The purpose of this study was to evaluate the effect of low Ca intake on the 10-year incidence of vertebral fractures in cohorts I and II of the Japan Public Health Centre-based Prospective Study. The baseline studies were conducted in 1990–1994, with the follow-up studies conducted after 10 years. We analysed 33 970 subjects aged 40–59 years in cohort I and 41 664 subjects aged 40–69 years in cohort II. At baseline, the intake of Ca was assessed as a predictor, using validated FFQ. A meta-analysis was performed to estimate a summary relative risk (RR) for the two cohort studies. The 10-year cumulative incidences of self-reported vertebral fractures were 0·38 % for cohort I and 0·56 % for cohort II. In women, lower Ca intake was associated with a higher incidence of vertebral fractures (P for trend = 0·001), with the lowest quartile of Ca intake having a significantly higher incidence (0·89/1000 persons per year or RR 2·10 (95 % CI 1·25, 3·55)) than that (0·42/1000 persons per year) of the highest. In addition, the RR calculated using energy-adjusted Ca intake (by the residual method) as an outcome was 1·92 (95 % CI 1·28, 2·88). However, no such association was observed in men. An increase of Ca intake should be considered as a preventive strategy for vertebral fractures in peri- and post-menopausal women with a low Ca intake.