Studies that compare pregnancy outcomes must have an adequate number of subjects to detect a difference between the exposed group and a control group. As the the table below illustrates large sample sizes are necessary to show a twofold increase in adverse pregnancy outcomes. To detect smaller differences requires a larger sample size. To detect moderately common birth defects such as cleft lip or palate requires a sample size that is prohibitive for most controlled studies.

Sample Sizes Necessary to Detect a Two Fold (100%) Increase In Selected Adverse Pregnancy Outcomes (80% power, 5% Level of Significance) Outcome Denominator Population rate Number of Exposed Pregnancies Needed Spontaneous Abortion Enrolled Pregnancies 15/100 266 Low Birth Weight Live Births 10/100 261 Fetal Death Live births plus fetal deaths 3/100 684 Any major birth defect Live Births 3/100 684 Cardiovascular defect Live Births 1/115 2196 Cleft lip with or without palate Live births 1/930 17311 Stickler syndrome (or other rare birth defect) Live births 1/10000 185,539 Adapted from: Draft Guidance for Industry on Establishing Pregnancy Registries . U.S. Department of health and Human Services June , 1999.