ABSTRACT
The purpose of this study is to predict pregnancy-induced hypertension (PIH) by standard
methods at the first antenatal visit prior to the 12th gestational week (GW). This
is a retrospective study of 1189 primipara and 957 multipara who delivered singleton
babies having no major malformations after the 32nd GW. Multiple regression analysis
for the development of PIH was performed using the mean arterial blood pressure (MBP),
body mass index (BMI), hematocrit count at the first visit prior to the 12th GW, a
family history of hypertension, number of pregnancies, number of deliveries, past
history of hypertension and evidence of threatened abortion as explanatory variables
in primipara and multipara, respectively. We obtained a significant regression equation
in both primipara and multipara. Multiple correlation coefficient for primipara was
0.40223 (p < 0.01) and for multipara was 0.50764 (p < 0.01). When MBP was over 80 mm Hg, BMI was over 23.6, a family history of hypertension
was present, or a past history of hypertension was present, these variables were significantly
correlated with the development of PIH both in primipara and multipara. By combining
these variables, we obtained 32.7% positive predictive value, 93.5% negative predictive
value, 31.7% sensitivity, and 93.7% specificity in primipara and 24.6%, 96.7%, 52.6%,
and 89.8% in multipara, respectively. As the incidence of PIH was 8.7% in primipara
and 5.9% in multipara, we could predict 3.7 and 4.1 times higher than the true incidence.
PIH can therefore be predicted by the simple combination of MBP, BMI, a family history
of hypertension, and a past history of hypertension prior to the 12th GW. It can also
be predicted by the multiple regression equation with the use of nine explanatory
variables.
Keywords
Pregnancy-induced hypertension - multiple regression analysis - body mass index
