The exposure was hypertensive disease during pregnancy or within 42 days post-delivery. The hypertensive diseases included preeclampsia, gestational hypertension, chronic hypertension, and chronic hypertension with superimposed preeclampsia. Preeclampsia was defined as persistent de novo hypertension that develops at or after 20 weeks of gestation, accompanied by either proteinuria or features of maternal organ or uteroplacental dysfunction. Features of maternal organ dysfunction include acute kidney injury (creatinine ⩾90 μmol/L or 1 mg/dL), liver involvement (elevated alanine aminotransferase or aspartate aminotransferase >40 IU/L) with or without right upper quadrant or epigastric abdominal pain, neurological complications (such as eclampsia, altered mental status, blindness, stroke, clonus, severe headaches, and persistent visual scotomata), and haematological complications (decreased platelet count <150,000/μL, disseminated intravascular coagulation, haemolysis). Uteroplacental dysfunction includes fetal growth restriction, abnormal umbilical artery Doppler waveform analysis, or stillbirth (16, 17). Gestational hypertension was defined as persistent de novo hypertension that develops at or after 20 weeks of gestation in the absence of features of preeclampsia. Chronic hypertension referred to high blood pressure predating the pregnancy or recognized at < 20 weeks of gestation. Chronic hypertension with superimposed preeclampsia was diagnosed when preeclampsia occurred in a pregnant woman who had pre-existing chronic hypertension.

We included randomized and non-randomized control trials, prospective and retrospective cohort studies, case–control studies, and cross-sectional studies performed in humans. The studies were included based on the following criteria:

There were no restrictions by the date of publication.

We excluded case series and reports, cost–benefit analyses, and qualitative research, as well as reviews, newspapers, books, conference abstracts, theses, commentaries, letters, editorials, and unpublished data. Animal and in vitro studies were also excluded. Studies that only assess arterial stiffness measurements during pregnancy, labour, and delivery or within 6 weeks post-partum were excluded.

The control group consisted of women under 60 years old with no history of HDP, cardiovascular diseases, renal failure, or diabetes.

Mendeley Reference Manager was used to store and manage searched items under the following headings in a separate table: Authors, publication year, title of the study, location of the study, participants ages, country, and main study findings. Two independent reviewers screened all the titles and abstracts using the predetermined inclusion and exclusion criteria. We followed PRISMA-P guidelines. Records identified as potentially eligible based on title, abstract, and full texts were obtained to screen. Where consensus on eligibility could not be achieved, a third review author was involved in the discussion. When the same cohort was reported in multiple articles, the study which contains the largest sample was selected. Data from selected studies were extracted and entered into a Microsoft Excel spreadsheet.

Two independent authors evaluated the methodological quality of all included studies. Where consensus on eligibility could not be achieved, a third reviewer was involved in the discussion. The Newcastle Ottawa Scale (NOS) and Grading of Recommendations Assessment, Development, and Evaluation (GRADE) were used to assess risk of bias and quality of evidence (18). Disagreements among the two authors was resolved by consultation with the other authors to reach a consensus. No studies were excluded based on the risk of bias assessment.

Potential sources of heterogeneity were first assessed using the I² statistic (19). Values of 25% or less were classified as low, around 50% as moderate, and 75% or more as high (20). The subgroup and sensitivity analyses using RevMan 5.4 and NOS assessment scale were performed. Since all subgroups contained fewer than ten studies, a funnel plot was only created for those with six or more studies.

Estimates of mean and ± SD for arterial stiffness indices (cfPWV, AIx, AIx@75) for women who had HDPs (for each subtype of HDP, as available) and normotensive women was obtained from the relevant studies. For studies where the estimates were reported as the median and interquartile range, approximate estimates of mean and ±SD were calculated using the available estimates of the median and, first quartile and third quartile. These data were summarized using a DerSimonian and Laird random-effects model (21). A separate meta-analysis was conducted for each hemodynamic indices of interest (cfPWV, AIx, AIx@75) for the combined outcome in those who had HDPs. For studies that provided data on the association between these hemodynamic variables and the outcome of interest, we meta-analyzed these effect measures using a random effects model. We performed a subgroup analysis for each subtype of HDP, as available.

Through our literature search, we identified 121 articles. After removing 22 duplicates and excluding 70 based on title and abstract screening (see Figure 1), Twenty-nine articles remained that were eligible for full-text review. Out of these, 12 studies were included in the final review. Specifically, there were 4 cross-sectional studies, 5 cohort studies, and 3 case–control studies (Table 2). The total number of participants across the studies was 856 women with HDP and healthy women.

The included studies, conducted between 2009 and 2023, came from various countries: South Africa (1), Argentina (1), Denmark (1), Portugal (1), United States (3), Austria (1), Canada (1), the United Kingdom (1), Norway (1) and Italy (1) (see Table 2). These studies evaluated arterial stiffness, with some utilizing carotid-femoral pulse wave velocity (cfPWV) and/or augmentation index (AIx) and/or Augmentation index adjusted for heart rate (AIx@75).

HDP encompasses various types (16). While only two studies, Ehrenthal et al. (22) and Moe et al. (23), examined composite HDP, other studies focused on specific types, including preeclampsia, early preeclampsia, and late-onset preeclampsia. As a result, the findings were categorized into four groups: composite HDP, preeclampsia, early preeclampsia, and late-onset preeclampsia.

The risk of bias was assessed using the NOS for cohort and case–control studies, and a modified NOS for cross-sectional studies. The assessment indicated that all studies had a low risk of bias (Tables 3, 4). Recognizing that NOS evaluates risk of bias but does not fully capture the overall certainty of evidence, the GRADE framework was further applied. This comprehensive assessment, which considers factors such as consistency, directness, and precision, rated the overall certainty of the evidence as moderate to high (Table 5).

A total of 7 studies assessed the AIx in patients with all various HDP compared to normotensive women (Figure 2A). The pooled data demonstrated a significantly elevated AIx among women with HDP, indicating a Mean difference (MD) of 11.63 and a 95% confidence interval (CI) of [1.72–21.54]. This analysis included 345 participants, comprising 167 women with HDP and 178 normotensive controls. The heterogeneity among the studies was high, with an I² value of 98%. Additionally, the analysis of funnel plots indicated the presence of publication bias, with a p value of 0.045 in the Egger’s test (Figure 3).

Furthermore, a total of 11 studies examined cfPWV in women with all various HDP compared to normotensive women (Figure 2B). The pooled data showed a significantly elevated cfPWV in women with HDP, with a MD of 0.53 and a 95% CI of [0.27–0.78]. This analysis included 738 participants, comprising 363 women with HDP and 375 normotensive controls. The studies showed moderate heterogeneity, with an I² value of 64%. Additionally, the analysis of funnel plots indicated the presence of publication bias, with a p value of 0.002 in the Egger’s test (Figure 4).

Four studies investigated AIx in patients with HDP compared to normotensive controls within 1 year after delivery (see Figure 5A). The results showed no significant increase in AIx for women with HDP, reporting a MD of 14.85 and a 95% CI of [−6.03–35.72]. Additionally, there was high heterogeneity among these studies, with an I² value of 97%.

In contrast, other studies focused on AIx in patients with HDP compared to normotensive controls more than 1 year after delivery (see Figure 5B). The findings indicated a significant increase in AIx for women with HDP, with a MD of 9.11 and a 95% CI of [4.20–14.02]. The heterogeneity in these studies was also high, with an I² value of 94%.

Six studies evaluated cfPWV in women with HDP compared to normotensive controls within 1 year after delivery (Figure 6A). The results demonstrated a significantly higher cfPWV for women with HDP, reporting an MD of 0.59 and a 95% CI of [0.32–0.86]. The heterogeneity among these studies was moderate, with an I² value of 41%.

Additionally, five studies focused on cfPWV in women with HDP compared to normotensive controls more than 1 year after delivery (Figure 6B). The findings also revealed a significant increase in cfPWV for women with HDP, with an MD of 0.45 and a 95% CI of [0.03–0.88]. Heterogeneity in these studies was also moderate, with an I² value of 67%.

A total of 4 studies assessed the AIx in patients with preeclampsia compared to normotensive women (Figure 7A). This pooled data demonstrated a significantly elevated AIx among women with preeclampsia, with a mean difference (MD) of 8.57 and a 95% confidence interval (CI) of [4.22–12.92]. The analysis included 207 participants, comprising 102 women with preeclampsia and 105 normotensive controls. The heterogeneity among the studies was high at 91%, indicating varying results across the studies. In two studies (12, 14) that adjusted the AIx for heart rate (AIx@75), the augmentation index remained significantly elevated after the adjustment, with a MD of 7.50 and a 95% CI of [6.09–8.90] (Figures 7A,B). These two studies showed no heterogeneity.

Additionally, an examination of cfPWV across eight studies indicated a significant increase in cfPWV among women with preeclampsia, demonstrating an MD of 0.47 with a 95% CI of [0.20–0.74] (Figure 7C). This analysis involved 600 patients (298 preeclamptic and 302 normotensive). The studies displayed moderate heterogeneity, with an I² value of 62%.

The funnel plot was generally symmetrical; however, a few studies deviated from the main axis, indicating the possibility of a small sample effect or publication bias. To further investigate potential bias, Egger’s test was conducted, which showed no significant publication bias (p = 0.144) (Figure 8).

One study focused on AIx in patients with composite HDP against normotensive controls (22). The results indicated no significant increase in AIx for women with HDP (MD 5.40, 95% CI [−0.66–11.46]). The heterogeneity assessment was not applicable due to the design being based on a single study. This analysis included a total of 74 participants, with 33 diagnosed with HDP and 41 who were normotensive. However, AIx was found to be significantly elevated after adjusting for heart rate (AIx@75) (MD 6.40, 95% CI [0.52–12.28]) (15).

Two studies, Ehrenthal et al. (22) and Moe et al. (23) analysed cfPWV in the context of HDP, reporting no significant increase in cfPWV among affected women (MD 0.35, 95% CI [−0.13–0.83]), with zero statistical heterogeneity noted (Figure 9). This part of the analysis comprised 305 individuals (169 with HDP and an equivalent number of normotensive controls).

In the context of early onset preeclampsia (EOP), two studies, conducted by Christensen et al. (26) and Franz et al. (24), evaluated the AIx and found no significant increase associated with EOP (MD 1.55, 95% CI [−0.74–3.84]), with a heterogeneity of 88% (Figure 10A). The studies included a total of 64 participants, comprising 32 EOP and 32 normotensive women. Orabona et al. (25) was the only study to evaluate AIx after adjusting for heart rate, and it indicated a significant elevation in AIx following this adjustment (MD 0.94, 95% CI [2.17–3.70]).

Additionally, Christensen et al. (26), Franz et al. (24), and Orabona et al. (25) also assessed cfPWV in patients with EOP. The results indicated a significant increase in cfPWV (MD 1.86, 95% CI [0.25–3.47]) and a heterogeneity of 92%. A total of 124 participants were included in these studies, consisting of 62 EOP and 62 normotensive women (Figure 10B).

Christensen et al. (26) and Franz et al. (24) also studied the AIx in cases of late-onset preeclampsia. Their analysis showed no significant difference (MD of 2.44, 95% confidence interval [−8.82–13.70]) and indicated a heterogeneity of 65% (Figure 11A). This study included a total of 64 participants, comprising 32 individuals with LOP and 32 normotensive controls. However, after adjusting for heart rate, Orabona et al. (25) found that the augmentation index was significantly higher in the LOP group (MD of 4.80, 95% confidence interval [0.60–9.00]).

Furthermore, Christensen et al. (26), Franz et al. (24), and Orabona et al. (25) also assessed cfPWV in patients with LOP. Their findings indicated no significant increase in cfPWV (MD 0.10, 95% CI [−0.42–0.62]), with a heterogeneity of 56% (Figure 11B). The total sample size in this analysis was 124 participants, including 62 with LOP and 62 normotensive individuals.

A total of three studies assessed the AIx in patients with HDP compared to normotensive women (22, 24, 26) (see Figure 12A). The pooled data demonstrated a significantly elevated AIx among women with HDP, indicating a MD of 16.82 and a 95% CI of [1.11–12.92]. This analysis included 138 participants, comprising 65 women with HDP and 73 normotensive controls. The heterogeneity among the studies was high at 93%, suggesting varying results across the studies. In the only study that adjusted the AIx for heart rate (AIx@75), the augmentation index remained significantly elevated after adjustment, with a MD of 6.40 and a 95% CI of [0.52–12.28] (22).

Furthermore, an examination of cfPWV across five studies showed no significant increase in cfPWV among women with HDP, demonstrating a MD of 0.70 with a 95% CI of [−0.10–1.49] (see Figure 12B). The heterogeneity was again high at 93%, and this analysis involved 243 patients (105 with HDP and 138 normotensive).

Two studies focused on AIx in patients with HDP compared to normotensive controls (12, 27) (see Figure 13A). The results indicated a significant increase in AIx for women with HDP, with a MD of 6.85 and a 95% CI of [5.31–8.39]. The heterogeneity among these studies was low at 40%. This analysis included a total of 137 participants, with 67 diagnosed with HDP and 70 who were normotensive. In one study, the AIx remained significantly elevated after adjusting for heart rate, with a MD of 7.50 and a 95% CI of [0.68–8.92] (12).

Four studies analysed cfPWV in the context of HDP, reporting a significant increase in cfPWV among affected women, with a MD of 0.59 and a 95% CI of [0.35–0.82] (12, 23, 27, 28). No statistical heterogeneity was observed in this analysis (see Figure 13B). This part of the analysis comprised 305 individuals (223 with HDP and 203 normotensive controls).

This review had a limited number of studies, which could introduce bias. Additionally, a high level of heterogeneity was observed in most group analyses. However, assessments using the NOS revealed a low risk of bias in all studies, with the GRADE framework rating the evidence as moderate to high certainty. Furthermore, the review adhered to the PRISMA-P guidelines, and a comprehensive search was performed across several databases.