Several systematic reviews have been published recently on birth outcomes of infants conceived through assisted reproductive technologies (ART), compared with infants conceived spontaneously. These outcomes include perinatal mortality, preterm birth, low birthweight and birth defects. Methodological limitations of many of the individual studies (including small sample size, potential for bias in ascertainment of outcomes and considering singletons and multiples together) were obviated in these reviews by excluding studies where methods were considered inadequate, by conducting meta-analyses using data from all methodologically sound studies (small and large) and by examining singletons separately. Overall, the reviews indicate few differences between outcomes in ART twins compared with twins conceived spontaneously. However, in singleton ART infants, there are around two-fold increases in risk of perinatal mortality, low birthweight and preterm birth, about a 50% increase in small for gestational age and a 30–35% increase in birth defects, compared with singletons conceived spontaneously. Couples considering ART should be counselled about the increased risk of adverse outcomes. Epidemiologists, in conjunction with clinical and laboratory colleagues, should now focus on large, methodologically sound studies with long-term follow up that seek to identify the reasons for these increased risks and their long-term consequences, whether they are associated with particular technologies and causes of infertility, and how they might be reduced.