In another presentation at the same oral session,
Sonneveld described findings from a study of response-guided interferon therapy
using stopping rules based on HBsAg levels (<1500, 1500-20,00 or >20,000
IU/mL) or HBsAg decline (yes or no) at 12 and 24 weeks.
The researchers noted as background that several
factors are associated with response to interferon, including HBV genotype,
baseline viral load, ALT level and possibly IL28B gene variants, which predict
response to interferon-based therapy for hepatitis C. Serum HBsAg levels appear
to reflect the amount of covalently closed circular DNA
(cccDNA) in the liver – a
by-product of viral replication – so
serum HBsAg levels during treatment may predict sustained response.
This pooled analysis included 803 participants in three
trials. Treatment response was defined as both virological response (HBV DNA below
200 IU/mL) and HBsAg loss at 24 weeks after the end of treatment.
A majority of patients (75%) were men and about 80%
were Asian. Nearly half had HBV genotype C, followed by 25% with genotype B. For
treatment, 58% used pegylated interferon monotherapy while the rest used
pegylated interferon plus lamivudine (Epivir).
HBsAg decline during treatment varied strongly
according to HBV genotype, with genotype A showing the greatest decline
followed by B, C and D, in that order. In all cases, HBsAg levels were lowest at
the end of treatment and then started to rise again, but did not reach baseline
levels by the end of follow-up.
Across all genotypes, however, the researchers
observed that responders had sustained lower HBsAg levels than non-responders,
whose levels returned nearly to baseline values after finishinging treatment.
At 12 weeks, both HBsAg level
and HBsAg decline predicted treatment response. There were 45% of patients with
HBsAg <1500 IU/mL, 22% with HBsAg 1500-20,000 IU/mL and 6% with HBsAg
>20,000 IU/mL who went on to experience combined response six months
post-treatment.
Overall, people who
experienced HBsAg decline at 12 weeks were about twice as likely to achieve
combined response (26 vs 14%,
respectively). However, the predictive value of the stopping rules varied by
HBV genotype, necessitating genotype-specific algorithms at week 12.
At 24 weeks, HBsAg level was a
better predictor of combined response than HBsAg decline. There were 46% of
participants with HBsAg <1500 IU/mL, 16% with HBsAg 1500 to 20,000 IU/mL and
only 3% with HBsAg >20,000 IU/mL who experienced combined response six months
after finishing treatment.
Again, people who experienced
HBsAg decline at 24 weeks were about twice as likely to achieve combined
response (25 vs 12%, respectively).
But at this time point, negative predictive values were 96% for HBV genotype A
and 100% for genotypes B, C and D, so genotype-specific stopping rules were not
necessary.
Overall, the researchers concluded that the chances
of combined response were best for people with HBsAg below 20,000 IU/mL at week
24 of treatment, supporting a stopping rule that all patients with HBsAg above
20,000 IU/mL at that point should discontinue treatment.