Background: Recent revelations of infanticide among SIDS cases diagnosed in multiple siblings have aroused concern that more misdiagnosed cases of infanticide may exist. Because twin births provide the opportunity to observe the fate of two offspring within a year's time the characteristics of SIDS and homicide among twin births were examined. Methods: U.S. Vital Statistics Linked Birth/Infant Death Certificate files for the years 1987 to 1991 were used. The population was limited to live births with birth weights ≥ 500 grams and gestational ages ≥ 24 weeks. A SIDS death was defined as a death attributed to an ICD-9 Code of 7980 that had an autopsy performed. A homicide was attributed to an ICD-9 Code of E9600 to E9689. As part of the analysis, an algorithm to match twins to their co-twins in order to measure differences in birth weight and age of death was used. Results: There were 1,056 SIDS deaths among twins over the 5 year period for a SIDS rate of 2.41/1000 live births. This compares to a SIDS rate among singleton births of 1.20 (risk ratio= 2.0). There were 60 deaths attributed to homicides among twins for a rate of 0.137/1,000 live births. This compares to 1,446 homicides among singleton births for a rate of 0.074(risk ratio=1.85). Among the matched twins the occurrence of a subsequent SIDS death in a twin pair was a rare event occurring among only 7 of 760 twin pairs that experienced SIDS(ratio=0.0092) while the ratio for subsequent homicide in twin pairs in which one twin had died of homicide was 3 of 36 (ratio=0.083). Simultaneous homicide occurred among 2 of the 3 twin pair, while among the 7 pair of SIDS deaths there was only 1 instance of a simultaneous death. Conclusions: These data suggest that the occurrence of both twins dying of SIDS is a rare event and that the occurrence of a simultaneous SIDS in a twin pair is very rare. When simultaneous SIDS occurs, closer scrutiny of the circumstances surrounding the deaths may be merited to rule out the possibility of infanticide.