ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1990, p. 702-7080066-4804/90/050702-07$02.00/0Copyright © 1990, American Society for Microbiology

Increased Aminoglycoside Dosage Requirements inHematologic Malignancy

RAJA G. ZEITANY,lt* NAGI S. EL SAGHIR,2t CHERUPPOLIL R. SANTHOSH-KUMAR,2AND MARSHA A. SIGMON'

Departments of Clinical Pharmacy' and Medicine,2 King Khalid University Hospital andKing Saud University College of Medicine, Riyadh, Saudi Arabia, 11472

Received 8 August 1989/Accepted 9 February 1990

Aminoglycoside pharmacokinetic parameters were studied prospectively in 27 patients with an underlyinghematologic malignancy and fever associated with neutropenia and in 18 control patients. Pharmacokineticparameters and dosages were determined by linear regression analysis of a one-compartment model by themethod of Sawchuk et al. (R. J. Sawchuk, D. E. Zaske, R. J. Cippolle, W. A. Wargin, and R. G. Strate, Clin.Pharmacol. Ther. 21:362-369, 1976). Significant differences between the study and control groups were foundfor aminoglycoside volume of distribution (0.40 0.1 versus 0.27 0.05 liter/kg [mean ± standard deviation],respectively; P < 0.0001), clearance (116.6 48.9 versus 68.6 26.7 ml/min, respectively; P < 0.0001),half-life (2.27 + 0.66 versus 3.5 1.8 h, respectively; P < 0.0001), and elimination rate constant (0.33 0.11versus 0.24 0.09 h-1, respectively; P < 0.001). The percentage of bone marrow blast cells (at the time ofdiagnosis) in patients with acute leukemia significantly correlated with increased aminoglycoside clearance (R12= 36.98%; P = 0.0001). Patients with stage IV lymphomas (Hodgkins disease and non-Hodgkins lymphoma)had a significantly increased clearance compared with patients with lower stages of lymphomas (105.1 _ 18.5versus 84.1 14.9 ml/min; P = 0.014). Fever, leukocyte count, or chemotherapy, among other clinical andlaboratory parameters that were studied, had no significant correlation or effect on aminoglycoside disposition.The average dose of amikacin required to maintain peak concentrations in serum above 20 ,ug/ml in patientswith a hematologic malignancy was 27.5 8.43 mg/kg per day. Pharmacokinetic parameters and dosages forthe control patients were comparable to general literature standards. We conclude that the dosagesrecommended by the manufacturers or those derived from nomograms underestimate the aminoglycosidevolume of distribution and clearance in patients with a hematologic malignancy and result in suboptimal peakaminoglycoside concentrations in serum. We recommend that in febrile neutropenic patients with anunderlying hematologic malignancy, amikacin be initiated at 7.5 to 10 mg/kg per dose every 8 h (2 to 2.5 mg/kgper dose every 8 h for gentamicin) and adjusted within 24 h based on individual pharmacokinetic analysis.

Aminoglycosides remain one of the most important groupsof antibiotics in the treatment of gram-negative infections(25). To maximize efficacy and minimize toxicity with theseagents, determination of the drug concentration in serum isnecessary (1, 30). Achieving a peak concentration in serum(Cmax) for gentamicin or tobramycin of 6 to 8 ,ug/ml and foramikacin of 20 to 28 ,ug/ml is associated with improvedsurvival in patients with gram-negative bacteremias andpneumonias (1, 17, 25-27). The two major factors thatdetermine the value of the Cma are the dose and the apparentvolume of distribution (Vss) of the drug. The Cm. is directlyrelated to the dose and is inversely related to the V.,Aminoglycoside dosage nomograms as well as manufac-

turer-recommended doses assume a constant Vss of approx-imately 0.25 liter/kg of body weight (3, 6, 16). It has beenobserved that in certain patient groups (i.e., those with burns[41], with cystic fibrosis [18], in intensive care units [14], inthe postoperative period [29]) the Vs, is increased. Dosageschedules based on nomograms and manufacturer-recom-mended guidelines lead to underestimation of the Vss and,

* Corresponding author.t Present address: 1722 North Avalon Boulevard, No. 2, Wilm-

ington, CA 90744.t Present address: Section of Hematology/Oncology, Oakwood

Hospital, Dearborn, MI 48123.

hence, subtherapeutic aminoglycoside Cmax values in thesepatient populations (14, 18, 22, 29, 41).

Optimal aminoglycoside dosage is particularly importantin febrile neutropenic patients. It has been shown thatneutropenic patients with gram-negative bacterial infectionsrequire higher peak bactericidal concentrations in serum toimprove their outcome when compared with those requiredin nonneutropenic patients (34).

Despite the widespread use of aminoglycosides in febrileneutropenic patients with an underlying malignancy, onlylimited data are available on the pharmacokinetic parametersof aminoglycosides in these patients (15; R. P. Manny andP. R. Huston, Letter, Clin. Pharm. 5:629-630, 1986; J. K.Phillips, R. L. Spearing, D. J. Crome, and J. M. Davis,Letter, N. Engl. J. Med. 319:1290, 1988). In a study ofinfectious complications in patients with acute leukemia(10), we adjusted the aminoglycoside dosage by the methodof Sawchuk et al. (32) and observed that doses higher thanthose recommended by the manufacturer were needed inorder to achieve therapeutic aminoglycoside concentrationsin serum (unpublished data). Hence, we set out a prospec-tive controlled study to determine the aminoglycoside phar-macokinetic parameters in patients with an underlying he-matologic malignancy. In this report we describe the detailsof our results and propose new aminoglycoside dosageguidelines for adult patients with an underlying hematologicmalignancy.

702

Vol. 34, No. 5

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

AMINOGLYCOSIDE DOSAGE IN HEMATOLOGIC MALIGNANCY

TABLE 1. Patient characteristics

Patient No. of No. of Age (yr Sex (no. Wt (kggroup patients vations [mean ± SD]) M/no. Fa) [mean ± SD])

StudybAML 7 40 32.3 ± 15.6 5/2 50.2 ± 7.8ALL 3 6 24 ± 19.6 2/1 53.9 ± 11.3CLL 2 4 57.5 ± 20.2 1/1 57.5 ± 8.7NHL 11 19 35.1 ± 14.8 10/1 53.8 ± 14.8HD 4 5 51.4 ± 3.1 2/2 50.8 ± 6.4

Total 27 74 38.2 ± 17.7 20/7 50.8 ± 11.6

Control 18 20 31.6 ± 15.6 11/7 68.3 ± 20.9

a M, Male; F, female.b Abbreviations: AML, acute myeloblastic leukemia; ALL, acute lympho-

blastic leukeniia; CLL, chronic lymphocytic leukemia; NHL, non-Hodgkinslymphoma; HD, Hodgkins disease.

MATERIALS AND METHODS

Patient population. Forty-four febrile neutropenic epi-sodes requiring aminoglycoside therapy in 27 consecutivepatients with an underlying hematologic malignancy be-tween February 1988 and February 1989 were studied.Patients in the study group were newly diagnosed, in docu-mented relapse, or at various stages of their treatment(induction, reinduction, consolidation, maintenance). Eigh-teen randomly selected age- and sex-matched adults receiv-ing aminoglycosides for gram-negative bacterial infections(urinary tract infections, diabetic foot ulcers, osteomyelitis,bacteremia, and hospital-acquired pneumonias) admitted tothe general medical wards, who did not meet the exclusioncriteria, served as controls. Control patients had no under-lying malignancy and were otherwise healthy adults, asidefrom their primary medical problem. All patients were ofArab origin, and their characteristics are described in Table1. Neutropenia was defined as an absolute neutrophil countof <1,000/mm3. Fever was defined as an oral temperature of>38.5°C on one occasion or >38°C on two occasions in a24-h period at least 4 h apart that was unrelated to the use ofdrugs or infusion of blood products.

Patients with ascites, pleural effusions, major surgery,creatinine levels in serum of >196 ,umol/liter, or septic shockprior to the initiation of aminoglycoside therapy were ex-cluded from the study and control groups, as these condi-tions may alter aminoglycoside disposition (24, 39).

Antibiotic treatment. Empiric therapy consisted of intra-venous ceftazidime (2 g every 8 h) in combination withintravenous amikacin (7.5 mg/kg every 12 h) or gentamicin(1.3 mg/kg every 8 h) (actual body weight was used in all butone patient, who was obese; lean body weight was used inthe obese patient). Vancomycin, metronidazole, amphoteri-cin B, or acyclovir was added as indicated.

Pharmacokinetic analysis. Serum samples were collectedafter at least five estimated half-lives of the aminoglycosidehad elapsed, usually 24 h, to ensure that steady-state con-centrations in serum were achieved. A predose sample(Cmin) was taken immediately prior to administration of thenext scheduled dose. A second sample (Cm.) was taken 1 hfollowing the start of a timed 30-min infusion. All drugs wereadministered either in 50 or 100 ml of a 5% glucose solution.Actual times of all serum sample collections and drugadministration were recorded and used for pharmacokineticcalculations. Aminoglycoside concentrations in serum weremeasured by a fluorescent polarization immunoassay (TDx;

Abbott Diagnostics Division, Irwing, Tex.), and analyseswere conducted within 8 h of sample collection. Serumsamples were obtained through peripheral venipunctures inall the control patients and in 64 (87%) instances in the studygroup. The remaining 10 (13%) blood samples were drawnthrough Hickman catheters, as peripheral venipunctureswere technically difficult.

Calculation of individual aminoglycoside pharmacokineticparameters was performed by the method of Sawchuk et al.(32) by using two serum samples. The following equationswere used in the calculation of the pharmacokinetic param-eters: kel = [In (Cm,x/Cmin)]/time difference, V.s = [K (1 -e-keI)]/[kel (Cmx - Cmin e-ke), t1/2 = 0.693/kel, Ko = [kei Vsspeak D (1 - e-ke7)/(1 e-kee), Cma = [KO(l - e-kel)]/[VSke, (1 - e-kej), Cmin = Cmax [e-kel(T )I, and CL = kel V"16.67, where kei is the elimination rate constant, in hour-';Ko is the infusion rate, in milligrams per hour; t is the timeperiod of infusion, in hours; t1/2 is the half-life, in hours; T isthe dosing interval, in hours; peak D is the desired Cmax, inmicrograms per milliliter; and CL is clearance, in millilitersper minute.The results obtained from the initial pharmacokinetic

studies for each patient were used to determine the dosagerequired to achieve a Cma, of >20 and <30 ,ug/ml foramikacin or >6 and <10 ,ug/ml for gentamicin. Desired Cminconcentrations were <8 ,ug/ml for amikacin or <2 ,ug/ml forgentamicin.

After attainment of steady state with the calculated dos-age, determination of the aminoglycoside level in serum wasrepeated by the method described above. If levels in serumwere inappropriate, the dosage was recalculated. Aminogly-coside levels in serum were determined weekly to ensurethat optimal therapeutic concentrations were maintained.Concentrations of ceftazidime and vancomycin in serumwere not measured in any of the patients.

Calculation of pharmacokinetic parameters was con-ducted by using a proprietary spreadsheet computer pro-gram (SuperCalc 4 version 1.0; Computer Associates Inter-national) into which the equations were entered. Along withpharmacokinetic data, relevant clinical and laboratory infor-mation was recorded within the spreadsheet. Informationconsisted of weight; height; age; treatment protocol; numberof days postchemotherapy; temperature; concomitant ad-ministration of parenteral nutrition and amphotericin B; totalnumber of platelet and blood transfusions; creatinine level inserum; leukocyte count; hemoglobin; platelet count; bloodurea nitrogen; and serum glutamic oxalacetic transaminase,serum glutamic pyruvic transaminase, and bilirubin concen-trations. Development of ascites or pleural effusion; positiveblood, sputum, or urine cultures; and, where appropriate,the percentage of bone marrow blast cells (defined as thepercentage of 1,000 counted nucleated bone marrow cells)present at the time of original diagnosis were also noted. Allclinical and laboratory data were recorded concurrently witheach determination of the aminoglycoside concentration inserum in order to correlate any changes in pharmacokineticparameters with the variables mentioned above within eachcourse and between subsequent courses of aminoglycosideadministration.

Nephrotoxicity was defined as a rise of the creatinine levelin serum of >44 ,umol/liter from the base line. Ototoxicitywas not assessed because of the difficulty in performingaudiometric examinations in this patient population.

Creatinine CL was estimated by using the method de-scribed by Cockcroft and Gault (5). Estimation of kei was

VOL. 34, 1990 703

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

ANTIMICROB. AGENTS CHEMOTHER.

based on the following equation: 0.0024 (creatinine CL) +0.01 (9).

Statistical analysis. A two-sample t test (two-tailed) andone-way analysis of variance (ANOVA) were used to deter-mine statistical differences between group means. The Pear-son product correlation coefficient (r), simple regressionanalyses, and coefficient of determination (R12) were used todetermine correlations and associations. A probability of<0.05 was considered significant. All statistical analyseswere conducted on a proprietary statistical software pro-gram (Statgraphics version 2.1, Statistical Graphics Corp.).

TABLE 2. Aminoglycoside pharmacokinetic parametersa

Patient V.. k.t1,2 CLgroup (liter/kg) (h-1) (h) (ml/min)

StudybAMLALLCLLNHLHD

0.41 ± 0.090.40 ± 0.060.35 ± 0.040.40 ± 0.140.38 ± 0.13

0.36 ± 0.070.31 ± 0.040.25 ± 0.060.31 ± 0.070.30 ± 0.02

2.17 + 0.752.23 + 0.302.92 ± 0.782.39 + 0.572.36 ± 0.18

131.4 ± 58.9113.4 ± 38.980.7 ± 18.4

100.6 ± 19.093.0 ± 23.5

Total 0.40 ± 0.10C 0.33 ± 0.11d 2.27 + 0.66e 116.7 ± 48.9f

RESULTS

A total of 74 pairs of determinations of aminoglycosideconcentrations in serum (60 for amikacin and 14 for genta-micin) were performed during 44 treatment episodes in 27consecutive adult patients with a hematologic malignancy.Twenty pairs of determinations of aminoglycoside concen-trations in serum were recorded for the 18 control patients.No patients with an underlying hematologic malignancyadmitted to the hematology-oncology unit for fever andneutropenia met the exclusion criteria or were denied entryinto the study. For analysis of pharmacokinetic parameters,results from patients receiving amikacin and gentamicinwere combined, on account of their virtually identical dis-positions (24). For Cma and dosage analyses, the two groupswere analyzed separately.There were a total of 17 positive cultures obtained in 15 of

the study group patients. Infections included six pneumonias(two gram negative, one gram positive, two fungal, and oneacid-fast bacillus), one urinary tract infection (gram nega-tive), five bacteremias (three gram negative and two grampositive), and five catheter-related infections (four grampositive and one gram negative). In the control group, therewere 14 positive cultures in 12 patients. Infections includedtwo pneumonias (gram negative), six urinary tract infections(gram negative), two bacteremias (one gram positive and onegram negative), one osteomyelitis (gram negative), and onediabetic foot ulcer (mixed infection).

V.S. Aminoglycoside Vrr values were consistently andsignificantly elevated in those patients with hematologicmalignancies as compared with V., values in the controlgroup. Within the group of patients with a malignancy, therewas no statistically significant difference in the mean Vs,between the various diagnoses (ANOVA, P = 0.77; Table 2and Fig. 1). Patients with acute leukemia had several epi-sodes of fever and neutropenia during their treatment whichrequired multiple courses of aminoglycoside therapy. Nosignificant variation in the Vss was observed from course tocourse of therapy.The presence of fever, microbiologically documented in-

fection, chemotherapy administration, leukocyte count(both at nadir and after recovery), or any other clinical orlaboratory data recorded (mentioned above) had a significanteffect or correlation with the Vss.

t112. The t1/2 values were significantly shorter in thosepatients with a hematologic malignancy compared with thet1/2 values in the control group (Table 2). There was nosignificant difference in the mean aminoglycoside t1/2 be-tween patients with the various hematologic malignancies(ANOVA, P = 0.28). The t1/2 had a significant directcorrelation with age (R2 = 22.28%; P = 0.00002) but not withthe Vss (R2 = 4.25%; P = 0.08) in patients with a malignancy.There was no correlation or association between t1/2 and anyof the other clinical or laboratory data recorded.

Control 0.27 ± 0.05C 0.24 ± 0.09d 3.50 ± 1.81e 68.6 ± 26.7fa Values are means + standard deviations. Probabilities were determined

by the two-sample t test." See footnote b of Table 1 for abbreviations.c P = 0.0000004.dp = 0.0005.' P = 0.000005.fp = O.oOW05.

CL. Patients with a malignancy had a significantly greateraminoglycoside CL compared with the CL in the controlgroup (Table 2). Patients with acute leukemia had increasedaminoglycoside CL compared with the CL in other patientswith hematologic malignancies (P = 0.01); however,ANOVA of the CL per diagnosis resulted in a probability of0.055. In those patients with acute leukemia, there was asignificant correlation between CL and the percentage ofblast cells in the bone marrow at the time of diagnosis. Areciprocal simple regression model best fit the data, with R2= 36.98% and P = 0.0001 (ANOVA), although a simplelinear regression model was also significant (Fig. 2). Thenumber of days of drug administration postchemotherapyand, therefore, the subsequent percent reductions in periph-eral and bone marrow blasts had no significant effect on CL.Patients with stage IV Hodgkins disease and stage IVnon-Hodgkins lymphoma had a significantly increased ami-noglycoside CL compared with patients with a lower stage oflymphoma (105.1 ± 18.5 versus 84.1 ± 14.9 ml/miin [mean ±standard deviation], respectively; P = 0.014). Total body CLof the aminoglycoside (116.3 + 48.9 ml/min [mean + stan-dard deviation]) was significantly greater than creatinine CL(102.7 + 27.8 ml/min) in patients with a hematologic malig-nancy (P = 0.04); however, creatinine CL had a significant

15 -

a

* 10-az-: oA

Z .

Normal Range =]Above Noral

0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7

Volume of distribution (L/kg)

FIG. 1. Aminoglycoside V., observed in the group of patientswith a malignancy. A total of 89% of the observations were greaterthan the population average of 0.2 to 0.3 liter/kg.

704 ZEITANY ET AL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

AMINOGLYCOSIDE DOSAGE IN HEMATOLOGIC MALIGNANCY

c

E2:Ei

a

-

0

0

E

250

200

150

100

50

0% 20% 40% 60% 80%Bone marrow blata at presentation

100%

FIG. 2. Linear regression analysis of CL regressed on the per-centage of bone marrow blast cells at presentation. Omission ofoutliers did not significantly influence the value of the correlationcoefficient.

direct correlation with aminoglycoside CL (R2 = 15.79%; P= 0.0006). There were no significant correlations or effectsupon aminoglycoside CL by any of the remaining clinical orlaboratory data recorded.

kei. Patients with an underlying hematologic malignancyhad a significantly increased ke, compared with the ke, values

in the control group (Table 2). There was no significantdifference between the means of the estimated and actual kejvalues for the patients with a malignancy (P = 0.95), andthere was a significant correlation between the two (RJ2 =

13.38%; P = 0.001). Actual kel values had a significant butweak correlation with creatinine CL (R2 = 10.54%; P =

0.005).Dosage and C,,.,,. Those patients with an underlying he-

matologic malignancy who achieved a steady-state Cmaxequal to or above 20 ,g/ml and Cmin values less than 8 ,ug ofamikacin per ml required 27.50 + 8.43 mg/kg per day (Fig. 3)versus 19.48 + 4.77 mg/kg per day in those patients withCm.X values of less than 20 ,ug/ml (P = 0.00003).Of the 14 observations in patients receiving gentamicin,

there were only three with a Cm. ofgreater than 6 ,u g/ml. Inthese three observations, the average dosage was 6.8 mg/kgper day.The mean interval required to maintain Cmax values of

amikacin above 20 ,ug/ml and gentamicin above 6 ,ug/ml was8.67 h, with a median of 8 h and a range of 6 to 12 h.

0

1-A

0

z

10 1S 20 25 30 35 40 45

Daily do.e of amikacin (mg/kg/day)

FIG. 3. Total daily dose of amikacin required by the patientswith a malignancy to achieve amikacin Cm;,, values of >20 ,Lg/ml.Only one observation was recorded in which the patient achievedtherapeutic amikacin Cmax with the recommended dosage of 15mg/kg per day.

Toxicity. No patient had a significant rise in creatinineconcentration in serum that was directly attributed to ami-noglycoside administration. One 60-year-old patient withacute myeloblastic leukemia, who received multiple coursesof high-dose aminoglycoside therapy, did have a rising Cminwithout a rise in the creatinine level in serum. The dosagewas recalculated without further sequelae. Another patientwith non-Hodgkins lymphoma did have a significant rise inhis creatinine level in serum, but he was also receivingamphotericin B. When amphotericin B was withheld for 2days, the creatinine level in serum returned to base-linelevels.

DISCUSSION

Infectious complications account for 50 to 70% of deathsin patients with a hematologic malignancy (2, 4). Aminogly-cosides are widely used in empiric treatment of febrileepisodes in neutropenic patients (33). Because of the narrowrange between subtherapeutic, therapeutic, and toxic con-centrations, appropriate dosage calculation can only beachieved through determinations of the concentration inserum in these patients, who are often seriously ill (23, 30).Several studies have demonstrated that an improved out-come in patients with gram-negative infections is directlyrelated to the Cmax of the aminoglycoside (1, 17, 25-27).Likewise, attainment of an adequate Cm,xfrom the first dosecontributes to therapeutic efficacy (14).

Klastersky (19) has shown that as the absolute neutrophilcount decreases, higher bactericidal activity in serum isrequired for successful treatment. A peak bactericidal activ-ity in serum of at least 1:8 is required in neutropenic patients(absolute neutrophil count, <1,000), while in severely neu-tropenic patients (absolute neutrophil count, <100) peakbactericidal activity in serum of 1:16 is required. At loweraminoglycoside concentrations, synergism between the be-ta-lactam and the aminoglycoside may not occur. This canbe overcome by increasing the dose of the aminoglycoside toattain a higher Cmax (19). Trough aminoglycoside concentra-tions in serum have not been correlated with an improvedoutcome in patients with gram-negative infections (25-27).Dosage nomograms and manufacturer-recommended

doses are derived from either a specific patient populationor, more commonly, healthy adults, rather than the popula-tions for which the drug is intended (8). The aminoglycosidesexhibit wide inter- and intrapatient variabilities in theirpharmacokinetic parameters (38) and are known to be af-fected by various disease states (30). Our data show thatdosages recommended by the manufacturer and those basedon nomograms are not appropriate for patients with anunderlying hematologic malignancy. The Vss and CL valuesof aminoglycosides in our patients with a hematologic ma-lignancy were significantly elevated compared with those inthe control population, who did not have a malignancy. Theuse of manufacturer-recommended doses or those of dosagenomograms underestimated the Vss in patients with a hema-tologic malignancy and, hence, led to a reduced Cmax. The500-mg twice-daily dose of amikacin, which is intended foruse in adult patients with a normal Vss (0.25 liter/kg), wouldresult in a 41% reduction in Cmax if the Vss is increased to 0.4liter/kg (study group mean). If the Vss is 0.67 liter/kg (studygroup maximum), a 63% reduction in Cmax occurs. Peakconcentrations in serum in our patients with a malignancywere raised to optimum levels when the dosage was in-creased nearly 100% over manufacturer recommendations tocompensate for the increased V~s and CL.

r- 0.483R-aquarmd. 2143%

P P0.005 '~~~~~~~~~~

S Normal Rang. z Abova Normal7Z _ _

VOL. 34, 1990 705

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

ANTIMICROB. AGENTS CHEMOTHER.

Usually, as the V.. increases, the t,/2 increases; thisassumes that CL remains constant (39). Our patients with anunderlying malignancy did not have an increase in theaminoglycoside t1l2 with increasing Vs.. The tln was, infact, significantly shorter in the patients with a malignancythan in the control group, despite their increased V.. values.Total aminoglycoside CL from the body exceeded creatinineCL significantly. These observations raise the possibilityof enhanced renal or extrarenal elimination of aminoglyco-sides in patients with an underlying hematologic malig-nancy.As mentioned above, no single factor that we recorded

was associated with the observed increase in aminoglycosideV.S. Altered metabolic states (21) produced by the diseaseprocess may explain the increase in Vs.. An increased CLassociated with a high percentage of blast cells in the bonemarrow at the time of presentation in patients with acuteleukemia and in patients with advanced lymphoma (stage IV)represents a previously unreported finding. It is also ofinterest that none of the pharmacokinetic parameters, in-cluding CL, showed significant changes after chemotherapyadministration. Therefore, it is unlikely that the tumorburden alone is directly responsible for the observed alter-ations in aminoglycoside disposition. Increased CL may beassociated with the severity or advanced state of the disease,but in order to establish a significant correlation, a largeprospective trial analyzing CL and complete remission oroverall survival would be required. We are, however, reluc-tant to draw any firm conclusions at this time from the smallsample of patients that we studied.The TDx assay (Abbott) has been shown to result in

aminoglycoside pharmacokinetic parameters which differslightly from those obtained by radioimmunoassay (31). Thiswas not a factor in this study since the TDx was also used inthe control group. We used two blood samples to determinethe pharmacokinetic parameters. The two-sample method iscomparable to the three- and four-sample methods describedby Sawchuk et al. (32), and no difference has been shown toexist between the two methods if appropriate steps are taken(35). We realize that the two-sample method, despite theeconomic and time-saving benefits, is more apt to result inerroneous pharmacokinetic parameters than are the three- orfour-sample methods if timing of administration and sam-pling are inaccurate. Nevertheless, we must point out that inall our patients, we waited until steady state was achievedand strict timing, both for administration and sampling, wereadhered to.The calculated parameters for the aminoglycoside V.,

CL, and tj,2 in patients without a malignancy correspondedto the results obtained in Western centers for similar patientgroups (28). The primary purpose for the inclusion of acontrol group in the study design was to eliminate ethnicdifferences as a cause of alteration in aminoglycoside dispo-sition. Since the pharmacokinetic parameters obtained in thecontrol group were similar to those obtained in Westernpatients, there is no reason to believe that patients of MiddleEastern origin have distinct aminoglycoside dispositions.Most reported studies on empiric treatment of neutropenic

febrile episodes, using an aminoglycoside, make no mentionof pharmacokinetic analysis to optimize aminoglycosidetherapy (11, 12, 20, 36). Some investigators have usedamikacin at a dose of 15 mg/kg per day administered every 6or 12 h, with the average adult receiving 250 mg four timesdaily or 500 mg twice daily (11, 12, 20, 36). It was pointed outby the authors ofthe European Organization for Research onTreatment of Cancer (EORTC) trial III (20) that there were

no specific study-wide criteria for evaluating aminoglycosidelevels in blood to ensure adequate concentrations of amika-cin in serum. Some of the treatment failures in the amino-glycoside, beta-lactam-containing arm of the studies men-tioned above may have been due to subtherapeuticaminoglycoside Cm,. If peak concentrations in serum arelow, then the goal of combination therapy to achieve maxi-mal synergistic and bactericidal activities is not fulfilled (19,33). Our findings raise a question as to the results of studieswith aminoglycosides in patients with a hematologic malig-nancy in which the aminoglycoside dosage is not individu-alized.One center, participating in the EORTC trial III, estab-

lished a protocol to maintain the amikacin Cm.., above 25,g/ml. In this center, the average patient required 28.7 mg ofamikacin per kg per day to maintain the predeterminedconcentrations in serum (20). A previous study, which wasconducted at the same center and which used a pharmaco-kinetically derived amikacin dosage, revealed that theirpatients with granulocytopenic cancer required an averagedose of 29 mg/kg per day (7, 13). Another study (15)observed increased V.. as well as CL values in their adultand pediatric patients with cancer. However, as mentionedby the investigators, there were several limitations withinthe study design. Two recent letters (Manny and Huston,Letter, 1986; Phillips et al., Letter, 1988) noted increased V..values in their patients with a variety of hematologic andoncologic disorders. Our findings and these observations (7,13, 15, 20; Manny and Huston, Letter, 1986; Phillips et al.,Letter, 1988) establish that aminoglycoside Vs, CL, anddosage requirements in patients with an underlying hemato-logic malignancy are increased.During the later part of our study, when empiric treat-

ment with an aminoglycoside was required for a febrileneutropenic patient with an underlying hematologic malig-nancy, amikacin was initiated at a dosage of 7.5 to 10 mg/kgper dose every 8 h (if the patient was obese, the dosage wasinitiated by using lean body weight). This dosage compen-sates for the increased V.. and CL and establishes maximalaminoglycoside concentrations in serum at the initiation oftherapy. The dosage was subsequently individualized within24 h.We observed no signs of nephrotoxicity from these large

doses. None of the patients had a significant rise in creati-nine in serum, despite multiple courses of high-dose amino-glycoside therapy. It has been shown that the use of individ-ual patient pharmacokinetic parameters to optimize theaminoglycoside dosage is not only associated with improvedoutcome but decreased toxicity as well (30, 37, 38, 40).

In summary, we reported aminoglycoside pharmacoki-netic analysis to show increased V, CL, and dosage re-quirements in patients with an underlying hematologic ma-lignancy. Our findings indicate that doses recommended bythe manufacturer and those of nomograms underestimateaminoglycoside V.. and CL. Based on our results, werecommend that higher doses (7.5 to 10 mg of amikacin perkg per dose every 8 h) be used initially and adjusted within 24h, according to individualized pharmacokinetic analysis, infebrile neutropenic patients with a hematologic malignancy.We suggest similar guidelines for gentamicin, with initialdoses of 2 to 2.5 mg/kg per dose every 8 h. Studies withaminoglycosides for the treatment of febrile neutropenicpatients with a hematologic malignancy which do not explic-itly describe in detail individualized aminoglycoside dosageshould be interpreted cautiously.

706 ZEITANY ET AL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

AMINOGLYCOSIDE DOSAGE IN HEMATOLOGIC MALIGNANCY

ACKNOWLEDGMENTS

We thank Robert Buchanan of the Clinical PharmacokineticsLaboratory, for processing our samples in a timely manner, and thenursing staff of the hematology-oncology unit for their excellentcooperation in timing drug administration and sampling.

LITERATURE CITED

1. Barza, M., and M. Lauermann. 1978. Why monitor serum levelsof gentamicin. Clin. Pharmacokinet. 3:202-215.

2. Bodey, G. P. 1975. Infections in cancer patients. Cancer Treat.Rev. 2:89-128.

3. Chan, R. A., E. J. Benner, and P. D. Hoeprich. 1972. Gentami-cin therapy in renal failure: a nomogram for dosage. Ann.Intern. Med. 76:773-778.

4. Chang, H. Y., V. Rodriguez, G. Narboni, G. P. Bodey, M. A.Luna, and E. J. Freireich. 1976. Causes of death in adult patientswith acute leukemia. Medicine (Baltimore) 3:259-268.

5. Cockcroft, D. W., and M. H. Gault. 1976. Prediction of creati-nine clearance from serum creatinine. Nephron 16:31-41.

6. Cutler, R. E., A. M. Gyselynck, W. P. Fleet, and A. W. Forrey.1972. Correlation of serum creatinine concentration and genta-micin half-life. J. Am. Med. Assoc. 219:1037-1041.

7. De Jongh, C. A., J. C. Wade, S. C. Schimpff, K. A. Newman,R. S. Finley, P. C. Salvatore, M. R. Moody, H. C. Stadiford,C. L. Fortner, and P. H. Wiernik. 1982. Empiric antibiotictherapy for suspected infection in granulocytopenic cancerpatients: a comparison between the combination of moxalactamplus amikacin and ticarcillin plus amikacin. Am. J. Med. 73:89-96.

8. Dettli, L. C. 1973. Translation of pharmacokinetics to clinicalmedicine. J. Pharmacokinet. Biopharm. 1:403-418.

9. Dettli, L. C. 1974. Drug dosage in patients with renal disease.Clin. Pharmacol. Ther. 16:274-280.

10. El Saghir, N. S., A. D. Hall, C. R. Santhosh-Kummar, andR. G. Zeitany. 1989. Infectious complications during therapy ofacute leukemia in Saudi Arabia. J. Hosp. Infect. 14:209-215.

11. EORTC International Antimicrobial Therapy CooperativeGroup. 1978. Three antibiotic regimens in the treatment ofinfection in febrile granulocytopenic patients with cancer. J.Infect. Dis. 137:14-29.

12. EORTC International Antimicrobial Therapy CooperativeGroup. 1987. Ceftazidime combined with a short or long course

of amikacin for empirical therapy of gram-negative bacteremiain cancer patients with granulocytopenia. N. Engl. J. Med.317:1692-1698.

13. Finley, R. S., C. L. Fortner, C. A. De Jongh, J. C. Wade, K. A.Newman, E. Caplan, J. Britten, P. H. Wiernik, and S. C.Schimpif. 1982. Comparison of standard versus pharmacokinet-ically adjusted amikacin dosing in granulocytopenic cancer

patients. Antimicrob. Agents Chemother. 22:193-197.14. Hassan, E., and J. D. Ober. 1987. Predicted and measured

aminoglycoside pharmacokinetic parameters in critically ill pa-

tients. Antimicrob. Agents Chemother. 31:1855-1858.15. Higa, G. M., and W. E. Murray. 1987. Alterations in aminogly-

coside pharmacokinetics in patients with cancer. Clin. Pharm.6:963-966.

16. Hull, J. H., and F. A. Sarubbi. 1976. Gentamicin serum concen-

trations: pharmacokinetic predictions. Ann. Intern. Med. 85:183-189.

17. Jackson, G. G., and L. J. Riff. 1971. Pseudomonas bacteremia:pharmacologic and other bases for failure of treatment withgentamicin. J. Infect. Dis. 124(Suppl.):S185-S191.

18. Kearns, G. L., B. C. Hilman, and J. T. Wilson. 1982. Dosingimplications of altered gentamicin disposition in patients withcystic fibrosis. J. Pediatr. 100:312-318.

19. Klastersky, J. 1986. Concept of empiric therapy with antibioticcombinations; indications and limits. Am. J. Med. 80(Suppl.5C):2-12.

20. Klastersky, J., M. P. Glauser, S. C. Schimpff, S. H. Zinner, H.Gaya, and the European Organization for Research on Treatmentof Cancer Antimicrobial Therapy Project Group. 1986. Prospec-tive randomized comparison of three antibiotic regimens forempirical therapy of suspected bacteremic infection in febrilegranulocytopenic patients. Antimicrob. Agents Chemother. 29:263-270.

21. Kreuzer, F., and S. M. Cain. 1985. Regulation of the peripheralvasculature and tissue oxygenation in health and disease. Crit.Care Clin. 1:453-470.

22. Lesar, T. S., J. C. Rotschafer, L. M. Strand, S. D. Lynn,and D. E. Zaske. 1982. Gentamicin dosing errors with fourcommonly used nomograms. J. Am. Med. Assoc. 248:1190-1193.

23. Mann, H. J., D. W. Fuhs, R. Awang, F. A. Ndemo, and F. A.Cerra. 1987. Altered aminoglycoside pharmacokinetics in criti-cally ill patients with sepsis. Clin. Pharm. 6:148-153.

24. Mathews, A., and G. R. Bailie. 1987. Clinical pharmacokinetics,toxicity and cost effectiveness analysis of aminoglycosides andaminoglycoside dosing services. J. Clin. Pharm. Ther. 12:273-291.

25. Moore, R. D., P. S. Lietman, and C. R. Smith. 1987. Clinicalresponse to aminoglycoside therapy: importance of the ratio ofpeak concentration to minimal inhibitory concentration. J. In-fect. Dis. 1:93-99.

26. Moore, R. D., C. R. Smith, and P.S. Lietman. 1984. Associationof aminoglycoside plasma levels with therapeutic outcome ingram-negative pneumonia. Am. J. Med. 77:657-662.

27. Moore, R. D., C. R. Smith, and P. S. Lietman. 1984. Theassociation of aminoglycoside plasma levels with mortality inpatients with gram negative bacteremia. J. Infect. Dis. 149:443-448.

28. Neu, H. C. 1982. Pharmacology of aminoglycosides, p. 125-142.In A. Whelton and H. C. Neu (ed.), The aminoglycosides:microbiology, clinical use, and toxicology. Marcel Dekker, Inc.,New York.

29. Niemiec, P. W., M. D. Allo, and C. F. Miller. 1987. Effect ofaltered volume of distribution on aminoglycoside levels inpatients in surgical intensive care. Arch. Surg. 122:207-211.

30. Noone, P., T. M. C. Parsons, J. R. Pattison, R. C. B. Slack, D.Garfield-Davies, and K. Hughes. 1974. Experience in monitoringgentamicin therapy during treatment of serious gram-negativesepsis. Br. Med. J. 1:477-481.

31. Rotshafer, J. C., H. G. Berg, R. B. Nelson, L. Strand, and D. J.Lakatua. 1983. Observed differences in gentamicin pharmaco-kinetic parameters and dosage recommendations determined byfluorescent polarization immunoassay and radioimmuno-assaymethods. Ther. Drug Monit. 5:443-447.

32. Sawchuk, R. J., D. E. Zaske, R. J. Cippolle, W. A. Wargin, andR. G. Strate. 1976. Kinetic model for gentamicin dosing with theuse of individual patient parameters. Clin. Pharmacol. Ther.21:362-369.

33. Schimpff, S. C. 1985. Overview of empiric antibiotic therapy forthe febrile neutropenic patient. Rev. Infect. Dis. 7(Suppl. 4):S734-S739.

34. Sculier, J. P., and J. Klastersky. 1984. Significance of serumbactericidal activity in gram-negative bacillary bacteremia inpatients with and without granulocytopenia. Am. J. Med. 76:429-435.

35. Torress, M. J., J. W. Kern, M. A. Gill, T. V. Berne, P. N. R.Heseltine, A. E. Yellin, and F. C. Chenella. 1983. Comparison ofserum sampling methods for estimating gentamicin pharmaco-kinetic variables. Clin. Pharm. 2:353-355.

36. Wade, J. C., S. C. Schimpff, K. A. Newman, C. L. Fortner,H. C. Standiford, and P. H. Wiernik. 1981. Piperacillin orticarcillin plus amikacin: a double blind prospective comparisonof empiric antibiotic therapy for febrile granulocytopenic cancerpatients. Am. J. Med. 71:983-990.

37. Zaske, D. E., J. L. Bootman, L. B. Solem, and R. G. Strate.1982. Increased burn patient survival with individualized dos-ages of gentamicin. Surgery 21:407-411.

38. Zaske, D. E., R. J. Cipolie, J. C. Rotschaffer, L. D. Solem, N. R.Mosier, and R. G. Strate. 1982. Gentamicin pharmacokinetics in

VOL. 34, 1990 707

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.

708 ZEITANY ET AL. ANTIMICROB. AGENTS CHEMOTHER.

1,640 patients: methods for control of serum concentration.Antimicrob. Agents Chemother. 21:407411.

39. Zaske, D. E., R. J. Cipolle, R. G. Strate, and W. F. Dickes. 1981.Increased gentamicin dosage requirements: rapid elimination in249 gynecology patients. Am. J. Obstet. Gynecol. 139:896-900.

40. Zaske, D. E., P. Irvine, L. M. Strand, R. G. Strate, R. J. Cipolle,

and J. Rotschafer. 1982. Wide interpatient variations in genta-micin dose requirements for geriatric patients. J. Am. Med.Assoc. 248:3122-3126.

41. Zaske, D. E., R. J. Sawchuk, and R. G. Strate. 1978. Thenecessity of increased doses of amikacin in burn patients.Surgery 84:603-608.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

ac o

n 25

Feb

ruar

y 20

22 b

y 20

0.11

5.55

.78.