This study was conducted at Marmara University Pendik Training and Research Hospital, and ethical approval was obtained from the Ethics Committee of Marmara University (Approval No: 09.2018.162, dated 06 January 2018). The study was performed in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its subsequent revisions. This was a retrospective cohort study of adult patients with hypertensive disorders of pregnancy. We included pregnant women aged $>$18 years who underwent caesarean section at our university hospital under regional or general anaesthesia following admission for GHT, preeclampsia, or eclampsia between 2012 and 2017. Patients who underwent surgeries other than caesarean section were excluded, as were those with incomplete data. Also excluded were patients admitted to the intensive care unit (ICU), since these represent a distinct population with a greater severity of illness, as well as different management protocols and outcome trajectories. The inclusion of such patients would likely have introduced confounding factors, thus reducing the internal validity of our study.

The requirement for patient consent was waived because of the retrospective nature of this study. Data was entered into a standardised data collection sheet on Microsoft Excel for Mac, Version 16.83 (Microsoft Corporation, Redmond, WA, USA). The following information was collected: (1) baseline demographic characteristics including age, body mass index (BMI), gestational age, concomitant maternal disorders, and laboratory results for creatinine and haemoglobin; (2) SBP and DBP readings at admission and delivery; (3) fetal and neonatal data including fetal birth weight, Appearance, Pulse, Grimace, Activity and Respiration (APGAR) scores at 1 and 5 minutes, prematurity, and IUGR; and (4) postpartum and postoperative outcomes including length of hospital stay, presence of postoperative severe arrhythmia, and pulmonary embolism. The length of hospital stay was defined as the duration between the first day of patient admission to the hospital and discharge. Gestational age was recorded as a continuous variable, but was categorized into the following clinically meaningful groups: preterm (37 weeks), term (37–41 weeks), and post-term ($\ge$42 weeks). This categorization facilitated the interpretation of outcomes and also aligns with standard obstetric practice.

Both SBP and DBP values were recorded to evaluate potential relationships between hypo- or hypertension and various outcomes, including length of hospital stay, fetal birth weight, and risks for pulmonary embolism, IUGR, premature delivery, and arrhythmia. Blood pressure was measured using an automated oscillometric device (Model HKD-02, manufactured by Hunkar Ltd. A.S., Denizli, Turkey). Measurements were obtained in the seated position after at least 5 min rest. For each patient, three consecutive readings were taken at one-minute intervals, with the average of the last two readings used for analysis. All measurements were performed by trained personnel using appropriately sized cuffs to ensure consistency and accuracy.

Participants were categorized into three different exposure groups to assess outcomes across various hypertensive and blood pressure parameters. The first categorization grouped patients by their specific diagnosis of hypertensive disorder of pregnancy (GHT, preeclampsia, or eclampsia). The second categorization grouped participants based on their DBP at delivery (90 mmHg or $\ge$90 mmHg), while the third grouped them by their SBP at delivery (140 mmHg or $\ge$140 mmHg).

Preeclampsia was diagnosed based on the presence of new-onset hypertension after 20 weeks of gestation, along with proteinuria $\ge$300 mg/24 h, or a urine dipstick reading $\ge$2+. In cases without significant proteinuria, diagnosis was made if there were systemic manifestations, including thrombocytopenia, renal or hepatic dysfunction, pulmonary edema, or neurological symptoms. To prevent the overestimation of cases, a urine dipstick reading of 1+ was not considered diagnostic.

Postoperative severe arrhythmia was defined as any clinically significant disturbance of cardiac rhythm occurring within the first 48 h after surgery and requiring intervention, including pharmacologic therapy or electrical cardioversion, or causing hemodynamic instability. Severe arrhythmias included sustained ventricular tachycardia, ventricular fibrillation, torsades de pointes, new-onset high-grade atrioventricular block, or rapid atrial fibrillation/flutter with hemodynamic compromise. All events were confirmed by continuous electrocardiogram (ECG) monitoring, telemetry, or 12-lead ECG.

Statistical comparisons were performed to evaluate differences across three groups: GHT, preeclampsia, and eclampsia. Numerical data were summarized as the mean and standard deviation if normally distributed, or as the median and interquartile range if they were not normally distributed. The distribution of continuous variables, such as age, birth weight, and BMI, was assessed using the Shapiro-Wilk test. Groups with normally distributed numerical data were compared with the Student’s t test and analysis of variance (ANOVA) test, while groups with skewed data were compared with the Kruskal-Wallis test and the Wilcoxon rank-sum test. Categorical data were presented as frequencies and percentages. Pearson’s Chi-squared test and Fisher’s exact test were used to test hypotheses and compare categorical data.

The relationships between exposures and outcomes of interest were evaluated using three models of multivariable logistic regression and linear regression. The first model examined associations between hypertensive disorders of pregnancy and maternal and fetal outcomes, the second model examined associations between DBP at delivery and the same outcomes, while the third model examined associations between SBP at delivery and the same outcomes. The logistic regression and linear regression models were adjusted for age and gestational diabetes mellitus (GDM). Logistic regression was utilized for categorical outcomes, which included the development of pulmonary embolism, IUGR, premature delivery, and postoperative arrhythmia. Linear regression was employed for continuous outcomes, including fetal birth weight and the length of hospital stay. For categorical outcomes, modified Poisson regression with robust variance estimation was used to calculate adjusted risk ratios (aRR). We chose this approach because several outcomes were common, and estimating aRRs provides a more interpretable measure of relative risk than odds ratios, which can overestimate the effect size in such scenarios.

The 95% confidence interval (95% CI) for aRRs and regression coefficients (Coef.) was also reported. Exact p-values were reported and interpreted as evidence against the null hypothesis. All statistical analyses were carried out using IBM SPSS Statistics 22 (IBM Corp, Armonk, NY, USA) and Stata 18.0 (StataCorp, College Station, TX, USA).

A significant difference in patient age was found between groups (p = 0.04). Pairwise analysis revealed that patients with preeclampsia were significantly younger compared to those with GHT (p = 0.015). Differences in the level of 24-h urine protein between groups were highly significant (p 0.001). Post-hoc analysis revealed that preeclampsia patients had significantly higher protein levels than GHT patients (p 0.001), while eclampsia patients had significantly higher protein levels than both of the other groups (p = 0.001).

A significant difference was found in the gestational age at delivery (p 0.001). Specifically, patients with preeclampsia had significantly earlier deliveries than those with GHT (p 0.001). Birth weight was also significantly different between the groups (p 0.001). Pairwise comparisons showed that preeclampsia patients had infants with a lower birth weight than those of GHT patients (p 0.001). Similarly, eclampsia patients also had infants with a lower birth weight than those of GHT patients (p = 0.019).

The duration of hospital stay differed significantly across the groups (p 0.001). Both preeclampsia (p 0.001) and eclampsia (p = 0.005) groups had significantly longer hospital stays compared to the GHT group. Primigravid status was significantly associated with the diagnostic group (p = 0.039), and was attributable to a significant difference between the preeclampsia and GHT groups (p = 0.049).

DBP at delivery demonstrated a significant overall difference between groups (p = 0.036). This was driven by a difference between the preeclampsia and GHT groups (p = 0.016), with no significant difference observed in the other pairwise comparisons. Creatinine levels were significantly different among the groups (p = 0.005). Pairwise comparisons confirmed that both preeclampsia (p = 0.005) and eclampsia patients (p = 0.040) had significantly higher creatinine levels compared to GHT patients.

Table 2 presents the blood pressure measurements, as well as the maternal and fetal outcomes, in each of the three hypertensive groups. No cases of maternal mortality occurred in this cohort. All groups had a similarly low incidence of postoperative arrhythmia and pulmonary embolism. Our analysis also revealed a significant association between prematurity and the diagnostic group (p 0.001), with the difference being primarily between the preeclampsia and GHT groups (p 0.001).

Further, the majority of patients had elevated SBP and DBP readings at delivery. The incidence of seizures was very different across the groups (p 0.001). As expected, a significant difference was observed between the preeclampsia and eclampsia groups (p 0.001), and between the eclampsia and GHT groups (p 0.001). The rates of seizure were negligible in the preeclampsia and GHT groups (Table 2).

We examined the associations between the type of hypertensive disorder including gestational hypertension, preeclampsia, and eclampsia together with maternal and fetal outcomes. Multivariable logistic regression analysis showed that compared to patients with GHT, patients with preeclampsia and eclampsia had 2.77-fold and 2.92-fold higher risks of developing IUGR, respectively (Table 3). Similarly, patients with preeclampsia and eclampsia had 3-fold and 4-fold higher risks of premature delivery, respectively, compared to those with GHT. The risk for postoperative arrhythmia was lower in the preeclampsia group, but this difference did not reach statistical significance. There was no difference in the risk of pulmonary embolism between patients with preeclampsia and GHT. Linear regression showed that both preeclampsia and eclampsia were both associated with significantly lower fetal birth weight compared to GHT. Further, they were both associated with a longer hospital stay compared to GHT. The association between eclampsia and the development of pulmonary embolism or postoperative arrhythmia could not be assessed due to an insufficient number of cases with these outcomes.

We evaluated the impact of diastolic blood pressure at delivery, comparing women with DBP 90 mmHg versus $\ge$90 mmHg. Patients with a DBP of 90 mmHg during delivery had consistently lower frequencies of pulmonary embolism, IUGR, premature delivery, postoperative arrhythmia, and seizures compared to patients with a DBP of $\ge$90 mmHg during delivery (Table 4). Adjusted linear regression revealed that a DBP of $\ge$90 mmHg during delivery was associated with a significantly lower fetal birth weight compared to patients with a DBP of 90 mmHg at delivery (Coef. –381.5, 95% CI: –739.6 to –23.4, p = 0.04). On the other hand, there was no difference in hospital length of stay or between the high and low DBP groups. Despite weak evidence against the null hypothesis, we observed 20% and 21% increases in the risks of IUGR and premature delivery, respectively, for patients with a DBP of $\ge$90 mmHg during delivery (Table 4). The risks for pulmonary embolism and seizures could not be analyzed, as none of the patients with a DBP of 90 mmHg at delivery developed these outcomes.

We analyzed the relationship between systolic blood pressure at delivery as 140 mmHg vs. $\ge$140 mmHg and the same set of outcomes. The effects of SBP during delivery on maternal and fetal outcomes are presented in Table 5. The mean fetal birth weight was slightly higher in patients with SBP $\ge$140 mmHg during delivery (mean 2423.4 g) compared to patients with SBP 140 mmHg (mean 2269.1 g). Both groups had an equal median hospital length of stay of three days. Almost no differences were observed in the frequencies of maternal and fetal outcomes between the two groups. Multivariable logistic regression showed no differences in the risk of developing IUGR or premature delivery in patients with SBP $\ge$140 mmHg during delivery compared to those with SBP 140 mmHg at delivery (Table 5).

The present study has several strengths. It included all eligible adult patients diagnosed with hypertensive disorders of pregnancy and who underwent caesarean section at our university hospital. Additionally, it investigated the impact of high and low SBP and DBP readings on pregnancy outcomes in a high-risk population. This provides valuable insights into the complications associated with uncontrolled blood pressure, highlighting the importance of proper blood pressure management. Our findings contribute to the development of management algorithms and guidelines for patients with hypertensive disorders of pregnancy.

Nonetheless, certain limitations should be acknowledged. First, this was a single-center study, which may limit the generalizability of our findings to other settings or populations. In particular, the sample size for the eclampsia group was small, and hence the findings for these patients should be interpreted with caution. Second, the retrospective cohort study design is subject to incomplete data and does not permit the establishment of causality. Future multicenter studies with larger sample sizes and longitudinal designs are recommended to further explore this topic, establish causality, and examine long-term effects. Although our regression analyses were adjusted for age and GDM, other important potential confounders such as BMI, parity, use of antihypertensive medication, and severity of proteinuria were not included. These unmeasured factors may have influenced outcomes, particularly since the eclampsia patients in our cohort were younger and more likely to be primigravida, both of which may independently affect maternal and fetal outcomes. The study draws exclusively on data collected 7–9 years ago (2012–2017). Over this extended interval, diagnostic algorithms, therapeutic guidelines, and demographic characteristics have evolved substantially. Contemporary clinical practice now incorporates newer pharmacologic agents, refined monitoring modalities, and earlier intervention strategies that may materially alter patient outcomes.