gentamicin or tobramycin 5–8 mg/L) correlate best with therapeutic efficacy,

whereas toxicity tends to correlate with elevated trough levels (Cptrough

), which

reflects prolonged exposure to high drug concentrations. To minimize the risk of

toxicity, trough levels of less than 2 mg/L should be maintained. In patients with

normal renal function, these target serum aminoglycoside concentrations are usually

obtained after standard doses (e.g., 1.5 mg/kg) administered every 8 hours. Peak and

trough levels are typically measured once steady state is achieved, which is typically

within 24 hours.

41–44

Many clinicians now use once-daily dosing of the aminoglycosides (e.g., 5 mg/kg

every 24 hours) for patients with normal renal function in an attempt to minimize

aminoglycoside accumulation and nephrotoxicity. The rationale for this regimen is

based on the aminoglycosides’ concentration-dependent killing and postantibiotic

effect. This approach is not recommended for patients with advanced renal

impairment, however. When once-daily dosing is used, peak concentrations are less

helpful; however, trough concentrations should be monitored with a target of being

below the limit of analytic detection (<1 mg/L). The discussion regarding

aminoglycoside dosing in renal impairment that follows is based on the traditional

every 8 hours dosing regimen.

Aminoglycosides are almost completely eliminated by the kidneys; thus, the

clearance of these drugs essentially is equal to the glomerular filtration rate (GFR).

The pharmacokinetic properties of gentamicin and tobramycin are similar. A close

correlation also exists between CrCl (a surrogate for GFR) and gentamicin total body

clearance. As renal function deteriorates, aminoglycoside doses must be modified to

achieve the desired peak and trough plasma concentrations. Failure to appropriately

adjust the dosage of aminoglycosides in renal insufficiency can lead to high drug

plasma levels that can result in ototoxicity and nephrotoxicity.

In many cases, the aminoglycoside dose can be modified by extending the dosing

interval rather than simply reducing the dose. This permits maintenance of adequate

peak plasma concentrations to ensure efficacy, while allowing for sufficient

elimination between doses to produce trough levels less than 2 mg/L. The advantages

and disadvantages of adjusting the dosing interval versus reducing the dose are

summarized in Table 31-2.

Figure 31-1 illustrates the effect of increasing the dosing interval in a patient such

as G.G. with renal function that is 30% of normal. Although this is the preferred

method for adjusting the dose of aminoglycosides, for many other drugs requiring

dose adjustments in renal disease, simple dosage reduction is sufficient. Commonly

used drug references such as Facts and Comparisons can be used for dosing

guidelines for drugs used in patients with renal failure.

45

DETERMINATION OF APPROPRIATE DOSE

A number of methods have been developed to determine the appropriate

aminoglycoside dose for patients.

46 One method is Bayesian forecasting, in which

pharmacokinetic data obtained in the individual patient are integrated with

population parameters. Initially, a dose is used that is based on population parameter

values adjusted for characteristics such as increased SCr. Drug concentrations for the

individual patient are measured at specific times (e.g., peak and trough

measurements), and these are compared with the expected values from the population

data. Individualized pharmacokinetic parameter estimates are subsequently derived

using Bayes’ theorem to calculate a more patient-specific dosing regimen.

47

Because of the wide interpatient variability in aminoglycoside pharmacokinetic

parameters and the narrow therapeutic index for these drugs, doses should be

adjusted based on pharmacokinetic principles (e.g., Bayesian calculations or

methods described later in this chapter) and plasma concentrations that are specific

for this patient.

p. 668

p. 669

Table 31-2

Advantages and Disadvantages of General Approaches to Dosing Adjustments

in Renal Disease

Method Advantages Disadvantages

Variable Frequency

Use the same dose but ↑ the dosing

interval

Same Cpave

, Cpmax

, Levels may remain subtherapeutic for

prolonged periods in patients requiring

Cpmin

Normal dose

dosing intervals >24 hours

Variable Dose With Fixed Cpave

↓ Dose to maintain a target Cpave

;

keep the dosing interval the same

Same Cpave

Normal dosing interval

↓ Peak levels, which may be

subtherapeutic; ↑ trough levels, which may

↑ potential for toxicity

Cpave

, average plasma concentration; Cpmax

, maximum plasma concentration; Cpmin

, minimum plasma

concentration.

Figure 31-1 Serum concentration versus time profile for a patient with normal renal function (dotted line), and for

patient G.G. (Case 31-1) whose estimated creatinine clearance is 27 mL/minute (solid lines).

Patient-Specific Methods

Sawchuk et al. developed a method to derive patient-specific estimates of Vd and

clearance based on the patient’s size and estimated CrCl.

43 These parameters can be

used to calculate a specific dose for G.G. that will produce the desired gentamicin

peak and trough concentrations. If steady-state serum concentrations of gentamicin

are known, they can be used to calculate even more specific parameters. To initiate

gentamicin therapy, pharmacokinetic parameters should first be estimated from

population values.

The clearance of gentamicin (Clgent

) can be calculated based on G.G.’s CrCl.

Using the Cockcroft–Gault equation,

48

the CrCl can be estimated as follows:

where IBW is ideal body weight in kilograms, age is measured in years, and SCr is

serum creatinine in mg/dL.

With a SCr of 3.4 mg/dL, an ideal body weight of 70 kg, and an age of 31 years,

G.G.’s estimated CrCl, is 27 mL/minute.

For practical purposes, Clgent

is usually considered equivalent to CrCl. Therefore,

Clgent also is approximately 27 mL/minute or 1.6 L/hour. The Vd of gentamicin

(Vdgent

) is approximately 0.25 L/kg in patients with normal or impaired renal

function.

43,48,49

The Vdgent will be different in obese patients or those who have fluid overload.

Although G.G. does have some fluid retention, this is minimal and should not affect

her Vdgent significantly. Therefore, the Vdgent

for G.G. is as follows:

The loading dose of gentamicin (LDgent

) can be determined using the following

equation:

For treatment of infections caused by Pseudomonas species, a peak level of

approximately 6 to 8 mg/L is desired:

Using Clgent and Vdgent

, the elimination rate constant (Kd) and half-life for

gentamicin can be estimated as follows:

For the aminoglycosides, the dosing interval (τ) is determined by doubling the

half-life because by the end of two half-lives, 75% of the drug will have been

eliminated. This will usually lead to a desired trough level of less than 2 mg/L.

Therefore, gentamicin should be administered at least every 16 hours. For

convenience, an interval of 24 hours can be used, which also will achieve the

desired trough concentration.

Gentamicin is usually infused over 30 minutes. To determine the peak gentamicin

concentration, serum samples are drawn 30 minutes after the infusion has been

completed. Because the

p. 669

p. 670

estimated elimination half-life of gentamicin in G.G. (7.6 hours) is much longer

than the infusion time (0.5 hours), the intravenous (IV) bolus model can be used to

calculate an appropriate maintenance dose.

To achieve the peak concentration of 7 mg/L, the following equation can be used:

where tsample usually equals 1 hour (30 minutes after a 30-minute infusion).

The expected trough level in G.G. can now be estimated by the following equation:

Although not the case for G.G., patients with normal renal function may eliminate a

significant amount of gentamicin during the 30-minute infusion. In these patients, the

intermittent infusion model should be used to account for this loss of drug, where t

in

is the duration of the infusion:

REVISED PARAMETERS

CASE 31-1, QUESTION 3: After 72 hours of gentamicin therapy, G.G.’s peak and trough levels are 7.6 and

2.6 mg/L, respectively. Her physician attributes this to a gradual decline in renal function. (Her most recent SCr

is 4.8 mg/dL.) How would you revise G.G.’s dosing regimen based on these levels?

A gentamicin trough level of more than 2 mg/L suggests that G.G.’s dosing interval

is too short. Although her peak concentration is within the normal range of 5 to 8

mg/L, her trough concentration indicates that she is at a potentially toxic level. Her

pharmacokinetic parameters can be revised based on these values, and a new Kd can

be estimated from the following equation:

Because little change in G.G.’s Vdgent

is expected, a new Clgent

(Cl

revised

) can be

estimated from her revised elimination constant (if necessary, a revised Vdgent could

be calculated, keeping Clgent constant, although the clearance is more likely to change

than the volume of distribution):

These revised values for Kd and Cl can now be used to calculate a revised

maintenance dose to maintain the Cptrough at less than 2 mg/L using Equation 31-11:

The revised dose is now 115 mg (or ~110 mg) every 48 hours.

CASE 31-1, QUESTION 4: What are some limitations in calculating G.G.’s CrCl based on her SCr? Can this

estimate safely be used to predict gentamicin clearance?

See Chapter 28, Chronic Kidney Disease, for information about equations used to

calculate CrCl and estimate GFR. For patients with stable renal function, CrCl can

be estimated from SCr using the Cockcroft–Gault equation (see Eqs. 31-4 and 31-5).

In a patient such as G.G., however, whose renal function continues to decline during

the hospital course, estimation of renal function based on her increasing SCr becomes

more difficult. Because her SCr does not reflect a steady-state level, the previous

equations can no longer be used to accurately estimate her renal function. Because

G.G.’s SCr has increased rapidly from 3.4 to 4.8 mg/dLduring the past few days, her

CrCl is probably much lower than that estimated using the Cockcroft–Gault method.

A rising SCr may represent a decline in renal function manifesting as an accumulation

of creatinine.

Although prediction equations such as the Modification of Diet in Renal Disease

(MDRD) equations are a good measure of GFR,

50

they were developed in patients

with chronic kidney disease, therefore limiting its use in healthy patients. In addition,

the MDRD equation has not been validated for the dosing of most drugs in the setting

of renal dysfunction.

51,52 There can be significant differences in drug dosing regimens

when the MDRD and Cockcroft–Gault methods are used to estimate renal

function.

53,54