A comparative systematic review and meta-analysis of uterine artery resistance in pregnant women with and without previous history of cesarean section

Cesarean section rates have been increasing globally, resulting in a rising number of women with uterine scars from previous cesarean deliveries [ 1 ]. The necessity of assessing complications and side effects of cesarean sections is crucial due to the significant impact they can have on maternal and fetal health. Cesarean sections can lead to several short-term complications, such as infection, hemorrhage, and anesthesia-related issues, as well as long-term complications affecting subsequent pregnancies, including uterine rupture, abnormal placentation, and increased risk of preterm birth [ 2 – 4 ]. These complications not only pose risks to maternal health but can also result in adverse neonatal outcomes such as low birth weight and respiratory distress [ 5 , 6 ]. However, the literature presents conflicting reports on this association. Some studies have found a positive correlation, indicating increased uterine artery resistance in women with previous cesarean sections [ 7 , 8 ], even if the differences were not statistically significant [ 9 ]. Conversely, other studies have found no significant differences or negative correlations [ 10 , 11 ]. Several theories have been proposed to explain this association. First, it is hypothesized that the uterine scar tissue resulting from a Cesarean section may lead to impaired vascular remodeling. This process is crucial during early pregnancy to ensure adequate blood flow to the placenta [ 12 , 13 ]. The scar may cause fibrosis and localized ischemia, reducing the elasticity of the uterine wall and increasing vascular resistance [ 14 ]. Another explanation proposed using animal studies is poor trophoblastic invasion, a process essential for establishing a healthy placental blood supply, which may be disrupted by cesarean-induced uterine scarring [ 15 , 16 ]. Based on the existing literature, we hypothesized that uterine artery resistance—particularly as measured by the Pulsatility Index (PI)—would be significantly higher in pregnant women with a history of cesarean section compared to those without. This hypothesis stems from studies suggesting impaired vascular remodellingand altered uteroplacental perfusion in scarred uteri. Clinically, identifying such associations could inform obstetric monitoring strategies, especially in high-risk pregnancies.

A total of 958 records were identified through database searches, and one additional record was identified through other sources. After removing duplicates, 441 records were screened by title and abstract. Of these, 67 full-text articles were assessed for eligibility, and 61 were excluded with reasons. Ultimately, six studies met the inclusion criteria and were included in the final analysis. One of the included studies was conducted by our research team. Although we were aware of its findings during the review process, we followed a rigorous, predefined eligibility protocol and objective data extraction procedures to minimize bias. The six included studies were published between 2011 and 2023 and originated from Turkey (n = 3), Iran (n = 1), Thailand (n = 1), and Norway (n = 1). All studies employed observational designs—either retrospective or prospective cohort studies. Collectively, they encompassed a total of 1,656 pregnant women, with individual study sample sizes ranging from 64 to 503. Uterine artery resistance was assessed using Doppler ultrasound, and outcomes were reported as either PI, RI, or both. All studies used standard gestational windows (11–32 weeks) for Doppler measurements. ( Fig 1 , Table 1 ). The PRISMA flowchart illustrates the stringent selection process employed to ensure the inclusion of high-quality studies in our analysis. All six studies included in the meta-analysis were assessed for quality using an NOS checklist and were rated as “Good” based on their selection, comparability, and outcome criteria. ( Table 2 ). The certainty of evidence for each outcome was assessed using the GRADE approach, both manually and through the GRADEpro GDT software ( https://gradepro.org ). For the association between prior cesarean section and increased uterine artery pulsatility index (PI), GRADEpro initially rated the certainty as low, consistent with default ratings for observational studies ( S2 Table ). However, based on consistent effect direction across studies, low heterogeneity (I² = 26.6%), large pooled sample size (n = 1,656), and low risk of bias, we manually upgraded the certainty to moderate. Although the pooled effect estimate (Hedges’s g = 0.15; 95% CI: 0.03 to 0.26) excluded the null, its small magnitude and use of log-transformed MoM values warranted cautious interpretation ( Table 3 ). In contrast, the certainty of evidence for the resistance index (RI) was rated as very low in both manual and GRADEpro assessments, due to reliance on only two studies (n = 656), wide and non-significant confidence intervals (Hedges’s g = 0.19; 95% CI: –0.06 to 0.43), possible inconsistency, and an inability to assess publication bias. Gestational age was presented as Weeks (Day) Studies scoring 7–9 stars were rated as good quality, 4–6 as fair quality, and 0–3 as poor quality. PI: Pulsatility Index, RI: Resistance Index. The meta-analysis included six studies examining the PI in 1656 pregnant women with and without a history of cesarean section. Using a random-effects model (DerSimonian–Laird method), the pooled effect size showed a statistically significant increase in PI for women with a history of cesarean section (Hedges’s g = 0.15, 95% CI [0.03, 0.26], p = 0.01). The analysis revealed low heterogeneity among the studies (I² = 26.60%, p = 0.23), indicating that the variability in effect sizes was primarily due to sampling error rather than true differences between the studies. The test for overall effect was significant (z = 2.45, p = 0.01), suggesting that the history of cesarean section is associated with an increased PI ( Fig 2 ). Assessment for publication bias using Egger’s regression test indicated no significant small-study effects (p = 0.81), supporting the robustness of the results. The funnel plot did not show significant asymmetry, consistent with Egger’s test results, further confirming the absence of substantial publication bias in the main analysis ( Fig 3 ) and additional sensitivity analysis ( S1 Fig ). The studies were evenly distributed around the mean effect size, reinforcing the reliability of the meta-analysis findings. Meta-regression was performed to explore the influence of potential confounders on the observed effect sizes, specifically considering age, BMI, and weeks of gestation. The meta-regression results indicated None of these factors showed a statistically significant impact on the effect size (p > 0.05), indicating that the association between cesarean section history and increased PI was not significantly influenced by these confounders. The residual heterogeneity was significant (Q_res = chi2(5) = 113.96, p < 0.0001), suggesting that other unmeasured factors may contribute to the variability in PI outcomes ( S3 Table ). The sensitivity analysis for the PI indicates a statistically significant increase in PI for women with a history of cesarean section (Hedges’s g = 0.18, 95% CI [0.08, 0.29], p = 0.00), with no observed heterogeneity among the studies (I² = 0.00%) ( S1 Fig ). The meta-analysis of the RI included two studies, with a total sample size of 326 participants in the treatment group (history of cesarean section) and 330 participants in the control group (no history of cesarean section). The pooled effect size, calculated using a random-effects model, showed a small increase in RI for women with a history of cesarean section (Hedges’s g = 0.19, 95% CI [−0.06, 0.43], p = 0.13). The heterogeneity among the studies was moderate (I² = 48.08%, p = 0.17), indicating some variability in the results across studies. The test of overall effect (z = 1.50, p = 0.13) suggests that the increase in RI was not statistically significant ( Fig 4 ). For the RI outcome, no definitive interpretation could be made regarding publication bias due to the small number of studies (n = 2).

This meta-analysis suggests that women with a history of cesarean section may exhibit slightly increased uterine artery resistance, as measured by Doppler indices. While the observed association is modest, it may have implications for uteroplacental perfusion and pregnancy outcomes. These findings highlight the potential value of closer monitoring in pregnancies following cesarean delivery, although further research is needed to clarify the clinical significance and guide management strategies.

We formulated the study question using the PECO framework [ 17 ] (Population, Exposure, Comparator, Outcome) as follows: Is a history of cesarean section associated with increased uterine artery resistance—measured by PI or RI—in subsequent pregnancies compared to women without such a history? Where Population (P): Pregnant women undergoing second-trimester Doppler ultrasound assessment; Exposure (E): History of previous cesarean section; Comparator (C): Women with no prior cesarean section; Outcome (O): Uterine artery resistance indices, including Pulsatility Index (PI) and Resistance Index (RI), assessed using color Doppler ultrasonography. This systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards [ 18 ] and registered in Prospective Register of Systematic Reviews (PROSPERO) (CRD42024524343). We included all peer-reviewed original observational and interventional studies conducted in humans, as well as relevant academic theses (both published and unpublished) and full-text conference papers, provided they reported on uterine artery resistance measured by color Doppler ultrasonography in healthy pregnant women with and without a history of cesarean section. Eligible studies could be in any language, provided sufficient methodological and outcome data were accessible. Studies were excluded if they lacked a comparison between women with and without prior cesarean section, did not report uterine artery resistance indices (PI or RI), provided insufficient data for analysis, or were non-original publications such as reviews, case reports, or abstracts without full texts. The databases PubMed, Scopus, Web of Science, and Embase were systematically searched for studies up to 6 th April 2024. The authors concurred that Cesarean, Caesarean, “abdominal delivery”, Doppler, “Uterine artery” and UtA be in the search query. Literature search was conducted in the title and abstract by mentioned search terms. The search strategy was tailored for each database to capture the most relevant studies (See S1 Table ). Two independent authors (Az.M and B.N) reviewed the title, abstract, and full text of retrieved studies after identifying and deleting duplicates by EndNote software version X8(Clarivate Analytics, Philadelphia, PA, USA). A third researcher (Ar.M) was consulted to resolve the conflict. Studies were included in the analysis if they reported Color Doppler ultrasonography results for uterine artery resistance in both patients with and without a history of cesarean section, either in the abstract or full text. To access non-English or unavailable studies, we email authors and journals [ 19 – 24 ]. All selected articles were assessed for the ROB using the Newcastle-Ottawa (NOS) checklist for observational studies [ 25 ]. Study quality was categorized into Good quality, Fair quality, and Poor quality based on the NOS checklist instruction. This tool evaluates three domains: A: Selection (up to 4 stars): Representativeness of the exposed cohort, selection of the non-exposed cohort, ascertainment of exposure, and confirmation that the outcome was not present at study start. B: Comparability (up to 2 stars): Assessment of whether the study controlled for confounding variables. C: Outcome (up to 3 stars): Assessment of outcome, adequacy of follow-up, and length of follow-up sufficient for outcomes to occur. Two reviewers (Ar.M and Az.M) independently assigned stars to each item using a standardized NOS form. Disagreements were resolved through discussion or consultation with a third reviewer (Sh.Ch). Based on the total score out of 9 stars, studies were categorized as:Good quality (7–9 stars), Fair quality (4–6 stars), Poor quality (0–3 stars). All included studies scored 7 or higher and were rated as “Good” quality. The certainty of evidence for each outcome was assessed using the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) approach [ 26 ] b (Ar M). He performed both a manual GRADE assessment based on GRADE guidelines and a software-supported evaluation using GRADEpro GDT ( https://gradepro.org ). The manual assessment was used to provide contextual judgments for domains such as imprecision and inconsistency, particularly in interpreting clinical relevance and methodological consistency across studies. The PI and RI (Resistance Index) were considered the study outcomes and extracted from the papers. PI is a Doppler ultrasound measurement that reflects the pulsatility of blood flow, calculated as: PI = (Peak Systolic Velocity − End Diastolic Velocity)/ Mean Velocity. RI, also known as the Pourcelot index, is calculated as: RI = (Peak Systolic Velocity − End Diastolic Velocity)/ Peak Systolic Velocity [ 27 ] The author’s name, publication years, age, Body Mass Index (BMI), Doppler week, PI, and RI were other variables extracted from the eligible studies and imported into the Excel database by two authors (B.N and E.S) independently and the discrepancy was dissolved by (Ar.M). Meta-analysis was conducted using STATA version 17 (StataCorp LLC, College Station, TX, USA) to synthesize data from the included studies. Each study’s primary outcomes, including the PI and Resistance Index (RI), were extracted. We employed a random-effects model, specifically the DerSimonian-Laird method, to account for potential heterogeneity among studies. Hedges’s g was calculated to measure the effect size, which estimates the standardized mean difference between the groups (women with and without a history of cesarean section). Heterogeneity among the studies was assessed using the I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than chance. An I² value greater than 50% indicates substantial heterogeneity. Additionally, Cochran’s Q test was used to test for heterogeneity. Potential publication bias was assessed using Egger’s regression test. Given the small number of included studies, we supplemented Egger’s regression test with visual inspection of funnel plots for each meta-analysis outcome, in accordance with PRISMA recommendations. A non-significant p-value (p > 0.05) in Egger’s test suggests that there is no significant publication bias. Meta-regression was performed to explore the influence of potential confounders on the observed effect sizes. This included variables such as gestational age at the time of Doppler assessment, maternal age, and BMI. Sensitivity analyses were also conducted by excluding studies that are the source of the heterogeneity to assess the robustness of the findings. Since some studies reported the mean and standard deviation for the multiples of the median (MoM) of the PI, we generated simulated datasets reflecting the MoM distribution adjusted for gestational age using Gomez et al.‘s reference values to estimate the mean and standard deviation of the PI values [ 28 ]. This Bayesian simulation process involved adjusting the MoM values for gestational age through regression analysis and summarizing the simulated PI values. The results from 2,000 simulations provided robust estimates for the mean and standard deviation of the PI, adjusted for gestational age.

(DOCX) (DOCX) (DOCX) A: Forest plot of the PI meta-analysis following removal of the source of the heterogeneity. B: Funnel plot of the PI meta-analysis following removal of the source of the heterogeneity. (PDF)