A population-based case-control study of risk factors for childhood cancer was conducted for 593 cases diagnosed over the period 1986-1988 in Moscow children 0 to 14 years of age. Two healthy controls to every case were selected from registers of local pediatric polyclinics by age, gender and residence. The parents of 593 cases and 1181 controls were interviewed face-to-face. Significantly higher odds ratios (OR) were associated with cancer in close relatives [OR 1.6; 95% confidence interval (CI) 1.3-1.9], any pathology associated with pregnancy (OR 2.9; 95% CI 2.4-3.6), including threatened miscarriage (OR 2.1; 95% CI 1.5-3.0), toxemia (OR 2.2; 95% CI 1.8-2.8) and hormone treatment during pregnancy (OR 2.2; 95% CI 1.0-4.5). Pre-term births were significantly associated with brain-cancer risk (6/1; OR 13.3; 95% CI 1.5-301.2). For low birth weight (< or = 2500 g) children born from full-term pregnancy, the OR for all cancers combined was 2.5 (23/22; 95% CI 1.4-4.7) and for leukemias 4.7 (9/4; 95% CI 1.4-16.5). In all, 100 cases and 151 controls had birth weight > or = 4000 g (OR 1.4; 95% CI 1.1-1.9). Risk of nephroblastoma was also significantly related to this factor (11/5; OR 5.1; 95% CI 1.6-16.4). A positive trend of OR with decreasing duration of breastfeeding was significant for all cancer combined (p < 0.05). Significantly higher OR were observed for dermatitis (12/6; OR 4.0; 95% CI 1.4-12.1) and viral hepatitis (40/22; OR 3.8; 95% CI 2.3-6.3) in child medical history.