Seroepidemiology of Toxoplasma gondii among pregnant women in different parts of Sweden.

Surveys of pregnant women in four areas of Sweden in 1987-88, reveal a significant trend for decrease in Toxoplasma seroprevalence from Gotland island (26%, n = 467) in the south through Orebro county (18%, n = 1413) and Stockholm area (18%, n = 939), to Northern Sweden (12%, n = 837). No within area differences were observed between samples from rural and urban localities. Quantitative antibody data indicate marginally higher levels in the north than in the south, and a significant declining trend by age only in Orebro county. Incidence models are used to describe age-seroprevalence profiles for each area, using different assumptions about age- and time-specific infection rates, and to estimate the risk of maternal infection at the time of the survey. It is shown that the patterns of seroprevalence with age in Orebro county and Northern Sweden, but not Gotland island or Stockholm, strongly implicate time-dependent changes in Toxoplasma incidence, consistent with a declining incidence in the past which has possibly been reversed in recent years. The estimates of Toxoplasma incidence and risk of maternal toxoplasmosis are strongly dependent upon the underlying assumption of temporal change in incidence, with wide ranges in the predicted values. These studies demonstrate the difficulties in interpretation of horizontal cross-sectional data and the need for longitudinal studies of age-prevalence and seroconversion in the determination of the true risk of maternal toxoplasmosis.