Factors affecting plate assay of gentamicin

Transcription

1 Journal of Antimicrobial Chemotherapy (1977) 3, Factors affecting plate assay of gentamicin II. Media D. C. Shanson* and C. J. Hince Department of Medical Microbiology, The London Hospital Medical College, Turner Street, London El 2AD, England The performance of 11 different media has been studied using Klebsiella NCTC for serum gentamicin plate diffusion assays. Oxoid was the only medium giving reliable results for all 3 types of assay: 5 and at 37 C and 2$ h at 40 C. Two media, and Sensitest, were associated with highly reliable 5 and assays. Six media gave good but slightly less reliable results. Five of these showed good discrimination, while the other, Difco no. 11, had poor discrimination but sharply defined inhibition zones. Oxoid no. 5 and Difco no. 5 media had poor discrimination and gave poor results. The least discrimination and worst results were obtained with medium, with errors of ±69%. Reliable 2\ h assays at 40 C were achieved with agar, and Blood Agar Base. Introduction Plate methods of gentamicin assay are used with incubation over, 5 h (Reeves, 1972) and 2\ h (Shanson, Hince & Daniels, 1976). The purpose of this study was to compare the discrimination, sensitivity, appearance of inhibition zones and of assay with 11 different media at each assay time. The 'correct' values of gentamicin for the experiments were derived from the mean results of assays with medium. This has been shown to produce results very similar to the acetyltransferase assay (Broughall & Reeves, 1975). The media used are tabulated in the Methods section. The Klebsiella NCTC which was used to surface seed all the media was chosen because its antibiotic resistance is useful for gentamicin assays from patients (Waterworth, 1973) and because the same organism was used in a previous study of the effect of diluents on gentamicin assays (Shanson & Daniels, 1975). Materials and methods The media used were: Oxoid Diagnostic Sensitivity Test Agar () CM261 ( ) Oxoid Sensitest Agar, CM409 ( ) Oxoid Agar Base, CM331 ( ) Oxoid Blood Agar Base no. 2, CM271 ( ) Present address: Department of Microbiology, St. Stephen's Hospital, Fulham Road, London, SW10, England, 17

2 18 D. C. Shanson and C. J. Hince Oxoid Antibiotic Medium no. 5, CM335 ( ) Difco Bacto Antibiotic Medium no. 2 ( ) Difco Bacto Antibiotic Medium no. 5 ( 80) Difco Bacto Antibiotic Medium no. 11 ( ) Lab M Sensitivity Test Agar, lab 12 ( ") Lab M Tryptone Soy Agar, lab 11 ( ) Lab M Mueller-Hinton Medium, lab 39 ( ) Thirty ml agar was poured into 100 mm square plastic plates which were levelled, stored at room temperature for 1 to 4 days, and dried for 30 min at the temperature of assay before use. Assay organism and plate preparation Aliquots of an overnight peptone water culture of the assay organism, Klebsiella edwardsii var. atlantae NCTC 10896, were stored at 70 C, and each week a fresh sample was thawed and subcultured on to blood agar. The following day several colonies were emulsified in peptone water and a few drops of this suspension were added to 100 ml nutrient broth or 3 ml peptone water, which were incubated at 37 C overnight. The overnight nutrient broth culture was diluted 1 in 10 in nutrient broth for 2\ h assays and the overnight peptone water culture was diluted 1 in 50 for 5 and assays, to surface seed plates. After seeding plates were drained well and those for 5 and assay left at room temperature for about 15 min. For 2\ h assays seeded plates were dried for exactly 30 min at 40 C. Immediately after drying 7 mm wells were cut. Test sera and standards Standards of Wellcome Horse Serum no. 5 (Inactivated) containing 1-25, 2-5, 5-0, 100 and 20-0 ug/ml gentamicin were used for all the assays (Shanson & Daniels, 1975). Test sera containing 1-9, 2-1, 3-4, 40,, 7-7 and 160ug/ml gentamicin were prepared in pooled normal human serum. Standards and test sera were stored at 20 C. Four standard and 2 test sera were included in duplicate in each plate. Plates for 2\ h assay were incubated at 40 c C and plates for 5 and assay at 37 C. Zone sizes were measured with a millimetre rule on a Leebrook zone reader (Leebrook Instrument Co. Ltd., Sarum Farm, Winchester, Hampshire). A standard curve was constructed by relating the zone diameter to log concentration of gentamicin. Results Reliability of 5 and assays with different media The of human serum gentamicin 5 and assays are shown for each medium in Table I. Reliability was defined as [mean % error -f (2 x standard deviation % error)]. The different media were placed into groups in the table in order of the of the assay results obtained with 9 assays of each of 3 human sera containing 1-9, 3-4 and ug/ml gentamicin. The of human serum gentamicin assays was very good for and Sensitest media (±11%), good (±22%) for Lab M Sensitivity Test, Mueller-Hinton, Difco 11, and Blood Agar Base media, poor (±29%) for Oxoid no. 5 and Difco no. 5 media and very poor (±69%) for medium. Five and assays gave similar results. The results of assays of sera containing 160 ug/ml using an overnight culture diluted 1 in 100 are shown in Table II. The 3 media giving results of good,

3 Media and gentamicin assays 19 Table I. Reliability of results of 5 and human serum gentamicin assays with 11 different media Media Media with very good Sensitest Media with good Lab M Sensitivity Test Agar Lab M Tryptone Soy Agar Mueller-Hinton Difco no. 11 Blood Agar Base Media with poor Oxoid no. 5 Difco no. 5 Media with very poor 80 assay Range iaf results ([ug/ml) for test sera 1-9 ug/ml 3-4 ug/ml ug/ml ^ ^ No. of assays=4 and 5 for each serum assay at 18 and 5 h respectively. s.d.=standard deviation Errors of assay (3 test sera combined) (Mean) (Mean-I- 2XS.D.) 2% 7% 7% 6% 14% 10% 9% 9% 10% 7% 14% 10% 12% 6% 13% % 37% ± 4% ±11% ±11% ± 9% ± 7% ±13% ±16% ±18% ±18% ±19% ±13% ±20% ±11% ±22% ±22% ±18% ±12% ±2 ±2 ±29% ±33% ±69% within ±14%, were, and Sensitest. The other 3 media, Difco no. 11, Oxoid no. 5 and all gave unreliable results. The mean result of assay was higher with medium than with the other media and the standard deviation was the largest. Discrimination, sensitivity and sharpness of edges of inhibition zones Discrimination was defined as the difference in size of the inhibition zones produced by the 1-25 and 100 ug/ml gentamicin standards. The 7 media at to showed good discrimination, in contrast to the 4 media with non-physiological which showed

4 20 D. C. Shanson and C. J. Hince Table n. Reliability of results of assay of 160 ng/ml gentamicin serum with 6 different media Errors of assay Media Mean ± s.d. (ug/ml) Mean % ± error Mean % error + 2X(S.D. % error) Media with good Sensitest Media with poor Difco no. 11 Oxoid no ± ± ± ± ± ±5-9 No. of assays=5 for each serum. s.d.=standard deviation. 2% 8% 4% 18% 20% 33% ± 7% ±14% ±10% ±28% ±39% ±9 Table HI. Characteristics of 5 and gentamicin plate assays with different media Media with good discrimination Lab M Sensitivity Test Agar Mueller-Hinton Sensitest Blood Agar Base Lab M Tryptone Soy Media with poor discrimination Difco no. 5 Oxoid no. 5 Difco no Discrimination* (mean±s.d.) 69±3 64±5 61 ±3 61 ±5 55±3 52±3 51±8 33±1 33±1 29±1 27±1 46±2 42±3 45 ±2 49 ±4 44±2 42±3 37±3 25±1 26±2 25±2 22±1 Sharpness of edge Very good Very good Very good Sensitivityt No. of assays=6 for each serum. s.d. =Standard deviation. 'Discrimination was defined as the difference in zone sizes produced by the 100 and 1-25 ug/ml gentamicin standards. tsensitivity was defined as the diameter of the mean zone produced by the 1-25 ug/ml gentamicin standard. The diameter of the well in the agar=55 units on the zone reader poor discrimination. With all media, discrimination decreased when the assay time was reduced from 18 to 5 h. Sensitivity was defined as the diameter of the inhibition zone produced by the 1-25 Ug/ml gentamicin standard. Sensitivity was greatest with the 3 media at to 80

5 Media and gentamicin assays 21 (Oxoid no. 5, Difco no. 5 and Difco no. 11) and least with Lab M Tryptone Soy Agar (Table III). Although medium was acidic the sensitivity was not reduced compared with the media at normal. Most of the media were associated with double or blurred edges of inhibition zones and varying numbers of colonies within zones after incubation. When double edges appeared the diameters of the outer zones were measured. Difco no. 11 medium gave the best inhibition zones after assay with no colonies within the zones, and very sharp zone edges at both 5 and., Sensitest and Blood Agar Base were the only media associated with good discrimination and sharp edges at 5 or incubation of assays (Table III). 2\ h assays at 40 C Only 3 of the 11 media surface seeded with a heavy inoculum of Klebsiella showed sufficient growth after 2\ h incubation at 40 C for the assays to be read easily. These were, and Blood Agar Base. The discrimination, sensitivity and sharpness of zone edges were similar, but growth was slightly better on Agar. medium gave very sharp zone edges, but growth was light and discrimination poor (Table IV). The of 2\ h assays was poor with but good with, and Blood Agar Base (Table V). Table IV. Characteristics of 2i h gentamicin plate assays at 40 C with different media Blood Agar Base 30±2 3O±2 28 ±2 16±2 Very good No. of assays=6 for each serum, S.D. = Standard deviation. *Discrimination was defined as the difference in zone sizes produced by the 100 and 1-25 ug/ml gentamicin standards. tsensitivity was defined as the mean diameter of the zone produced by the l-25ug/ml gentamicin standard. The diameter of the well in the agar was 55 units on the zone reader. Table V. Reliability of results of 2 h serum gentamicin assays at 40 C with different media Media 2-1 ug/ml Human test sera (ug/ml) (mean ± S.D." 1 40 ug/ml 7-7 ug/ml Mean % ± error (assays of 3 combinled) Mean % error + 2X(S.D. % error) Blood Agar Base 2-2 ±004 20±017 20± ± ± ± ± ±0-42 No. of assays=5 for each serum. s.d.=standard deviation. ± ± ± ±0-27 4% 6% 9% ± 9% ±16% ±1 ±24%

6 22 D. C. Shanson and C. J. Hince Discussion The acetyltransferase enzyme method for the assay of gentamicin gives accurate results and correlates well with an Klebsiella plate assay using Oxoid agar (Broughall & Reeves, 1975). The results of assays were used to compare the results of assays using the other media. The standard deviation of the per cent error of 12 assays was only ± 1 %. Media could be arranged into 4 groups according to the of human serum gentamicin assays, very good, within ±11 %, good, within ±22%, poor, within ±29%, and very poor, within ±69%, using horse serum gentamicin standards. If the experiments had been performed using human instead of horse serum gentamicin standards the expected results would be similar since the standard curves for each serum are virtually identical (Shanson & Daniels, 1975; Jarvis & Leung, 1976). Both media in the first group, and Sensitest, were associated with good discrimination and sharp zone edges in 5 and assays. The best medium in the second group was Lab M Sensitivity Test Agar which had the greatest discrimination but poor zone edges. There was one medium of poor discrimination in the second group with good, Difco no. 11. The other media with poor discrimination, Oxoid no. 5, Difco no. 5 and, were in the third and fourth groups giving unreliable results. Discrimination and sharpness of zone edges both appeared to be important factors determining the of assay results. Media with good discrimination generally gave more reliable results than those with poor discrimination. Very sharp zone edges were seen with Difco no. 11 and this probably compensated to some extent for the poor discrimination so that assay results of good were obtained. The sensitivity of all the media was adequate for clinical serum gentamicin assays. Most serum samples for gentamicin assay are at to 80 which is similar to the of horse and human serum standards and to all the media except, at. This medium was often associated with raised assay results compared with the other media, especially with the sera containing 160 ug/ml gentamicin. had the least discrimination of all the media, and gave highly misleading results. A recent study of factors affecting gentamicin plate assays used as the assay medium, but no figures are given for the errors of assay (Deacon, 1976). In Deacon's study the was adjusted to 9 without any useful increase in discrimination. An alteration of the of Sensitest medium from to 6-9 and in an earlier study also failed to affect the slopes of the gentamicin standard curves (Shanson & Daniels, 1975). In the latter study it was noted that acid changes in the of sera containing gentamicin affected the assay result much more than alkaline changes and this effect was more marked with Difco no. 11 (poor discrimination and ) than with Sensitest medium (good discrimination and ). Sensitest has also been found to be less susceptible to the changes of sera containing gentamicin than Difco no. 5 medium (Jarvis & Leung, 1976). Assays of gentamicin in serum from patients with serious Gram-negative sepsis may be associated with errors unless a suitable medium is used (Shanson, Kensit & Hince, 1975). Higher assay results were obtained with Difco no. 11 compared with or Sensitest media. The differences were both statistically and clinically significant. Experiments where known amounts of gentamicin were added to sera from patients with Gram-negative sepsis showed that the correct assay results were obtained with and Sensitest media. Falsely low serum gentamicin assay results were reported with uraemic

7 Media and gentamicin assays 23 sera (Stressman, Michel, Licht & Sacks, 1975) but this effect was shown to be mediadependent and satisfactory results were obtained with and Sensitest media (Shanson, Kensit & Hince, 1975). Subsequently Stressman and his colleagues repeated their assays with uraemic sera on Difco no. 5 and media and confirmed that medium gave good results (Stressman, Michel & Sacks, 1976). Satisfactory serum gentamicin assay results with uraemic sera were also obtained by Broughall et al., who used medium (Broughall, Pugsley & Reeves, 1975). The results of the present study and the experiences described in the other studies suggests that and Sensitest are the most reliable media for 5 and serum gentamicin assays. Increasing the temperature of incubation of gentamicin assay from 37 to 40 C results in larger and sharper inhibition zones as well as more rapid growth of the Klebsiella (Shanson, Hince & Daniels, 1975). Three media,, and Blood Agar Base were associated with 2\ h assays at 40 C of good in the present study. The most reliable medium for 2\ h assay was agar, which showed the best growth of Klebsiella. The agar was the only medium associated with accurate gentamicin assays for all three conditions of assay, 2\ h at 40 C, and 5 and at 37 C. Acknowledgements We are grateful for the statistical help given by Mr Steven Evans of the Hill Computer Centre, The London Hospital Medical College and to Roussel for supplying gentamicin. References Broughall, J. M. & Reeves, D. S. The acetyltransferase enzyme method for the assay of serum gentamicin concentrations and a comparison with other methods. Journal of Clinical Pathology 28: (1975). Broughall, J. M., Pugsley, D. J. & Reeves, D. S. Potential pitfall in bioassay of serum gentamicin Lancet ii: 1095 (1975). Deacon, S. Factors affecting the assay of gentamicin by the plate diffusion method. Journal of Clinical Pathology 29: 54-7 (1976). Jarvis, J. D. & Leung, T. W. C. Some factors influencing the plate assay of gentamicin. Chemotherapy progress. In Proceedings of the 9th International Congress of Chemotherapy (1975). Plenum Press, New York (1976), Vol. 2, pp Reeves, D. S. Assay of gentamicin. Lancet ii: (1972). Shanson, D. C. & Daniels, J. V. Factors affecting plate assay of gentamicin. I. Diluents. Journal of Antimicrobial Chemotherapy 1, (1975). Shanson, D. G, Hince, C. & Daniels, J. V. Assay of gentamicin and tobramycin by a reliable 2-J- hour Klebsiella plate method. Chemotherapy. In Proceedings of the 9th International Congress of Chemotherapy (1975). Plenum Press, New York (1976), Vol. 2, pp Shanson, D. C, Kensit, J. & Hince, C. Uraemia, Gram-negative sepsis and gentamicin assays. Lancet ii: 875 (1975). Stressman, J., Michel, J., Licht, A. & Sachs, T. Potential pitfall in bioassay of serum gentamicin. " Lancet ii: 563 (1975). Stressman, J., Michel, J. & Sachs, T. Potential pitfalls in bioassay of serum gentamicin. Lancet ii: 100 (1976). Waterworth, P. M. In Antibiotics and Chemotherapy. Garrod, L. P., Lambert, H. & O'Grady, F. W. Eds. Churchill-Livingstone, London (1973), pp (Manuscript accepted 14 April 1976)