Interpretation of studies that have examined parity as a risk factor for Down's syndrome has been hindered by inadequate control for maternal age and/or failure to account for the differential use of prenatal diagnosis and pregnancy termination between low-parity and high-parity women. In this case-control study, the authors used exact matching on maternal age-minimize confounding and evaluated the potential impact of differential termination. A total of 898 cases of Down's syndrome and 4,488 controls were identified using Washington State birth certificates from 1984-1998. There was a trend towards increasing risk of Down's syndrome with increasing parity in both younger (age <35 years) and older (age > or =35 years) mothers. Restriction to women with no indication of amniocentesis (for whom differential termination is unlikely) resulted in a blunting of the odds ratios; however, a trend for parity remained. After restriction, odds ratios were as high as 1.65 (95% confidence interval: 1.13, 2.40) in younger women with a parity of three (compared with a parity of zero) and 2.41 (95% confidence interval: 1.41, 4.12) in older women with a parity of four or more. Although the odds ratios for older women were probably biased upwards because of underreporting of amniocentesis on birth certificates, these data support an association between parity and Down's syndrome.