Principles and Practice of Pharmaceutical Medicine

is usually done when the circulating concentration
response relationship for the wanted effects of the
investigational drug is well understood, of orthodox
sigmoid form, when there is no active metabolite(s),
and when the dose or concentration–response rela-
tionshipforunwantedeffectsisalsowellunderstood
and a long way to the right of the curve for efficacy.
Small studies with dense sampling, or population
kinetics with sparse sampling during the phase III
program can accomplish this. Note that the latter
cannot be accomplished when renal failure has been
a routine exclusion criterion during the clinical
development program. When study sponsors are
confident, then a small study (sayn¼4–6 subjects
with severe renal insufficiency) may serve to
exclude an effect on the drug.

Quantifying the effect of renal failure

Assuming that studies are needed, then patients
with mild or moderate renal failure (estimated
creatinine clearances 50–80 and 30–50 ml min
^1 ,
respectively) must be studied. Age-, sex-, diet- and
smoking-matched controls for the disease state
being treated should be used, rather than young,
fit volunteers, in order to avoid false-positive study
conclusions.
A single-dose study will usually be acceptable to
regulatory authorities, provided that there is clear
evidence elsewhere in the dossier/NDA that single-
dose data can predict multiple-dose pharmacoki-
netics. If multiple-dose studies are needed then
these should include an observation period at steady
state. The study size (i.e. number of patients per
group) will be determined by a power calculation
using the known variability of the pharmacokinetic
parameters of the drug in question; in practice there
is seldom the need for more than 15 subjects per
renal function stratum, unless a population kinetics
approach has been preferred. The nonlinear and
non-compartmental modeling procedures for use
in a population kinetics scheme are beyond the
scope of this chapter, and should certainly be dis-
cussed in advance with the relevant regulatory
authorities.
When possible, pharmacodynamic assessments
should be made in conjunction with the pharmaco-

kinetic estimates. The reason for this is that it can

check that there has not been some supersensitivity

state induced in the biophase by the disease causing

the renal insufficiency, and thus to exclude a false-

negative conclusion when purely pharmacokinetic

data are analyzed.

Assayswill usually be on plasma and urine.

Plasma protein binding should be estimated simul-

taneously because renal disease can alter plasma

protein binding of some drugs.

Note on estimation of creatinine

clearance

Formal measures of glomerular filtration rate

(GFR), using intravenous inulin or radio-iodinated

sodium iothalamate, will not have been performed

on most patients with relative renal failure in ordin-

ary clinical practice. Of several alternatives in

adults, creatinine (Cr) clearance (CrCL) is prob-

ably the most commonly employed, and uses the

familiar formula:

CrCLðml min
^1 Þ¼

ð½CrŠurineðmg dl
^1 Þurine excretion rate

ðml min
^1 ÞÞ=½CrŠplasmaðmg dl
^1 Þ

This requires a timed urine sample (e.g. a 24-h

collection). The Cockroft–Gault estimate of

CrCL uses only a point measure of serum creati-

nine and currently enjoys wide acceptance by reg-

ulatory authorities.

For men, the Cockroft–Gault estimate is:

CrCLðml;min
^1 Þ¼

½ 140 ageðyÞweightðkgފ=

½ 72 serum creatinineðmg dl
^1 ފ

For women, the Cockroft–Gault estimate is the

same as for men, except that the result is multiplied

by 0.85.

In infants, the Cockroft–Gault estimate is inac-

curate. Currently, the US FDA Guidance for

19.2 RENAL INSUFFICIENCY 251