Plasma,82 Microglobulin as a Means of Predicting Gentamicin

Transcription

1 ANTIMICROBIAL AGNTS AND CHMOTHRAPY, Apr. 197, p //4-53/7$2./ Vol. 17, No. 4 Plasma,2 Microglobulin as a Means of Predicting Gentamicin Serum Concentrations J. VALSAMIS,* H. BRAUMAN, AND. YOURASSOWSKY Department of Clinical Biology, Brugmann University Hospital, 12 Brussels, Belgium A novel and accurate method for predicting gentamicin peak serum concentration is described. The method considers body weight and renal function as determined by the concentration of,b2 microglobulin in plasma. In 32 subjects whose renal function ranged from normal to severely impaired, the peak serum concentration of gentamicin was more closely correlated with,b2 microglobulin (r =.69) and with creatinine clearance (r =.69) than with serum creatinine (r =.53). A nomogram was constructed which related plasma /32 microglobulin concentrations and body weight to predicted gentamicin peak serum concentrations. When the nomogram was clinically applied, the predicted peak gentamicin concentrations corresponded closely to the actual measurements obtained (t =.64; P is not significant). We suggest that plasma,b2 microglobulin concentrations compare favorably with serum creatinine values in the prediction of gentamicin peak concentrations in patients with renal impairment. The use of gentamicin is complicated by its narrow margin of safety (4,, 11, 12). Gentamicin is distributed in the extracellular fluid, is not catabolized, and is excreted via glomerular filtration by the kidney (11, 14). Thus, any impairment in renal function can lead to the accumulation of the drug if the administration schedule and the dose are not modified accordingly. A precise knowledge of renal function is therefore highly important. However, a reliable measurement of the glomerular filtration rate (GFR) by means of inulin clearance (Cln) or endogenous creatinine clearance (CCr) frequently cannot be obtained in a short period of time. In addition, the accuracy of clearance collections may be affected adversely by the presence of lower urinary tract obstruction. Usually clinicians rely upon the serum creatinine concentration. This measurement, which is easy to perform, is informative in most cases; however, it may be misleading in predicting the actual GFR. For instance, in elderly or bedridden patients, it tends to overestimate renal function because serum creatinine concentrations are dependent on the muscular mass which is reduced in such patients. From data published in the literature, as well as from our own experience, it appeared that a simple determination of plasma,2 microglobulin (f2m) is a reliable substitute for tedious clearance measurements and gives information more reliable than that obtained from the serum creatinine concentration (1, 17). Since advances in radioimmunoassay have facilitated the precise measurement of circulating gentamicin, it is now possible to rapidly adjust the dosage of that drug. The aim of the present study was to apply plasma,2m measurements in the preparation of a nomogram which reliably estimates peak gentamicin serum concentrations. Our results suggest that fi2m is superior to serum creatinine in this regard. MATRIALS AND MTHODS Details of 32 patients who received gentamicin for various infectious diseases are shown in Table 1. The dose and frequency of administration of the antibiotic in these patients were determined by the scheme proposed by Gingell et al. (7) and are included in the table. Venous blood samples were drawn 1 h after the intramuscular (i.m.) injection of gentamicin when the drug achieves a peak serum level (PSL) (11, 14). In all patients, a venous sample for the simultaneous measurement of gentamicin and l2m was obtained on at least one occasion and in most patients at more frequent intervals on the same or subsequent days. In 3 of the patients, one or more corresponding serum creatinine concentrations were measured. In 16 of the patients, a 24-h CCr determination was available. In 24 patients, a trough serum level (TSL) of gentamicin was also measured on one or more occasions immediately before the next injection of the drug. The number of blood samples from each patient for the various parameters which were measured are presented in lable 1. Heparinized blood was obtained for fl2m determination; creatinine and gentamicin were measured in serum samples. Heparinized samples were immediately centrifuged, and plasma samples were assayed for 21m or preserved at -2 C until assayed. Preservation at -2 C does not alter the results. Serum creatinine and gentamicin were assayed immediately. The Rianen kit was used for gentamicin assay. The sensitivity of the assay as judged from the minimum assayable level in plasma is.2,ug of plasma, 53 Downloaded from on August 14, 21 by guest

2 VOL. 17, MICROGLOBULIN AND GNTAMICIN CONCNTRATIONS 531 diluted 1,1, per ml. The intra-assay coefficient of variation of duplicates is 3.5%, the inter-assay coefficient of variation is 13% and the recovery of known amounts of gentamicin added to plasma approaches 1%. Plasma l2lm was assayed by radioimmunoassay (RIA), using the Phadebac,2m micro test. The minimum detectable level in plasma is far below the minimum plasma concentrations, and plasma was diluted 2 times before the assay. The qualitative aspects of the assay have been reported elsewhere (5). The normal range in a population of 137 men and women was. to 2.9 mg/liter. There was no difference between the results of the two sexes. If there is no modification of renal function, the fl2m level is stable, indicating a constant production. The level tends to increase as Cln, the best known measurement of GFR, decreases. The angular coefficient between log f)2m and log ClG is close to -1 (17); l2m levels are, thus, very representative of GFR. In contrast, although highly significant, the relationship between log CCr and log CIn shows an angular coefficient less than -1, indicating that CCr is less representative of the "true" GFR (17). Creatinine was measured by an automated technique (2). Statistical analyses followed the methods described by Snedecor and Cochran (16). RSULTS Individual gentamicin PSL, TSL, f2m, creatinine, and CCr values are displayed in Table 1. From these data, correlations were calculated and are presented below. The correlation between the gentamicin PSL and serum creatinine is shown in Fig. 1. The coefficient of correlation (r) was.529 and approached the.1 level of probability. There was considerable scattering of the PSLs around the regression line, especially at the lower range of serum creatinine, indicating the relative independence of PSL in the normal or slightly subnormal range of serum creatinine. The 24-h CCr values measured in 16 of the 3 patients receiving mg of gentamicin per injection ranged from to 15 ml/min. The correlation between PSL and the corresponding CCr (r =.695, P <.1) is shown in Fig. 2. Gentamicin PSL was highly correlated with plasma,2m (r =.693, P <.1). The relationship, which is depicted in Fig. 3, resembles that observed with CCr and is an improvement over that encountered with serum creatinine, although the correlation coefficients do not differ significantly. Figure 4 outlines the relationship observed between PSL and TSL in 24 of our patients. The two were closely correlated, and a high degree of significance was observed. The correlation between plasma,2m and CCr was studied in 24 paired determinations from 16 patients (Table 1). The relationship is defined by the expression log,2m = -.77 log CCr (r =.76; P <.1) and exhibits an angular coefficient less than -1. Plasma i2m and serum creatinine were also highly correlated (r =.73; P <.1). The expression log l2m =.65 log creatinine defines the relationship. To consider the body weights of the patients in the analysis, we normalized our data by calculating a factor R (Table 1) according to the following relationship: R = (PSL)/(dose/kg). Thirty-six R values were utilized to define the relationship between R and corresponding f)2m levels and resulted in the following linear equation: R =.53,2m (r =.91; P <.1). From this equation, the dose of the drug to be administered every h to achieve any desired PSL was derived as follows: Dose to be injected (mg) = [desired PSL (ug/ml) x weight (kg)]/ [.53 x l2m (mg/liter) ]. This last equation can be solved for a series of body weights and series of f2m levels for any desired PSL. Table 2 gives the milliliters of gentamicin (1 ml = 4 mg) to be injected for different body weights and the fl2m values required to achieve a PSL of 4,ug/ml. Table 3 provides the same information for a PSL of 7,ug/ml. Tables 2 and 3 are given as examples because they correspond to our current clinical practice. Figure 5 is a graphic representation of the dose of gentamicin, in milligrams, to be given to achieve a PSL of 7 Ag. The reliability of the nomograms has been further tested on 19 occasions in 17 patients. The expected PSLs according to calculation did not differ significantly from the actual measured PSLs (Fig. 6) confirmed by the paired t test (t =.63; P is not significant). DISCUSSION An abundance of literature stresses the importance of a precise knowledge ofrenal function when gentamicin is to be administered. Thus, several schedules of administration have been proposed to avoid the toxic effects of gentamicin (3, 9, 1). Since clearance measures are cumbersome, time consuming, and frequently imprecise, we turned to a simple and seemingly more precise indicator of GFR. From the present data, as well as from previously published observations (1, 17), it appears that /32m offers a good alternative, particularly since gentamicin PSL is more highly correlated with fl2m than with serum creatinine concentrations. In most hospitals, it is not difficult to obtain gentamicin measurements to correct the treatment accordingly. However, it is more appropriate to predict the therapeutic level prospectively to avoid toxic levels. For this purpose, fl2m Downloaded from on August 14, 21 by guest

3 532 VALSAMIS, BRAUMAN, AND YOURASSOWSKY ANTIMICROB. AGNTS CHMOTHR. ~~~.4 q D.Y Q 14 z 5. pa 55SO '5c-B5 5D. as - Q.S =; *s, S.. 1t cz o. _b ".2 1z c~ r._ U, Q_. "% CCY) oo o= co 4 4,-v m 4'-4 P- o L m LO~~~ LO cs - cl co o IV C4 o LO b UN N l t:>c,-1c"'eit >X -_ LO ObO 66;C6 C6 cvi v 4 o v~~lr -1MCli OR cq1 ~o r-4 "m C tzci~6 imu l " r r- r P- LtOOI t- c L 13ul 4 L es L4 4 4 r r- C-i CI Downloaded from It4.6 ms.4 la I= C 1 co to bo CD~ ~ ~ N1 M t- O m C1 N cxo -s N1 t C CqLN CN N t- t-z 4 4 v 4 tt O O t N_c N N VD - o; - - o 4,4 - x x xx x x X x x x CCD C -4 m C4 O -_ u COb CD b. t- LO '-4C w to V-4 3 z C LO -A -- A - q....." O- OrH ; C ( 4r44 C O4r4r4 x x x xx x - cli X X X X CO C" er _.. x c.. on August 14, 21 by guest -a. t5:- C:t- o 9 il C's C _ i Ue CD :- Ca co 9 O O /5; g CT; 't Les Cq ci 4" coq UD to CO Ct- t- -4 C-- LO Cq s z o "-4 Cq1 MI"L CD~ 7-1- P- " P-4 t Cq

4 VOL. 17, 19 #2 MICROGLOBULIN AND GNTAMICIN CONCNTRATIONS 533.) ~ 9:. 3,A~ 2.~c '3 f o o V-- el CD ed 4 e6 c..z* C 4' --e v- - -b C14.s cq cq (O (O 9 LO r4 P-4 s.. (M m oi 6 4 ci 4 Ci v v C'i C'i r4 Les P-4 m _ o~. c I" LC' ' Downloaded from - (M~ ~ ~ v* xx x x x x x xx x on August 14, 21 by guest -C~~~~~~ ~ t ul r- - 2 Co v>1 $4 4) 4),; r. 4 4).>,q - a Q 4) % 4-'-, CL) 1 1 to.d.5. C3.

5 534 VALSAMIS, BRAUMAN, AND YOURASSOWSKY ANTIMICROB. AGNTS CHMOTHR. -g N -W C- ap _- c * r=29 1 pv 1 o Serum creatinine ( mg per 1 ml ) FIG. 1. Relationship between serum creatinine and PSL observed after administration of a fixed dose of mg of gentamicin i.m. regardless of the weight of the patient. ", 6 \;\ ~~~~~ K * \ fo Y=-.37 X 7.3 r =.695 _ p <.1\ 5 1 1S Clearance creatinine (ml/min) FIG. 2. Relationship between PSL of gentamicin and Cc, for patients receiving a fixed dose of mg ofgentamicin i.m. proved to be an excellent tool, particularly since it can be measured in 4 h by radioimmunoassay. However, it should be stressed that in certain special instances f2m may be increased in the absence of any impairment of GFR (6, 13). In these cases, GFR will be underestimated, and lower-than-expected gentamicin serum concentrations will be obtained. There are no previ- S cm Y= 2X r=.693 pc Beta2 microgiobulin ( my/ I) FIG. 3. Relationship between PSL of gentamicin and f32m plasma levels for patients receiving a fixed dose of mg ofgentamicin i.m. -..5_ Y= 135X 3.65 r = 766 P Gentamicin trough serum level (,ug/mil ) FIG. 4. Correspondence between PSL and TSL in patients receiving a fixed dose of mg ofgentamicin i.m. ously described cases of diminished production of l2m, and therefore the danger of overdosage is highly unlikely. In patients undergoing hemodialysis, in which dialyzing membranes porous to f2m (molecular weight 11,) are used,,2m tends to decrease with dialysis but still remains higher than 2 mg/liter. This is the limit above which one should consider only body weight, not renal function, when calculating the dose of gentamicin (Fig. 5). The consideration of body weight is especially 4 Downloaded from on August 14, 21 by guest

6 VOL. 17, 19 2 MICROGLOBULIN AND GNTAMICIN CONCNTRATIONS 535 TABL 2. Gentamicin required to achieve a PSL of 4 pg/ml, according to weight and 132m plasma level Dose (ml)a for the following body wt (kg): #2m (mg/liter) a mg = 2 ml. TABL 3. Gentamicin required to achieve a PSL of 7 pg/ml, according to body weight and,12m plasma level,2m Dose (ml)a for the fouowing body wt (kg): (mg/liter) a mg = 2 ml. important in obese patients because the extra- l2m per liter. It appears that in children the full cellular volume in which gentamicin is distrib- scale of the nomogram can be safely applied, uted is more proportional to the "ideal weight" since the extracellular volume in children is (15). However, introducing this correction in our greater than expected from the body weight (11). PSL predictive formula did not alter its expres- In renal insufficiency, using the nomograms, the sion because our subjects were not obese. Nev- schedule of gentamicin administration every h ertheless, the notion of ideal weight should not may be maintained instead of extending the be neglected. interval between injections as has been proposed In clinical practice, the nomograms are used by others (11). between the lower limit of 4 mg/liter and the For reasons of safety, we chose to express the upper limit of 2 mg/liter. For f2m values lower dose to be given in milliliters rather than millithan 4 mg/liter or greater than 2 mg/liter, the grams (Tables 2 and 3). In our experience, the dose given is calculated by the weight and by risk of confusion by nurses is greatly diminished the fixed limits of 4 or 2 mg/liter. The reason when the dose is expressed in milliliters. of the choice of 2 mg/liter as an upper limit has In conclusion, we present convincing argualready been given. We chose the lower limit of ments in favor of the use of 12m as a parameter 4 mg/liter as a safeguard to avoid potential for the adjustment of gentamicin treatment in overdosage. As long as the problem of nephro- renal failure cases. This parameter seems to be toxicity has not been settled, it is wise to limit more reliable than serum creatinine, especially the dosage of gentamicin according to the weight in elderly and bedridden patients with loss of and to have a GFR corresponding to 4 mg of muscle mass. The use of the proposed nomo- Downloaded from on August 14, 21 by guest

7 _ ~~~~~~~~~ VALSAMIS, BRAUMAN, AND YOURASSso...~~~~ ~~~~ Li....~~~~~~~~...;.; ~ ~~~~~~~~~~~... ~. FIG.5. Nomogram, ti in accountt.h --m.in A-+ ~ ~ ~ t ca t. d.os t g i (n. l...: m.: p curve. Forreasons...; of clrity,... ony ac o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... -.t beta2 microgl...b...n....- i - - plsm leelan th weigh of th pains enabliing' u 6 D.. FIG.~~~~~~~~~~~~~~~~~~~~~~~.oorm takin...i co n ~~~~~~~~~~ : ~~~~~~~~~~~~~~~~~~~~~ ceraina numberofdth weight cuvetheaviensbenarepredsented but gintrediatrea weigth curest cadnabe dorawn by applying the second equation in the text. -g l-~ z CLA iz o Y= 1. X-.3 r =.92 p <.1,. T B MASURO GNTAMICIN SRUM VALUS ug/ml FIG. 6. Relationship between predicted gentamicin serum values and measured serum values. grams has proven to provide PSLs close to those predicted. The number of patients with elevated f2m not explained by impairment of renal function is relatively low in the general hospital and ANTIMICROB. AGNTS CHMOTHR. will not lead to overdosage of gentamicin. Our method allows the safe use of gentamicin in clinical practice. LrTRATUR CITD 1. Brauman, H., J. tienne,. Dupont, J. Van Geertruyden, P. Vereerstraeten, and C. Toussaimt Plasma /32-microglobulin in kidney transplant patients. Acta Clin. Belg. 31(Suppl. ): Chasson, A. L., H. J. Grady, and M. A., Stanley Determination of creatinine by means of automatic chemical analysis. Am. J. Clin. Pathol. 35: Chow, M., J. Deglin, A. Harralson, R. Bartlett, and R. Quintiliani Prediction of gentamicin serum levels using a one-compartment open linear pharmacokinetic model. Am. J. Hosp. Pharm. 35: Dahlgren, J. G.,. T. Anderson, and W. L. Hewitt Gentamicin blood levels: a guide to nephrotoxicity. Antimicrob. Agents Chemother. : vrin, P.., P. A. Peterson, L. Wide, and L. Berggard Radioimmunoassay of,b2 microglobulin in human biological fluids. Scand. J. Clin. Lab. Invest. 2: vrin, P.., and L. Wibell Serum 132 microglobulin in various disorders. Clin. Chim. Acta 43: Gingell, J. C., and P. M. Waterworth Dose of gentamicin in patients with normal renal function and renal impairment. Br. Med. J. 2: Goodman,. L., J. Van Gelder, R. Holmes, A. R. Hull, and J. P. Sanford Prospective comparative study of variable frequency regimens for administration of gentamicin. Antimicrob. Agents Chemother. : Hull, J. H., and F. A. Sarubbi, Jr Gentamicin serum concentrations. Pharmacokinetic predictions. Ann. Intern. Med. 5: Jackson,. A., and Don C. McLeod Pharmacokinetics and dosing of antimicrobial agents in renal impairment. Part i. Am. J. Hosp. Pharm. 31: Jackson, G. G Gentamicin and tobramycin. Antimicrobial therapy, 2nd ed., p W. B. Saunders, Philadelphia. 12. Jackson, G. G., and G. Arcieri Ototoxicity of gentamicin in man: a survey and controlled analysis of clinical experience in the United States. J. Infect. Dis. 124:S13-S Manicourt, D., H. Brauman, and S. Orloff Plasma and urinary levels of /32 microblobulin in rheumatoid arthritis. Ann. Rheum. Dis. 37: Riff, L. J., and G. G. Jackson Pharmacology of gentamicin in man. J. Infect. Dis. 124:S9-S Schawartz, S. N., G. J. Pazin, J. A. Lyon, M. Ho, and A. W. Pasculle A controlled investigation of the pharmacokinetics of gentamicin and tobramycin in obese subjects. J. Infect. Dis. 13: Snedecor, G. W., and W. G. Cochran Statistical methods, 6th ed. Iowa State University Press, Iowa City. 17. Wibell, L., P.. vrin, and I. Berggard Serum,/2 microglobulin in renal disease. Nephron 1: Downloaded from on August 14, 21 by guest