The second international standard for tetracycline*

Bull. Org. mond. Sant,6 1973, 48, 99-105 Bull. Wid Hith Org. The second international standard for tetracycline* J. W. LIGHTBOWN,1 A. H. THOMAS,1 & I. L. BERRYMAN 2 A second international standard has been established to replace the first International Standard for Tetracycline, stocks of which were depleted. The International Unit of tetra- cycline is defined as the activity contained in 0.00101833 mg of the second International Standard, corresponding to a potency of 982 IU/mg. These values were assigned on the basis of the results of an international collaborative assay carried out by laboratories in 19 countries. Most of the assays were by diffusion techniques; afew were by turbidimetric methods. Both procedures gave statistically valid assays but there was some heterogeneity of repeated estimations within a laboratory and also some heterogeneity of the mean estimates from each laboratory. The agreed defined potency was the unweighted mean of 281 individual assays; the confidence limits were 978-986 IU/mg. Since in 1969 stocks of the first International Stan- dard for Tetracycline were becoming depleted, mate- rial suitable to serve as a second international stan- dard was obtained and a collaborative assay ar- ranged. The sample, consisting of approximately 430 g of tetracycline hydrochloride, was made avail- able through the generosity of The Pfizer Group, United Kingdom, and through the good offices of Mr E. Addison of that company. In October 1969 the WHO Expert Committee on Biological Stan- dardization (1970) authorized the National Institute for Medical Research, London, to establish this mate- rial as the second International Standard for Tetra- cycline on the basis of the results of the collaborative assay and to define the international unit with the agreement of the participants. THE PROPOSED SECOND INTERNATIONAL STANDARD FOR TETRACYCLINE The sample of tetracycline hydrochloride was received at the National Institute for Medical Re- search, London, in a single container in May 1969. The following analytical data were supplied by the manufacturer: * From the WHO International Laboratory for Biological Standards, National Institute for Medical Research, Mill Hill, London, England. Division of Biological Standards. 'Statistical Services Section. lot No. 903-7172 microbiological assay 3 977 "&,ug "/mg chemical assay 974 ,ug/mg volatile matter 0.48% sulfated ash 0.01 % chromatography: anhydrotetracycline 0.7% 4-epitetracycline 1.5% The sample was kept in its original sealed con- tainer in the dark at -20°C until later in May 1969, when the material was mixed and distributed into approximately 5 000 nonactinic, neutral glass am- poules so that each contained approximately 75 mg. The ampoules were fitted with vented polyethylene plugs and dried over phosphorus(V) oxide in vacuo to constant weight in 7 days. After filling with pure dry nitrogen, the ampoules were sealed by fusion of the glass, tested for leaks, and stored in the dark at -20°C. The material in the ampoules was examined by heating at 60°C over phosphorus(V) oxide at a pressure of <0.05 mmHg for 11 hours; no sig- nificant loss of weight was observed. When exposed to the atmosphere at 60%o relative humidity in an open weighing bottle, a sample of the material in the ampoules increased in weight by approximately 0.1 % in 3 hours. a When jg is placed between quotation marks ("g "), it refers to a certain amount of activity (potency) and not to its accepted usage as a unit of mass. 2985 - 99 Table 1. Details and results of bioassays used by participating laboratories Labor- No of Mean Total Average atory Assay method Design Test organism N.- potency Heterogeneity Toa weight/ No. asa;(lU/mg) weight assay 1 Petri dish 2 Petri dish 2 large plate 3 Petri dish (6 assays) large plate (3 assays) 3 Petri dish (4 assays) large plate (3 assays) 4 Petri dish 5 turbidimetric 6 Petri dish 7 Petri dish 8 large plate 8 large plate 9 large plate 1 1 turbidimetric 12 large plate 13 turbidimetric 14 large plate 15 large plate 15 large plate 17 Petri dish 17 Petri dish 1 8 turbidimetric 18 Petri dish 18 Petri dish 20 Petri dish 21 large plate 21 large plate 21 large plate 21 turbidimetric 22 large plate 23 Petri dish 23 Petri dish 3 + 3 3 + 3 2 + 2 3 + 3 3 + 3 3 + 3 3 + 3 2 + 2 3 + 3 3 + 3 3 + 3 3 + 3 (4 assayS) 20 assays) 3+3 (1 assay) 2 + 2 (3 assays) 3+3 6 + 6 3+3 (6 assays) 2 + 2 (22 assays) 3 + 3 3 + 3 3 + 3 3 + 3 7 + 2 2 + 2 2 + 2 4 + 4 4 + 4 4 + 4 4 + 4 4 + 4 3 3 3 + 3 3 + 3 3+3 Bacillus subtilis 9 993 ATCC 6633 Bacillus subtilis 8 993 ATCC 6633 Bacillus pumilus 10 986 ATCC 8241 Sarcina lutea 9 1 004 ATCC 9341 Bacillus pumilus 7 984 NCTC 8241 Bacillus subtilis 18 975 var. L2 No. 21 Staphylococcus aureus 6 969 ATCC 6538-P Sarcina lutea 24 986 ATCC 9341 Sarcina lutea 12 977 ATCC 9341 Sarclna lutea 4 993 NCTC 8340 Staphylococcus aureus 3 983 TH6 Bacillus cereus var. 24 978 mycoides ATCC 11778 Staphylococcus aureus 4 953 ATCC 6538-P Bacillus pumilus 11 1 001 NCTC 8241 Staphylococcus aureus 11 1 044 ATCC 6538-P Bacillus subtilis 28 991 No. 201 Bacillus subtilis 14 ATCC 6633 Bacillus cereus 1 ATCC 9634 Bacillus subtilis 15 Sarcina lutea 8 ATCC 9341 Staphylococcus aureus 5 ATCC 6538-P Bacillus cereus var. 4 mycoides ATCC 11778 Sarcina lutea 5 ATCC 9341 Bacillus cereus 10 ATCC 9634 Sarcina lutea 1 ATCC 9341 Bacillus cereus var. 9 mycoldes ATCC 11778 Bacillus pumilus 4 NCTC 8241 Klebsiella pneumonlaa 5 ATCC 10031 Bacillus cereus 4 NCIB 8122 Bacillus pumilus 4 NCTC 8241 Bacillus cereus 4 NCTC 10320 N.S. N.S. N.S. N.S. N.S. N.S. P<0.001 N.S. N.S. P<0.01 P<0.01 N.S. 59 969 277 256 1 487 861 55 477 18 456 182 650 13720 42 703 155 400 11 878 6 671 543 759 6 663 34 657 148 786 6 164 2 637 10 147 2 287 1 779 12 950 2 970 2 224 22 657 N.S. 30773 7 693 N.S. 321 109 29192 P<0.001 25 317 2 302 P<0.001 332 979 11 892 950 N.S 1 054 - 969 P<0.01 970 N.S. 923 P<0.02 959 N.S. 991 982 946 974 980 962 987 1 006 995 N.S. P<0.001 N.S. N.S. P<0.001 N.S. N.S. N.S. 89 315 4754 77 900 109 073 5 757 8018 4 494 32 088 6416 154 471 16 521 20 899 174 608 200 511 94 254 6 380 4754 5 193 13634 1 151 2 005 899 3 209 6416 17 163 4130 4180 43 652 50128 23 564 TETRACYCLINE THE FIRST INTERNATIONAL STANDARD FOR TETRACYCLINE This was established in 1957 (WHO Expert Com- mittee on Biological Standardization, 1958) and has been described by Humphrey et al. (1957). Its potency is 990 IU/mg. THE COLLABORATIVE ASSAY The participating laboratories were asked to esti- mate the potency of the proposed second inter- national standard for tetracycline in terms of the International Standard for Tetracycline by biologi- cal assay. They were supplied with 5 ampoules of each of the two materials. As is customary in these collaborative assays, laboratories were asked to design their assays in such a way that each assay would contain sufficient information to provide from its own internal evidence an estimate of potency and the fiducial limits to that estimate. It was also requested that the assays should provide data to allow evaluation of linearity and parallelism of the log dose-response curves. The dilutions of the stan- dard and unknown, which formed the basis of the assay, were to be tested with or without replication in a single unit of the experiment. Participants were asked to provide additional information by comparing the two materials by any chemical method with which they were familiar. Altogether 23 laboratories in 21 countries agreed to take part in the collaborative assay, and assay results were received from 21 laboratories, one of which submitted results from chemical assays only. In this report the laboratories are referred to by number only, the number not necessarily bearing any relation to the order in which the laboratories are listed in Annex 1. STATISTICAL ANALYSIS In all, 281 assays were carried out by biological methods and 33 by chemical methods. The biological assays were carried out using either the plate diffusion method (250 assays) or the turbidimetric method (31 assays) and with one of six different test organ- isms. Table 1 lists, for each laboratory, the number of assays and the method and test organism used. The chemical assays are reported in Table 2. In accordance with the usual practice, only the biological assays have been used to determine the potency of the proposed second international standard. For each assay, the log potency was estimated Table 2. Details and results of chemical assays reported by participating laboratories Laboratory MehdNo. of Mean No. Method assays potency(Mg/mg) 1 colorimetric a 4 976 spectrophotometric b 4 981 10 colorimetric a 4 983 spectrophotometric 4 975 11 spectrophotometric b 2 993 15 spectrophotometric b 1 983 17 spectrophotometric b 4 999 20 colorimetric c 10 988 All assays 33 985 a lron(lll) chloride method. b Ultraviolet absorption. CAcid colorimetric method: for further details of the methods see, for example, Grove & Randall (1955). from the horizontal distance between the two log dose-response lines, an analysis of variance appro- priate to the design of the assay was calculated, and statistical tests of significance for parallelism and curvature were carried out. For Laboratories 1, 4, and 9, the square of the zone diameter was used as the response metameter, since this procedure improved linearity and parallelism. In analysing the assay data for the turbidimetric assays from Labora- tories 13 and 18, data relating to some dose levels were omitted from the statistical analysis; both labo- ratories had used a large number of dose levels (Laboratory 13 used a 6+6 design, Laboratory 18 a 7+2 design). It was found that the log dose- response lines were not linear over their entire length. A graph of response against log dose was plotted for each assay and the data that were considered to relate to the straight portion of the line were included in the statistical analysis. Laboratory 21 used a 4+4 assay design, but for four diffusion assays did not submit data for all dose levels, omitting those data that it was felt would lead to curvature. In the assays from Laboratory 1, which were 3+3 block assays, the responses had been measured independently by each of three operators, but for most assays it was not possible to correlate, for a given zone diameter, the three replicate read- ings with the three operators since the necessary information had not been recorded. The nine assays 101 J. W. LIGHTBOWN ET AL. from this laboratory were based on three sets of weighings of the test material and of the standard, each set being assayed by each of the three opera- tors, and the mean potency that is reported (Table 1) was obtained by analysing the readings made by the operators who actually performed the assay. RESULTS Statistically significant departures from parallel- ism of the log dose-response lines were found in four assays, a number no greater than would be expected by chance, and these particular assays were not rejected. No laboratory found the two preparations to produce log dose-response lines that were consistently concave or consistently con- vex, nor was the slope of the line for one prepara- tion consistently greater than that for the other preparation. For each laboratory, the residual mean squares from the analysis of variance were pooled over assays of the same design that used the same test organism. These pooled variances were then used within that laboratory, when weights were calculated for the log potencies from the assays from which the pooled variances had been derived. Using these weights, tests of homogeneity were made within laboratories on the log potencies of assay that had employed the same method and test organism. If the log potencies were homogeneous, the weighted mean was calculated and its weight was taken as the sum of the individual weights; if the log potencies were heterogeneous an unweighted mean was calculated and its weight assigned from the variability between the unweighted log potencies. The results from these calculations are set out in Table 1, where it can be seen that 20 laboratories provided 31 estimates of the potency of the proposed second international standard. These 31 potencies were found to be significantly heterogeneous (X2=128.4, d.f.=30, P<0.001) and the possible sources of this heterogeneity were examined. The 26 potencies obtained using the plate (diffusion) method were a heterogeneous group, as were the 5 mean potencies for the turbidimetric method (P<0.001 for both methods), and the un- weighted means were very similar-982 IU/mg for 250 plate assays, 984 IU/mg for 31 turbidimetric assays. The results in Table 1 were also tested to see whether use of a common test organism had led to greater consistency in the results from a group of laboratories. For the plate diffusion assays, the results derived from the four test organisms that had been used by more than one laboratory (Bacillus subtilis, Bacillus pumilus, Bacillus cereus, and Sarcina lutea) were tested for heterogeneity, within organ- isms, between laboratories; but this was found to be no less than that between organisms. Four of the five laboratories that performed turbidimetric assays employed the same test organism (Staphylococcus aureus) so that it was not possible to examine such effects in this case. In combining the results to determine the potency of the second international standard the unweighted mean of all 281 assays was taken and confidence limits were based on the variability between these assays. The mean log potency obtained in this way was 2.992, equivalent to a potency of 982 IU/mg, with 95% confidence limits of 978 to 986 IU/mg. Although the overall mean potencies obtained by the two methods (plate diffusion and turbidimetric) were in close agreement, the data from the study indicate that turbidimetric methods gave results that were more variable, both among laboratories and among assays within a laboratory, than those ob- tained using plate diffusion methods. The five mean potencies obtained with turbidimetric methods range from 923 to 1 044 IU/mg, in contrast to the narrower range of 950 to 1 006 lU/mg for the 24 mean poten- cies obtained using plate diffusion methods (Table 1). Heterogeneity within laboratories was also more common for turbidimetric assays (4 out of 5) than for diffusion assays (5 out of 26). This is reflected in the average weight per assay; from Table 1 it can be seen that, in general, the plate assays have a higher average weight than the turbidimetric assays. Assays from Laboratory 1 The data from Laboratory 1 were examined, for the one set where this was possible, to see whether there was any effect related to the fact that different operators assayed the same solutions, or read the same plate. The following observations were made. (1) When the three operators (A, B, and C) read the same plate, A and C obtained virtually identical readings (zone diameters), but B's readings were consistently higher, the average difference being as high as 7% in some assays. (2) Probably because these differences in reading the zone diameters were consistent, there were no significant differences in the potencies calculated from the readings of the different operators. When each operator's readings were analysed as separate 102 TETRACYCLN assays, the three potencies varied by as much as 4.5% in some cases but the estimates were not consistently different over all assays. Mean values of 996 IU/mg for A, 988 IU/mg for B, and 990 IU/mg for C were obtained. Chemical assays Three different chemical assay procedures were used in 6 laboratories but the mean value for all assays (Table 2) agreed closely with the mean potency obtained by biological assay. DISCUSSION Both intra- and inter-laboratory heterogeneity of the potency estimates was observed in this collabora- tive study, but the intra-laboratory variations were no greater than is usually encountered with this type of high precision assay of antibiotics. The mean esti- mate of potency of the second international stan- -dard was based on a greater number of assays than was used to establish the first international standard (Humphrey et al., 1957). Although there was inter- laboratory heterogeneity in both international col- laborative studies, the variation about the mean of ±7.5% in the present study was much less than the corresponding range observed in the collabora- tive assay of the first international standard for tetracycline, i.e., ± 15%. The present results show that the internal evidence of the turbidimetric assays provided a less reliable indication of the true assay variability than did the internal evidence of the plate diffusion assays, i.e., the weight of the log potency overestimated the reproducibility between assays, particularly for turbi- dimetric assays. This overestimation reflects the very small variability between the measurements of repli- cate responses at the individual dose levels within an assay, and it seems possible that the designs, techniques, and procedures that are commonly used inturbidimetric assays are such that the true variabil- ity of the replicate responses may not be revealed. The chemical estimates of the purity of the two preparations were very similar: the mean for the first international standard was 987 ,ug/mg and that for the second international standard was 985 &g/mg. Thus the estimates of the mean potency of the second international standard determined biologically and chemically were in close agreement. The second international standard was found to be approximately 1 % less potent than the first inter- national standard, a slight, but nevertheless signifi- cant difference. The two materials have been exam- ined for related impurities such as chlortetracycline and tetracycline degradation products by the chro- matographic methods of Addison & Clark (1963), the British Pharmacopoeia 1968, the Code ofFederal Regulations, 1969, and Fernandez et al. (1969). The only detectable differences were the slightly higher content of epitetracycline in the first international standard and the presence of a small proportion of anhydrotetracycline (<1.0%) in the second inter- national standard. Traces of epianhydrotetracycline, but no chlortetracycline, were detected in both standards. ESTABLISHMENT OF THE SECOND INTERNATIONAL STANDARD AND DEFINION OF THE INTERNATIONAL UNIT In accordance with the instructions of the WHO Expert Committee on Biological Standardization (1970), the above material was established as the second International Standard for Tetracycline and with the agreement of the participants in the col- laborative assaythe potencywas defined as 982 IU/mg. The international unit for tetracycline is defined as the activity contained in 0.00101833 mg of the second International Standard for Tetracycline. IUME DEUXItME tTALON INTERNATIONAL DE TETRACYCLINE Un deuxieme etalon international de tetracycline a et6 constitu6 afin de remplacer le premier etalon international dont les stocks 6taient presque epuises. Le National Institute for Medical Research, de Londres, s'est charge de r6partir le materiel propose en ampoules et d'organiser un titrage comparatif auquel ont participe 21 laboratoires de 19 pays. On a proc6d6 au total A 281 essais biologiques sur six micro-organismes differents, par diffusion ou par turbidimetrie, et A 33 analyses chi- miques. Seules les donn6es fournies par les 6preuves bio- logiques ont servi A definir l'activit6. L'analyse des resultats a fait ressortir une certaine h6t6rog6n6it6 des estimations de l'activit6 A la fois dans un meme laboratoire et entre les laboratoires. Cette 103 104 J. W. LIGHTBOWN ET AL. heterogen&et n'a pu etre attribuee A l'utilisation de methodes ou de micro-organismes differents. La combi- naison des resultats des 281 essais a conduit A adopter une valeur moyenne d'activit6 de 982 UI/mg, avec des limites de confiance de 978 A 986 UI/mg. Conformement aux instructions du Comite OMS d'experts de la Standardisation biologique (1970), les participants au titrage comparatif ont constitue le mate- riel propos6 en deuxieme etalon international de tetra- cycline dont 1'activite a et6 fix6e a 982 UI/mg. L'unit6 internationale de tetracycline a etW definie comme l'acti- vite de 0,00101833 mg du deuxieme etalon international de tetracycline. REFERENCES Addison, E. & Clark, R. G. (1963) J. Pharm. Pharmacol., 15, 268-272 British Pharmacopoeia, 1968, Addendum 1969, London, Pharmaceutical Press, pp. 76-77 Code of Federal Regulations, 1969, Title 21, Chapter 1, Section 141.550, Washington, Federal Register Fernandez, A. A. et al. (1969) J. pharm. Sci., 58, 443-446 Grove, D. C. & Randall, W. A. (1955) Assay methods of antibiotics, New York, Medical Encyclopedia Inc., pp. 53-56 (Antibiotics Monograph, No. 2) Humphrey, J. H. et al. (1957) Bull. Wld Hlth Org., 17, 521-526 WHO Expert Committee on Biological Standardization (1958) Wld Hlth Org. techn. Rep. Ser., No. 147, p. 6 WHO Expert Committee on Biological Standardization (1970) Wld Hlth Org. techn. Rep. Ser., No. 444, p. 9 Annex 1 PARTICIPATING LABORATORIES National Biological Standards Laboratory Department of Health Canberra, Australia (Dr L. F. Dodson & Mr N. M. Semple) Institute of Hygiene and Epidemiology Brussels, Belgium (Dr A. Lafontaine, Dr A. Vanden Bulcke, & Mr R. Bogaerts) Drug Control Department Belgian Pharmaceutical Association Brussels, Belgium (Dr J. Dony & Miss I. Boudru) Biologics Control Laboratories Laboratory of Hygiene Department of National Health and Welfare Ottawa, Ontario, Canada (Dr L. Greenberg & Miss K. Fitzpatrick) State Institute for the Control of Drugs Prague, Czechoslovakia (Dr J. R. Burianek, Dr I. 0. Hrdy, & Mr Z. Vesely) Microbiology Section National Public Health Laboratory Montpellier, France (Professor J. Desbordes & Mrs Esteve) Laboratory for the Control of Pharmaceuticals Helsinki, Finland (Dr E. Nieminen) Research Institute for Pharmaceutical Chemistry Budapest, Hungary (Dr 1. Horvath, Dr 1. Koczka, Susan Kerenyi, Rita Szlauk6, I. Inczefi, & Barbara Derfoldi) Central Drugs Laboratory Calcutta, India (Dr D. Ghosh & Dr J. N. Ghosh) Department of Antibiotics National Institute of Health Tokyo, Japan (Dr H. Umezawa & Dr S. Yamazaki) Laboratory of Chemotherapy National Public Health Institute Utrecht, Netherlands (Dr A. Manten & Dr B. van Klingeren) Drug Control Laboratory Oslo, Norway (Mrs E. D. Aarnes & Miss E. Holum) Technical Department of Pharmacy and Control of Drugs Lisbon, Portugal (Dr Manuel Godinho de Matos, Jr & Dr (Mrs) Maria Jose Mendanha de Sa Lemos) State Pathology Laboratory State Health Department Cape Town, South Africa (Dr J. E. C. Mullen & Miss P. 0. Wessels) TETRACYCLINE Drug Analysis Division Drug Standard Institute Bangkok, Thailand (Dr P. Tamprateep, Mrs R. Pintavom, Mrs S. Srichaiyanta, Miss W. 0. Nana, Miss P. Chanruang, Miss V. Khummuang, & Mr V. Kanavuti) Drug Control Department Central Institute of Hygiene Ankara, Turkey (Dr I. Tuna, Dr Mithat Kiper, & Miss Olkil Onal) State Control Institute for Medical Biological Preparations Moscow, USSR (Professor S. G. Dzagurov & Professor L. M. Jacobson) The Pfizer Group Sandwich, Kent, England (Mr 0. Hughes) Division of Biological Standards National Institute for Medical Research London, England (Mr J. W. Lightbown, Mr P. Isaacson, & Dr A. H. Thomas) Food and Drug Administration Department of Health, Education, and Welfare Washington, D.C., USA (Dr W. W. Wright & Mr A. Kirshbaum) Institute for the Control of Drugs Zagreb, Yugoslavia (Dr Milan Grims, Mirjana Kupini6, & Ana Radogevi6) 105