Background/objectives: Smoking is an established cofactor for cervical carcinogenesis, but evidence on how Smoking Pattern Changes around cohort entry relate to risk in Japan is limited. We quantified cervical cancer risk by baseline smoking status and by changes between two routine health checkups in a nationwide cohort.
Methods: We used the Japan Medical Data Center claims-checkup database between January 2005 and July 2022. Women with ≥2 pre-index checkups were included; the index date was the second checkup. Self-reported smoking at each visit defined never, former (quit), new (initiated), and current (persistent) smokers; checkup pairs >36 months apart were excluded. Incident cervical cancer required ICD-10 C53 plus cancer-directed treatment (surgery, radiotherapy, or systemic antineoplastic therapy). Multivariable Cox models estimated hazard ratios (HRs) with 95% CIs, adjusting for age, BMI, alcohol, exercise, hypertension, diabetes, cerebrovascular and cardiovascular disease, and cholangitis.
Results: Among 1,330,797 women, incidence rates (per 100,000 person-years) were 151.4 in never smokers and 244.9 in ever smokers. Ever versus never smoking was associated with higher risk (HR 1.53, 95% CI 1.43-1.62). A graded risk was observed across Smoking Pattern Change categories versus never: former HR 1.44 (1.15-1.79), new HR 1.51 (1.20-1.90), current HR 1.54 (1.44-1.64). By age, HRs were 1.58 (1.47-1.70) for <50 years and 1.35 (1.17-1.55) for 50-64 years; ≥65 years was not statistically significant (HR 0.69, 0.30-1.59).
Conclusions: Smoking was associated with substantially higher cervical cancer risk, with a clear risk gradient from former to new to current smoking. The rapid elevation in new smokers and residual risk after quitting support integrating proactive cessation and initiation prevention into risk-stratified screening and routine health-check programs in Japan.
Keywords: cancer prevention; cervical cancer; epidemiology; nationwide cohort study; risk stratification; screening; smoking pattern changes.