Quantitation of Renal Function

Accurate assessment of renal function in critically ill patients is imperative. Serial estimates or measurements of renal function routinely are recommended to guide individualization of drug dosage regimens to optimize clinical outcomes. The calculation of creatinine clearance (CL_cr) from a timed urine collection with creatinine measurement in serum and urine has been the standard clinical measure of renal function for decades. Urine is difficult to collect accurately in the ICU. Furthermore, many commonly used medications interfere with measurement of creatinine, especially when colorimetric assay methods such as the Jaffé method are used. Thus, measurement of CL_cr by this approach is not always the best way to assess glomerular filtration rate (GFR).^6,7 The administration of radioactive (¹²⁵I iothalamate, ⁵¹Cr-EDTA, or technetium-99m DTPA) or nonradioactive (aminoglycosides, iohexol, iothalamate, and inulin) markers of GFR, although scientifically sound, is clinically impractical because intravenous (IV) or subcutaneous (SQ) administration of the marker and the collection of multiple timed blood and urine collections make the procedures expensive and difficult to perform.

Estimation of CL_cr or GFR requires only routinely collected laboratory and demographic data and is inexpensive and clinically feasible. The Cockcroft and Gault (C-G) method for estimation of CL_cr⁸ and the Modification of Diet in Renal Disease (MDRD) method for estimation of GFR⁹ correlate well with CL_cr and GFR measurements in individuals with stable renal function.⁹ These methods lose their predictive performance, however, in patients with liver disease,^10–12 unstable renal function,^6,13,14 estimated GFR above 60 mL/min,¹⁵ or obesity.^16,17 In critically ill patients with AKI, both MDRD and C-G overestimate estimated GFR by 33% and 80%, respectively.¹⁸ Finally, although several methods for CL_cr estimation in patients with unstable renal function have been proposed,⁷ as well as equations for “adjusted” weights (for use in the C-G equation) in the obese,^16,17 the accuracy of these methods has not been rigorously assessed, and at present their use cannot be recommended.

There is considerable controversy at this time regarding whether the C-G or MDRD equation should be utilized to guide drug dosing adjustments in patients with CKD. The MDRD equation was developed to estimate GFR in patients with CKD.¹⁵ However, renal dosing recommendations for currently approved drugs are predominantly (>95%) based on relationships between drug clearance and CL_cr estimated by the C-G equation.¹⁹ The literature now suggests that the two equations cannot be utilized interchangeably, as several studies have shown that the use of the MDRD equation results in discordant dosing recommendations (compared to dosing based on C-G) in up to 40% of patients.^19–21 Complicating these comparisons even further is another problem: several versions of the MDRD equation (and most recently, an equation called Chronic Kidney Disease Epidemiology Collaboration [CKD-EPI]) have been reported since the publication of the original formula.²² As such, clinicians should continue to utilize CL_cr estimated by C-G for drug dosing.²³

Altered Drug Disposition in Critically Ill Patients with Renal Insufficiency

Effect On Drug Distribution

The volume of distribution of several drugs is increased significantly in patients with AKI or severe CKD.^25,28–30 Increases may result from fluid overload, decreased protein binding, or altered tissue binding (Table 174-1). The volume of distribution of only a few drugs is decreased in patients with CKD, and the mechanism proposed for this change is a reduction in tissue binding. Digoxin and pindolol are two prime examples, and for both of these drugs, a significant relationship exists between the decrease in distribution volume and CL_cr.^29,30

TABLE 174-1 Effect of End-Stage Renal Disease on Volume of Distribution (L/kg) of Selected Drugs Used in the ICU*

Strategies for Drug Therapy Individualization

Secondary references such as the American Hospital Formulary Service Drug Information,³⁵ Drug Prescribing in Renal Failure by Aronoff and colleagues,³⁶ and Goodman and Gilman’s The Pharmacological Basis of Therapeutics³⁷ are excellent sources from which one can acquire information on the pharmacokinetic characteristics of drugs in subjects with normal renal as well as impaired renal function. However, these references often do not provide the explicit relationships of kinetic parameters with CL_cr or GFR. In addition, since the references employ different definitions of renal impairment and utilize varying methodologies for deriving recommendations, drug dosing recommendations occasionally differ significantly.³⁸ This section provides a practical approach for drug dosage individualization in critically ill patients with AKI or CKD and patients receiving continuous renal replacement therapy (CRRT) or intermittent hemodialysis (IHD). Basic pharmacokinetic principles (see Chapter 169) combined with the disposition properties of a particular drug and a quantitative measure of the patient’s degree of renal function enable the clinician to design an individualized therapeutic regimen.

If the relationship of a drug’s total body clearance with CL_cr or GFR is not known, one can estimate the patient’s total body clearance, provided that the fraction of the drug that is eliminated renally unchanged (f_e) in subjects with normal renal function is known. The following approach makes six assumptions: (1) the volume of distribution is unchanged, (2) the change in total body clearance is proportional to CL_cr, (3) renal disease does not alter the drug’s metabolism, (4) metabolites, if formed, are inactive and nontoxic, (5) the drug obeys first-order (linear) kinetic principles, and (6) the drug’s pharmacokinetics can be described adequately by a one-compartment model. If these assumptions are valid, the kinetic parameter/dosage adjustment factor (Q) can be calculated as Q = 1 − (f_e [1 − KF]), where KF is the ratio of the patient’s CL_cr to an assumed normal value of 120 mL/min. The estimated total body clearance (CL_PT) can be calculated as follows: CL_PT = CL_normT × Q, where CL_normT is the value in patients with normal renal function (i.e., patients with a CL_cr of ≥ 120 mL/min). The elimination rate constant of the drug can be calculated as the quotient of the estimated total body clearance and volume of distribution. When these three key kinetic parameters are estimated, the individualized dosage regimen can be calculated as described in Chapter 169.

The optimal dosage regimen for an ICU patient with AKI or CKD depends on the desired goal. If there is a significant relationship between maximal plasma concentration and clinical response^39,40 (i.e., aminoglycosides) or toxicity^40,41 (i.e., quinidine, phenobarbital, and phenytoin), the dose and dosing interval may have to be modified. If the dosing interval is increased, the maximal plasma concentration and minimal plasma concentration are similar to values in individuals with normal renal function, but the desired target concentrations may not be precisely attained. In this case, consultation with a clinical pharmacist/pharmacologist may be warranted to facilitate the design of a revised dosage regimen. If no specific target values for maximal plasma concentration or minimal plasma concentration have been reported, attaining the same average steady-state concentration may be appropriate (i.e., cephalosporins). This goal can be achieved by decreasing the dose (D_PT = D_NORM × Q) or prolonging the dosing interval (τ) (τ_PT = τ_NORM ÷ Q).* If the dose is reduced while the dosing interval remains unchanged, the maximal plasma concentration becomes lower and the minimal plasma concentration higher (Figure 174-1). This dosage adjustment method, if taken to the extreme, results in maintenance of the desired average steady-state concentration by continuous infusion of a parenteral product. These principles have been used to derive dosage recommendations for commonly used drugs in the ICU for patients with mild, moderate, and severe kidney injury (Table 174-3).

Figure 174-1 Although the average steady-state concentrations (C_ave) are identical, the concentration-time profile would be markedly different if one changes the dose and maintains the dosing interval constant (scenario A) versus changing the dosing interval and maintaining the dose constant (scenario B) or changing both (scenario C). C_max, maximal concentration; C_min, minimal concentration.

(From Matzke GR, Frye RF. Drug therapy individualization for patients with renal insufficiency. Adapted from Dipiro JT, Talbert RL, Yee GC et al., editors. Pharmacotherapy: a pathophysiologic approach. 7th ed. New York: McGraw Hill; 2008, p. 833-44. Copyright McGraw Hill.)