Smoke before birth, and rheumatoid arthritis decades later

Design

prospective cohort analysis in UK Biobank with Cox models, mediation analysis and gene-environment interaction testing

Population

427,037 participants for intrauterine exposure and 414,318 for age at smoking initiation

Primary outcome

incident rheumatoid arthritis

Effect

intrauterine exposure HR 1.15 (95% CI 1.08 to 1.21); childhood initiation HR 1.74 (1.39 to 2.17); high genetic risk plus childhood initiation HR 2.89 (2.17 to 3.86)

UK Biobank questionnaire, biomarker, genetic and linked health-record data were used to test whether early-life tobacco exposure predicts incident rheumatoid arthritis — intrauterine exposure in 427,037 participants and age at smoking initiation in 414,318 — with mediation analysis for allostatic load and interaction testing against polygenic risk.

Intrauterine tobacco smoke exposure was associated with a modest increase in rheumatoid arthritis risk (HR 1.15, 95% CI 1.08 to 1.21). Age at starting to smoke mattered in the expected direction: against never-smokers, starting in childhood carried HR 1.74 (1.39 to 2.17), in adolescence 1.56 (1.29 to 1.88) and in adulthood 1.47 (1.22 to 1.76). Allostatic load mediated a statistically significant but trivially small share — indirect effects between 1.001 and 1.009. Genetic risk compounded the exposure: high polygenic risk with intrauterine exposure gave HR 1.93 (1.75 to 2.13) and with childhood initiation 2.89 (2.17 to 3.86).

Smoking as a rheumatoid arthritis risk factor is old news; the timing gradient is what is new here. Starting younger carried more risk than starting later, and exposure before birth registered at all. Intrauterine exposure was self-reported decades afterwards, which is a serious limitation, and UK Biobank participants are not representative. But the message for a clinic conversation is clear enough: with a family history of rheumatoid arthritis, the age at which a young person starts smoking is not a neutral detail.

• Take a smoking history that includes the age of starting, not just current status and pack-years
• Raise smoking specifically with young relatives of patients with rheumatoid arthritis
• Ask about smoking in pregnancy when taking a family history in a young patient
• Do not present the genetic interaction as individual prediction — polygenic scores are not a clinical test
• Continue to treat smoking cessation as the single most useful modifiable factor in established disease

Why it matters

It puts the modifiable exposure decades before the disease, in the age group nobody counsels about arthritis.

Don't overread it

Observational, with intrauterine exposure recalled in adulthood — it cannot establish that prenatal smoke causes rheumatoid arthritis.

The statistics, in plain English

A hazard ratio of 1.15 for intrauterine exposure is small — it shifts population risk, not an individual's prognosis. The mediation figures of 1.001 to 1.009 are statistically significant only because the cohort is enormous; they describe essentially no mechanism. The combined genetic and exposure hazard ratios describe groups defined after the fact and are not a risk calculator.

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.