Antenatal corticosteroid prophylaxis in women with increased sFlt-1/PlGF ratio in the clinical routine – A retrospective analysis

Introduction

Preeclampsia (PE) and fetal growth restriction (FGR), particularly when occurring before 34 weeks’ gestation, remain major causes of maternal and perinatal morbidity and mortality worldwide [1], 2]. First-trimester-screening followed by prophylactic aspirin in high risk women can reduce the incidence of PE-related preterm births. However, specific therapeutic options to improve outcomes remain limited [3], 4]. When maternal and fetal symptoms deteriorate in PE and FGR, close monitoring and timely iatrogenic preterm delivery are essential [5]. The administration of antenatal corticosteroids (ACS) should be discussed with the expectant parents, as it significantly reduces fetal morbidity and mortality by enhancing fetal lung maturity, thereby lowering the risk of distress syndrome, as well as reducing the rates of intraventricular hemorrhage and necrotizing enterocolitis when delivery occurs at ≤34 weeks’ gestation [6], [7], [8], [9], [10]. In patients with FGR, the TRUFFLE-study provides guidance for delivery <32 weeks based on Doppler examinations and fetal heart tracings [6], [11], [12], [13]. Currently, no universally accepted clinical, laboratory, or hemodynamic thresholds reliably indicate when immediate delivery is indicated in PE. International guidelines recommend individualized decisions based on disease severity, which can make clinical management uncertain [14], [15], [16]. To improve risk-stratification, the German-Austrian-Swiss guideline for hypertensive pregnancy disorders recommends the use of the angiogenic biomarkers soluble fms-like tyrosinekinase-1 (sFlt-1) and Placental Growth Factor (PlGF) [17]. An sFlt-1/PlGF ratio ≤38 has a high negative predictive value for ruling out PE, while a ratio >85 is strongly associated with a significantly reduced remaining pregnancy duration [5], 18]. However, the guideline currently advises against using this ratio to determine timing of delivery or ACS administration [13]. Recent studies by Stepan and Siepen et al. suggest that an sFlt-1/PlGF ratio >655 at ≤34 weeks indicates the need for ACS, and Stolz et al. suggested ratio’s >1,000 may benefit neonatal outcomes [19], [20], [21].

In our tertiary perinatal center, the sFlt-1/PlGF ratio measurement has been routinely implemented since 2010 to identify women at risk of PE-related pregnancy outcomes. This study aimed to retrospectively analyze a 10-year-cohort of women with suspected PE or FGR <34 weeks’ gestation to assess pregnancy outcomes, indications for ACS, and the predictive value of the sFlt-1/PlGF ratio for PE-related preterm delivery.

Materials and methods

Results

Of 1,928 eligible patients, n=801 patients met inclusion criteria and presented between August 2010 and December 2019 with signs and symptoms of PE and FGR between 22+0- and 33+6 weeks’ gestation. As shown in Figure 1, after excluding women who had received ACS before sFlt-1/PlGF measurement, as well as those with intrauterine fetal death or elective termination, n=702 were included (n=530 (75.5 %) with low-risk according to an sFlt-1/PlGF ratio <85, n=172 (24.5 %) high-risk with an sFlt-1/PlGF ratio ≥85). Among women classified with low-risk, n=67 (12.6 %) received ACS, whereas n=463 (87.4 %) did not. In the high-risk group, 128 (74.4 %) received ACS and 44 (25.6 %) did not. Correct ACS administration-defined as PE-related birth ≤34+0 weeks after treatment-was achieved in 118 of 129 (92.2 %) high-risk and 33 of 67 (49.3 %) low-risk women.

Figure 1:

Inclusion procedure of the retrospective cohort.

Among the n=41 women classified as false positives – defined as high risk according to an sFlt-1/PlGF ratio ≥85 but without a PE-related delivery at ≤34+0 weeks – only 10/41 (24.4 %) received ACS. These patients delivered at a median gestational age of 34+6 (31+2–35+4) weeks, and none delivered beyond 37+0 weeks. In contrast, among the 34/530 “false-negatives” (low-risk with sFlt-1/PlGF ratio <85 but PE-related birth ≤34+0 weeks), 33/34 (97.1 %) received no ACS (Figure 1).

A total of n=196 (27.9 %) women experienced preterm delivery at ≤34+0 weeks, in n=165 of these cases, the preterm delivery was related to PE. Baseline maternal characteristics (age, body-mass-index, ethnicity, smoking) were comparable between the groups (Table 1). High risk women were more often nulliparous (high risk with ACS 50 %, without ACS 47.7 %, low risk with ACS 28.4 %, without ACS 30.7 %, p<0.001), while a history of hypertension was more frequent among those receiving ACS (ACS and high risk 20.3 %, ACS and low risk 20.9 %, no ACS and high risk 2.3 %, no ACS and low risk 13.8 %, p<0.001). Median uterine artery pulsatility indices (p<0.001) as well as median umbilical artery pulsatility indices (p<0.001) were highest in the high-risk ACS group (median mean uterine artery PI 1.67, IQR 1.31–1.92) (median umbilical artery PI 1.39, IQR 1.16–1.80). Classical PE-symptoms, including new-onset of hypertension (44.5 and 47.5 %) and proteinuria (27.3 and 29.5 %) were most frequent in the high-risk groups (both <0.001). Suspicion of FGR was most common in high-risk women with ACS (46.9 %, p<0.001).

Pregnancy outcomes are summarized in Table 2. Nearly all women in the high-risk ACS group (98.4 %) had a PE-related indication for delivery, with 92.2 % delivering ≤34+0 weeks. In comparison, 84.1 % of the high-risk women without ACS (29.5 % ≤34 weeks) and 62.7 % of low-risk women with ACS (49.3 % ≤34 weeks) delivered due to PE-related complications (p<0.001). The median interval from ACS to delivery was shorter in the high-risk vs. low-risk group (4.0 vs. 8.0 days (p<0.001)); 49.2 % of high-risk and 39.2 % of low-risk women delivered within the optimal window of 48 h to 7 days after ACS (p=0.232) (Table 3).

Median remaining pregnancy duration and gestational age at delivery differed significantly (p<0.001): High-risk women with ACS delivered after 7 days (IQR 3.0–17.8) at 29+5 weeks (IQR 26+6–32+0), high-risk women without ACS after 16.5 days (IQR 3.0–25.8) at 34+3 weeks (IQR 32+6–36+3), low-risk women with ACS after 28 days (IQR 11.0–43.0) at 32+3 weeks (IQR 29+5–34+0), and low-risk women without ACS after 60.0 days (IQR 42.0–80.0) at 38+1 weeks (IQR 37+0–39+3) (Figure 2). The earlier deliveries high-risk groups explained the high rate of neonatal respiratory distress syndrome (RDS): 70.3 % in the high-risk with ACS, 44.8 % in low-risk with ACS, 15.9 % in high-risk without ACS and 0.4 % in low-risk without ACS, (p<0.001). Table 4 summarizes the indications for preterm delivery ≤34 weeks’ gestation: In the high-risk group without ACS, 5 of 16 women (31.3 %) experienced placental abruption, and 3 of 16 (18.8 %) developed HELLP syndrome.

Figure 2:

Kaplan-Meier-curve showing (A) the remaining pregnancy duration between high- and low-risk-patients with and without ACS and (B)  Gestational age at delivery in these groups. Differences have been calculated with Log-Rank-test. The significance level is p=0.05. Red line: high-risk-group with sFlt-1/PlGF ≥85 with ACS; grey line: high-risk-group with sFlt-1/PlGF ≥85 without ACS; blue line: low-risk-group with sFlt-1/PlGF <85 with ACS; green line: low-risk-group with sFlt-1/PlGF <85 without ACS.

In contrast, low-risk groups more frequently had non-PE-related indication for delivery, such as preterm contractions (1 of 9 (11.1 %) without ACS; 7 of 51 (13.7 %) with ACS) or premature rupture of membranes (1 of 9 (11.1 %) without ACS; 3 of 51 (5.9 %) with ACS). Supplementary Table 1 shows the delivery indications for the 165/196 patients with PE-related delivery at ≤34 weeks’ gestation and the corresponding reasons for ACS administration.

The sFlt-1/PlGF ratio predicted PE-related delivery ≤34+0 weeks with an AUC of 0.95 (95 % CI 0.94–0.97), optimal cut-off of 52.5, sensitivity 87.9 % (95 % CI: 82.0–92.0), specificity 87.9 % (95 % CI 94.9–90.4), positive predictive value (PPV) of 69.1 % (95 % CI 62.5–74.9) and a negative predictive value (NPV) of 95.9 % (95 % CI 93.8–97.4) (Figure 3). Median sFlt-1/PlGF ratio was 252.0 (IQR 101.5–515.0) in women with PE-related delivery ≤34+0 weeks vs. 5.0 (IQR 3.0–12.0) in others (p<0.001). The corresponding time to delivery was 8.0 days (IQR 3.0–21.0) vs. 56.0 days (IQR 36.0–78.0) in the two groups (p<0.001).

Figure 3:

sFlt-1/PlGF-ratio of women with- and without PE-related delivery ≤ 34+0 weeks. (A) Box plots showing the median sFlt-1/PlGF ratio of the cohorts with- and without PE-related delivery −34+0 weeks. In blue: n=537 patients without PE-related delivery −34+0 weeks, in red: n=165 patients with PE-related delivery −34+0 weeks. (B) Receiver-operating-curve (ROC) showing the predictive value of the sFlt-1/PlGF ratio for an indication to ACS because of a PE-related preterm birth ≤34+0 SSW generated an AUC of 0.95 (0.94–0.97), p<0.001.

Applying previously suggested cut-off-values (>655 or >1,000) yielded specificities of 100 % but sensitivities of only 17.6 and 7.3 %, respectively [19], [20], [21]. In this post-hoc validation of the present cohort, a specificity of 95 % (95 % CI 0.93–0.97) was achieved at a cut-off of 120 (sensitivity 71 %, 95 % CI 0.63–0.78), and a maximal specificity of 99 % (95 % CI 0.98–1.00) at a cut-off of 255 (sensitivity 49 %, 95 % CI 0.98–1.00).

Discussion

Conclusions

In this real-world cohort, one third of women with suspected PE or FGR required ACS, and appropriate decisions were made in 90 % of cases. The sFlt-1/PlGF ratio showed excellent predictive accuracy for PE/FGR-related delivery ≤34+0 weeks (AUC 0.95), confirming its clinical utility as a decision-support biomarker. An sFlt-1/PlGF ratio >255 should prompt counseling and consideration of ACS administration, balancing high specificity with timely intervention to optimize perinatal outcomes.

For prospective evaluation, as the sFlt-1/PlGF ratio is already well established as a standard tool for assessing the risk of PE-related adverse outcomes, a fully blinded assessment of its role in guiding ACS administration would be ethically challenging. Nevertheless, the present study may help support the ethical justification for a future prospective design. As a long-term goal, an algorithm integrating clinical signs, biomarkers, and their dynamic changes should be developed to provide structured guidance for ACS administration.