Whether reduced sugar exposure during the first 1,000 days of life influences later rheumatoid arthritis (RA) risk is unclear. We examined this question using the end of sugar rationing in the United Kingdom as a natural experiment.

We analyzed 63,051 UK Biobank participants born between October 1951 and March 1956. Birth windows were defined to represent prenatal and early postnatal exposure to sugar rationing within the first 1,000 days. Participants with prevalent RA before baseline were excluded. Incident RA was identified from first-occurrence records; medically ascertained RA was examined in sensitivity analyses. Follow-up began at recruitment with delayed entry, and attained age was used as the underlying time scale. The primary reference group comprised births from July to December 1954, which were unexposed during the first 1,000 days while remaining temporally close to exposed cohorts. Models were adjusted for sex, month of birth, survey year, birthplace, race (White versus non-White), and parental cardiovascular disease or diabetes, with additional adult covariates examined in sensitivity analyses.

Among 63,051 participants, 778 incident RA events were identified. No clear association was observed between early-life sugar rationing exposure and adult RA risk. In the main binary model, the hazard ratio (HR) for rationed versus non-rationed participants was 1.08 (95% confidence interval [CI] 0.93–1.26; p=0.300). Exposure categories ordinally coded from greatest to least rationing exposure also showed no protective dose-response pattern (trend HR 1.03, 95% CI 0.98–1.07; p=0.235) (Figure 1). Sensitivity analyses yielded estimates broadly consistent with the main analysis. Using medically ascertained RA, the HR was 1.08 (95% CI 0.92–1.27; p=0.368). Expanding the reference group moved the earliest exposed-window estimate closer to the null, from HR 1.26 to 1.10. Placebo cutoff analyses yielded no consistent pseudo-effect. Because Schoenfeld residual-based tests suggested some departure from proportional hazards, we additionally estimated age-restricted RA-free time. The RA-free time difference was 0.017 years (approximately 6 days) at age 70 and 0.015 years (approximately 5 days) at age 72, both non-significant (Table 1).

In this UK Biobank natural experiment, we found no clear evidence that reduced sugar exposure during the first 1,000 days of life was associated with lower adult RA risk. If an effect exists, our findings suggest that it is unlikely to be large. More precise exposure assessment or mechanistic intermediate phenotypes may help clarify whether a small effect exists.