Cochrane on thymosin-alpha 1 in hepatitis B: favourable numbers, very low certainty, no ongoing trials

Design

Cochrane systematic review and random-effects meta-analysis of randomised trials, RoB 2 and GRADE assessed

Population

1,349 adults with chronic hepatitis B across 10 trials in six countries, published 1991–2018

Primary outcome

all-cause mortality, serious adverse events and health-related quality of life

Effect

all-cause mortality RR 0.53 (95% CI 0.29–0.96), very low certainty; serious adverse events RR 0.72 (0.53–0.99), low certainty

Thymosin-alpha 1 has been used as an immunomodulator in chronic hepatitis B for three decades, mostly in Asia, often alongside nucleos(t)ide analogues or interferon. This Cochrane review assembled the whole randomised evidence base: 10 trials from Bangladesh, China, Italy, Korea, Singapore and Taiwan, 1,349 participants, published between 1991 and 2018, with follow-up from six months to five years.

The pooled estimates read well. All-cause mortality risk ratio 0.53 (95% CI 0.29 to 0.96), serious adverse events 0.72 (0.53 to 0.99), hepatitis B-related mortality 0.53 (0.29 to 0.96), non-serious adverse events 0.47 (0.27 to 0.83) — all with no heterogeneity. Little or no effect on quality of life, from a single trial, and on histological improvement, where heterogeneity was 74%. But GRADE certainty was very low for every outcome except serious adverse events, which was low, downgraded for risk of bias, imprecision and inconsistency. Four trials were industry funded. Sixteen further studies are awaiting assessment because of incomplete reporting and the authors received no reply to their enquiries. No ongoing trials exist.

The honest conclusion is the one Cochrane reaches: we do not know whether this drug does anything. What makes it a practice point rather than an academic one is that a mortality risk ratio of 0.53 will keep being quoted without the certainty rating attached. If you prescribe thymosin-alpha 1, the defensible position is that the evidence is compatible with benefit and too weak to establish it — and that no trial is running that will settle it.

• Do not quote the mortality risk ratio without the very low certainty rating that goes with it
• Prioritise nucleos(t)ide analogue therapy, where the evidence base is not in question
• Tell patients plainly that the evidence for this agent is very uncertain, not that it does not work
• Note that no ongoing trials were identified — this will not resolve on its own
• Sixteen further studies could not be assessed because of incomplete reporting; the picture may be incomplete

Why it matters

A widely used agent turns out to rest on an evidence base that Cochrane rates as very low certainty on every important outcome.

Don't overread it

Very low certainty means we do not know, not that the drug is ineffective — the intervals are compatible with real benefit.

The statistics, in plain English

A risk ratio of 0.53 with an interval reaching 0.96 already sits close to no effect at its upper bound, and that is before the certainty rating is applied. GRADE downgrades for reasons a confidence interval cannot show: high risk of bias within the trials, and small participant numbers behind each estimate — 907 people across three studies for the mortality figure. An I-squared of zero looks reassuring but only says three small, similarly biased trials agreed with each other.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.