View PDF Version - RePub - Erasmus Universiteit Rotterdam

INTRODUCTION
Dynamic prediction of response to PEG-IFN 63
With the emergence of new antiviral agents, treatment options for chronic hepatitis B
virus (HBV) infection have changed considerably. At present, the Food and Drug Administration
(FDA) and the European Medicines Agency (EMEA) have approved six agents
for the treatment of chronic hepatitis B: pegylated interferon (PEG-IFN), lamivudine,
adefovir, telbuvidine, entecavir and tenovovir. Current guidelines do not recommend
any particular agent as fi rst-line therapy and the preferred initial treatment of individual
patients remains controversial.
Response to treatment of chronic hepatitis B patients with positive hepatitis B e antigen
(HBeAg positive), defi ned as HBeAg seroconversion to anti-HBe, suppression of
serum HBV DNA and normalization of ALT, is associated with a favourable long-term
prognosis, independent of the drug by which the response has been induced. 1-3
PEG-IFN treatment results in HBeAg seroconversion in 30% of treated patients after
one year of therapy. 4-5 Response is sustained in over 70% of these patients6-7 and has
shown to prolong survival and to reduce the incidence of hepatolocellar carcinoma. 1
Unfortunately, PEG-IFN therapy is associated with side-effects ranging from fl u-like
syndromes to neuropsychiatric disorders such as depression in a considerable number
of patients. 8 Therefore, patients should be selected for PEG-IFN therapy on the basis of
their pre-treatment probability of response, in order to optimize the balance between
potentially benefi cial effects and harmful side-effects. A prognostic model, that uses
readily available baseline variables, has recently been developed for these patients.
This model is based on the data from two large trials (n=721) investigating the effect
of PEG-IFN in HBeAg positive chronic hepatitis B4-5 and has shown that response to
PEG-IFN depends on HBV-genotype, sex, age, prior treatment with IFN, baseline HBV
DNA and ALT levels. 9 The model may be used to select patients at baseline who have
a relatively high probability of response. However, for an individual patient uncertainty
remains as to whether he or she will actually benefi t from this treatment. On-treatment
HBV DNA levels may help to refi ne the prediction of response, and an optimal stopping
rule based upon HBV DNA viral load decline during therapy may help to establish an
early indication for cessation of treatment.
In the current study the prediction model described by Buster et al9 was extended by
including on-treatment HBV DNA and ALT measurements. This enabled an individual
update of the response probability based upon new data acquired after each follow-up
visit.
Such a model can be benefi cial in guiding and supporting the patient through therapy
and may help to identify those patients for whom treatment cessation should be considered.
Different models were fi tted and compared in order to establish a model that not