A record-linkage study on risk factors for cause-specific infant mortality

Infant death certificates were linked with birth certificates for infants born to residents of Tohoku, Tokai and Kyushu regions in 1989 (n = 409, 679, or about one-third of all births in Japan), to examine the effects of variables, as reported on birth certificates, on cause-specific infant mortality. "Certain conditions originating in the perinatal period" and "congenital anomalies" accounted for nearly 90 percent of neonatal deaths, while "congenital anomalies", "injuries and poisoning" and "sudden infant death" were responsible for about 65 percent of postneonatal deaths. Mortality rates for almost all causes of infant deaths, except injuries and poisonings, increased as birth weight decreased not only in the neonatal period but also in the postneonatal period. This suggests that low birth weight places some infants at higher risk of death, and conditions that lead to low birth weight independently contribute to the risk of infant death. Cox's proportional hazards linear model was used to assess the effects of variables on infant mortalities by causes of death. An extremely strong birth weight effect was noted for "certain conditions originating in the perinatal period" and "congenital anomalies". Being a male infant and late order of birth in multiparity were other risk factors for deaths from "congenital anomalies", while being a male infant, resident of Tohoku region and maternal stillbirth experience related to deaths from "certain conditions originating in the perinatal period". Elevated risks of sudden infant death syndrome (SIDS), of which mortality rate in Japan was considerably lower than those in most developed Western countries, i.e. 0.23 per 1,000 live births in 1989, were associated with low birth weight, being a male infant, low maternal age, late order of birth in multiparity and illegitimacy. Low maternal age, late order of birth in multiparity and illegitimacy, also, related significantly to increased risk of infant deaths for "injuries and poisoning". These results suggest the independent contributions of socioeconomic factors to infant mortality, especially postneonatal mortality, from SIDS, "injuries and poisonings".