BACKGROUND: To further assess the utility of targeted blood lead screening for children from households with members having occupational lead exposures, we conducted a meta-analysis of all available reports of take-home lead exposures. Our objective was to estimate the blood lead levels among U.S. children (ages 1-5) from households with lead-exposed workers. METHODS: Reports considered for inclusion were cited in Medline, Toxline, Excerpta Medica, and Bio-Med plus all unpublished reports available at the National Institute for Occupational Safety and Health through 1994. The a priori criteria for inclusion of U.S. reports required their having data on: (1) venous blood lead levels for children, (2) children's ages, (3) data for at least five children, (4) workers' occupations, (5) workers' blood lead levels, and (6) data collection methods. RESULTS: Based on a meta-analysis of 10 reports from 1987 through 1994, the children (n=139) of lead-exposed workers (n=222) had a geometric mean blood lead level of 9.3 microg/dL compared to a U.S. population geometric mean of 3.6 microg/dL (P=0.0006). Also in this group, 52% of the children had blood lead levels (BLLs) >/= 10 microg/dL compared to 8.9% in the U.S. (P=.0010), and 21% of the children had BLLs >/= 20 microg/dL compared to 1.1% in the U.S. (P=. 0258). CONCLUSIONS: We estimate, based on 1981-83 survey data, that there are about 48,000 families with children under six living with household members occupationally exposed to lead. If the findings from this meta-analysis (admittedly limited by small numbers) are generalizable, about half of the young children in these families may have BLLs >/= 10 microg/dL. Data were too sparse to determine if children of workers with elevated blood leads were at greater risk than children whose parents were only known to be lead exposed. Our findings support the position that children of lead-exposed workers should be targeted for blood lead screening.