Despite many reports of cocaine's ill effects on the developing fetus, scientists lack definitive evidence specifically linking cocaine to adverse reproductive effects (SN: 9/7/91, p.152). Using a powerful statistical technique, a Canadian research team has found that cocaine by itself causes very few problems during pregnancy.

Gideon Koren of the University of Toronto and his colleagues identified 20 previously published cocaine studies that in- volved pregnant women and yielded mixed results. Those studies often relied on small samples of cocaine users -- a problem that limited each study's statistical power.

To home in on cocaine's reproductive risks, his team turned to a method called meta-analysis, which statisticians use to assess data by pooling a number of similar studies. Koren and his colleagues identified women in the 20 studies who used cocaine during pregnancy but did not use other illicit drugs or alcohol, and compared them with those who reported no drug or alcohol use during pregnancy. They found no statistical link between prenatal cocaine use and premature delivery, low birthweight or congenital heart defects in babies -- problems often thought to result from cocaine.

The meta-analysis suggests that confounding factors -- such as other drugs, alcohol and smoking -- may account for the fetal growth retardation or prematurity commonly ascribed to cocaine, the researchers assert in the October _TERATOLOGY_.

Koren says women who use cocaine tend to smoke more cigarettes than women who use other illicit drugs and are more likely to drink alcohol and take additional drugs.

The meta-analysis did reveal a chance that a pregnant woman's cocaine use by itself might cause malformations of the genito-urinary tract in a small number of infants. Koren says this effect may trace to cocaine-induced constriction of the placental blood vessels.