both the Child-Pugh score and the MDF at all
relevant time points.
Methods that involve studying the disposition of
some exogenously administered agent (e.g. indo-
cyanine green, antipyrine, galactose or dextro-
methorphan) have now been superceded by
functional (often multicomponent) tests. Mono-
ethylglycinexylidide formation has not found
wide acceptance. More complicated Cox pro-
portional hazards models may exist for other liver
diseases, but are only used specifically for them
(e.g. the Mayo Clinic Survival Model for primary
biliary cirrhosis; see the US FDA Guidance).
When studies are needed
In general, studies are needed for all drugs to which
the liver is exposed, unless
drug excretion is entirely renal
hepatic metabolism accounts for<20% of clear-
anceandthe drug has a clearly demonstrated
wide therapeutic window or
the drug is volatile and it (and its metabolites) is
readily excreted by the lung
Most regulatory authorities allow some relaxation
of the requirement for studies when the drug is for
single-dose only, and when adverse events areCmax
rather than AUC-related (because Cmaxusually
varies little with reduce rate of clearance under
these conditions).
Study design
Generally, there should be a clear understanding
of the pharmacokinetics of the drug of interest in
all three Child-Pugh classes of liver disease,
unless the Sponsor is willing to accept strong
labeling against administration in severe liver
failure (and then has to study only the other two
grades of severity). The size of each treatment
group depends upon the variability in the phar-
macokinetics of the agent, although, regardless of
this fact, the US FDA Guidance recommends that
there should be at least six patients. As before,
single-dose studies may suffice when there is
reason to believe that all the stages of liver dis-
ease that are being studied, the pharmacokinetics
of a single dose are indeed predictive of the multi-
ple dose/steady state situation. When drugs are
being developed for more than one route of
administration, then usually one can be chosen
that provides the maximum information, and the
need to study the second route is obviated. Popu-
lation kinetics approaches are also sometimes
feasible if incorporated into phase III develop-
ment schemes, and when appropriate nonlinear
and non-compartmental models can be defined.
Conclusions of no effect are based upon the
pharmacokinetic tolerances accorded to generic
versus innovative products. However, small
numbers of patients usually make this quite
difficult.
In general, regulatory authorities are keen to
provide advice on particular study designs that
are appropriate on a case-by-case basis, and this
should always be an agenda item at end-of-phase IITable 19.1 Child-Pugh Point-scoring systemPoints scored
12 3
Encephalopathy grade 0 * 1or2 3or4
Ascites None Slight Moderate
Serum bilirubin (mg dl^1 ) <22–3> 3
Serum albumin (mg dl^1 ) >3.5 2.8–3.5 <2.8
Prothrombin time prolongation (s) <44–6> 6
* 0 ¼normal; 1¼restlessness 5 Hz waves on EEG; 2¼lethargic, disoriented, asterixis; 3¼somnolent/stuperous
rigidity; 4¼unrousable coma (each grade>0 can have other symptoms). Total points: good operative risk:6;
moderate risk: 7–9; poor risk:>9 points.19.3 HEPATIC INSUFFICIENCY 253