The international standard for rolitetracycline*

Bull. Org. nwnd. Santd | 1973, 48, 229-234Bull. Wld Hith Org. The international standard for rolitetracycline * J. W. LIGHTBOWN,' JILLIAN M. BOND,1 MARJORIE V. MUSSETF,2 & LAURA L. TACKETT2 An International Standardfor Rolitetracycline has been established and the international unit of this antibiotic defined as the activity contained in 0.001004 mg of the international standard. The definition of the international unit was based on the results of a collaborative assay in which 8 laboratories in 6 different countries participated; a total of 133 assays were performed. The assay was in terms of the Working Standard of the USA Food and Drug Administration; mean potencies for individual laboratories varied within a range of only 2 % of the mean for all assays although 7 different test organisms were used in both diffusion and turbidimetric assays. Individual assays, however, provided potencies that varied within a range of40 %. The WHO Expert Committee on Biological Stan- dardization (1964b) noted that, in accordance with an earlier request (WHO Expert Committee on Biologi- cal Standardization, 1964a), national control labora- tories in 20 countries had been consulted on the need for an international reference preparation ofrolitetra- cycline. Thirteen laboratories were in favour of set- ting one up, and a sample suitable for that purpose had been obtained. The Committee requested the National Institute for Medical Research, London, England, to arrange for a collaborative assay of this material. The proposed international reference preparation was first assayed against the Working Standard for Rolitetracycline of the USAFood andDrug Adminis- tration (FDA) by the FDA laboratories and by the National Institute for Medical Research, London (NIMR). These assays showed that its potency was 1.09 times that of the FDA standard, which had a potency of 1 000 " ug "/mg,8 implying a purity of 100%. On this basis the proposed international refer- ence preparation would have had a purity of 109%. In view of these results, which were reported to the WHO Expert Committee on Biological Standardiza- tion (1965), it was decided to assay the preparation against other national standards before continuing the collaborative assay. * From the WHO International Laboratory for Biological Standards, National Institute for Medical Research, London, England. Division of Biological Standards. 2 Statistical Services Section. 'When pg is placed between quotation marks ("jig "),it refers to a certain amount of activity (potency) and not toits accepted usage as a unit of mass. The preparation was assayed against the Canadian National Standard for Rolitetracycline and against a second working standard that had been established by the FDA, and no significant differences were found. The assay against the FDA first working stan- dard was repeated and failed to confirm the previous findings, no difference being found between the two preparations. It was therefore decided to proceed with an international collaborative assay of the proposed international reference preparation in terms of the FDA second working standard. MATERALS The proposed international reference preparation of rolitetracycline The material was supplied through the generosity of Farbwerke Hoechst AG, Frankfurt am Main, Federal Republic of Germany, and the good offices Dr Fritz Lindner of that company. The following data were supplied by the manufacturer. batch No. P126/4/65 sterility sterile biological activity 1 mg corresponds to 904 " ,ug" of tetracycline hydrochloride substances similar corresponds to the require- to histamine ments of FDA toxicity corresponds to the requirements of FDA pyrogen corresponds to the requirements of FDA colour light yellow odour very little odour clarity clear as water (yellow solution) 3000 - 229 230 J. W. LIGHTBOWN ET AL. pH 8.2 water content (Karl Fischer) 0.89 % water content (60°C in vacuo) 1.08% Cl content 0.14% Sulfated ash 0.02% [a] '+52.760 (500 mg of substance in purest dioxane filled up to 25 ml) content of pyrolidinyl 99.6% (UV test, calculated methyltetracycline (PMT) from the extinction of a 1 % PMT solution of 1 cm thickness at 380 nm) The sample was received at the NIMR in August 1964 and was stored, protected from moisture, at -10°C in the dark until May 1965, when it was distri- buted in approximately 100-mg amounts into neutral glass ampoules fitted with vented polyethylene plugs. In accordance with the suggestion of the manufac- turers, the material in the ampoules was dried for 24 h over phosphorus(V) oxide in vacuo; this proce- dure was expected to leave a residual moisture content of approximately 1 %, which was considered desir- able to maintain stability. After drying, the ampoules were filled with nitrogen, sealed by fusion of the glass, and stored at -20°C. The moisture content of the material in the am- poules, measured as loss in weight when dried to constant weight at 56°C at a pressure of <0.05 mmHg over phosphorus(V) oxide, was found to be 1.7 %. When exposed to an atmosphere of 50% rela- tive humidity, the material in the ampoules picked up approximately 0.2% (by weight) of moisture in 10 min and 1.0% in 40 mi. FDA Second Working Standard for Rolitetracycline A sample (2.0 g) of the FDA Second Working Standard for Rolitetracycline was obtained 1 in July 1966 and was stored for 3 weeks at -20°C, protected from moisture. It was then distributed in approxi- mately 50-mg quantities into neutral glass ampoules, which were then filled with nitrogen and sealed without previous drying. The potency of the material was defined as 994 " jg "/mg when dried for 3 h at 60°C at a pressure of 5 mmHg or less. THE COLLABORATE ASSAY Six ampoules of the proposed international refer- ence preparation and four ampoules of the FDA 1 From Dr W. W. Wright, Food and Drug Administration' Department of Health, Education, and Welfare, Washington D.C., USA. second working standard were made available to each of the eight laboratories participating in the collaborative assay. Throughout this report, labora- tories are identified by code numbers, which do not represent the order in which their addresses are listed in Annex 1. The laboratories were asked to ensure that each of their assays yielded sufficient information to provide from its own internal evidence an estimate of the potency of the preparation in terms of the FDA second working standard, and fiducial limits to that estimate. It was also required that the design should allow tests of validity (parallelism and linearity) to be made. Statistical analysis of the results The participating laboratories provided the results of a total of 133 assays of the preparation in terms of the FDA standard. These assay data were analysed statistically by standard methods for parallel-line assays, relating response (or transformed response) to log dose. Eight assays were discarded as invalid be- cause the " departure from parallelism " term in the analysis of variance was significant at the 1% level. A single assay from Laboratory 1 was rejected be- cause the log dose-response lines were significantly curved (P< 0.01) and the potency estimate was differ- ent from other values obtained by the same labora- tory. In 10 of the 124 assays considered to be valid, the parallelism term was of borderline significance (0.01 < P < 0.05). A breakdown of assays by labora- tory, giving the assay method, test organism, response metameter used, and assay validity is given in Table 1. For most of the laboratories a linear relationship existed between the response (as measured) and the log dose, but for Laboratory 4 the square of the zone diameter was more satisfactory. Laboratory 8 mea- sured the area of the zone of inhibition in one set of 8 plates but used the zone diameter as the response in its other assays. When area was the measured response, it was decided to use a square-root trans- formation, since this reduced the response metameter to the same dimensions as the zone diameter in the other assays and also was nearer to a linear log dose relationship than the untransformed response. Laboratory 2 used 18 doses of each preparation over a wide range of concentrations. A number of the doses produced negligible responses whereas, at the other end of the response curve, some doses per- mitted no growth of the test organism. The responses to these doses at the ends of the range were omitted ROLITETRACYCLINE Table 1. Details of assays and methods used No. of No. ofLab. Assay method Test organism Response metameter assays, valid assays a5 1 plate diffusion Bacillus subtilis zone diameter 1 1 ATCC 6633 Sarcina lutea 5 2 (1) ATCC 9341 2 nephelometric Staphylococcus aureus turbidity 6 5 (1) NCTC 6571 3 turbidimetric S. aureus transmittance 5 5 ATCC 6538-P 4 plate diffusion B. pumilus 19 17 NCTC 8241 square of zone B. cereus diameter 22 20 (3) ATCC 11778 5 plate diffusion B. subtilis zone diameter 25 24 ATCC 6633 6 turbidimetric S. aureus absorbance 7 7 ATCC 6538-P 7 plate diffusion S. lutea zone diameter 11 11 ATCC 9341 8 plate diffusion B. cereus square root area; 8 8 (1) NCTC 10320 zone diameter 8 8 (2) B. pumilus zone diameter 16 16 (2) NCTC 8241 a The figures in parentheses refer to assays with departures from parallelism of borderline signi- ficance (0.01 <P <0.05). from the calculation and 4-7 doses of each prepara- tion, in the most effective part of the response curve, were finally used to obtain estimates of potency. In order to reach a decision about the evaluation of these data, the results of each assay were analysed several times, including or excluding doses near the end of the effective range; the square-root transfor- mation of responses was also tried. No way could be found to obtain parallelism of the log dose-res- ponse lines for one of the six assays, which was there- fore rejected. Three of the remaining assays of this laboratory showed significant curvature of the lines, but were not excluded because the potencies did not differ greatly from those obtained when certain doses were included or excluded, or when a square-root transformation was used. The log potencies (Mi) for individual assays were tested for homogeneity, within each laboratory, using the formula: X2 = £WI(M1-M)M where the weights (WI) are the reciprocals of the variances of the log potencies and M is the weighted mean log potency. Sets of log potency estimates were homogeneous for each of the Laboratories 1, 5, 6, and 7 when tested in this way-i.e., the probability of X' was greater than 0.05. Weighted means were calculated for these laboratories, together with weights equal to the sums of weights for the separate assays. Log potencies from Laboratories 2 and 3 were shown to be heterogeneous by the X2 test. Unweighted mean log potencies were evaluated and assigned weights equal to the reciprocals of the variances of these means, the variances being calculated direct from the distributions of individual log potencies. For Laboratory 4 there was heterogeneity among the 20 assays performed with Bacillus cereus. This resulted from variations within the set of 6 assays carried out on one of the 4 days of testing; assays on the other 3 days were homogeneous. An unweighted mean log potency was obtained for the heterogeneous set. As it was homogeneous with the weighted values for the other 3 days, it was combined with them to give a single value for the whole group of 20 assays. This potency was also homogeneous with the weight- 7 231 J. W. LIGHTBOWN ET AL. ed log potency obtained from the 17 assays carried out with Bacillus pumilus, and a final weighted value for Laboratory 4 was calculated by combining the results from the two test organisms. The design for Laboratory 8 was such that each of the two test organisms was used on two different days. On each of these days 4 replicate assays were carried out for each of 2 separate sets of weighings and dilutions of the preparations under comparison. There were thus 16 assays for each organism, in sub- sets of 4. The assays using B. pumilus formed a homogeneous set but the assays using B. cereus, which were homogeneous within the subsets (weigh- ings), were heterogeneous between sets. The overall potency value obtained with the latter organism was therefore estimated as the unweighted mean log po- tency for the 4 subsets, and a weight was estimated from the " between weighings " variance. The log potency value thus obtained was homogeneous with the mean log potency obtained with the other organ- ism and was combined with it to give the final weighted mean potency for Laboratory 8. The results of the homogeneity tests for each labo- ratory are given in Table 2 together with the final mean potencies and their 95% confidence limits and total weights. The final log potencies for the 8 laboratories were homogeneous and were combined to give an overall weighted mean potency ratio of 1.002 with confi- dence limits (P = 0.95) of 0.999-1.005, i.e., 996 "ug "/mg (993-998). In addition to comparing the proposed interna- tional reference preparation with the FDA working standard, Laboratories 5 and 7 also compared it with their own national standards. The results of these comparisons are given in Table 3, details of the assay methods being the same as those given for the corresponding laboratories in Table 1. DISCUSSION The final mean potencies obtained by the 8 labo- ratories participating in this collaborative assay agreed remarkably well, being within a range of approximately 2%, but it should be recognized that individual potencies extended over a range of 40%. A certain degree of heterogeneity within laboratories was observed, but this was no greater than is usually found with this type of high-precision assay of anti- biotics. The potency of the proposed international reference preparation did not differ significantly from that of the FDA working standard, and it was sug- gested that the international unit be defined so as to be equivalent to the " Mug " of activity defined by the FDA standard. If the international unit were defined in this way, it would not differ significantly in value from the units defined by the other two national standards that were available. The potency of the Table 2. Results of valid assays No. of P value for x2 Mean Weight 95% confidenceLab. Category assays homogeneity potency eig limits 1 3 > 0.05 0.990 68 134 0.973-1.008 2 5 < 0.001 0.986 1 125 0.860-1.129 3 5 < 0.001 1.008 89 095 0.993-1.024 4 B. pumilus 17 > 0.05 1.005 1 153 835 1.004-1.006 B. cereus a 20 > 0.05 1.005 332 913 0.997-1.013 combined 37 > 0.05 1.005 1 486 748 1.001-1.009 5 24 > 0.05 0.997 36 459 0.973-1.021 6 7 > 0.05 0.995 425 566 0.988-1.002 7 11 > 0.05 0.997 117 652 0.985-1.010 8 B. pumilus 16 > 0.05 0.999 195 528 0.994-1.004 B. cereus b 16 < 0.001 1.002 15 006 0.944-1.064 combined 32 > 0.05 0.999 210 534 0.990-1.010 1-8 124 > 0.05 1.002 2 435 310 0.999-1.005 a Based on weighted means for days 1, 2, and 4 and on a direct estimate for day 3. b Based on a direct estimate of weighted means for 4 weighings. 232 ROLITETRACYCLINE 233 Table 3. Results of assays of proposed international reference preparation against national standards. P value Potency ofValid P Mean Weigh 95% confidence nationalLab. assays homogeneity potency e t limits standardhomogeneity (~~~~~Csg /Mg) 5 25 a > 0.05 1.007 35 381 0.983-1.031 995 7 11 > 0.05 1.019 121 425 1.006-1.033 970 a Including 1 assay with a departure from parallelism of borderline significance. FDA standard was 994 " ug "/mg and, since the proposed international reference preparation was not from the same batch as the FDA standard, con- fusion might have resulted from defining its potency as 994 international units/mg (lU/mg). It was there- fore suggested that a potency of 996 IU/mg should be assigned to the proposed international reference preparation. This proposal was accepted by the colla- borating laboratories. The WHO Expert Committee on Biological Standardization (1969) established the material as the International Standard for Rolitetra- cycline and defined the International Unit of roli- tetracycline as the activity contained in 0.001004 mg of the International Standard for Rolitetracycline. RItSUMIP ETALON INTERNATIONAL DE ROLITtTRACYCLINE On a etabli un 6talon international de rolitetracycline pour r6pondre a un souhait exprime par les laboratoires nationaux de treize pays sur les vingt consultes a ce sujet. Le mat6riel 6tudie, en provenance de la R6publique federale d'Allemagne, a e reparti en ampoules au National Institute for Medical Research, de Londres, et sounmis a un titrage comparatif avec l'6talon de travail de la Food and Drug Administration (FDA) des Etats- Unis d'Am6rique. Huit laboratoires de six pays ont par- ticip6 & cette 6tude collective au cours de laquelle on a procede a un total de 133 epreuves par diffusion et turbi- dim6trie portant sur sept micro-organismes differents. Bien qu'on ait releve des variations allant jusqu'a 40% entre les activit6s mesurees dans un meme laboratoire, les activites moyennes obtenues dans les divers labora- toires se sont r6v6l6es remarquablement homogenes, les ecarts ne depassant pas 2% environ. L'analyse de 1'en- semble des r6sultats a montr6 que I'activit6 de la prepa- ration de r6f6rence proposee ne differait pas sensiblement de celle de 1'6talon FDA. Les participants au titrage comparatif sont donc convenus d'accorder a l'unite internationale une activite equivalente a celle du a,g d'activit6 * d6fini par 1'etalon FDA et d'attribuer a la pr6paration de r6f6rence proposee une activite de 996 Ul/mg. En consequence, le Comit6 OMS d'experts de la Standardisation biologique (1969) a constitue ce materiel en 6talon international de rolit6tracycline et a d6fini l'unit6 internationale de rolit6tracycline comme I'activite de 0,001004 mg de 1'etalon international. REFERENCES WHO Expert Committee on Biological Standardization (1964a) Wld Hlth Org. techn. Rep. Ser., No. 274, p. 10 WHO Expert Committee on Biological Standardization (1964b) Wid Hlth Org. techn. Rep. Ser., No. 293, p. 9 WHO Expert Committee on Biological Standardization (1965) Wld Hlth Org. techn. Rep. Ser., No. 329, p. 7 WHO Expert Committee on Biological Standardization (1969) Wld Hlth Org. techn. Rep. Ser., No. 413, p. 10 J. W. LIGHTBOWN ET AL. Annex I PARTICIPATING LABORATORIES National Biological Standards Laboratory Department of Health Canberra, Australia (Dr L. F. Dodson) Biologics Control Laboratories Laboratory of Hygiene Department of National Health and Welfare Ottawa, Ontario, Canada (Dr L. Greenberg & Miss K. M. Fitzpatrick) Division of Biological Standards National Institute for Medical Research London, England (Dr J. M. Bond & Mr. P. Isaacson) Farbwerke Hoechst AG Frankfurt a. M., Federal Republic of Germany (Dr A. Haussler & Dr K. H. Wallhauser) Paul-Ehrlich Institute Frankfurt a. M., Federal Republic of Germany (Dr G. Siefert) National Institute of Health Tokyo, Japan (Dr H. Umezawa) Bristol Laboratories International Corporation Syracuse, N.Y., USA (Dr P. J. van der Laan) Food and Drug Administration Department of Health, Education, and Welfare Washington, D.C., USA (Dr W. W. Wright) 234