Micromethod for Assaying Serum Levels of Erythromycin

Transcription

1 APPuT MICROBIOLOGY, Jan. 1969, p Copyright 1969 American Society for Microbiology Vol. 17, No. 1 Printed in U.S.A. Micromethod for Assaying Serum Levels of Erythromycin STANLEY C. BELL, J. WAYNE HAMMAN, AND WALTON E. GRUNDY Scientific Divisions, Abbott Laboratories, North Chicago, Illinois Received for publication 13 November 1968 A micromethod for assaying serum levels of erythromycin is described. The assay had the following characteristics: detection of 0.03 to jsg/ml, a long-range curve which minimizes sample dilution, 0.04 ml for a single measurement (0.3 ml required for full coverage), and utility for a variety of body fluids. The method employs radial diffusion from small paper discs which were saturated by capillary action rather than by dipping or pipetting. Although the method was designed to handle serum in which the sample volume is limited, statistical analysis demonstrated that the method has satisfactory precision for routine use. A study of 10 consecutive assays indicates a precision of 412% at the 95% confidence limits. The method of least squares was used to calculate the line of best fit, and the statistics were developed on this basis. The assay method is applicable to a variety of antibiotics. Numerous methods for microbiological assay of antibiotics in small volumes of serum and other body fluids have been proposed: tube dilution (9), dilution in capillary tubes (7), linear diffusion in capillaries (8, 13), radial diffusion from paper discs (4, 5, 10), radial diffusion from agar wells (2), and microtube dilution (1). Several of these methods are not readily adaptable to routine use because of their laborious nature. In addition, the limited range of the standard curve may require one or more dilutions of the samples before assay. The assay for erythromycin described in this paper fulfills the requirements of a method of acceptable accuracy and simplicity which is readily adaptable to other antibiotics. MATERIALS AND METHODS Standard. The working standard of erythromycin base was dissolved in methanol to a concentration of 1,000,ug/ml. This solution was diluted further to 100 pzg/mi in phosphate buffer, ph 7.0 (KH2PO4, 4.0 g; K,.IPO4, 13.6 g; water to 1,000 ml). This reference standard is stable for at least 1 week under refrigeration. Dilutions containing 0.25, 0.50, 1.0, 2.5, 5.0, 10.0, 20.0, and 40.0 jg/ml were made in distilled water from the buffer solution (100pug/ml). These solutions were then diluted with pooled normal human serum to concentrations of 0.025, 0.05, 0.10, 0.25 (reference), 0.50, 1.0, 2.0, and 4.0,pg/ml. A reference standard of either 0.5 or 1.0,gg/ml may be used in place of 0.25 pug/mi if the higher concentration is more appropriate for the samples being assayed. Inoculum preparation. The test organism Sarcina lutea ATCC 9341 was maintained by biweekly transfer on agar slants of Antibiotic Medium one (Difco). Growth was washed from a slant with 3 ml of sterile distilled water. This suspension was used to inoculate a 2-liter Povitsky bottle containing 300 ml of the same medium supplemented with 7 g of agar per liter. The bottle was incubated for 24 hr at 28 C. Growth was harvested by washing with 20 ml of phosphate glycerol buffer (KH2PO4, 0.79 g; K2HPO4, 1.0 g; glycerol, 150 ml; water to 1,000 ml; ph adjusted to 7.0 4t 0.1). The turbidity of the suspension was adjusted so that a 1:200 dilution of the bulk suspension gave a 45% light transimission on a Beckman Model B spectrophotometer, sensitivity no. 3 at 650 nm. Cell count was approximately 4.5 X 109 per ml for the bulk suspension. The optimal amount of suspension used in the assay was determined by trial assay plates. Usually 0.08 to 0.10 ml of the concentrated suspension was used to inoculate 1 liter of agar medium. Small portions of the bulk suspension can be frozen at -60 C, stored at that temperature, and used for at least 1 year (11, 12). Preparation of assay plates. Antibiotic Medium, one with ph adjusted to 8.0 before sterilization, was cooled to 48 C and inoculated with S. lutea. Quantities (10 ml) of the inoculated medium were poured into 100- mm sterile plastic petri dishes on a level surface and allowed to harden. The use of both base and seed layers offers no significant advantage. The plates were refrigerated at 4 C for at least 30 min before use and used only on the same day. Collection of serum. Blood from subjects was collected either in test tubes or capillary tubes and proc- 88

2 VOL. VOL.17, ASSAYING SERUM LEVELS OF ERYTHROMYCIN essed to obtain serum. If the serum must be maintained for more than 1 or 2 days before assay, freezing is preferable to refrigeration. Plating procedure. One-quarter-inch (6.35 mm) paper discs (E-740; Schleicher & Schuell Co., Keene, N.H.) held by forceps were saturated by capillary action and placed on opposite sides of the agar plate. Dipping the discs into the serum was not satisfactory. The reference discs (0.25,g/ml) were placed in position 90 from the first two discs, a total of four discs per plate. This was done to prevent overlap of the resulting inhibition zones from samples of high potency. All levels of the standard curve were plated as described above, six replicate plates per level. Samples were plated similarly, three replicate plates per sample, six discs of sample, and six of reference standard where volume of sample allows. If the sample is limited, as many discs as can be saturated are used. Assay plates were inverted and incubated at 30 C for 16 to 18 hr. Reading and manual calculation. Zones of inhibition were read to the nearest 0.25 mm by projection or other suitable method. The reference zone measurements from the plates of the standard curve were averaged, and that average was used to adjust for plate variation in the 0.025, 0.05, 0.1, 0.5, 1.0, 2.0, and 4.0 Jzg/ml standards. The adjusted zone sizes plus the reference were used in the construction of the standard curve. The curve was drawn with the zone diameters plotted arithmetically and standard concentrations logarithmically on a 3-cycle semilog scale. The zones from the test samples were adjusted, and potency values were read directly from the standard curve. Experimental designs. An initial study (A) was conducted to determine the precision of the saturated disc assay method for serum with a varied number of plates and zones per plate. The antibotic standard concentrations used were as described above. Five plates with four zones per plate were used at each of the standard concentrations. This was repeated for each of the 3 days. In study A, an analysis of variance was made at each of the six concentrations to determine the relative magnitude of variation of zones and plates. The plateto-plate variance component was determined to be very slight, i. e., most plate-to-plate variation can be accounted for by zone-to-zone variation. Therefore, in subsequent analyses of precision, the total number of zones was considered irrespective of the number of plates. Also, a day-to-day variance component was considered to be not relevant, since a new reference standard was used each day, and unknowns were related to the standard curve for that day. Subsequent to study A, this assay method was put into general usage. The precision from 10 consecutive routine blood level assays, study B, was determined and compared with that of study A. The standard antibiotic concentrations used were the same as those for study A. Also five plates with four zones per plate were used at each concentration. Statistical methods. A linear relationship was established between mean zone diameters of inhibition TABLE 1. Erythiromycin (ug/ml) Comparison ofzone sizes for two methods of applying serum to discs Zones of inhibition diameter (mm) Capillary method Pipetting method against the log of the antibiotic concentration of each standard. The statistical technique used was that of linear regression, i.e., fitting a line by the method of least squares. Brownlee (3) provides methodology for the determination of confidence limits when the regression line is used in reverse. This is the case when the antibiotic concentration is estimated from an average zone diameter. Since the log of the concentration is used for the abscissa, the assay error can best be expressed in per cent as delineated by Finney (6). More precise estimates of concentration can be obtained by the use of methods of polynomial regression as delineated by Bennett et al. (2) and the use of computers for curve fitting. Further study has been carried on to develop such methods; the gain in precision is worth the additional effort required. RESULTS Several methods were investigated to determine the best procedure for applying a constant amount of serum to the discs. Pipetting a precise volume was slow and laborious. The amount of serum was limited to 0.02 ml because larger amounts were taken up too slowly or tended to flood the discs. Saturating the discs by capillary action proved to be more reproducible and faster than dipping the discs and draining the excess. A direct comparison of discs, containing 0.02 ml of serum applied with a pipette and discs saturated by capillarity, is presented in Table 1. Each zone measurement in the table represents an average of at least 10 observations. The results indicate a larger sample volume by the capillarity method as well as an easier procedure than pipetting. An experiment was carried out to measure the degree of variation in the absorption of serum as determined by the weights of individual discs. Each of two operators saturated 20 discs by the capillary method. The average of 20 discs of the pooled serum was so close for each of the two operators that no component of variation existed between them beyond disc-to-disc variability. The grand average for the 40 discs was determined to be g per disc with a standard deviation of 89

3 90 BELL, HAMMAN, AND GRUNDY APPL. MICROBIOL. Assay no. TABLE 2. Zone diameter (mm) for each point ofstandard curve from the 10 consecutive assays of study B Concn (ug/ml) a c' Elz 0N Avg Reference..&p W..-..&...P..A.w.% ERYTHROMYCIN CONCENTRATION, MCG/ML FIG. 1. Composite curve from 10 consecutive blood level assays ofstudy B The inherent variation reflected for a single disc was determined to be 14% deviation from the grand average to 95% confidence limits. When more discs are used, an averaging effect is expected. For example, the inherent variation of serum weights from the average is expected to be reduced from 14% for a single disc to 6% when six discs are used. Although this variation in serum absorption among discs is not directly translatable to assay error, it is important that this variation be minimized. Average zone size data of study B are presented in Table 2. The intent of this table is to give laboratory personnel a basis for approximating zone sizes at various concentrations and to indicate the variation in zone sizes between days. The grand average -figures are those used in the construction of the composite standard curve (Fig. 1). Such a curve can be used for manual calculation with proper adjustment accorded to the reference value on any given day. TABLE 3. Summary statistics for standard curve in erythromycin microdisc assay studies No. of Men Ma b Standard Study zones for MeancMean Slo p error Stuy aznesr for c diameter (mm/lgso of estcusnes (Jig/ml) (mm) dose) imate (mm) A B Graphical and statistical calculations demonstrate an excellent linear relationship of average zone size to the log concentration over the range of 0.05 to 2.0 gg/ml. This is the range over which the statistical computations were made, although the graphical relationship (Fig. 1) is shown over range of to 4.0 IAg/ml. In study A, 120 zones were used in computing a standard curve (20 at each of the 6 concentrations). Computed results were highly compatible for the 3 days. Results for the standard curves of study B likewise were homogeneous. The standard errors of estimate for study A and study B were determined to be of similar magnitude (0.63 and 0.54 mm, respectively). Data from both studies are given in a pooled form in Table 3. By using the information from the regression analysis of study B we can ascertain the expected precision of an assay for a specified number of zones. Such precision for predicting the concentration of the unknown in the proximity of the midpoint of the standard concentrations used was determined. The equations referred to in Brownlee (3) were solved to determine the precision when the number of zones is varied in the determination of the concentration of the unknown. For a single zone, the assay error confidence

4 VOL. 17, 1969 ASSAYING SERUM LEVELS OF ERYTHROMYCIN 91 TABLE 4. Microdisc assays of other antibiotics employing one-quarter inch (6.35 mm) paper discs Antibiotic Culture ATCC Medium voaumeg Concn m units orog/ml IncubatCon. ~~~~~~ml Penicillin G B. subtilis 6633 Difco 1, ph unad ,0.05, 0.10, 0.25, 30 justed 0.50b, 1.0, 2.5, 5.0 Penicillin V B. subtilis 6633 Difco 1, ph unad ,0.10,0.25, 0.50b, 30 justed 1.0, 2.5, 5.0 Chlortetracycline S. epidermidis Difco 2, ph adjusted ,0.02,0.04,0.08, 30 to 5.7 prior to ster- 0.16, 0.32b, 0.64, ilization Oxytetracycline S. epidermidis Difco 2, ph adjusted , 0.20, 0.40b, 0.6, 30 to 5.7 prior to ster- 0.8, 1.6, 2.0, 3.0 ilizationc Tetracycline HCl S. epidermidis Difco 2, ph adjusted , 0.20, 0.4ObO, 0.6, 30 to 5.7 prior to ster- 0.8, 1.6, 2.0 ilizatione Ampicillin S. lutea 9341 Difco 1, ph adjusted ,0.20,0.40, 0.60b, 37 to 8.0 prior to ster- 0.80, 1.0, 2.0 ilizationd a Plate, 100 mm. breference concentration. c With HaP04, 14.7 M. d With KOH, 10 N. limits in the estimation of antibiotic concentration is of the magnitude of 30%; for six zones (three plates and two zones per plate), it is of the magnitude of 12%. Although the precision given above is for the proximity of the midpoint of the standard concentrations, the increase in error was not substantial at the extreme concentrations. With four zones the 14% error computed at the midpoint increases to 15% in the vicinity of 0.05 and 2.0 pg of serum per ml. This can be explained by the fact that the slope of the curve is very large relative to the zone variation exhibited. Hence, assay error is changed very little over the range of standard concentrations. The assay technique is applicable to a variety of other antibiotics (Table 4). It appears that the use of polynomial regression and a computer for curve fitting is readily adaptable to several, if not all, the antibiotics. DISCUSSION The system described has sensitivity of 0.03 to 0.035,ug of serum per ml, with a curve up to 4.0,gg/ml that reduces the need for intermediate dilutions. Small sample volume, 0.04 ml per disc, makes it possible to obtain a satisfactory estimate on a limited sample or a more precise figure with increased coverage. The capillary saturation procedure is adaptable to large-scale routine use of theassayin blood serum studies without the tedium associated with the use of pipettes for applying samples and standards to the paper discs. The procedure can also be used to assay materials other than serum; namely, spinal fluids, ocular fluids, pleural fluids, and ear fluids to obtain estimates with known statistical precision. In some of these applications, it may be desirable to employ a diluent other than serum for the standard curve, i.e., a diluent which more closely resembles the material being assayed. Although the method described was developed to handle serum where the sample size was limited, it has proved generally applicable for antibiotic assays. A total of three assay plates, six reference discs and six unknowns have been adopted as standard practice. If greater precision is desired, the cylinder plate method with larger samples may be employed. However, the macromethod requires a much greater quantity of serum. The assay method is readily adapted to computer calculation; it eliminates the labor of manual calculation with its attendant errors. The method of least squares can be used to calculate the line of best fit; the statistical data were developed on this basis. The method of polynomial regression can be used to better fit the curve to the data. The micromethod is generally adaptable to antibiotics other than erythromycin. Sufficient data were developed with six antibiotics to determine the sensitivity and the range of the standard curves. Although the method does not provide a 100-fold range for all the antibiotics, it does have

5 92 BELL, HAMMAN, AND GRUNDY APPL. MICROBIOL. a range of at least 25-fold, thus minimizing the need for making dilutions. If the assay is run routinely and data are accumulated to construct a composite standard curve, it is useful to employ the composite curve with only a reference value obtained from the assay plates. The reference is placed in position on the composite curve graph, and a line is drawn through the reference point parallel to the composite curve. Samples are calculated from the parallel curve. This technique is convenient when assaying small numbers of samples or confirming previous assays when a high degree of precision is not required. ACKNOWLEDGMENT We thank S. Liisa Pennanen and Darlene Davis for technical assistance. LITERATURE CITED 1. Abramowicz, M., J. 0. Klein, D. Ingall, and M. Finland Levels of penicillin in serum of newborn infants. Am. J. Diseases Children 111: Bennett, J. V., J. L. Brodie, E. J. Benner, and W. M. M. Kirby Simplified, accurate method for antibiotic assay of clinical specimens. Appl. Microbiol. 14: Brownlee, K. A Statistical theory and methodology in science and engineering. John Wiley & Sons, Inc., New York, p DE Beer, E. J., and M. B. Sherwood, The paper-disc agar-plate method for the assay of antibiotic substances. J. Bacteriol. 50: Ericsson, H Assay of antibiotics in small amounts of fluid. Scand. J. Clin. Lab. Invest. 12: Finney, D. J Statistical method in biological assay. Hafner Publishing Co., Inc. New York, p Fleming, A Micro-methods of estimating penicillin in blood serum and other body fluids. Lancet 2: Fujii, R., M. Grossman, and W. Ticknor Micromethod for determination of concentration of antibiotic in serum for application in clinical pediatrics. Pediatrics 28: Rammelkamp, C. H A method for determining the concentration of penicillin in body fluids and exudates. Proc. Soc. Exptl. Biol. Med. 51: Sokolski, W. T., and G. M. Savage A microtechnique for the assay of novobiocin in whole blood or serum, p In H. Welch and F. Marti-Ibanez (ed.), Antibiotics annual, Medical Encyclopedia, Inc., New York. 11. Tanguay, A. E Preservation of microbiological assay organisms by direct freezing. Appl. Microbiol. 7: Tanguay, A. E., J. C. Martin, and M. H. Blanchard Survival of microbiological assay organisms after prolonged storage at -40 C, p In C. F. Koda (ed.), Developments in industrial microbiology, vol. 6. Am. Inst. Biol. Sri., Washington D. C. 13. Torii, T., and H. Tojima On a micromethod of assaying blood penicillin level. J. Antibiotics (Tokyo) 2: Downloaded from on March 8, 2019 by guest