Results
Between 27 April 2016 and 24 December 2018, we recruited 501 women to the PEACOCK study, across 36 maternity units in England and Wales (Figure S1). There were no statistically or clinically relevant differences across the participants who received expectant management as usual care outside the trial (n=182) and within the PHOENIX trial allocation arm (n=159) (Table 1, Table S1). Women in the two PHOENIX trial allocation arms (presented here as women in planned delivery group included for secondary analysis) were, as expected, balanced for baseline characteristics (Table S2).
There were similar outcomes in women randomised to expectant management group and those participating in the non-randomised expectant management group (Table 2), whilst outcomes in the planned delivery and randomised expectant management groups (Table S3) reflect those in the larger PHOENIX trial, with earlier gestation at delivery, as expected.(2) In women managed expectantly, 211 of 341 (61.9%) women delivered within seven days; in these women, 81 (50.9%) of those randomised and 95 (52.2%) of those non-randomised (i.e. outside the trial) had indicated delivery due to clinical concerns for maternal or fetal wellbeing. There were no perinatal deaths in the study.
For the PlGF test by the Quidel assay, the test performance for PlGF in determining need for delivery within seven days at low (<100 pg/mL) and very low (PlGF <12 pg/mL) is shown in Table S4. The test performance for PlGF <100 pg/ml in determining need for delivery within seven days had sensitivity of 97.9% (94.8-99.4%) for delivery within seven days; the negative predictive value was 71.4% (41.9-91.6%) and the specificity was 8.4% (4.1-14.9%). Similar test performance statistics for determining need for delivery within 14 days are shown in Table S5 and within two days in Table S6.
For the sFlt-1:PlGF ratio by the Roche assay, the test performance for sFlt-1/PlGF ≥83 in determining need for delivery within seven days had sensitivity of 91.4 % (86.3-95.1%) for delivery within seven days; the negative predictive value was 60.5% (43.4-76.0%) and the specificity was 20.9% (13.7-29.7%). Similar test performance statistics for determining need for delivery within 14 days are shown in Table S5 and within two days in Table S6.
For evaluation of the PREP-S prognostic model in this cohort, baseline predictor variables were assessed in the PEACOCK study cohort and compared with the original PREP-S cohort (Table S7). There were important differences between the two cohorts, particularly relating to gestation at enrolment (as different inclusion criteria were used), definitions used for (and therefore incidence of) adverse maternal outcomes.
The Receiver Operator Curve (ROC) areas for PlGF (Quidel), sFlt-1:PlGF ratio (Roche) and PREP-S are shown in Table 3 (and Figure 1), with consideration of the PREP-S model for a dichotomised endpoint (delivery within seven days). PREP-S was also assessed in combination with PlGF (Quidel), and with the sFlt-1:PlGF ratio (Roche), treating PREP-S as a single predictor.(4) The corresponding ROC areas for the clinical prediction model (PREP-S), PlGF (Quidel) and sFlt-1:PlGF (Roche) in this cohort in determining need for delivery within seven days was 0.64 (standard error (SE) 0.03), 0.60 (SE 0.03), and 0.63 (SE 0.03) respectively. The ROC area for PREP-S in combination with PlGF (Quidel) in determining need for delivery within seven days was 0.65 (SE 0.03), and PREP-S in combination with sFlt-1:PlGF (Roche) was 0.64 (SE 0.03). PREP-S (when used to determine a binary outcome), and PlGF have limited clinical applicability in this cohort in determining need for delivery within seven days.
Performance of the PREP-S model, PlGF (Quidel), and sFlt-1:PlGF ratio (Roche) was similar in determining delivery in 14 days and in 2 days in this cohort (Table S8 and Figures S2 and S3), with these predictors having limited clinical applicability in this setting.
The Kaplan-Meier time to delivery estimates for women in the expectant management groups, stratified by four PREP-S risk categories (as observed) are shown in Figure S4 and the recalibrated estimates in Figure S5.
Calibration of the PREP-S model is shown in Table S9, with calibration in the large and of the slope assessed for predicting delivery for preeclampsia within seven days. Calibration of the PREP-S model in this cohort was less good than that achieved in the original PREP-S cohorts. Overall, approximately the same number of women had the outcome as predicted by the model (expected value 0; calculated value -0.13; not significantly different). However, calibration of the slope was 0.375 (expected value 1.0), suggesting that the difference between adverse outcome event rates between low and high risk groups was not as great as the PREP-S model suggested, with PREP-S consistently over-predicting the adverse event rate in the higher risk groups. Recalibration of the model slightly improved the calibration of the PREP-S probabilities (Tables S10-S12), so that the notional probabilities were slightly closer to the actual event rate in the various subgroups.
Calibrations plots are shown for PREP-S for delivery within seven days (Figure S6, Table S10), two days (Figure S7 Table S11) and 14 days (Figure S8, Table S12), used as a prognostic model to determine time to delivery within a certain number of days as a binary outcome, not as a time-to-survival model. Evaluation of other thresholds for PlGF did not substantially improve test performance (Table S13) over and above the pre-specified thresholds.