aflatoxin 3^ level and cancer incidence, expressed as follows: - s.--`9 uo6 -t - u.---o where y total cancer rate/lC^/ysar and x aflatoxin concentration in ng/kg/bocy veight/day. CMA. 049517 _1 TAIL I "I: Sunnaricec Results :: loidanitlcgy Studies Measuring rrirary ,. Liver Cancer rate and Arlatcxir Intake (ng/kg body vaight/day). (?rc= Reference 24.) Locality Liver Cancer Rate (10^/year) Estimated Aflatcxia Intake (tg/'kg body veighi/day Kenya - Eigr. Altitude Thailand - 3cr.gkn.la Swi2iland - Sighveid Kan7a - Middle Altituc Svarilard - Midcleveld Kenya - Lav Altitude Thailand - Katburi Svaailar.d - Lcvveld Me ranticue - Inhanbar.e 0.7 2.0 2.2 2.9 4.0 4.2 6.0 3.0 3.5 5.0 5.1 5.2 3.9 10.0 45.0 43.1 222.4 CMA 049518 Using the estimated average and ns:-:inur chronic exposures for dieta--- levels :: ailarov.irs in the scutheastem stares (2.7 and 5.3 ng/kg body waiter /da y, respectively, table I*') *r.d the lag relation between diecar;: aflatoxir. and csccer"ircider.ce derived free epiderioiogical studies, ;he calculated range in the expected crude cancer incidence/year is 0.37 to 3.74 per 100,000. the actual crude rares/lO^/year in Alabama and Georgia are 0.'97"aad 1. IS,~ respectively (18). _______ It should be noted that the abeve calculations involve a significant assumption that suite possibly is not the case, they assume that the back ground liver cancer rats (i.e., that rate cue to unknown factor(s) other than aflatoxir present in the areas of the epidemiological studies) is- ec/ua to that of the United States. Of course, there is also the inplicit assure tian that aflatoxir contributes to the rata of liver cancer in this country there is no evidenca that this is true. And certainly there are other sub stances in the envirenaenr, including the verxplace, that are known to be heparocoxic and could contribute to this disease rate. In an attempt to reduce the effects of this possible error, a differen type of risk estinate has been nade. Specifically, a linear, ror-logarichnic relationship, derivad fros the human epidemiological data tabulated by Peers and Linsell (25), allows estimation of the human cancer rata related to dietary aflatoxins (cr to ether interfering factors, since these epideriologic studies cannot conclusively show that the increased rate of liver cancer above background is due only to dietary aflatoxin). the relations;-.! is: y .106 x - 2.2, where y ir.d :: have the same reining 13 above and 2.2 is the oac/.grour.c cancer incidence/10"/year when x 0. CMA 049519 -- * 0" isit-laiis cf aisics at aitzstattvi toiz?.a:;ii3 Using the last derives relationship, theoretical risk ts::r:is have * made for various aflatoxin levels chat sigh: result froa alternative toleraaces. Tr.ese are shown ia Table VII and are based on the assumption that -the legal maximum level of aflacoxias in peanut products is directly pro portional to the average level of aflaroxins actually found in these product For example, the current FDA action level of 20 ??b leads to surrey estimate of average levels o: 2 ppb in peanuts and peanuts products and 10 ppb in dry milled ton products. This information is based or. data collected while the 20 ppb action level vas in effect for all food commodities. If one assumes a diract proportional relationship, a reduction of the 20 poo level to 15 ?p' for peanuts and peanut products could produce an avarage leva! of 1.5 ppb aflacoteir.s in this type of food. Similarly, a reduction of the tolerance to 10 pro and 5 ppb oould produce average levels of 1.0 ppb and 0.5 ppb, respectively. Since, for purposes of this assessment, the legal limit for com is net under consideration, the estimated average level of aflatoxins in this food is held constant. CMA 049520 CMA 0 4 9 I'AIII.K VII: Ks limaled l.l let Imp Oniierr Halm llased mi K|ildr.iiilul<i|>,y Similes tneldeiire ll;il ru IV r inn,mm U if till MM III l.eve | (I'l'1') I'KANIITS l- It Inks llenul l lii|; From tin i 1 y hilnkc^ Ave r.tj'e Man 1 iniini At 1 ill 11 Mill l.eve 1 <lt*) COHN - ttlr.kn Kecul | ln|> Pimm Dally C I *w1 Ave liiye Max 1 ihiiiii TOTAI. Hlnkn Hemiltlii|; Fiiim ll.-illy Average Max Imiim ) 1 . ' 1.1) n.'i l.l lt.lt n.'i 0. 3 2.7 2.1 l.l 0.7 in 19. I fi'i. .1 to 19. 1 fi'i. 1 in 19.1 f.A. 1 in 19. 1 fiA. 1 70.2 19.9 1 9. fi 19. A U 7.0 lift. A fiS.li tri.n S / tinned mi IVers and Linnet 1 (2fi) . See texl . -- / A II ill OM i II levels air Lltnse whlr.1i ml|*hl Iio expelled In |irainil n amt jieaiinl prodnel n If Die tiller nuee we re redne 1 1 Mill 1 III' t in real 20 |i|ili Cii IS, 10, and S ppli, respei:l Ively. / A f 1 ;ii ox 1 ti levels In eiirn |irudne 1 n assumed la lie an nvernp.c of 10 l>l>li. / nr I ili'iii < ill r per IUO, (ton |io|Mi I nl I mi. 20- ir. Ti:ls "II, :.'t :i:ti ever a d: exposure (72 years) are 5:i=i:tc t'tr zhe Southeast to range from 20.2 to -C 6T.G/IC3 i: the current 20 ppb accicn level; 15.5 to 56.1/I03 t: e 15 ppb tolerance, 19.6 to 65.6/105 at a 10 opb tolerance ana 19.A to 65.C/105 at a 5 opb tolerance. Thus, based on the epidemiological data (26), assessment of the estiaar risk of human e:cposure to varying levels of aflacoxins ir. peanuts and peanut products shevs relatively no significant gain in the protection of the oubli health (i.e., reduction in liver cancer) by adopting a tolerance of 5 ppb (currently tbs lovest enicrcasble tolerance) as opposed to FOA's proposed tolerance of 15 ppb. This analysis applies tc the population at highest risk, namely that of the southeastern United States. In ocher areas of the country, vfeere aflatoxln contamination of conn is not a significant prcblem, the health risks from aflatoxins art due, for the most part, to exposure to contaminated peanuts. The estimated lifetime risk to humans from this source of exposure only, based cn the same type of cal culations used to derive Table "I, ranges from 0.3 to 2.0/10^, if it Is assumed that all peanut products contain 1.5 ppb aflacotclns (under the pro posed 15 poo tolerance). This theoretical lifetime risk of cancer vould be reduced to 0.5-1.3/10^ at a 10 ppb tolerance and 0.3-0.7/10^ at a 5 ppb tolerance. ::: the convenience of the reader, the lifetime liver cancer incidence rates listed in the foregoing tables are summariced in Table "III. CMA, 049522 -51 , cirum srpected lifetime liver Career Incidence Pates Oca :o -.5 laccyir. I:c?cscTes Vr.der Curran: PDA Guideline c: 20 ??5 Incidence Pac-es1--/ Total S. - 1S1 per 100,000 Alabama - 100 per 100,000 Georgia - 103 per 100,000 Estimated Incidence Paces Due to Aflacoy.ins Trcm Cccbir.ed Pa: Studies --If . Total Consumption: Peanut Products Only: 240 to 1100 per 100,000 2 to 7 per 100,000 Prom Combined Animal Studies 2/ Total Consucpcion: 17 to 126 per 100,CCC Peanut Products Only: less than .1 per 100,000 rroa .ip...i.a..e...c..i.O....i.O....g..y.....s. .t.udies --3/ Total Consumption: Peanut Products 3r*i7: 20.2 to 67.0 per 100,000 1.1 to 2.7 per 100,000 see leitt, pages lo, Set Text, Table V. See Tent, Table Vi; CMA. 04 9523 (i) "W w # * * * Hesselrire, C. 7;., i-i Merino. (Ids), M*rtttxins ir. Hunan a?, z Ar.lna 1 Ha el re , Parha ca>: Pualishers, Par): Fores : Sauer.,-llluro 1?77; pc. 5-i;o. Q) Voean, C. N. CHj) AfUcoxtn Carcinogenesis, in Methods ir. Canter Research. Vol. VII, H. 3usch (id.), Acadenic Press, New York, New York, 33. 309-344. (3) Scalaff, L., Af la toxins--At. Overview, in Reference 1, ?p. 7-23. (4) Rodricks, J. V., and Scalaff, L., AfUtaxir. Residua* Froffl Contiainaced Feed in Idible Tissues of Faoc-Produting Ar.inals, ir. Reference 1, ??. 3 (3) "denser*, ?. Aflacaxin Control During "acd Processing af Peenucs, i Reference I, pa. 121-13S. (5) Lilieho;, 0. 3., end Sesselcire, C. V. , Aflacaxin Canrral During Plane Graven end nerves; c; Cam, in Reference 1, pa. lO'-llF. (7) Scalar:, 1., Occurrence af Mycacaxirs in "acd end Feed, in M*.*ta toxins a Ocher Funzsl Related Food Prablens, J. *. Rc cricks (Id.), Advences in Cheaiscry Series, No. 149, As. Chen. Soc., -eshingron, D.C., ??. 23-50. (S) U.5. Department c: Agriculture Surrey of Household "cod Ccnsunpcian in Indeed Scares, 1965-1556; Food Intake ef Individuals, Spring 1955 (con cur er capes). (9) Marcel, N. , and Bryan, R. , "Safery" Tesring cf Carrinager.ic Agents, J. 77a;1. Cancer Ins: , 27 , 455-470 (1951) . (10) Mantel, M., _e_r al_., An Inaraved Mancei-3r.*en Procedure far "Safery" Testing of Carcinogens, Cancer Research, 25. 36i-372 (1575). (11) Alfir-5carer, R. 3., r: el., Studies of Leng-Tem Adniniscration af Afl toxin ra Rat* is a Nature! Food Canranir.anr, J. Aner. Oil Chen. Soc.. 45, 493-497 (1959). (12) Nevbeme, ?., and 3ucler, if. H., Acute and Chronic Iffecrs of Aflacaxin cn the Liver of Doses tic and Laboratory Aninals, Cancer Res., 2?, 236-25C (1559). (13) Vegan, career. N., e_r al. , Carcinogenic Iffects af Law Dieter* Levels of Af. in Rats, Face Casnet. Toxical., 12, 531-555 (15*4). (14) Semes, J. Id., end 3utler, Cartinasenic Acriviry a: Afiatoxin to Rats, Nature 202, 1015 (1964). CMA 049524 :i5) Ne-.-beme, and bv 2letnylstilbastool in MaIla, Rats, - r t r. ' (1969) Cirzijr.L5 environ. r.azlzh, 1?, *o?~~7C (16) Cornfield, J., e_t _ai, Private Coenunitatior., 1577. (17) 3ioaarry Branch, National Car.car Institute, Preliminary Report, Taira National Cancar Survey, 1969 Incidence. National Institutes or Health, DET, 1971. (13) Stole::, L. , ar.d rriedmar., I., Information Bearing or. the evaluation of the Harare :o Man From Ariarosin Ingestion, PAG Bulletin, Vol. VI, No. 2, 00. 21-32 (1975). (19) Hsisi, 3., e_t _sl.. Comparative Metabolism of Aflatoxin, in Reference 1, oo. 37-30. (20) Shank, 3... e_t al_. , Dietary Aflatoxins ar.d Hunan liver Cancer IV, -cod Ccsr-et. Tsxitcl. . 12, 171-179 (1972). (21) (22) Peers, . , and linsall, C., Dietary Aflatoxins and liver Cancer--A Popu lation Based Study in Sanya, Brit. J,. Car.car. 27. 173-431 (1973). 4 ' `P` " ^ van Rarsburg, S. J. , e_t _al., Primary Liver Cancer and Aflatoxin in a Sign Cancer Area, _S. Afr Med. J_., _43, 2508a-230Sd (1974). (22) Peers, tz_ il. , Dietary Aflatoxins ar.d Hunan Liver Cancer, A study in Swaziland, I.-.s t. J_. Cancer, 17, 157-175 (1575) (24) van Rensburg, 5. J., Role of epidemiology in elucidation of Mytotcxir, Health Risks, in Reference 1, op. 595-712. ;axunle, T. M., et_ al. , Prinary Liver Cell Carsinctaa (P1CC) in the Norths Guinea Savanna of Nigeria, Irar.s. Roval Soc. Trso. Mad. Hva.. 71. 335-337 (1977). (25) Peers, and Lir.sall, C. A., Dietary Aflatoxins and Hunan Prinary Liver Cancer, Presented a: III IIP AC Symposium or. My to toxins in Foodstuffs, Paris, Sept. 1975. CMA 049525 CHEMICAL MANUFACTURERS ASSOCIATION EXHIBIT C Consultants Comments on Individual Criteria Documents 049526 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON ARSENIC Conclusion It is suggested that the fish data be utilised to generate the freshwater criteria. There should be no criteria established for saltwater organisms. Prepared By Albert C. Hendricks Route 4, Box 1023, Christiansburg, Virginia 24073 and John Cairns, Jr. 501 Bishop Road, Blacksburg, Virginia 24060 April 17, 1979 CMA. 049527 Sunimary Sheet ARESENIC EPA Criteria Freshwater - 57 ug/1 as a 24 hour average and 130 ug/1 as the maximum concentration. Saltwater - 29 ug/1 as a 24 hour average and 67 ug/1 as the maximum concentration. Residues - No Residue Limifed Toxicant Concentration could be determined for fresh or saltwater since no maximum permissible tissue concentration for arsenic is available. Conclusions It is suggested that the fish data rather than the invertebrate data be used for deriving a freshwater criterion for the following reasons: 1) There is larger data base for the fish and 2) A porti.on of the invertebrate data may or may not be LC50 data. Therefore, the following criteria is proposed to protect freshwater organism. 1331 ug/1 as a 24-hour average and the concentration should not exceed 3025 ug/1 at any time. There should be no criterion established for saltwater organisms since the data are meager and lack credibility. General Comments There is a fair data base for freshwater fish and a meager amount for freshwater invertebrates and saltwater organisms. The acute value for the freshwater invertebrate are used to establish the criterion for freshwater organisms rather than the chronic data. 1 CMA 049528 Specific Comments Page A-l. In surface freshwater in the U.S. the mean As concentration is 64 ug/1. 57 ug/1 of As is the 24 hour average. Those discharges using average surface water will obviously exceed the 57 ug/1 as a 24 hour average. Those discharges that utilize surface waters with As concentrations greater than 130 ug/1 would never be able to discharge. Page A-5. The top line of this page states that As has been shown to'bioconcentrate in fresh and saltwater organisms, however, on page B-6 the documents States that As is not bioconcentrated. Also, on C-4 there is a table that shows no bioconcentration. The statement on page A-5 is incorrect and should be deleted. Page B-3. The report by Hughes and Davis (1967) should be considered suspect since the As levels were not measured in the tests. With the large differences between their results and others it may be that they simply calculated incorrectly their final concentrations. Page B-3. Line 5 from the top states that "All results are express as arsenic." In the original articles the toxicant were expressed usually as arsenate organic arsenic, etc. Page B-4. Line 11 from the bottom begins a discussion of the geometric mean for Table 1. The EPA calculated a geometric mean of 17,239 ug/1 for freshwater fish. With the same adjusted values, I arrived at a geometric mean of 11,860 ug/1. It is not clear as to how EPA arrived at their geometric mean. Page B-4. Line 4 from the bottom begins a discussion of the data in Table 2 of the document. There are two points that need to 2 CMA 049529 be made about the table. 1) Sanders and Cope (1966) did not generate LC50 values for pulex or S. serrulatus rather they estimated 48- hour EC50 immobilization values for these organisms. 2) The geometric mean for their adjusted values should read 3,166 ug/1 and not 2,784 ug/1. Page B-5. The chronic toxicity tests gave values higher than the adjusted acute values, therefore, the acute values were used to establish the criterion. This again appears to be a case of attemptin,' to place the criterion at unnecessarily low levels. Page B-6. The bioconcentration factors were all derived from unpublished data, therefore, little value can be placed on them. Page B-S. The Final Fish Acute Value should read 3025 ug/1, The Final Invertebrate Acute Value should read 150 ug/1, The Final Acute Value should read 150 ut/1 and 0.44 times Final Acute Value should read 66 ug/1. Page B-12. The saltwater data are meager and lack credibility (no measurements were made), therefore, no criterion should be formulated Criterion Formulation It is suggested that the fish data rather than the invertebrate data be used for deriving a freshwater criterion for the following reasons: 1) There is larger data base for the fish and 2) A portion of the invertebrate data may or may not be LC50 data. Therefore, the following criteria is proposed to protect freshwater organism. 1331 ug/1 as a 24-hour average and the concentration should not exceed 3025 ug/1 at any time. There should be no criterion established for saltwater organisms since the data are meager and lack credibility. CMA 049530 AN EVALUATION OF THE EPA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON ARSENIC CONCLUSION: The criteria is appropriate Tor inorganic arsenic but the data does not support such a criteria for the organic arsenicals. The document should be revised in order to treat these two classes of arsenicals separately. Prepared by: Dr. Sorell L. Schwartz and Dr. Joseph F. Borzelleca Division of Toxicology Department of Pharmacology Medical College of Virginia Richmond, Virginia 23298 - 23 April 1979 CMA 049531 GENERAL COMMENTS The document, in the largest sense, represents a rather complete review of the literature on arsenic. Unfortunately, there appears to be an operating sense that everything ever written about the toxicity of arsenic must be reviewed. This makes the document cumbersome, causing substantial difficulties in interpretation. It is really ynnecessary to constantly repeat acute toxic effects and subchronic effects of As by individual reference to all of the studies reporting this information. The major problem with the document is a reflection of the attempts to be encyclopedic. That is, there is no clear distinction made between inorganic ana orgunic As. Whereas evidence for carcinogenicity is alleged for inorganic As, no such information is available on the organic arsonicals such as disodium methyl arsenic and monosodium methyl arsenic. A major problem in this regard is the problem of specition of arsenicals in soil, water, etc. This should be addressed in the document so tiiat the problem of criteria applying to inorganic As and those to organic As are dealt with separately. It might even be a good idea to deal with the. two general forms in separate sections, bringing the sections together when establishing criteria. This is a critical point and the clarification must be made if the document is to be useful. The means of derivation of the criterion per se seems to be O.K. con sidering the difficulty in dealing with an apparent human carcinogen in the absence of confirming animal data. -1- CMA. 049532 SPECIFIC COMMENTS The citation of Holland (1904) is a good example of overfill. Under Absorption, the statement in,if ['articles ^rnllcr than 0.1 urn do not cci'i'c not of suspension in tne innalod air stream is wrong. rhere is much newer data than that of Falk one! Kotin which shows that smal ici particles will impact on the alveolar surface. The estimate by Stockinyer and Woodward that 80T of inyested inorganic arsenic is absorbed, vs. 20of inhaled As, is not an estimate -- it is a ojgss. And one made with no evidence to suggest that it is a good one. -2- CMA. 049533 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON BENZENE * The criterion for freshwater is of dubious value and a criterion should not be set until better data are available. The criterion for saltwater should be based on empirical data rather than on extrapolation of freshwner data. This draft document is not as valuable as the previous document. Major criticisms include: 1) use of data not available for peer review; 2) high error in estimating final criterion values; 3) poor review of the pertinent literature; 4) bias in data selection, expecially for plants; and 5) mathematical errors in calculating geometric means. prepared by Arthur L. Buikemu, Jr. 606 Alleghany St. Blacksburg, Va. 24060 and John Cairns, Jr. 501 Bishop Rd. Blacksburg, Va, 24060 May 10, 1979 CMA. 049534 Aquatic Criterion For benzene the criterion to protect freshw.iter aquatic Life as derived using the Guidelines is 3,100 ug/ as a 24-hour average and the concentration Ghould not exceed 7,000 ug/l at any time. * The data base for saltwater aquatic life is insufficient to allow use of the Guidelines, The following recommendation is inferred from toxicity data for freshwater organisms. For benzene the criterion to protect saltwater aquatic life as derived using procedures other than the Guidelines is 920 ug/2 as a 24-hour average and the concentration should not exceed 2,100 ug(l at any time. General Comments EPA ia requested to read three documents previously submitted by the American Petroleum Institute (API) in response to the Water Quality Criteria "adjustment factors" published in the Federal Register (43[97Jr21506; 1978). These comments from API are on the public record and they address the prob lems in trying to correct an inadequate toxicity data base. It is unfortunate that EPA has chosen not to consider fate and effects which were included in the earlier draff document. Ambient levels of benzene that will impact on aquatic life will rarely occur because of its high vola tility and the rapidity of microbial degradation (e.g., Howard and Kurkin, 1974). The natural degradation of benzene in water was estimated at 200 to a 330 x 10 ug/t/day (Lee, 1977, Prevention and Control of Oil Spills 1977: Ml). The half life of benzene in a cubic meter of waLcr was estimated to be 37.3 minutes if mixing was complete (McKay and Wolkoff, 1973, Environ, ici. Tech. 7;611). To ignore these data while establishing water quality criteria based on "adjustment factors" is incredulous. Some of the toxicity data are from EPA documents which have not been 1 CMA 049535 I subjected to peer review. These data also lack credibility because they were not collected following published EPA methodologies. Specific Comments Freshvdter Organisms p B-l -* In reviewing the acute fish toxicity dnta the following references wore ignored: Liebmann, I960, Handbuch der Frischwasser und Abwasser Biologie. II. Oldenburg, Munich. Hubault, 1936, C. R. Seanc. Acad. d'Agric. France 22:130. Reichcnbuck-Klinke, 1967, Arch. Fish. Wiss. 11:122 Shelford, 1917, Bull, 111. State Lab. Nat. Hist. 11:381. Turnbull et al., 1954, Indust. Engineer. Chemistry 46:324. Even if the above data were included in the final document the questionable test methods used by the cited authors may invalidate their use. Table 1 Because the 24-hr LC50 values reported by Pickering and Henderson arc not significantly different from the 96-hr LC5Q values, why weren't the data corrected accordingly? If adjustments of data are to be used, they should be used accordingly, P B-l and Table 1 The geometric mean for fish acute toxicity data is wrong. It should be 22,348 pg/2. Consequently the final fish acute value should be 5,730 ng/t. The final fish acute value calculated to be 7,000 pg/i lias an approximate error in prediction of 1282. P B-l The Daphnla magna data should be available for public comment because this EPA funded study did not follow published EPA test methodologies. CM* 049536 3 In our opinion this is academically unsound, D. manna were not the only invertebrate tested. The mosquito Aedes aegyptl (Berry and Brammer, 1977, Environ, Pollut. 13:229) has also been exposed to benzene. p B-l and Table 2 How do you derive a mean for one value? After adjustment the error in prediction nay be as high as 293%. p B-2 and Table 3 Again the study is not readily available for public comment. Presumably the test methodology is also questionable (e.g., benzene concentrations may have been nominal). What good is the chronic value if no inhibiting levels were obtained? The possible error in predicting a final invertebrate chronic value is 95%. The explanation proposed to account for a chronic value higher than an adjusted acute toxic value is unbelievable. One of the two studies must be in error or this illustrates the problems in applying "adjustment factors". Another geometric mean of one number. p B-2 and Table k While the validity of the Chlorella data was questioned (Atkinson et al., 1977, Water, Air and Soil Pollut. 8:235), because ot poor test methodolog> it is our opinion that algal tests conducted with cotton stoppered bottles (which allows exchange with the air) more closely approximate what may happen in nature. Lastly, the lowest level of benzene reported to exert an effect on Chlorella was 500,000 and not 525,000 ug/I. CMA O49537 4 There are three other studies on the effects of hctizi.no to plants whfcli should be examined for applicability in setting water quality criteria. These include studies on yeast (Levan, 1947, Hereditas 33:457), Elodoa and Potamoftetoh (Frank et al.t 1961, Weeds 9:515; Currier and Peoples, 1954, Hilgardia 23:155). P B-2 Residues - The n-octanol/water partition coefficient is 134.9 and not 132 as reported (Chiou et al., 1977, Environ. Sci. Tech. 11:475; friend, et.al., 1977, Environ. Health Perspcct. 20:55). Bioconcentration (BCF) data for benzene do exist for marine fishes (see below). To discuss BCF daLa for benzene may be unwarranted because benzene and/or its metabolites are rapidly depurated (see below). p B-3 and Table 5 Why is the datum for Salmo trutta placed under miscellaneous when the 24-hr value could be corrected and placed in Table 1? p B-4 The final chronic value derived by multiplying 7,000 ug/l by another adjustment factor of 0.44 has an estimated error in prediction of 68X. The final 24-hr average concentration criterion has on accumulated error in estimation of up to 175Z. As such the calculated value should be raised accordingly if EPA persists in using these correction factors. Because the maximum concentration is estimated to be 7,000 yg/t and because the probable error of estimating this value is up to 1287., the final value should also be raised accordingly. CMA. 049538 5 Saltwater Organisms p B-10 We agree with the statement that unique effects may occur at benzene concentrations as low as 700 pg/i; however, more research is needed, Struhsaker (1977) did admit that adult fish were collected from San Francisco Bay which is known to be polluted. p B-10 and Table 6 In reviewing the acute fish toxicity, two references were omitted: Brocksen and Bailey, 1973, Proc. Joint Conf. Prev. Control Oil Spills: 783. Morrow et al,, 1975, Copeia 1975:326. There is an error in the geometric mean and it should be 8,147 pg/f.. Con sequently, after dividing by 3.7 the value should he 2,201 ng/C. The approxi mate error in predicting the Final Acute Pish Value is up to 175%. p B-10 and Table 7 Under invertebrate acute toxicity the following references were Ignored: Barnett and Kontogiannis, 1975, Environ. Pollut. 8:45. Bcnville and Korn, 1977, Calif. Fish and Came 63:204. Donohue et al., 1977, Environ. Pollut. 13:187. Eldridge et al., 1977, Calif. Nevada Wildlife Trans. 1977:90. Hubault, 1936, C. K, Seanc. Acad. d'Agric. France 22:130. LeGorc, 1974, Diss, Abst. B 35(7):3168, Potera, 1975, Diss. Abst. B 36(5):2010. Price et al., 1974, JWTCF 46:63. Neff et nl., 1976, l_n Sources, Effects and Sinks of Hydrocarbons in the Aquatic Environment, p. 66-83. Tatem et al., 1978, Est. Coast. Mar. Sci. 6:365. The estimated error in predicting the Final Invertebrate Acute Value is up to 757%. The datum in Table 7 for T1griopus is wrong. The correct reference is Barnett and Kontogiannis (1975) and not Korn et nl, (1976) who never studied CMA. 049539 this copepod. The calculated 96-hr LC.`>0 may be in error. Given the poor experimental design of this assay we feel that a 24-hr LC50 would be more accurate. p B-ll and Table 8 * The research reported f'or marine algae will not realistically apply to the real world because this research was conducted with sealed flasks. Volatile compounds, especially benzene, are rapidly lost to the atmosphere, etc. Why were the data for Dunaliclla left out of the document? At low concen trations benzene was stimulatory and at concentrations up to 100 mg/l benzene had no effect on Dunaliclla. The fact that these data were left out implies that there was bias in the selection of data. It is also possible that Amphldinlum growth was inhibited at concentrations of benzene as low as 0.001 yg/Z. The authors of the document selected the data by Atkinson et al. (1977) for Skeletonena for the lowest effect on plants (20,000 yg/Z). This is interesting because the authors are the same as those in the other paper referenced (Dunstan et al., 1975). Because a) the research in both studies was conducted in the same laboratory by the same workers, and b) Atkinson et al.'s research was conducted at the same time as Dunstan ec al.'s, why do these differences exist? This problem should be reconciled before placing confidence in the 20,000 figure. ilJL-A1. While no steady state BCF is available for benzene there are limited data to suggest that the estimated LCF of 24 may be too high. For example, after seven days marine eels concentrated benzene 15 times above ambient (Ogata and Miyake, 1975, Water Res. 9:1075). Other data are available for benzene and/or CMA 049540 7 its metabolites and these were included in the first draft but ignored in this draft. Why? Lastly, depuration of benzene is quite rapid no the dis cussion of a BCF of 24 probably is not warranted (Korn et al., 1976; Roubal et al.t-1977, Arch. Environ. Cont. Toxicol. 5:513; Struhsaker, 1977). The partition coefficient*for benzene is higher than reported in this document (see above). p r.-ll and Table 9 There is a myriad of other literature on benzene which has been Ignored by the authors of this document and they may be of value in setting wjter quality criteria. More data could be extracted from Struhsaker et al. (1974) and Struhsaker (1977). The paper by Korn ct al. (1976) was misquoted. The concentration of benzene was 6 y/. When converted to yg/ the value should he 5,300 and not 6,000. r B12-13 We agree that no salLwater criterion can be derived for benzene "because no Final Chronic Value for either fish and invertebrate species or a good substitute for either value is available." Why then do the authors propose to apply Dnphnla data which are both questionable and unavailable for peer review? To multiply the Final Invertebrate Acute Value by 0.44 is to increase the potential error in predicting the final value by up to 804%. The fl.nal 24-hr average concentration criterion has an accumulated error in estimation of <p to d04%. We suggest that the final overage con centration be increased accordingly. The final maximum concentration has an accumulated error up to 125% so CMA 049541 6 the final maximum value should also be increased accordingly. Summary in spite of earlier problems, this document is not as valuable or as scientifically valid as was the earlier draft. The major criticisms include: a) reliance on data for Daphnla which were poorly generated and not available for peer review, b) high additive errors in estimating final criterion values because of the inadequate data base used to generate "adjustment factors", c) poor review of the pertinent literature. Conclusion EPA should not persist in using "adjustment factors" to correct an inadequate data base. The errors in their calculations and data base are scientifically incredulous and indefensible. li EPA persists in using "adjustment factors" to correct an inadequate data base, they should increase the final criterion values by the* compounded error terms. Further, EPA should support research comparing predictions of final criterion values and empirically derived data to determine if their "adjustment factors" arc valid. Fate and effects of benzene should be considered before setting water quality criterion. CMA 049542 AN EVALUATION OF THE EPA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON BENZENE CONCLUSION: Though this is a well-written document, there is no scientific information presently available which can justify setting the criteria at this proposed level. Prepared by: Dr. Richard A. Carchman and Dr. Joseph F. Borzelleca Division of Toxicology Department of Pharmacology Medical College of Virginia Richmond, Virginia 23098 23 April 1979 CMA. 049543 GEN'SEAL COMMENTS This report presents a comprehensive and critical evaluation of the data on benzene. However, it. i4 inappropriate to set the criteria limit on the basis of ambient water level alone, since oral exposure represents only a minor route. Inhalation of benzene is a more significant problem. Oral intake may be important to toxicologically since adverse health effects of benzene are cumulative. Therefore ingested benzene, together with inhaled benzene, may poso a health problem. The authors indicate it would be appropriate to set the criteria level at zero. However, the authors suggest criteria based on the data available. SPECIFIC COVME.NTS C-7 Tr.e data demonstrate that only trace levels of benzene have been found in water, and then in only a small percentage of the sources analyzed. C-2 The reports on dietary intake of benzene conclude that this route of exposure is not considered to be a problem. However, fairly high levels are found in certain foods. This suggests a potential problem. C-7 Benzene is extensively metabolized mostly via the MFO in the liver. It is these metabolites that are responsible for its toxicity. Since this system is inducible by benzene itself, as well as other compounds, a potential problem that warrants important consideration is the interaction of benzene with other compounds. Benzene is metabolized to a highly reactive arene oxide intermediate. Since this compound covalently binds with cellular macromolecules, there is a CMA 049544 C-ll C-ll -12 C-12 C-13 C-15 C-15 C-15 C-16 C-I6 reasonably large potential for mutagenesis and carcinogenesis. This aspect needs further elucidation. What types and at what concentration were the effects observed on the G-I tract? At what exposure levels and doses were these effects observed? In Ahsoy's study, the workers were exposed to benzene and to other compounds. In Lange's study on the chronic effects of benzene on the immune system, the range of concentrations used was not presented. 'The study on the effects of benzene on cell renewal rate is questionable, without more details. Evaporation of benzene may have occurred during the incubation thus changing the concentration to which the cells were exposed. Also, it is important to know whether the cells were constantly fed and new drug added or just counted daily. As the authors point out, most human studies on health effects involve workers exposed to high concentrations of benzene as well as other compounds. The potential interaction of benzene acting either syncrgistically or antago nistically with other compounds is very important and warrants further investi gation. This could be a significant factor in setting the criterion level. No indication is given in Gofmekler's study on the route of exposure. It is not clear how subcutaneous injections of benzene is relevant in the evaluation of its health hazard. The primary route of exposure is inhalation, and the criterion is for ambient water levels. The fact that benzene was not mutagenic in the Salmonella/microsome in vitro assay is not a definitive evaluation of its mutagenic potential, since other known carcinogens are also false negatives in this system (i.e. metals). The relevance of subcutaneous injections of benzene as a test for mutage nicity in Kissling and Speck's study is unclear. The fact that the effects of benzene and toluene are additive is very CMA 049545 -3 C -16 C-I7 C-17 -18 C-IS -19 C-20 C-21 C-21 -22 C-24 important and should be considered in setting the final criterion levels. Ip. and subcutaneous injections do not seem relevant to benzene health hazards, except as potential j_n vivo assays. Under what conditions (i.e. dose, time, duration, etc.) were the experiments on cultured human leukocytes conducted? How are the workers exposed to benzene and toluene separated, if they worked in the same plant? Lignac's study should not be included in evaluation of benzene since the routes of administration were inappropriate for the comparison to incidence of leukemia in humans. The fact that no controls were run severely limits the validity of data. The data on leukemia induced by injection of benzene into the animals is not appropriate for comparison to human incidence of leukemia induced by inhalation of benzene. The animal studies on leukemia, induced by inhalation of benzene, have proven negative. On the other hand, the human epidemiological evidence is strongly suggestive of a causal relationship. This may be due to the fact that most human cases involved occupational exposure to benzene mixed with a variety of other agents. As reported earlier, the effects of benzene are additive with toluene in toxicity studies. It may be a combination of benzene with some as yet unknown agent that is responsible for the induction of leukemia. Although the authors suggest a convincing association between benzene exposure and leukemia, they indicate there arc problems with the data. A significant problem is the role of related compounds and/or metabolites in the exposed patients. The case histories are only suggestive since the subjects were exposed to benzene and other compounds. That benzene can produce Hodgkins disease at lower concentrations (150-210 ppm) suggests a potential synergism with other agents, CMA 049546 C - 26 C-2Q C - 31 The retrospective studies of Girard and Revol, looking at history of benzene exposure and leukemia, may be limited. The report by Thorpes is questionable since "potential exposure to benzene does not necessarily effect true exposures. These patients might also have been exposed to other compounds that were * synergistic with benzene. CMA. 049547 AN EVALUATION OF Tllb PROPOSED AQUATIC LIFb CRITKKION ON BERYLLIUM The proposed criterion using logarithmic extrapolations as a measure of relative species sensitivity with water hardness is un acceptable for beryllium. This recommendation was based upon acute toxicity -using the least squares regression of the natural logarithm for three fish species combined with the chronic value of one inverte brate (Daphnia magnu). Data used from the two Daphnia studies could not be evaluated since both studies were not published m the open literature or available for review. Those assumptions or .judgments using slopes obtained from the relationship of hardness and toxicity require more testing before use in the establishment of water quality criteria. Prepared by Donald a. Cherry 20-3 Maywood Street Blacksburg, Virginia 2-1000 and John Cairns, Jr. 501 Bishop Road Blacksburg, Virginia 240G0 May 1, 1979 CMA 049548 GLNLKAL COMMENTS The criteria to protect freshwater aquatic life, which include the 24-hour average concentration, (1.24 e In [hardness] - (j.bbl , and the maximum concentration at any time, e'(1.21 In [hardness] - 1.40; are unacceptable for beryllium. These data wore based upon acute studies of three fish combined with chronic: studies of one inverte brate (Daphnia magnu), The experimental design and methodology lor the Daphnia studies are not available for review and have not been published in the open literature. Until such data are reviewed by the scientific community and more organisms including fisu are tested, the criterion document as proposed lacks credibility. The criterion to protect freshwater aquatic life is based on only one invertebrate or cladoceran species and should be so stated. It is agreed that no criterion for beryllium can be derived (min the insufficient data available for saltwater aquatic life. Specific Comments p A-2. 1) The reference Sutt et al . i'JeO, which js missing in all reference sections, lias not been located in the literature. Is the spelling and/or date of the reference correct? p D-2. 1) The Slonini and Slonim 11)73 study, as cited in the second paragraph, is not listed in any reference section ol this document. CMA. 049549 .p B-3. 1) In paragraph 2, line:; 7-9, the calculation loading to the assumption that there is "a range of species sensitivity to beryllium of only two-fold when corrected for hardness effects", is misleading. These corrected values were based on only three fish species. The slopes for the derived equations exhibit some degree of variability and are presented in the preceding paragraph. Therefore, when a geometric mean of these variable results is cal culated and then used to calculate the intercepts for each species, the variability in intercept values is reduced. 2) In lines 6-8 of paragraph 3, the statement that "a similar relationship exists between acute beryllium toxicity and hardness for invertebrate species as was demonstrated for fish" may be true. However, the application of the lisli acute mean slope (1.24) to invertebrate acute and chronic data on p. B--1 is questionable. This distinction is important because the final freshwater criterion is based on invertebrate chronic values that are modified by fish acute toxicity data. 3) The adjusted 48-hour LC50 value of 1,170 is 1,770 ug/i in Table 2. 4) Acute toxicity data and chronic toxicity inlonnatiou (on p B-4) were taken from one species (Jiaphn i a mngna) in two studies (U.S. EPA 1978, and Kimball-Manuscript). Neither paper was a primary reference, being either an internal EPA document cl one of unknown origin which was unavailable tor review. Establish es 049550 iny a standard from the response of only one species is considered inadequate, even if the studies in question were available for scrutiny. p. B-4. 1) The word "realtionship" in the second and third lines of the last paragraph should be "relationship". There are many errors in the citing of references in the text relative .^to their omission in the reference sections. It is assumed that for the sake of repetition, a reference was listed once m either reference section A, U, or C, instead of listing all references each time they appeared in their respective sec tion of the text. If so, references cited in the text would supposedly be found just once in any of the three sections of reference. However, citations were referenced in more than one section (e.g., Wagner ct al. I960, Wagoner et al. 1978, Van Ordstand et al. 1945, U.S. EPA 1978, Van Cleave and Kay lor 1955, Gardner and Heslington 1946, Grier 1949, Berg and Burbank 1972, Barnes 1948, Branion et al. 1931, Loeb and Sirover 1977, Luke et al. 1975, Topper 1972b, Toda 1968, Thornton 1950). These illustrations were taken by randomly selecting the first letter, "W", "V", "G", "B", "L", and "T", of authors' last names listed in both reference sections A and C. Since references were listed more than once, it was assumed that each section had all references listed that were cit ed in that section. Unfortunately, reference section A failed- to have CMA 049551 .citations for: National Academy Science 1958, Teppor 1972a, Haven and Spronk 1953, Slonim 1973, Reeves 1965, Warner et al. 1965, Sutt et al. 1950, and Mullen et al. 1972. In reference section U, the following were not listed: Tarzwell and Henderson I960, Slonim and Slonim 1973, U.S. EPA 1978, Slonim and Hay 1975, Slonim 1973, and Jackim et al, 1970 (cited as Jacki_n). There were also several important studies from which data were used in the establishment of guidelines which were not referenced in any section of the document. These included: Slonim 1973 (nine bioassays in Table 1), Slonim and Slonim 1973 (four bio assays in Table 1), Slonim and Ray 1975 (eight bioassays m Table 6), all of which were discussed and cited in the text. The basic premise for protection of freshwater aquatic life was based upon the mean slope (1.24) through the geometric mean toxicity value and hardness for each fish species. Much of this information was based upon Slonim's studies. Other missing references included Jackim et al. 1970 (reduced alkaline phosphatase activity in the mummichog) which is presumably Jacking et al. 1970. The Kimball (Manuscript) reference in Table 2, concerning Daphnia magna mortality included no address of the authority in the lit erature reference. However, chronic effects of beryllium at 30 ug/i. and reproductive data for D. magna were utilized, adjusted, and manipulated to obtain slopes for chronic toxicity. This reference and the EPA 1978 study are unacceptable until made avail- lable for assessment of the experimental design. The reference CMA 049552 .to Suit et al. 1950 concerning human health effects was missing in all reference auctions. It is assumed that Knejci and Schell I960 (ionic behavior of beryllium to pH) is Krejci and Schell 1966, and that N'as 1958 (concerning human health) is K'asji 1958. The document is difficult to evaluate and interpret from these discrepancies listed above. A general review paper by Reeves 1977 ("Beryllium in the environ ment") slrould be included into the document. Another uncited paper by Lislor 1974 ("Radiocadmium exchange with seawater by Fundulnx he loroc li tus [L.] [Pisces ; Cypnnodontidae ] ") related the exchange of beryllium from water into whole body residues relative to the concentration factor (0,66-0.85 from 5 to 50 mg/ in water) >f fish. Hildebrand and Cushman 1978 ("Toxicity of gallium and eryllium to developing carp eggs [Cvpnnus enrpio] utilizing pper as a reference") reported that hatching success of carp :s was not inhibited by beryllium concentrations up to O.OS- 9 ppm while no hatch was reported at 0.20 ppm. This studyid also be incorporated into the document. CMA 049553 CRI TER I ON FuKMULAT I ON The guidelines used for establishing a final acute and chronic toxicity criterion from the baphnia magna data represent only two values which are the minimal number which contribute to a mean (Tables 2-u). The adjusted LC50 value of G,691 ugjl (U.S. LPA 1978) from the LC50 of 7,900 ug/Jt may be erroneous depending upon how it was calculated. The adjusted LC50 of 1,775 ug/t from Kimball (Manuscript) differs from the value quoted in the text (1,175 ug/i). It docs not seem logical to assume that the Final Invertebrate Chronic Value derived from Duplin i a data should become the Final Chronic Value since no chronic test data were available for fish. It is possible that Paphnia represent the most sensitive species. Evaluation of the guidelines for establishing a final toxicity criterion cannot be completed at this time since important key data were taken from, two improperly documented (manuscript form) or unreferenced studies which represented approximately 50% of the bioassays listed in Tables 1-6. Until these studies arc properly referenced and/or published in the open literature with peer review, the criterion proposed lacks credibility. CMA. 049554 -7- MISSING AND ADDITIONAL REFERENCES 1. Eisler, It. 1974. Radiocadmium exchange with seawater by Kundu1 us hcterocl i tus (L. ) (Pisces : Cypnnidont idae). J. Kish Biol. 6 ; 601-G12. 2. Hildebrand, S. G. and R. M. Cushman. 1978. Toxicity of gallium and beryllium to developing carp eggs (Cyprinus car pro) utilizing copper as a reference. Toxicology Letters 2:91-95. 3. Reeves, A. L. 1977. Beryllium in the environment. Clinical Toxicology 10:37-48. 4. Slonim, A. R. 1973. Acute toxicity of beryllium sulfate to the common guppy. J. Water Poll. Conti. Fed. 45:2110-2122. 5. Slonim, C. B. and A. K. Slonim. 1973. Effect of water hard ness on the tolerance of the guppy to beryllium sulfate. Bull. Environ. Contamin. Toxicol. 10:295-301. 6. Slonim, A. R. and L. E. Ray. 1975. Acute toxicity of beryllium sulfate to salamander larvae (Ambystoma spp.). Bull. Environ. Contamin. Toxicol. 13:307-312. CMA 049555 a;; evaluation of the EPA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON BERYLLIUM CONCLUSION: The proposed Water Quality Cruena based on beryllium as a carcinogen is inappropriate. Expcrinum'ta I evidence does not show beryllium to be carciiiGdcn when ingested. *. The document should be thoroughly revised. Prepared by. Or. Harry L. Skalsky and Dr. Joseph F. BorzeUeca Division of Toxicology Department of Pharmacology Medical College of Vi mini a Richmond, Virginia P320R 23 April 1070 CMA 049556 flF. NEPAL COMMITS A clearly stated purpose of these criteria documents is to establish levels of "exposure from all sources, not limited to drinking water and ingesting fish" (Fed. Re(J., Vo 1. 44, p. 18975. 1979). This has not been accomplished for beryllium. Estimates for the abundance of beryllium in the earth's crust range from 2 to 10 ppm (Merck Index, 1978). The proposed standard is 0.037 ug/1 (risk level of 10'^) which is'vO.QS? ppb or10,000 times lower than the element exists in nature. No estimate of tota1 dietary intake of beryllium was made. Estimates were made for shell fish consumption, hut no allowances were made for other dietary sources. For example, potatoes and tomatoes contain 0.17 mg/kg and 0.24 mg/kg, respectively, of beryllium (page C-l) and contribute to overall dietary intake. To establish a water criteria standard for beryllium, a "natural hu.kground" level for exposure must be established. Thu human population lias been exposed to low levels of beryllium since the beginning of time, wnen life originated on earth. Tims, a question is raised as to tne basic validity of this document. EPA clearly states (red. Reg.. Vni. 44, p. 18977, 1979) that risk levels are based on the extrapolation from a known effect level tn zero exposure level. The major point is that a zero level of beryllium has never existed. In order to set any valid criteria, a "natural background" level nf beryllium must bo used and not zero. The formulation of any legal criteria should be based on the best data available. This has not been done for the proposed standard on beryllium. The calculation of a safe exposure level for man is based on the consumption of two liters of water per day; therefore, the most appropriate route of compound udn.inistration in experimental animals must be by ingestion, ihr -1- CMfc 049557 proposed water standard has been calculated as if beryllium is a carcinogen. However, there is no recorded instance of cancer being produced by the ingestion of beryllium. Definitive feeding and drinking water studies (Schroeder and Mitchener, 1975, and others C-19) have demonstrated that beryllium at 5 ppm. caused no change in the growth rate of mice or rats, longevity, or incidence of tumors. This is a no effect level at ''10.00') times higher concentration than the proposed standard. These negative data are extremely important hecause the oral route of exposure was used and this is the route of human exposure. - I- CMA. 049558 SPECIFIC CO.`*MN'TS Reeves (1965) not included in reference list on page A-5. Orel EO^q of 9.7 mg/kg mentioned, but no identification of test species'. In the third and fourth paragraphs, a variety of effects of beryllium are cited, but routes of exposure are not given. The point that ingestion of beryllium has not been shown to cause cancer is important and is not accorded enough attention. In establishing a <,t mul.ir<j for water, any information dealing with ingestion should be examined <. I 'sely. This is the route by which the human population would be exposed. It is also important because tne results of oral versus inhalation tend to be very different. Methodology for detection of beryllium bus improve'; since l'n/, and the 5.A percent cited by Kopp and Kroner (page A-1) is probably on the In./ side.However, for those waters containing beryllium, tne mp.m tunccntrntiun is 0.19 ug/1. To meet the highest proposed criteria level (0.067 uq/1), these average concentrations would have to be lower by one half. The final chronic value for beryllium has been calculated using some unusual assumptions, due to the lack of adequate data. Hardness data from acute fish studies (no chronic fish data) was combined with chronic data from Daphnia mogna (no hardness data available) to establish the relationship of chronic toxicity values and hardness. No extrapolation factor or error calculation was made. Data derived from an invertebrate was applied directly to a vertebrate. No supporting data for this type of cxtrapolalion was provided. Extrapolation from one fish to another is acceptable, but to consider Daphnia and fish as equal requires substantiating evidence. 3- CMA. 049559 [j-5 The sLiiitrury is an overt misrepresentation of the facts. The toxicity values and their relationship to hardness are presented as established fact when they are only rouejh estimates. C-5 Under Residues - Table 4 should hr chunked to Table b. C-] Some foods contain rather s i rjni f i tun t amounts of hr'", I i i uni i.pni.u i , tomatoes, etc.). Thmc values are mentioned but are never utilized in do: -runn- inn a told! ex;.ns urn level. This is n_<U conn is mnt with the clearly laifl purpose*#* these criteria docui.cnts (fed. . , boi. 44, p, 199/5, 1 5/m C-2 A major error bus been made in calculation ui the bio-concentration factor. liCU has been measured only in blucjills. The U'A crronerously assumes tnal "based on data for lead and cadmium, beryllium would .probably have a lower 11CF for fish and decapod muscle man for fish whole body, but probably would huvn a higher CCF for mo Husks." It is rediculous to assume mat beryllium distri bution can be dL-nved from distribution data ot lead or cadmium. I am not aware of any evidence mat supports this assumption. In fact, data indicate that beryllium, will behave more like maqncsium or calcium. This is a critical error, since the bCF_ is used to calculate exposure. C-19 IMPORTANT POINT; No record of any evidence of cancer due to bciyllium ingestion. '.C-2b Current levels of beryl liuai exposure am nut mi rr( 1 . dim to laid nt i ncorpora tion or data concerning dietary sources other LI.an seaiooJ. C-?`) The use of a "one-hit" extrapolation model is 1 nappropriam on two lounis; 1. A aero level of beryllium has never existed. 2. beryllium is not an established carcinmjen orally. -4- CMA. 049560 7he latter point n.ay be related to a !<jck of jbburption o! beryllium in the put. j - CMA 049561 Comments on the Aquatic Toxicity Section of the EPA Water Quality Criteria Document (March 15, 1979) on Cadmium Summary It is important that the relationship of cadmium toxicity to water hardness has been used in establishing the freshwater criterion. The fresh water criterion values are properly derived under the proposed guidelines and appear to be approxi mately protective of aquatic life under laboratory conditions. It is possible that a relationship between cadmium toxicity and salinity should have been derived. Otherwise, the saltwater criterion values are prop erly derived under the proposed guidelines; both the maximum concentration and the 24-hour average concentration were established on the basis of studies of Mysidopsis oahia, a mysid shrimp, and should be approximately protective of all tested species under laboratory conditions. Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and James L. Peterson, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19103 May 1979 CMA 049562 Proposed Criteria - Aquatic Life For cadmium the criterion to protect freshwater aquatic life as derived using the Guidelines is e(Q-87 In (hardness)-4.38) as a 24-hour average (derived from the fish chronic slope and the invertebrate chronic datrum) and should not exceed e(1.30 In (hardness)-3.92) (the final fish acute value) at any time, The following table illustrates the criterion for repre sentative hardness values. Hardness mg/1 20 40 100 300 24-hour average yg/l 0.17 0.31 0.63 1.76 Maximum yg/1 0.97 2.4 7.9 33. For cadmium the criterion to protect saltwater aquatic life as derived using the Guidelines is 1.0 yg/1 (the final invertebrate chronic value) as a 24-hour average and the con centration should not exceed 16 ug/1 (the final invertebrate acute value) at any time. Our Conclusions It is important that the freshwater criterion values were adjusted for the apparent relationship between cadmium toxicity and water hardness. In the freshwater criterion, the maximum concentration is derived properly according to the proposed guidelines, but this involves a considerable downward extrapolation of the slope from the range of cadmium concentrations for which it was der ived. The maximum concentration appears to be at realistic excursions above the 24-hour average concentration in most of the range of water hardness. The 24-hour average concentration is set as high as the available data will permit, and appears to be approximately protective under laboratory conditions for all the species tested. Although no scientific support is pro vided for applying the fish chronic slope to the invertebrate chronic datum, failure to extrapolate some slope would result in a much more stringent 24-hour average value at high levels of water hardness. CMA 049563 In the saltwater criterion formulation, it is likely that a relationship b'etween cadmium toxicity and salinity could have been derived using published data, not all of which was tabled in the document, at least for fish acute and invertebrate acute values. The criterion formulation process was otherwise followed properly for both the maximum and the 2-1 -hour average values. Both components of the saltwater criterion are based on studies of .*lysidopsis bchia , a mysid shrimp. The criteria appears to be approximately protective of tested species under laboratory conditions. CMA 049564 General Comments The literature on cadmium toxicity to aquatic life seems to be both extensive and difficult to summarize. The authors of this document seem to have accomplished a great deal of good work and deserve to be commended. There arc, however, a number of loose ends, some of which warrant comment here. The relationship of cadmium toxicity to other water quality parameters, such as hardness, alkalinity, pH and salinity seems to warrant complex as a description. Specific comments will be addressed to several of those points. The separate consideration of freshwater and saltwater studies tends to obscure an interesting relationship (backed up by very few data) between cadmium concentrations and the developmental stages of striped bass, which arc normally anadromous, although land-locked populations exist in the United States. The available data suggest that some other anadromous fish are very sensitive to cadmium, at least in freshwater. Thus, that topic may warrant discussion in terms of use clas sifications for specific water bodies before any application of the criteria is made in developing water quality standards. The document authors seemed to presume that inconsistent relationships of cadmium toxicity with other water quality parameters were "contradictory." At least in terms of logic, such relationships are merely subcontrarics; as such, it is conceivable that rather different physiological mechanisms are involved in fairly distantly related taxa. CMA. 049565 In comparing this document with the equivalent document for lead, it appears that there are inconsistencies in the flexible application of the guideline procedures for develop ing criteria. This perhaps warrants reconsideration of the procedural flexibility both within and between documents. It appears that rather rigid application of the guideline procedure for rejecting literature may have been made. For example, a study of the Australian subspecies of a mussel is retained (same "species" as in North America), but a study in which deionized distilled water was used in diluting seawater (rather than as the dilution medium for the toxicant) was not included. In summary, this is a respectable draft of the criteria document. Perhaps comments by us and others will aid the authors in developing a revised document. Specific Comments Introduction (ppA-1 to A-4) The seawater value from page C-7 should be incorporated in the first paragraph. The only values attributed to Eaton 1974 in the Aquatic Toxicology section are for chronic toxicity; Carroll ct al., 1979 does not appear in the aquatic toxicology references. Perhaps better references for acute toxicity could be used here if they are needed at all. CMA. 049566 Aquatic life Toxicology (B-l to B-60) Freshwater Organisms (B-l to B-27) Introduction (B-l to B-4). The discussion of the aqueous chemistry of cadmium warrants the citation of references, and perhaps some expansion to consider the work by McCarty, et al., 1978. Acute Toxicity (B-4 to B-7) The work by McCarty et al., 1978 and other papers warrant comment* as to the relationship between toxicity and hardwatcr. There is some basis for suspecting that alkalinity rather than hardness is related to the mechanism of reduced hardwatcr toxicity. Moreover, the data suggest that, at least in static tests, the average concentration to which organisms arc exposed* over 96-hours is much less than the initial measured value in hardw3ter (perhaps less than 101 of the initial measured value). This is at least enough to question the correction factor for static tests, and perhaps also to require time averaging of the exposure concentrations for studies done in hardwatcr. The studies by Hughes (1973) on early stages of striped bass do not appear to be readily accessible. It could be use ful to know whether these are from an anadroinous or a land locked population. Studies on fish which can be either landlocked or anadromous could usefully be categorized by the source population, since this also applies to steelhead. Anadromous fish may lie under greater physiological stress in freshwaters, and hence CMA 049567 more sensitive t-o toxic stresses. All the very low l.CSo values are from species which are (at least partly) anadromous, and these determine the final fish acute, which is also the Final Acute Value. The fresh acute values for cadmium are much more valuable than the guideline species sensitivity factor provide f or. In the tabulated invertebrate studies, there is no posi tive support relating hardness to toxicity, and the rotifer data slfow no evidence that this relationship would hold. Marshall (1979) reports as 4S hour ECSO for Dcphnia galcatanendotae which could be considered for inclusion here. The invertebrate data may be only about half as variable as the guidelines sensitivity factor suggests. The application of slopes derived for fish to invertebrates is unsupported by this document, although required by the guidelines. Chron ic Toxicity (B-7 to B-9) The chronic fish values arc very low relative to the same species in acute tests, and the tabulated range of chronic values is only a 56-fold difference, compared to several or ders of magnitude in the acutes. Hence, these data reflect a very different portion of the cquiblibrium chemistry of cadmium than the data for which multiple hardness values were used to adjust acute tests results. Thus, it does not seem unreason able (nor is it contradictory) to have only a possible (but non-significant) hardness to toxicity relationship for brook trout and no apparent relationship for channel catfish. Indeed CMA 04956S the adjustment may not be warranted at such low cadmium con centrations, in which case there is a question as to whether an "overall relationship" can be established under the guide lines, when this is coupled with our previous concerns about the actual exposure concentrations in hard water in the fish acute studies. Given the data and procedures used in Table 3, the spec ies sensitivity factor of 6.7 does not appear unreasonable for the observed range of variation. It is not clear, however, how the application factor comparison is made, since section X of the guidelines is bypassed if the Final values arc related to water quality. The final invertebrate chronic value depends solely on a very low value obtained with Dcphnia nagna which may well be even further out of the range of a cadmium to water hardness relationship than the chronic fish values. We consider it to be unscientific to apply the fish chronic slope to the inverte brate chronic data, both because the change in concentrations involved modifies the thermodynamic equilibrium and because no relationship of water hardness to invertebrate toxicity has been established in this document. (Separate comments ad dress this aspect of the guidelines.) We note that because this was one of the sensitive species in acute studies, the species sensitivity factor was not applied by the agency in determining the final invertebrate chronic value in freshwater. Marshall (1379) notes that studies on Dsphnij. julcatu no:far.re give results comparable to those for Dapr.-.tic. ,Tep>tc. CMA 049569 (1 Table 7. Some of the longcr-than-acute values, such as for rainbow trout, warrant comment since they arc as large as the LC50 value reported in acute studies, sometimes by the same authors. This is just another aspect of the overall con fusing picture of cadmium toxicity. The three spine stickleback poses an opposite problem, perhaps suggesting that the final chronic may not be fully protective. Saltwater Organisms (B-28 to B-47) % Acute Toxicitv (B-2S to B-50) The salinity to toxicity relationship is more complicated than the document indicates, in that at salinities below 20% (Voycr, 1975) or 15%(Eisler, 1971), mummichogs arc also less resistant to cadmium than at 15% or 20%;moreover, Middaugh and Dean (1977) did not regard the differences in cadmium toxicity from 20% to 50% salinity as meaningful for mummichogs or Atlantic silvcrsidcs. Hence, it may be that high salini ties are protective in more nearly an asymptotic than an exponential fashion, as would be consistent with data on Palaemcnetes pu$io (Sunda, et al., 197S). The development of additional data across the range of estuarine salinities may help to clarify whether this relationship, which could be ex pected on the basis of the chemistry of cadmium, can be established. The range of 7.S-fold for only four species is not in consistent with the species-sensitivity factor, but the document does point out that these species represent only two CMA- 049&70 families of fish. Hence the evaluation of the species sensi tivity factor for this application should, as stated, await more data. In the invertebrate data, one could question the inclu sion of Ahsanullah's (1976) d<ata on the Australian subspecies of the edible mussel, even though it is the same nominal species, but this certainly has no meaningful affect on the final invertebrate value. The omission or exclusion of data on rclzc-.onencs pupia from Sunda, et al., 197S is perhaps more significant in that differences in cadmium toxicity were found at different salinities; hopefully this paper was overlooked rather than excluded by a very rigid interpretation of the last sentence in Section II.E of the Guidelines. Their data do suggest that cadmium is relatively more toxic to ?. pupi? at low salinities (ca. 50) than at salinities near that of seawater. The Final Invertebrate Acute Value is based on the single flow-through study by Nimmo, et al., 1977a. Chronic Toxicity (B-50 to B-51) The document comments on the relatively small difrcrcncc found between the acute and chronic values for X'jsidopsic bekia. However, similar small acute to chronic shifts are indicated for the very sensitive freshwater fish and inverte brates, which can perhaps be coupled with knowledge of the equilibirium chemistry of low versus high concentrations of total cadmium. 049571 In developing the Final Invertebrate Chronic Value in freshwater, the species sensitivity factor was not used be cause one of the most acutely sensitive species was the basis of the chronic value. Here, in the same setting, the species _ sensitivity factor has been applied. It seems reasonable to evaluate the invertebrate chronic without the sensitivity factor prior to setting criterion value potentially lower than necessary for the protection of saltwater life. Biconcontration: It appears from Guideline Section XIV.B that the oyster bioconcentration limit, relative to human emetic response, of 5 ug/1 would be a candidate for setting the Final Chronic Value if the Final Invertebrate Chronic Value were set at 5.5 ug/1 instead of at 1 yg/1. Miscellaneous: Dawson et al., 1977 did not regard the decrease in enzy matic activity as a toxic response, but rather as an adaptive response within the range of cadmium concentrations which would be non-toxic to juvenile striped bass in estuaries. Hence, the only potential question from Table 15 data raised below 5 ug/1 is for larval fiddler cra"bs (Vcrnberg, ct al., 1974); that paper suggests that most of the potential for re duced survival from exposure to 1 ug/1 of cadmium would be under suboptimal temperature and salinity conditions. Perhaps that study is sufficient justification for lowering the Final CMA. 049572 Chronic Value in saltwater from S ug/1 to 1 pg/1, as was done by invoking the species sensitivity factor in computing the Final Invertebrate Chronic Value. Table 8. Eislcr (1971) reports additional 96-hour LCS0 values for mummichog not given in this table. Additional References Sunda, W.G., et al. 1978. Effect of Chemical Spcciation on Toxicity of cadmium to grass shrimp, Palacnonetcz pugio: Importance of free cadmium ion. Environmental Science and Technology, 1 2 (4) : 409 - 415. Marshall, J.S. 1979. Cadmium toxicity to laboratory and field populations of Daphnia ga leaza-nendotee. Bulletin of Environmental Contamination and Toxicology, 21:4S3-4S7. CMA 049573 AN EVALUATION OF THE EPA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON CADMIUM CONCLUSION: This criteria document appears to be scientifically sound and requires only minor revision. Prepared by: Dr. Harry L. Skalsky and Dr. Joseph F. Borzelleca Division of Toxicology Department of Pharmacology Medical College of Virginia Richmond, Virginia 23298 23 April 1979 049514 General Comments This document contains a number or errors. These will be addressed under the specific consents section. The subjects of mutagenesis and carcinogenesis (C-39 to C-57) were well addressed in this criteria document. The conclusion that "the case for cadmium as a carcinogen is not persuasive when the existing data are critically reviewed" (C-62) and the recommendation that cadmium not be considered a suspect human carcinogen fC-62) appears to be scientifically sound on the basis of current data. It is also appreciated that this document recognizes that "since cadmium is an element, it will not be destroyed and may be expected to persist indefi nitely in the environment in some form" (A-2). This point is often lost in regulatory activities. Although thorough in some areas, this document has shortcomings in others. No mention has been made as to the effect of water treatment on levels of cadmium in finished water. Linstedt et.ll (J- Water Pollut. Control Fed. 43:1507-1513, 1971), Nilsson (Water Res. 5:51-60, 1971), Furukawa (EPA 902/9-34-001, 1973) and others have indicated that cadmium is efficiently removed (greater than 90"') under a variety of treatment conditions. Since most surface water supplies for large populations are subjected to treatment that includes coagulation, sedimen tation, filtration and disinfection, it is necessary to know what effect these processes would have on the concentration of cadmium in finished water. It is surprising that there is no mention of the recent (1977) NAS Safe Drinking Water Committee report entitled "Drinking Water and Health". Cadmium is discussed on pages 207-208, 211-213, 220-222, 236-241, 302-303, and 305-306 of this document. -1- This NAS report has been used and accepted as a guideline for drinking water standards. It should at least be accorded a reference in this criterion document. Other references have also been neglected and some have not been appropriately quoted. Examples of this are given in the special comments section of this report. It appears that this criterion document should be revised to insure that the data concerning cadmium is thoroughly covered. Specific Comments A-2 Good point. Cadmium can not be destroyed and may be expected to persist indefinitely. B-l Tirst sentence reads as if no natural fresh waters have concentrations higher than 0.01 ug/1 - If this is true, it should be clearly documented. In an EPA survey of 14 ground water and 60 surface w.:;er supplies, the levels of cadmium ranged from 0.2 to 12 i.g/1. C-7 Good point. Orinkinq water contributes relatively little to man's total daily cadmium intake (3-4 ;ig/d). C-8 Only 5^ of ingested cadmium is absorbed. C-7 This page contains the only mention of removal of cadmium from drinking water in the entire document. This section clearly should be expanded. NAS' report on Drinking Water and Health and general conments section of this report should be consulted. C-20 Most of this data concerns health effects of industrial exposure to cadmium. -24 While interesting, it should have little impact on standard formulation because it is an inappropriate route of exposure (inhalation vs oral). - 2- CMA 049576 C-33 Reproductive effects. I am somewhat familiar with this area of literature and, without effort, there were at least four pertinent articles concerning' cadmium's effect on reproduction that were not cited or discussed. Lee and Dixon, 187:641-652, 0. Pharm. fc Exp. Therap. 1973 Chiquoine, 149:23-36, Anat. Rec. 1964 Chiquoine, 10:263-265, <1. Reprod. (r 't. 196!. Allanson, 24:463-462, J. Endocrinol. 1962 C-40 A paragraph is accorded the Friedman and Staub (1975) in vivo mutagenic assay. The results are explained away "due to cadmium's ability to impair testicular blood supply". However, if one reads this paper, one finds in addition to the lOmg/kg acute exposures which are discussed in this document, Friedman also treated animals at 1 mg/kg for four weeks which is not mentioned in this document. Although I do not believe this discrepancy should have any significant inpact on the proposed standard, it does indicate that the document should be thoroughly revised to insure that all information from a cited reference is used and evaluated properly. C-63 From the data presented, I feel the concluding statement of the document is appropriate. "From this analysis it appears that a water criterion needs to be no more stringent than the existing Primary Drinking Water Standard (in ug/l) to provide ample protection for human health. - 3- CMA. 049577 I A review of proposed ambient water quality criteria for carbon tetrachloride Recommendation: A weak, supporting data base leaves the exact freshwater criterion open to debate. Data for marine life are significantly less complete and the methods utilized in developing the criterion are neither scientifically nor statistically defensible. Alan U. Maki, Ph.D. Environmental Safety Department The Procter & Gamble Co. Cincinnati, OH 45217 CMA. 049578 A Review of Proposed Ambient Water Quality Criteria for Carbon Tetrachloride Freshwater and Marine Aquatic Life The criterion to protect freshwater aquatic life for carbon tetrachloride was derived using procedures other than the recommended guidelines. The criterion is established as 620 ug/1 for a 24-hour average and the concentration should never exceed 1400 ug/1 at any time. The supporting data base leaves the exact criterion open to debate. Data for marine life are significantly less complete and the methods utilized in developing the criterion are neither scientifically nor statistically defensible. Specific Comments Freshwater species data consist of one Daphnla and two bluegill static acute tests, and a fathead minnow embryo-larval assay. LC50 data for the two bluegill tests vary from 27,300 ug/1 to 125,000 ug/1 for 96 hr. exposures indicating that volatilization of the test material significantly affects results. Although the unmeasured test concentration correction factor was uso,<J for all test results, the extreme volatile nature of this compound significantly detracts from the interpretation of these data. Additionally, employing geometric means based on one or two LC50 values in the formulation of the criterion number can result in an error term approaching 200-3002. A minimum of toxicity data indicate that carbon tetrachloride is apparently not as toxic to marine life as freshwater life, however, the procedures outlined in the criterion document for development of the Marine Final Acute Value are not defensible. Averaging toxicity data for bromoform, methylene chloride and chloroform to develop a correction factor to be applied to carbon tetrachloride is a procedure unspecified in any of the guidelines and it can be argued that the resultant criterion is in no way indicative of effects of carbon tetrachloride. We caution that even structure: activity correlations have their predictive limits. CMA 049579 EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR Carbontetrachloride Conclusions: The methodology used for the derival ion of the health based criterion for carbontetrachloride .ay not be appropriate. The significant toxicological literature on carbontetrachloride is veil reviewed. Prepared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 10, 1979 CMA. 049580 General Comments: The criterion document is very well written and represents an excellent review of carbontetrachloride (CCl^) toxicity. However, the criterion developed on the basis of animal models and an extrapolation model for carcinogenic risk may not be appropriate. Specific Comments: Page C-5, Second Paragraph. The units describing the application rate of CC14 (80 mg/1) do not make any sense. Pages C-64 to 0 67. If one evaluates the proposed criteria 3 in light of occupational exposures at the TLV level of 10 mg/m , the following apparent paradox occurs. Exposure at TLV: 33 5 10 mg/m (TLV) x 10m (inhaled air/8 hr) x y (work week/total week) x 0.75 (assumed efficiency of absorption) = 53.6 ma/day or 53,600 ug/day. Exposure at extrapolated risk of 10 2-6 ug/1 (21 + 69 (BCF) x 0.0187) = 8.55 ug/day Risk/dose = 10-5/8.55 = 1.77 x 10"6 Extrapolated risk at a rate of intake equivalent to the TLV for a life time: P = 1 - exp (-1.17 x 10'6 x 53,600) * 0.061 Thus, about 6% of all people exposed for long periods to CC14 at the TLV are extrapolated to develop cancer within their life time. Such an increase might or might not be readily detected by ordinary epidemiological methods. However, many people have CMA 049581 been exposed for long periods to CCl^ at concentrations even higher than the present TLV. Also, all experimental data indicate that CCl^ produces malignant tumors almost exclusively in the liver in all species and strains in which it has been found to be carcinogenic. However, the incidence of primary liver cancer in man is very low, even in most industriallyexposed populations. Consequently, one must conclude that an anomaly exists in either the animal model or the extrapolation model. The process selected to establish criteria for compounds found to be carcinogenic in animal models may not be directly applicable to CCl^. CMA 049582 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON CHLORDANE The document presents criteria that arc derived from a rela tively large data base for aquatic organisms. A major criticism concerns the applicability of using the FDA limit for chlordane in animal feed to calculate the RLTC for both fresh and saltwater life. The final values for the 24-hour average criteria are several orders of magnitude below the reported acute and chronic values for fish and invertebrates. This information, in conjunction with the cited evi dence that a rapid turnover of certain chlordane isomers occurs in aquatic organisms, suggests that the reported RLTC values ma> be overprotective. Until more work is reported concerning the biological half-life and toxicity of chlordane isomers, the iinal 24-hour average criteria should be listed as tentative or amended to include final chronic values. Prepared by Donald S. Cherry 204 Maywood Street Blacksburg, Virginia 24060 and John Cairns, Jr. 501 Bishop Road Blacksburg, Virginia 24060 May 1, 197`J CMA. 049583 GENERAL COMMENTS The key data used to establish the criterion to protect freshwater aquatic life (0.024 pg/ chlordane as a 24-hour average) is esti mated from the calculated Residue Limited Toxicant Concentration (RLTC). This RLTC value is based upon a calculation of the FDA limit for chlordane in animal feed and the bioconcentration factor of invertebrate data which was lower than the Final Fish and Inverte brate Chronic Values. The FDA limit for chlordane in animal feed applies to substances eaten by animals that may ultimately be con sumed by man. The maximum concentration (0.3G yg/) for the protection of 1reshwater aquatic life is based upon a calculated final acute toxicity value for invertebrates taken from several studies since values for fish were higher. To protect saltwater aquatic life, the 24-hour average concentra tion of 0.0091 yg/ was estimated from the calculated RLTC taken from only one fish species and no invertebrate data and should be so stated. Limitations in the data base available should also be mentioned. The maximum concentration (0.18 yg/) for protection of saltwater aquatic life is represented by the final acute toxicity value which is taken from six species of invertebrates reported from four ref erences. Although a final acute toxicity value (3.2 yg/) was available for fish, the invertebrate value was lower ana used instead. 049584 CMA 2- - Specific Comments p B-l, Paragraph 2 1) Some recent work has reported bioconcentration data of chlordane for freshwater fish (see Roberts et al. 1977, Veith et al. 1977, Moore et al. 1977). p B-l, Paragraph 3 1) The implication that a chemical formulated as a pesticide will not be harmful to aquatic plants is unjustified. Moore et al. 1977 demonstrated that some algae (Ankistrodesmus amalloides) have a maximum bioconcentration value which greatly exceeds that of freshwater fish (5,500 vs 162, respectively). Biggs et al. 1978 showed that 10 vg/l chlordane reduced algae growth rate and 14 C uptake per unit of chlorophyll a. There fore it cannot be stated that the effect of chlordane on aquatic plants is "not important". p B-3, Paragraph 3, Line 2 1) The sentence should read; ... "showing 2.5 times greater toxicity in 96 hour intermittent flow-through tests," A pulsed injection of chlordane should not be confused with a continuous application, for lethal responses would be different. p B-3, Paragraph 3, Lines 5-8 1) This sentence needs slight modification, such as: "The adjustments for static conditions, unmeasured concentrations, t* and a 24-hour instead of a 96-hour test were made, which resulted CMA 049585 in an LC50 equal to 3 yg/i for freshwater fish." The calculations from Naqvi and Ferguson 1970, which standardize the reported 24-hr LC50,varied from 7.8-436 ppb but are not presented. These values are quite important for an adequate formulation of the criteria. p B-2, Paragraph 2 1) A recent article by Rao et al. 1975 may contribute to the knowledge concerning the susceptibility of common carp to chlordane. The 96-hr TLm of 2.9 yg/t for carp is much lower than that of Ludemann and Neumann 1960. Unfortunately the jour nal (Anz. Schadlingskunde) was not readily available, so a comparison of methods was not possible. p B-4, Paragraph 2, Line 15 1) The Guidelines say that bluegill "are among the less sensitive of the species acutely tested" to chlordane. However, according to Katz 1961, they are more sensitive than Chinook sal mon, coho salmon, rainbow trout, fathead minnow, goldfish and guppy. For reasons unknown, LC50 of 22 ugJi for bluegill was omitted from Table 1 as were the last three species mentioned above. These data should be included. p B-19, Plant Effects 1) There is a reference for aquatic plants, Biggs et al. 1978. In this study, 10 yg/i adversely affected plant phy siology. The final plant value of 1,000 yg/t may therefore be too high, although the final criterion of 0.0091 yg/l should pro vide adequate protection for aquatic plants. CMA 049586 -4- P B-6, Residues 1) Roberts et al. 1977, in a study not referenced in the document, demonstrated that the tissue retention of chlordane in fish is directly proportional to the adiposity of the fish. However, it was also shown that transfer of pre-exposed fish to insecticide-free water resulted in a rapid elimination of chlor dane, perhaps through the gills, feces or kidneys (Moore et al. 1977). This rapid turnover raises the possibility that an RLTC of 0.024 ug/t may be overprotective. However, a radical reap praisal cannot be made until the toxicity of the different iso mers, which vary in biological half-life (Roberts et al. 1977), is documented. p B-14, Table 5 1) The 24-hour average concentration is the Final Chronic Value of 0.024 yg/i. This number was based on Residue Limited Toxicant Concentrations from Cardwell et al. 1977. If one inspects Table 24 on page 67 of that study, it is difficult to understand how all the concentration factors (CF) were derived. For example, the CF factors for heptachlor were derived from the heptachlor content per measured chlordane concentration in water. However, in using the same approach for the remaining columns (chlordenes, cis-chlordane, trans-clilordane, cis-nonachlor, and t rans-nonachlor) different CF values are obtained from those found in Table 24. p B-18. 1) The statement, "The LC50 value of 0.18 yg/t, which l* is about two times lower than the lowest value in Table 2", should be corrected in that Table 2 should be Table 8 (12-13th % CM* lines of paragraph one). 2) On lines 13-14 of the same paragraph, the statement, "Since there are data for only four species", should be corrected to read, "Since there are data from four studies" ... p 13-20. 1) A new paper recently published by yitko 1978 demon strates bioconcentration in saltwater organisms; measured levels for chlordane (ug/g in lipid tissue) were reported for several organisms: Lobster White Shark Liver 0,078-0.I 2.0 Herring 0.039-0.114 Rcdf i sh Atlantic Salmon 0.179 0.220 Several of these residues could exceed the accepted limit for chlordane in animal feed. However, the data are limited in that these numbers represent single determinations. It is important to note that lobster residues were similar to levels measured in commercial fish. p B-29. 1) The volume number of the Henderson et al. 1959 ref erence in the Trans. Amer. Fish. Soc. should be volume1 88. p B-30. 1) The Mt'hrlo et al. 1974 reference is incorrect. Volume 2 of the Bull. Env. Con Lam. Toxicol, does not include a page num bered 513. CMA 049588 r -6p B-31. 1) The Naqvi and Ferguson 1970 reference is incorrect. Volume 4 should be changed to volume 99. ^ CMA 049589 CHIItHION FORMULATION Guidelines for establishing a 24-hour average concentration (0.024 yg/t) for protecting freshwater aquatic life were bused upon the maximum permissible tissue concentrations established by the FDA in animal feed divided by the geometric mean bioconcentration factor for two invertebrate species (Daphnia magna and Hyal 1 el a ii/.tf'ca). These RLTC data, which represented only one study, were one to two orders of magnitude lower than fish and invertebrate final chronic values. Although no data in Table G contruriictod the conclusion that 0.024 ag/ will adequately protect invertebrates, the final chronic value calculated for fish (0,24 rg/Z) was also lower than any reported effect upon aquatic biota in the studies relerenceu and may also protect aquatic life. The maximum concentration (0.3b ug/t) not to he exceeded at any Lime for the protection of freshwater aquatic life was lower for inverte brates than for fish. Data for ten fish species and five invertebrate species were available which represented several studies. This con centration appears to be reasonable from the data base present. In developing the 24-hour average concentration of U.00&1 ug/ for the protection of saltwater aquatic life, no RLTC or final chronic value information was available for invertebrates. The bioconcentration factors comprising the RLTC were taken from juvenile and adult fish of one species reported from three studies. It is recommended that these limitations in the data base be CMJl 049590 9f 8- - stated. The recommendation that the maximum concentration (0.18 ug/t) to protect saltwater aquatic life, as estimated from the final acute toxicity value for invertebrates, is reasonable from the data base of fish and invertebrate studies available. VCMA 049591 -9- MI5SING OR ADDITIONAL REFERENCES 1. biggs, D. S., R. G. Rowland, 11. D. O'Conners, Jr., C. D. Powers and C. F. Wurster. 1978. A comparison of the effects of chlordane and PCS on the growth, photosynthesis, and cell size of estuarine phytoplankton. Fnviron. Poll. 15:253-263. 2. Moore, R., E. Toro, M. Stanton and M. A. Q. Khan. 1977. Absorption and elimination of 14 C-alpha and gamma-chlordane by a freshwater alga, daphnid and goldfish. Arch. Environ. Contam. Toxicol, 6:411-420. 3. Rao, T. S., M. S. Rao and S. 13. S. K. Prasad. 1975. Median tolerance limits of some chemicals to the fresh water fish "Cyprinus-carpio". Indian J. Environ. Hlth. 17:140-146. 4. Roberts, J. R., A. S. V. DeFrietas and M. A. J. Gidney. 1977. Influence of lipid pool size on bioaccumulation of the insec ticide chlordane by northern redhorse suckers (Moxostoma macro lepidotum). J. Fish. Res. Ed. Canada 34:69-97. 5. Veith, G. D., D. W. Kuehl,F. A. Puglisi, G. E. Glass and J, G. Eaton. 1977. Residues of PCB's and DDT in the western Lake Superior ecosystem. Arch. Environ. Contam. Toxicol. 5:487-499 1* 6. Zitko, V. 1978. Nonachlor and chlordane in aqUatic fauna. Chemosphere 1:3-7. CMA 049592 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR Chlordane and Heptachlor by John DouU , M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: The ambient water quality criterion based on the NOEL-Safety Factor approach (30 u/U appears to me to be more appropriate for these agents than the risk extrapolation values based on the questionable NCI mouse hepa toma study. CMA 049593 GENERAL COMMENTS: The key issue in establishing a human health criterion for both chlordane and heptachlor is the question of whether either or both of these pesticides are potential human carcinogens. The NCI data which is reviewed in this docu ment demonstrates that both agents are capable of producing hepatomas in mice, and the NAS expert pathology group convened to re-examine the pathology slides and conclusions of the NCI agreed with this effect in mice. The NAS committee concluded further, however, that neither of these agents produced liver tumors in rat studies and this, together with questions about the predictability of carcinogenesis assays in mice, raises doubts about the carcinogenicity of chlordane-heptachlor in man. If these agents are not human carcinogens, then the proposed standard is inappropriate and the human health criterion can be t established using the conventional safety factor approach (a criterion of 30 ug/1 is calculated in this report using this NOEL-safety factor approach). Even if chlordane and heptachlor are assumed to be potential human carcinogens, there are still questions as to how the mouse data should be extrapolated to estimate human risk. The approach used in this series of criterion documents has been discussed elsewhere and need not be repeated here. In view of the controversy concerning the potential carcinogenicity of chlordane and heptachlor for man, it would seem to be preferable at present to use the criteria based on the NOEL for setting the chlordane-heptachlor ambient water quality criterion. CMA 049594 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON CHLORINATED NAPHTHALENES Document is of greater value than the previous document because the criterion was based on 1-chloronaphthalone and not on mixtures as was the original criterion. Major criticisms include the use of data not available for peer review and high additive errors in estimating the final criterion values because of using the "adjustment and sensitivity factors" in deriving the final numbers.Because the criterion are based on single data points without confidence limits the criterion should be listed as tentative or not listed at all. prepared by Arthur L. Buikema, dr. 600 Alleghany Blacksburg, Va. 2400O and John Cairn;;. Jr. 001 Bishop Rd. Blacksburg, Va. 24()Oo May 10, 1979 CMA 049595 AQUATIC CRITERION For 1-chloronaphthane, the criterion to protect freshwater aquatic life, as derived using procedures other than the Guidelines, is 29 ug/1 as a 24-hour average and the concentration should never exceed 67 ug/1 at any time. For 1-chloronaphthalene, the criterion to protect saltwater aquatic life, as derived using the Guidelines, is 2.8 ug/1 as a 24-hour average and the concentration should never exceed 6.4 ug/1 at any time. GENERAL COMMENTS EPA is requested to read three documents previously submitted by the American Petroleum Institute (API) in response to the Water Quality Criteria "adjustment factors" published in the Federal Register (43(97) : 21506; 1978). These comments from API are or. the' public record and they address the problems in trying to correct an inadequate toxicity data base. It is unfortunate that EPA uses data collected for an EPA contract that was not available for peer review. Further, these data lack credibility because they were not collected following published EPA methodologies (i.e., static tests and concentration of toxicant was unmeasured). EPA should be commended for setting criteria for 1-chloronaphthane instead of for chlorinated napthalenes in general. They appear to have corrected their previous lack of adequate literature coverage. SPECIFIC COMMENTS Freshwater Organisms p B-l, Introduction The available LC50 and EC50 values for the bluegill, Daphniu, CMA 049596 and algae are similar only after the animal data were "adjusted." These animal data were bbtained from experiments not using recommended EPA procedures. Because of the gross errors with the "adjustment factor" concept and data base, EPA should refrain from making conclusions from "corrected" data. p B-l, Acute Toxicity Acute toxicity data which were not available for peer review and which had to be "adjusted" should not be used to set water quality criteria for 1-chloronaphthalene. The compounded error in predicting the bluegill value is up to 128% while for the Daphnia datum it is up to 295%. How can you have geometric means of single numbers? p B-2, Plant Effects Again, we are given data that were not available for peer review. This is scientifically unjustified. The use of a "lowest" plant value even though the two numbers presented in Table 3 aren't different is not justified. Are the differences significant? Given the problems in counting algae we would like to see confidence intervals for the algal data. What happened to the datum for yeast that was presented in the earlier version? Just because the LC100 was 48.8 milligrams per liter the datum shouldn't be ignored. p B-2, Miscellaneous Again, the data are from an EPA document not readily available for peer review. -2CM* 0495^ p B-3, Summary of Available Data We agree that there are no data available for establishing freshwatri criterion. To multiply the invertebrate acute datum (with up to 295% error) by 0.44 increases the error of prediction by up to 340%. p B-4 It is true no adverse effects of 1-chloronaphthalene have been reported - because they haven't been studied. p B-4, Criterion The 24 hr average value of 29 micrograms/1 has an error in prediction of up to 340%. If EPA persists in using "adjustment factors," EPA should raise the predicted 24 hr average value by the error terms. More importantly, EPA should use predicted concentrations with caution because empirically obtained data are lacking. The maximum concentration has an error of prediction of up to 295%. If EPA persists in using "adjusted" data, then they should raise the predicted value accordingly. Reliance on data not available for peer review is scientifically unjustified. Saltwater Organisms p B-9, Introduction EPA should be commended for admitting that data for commercial mixtures are not useful in setting criterion. Again, the toxicity data for fish, invertebrates, and algae are said to be comparable but only after "adjustment." Given the errors in prediction by using the "adjustment" value it is doubtful that the numbers are comparable. -3- 049598 p B-Dt Acute Toxicity Again, there are geometric means of single values. The error in predicting the fish value is as high as 125%, while for the invertebrate;;, it may be as high as 757%. p B-10, Chronic Toxicity Again this study is not available for peer review. To set a criterion based on these data is academically unsound. Because comparable MATC values were not available for saltwater fishes why did EPA persist in dividing the datum by 6.7 (See Table 5; FR 43(97) :21514 ) , which was for freshwater fish. The possible error in dividing by a 6.7 "adjustment factor" is up to 180%. p 3-10 Plant Effects Again the reference is not available for peer review. If there are no significant differences in the EC50 values, why was the lowest value chosen? p B-10, Miscellaneous The data in Table 10 indicates that EPA has reviewed the literature but if the criterion is for 1-chloronaphthalene then are these data necessary? p B-ll and 12, Criterion To set a criterion from one datum that is not available for peer review and that lias been conducted using methods other than those recommended by EPA is academically questionable. The error in predicting -4- CMA 049599 the 24 hr average value is as high as 804% and for the maximum con centration it is as high as 757%. If EPA persists in using "adjustment factors", then EPA should raise the final values accordingly. SUMMARY This document is of greater value than the previous document because the criterion was based on 1-chloronaphthalene and not on mixtures, etc. as was the original criterion. However, major criticisms of this document are: a) reliance on toxicity data which were not available for peer review and were not collected using published EPA methods. b) high compounded errors in estimating final criterion values because of the inadequate data base used to generate "adjustment factors." CONCLUSION EPA should not persist in using "adjustment factors" to correct an inadequate data base. The errors in their calculations and data base are scientifically incredulous and indefensible. If EPA persists in using "adjustment factors" to correct an inadequate data base, they should increase the final criterion values by the compounded error terms. Further, EPA should support research comparing predictions of final criterion values and empirically derived data to determine if their "adjustment factors" arc valid. -5CMA 049600 EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR Chlorinated Naphthalenes Conclusions: The basis for the criteria is very weak, The bioconcentration factor was not appropriately derived, There are insufficient data to establish numerical criteri a. ProDared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 40109 May 10, 1979 CMA 049601 General Comments: The section on human health and mammalian toxicology clearly summarizes most of the known toxicology of the chlorinated naphthalenes. The paper appropriately attempts to establish criteria for five different chlorinated naphthalenes rather than to establish a categorical criterion. However, the basis for the derivation of the proposed criteria is very weak. This is especially true for the bioconcentration factor. Another deficiency of the criterion document is the lack of an evaluation of the contribution of impurities, such as the dibenzofurans, to the observed toxicity of the chlorinated naphthalenes. Detailed Comments: Page C-7. A bioconcentration factor of 4,800 for Halowax 1014 is most certainly too low by more than one order of magnitude. The study in brown shrimp is inadeauate to derive a meaningful bioconcentration factor. Page C-20. There is no indication how the threshold limit values were derived and how they relate to levels of chronic exposure which have produced marginal effects and to levels which have produced no effects. Page C-22. As a result of the uncertainties associated with the TLV's the acceptable daily intakes, which were cal culated using the Stockinaer-Woodward model, must be considered very tentative. CMA 049602 Page C-23. Because of the known deficiencies in the data for bioconcentr'ation coefficients, it is not possible to derive believable criteria that include fish consumption. The criteria given are probably too high, but it not possible to derive more accurate criteria at the present time. CMA. 049603 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON CHLOROFORM Conclusion Due to the extreme lack of data for Chloroform in regards to aquatic toxicity testing and the fact that the meager amount of data did come from unpublished reports make it impossible to generate acceptable criteria for fresh and saltwater organisms. Prepared by Albert C. Hendricks Route 4, Box 1023 Christiansburg, Virginia 24073 and John Cairns, Jr. 501 Bishop Road Blacksburg, Virginia 24060 April 12, 1979 ^ 049604 Summary Sheet CHLOROFORM EPA Criteria: Freshwater - 500 ug/1 as a 24-hour average and 1,200 yg/1 as the maximum concentration. Saltwater - 620 yg/1 as a 24-hour average and 1,400 ug/1 as maximum concentration. Residues - Shown not be an environmental hazard to aquatic life. Conclusions: Due to the extreme lack of data for Chloroform in regards to aquatic toxicity testing and the fact that the meager amount of data did not from unpublished reports makes it impossible to generate acceptable criteria for fresh and saltwater organisms. CMA, 049605 General Comments The key data used to establish the aquatic life criteria are from unpublished EPA contracts. Until such data are published in the open literature, the criteria as proposed lack credibility. The "fudge factors" utilized and the concept behind these "fudge factors" are rather maddening. The numbers were derived from a wide variety of tests utilizing a variety of test organisms and compounds. It is simply unscientific to apply these factors to the data presently available for chloroform. The EPA in applying the "95% factor" often use a geometric mean derived from only one value. Not only is it silly to say a mean has been derived from one number, but statistically it is completed unsound to rely on one number. Fortunately, the geometric mean and the number are identifical. The fact that the EPA is relying on one data point to arrive at the Final Acute Value for both fresh and saltwater and the Final Chronic Value for freshwater renders the data quite suspicious. It is conceivable that a rather low number was generated at the invertebrate tests and if more tests had been performed values similar to the unadjusted LC50 fish values may have been generated. Also, it appears that the large numbers used to divide the geometric mean of the adjusted values will insure that the invertebrate data will always generate the Final Acute Value. Specific Comments p. B-l. (1) The fact that the chloroform levels in all the tests were not determined reduces the credibility of the tests a great deal. (2) The work by Bentley, et al. (1975) is not available and cannot be properly evaluated. This is true, also, for the EPA (1978) work. CM* 049606 p. B-3. Fish Invertebrate Acute Value should read Final Invertebrate Acute Value. p. B-9. I find it ludicrous that EPA is attempting to formulate a criterion for chloroform in saltwater from one unpublished data point. Criterion Formulation Due to the extreme lack of data for chloroform in regards to aquatic toxicity testing and the fact that the meager amount of data did come from unpublished reports makes it impossible to generate acceptable criteria for fresh and saltwater organisms. CMA 049607 EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR Chloroform Conclusions: The methodology used for the derivation of the health based criterion appears to be inappropriate for chloroform. The significant toxicological literature on chloroform in adequately reviewed. Prepared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Aroor, MI 48109 May 10, 1979 CMA 049608 General Comments: The criteria document on chloroform constitutes an appro priate summary of the known toxicological information. The methodology used for the derivation of the health based criterion may be inappropriate for the reasons noted under Specific Comments. Specific Comments: Page C-6. The calculation of the hypothetical levels of chloroform in beef cannot be supported because it involves too many assumptions. The calculation violates many principles of pharmacokinetics, especially in ignoring excretion and compost mentalization into lipids. Page C-23, First Complete Paragraph. This is a very signi- . ficant statement. But it is presented without data on the incidence of congenital abnormalities or the identity of the anaesthetic vapors involved. Pages C-42 to C-44. The total daily exposure for a risk of 10-5 is 2.1 ug/1 (21 + 14(BCF) x 0.0187) = 4.75 ug. There fore, the risk/dose is 2.11 x 10 At the current TLV of 9.8 mg/m3, the adjusted daily exposure is 9.8 mg/m3 x 10m3 x 0.75 x y = 52.5 mg or 52,500 ug/day. According to the risk estimate provided based on the criteria document, the single-hit model applied to life-time exposures equivalent to the TLV would result in a risk of cancer of P = 1 - exp (-52,500 x 2.11 x 10-6) = 0.10 or a 10% chance of cancer due to TLV level exposures. It should CMA 049609 3 be considered that the TLV has been dropped to 9.8 mg/m only recently, and that many people have had higher exposures than this. However, there is no evidence that chloroform has resulted in a 10% incidence of cancer in occupationally-exposed people. Consequently a discrepancy exists in the relationships between the animal model, the extrapolation model, and the human experience. The overall approach for setting criteria for substances found to be carcinogenic in animal tests may not be applicable to chloroform in the exact form that has been pro posed by the EPA. CMA 049610 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON' 2-CHLORO PHENOL Conclusion f protecti on of a cuatic life t he criter ion should be based cr. tonicity and not taste . From the data presented a value cf 1800 p?b is dete rmir.ed to be the numbe r that should not be exceeded in a 24 hour period. A final chronic value needs to be derived in order to finish the criteria. Prepared by W. B.' Neely Environmental Sciences Research The Dow Chemical Company Midland, MI 48 6-10 June 1, 1979 CMA. 049611 GENERAL COMMENTS The criteria state 60 ug/L as a 24 hour average for protection cf freshwater aquatic life. This is misleading as the number is based on fish taste, hence, has little or r.o bearing on tonicity. If taste is to be used as a limiting number, then it should be so stated in the criteria. Basing the excursion number that should not be exceeded at any time on invertebrate data is overprotective. Kith the rapid generation time of freshwater invertebrates it is inconceivable that a brief excursion from the ambient level will significantly alter the quality of water. A more defensible excursion n-umber would be either: 1) the final fish acute value of 1800 ug/L, or 2) an invertebrate number that is designed with a safety value of 10-15% of the population rather than 95%. CMA 049612 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR 2-Chlorophenol by John Doii 1 1 , M.D. , Ph.D. Un iversity of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: Organoleptic effects are not toxic effects, and thus are in appropriate as a basis for human health criterion setting. If the toxicity data for this agent is inadequate to establish an ambient water quality criterion for 2-chlorophenol, it should be so stated. Organoleptic effects are appropriate for establishing regulatory criteria but should not be used to establish human health-related criteria based on presumed toxic effects. I CMA. 0496X3 r. GENERAL COMMENTS: The human health criterion for 2-chloro phenol is stated to be based on the organoleptic properties of this agent because of the lack of toxicity data in animals or humans. Organoleptic effects do not constitute toxic effects and in my opinion do not provide a sound basis for establishing human health standards. There is fairly complete data on the acute toxicity of 2-chloro phenol human and there is, in addition, information on the effects of/exposure to this agent as the result of spills, etc. Thus, the statement on page C-20 that, "There are no reports of human or domestic animal toxicoses from accidental or intentional exposure to 2-ch1oropheno1." is incorrect. The paper by Baker et al. in the Archives of Environmental Health (March, 1978, pg. 84 to 94) refers to human health effects of a phenol spill in which other phenol deriva tives were present and a similar situation has occurred in Sturgeon, Missouri, a few months ago (documented in the files of the Regional EPA office in Kansas City, Mo.). The references in the Deichman papers listed in the criterion document also contain some human exposure data. It appears that the authors of this criterion document were focusing on papers dealing with industrial exposure to 2-chloro phenol and thus did not look to the poison control or accidental spill reports in the medical literature as a source of toxicity data for 2-chloro phenol in man. Poison control reports usually are not very satis factory for establishing toxicity standards in the absence of supportive animal toxicity studies, but in this case it is possible to make some fairly firm predictions on the basis of the documented structura 1-act i vity correlations within the phenol derivatives. Such predictions would be more meaningful than predictions based on subjective criteria such as odor and taste which have no clear relationship to injury, tissue damage or adverse effects on human health. If the animal toxicity data and the human exposure information are inadequate, CMA. 049614 2 an appropriate recommendation would be to obtain additional toxicity information before attempting to establish such criteria. SPECIFIC COMMENTS: Page C-l. Reference to potential industrial exposure does not indicate the route. Page C-U. Suggests that inhalation exposure is not a general population threat. Inhalation exposure is a major factor in most spill situations since dermal exposure is well documented and generally protected against. In the Sturgeon, ,*Mi ssour i , situation, for example, only one of the workers involved in cleaning up the contaminated track area exhibited dermal effects, but all of these workers plus the 800 people living in the town inhaled the fumes from the spill for a period of several weeks. The document suggests that chloracne could occur as a result of exposure but there is no documentation for this and such an effect has not been reported,t to my knowledge. The proposed criterion for human health is over two orders of magnitude below that proposed for aquatic life. The data referenced in this report does not support this magnitude of sensitivity for aquatic life when compared to mammals. There is a lack of consistency in the recommendations for the phenol derivatives included in this series of ambient water quality criteria documents. It would seem to be more appropriate to consider all of these agents as a group in view of their metabolic and toxicologic similarities and to make general recommendations for the group. CMA 049615 EVALUATION OF THE PROPOSE D AQUATIC LIFE CRITERIA ON DICHLOROBENZENES Conclusion: The alternative procedures (44 FR March 15, 1979; Appendix B-l) used to establish water quality criteria for dichlorobenzenes have not been adequately verified. Consequently, criteria based on these procedure are not scientifically defensible. Prepared by: Carl F. Muska E. I. du Pont de Nemours and Company Haskell Laboratory for Toxicology and Industrial Medicine Newark, Delaware 19711 May 23, 1979 CMA 049616 General Comments. Due to the general lack of aquatic toxicity data for dichloro benzenes, EPA has relied heavily on the alternative procedures (44 FR March 15, 1979; Appendix 3-1) for estimating water quality criteria for this group of compounds. The assumptions used to justify these alternative procedures have not been adequately verified. The practice of extrapolating toxicity data for one compound" (in this case 1,2-dichlorobenzene) to structurally similar compounds (1,3-ar.d 1,4-dichlorobenzene) is a question able validity and should not be considered as a substitute for actual test data. Specific Comments Page B-l When generalizations are made concerning tne toxicity of a compound, the type of toxicity should be specified (i.e., acute, chronic). For example, comparable data for the biuegill, Daphnia magna and Selenastrum capricornutum suggest no appreciable differences in the acute toxicity of 1,2-, 1,3 and 1,4-dichlorobenzene. Chronic toxicity data are available for only 1,2-dichlorobenzene. Similarly, the data in Table 1 indicate that the position of chlorine atoms on the benzene ring does not appreciably influence the acute toxicity of dichlorobenzenes. The influence of chlorine atom position on the chronic toxicity of dichlorobenzenes has not been studied. CM*. 049617 -* Page B-2 The document refers to the chlorinated benzene criterion document for details concerning the relationship between degree of chlorination and acute toxicity. This document was not available for review. The adjusted 48-hour EC50 (Daphnia rr.acna) for 1,3-dichloro benzene (23,800 ug/1) is more than an order of magnitude greater than the adjusted 48-hour EC50 for 1,2-dichlorobenzene (2,070 ug/1). Yet, in the Introduction section (p. B-l), the document states that the data suggest ... "no appreciable differences in the toxicity of 1,2-, 1,3- and 1,4-dichlorobenzene." Order of magnitude differences in acute toxicity are generally considered "appreciable." Page 3-6 The data used to formulate criteria are from an unpublished EPA source (U.S. EPA, 1973). The Final Acute Values are based on data for only two aquatic species - bluegill sunfish and Daphnia m.nnna. This is an extremely limited data base;' toxicity data on more species are needed. The data used to derive the sensitivity factors for fish (3.9) and invertebrates (21) should be carefully revaluatcd to determine whether the difference between these factors actually reflects a greater range in sensitivities to toxic materials for invertebrates as compared to fish or is a function of the methods used in the Guidelines for calculating the sensitivity factors. CMA, 049618 1 Pace B-S No Final Chronic Value for either 1,3- or 1,4-drchlorobenzene is derived using the Guidelines because no chronic toxicity data are available. The rationale used by EPA for estimating Final' Chronic Values for these compounds is as follows: since the experimentally determined Final Fish Chronic Value for 1,2dichlorobenzene (150 ug/1) is greater than .44 times the Final Acute Value, the 24-hour average criteria can be estimated for other .-dichlorobenzenes by multiplying 0.44 times the Final Acute Value. This procedure (44 FR March 15, 1979: p. 15373 , 2.A.) implicitly assumes that because 1,2-dichlorobenzene is structurally similar to 1,3-and 1,4-dichlorobenzene it can be expected that 0.44 times the Final Acute Value will be less than a Final Chronic Value based on actual test data. There is no basis for this assumption. Criteria should not be formulated until appropriate chronic toxicity data are available. Page B-20 No chronic toxicity data with saltwater species are available; yet, 24-hour average criteria are estimated based on alternative procedures (44 FR March 15, 1979; Appendix 3-1). The assumptions used to develop these alternate procedures have not been adequately studied. CMA 049619 AN EVALUATION OF THE EPA AMBIENT WATER QUALiTY CRITERIA DOCUMENT ON nichlorobenzenes CONCLUSION: The criteria established for dichlorobenzene is too low and is not defensible based upon the paucity of experimental toxicity evidence cited. Prepared by: Ur. Richard A. Carchman and Dr. Joseph F. BorzoMcca Division of Toxicology Department, of f'harmarolo'jy Medical College of Virginia Richmond, Virginia Nbi.'OfJ 23 April lo;o q49620 CWA General Co^ents The present criteria document contains a very comprehensive (i.e. in terms of the published literature), although superficial, review of the specific literature on dichlorobenzenes. It is difficult to evaluate the usefulness and validity of many of the studies reported since the information presented is inadequate. For example, details of the experimental protocols (i.o. dose of OCG, duration, number of animals, route of administration, etc.) are o'ten lack ing. The data on carcinogenicity and mutagenicity is particularly noteworthy in this regard. The potential carcinogenicity or mutagenicity of DCC is significant since it may be converted to reactive intermediates. However, the data are lacking. The tests systems reported were inappropriate (compared with the current standard testing procedures), uncontrolled, and the doses used very high (greater than those used in the acute and chronic animal studies reported). The criteria value presented is low, and probably not defensible due to the paucity of experimental details on the acute toxicity of DCU, and the lack of well-controlled chronic toxicity tests. It is important to note that the authors of this document conclude that "there is not sufficient evidence from human or animal tests to qua!itatively suggest that DCBs are carcinogenic or mutagenic in mammals." A major potential health concern might bo the interaction of the llfnn with other exogenous and endogenous compounds, since iirIts can induce: K'O (mi 'od function oxidases) in the livin'. This could bo a latter in sotting the (>iLorin values. -1- CMA. 04 9621 A major concern with most of the studies reported is that very high doses were required to produce toxic effects. Lxpected exposure levels were lower by orders of magnitude. C-l Although more highly halogenated benzenes Imve been shown to be persistent in the environment, the data suggests that this is unlikely with OCbs. C-2 Table 2 is difficult to comprehend without more detail. C-S T | rC1 data includnd on air contamination is m i n ima 1 . Other pes 11 l idm, ippcur to be a much yreoter problem than DCCs. C-1J -lb There is a paucity of experimental details, [.valuations cannot be made. The doses used in these studies were very large. C-16 A potential concern with DCBs is their lipid solubility and persistence in the body (cumulative toxicity). Concern is lessened since these compounds are metabolized extensively. C-Ui The animals were intubated with 1 ,2-DCB dissolved in water. It has been previously reported that these compounds are highly lipid soluble. The water solubility data arc questionable. C-22 As the authors indicate, no data are reported, just supposition. However, the fact that unmetabolized DCBs may mm active nucleophilic ai er.e omlP intermediates warrants further investigation, pariicularly sumo if is those compounds that would be involved in c arc i noyrn i r i fy ami ii.ut agm u. i fy. C-23 -24 The number of reported cases of OCD poisoning is very small (22) and these were mostly occupationally related. The toxicitics resulted from acute and -2- CMA 049622 chronic exposures to very high levels of DCBs. The intoxicated individuals were exposed to other substances. C-32 -33 C-3fi -41 The details on drug metabolism presented are not relevant. The discrepancy between Hollingsworth's and Varsharskaya1 s data is sigmticant. Since the mute of exposure in the latter study is not mentioned, an evaluation cannot be made. C-39 R C-42 The relevance of using subcutaneous or i.in. infections of UfCs to evaluate their health hazard is not clear, since the prim,,r/ routes of exposure to these compounds arc innalation and oral. C-4S A potential toxic effect of these compounds is in their interaction with other compounds. This needs more evaluation. r.-V -5H The available relevant data on teratology, carcinogenicity, and mutagenicity is minimal and a formultion of a criteria value based on the data presented would be limited. No information is available on teratology, and no biochemical studies on mutagenesis have been conducted. Chromosomal aberrations wetc only reported at very high doses of the chemicals and many of the studies appear to have been done quite crudely without regard to established scientific protocol. C-54 1,2-DCB was not found to be mutagenic in the Ames assay (Salmonella typhimuriurn). This is not conclusive evidence (see recent review by !'. Jacobs, ,1. Environ. Path, fc Tox. 2:1205-1215, 19/9). These compounds, in the appinpriule concentration ami duration oi exposure may lie mutagens, I url in-1 mom . some false negatives iiave been reported in the Ames test. The tests reported here did not include a drug metabolizing or activating system, and si nee it is the reactive oxide intermediates of DCBs that are suspect of carcinogenic and -3- CMA. 049623 mutagenic activity, you would not e-M <'Ct those impounds to be nusitivn. C_54 The tests for carcinogenicity are also lacking. The skin test is not an appropriate test for either the tumor-produciny or tumor-pronoting activity of potential carcinogens. Although the studios suggest that there is a potential for carcinogenicity, a more rigorous evaluation of the activity of the DCBs using established test' ing protocols is essential in order to evaluate these compounds properly and establish an appropriate criteria. * -n- CMA 049624 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON DICHLOROETHYLENES Conclusion The only numbers that can be developed using the guidelines are the maximum concentrations. Chronic numbers need to be developed using the procedure outlined by the Agency. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical Company Midland, MI 48640 June 1, 1979 CMA 049625 1,1 Dichloroethylene The final acute value based on two experiments using Daphnia r.agna must be suspect until the differences between the experi ments are resolved. As a general comment it is difficult to accept Daphnia magna as a representative of the total inverte brate population and using the safety factor of 21 to account for species variation. At best the final value as listen m the criteria should have a confidence interval of about 6, i.e., the 1200 could just as well be 7200. Much more confidence can be placed on the final acute value of 16,000 yg/L. As the guidelines indicate, no final chronic value can be estimated until the experiment is conducted. We do not agree that the factor 0.44 should be used until this factor is sub stantiated by further analysis. In the case of the saltwater organisms the only value that has any degree of confidence is the 37,000 yg/L for fish. Again, if the invertebrate data on mysid shrimp is used then the associated criteria should indicate the high degree of uncertaint CMA 049626 1,2 Dichloroethylene As we indicated in our general comments, we do not believe that the alternate guidelines as expressed in Appendix B-l of the March 15 Federal Register should be used. Accordingly, the only value that can be accepted is a final freshwater acute value of 19,000 yg/L. Our recommended criteria would be as follows. 1.1 Dichloroethylene - The criterion to protect freshwater aquatic life should not exceed 16,000 yg/L in any 24 hour period The ambient level will be established at a later date. For salt water species the concentration should not exceed 37,000 ug/h in any 24 hour period. 1.2 Dichloroethylene - The criterion to protect freshwater species is a concentration of 19,000 yg/L in any 24 hour period. For saltwater species a value cannot be derived at the present time. CMA 049627 EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR DICHLOROETHYLEN ES Conclusions: The evidence regarding the carcinogenicity of 1,1-dichloroethylene is overstated. The methodology used to derive the health based criterion is probably inappropriate. The significant literature on the toxicology of these compounds is adequately reviewed. Prepared byt Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 14, 1979 CMA 049628 G leral Comments; The criteria document on the dichloroethylenes cites most of the significant information on the toxicology of these three compounds. However, the evidence regarding the carcinogenicity data of 1,1-dichloroethylene seems to be overstated. The evidence for the carcinogenicity of this compound is qualitatively similar to that of selenium which was considered to be non-carcinogenic. The methodology used for the derivation of the health based criterion may be in appropriate for 1,1-dichloroethylene for the reasons noted under Specific Comments. Specific Comments'. Pages C-19 to C-21: How were the p-valuec for Tables l'J and 14 derived? The incidence rates for tumors do not demon strate a sizeable change from control values and their signi ficance should be evaluated with the Fischer Exact Test. Thus, the p-value for the incidence of mammary adenocarcino- 3 mas at a concentration of 40mg/m is listed as p=0.068, while by the Fischer Exact Test this value should be p=0.32. It is doubtful whether a substance should be called a carcinogen for the purpose of setting water quality criteria, if the only positive tests can be found in inhalation studies and the positive results are confined to mice. In addition, the kidney adenocarcinomas, which form the basis of the pro posed criterion, are only found in male mice and not to any CMA. 049629 4 significant extent in female mice. The incidence of mammarytumors in rats does not show a convincing dose response rela tionship (figure 14), and hamsters failed to produce any tumors in an inhalation study (Maltoni, 1977). Rats did not develop a significant incidence cf tumors when 1,1-dichloroethylene was administered in the drinking waterCRampy et al. 1977). In addition Ott et al.(1976) did not find any abnor malities associated with exposures of 9 to 280 mg/m3 in a population of 138 workers. If one calculates the absorbed dose of 1,1-dichloroethylene at the TLV averaged over a week (see page C-24), then a worker should absorb approximately 114 mg, or 114,000 ug per day. According to the proposed criterion, the risk _ct _6 per dose is 10 /2.77 = 3.61 x 10 . Therefore the estimated risk to humans exposed to the TLV level according to the single hit model can be calculated as P l-exp(-114,000 x 3.61 x 10"6) = 0.337. Considering that some of the exposures reported by Ott even exceeded the present TLV, it is astonishing that he did not find any. excess cancers. It appears likely that the combina tion of animal models used, and the single hit extrapolation model have produced a risk estimate that is excessive. CMA, 049630 AN EVALUATION' OF THE PROPOSED AQUATIC LIFE CRITERION ON 2,4-DICHLOROPHENOL on e;:ion of aquatic life the criteria should he based on and not taste. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical Company Kid land, MI 4 S 64*0 June 1, 1979 CMA 049631 GENERAL COMMENTS Kirk and Othmer is not an appropriate reference source for key data. V.'her.ever ' possible primary references should be used. For example, using Kirk and Othmer*s encyclopedia to establish that 2,4-Dicnlorophenol is a metabolic intermediate lacks the credibility that citing the original reference would have. In % this same context the reference for the octanol/vater partition coefficient should be given. The introduction admits that there is little data on 2,4-DCP levels in either effluents or natural waters. In addition, it is also admitted that there is a paucity of information oh aquatic toxicity. In spite of these admissions the conclusion is drawn (p. A-4) that 2,4-DCP represents a potential threat to aquatic and terrestrial life, including man. While it may be a "potential threat" (in those cases of ar. inadvertent loss) there certainly is not a threat under the present or past use conditions. CMA. 049632 CRITERION' FOR;ILLATION (p B-5) If tainting is to be used for setting a 24 hour average number, then it should be so stated in the criterion statement. Other wise the value could be misrepresented as having some basis for toxicity which it does not have. I-- is our conclusion that a chronic value needs to be established by the Guidelines to be used as the 24 hour average. The excur sion as we have said before (see comments on 2-chlorophenol) should be based on either the final fish acute value (700 ug/L) or an invertebrate value designed to protect 10-151 of the popu lation. CMA 049633 EVALUATION OF OF TLE HEALTH EFFECTS THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR SECTION 2,4 Dichlorophcnol conclusions: The criterion is appropriately formulated. The document contains some misinterpretations, but otherwise adequately reviews the significant toxicological literature. Prepared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 .May 10, 1979 CMA 049634 General Comments: With the exception of some of the data noted below, the criterion document for health effects of 2,4-dichlorophenol is well-written and takes appropriate notice of the experimental results that are currently known. It should be noted that this criterion for human health is based on the prevention of taste and odor problems from 2,4-DCP and that the criterion for the direct protection of human health lies significantly above the ta^te and odor threshold. Detailed Comments: Pages C-l to C-3. The dilution of a 2,4-DCP effluent in water containing 100,000 mg/liter total solids to a fresh-water total solids concentration of 1,000 mg/liter constitutes only a 100-fold dilution. It is assumed that this worst case dis- ' charge after only 100-fold minimal dilution will serve as a source for drinking water and that it would be treated in a fashion that would not permit the removal of any 2,4-DCP. These assumptions are highly unrealistic and it is dubious that this example can stand even as a worst case exposure level. Page C-ll. 0.51 ug per g of kidney were found in cattle that were fed 300 ug of 2,4-D per g of feed, which corresponds to approximately 9 mg of 2,4-D per kg of body weight. A con centration of 300 yg of 2,4-D per kg of feed cannot be considered to be realistic in a field exposure. CMA 049635 Page C-14. The consumption of chicken liver and cattle kidney by man would result in approximately equivalent exposures to 2,4-DCP only if those cattle had been fed relatively high concentrations of 2,4-D and the chickens had been fed high concentrations of Nemacide. Whether the concentrations of , these chemicals fed to chicken and cattle have any relation to normal environmental exposures is very questionable. Pages C-14 to C-15. The worst case estimates regarding the intake of 2,4-DCP from contaminated meat and water must be questioned. It is unlikely that a human could consume 0.5 kg of kidney daily. The applicability of the estimate for 2,4-DCP levels in water has been questioned previously. It must therefore be concluded that the worst case estimate of consumption of 4 yg of 2,4-DCP per kg of body weight per day represents an over-estimate, even for a worst case. Page C-17. The calculated bioconcentration coefficient includes so many assumptions that it must be considered to very tentative at best. Pages C-16 to C-17. The available data do not permit an estimation of the ingested 2,4-DCP. No marketbasket surveys for 2,4-DCP are available and any estimates based on laboratory studies involving precursors for 2,4-DCP can provide only the most general indications of even a worst case incidence. Table 7 should appear immediately after page C-26. Pages C-25 to C-28. The promoter study by Boutwell and Bosch (1959) is clearly not designed as a carcinogenicity study. 04963g The protocol is not remotely related to the ingestion of 2,4-DCP and the degree of emphasis given to this study is totally inappropriate. The study is not even applicable to the evaluation of the hazard of a dermal or respiratory exposure to 2,4-DCP. Page C-23. while it is appropriate to cite the studies on taste and odor by Burtschell and by Hoak, it should also be pointed out that the protocols for the determination of taste^and odor thresholds are very poorly described in both studies, even though it is well known that the determination of the thresholds is highly dependent upon the exact protocol. Regretably these poor descriptions of protocol are shared by most other publications dealing with taste and odor thres holds. Consequently, the certainties attached to taste and odor thresholds are relatively poor. In .addition, it not often clear whether a taste and odor threshold refers to an unaccept able taste or odor which might affect the consumption of water. Pages C-30 to C-33. The criterion for 2,4-DCP is appro priately formulated. CMA 049637 EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERIA ON 2,4-DIMETHYLPHENOL Conclusion: Freshwater criteria for both maximum and 24-hour average concentrations are based on one static acute test. Basing national water quality criteria on such a limited data base is not scientifically defensible. Prepared by: Carl F. Muska E. I. du Pont de Nemours and Company Haskell Laboratory for Toxicology and Industrial Medicine Newark, Delaware 19711 May 23, 1979 CMA 049638 General Comments: Proposed ambient water quality criteria for 2,4-dimethylphenol for the protection of aquatic life are based on data from two unpublished studies. The data base includes only three acute and one chronic test. Water quality 'criteria for 2,4-dimethylphenol should not be derived until more data are available. The 24-hour average criterion is calculated by multiplying the Final Invertebrate Acute Value by a 0.44 sensitivity factor. The Final Invertebrate Acute Value is based on only one static acute test with Daphnia magna. The 48-hour LC50 is corrected by an "adjustment factor" (unmeasured to measured concentrations) and a "sensitivity factor." The end result is a proposed national water quality criterion based on one acute invertebrate toxicity . test to which one "adjustment factor" and two "sensitivity factors" are applied. These mathematical procedures are used in preference to empirically derived chronic toxicity data on fish. Because of the mathematical manipulations used to derive the criteria for 2,4-DMP, the results must be regarded as unacceptable until additional aquatic toxicity data are available and until the Guidelines (44 FR March 15, 1979; Appendix B) have been thoroughly evaluated. Specific Comments: Page A-l The document states that ... "its [2,4-dimethylphenol] toxicity to aquatic life, and its carcinogenic effect in mammals clearly indicate the potential hazard of 2,4-DMP to aquatic and CMA. 049639 2 Page A-l (Continued) terrestrial life." This statement implies that 2,4-DMP is carcinogenic; however, on page C-35 of the document, a report by the Carcinogen Assessment Group concludes that ... "the data in Boutwell and Bosch regarding possible carcinogenic effects (via initiation) of 2,4-dimethylphenol are inconclusive." An apparent contradiction exists in the document concerning the carcinogenic effects of 2,4-DMP. Page 3-1 The Final Fish Acute Value (2,200 ug/1) is based on the geometric mean of only two LC50 values, both of which are from unpublished sources (Phipps, et al., manuscript; U.S. EPA, 1978). The manuscript by Phipps is not readily available and therefor? could not be reviewed. The Final Invertebrate Acute Value (86 yg/1) which becomes the Final Acute Value is based on only one static acute test with Dapnnia magna. The 48-hour EC50 was "adjusted" from 2120 yg/1 to 1796 yg/1 and then divided by the "sensitivity factor" of 21. Basing criteria on such few data is not scientifically defensible. Page B-2 The relevance of the information on the relationship between the acute toxicity of chlorinated phenols to bluegills and the degree of chlorination should be explained. Unless these data can be related to developing criteria for 2,4-DMP, it should be deleted. CMA. 049640 3 Page B-2 (Continued) The statement that "since Daphnia magna appears to be more acutely sensitive than the fathead minnow or bluegill, a chronic test for this species would probably be lower than the Final Fish Chronic Value ..." is total speculation and has no place in a criterion document. The relationship between acute and chronic toxicity data for fish is not necessarily applicable to invertebrates such as Daphnia magna. Page B-3 Data are presented which indicate that exposure of fathead minnows to 2,4-DMP for 192 hours does not cause any appreciable cumulative mortality as compared to the 96-hour LC50. Data of this type suggest that sensitivity or application factors as derived in the EPA Guidelines (44 FR March 15, 1979; Appendix B) may not be applicable for 2,4-DMP. Additional information, particular chronic toxicity data for invertebrate species, is necessary before defensible water quality criteria for 2,4-DMP can be established. CMA 049641 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR 2,4-Dimethylphenol by John Doull, M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: It is not clear why organoleptic criteria are used for some phenol derivatives and not for others. Since toxicity data rather than esthetic observations should be used to establish human health ambient water quality criterion, I would agree that there is insufficient toxicity data to establish a criterion for 2,4-dimethylphenol as proposed in this document. CMA 049642 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON 2,3,7,8-TETRACHLORODIBENZO-P-DIOXIN Conclusion He agree that there is insufficient date to develop a criteria ai this time. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical Company Midland, MI 48640 June 1, 1979 CMA 049643 GENERAL COMMENTS: This document concludes that no criterion for human health effects due to exposure to 2, *4-dimethylphenol can be established because of the lack of mammalian toxicity data and human health effects. It should be pointed out, however, that the document does contain organoleptic data for this agent which is, in fact, similar to that for 2-chlorophenol, and that the human health criterion for 2-chlorophenol was based on the organoleptic effects of that agent. The establishment of human health criterion for phenol and its deriva tives should be consistent and should reflect the state of our toxicologic knowledge about this series of agents rather than the lack of information about specific members of the series. The aquatic organism criterion for 2,4-dimethylphenol is less than that for 2-chlorophenol, which suggests that the methyl derivative is more toxic, but comparison of the two documents does not support this difference. This document does contain information on the acute toxicity of phenol derivatives (page C-2M and subchronic data for such derivatives (page 28). This kind of information, together with our rather considerable knowledge of how these agents are metabolized, would provide a reasonable approach to making recommendations for the entire series. It is surprising, therefore, that this approach was not used. SPECIFIC COMMENTS: Methyl is misspelled on pages C-19, C-22 and C-23- Cm 049644 GENERAL COMMENTS GCDD is not a contaminant of 2,4-dichlorophenoxy acetic acid (2,4-D) and references to 2,4-D should be removed from the document. References to the reported persistence of TCDD of 10 years should be included as well as the reported bioaccumulation factor of 3000 . CRITERION DEVELOPMENT Since no criteria are reported no comments can be made. CMA, 049645 EVALUATION OF OF THE HEALTH EFFECTS THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR SECTION 2,3,7,8-Tetrachlorodibenzo-p-dioxin Conclusions: The criterion for human health is appropriately derived, assuming that the single-hit extrapolation model is applicable. The significant toxicological literature is adequate ly reviewed. Prepared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 10, 1979 CMA 049646 General Comments: ' The document on TCDD is, in general, well written. TCDD clearly is an extremely toxic compound. The criterion for human health is appropriately derived, if one assumes that a single-hit extrapolation model for the estimation of carcinogenic risk in man is applicable. Detailed Comments: Pag'e C-4. The calculated dose for TCDD for man represents an unrealistic over-estimate. The concentration of 4 lbs 2,4,5 T per 10 gal of water per acre obviously constitutes a highly concentrated spray suitable only for aerial appli cation. The assumption that a person would spend 8 hours continually underneath the application pattern of the spray plane is highly unrealistic. Page C-16. The statement on carcinogenicity and feto- toxicity on the bottom of page C-16 is incomplete and does not make any sense. Pages C-20 to C-23. The sections on the inductions of various metabolic pathways should be more appropriately incor porated into the section on metabolism rather than on muta genicity where they are at the present time. Page C-29. At a postulated risk level of 10 the concentration of TCDD in drinking water combined with consumption of fish should be 4.55 x 10 7 ug per liter, rather than 455 x -7 10 ug per liter. CMA 049647 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON FLUORANTHENE Conclusion Acceptable criteria for fresh and saltwater organisms could not be generated due to the extreme lack of data and the fact that the data is in an unpublished report. Prepared by Albert C. Hendricks Route 4, Box 1023 Christiansburg, Virginia 24073 501 Bishop and John Cairns, Jr. Road, Blacksburg, Virginia 24060 April 12, 1979 CMA 049648 EPA Criteria: Freshwater Saltwater Res*idues Summary Sheet FLUORANTHENE 250 yg/1 as a 24-hour average and 560 yg/] as the maximum concentration. 0.30 y/1 as a 24-hour average and 0.69 yg/1 as the maximum concentration. No steady-state bioconcentration factor (DCF) is available for this compound. An estimated DCF of 3,100 was generated. Conclusion: Due to the extreme lack of data (one data point per organism) for fluoranthene in regard to aquatic toxicity testing and the fact that the data came from one unpublished report, makes it impossible to generate acceptable criteria for fresh and saltwater organisms. . CMA 049649 General Comments The only aquatic toxicity data for fluoranthene comes from one unpublished EPA contract. Until these data are published in the open literature, the criteria as proposed lack credibility. The EPA is attempting to establish criterion from one data point for this compound. Statistically this approach cannot be justified and scientifically it cannot be supported. Specific Comments p. B-l. , (1) The fact that the fluoranthene level in all the tests were not measured reduced the credibility of the test a great deal. (2) The work of the EPA (1978) has not been published and cannot be properly evaluated. p. B-2. The document states that the bioconccntration factor (DCF) has not been measured but it goes on to state that an octanol-water partition coefficient could be calculated and indeed it does arrive at the number 79,000. However^ the document does not state how this number was derived. It is assumed that the calculated octanol to water partition coefficient came from the reference sited in the Federal Register, Vol. 43 #97 p. 21509 which is "Leo, A. J., 1975, In Symposium on Structure-Activity Correlations in Studies of Toxicity and Bioconcentration with Aquatic Organism. G. D. Vcith and D. E. Konosewich, eds. International Joint Commission, Windsor, Ontario, pp. 151-17G. This reference is not specifically referred to, therefore, one is not sure that this was the one used. p. B-3. Line 10 up from bottom of page states that no Final Chronic Value was obtained for freshwater organisms. However, three lj.nes above this we find a Final Chronic Value of 54,000 ug/1. This is very confusing. Also, it appears that this 54,000 yg/1 comes from plant tests, and it is not clear whether the plant tests are acute or chronic. p. B-3. The EPA is attempting to estimate a criterion for freshwater from saltwater organism. However, they have already arrived at a criterion from the freshwater data. I assume that they are attempting to justify the freshwater criterion with the saltwater data rather than estimate it. The freshwater criterion levels are so much greater than the saltwater levels that it become very difficult to even try to justify one with the other. CMA 049650 p. B-8. For saltwater organism:;, i^e. , the myrid shrimp, the Final Invertebrate Chronic Value (also the Final Chronic Value) was 3.1 wg/1 while the Final Invertebrate Acute Value was 0.69 ug/1. In this instance we have a final chronic value that is higher than the Acute value (approximately 4.5 times greater). The chronic value is, also 10 times' greater than the 24-hour average value. Obviously, there is a great deal of unnecessary overprotection by using these "fudge factor" derived acute values as opposed to the chronic values. Criterion Formulation Due to the extreme lack of data (one data point per organism) for fluoranthene in regards to aquatic toxicity testing and the fact that the data came from one unpublished report, makes it impossible to generate acceptable criteria for fresh and saltwater organisms. CMA 049651 AN EVALUATION OF THE EPA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON fluoramhlni: CONCLUSION: There is not sufficient data to justify the formulation of a water quality criteria level of fluoranthcne. Prepared by: Dr. Harry L. Skalsky and Dr. Joseph T. llorrol locn Division of Toxicolnii/ Department nf Phaniwu o Imp/ Medical College of Virginia Richmond, Virginia 2329H 23 April 1979 CMA 049652 GENERAL COMMENTS It appears that there are not sufficient data to justify the formulation of a water quality criteria level for fluoranthene. This lack of data is further illustrated in the "Specific Comments" section of this report. The formulation of a water quality criteria level for fluoranthene is not justifiable on two major points. First, there are not sufficient data available to scientifically set an exposure level. For fluoranthene, there are no steady state bioaccumulation data (C-16), no pharmacokinctic data (C-21), no metabolic data (C-22), no data on excretion (C-23), and no teratogenic data (C-27). Secondly, the data that are available do not indicate that fluoranthene is a hazardous material. It is not particularly toxic; the LD^q in experimental animals is 1-5 grams/kg (C-24). Fluoranthene has displayed no mutagenic activity in in vitro systems (C-28) and has not been shown to be carcinopcmc-in experimental animals (C-29). Why is it being regulated? In the criteria document, constant reference is made to fluoranthene being a polycyclic aromatic hydrocarbon (PAll) and it is implied that it is toxic. There is a need to regulate total PAH concentrations because of the presence of known mutagens and carcinogens, such as bcnzo(a)pyrene. Individual PAHs should be considered separately. Fluoranthene does not appear to be a hazardous chemical. This is not true of some other PAHs. CPA should continue to regulate known carcinogens, such as benzo(a)pyrene or benz(a)antliracene. Another point that must be addressed is the proposed figure of 200 ug/l. This limit has been recommended and calculated as a drinking water standard (C-45). However, on page C-46, it is clearly stated that "Inadequacies in the current data base prevent the formulation of a drinking water criterion . . -1- CMA. 049653 This contradiction within the document demonstrates that there are insufficient data for setting a criteria for fluoranthene. In addition, fluoranthene is effectively removed (87.5 to 1002) during the treatment of drinking water (C-5). CMA. 049654 SPECIFIC COMMENTS 13-1 Acute data in Fresh water organisms were collected using static test procedures and unmeasured concentrations. Data from static tests arc question able. For meaningful data, these tests should he repeated in flow-through systems with a known concentration of fluoranthene. B-2 There are no data on the chronic effects of fluoranthene on fresh water organisms. 13-2 There are no data on the steady state bioconcentration of fluoranthene. 13-3 "No fresh water criterion can be derived for fluoranthene using the guidelines - ESTIMATE of criterion calculated through data from salt water organism." The fact that these criterion are estimates has been overlooked, in many sections of the document, and these numbers are erroneously presented as being soundly established. This can be demonstrated in the abstract Page B-1 and B-3. Li-9 There is no measured steady state bioconcentration factor available for fluoranthene. C-3 Removal of fluoranthene from water hy convention^ 1 sewage treatment. 91-99% effective. C-5 Removal of fluoranthene during drinking water treatment 87.5-100"' effective. C-2 There are no data available concerning the pharmacokinetics of fluoranthene. C-22 There are no data on the metabolism of fluoranthene. - 3- CMA 049655 C - 23 The statement, "It is reasonable to assume that fluoranthene is metabolized in a manner which is consistent with the general biochemical scheme for biotrans formation of PAHs." This may indeed be possible, but there is no evidence to indicate that this is so. C-23 There are no data available concerning the excretion of fluoranthene. C-25 Fluoranthene presents relatively low degree of toxicity. C-27 There arc no data available on the potential teratogenic effect'; of f1uoranthene. C-28 Fluoranthene displayed No flutagenic Activity in the Ames assay. Duplicated in separate laboratories. C-?9 Fluoranthene has been tested for carcinogenic activity in several studies with negative results. C-33 -35 The data review does not indicate that fluoranthene is either an inhibitor or tumor promoter. SUMMARY There are not adequate data to establish a water quality criterion for fluoranthene, either on the basis of effects on human health or fresh water organisms. - 4- 049656 CM* ' An Evaluation of the Aquatic Life Toxicology Section of the EPA Ambient Water Quality Criteria Document (March 15, 1979) on Heptachlor Summary The Freshwater criterion to protect aquatic, organisms is based on dubious ciironic and biocon centration data and should not be considered as completely formulated. The Saltwater criterion to protect aquatic organ isms is correctly formulated, but the recommended 24-hour average concentration is not detectable in saltwater with current technology. Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and James L. Peterson, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19105 May 1979 CMA 049657 Proposed Criteria For heptachlor, the criterion to protect freshwater aquatic life as derived using the proposed guidelines is 0.0015 ug/1 (Residue Limited Tissue concentration derived from the FDA maxi mum heptachlor concentration for animal feed and the freshwater bioconcentration factor) as a 24-hour average and the concentra tion should not exceed 0.45 ug/1 (the final invertebrate acute value) at any time. The criterion to protect saltwater aquatic life as derived using the proposed Guidelines is 0.0036 ug/1 (Residue Limited Tissue concentration derived from the FDA maximum heptachlor concentration for animal feed and the saltwater bioconccntration factor) as a 24-hour average and the concentration should not exceed 0.05 ug/1 (final invertebrate acute value) at any time. Our Conclusions h'e find that the Criterion Formulation for freshwater aquat ic life is faulted by the inclusion of dubious chronic toxicity data and an inappropriate bioconcentration factor. It appears that no Final Chronic Value or estimated Final Chronic Value can be derived uder the Guidelines or published alternative proce dures . Ke find that the Criterion Formulation for saltwater aquatic life is apparently in proper form, with a minor correction (0.44 x Final Acute Value is 0.022 ug/1), but we note that the recom mended 24-hour average concentration is at a level below that detectable in saltwater with current technology. CMA 049658 General Comments The section on aquatic life toxicology is clearly documented and relatively easy to review. In general the studies which were included correspond to those which ought to have been retained in the process of screening the data base (Guidelines', Section III); we do find one significant exception, which is discussed below. We find that the Criterion Formulation for freshwater aquatic life is faulted by the inclusion of dubious chronic toxicity data and an inappropriate bioconcentration factor. It appears that no Final Chronic Value or estimated Final Chronic Value can be derived under the Guidelines or published alternative procedures. We find that the Criterion Formulation for saltwater aquatic life is apparently in proper form, with a minor correction (0.44 x Final Acute Value is 0.02 2 ug/1) , but we note that the recommended 24-hour average concentration is at a level below' that detectable in saltwater with current technology. While we agree with the expressed concern that this recommended 24-hour average concentration may not afford fully adequate protection to penaeid shrimp, that value is less than the un measured concentration reported to cause only IS" mortality in penaeid shrimp and perhaps is sufficiently small to protect all but 0-10" of penaeid shrimp in accordance with the Guidelines. We see no immediate need to modify the Guidelines in this case. Specific Comments Freshwater Organisms Chronic Toxicitv The chronic toxicity study by Macek, et al.(1976) with heptachlor and the fathead minnow should probably be regarded as unavailable under the Guidelines because it was subject to the same analytical difficulties in measuring heptachlor con centrations in water as other studies in that report which were not regarded as available for the criteria docunent. Further, the available statistical information on heptachlor concentra tions in Table 10 of Macek et al. (1976) shows that there has been some erroneous compilation of the heptachlor data in that table; specifically, some of the standard deviations are numer ically impossible if the means, sample sizes and ranges are correct. The exclusion of this study would then mean that no valid chronic test data were available for any fish species. CMA. 049659 Heptachlor Residues The doubts regarding the heptachlor concentrations in water in the fathead minnow chronic study extend to the bio concentration factors computed for the fish residue levels, which depend on accurate concentration measurements. Further, the concentration data from day 210 to 276 are stated to be unknown, and hence the last 28 days (the minimum period of interest for a bioconcentration study under the guidelines) has no available measurements of exposure concentrations in test water. Therefore the bioconcentration factors, even if correctly recalculated from Table 27 of Macek, et al. (1976), are unavailable under the Guidelines. Further the Residue Limited-Tissue Concentration (RLTC) , if it were to be based on the FDA animal feed action-level (0.05 mg/kg), would require residue measurements of whole fish, not eviscerated fish. Thus, the RLTC is apparently not available. Miscellaneous Perhaps the Dcphnia nagr.a study by Macck, ct al . (1970) warrants mention here, and in Table 5, even though the concen trations can only be reported as nominal. Criterion Formulation Freshwater -Aquatic Life Final Fish Chronic Value should be treated as not available Residue Limited Toxic Concentration should be treated as not available. Final Chronic Value should be treated as not avail able. It does not appear that any of the published alternative procedures can be satisfied as a means for estimating a Final Chronic Value for freshwater. Thus, the freshwater criterion cannot be completed under the Guidelines or the published alter native procedures. Saltwater Organisms Acute Toxicitv - The last paragraph discusses the frequent appreciable reductions of measured heptochlor concentrations relative to nominal concentrations. It is true that most tests 049660 CM&- Hep tachlor show markedly lower measured (than nominal) concentrations on average, but there are counter-examples, e.g., a few tests run by Schimmel, et al. (1976a, b) , in which mean measured concentrations exceeded nominal. Thus, it may not be warrant ed even to consider deriving an adjustment factor specific to heptachlor for measured versus unmeasured concentrations. Chronic Toxicitv Presumably the next to last sentence of page B-24 (begin ning "No other fish...") refers to the current absence of alternative saltwater fish for which life cycle tests have been developed. Residucs Schimmel et al. (1976b,) although only a 24-day study of spot, achieved steady-state bioconcentration values and is indeed available. Miscellaneous Concern was expressed over the adequacy of protection oifered to pink shrimp at the Final Chronic Value. This appears to be a possible issue, but as Stephan (1977) has pointed out, a generally different level of effort in testing would be required to obtain adequate LCij values. This may qualify as a case in which the species sensitivity factors warrant review. With other toxins, examples of concern in the other direction can be found. Criterion Formulation 0.44 x Final Acute Value = 0.022, not 0.22 ug/1. Additional Reference Stephan, C.E. 1977. Methods for calculating an I,C<o. pp. 6 5 - S 4 in Mayer, F.L. and J.L. Hamel inch (ed.O Aquatic Toxicology and Ha card Evaluation. ASTM STP 632. CMA 049661 EVALUATION OF THE PROPOSED ACUATIC LIFE CRITERIA ON HZNACHLGRCBUTADIENE Conclusion: Available data do not support the derivation of numerical water quality criteria. Prepared by Carl F. Muska E. I. du Pont de Nemours and Company Haskell Laboratory for Toxicology and Industrial Medicine Newark/ Delaware 19711 May 23, 1979 CMA 049662 General Comments: No criterion for the protection of either freshwater or salt water aquatic life was derived for hexachlorobutadiene because ... "no Final Chronic Value for either fish or invertebrate species or a good substitute for either value is available, and there are insufficient data to estimate a criterion using other procedures (p. 3-3; p. B-9)." Given the general lack of aquatic toxicity data for this compound, the criterion statement is appropriate. * The criterion of 0.9 yg/1 derived in the first draft of the EPA water quality criteria document for hexachlorobutadiene was not scientifically defensible. This conclusion is supported by the current document. Additional aquatic toxicity data are required before a numerical criterion for hexachlorobutadiene can be formulated. CMA 049663 EVALUATION or THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR Hexachlorobutadiene Conclusions: The health based criterion for hcxuch 1 orobut.udiene is defensible, assuming that a sjnqlo-hit ext rapulat ion model is applicable. The significant toxicological 1 i t c ra tun on hexachlorobutadiene is adequately reviewed. Prepared by: Rolf Ilartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 10, 1979 CMA. 049664 General Comments: In general, this document is well written. The survey of the available data is reasonably complete and the criterion which has been derived for protection of human health is appropriate, assuming that a single-hit model is applicable for this extrapolation. Specific Comments: Pace C-25. The statement on mutagenicity in S. tyohirnurium is misleading. Taylor (1978) classified HCBD as not proven to be mutagenic because of the low number of induced reversions. The existence of a slight dose-response relationship in the number of reversions, which did not reach twice the number of background reversions at the highest dose level tested, is insufficient to establish the mutagenic potential of HCBD. 049665 An Evaluation of the Aquatic Life Toxicology Section of the EPA Ambient,Water Quality Criteria Document (March IS, 1979) on Hexachlorocyclopentadiene Summary The freshwater criterion for the protection aquatic life is properly derived, but based data for only one species, Pir.e-pha.le3 prcr (fathead minnow). There are insufficient for a saltwater criterion to be derived for protection of aquatic life, and none was attempted. Cl of on Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and Janes L. Peterson, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19103 May 1979 OAK 049666 Proposed Criteria - Aquatic Life For hexachlorocyclopentadiene, the criterion to protect freshwater aquatic life, as derived using the Guidelines is 0.39 ug/1 (based on the final fish chronic values) as a 24hour average and the concentration should not exceed 7.0 ug/1 (based on the final fish acute value) at any time. For saltwater aquatic life, no criterion can be derived using the Guidelines, and there are insufficient data to estimate a criterion using other procedures. Our Conclusions The freshwater criterion is based entirely on data for fathead minnow. The recommended maximum concentration is properly derived and is presumably protective of fathead min now under laboratory conditions. The 24-hour average concentration has been adjusted well below the no-effect level for fathead minnow, but we have no basis for evaluating this as an extrapolation to other organisms or to field condi tions . No suitable data are available for saltwater criterion formulation. CMA 049667 General Comments Hexachlorocyclopentadiene The freshwater criterion for hexachlorocyclopentadiene is based essentially on the work of Spehar, et al. (in press), much of which was reported by Spehar et al. (1977). The work gives evi dence of careful quality control, as did the earlier work by Henderson (1956) . Soecific Comments Freshwater Chronic Toxicitv The third sentence here should begin "The Final Fish Chronic (not "Acute"). Admittedly there is limited evidence for hexachlorocyclo pentadiene, but this seems to be a case in which the guideline sensitivity factor is perhaps excessively stringent. This is much more stringent than necessary to protect the tested organ isms, so we cannot evaluate the suitability of the final chronic value for other organisms and conditions. Additional Reference Spehar, D.L., et al. (1977). A rapid assessment of the toxicity of three chlorinated cyclodiene insecticide intermediates to fathead minnows. Ecological Research Series, Report EPA-600/5-77-099. 50pp. CMA 049668 AN EVALUATION OF THE ERA AMBIENT WATER QUALITY CRITERIA DOCUMENT ON HEXACHLOROCYCLOPENTAPIENE CONCLUSION: It is evident that the information needed to accurately establish a criterion for HEX is lacking. Prepared by: Dr. John L. Egle, Jr. and Dr. Joseph F. Borzelleca Division of Toxicology Department of Pharmacology Medical College of Virginia Richmond, Virginia 23298 23 April 1979 CMA 049669 GENERAL COMMENTS The only relevant study of chronic effects was a 6-month study in rats involving 3 dosage levels. Neutropenia and a tendency toward lymphocytosis were the only adverse effects observed and only at the highest dose. Longer term studies and adequate evaluation of carcinogenic potential are lacking. Since the limited toxicity observed by Naishrtein and Lisnvslaya described in the document (C-7c>) as being possibly attributable to tub compound, selecting the higher value of the proposed range does rot see'' inappropriate. Thus 4.0 pg/1 appears to be as teasonablc a figure as 0.4 rg/1, bearing in mind that the data for establishing a criterion is totally inadequate at the present time. -1- 049670 SPECIFIC COMMENTS C-l "Many commercially important organochlorine pesticides" - This is no longer true. Compounds have been banned or use severely restricted. Main use now is in manufacture of flame retardants. Thus present exposure potential is considerably less than implied here. C-2 It is not clear from the single reference (Bell, 1978) whether use of HEX is declining or not. It is stated that use in pesticide manufacture is declining, but use in manufacture of flame retardants is incteasing. Is overall use (and hence environmental exposure potential) of HEX increasing or decreasing? It is assumed the latter is true, but this should be made clear. C-3 "Scores of workers" - Granted this section is intended to be a brief description of the Louisville problem, but a more quantitative term than "scores" would be appropriate. "Scores" is a dramatic term "indicating indefinitely large numbers" (Funk and Wagnall's Dictionary). This implies a massive disaster. C-4 Chemical properties are described on page A-l. No need to repeat here. C-5 Were lines or words left out of last sentence? Meaning is unclear as it now appears. C-9 Discussion of Spehar study is extremely vague. First, are sentences 2-4 in the 2nd paragraph related to the Speh'ir reference? (Hiere is no other here). Sentences 2-4 raise such questions as: What percentage of fish contained HEX? What were the exposure levels? What were the HEX levels found in fish? - 2- CMA 049671 C-13-15 The Kommineni study is not really a pharmacokinetic investigation. Devoting 3 pages to gross and histopathological observations under the heading of pharmacokinetics is inappropriate, incorrect, and contributes to the excessive length of this document. C-19-25 These acute toxicity findings have no bearing on the criterion. They should have been summarized briefly. C-28 As noted, the deaths in the 1972 IRDC study were not necessarily due to Hx. C-30-31 Dermal study. It is unclear how many concentrations of HEX were used in this study. Apparently, only two were involved which is unfortunate since the lower level was toxic, especially to mice. C-54 Item 4? Where are items 1-3? C-69 C-77 Elevations ... in what? -3 3 8 hr. resp vol of 10M too high; Int. Comm Radiol. Protect, says 7.(1 M . C-78 0.7857 tmj/day should be termed ADI (is 0.60 if 7.6 n:' is used). Assumption of 100" resp. absorption - real value almost surely less. Criterion is calculated in very confusing and perhaps incorrect manner. The relevant formula from the "Guidelines and Methods" is: CR = ACI/WC + (R x F) where CR = criterion ADI = acceptable daily intake (0.7857 mg/day) WC = daily water consumption (2 liters) R = Bioconcentration factor for fish (3.2) F = Avg. wt. of fish consumed/day (0.0187g) CMA 049672 C R = 7857 + 0.06 2 CR = 0.45 mg/1 Thus the figure should be T5.. higher than shown, when derived from the TLV value. -4- 049673 CMfv Comments on the Aquatic Toxicology Section of the EPA Ambient Water Quality Criteria Document (March 15, 1979) on Lead Summarv The modification of the proposed guidelines per* mitted proper adjustment for lead toxicity as a function of hardness in freshwater. The fresh water criteria are properly derived and reasonably valued under tested conditions. The data base' is insufficient to complete saltwater criterion formulation. Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and James L. Peterson, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19105 May 1979 CMA. 049674 Proposed Criteria - Aquatic Life For lead, the criterion to protect freshwater aquatic life as derived using the Guidelines is "cO-51 In (hardnes$)-3.37)M ns a 24-hour average (based on the slope for fish acute studies and the adjusted mean intercept for fish chronic studies') and the concentration should not exceed "eO`51 ln(hardne$s)-l.39) " (based on the slope for fish acute studies and the adjusted mean intercept for invertebrate acute studies) at an/ time. The following table illustrates the meaning of this cri terion,for waters at specified hardness. Hardness (mg/1) 24-hour Average Cue/1) Maximum C u g /1) 20 5.2 40 9.0 100 56. 500 189. 25 65 261 1570 Our Conclusions It is important that the proposed guidelines have been modified to permit adjusting the criteria for hardness or re lated water quality parameters. The application of the fish acute slope to other kinds of data, although required by the proposed guidelines, is an extrapolation without scientific support in the document. The 24-hour average and maximum concentrations have been properly derived under the proposed guidelines. The 24-hour average appears to be properly protective of aquatic life under laboratory conditions. There arc insufficient data to complete the saltwater criteria formulation process. 049675 General Comments Lead This document generally makes full use of the available data on the toxicity of lead ko aquatic life. Our comments are generally concerned with the adequacy of the data base for criterion formulation, particularly since water quality stan dards, for particular uses and bodies of water, may be established on the basis of section 304(a) criteria. Lead is one of the metals with highly variable toxicity in different freshwaters; this variability has usually been assoc iated with the protection afforded to aquatic life by increased water hardness. It is at least plausible that alkalinity, rather than hardness, is directly responsible for (and predic tive of) the reduced toxicity of lead in freshwaters of high hardr.es's or high alkalinity. The data summarized in the criteria document do not provide sufficient information to distinguish between these two correlated water quality parameters in freshwaters. The supposed relationship of toxicity to dissolved lead, to tlac extent that this had been or might become clearly documented, deserves some treatment in the criteria document, even though the criteria themselves are to be based on total lead. Specific Comment's IntroJuetion (pp. A-l to A-5) - The third sentence is not a very useful summary of the solubility of ionic lead, since pH is not the only major influence; also, Durum, 1973, is not in the list of references. Perhaps the low lead concentration reported in seawater deserves more comment, since the usual table (reprinted from Sverdrup, ct al. (19421) gives the average lead concentration in seawater as 4-5 yg/1. It might be useful to summarize the Kopp and Kroner (1967) data, perhaps by plotting them against hardness ns an indication of how well water quality conforms to the recommended freshwater criterion for lead for the protection of aquatic organisms. Perhaps the Davies ct al. (1976) paper slum Id he referenced on the second line of p. A-2 by way of providing a thermodyna mic explanation for the reduced toxicity of total lead in hard water. cMA 04967S Lead The acute and chronic toxicity to vertebrates mentioned on p.A-1 are not ci.ted in the aquatic toxicology section. Freshwater. (3-1 to B-20) Acute Toxicity (B-l to B-S) On page 3-3, line 9, the "test water" should be the soft test water. The question can be raised, on the basis of the aqueous chemistry of lead, as to whether alkalinity might be more pre dictive,than hardness in selecting a water quality measurement on which to form, a dependent relationship for lad toxicity. This would permit a partial evaluation of the exponential relationship, since rainbow trout have been studied at four alkalinities (30, 103, 22S and 267 mg/1). (Hale, 197"; Davies, et al., 1976), three of which have measured concentrations. For blucgill and fathead minnow, no intermediate values of hardness are available, and the criterion document did not list alkalinities (since these are not needed if hardness is retained) . The Davies, ct al. (1976) data for hard and soft-water measured total lead compared with nominal total lend sugge-t that the guideline value of 0.77 may be inappropr into for lend for converting unmeasured to measured concentrations, at least in static tests. The reported (p.B-i) 14-fold range in species sensitivity appears to be in general agreement w'ith the approximately 11fold range which can be crudely estimated by bacK-calculat ing from the guideline value of 3.9. The calculated log intercepts for invertebrates, the ranges of which was given as a measure of relative species sensitivity, shows approximately a 500-fold difference which is in general agreement with the species sensitivity factor of 21, from which back-calculations suggest a typical 260-fold difference from most to least sensitive. Chronic Toxicity (S-6 to B-S). Davies, et al. (1976) give additional chronic data, men tioned on p.A-1, but not reported in this section in Table 7. CMA 049677 Lead Using the guidelines sensitivity value for freshwater fish chronic studies of 6.7 results in a much greater reduction in' computing the Final Fish Chronic Value than would seem to be warranted by the 2.2-fold range after adjustment for hardness in relative species sensitivities in Table 3, especially since the rainbow trout value (the most sensitive species in Table 1) Is right in the middle of the available fish chronic values. If the application factor were invoked, the estimated final fish chronic value would have been e(1.51. In (hardness)-1.40) , which is very close to the value before adjusting by 6.7, namely e (1.51. In (hardness)-1.47). The above reasoning seems to be in line,with the document's rationale for not using the species sensitivity value for the final invertebrate chronic value (p. B-7). (It is noteworthy, however, that data in Table 7, suggest that adverse effects occur at concentrations very lit tle higher than those given by the final fish chronic value. Those data should be discussed here or under Miscellaneous.) Table 1. The first entry (rainbow trout in hardwater) combines the results of two tests at two different hardnesses (555 mg/1 is neither the arithemetic nor the geometric mean) with the arith metic mean of the two LC53 values given; moreover, measured lead concentrations were published for both of these tests. The tests should be shown separately as were the multiple soft water tests on fathead minnow by Pickering and Henderson (1966). Saltwater (pp. 3-21 to 3-23) Criterion Formulation - The RLTC and Final Chronic Value entries presumably should be deleted (p.B-24), ',\'e cannot locate the source of the 3.5 pg/1 values. CMA. 049678 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR Lead by John Doull, M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: The ambient water quality criterion recommended in his document for lead (50 ug/1) appears to be appropriate on the basis of the animal and human toxicity data. CMA. 049679 GENERAL COMMENTS : The human health criterion recommendation in this report for lead (50 ug/1) is based primarily on the EPA air quality criteria for lead document (which is listed four times in the reference), and on the WHO document (Environmental Health Criteria 3, lead). Other recent important documents dealing with the setting of standards for lead have been reviewed, including the NAS Safe Drinking water report and the NAS report on Airborne lead in perspective (which is included twice in the reference list). The proposed recommendation is consistent with the recommendations for lead standards proposed by other agencies and organizations, although, the NAS Safe Drinking Water Committee suggested that consideration be given to lowering the limit if additional research demonstrates that the current levels do not provide a safe limit for fetuses and growing children. ! The toxicology of lead in man and animals has been extensively reviewed in a number of recent documents, criteria reviews, committee reports and books,, and this document appears to have covered essentially the same ground and to have reached similar conclusions about the human health hazards associated with lead in our environment. Comparison of the ambient water quality criteria document for lead with the air quality criteria document for lead, suggests that the criterion formulation section of the water report could have been appended to the air quality document or even to one of the many NAS reports on lead without losing any of the important toxicologic studies or missing any of the current controversies involving environmental lead. In the case of this metal, at least one has the impression that we are spending more time evaluating the toxicologic data available for lead than we are in generating new data which is needed to resolve some of the remaining problems associated with lead in our food, air, and water. Other than the fact that it generates a distinct feeling of deja vu in this reviewer, there is nothing wrong with this document or its conclusions regarding the human health criterion for lead in ambient water. CMA. 049680 Comments on the Aquatic Toxicology Section of the EPA Ambient Water Quality Criteria Document (March 15, 1979) on Nitrosamines Summary So criteria for the protection of aquatic life were derived'for N-nitrosodiphenylamine under the proposed guidelines since there are insuf ficient data in both freshwater and saltwater. N-nitrosodiphenylamine is an appropriate test compound in this class. Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and James L. Peterson,, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19105 Mav 1979 049681 Proposed Criteria - Aquatic Life N-nitrosadiphenylamine For freshwater aquatic life, no criterion for N-nitro- sodiphenylamine can be derived using the Guidelines, and there are insufficient data to estimate a criterion using other procedures. For saltwater aquatic life, no criterion for N-nitrosodiphenylamine can be derived using the Guidelines, and there are insufficient data to estimate a criterion using other procedures. Our Conclusions No criteria for the protection of freshwater aquatic life were formulated since a final chronic value for N-nitrosodiphenylamine could not be derived under the guidelines. If the final chronic value were to be established in the vicinity suggested by available data, it would lie within the ambient range for total nitrosamines of natural origin. The selection of N-nitrosodiphenylamine for tests of toxicity to aquatic organisms is reasonable in view of the statement that this is the most prevalent anthropogenic nitrosamine. There are inadequate data to derive a criterion for the protection of saltwater aquatic life under the proposed guide lines . CMA. 049682 General Comments Nitrosamines The primary concerns with nitrosamines (and therefore available data) seem clearly to be in terms of human health rather than the immediate protection of aquatic life. Moreover, the evi dence presented suggests that nitrosamines of natural origin are likely to predominate over anthropogenic nitrosamines in ambient water. The introduction (p.A-2) does mention very low concentrations of nitrosamines ("less than 0.1 ug/1") in industrial pipe sources. Hence, it would be desirable if data can be developed to establish at least the freshwater criterion values. The tentative upper bound to a Final Chronic Value in freshwater is based on very recent testing the report of which we have not seen. '* The current separation of this group into individual compounds for criteria consideration is far more desirable than lumping them together as was previously done. Specific Comments Introduction - The use of N'-nitrosodiphcnylamine for acute tests of aquatic toxicity seems reasonable in view of the state ment (p.A-2) that this is the most prevalent anthropogenic nitrosaminc. Chronic Toxicity Apparently the second clause of the first sentence should read, "but adverse effects were observed at every test concen tration, even the lowest at 4S ug/1(Table 3)." The suggested Final Invertebrate Chronic Value of "less than 9.4 ug/1" is certainly in the range reported for ambient stream measurements, "less than 10 ug/1" (p.A-2). If an MATC value is eventually established which yields a definitive Final Chronic Value in this range, there may be some question as to what a regulatory agency could accomplish on the basis of such a criterion value, given that most exposure to nitrosamines is not anthropogenic in origin (p.A-2). Mi seel 1ancous Ferraro, ct al. (1977) cite histopathology work on a neuro tic crayfish reported by Harschbargcr et al. (1971); perhaps this should be considered for possible citation. CMA. 04 9683 Nitrosamincs Additional Reference Harshbarger, J.C., et al. 1971. Effects of N-nitrosodimethylamine on the crayfish, Procambarus clarkii. In Proceedings of the Fourth International Colloquium on Insect Pathology with the Society for Invertebrate Pathology, College Park, Maryland, August 25-28, 1970, pp. 425-430. CMA. 049684 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR Nitrosamines by John Dou11, M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION; Despite some reservations on the rationale for transferring data calculated for food intake to the ambient water quality situation, the use of the risk extrapolation approach to setting criteria for nitrosamines appears to be justified. CMA 049685 \- GENERAL COMMENTS; The decision of the authors of this document to include both the nitro- samides and the various nitrosamines in a single criterion document is appro priate in view of the present state of development of the toxicity data for these agents. It is probable, in fact, that the final decision on the human health criterion for this group of agents will also need to consider nitrites, nitrates and amines in the ambient water supply in order to establish satis factory criteria recommendations. The data on page C-7 concerning the short ha 1 f-1 ife*-*of nitrosamines is reassuring as is the intake data for food versus water, etc. In this respect the statements on C-9 concerning percentage intake should be referenced. It would also be helpful if the authors would expand the statements regarding the analytical difficulties in detecting nitrosamines and nitrosamides (page C-10). Most of the data in this report relate to the nitrosamine content of foods, and there is some question as to the transferability of such data to a water situation. The use of the risk-extrapolation procedure to set ambient water quality criteria on this type of data needs to be evaluated carefully since it appears from this document that there are important chemical and physica1-chemica1 considerations for which definitive data is lacking. CMfc 049686 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERION ON PENTACHLOROPHENOL Conclusion Chronic values for freshwater species need to be developed. At this time the only valid number is the final acute number. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical Company Midland, MI 48640 June 1, 1979 CMA 049687 CRITERION FORMULATION B-8 Freshwater The only acceptable value for criteria formulation is the final fish acute value of 25 yg/L. The invertebrate value is based on three experiments with the' Tubificid worm at three different p;-;' s and for a 24 hour period. To extrapolate this to 9 6 hours is unreasonable, further to assume that the three values repre sent three different species for purposes of the guidelines is also unreasonable. Until further work has been done the only value that should be used is the fish acute. Since chronic studies have not been performed, it is impossible to derive a final chronic value. The value based on Chlorella , pyrenoidosa must be reserved until a decision can be made on both the importance and the sensitivity of this species to reflect the aquatic plant world. Saltwater The value of 25 yg/L appears to be reasonable for a final fish acute value. Since we do not accept the data base for the safety factor of 49 in converting invertebrate acute to a value for protecting 95% of the species, we cannot accept 8.5 yg/L as a reasonable acute value. In this connection it is interesting io note that the final fish value of 25 yg/L is more than adequate to represent the acute toxicity for the invertebrate species studied in Table 7 (p. B-27). criterion for oentachloroohanol to orotect saltwater n-u:.tic should be 9.6 yg/L as a 24 hour aver if- nr4- '*'1 .J V ^ CMA 049688 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR Pen tach 1oropheno1 by John Dou11, M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: The ambient water quality criterion proposed in this document (1A0 ug/1) is appropriate for PCP itself but is probably too high for dioxin containing PCP. Since a separate criterion is being proposed for dioxin, it is probably acceptable to establish a criterion for "pure" PCP, and the value recommended in this document is reasonable. CMA 049689 GENERAL COMMENTS: The human health criterion for pentachlorophenol recommended in this document is 1^0 wg/1, which is roughly 7 times the no-adverse-effect level suggested by the Safe Drinking Water Committee in 1977. The difference is primarily due to the use of a safety factor of 100 in this document in contrast to the use of a safety factor of 1000 in the NAS document. Although it is stated in this document that the data of Schwetz et al. and that of Innes et al. were apparently not available or not considered by the NAS committee (which is incorrect in both assumptions), the major reason for the use of the large uncertainty factor by the NAS committee was the concern about the presence of dioxin as a contaminant in commercial samples of pentachlorophenol. The NAS committee was justifiably much more concerned about the chronic toxic effects of dioxin as a contaminant in PCP than the possible chronic toxic effects of PCP itself and reflected this concern in their choice of an un certainty factor for this agent. There is almost no data in this document to support the argument that the human health criteria for PCP should be established on the toxicity of the pure chemical PCP rather than on the toxicity of PCP as it is likely to be found in the "real world". This would be appro priate if an ambient water quality criterion were also being established for the various dioxins which occur as contaminants of commercial PCP samples. CMA. 049690 TAABB1I.K VI-5555.. Summary of Chronic Toxicity Data for Pentachlorophenol "1-299 Species Duration Of Study Dosage Level:-, and Ho. of Animals per Groun Highest No-ndverse effect Level or Lowest Minimal-effeet Level affect Measured Reference Rat 1? weeks 0-200 ppm in diet, 25 ppm (1.25 mg/kg/day]| Ho toxic Knudson et al., DO animals per effect 3 VV'< group Hamster days 5-10 1.25-20 mg/kg/day orally 2.5 mg/kg/day Rat 90 days 0-30 mg/kg/day in 3 mg/kc/dayC dieta Ho fctotoxic or em'oryotoxie eff ct Hinkle, 1973 No toxic effect Johnson et al., 1973 Rat 90 days O-RO^ng/kd/day in 3 mg/kg/day diet Increased Johnson et al., organ weii ht e, 1973 scrum enzyme ehang.es Rat days 6-15 0-50 mg/kg/day, 5 mg/lg/day ornllv, 20-20 a animals per group Teratogenic effects Schwct2 et al., 197't Rat 6 months 0-50 ppm in diet, 100 ppm 20 animals per (5.0 mg/kg/day) eroup Rat 8 months 0-500 ppm in diet* 500 ppm (25 mg/kg/day) No toxic effect Mo toxic effect Kimbrough and Linder, 1975 Goldstein et al. 1976 Using an uncertainty factor of 1,000, suggested no-odvcrse-effcct level in drinking water is calculated as follows: &pure pep 3 0.003 mg/ke/day (ADI), 0.003 x 70c x 0.1d 0.021 mg/liter. . 1,000 technical pep Q Assume average weight of human adult 70 kg. ^Assume average dally Intake of water for man is from water. 2 liters, and that 20,i of total intake eTost study from which to calculate suggested no-advcrse-cffcct level. IIIW U^W ^JI ^'{J5 a it* f. * - _ * * ', *' 4 4it* If* i d! ! :V / CMA 049691 IP* Comments on the Aquatic Toxicology Section of the EPA Ambient Water Quality Criteria Document (March 15, 1979) on Selenium Summary The freshwater criterion is derived under one of the proposed alternative procedures, using un published data to permit criterion formulation. The freshwater criterion appears to be fully pro tective of all tested species under laboratory conditions. The saltwater criteria is properly derived under the proposed guidelines, but is numerically based on a single acute study of Mysidopsis bz'nia, a mysid shrimp. The saltwater criterion is fully protective of all tested species. Prepared by John A. Hendrickson, Jr., Ph.D. Associate Curator Division of Limnology and Ecology and James L. Peterson, Ph.D. Director Division of Limnology and Ecology Academy of Natural Sciences Nineteenth and the Parkway Philadelphia, PA 19105 May 1979 CMA 049692 Proposed Criteria * Aquatic Life For selenium the criterion to protect freshwater aquatic life, as derived using procedures other than the Guidelines, is 9.7 yg/1 (derived as 0,44 times final invertebrate acute value) as a 24-hour average and the concentration should never exceed 22 yg/1 (the final invertebrate acute value) at any time. For selenium the criterion to protect saltwater aquatic life as derived using the Guidelines is 4.4 yg/1 (derived as 0.44 times the final invertebrate acute value) as a 24-hour average and the concentration should not exceed 10 yg/1 (the final invertebrate acute value) at any time. Our Conclusions The freshwater criterion was derived according to the first of the published proposed alternative procedures since a fish or invertebrate chronic value could not be derived under the proposed guidelines. These chronic values did not direct ly affect the criterion values, but other unpublished data was used to modify the final invertebrate value, raising this con centration significantly. The most important determinants of the criterion components are the species sensitivity which cannot be directly evaluated in this case. The freshwater criterion appears to be fully protective of all tested species under laboratory conditions, but does not address ambient conditions. The saltwater criterion is properly derived according to the proposed guidelines, but is numerically based on a single study of A'ysidopsis bahia, a mysid shrimp, and the species sensitivity factor for invertebrate acute studies. The cri terion appears to be fully protective of all tested species; further evaluation is not currently possible. CMA. 049693 General Comments- Selenium The available data indicates that selenium can be acutely toxic to aquatic invertebrates and fish. The criterion document does not indicate the levels of selenium concentrations encountered in ambient waters; the discussion of health effects permits the inference that selenium concentrations in freshwaters are high ly variable. An old value (which may not be reliable in view of analytic improvements) gives the average concentration in seawater as 0.004 mg/1, or 4 ug/1 (Sverdrup, et al. (1942)). The introduction (p.A-1) indicates that the major sources of environmental selenium are natural rather than anthropogenic. The health effects data suggest that dietary selenium is essen tially in low concentrations for most species (p.C-46) and that there are significant interregional differences in ambient selenium concentrations (p.C-47); this could suggest that native or naturalized populations of aquatic organisms could be adapted to different ambient concentrations of selenium. Those points suggest that it might be unreasonable to establish a nation wide selenium criterion for the protection of aquatic life if such criterion values were to be in the range ordinarily encoun tered by aquatic life in the absence of anthropogenic sources. A similar pattern of reasoning would suggest that the proposed 24-hour average value for selenium in seawater is very possibly, within the range ordinarily encountered by marine organisms, and that marine organisms in some coastal regions may be adapted to higher concentrations of selenium than the proposed 24-hour average concentration. The aquatic life criteria here not developed using bioconcentra tion data, presumably because it is difficult to establish an appropriate limit for selenium concentrations in food. (pp.C48, 49). Specific Comments Freshwater Acute toxicitv On p.B-2, the text refers to "two test results with Parhnic. magna," but Table 2 indicates that one of these is with Hyallcla aztsaa instead; therefore, the comments on Guideline adjustment factors should not be inserted at this point. CMA. 049694 Selenium Criterion Formulation The first attempt failed for lack of a suitable fish or or invertebrate chronic value; when these were supplied from very new data, a second attempt was made. There is one computational error in the second attempt, namely, that the geometric mean of the two available fathead minnow (juvenile] flow-through tests with measured concentrations should have been used instead of the single lowest value; since this geo metric mean does not indicate different sensitivity for fry and juveniles, the geometric mean for fry and juveniles should be taken, and this yields a Final Fish acute Value of 2100 ug/1. The correct procedure was followed for recomputing the Final Invertebrate Acute Value, and these two Dapknia negne tests do not clearly support the Guideline adjustment factors. Although the Final Fish Chronic Value is less than 1' of the Final Fish Acute Value, the reduction does not seem unrea sonable in view of the long-term studies presented in Table 5. Saltwater All the available marine data are from studies just complet ed for EPA. Tests have been completed on only one fish, one invertebrate and one algal species. Acute Toxicitv The Guideline sensitivity factor of 49 for saltwater invertebrates results in a Final Invertebrate Acute Value much lower than the equivalent calculation, with a sensitivity factor of 5.1, for the same invertebrate, in arriving at the Final Invertebrate Chronic Value. The saltwater criterion values are both based on the Final Invertebrate Acute Value. This sensi tivity factor is subject to review as more sets of data become available. Criterion Formulation Page B-14 was omitted in our copy of the criterion document, but the information should be identical to that presented in the Federal Register (44:15962, 3rd column). The 24-hour average concentration of 4.4 ug/1 is uncomfortably close to the reported average of 4 yg/1 in seawater (Sverdrup, et al., 1942). This CMA 049695 Selenium suggests that the sensitivity factor may be unnecessarily large in this case. Additional References Sverdrup, H.U., et al. 1942. The Oceans. Prentice-Hall. CMA. 049696 EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR Selenium Conclusions: The health based criterion selected for selenium is appropriate. The significant toxicological literature on selenium is adequately reviewed. r Prepared by: Roj.f Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 10, 1979 CMA. 049697 General Comments: The proposed criterion for the protection of human health is approximately 10 yg selenium per liter, which is similar to the recommendation made by the NAS Committee on Drinking Water and Health and is identical with the present EPA drinking water standard. The document is, in general, very well written and takes account of most of the known literature on selenium. The document discusses the role of selenium as a required trace element, its toxicity at elevated levels, and the known information on the possible carcinogenicity of selenium. For selenium, the experimental data on cancer are equivocal. There clearly are data that argue for both sides of the question of carcinogenicity or non-carcinogenicity. It is interesting to note that in this case EPA chooses the position to call selenium non-carcinogenic and does not employ the carcino genicity extrapolation model for selenium, while for other chemicals it takes any indication for carcinogenicity as a sufficient basis to label a chemical a carcinogen. If such a position were chosen for selenium, then the resulting criterion for selenium would be less than the con centrations required to prevent deficiencies. Detailed Comments: Page A-2. The statement regarding the FDA position on selenium appears to be out of date. In fact, as indicated on page C-42, the FDA has ruled that sodium selenite may be used as a feed additive with certain restrictions. q4969 CM* Page c-1. The units utilized to express the selenium contents in foods differ significantly between different food commodities. The concentration of selenium in eggs is quoted incorrectly from the NAS document; the concentration should be 10 ug per gram rather than 10 ug per liter. Page C-25. The dilutions cited for the various tests done in tissue culture are meaningless because no units are given. CMA, 049699 jiill 7' TO Comments on the EPA Water Quality Criteria For Silver For the Protection of Aquatic Life (FR AA, pg- 1596A, March 15, 1979) Prepared by: Haines B. Lockhart, Jr., Ph.D. Senior Chemist Health, Safety and Human Factors Laboratory Eastman Kodak Company Conclusions May 1A, 1979 On the basis of the data and discussion given in the Silver Criteria Document (PB 292 AA1), the criteria to protect freshwater aquatic life of 0.000009 parts per million total silver as a 2A-hour average, and 0.0019 parts per million total silver as a maximum, cannot be justified. The 2A-hour number is based on one egg-fry sub-chronic fish bioassay in which water of a hardness of only 26 mg/l, as CaCO^, was used, and the test chemical was silver nitrate. The aquatic toxicity of this soluble silver salt is unrepresentative of the toxicities of silver species actually present in most silver-bearing wastes discharged to the en vironment, and the dilution water used is unrepresentative of most natural waters. CMA 049700 General Consents It is disturbing to read in this criteria document the way in which the EPA has repeatedly ignored considerations of silver toxicity as a function of the environmental chemistry of silver. The EPA's treatment of all silver species which could be present in the environment, as if these species were soluble, uncoraplexed (free) silver ion, disregards volumes of published research on the inorganic chemistry of silver. The cri terion document avoids the important facts in the scientific literature (Terhaar et al, 1972, 1977; Bard e_t al, 1976; and Cooper and Jolly, 1970) that point to the effects of silver on aquatic organisms as being a function of the concentration of free silver ion, not a function of , the concentration of total silver. CMA 049701 The method 'used by the EPA to develop the silver water quality criteria for the protection of aquatic life, described in the May 18, 1979 Federal Register, is long on statistics, short on biology, and lacking in environ mental chemistry. The method used to calculate water quality criteria concludes that the analytical chemistry, physico-chemical properties, and aquatic effects, as functions of biological species exposed and chemical species of silver present, can be determined with "statistical validity" from the results of previous studies on a random set of un related chemical compounds. Such conclusions by the writers of the EPA method are incorrect. The criteria numbers for total silver should be subjected to a very important test: reasonableness. One good test for reasonableness would be: are the numbers put forth consistent with what is found in natural unpolluted waters where the aquatic species to be protected are known to flourish? The numbers proposed by the EPA for total silver criteria fail this simple test of reasonableness. Significant data exist in EPA's own STORET data base which show that concentrations of total silver more than two orders of magnitude greater than the proposed 0.000009 parts per million criterion value are present in natural waters, with no measurable adverse effects on aquatic life. In fact, many of these same natural waters are recommended by the respective local wild life and conservation groups as being excellent fishing areas. However, a strict interpretation of the EPA criteria for total silver would be that nature has erred by placing harmful concentrations of total silver in natural waters. CMA 049702 -3- To be scientifically valid, the criteria for silver must be written in terms of free silver ion, not total silver. The method for calculation of water quality criteria, as set forth in the May 18, 1978 Federal Register, is inappropriate to use for the establishment of criteria on silver. Specific Comments (Page numbers refer to the criteria document) A- 1 Only those silver salts which readily dissolve to give appreciable free, soluble silver ion concentrations, such as silver nitrate and silver sulfate, are toxic to a large number of microorganisms. Silver compounds of environmental significance, such as silver sulfide and silver thiosulfate complex, have been shown by Tcrhaar et al, (1972, 1977) and Bard et al (1976) to be without significant toxicity to aquatic life. A- 1 Free silver ion, not "silver salts", combine with sulfhydryl groups on biological molecules. Free silver ion, not "many silver salts", can be absorbed into the human circulatory system resulting in argeria. Such effects caused by silver ions are discussed by Cooper and Jolly (1970). A- 2 Effects on aquatic organisms have been demonstrated by free silver ion at concentrations > lppm. Effects have not been shown by silver species in which the silver ion is tightly completed (see Bard, Terhaar, above). CMk 049703 A- 2 The reference by the EPA to Kopp's finding that silver concentrations in well and surface waters ranged from 0.1 to 38 ppb, with a median of 2.6 ppb, is interesting in light of EPA's 24-hour average criterion value in fresh water of 0.009 ppb. A reasonableness test should be applied here. B- 1 The statement that " there are no data available on the toxicity of different forms of silver." points up the inadequacy of the literature search on silver effects on aquatic organisms. Data on the effects of silver species are found in such journals as the Journal of the Water Pollution Control Federation (Bard et_ al^, 1976) and Water Research (Terhaar e a_l, 1977). B- 1 The acute and chronic effect data on fish, specifically rainbow trout, are from Davies et al (1978). It should be noted that the chloride content of the water used in the Davies studies represents about 5% of the water in the U.S. supporting a good fish fauna (McKee and Wolf, 1963); 95% of all such waters have > 170 mg/1 of chloride ion present. Thus, the criterion value of 0.009 ppb may be valid only to protect rainbow trout in Colorado-like mountain streams from the adverse effects of silver nitrate. B- 3 The use of the lowest plant toxicity value instead of a geometric mean of all data when calculating the final plant value seems statistically inconsistent with the manner in which the other values were calculated. CMA. 049704 -5- B- 5-10 Tables 1-6 contain some studies performed with silver species other than silver nitrate (free silver ion). All of these chemical species should not be lumped together as "silver" when reporting out effects. B- 6 Table 2. Only the extremes of the data are reported here, showing that there is little correlation of toxicity between two invertebrate species. B- 7 Table 3. Davies _et al (1978) reported the "no-effect" limit as 0.09 yg/1 based on adjusted control mortality. However, actual control mortalities were never reported. If the control mortalities exceeded 10%, by the EPA's own definition, this test would be considered invalid. B-10 Table 6. Data from this table were evidently not used for the criterion calculation. Terhaar _e al (1972) is quoted incorrectly; the 96-hour mortality to fathead minnows using silver thiosulfate at 250 mg/1, as Ag, is 25%, not 75%. It should be noted in this table that Coleman and Cearley (1974) exposed bluegills to 70 ug/1 silver as AgNO^ for six months, with no effects. B-13 It is an indication of a problem with the method when the Final Acute Value is less, 0.58 pg/1, (based on invertebrate data) than 049705 the Final Chronic value, 3.6 ug/1 (also based on invertebrate data). A reasonableness test should be applied here. CMA 049706 References Bard, C. C., Murphy, J. J., Stone, D. L. and Terhaar, C. J., "Silver in Photoprocessing Effluents", J. Water Pollution Control Federation 48. 389-394 (1976). Coleman, R. L. and Cearley, J. E., "Silver Toxicity and Accumulation in Largemouth Bass and Bluegill", Bulletin of Environmental Contamination and Toxicology _12, 53-61 (1974). Cooper, C. F. and Jolly, W. C., "Ecological Effects of Silver Iodide and Other Weather Modification Agents: A Review", Water Resources Research 6, 88-98 (1970). Davies, P. H., Coettl, J. P. Jr., and Sinley, J. R., "Toxicity of Silver to Rainbow Trout", Water Research JL2, 113-117 (1978). Kopp, J. F. "The Occurrence of Trace Elements in Water", 3rd Annual Conference on Trace Substances in Environmental Health. D. Hemphill, ed., V. of Missouri, Columbia, Mo. McKee, J. E. and Wolf, H. W., "Water Quality", California State Water Resources control Board, Pub. 3-A (1963). Terhaar, C. J., Ewell, W. S., Dziuba, S. P. and Fassett, D. W., "Toxicity of Photographic Processing Chemicals to Fish", Photographic Science and Engineering 16, 370-377 (1972). CMA 049707 Terhaar, C 1 > Ewell, V* S *, 0 2 iuba , S. P, White , W * W., end Murphy , P. J., "A Laboratory Model for Evaluating the Behavior of Heavy Metals in an Aquatic Environment", Water Research H.. 101-110 (1977). I CMA. 049708 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR Silver by John Doul1, M.D., Ph,D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: The proposed ambient water quality criterion for silver appears to be justified on the basis of the animal and human toxicity data presented in this document. The use of a safety factor between 10 and 100 is rather unusual, but the proposed criterion (10 ug/l) seems to be a reasonable criterion for this meta1. CMA. 049709 GENERAL COMMENTS: This document provides a great deal more information than is needed to establish a human health criterion for silver in ambient water supplies. The first 20 pages of the human health section, for example, could be appropriately summarized in a few paragraphs with short tables listing the levels of silver in foods, human tissues, and geographical locations. Most of the rest of the human health section of this report (160 pages) is also unnecessarily redundant, repetitive and full of extraneous material. The toxicity data, which is on page 68, demonstrates that both the acute and chronic toxicity of silver has been extensively investigated in a variety of species, including man. The synergism/antagonism section appears to be more focused on selenium than on silver, although the silver-selenium inter action is well characterized. The mutagenesis and carcinogenesis sections provide good research reviews of the work done with silver but contribute little to the criterion-setting requirements for toxicity data. In the criterion formulation section of this report, considerable emphasis is given to the derivation of the existing guidelines and standards for silver in drinking water and work place air. The dependence of these criteria on argyria is described and the problems associated with the use of a cosmetic effect rather than a toxic effect as the basis for criterion-setting is discussed. Evaluation of the reported toxic effects in rats chronically exposed to silver in their drinking water leads to the conclusion that the NOEL for chronic exposure to silver would be about 1*00 yg/1 and, using a safety factor of 10 to 100, the criterion level was calculated to be 10 yg/1 (1.6 to 16 yg/1). The recommen dation of the NAS Safe Drinking Water Committee for silver was to give consider ation to taking silver off of the list of substances to be included in primary drinking water standards in view of the low natural concentrations of this agent in public water supplies. CM* 049710 AN EVALUATION OF THE PROPOSED AQUATIC LIFE CRITERIA ON TETRACHLOROETHYLENE sion ta base cn fish toxicity is rr.ore credible than ths data hr.la magna, hence it should be used for criteria setting, the guidelines, a value of 2600 ppb should not be exceeded 4 hour.period. A final chronic value cannot be derived present time. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical Company Midland, MI 48640 June 1, 1970 CMA 049711 GENERAL COMMENTS A partition coefficient of 339 indicates that tetrachloroethvlene has a modest affinity for lipids. I believe that the statement chat FEE has a "high affinity" should be changed (p A-l). The data base for freshwater organisms p B-l indicates that the irverteorate data on one species is unpublished and should net ut used for criteria development. The more credible data is the toxicity on fish. Since there is no chronic data, then no chronic criteria can be established at this time. The use of the alternate guidelines as suggested on p B-4 is not acceptable. The data base on saltwater organisms (B-ll) is also very weak. To use one value for nysid shrimp using 96 hour unmeasured, static conditions is inconsistent with the guidelines. Since this value was determined in 1978 under an EPA contract, the experimental procedure should have called for a flow-through protocol where the concentration was measured. 497i2 EVALUATION OF or THE health effects THE CRITERIA DOCUMENT AMBIENT WATER FOR section FOR Tetrachloroethvlene Conclusions: The criteria document contains some misinterpretations, and the methodology used for the derivation of the health based criterion is probably inappropriate for tetrachloroethylene. The significant toxicological literature is adequately reviewed. Prepared by: Rolf Ilartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Ivrbor, MI 4 81U9 May 10, 1979 CMA 049713 General Comments: In general, the document covers the literature on tetra- chloroethylene (PCE) well. However, the criteria document contains some questionable interpretations, and the method for deriving the criteria is probably inappropriate as noted under the specific comments. Specific Comments: Page C-3. Some data on concentrations found in dry-cleaning establishments and other occupational settings should be cited here. The present TLV for PCE is 670 mg/m"*. Page C-ll. The significance of electrical impedance measurements on the cerebral cortex is obscure to me. Was this information gleaned from an abstract of the Russian paper or from a full translation? Pace C-14. It is very questionable that PCE is "closely related to vinyl chloride"; there are significant differences in physical, chemical and toxicological properties between these two substances. Page C-15. The last full sentence on this page is purely speculative, and appears to reveal a bias on the part of the original author, especially when read in context with the last paragraph on Page C-14. Page C-16. Mixtures whose effects are less than additive are defined as being antagonistic by the joint action model. In order for potentiation to occur the components of the mixture must cause greater than additive effects according to a joint CMA 049714 action model. These mixtures can be called synergistic only according to an independent action model which expects that the damage produced by one chemical has no influence whatever on the effects produced by the second chemical in the mixture. Since non-specific interactions usually occur, the independent action model scores the interactions of most chemicals as synergistic. Page C-22. The application of the Stockinger-Woodward model to derive a "safe level" for ingested PCE in humans from a rat inhalation study is probably inappropriate. This is especially true for an effect such as "altered electrical impedance of the cerebral cortex" whose significance is questionable. Pages C-23 to C-24. If one calculates the amount of PCE inhaled at the TLV, assuming the inhalation of 10m^ per 8-hour day, a ratio of absorption by ingestion vs inhalation of 0.57, and adjusts from a 5-day work week to a full week, then the estimated daily intake per 70 kg person would be 670 mg/m^ x 10m^ x 0.57 x y = 2728 mg/day. At an extrapolated risk of 10 V the daily exposure according to the criteria document would be 2.0 ug/1 (21 + 110(BCF) x 0.0187) = 8.11 ug/day. This would result in a risk/dose of 10~V8.11 = 1.23 x 10-V This can be applied to estimate the risk from exposure at the TLV using the single-hit model. P = 1 - exp (-2,728,000 x 1.23 x 10-6) = 0.9653 The single-hit model predicts the occurrence of cancer in . virtually all people exposed to PCE at the equivalent to a TLV exposure for a life-time. CMA 049715 However, PCE has been used for a long time, and many people have been exposed to it without any obvious exposure-related cancers being noticed. There is an obvious discrepancy between the animal model, the extrapolation model and the human experience! The conclusion must be that the use of the EPA approach to developing water quality criteria for compounds that have been found carcinogenic in animal testing is inappro priate for PCE. CMA. 049716 AN EVALUATION OK THE PROPONED AQUATIC LIFE CRITERION ON THALLIUM The recommendation that no criterion for thallium can be derived for freshwater or saltwater aquatic life is justified from the scarcity of data in the literature. Chronic toxicity data for freshwater fish were presented from a U.S. EPA study that had lim ited peer review and circulation and was not clearly referenced in the text. No recommendation is justified at the present time from the additional studies found and added to this document. Prepared by Donald S. Cherry 204 Maywood Street Blacksburg, Virginia 24060 and John Cairns, Jr. 501 Bishop Road Blacksburg, Virginia 24060 May 1, 1970 GENERAL COMMENTS The recommendation that no criterion for thallium can be derived for freshwater aquatic life is justified. There are insufficient data to estimate a criterion, and only a few general references on thallium studies have been cited in the criterion document. Specific Comments p A-l. 1) The Kith line, "o trivalent" should road "o_r trivnlent". p A-3. 1) In the 2nd sentence, the authority or reference that reported chronic toxicity to fisli at thallium concentrations as low as 20 ug/l and to bioconcentrate by a factor of up to KiO is missing. It is assumed to be U.S. EPA (1978) as referenced in Table 3 (p U-7). If so, it is difficult to evaluate the methodology and potential mor tality between controls and treated groups because of the limited distribution of this study. 2) Additional studies that may be incorporated into the introduction section include Mathis and Kevern 1975 t"Distribution of mercury, cadmium, lead and thallium in a eutrophic lake"), Eisler and LaRoche 1972 ("Elemental composition of the estuarine teleost Fundulus heteroclitus L."), Norris et al. 1976 ("Toxicity and accumu lation of thallium in bacteria and yeast"), and Nehring 19C2 ("Experi ments of the toxicological effect of thallium ions on fish and fishfood organisms"). p A-5. 1) In the last reference, Zitko 1975 is cited twice on p A-l, but not as Zitko et al. 1975. Which of these references is Zitko et al.? CMA. 049718 2- p B-l. 1) In the 2nd sentence, again, what is the reference to chronic toxicity of thallium to fish at 20 pg/A? The same question applies in the next sentence at 100 pg/A for algae. p B-4. 1) The summary of available data for freshwater aquatic life is basically represented from the U.S. Environmental Protec tion Agency (1978) which has limited peer review and circulation. A similar trend is observed on p B-13 for saltwater aquatic life. Since this unpublished EPA document is not a primary reference, it is recommended that the methodology be made available for review before evaluation of the data is completed. V- CMA 049719 -3CRITERION FORMULATION The guidelines for establishing a final acute fish toxicity number may have been erroneously calculated. The geometric mean of adjusted LC50 values of 2,000 ug/t should be readdressed to see if 16,012 pg/i is correct. No formulation of criteria for freshwater or saltwater aquatic life should be attempted for thallium at this time until sufficient data are generated in the literature. v- CMA. 049720 -4MISSING AND ADDITIONAL REFERENCES 1. Eisler, R. and G. LaRoche. 1972. Elemental composition of the estuarine teleost Fundu I us heteroc-1 i tus (L.). .J. Experi mental Marine Biol, and Ecol. 9:29-42. 2. Mathis, B. J. and N. R. Kevern. 1970. Distribution of mer cury, cadmium, lead and thallium in a cm trophic lake. Hydrobiologia 40:207-222. 3. Nehring, D. 1902. Experiments on the toxicological effect of thallium ions on fish and fish-food organisms. Z. Fisch. 11:557-562. 4. Norris, P. , K. M. Wing, M. N. Hughes and D. P. Kelly. 1976. Toxicity and accumulation of thallium in bacteria and yeast. Arch. Microbiol. 110:279-286. CMA. 049721 COMMENTS ON THE PROPOSED HUMAN HEALTH CRITERION FOR ThaIlium by John DouI 1, M.D., Ph.D. University of Kansas Medical Center Kansas City, Kansas 66103 CONCLUSION: Although the safety factor used to calculate the ambient water quality criterion for thallium seems to be higher than justifiable, I would have no serious objection to the final criterion proposed in this document ug/1). CMA. 049722 GENERAL COMMENTS: In following the arguments, assumptions and calculations used to calculate the human health criterion for thallium in ambient water supplies, it is evident that the recommended level of A ug/1 provides a generous margin of safety. The rat subchronic study (Downs et at.) isreported to demonstrate a NOEL of less than 12 ppm in the diet (calculated to be less than 1.6 mg/kg/day). Reducing the dose by 1/3 results in no symptoms or adverse effects, which suggests a NOEL of about I mg/kg. Using the NAS Safe Drinking Water Committee approach for data of this type (no 2-year chronic, scanty subchronic, no acute human data) with a safety factor of 1000, the calculationwould be as follows: jQpp = 1.0 ug/kg/day (AD I) X 70 X 0.1 = 7 ug 'iter In the case of thallium, this safety factor seems large since there is rather good data on the acute lethal dose for thallium in children and adults, and the toxic effects of thallium arc similar following single or fractionated doses. ' The human health criterion calculated in this document includes a further re duction (additional safety factor) to take care of the differences in intake required by the rat and man to attain a Steady state condition. In the NAS SDW approach, such additional factors are not included since they are assumed to be more than adequately provided for by the use of the 3-fold safety factor approach. Thus, the human health criterion calculated in this document is lower than the HAS SDW toxicologists would have recommended. However, the difference is not large and there are some arguments for recommended levels in this document, which arc not directly related to the safety factor approach. In addition, the data in this report indicate that this level will be reasonable in terms of detection methodology and the "real world" situation. This report demonstrates one of the current concerns of many toxicologists in the use of safety factors by regulatory agencies to establish hazard standards. When safety factors are applied sequentially or "pyramided," the CMA 049723 no: result is often a gross exaggeration of the level of protection. For agents with a steep dose-response relationship, this approach is almost invariably inappropriate, both from a toxicologic viewpoint and for maximizing benefits and minimizing the risks associated with a specific chemical entity. CMA 049724 / AN EVALUATION OF THE PROPOSED AQUATIC LITE CRITERION ON' TRICHLOROETHYLENE Ccr.rlusicr. defensible number for the ma;cr.:-r.-.ra:i:r. char sr.culd nor be exceeded is 7900 ppb. A chronic number needs no be developed. Prepared by W. B. Neely Environmental Sciences Research The Dow Chemical CompanyMidland, MI 48640 June 1, 1979 CMA 049725 GENERAL COMMENTS The key data used to establish the aquatic life criterion are free, unpublished EPA documents. Until such data is published :r. uhe open literature, the criterion as proposed lack credibility. The criterion to protect freshwater aquatic life is based on invertebrate species and should be so stated. Specific Comments p A-l. 1) Trichloroethylene is not used as an inhalation aesthetic at this time and the statement indicating such a use should be deleted. 2) The paper by Pearson and McConnell is a general overview paper and should not be used as primary reference. 3) The paper by Reading et al on bioconcsntratior. was not available for review. Again, this is an internal EPA document and is not suitable as a primary reference. The statement "may bio accumulate" implies a potential problem. At the expected water concentration it is inconceivable that bioconcentration will be a problem. Hence, this statement implying bioaccunulatior. should ue deleted. CMA 049726 2- CRITERION FORMULATION B-5 The guidelines for establishing a final acute invertebrate tonicity number suggest that the geometric mean of all acute values be estimated. Since there is only one acute value, it is impossible to determine a mean. Hence, this method should not be used. A more defensible number is 7900 ug/L based on the fish acute studies . CMA 049727 f EVALUATION OF THE HEALTH EFFECTS SECTION OF THE CRITERIA DOCUMENT FOR AMBIENT WATER FOR TRICHLOROETHYLENE Conclusions: The evaluation of the carcinogenic effects of tri chloroethylene is compromised and the derived criterion is very questionable. The significant toxicological literature on tri chloroethylene is adequately reviewed. Prepared by: Rolf Hartung, Ph.D. Professor of Environmental Toxicology The University of Michigan Ann Arbor, MI 48109 May 14, 1979 CMA 049728 General Comments; The criterion document on trichloroethylene (TCE) covers the available literature relative to toxicology and health effects well. The evaluation of the carcinogenic effects of TCE is compromized and the derived criterion is very questionable. Sp(r' f jc Comments; Pact: C-2 (top) What is trichloroethylene chloride? Pago c-21 (bottom) The interactions of TCE with CNS stimulants or depressants arc speculative. Paqc C-22. Considering the weakness of the mutagenicity data the introductory statement in the teratogenicity section is inappropriate. Puqes C-23 to C-24. The last sentence of the mutagenicity section should be used to introduce this section. The occur rence of impurities in the TCE samples tested has so compro mized many mutagenicity studies and the carcinogenicity study, that it should be treated as a major point of discussion. Page C-2d (center). The reported concentration of IN TCE in tissue culture is either reported in error, or the resulting changes bear no relationship to any real world changes. Pages C-2-) to C-2b. Since the proposed criterion is for TCE and not for epichlorohydrin or epoxibutune, the NCI carcino genicity :ala cannot be used to establish a criterion for TCE CMA 049729