A prospective study of malaria during pregnancy was conducted between September 1986 and December 1989 in an area of unstable (mesoendemic) malaria transmission on the Thai-Burmese border. Antenatal clinics were set up in camps for displaced persons of the Karen ethnic minority and 1358 pregnant women were enrolled at a mean estimated gestational age of 23 weeks (standard deviation 5.7 weeks) and were followed weekly until delivery. Malaria developed in 505 women (37.2%); 80.2% of infections were Plasmodium falciparum, 17.1% were P. vivax, and 2.7% were mixed. Primigravidae were infected more commonly than multigravidae: 153/322 (47.5%) compared with 318/953 (33.3%) (P less than 0.001). The incidence of malaria declined from the 20th week of gestation (12%) towards term (4.4%). Most infections were detected before symptoms developed, and there were no deaths associated with malaria. Despite this, malaria was associated with an overall 123 g reduction in birthweight (95% confidence interval [CI] 34-212 g). This reduction was largely accounted for by lower birthweights of babies born to infected primigravidae (mean reduction 151 g, 95% CI 21-282 g) and women in their 2nd and 3rd pregnancies (mean reduction 185 g, 95% CI 84-286 g). The incidence of anaemia requiring treatment was higher in women who developed malaria, 149/420 (35.4%) compared with 191/670 (28.5%), and was proportional to the number of parasitaemic episodes. Thus, despite regular antenatal clinic attendance with prompt detection and treatment of malaria (the currently employed antimalarial strategy in areas with multidrug-resistant P. falciparum), malaria still had a significant adverse effect on pregnancy.(ABSTRACT TRUNCATED AT 250 WORDS)