Dysmenorrhea is a common gynecologic disorder in women of reproductive age. Previous studies have found an association between current cigarette smoking and prevalence of dysmenorrhea. This study investigated the association between exposure to environmental tobacco smoke (ETS) and the occurrence of dysmenorrhea among women without a history of this disorder. The study population consisted of 165 newly wed, nonsmoking Chinese women (in Shenyang, China), who intended to get pregnant and who had no past history of dysmenorrhea at the time of enrollment. These women completed a baseline questionnaire interview upon enrollment and were prospectively followed by daily diary. Dysmenorrhea was defined as a diary recording of abdominal pain or low back pain for at least 2 days during a menstrual period. A subject's ETS exposure was defined as the mean number of cigarettes smoked per day at home by household members over an entire menstrual cycle before the menstrual period. A logistic regression model was used to assess the effect of ETS on the risk of dysmenorrhea, with adjustment for age, body mass index, education, season, area of residence, occupation, shift work, perceived stress, passive smoking at work, and occupational exposure to chemical hazards, dust, and noise. Generalized estimating equations were used to account for autocorrelations as a result of multiple cycles per subject. This report is based on 625 prospectively followed menstrual cycles with complete baseline and diary data. ETS exposure was reported in 77% of cycles, within which average daily exposures throughout the cycle ranged from 0.02 to 10. 3 cigarettes. The incidence of dysmenorrhea was 9.7% and 13.3% among nonexposed and exposed cycles, respectively. Among ETS-exposed cycles, there was a positive dose-response relationship between the numbers of cigarettes smoked and the relative risk of dysmenorrhea. The adjusted odds ratios of dysmenorrhea associated with "low," "middle," and "high" tertiles of ETS exposure versus no exposure were 1.1 [95% confidence interval (CI), 0.5-2.6], 2.5 (CI, 0.9-6.7), and 3.1 (CI, 1.2-8.3), respectively. The findings were consistent with those of analyses limited to the first follow-up menstrual cycle from each woman. These data suggest a significant dose-response relationship between exposure to ETS and an increased incidence of dysmenorrhea in this cohort of young women.