Background/Purpose:
Allopurinol is a first line urate lowering therapy
in the management of gout. The observation that allopurinol hypersensitivity
was associated with allopurinol dose, renal impairment and diuretics led to the
suggestion that lower doses should be used in patients with renal impairment.
It is well documented that many patients receiving creatinine clearance (CrCL)-based
allopurinol doses fail to achieve target plasma urate. Allopurinol doses above
300mg daily are used rarely despite increasing evidence that doses higher than
those based on CrCL may be safe and effective. With the introduction of new
urate lowering therapies the ability to predict the dose at which a particular
patient may achieve target urate may influence a clinician’s decision on first
line therapy.

The aims of this
study were to explore factors that predict allopurinol response in patients
with gout, to determine the probability of achieving the recommended plasma
urate of ≤0.36mmol/L using CrCL-based doses and to predict the maintenance
doses required to achieve the target plasma urate.

Methods : Data were sourced from five
studies. A population analysis was conducted using NONMEM® v.7.2. Covariates
included renal function, body size, sex, ethnicity, concomitant drugs, and
renal transporter genotype. The final PKPD model was implemented in MATLAB
(2014a). Stochastic simulations were performed under two scenarios; 1) using
doses recommended by CrCL-based dosing recommendations, and 2) using daily
doses sufficient to achieve target plasma urate concentrations in >75% of simulates.
The model was evaluated against previously published data.

Results: A total of 1135 oxypurinol and
1178 urate plasma concentrations from 134 patients were available. A one
compartment PK model with first order absorption and elimination was the best
fit to the oxypurinol data. A simple direct effects (Emax) model provided an
adequate description of the steady-state plasma urate data. Renal function,
diuretic use, and body size were found to be significant covariates. Dose
requirements were found to increase approximately 2-fold over a 3-fold range of
weights and were 1.25-2 times higher in those taking diuretics. Renal function had
only a relatively minor impact on allopurinol dosing. Table 1 shows the
probability of achieving urate with CrCL based dose for a 70kg patient not
taking diuretics and the dose predicted to achieve target urate. The model
performed well when evaluated against external urate data.

Conclusion: This population PKPD model
for allopurinol supports the contention that patients receiving CrCL-based
allopurinol doses frequently fail to achieve target urate. The model provides a
means of predicting allopurinol maintenance dose requirements for individual
patients but needs to be validated against prospective data.

Table 1: Model
predictions for percentage of patients achieving target urate on CrCL based
doses and dose required to achieve target