It has been reported that parental human leukocyte antigen (HLA) compatibility is associated with certain adverse reproductive outcomes such as multiple spontaneous fetal losses and malformations, particularly among women without a prior livebirth. In order to study the relation between parental HLA compatibility and the risk of SIDS, a population-based case-control study was conducted using 1982-1990 Washington State linked birth and death certificate data. Concordance of parental race/ethnicity combined with prior maternal fetal loss was used as surrogate measures of parental HLA compatibility. The relation was evaluated separately among infants with and without a prior live-born sibling. Among first liveborn infants, those of racially discordant parents were at reduced risk of SIDS compared with those of racially concordant parents (relative risk [RR] = 0.57, 95% confidence interval = 0.34-0.96). Infants of white-minority parents were all at lower risk of SIDS than infants of white-white parents. Infants of racially concordant parents with two or more prior fetal losses were at increased risk of SIDS (RR = 2.44, 1.07-5.56), relative to infants of racially discordant parents. No such associations were observed among infants with prior live-born siblings. In fact, in this latter population, infants of racially concordant parents tended to have a lower risk of SIDS than those of discordant parents. This study suggests that a portion of SIDS cases, particularly among first live-born infants, may be due to parental HLA sharing. However, final determination of the existence of this relationship requires actual HLA typing of parents of both cases and controls.