Drug Dosing in Renal Disease

Abstract

Renal disease alters the effects of many drugs, particularly when active drug moieties are renally cleared. Drug doses should usually be reduced in renal disease in proportion to the predicted reduction in clearance of the active drug moiety. Patient factors to consider in adjusting drug doses include the degree of renal impairment and patient size. Drug factors to consider in adjusting doses include the fraction of the drug excreted unchanged in urine and the drug’s therapeutic index. Estimates of renal function are useful to guide dosing of renally cleared drugs with medium therapeutic indices, but are not precise enough to guide dosing of drugs with narrow therapeutic indices. This article discusses principles of drug dose adjustment in renal disease.

Introduction

Renal disease alters the effects of many drugs, sometimes decreasing their effects but more often increasing their effects and thus potential toxicity. Many of these changes are predictable and can be mitigated by changing drug doses.

Renal disease interacts with drugs in three main ways. Firstly, patients with renal disease may be more vulnerable to a given drug effect (patient susceptibility). Secondly, a drug effect may be exaggerated or attenuated in patients with renal disease (pharmacodynamic change). Thirdly, and most importantly, some drugs have higher steady-state concentrations when given at usual doses to patients with renal disease (pharmacokinetic changes). This article describes how to adjust drug dose to account for the pharmacokinetic changes that occur in renal disease.

This can be summarised in one sentence: The drug dose should be reduced proportionally to the predicted reduction in drug clearance.

Increased drug clearance results in lower drug concentrations, while decreased drug clearance results in higher drug concentrations and hence greater drug effects. To avoid harm when drug clearance is significantly decreased, the dose of renally cleared drugs should usually be reduced in patients with renal disease.

Estimates of GFR Based on Creatinine

Estimates of glomerular filtration rate (GFR) are used to estimate renal function, in diagnosing renal disease and are also used to estimate renal drug clearance. Some active drug moieties are wholly or partly cleared from the body by the kidneys and this is the physiological rationale for using eGFR to estimate drug clearance. Similar to serum drug concentration, serum creatinine is dependent on both creatinine production (equivalent to drug dose) and creatinine clearance (equivalent to renal drug clearance).

Serum creatinine is the most commonly used analyte in the evaluation of renal function, and equations using serum creatinine concentration are the basis of most estimates of GFR (as discussed in detail elsewhere in this issue).¹ Estimated GFR (eGFR) has the units mL/min/1.73m²,² whereas creatinine clearance and drug clearance are both measured in mL/min. To avoid confusion, units should be carefully noted and their implications considered. This is discussed in more detail below.

Drug Dosing

The units of drug dose are amount per unit time e.g. 500 mg twice daily. For most drugs, prescribing information recommends a standard dose and provides some guidance on when this should be changed.³ The advice is necessarily imprecise as most drugs have large inter-individual variability in clearance and response. Consequently, dose adjustment is crude for most drugs in most circumstances, for example by doubling or halving doses. This is reflected in the available drug preparations, with most provided in a limited number of strengths which limits dose adjustment options. In most cases having a means to identify when drug dose should be halved or doubled is important, whereas a 20% change in dose is usually impractical or unnecessary. However there are several drugs for which small changes in dose or concentration may have an important effect, commonly known as a narrow therapeutic index.

Therapeutic Index

$$\text{The}\hspace{0.17em}\text{therapeutic}\hspace{0.17em}\text{index}=\frac{\text{minimum}\hspace{0.17em}\text{toxic}\hspace{0.17em}\text{dose}}{\text{minimum}\hspace{0.17em}\text{effective}\hspace{0.17em}\text{dose}}$$

For drugs with narrow therapeutic indices (Table 1), a small change in drug concentration can cause toxicity or loss of efficacy. Narrow therapeutic index drugs should be dosed using robust biomarkers, as estimates or empirical calculations of dose are not reliable enough to be safe. For example, lithium is a renally cleared drug with a narrow therapeutic index; therapeutic drug monitoring is used together with clinical response to guide dosing. Warfarin is a narrow therapeutic index drug that is metabolised rather than renally cleared; international normalised ratio (INR) is used as a biomarker to guide dosing.

Table 1.

Examples of drugs with narrow therapeutic indices.

Renally Cleared		Metabolised
Aminoglycosides	amikacin	Anticoagulants	warfarin (INR)*
	gentamicin	Anticonvulsants	lamotrigine
Glycopeptides	vancomycin		phenytoin
Other	digoxin	Cardiac drugs	amiodarone
	lithium		perhexiline
	morphine 6-glucuronide	Hormones	insulin (glucose)*
			thyroxine (TSH)*
		Immunosuppressants	mycophenolate
			tacrolimus

^*items in brackets represent recommended biomarker for dose titration.

Doses of these and other narrow therapeutic index drugs should be based on clinical response, biomarkers of effect and/or drug concentrations; empirical dose estimates are not sufficiently robust. Note that these are examples only and that this is not a comprehensive list.

Conversely, for drugs with a wide therapeutic index, even large changes in drug clearance may have only a modest impact on response, and therefore dose adjustments are less important. Beta-lactam antibiotics constitute a class of renally cleared drugs with wide therapeutic indices.

Drugs with narrow therapeutic indices should be dosed using biomarkers of effect and/or therapeutic drug monitoring.

Drug Clearance

Drug clearance (CL) and bioavailability (F) (the fraction of the drug dose that reaches the systemic circulation) determine the steady state plasma concentration (Cp) at a given dose (Equation 1 and Figure). CL has the units of volume/time and F is dimensionless (%). Note that CL is not the same as drug elimination (which, like dose, has units of amount/time and becomes equal to dose at steady state).

Figure.

Pharmacokinetics: the relationship between drug dose, drug concentration and pharmacokinetic variables. Adapted from Begg.⁴ CL = (apparent) clearance; F = bioavailability; Cp = steady state plasma concentration; Vd = (apparent) volume of distribution; k = elimination rate constant; t½ = elimination half-life.

$$\text{Dose}=\text{Cp}\times \stackrel{\text{CL}}{}/\underset{\text{F}}{}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\text{units}\hspace{0.17em}\text{of}\hspace{0.17em}\text{equation}\hspace{0.17em}1:\hspace{0.17em}\hspace{0.17em}\stackrel{\text{mg}}{}/\underset{\text{h}}{}=\stackrel{\text{mg}}{}/\underset{\text{L}}{}\times \stackrel{\text{L}}{}/\underset{\text{h}}{}$$ Equation 1:

From Equation 1 it can be seen that if CL is halved, drug dose should be halved to keep the drug concentration the same. Thus, if a drug is 100% renally cleared and renal function is half-normal, the drug dose should be halved, all other things being equal. However, many drugs are inactivated by metabolism (in the liver predominantly), and hence doses of metabolised drugs do not usually require changing in renal disease.

There are some drugs that are partially cleared by the kidneys and partially metabolised (e.g. low molecular weight heparins). For these drugs the dose reduction needed in renal disease is less than that for drugs that are 100% renally cleared. For example, in a patient with half-normal renal function, the dose of a drug that is half renally cleared and half metabolised would typically need to be reduced by a quarter.

Fraction Excreted Unchanged

The fraction excreted unchanged (fe) is the proportion of the active drug cleared renally in an average healthy person. The doses of drugs with fe ≥0.5 (50% or more renally cleared) should usually be reduced in patients with renal disease. This is particularly important for drugs with an intermediate therapeutic index (Table 2).

Table 2.

Examples of drugs with fraction excreted unchanged (fe) ≥0.5.

Intermediate therapeutic index drugs	fe
acyclovir and related antivirals	0.6–1
allopurinol	1 (oxypurinol)
atenolol	1
dabigatran	0.9
DPP-IV inhibitors* (e.g. sitigliptin)	0.8–1
gabapentin and pregabalin	1
low molecular weight heparins	0.7
metformin	1
morphine-6-glucuronide**	1
Wide therapeutic index drugs
ACE inhibitors†	1
beta-lactam antibiotics††	0.7–1
carbapenems	0.6–1

^*Dipeptidyl peptidase-IV inhibitors, used in treating type 2 diabetes.

^**The major active metabolite of morphine.

^†Most angiotensin-converting enzyme (ACE) inhibitors are prodrugs (suffix -pril, e.g. enalapril) of a renally cleared active moiety (suffix -prilat, e.g. enalaprilat).

^††With some exceptions e.g. ceftriaxine with fe ∼ 0.5.

$$\begin{array}{l}\text{patient}\hspace{0.17em}\text{dose}=\\ \text{usual}\hspace{0.17em}\text{dose}\times ((1-\text{fe})+\text{fe}\times \frac{\text{estimated}\hspace{0.17em}\text{patient}\hspace{0.17em}\text{renal}\hspace{0.17em}\text{function}}{\text{normal}\hspace{0.17em}\text{renal}\hspace{0.17em}\text{function}})\end{array}$$ Equation 2:

From Equation 2 it follows that dose should be reduced in proportion to a patient’s relative renal function for drugs that are 100% renally cleared, all other things being equal. The calculated dose can usually be rounded to the nearest whole or half tablet or the dose interval can be changed.

Estimates of Renal Function

For chronic kidney disease, estimates of renal function are used to predict disease outcome. However for drug dosing, estimates of renal function are used to estimate the renal clearance of the drug in question. When evaluating estimates of renal function for prescribing guidelines it is important to consider the gold standard of drug effect or, alternatively, drug concentration. Pharmacokinetic studies in patients with renal disease are usually conducted as part of preregistration evaluation. While not all these studies are published in peer-reviewed journals, the results can usually be obtained from the US Food and Drug Administration (FDA) databases.⁵

Patient size is a particularly important factor to consider when using creatinine-based estimates of renal function. The units of drug clearance are volume/time (mL/min), whereas the units of eGFR for chronic renal disease are volume/time/standard size (mL/min/1.73m²). How to consider the effect of size in drug dosing is discussed in detail elsewhere and has been the subject of interventional studies.^6–8 This aspect is a major limitation in using eGFR for drug dosing.

In many other respects the limitations of using creatinine-based measures both for drug dosing and for monitoring renal disease are similar. For example, creatinine-based estimates of renal function are not reliable in pregnancy. Methods of estimating renal function are discussed elsewhere in this journal.^1,9

Drug Product Information

Each drug is registered with product information for prescribers which encapsulates large amounts of research specific to that drug. When dosing advice for renal impairment is provided in the product information, it is important to know what method of defining renal impairment is used in that particular product information if a patient’s dose is to be reviewed. In most cases it is calculated creatinine clearance using the Cockcroft and Gault equation,¹⁰ as for many years this has been recommended by the FDA for pharmacokinetic calculations.¹¹ Although this situation may change, at the time of writing the Cockcroft and Gault equation underpins the majority of drug product information.

It should also be noted that product information sheets often have only crude dosing guidelines for renal impairment and sometimes specifically exclude patients with renal impairment (usually because these patients were excluded from the primary studies). In these situations, applying first principles is the only option if the drug is to be used, and such considerations allow for more appropriate prescribing in many situations. For example, a small change in creatinine clearance (or eGFR) from 31 mL/min to 29 mL/min may appear to cross a treatment threshold (of 30 mL/min) but be due solely to imprecision of measurement. Applying the letter of the product information may result in a large change in dose or stopping the drug altogether, whereas in reality the dose may not need changing.

Examples

Lithium is used primarily in the treatment of bipolar disorder and is 100% renally cleared. It has a narrow therapeutic index. An estimate of renal function helps select a starting dose for treatment. Subsequent dosing should be guided by measured drug concentrations and clinical response. Estimates of renal function can also be helpful in interpreting drug concentrations (e.g. assessing whether a low concentration is due to under-dosing or poor compliance).

The anti-gout drug allopurinol is the prodrug of oxypurinol which is 100% renally cleared. It has an intermediate therapeutic index. An estimate of renal function as an estimate of drug clearance provides useful guidance to dosing and can be used together with clinical and biochemical measures of effect (e.g. serum urate concentration).

Amoxicillin is a beta-lactam antibacterial drug that is 100% renally cleared. It has a wide therapeutic index. It is usually used acutely i.e. short-term. Most dosing guidelines do not discriminate on the basis of renal clearance except for patients with end-stage renal disease. Estimates of renal function are helpful in identifying patients who may need shorter dose intervals (high clearance) or may be adequately treated with smaller dose amounts or longer dose intervals (low clearance).

Compliance

No discussion on drug dosing is complete without a mention of compliance. Compliance is the greatest cause of therapeutic failure in the treatment of chronic diseases.¹² In prescribing any drug, the possibility of current or future non-compliance with the prescribed regimen should be considered. For example, a low digoxin concentration may be due to either underdosing or appropriate dosing with poor compliance. An estimate of GFR together with the digoxin concentration helps to discriminate between these two possible causes of treatment failure.

Other Information

The effects of other variables that may influence drug dosing in renal disease, such as changes in urine pH and the role of drug transporters for secretion, are beyond the scope of this article. Further information can be found in clinical pharmacology textbooks (such as Begg,⁴ Goodman and Gilman,¹³ and Cervelli¹⁴) or the primary literature. Online resources are increasingly available for further learning.^15–17

Conclusion

Important variables relevant to prescribing decisions are often not recognised in clinical care. The primary message of this article is that renal function is an important physiological variable that predictably affects the pharmacokinetics and thus clinical effectiveness of some drugs. Recognition of renal function as a prescribing issue is more important than the precision of different estimates of renal function. When prescribing for a patient with decreased renal function the characteristics of both the patient and the drug should be considered.

Principles that can be applied to any drug dosing situation have been described in this article. These are predicated on pharmacokinetic principles, patient physiology and skill in utilising drug information. Drug prescribing protocols do not cover all scenarios. Understanding and applying principles of good prescribing, such as those outlined above, allow recognition of potential problems, avoidance of drug toxicities and improvements in treatment efficacy.