So. Larsen et M. Lebech, MODELS FOR PREDICTION OF THE FREQUENCY OF TOXOPLASMOSIS IN PREGNANCY IN SITUATIONS OF CHANGING INFECTION-RATES, International journal of epidemiology, 23(6), 1994, pp. 1309-1314
Background. Estimation of the number of women infected during pregnanc
y with Toxoplasma gondii from seroconversion or seroprevalence data me
ets with various difficulties. Because of the high risk of transmissio
n of the infection to the fetus such infections are however a major co
ncern in pregnancy-related health planning. Methods. The expected annu
al percentage of pregnant women infected with Toxoplasma was calculate
d using models with varying assumptions with regard fo the infection r
ate, assumed to be independent of age but dependent on calendar time.
Three situations were studied: a stable situation, a sudden fall in th
e infection rate and a gradually declining (slower or faster) infectio
n rate over the lifetime of the pregnant women. Results. With a consta
nt infection rate, a maximum number of affected pregnancies occurs at
a yearly infection rate of 4%. In countries with a strongly decreasing
annual infection rate, estimates based on data on the relation betwee
n age and seropositivity related to only one period of time tend to ov
erestimate the number of affected pregnancies by as much as 60%. Concl
usions. In countries in transition from high to low infection rates, i
t is likely that the influence of decreasing immunity will, at least t
emporarily, more than outweigh the influence of the falling infection
rates, resulting in a higher number of infected pregnant women. The mo
dels used can also describe situations with age-dependent variation in
the infection rate, and may well apply to other infectious diseases r
elevant to pregnancy.