RESEARCH/PRACTICE REPORTS

PREDICTION OF GLOMERULAR FILTRATION RATE USING AMINOGLYCOSIDE CLEARANCE IN CRITICALLY ILLMEDICAL PATIENTS BarbaraJ. Zarowitz, Sylvie Robert, and Edward L. Peterson

OBJECTIVE: The aim of this preliminary investigationwas to evaluate the use of aminoglycosideserum concentrationsas a surrogate measure of the glomerular filtration rate (GFR) in comparison with other measured and empiric methods against inulin, the criterion standard measure of GFR. DEStGN: A consecutive sample of all eligible patients. SETIING: An eight-bed medical intensivecare unit in a universityaffiliatedtertiary-care teaching hospital. PATIENTS: Ten critically ill medical patients receiving gentamicin or tobramycin for presumed or documented gram-negativebacillary infection were enrolled in the study. The patients were mechanically ventilatedand had underlyingorgan system dysfunction. All ten patientscompleted the study. INTERVENTION: Patients underwent renal functional assessment by measured inulin (CI;n) and 24-hour urinary creatinineclearance (CI.,.). Aminoglycosideserum concentrationswere used to estimate GFR and were compared with the two measured methods and a creatinineclearance calculated with the Cockcroft-Gaultmethod (CIClJ All evaluations were performed the same day. RF.5ULTS: CI;n averaged 51.6 ± 35.0 mL/min and serum creatinine ranged from 0.3 to 5.4 mg/dL (26.5 to 477.3 umol/L), Steady-state peak and trough aminoglycosideconcentrationswere 6.1 ± 1.4 and 1.3 ± 0.9 J.lg/mL, respectively. There were no statistically significant differencesbetween the various methods, although the aminoglycoside-calculated GPR (Clamg ) 95 percent confidence intervals were smaller than Cl., and ClcGcompared with CI;n' Mean absolute errors were smaller with Clamg than with Cl., and CICI ; . Regression results indicated that only Clam. and ClcG demonstratedagreement with Cl; (lines not different from y=x). However, the Clamg showed closer agreement, with a mean square error almost half that of ClcG (9.6 vs. 18.1). CONCLUSIONS: Clam. can be used routinely as an estimate of GFR in critically ill patients, with less error than empiric methods.

AnnPharmacother 1992;26: 1205-10. BARBARA J. ZAROWITZ, Phann.D.. FCCP. is the Clinical Manager. Department of Pharmacy Services. Henry Ford Hospital. and an Adjunct Associate Professor of Pharmacy Practice. Wayne State University. Detroit. MI; SYLVIE ROBERT, D.P.H.• Phann.D .• BCPS, was a Pharmacy Research Fellow at Henry Ford Hospital and Wayne State University; she is now an Assistant Professor of Pharmacy, Universite de Montreal, and a Clinical Researcher, Centre de Recherche de l'Hotel-Dieu de Montreal; and EDWARD L. PETERSON. Ph.D .. is a Senior Biostatistician, Division of Biostatistics and Research Epidemiology. Henry Ford Hospital. Reprints: Barbara J. Zarowitz, Pharm.D.• FCCP. Department of Pharmacy Services, Henry Ford Hospital. 2799 West Grand Blvd., Detroit, MI 48202.

ACCURATE ASSESSMENT OF RENAL FUNCTION is desired in critically ill patients to facilitate the development of optimally individualized drug therapy regimens. Organ system dysfunction; malnutrition; hypoperfusion secondary to compromised cardiac output, dehydration, or renal vascular disease; and direct renal injury due to nephrotoxins predispose critically ill patients to compromised, dynamically changing renal function that is difficult to characterize. t·4 The use of urinary creatinine clearance (CIcr) measures to estimate glomerular filtration rate (GFR) are inherently flawed, as creatinine is both filtered by the glomerulus and secreted into the tubules. This often leads to increasingly disparate overestimations of GFR by urinary CI.:r measures as renal function declines.t" Additionally, the use of serum creatinine in existing formulas or nomograms may produce grave overestimates of GFR. Patient immobility and reduced muscle mass, which decrease creatinine production, yield low serum creatinine values and high estimates of Cl cr.9,l o Rising serum creatinine values, consistent with deteriorating renal function, lag behind functional decline as creatinine elimination and half-life become prolonged, limiting the accuracy of CI.:r estimates of GFR from serum creatinine." Thus, critically ill patients represent a population in which accurate estimates of renal function are necessary but difficult to obtain. A thorough review of complicating factors and considerations has been published." Aminoglycoside antibiotics are primarily eliminated unchanged in the urine following glomerular filtration." Active renal tubular secretion and biliary excretion are minor elimination pathways, even when the GFR is low," As a result of the predominance of glomerular filtration, aminoglycoside clearance has long been considered to be a clinical measure of the GFR. It is common practice for pharmacists to adjust dosage regimens of renally eliminated medications based on the degree of renal impairment evident from the pharmacokinetic assessment of aminoglycosides. Although the validity of this practice has been assessed previously in surgical and general intensive care units, using Cl., (both filtered and secreted) as a marker of renal function, neither evaluation used a criterion standard measure of GFR.I4-16 The purpose of this preliminary investigation is to report the comparative accuracy of the GFR calculated

The Annals ofPharmacotherapy • Downloaded from aop.sagepub.com at Bobst Library, New York University on January 28, 2016

1992 October, Volume 26

•

1205

from aminoglycoside clearance, 24-hour urinary Cl cr ' and estimated Cl., with the gold-standard measure of inulin clearance as a measure of the GFR. Inulin, which is filtered but neither secreted nor reabsorbed, serves as an excellent marker substance for the assessment of the GFR.17,18

Methods We evaluated ten critically ill patients in a medical intensive care unit. All ten were receiving gentamicin or tobramycin as part of their normal course of therapy and were participating in a study of inulin used for the assessment of renal function." The inulin study methodology was approved by the institution's human rights committee and written informed consent was obtained from all patients. These ten patients represented all of the inulin study participants who were receiving therapy with an aminoglycoside antibiotic. Patient entry criteria for the inulin study included a patent indwelling urinary catheter and the availability of an arterial line for blood access. Patients were excluded if they were receiving trimethoprim/sulfamethoxazole, Hj-receptor antagonists, flucytosine, aspirin, vasopressors, were oliguric «400 mlzd), were undergoing dialysis or ultrafiltration, or were fluid-restricted.P'" All nutritional components, laboratory values, intake/output, and drug therapy were carefully documented for two days preceding, during, and two days following investigation. On the morning of the investigation, urine collection containers and urinary catheters were completely emptied. The inulin clearance technique was carried out according to well-described, standard procedures." The patients were placed in a supine position to receive a loading dose of inulin 50 mg/kg iv over ten minutes followed by an inulin maintenance infusion calculated to maintain an expected inulin plasma concentration of 0.25 mg/mL (0.05 umol/l.) whe~: dose = (C",) (Ch)

nine clearance(mUmin), Cp mid = plasma concentration at the midpoint of the collection interval (mg/dL), and time = collection interval (min). The mean of the 30-minute inulin clearances was used for subsequent comparison with the 24-hour Cl; and the Cl cGcalculated by Cockcroft and Gault." The lesser of total body weight or ideal body weight and either actual or normalized serum creatinine values of 1.0 mg/dL (88.4 umol/l.), when actual creatinine values were < 1.0 mg/dL, were used in the calculated ClcGequation." Some patients had falsely low serum creatinine values due to malnutrition or age causing decreased creatinine production, rather than enhanced elimination. In these patients, a creatinine value of 1.0 mg/dL was used rather than the actual creatinine concentration to test the improvement in predicted C~,. Ideal body weight was calculated by the method of Devine." Blood samples for aminoglycoside serum concentrations were drawn, on the same day as the inulin and creatinine studies as ordered by the physician for therapeutic purposes, as a peak at one hour following a 3D-minute infusion of the dose, and a trough within one hour of the next administered dose. Each dose of gentamicin or tobrarnycin was delivered in intravenous admixtures of the dose in 100 mL of NaCi 0.9%. The prescribed doses were administered over 30 minutes by gravity infusion and the lines were completely cleared after the infusion was completed. Patients were assessed to have steady-state drug concentrations as determined by half-lives obtained from the most recent pharmacokinetic assessment. These blood samples were obtained as a part of routine care. The serum was harvested and analyzed in duplicate by fluorescence polarization with interday and intraday coefficients of variation of ~3 percent at I, 4, and 8 ug/ml., The concentrations obtained were used to calculate steady-state parameters of the elimination rate constant (k.), volume of distribution (Vd ) , and maximum (Cmu ) and minimum (C nUn) concentrations by the method of Sawchuk et al." The GFR from aminoglycoside serum concentration data was calculated as: (k-O.015) gentamicin Clamg = - - (mL/min) 0.00285 (k-O.OI) tobramycin C1amg = - - (mlzrnin) 0.0031

dose = inulin maintenance infusion rate (mg/min) C",= steady-state inulin plasma concentration (0.25 mg/mL, 0.05 Jlrn0I!L) ClcG= Cl., estimated by Cockcroft-Gault equation (mUmin).JO After an initial equilibration period of 60 minutes, three 30-minute urine collections were performed, one following another, and the inulin infusion then was discontinued. All other intravenous solutions were continued throughout the procedure in order to maintain adequate urine output. A 10-mL heparinized blood sample was obtained at baseline prior to inulin administration and 5 minutes before the midpoint of each 3D-minute collection period (i.e., at 10 minutes), according to our inulin protocol. Blood samples were centrifuged and the plasma was harvested and frozen along with a urine aliquot from each collection period, at _70°C in cell culture tubes. Immediately following the end of the third 30minute urine collection period, a 24-hour urine collection was initiated. A blood sample was obtained at the midpoint of this period for determination of plasma creatinine and blood urea nitrogen. Total urine volume was measured and an aliquot frozen, with the midpoint plasma sample, for future analysis. Plasma creatinine concentrations were analyzed by an enzymatic method on the Ektachem 400 analyzer (Eastman Kodak, Rochester, NY). Urine samples were assayed for creatinine by a rate-dependent modification of the Jaffe reaction with an Astra-4 Beckman creatinine analyzer (Beckman Instruments, Brea, CA).3' Linear ranges for serum and urinary creatinine were 0.05-14 mg/dL (4.42-1237.6 umol/L) and 0-400 mg/dL (~35360 umol/L), respectively. The coefficients of variation for serum and urinary creatinine measurements were