Fetal macrosomia (birth weight >/=4,500 g) is known to increase a number of adverse maternal and perinatal outcomes. Although there is a clear association between maternal diabetes mellitus and fetal macrosomia, the majority of macrosomic infants are born to non-diabetic mothers. We wished to determine the recurrence rate of macrosomia in non-diabetic pregnancy and to see if a history of multiple prior macrosomic infants confers additional risk. A retrospective analysis of 14,461 term, singleton, infants born to non-diabetic mothers in 1997 and 1998 was performed, using a computerised hospital database. Among 14,461 term pregnancies, 529 infants (3.7%) were macrosomic, and the incidence was significantly higher in parous women (4.6%) compared with nulliparas (2.4%, p < 0.0001). Over the next 5 years, 164 of these women returned for another delivery. Women with a history of one macrosomic infant are at significantly increased risk of another macrosomic infant in a subsequent pregnancy (OR 15.8, 95% CI 11.45 - 21.91, p < 0.0001). For women with two or more macrosomic infants, the risk is even greater (OR 47.4, 95% CI 19.9 - 112.89, p < 0.0001). Macrosomia was associated with increased rates of instrumental delivery and anal sphincter injury regardless of parity, and additionally with increased rates of caesarean delivery and shoulder dystocia among nulliparas. Overall, 88% of women who laboured with a macrosomic infant achieved vaginal delivery.