Document KGyD0ZzkkKLJv5GGwN2b4KOmQ

Urn J. mduMr. McU. (V71. J, 226-2JA insulation workers in Belfast 3. Mortality 1940-66 P. C. ELMEStnd MARION J. C. SIMPSON 0pimnt of Therapeutic* and Pharmacology. Queer.'* wnivornty, uifu*i GO --f o o CO cn co cr* cn Kim**, V. C,, m4 SImpute. Marfoa J. C. (1971). Hrit. J. uuiustr. Med.. 22, 226*236. insuij;,o(i when fa BeKml. i. Mertallty 194046, One hundred and seventy men jweM ider.lilied ax making up the total population <tf insulation workers iaeilast ,r, "..s n an analysis of nil the information about deaths (hut h*s emerged from i.'ucizitf ;hcse men up to inc end of )%6. Five ranwm uthrmitd. end the nutfidUiy experience of ;he tertu mdezi* compared wifli that of other men in Sorthern Ireland over (he peraid, There were V* deaths when .wly .17 were espected. The number of deaths occurring escccded humc epe.ied ; hr. ,ei live p*rmd ISMO-M and the increase was statistically significant during the per.od iWtfuSJ and onwards. There wet an especially high mortality (compared with other Northern Ireland mates) due to center of the lung, mesothelioma of the pic -ru and pc.M.i.'eurr;. caret, c.f the gastroinicstinul trect. end lihrotie lemons of Ihc lungs. The mho of observed over expceicd deaths was 2 6 for all causes. 19 for all canters, and i 7-6 for cancers of the lower respir-iory tract and pteuru. Those men finally classified as dying from lung cancer showed evidence rf lung fibrosis whereas those cla**ifard asdying from mes<Khciioma die not,('.wiparisorts w i.h.r, the group failed to show any relationship between age a. hrsi e<>>siire or dar.i;,on o.'cs.vajm and the esevssivv mortality, There were too few rt.m-s.y.okerx to show the s,g,p,,,ea.-.e of smoking. The first two studies (Wallace and Langlunds. 1971; i-anglands, Wallace, and Simpion, 1971) were con cerned with the health of men ilill employed as insulators; the third study is an attempt to evaluate the deaths which have occurred in this occupational group. For a complete picture of the occupational hazard it would also he necessary to find out the proportion of men who had to leave this occupa tion for medical reasons other than death. But records giving the reason* why men changed t.i lighter work or became unemployed were nonesisient and information about those who drew sickness benefit was scanty and hard to cook by. Therefore, a was decided to restrict the enquiry to the frequency and causes of death. As there were no specific regulations concerning the handling of asbestos in Northern Ireland up to the time of the enquiry, there were no offica.1 records of the numbers of mer. employed. Because of this lack of official recognition, few patients were referred to the pneumoconiosis panels and no special arrangements were in existence for recording (he causes of death tn this group. The enquiry was conducted m three stages -- ,nc identification of the popula.ion at risk, the idc/.tificaiior. of ;he men wno had .vd and. finally, tnc analysisof ;he ;r'nrma;.t.T. which eouiu be obtairatu a.xtiii .hese cicatns. To define the population -i r:w, a preliminary investigation of the employers' and trades anion records was ,na-c. Work or, the other two studies had indicated that tlscse records were r.o: complete arid became progressive.y isss so inc fer.her back they were traced, But other work on the malignant hazards of asbestos exposure indicated a long latent period ,'Xno.. dxxi. and Kdi. .965; times, KM',, so thus .vowttvv rcs-iis could au ex.-vetod only if a d26 wiirkir.g period, A -i , ,`Mil i.s title - I fate of ali .'i liWui.titttl v - t previous cape `iUi fuplaulaa. \ list of il; : ' enve/erx iiuf... i,,fol.'aaiiiili i inis part.ee'. t Transport a^.. ..re care i.vtle,. were meomp, .: ofliuincd Py study of .tin ' extremely tun i .hint of (Jit > -ontined v v... : wT.u lived sio-e i l-wliow.'.ip . were ;u;cr t c.eur,'otee, mu -f ;>VX4 Ctrl' JS'I Af.'snyerr*,'.; . arid ItiiHi' n'l w. / .Tier, :o ke.,;- > wCurr.ny afivi s (o the corn,i-:: i prisvhle. Wrap ; was ubl.iin. . < Oallv' jrVsa I jJi'fiS'. * *;?: {he , (h'hinu t'lMf . i ^ycpa/i/n^n; ^ . jtmn. Hospit.s1 i where d wjn ; were re-eurr;r-.o in Inc !ij;h. o,' e- Both Cor 1.V i utJOr. Wjs u'. : oeaipu;ioA*ki r irsj; habits;. A * ihui no rei;iv\ rr.i%s:Uv .Src*WoJ >>.' e:~st / / .if i` i hy LOfnp.ti'iht' y uroep wiih ;r.v . i whiiie. <: Uu% r.el i. il\ aiid Wk'i S(V "W . i s working popui,. i way. :i s\: } ST0095367 Insulation A'urksrs in Xei/ost. Motl;o,:y /'AYj-A/j +Z7 working population could be studied over - long period. A compromise was achieved by choosing 1940 as the starting point, and this is a study of the fate or ail men known to have been employed 0,1 insulation work .n that year, irrespective of age and previous experience. groups according to tihc age a; which they became insulators _r.d this was taker, to be their first sign:hear,: exposure to asbestos. On examining the material it was found that very few of the men employee ir. 1940 kefi after a short period of employment. The majority continued w orking until they become disabled and died, or until they retired.Therefore, those whojoined the industry young haa incioiyesi exposure.ir.d lfio*e who joined late the shortest NlMfecd* Th* puputulioa A list of all tha it- employed us insulators and pipe covcrcn during 1940 was drawn up. The sources of information were the lrude union membership lists for this par cufar trade (11/72 Insulation Branch of the transport and General Workers' Union! and the boohs and card indexes of the various employers. These records were incomplete but were supplemented with information obtained hy interviewing other insulators. An inter.s.ve siody of ibis population irr lh last live years makes n exp,.sure. The distribution of men according to a>e in 1940 and age at entry into the industry is shown trt Table I. Age survival carves were eiinstrdelcd for each of the four group* of men who joined up to Ihe age of 2o (inclusive), 21 to 50, .11 to 40, and over 40 yean (Fig. I and Table 2). Age survival curves dividing the men into groups according to i.'teir Juration of exposure were also constructed but showed similar results. In the eonsirueiion of these curves it is the propo.-uon of cacn group at risk surviving each five-year age period that is plotted. estremely unlikely that ihe final list falls as mush as i shun of Ihe aelaal total particularly as (bis trade was confined to men from a relatively small number of families who lived close to the shipyard in Heifast. hollow-up was not easy because the original addresses were often useless because of war damage and slum TAux,}'. i Dl.sTalbUTIOV Of VltN ACCOXWV. TO Axjr. IN 1940 AND Aon k~ ENTRY clearance, but family relationships often helped in tracing ihose still living. All those still alive were interviewed. Arrangements were made with the general practitioners and other doctors concerned with the care of the iiving mm to Uep us informed about (heir health. Deaths occurring after she beginning of the entfutry were reported in ihe coroner and Mtropsy was carried utn wherever possible. When deaths had already occurred, information was obtained from the nearest available relative as to ihe slate and plttx of tleaih and previous illnesses involving hospual admission or investigation. This was followed up with the co-opcralion of the Registrar General who con firmed ihe place and dale and pv the registered cause ticaih. Records of the hospitals, Queen's Univcrsi;/ Department of Pathology, and the Korrnstc Pathology Department were then searched for addtikmai inform Ayr in !</#, ! - 'i/p,o l.p in i'/ U-U ' M* rt>-44 4,-.*<. i fcV'-i 55-54 AO jr.J ever j 4 il A 4 .H y Twwi 5* u-r entry -- -- _ -------- ll-JO JI-40 41 i i i Tutci \ s lit )< t V f? y ivl .S 4 5 i i : ./* 1 *2 T! 4 4 2* }; *i * fo \h 1 1 bS ation. Hospetal records, e-rays, pathological reports, and. where it was available, Ihe actual histological material were reoaimintd in order to determine live cause of dca.it in the light of current diagnostic criteria. bosh for (he Irvtng and for the dead as much >niV,r,Ti- asnm as was andaWr was menial about each man (to ocsupa*n>nal history, mnhcai hntoey. abode, and wnoitng habost. A family history was also taken to nuke sen: that no relatives who were insulation workers had been missed. CaKapcrjveti wilb ilia fnuttvi turpcU.kM, The Northern Intiisl kcpslrjr General fondly provided uMcs u4 nude deads* for Use wtsofo prosaaee for rack yur foosc iv-al onwards. The deoihs were lasted in fisc-year tigs groups With ar, estimate of., e total popiiiaiicr. for this a tie group for the relevant year based or. ir.e ;-s. The tahtes gave ng-res for cancers of the lung, mediastinum, and piet.ra Methods of analysis I videnee of occupational effect on mortality was sought hy compartsons within the group and by comparing the group with the male population of Northern Ireland as a whole. It was not thought to be possible, with the resources available, to select and investigate a control group of the working population and follow its mortality ,n the same way. (Nos. 1A2-4) a.-,d for cancer of aii types as well as for deaths ace to ail causes. The insulators were divided ir.to the same five-year age group* for each year of the study to give the `pop.elation at risk', and the number of deaths in each group obser.'ad ;n each year was recorded as ihe'obseTved deaths'. Fur the pur,vise of the analysis both the Registrar General's figures and the figure., .fir the insulators were aggregated into five-year period* (Table ft;. The figure forpeeled' death* for each S>vy-:ar period wat obtained ny calculat ing the expected deaths for ever, individual year and the.', Cammrham witUa dw group The men were u.v.dcu ..-no -ddmg the five years together. Ir. ,ciing (he i.e-ace 229 r < tJmn ttmJ 1/ J C Stmfum i.u. i. influence of aye at entry into insulation work on survival at u.lTeren: uges: <u> ul; duoi/is in ;i-,e /.lunger entry groups are included; (A) deaths occurring in the younger entry group. before the aye of 45 are excluded vo that ell groups start at 100% at the aye of 45 and their survival from then is more directly comparable. point on each curve represents the pctccntaye of those at risk surviving :o that age. ol JaftROon between ofetcnnl and opened aurahm of deaths j* was calctdaiesl directly from this difference and the expected values. Althouyh this method makes due allowance foe aye and to and for the changing palten for mortality ia the province during the last 25 years, no allowance is possible for more precise doamatc, for social class or for smoking habits. Where Agues /or .ic county foroegh o/ oc./->; were available those are shown in the tables, ia; on!\ fragmentary .nformauon about :hc eh'cct a: sea*! tic . (Park. i9f>5-6i and cigarette sooiuitg (Wicic.n. ]/>*, or the paiteni of monad., was ascdooic. '.Tirs ituonval ton was used to test the valw.ly of the couciusioro drawn from the comparison with the general ;sopuh.tton. TABLE 2 Deaths and Ace at Evntv 4fr mt mtr\ l > M hlluM li| \< mA1 Dserf Rate < ptt ymr> 1. j t/pi-45 1 jM 4? 7s fMo j 2 5 .Coopers*/ 5/./.1 ' "i ii -- -- * -- Ji x : S i 7 */ ; 2 1*5 J 5-6 ;o a) 29<V; M rM fisrsl Rate l *. per ycafl l)-40 tnwn 13 Si At frAHk fHfcJ Kak- 1 *. per >eri I7 '( 0-6 1 1 OJ1 : 7 f .1-2 35 % iJ r< :? (S 5 4-5 : 6-2 42 5k j.) a i.i s *S as 2 A ) U rM OvO ate 45 5* per xui ? . .3 i-i i 1 i1 i 4 ! ~ i 1 J i ^ : i 4 <U ` Ai in.' tanw af --a u wlr>m *ayt a aaary* yny wnnur w the htuaa; udiag agr period uiiocjs44 lew l/'wl * *en4 rf A* aye yanad by *r bear rfer wady Was mad* ?au; ut "ami i`i i : < Asr s'ox? I 'A.' ; C/> :o-2. 25-29 ' i C O i 12; o : n. .*> vo II !S; . 5 c/l to w-je cn ( ! oo '.s> V<r sy i (i i'vLunuickl yH>fkiUi`i ' s>v group m v-f .'CilJiftV ^UC <0 s/ll ssi.;s * Itik >fM\ n^s' ^r.kuij't Js\r > (Si 'hliO.Aor of UVkka^Jl... k !<mrttnf vt .'iw- . <*44 ; Ssr.Ti.KvU nuiV'tv* . * sSouUl Ji*VC OkVu.Vs CUrCsiU.iwJ OV sks^Aeisv ' ' y;ufs y,\; Oh>*-rs"*svSl oc*'`MN v.i* l <rivci*tjun u\<hcr> * Kasii mc>;.cirdsY prsHV fulC Cm*t iK* foJW>*V\g 0\ . ' l ooun :hc i-thw u r.V; r.^ii: itirwvinv: tf*c *,'-**>** N.v c dupMdil) Jsmr p :,. i "7*V* IW >7W. v'Npccta! Jcjift' forth* \kV. <wc 7.i.no 4:. jnJ ;bc i cxpwvicsl Juths r. iij. \ ; Institution workers in Belfast. 3. Mortality IV40-M 229 ST0095369 TABLE 3 Pakt os Tabli: lo snow Method os calculating Exacted Deaths ttroap 20-24 25-29 30-34 .15-39 1W-9 Year !9i0~* IVJ.l-V (1) 57 000 (2) 112 (1) 15 (4) 0-155 (5) 1 N. 11) 4 7 500 <2l 121 (1) IS (4) 0 25a (Si 1 'V 71) 47 700 (2) 102 (1) 14 14) 0-136. (5) 1 N. 11)45 000 (2) I3S (3) 17 (4) 0-23v (3) 1 N. (1) 40 600 (2) 101 (3) 11 (4) 0 146 (5) 0 About onv-lhlrd of the men Joined IH Industry before they were 21 years of age and another third joined before they were 31. Although most of the rest joined before they were 41.16 did not join until they were older. There was no evidence that this pattern of widely differing ages at entry had changed over the previous 2S lo 30 years. Deaths Between 1940 and 1966 inclusive 98 (39-3 *.) of those traced had died. Table 4 shows the distribution of these deaths in time. As the population grew older the mean ages of those dying tended lo rise, and the proportion of the survivors who died in each year also rose. As this pattern of changes was what might be expected from following any group of assorted ages over a period of years, the control studies were needed to determine whether the death rate as a whole, the rate of increase in death rale over the years or the sudden increase front I % to over 3 per year after 1950 differ significantly from that expected. 40-44 71) 44 300 (2) 169 (3) 16 (4) 0 285 (5) 2 (1) Estimated population in Northern Ireland of men in this u|t iroup in one of the live >eurs concerned (2) Deaths due to all causes among men in Northern Ireland in this age group for the same year 13) Number of insulation workers in this age group at the beginning of the five-year period 14! Estimated number of deaths due to all causes which should have occurred during the five years. This is calculated by adding (2) (it for each of the individual (I) years (3) Observed deaths due lo all causes in this group of insulation workers in the five years Each five-year age group of men in I MO forms a cohort whose fate can be follosvcd over the yeats by moving diagonally down the table to the right (see arrows). The number of men surviving the group's five-year experience enters the square diagonally down to the right. The totals of the vertical columns give the observed and expected deaths for the whole group for each five-year period Iscc Table 4). and the horuonlal totals give the observed and expected deaths in each five-year age group (Table 3). Comparisons within the group The influence of age at entry and consequently the duration of exposure at a given age is shown by the survival curves (Fig. I) and tabulated figures (Table 2). A cohort of men joining an industry at an early age will contain some sick men by the time they reach the age of 45 or 30. These sick men are being carried by their males. But a cohort of men joining at the age of 40 or later will have to be fit enough to do their share of work. Consequently the late entry group, being healthier, will show a low initial mortality compared with earlier entries (Hammond, Selikoff. and Churg, 1963; Newhouse, 1969). Thus, ever adjusting for those deaths which occur before the age of 43, as has been done in Fig. I b, the slope of the survival curve of the late entry group (over 40 years) shows the expected difference from the others for the next 10 years There is no consistent difference relating age of entry or duration of exposure between the other three groups. All four groups show a high mortality beyond the age of 33 and, considering the small numbers involved, the rates are remarkably similar. At the timeof thestudyo.ilytivc men had lived beyond the age of 70 and these had died at 71 (4) and 72 (11. Results f*npahllnu One hundred and seventy men were identified as being employed as insulators or pipe coverers in Northern Ireland in 1940, Five were uniraced, leav ing 163 for analysis. The distribution by age at entry and age in 1940 is shown in Table I. From this it is seen that there was a wide scatter of both age and previous exposure at the time when the study started. Comparison with the general population When the deaths occurring in each five-year period are com pared with those expected for the general population adjusted for age, the significance of the figures is revealed. Compared with Northern Ireland males (Fig. 2) there is a progressive divergence throughout the period of observation. At no stage is the experi ence of ihc insulators, who were all fit enough to work in 1940, better than that of the population as a whole. The divergence increases with time. This ( t -tf k-< hmilaliuti workrn in Hel/utl. 3. Murlainy /Wtt-M 229 tabu: 3 P*at tit Ta*u i< ui Muwt) it ( AirtiAiiVj Expinio OtAtii* Agr group 2u-U (WJ-V IVflf /KM /VJi.v Ahum one-thud of the men joined the unluuiy Mur they eat 21 >cart at age and another tlurd joined before they were 31. Although mo*! of the rev' joined before they mere 41,16 did not join until they nere older. There oat no evidence that lhit pattern of vt idcly differing agev at entry had changed over the previous 25 to 30 yean. 2M* M-.M Hi JT eu Ci IC iJi II 14) 0 III Ui i N. Ill 47 SOU Cl 12.x Ul It Uj 0 25f. (Jl 1 \ (ll 47 700 Cl 102 U) 14 (4| 0IJ6. (ll 45 000 Cl Ut Ol 17 14) 0 2). <J) 1 N. traced Bad dud. TaMr 4 how* ike hiirdumna at Ihcte dcalkt mi ume. At the popubuoa gen odder the mean age* of those dying tended to me. and the proportion of the turvtvort ho died ui etek year ulto rote. At thit paiicm of change* was what might be etpeclcd from following any group of attuned aget over a period of yean, the control uudict were needed to determine whether the death rate a* a (ll 40MSJ whole, the rate of increatc in death rule over the (2) 101 year* or the Hidden mereate from IT,, to over 3*,, U U (4) 0 146 (J) 0 per year after 1950 differ tignilicanlly from that expected. M4 711 44 JOO Cl 16V (J) 1* (4) o :t; (3) 2 (ll I'tiinuicd population in Northern Ireland of man in (hit ate troup in one of th< due >car> concerned ill IScatht due io all cuutct among men in Northern Ireland in thi age group for the tame year <.<l Numher of uituUlion worker* in thit age group at the beginning of the rive-year period |4| Ittunated numher of deatht due to all eatna thth thould hate occurred during the h\c >earv TKn it < ikuletrd h> adduag1,1 ' -* for each of the mdonlual teart (St Ofctcncd dntht due u all camct m ihit group of intulauoo workert m the Ate >cart Each Ave-yew age group of men in 19*0 form a cohort nhoie fate can hr fatfoued oner the ywait he muting tupMdb du*n the laMr ta the mpa tme trwau The numher admen winng the gruup't hcaeu opmen.e enaert ihc ogntac lngnmfo <lo*a to the ngiu. The tatah of the ccnacat oahamaK gw the efourve* ami eyoud hi phi foe for uhrfo group foe nnchhee-juag pater I or TgOOc At. amt the hintimil totah gne Iht thieei ed ami ctfmtaJ fcuahi m mth foe-year apt png iToaOc 5k Comparham wtlhia the group The influence of age ut cniry and contcqucntly the duiatwn of expoture ui a given age ix thown by the turvival curv>rt (hig. 11 and tabulated ligures (Table 2). A cohort of men joining an induttry ai an early age will contain tome tick men by the lime they reach ihc age of 4$ or SO. These tick men arc being carried by lheir mulct. Bui a cohort of men joining al the age of 40 or taler will have lo be lit enough to do thar there of work. Conteifuemly the late entry group, hang healthier, will thow a low initial mortality compared with earlier annex I Hammond, Scbkotf. and Oturg. I9&S; Newhouse. IM4k Thu*. eve* adpnliag lor thuuc death* which occur before the age of 4S. at h* been done in Fig. lb. the dope of the turvival curve of the late entry group tuver 40 yeani thow* the expected difference from the other* for the next 10 yean There it no cunutkrw difference relating ope of awry or duration of ctpxoR hetween ike other three group*. AM four group* thow a high aajrtahty beyowd Ok age of 55 and. ooaudenng the tmall number* omibed. the rate* are remarkable unmiar. At thetuneof ihettwiy safe hoe men had bted beyond the age of % and these hod shed at 7lt4iand'2lt Oacbunked oad tcveniy gnat were ntettidted u hens employed at anulikn ar pgx ooMcrer* ua ? whru Inrtiml IWl Kmc nar ftri nl. leut- mH3faiad(hh.TktetahM)T4taa} anal age iWhthangaToHc I. Fiat Am < it nets that iat n a nide taatr aff hock age and ptettomnpa --adciaacnhadgdnj} itgitl death* worm n nock kve-twor poxd one owe pared with those expected for the general ppibixa adjatied for age. the ognthcoitcc of the kprti it receded. Cinpuel wnh Norther* Ircfead ado ifg. 2i thcoe it a pwpciute hvaytMC Or up w *e pervd of ohovowo. At ao tup r> foopm- eaoe at he eWhUv who woe ah at enough to week m IMO. hens than hu of he pgdinw at a mhoAe. The dnrrgrarr maeun *ek Mae That ST 0095370 2)0 /*. C. EJmes anti M. J. C. Simpuin V-O cn CO TABLE 4 DEATHS IN FlVE-YEAR DIKII JDS Cl ASSII l< /MIM At n&k Death* 1ft3 10 Mean age at death t > r > *. dying year sot 12 Deaths expected <N.I. males) Significance SMK* Deaths expected 1 Belfast onJ>) 5 54 0 1 > P > 005 180 /wrs-v 155 7 51-4 o-v 541 P> 01 12V Pvtiod iVSo-4 /VJJ.V 14K 122 Jh 2U u* 5*0 5 5 .1.1 ftW H . 0-UU0J lft.1 7-2 77 P - 0 001 274 Jau-4 1112 1H IV6S-6 71 A SY-4 57 41 V-0 P < OUUOI 3.15 10 7 11 01 > P -> 0 05 162 *SMR standardized mortality ratio. Although the number* are too small to give the precise meaning to this ratio which it has for large populations, it gives an idea of the order of the difference observed. Malignant Upper respirator* lower respirator* Pleural mesolhelioi Penumcal mesothe (asiroimcsinial (icmiounnar* Non-nialign.iiti Respirator* 1 uhereuluMv Other lung disease ( urdimaseul.tr Other information is shown in more detail in Table 4 and a similar comparison by age in Table 5. To show the order of the difference, standardised mortality ratios (SMKs) are also shown in these tables although the numbers involved are insufficient to give the ratios much precision. The significance of the differences is statistically valid but no allowance is made for the environmental and social factors. The divergence becomes of obvious significance after 1930, as might be expected from the sudden increase in mortality in the workers ut that time. Table 5 shows that the greatest difference between the workers and the general population occurs between ihc ages of 30 and 39 years. Yeor fig. 2. Survival of inflation worker* at S-year interval* compared with the male population of Northern Ireland adjusted for age. By the end of 1966 only 41 * (67 men) survived compared with 7)*i (121 men) expected to survive. Causes of death Table 6 shows the causes of death observed both as given on the death certificate and as revised in the light of all available information. There were a number of differences in detail between these two lists and there were also a number of cases where Ihc TABLE 5 Deaths by Age Tlic death .cr delituie evidence . lower rcspirninrv c.i plcurjl mesntheltnm two which may h.iv evidence was oiMilhc Hon. Iheelleci ol'tr.i to malignant and ok nd plcurjl c.mccr I discussed in tlie lev! linal diagnosis n particularly the Registrar Cicncra but three other i were diagnosed oi There were also m review could haw one case dassiiie which might havv Because of these classification was this classification grouped together included in the g.r Using this broai between the dead Apr At risk Deaths observed Deaths expected INI) Significance SMK1 Up to JV `44 45-4V 30-34 33-3V 60*64 63aIV 151 I.W 116 III *2 4V IK 5 8 12 21 21 11 12 0-5>P>0 2 156 10 0 0l>p>0 00t 267 4 21 Pel) 001 2*4 62 P< 0-001 ,12V 70 P<0 00l 32v ft 0 oot > p>o-ooi 216 47 0 1 > P 0 05 170 7tt utui our K 124 ' 246 `S* (uoinoic to T*W 4. l amer tail tspCM TubcrcuUoi** Kon*iuhcrcuU>iift uJw Other ouw> n land j km the > ratios ugh (he Ithe ratios Iferences in kde fur the f significance 1 (he sudden Jt that link ; between il . curs betwee i rved holh .vised in 11 icrc were n ihese t cn where i! hiuihiliiiii workers in HeUust. J. Mortality IV40-66 231 TABLE 6 C l ASMMCAIIIIN III C'aUSISIII l)l A III OttilJi 4 trlih* i alt kfMSiil t lawifti ution 1 UIM' liU'if ill uimtyus Malignant l pper respirator) l.ocr respirator) Pleural mesothelioma Peritoneal mesothelioma Oasirointeslinal (cnitourinaf) 1 25 l 0 l.t 1 1 241 4/ n 12/ 1 45 1 2M 15 l Non-mulignunl Kespirutory (ubcrculosis Other lung disease Cardiovascular Other 3* bb b 27 2b 26 IP in K f A*nr: The death certilieale diagnosis was revised onl> when detinue evidence was available. There remain among the lower respiratory cancers nine cases which may have been pleural mesotheliomas, and among the gastrointestinal cases two which may have been peritoneal mesotheliomas. The esidencc was insullicvcnt in these cases to justify reclassilicalion. The clTect of transferring four cases from nun-malignant lo malignant and increasing the number of lower respiratory and pleural cancer from 26 to 2lf on stalislieal analysis is discussed in the test. arc less important. The revised diagnoses were used in the comparisons which follow. 1 he cirecl of raising the number of cancer deaths from 41 lo 45 and lower respiratory cancers from 26 lo 28 on Ihe statistical significance of any differences observed will be shown to he negligible. During the first 10 years there were six deaths due lo tuberculosis. Evidence of healed disease was found in some later deaths but Ihe disappearance of significant tuberculosis coincides with the availability of satisfactory anliluherculosis trculmcnl, The other change in the pattern of causes of death with time (Table 7) is the increase in the proportion of mulignunl deaths. This increase with time is to he expected in un ugcmg population and il is necessary to compare this group with the general population to determine its significance. Malignant disease The comparison between the observed deaths due to cancer of all types and deaths due to the same cause in the populations holh of Northern Ireland and of Belfast in particular is shown in Figure 3. There is a marked increase from 1950 final diagnosis remained in doubl. This allcclcd particularly the diagnosis of mesothelioma; ihe Registrar General recorded only one pleural case, hut three other pleural and three peritoneal cases were diagnosed on ihe basis of postmortem findings. There sscrc also nine carcinomas of the lung which on review could have been pleural mesotheliomas, arid one case classified as a carcinomatosis peritonei wlilch might have been a peritoneal mesothelioma. Because of these dillicultics only a much broader classification was used for the statistical analysis. In this classification cancers of Ihe lung and pleura are grouped together and peritoneal mesotheliomas are included in the gastrointestinal group of malignancy. Using this broader classification, the differences between the death certificates and revised diagnoses nt.. 3. Deaths due to eanccr of all types occurring in insulation workers in each live-year period compared with deaths expected for age-matched males in Northern Ireland as a whole and in the county Borough of Belfast. TABLE 7 C'ltANOtS IN CaLSI-S <)! DtATH nitwits 1940 AM] 1966 C jnccHalJ tspes) Tuberculosis Son-tuberculous chronic lung disease Other causes /V.l/J-V /VjM-V I 2 1 1 1 1.' \ 000 517 t7 I I r> h 7 i titul t 45 n IS \ 2b 0 21 ST0095372 CO --H O CD <JD cn CO -4 CO 232 /*. ('. t ime am! M. J. Snnpuut onwards and from then on cancer dcalhs arc nearly eight times more frequent (han expected. Die ligutes arc given in Table 8. Twenty-eight of (he 45 deaths were classified as cancers of Ihc larynx, lung or pleura, which repre sent 75*^ of (he excess deaths due lo malignant disease. The distribution of these deaths hy age as compared with Northern Ireland males is shown in Table 9. For the whole group these cancers were 17 times more frequent than expected. Mesotheliomas Towards the end of the period covered by the survey an intensive search was being made for mesotheliomas in Northern Ireland, but a maximum of 12 new cases w as recorded lor any year. As these cases were derived from the whole popula tion of Northern Ireland the expected incidence for a group of 165 men is negligible even when considered us a proportion or the 98 deaths. Therefore all the seven confirmed cases (four pleural and three peritoneal) must he considered in excess of expect ation and this incidence is highly significant. Other cancers There remain 17 deaths due to nonrespiratory malignant disease. For Northern Ireland males the expected figure was 5-16 deaths. The threefold excess is highly significant <y` 2X04. I* . IMXJ5). Comparisons with Ihc general popula tion under more precise diagnostic headings were not made because of doubts about the precise diagnosis in many cases. Hut 15 (including three peritoneal mesotheliomas) were classified us gastrointestinal in origin. The revised diagnoses in these cases were three mesotheliomas, four carcinomas of the stomach, four of the colon, two of the rectum, one of the head of pancreas, and one lymphosarcoma of the small intestine. Of these, one classified as gastric and one us rectal carcinoma could have been meso theliomas. Non-malignani disease There were 52 deaths due to noil-malignant disease when only 30 9 were expected. This difference is highly significant ly' - 181: P c 0-005;. The mortality was higher than expected throughout the period of observation; the sharp rise from 1950 onwards noted for malignant disease was not seen. The 13 cardiovascular and eight other nonrespiratory deaths were rather fewer than expected, and the excess lies among those 26 cases classified as lung disease other than tuberculosis. It is difficult to put these into precise diagnostic categories so as to compare them with the expected death rales for the TABLt8 DtxiHs om mi All Tvnsoi Canok in Fivl-yia* Pmums Af risk Death* I xpectcd Northern Ireland male* SMR1 Belfast *>nl> /WO-4 Ift 5 1 0 605 us /WJ-V 155 o-vn ,vu IWi-4 |4K n 1 275 l2o 1*466 /V.U-V 122 M 1 52V 12U tVhO-4 102 14 1 747 2-24.1 bor ihc whole group 2J4; H*.0-0005 `See* fooinow \oTaV4t 4. ;4I registered dejths due lo all i>po of vaneer. mJ- 71 \ 0-774 >40 total I#i5 4V f K TABLfc 9 Dlaths di l >o Canok oy TIIL l.AKVNX. LllMi, AND Pu.CKA uv Ac.t At risk Observed deaths 1 xpeeled (N 1 l SMK* Up lo 44 1w 1 0 006 4S-4V Hr. s o-ui 14(10 5D-S4 Ml X 0 245 V40 J5-.lv K2 4 a 442 mu 4V 7 a xm IK44I 4.1-A V IK 0-222 >Ul 7ti H 5 o-nn 4*50 Tor Ihc whuW fruup evin il'ihriv ik-jlhs am* v-vp.-cli-d xJ *Scc I'cHUHv'tc to Table 4. *2( registered deaths due lo i!k*x* saueers 212, P- IIINNI3 itOtii I m# 17(a) population .r the dealt) cci cause of dcu tuberculosis logical cudc mformalion. whether ushcnoi. So that . evidence (Kit coiitrihulcd ti group of men assessed. Pulmonary fihi the deaths tin evidence ol'esi easily sulhcicm eases of astx* patients died gasiroiniestin.d cases ot plei Conversely, ol carcinoma of I about live, and ing eight all v histological esu carcinoma of it fibrosis. Of the case showed no and three i>erili on histological i radiological cviv Influence of otiic The prifflurv ana group of men iwhole male pop ever, these men populated area o and semiskilled s domicile and sou forty their inori. expected lo he Northern Ireland possibility thoi t cigarettes llun > Taking each of il and considering . causes of death. I as follows: Domicile flic u males lor 1950-4 a sulfieienily dillcic o 'r tf *' >*ui %igniiK.p>i *\ `mM; .ith thv Ail' *' ; .'u/u* nn%lic Ju 'sue not >ut the i`u. ^ Ji.i , .i\j\ .'ludini! in., itii.'ucal cd as ..... tin.il in se* in 11ich* n.- v' were f Can.numuN the *o of lIk' kxIuiu. iwic of c l>nipluiNauomj i>i rtvc ' classiha! as k ind Juid h.oc been iiicmi- e were 'ideaths dm- w kpccled. IK I. kpcvlcd Insulation workers in Ht'l}a\t. J. Mortality MO-to IW population us u whole. Ashestosis was mentioned on for 1960-2. He gives a ligure of 197 deaths due to the deulh eertilicute und regarded us u contributory lung cancer per KXHXXJ living for men over the age of c.iusc of death in nine cases, one being a case of 25 years living in `Inner Belfast' compared with ItMi tuberculosis. There Mere 11 other cases vs ith radio per Northern Ireland as u whole. (Inner Belfast logical evidence of tibrosis but no histological produced 25b of the H73 lung cancer deaths analysed.) information. In eight cases there was no evidence This would bring the expected ligure for cancer of the whether ashestosis or other librosis was present or lung in insulation workers from I 64 to 3 23 where 2b not. So that although there is strong presumptive were observed (Table 9). The difference remains evidence that pulmonary librosis due to ashestosis highly significant (y1 - 193; P < 0UX)5)and would contributed to the excessive mortality among this not he less significant if the death certificate ligure of group of men. its numerical importance cannot be 2b observed cases were used. assessed. Social class insulation workers arc a mixed group Pvtnonary fibrosis and malignant disease In 40 of the deaths there was radiological or postmortem evidence of extensive tibrosis (excluding tuberculosis) easily sufficient to have justified their certification us cases of ashestosis. Twenty-two (55 "Al of these patients died of cancer; 16 (40U| of lung, live gastrointestinal, and one of larynx. There were no cases of pleural or peritoneal mesothelioma. C onversely. of those 24 cases finally classified as carcinoma of the lung, there was no information jhoul live, and extensive librosis in II. The remain of semiskilled and skilled workers and if it is assumed that they represent a group, half of whom arc skilled and half semiskilled, then allowance can be made for social class based on the data for Northern Ireland given hy PurMI965-6)andWicken(l966), hoihha-cd on the 1961 census. Their figures for class 111 and IV manual workers give an average SMK of 97 5 for all deaths, I(XI for all cancers, and 114 for cancers of ihc lung. Therefore adjustment of the expected figures for social class alone would nol allcvt I ho sigmlicancc of I lie dillcrcnccs observed in insulation workers. ing eight all showed some librosis on .v-ray or histological examination. There were no cases of carcinoma of tl'C bronchus known to have no lung librosis. Of the seven mesotheliomas, one pleural case showed no lung librosis at all. the other pleural und three pentonvul mw. showed inlnlmul flhhi.l. mi histological examination only, and none showed radiological evidence of librosis. Smoking lubits There were 46 men in the group whose smoking hahiis were not known, and of the remainder only live did not smoke. Comparisons within the group (see Table 101 showed lhal neither of the two deaths in non-smokers was doe to earner I lieu) was no significant difference between the death rates observed due to respiratory cancer between those who smoked more and those who InHuenct of other environmental factors The primary analysis of the deaths among this small group of men is based on a comparison wiih the whole male population of Northern Ireland. How ever, these men lived and worked in a densely populated area of Belfast und belonged to the skilled and semiskilled social class. On the grounds of both domicile and social class their mortality and particu smoked less than 20 cigarette- per day. Taking those whose smoking habits were known and comparing them with the people surveyed h> Wickcn (I966| in Inner Belfast (Table II). it is evident that the insulators were heavy sinokets. These figures indicate that the expected frequency of lung cancer for the insulators should he 1-26 times lhat for Inner Belfast. larly their mortality from lung cancer might be These three faclors arc interrclaled in that the expected to be above that of the population of population of Inner Belfast contains a higher Northern Ireland as a whole. The third factor is the pioportion of social classes IV, V. and VI than possibility that this group of men smoked more Northern Ireland as a whole and that these social eigarctles than other men in Northern Ireland. classes also smoke more. Therefore to make an Taking each of these environmental factors in turn estimate of the combined ell'ccts of the three factors und considering also the clfect of reclassifying the by regarding them as independent lends to exaggerate causes of death, their importance can he estimated the result. Theligu.^sgiven in Table I2arcasneai an us follows: approximation as is possihle with the inu-rmalion available hut must still he an overestimate of the Domicile The total deaths expected for Belfast combined effect. The numbers of deaths expected aic males for 1950-4 und 1900-4 shown in Table 4 arc not increased hy factors of I-OH, I 22. and 2 M3 for all sufficiently dilfeieni from Northern Ireland as a deaths, all cancers and respiratory cancels, hul m whole to alter the significance of the differences spile of this the dilferenccs between observed and observed. No figures were available m lhe Registrar cx|ieclcd would remain hiuhly significant even if ihc General's tables lor cancer of the lung in Belfast death certificate figures of 41 due to all cancers and males, hut Wickcn (I966J analysed the data available 26 due lo lower respiratory cancer bad been used. ST009537I* ST0095375 2J4 P. ( . /iW' mill \ t. J. ('. Siniptmi tabu: in Smoking Habits am) Miwtai.ity Non-snmkcrs 1 fan ilmn 20 tins ith pip* nml o-*innkrn Mon? Ilian 2o day \ot Vmmn At ri\k 5 hj 52 4h Heaths (aft (awes) 2 1ft 2V 2V Heaths (rr%p. tamer) Observed i.vpet ted or. 0 in tin? 14 V 0 5' 17 V 0 462 IV 5 Si/r There in no significant dilTererKe between the frequency of cancer of the lung nr death due t* all causes in the t*n smoking groups it) (>> P>0 5 m both cases). TABLF. II Knows Smoking Habits comparm) with `Issrn Bh past' Malis Ciiuimte\ itat None li n II 22 ;\ Other (pipe. etc) Inner fii lloit 1 Slain \r SUM Im titof mn ) tana i am er\ 15". 22". 51 \ 1 (>\ 1 7 **,, IX 4 2". (S| 74 22-7',, (27i i: 50 5 (n '02 2 W)\, r25 XI 1 X (2) TABLK 12 Sk.mi i< anc r t >i I'xciss MnarMirv Anjisnn i>>k Six IAI ( 1 ass. Domiciii, ami Smoking Duntfon of (h rhk Although the higher mortality in insulators appears to have been present throughout the whole period, it became significantly different from the rest of the population from 1950 onwards. If the exposure to asbestos dust look a finite time (say 25 ycarsi to produce death, then the increase in deaths from 1950 onwards might have been due to a deterioration in working conditions in about 1925. In big. 4 the deaths are plotted at live-year intervals, comparing the experience of those who started before 1920 with those who started between 1920 and 19.10 and ihosc who started after 1910. If there was a marked change in conditions of w ork in the period before 1940, then there would he a synchronous change in all three curves, figure 4 shows that there was a progressive increase in mortality in all three groups as the lime K.imc prediction N Ireland males) \d)UNtmcnt factor* \diuslcd prcdichoo < )hscrxcd Y: l Aft 4 <IMW AH < ant rrt Resfunttfir t am vr 57 ft \ OX 4(1 6 /X 107 l 0 0005 A-X4 1 22 M5 45' 14.7 `) ' ooooi 1 (4 2xi 4 (4 2*J 121 (1 - 1)0005 ' \iliu.rtncni r.Kinr is produced h;. mnliipl\in|t Inpclher ihc `nilisidu.il r.iilisrs J f`,r tUss, donmile. .ind ."ii.kmiT In iKc s.isc .if smnting. *'nl> Ihc dilU-rcnic m smnAinil h.iHils between men di'mcitril m Inner lii`lf.isl .md ihc iiimiIiIiiiii wnrkers ss.is l.iken into cnnsHlcr.ilinii If ihc Humes 4 1 .md 26 .ire suhsi.mled Tnr Ihe 4X nml 'H slnmn (nr .ill i..msers ,md rs-spir.iinrs s.inver ihe snrrespnrHl- 'muri's Im t .ire I1' ,MU| 111r k,| p . nimiK bnih Yeoi IP.. 4. ('omp.iriMiti uf number of deaths due lo all eauws per 10t) lump insulation workers oeetirring in each llsc-seor period for those men who started work up lo I'W |s*2i-!>, and tvt|..m. passed, the increase been in the industry older. There is no cvi. in Ihc direction of the cd change in working r llte follow-up of I from 1940 to 1966 . mortality compared w Ireland. This mortalil ation that it cannot he factors such as smokir excess of deaths is tho related to their work described by Hammo workers in New York the work is related rcscmblcsthat shown ir factories before the measures and who ha era (Knox ft at.. I' excessive mortality ap ation of asbestos duinformation about the. insulation workers rcc from the severity of iht arc known to have heer commercial types, nan crocidolitc. They arc exposure to anlhophyl The complete patten workers will not he rev This studv shows tha already occurringat the study and had rcac proportions after only due lo chronic non-h occurring throughout ( these cases ashestosix contributory cause of d this cause may have co of ihc 26 deaths in ihifatal lung lihrosis is lot certain factory groups h hul higher than that pre in factories only afli tKnox ci ill.. 1965; Smi Numerically more im lure death was malignar cancer of all types lirs nilicamly in ihc period IK deaths m excess of . appears to he no parlict began it' exceed expect; the fact (hat the men w age in increasin'' numlx lasttlahan workers ia Belfast. .1. Mortuitty /WfJ*W 235 ST0095376 passed, ihc increase being steepest m those who hail heen in the industry tonnes) ami who were therefore older, there is no evidence of a synchronous change in the direction of the curves compatible with a mark ed change in working conditions. lung lihrosis (Huchanan, 1965), the commonest malignant cause of death was lung cancer. The precise numbers cannot he given because of diflicullicx in retrospective diagnosis. There were 2H eases classified as cancers of Ihc larynx, lung or pleura of whom four were probably mesotheliomas. The DKcussk* excess of respiratory neoplasms hegan to appear in Ihc follow-up of 165 men out of a group of 170 men over Ihc age of 45 and increased in frequency from 1940 to (9P6 has shown a high premature with age. Making every possible allowance for other mortality compared with the population of Northern factors the mortality was 7K (or 76 if the registered Ireland. This mortality is set far in excess of expect cause of death is used) instead of the prcdiclcd 4 64. ation that it cannot he accounted for by other know n i.c.. nearly 6 (5! I limes that expected (Table I7l. The factors such as smoking, domicile or social class. The findings suggest that men developing lihrosis are excess of deaths is thought to he due to some factor more likely lo die with primary hronchial cancer related to their work. Similar mortality has been whereas mesotheliomas may occur with minimal or described by Hammond a at. (1965) in insulation no evidence of ashestosis. There were too few non- workers in New York where a smaller proportion of smokers to confirm the suggestion thal lung cancer the work is related to ships. The pattern also seldom, if ever, occurs in non-smokers iSclikoff. resemblesthalxhowninworkcrscmplnycd in asbestos t'hurg. Hammond, and KafTcnhurgh. 196#). factories before the introduction of protective Mesotheliomas appear as a relatively frequent measures and who have survived into Ihc present fatal tumour. Although only one case was recorded era (Knox yt al., 1965; New house. 1969). The on the death certificate, three other pleural and three excessive mortality appears to he due to the inhal peritoneal cases were diagnosed in retrospect and ation of asbestos dust. There is no quantitative there may have been others among thiise in whom information about the amount of dust exposure these the information was insullicicnt. This is similar lo insulation workers received other than hv inference the experience of Newhousc and Wagner (19691. I rum the severity of their occupational disease. They Although the proportion of bronchial carcinoma to .re known to have been exposed to three of the main pleural mesothelioma may He similar to the I ondon cinunercial types, namely, chrysotilc. anioxite. and study, the proportion of peritoneal mesotheliomas is crocidolitc. They arc unlikely to have had much lower. Ascertainment of peritoneal mesotheliomas jx exposure to anthophyllite. even lower in the general population in Relfasl I lie complete pattern of deaths in these insulation workers will not he revealed until they have all died. I Ins study shows that the excess of deaths was 0 limes and Wade. 19651. Cancer or the lung and mesotheliomas do not account lor all the excess of deaths; a significant already occurringat the beginning of the period under excess of other cancers remain, as was found by study and had reached statistically significant New house (I969t. and most of these appear lo he in proportions after only 15 years. An excess of deaths thegaslromtestmal iraet (Hammond rial. I965i. due to chronic non-mherculous lung disease was As the men grow older the pattern of causes of occurung ihroughoul the period. In at least eight of death changes. In the linal analysis there may he a these cases ashestosis appears to have been a larger proportion of deaths clue lo 'other causes', contributory cause of death and lung fibrosis due to this cause may have contributed in at least II more of the 76 deaths in this category. This incidence of latal lung lihrosis is lower than that encountered in particularly lo cardiovascular disease. The excess deaths due to lihrolic lung lesions and lo cancers of the lung and gastrointestinal iracl may he less frequent and mesotheliomas ipartly because of certain factory groups before precautions were taken necropsy and therefore more precise diagnosis! may but higher than that predicted for workers employeu appear more frequent ll will heuf interest in see how m lactones only after precautions were taken many of tins group of men lire beyond the age of 70. i kuos yi al.. I'lhS; Smitlicr. t`)6S, The Merewelher and I'rice report was issued in Numerically more important as a cause of prema 195(1 and Hie regulations for factory asbestos workers ture death was malignant disease. The deaths due to came inlo force in I9.VV If the insulation workers had cancer of all types lirsl exceeded expectation sig been followed prospectively even from 1940 their nificantly in the period 1950-4 and altogether about lia/ard would have been recognized by 1955. tx deaths in excess of expectation occurred. There appears to be no particular reason why these deaths is gau io eviui'd t'v|<v> tiiiiou at llilx lam' otliei Hunt the lad that the men wete reaching the susceptible nee m increasing nuniheis 1 lux imcstigaliiin wav made possible by die provision of I mills It mu iIh III noil I iit]iio- ( iiito i I niupmi'n anil lo I hew illmg helpol die w or kern diemxebes. divii employers, then union, die si,Of of die Kegisiiar (iener.il iNIi. anil \s might be expected fnuu the high incidence of tnnnnieiable doclois and linspnal records personnel. 236 P. (. Etmrs am! if. J. (\ Simpson Statistical advice given hy Ihe Denari ment of Medical Statistics of Queen's University, Belfast. References Buchanan. W. D. I l965)..Ashestosis and primary inlrathoracic ncopla>ms. Ann. V V. Atad. An'., 132, 507-5UL I Imes. P. C. (1966: *che epidemiology and clinical features of axhcslosts and iclalcd diseases. Posttrad. mcd. J., 42, 625-6.15. and Nk ade. O. L. (1965). Relationship between exposure to asbestos and pleural malignancy tn Belfast. Ann. V. >. Atad 5*L, 132, 549-557. Hammond, L. C\, SelikofT, I. J.. and Churg. J. (1965). Neoplasia among insulation workers in the United States with special reference to intra-abdominal neoplasia. An*. V I Atad An., 132, 519-525. Knox. J I.. Hull, R. S.. and Hill. I. D. (19651. Cohort analysts of changes in incidence of bronchial carcinoma in a textile asbestos factors. Ann. .V. y. Atad. Sa . IJ2, 526-555. I.anglands. Jean II. M . Wallace. W. F. M., and Simpson, Marion J. C. (1971). Insulalion Workers in Belfusl. 2. Morbidity in men still at work. Bnt.J. mdustr. Med2*, 217-225. Merewether, l:. R. A., and Price. C. W. (1950). Report on ( fleet* of Asbestos Dust on the l ungs and Dust *uprcswon in the Asbestos Industry. tl.M.S.O., London. Newhouse. M. 1. < 1969). A studs of the mortality of workers in an asbestos factory. Brit. J. mdtnfr. Met/.. 26, 294-5W , and Wagner. J. C (1969). Validation of death cemlicalcs in jsbeslos workers Brit. J. tndnstr. Med.. 26, 502-107. Park. A. T. (1965-6). Occupational mortality in N. Ireland (1960-2). J. Statni. and Social b.nyuirv Society of trrlami. 21, Pi. IV. 24-42. Sclikolf. I. J., Churg. J., Hammond. I C . and Kairenhurgh. J. (1968). Prospective studies of asbestos neoplasia in asbestos workers. Paper 72 Proc. 2nd tnt Conf. Biological I fleets of Asbestos. Dresden. To be published. Smither. W. . (1965). Secular changes in ashcsti>sis in an asbestos factors. Ann. V. )\ Atad. .Vi., 132, 166-IKI. Wallace. W. I. M. and I anglumlv, Jean H. M. <1971) Insulation workers in Belfast. I. Comparison of a random sample with a control population Brit. J. imitntr Med . 28, 2H-2I6 Wickcn. A. J. (1966). I nv nmcntul and Personal Factors in l-i/ng Cancer and Bronchitis Mortality in Northern Ireland 1960-62. Tobacco Research Council Research Paper No. 9. London. Received for publication July 21. 197(1. Brit. J indt* Nutrr on ar 1. The K.J.COl MRC Envir London Sc London W. Esso Hous< OO O O UD cn co roiiN 2.17-: win (S.S. them to ex; paper the ei Th. hulh. Pcrsi. Ihron a prerecori radial hccan skillet Tlx and I' tempt if adv. weree. a sma Globe