The second international reference preparation of lymecycline*

Bull. Org. mond. Sante 11973, 48, 81-84 Bull. Wid Hith Org.J The second international reference preparation of lymecycline* J. W. LIGHTBOWN,1 A. H. THOMAS,1 & I. L. BERRYMAN' The second International Reference Preparation of Lymecycline was established. Nine laboratories in 8 countries collaborated in assaying its potency in terms of the first International Reference Preparation, of which stocks had become exhausted. The best estimate ofpotency was taken to be 948 lU/mg, which was the unweightedgeometric mean of a total of 124 assays. On this basis, the International Unit oflymecycline was defined as the activity contained in 0.0010548 mg of the second International Reference Preparation. As stocks of the first International Reference Preparation of Lymecycline were nearing depletion, the WHO Expert Committee on Biological Standard- ization (1969) requested the National Institute for Medical Research, London, to obtain material suit- able to serve as the second international reference preparation and to arrange a collaborative assay. The proposed second International Reference Preparation ofLymecycline (PIRP 2) A 500-g sample of lymecycline from batch No. 944779 was generously donated by Carlo Erba SpA., Milan, Italy, through the good offices of Dr B. Rocchi of that company. The sample, in a single container, was received by the National Institute for Medical Research in June 1969. The following analytical data were supplied by the manufacturer: description yellow, very hygroscopic powder alkalinity pH 8.1 (1% w/v in water) light absorption 280.3 (method of British Pharma- (380 nm) ceutical Codex, 1968) [a] 20 -1950 (0.5% w/v in water)(c]D water 0.94% potency 922 IU/mg (confidence limits 881 to 966 IU/mg) undue toxicity pyrogens conforms to requirements of Brit- ish Pharmaceutical Codex, 1968 From the WHO International Laboratory for Biological Standards, National Institute for Medical Research, Mill Hill, London, England. Division of Biological Standards. 'Statistical Services Section. histamine-like substances free tetracycline less than 1% epimers 4.2% anhydrotetracycline none detectable potency (anhydrous basis) 931.4 IU/mg The sample was kept in its original sealed con- tainer in the dark at -20°C until October 1969, when the material was reduced to powder, mixed with a Turbula mixer, and distributed into approxi- mately 4 200 ampoules of neutral, nonactinic glass so that each contained about 100 mg. The ampoules were fitted with vented polyethylene plugs and kept over phosphorus(V) oxide in vacuo until the mate- rial contained in them had dried to constant weight (5 days). After having been filled with pure dry nitrogen, the ampoules were sealed by fusion of the glass, tested for leaks, and stored in the dark at - 200C. The first International Reference Preparation of Lymecycline (IRP) The first international reference preparation, established by the WHO Expert Committee on Bio- logical Standardization (1969), had a potency of 903 IU/mg. THE COLLABORATIVE ASSAY Of the 12 laboratories invited to take part in the collaborative assay, 9-representing 8 countries -sent data. In the present article, the laborato- ries are referred to by a code number not neces- 2982 -81 J. W. LIGHTBOWN ET AL. sarily bearing any relation to the order in which the laboratories are listed in Annex 1. The participants were requested to determine the potency of the proposed reference preparation in terms of the first international reference preparation, and were supplied with 5 ampoules of each prepara- tion. They were asked to design their biological assays in such a way that each assay should provide evidence of (1) the linearity and parallelism of the log dose-response lines of both preparations, and (2) the precision of the potency estimate. STATISTICAL ANALYSIS Data The 9 laboratories contributed data on 124 bio- logical assays 115 of which were carried out by plate diffusion methods and 9 by turbidimetric methods. No laboratory reported quantitative chemical or physical determinations of the relative lymecycline content of the 2 preparations. Chromatographic examination of the PIRP 2 by Laboratory 2 showed that it contained no detectable anhydrotetracycline or 4-epianhydrotetracycline, and less than 4% of epitetracycline. Methods of analysis All assays were analysed by means of the conven- tional methods for parallel line bioassays-that is, by calculating the relative potency from the horizontal distance between the parallel log dose-response lines of test and standard, and calculating an analysis of variance appropriate to the design of the assay so as to test for the linearity and parallelism of the dose- response lines. For the plate assays from Labora- tories 2, 5, 9, and 11, the square of the zone diameter was used as the response metameter so as to elimi- nate curvature of the dose-response lines; for the 4 + 4 plate assay from Laboratory 11, it was found that the log of the zone diameter gave a linear dose-response line, in contrast to the untransformed or squared responses, both of which gave curved dose-response lines. For the turbidimetric assays of Laboratory 5, log response was found to give better linearity than untransformed response. The turbidi- metric assays from Laboratories 4 and 5 both em- ployed a large number of doses (Laboratory 4 used a 6 + 6 design and Laboratory 5 a 5 + 5 design), and for some assays the dose-response lines were not linear over the entire dose range. Where this situa- tion occurred, a graph of response (or log response for Laboratory 5) against log dose was drawn, and data that were considered not to lie on the linear part of the dose-response line were discarded from the statistical analysis. In the assay performed by Labo- ratory 8, solutions made from replicate weighings from replicate ampoules were included in each assay, but because of the design of the experiments it was not possible to isolate the effects of variations be- tween ampoules. Of the 124 assays in the study, 20 showed signifi- cant departures from parallelism (12 at P< 0.05, 4 at P< 0.01, and 4 at P< 0.001) and the slope of the IRP was steeper than that of the PIRP 2 in 16 of those 20 assays. However, since in exactly half of the assays the dose-response line for the PIRP 2 had a steeper slope than that for the IRP, there did not seem to be any justification for assuming that the two prepara- tions differed significantly in this respect. A number of assays showed significant departures from linear- ity, but no laboratory found that the two prepara- tions produced consistently curved dose-response line. No assays had to be rejected on the grounds of invalidity. The first step in the combination of results was to pool, for each laboratory, the error mean squares from the analysis of variance from assays that had been performed by the same operator or in which the same test organism had been employed. Using these pooled variances, a weight was assigned to the log potency from each assay within a group, and tests were carried out within each laboratory group for the homogeneity of the log potencies and their weights. If the log potencies were homogeneous, the weighted mean was calculated for the group and weighted from the sum of the individual weights; if they were heterogeneous, the unweighted mean of the log potencies was taken and a weight was assigned from the variability between the log potencies. The results obtained in this way for each laboratory are pre- sented in Table 1. A test of homogeneity of these 13 estimates of potency showed them to be heterogeneous (P< 0.001). Because of the small number of labora- tories, it was not practicable to attempt to isolate statistical differences owing to the use of different test organisms or of different assay methods. The unweighted geometric average of all 124 assays was considered to be the best estimate of potency under the conditions of the trial. This value was found to be 0.0212, corresponding to a potency ratio of 1.050, which is equivalent to a potency of 948 IU/mg. Confidence limits (P = 0.95) based on the variability between assays were 942-954 IU/mg. 82 LYMECYCLINE Table 1. Methods and results Labor- Assay method Design Test organism No. of Mean Potency Hetero- Total atory No. assays ratio (lU/mg) geneity weight 1 diffusion 2 + 2 (2 assays) Bacillus pumilus 4 1.011 a 912 a N.S. b 61 467 large plate 3 + 3 (2 assays) NCTC 8241 fish spine beads 2 diffusion 2 + 2 (15 assays) B. pumilus 18 1.072 a 968 a N.S. 318 625 large plate 3 + 3 (3 assays) NCTC 8241 paper disks diffusion 2 + 2 (15 assays) Bacillus cereus 18 1.065 962 P<0.01 111 303 large plate 3 + 3 (3 assays) ATCC 11778 paper disks 3 diffusion 3 + 3 (9 assays) B. cereus 20 1.067 963 P<0.001 47 802 Petri dish 4 + 4 (11 assays) ATCC 9634 punch holes 4 turbidimetric 6 + 6 Klebsiella pneumoniae 5 1.052 950 P<0.001 8 988 ATCC 10031 5 diffusion 3 + 3 Bacillus subtilis 14 1.043 a 942 a N.S. 215 930 Petri dish ATCC 6633 steel cylinders turbidimetric 5 + 5 Staphylococcus aureus 4 1.019 920 P<0.02 9 455 ATCC 6538P Tech- nician 8 diffusion 3 + 3 B. subtilis 10 A 1.023 a 924 a N.S. 689 339 Petri dish 3 + 3 ATCC 6633 10 B 1.036 936 P<0.01 132 330 steel cylinders 9 diffusion 3 + 3 B. cereus 4 1.062 a 959 a N.S. 6766 large plate ATCC 9634 punch holes 11 diffusion 2 + 2 (8 assays) B. pumilus 9 1.050 a 948 a N.S. 82190 large plate 4 + 4 (1 assay) ATCC 14884 punch holes 12 diffusion 3 + 3 B. cereus 4 1.022 922 P<0.05 11 565 Petri dish NCTC 10320 fish spine beads diffusion 3 + 3 B. pumilus 4 1.020 a 921 a N.S. 267 864 Petri dish NCTC 8241 fish spine beads a Weighted value. b N.S. = not significant. ESTABLISHMENT OF THE SECOND INTERNATIONAL REFERENCE PREPARATION AND DEFINITION OF THE INTERNATIONAL UNIT In accordance with the instructions of the Expert Committee on Biological Standardization (1971), the material assayed in the present collaborative study was established as the second International Refer- ence Preparation of Lymecycline and, with the agree- ment of the participants, its potency was defined as 948 IU/mg. The International Unit of lymecycline is defined as the activity contained in 0.0010548 mg of the second International Reference Preparation of Lymecycline. ;UMt DEUXIEME PREPARATION INTERNATIONALE DE REFERENCE DE LYMECYCLINE Les stocks de la premiere pr6paration internationale de reference de lymecycline etant epuises, le Comite OMS d'experts de la Standardisation biologique (1969) a pri6 le National Institute for Medical Research, de Londres, de se procurer un materiel pouvant servir de deuxieme preparation internationale de reference et d'organiser un titrage comparatif. La preparation propos6e, consistant en 500 g de lyme- cycline, a ete repartie en ampoules et a fait l'objet d'un titrage collectif dans neuf laboratoires de huit pays. 83 84 J. W. LEIGBTEOWN ET AL. Les epreuves, utilisant la diffusion sur plaque et la turbi- dim6trie, ont port6 sur huit souches diff6rentes de micro- organismes. L'analyse des activit6s relatives obtenues lors de 124 essais a montre des r6sultats souvent h6t6rogenes dans un m&me laboratoire lors de la r6p6tition des epreuves, et des estimations variables de l'activit moyenne suivant les differents laboratoires. Les participants sont convenus d'6tablir la preparation propos6e en deuxieme preparation internationale de r6f6rence de lym6cycline dont l'activit6 a et6 fixee a 948 UI par milligramme. L'unit6 internationale de lyme- cycHine a ete d6finie comme I'activit6 de 0,0010548 mg de la deuxieme preparation internationale de r6f6rence. REFERENCES WHO Expert C(ommittee on Biological Standardization (1969) Wid flth Org., techn. Rep. Ser., No. 413, p. 12 WHO Expert Committee on Biological Standardization (1971) Wld Hlth Org. techn. Rep. Ser., No. 463, p. 11 Annex PARTICIPATING LABORATORIES National Biological Standards Laboratory Canberra, Australia (Mr N. N. Semple, Miss R. S. Thrum,& Miss H. McInnes) Institute of Hygiene and Epidemiology Brussels, Belgium (Dr A. Vanden Bulcke) Drug Control Department Belgian Pharmaceutical Association Brussels, Belgium (Dr J. Dony & Miss Phien I. Boudru) State Serum Institute Copenhagen, Denmark (Dr J. Bang, Dr 0. B. Jepsen, & Dr 0. Sebbesen) Laboratory for the Control of Pharmaceuticals Helsinki, Finland (Dr E. Nieminen) Carlo Erba SpA. Milan, Italy (Dr B. Rocchi) Drug Control Laboratory Oslo, Norway (Mrs E. D. Aarnes) Institute of Medical Microbiology University of Geneva Geneva, Switzerland (Professor R. H. Regamey, Miss M. Heitz, & Mrs N. Keller) Division of Biological Standards National Institute for Medical Research London, England (Mr P. Isaacson, Miss C. Steven, & Miss P. Newland)