This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and Cochrane handbook [19, 20]. The study was registered in PROSPERO (CRD420261306923).

A comprehensive literature search was performed in PubMed, EMBASE, Web of Science, and SCOPUS from database inception to January 2026. Search terms combined keywords and controlled vocabulary related to armed conflict (e.g., war, bombing, invasion, armed conflict) and neonates. The search strategy was adapted for each database. Reference lists of included studies were also screened to identify additional relevant articles Supplementary Table 1.

Studies were included if they examined populations exposed to armed conflict, incorporated a comparator group (such as pre-war versus wartime periods, exposed versus non-exposed groups, or temporal comparisons), reported neonatal outcomes, and used observational study designs including cohort, cross-sectional comparative, or registry-based approaches. Studies were excluded if they lacked a comparator group, were not related to war or armed conflict, were conference abstracts without sufficient data, or reported outcomes not relevant to neonatal health.

All identified records were imported into reference management software and duplicates were removed. Two reviewers independently screened titles and abstracts for eligibility, followed by full-text assessment of potentially relevant studies. Disagreements were resolved through discussion until consensus was reached.

Data were extracted independently by two reviewers using a standardized data extraction form. Extracted information included study characteristics, population description, group sample sizes, comparator definition, and neonatal outcomes. Dichotomous outcomes were extracted as event counts and total sample sizes. Continuous outcomes were extracted as means and standard deviations. When outcomes were reported as rates per 1,000 births, event numbers were calculated using the reported denominators to enable meta-analysis.

Methodological quality was assessed independently by two reviewers. Cohort and historical cohort studies were evaluated using the Newcastle–Ottawa Scale [21], which assesses selection, comparability, and outcome domains with a maximum score of nine. Studies were categorized as high, moderate, or low quality based on established thresholds. The cross-sectional study was assessed using the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [22], which includes fourteen methodological domains and provides an overall rating of good, fair, or poor.

Meta-analyses were conducted using Review Manager (RevMan 5.4). Dichotomous outcomes were pooled using odds ratios (ORs) with 95% confidence intervals (CIs) for preterm birth, stillbirth, and neonatal mortality, while continuous neonatal anthropometric outcomes, including birth weight, birth length, and head circumference, were pooled using mean differences (MDs) with 95% CIs. Statistical heterogeneity was evaluated using the Chi-square test and quantified with the I² statistic. A fixed-effects model was initially applied; when substantial heterogeneity (P < 0.10) was detected, a random-effects model was used. Sensitivity analyses were performed by excluding studies contributing most to heterogeneity. Outcomes that could not be quantitatively pooled due to inconsistent definitions or reporting were synthesized narratively. No subgroup analysis was conducted. A p-value <0.05 was considered statistically significant.

The studies were conducted in Ukraine, Poland, Serbia, Croatia/Bosnia, and Ethiopia, and employed retrospective cohort, historical cohort, cross-sectional comparative, and national registry designs. Sample sizes varied substantially, ranging from small hospital-based comparisons (n = 39) to national registry data exceeding 500,000 live births. Most studies compared pre-war and wartime periods, while one study contrasted residents with internally displaced persons. Baseline maternal characteristics were inconsistently reported. Table 1.

According to the Newcastle–Ottawa Scale assessment, cohort studies demonstrated moderate to high methodological quality, with scores ranging from 6 to 9 out of 9. Three studies were rated high quality (scores 7–9), reflecting strong selection procedures and objective outcome assessment, particularly in large registry and hospital-based cohorts [14, 15, 25]. Moderate-quality studies (6/9) primarily lacked adequate control for confounders, as they did not perform multivariable adjustment or matching for key maternal characteristics such as age, parity, socioeconomic status, and pre-existing medical conditions, thereby limiting comparability between exposed and unexposed groups [13, 24, 26]. Overall, the quality assessment indicated generally acceptable methodological quality, although important limitations were identified in confounding control and comparability Supplementary Table 2.

The cross-sectional study was rated as fair quality due to limited sample size, lack of confounder adjustment, and absence of sample size justification, despite clearly defined exposure and outcome measures [23]. These quality limitations were considered when interpreting the pooled findings. Supplementary Table 3.

Beyond the pooled outcomes, several neonatal indicators were reported but not meta-analyzed due to inconsistent definitions or limited reporting. Lakhno et al. [23] reported significantly shorter birth length (44.84 ± 7.64 cm vs. 52.0 ± 7.22 cm) and lower Apgar scores (6.74 ± 2.05 vs. 7.7 ± 2.08) among internally displaced infants compared with residents [23]. Lakhno et al. [13] documented variations in gestational age (GA) distribution across 2021–2023 without detailed anthropometric stratification [13]. Liczbinska et al. [24] additionally reported changes in miscarriage rates (895 pre-war vs. 851 during WWII) [24].

Our systematic review and meta-analysis of seven studies across diverse geographic settings found no significant associations between armed conflict exposure and preterm birth, stillbirth, or neonatal mortality after addressing heterogeneity. Unexpectedly, exposed neonates showed significantly higher BW and head circumference than non-exposed neonates.

These counterintuitive findings warrant careful interpretation. Several mechanisms may explain the paradoxical increase in anthropometric measures. First, survival bias likely plays a substantial role, infants who survive to delivery in conflict settings may represent a more reliable subset of pregnancies, as those with compromised fetal health may be more likely to result in early pregnancy loss or SB that goes unreported in conflict zones with weakened surveillance systems [7, 27]. Second, population displacement patterns may concentrate healthier, more socioeconomically advantaged families near functioning health facilities, while those unable to relocate face worse outcomes that remain invisible to hospital-based studies [17, 28]. Third, the included studies demonstrated considerable methodological heterogeneity in exposure definition, timing assessment, and comparison group selection, which may affect true associations.

The absence of significant findings for SB and neonatal mortality contradicts both theoretical expectations and prior observational evidence suggesting heightened perinatal risks in conflict settings [4, 17]. However, this lack of association may reflect the limitations of available data rather than true biological protection. Most included studies relied on facility-based data, potentially missing the most vulnerable populations who cannot access healthcare. Finally, the moderate quality of several studies, particularly the absence of multivariable adjustment for critical confounders such as maternal socioeconomic status and nutritional status, limits our ability to isolate the independent effects of conflict exposure [13, 24, 26].

The included studies had strengths such as large registry or hospital-based datasets, defined comparison groups, and objective outcome measures. However, several were retrospective, single-center, or historical studies, with potential selection bias, temporal confounding, and limited adjustment for maternal and socioeconomic factors. These differences should be considered when interpreting the pooled estimates.

Our findings diverge substantially from prior research examining armed conflict and neonatal outcomes. A comprehensive study by Le and Nguyen analyzing 53 developing countries over three decades found that intrauterine exposure to armed conflict reduced BW by 2.8% and increased LBW incidence by 3.2 percentage points [29]. Similarly, Mansour and Rees, investigating the al-Aqsa Intifada’s impact, documented modest increases in LBW risk among conflict-exposed Palestinian infants [30]. Studies, utilizing large population-based datasets and difference-in-differences designs likely captured more representative samples than the predominantly hospital-based studies in our review.

A systematic review by Keasley et al. [31] identified evidence for increased LBW risk across nine studies of mothers exposed to armed conflicts in diverse settings including Libya, Bosnia, Palestine, and Afghanistan [31]. Their narrative synthesis highlighted consistent patterns of adverse pregnancy outcomes, though methodological heterogeneity limited quantitative pooling of all outcomes. Importantly, that review encompassed studies examining LBW as a binary outcome rather than continuous BW measures, which may explain some discordance with our findings.

More recently, Jawad et al. [11] conducted a regression analysis of 181 countries from 2000 to 2019, demonstrating that conflicts classified as wars were associated with increased maternal and infant mortality, though notably, they found no evidence of association with neonatal mortality, a finding consistent with our results [11]. This suggests that conflict’s impact on neonatal survival may be more nuanced and context-dependent than its effects on broader maternal and child health indicators.

A meta-analysis by Behboudi-Gandevani et al. [17] examining immigrants from conflict zones found increased odds of small for GA and neonatal mortality compared to host populations, highlighting that conflict effects may extend beyond the immediate war zone through displacement and refugee experiences [17].

The clinical implications of our findings must be considered cautiously given the methodological limitations and potential for selection bias. While the absence of significant adverse associations may offer superficial reassurance, this likely reflects data gaps rather than genuine protection. Healthcare providers caring for conflict-affected populations should maintain heightened awareness of adverse outcomes regardless of these pooled estimates.

The paradoxical anthropometric increases highlight that hospital-based datasets systematically undercount the most vulnerable, underscoring the need for community-based surveillance and mobile health initiatives [6, 18]. Humanitarian organizations must prioritize establishing parallel data collection mechanisms outside hospital settings to capture the full spectrum of outcomes.

Beyond immediate obstetric outcomes, the health effects of war may also be mediated by exposure to hazardous munitions and toxic conflict remnants. Reported use of incendiary substances such as white phosphorus in populated areas has raised serious humanitarian and public health concerns, particularly regarding possible long-term reproductive, developmental, and intergenerational consequences [32]. These risks are rarely measurable in routine hospital-based datasets, reinforcing the likelihood that current pooled estimates underestimate the true burden of conflict on maternal and neonatal health.

From a public health perspective, these findings highlight the critical importance of maintaining essential maternal and newborn care services during armed conflicts. Even if hospital-based outcomes appear preserved, this may reflect successful triage and resource allocation rather than absence of need. Interventions should focus on ensuring access to midwifes, emergency obstetric care, and nutritional support for all pregnant women in conflict zones, particularly those unable to reach formal healthcare facilities [12].

This systematic review and meta-analysis represents, to our knowledge, the first quantitative synthesis examining the association between armed conflict exposure and neonatal outcomes. Studies from diverse geographic regions were included, enhancing the relevance and applicability of findings across different conflict settings. Methodological quality was systematically evaluated using validated assessment tools, allowing transparent appraisal of study strengths and limitations.

However, several important limitations constrain interpretation. First, the small number of included studies (n = 7) and their predominant reliance on hospital-based data limit representativeness. Second, Substantial statistical heterogeneity across outcomes reflects methodological diversity in exposure definitions, comparison groups, and confounder adjustment. Third, most studies demonstrated moderate quality with inadequate control for socioeconomic and nutritional confounders that may independently affect neonatal outcomes. Fourth the absence of individual patient data precluded subgroup analyses by conflict intensity, exposure timing, or maternal characteristics. Fifth, publication bias assessment was not feasible given the limited number of studies. Sixth, many studies do not fully reflect the context of long-standing conflicts in the region (e.g., Syria, Libya, Palestine, Lebanon, Iraq, Iran, Sudan, and Yemen). Seventh, maternal and neonatal outcomes in these settings are influenced not only by conflict but also by economic collapse, displacement, food insecurity, and disruption of healthcare systems, which may produce large-scale ramifications that are difficult to disentangle from the direct effects of war. Finally, the ecological nature of conflict exposure measurement in many studies prevents definitive causal inference about individual-level effects.

We observed no significant associations between conflict exposure and preterm birth, stillbirth, or neonatal mortality, alongside paradoxically increased birth weight and head circumference. Rather than biological protection, this pattern likely indicates facility-level selection bias, missed out-of-hospital deliveries, and fragile surveillance in war zones. Therefore, the apparent lack of adverse associations likely reflects data limitations rather than the absence of serious conflict-related effects. Future research requires community-based prospective cohorts with multivariable adjustment (nutrition, socioeconomic indicators, healthcare access) and exposure timing. Clinicians should maintain heightened risk awareness, while policymakers and humanitarians should enforce health-facility protections and invest in community-level outreach and parallel surveillance to reach the most vulnerable pregnant populations. Ultimately, preventing armed conflict and spreading peace remains the most effective intervention for protecting maternal and neonatal health globally.