A meta-analysis was conducted to compare the lactose intolerance symptoms of lactose maldigesters after consuming lactose (as milk, lactose dissolved in water, milk products, or commercial product) with responses after a placebo under masked conditions. An English language MEDLINE search was conducted using the medical subject heading of "lactose intolerance" from 1966 to January 2002. From an initial 1,553 citations, 2 independent reviewers selected 21 studies based on study design (randomized, crossover, blind) and use of an amount of lactose likely to be found in a meal (7-25 g) and a placebo among subjects free of gastrointestinal problems and >4 years old. Mean severity of symptom responses were analyzed as standardized differences, and the presence or absence of a symptom was estimated as pooled incidence differences (ID). For severity of flatulence, the standardized difference was 0.18 (95% confidence interval [CI] -0.16 to +0.52). The CIs for abdominal bloating and pain, degree of diarrhea, frequency of bowel movements per day, and frequency of diarrhea per day also included 0. For abdominal bloating, the ID was 5.9 more people per 100 with symptoms after lactose than placebo (CI -0.07 to +0.19). This same nonsignificant relationship was found for abdominal pain. The ID for diarrhea or loose stools was 0.15 (CI 0.03 to 0.28). Although the incidence of diarrhea was significantly higher, the size of the effect was very small. The results indicate that lactose is not a major cause of symptoms for lactose maldigesters following usual intakes of dairy foods, that is, 1 cup.