We conducted a hospital-based observational, analytical cross-sectional study over a six-month period, from 01 March to 01 September 2026, at the Feto-Maternal Unit, Department of Obstetrics and Gynaecology, Dhaka Medical College Hospital, Dhaka, Bangladesh, a tertiary referral centre for high-risk pregnancies.

Consecutive women admitted with preeclampsia with severe features during the study period were enrolled until 216 participants were included. With this sample, the prevalence of any coagulation abnormality could be estimated with an absolute precision of approximately ±6.4% at the 95% confidence level. Preeclampsia was defined as new-onset hypertension after 20 weeks of gestation in a previously normotensive woman, with systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg on two measurements at least 4 hours apart, accompanied by proteinuria or evidence of maternal end-organ dysfunction [20]. Severe features included severe-range blood pressure (systolic ≥160 mm Hg or diastolic ≥110 mm Hg), platelet count <100,000/µL, renal insufficiency (serum creatinine ≥1.1 mg/dL or doubling of baseline in the absence of other renal disease), liver transaminases more than twice the upper limit of normal or persistent right upper-quadrant/epigastric pain not explained by another diagnosis, pulmonary oedema, new-onset headache unresponsive to medication and not accounted for by an alternative diagnosis, or visual disturbance. Women with preeclampsia with severe features who also had an eclamptic seizure (n = 12) were included in the study group, and seizure episode was analysed as a marker of maternal disease severity (Section 2.4). Women with known bleeding disorders predating pregnancy, current anticoagulant therapy, chronic liver disease, platelet disorders unrelated to preeclampsia, or refusal to participate were excluded.

After written informed consent, maternal sociodemographic characteristics, obstetric history, relevant comorbidities, presenting clinical findings, blood pressure, urine protein, antenatal ultrasonographic findings, gestational age at admission and at delivery, mode of delivery, and neonatal birthweight were recorded using a structured data-collection form. Venous blood obtained at the index admission was analysed in the hospital laboratory by standard automated laboratory methods for complete blood count, platelet count, prothrombin time (PT), activated partial thromboplastin time (aPTT), D-dimer, serum creatinine, and alanine aminotransferase. D-dimer was measured in citrated plasma by a quantitative latex-enhanced immunoturbidimetric assay on an automated coagulation analyser and is reported in mg/L fibrinogen-equivalent units (FEU).

The antenatal scan used for outcome classification was the ultrasound and Doppler examination performed during the index admission, at or after admission blood sampling and before delivery. Coagulation status was therefore established before, or at the same time as, the scan findings and never after them. Antenatal scan categories were abstracted from the contemporaneous ultrasound/Doppler report according to the criteria given in Section 2.4. There were no missing observations for variables included in the reported analyses.

For the coagulation analysis, thrombocytopenia was defined as platelet count <150,000/µL, the conventional definition of thrombocytopenia in pregnancy [21]; platelet count <100,000/µL was treated as a more severe, nested subgroup. The prespecified study cut-offs for prolonged PT, prolonged aPTT, and elevated D-dimer were >16.1 seconds, >40 seconds, and >1.7 mg/L FEU, respectively; each is the upper reference limit issued by the hospital laboratory for the assay in use. These laboratory limits are not gestation-specific, and the implications for D-dimer are considered in the Discussion. Any coagulation abnormality was defined as the presence of at least one abnormality among platelet count, PT, aPTT, and D-dimer. Participants with none of these abnormalities were classified as having no predefined coagulation abnormality.

Maternal disease-severity markers were severe-range blood pressure, urine protein 3+/4+, serum creatinine ≥1.1 mg/dL, alanine aminotransferase ≥70 U/L, and seizure episode. Obstetric outcomes were preterm delivery (<37 weeks) and early preterm delivery (<34 weeks). Fetal and placental outcomes included fetal growth restriction, any abnormal antenatal scan, low birthweight (<2500 g), and birthweight in kilograms. Any abnormal antenatal scan was a composite outcome defined as fetal growth restriction, oligohydramnios, or placental insufficiency documented on antenatal assessment. Fetal growth restriction was defined as an ultrasonographic estimated fetal weight or abdominal circumference below the 10th centile for gestational age [22]. Oligohydramnios was defined as an amniotic fluid index ≤5 cm or a single deepest vertical pocket <2 cm [23]. Placental insufficiency was defined as abnormal umbilical artery Doppler velocimetry, that is, resistance indices raised for gestational age or absent or reversed end-diastolic flow [22]. These were the criteria applied by the reporting sonologists in routine practice, and each category was abstracted as documented in the report.

Clinical history/comorbidity burden was calculated as the number of recorded background factors present: family history of cardiovascular disease, hypertension in a previous pregnancy, dyslipidaemia, pregestational diabetes mellitus, polycystic ovary syndrome, chronic kidney disease, hypothyroidism, inherited/acquired thrombophilia, and autoimmune disease.

Data were entered, cleaned, and analysed using SPSS version 25. Continuous variables were summarised as mean ± standard deviation (SD) when approximately normally distributed and as median (interquartile range, IQR) when skewed; categorical variables were summarised as frequency and percentage. The prevalence of any and individual coagulation abnormalities was estimated for the full cohort; exact binomial 95% confidence intervals were used for the prevalence estimates displayed in Figure 1.

Participants were categorised as having no coagulation abnormality or any coagulation abnormality using the prespecified thresholds. Categorical maternal, obstetric, and fetal variables were compared with the Pearson chi-square test; Fisher’s exact test was used when expected cell counts were <5. Where the chi-square P value lay close to the significance threshold (0.04 to 0.06), the Fisher’s exact P value is reported alongside it as a check of robustness, and such results are interpreted as borderline. Continuous variables were compared using the independent-samples t test or Mann-Whitney U test, as appropriate. Crude odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for categorical outcomes.

Multivariable logistic regression was used to estimate the association between any coagulation abnormality and any abnormal antenatal scan, adjusting for age, body mass index, severe-range blood pressure, proteinuria 3+/4+, serum creatinine ≥1.1 mg/dL, and clinical history/comorbidity burden. Multivariable linear regression was used to estimate the association between any coagulation abnormality and birthweight, adjusting for the same maternal covariates and gestational age at delivery. Covariates were selected a priori as potential confounders reflecting overall maternal disease severity; because some may also lie on the pathway between haemostatic derangement and placental dysfunction, the adjusted estimates represent associations independent of concurrent disease severity and may be conservative. Adjusted associations are reported as adjusted ORs (aORs) or regression coefficients (β) with 95% CIs. All tests were two-sided, and P < 0.05 was considered statistically significant. P values are reported to three decimal places, and values below 0.001 are reported as P < 0.001.

Gestational age was handled differently in the two models. In the scan model, gestational age at delivery was not entered because delivery follows the scan and its timing may be determined by the scan findings; adjusting for it would condition on a consequence of the outcome. Gestational age at admission, although recorded, was not part of the prespecified covariate set, which was restricted to maternal characteristics and markers of concurrent disease severity, and two of the three components of the composite scan outcome are themselves referenced to gestational age (centile-based fetal biometry and gestation-specific Doppler indices). In the birthweight model, gestational age at delivery was included because it is the principal determinant of birthweight, and the aim was to estimate the association with fetal size that is not explained by gestational duration. Because gestational age at delivery may itself be shortened by disease severity, it may act as an intermediate variable; the gestational-age-adjusted coefficient is therefore interpreted as the association that does not operate through earlier delivery, and the unadjusted mean difference, which does not condition on gestational age, is reported alongside it.

Two sensitivity analyses were performed. First, because a platelet count <100,000/µL is both a severe-feature criterion and a component of the exposure, the cohort was restricted to women with a platelet count ≥100,000/µL, all of whom necessarily qualified for severe features through a criterion other than platelet count, and the prevalence of any coagulation abnormality was re-estimated. For the outcome associations, a tipping-point (bounding) analysis was applied to the crude estimates in this restricted cohort: we determined how many of the excluded women would need to have had the outcome for the association to lose statistical significance, and we report the estimate under the most extreme allocation. Second, to gauge the potential influence of gestational age at admission and of other omitted or unmeasured confounders on the scan model, we calculated the E-value for the adjusted odds ratio and for the confidence limit closer to the null, using the square-root transformation for a common outcome [24].

The study protocol was reviewed and approved by the Institutional Review Board of Dhaka Medical College (IRB No. IRB-DMC/2026/111) and was conducted in accordance with the Declaration of Helsinki (2024). Written informed consent was obtained from all participants prior to enrolment. Participation was voluntary, and participants’ privacy and confidentiality were maintained throughout the study. Participants were informed of their right to withdraw at any time without penalty or any effect on their clinical care.

The mean maternal age was 28.5 ± 5.4 years and mean BMI was 30.9 ± 4.4 kg/m²; 25.0% were nulliparous, and the median gravidity was 3 (IQR, 2 - 4). Mean gestational age at delivery was 35.2 ± 2.7 weeks. Anaemia was present in 58.8%, and 27.3% reported a family history of cardiovascular disease; the most frequent background comorbidities were hypertension in a previous pregnancy (23.6%) and pregestational diabetes mellitus (15.7%) (Table 1).

Table 1. Baseline maternal and obstetric characteristics (N = 216).

Among the 216 women with preeclampsia with severe features, 147 (68.1%) had severe systolic hypertension (≥160 mm Hg), while 112 (51.9%) had severe diastolic hypertension (≥110 mm Hg). Proteinuria of 3+/4+ was present in 70 (32.4%) women. Thrombocytopenia with a platelet count <150,000/µL was observed in 105 (48.6%) participants, including 30 (13.9%) with a platelet count <100,000/µL. Prolonged PT (>16.1 s) and prolonged aPTT (>40 s) were found in 36 (16.7%) and 42 (19.4%) women, respectively. An elevated D-dimer level (>1.7 mg/L) was observed in 85 (39.4%) participants, while elevated serum creatinine (≥1.1 mg/dL) was noted in 92 (42.6%). Overall, 140 (64.8%) women had at least one coagulation abnormality (Table 2).

Table 2. Severity features and coagulation profile of women with preeclampsia with severe features (N = 216).

Compared with women without a coagulation abnormality, those with any coagulation abnormality more often had severe-range blood pressure (93.6% vs 60.5%), serum creatinine ≥1.1 mg/dL (55.7% vs 18.4%), and alanine aminotransferase ≥70 U/L (25.7% vs 2.6%) (all P < 0.001), as well as proteinuria 3+/4+ (37.1% vs 23.7%; P = 0.044; Fisher’s exact P = 0.049). Seizure episodes were numerically more frequent (7.9% vs 1.3%), but the difference was not statistically significant (Fisher’s exact P = 0.060) (Table 3).

Women with coagulation abnormalities had higher rates of abnormal antenatal scans (62.1% vs 39.5%; P = 0.001) and low birthweight (82.1% vs 68.4%; P = 0.021). Fetal growth restriction was numerically more frequent (38.6% vs 25.0%); the prespecified chi-square test gave P = 0.044 (crude OR 1.88; 95% CI 1.01 - 3.50), but the result was not robust to the choice of test (Fisher’s exact P = 0.051). We therefore regard this difference as borderline and do not interpret it as an established association. After adjustment, coagulation abnormalities remained associated with increased odds of an abnormal antenatal scan (aOR, 3.20; 95% CI, 1.60 - 6.40; P = 0.001) and lower birthweight (β, −0.151 kg; 95% CI, −0.270 to −0.032; P = 0.013). The birthweight difference was −0.22 kg (95% CI −0.32 to −0.12) before any adjustment and −0.151 kg after adjustment for maternal covariates and gestational age at delivery; the association was therefore attenuated, but not removed, after adjustment that included gestational age at delivery. Preterm and early preterm delivery were not significantly different between groups (Table 4).

Table 3. Maternal clinical and biochemical characteristics according to coagulation status.

Table 4. Fetal and obstetric outcomes according to coagulation status and adjusted associations.

Crude P values for categorical outcomes are from the Pearson chi-square test; for fetal growth restriction, Fisher’s exact P = 0.051. OR, odds ratio; aOR, odds ratio adjusted for age, body mass index, severe-range blood pressure, proteinuria 3+/4+, serum creatinine ≥1.1 mg/dL and clinical history/comorbidity burden; MD, unadjusted mean difference; β, coefficient adjusted for the same covariates and gestational age at delivery.

Figure 1 presents the frequency of each coagulation abnormality with exact binomial 95% CIs: any coagulation abnormality 64.8% (95% CI 58.0 - 71.2), platelet count <150,000/µL 48.6% (41.8 - 55.5), D-dimer >1.7 mg/L 39.4% (32.8 - 46.2), aPTT >40 s 19.4% (14.4 - 25.4), and PT >16.1 s 16.7% (12.0 - 22.3).

Figure 1.Frequency (with exact 95% confidence intervals) of coagulation abnormalities in preeclampsia with severe features (N = 216).

All 30 women with a platelet count <100,000/µL were, by definition, in the coagulation-abnormality group. Among the remaining 186 women, who had a platelet count ≥100,000/µL and therefore qualified for severe features through a criterion other than platelet count, 110 (59.1%; 95% CI 51.7 - 66.3) had at least one coagulation abnormality, compared with 64.8% in the full cohort. In the tipping-point analysis of this restricted cohort, the crude association with an abnormal antenatal scan remained statistically significant unless at least 28 of the 30 excluded women (93.3%) had an abnormal scan, a proportion well above the 62.1% observed in the coagulation-abnormality group as a whole. Under the most extreme allocation, in which all 30 excluded women had an abnormal scan, the association was attenuated but remained in the same direction (51.8% vs 39.5%; crude OR 1.65; 95% CI 0.91 - 2.98; P = 0.097). The crude birthweight difference would have been abolished only if the mean birthweight of the 30 excluded women had been approximately 1.27 kg or less, against a group mean of 2.08 kg. The crude association with low birthweight was less robust: it would lose statistical significance if 26 or more of the 30 excluded women (86.7%) had a low-birthweight infant, a proportion close to the 82.1% observed in the group overall.

For the scan model, the E-value was 2.98 for the adjusted odds ratio and 1.84 for its lower confidence limit. An omitted or unmeasured confounder would therefore need to be associated with both coagulation abnormality and an abnormal scan by a risk ratio of at least 2.98, beyond the measured covariates, to explain away the observed association, and by at least 1.84 to move the confidence interval to include the null. For comparison, delivery before 34 weeks and before 37 weeks, the available proxies for earlier gestational age at presentation, were only 1.11 and 1.20 times as frequent, respectively, in women with a coagulation abnormality as in those without (Table 4).