Evaluation of dose-response relationships between smoking tobacco, alcohol consumption and oral cancer: a systematic review and meta-analysis
A partir d'une revue systématique de la littérature publiée jusqu'en mars 2025 (34 études), cette méta-analyse évalue les associations entre la fréquence, la durée, l'âge de début du tabagisme, la durée de la consommation d’alcool et le risque de cancer de la cavité buccale
Background: Tobacco smoking and alcohol consumption are established independent risk factors for oral carcinogenesis, yet the quantitative dose-response relationships for key exposure variables remain unclear.
Methods: A systematic review and dose-response meta-analysis (PROSPERO registration number: CRD42024531040) was conducted following the PRISMA 2020 guidelines. Cohort, case-control, and cross-sectional studies reporting relative risks (RRs) or odds ratios for at least one of the exposure variables, such as smoking frequency, smoking duration, alcohol duration, and age at smoking initiation were included. Random-effects dose-response models with restricted cubic splines were applied. Quality of the included studies was assessed using the Newcastle-Ottawa Scale.
Results: Data from 34 eligible studies (27 on smoking frequency, 22 on smoking duration, 11 on alcohol duration, and 11 on age at smoking initiation) were analysed. Smoking frequency showed a strong non-linear association, peaking at 30 smoking units/day (RR: 5.98; 95% CI: 3.37–10.60). Smoking duration was positively associated with risk, reaching a maximum at 40 years (RR: 2.19; 95% CI: 1.48–3.23). Duration of alcohol consumption showed a positive but weaker association, with the highest risk occurring at 50 years (RR: 3.13; 95% CI: 1.40–6.97). Age at smoking initiation was not independently associated with risk.
Conclusions: Smoking frequency and duration each exhibit strong and independent dose-response relationships with oral cancer, while alcohol duration independently confers moderate risk. These findings highlight exposure-specific risk thresholds that can inform targeted prevention and screening strategies, particularly in populations with high incidence rates.
BMC Public Health , article en libre accès, 2026