Hypertensive disorders of pregnancy, particularly preeclampsia (PE), are major contributors to adverse outcomes, causing approximately 76,000 maternal deaths and over 500,000 perinatal and infant deaths each year worldwide [1, 2, 3]. Platelet activation, aggregation, and consumption at sites of endothelial injury are central to PE pathobiology, supporting the exploration of platelet-based metrics as accessible tools for risk stratification and screening [4]. A meta-analysis of 25 studies found lower platelet counts [pooled weighted mean difference (WMD): –41.45 $\times$ 10⁹/L] and higher mean platelet volume (MPV; pooled WMD: +0.98 fL) in PE versus normotensive pregnancy, reinforcing biological plausibility [5]. Complementarily, pooled diagnostic performance estimates indicate an MPV sensitivity of 0.676, specificity of 0.710, and an area under the curve (AUC) of the summary receiver operating characteristic curve (SROC) of 0.7889 for identifying PE [6]. Altogether, these findings position platelet indices as potential candidates for surveillance in obstetric care.

Despite these advances, uncertainties remain regarding optimal timing, thresholds, and clinical contexts for applying platelet indices [7]. In a large population-based analysis, higher first-trimester MPV categories were associated with gestational hypertension [9.8–10.3 fL: odds ratio (OR): 1.23; $\ge$10.4 fL: OR: 1.46], whereas associations with PE did not reach statistical significance, raising questions about phenotype-specific utility [8]. At the hospital level in Nepal, a platelet count–to–platelet distribution width (PC/PDW) ratio $\le$15.1 yielded an AUC of 0.767, with 70.8% sensitivity and 81.9% specificity for PE detection, and an 11.02-fold increase in odds at this cutoff. These findings underscore the potential of this low-cost triage while highlighting variability across clinical settings [9]. Community and clinical risk profiles also differ. Indeed, in Odisha, India, family history of hypertension and maternal anemia increased the odds of PE by 3.15 and 2.14, respectively, emphasizing context-specific baselines that may interact with biomarker performance [10]. Further, comparative policy analyses note inconsistency in how high-risk criteria are defined and when they are applied, complicating standardized implementation of prediction strategies [11]. In summary, applicability and thresholds for platelet indices likely vary across populations, timelines, and healthcare-system contexts.

Further heterogeneity is observed in cross-sectional data from Ethiopia, where higher MPV (10.6 $\pm$ 1.2 vs. 8.5 $\pm$ 1.0 fL), lower platelet counts (202.4 $\pm$ 71.5 vs. 259.9 $\pm$ 58.4 $\times$ 10³/µL), and increased PDW (15.2 $\pm$ 2.7 vs. 12.8 $\pm$ 2.5 fL) were documented among women with PE [12]. At the mechanistic level, trimester-specific shifts in platelet indices have been linked to aspirin nonresponsiveness, with increased placental CD62P/CD42b expression observed in PE and persisting uncertainties at the maternal–fetal interface [13]. Overall, the literature indicates clinical promise yet exposes heterogeneity in effect sizes, endpoints, and practice frameworks, which must be addressed for real-world use. Therefore, the aim of the present study was to evaluate the association between MPV and the risk of PE in a case-control cohort from the Peruvian National Maternal Perinatal Institute.

This case-control analysis indicates that MPV is higher in women with PE and remains independently associated with the condition after adjustment for maternal factors. Furthermore, gestational obesity and inadequate prenatal care show independent associations, whereas parity and age do not. Prior PE differentiated groups in unadjusted comparisons but was excluded from multivariable modeling due to quasi-complete separation. As a single biomarker, MPV demonstrates moderate discriminatory capability with a clinically plausible threshold, supporting its value as part of a multivariable risk profile rather than an individual diagnostic test.

Our study indicates that higher MPV is independently associated with PE, in alignment with multiple reports. Li et al. [18] observed higher MPV in both mild and severe PE versus healthy pregnancies, as well as a significant univariate association with PE. Similarly, Bawore et al. [19] reported elevated MPV in PE with a clear severity gradient. Prospective data corroborate the pattern: Udeh et al. [20] found increased MPV among women who later developed PE at 14–18 gestational weeks and at end-of-follow-up, and Thakkar et al. [21] reported higher baseline MPV at 14–18 gestational weeks in early-onset PE. Severity-related increases were also confirmed in a prospective study with ROC verification [22]. However, Lin et al. [23] observed a partial divergence in timing, as found no group differences in the first trimester, with separation appearing later (e.g., 16–19 weeks), suggesting gestational-age dependence. Methodological factors (prospective vs. case-control design), PE phenotype (early/severe vs. mixed), and analytical variability in hematology platforms may underlie residual heterogeneity. Collectively, converging evidence supports MPV as a biologically plausible marker of platelet-activation in PE. MPV is most informative after the early first trimester and offers an accessible metric with potential to enhance risk stratification in antenatal care, particularly resource-limited settings.

In our settings, MPV alone offered only moderate discrimination for PE using a single cut-point rule. Performance varied across studies, with early-onset cohorts reporting strong accuracy. Udeh et al. [20] showed high AUC with a 9.4 fL threshold and near-perfect accuracy in severe early-onset PE at ~10.7 fL, and Thakkar et al. [21] reported similarly high AUC with a $>$10.2 fL threshold. By contrast, Sachan et al. [22] found moderate AUCs with a 9.05 fL cutoff, balancing modest sensitivity against higher specificity, and Lin et al.’s [23] preterm-PE prediction yielded lower AUCs, closer to moderate ranges. These discrepancies likely reflect phenotype (early/severe disease amplifies signal), gestational-age at sampling (mid-second trimester yields clearer separation than first trimester), and inter-laboratory MPV calibration differences that shift optimal thresholds across platforms. Taken together, MPV’s individual diagnostic value appears context-dependent, showing greatest utility in early or severe PE. Its effectiveness is maximized within multivariable, trimester-specific algorithms rather than as a solitary screening tool, a strategy with practical relevance for scalable, tiered PE risk assessment across diverse health systems.

We observed that gestational obesity was independently associated with PE. This aligns with studies showing higher mid-pregnancy adiposity among women with PE. Li et al. [18] reported a higher second-trimester BMI in PE alongside elevated systolic and diastolic pressures, and Thakkar et al. [21] found higher gestational BMI among PE cases while other baseline factors were comparable. Differences in effect sizes across reports likely reflect the timing and method of BMI assessment, PE phenotype mix (early vs. late onset), and residual confounding by hemodynamics or glycemic status that track with weight gain. From a policy perspective, these convergent findings support integrating routine mid-trimester weight surveillance and metabolic risk counseling into antenatal pathways to control preventable PE burden in resource-constrained settings.

In our analysis, the crude association between pregestational obesity and PE lost significance after multivariable adjustment. External literature largely documents higher baseline BMI among PE cases. Lin et al. [23] reported a significant pregestational BMI difference, but does not explicitly test for attenuation after adjustment. Moreover, one cohort restricted pre-pregnancy BMI ($\ge$30 kg/m² excluded), limiting evaluation of true baseline effects [18]. Attenuation in our context may arise from collinearity with gestational weight gain or blood pressure, measurement error in recalled pre-pregnancy weight, and phenotype heterogeneity. Clinically, this distinction underscores that dynamic gestational adiposity may be a more actionable target than static preconception BMI alone, informing antenatal counseling strategies with greatest translational impact in primary care.

Parity and maternal age were not associated with PE in our study. The literature is mixed but broadly consistent with a limited role of these demographics in certain cohorts. Lin et al. [23] observed older age associated with PE cases but a similar parity distribution without significance. Udeh et al. [20] observed no differences in age or gravidity between women who developed PE and those who did not. Thakkar et al. [21] also found no group differences among these covariates. Variability likely reflects sampling frames, age ranges, and the dominance of metabolic or placental factors over demographic predictors. Public-health programs should therefore prioritize modifiable cardiometabolic risks over demographic screening alone to improve PE prevention yield.

A prior history of PE differentiated groups in bivariate analysis but could not be retained in multivariable modeling due to quasi-complete separation. The external evidence supports a similar direction of effect. In fact, Lin et al. [23] reported a higher proportion of multiparouss with prior pregnancy-induced hypertension in the PE group, whereas Bawore et al. [19] mitigated this issue by excluding women with prior PE, precluding its analysis as a predictor. Such handling decisions reflect the well-known strength of recurrence risk and its analytic challenges. In practice, meticulous obstetric history-taking remains a low-cost, high-value measure to identify elevated PE risk early, guiding intensified surveillance in settings with limited health-system resources.

Our case series predominantly comprised later-onset presentations. In contrast, prospective cohorts quantifying early-onset PE report relatively low prevalence but provide detailed distributions of onset timing. Thakkar et al. [21] estimated an overall early-onset PE of 5.3%, with most cases clustering between 28–32 weeks, and Udeh et al. [20] reported a comparable early-onset frequency of 5.9% with severity stratification. Differences between our hospital-based series and prospective cohorts likely reflect referral patterns, triage thresholds, and regional epidemiology. Recognizing a primarily late-onset profile in our setting argues for scalable second-trimester risk assessment and timely escalation pathways, aligning limited resources to the period of greatest preventable morbidity.

Taken together, our findings support MPV as a pragmatic, laboratory-ready biomarker that captures platelet activation relevant to PE pathophysiology and can be leveraged within routine antenatal workflows. Because MPV is generated automatically by standard hematology analyzers at no additional cost, it offers a scalable tool for risk stratification in settings with limited resources and high disease burden. Clinically, MPV is unlikely to serve as a stand-alone test; rather, its value lies in trimester-specific algorithms that integrate maternal characteristics and blood pressure with hematologic and metabolic indices to identify women who may benefit from closer surveillance or preventive measures. By enabling earlier identification of at-risk pregnancies, particularly beyond the earliest weeks when predictive accuracy improves, MPV-informed pathways could help reduce preventable maternal and perinatal complications and align with global efforts to strengthen low-cost, point-of-care diagnostics in maternal health programs.

This study was conducted at a national tertiary referral center with standardized clinical records and laboratory workflows, enhancing internal validity and consistency of MPV assessment. The case-control design enabled efficient capture of clinically adjudicated PE and appropriate normotensive controls from the same source population and time frame. We complemented group comparisons with multivariable modeling to isolate the independent contribution of MPV while accounting for key maternal factors. Diagnostic performance was reported transparently using ROC analysis and likelihood ratios at a pre-specified cut-point, with clear acknowledgment of the design-specific constraints (e.g., not estimating predictive values). Together, these features provide a coherent and clinically interpretable evaluation of MPV’s association with PE and its potential utility as an accessible component of antenatal risk assessment.

In this hospital-based case-control study, higher MPV was independently associated with PE, and as a single biomarker, MPV showed moderate discrimination using a pragmatic cutoff. Gestational obesity and inadequate prenatal care were also associated with PE, whereas parity and age were not. Given that MPV is routinely generated by hematology analyzers at no additional cost, integrating MPV into trimester-specific, multivariable antenatal risk models could help prioritize surveillance and preventive measures, especially in resource-limited settings. Prospective, multicenter studies should establish analyzer- and trimester-specific thresholds and quantify absolute risk to guide implementation and decision-curve–based clinical use.

The datasets generated and analyzed during the current study are not publicly available due to institutional and patient confidentiality policies, but are available from the corresponding author upon reasonable request.

RJR-Y: Data Curation, Investigation, Methodology, Writing Original Draft, and Writing Review & Editing. GD-B: Data Curation, Investigation, Methodology, and Writing Review & Editing. MAA-H: Conceptualization, Methodology, Formal Analysis, Supervision, Validation, Writing Original Draft, and Writing Review & Editing. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was conducted in accordance with the Declaration of Helsinki and approved by the Comité Institucional de Ética e Integridad Científica (CIEI) of Universidad Privada Norbert Wiener (Expediente 0995-2024; approval issued 12 November 2024). Informed consent was waived due to the retrospective nature of the study and the use of anonymized medical records without direct patient contact.

We thank the clinical and laboratory staff of the Obstetrics Hospitalization Service at the Instituto Nacional Materno Perinatal (INMP), Lima, for their support during data abstraction and verification.

This research received no external funding.

The authors declare no conflict of interest.

During the preparation of this work, the authors used ChatGPT-3.5 in order to check spelling and grammar. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.