Radioimmunoassay for Measurement of Gentamicin in Blood

3 downloads 0 Views 775KB Size Report
Jan 31, 1973 - The Division of Clinical Pharmacology, Department of Medicine, Toronto General Hospital,. Toronto, Ontario, Canada. Received for publication ...
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1973, p. 585-589 Copyright 0 1973 American Society for Microbiology

Vol. 3, No. 5 Printed in U.SA.

Radioimmunoassay for Measurement of Gentamicin in Blood W. A. MAHON, J. EZER, AND T. W. WILSON The Division of Clinical Pharmacology, Department of Medicine, Toronto General Hospital, Toronto, Ontario, Canada

Received for publication 31 January 1973

Gentamicin toxicity has been shown to be related to high concentrations in Because there is a narrow range between its therapeutic and toxic levels, serial monitoring of gentamicin is the most reliable method of guiding therapy. Microbiological assays commonly in use do not afford the desired speed and accuracy, and results may be difficult to interpret in the presence of other antimicrobials. Hence, a rapid, sensitive, and highly specific radioimmunoassay for measurement of gentamicin in serum has been developed. Antibody to gentamicin was raised in rabbits by using a gentamicin-albumin conjugate. Tritiated gentamicin (specific activity 1.0 Ci/mM) competes with unlabeled gentamicin for binding sites on the antibody. Dextran-coated charcoal separates the unbound from antibody-bound gentamicin. Serum levels of gentamicin are determined by comparison with a standard curve. This method can detect concentrations as low as 0.01 yg/ml. Results of a 24-tube run can be obtained in 1 h, thus allowing modification of gentamicin dosage to advantage.

serum.

Gentamicin has been found to be a useful antibiotic for the management of serious infection due to aerobic, gram-negative bacilli. However, it has a relatively narrow therapeutic index, because toxicity to the drug has been found at concentrations abve 12 ug/ml (8); a common therapeutic peak concentration is 6 to 8 jg/ml (17). Elimination of the drug is by the kidneys (1, 9), and any reduction in renal function reduces the elimination of the drug and results in a high blood concentration. According to Riff and Jackson (14), one of the principle problems associated with the pharmacology of gentamicin, even with patients who have normal renal function, is individual variability in the rate of elimination. Because of this it seems evident that therapy with this drug should be monitored by measurement of the drug in serum. Antimicrobial assays currently in use may not afford the desired precision and generally require at least 4 to 6 h of incubation. The results from these assays are difficult to interpret in the presence of other antimicrobials, although the use of bacteria resistant to other antibiotics has reduced this problem to some extent. We report the details of a rapid, sensitive, and highly specific radioimmunoassay for the measurement of gentamicin in serum. (This paper was

presented in part at the 12th Interscience Conference on Antimicrobial Agents and Chemotherapy, Atlantic City, N.J., 26-29 September 1972.) MATERIALS AND METHODS Preparation of antigen. Human serum albumin (10 mg; Connaught Medical Research Laboratories) was dissolved in 1 ml of a 40 mg/ml gentamicin base solution (Garamycin, Schering). To this solution, 0.5. ml of water containing 400 mg of freshly dissolved 1-ethyl-3-(3-dimethyl-aminopropyl) carbodiimide hydrochloride (ethyl carbodiimide) (5, 11, 15) was added. The reaction was allowed to take place at room temperature with gentle stirring for 1 h and then was dialyzed against running tap water overnight. Evidence of successful conjugation was obtained by adding labeled gentamicin as a tracer to the reaction mixture and by obtaining counts after dialysis which were 4 to 5 times greater than those produced by the control dialysate containing no ethyl carbodiimide. It was also possible, by using the tracer gentamicin, to calculate the number of gentamicin molecules coupled to one albumin molecule, and this was found to be 15. Immunization procedure. Approximately 2 mg of conjugate was suspended in complete Freund adjuvant (Becton-Dickinson) and injected into toepads and leg muscles of six albino New Zealand rabbits. Intramuscular injections of 2 mg of conjugate were given weekly for 6 weeks, after which the challenge 585

586

t*->~ chlorMpeni

ANTIMICROB. AG. CHEMOTHER.

MAHON, EZER, AND WILSON

dose was reduced to 1 mg every 2 weeks. Gentamicinspecific antibody was detected 4 weeks after the initial immunization, and the titer increasedprogressively thereafter. Purification of tritiated gentamicin. Gentamicin was tritiated by a labeling service (Amersham-Searle, Des Plains, Ill.). The material received from this service was found to contain only 5 to 10% of the radioactivity as gentamicin and was therefore purified. Separation of gentamicin from its impurities was carried out by using Sephadex G-10 gel filtration. Sephadex G-10 (25 g) was suspended in 50 ml of 0.02% aqueous NaCl, placed in a Sephadex column, (15 x 30 cm), and washed with 1 liter of 0.02% NaCl. Raw gentamicin (2 ml) was added to the column and eluted with 0.02% NaCl at a flow rate of 0.5 ml/min. Fractions (1 ml) were collected, and the separation was evaluated by thin-layer chromatography (TLC) by using lower-phase chloroform-methanol-17% ammonium hydroxide (2:1:1, vol/vol/vol) on silica gel. Comparisons were made with pure components Cl, C2, and C10 in the TLC system which showed three radioactive peaks on the radiochromatogram corresponding to the three gentamicin components. Based on the antibacterial activity, the final product had a specific activity of 1.02 Ci/mmol of gentamicin sulphate. Assay procedure. A series of standard concentrations of unlabeled gentamicin was prepared in pooled, normal, human serum. A 0.01 ml portion of the standard solutions was added to 1 ml of phosphatebuffered saline (0.01 M, pH 7.4) to give a 100-fold dilution of the serum. Tritiated gentamicin was added in a volume of 0.05 ml (12 ng) followed by 0.1 ml of diluted antiserum. A solution of antiserum in buffer adequate to bind 40 to 50% of the tritiated gentamicin was used. The final antiserum dilution in the reaction so

mixture of 1:300 to 400 was required to achieve this. The test tubes containing the above mixture were shaken in their racks and incubated for 15 min at 4 C. At the end of the incubation, 0.2 ml of dextran-coated charcoal (Dextran 150) suspended in phosphate-buffered saline was added (7). It was found that this technique removed 98% of the radioactivity in a tube containing no antibody. After mixing, the test tubes were centrifuged for 10 min at 4,000 rpm. The supernatant fluids were transferred quantitatively into counting vials with 12 ml of Bray solution (10). Counting was carried out for 1 min in a NuclearChicago Corp. liquid scintillation spectrometer, and the percentage of 3H-gentamicin bound to antibody was calculated by using a programmable calculator (Hewlett-Packard 9810A). The quantities of tritiated gentamicin used gave a maximum of 5,000 counts/ min. All samples were carried out in duplicate. Results from a 24-tube run could be obtained in 1 h.

RESULTS

Standard curve. Figure 1 shows a standard curve obtained by the addition of known concentrations of unlabeled gentamicin. Unknown samples were determined by comparing the percentage bound with the standard curve, and the final concentration was obtained by multiplication by 100. Specificity. Specificity of gentamicin antibody was tested separately against the three components of gentamicin, the other commonly used aminoglycosides, penicillins, cephalosporins, lincomycin, and chloramphenicol. The binding capacity of any one of gentamicins Cl, C2, and C was observed to be less than the

> ~~~~~~~~~~~~streptomycin carbenicillin a~~~~~~~~~~ ~noycin ampicillin -_.i _ ~~~~~~~~~lincomycin methicillin penicillin G\ A inmyi ~~~~~~~~~~~~~cepholothh 40 ) >

c

.2 30

E

I-

Gentomicin

20

Qc

C 2~ ~ ~

25

5

10

25

50

100

250

500

10'

2500

~~~~~~~~

10

25000

io-.o

ligand concentration (ng/ml)

FIG. 1. Typical standard curve is shown by the dashed line. The components of gentamicin are shown by UCl, SC2, and AC1. A decrease in specificity to the antibody is evident by the displacement of the component curves to the right at the lower concentrations. The specificity is shown by the solid line. Negligible reduction in antibody binding was shown against concentrations of up to 2.5 mg/ml for kanamycin, and there was no inhibition of binding by streptomycin, neomycin, lincomycin, chloramphenicol, various penicillins, and cephalothin against concentrations of up to 10 mg/ml.

VOL. 3, 1973

RADIOIMMUNOASSAY FOR GENTAMICIN MEASUREMENT

gentamicin mixture at most concentrations (Fig. 1). The aminoglycosides, streptomycin and neomycin, showed no inhibition of the binding of tritiated gentamicin in concentrations up to 10 mg/ml of serum. Kanamycin reduced the binding at concentrations greater than 2.5 mg/ml of serum. The other antibiotics tested, ampicillin, carbenicillin, methicillin, penicillin G, cephalothin, and chloramphenicol did not inhibit the binding of tritiated gentamicin at concentrations up to 100 mg/ml. Comparison with antimicrobial assay. In order to compare the standard antimicrobial assay used in hospitals with the radioimmunoassay, sera from 24 patients receiving gentamicin were divided into two samples. A cup plate assay was carried out by incubation over 6 h in antibiotic medium 11 (Difco), pH 7.9, against Enterobacter. Figure 2 shows the comparison. Both methods were run in duplicate, and each point is a mean of at least two determinations by each assay. The line of regression and its 95% confidence limits are shown. A correlation coefficient of 0.64 was found by regression analysis (P < 0.001); thus, there was no significant difference between the two methods. However, it is evident that in some samples the measurement by one assay was considerably different from the measurement by the other assay. We do not have the complete explanation for these differences, but many of these discrepancies occured in patients receiving more than one antibiotic. It appears likely that the presence of a second antibiotic in serum gave a falsely large zone of inhibition, E

i1

Y= 0 74+1 76 °r 8

_

~

r=064

E=95% c onfid lence

z

v

gentomicin

conce

ntration

radi oimmunoassay

(uq/mnl

FIG. 2. Comparison between measurements of gentamicin in serum by antimicrobial assay and radioimmunoassay.

587

although the organism was deemed resistant to the other antibiotics. This does not explain the few circumstances where the concentration of gentamicin was found to be high by radioimmunoassay and low by antibacterial assay. Reproducibility. Table 1 shows a detailed presentation of the reproducibility results obtained. Each of five different serum pools was divided into four parts to which gentamicin was added to give concentrations of 2.0, 5.0, 7.0, and 10.0 itg/ml. Repeated determinations were made on each of these 20 samples on different assay days. An analysis of variance was carried out at each concentration and revealed that there were no significant differences from the means at any concentration carried out on each of the 5 days. Representative case. Mr. H., a sixty-nineyear-old male, was admitted to Toronto General Hospital because of hemoptysis of 3-day duration. He had a fever of 38 C, intermittently with cough, and persistent hemoptysis. The chest x ray showed left lower lobe consolidation with probable cavitation. After his admission to the hospital and during this investigation, he developed a pleural effusion and became more ill. He was started on gentamicin and cephalothin prior to the availability of culture results. His creatinine clearance was found to be 40 ml/min, and the initial dose of gentamicin therefore was limited to 60 mg. This was given by intravenous infusion over 1 h; Fig. 3 shows the plasma disappearance curve as measured by radioimmunoassay. It can be seen that the peak concentration was 3 Ag/ml and that the calculated half-life in plasma was 3 h. The dose of gentamicin was therefore increased despite the impaired renal function, and he was given 80 mg intravenously at 8-h intervals. Sputum culture showed Pseudomonas species, and cephalothin was discontinued. After three doses, a second study was carried out and is shown in the upper curve of Fig. 3. A peak concentration of 5 ug/ml and a calculated half-life of 2.8 h were found. These studies suggested, therefore, that the gentamicin would not accumulate in the patient's plasma despite his reduced renal function, and it was felt that the peak concentration of 5 gg/ml might be a satisfactory one for the treatment of the patient's lower-lobe pneumonia. Clinically he responded slowly, and sputum cultures became negative, but subsequent investigation revealed that he had a left-lowerlobe, squamous-cell carcinoma. The measurement of gentamicin by radioimmunoassay in this patient revealed that it was possible to increase the dose of gentamicin,

588

ANnMICROB. AG. CHEMOTHER.

MAHON, EZER, AND WILSON

TABLE 1. Gentamicin concentration in human plasmaa Obtained Concn

Expected concn

(pg/mi)

PP1

2 5 7 10

1.88 ± 0.22 5.28 + 0.25 7.20 ± 0.36 9.60 + 0.46

|

PP2

PP3

1.9 + 0.22 5.23 ± 0.50 7.26 + 0.67 9.68 + 0.28

2.00 + 0.31 5.04 + 0.39 7.20 ± 0.50 9.92 + 0.48

PP4

J

PP5

1.94 X 0.17 2.00 ± 0.17 5.12 + 0.28 5.12 ± 0.36 6.98 X 0.34 7.22 + 0.76 10.10 + 0.50 110.00 ± 0.27

PP1 to 5

1.95 + 0.11 5.16 + 0.17 7.17 + 0.22 9.86 + 0.25

aThe concentration of gentamicin found in different plasma pools was carried out at four different gentamicin concentrations. For each pool number, the mean i standard deviation carried out on 4 different assay days is shown. An analysis of variance for each concentration showed no statistically significant difference between means. PP, Plasma pool.

in the serum of children. Furthermore, repeated samples can be taken, and it is relatively easy to carry out individual kinetic studies of the drug based on the decrease in plasma concentration E with time. The rapidity of the assay allows modification of initial dosage to be carried out quickly and, in addition, permits the patient to z be monitored closely and repetitively where changing renal function might require changes in dose of gentamicin. The preliminary report of Lewis et al. (12) on a gentamicin radioimmunoassay differs from this report in several important ways. Notably, no attempt had been made on their part to purify the labeled gentamicin. The presence of radioactive impurities resulted in about 10% more nonspecific binding and caused, as well, a HOURS FIG. 3. Serum concentration of gentamicin in a general loss in assay sensitivity. patient who received two different doses of genWe have shown that about 90% of the raditamicin LV. The lower curve was taken after the oactivity in raw, tritiated gentamicin is due to initial dose of 60 mg. whereas the upper curve was impurities, and our attempt to purify it proved taken after three doses of 80 mg. LV. to be rewarding. in that respect. As a consequence, the steepness of the slope in the standespite his impaired renal function, and to dard curve in Fig. 1 is about three times greater maintain him on a dose which produced an than that reported by Lewis et al. adequate blood concentration. The initial dose Despite the variability of most microbial of 60 mg given at 12-h intervals, which was based assays, their precision was adequate for the on measurement of creatinine clearance, would measurement of antibiotics in serum until drugs have produced levels of gentamicin for consider- that give only a relatively small margin between able periods of time that could have been less toxic and therapeutic blood concentration came than adequate. The initial study of the plasma into use. However, because the precision of disappearance curve enabled rapid adjustments most antimicrobial assays is usually no better in dosage to be carried out, whereas the second than 10% and because the incubation has to be study conducted after three doses of gentamicin carried out over 4 to 6 h, it is difficult to modify revealed that no accumulation was occurring. dosage schedules rapidly. When the patient receives an additional antibiotic, the concentraDISCUSSION tion of gentamicin has to be assessed by using The measurement by radioimmunoassay of an organism resistant to the other drug, and gentamicin has been shown to be a practical, such bacteria may be slow growing, thereby rapid, and precise method for the measurement increasing the delay in measurement of the of gentamicin in serum. Because of the small concentration of gentamicin. volume of 10 uliters of serum required, the test There have been many nomograms devised is applicable for the measurement of gentamicin (3, 4, 6, 10, 13) that facilitate the estimation of ..

VOL. 3, 1973

RADIOIMMUNOASSAY FOR GENTAMICIN MEASUREMENT

589

Studies were funded in part by a grant from Schering suitable dosage adjustments of gentamicin to be used in clinical situations. These nomograms Corporation Ltd., Pointe Claire, Quebec. are usually based on a single-compartment LITERATURE CITED model, with the factor determining gentamicin 1. Black, J., B. Calesnick, D. Williams, and M. J. Weindosage being a measure of renal function. At the stein. 1963. Pharmacology of gentamicin, a new broadpresent time, we have not carried out a comparspectrum antibiotic. Antimicrob. Ag. Chemother. 1964, 138-147. ison of blood concentrations measured by radiA. G. 1960. A simple efficient liquid scintillator for oimmunoassay versus the estimate of steady- 2. Bray, counting aqueous solutions in a liquid scintillation state plasma concentrations which would be counter. Anal. Biochem. 1:279-285. obtained by using nomograms. Furthermore, 3. Chan, R. A., E. J. Benner, and P. D. Hoeprich. 1972. Gentamicin therapy in renal failure: a nomogram for because a comparison of the reliability of the dosage. Ann. Int. Med. 76:773-778. chemical method (16) with that of the radioim- 4. Cutler, R. E., A. Gyselynck, W. P. Fleet, and A. W. Formunoassay has not been carried out, it is not rey. 1972. Correlation of serum creatinine concentrapossible to state which method is the preferred tion and gentamicin half-life. J. Amer. Med. Ass. 219: 1037-1041. one. G., and K. Lempert. 1963. On the mechanism Commercially available gentamicin is com- 5. Doleschall, of carboxyl condensations of carbodiimides. Tetraheposed of three components, and it is therefore dron Lett. 18:1195-1199. interesting to note the lesser binding capacity 6. Gingell, J. C., and P. M. Waterworth. 1968. Dose of gentamicin in patients with normal renal function and found for each of the components, C1, C2, C15, of renal impairment. Brit. Med. J. 2:19-22. gentamicin relative to that of the gentamicin 7. Herbert, V., K. S. Lau, C. Gottlieb, and S. J. Bleicher. mixture at the same concentration. Metabolic 1965. Coated charcoal immunoassay of insulin. J. Clin. breakdown of gentamicin has not been previEndocrinol. 25:1375-1384. ously described but has been suggested from 8. Jackson, G. G., and G. Arcieri. 1971. Ototoxicity of gentamicin in man: a survey and controlled analysis of other studies with 3H-gentamicin from this clinical experience in the United States. J. Infect. Dis. laboratory. However, the possible metabolic 124(Suppl.):S130-S137. degradation products have not been found in 9. Jao, R. L., and G. G. Jackson. 1964. Gentamicin sulfate, new antibiotic against gram-negative bacilli. J. Amer. blood and thus, even if shown to be present in Med. Ass. 189:817-822. urine, should not interfere with the measure- 10. Jelliffe, R W., R. Knight, J. Buell, R Kalaba, and R ment of the drug in the blood. The comparison Rockwell. 1970. Computer assistance for gentamicin carried out between the usual antimicrobial therapy. Clin. Res. 18:441. assay for gentamicin used in this hospital and 11. Khorana, H. G. 1953. The chemistry of carbodiimides. Chem. Rev. 53:145-166. the radioimmunoassay shows that there is 12. Lewis, J. E., J. C. Nelson, and H. A. Elder. 1972. two asbetween the broad, general agreement Radioimmunoassay of an antibiotic: gentamicin. Nasays. However, in some instances there were ture N. Biol. 239:214-216. significant discrepancies between the two as- 13. McHenry, M. C., T. L Gavan, R. W. Gifford, N. A. Gemkink, R. A. van Ommen, M. A. Town, and J. G. says. Most of these occurred in patients who Wagner. 1971. Gentamicin dosages for renal insuffiand this were receiving more than one antibiotic, ciency. Ann. Int. Med. 74:192-197. may well be responsible for the difference. The 14. Riff, L. J., and G. G. Jackson. 1971. Pharmacology of gentamicin in man. J. Infect. Dis. 124(Suppl.):S98exact frequency of such variability and the S105. factors responsible for it have not been fully 15. Sheehan, J. C., and G. P. Hess. 1955. A new method of elucidated. forming peptide bonds. J. Amer. Chem. Soc.

ACKNOWLEDGMENTS We thank J. F. MacDonald, Medical Director, Schering Corporation Ltd., Pointe Claire, Quebec, for advice and

encouragement. A. Waitz of Schering Corporation, Bloomfield, New Jersey kindly provided pure gentamicin powder as well as the separate components of gentamicin.

77:1067-1968. 16. Smith, D. H., D. Van Otto, and A. L. Smith. 1972. A rapid chemical assay for gentamicin. N. Engl. J. Med.

286:583-586. 17. Winters, R. E., and K. D. Litwack. 1971. Relation between dose and levels of gentamicin in blood. J. Infect. Dis. 124(Suppl.):S90-S95.