Document pyvOy4VrZpKR9R88kokk1xNB

 As Ah a i i l B i t ? - - .... ,,........................................................................ ir - , , ,, 1-v-.i frrtft-rtafirrin r 'f riTMr<tf-*fnifT rntff'---*- v - y - T iff: PREPARATIO N ' OF X OX PARO GEN IC INFUSIO N AND O T H E R IN T R A V E N O U S F L U ID S 15Y A D S O R P T IV E F IL T R A T IO N * R E PO R T O F 4 i M O N TH S' TR IA L C o T (- 1, M .D ., a n d A. M . W r ig h t , M .I). N ew York, N . Y. F R O * T H E LA BO RA TO RY O ' T.XJ E R IV E S T a L -V iO - lty AXO T H E D E PA R T V E X T O F S U ltO E E Y , N E W Y O R R i N : u f 2 y : Y c o l l e o e o f m e j ic ix e . n e w y o r i^ n . y . P a r t I -- B iologic A sp ec t s Cj ? R eprinted from Annals of Surgert, September, 1M 2 0 7 South Sixth Street, Philadelphia, P enn a. VO. l i t SEPTEMBER, 1 A N N A L S of SURGERY A MONTHLY REVIEW of surgical science and practice O f fist Pxisifsfiia t f i t J * 2n V n JjwAW-*>n LEWIS STEPHEN PILCHER. JLDu LUX tv^wu on cnn Kmnil nmaxnrrr iw t EiMTvmivi B o o a WALTER E- LEE, 1U>. C>MAS A R X E Y P * 0 0 K 5^ JUX CtORCE P-^ V U L int. ROL EDW. IX CiflRCHJLI. MJX KTfcTV'.' Uu STARSTTS.-AAS GRAHAM. XUX *A SA34LXL C BARYTY. JUX ROYrcDra.cttX cC L l'K Ek.ws1UX HOTvUaU\ >f tCAVJXoFTZ1CF.Ra HJX IULIA> R FHE3C5TFR. JUX .M Y R kFv.*RwITmVHCVFT. J*x..AUIX ALLEXNT Ti\OMT S m t tJUX )M8 T tn h ic a ii. m a . T. H tm ti. l. m o l u*kun crv^ i.-vRt J". B. L I P P I N C O T T CkdOkTMUkPi. A N Y . P I B L I S H E R S C op y rig h t 1^4^ b y J . R. L jp p:ncott C ompany. South Sixth S*. F'HnrVu^iVrffl- i - . , . r^ : k^ rtyiUiri I i G O i I G OO O ph TH E PREPARATIO N OF N O N PYRO G EXIC INFUSIO N AND O TH ER INTRAVENO US FLU ID S B Y AD SO RPTIVE FILTRATIO N * REPORT o y 4 2 m o n t h s ' t r i a l C o T ut, M .D ., an d A . M . W r ig h t , M .D . N'tw Vo*iL, X . V. ROM THE LABORATORY OF EXPERT MENTAL SURGE* T AND THE DEPARTMENT OF SURGERY. N EW V O ** U N IV E R SIT Y CO CILiG E OF M E D IC IN E , N E W Y O R E , X . Y. P art I --B iolo gic A spec ts l x 1937. there was reported from this laboratory a method of preparing intravenous fluids based on a new principle.1- - A s pointed out in that report, it was S ieb ertV -4 extension 01 Hort and PenioldV' work which laid the foundation for the various methods of preparing intravenous fluids at pres ent practiced in commercial houses and hospitals. Even the method described by Rademaker15- 7 and the more recent and mechanically ingenious elabora tion by W a lt e r .* - 10 stem from Siebert's original principle. The principle, however, suffers from a critical weakness in the assumption .that if the water used in the manufacture of the fluids is pyrogen-free, then the resulting m ix ture must also be pyrogen-free. But should the water become contaminated, say as a result of failure to clean the still, or should pvrogen be present in the chemicals used in making the solutions, the lack of provision for its removal is a notable disadvantage of the method. It is. in fact, the lack of security behind the old working principle which impels manufacturers to test their intravenous fluids on the rabbit. Unfortunately, not even this precaution has entirely eliminated the occurrence of pyrogenic reactions. A s pointed out in a previous communication.11 this failure is perhaps due to the inade quacy of the test doses in use. That chemicals occasionally are contaminated with pvrogen has not here tofore entered into consideration. In previous papers1- we have reported on the pvrogenicity of samples of inulin as they have come fresh from the manufacturer. Table I shows the effect on test animals of samples ot sodium chloride, of dextrose, and of sodium citrate, all " commercially" and " technicallv pure." which had been made into solutions with nonpvrogenic water and tested immediately after the containers had been opened. The sodium chloride sample was the more reactive of two found pyrogenic in a test of 38 samples, an incidence of slightly more than 7 per cent. The dextrose was the most reactive of four positive samples of 32 tested, an incidence of 14 per cent, and the sodium citrate was the most reactive of four of 18 tested samples, an incidence of over 22 per cent. It may also be mentioned that sodium citrate * Aided by the Departments ot Hospitals and Purchase of the City of Xew York, X . Y , and a grant from the Xew York Foundation. MINTS IS T. i A. 412 ` HP m m m rn m m m m m m m ~-'F t ; N\ utAuiiittietfr I3K PREPARATION OF INFUSION FLUIDS is usually more heavily contaminated than dextrose which, in turn, is usually more heavily contaminated than sodium chloride. Because of the fact that test animals injected with such large doses of sodium citrate are liable to die in cardiac arrest, perhaps a-, a result of the blood calcium being thrown out oi ionization, it is necessary to add an equivalent dose of calcium chloride to the sodium citrate test solution before it is injected. A T able I EFFECT OF VARIOUS COHWERCIAL SOLUTION'S T est Dose G a W eight Tem p. Rise E x per. K o. Chem icals t Sodium ch lo rid e.. . . Vol. Cc. .4 A nim al Used R abbit K Z. s Deg. F. 1.2 50 2 Sodium chloride . . . 12 Dog IS 2 .0 1500 3 D e x tr o s e ................. . . 7.5 R abbit 3 2.0 150 4 D e x tro s e ................. . . 75 Dog 15 2 .4 1500 3 Sodium c itra te * ..... 0 .2 8 R abbit 3 2.2 Leukocyte Count C h an g e x 1000 -- 5.7 -- 9.4 6 Sodium citrate* . . . . 2 .8 Dog 15 2 .8 -- 8.2 115 T hese dosages w ere calculated on th e bases th a t th e m axim um dose of sodium citrate given to a 60-Kg. a n m is 5.6 G o . (3 lite rs o f b lo o d ), a n d t h a t th e r a b b it is a p p r o x im a te ly th r e e tim e s a n d th e d o g s-r tim e s less sensitive to pyrogen than m an. An equivalent am ount of nonpyrogenic calcium chloride w as added to each odium citrate solution. The chemicals in these cases were perhaps contaminated by pyrogen in the process of manufacture. Nonpyrogenic chemicals even in the " dry" state can, however, become pyrogenic on standing under unsterile conditions, as has been found by Smith13 to be the case with inulin. and by us with dex trose, sodium citrate and heparin. Apparently the better medium the chem ical is for bacterial growth, the more easily it becomes pyrogenic. Humidity and a warm environmental temperature seem to accelerate pyrogen produc tion, probably by acceleration of bacterial growth. Among aerial contami nants, Co Tui and Schrift14' 9 have found B . protcus vulgaris and B . subtilis fairly prolific pyrogen producers. In the light of these findings, therefore, the assumption that to guard against the occurrence of pyrogenic reactions it is necessary only to keep the water pyrogen-free can be costly as well as unsafe. The problem will perhaps become increasingly acute with the increasing use of blood and plasma transfusions. The adsorptive filtration method of preparing intravenous fluids differs from the older method in that it is based on the principle of filtering the resulting mixture of chemicals and distilled water through specially prepared compressed asbestos pads proved to remove pyrogen. This method is effi cacious not only for infusion fluids but for all crystalloidal intravenous medica ments. With colloidal substances such as sera and gum acacia, however, it is ineffective. The method has been in experimental use on the New York University Surgical Division of the Bellevue Hospital continuously since July, 1937 except ior an intermission of 16 months. I: is the purpose of the present communication to summarize the results of the experiment as well as to describe the practical method of manufacture which has been evolved. 0 0i 0! 1 0! 1 i ) i0 1 -S 0 r -Nt "w* 0 a ' 413 II o c c CO TCI AND WRIGHT ****>' <* c The practical difficulties encountered in the course of this work may be briefly discussed rince they are difficulties which many hospitals and similar institutions which undertake the manufacture of their own infusion flu id s c will have to face. 1. The first problem was the establi-hment of the criterion of pyrogenidty. With fluids which are frankly pyrogenic, this is an easy matter since the c symptom-complex of the pyrogenic reaction is a familiar one. The difficulty was with fluids containing a borderline dose of pyrogen, which would elicit only a slight temperature rise easily mis-ed in the four-hour routine ward c temperature readings but might, when administered in doses two or three times larger than the usual one. give rise to a fairly severe reaction. It was. therefore, necessary during the first six months of this experi mental period to subject the fluids made to animal tests, using the rabbit and the dog. according to the technic described elsewhere.11 If the fluids were 0 found nonpyrogenic according to the tests, they were then released for ward use. The temperature readings of all patients receiving the fluid were taken hourly for four hours, both during and after the administration of the fluid. 0 In this way transient temperature rises could be detected. 2. The second problem was the elimination of contamination from exter nal sources, such as the rubber tubings, glass connections and needles used in the dispensing set. Originally, the latter were sterilized in the ward instrument sterilizers in common with contaminated pus basins and surgical instruments. 0 i Q The occurrence of four reactions on patients in the first three weeks of this period led to tests of the water from the sterilizer. Ten and 15 cc. samples of this water gave unmistakable reactions in both rabbits and dogs. This result led to the institution of a method of washing, packing and sterilizing the dis pensing sets which will be described in a later section of this paper. 3. A third difficulty was the presence of a yellowish-brown coloration in the solutions after they had been autoclaved. This was first thought to be due to caramelization but has been attributed by K m o 15 to the production of c levulinic acid.10 In our preparation two factors were found responsible for the coloration: <a ) The asbestos pads originally used were made for the filtration of fluids which did not have to be autoclaved. They imparted some principle to the filtrate which caused the formation of this coloration on heat ing. The adoption of a pad specially prepared for infusion fluid use has elim inated this factor; and (b i the second iactor was overheating of the solution. This occurred when, after sterilization, the autoclave pressure was permitted to sink so fast that the fluid inside the flasks began to boil over. The film of fluid on the inner wall of the neck of the flasks evaporated and the residue became browned, imparting the coloration to the rest of the solution. The remedy to this was obviously to lower the autoclave pressure so slowly that no boiling over took place. 4. The fourth and most serious difficulty was the presence of shreds and particles in the filtrate. While the injection of solutions containing these par ticles into numerous experimental animals caused no detectable symptoms, \ ' 414 t ) Vijumt Hi Nua3 PREPARATION' OF INF ESI OX FLUIDS their presence was none the les5 objectionable. ]t was becau-e of this difficulty > that the use of these fluids was -uspended tor i O months in 1938 and the firs: part of 1939, until a manufacturer could be found who would cooperate in devising means of eliminating these particle^. Jena filter-. Chamberland and ) lierkefeld filters were interposed at the outlet of the asbestos filter without avail. Finally, the difficulty wa- overcome by W h improving the quality of ) the filter pad and interposing an aloxite fiber eliminator between the asbesto- filter and the collecting bottle. v 5. Standardization of filter pads: The compressed asbestos pads, as now . > prepared for our use by the E n el Engineering Corp.. are calibrated with regard to two properties: <a> The speed of filtration; and ( b) degree 01- > pyrogen retentiveness. J y Speed of Filtration.-- This is controlled by three factors: The filtering pressure applied: the filtering are a : and the composition of the pad. The filtering pressure is sajest at 12 lbs., and the filtering area is limited by the y size of filter practicable in hospitals. Our present filter takes pads of 12 inches, with an actual filtering area of 29.5 cm. in diameter. The pads are made of asbestos and pulp. The more asbestos used, the greater the pyrogen ) retentiveness, hut the slower the filtration rate. The interplay of these two factors, therefore, determines not only the rate of filtration but also the efficiency of the pads to remove pyrogen. The pad now in use has a delivery volume of two liters every 2.5 minutes. A six-hour filtration would, thus, yield 288 liters, which would be more than enough to meet the daily require ments of a fairly large hospital. Hospitals using larger amounts oi fluid may X J 0 enlarge the filtering area of the filter by the method to be described later or install two such filters. Pyrogen Retentivencss of the Filter Pads.-- The ability oi these pads to 0 ' remove pyrogen was shown in a previous work16 to be due to adsorption and . not to sieving. This means that the property can be exhausted. It is. there fore. necessary so to standardize the pads that they are not only uniform but y also possess such a margin of safety in pyrogen retentiveness that they remove the maximum amount of pyrogen which can occur in the distilled waters and the chemicals in use. y In order to calibrate the pads in this respect it was necessary to secure a sample of pyrogen and use it as a standard. The pyrogen used in this work was isolated from the cell-free Berkeield filtrate of the supernatant portion y of concentrated stock suspen ;on of typhoid vaccine ( 1 cc. == 88 billion c e lls!. From seven liters of this suspension, four grams of the substance was yielded. The method of isolation as well as the chemical and biologic aspects of this y substance will be reported elsewhere. It is necessary, however, to state that the substance is still in the crude stage, containing six per cent nitrogen, which on further purification has been reduced to 1.6 per cent. The biologic 0 (pyrogenic) potency of this crude sample, called "pyrogen-A ." has been determined on rabbits, dogs, and men. Table II shows the effect of different doses oi this preparation. It the minimum effective dose (M E D ) is arbi- 0 o4IS 0 ccc cc c Ii .0 0 0 o CO TUI .v .r o W RIG H T V ^'T'lrry ~^d*-SWb*T. 2i t : irari'v defined a- that do-e. / r kilo I* Ey--/.eight. which on intravenous in jection will can-o- a temperature rbe of 0.81 :o 1 . 0 the M E D in rabbits is 0.1 gamma; in dog-. 0.4 gamma; and 0.03 gamma in men. The reason for using this crude substance as a standard ir.-tead of the more purified sub stance is that the process of purification is -o tedious that it may be a long time before enough of the purer substance can be available for this purpose. In any case, the biologic calibration o: "jn n -gen-A" makes it unnecessary- to wait for the purer preparation. T a BLF II EFFI.C1 OF DOSFAs OF 1*YROGEV-"a ' IS ABBITS. OOGS. AKD MEN W eight Tfem p. C h an g e S u b je c t iv'2 - D g* p e r K g . Deg F. R abbit 1. . .. R a b b it 2 ............ 0.12 7 ; 0.12 0 . 1.0 R a b b it 3 ............ ............................................................... 3 0 0.12 0.9 R a b b it 4 ............ ............................................................... 3 . 0 2.0 3.5 D o g 1.................. ............................................................... t ? . o 03 1.0 D og 2 ................ ............................................................. 1 6 .0 0.3 1.1 D o g 3 ................ ............................................................... 15.5 0.3 0.8 D og 4 .............. ............................................................. 1 6 .0 2.0 3.2 M a n 1 ................. ............................................................... 5 2 . 0 0.04 1.2 M a n 2 ................. ............................................................... 6 4 . 0 0.04 0.9 M a n 3 ................. ............................................................... 7 2 . 0 0.04 1.1 M a n 4 ................. ............................................................... 70 .5 2.0 5.0 Leukocyte C ount C hange x 1000 --4 . 1 -- 5.5 -- 10.2 Process of Calibrating the Pads.-- One milligram of `'pyrogen-A'' (1000 gammas) is weighed out on a sensitive analytic balance and dissolved in 100 cc. of pyrogen-tree distilled water. Ten cubic centimeters ot this solution contains 10 gammas of pyrogen. Amounts of from five cubic centimeters up. in multiples of five, are filtered through test pads of a convenient size mounted in convenient filters. The filtrates are then tested in both rabbits and dogs. Table I I I shows the protocols of some of these tests. E xperim ental W eight A nnual Kg- R a b b it t ...................... .......... 2 .5 D o g 1............................ .......... 1 6 .0 R a b b it 2 ..................... .......... 3 .0 D og 2 ............................ .......... 1 5 .0 R a b b it 3 ...................... .......... 3 .0 D o 3 ............................ l> > R a b b it 4 ...................... .......... 2 .5 D o g 4 ............................ .......... 1 6 .5 R a b b it 5 ...................... .......... 3 .0 D og 5 .......................... T abve III CAUBRAT1PS OF FH.TEX PM'S Volum e T .x a l D-<se P y ro g e n P re tiltra te m P rei-'trate C c. (7<?nc.?ea 30 300 30 300 35 350 35 350 40 403 40 4CO 45 *50 45 *50 50 5CO 50 500 Temp. Change Deg. F. -- .2 -F-1 -r.2 .25 ~ r. 15 -f-1 -r.4 -f .2 -r.4 + .8 Leukocyte Count C han g e x 1000 -- .2 -- .3 + .i -- .5 -- 1.2 It will be seen from Table I I I that the rad removes nearly 500 gammas of " pyrogen-A." Since the pad is 2.5 cm. in diameter, it has a surface of 4.9 sq. cm., and. therefore, each square centimeter removes over 100 gammas of `pyrogen-A." The filtering surface of the four large filter pads used in the hospital filter assembly are each 29.5 cm. in diameter, making a total filtering area of 2734 sq. an . The.amount of pyrogen removable by the entire filter 416 1 Vtfyv* 1)4 N au .^ 2 PREPARATION' OT IN F U S IO N FLUIDS assembly iN approximately 273.400 (jatnums. Since about 140 yatnmas are required to cau-^e a temperature rise of 5 / / F . in an average man of 70 kilos, this filtering area will remove about 1953 such doses. When applied to solutions of sodium chloride or dextrose in distilled water this range of safety is more than ample as the amount of pyrogen contained in any combination of these three substances in the usual dose is tenuous and. further, the filter-pad assembly is made to deliver not more than 288 liters with each filtration. The case of sodium citrate requires special consideration. A s shown in Table I. gram for gram, the contamination of pyrogen in sodium citrate is heavier. This, however, is more than offset by the fact that the dose of -odium citrate used per 500 cc. of blood is usually in the neighborhood of 1S75 mg. in 75 cc. of water. This, by weight, is only about one-fifth of the dose of sodium chloride in one liter of physiologic salt solution and less than one-:\\enty-seven:h of the dose of one liter of 5 per cent dextrose solution. Therefore, in making sodium citrate solutions, it is only necessary to use two filter pads and restrict the manufacture to not more than 300 doses. These calculations are admittedly based on biologic effects, and assume that pyrogen from sources other than the typhoid bacilli have the same potency per unit-weight as typhoid pyrogen. This assumption has no direct experi mental basis at present. The only justification for applying the figures derived from `"pyrogen-A" to the pyrogenic substances found in these fluids, whose exact source is unknown, is that they have so far held true in the samples whose pyrogenic content was roughly determined by the biologic test, then filtered through asbestos pads of the required area and then retested and found negative for pyrogen. In the absence of more exact knowledge, this approximate correspondence for the present can be used as a rough working principle. This method of pretesting before filtration may be used for any intravenous material whose maximum pyrogenic content has not been as thoroughly studied as have those of dextrose, sodium chloride, and sodium citrate. R E SU L T S O F 4 5 M O N TH S' TR IA L During this experimental period. 3S67 liters of intravenous fluids pre pared by the nonpyrogenic method have been administered. They may be tabulated as follows: No. of Administrations 1815 652 125 64 Amount Given Liters 1 2 5 4 Total Given Liters 1*15 1504 492 256 Totals 2654 5667 titers In none of these was there a reaction. One patient who complained of a chilly sensation after receiving toco cc. did not have a temperature rise. It 417 C T CO T U I A XL) W RIGH T Aroeu'.t ni &m\w7 ii \z is clear that if any pyrogen was present in the solutions, it was in such minute 0 amounts as to be below the M E D even in three to four liters. P art I I -- M ech an ical A spects 0 The parts which comprise the entire filter assembly may now be enumer ated and the more important items described. The filtering portion of the assembly is diagrammatically represented in Figure i. It consists of a com 0 pressed nitrogen or air tank. A . which iurnbhes filtration pressure: the reser voir, B. which holds the prefiltered m ixture; the filter itself, C ; the aloxite fiber eliminator. D : and the collecting flask, E. c 0 0 0 0 0 0 o lO F ig . i . The reservoir, usuallv of 250 liter capacity, is made of stainless steel, and may be equipped with an electric stirrer. Figure 2 illustrates the essential parts of the filter as well as the path taken by the solution passing through it. It may be seen that each filter pad is held between two perforated Monel metal plates or frames, the three form ing a pad-frame unit. Two such units form a compartment. The fluid flows from the solution inlet. A . into a conduit system. B. which splits it into two streams. One stream filters through one pad downward and the other upward to meet in the collecting chamber. C. situated between the two pad-frames. From C the solution is conducted into the second compartment to liave the process repeated, after which the conjoint stream front the second collecting chamber, D. passes through the fiber eliminator into the collecting flask. Both the latter are to Ite descril>ed in a subsequent section. * 41S 1 Volw** 11 dumber 3 PREPARATION OF INFUSION FLUIDS Since provision i- made in ;he present filter to accommodate two addi tional compartment', the filter has a still wider range of u-efulnes^ than has been calculated. Any one of the three factors-- capacity, speed of filtration, or retentiver.e---- can be increased two-fold by this doubling of the filtration area. The stream in this case would be split in four instead of two as in the filter illustrated in Figure 2. Figure 3 is a photograph of the filter ensemble. A is the spindle with which the pads are loaded and unloaded and the filtering system ren dered watertight. B is the filtering system. C is the pressure gauge. D the housing for the aloxite fiber eliminator, and E the outlet. The self-autoclav- F ig . 2. ing system, F . consists o f a steam generator in which the source of steam used is pyrogen-free distilled water. It may be heated electrically, by gas. by housesteam. or by an oil burner. The steam generated is led into the interior of the filter by the tiexible hose. G. The course of the steam through the filter is identical with that of the fluid which follows it and whose course has been described in the foregoing paragraph. The aloxite fiber eliminator is a thick, finely porous filter made of car borundum. Because of its composition it is sturdy and may be heated to a temperature high enough to destroy any pyrogen which may have developed in the interstices of the stone. The dispensing pan of this system, mainly the collecting and storing flasks and accessories, is essentially the same as that described bv W alter/ s- 10 It is the most satisfactoix- one available at present, made to withstand hard 419 CO T U I A.b'D W K IG H T ,4nr-J 4-,*r*'rT u * ~ S r <.+~ 5,4; u-age, and in three year-* oi u-e ha.* -uttered the minimum amount of break age. It ha*, however, the -light inconvenience of having too many parts. Figure 4 illustrates the part of thi- ^y-tem which cor.-ists o f: A . pyrex flasks of specially rugged construction of one to two liters capacity, preferable graduated: B. a one-hole rubber bu-hmg to fit the fla-k : C. a special -tain'.e-s steel stopper with a partially grooved -tern; D. a two-wav vent tube for the administration of the fluids and tor the adntis-ion of air into the flasks to dis place the fluids. F ig. 3. The bushing. B. is oi special rubber which can stand 500 sterilizations without shrinking. It is provided with a hole in the center for the insertion of the two-way vent tube when fluid is being administered, and for the inser tion of the steel stopper when the flask is not in use. The stopper is provided with a partially grooved stem which is fitted partly into the hole of the bush ing while sterilization takes place. The groove allows equalization of inside pressure with outside pressure. When sterilization is completed, the stent of the stopper is pushed all the wav in. as in E . the umbrella of the stopper serving as an hermetic seal for the rubber bushing. The original Fenwal-tube is short, so that the flask end is submerged in the fluid when the flask is inverted for administration. There are two obiec- ' 420 V otes* 115 Nat*r Z PREPARATION' OF INFUSION FLUIDS tions to thi3 feature: Fir^t, when the infusion is being given, the air which enters the fla-k to displace the fluid has to be bubbled through the fluid. This, in effect, means the washing of 1000 to 2000 cc. of unsterile air in the infusion fluid. Bacteriologic cultures of ten samples of fluid subjected to thi.~ kind of air-washing gave three positive cultures. By the simple device of r !, f bc~\ lb c c o ) t lengthening the tube so that the flask end extends above the surface of the fluid when the flask is inverted, this washing effect is avoided. X o positive cultures were found in ten fluids subjected to this system of air displacement. If it is desired to admit absolutely sterile air into the flask, the air vent may be modified in such a way as to accommodate a small sterile Berkefeld candle. This was originally used by us but later discontinued because of inconvenience. 421 11 X d $ 0> 0 0 CO TUI AMD 'ARIGHT Ar_j*lt ft ? * ir i*77 The second objection to the -bort original Femval-vent-tube is that unless sjKrcial care i- taken, the fluid i- as liable to flow through the air vent as through the fluid vent. The split ring bracket. F. i\ simply a wire basket for suspending the flask in an inverted po-ition when the flask is in u-e. Some reference must be made to the still which delivers distilled water for making the solutions. Since the still is the heart of all other processes, much emphasis has been laid or it. While it is our opinion that the undue concentration on the still has lulled to re.-t suspicion of other possible sources of pyrogen, we believe that the still must be cleaned periodically so that it adds no pyrogen to the distilled water to lower the factor of safety of the filter. Moreover, it must be large enough to deliver the necessary amount of distilled water to be used for clean-ing as well as to make the solutions themselves. A still that delivers io gallons of water per hour <40 liters), will be sufficient to meet the needs of the average hospital. The periodic cleaning also does away with incrustrations which lower the output of the still. CARE O F TH E PARTS The R eservoir: The reservoir when new must be rinsed first with soap and water and then copiously with distilled water. Subsequent care after every use consists in rinsing it out thoroughly with distilled water and then drying it. The F ilte r: When new. the conduit system of the filter is rinsed thor oughly with distilled water. Thereafter, the same rinsing process is repeated after each use. The A loxite Fiber Elim inator: After every use. the surface of the stone fiber eliminator is simply flushed thoroughly in order to wash away the par ticles and shreds which may be caught on its surface. A fter washing, it is placed in an electric oven at a temperature of 120 C. for one hour to dry. This temperature sterilizes the stone so that pyrogen production within the interstices of the stone is halted. Fcn tivl-Flasks. etc.: The initial cleaning and subsequent care of the Fenwal-flasks. rubber bushings, steel cages, and the modified vent tube have been described by W alter and need no extended treatment here. Suffice it to say. that the flasks are washed and brushed with hot soapsuds and rinsed first in tap and then in distilled water until the walls run homogeneously clear. The same treatment should he applied to the other glassware, namely, the vent tube and the drip as well as to the steel stopper. An occasional treatment of the glassware with a cleaning mixture of 1.5 per cent solution of sodium dichroniate ir concentrated sulfuric acid may have to be resorted to in order to get rid of spots in the glassware difficult to eliminate other wise. The rubber bushings when new are first scrubbed with soap and water until no further powdery substance is given ott. then rinsed with 422  - i l I I rr PREPARATIO N OF INFUSIO N FLU ID S distilled water. The rubber tubings when new are likewise rinsed well with hot soapsuds and then with distilled water. Immediately after each washing, the rubber tubings, the vent tubes, the drip and the needles are wrapped and packed in a convenient container. W e have used for this purpose a specially constructed copper cylinder, one toot long and 2 1 j inches in diameter, pro vided with a detachable cap as well as holes in the wall to allow penetration of the steam in the autoclave. After sterilization, the holes are closed by a special sliding device and the set stored for future use. Preparation for Filtration: The flasks without the rubber bushings in place are sterilized in the hospital autoclave. The rubber bushings and the steel stoppers are also sterilized but separately from the flasks. The filter with pads mounted on the frames is sterilized by its own self autoclave for 30 minutes at iS pounds pressure. A fter that it is allowed to cool to a temperature convenient for handling. The solution is mixed in the reservoir. The reservoir is then connected, on the one hand, with the nitrogen tank and. on the other hand, with the inlet of the filter. The collecting and storing flask is placed at the outflow tube of the filter and filtration is begun by turning on the gas gradually to 12 pounds. A fter discarding the first four liters, the flasks are then filled one by one. A fter collection, the flasks are partially stoppered with the stainless steel stopper in such a way that the groove in the stem of the stopper allows com munication of the inside of the flask with the outside (F ig . 4 ). The flasks are then put into the autoclave and sterilized at 250 F . for 30 minutes. After sterilization, special care is taken to have the pressure in the autoclave grad ually diminished so as not to cause boiling. The stem of the steel stopper may then be pushed home, in which case a partial vacuum is created as the solution cools. This partial vacuum, which helps retain the stopper in place and causes a water-hammer effect when the bottom of the flask is struck with the fist, may be used to inaicate whether the stopper is leaking after storage. After the flask is sealed, it is stored until needed. Tust before administering the fluid, the steel stopper is removed from the flask with a rocking and un-crewing miA->n. The flask end of the two way vent tube is now inserted through the hole in the rubber bushing to the point where one of the two filiations on the glass tube just slips into the flask beyond the rubber bushing. The flask is inverted, placed in the ring bracket, and the infusion started. Calculation of Costs: This method requires an initial equipment which may be divided into two pans, an irreducible item which con-ists of the reservoir and the filter, and an item which can be expanded according to need, namely, the flasks and their accessories. The rubber tubings and drip and needles do not enter into consideration here as they will have to be used in any case. The irreducible equipment co-t-> about Stooo. If the cost of the fia-ki for initial u.-,e is -f:Z at S600 and to this ri added part-time labor of $5 o a month, which can be expanded into full-time labor a- the need arises. 423 30 o0 oo o ,0 o [o i o0 o0 o0 qoo oo0 o0 o*0 o o0 o D0 Jo o0 CO TUI AND WRIGHT A'itiilt </ Hurgtrr the total may be set at S2200. Any hospital which expends this amount of money a year for inru-ion fluids will not incur a loss. even in the first year, bv adopting' this sy-tem. Beyond the fir-t year the -avings are substantial; the larger the amount of fluid.-, u=ed. the more substantial the savings, since the onlv items of expense v.ill he the com of replacement-, material and labor. But the saving- are r.ot the entire con-ideration. The safety from pyro genic reactions in intravenous medication, especially in blood and plasma banks, should be the major consideration, particularly during w ar times. Instances are known of tens of liters of blood and plasma which have had to 1*- discarded because the sodium citrate u-ed was pyrogenic. This type of saving has not entered into the above calculation but may be the major item. SU M M A RY A ND C O N C LU SIO N S 1. Methods of intravenous fluid manufacture which depend only upon the purity of distilled water used have no provision for dealing with pyrogen should this substance be present in the final mixture and, therefore, do not entirely eliminate pyrogenic reactions. 2. Chemicals have been shown to be occasionally contaminated with pyrogen. 3. The adsorptive filtration method of preparing assuredly nonpyrogenic fluids, described five t ears ago. has been developed into a practicable and convenient method for hospital use. 4. The compressed asbestos filter pads, around which this method is built, have been calibrated for speed of filtration and pyrogen retentiveness. The factor of safety is shown to be more than ample for the most exacting use in the manufacture of physiologic salt, dextrose, and sodium citrate solutions. 5. The uniformly nonpyrogenic results of 42 months' use of fluids manu factured according to this method are reported. 6. The mechanical aspects of the method are described. 7. The method not only eliminates pyrogenic reactions but is also more economical than the use of commercial fluids. The authors wish to acknowledge their grateful thanks to Dr. Edward M. Bemeker. Commissioner oi Hospitals, to Commissioner Albert Pleydell and to Mr. A. M. Moore of the Department of Purchase. City of N'ew York, for their encouragement and co operation in furnishing the equipment for the completion of this w ork: to Dr. William F . Jacobs. Medical Superintendent oi Bellevue Hospital, for the privilege oi installing the experimental niter in the Bellevue Hospital, to Mr. Francis X. Dealy. of the Ertel Engineering Corp. of Xew York, for his cooperation in the mechanical phases of the problem, and to Dr. Wilfred Ruggiero. Messrs. Arthur Yates. Angelo Benaglia. and Milton Schriit for their constructive suggestions; to Dr. Ralph S. Muckensus. of the Bureau of Laboratories. City of Xew Y o rk ; to Lederle Laboratories. Inc., and to E. R. Squibb & Co. we are indebted to for supplies oi the stock suspension of typhoid bacilli. REFEREXCES 1 Co Tui. McCloskey. K_ I_Schriit. M ,, and Yates. A . L . : A Xew Method of Pre paring Infusion Fluids. J.A .M .A .. 109. 250. lotp. - I d e m : A Xew Method of Preparing Xonpyrogenic Intravenous Infusion Fluids: Based ' 424 V/* jZi r, PREF'AKATION OF INFUSION' FLUIDS cm Removal of Pyrogen by Ad-orptive Filtration. A nnals or S urgery. 106, 1089- IO0M, December, J <>372 Sieben, FF B , Fever-producing Substance Found in ome Distilled Waters. Am. Jour. FTiysiol. 67, 90, 1923. * Idem : Cause of many Febrile Reactions folioteine Intravenous Injections. Am. Jour. Physiol.. 71, 62 1. 1923. s Hort. E. C . and Penfold, W. J . : The Dangers of Saline Injections. Brit. Med. Jour.. 2, 1589, December 16. 19 11. * Rademaker, L. : Cause and Elimination of Reactions after Intravenous Infusions. A nnals o f St.-jfoi.Kv. 92, 195. 1930. 7 Idem : Reactions after Intravenous Infusions: Further Report on their Elimination. Surg.. Gynec. and Obstet., 56, 956. 1933. v Walter, C. \ \ . : Economical Intravenous Therapy. J.A .M .A ., 104, Jbos 1935. * Idem : Preparation of Safe Intravenous Solutions. Surg_ Gynec. and Obstet., 63, 643. 1936. 10 Idem : The Relation of Proper Preparation of Solutions for Intravenous Therapy to Febrile Reactions. A nnals of S urgery. 112, 603. 1940. 11 Co Tui and Schrift. M. H .t A Tentative Test for Pyrogen in Infusion Fluids. Proc. Soc. Exper. Biol, and Med.. 49, 320-323. 1942. 17 Co Tui, Schrift. M. H., McCloskey. K. L., and Yates. A. L. : Filtrat'cm Studies on Pyrogenic Inulin. Proc. Soc. Exper. Biol, and Med.. 36, 227-230. March. 1937. 13 Smith. H. W. : Personal communication. 14 Co Tui and Schrift. M. H.t Production of Pyrogen by some Bacteria. J . Lab. and Clin. Med.. 27, 569-375. 1942. 17 Krno : Private communication quoted by Walter.1'' ** Co Tui, McCloskey. K. L.. Schrift. M. FF., and Yates, A. L. : Filtration Studies on Reactive Infusion Fluids. Proc. Soc. Exper. Biol, and Med., 38, 297-300, 1936. WALTER E. GREULE 426 TRANSPORTATION BUDG. CINCINNATI 2 . OHiO ,g M m m y f n X/ & 9 'T H E PREPARA TIO N OF NO N PVR GENIC INFUSIO N ANI) L O TIIKli IN TRAVEN O U S FI.('ID S HV A D SO RPTIVE FIL T R A T IO N ! 1 K H l'O K T O T M m o n t h s ' T U IA 1 , C o T m , M .1L , unii A. AL W hioht, M .D . N rw Yoiii*, H. Y. , -" FROM )111. I-ATIOKAIORY OF KXFLR! MENTAL SUKGJ RY A N IM F DI PAR IMKNT OF SURGERY, N EW YORK L 'N IV L R S n y L O L IM .C OF MI 1X U N I , N l \V YORK, N Y X' 1' aht I -- liio i.o ia c A k p k t s ;' X 1' ' ` jj l x 1937, there was reportai from tins laboratory a method of preparing'-fa intravenous fluids based on a new principle h 2 A s pointed out in tjiaf report,- it was SicbertV 1,4 extension of Hurt and Tenfold's-"' work which lai<jJ {he -,$'< foundation for the various methods of preparing intravenous fluids at pr^STfd||t|f ent practiced in commercial houses and hospitals Even the method described by Eademaker'*- 7 and the more recent and mechanically ingenious elabora-^||x tiou by W alter,8, y' 10 stem from Siebcrt's original principle. The principje, A,;' , however, suffers from a critical weakness in the assumption that if the w a t e r ^ i l1 used m the manufacture of the fluids is pyrogen-free, then the resulting m v p f' ' ture must also be pyrogen-free. But should the water become contammatedi:j*Jji say as a result of failure to clean the still, or should pyrogen be present in' the chemicals used in making the solutions, the lack of provision for its removal is a notable disadvantage of the method. It is, in fact, the lack p f , security behind the old working principle which impels manufacturers to test their intravenous fluids on the rabbit. Unfortunately, not even this precaution >y>r has entirely eliminated the occurrence of pyrogenic reactions. A s pointed out in a previous communication,11 this failure is perhaps due to the made- , f. quacy of the test doses in use. That chemicals occasionally are contaminated with pyrogen has not here tofore entered into consideration. Jn previous papers12 we have reported on the pyrogenicity of samples of inulin as they have come fresh from the manufacturer. Table I shows the effect on lest animals of samples of sodium chloride, of dextrose, and of sodium citrate, all " commercially" and " technically pure," which had been made into solutions with nonpyrogenie water and tested immediately after the containers had been opened. The sodium chloride sample was the more reactive of two found pyrogenic in a test of 38 samples, an incidence of slightly more than 7 per cent. The dextrose was the most reactive of four positive samples of 32 tested, an incidence of 14 per cent, and the sodium citrate was the most reactive of four of 18 tested samples, an incidence of over 22 per cent It may also be mentioned that sodium citrate Aided by the Departments of Hospitals and Purchase of the City of New Yoik, N. Y., and a grant from the New York Foundation. 412 & i. Pkl' P\l< \TlON OF INFUSION FLUIDS 5 1 is usually mure heavily contaminated than dextiose uhich, in turn, is usually M move heavily cotiUimmuicd than bodmm chloride. Because oi the tact that lest animals injected with such lar&e doses ot sodium citrate are liable to die in cardiac arrest, perhaps as a result of the blood calcium being thrown out oi ionization, it is necessary to add an equivalent dose of calcium chloride to the sodium citrate test solution before it is injected. Tabu: I ll FUCl OF VARIOUS COMMURCIAL SOLUTIONS C bcim caL T est D ose Gill. W eight T em p Rise Leukocyte C ount Vol Ct A nnual LFed Kg Deg. F C hange x 1000 Sodium chloride 4 R abbit 3 1.2 50 Sodium chloride . 12 Dog 15 2.0 --1 7 1500 D ex tro se 7.5 R abbit 3 2 0 150 D extrose.. 75 Dog 15 2 4 --9 4 1500 Sodium citrate* 0 28 Rabbit 3 2 2 11.5 Sodium curate* 2 8 Dog 115 IS 2 8 --8 2 These dosages were calculated on th e bases th a t th e m axim um dose of sodium c itra te given to a 60-Kg man Is 5.6 Gin. (3 liters of blood), and th a t th e ra b b it is approxim ately three tim es and th e dog six tim es less sensitive to pyrogen than m an An equivalent am ount of nonpyrogeme calcium chloride was added to each odium citrate solution. The chemicals in these cases were perhaps contaminated by pyrogen in the process of manufacture. Nonpyrogenic chemicals even in the " dry" state can, however, become pyrogenic on standing under unsterile conditions, as has been found by Smith13 to be the case with inulin, and by us with dex trose, sodium citrate and heparin. Apparently the better medium the chem ical is for bacterial growth, the more easily it becomes pyrogenic. Humidity and a warm environmental temperature seem to accelerate pyrogen produc tion, probably by acceleration of bacterial growth. Among aerial contami nants, Co Tui and Sehrift11-9 have found B . protcus vulgaris and B. subtUts fairly prolific pyrogen producers. In the light of these findings, therefore, the assumption that to guard against the occurrence of pyrogenic reactions it is necessary only to keep the water pyrogen-free can be costly as well as unsafe. The problem will perhaps become increasingly acute with the increasing use of blood and plasma transfusions. The adsorptive filtration method of preparing intravenous fluids differs from the older method in that it is based on the principle of filtering the resulting mixture of chemicals and distilled water through specially prepared compressed asbestos pads proved to remove pyrogen. This method is effi cacious not only for infusion fluids but for all crystalloidal intravenous medica ments. W ith colloidal substances such as sera and gum acacia, however, it is ineffective. The method has been in experimental use on the New York University Surgical Division of the Bellevue Hospital continuously since July, 1937 except for an intermission of 16 months. It is the purpose of the present communication to summarize the results of the experiment as well as to describe the practical method of manufacture which has been evolved. 413 *' ID II' WU \\ klUIT f imi 'I he |>tactical diilu nines encounleicd m the course of this work m^y bucfly discussed since it)cy .nc (hiiiuiliies which many hospitals and milaf institutions which undertake the manufacture of their own infusion fluids will have to hue * i The first problem was the establishment of the criterion of pyrogenicity. With fluids which are frankly pyrogenic, tins is an easy matter since the symptom-complex of the pyrogenic reaction is a familiar one, The difficulty was with fluids containing a borderline close of pyrogen, which would elicit only a slight temperature rise easily missed in the four-hour routine ward temperature readings but might, when administered in doses two or three times larger than the usual one, give rise to a fairly severe reaction. 1; ft was, thetefore, necessary during the first six months of this experi mental period to subject the fluids made to animal tests, using the rabbit and the dug, according to the technic described elsewhere 11 If the fluids were found nonpyrogenic according to the tests, they were then released for ward use The temperature leadings of all patients receiving the fluid were taken hourly for four hours, both during and after the administration of the fluid, in tins way transient temperature rises could be detected. 2. The second pioblem was the elimination of contamination from exter nal sources, such as the rubber tubings, glass connections and needles used in the dispensing set Originally, the latter were sterilized in the ward instrument sterilizers in common with contaminated pus basins and surgical instruments. The occurrence of tour reactions on patients in the first three weeks of this period led to tests of the water from the sterilizer. Ten and 15 cc. samples of this water gave unmistakable reactions in both rabbits and dogs. This result led to the institution of a method of washing, packing and sterilizing the dis pensing sets which will be described in a later section of this paper. 3. A third difficulty was the presence of a yellowish-brown coloration in the solutions after they had been autoclaved This was first thought to be due to caramelization but has been attributed by K rno1& to the production of levulinic acid.10 In our preparation two factors were found responsible for the coloration: (a) The asbestos pads originally used were made for the filtration of fluids which did not have to be autoclaved. They imparted some principle to the filtrate which caused the formation of this coloration on heat ing. The adoption of a pad specially prepared for infusion fluid use has elim inated this factor; and (b) the second factor was overheating of the solution. This occurred when, after sterilization, the autoclave pressure was permitted to sink so fast that the fluid inside the flasks began to boil over. The film of fluid on the inner wall of the neck of the flasks evaporated and the residue became browned, imparting the coloration to the rest of the solution. The remedy to this was obviously to lower the autoclave pressure so slowly that 110 boiling over took place. 4. The fourth and most serious difficulty was the presence of shreds and particles in the filtrate. While the injection of solutions containing these par ticles into numerous experimental animals caused no detectable symptoms, 414 u % l' K Fi'A K VflON OF INFUSION Fl.UlD S U jlicir presence \v;h none the less objectionable l l was because o! tin-, difficult) |liat (lie use ol these lluids was suspended for IP moulhs in 193K and llic lirst part of 19,59, until a m anufacturer could be found who would cooperate in dusiug means of eliminating these particles Jen a file rs, Chamherland and lierkefeld f ile r s w ere interposed at the outlet of the asbestos filter without avail. Fin ally, the difficulty was overcome by both im proving the quality of the filter pad and interposing an aloxite filter elim inator between tile asbestos tiller and the collecting bottle. 5. Standardization of filter pads - The compressed asbestos pads, as now prepaied for our use by the Krtel Engineering Corp., are calibrated with regard to two properties: (a) The speed of filtration; and (b) degree of pyrogen retentiveuess. Speed of Filtration.-- This is controlled by three factors: The filtering piessure applied; the filtering area; and the composition of the pad. The tillering pressure is safest at 12 lbs., and the filtering area is limited by the size of filter practicable in hospitals Our present filter takes pads of 12 inches, with an actual filtering area of 29 5 cm in diameter. The pads arc made of asbestos and pulp. The more asbestos used, the greater the pyrogen retentiveness, but the slower the filtration rate. The interplay of these two factors, therefore, determines not only the rate of filtration but also the efficiency of the pads to remove pyrogen. The pad now m use has a delivery volume of two liters every 2.5 minutes. A six-hour filtration would, thus, yield 288 liters, which would he more than enough to meet the daily require ments of a fairly large hospital. Hospitals using larger amounts of fluid may enlarge the filtering area of the filter by the method to be described later or install two such filters. Pyrogen Reteniivcness of the F ilter P ads.-- The ability of these pads to remove pyrogen was shown in a previous work10 to be due to adsorption and not to sieving. This means that the property can be exhausted. It is, there fore, necessary so to standardize the pads that they are not only uniform but also possess such a margin of safety in pyrogen retentiveuess that they remove the maximum amount of pyrogen which can occur in the distilled waters and the chemicals in use. I11 order to calibrate the pads in this respect it was necessary to secure a sample of pyrogen and use it as a standard. The pyrogen used in this work was isolated from the cell-free lierkefeld filtrate of the supernatant portion of concentrated stock suspension of typhoid vaccine ( 1 cc. = 88 billion cells). From seven liters of this suspension, four grams of the substance was yielded. Tlie method of isolation as well as the chemical and biologic aspects of this substance will be reported elsewhere. It is necessary, however, to state that the substance is still in the crude stage, containing six per cent nitrogen, which on further purification has been reduced to 1.6 per cent. The biologic (pyrogenic) potency of this crude sample, called " pyrogen-A," has been determined on rabbits, dogs, and men. Table I I shows the effect of diffeicnt doses of this preparation. If the minimum effective dose (M F D ) is ai bi ll 1 t wmmmm 0 0 TUI AND WKKJHT tranlv defined as dial close, per Kiln bodv-ueight, which on intravenous in jection wiii cause a temperature rise of o S to ].oF., the M E D in rabbits is j;.j o i g a m m a , in dogs, 0 4 g a m m a , and 0 0 4 gamma in men. T|ie-jea^on for XI using this crude subsiance as a standaul instead of the more purified'sub stance is that the process of purification is so tedious that it may be a long time before enough of the pum substance can be available for this purpose. In any case, the biologic calibration of " pyrogen-A" makes it unnecessary to . wait for the puier preparation. . ' ^ ' 5' Subject Rabbit 1 Rabbit 2 Rabbit 3 Rabbit 4 Dog 1. Dog 2 Dog 3 Dog 4 Man 1 Man 2 Man 3 Man 4. L in e r or nosj > or j\j<nc;rNW eight Kg 0 2.5 3 0 15 0 16 0 IS 5 16 Q 52 U 64 0 72 0 70 5 Dose per Kg 0 12 0 12 12 20 03 0.3 0.3 2.0 0 04 0 04 0.04 2.0 Tem p Change Leukocyte Count Deg F. C hange x 1000 0 .8 ' i} .jj 1.0 r V' S Q-9 3.5 1 1.0 - }.l ' i''' ; ' --6,1 , ` -+5 5 :I 'w IS 0.8 3.2 -- 10 2 1.2 0.9 1.1 )| 5.0 Process of Calibrating the Pads.-- One milligram o f " pyrogen-A" '. pQOO ' || gamnuis) is weighed out on a sensitive analytic balance and dissolved in ioo cc. of pyrogen-free distilled water Ten cubic centimeters of this solution contains 10 gammas of pyrogen Amounts of from five cubic centimeters up, in multiples of five, are filtered through test pads of a convenient size mounted in convenient filters. The filtrates are then tested in both rabbits and dogs. Table I II shows the protocols of some of these tests. lit! E xp erim en tal W eight A nim al Kg. R abbit 1 .................. ................2. S Dog t .. 16 0 Rabbit 2 30 Dog 2 15.0 R abbit 3, . . 30 Dog 3. 15 5 R abbit 4 25 Dog 4 . . 16.5 Rabbit 5 . . . . . 30 Dog 5 ,. . . . 15 5 T adli III tA L IbR A H O N OF H J-TE R PADS Volume T otal Dose Pyrogen P re fih ra te m Prefiltrate Cc Comma 30 300 30 300 35 350 35 350 40 400 40 400 45 450 45 450 50 500 50 500 Temp. Change Deg. F. -- .2 + .1 +2 25 + .15 + .1 +4 + .2 +4 +8 Leukocyte C Change x 1 -- .2 -- .3 + .1 -- .8 -- 1.2 It will be seen from Table I I I that the pad removes nearly 500 gammas of "pyrogen-A ." Since the pad is 2.5 cm. in diameter, it has a surface of 4 9 sq. cm., and, theiefore, each square centimeter removes over 100 gammas of " pyrogen-A." The filtering surface of the four large filter pads used in the hospital filter assembly are each 29.5 cm. in diameter, making a total filtering area of 2734 sq cm. The amount of pyrogen removable by the entire filter 416 PUKPAR Vl'toN Ob IM'USlON 1-1 UH)S assembly is approximately 273,400 gammas. Since about 140 yoiiiijidi rue required to cattle a temperature rise of 5 6 l*'. in an average man of 70 kilos, this filtering area will remote about 0333 such doses When applied to .solutions of sodium chloride or dextrose in distilled water this range of safety is more than ample as the amount of pyrogen contained in any combination of these three substances in the usual dose is tenuous and, further, the filter-pad assembly is made to deliver not more than jHH liters with each filtration The case of sodium citrate requires special consideration A s shown in Table I, gram for gram, the contamination of pyrogen in sodium citrate is heavier. This, however, is more than offset by the fact that the dose of sodium citrate used per 500 cc. of blood is usually in the neighborhood of 1875 mg. in 75 cc. of watei. This, by weight, is only about one-fifth of the dose of sodium chloride m one liter of physiologic salt solution and less than one-twenty-seventh of the dose of one liter of 5 per cent dextrose solution. Therefore, m making sodium citrate solutions, it is only necessary to use two filter pads and restrict the manufacture to not more Ilian 300 doses These calculations are admittedly based on biologic effects, and assume that pyrogen from sources other than the typhoid bacilli have the same potency per unit-weight as typhoid pyrogen. This assumption has no direct experi mental basis at present. The only justification for applying the figures derived from ``pyrogen-A" to the pyrogenic substances found in these fluids, whose exact source is unknown, is that they have so far held true in the samples whose pyrogenic content was roughly determined by the biologic test, then filtered through asbestos pads of the required area and then retested and found negative for pyrogen. In the absence of more exact knowledge, this approximate correspondence for the present can be used as a rough working principle. This method of pretesting before filtration may be used for any intravenous material whose maximum pyrogenic content has not been as thoroughly studied as have those of dextrose, sodium chloride, and sodium citrate. RESULTS OF 4 m o n t h s ' TRIAL During this experimental period, 3867 liters of intravenous fluids pre pared by the noupyrogenie method have been administered. They may be tabulated as follow s: No of Administrations 1815 6125,21 64 Amount Given Liters 2l 3 4 Total Given Liters 1815 1304 492 256 T o tals 2654 3867 lito In none of these was there a reaction. One patient who complained of a chilly sensation after receiving 1000 cc, did not have a tempei ature rise h 4- iniinwnmill l CO TUI AND WklCiHT ' , A.mi. surw/j. l.s'r;)'-. tf"" is clear ib.it if any pyrogen was present m the solutions, it was in such tninute amounts as to lie helow the M i'll) even in three to four liters. i.. a 1'a k t J I -- M e c h a n ic a l A s p e c t s / ' 'Mi/ The parts which comprise the entire liltcr assembly may now be enumer- aled anti the more important items described. The filtering portion of the assemhly is diagrainmatically represented in Figure I. .It consists of a com- piessed nitrogen or air tank, A, which furnishes filtration pressure; the reser voir, B, which holds the prefiltered m ixture, the filter itself, C ; the aloxitc fiber eliminator, D ; anti the collecting flask, K The teservotr, usually of 250 liter capacity, is made of stainless steel, and may be equipped with an electric stirrer. Figure 2 illustrates the essential parts of the filter as well as the path taken by the solution passing through it. It may be seen that each filter pad is held between two perforated Monel metal plates or frames, the three form ing a pad-frame unit. Tw o such units form a compartment. The fluid flows from the solution inlet, A, into a conduit system, B, which splits it into two streams. One stream filters through one pad downward and the other upward to meet in the collecting chamber, C, situated between the two pad-frames. From C the solution is conducted into the second compartment to have the process repeated, after which the conjoint stream from the second collecting chamber, D, passes through the filter eliminator into the collecting flask. Both tile latter are to he described in a subsequent section V -rf Vuimuc, u , ^ .Number d P RKPA kATfO N OK IXKUStON FLUIDS ft Since provisiun is made in the present filter to aceommodale two addi tional compartments, the filter has a still wider range of usefulness than has been calculated Any one of the three factois-- capacity, speed of filtration, or retentiveness-- can he increased two-fold by this doubling of the filtration , area. The stream in this case would be split in four instead of two as in the filter illustrated in Figure 2. ' Figure 3 is a photograph of the filter ensemble. A is the spindle with which the pads aie loaded and unloaded and the filtering system ren dered watertight, U is the filtering system, C is the pressure gauge, D die housing for the aloxite fiber eliminator, and E the outlet. The self-autoclav- ing system, F , consists of a steam generator in which the source of steam used is pyrogen-free distilled water. It may be heated electrically, by gas, by house- steam, or by an oil burner. The steam generated is led into the interior of the filter by the flexible hose, G. The course of the steam through the filter is identical with that of the fluid which follows it and whose course has been described in the foregoing paragraph. The aloxite fiber eliminator is a thick, finely porous filter made of car borundum. Because of its composition it rsr sturdy and may be heated to a temperature high enough to destroy any pyrogen which may have developed in the interstices of the stone. x The dispensing part of this system, mainly the collecting and storing flasks and accessories, is essentially the same as that described by W alter,6,0' 10 It is the most satisfactory one available at present, made to withstand bard 1 U J '] U A N D \ \ UK ii-.tgc, ami in thice (cars ul u-w lias siilicu a! the minimum amount'of break*, age h has, linwc\ cr, the --light nironvcnieik c of having too many parts, *1 ^ Figu re 4 illustrates the p ail of this st stem ubieli consists of. A , pyrex flasks of specially rua^cal construction of one lo tuas liters capacity, pieferably graduated ; 13, a one-hole rubber bushing to lit the ilask ; C, a special stainless steel stopper with a paitially grooved stem , I), a two-way vent tube foj- (be adm inistration of the fluids and for the adm ission of air into the flasks lo dis place the fluids The bushing, 11, is of special rubber which can stand 500 sterilizations without shrinking. It is provided with a hole in the center for the insertion of the two-way vent tube when fluid is being administered, and for the inser tion of the steel stopper when the flask is not in use. The stopper is provided with a partially grooved stem which is fitted partly into the hole of the bush ing while sterilization takes place, 'file gioove allows equalization of inside pressure with outside pressure. When sterilization is completed, the stem ol the stopper is pushed all the way in, as in E , the umbrella of the |topper serving as an hermetic >eal for the rubber bushing. i The original Fenwal-tubc is short, so that the flask end is submerged in the fluid when the flask is inverted tor administration. There are two olijec- 4in lions to this feature- F irst, when the infusion is hem*; given, the air which enters the flask to displace the fluid has to be bubbled through the fluid. T h is, in effect, means the w ashing of 1000 to 2000 cc. of im ^ erile air in the infusion fluid. B actcriologic cultures of ten sam ples of fluid subjected to this kind of air-w ashing gave three positive cultures B y the simple device of lengthening the tube so that the flask end extends above the surface of the fluid when the flask is inverted, this washing etTect is avoided. No positive cultures were found in ten fluids subjected to this system of air displacement. If it is desired to admit absolutely sterile air into the flask, the air vent may be modified in such a way as to accommodate a small sterile Bcrkefeld candle. This was originally used by us but later discontinued because of inconvenience. 421 I lie sauiiil (>bjt.< tion [u ihr siimi un^nul I'cmvahvent-lube is tin special c;uc is taken, the fiuul is as liable in flow through th( through the fluid vent The split ring bracket, 1\ is simply a wire basket for suspending the fiask in an in v e rte d po sitio n w hen tin iki-T is tn use f Some reference must he made to the still which delivers distil' ` for making the solutions. Since the still is the heart of all other much emphasis has been laid on it While it is our opinion that t concentration on the still has lulled to rest suspicion of other possible sources of pyrogen, we believe that the still must be cleaned periodically so that it adds no pyrogen to the distilled water to lowei the factor of safety of the filter. Moreover, it must be large enough to deliver the necessary amount of distilled water to be used for cleansing as well as to make the solqtioui^ themselves. A still that delivers 10 gallons of water per hour (40 lifer$)f will be sufficient to meet the needs of the average hospital. The periodic cleaning also does away with inerustrations which lower the output of the still. CARK OF T ilt: PARTS The Reservoir. The reservoir when new must he rinsed first with soap and water and then copiously with distil Itd water. Subsequent care after every use consists in rinsing it out thoroughly with distilled water and then drying it The Filter When new, the conduit system of the filter is rinsed thor oughly with distilled water. Thereafter, the same rinsing process is repeated after each use. The A loxitc F iber Eliminator After every use, the surface of the stone fiber eliminator is simply flushed thoroughly in order to wash away the par ticles and shreds which may he caught on its surface. A fter washing, it is placed in an electric oven at a temperature of 12 0 0 C. for one hour to dry. This temperature sterilizes the stone so that pyrogen production within the interstices of the stone is halted. /<>nwal-Flasks, etc. The initial cleaning and subsequent care of the Fenwal-flasks, rubber bushings, steel cages, and the modified vent tube have been described by W alter and need no extended treatment here. Suffice it to say, that the flasks are washed and brushed with hot soapsuds and rinsed first in tap and then in distilled water until the walls run homogeneously clear. The same treatment should be applied to the other glassware, namely, the vent tube and the drip as well as to the steel stopper. An occasional treatment of the glassware with a cleaning mixture of 1.5 per cent solution of sodium dichromate in concentrated sulfuric acid may have to be resorted to in order to get rid of spots in the glassware difficult to eliminate other wise. The rubber bushings when new arc first scrubbed with soap and watei until no further powdery substance is given off, then rinsed with 422 ^= t o n r r i f r i I W I ^ **? .^^T,r i '-ttw`tumvp inrum--#= dislilled water The rubber tubings when new au* likewiM.' 1 tilled well wuh hot soapsuds and then with distilled water Immediately after each washing, the rubber tubings, the vent tubes, the drip and the needles are wrapped and packed in a convenient container W e have used for this purpose a specially constructed copper cylinder, one foot long and 2 / inches in diameter, pro vided with a detachable cap as well as holes in the wall to allow penetration of the steam in the autoclave. After sterilization, the holes are closed by a special sliding device and the set stored for future use. Preparation for Filtration: The flasks without the rubber bushings in place are sterilized in the hospital autoclave. The rubber bushings and the steel stoppers are also sterilized but separately from the flasks. The filter with pads mounted on the frames is sterilized by its own selfautoclave for 30 minutes at 18 pounds pressure. After that it is allowed to cool to a temperature convenient for handling. The solution is mixed in the reservoir. The reservoir is then connected, on the one hand, with the'nitrogen tank and, on the other hand, with the inlet of the filter. The collecting and storing flask is placed at the outflow tube of the filter and filtration is begun by turning on the gas gradually to 12 pounds. After discarding the first four liters, the flasks arc then filled one by one. After collection, the flasks are partially stoppered with the stainless steel stopper in such a way that the groove in the stem of the stopper allows com munication of the inside of the flask with the outside (F ig 4). The flasks are then put into the autoclave and sterilized at 250 F . for 30 minutes. After sterilization, special care is taken to have the pressure in the autoclave grad ually diminished so as not to cause boiling. The stem of the steel stopper may then be pushed home, in which case a partial vacuum is created as the solution cools. This partial vacuum, which helps retain the stopper in place and causes a water-hammer effect when the bottom of the flask is struck with the fist, ma}' be used to indicate whether the stopper is leaking after storage. After the flask is scaled, it is stored until needed Just before administering the fluid, the steel stopper is removed from the flask with a rocking and unscrewing motion. The flask end of the twoway vent tube is now inserted through the hole in the rubber bushing to the point where one of the two dilations on the glass tube just slips into the flask beyond the rubber bushing. The flask is inverted, placed in the ring bracket, and the infusion started. Calculation of Costs: This method requires an initial equipment which may be divided into two. parts, an irreducible item which consists of the reservoir and the filter, and an item which can be expanded according to need, namely, the flasks and their accessories. The rubber tubings and drip and needles do not enter into consideration here as they will have to he used in any case. The irreducible equipment costs about $10130. If the cost of the flasks for initial use is set at $600 and to this is added part-time labor of $50 a month, winch can be expanded into full-time labor as the need arises, 423 C O I III \NI> W EH illT ,,bui>,lvuibr. 1lS8j1-T] llic total may be set at $2-?00 \ny hospital which expends lilis amount of money a year fui infusion fluids \\ill not incur a loss, even in the fir^t year, by adopting this .sNsteni Beyond the finsi year the savings are substantial; the larger the amount of fluids used,the more substantial the savings, since the only items of expense will Uthe cost of leplacements, materia! and labor. But the savings are not the eiitue consideration The safety from pyro genic reactions in intravenous medication, especially in bloo4 and plasma banks, should be the majoi consideration, particularly during w a rtim es. Instances are known of tens of liters of blood and plasma which have had to he discarded because the sodium citrate used was pyrogenic This type of sirin g has not entered into the above calculation but may he the major item. M bU M M A U Y AN D CONCLUSIONS T. Methods of intravenous fluid manufacture which depend only upon the purity of distilled water used have no provision for dealing with pyrogen should this substance be piescnt in the final mixture and, therefore, do (Jtjt entirely eliminate pyrogenic reactions. 2. Chemicals have been si own to be occasionally contaminated with pyrogen 3. The adsorptive filtration method of preparing assuredly nonpyrogenic fluids, described five years ago, has been developed into a practicable and convenient method for hospital use 4. The compressed asbestos filter pads, around which this method is lmilt, have been calibrated for speed of filtration and pyrogen retentiveness. The factor of safety is shown to be more than ample for the most exacting use in the manufacture of physiologic salt, dextrose, and sodium citrate solutions. 5. The uniformly nonpyrogenic results of 42 months' use of fluids piami- faclured according to this method are reported. 6. The mechanical aspects of the method are described. 7. The method not only eliminates pyrogenic reactions but is also more economical than the use of commercial fluids. The authors wish to acknowledge their grateful thanks to Dr. Edward M. Berneker, Commissioner of Hospitals, to Commissioner Albert Pleydell and to Mr. A. M. Moore of the Department of Purchase, City of New York, for their encouragement and co operation in furnishing the equipment for the completion of this work; to Dr. William F. Jacobs, Medical Superintendent of Bellevue Hospital, for the privilege of installing the experimental filter in the Bellevue Hospital, to Mr. Francis X. Dealy, of the Ertel Engineering Corp. of New York, for his cooperation m the mechanical phases of the problem, and to Dr. Wilfred Ruggiero, Messrs. Arthur Yates, Angelo Benaglia, and Milton Schrift for their constructive suggestions; to Dr. Ralph S. Muckensus, of the Bureau of Laboratories, City of New York, to l.edeile Laboratories, Inc., and to E. R. Squibb & Co. we are indebted to for supplies of the stock suspension of typhoid bacilli. REFERENCES 1 Co Tui, McCloskey, K I,.f SthnU, M., and Yates, A. L . : A New Method of Pre paring Infusion Fluids J.A .M .A , 109, 250, 1937. * d a n : A New Method of Preparing Nonpyrogenic Intravenous Infusion Fluids; Based 424 . WBWMWHfiiirut PRKPAR VI'ION' OF INFL S 1 0 N FLUIDS un Removal ol Pyrogen hy Adsorptive Filt iatum \ n n a i >. m S t 'kG)-K\, 106, oSu- 1094, Dcccmbei, 1937. J Siebei t, 1'. H., Fevci -producing Substance hound in some Distilled Waters, Ata. Join P h y sio !, 67, 90, 1923. i h ir iii` Cause of many Febiile Reactions following Intravenous Injections Am Jour Physiol. 71. 621. 1923. ' Hurt, E. C , and PcnfoUl, W J , . "Flic Dangers of Saline Injections. Bnt. Med Jour., 2, 15S9, Deeemher 16, 19 11 a Rademnkci, 1. Cause and Elimination of Reactions after Intravenous Infusions A n x .u s o f SuKOt wy, 92, 19 5 , 19 3 0 7 U n it Reactions after Intravenous Infusions . Further Report on their Elimination Surg , G ynee and O bstet, 56, 956, 1933 'W alter, C. \V .. Economical Intravenous Therapy. 104, j 688, 1935. ' I linn Preparation of Safe Intravenous Solutions. Surg , Gynec. and Obstet., 63, 643, ` 93f> Vl h i n ir The Relation of Proper Preparation of Solutions for intravenous Therapy to Febrile Reactions A n n \ ls of S urc.rry, 112, 603, 1940 11 Co Tui and Schrift, M. H .: A Tentative Test for Pyrogen m Intusion Fluids Proc. Sue. Kxper. Biol and Med , 49, 3-20-323, 1942 l -C o Tui, Schrift, M. H ., McCloskey, K L., and Yates, A. L . : Filtration Studies on Pyrogenic lmilin. Proc. Soe. Exper. Biol and M ed, 35, 227-230, March, 1937. 1J Smith, H W Personal communication 11 Co Tui and Schrift, M H . . Production of Pyrogen by some Bacteria. J Lab and (Jin M ed, 27, 569-57$. 1042. ' 'K rn o- Private communication quoted by Walter.10 IUCo Tui, McCloskey, K . L , Schrift, M H,, and Yates, A L Filtration Studies on Reactive Infusion Fluids. Proc Soc E xper Biol and M ed, 38, 297-300, 1936.