Association between the mid-trimester C-reactive protein-to-albumin ratio and subsequent preeclampsia: a retrospective cohort study

Preeclampsia remains a major cause of maternal and perinatal morbidity. We evaluated the association between the mid-trimester C-reactive protein-to-albumin ratio (CAR) and subsequent preeclampsia and assessed whether CAR added discriminatory information to routinely available maternal characteristics. This retrospective single-center cohort included 3,399 singleton pregnancies with C-reactive protein and albumin measured at 20–24 weeks before hypertensive disease diagnosis: 81 developed preeclampsia, 171 developed gestational hypertension, and 3,147 remained normotensive. Associations were evaluated using adjusted logistic restricted cubic spline models; performance was internally validated with 500 bootstrap resamples. A restricted cohort excluding 66 pregnancies with post-sampling GDM or IUFD and matched analyses within that cohort were evaluated as sensitivity analyses. Median CAR was 0.73 (interquartile range, 0.60–0.92) in normotensive pregnancy, 0.86 (0.67–1.07) in gestational hypertension, and 1.26 (0.98–1.59) in preeclampsia ( P < 0.001). Adjusted associations were nonlinear. At CAR 1.25 versus 0.75, adjusted odds ratios were 8.05 (95% confidence interval, 5.13–12.61) for preeclampsia versus non-preeclampsia, 6.52 (3.23–13.13) versus gestational hypertension, and 7.84 (5.15–11.94) versus normotensive pregnancy. On the g/L-albumin scale, these contrasts correspond to 0.125 versus 0.075 and are not clinical cutoffs. Optimism-corrected areas under the curve were 0.510, 0.459, and 0.518 for the availability-constrained clinical model and 0.859, 0.762, and 0.865 after adding CAR, respectively. Because the comparator was weak and lacked key established risk factors, these increases do not establish added value beyond a contemporary validated preeclampsia model. The restricted 3,333-pregnancy analysis produced closely similar spline estimates. Higher mid-trimester CAR was nonlinearly associated with subsequent preeclampsia, with consistent restricted-cohort and matched sensitivity analyses. Incremental value beyond a contemporary validated preeclampsia model was not evaluated or demonstrated. The reported CAR values are descriptive contrasts rather than clinical cutoffs; CAR remains an exploratory marker requiring prospective external validation.