Antenatal care (ANC) attempts to screen and provide surveillance and treatment to individuals according to the level of need. We assessed the value of antenatal risk allocation made at the first visit in identifying women who will experience pregnancy complications in a rural area in Zimbabwe. As part of an ANC trial women were allocated into low- and high-risk categories based on medical, demographic and obstetric history. All highrisk women were recommended hospital delivery. This evaluation is based on 5223 women who received traditional care from nurse-midwives in 12 rural health centres, of whom 2890 (55%) were classified as high risk by the traditional risk markers, including 1618 nulliparous women. Complications occurred in 924 (17.7%) women 577 (62.4%) of whom had risk markers identified at booking. Twenty per cent (577/2890) of women classified as high risk developed complications. There was a high recurrence of complications, such as hypertensive disorders, operative delivery and preterm delivery. Nulliparity was a risk for low birth weight, operative delivery and hypertensive disorders, whereas grandmultiparity (> or =6) was a risk for hypertension in pregnancy. Young age (< or =16 years) was)a risk factor for low birth weight and perinatal death. Age above 35 years was not an independent risk factor. The traditional risk allocation system, with a likelihood ratio of 1.16, was not effective in identifying women at risk of pregnancy complications and resulted in too large a risk group for referral.