Dosing of Aminoglycosides To Rapidly Attain Pharmacodynamic Goals and Hasten Therapeutic Response by Using Individualized Pharmacokinetic Monitoring of Patients with Pneumonia Caused by Gram-Negative Organisms

Abstract

Achieving a peak aminoglycoside concentration (C_max)/MIC of ≥10 within 48 h of initiation of therapy for pneumonia caused by gram-negative organisms results in a 90% probability of therapeutic response by day 7. Targeting an MIC of 1 μg/ml, empirical aminoglycoside loading doses of 348 (25th- to 75th-percentile range, 275 to 432) mg were calculated to obtain a C_max/MIC of 10 in our patient population. Individualized pharmacokinetic monitoring coupled with MIC data should determine subsequent dosing regimens to minimize the potential for toxicity and maximize the probability of clinical response.

Aminoglycosides have been used for treating a wide variety of serious infections for over 30 years. However, few data exist on how best to maximize dosages and scheduling to achieve the best therapeutic outcome for a given patient. As aminoglycosides exhibit concentration-dependent killing, peak concentrations (C_max) and C_max/MIC ratios have been postulated to be the best predictors of therapeutic efficacy (5, 6, 8–12, 14). Traditionally, aminoglycosides have been administered in clinician-determined doses (i.e., similar doses for patients of similar weight) or with individualized pharmacokinetic monitoring (IPM) with C_max targets associated with efficacy (8–10, 12, 14). However, since aminoglycosides show concentration-dependent killing, it appears more appropriate to target the pharmacodynamic parameter C_max/MIC in an attempt to optimize efficacy (7).

In patients with documented pneumonia caused by gram-negative organisms, data demonstrate that achieving a C_max/MIC ratio of ≥10 within 48 h of initiation of therapy results in a 90% probability of temperature and leukocyte count normalization by day 7 of therapy (7). Clinician-determined and C_max target dosing methods may lead to a delay in reaching appropriate C_max/MIC ratios. The aim of this analysis was to determine an initial aminoglycoside dosing regimen to immediately achieve pharmacodynamic parameters predictive of optimization of therapeutic response in patients with pneumonia caused by gram-negative organisms.

(This work was presented in part at the 98th Annual Meeting of the American Society for Clinical Pharmacology and Therapeutics, San Diego, Calif., 1997.)

This was a retrospective analysis of prospectively collected pharmacokinetic data for 78 consecutively treated adult medical and surgical patients admitted to Bassett Healthcare from 1983 to 1993. All data were collected by the Clinical Pharmacy Service. Patients with pneumonia caused by gram-negative organisms who received gentamicin or tobramycin for ≥72 h were eligible for analysis. Diagnosis of pneumonia was done according to the Centers for Disease Control criteria (2): (i) a new, otherwise unexplained pulmonary infiltrate on a chest radiograph, (ii) growth of a sole pathogenic organism in a purulent sputum culture, and (iii) leukocytosis (>10,000/mm³) and/or fever (≥38°C). Patients with cystic fibrosis or neutropenia were excluded.

Initial aminoglycoside dosing regimens were chosen by the patient’s physician (clinician-determined method) and only altered if the regimen was found to be a significant overdose or underdose by empirical calculations with hospital-specific patient population pharmacokinetic parameters (1). Pharmacokinetic analysis was performed within 72 h of initiation of therapy with the collection of one predose serum concentration, recording of the duration and time of dose infusion, collection of one postdistributional serum concentration at least 30 min after the end of the infusion, and collection of one postdose serum concentration at least one estimated half-life after the first postdose concentration.

Pharmacokinetic data were analyzed by the method of Sawchuk and Zaske (13), fitting the data to a one-compartment, intravenous-infusion model. Aminoglycoside doses were modified to obtain a 1-h C_max of 7 to 10 μg/ml and a minimum concentration (C_min) of <2 μg/ml before redosing (C_max target method). By the Microscan system (Dade, West Sacramento, Calif.) the MIC at which 90% of the isolates are inhibited (MIC₉₀) of both gentamicin and tobramycin for sterile body fluid isolates (nonurine) of Escherichia coli, Serratia marcesens, and Citrobacter, Klebsiella, Enterobacter, and Proteus species at Bassett Healthcare for 1996 was <1 μg/ml. The MIC₉₀ for P. aeruginosa was 5 μg of gentamicin/ml and 1 μg of tobramycin/ml. Using patient-specific pharmacokinetic parameters and the MIC₉₀ of 1 μg of tobramycin/ml, we modeled an empirical aminoglycoside loading dose to achieve a C_max/MIC of 10 (thus, a C_max of 10 μg/ml). The following equation was used to determine these doses (13): C_max = [k_o/(k_elV)](1 − e^−k_elt₁)e^−k_elt₂, where C_max is 10 mg/liter at time t₂, k_o is the aminoglycoside infusion rate in milligrams per hour, k_el is the elimination rate constant in hour⁻¹, V is the volume of distribution in liters, t₁ is the infusion duration in hours (set at 1 h), and t₂ is the length of the distribution phase (set at 1.7 h secondary to the prolonged distribution time for larger aminoglycoside doses) (4).

Definitive aminoglycoside regimens for achieving a C_max/MIC of 10 by using the MIC for the organism isolated in each patient (which in all cases was ≤8 μg/ml) were also determined by the following equations (13): τ = ln (C_max/C_min)/k_el, where τ is the dosing interval in hours and C_min is 0.5 μg/ml; and MD = [C_maxk_elV(1 − e^−k_elτ)]/[(1 − e^−kt₁)e^−k_elt₂], where MD is the maintenance dose of aminoglycoside in milligrams. Data are presented as medians and 25th- to 75th-percentile range.

Patient demographics are shown in Table 1. This was an elderly population with some degree of renal impairment. Measured pharmacokinetic parameters are shown in Table 2.

TABLE 1.

Patient demographics

Characteristic	No. (%) of patients
Age (yr)	69 (58–75)a
Sex
Male	52 (66.7)
Female	26 (33.3)
Aminoglycoside
Gentamicin	38 (48.7)
Tobramycin	40 (51.3)
Initial calculated creatinine clearance (ml/min/1.73 m2)b	67.2 (50.0–85.0)a
Concurrent antibiotic therapy
β-Lactam (sensitive)	56 (71.8)
β-Lactam (resistant)	17 (21.8)
None	5 (6.4)
Isolated organisms
Acinetobacter sp.	1 (1.3)
Citrobacter sp.	5 (6.4)
Enterobacter sp.	9 (11.5)
E. coli	1 (1.3)
Klebsiella sp.	9 (11.5)
Proteus sp.	1 (1.3)
P. aeruginosa	45 (57.7)
Serratia sp.	6 (7.7)
Stenotrophomonas sp.	1 (1.3)
MIC of aminoglycoside used (μg/ml)	1.0 (0.5–4.0)a

^aMedian (25th to 75th percentile). 

^bCalculated by the Cockcroft-Gault equation (3). 

TABLE 2.

Measured pharmacokinetic variables

Variable	Valuea
Initial clinician-determined dose (mg/kg)	1.6 (1.4–1.9)
Initial dosing interval (h)	12 (12–12)
First measured Cmax (μg/ml)	5.3 (3.9–6.3)
First measured Cmin (μg/ml)	0.6 (0.3–1.1)
First measured Cmax/MIC	3.6 (1.4–6.2)
Aminoglycoside clearance (ml/min/1.73 m2)	71.5 (50.4–91.3)
Half-life (h)	3.5 (2.6–5.0)
Volume of distribution (liters)	20.2 (16.3–27.8)
Volume of distribution (liter/kg)	0.32 (0.27–0.38)
Adjusted dose (mg/kg)	2.4 (1.9–2.8)
Adjusted dosing interval (h)	12 (12–18)
Second measured Cmax (μg/ml)	6.7 (5.2–7.6)
Second measured Cmin (μg/ml)	0.8 (0.5–1.1)

^aData presented as median (25th to 75th percentile). 

Assuming an MIC of 1 μg/ml, an empiric median tobramycin loading dose of 348 (275 to 432) mg (or 5.3 [4.3 to 6.7] mg/kg of total body weight) would have achieved a postdistributional C_max of at least 10 μg/ml and a C_max/MIC ratio of at least 10 in 50% of our patient population. A 7 mg/kg loading dose of tobramycin would achieve an initial C_max of at least 10 μg/ml in 90% of our patients. Utilizing patient-specific pharmacokinetic parameters and the isolated-bacterium-specific MIC, a definitive aminoglycoside dose of 460 (280 to 1,160) mg, or 7.0 (4.2 to 17.0) mg/kg, given every 8, 12, 18, 24, 36, 48, and 60 h in 14, 25, 26, 19, 10, 3, and 3% of our patients, respectively, would have achieved a C_max of 10 (5 to 40) μg/ml, a C_max/MIC of 10, and a C_min of 0.5 μg/ml in all patients.

Figure 1 compares the median first C_max/MIC ratio that would be achieved in these patients for a range of aminoglycoside MICs for three different dosing regimens: an empirical 7-mg/kg loading dose, the traditional C_max target method, and the clinician-determined dosing method. The traditional C_max target method and the clinician-determined method would only achieve the C_max/MIC target ratio of 10 for MICs of ≤0.5 μg/ml, while a 7-mg/kg loading dose would achieve C_max/MIC targets for MICs of ≤1 μg/ml in 90% of the subjects and would reach a median C_max/MIC of 6.6 (5.2 to 8.1) for an MIC of 2 μg/ml and 3.3 (2.6 to 4.1) for an MIC of 4 μg/ml. Using a large loading dose followed by IPM, and MIC data once available, should ensure the continued achievement of C_max/MIC targets.

FIG. 1.

Median initial C_max/MICs for three aminoglycoside dosing regimens for a range of typical MICs for gram-negative bacteria for 72 patients with pneumonia caused by gram-negative organisms.

Figure 2 represents a proposed dosing algorithm to maximize the probability of achieving the target C_max/MIC of 10 in patients with pneumonia caused by gram-negative organisms. This C_max/MIC target was chosen to achieve the maximum probability of response, considering the MIC₉₀s of our institution’s gram-negative organisms and the upper limit of aminoglycoside dose tolerability. However, other C_max/MIC targets may be more appropriate. The probability of temperature and leukocyte count resolution by day 7 of therapy has been calculated for the following C_max/MIC ratios: 4, 65%; 5, 71%; 6, 76%; 7, 81%; 8, 84%; 9, 88%; and 10, 90% (7). For example, in our patient population, more resistant organisms with aminoglycoside MICs of 2 and 4 μg/ml would require median loading doses of 10.6 (8.6 to 13.4) and 21.3 (17.2 to 26.9) mg/kg, respectively, to achieve the C_max/MIC target of 10. As the safety of these doses has not been established, targeting a C_max/MIC of 5 in this situation may be a more reasonable approach, even though this C_max/MIC does not give a ≥90% probability of a temperature or leukocyte response. In addition, the combination of aminoglycosides with an additive or synergistic time-dependent killing agent may be essential. As each institution has different patient population profiles and different bacterial organism sensitivities, early pharmacokinetic optimization is important.

FIG. 2.

Dosing algorithm for aminoglycoside therapy in nosocomial pneumonia caused by gram-negative organisms.

Achieving a postdistributional aminoglycoside C_max/MIC of 10 may decrease the time to therapeutic response in our patient population (7). Using an MIC₉₀ of 1 μg/ml, with an aminoglycoside volume of distribution of 20.2 (16.3 to 27.8) liters (0.32 [0.27 to 0.38] liter/kg) and a moderate degree of renal impairment, administering an aminoglycoside loading dose of 7 mg/kg with IPM with the first dose will rapidly achieve this goal. As all patients in this analysis had their aminoglycoside dosing intervals adjusted to achieve a C_min of 0.5 μg/ml (65% were dosed every 8 to 18 h), these data should not be extrapolated to empirical single-daily-dose regimens. The aggressive C_max/MIC target of 10 may not be practical for all institutions, as aminoglycoside doses for organisms for which the MIC₉₀s are greater than 1 μg/ml may be beyond a clinician’s acceptable range. However, these data stress the need for rapid pharmacokinetic optimization of individual aminoglycoside dosing.

By reducing the time to therapeutic response, overall courses of aminoglycoside therapy may be shorter; with a potential reduction in the incidence of nephrotoxicity. Furthermore, to reduce the risk of aminoglycoside toxicity, a reduction in dose (and thus overall exposure) can be achieved by utilizing the aminoglycoside with the lowest MIC for the isolated bacterium. Our institution’s microbiology data for 1996 illustrate the importance of rapid organism identification and accurate MIC data to optimize aminoglycoside regimens. This analysis has demonstrated that two commonly used methods for determining the dosages of aminoglycosides (clinician determined and traditional C_max targets) fall short of achieving optimal pharmacodynamic targets.

As this is a retrospective analysis, it will be necessary to conduct a prospective clinical trial to determine if an aggressive aminoglycoside regimen with subsequent adjustment based on an organism-specific MIC to achieve a C_max/MIC of 10 and a C_min of <2 μg/ml in patients with pneumonia caused by gram-negative organisms can reduce the time to therapeutic response and potentially reduce the length of stay in the hospital without increasing toxicity.