- Design
- Systematic review and meta-analysis of cohort, case-control and nested case-control studies
- Population
- 126 studies, 86,097,714 participants, pooled globally and by region
- Primary outcome
- Odds ratio for liver cancer by cause, compared with the odds ratio for lung cancer in current smokers
- Effect
- Smoking and lung cancer OR 6.0; HDV 23.5, HBV 15.2, HCV 12.8 (all higher, p<0.01); MASLD 2.6 and alcohol 2.4 (both lower, p<0.01)
Everybody counsels about smoking. Almost nobody frames chronic hepatitis in the same terms, and this meta-analysis of 126 studies covering 86,097,714 participants was built to make that comparison directly.
Against an odds ratio of 6.0 for lung cancer in current smokers, hepatitis D carried an odds ratio of 23.5 for liver cancer, hepatitis B 15.2 and hepatitis C 12.8 - all significantly higher than the smoking-lung comparison. Metabolic dysfunction-associated steatotic liver disease (2.6) and alcohol use (2.4) were significantly lower. There was no significant difference across world regions, which matters because it means the finding is not an artefact of where hepatitis happens to be prevalent.
The comparison is rhetorical as much as epidemiological, and it should be handled honestly. These are pooled odds ratios from heterogeneous observational studies with different comparator groups, not a head-to-head study, and an odds ratio does not tell a patient their absolute risk - most people with hepatitis B never develop liver cancer. Nor does a lower relative risk make MASLD unimportant: it is far more common, so it contributes more cancers in absolute terms than the ratio suggests.
Where it lands is in the consultation. A patient with hepatitis B who has been told they are 'a carrier' and does not attend for surveillance is the person this paper is about. India carries a large chronic hepatitis B population with poor surveillance uptake, so the framing is directly usable: this is a cancer-causing infection, and six-monthly surveillance is the point of knowing.
- Drop the word 'carrier' from hepatitis B consultations - it is heard as 'not a patient'
- State the cancer risk explicitly when a patient declines or defaults from surveillance
- Do not use the ratio to downgrade MASLD: it is commoner, so it causes more cancers overall
- Check that every hepatitis B and C patient in the clinic has a surveillance interval recorded, not just a diagnosis
- Screen household contacts and vaccinate - the risk here is preventable at source
The statistics, in plain English
These are odds ratios pooled across cohort and case-control studies, so they estimate relative risk, not absolute risk - an odds ratio of 15 does not mean 15% of patients get cancer, and most people with hepatitis B will not. Pooling studies with different control groups and adjustment sets makes the confidence in any single figure weaker than the precision implies. The comparison across two different cancers is also indirect: the studies were not designed to be compared, and residual confounding differs between smoking and hepatitis cohorts.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for gastroenterology & hepatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free