Analysis of peripheral blood variables in evaluation and prediction of adverse outcomes with hypertensive disorders of pregnancy

Abstract

Background

Hypertensive disorders of pregnancy (HDP) constitute a major obstetric complication globally, contributing significantly to perinatal morbidity and mortality, as well as to the incidence of severe pregnancy-related adverse outcomes. This study aims to evaluate the selected biomarkers and other parameters in the diagnosis and prediction of adverse pregnancy outcomes in women with HDP.

Methods

We enrolled a total of 165 pregnant women with HDP and 41 healthy pregnant women. Patients with HDP were subcategorized into gestational hypertension (GH), preeclampsia (PE), and severe preeclampsia (sPE). Using univariate analysis, we identified selected biomarkers and routine prenatal clinical indicators and further evaluated their associations with adverse pregnancy outcomes. Finally, a predictive nomogram was then developed using multivariable logistic regression. The predictive performance of the model was comprehensively evaluated through receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis, and clinical impact curves.

Results

Levels of N-terminal pro-brain natriuretic peptide (NT-proBNP), lactate dehydrogenase (LDH), and platelet distribution width (PDW) showed a positive correlation with HDP severity. Notably, concentrations of these three peripheral blood variables were significantly higher in patients who experienced adverse pregnancy outcomes compared to those HDP patients without complications (P < 0.01). Multivariate logistic regression analyses further confirmed that gestational age (GA), NT-proBNP, LDH, and PDW are independent risk factors for adverse pregnancy outcomes in HDP patients (P < 0.01). The ROC curves demonstrated that combining GA with NT-proBNP, LDH, and PDW provided stronger predictive value for adverse outcomes. Furthermore, the nomogram's calibration curve indicated excellent model calibration. The decision curve analysis (DCA) showed that the nomogram is of practical significance.

Conclusions

Serum NT-proBNP, LDH, and PDW expressions are strongly associated with HDP severity, and these peripheral blood variables, along with gestational age (GA), are independent risk factors for predicting adverse outcomes in pregnant women with HDP. Moreover, constructed a nomogram that can easily predict and evaluate the risk of adverse pregnancy outcomes in women with HDP.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-07954-w.

Background

Hypertensive disorders of pregnancy (HDP) are one of the most common health problems during pregnancy, affecting approximately 3%−5% of all pregnancies and contributing substantially to maternal and fetal morbidity and mortality [1]. Disease severity varies from mild to severe and may progress to life-threatening end-stage complications, such as hemolysis, elevated liver enzymes, and low platelet count syndrome, and maternal organ dysfunction [2, 3]. Although the pathophysiology of HDP is not understood fully, the current theory suggests that genetic and systemic inflammatory factors in HDP lead to maternal systemic inflammatory responses and endothelial dysfunction [4, 5]. Moreover, women with HDP experience cardiovascular alterations, such as increased blood pressure and peripheral vascular resistance, which increase the risk of premature cardiovascular disease. N-terminal pro-brain natriuretic peptide (NT-proBNP) is produced in the ventricle as a response to increased blood volume and left ventricular wall tension [6, 7]. Due to systemic inflammatory response, injured myocardial cells increase the release of NT-proBNP in patients with HDP, especially in those who later develop complications, preeclampsia (PE), and severe preeclampsia (sPE)with adverse maternal outcomes [7, 8]. Studies have demonstrated that NT-proBNP can serve as an effective biomarker for early cardiac dysfunction and may have potential diagnostic value in predicting HDP during pregnancy [9]. In addition, soluble fms-like tyrosine kinase-1 (sFlt-1) and placental growth factor (PlGF) are related to the onset and severity of HDP, and a higher sFlt-1/PlGF ratio indicates greater disease severity and worse outcome [10]. Despite its effectiveness, the use of the sFlt-1/PlGF ratio is limited by high cost and poor accessibility, especially in emergency settings, rural areas, and developing countries. In China, disparities in healthcare infrastructure further highlight the urgent need for reliable, low-cost prediction models that can be implemented at the primary care level. While traditional biomarkers such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), uric acid (UA), and blood urea nitrogen Alanine (BUN) have been previously evaluated for their association with severity of HDP [11, 12], the value of different combinations of various promising biomarkers has not been thoroughly studied within the setting of a routine clinical blood test. Notably, NT-proBNP is widely available in most of these settings, since it is part of diagnostic protocols when evaluating patients with heart failure. Therefore, this study intends to combine serum NT-proBNP and other clinical biomarkers that are relatively cost-effective and readily accessible to predict pregnancy outcome in patients with HDP, offering information for clinical evaluation and potential applications.

Materials and methods

A retrospective case–control study was conducted at Jiangyin People’s Hospital during the period from January 2020 to December 2021. According to standard criteria of HDP published by the American Obstetrics and Gynecology (ACOG, 2013) [13]: Gestational hypertension (GH) is the new onset of hypertension with blood pressure more than or equal to 140/90 mmHg at mid-pregnancy (≥ 20 weeks of gestation) without proteinuria. Preeclampsia is gestational hypertension with proteinuria of 300 mg/24 h in urine. Severe preeclampsia (sPE): systolic blood pressure (SBP) ≥ 160 mmHg or diastolic blood pressure (DBP) ≥ 110 mmHg on two occasions within a 2-h interval, proteinuria 300 mg/24 h in urine, severe headache, epigastric pain, visual disturbance, pulmonary edema, or HELLP syndrome, liver function, or thrombocytopenia (< 100 000/µL). Inclusion criteria: gestational age of 20–42 weeks, singleton pregnancy with a diagnosis of HDP on admission. Therefore, patients diagnosed with HDP were selected as the research objects and divided into GH (45 cases), PE (47 cases), and sPE groups (73 cases), respectively. 41 cases of healthy pregnant women were taken as controls. Exclusion criteria: Patients with chronic hypertension less than 20 weeks gestation (chronic hypertension), hyperthyroidism before pregnancy, chronic liver disease, thyroid disease, pregestational diabetes, metabolic syndrome, chronic kidney disease, multiple pregnancy, fetal chromosomal and structural anomalies, other cancer and blood diseases. The sample size consideration was estimated by the following formula n = Z_α/2 × P × (1-P)/d² based on the previous study [14]. Here, with the prevalence of preeclampsia of 3–5% and Z_α/2 is the level of significance at 95%, i.e., 95% confidence interval = 1.96, P = prevalence of HDP = 0.05 (after rounding), d = desired error of margin = 5% = 0.05. The minimum estimated sample size for the present study was 37 participants in each group. The number of participants in our study was above the required sample size. The data on age, gestational age, pregnancy weight, pregnancy height, SBP, DBP of cases and controls were obtained through medical records and the laboratory information system. body mass index (BMI) was calculated as kg/m². The outcomes of HDP patients were recorded at the end of pregnancy. Adverse pregnancy outcomes mainly included low birth weight (LBW), preterm, intrauterine growth restriction (IUGR), placental abruption, neonatal intensive care unit (NICU), postpartum hemorrhage, miscarriage, fetal distress, neonatal asphyxia, and fetal death in utero. Detection of serum NT-proBNP, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), uric acid (UA), blood urea nitrogen Alanine (BUN), serum creatinine (SCr), alkaline phosphatase (ALP) and lactate dehydrogenase (LDH): 4 mL of fasting venous blood was collected from patients woke up in the morning, the serum samples were separated after being centrifuged at 3000 rpm/min for 10 min. NT-proBNP was measured by the electrochemiluminescence technique, and ALT, AST, UA, BUN, SCr, ALP, and LDH were measured by enzyme colorimetry. The reagents and calibrators were Roche original reagents. The white blood cell (WBC), platelet (PLT), mean platelet volume (MPV), and platelet distribution width (PDW) were detected and calculated by an automatic hematology analyzer (Sysmex XN-9000). Our study was approved by the Academic Ethics Committee of the Affiliated Jiangyin Hospital of Xuzhou Medical University [Approval number: 2024ER (076)]. Due to the retrospective nature of this study, informed patient consent was waived by the Ethics Committee of the Affiliated Jiangyin Hospital of Xuzhou Medical University. All methods were performed in accordance with relevant guidelines and regulations.

Statistical analysis

SPSS (version 20.0) was used for statistical analysis. Continuous data of normal distribution was presented as mean ± S.D, and their differences were compared by Student’s t test. One-way analysis of variance (ANOVA) was applied for comparisons of differences among multiple groups. Normal distribution of continuous variables was analyzed by Levene's test. The chi-square test was performed for categorical variables comparison. Non-normal distribution data was expressed as M (P₂₅-P₇₅), the Mann–Whitney test was for two-group comparison, and the Kruskal–Wallis tests were used for comparison among different groups. Multivariate logistic regression analysis with a stepwise forward selection method was conducted to identify independent predictors (P < 0.05) for adverse pregnancy outcomes. These predictors were then used to construct a predictive nomogram. The diagnostic performance of individual parameters was assessed using receiver operating characteristic (ROC) curves to determine optimal cut-off values. The calibration curve was used to measure the calibration of the nomogram. Internal validation was initially analyzed by bootstrapping with 1000 random samples drawn with replacement. The net benefit of the model was evaluated by decision curve analysis (DCA). P < 0.05 was considered statistically significant.

Results

Baseline characteristics and clinical blood biochemical indicators in HDP patients

A total of 165 patients and 41 healthy pregnancy controls were retrospectively included in the study. Out of 165 HDP cases, 45 (27.27%) were classified as GH, 47 (28.48%) as PE, and 73 (44.24%) as sPE, respectively. Baseline characteristics were analyzed among women with HDP and healthy controls. As shown in Table 1, there was an increasing trend in BMI, SBP and DBP with the severity of disease, and the highest values were observed in the sPE group (P < 0.01). In comparison with the control and GH groups, women in the sPE group exhibited significantly lower GA (P < 0.01). Although the levels of BUN, SCr, ALT, and AST showed significant differences among the groups, their values remained within the normal reference range. In addition, there was no notable difference in age distribution, ALP, WBC, PLT, and MPV among the control and PE groups (P > 0.05). In Table 1, the levels of serum NT-proBNP, LDH, PDW and UA in sPE group were markedly increased than those in the PE, GH and healthy pregnancy groups [the median of NT-proBNP levels:266.8(160.60, 379.65), 144.60(77.80, 209.50), 42.50(30.00, 61.40) and 20.50(12.05, 28.00) pg/mL, respectively; mean ± SD of LDH levels: (255.26 ± 66.36), (228.54 ± 59.67), (208.77 ± 53.66) and (178.43 ± 37.7) IU/L, respectively; mean ± SD of PDW levels: (15.60 ± 2.01), (15.05 ± 2.01), (14.16 ± 2.02) and (12.93 ± 2.19) fL respectively; mean ± SD of UA levels: (397.04 ± 95.69), (383.55 ± 108.92), (325.37 ± 79.26) and (284.65 ± 60.78) IU/L], the differences were statistically significant (P < 0.01). Adverse pregnancy outcomes were observed in 8 cases of GH, 22 cases of PE, and 55 cases of sPE groups, respectively, with the sPE group showing the highest incidence.

Table 1.

General characteristics of the enrolled women

Variable	Controls	GH	PE	sPE	P-value
	(n = 41)	(n = 45)	(n = 47)	(n = 73)
Age, years	28.20 ± 4.79	29.24 ± 5.52	29.81 ± 5.83	30.52 ± 6.19	0.21
BMI (kg/m2)	28.17 ± 3.84	30.40 ± 3.30	30.70 ± 4.51	31.59 ± 4.72	< 0.001
GA	38.86 ± 0.90	38.39 ± 1.78	37.56 ± 2.25	35.64 ± 3.66	< 0.001
SBP (mmHg)	117.83 ± 19.64	141.13 ± 9.97	148.68 ± 9.60	165.55 ± 17.14	< 0.001
DBP (mmHg)	81.10 ± 6.60	95.20 ± 6.69	99.00 ± 4.78	109.04 ± 15.42	< 0.001
ALP (IU/L)	124.49 ± 36.70	139.94 ± 53.53	147.43 ± 47.68	136.00 ± 45.57	0.138
LDH(IU/L)	178.43 ± 37.75	208.77 ± 53.66	228.54 ± 59.67	255.26 ± 66.36	< 0.001
ALT (IU/L)	9.46 ± 4.63	9.76 ± 5.93	9.69 ± 5.61	13.12 ± 8.00	0.004
AST (IU/L)	14.93 ± 4.48	15.62 ± 6.10	17.06 ± 7.23	20.06 ± 7.68	< 0.001
BUN (mmol/L)	3.48 ± 1.23	3.37 ± 0.90	3.84 ± 1.30	4.17 ± 2.00	0.02
SCr (umol/L)	49.42 ± 9.55	55.24 ± 12.20	60.51 ± 14.78	62.87 ± 25.80	0.002
UA (umol/L)	284.65 ± 60.78	325.37 ± 79.26	383.55 ± 108.92	397.04 ± 95.69	< 0.001
WBC (109/L)	9.13 ± 2.17	9.28 ± 1.92	8.59 ± 1.98	9.43 ± 2.20	0.187
PDW(fL)	12.93 ± 2.19	14.16 ± 2.02	15.05 ± 2.01	15.60 ± 2.01	< 0.001
MPV (fL)	10.81 ± 1.08	11.14 ± 1.27	11.49 ± 1.28	11.25 ± 1.21	0.077
PLT (109/L)	219.73 ± 50.67	219.33 ± 67.04	199.38 ± 61.77	197.30 ± 68.13	0.128
NT-proBNP (pg/mL)	20.50(12.05,28.00)	42.50(30.00,61.40)	144.60(77.80,209.50)	266.8(160.60,379.65)	< 0.001
Adverse pregnancy outcomes	0	8 (17.8%)	22 (44.7%)	55 (75.3%)	< 0.001

GA Gestational age, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, ALP alkaline phosphatase, LDH lactate dehydrogenase, ALT alanine aminotransferase, AST aspartate aminotransferase, BUN blood urea nitrogen Alanine, SCr serum creatinine, UA uric acid, WBC white blood cell, PLT platelet, MPV mean platelet volume, PDW platelet distribution width, NT-proBNP N-terminal pro-brain natriuretic peptide. The data were analyzed by one-way ANOVA and Kruskal–Wallis tests

Adverse pregnancy outcomes of HDP and its correlation with NT-proBNP, LDH, UA and PDW

As shown in Table 2, levels of NT-proBNP, LDH, UA and PDW were elevated in patients with adverse pregnancy outcomes (NT-proBNP: 251.00(140.85,366.25) vs 72.10(39.45,147.70) pg/mL; LDH: 261.24 ± 66.83 vs 207.80 ± 47.73 IU/L; UA: 407.04 ± 95.81 vs 339.03 ± 92.07 IU/L; PDW: 15.66 ± 1.82 vs 14.42 ± 2.16 fL). Moreover, preterm birth (17/73) was the most common complication observed in the sPE group, followed by intrauterine growth restriction (IUGR) (11/73), when compared to the pregnancy-induced GH and PE groups. Additionally, 3 cases of placental abruption were observed only in the sPE group [low birth weight (LBW) + Preterm birth; Preterm birth + Placental abruption]; Placental abruption + LBW), 7 cases in this group had multiple complications (preterm + LBW + IUGR). No death occurred in these groups (Table 3).

Table 2.

Correlation of NT-proBNP, LDH, UA, and PDW levels with adverse pregnancy outcomes

Variable	BMI (kg/m2)	GA	SBP (mmHg)	DBP (mmHg)	NT-proBNP (pg/mL)	LDH (IU/L)	PDW (fL)	UA (IU/L)
Normal pregnancy outcome (n = 81)	31.16 ± 4.21	38.49 ± 1.36	147.62 ± 13.37	98.26 ± 12.43	72.10(39.45, 147.70)	207.80 ± 47.73	14.42 ± 2.16	339.03 ± 92.07
Adverse pregnancy outcome (n = 84)	30.86 ± 4.45	35.44 ± 3.54	160.32 ± 18.15	106.40±11.58	251.00(144.85, 366.25)	261.24 ± 66.83	15.66 ± 1.82	407.04 ± 95.81
Z/t	0.443	7.26	−5.11	−4.357	−7.583	−5.88	−3.971	−4.646
P-value	0.658	< 0.001	<0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001

BMI body mass index, GA Gestational age, SBP systolic blood pressure, DBP diastolic blood pressure, NT-proBNP N-terminal pro-brain natriuretic peptide, LDH lactate dehydrogenase, PDW platelet distribution width, UA uric acid

P-values were analyzed by Student’s t-test and the Mann–Whitney test

Table 3.

Distribution of adverse pregnancy outcomes among different groups

	GH (n = 45)	PE (n = 47)	sPE (n = 73)	P-value
Fetal complications				< 0.01
LBW	0	0	0
Preterm birth	2	8	17
IUGR	4	5	11
Fetal distress	0	0	1
Birth asphyxia	0	0	0
Macrosomia	0	1	2
NICU admission	0	3	1
Maternal complications
Placental abruption	0	0	0
HELLP syndrome	1	0	1
Multiple complications
Preterm birth + IUGR/	0	0	1
Preterm birth + Placental abruption	0	0	1
LBW + Preterm birth	0	1	4
Fetal distress + NICU admission	0	0	1
Birth asphyxia + NICU admission	0	0	1
Preterm birth + Placental abruption	0	0	1
Placental abruption + LBW	0	0	1
LBW + NICU admission	0	0	1
LBW + NICU admission + IUGR	0	0	2
LBW + Preterm birth + IUGR	1	1	7

LBW low birth weight (Birth weight < 2.5 kg), IUGR intrauterine growth restriction, NICU neonatal intensive care unit, HELLP hemolysis elevated liver enzymes low platelet count

P-values were analyzed using the chi-square test

Multivariate logistic regression analysis and Nomogram for adverse outcomes

A multivariate stepwise logistic regression analysis was performed with adverse outcomes as the dependent variables and the above residual variables, including GA, SBP, DBP, NT-proBNP, LDH, UA, and PDW as the covariates. The results revealed that GA, NT-proBNP, LDH, and PDW were identified as independent factors associated with the occurrence of adverse outcomes of HDP (P < 0.001, P = 0.001, P < 0.001, P = 0.017, respectively). The predictive factors in the model: GA (0.546), NT-proBNP (1.005), LDH (1.015), and PDW (1.309), ranked by odds ratios (OR) (Table 4). Furthermore, we constructed a prediction nomogram incorporating the independent variables GA, NT-proBNP, LDH, and PDW, which directly demonstrated the risk of developing adverse pregnancy outcomes in patients with HDP (Fig. 1). The nomogram serves as a predictive model, each variable was assigned a score on the point scale axis, and the total score was calculated by summing the individual item scores, which corresponded to a predicted risk of adverse pregnancy outcomes in patients with HDP. Finally, we performed internal validation using 1,000 bootstrap samples to assess the model’s performance and stability. The optimism-corrected C-index (Dxy) of 0.8095 and an R² of 0.5919, indicating good discriminative ability (Table S1).

Table 4.

Multivariate logistic regression analysis of risk factors for adverse pregnancy outcomes in HDP Patients

variable	β Coefficient	Standard error	P-value	Odds ration	95% CI of OR
NT-proBNP	0.005	0.002	0.001	1.005	(1.002,1.009)
LDH	0.015	0.004	< 0.001	1.015	(1.007,1.023)
PDW	0.269	0.113	0.017	1.309	(1.049,1.633)
GA	−0.606	0.131	< 0.001	0.546	(0.422,0.706)

GA Gestational age, NT-proBNP N-terminal pro-brain natriuretic peptide, LDH lactate dehydrogenase, PDW platelet distribution width, 95% CI 95% Confidence Interval, OR Odds ratio

Fig. 1.

Nomogram prediction model for adverse pregnancy outcomes in patients with HDP. Each predictor is assigned a corresponding point value on the top “Points” axis. The scores from all variables are summed to generate a total score, which is then mapped onto the bottom “Risk” axis to estimate the probability of adverse pregnancy outcomes. Higher total scores indicate a greater predicted risk. For example, a pregnant woman with a gestational age of 40 weeks, NT-proBNP level of 400 pg/mL, LDH level of 250 IU/L, and PDW of 15% would receive approximately 10 points for GA, 20 points for NT-proBNP, 20 points for LDH, and 11 points for PDW, yielding a total score of 61. According to the nomogram, this total score corresponds to an estimated risk of approximately 50% for adverse pregnancy outcomes. NT-proBNP: N-terminal pro-brain natriuretic peptide. LDH: lactate dehydrogenase; PDW: platelet distribution width

Internal validation of the nomogram

The ROC curves were used to derive cutoffs for GA, NT-proBNP, LDH, and PDW. The area under the curve (AUC) for predicting adverse pregnancy outcomes was 0.807, 0.842, 0.761, and 0.684, respectively. The analysis also revealed that NT-proBNP can distinguish adverse pregnancy outcomes from healthy pregnant ones at a threshold of 204.70 pg/mL, with a sensitivity of 63.1% and a specificity of 90.1%. LDH demonstrated an AUC of 0.761 (95% CI: 0.688–0.834), with 59.5% sensitivity and 84.0% specificity. PDW showed the lowest AUC (0.684, 95% CI: 0.601–0.767), but exhibited high sensitivity (94.0%) and low specificity (42.0%) at a cut-off value of 13.65 fL.GA had an AUC of 0.807 (95% CI: 0.739–0.875), with relatively high specificity (93.8%) but lower sensitivity (58.3%) at a cut-off of 36.64 weeks. Furthermore, the combination of these indicators yielded a higher predictive value, with an AUC of 0.915, 84.5% sensitivity, and 86.4% specificity (Fig. 2A and Table 5). DeLong’s test demonstrated that the combined model (AUC = 0.915) significantly outperformed NT-proBNP (AUC = 0.842, P = 0.008), LDH (AUC = 0.761, P < 0.001), GA (AUC = 0.807, P < 0.001), and PDW (AUC = 0.684, P < 0.001) in discriminative performance (Table S2). The calibration curve showed good agreement between predicted and observed probabilities of adverse pregnancy outcomes (Fig. 2B). The decision curve analysis (DCA) curve demonstrated that the nomogram provided a significantly greater net clinical benefit than either the treat-all-patients or treat-none strategies across a wide range of threshold probabilities (Fig. 2C). Furthermore, the clinical impact curve showed strong agreement between the predicted high-risk population and the actual high-risk population when the risk probability was ≥ 80% (Fig. 2D).

Fig. 2.

Performance and clinical utility of the nomogram model. A ROC curves showing the predictive value of GA, serum NT-proBNP, LDH, and PDW for adverse pregnancy outcomes in patients with HDP (GA: AUC = 0.796, NT-proBNP: AUC = 0.842, LDH: AUC = 0.761, PDW: AUC = 0.684, and combined indicators: AUC = 0.915). B Calibration curve showing the agreement between the predicted probabilities from the model and the observed probabilities. The ideal calibration line means an intercept of 0 and a slope of 1 for the calibration plot. C Decision curve analysis (DCA) of the nomogram model, compared with the “treat-all” and “treat-none” strategies. The curve illustrates the net benefit across a range of threshold probabilities, providing insight into the clinical utility of the model. D Clinical impact curve showing the net benefit, the number of individuals classified as high-risk by the nomogram, and the actual number of adverse pregnancy outcomes across different high-risk thresholds. GA: Gestational age; NT-proBNP: N-terminal pro-brain natriuretic peptide. LDH: lactate dehydrogenase; PDW: platelet distribution width

Table 5.

ROC curve analysis of individual biomarkers and the combined model

Parameters	Area	Cut off	Sensitivity	Specificity	95%CI
NT-proBNP	0.842	204.70	63.1	90.1	0.782–0.902
LDH	0.761	242.40	59.5	84.0	0.688–0.834
PDW	0.684	13.65	94.0	42.0	0.601–0.767
GA	0.807	36.64	58.3	93.8	0.739–0.875
Combined	0.915		84.5	86.4	0.871–0.959

NT-proBNP N-terminal pro-brain natriuretic peptide, LDH lactate dehydrogenase, PDW platelet distribution width, GA Gestational age, 95% CI 95% Confidence Interval

Discussion

This study evaluated peripheral blood variables of serum NT-proBNP with other clinical parameters levels in pregnant women with HDP, and their predictive values for adverse pregnancy outcomes. HDP is a multisystem disorder characterized by liver and renal function injuries attributed to vascular endothelial cell ischaemia and hypoxia, leading to elevated ALT, AST, SCr, and BUN [15, 16]. Consistently, our study yielded similar conclusions, but these values were within the normal reference range. Recent studies have recommended routine testing of UA as a prognostic marker, suggesting that UA levels rise with the risk of maternal and fetal complications [17, 18]. Despite renal function variables such as UA being elevated with the severity of HDP, our study suggested that it cannot independently predict adverse pregnancy outcomes in women with HDP. NT-proBNP is the active precursor of polypeptides, mainly secreted by cardiomyocytes and commonly used to monitor cardiac function in patients. Previous studies have shown that NT-proBNP is positively associated with disease severity in patients with HDP [9, 19]. Consistent with these findings, the present study demonstrated that NT-proBNP levels increase with the severity of HDP, more likely related to the incidence of patients with sPE and PE with adverse pregnancy outcomes. Evidence suggests that the heart load is elevated with the severity of hypertension, resulting in the serum NT-proBNP secretion by ventricular myocytes [20, 21]. However, the NT-proBNP levels between GH and healthy pregnancy groups were within the normal range [42.50 (30.00, 61.40) vs 20.50 (12.05, 28.00) pg/mL], limiting its diagnostic value in the early onset of pregnancy, similar to results obtained by Seong and Meera Kumari [22]. Moreover, a study by Zhang et al. [23] also showed serum NT-proBNP is an effective indicator for predicting newborn outcomes (AUC:0.788; cut-off:257.5 pg/mL; sensitivity:55.8%; specificity:87.8%). Similar to these findings, our results demonstrated that NT-proBNP levels were higher in adverse outcomes than healthy pregnancy outcomes in women with HDP, which can serve as a predictive biomarker for pregnancy complications with moderate sensitivity and good specificity (AUC = 0.842; cut-off:204.70; sensitivity: 63.1%; specificity:90.1%). Since the pathological mechanism of preeclampsia is thought to be associated with hypoxia and endothelial cell injury, and LDH gene is a well-characterized hypoxia-inducible gene reflecting the severity of inflammation [24, 25]. Demir et al. [26] showed the LDH level was significantly elevated in sPE and closely associated with poor complications. Dewi et al. [27] found that the placental LDH level was higher in the early stage of PE than that in normal-term placenta. Kozic et al. [28] displayed that LDH levels were useful in the diagnosis of the severity of HDP, modestly predicting for maternal and fetal complications(AUC: 0.74). Our results also suggested that LDH level was significantly correlated with the degree of HDP, especially elevated in patients with sPE and adverse outcomes during pregnancy, and they modestly predicted the adverse maternal outcomes in women with HDP (AUC: 0.761; cut-off:242.4; sensitivity:59.5%; specificity:84%). In addition, previous reports have also shown that PDW levels are significantly higher in PE compared with healthy control groups. Contact of platelets with the injured endothelium enhances thrombopoiesis, producing younger platelets, which are larger than older platelets, resulting in variable sizes of platelet parameters, including MPV and PDW [29, 30]. Our results revealed that the PDW levels in sPE were higher than those in other groups, and that was an independent risk factor for adverse outcomes in patients with HDP, suggesting that the platelet function variables should be tested throughout pregnancy. In contrast, many other findings have suggested that PLT parameters have little prediction value for the severity of PE, with a ROC of 0.62, sensitivity of 72% and a specificity of 52% [31]. Ceyhan et al. [32] claimed that PDW was not reliable to predict preeclampsia, because of differences in the methods and equipment. Freitas et al. [33] showed that PLT with a sensitivity of 68.97% and a specificity of 70.69% for PE. Consistent with these studies, although our study indicated that PDW level was higher in PE compared to pregnant patients with GH and healthy controls, yet was within the laboratory reference range, causing limited diagnostic value in predicting adverse outcomes (AUC: 0.684; cut-off:13.65; sensitivity:94%; specificity:42%). Wu et al. [34] have noted that the AUC of combined detection of serum TMAO, NT-proBNP and HIF-1a on prediction for pregnancy outcomes of patients is 0.91 (cut-off:4.198; sensitivity:53%; specificity:83%). Nonetheless, serum TMAO and HIF-1a are non-traditional biomarkers, and the sensitivity of combined biomarkers was poor, which restrict their application in the clinical laboratory. A longitudinal cohort study indicated that the addition of GA, NT-proBNP assessment improved the short-term prediction of delivery compared with soluble fms-like tyrosine kinase-1/Placenta growth factor (sFlt-1/PIGF) ratio alone (AUC = 0.845; 95%CI:0.787–0.896) [35]. On the other hand, Nan et al. [19] also reported that NT-proBNP demonstrated a predictive performance comparable to that of the sFlt-1/PlGF ratio for identifying PE with complications within one week of assessment, with AUCs of 0.818 (95% CI: 0.706–0.930) and 0.822 (95% CI: 0.723–0.921), respectively. Moreover, the combination of NT-proBNP and the sFlt-1/PlGF ratio showed excellent predictive accuracy for early-onset PE within one week, with AUCs of 0.971 (95% CI: 0.942–1.000) for NT-proBNP and 0.970 (95% CI: 0.932–1.000) for the sFlt-1/PlGF ratio. Zheng et al. [36] study reported that gestational age, 24-h urine protein (qualitative), and thrombin time (TT) were significantly associated with adverse pregnancy outcomes (P < 0.05), with an AUC of 0.781 (95% CI: 0.737–0.825), indicating moderate predictive performance. While PlGF and sFlt-1 have shown good predictive performance, their high cost and limited availability restrict their use in many primary care settings. However, our results indicated the combination of GA, NT-proBNP, LDH as well and PDW may yield higher predictive values for adverse pregnancy outcomes in HDP (AUC:0.915, sensitivity:84.5%, specificity:86.4%). Through our prediction model based on blood tests of NT-GA, NT-proBNP, LDH and PDW, doctors can early recognize the severity of HDP in clinical work, which will be helpful for managing the risk of adverse outcomes in patients with HDP. In addition, the nomogram and regression analysis further highlighted that NT-proBNP, LDH and PDW levels were risk indicators of poor prognosis in patients with HDP respectively, which were elevated in patients with adverse outcomes compared with those with normal pregnancy outcomes. Given that NT-proBNP, LDH, and PDW values were within the normal range and changed dynamically in the early pregnancy stage. Meanwhile, patients with sPE and PE rather than GH status demonstrated a percentage of complications such as preterm and LBW. Thus a close monitoring remains necessary, thereby preventing disease progression to overt sPE and adverse outcomes. To our knowledge, this is the first nomogram to assess the performance of the combination of these markers (GA, NT-proBNP, LDH, PDW) in the prediction of the need for adverse pregnancy outcomes in HDP. Moreover, NT-proBNP, LDH and PDW parameters are simple laboratory markers and easy to check during antenatal care; dynamic monitoring of multiple indicators can guide early interventions before progress to PE, particularly in high-risk patients exhibiting complications like preterm birth or low birth weight. Furthermore, we assessed the model's performance through internal validation, calibration curves, and clinical decision curve analysis, all of which demonstrated its good predictive efficacy and clinical utility. However, several limitations to our study should be considered. First, this was a single-center study with a relatively small sample size, which may limit the generalizability of the findings. Second, the absence of external validation is a major limitation of this study. Although internal validation was performed using bootstrap resampling, external validation with independent cohorts is essential to assess the generalizability and robustness of the model across different populations and clinical settings. Future studies are warranted to externally validate to determine if these biomarkers can improve maternal or neonatal outcomes. Finally, it is important to recognize that, being a retrospective study, several potential biases, including selection bias, recall bias, information bias, confounding bias, and reporting bias, may have influenced the validity of our results and findings. Therefore, by carefully designing the study, implementing clear inclusion/exclusion criteria, controlling for confounding variables, and employing appropriate statistical methods, we can minimize these biases. The following are the potential biases and solutions.

• Selection bias: mitigated by using strict inclusion and exclusion criteria.

• Recall bias: mitigated by using objective measures such as clinical records or laboratory results.

• Information bias: mitigated by relying on standardized diagnostic tools and medical records.

• Confounding bias: mitigated for through multivariate analysis.

• Reporting bias: mitigated by ensuring transparent reporting of all outcomes.

• Time bias: mitigated by standardize data collection within a specific time period or ensuring that time-related variables are considered in the analysis.

Conclusion

In conclusion, the expressions of GA, serum NT-proBNP, LDH, and PDW are associated with the severity of HDP, and the combination of these biomarkers enhances the efficacy of predicting adverse pregnancy outcomes. Notably, fetal complications such as low birth weight (LBW) and preterm birth, as well as maternal complications such as placental abruption, were observed at higher rates in the sPE and PE groups compared to the GH group. Moreover, we established a new nomogram predictive model of adverse pregnancy outcomes in patients with HDP, which demonstrated more accuracy in stratifying maternal and infant outcomes with HDP.

Abbreviations

HDP

Hypertensive disorders of pregnancy

PE

Preeclampsia

sPE

Severe preeclampsia

GH

Gestational hypertension

BMI

Body mass index

DBP

Diastolic blood pressure

SBP

Systolic blood pressure

NT-proBNP

N-terminal pro-brain natriuretic peptide

LDH

Lactic dehydrogenase

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

UA

Uric acid

BUN

Blood urea nitrogen Alanine

SCr

Serum creatinine

ALP

Alkaline phosphatase

WBC

White blood cell

PLT

Platelet

MPV

Mean platelet volume

PDW

Platelet distribution width

LBW

Low birth weight

IUGR

Intrauterine growth restriction

NICU

Neonatal intensive care unit

ROC

Receiver operating characteristic

Authors’ contributions

Yingdi Jiang and Lu Dai concepted and designed the study, Yingdi Jiang and Yanjie Zhou collected and analyzed data. Lu Dai wrote the draft of this manuscript. Yanjie Zhou and Jiang Lin edited, revised the manuscript. All authors contributed to approve the final manuscript.

Funding

This research was supported by the Development Fund of Affiliated Hospital of Xuzhou Medical University (No. XYFY202350).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Our study was approved by the Academic Ethics Committee of the Affiliated Jiangyin Hospital of Xuzhou Medical University [Approval number: 2024ER (076)], due to the retrospective nature of this study, informed patient consent was waived by the Ethics Committee of the Affiliated Jiangyin Hospital of Xuzhou Medical University. All methods were performed in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.