The impact of magnesium levels in pregnancy on adverse birth outcomes: a protocol for a systematic review and meta-analysis Authors: Naeera Abdul, Vinogrin Dorsamy, Chauntelle Bagwandeen Published: 2026 ABSTRACT AbstractBackgroundMagnesium plays a critical role in maternal and fetal physiology, influencing vascular tone, inflammation, and cellular metabolism. Despite its established biological importance, evidence linking maternal magnesium status to adverse pregnancy outcomes remains inconsistent, partly due to heterogeneity in measurement methods, reference ranges, and study populations. These uncertainties are of significant public health importance globally but are of particular relevance in low- and middle-income countries (LMICs), where comorbidities such as HIV infection, anaemia, and obesity are prevalent. This review will therefore include studies from all settings while prespecifying subgroup analyses relevant to LMIC contexts.ObjectivesThis systematic review and meta-analysis aims to synthesise global evidence on the association between maternal magnesium levels and adverse pregnancy outcomes. Secondary objectives are to explore trimester-specific patterns of maternal magnesium status, assess whether common comorbidities modify these associations, including HIV infection, anaemia, and obesity, and examine evidence from magnesium supplementation trials in a separate interventional evidence stream.Eligibility criteriaWe will include observational studies (cross-sectional, case–control, and cohort) and randomised controlled trials reporting quantitative maternal magnesium measurements and at least one adverse maternal or neonatal outcome. Peer-reviewed journal articles and postgraduate theses will be eligible, with no restrictions on language, setting, or publication year.Information sourcesElectronic searches will be conducted in PubMed, Embase, Scopus, Cochrane Library, DOAJ, LILACS, and African Index Medicus, alongside grey literature sources including OpenGrey and ProQuest Dissertations and Theses.Risk of biasRisk of bias will be assessed independently by two reviewers using design-specific tools: Hoy’s Risk of Bias Tool for prevalence studies, ROBINS-I for non-randomised observational studies, and the Cochrane Risk of Bias 2 tool for randomised controlled trials.Synthesis of resultsObservational and interventional studies will be analysed in separate prespecified analytic streams. Meta-analyses will be conducted where sufficient clinical and methodological homogeneity exists, using random-effects or quality-effects models. Where quantitative synthesis is not appropriate, findings will be synthesised narratively. Subgroup analyses will explore trimester of pregnancy, geographic region, laboratory measurement method, and comorbidity status. Certainty of evidence will be assessed using the GRADE framework.Systematic review registrationPROSPERO CRD42024518427.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13643-026-03204-y.Keywords: Pregnancy, Magnesium, Maternal health, Pregnancy outcomes, Comorbidities, Low- and middle-income countries Background The role of magnesium in pregnancy represents a research area of both significant promise yet ongoing controversy. Whilst the global consensus is that magnesium is vital for maternal and fetal health, inconsistencies in the literature regarding its optimal levels and association with pregnancy outcomes hinder its integration into prenatal care guidelines. There is variability in reference ranges used by different laboratories; commonly cited ranges for serum magnesium include both 0.75–0.95 mmol/L and 0.65–1.05 mmol/L with some studies suggesting 0.85 mmol/L as a suitable minimum threshold for defining adequacy [1–5]. This lack of standardisation in establishing the optimum magnesium range, together with conflicting results from existing studies, leaves critical knowledge gaps regarding the relationship between maternal magnesium status and pregnancy outcomes. These uncertainties are relevant across settings, but they have particular public health importance in low- and middle-income countries (LMICs), where maternal comorbidities and adverse pregnancy are disproportionately prevalent relative to available resources. This review therefore adopts a global inclusion strategy while prespecifying analyses that will examine whether findings differ by region and by comorbidity status. Magnesium is an essential trace mineral that supports over 600 enzymatic reactions, including those responsible for DNA synthesis, muscle contraction, energy metabolism, and nerve function [6, 7]. During pregnancy, there is an extension of these biochemical functions, directly influencing maternal–fetal health as magnesium requirements increase to accommodate fetal growth and maternal physiological adaptations. Sufficient magnesium levels help regulate vascular tone, reduce inflammation, and stabilize cellular energy pathways—key factors in preventing complications such as gestational diabetes mellitus (GDM), pregnancy-induced hypertension (PIH), and pre-eclampsia (PE) [8]. Magnesium sulphate, a cornerstone treatment for severe PE and eclampsia, further highlights the role of mineral’s clinical importance [9]. However, maintaining optimal magnesium levels is a delicate balance, as both hypomagnesemia and hypermagnesemia are associated with adverse maternal and neonatal outcomes [10, 11]. Compounding the challenges in understanding magnesium’s role is its complex interplay with other comorbidities, particularly those common in LMICs, such as HIV, anaemia, and obesity, which not only influence magnesium metabolism but also exacerbate pregnancy-related risks. For example, antiretroviral therapy in HIV-positive women may disrupt magnesium homeostasis, whilst anaemia and obesity independently contribute to adverse maternal–fetal outcomes [12–14]. Despite the potentially significant public health impact to be accrued, research addressing the interactions between magnesium levels and these comorbidities in vulnerable populations remains limited and inconsistent. The lack of comprehensive evidence impedes the development of targeted interventions and clinical guidelines that consider multifaceted challenges faced by high-risk populations, with few studies addressing the interactions between magnesium and these conditions in vulnerable populations [15–17]. To address these critical gaps in the literature, this article presents a protocol for a systematic review and meta-analysis aimed at synthesising current evidence on the relationship between maternal magnesium levels and adverse pregnancy outcomes. Unlike existing studies, which are often limited in scope and population specificity, this review seeks to provide a comprehensive and population-sensitive analysis of magnesium’s role in maternal and fetal health [18]. Importantly, to our knowledge, no systematic review has yet comprehensively evaluated maternal magnesium levels and their association with pregnancy outcomes. By consolidating existing data and identifying key gaps in current knowledge, the findings of this review will serve as a foundation for evidence-based clinical guidelines and public health strategies. The review will include studies from all geographic and income settings to ensure comprehensive evidence capture. However, prespecified subgroup analyses will examine regional patterns and comorbidity-specific findings, with particular attention to LMIC-relevant conditions such as HIV infection, anaemia, and obesity. In doing so, the review aims to provide evidence that is both globally informative and locally relevant to settings where nutritional deficiencies and pregnancy-related morbidity remain highly prevalent. Objectives Primary objective To evaluate the available evidence on the association between maternal magnesium levels and adverse pregnancy outcomes, including PE, GDM, preterm birth, and low birth weight (LBW). Secondary objectives To describe trimester-specific distributions and patterns of maternal magnesium status across pregnancy. To describe trimester-specific distributions and patterns of maternal magnesium status across pregnancy. To assess whether common comorbid conditions, including HIV infection, anaemia, and obesity, modify the association between maternal magnesium levels and adverse pregnancy outcomes. To assess whether common comorbid conditions, including HIV infection, anaemia, and obesity, modify the association between maternal magnesium levels and adverse pregnancy outcomes. To explore evidence from randomised controlled trials evaluating magnesium supplementation during pregnancy, analysed separately from observational evidence. To explore evidence from randomised controlled trials evaluating magnesium supplementation during pregnancy, analysed separately from observational evidence. Methodology This protocol for a systematic review and meta-analysis adheres to the Preferred Reporting Items for Systematic Review and Meta-analysis for Protocols (PRISMA-P) 2015 guidelines [19]. The completed PRISMA-P checklist has been submitted as a supplementary file. The final systematic review will be reported in accordance with the PRISMA 2020 statement to ensure transparency and completeness of reporting [20]. Framing the research question using the PICOS framework This review adopts a PICOS framework to guide the development of the research question and eligibility criteria, with exposure-based comparisons used for observational studies (Table 1): Table 1PICOS framework for eligibility of research questionPopulationPregnant women at any gestational age, including subgroups with comorbidities such as HIV infection, anaemia, or obesityExposureMaternal magnesium status measured using validated laboratory methods (e.g. serum, plasma, erythrocyte magnesium)ComparatorWhere applicable, comparisons across magnesium exposure categories (e.g. deficient vs adequate, quartiles, continuous measures)OutcomesAdverse maternal and neonatal outcomesStudy DesignObservational studies and randomised controlled trials, analysed separately as prespecifiedThe PICOS framework informed the inclusion criteria and helped ensure a focused and structured literature search PICOS framework for eligibility of research question The PICOS framework informed the inclusion criteria and helped ensure a focused and structured literature search Eligibility criteria Inclusion criteria ◦ Original research articles reporting quantitative maternal magnesium measurements obtained using validated laboratory methods. ◦ Original research articles reporting quantitative maternal magnesium measurements obtained using validated laboratory methods. ◦ Studies reporting at least one predefined maternal or neonatal outcome, including pre-eclampsia, eclampsia, gestational diabetes mellitus, pregnancy-induced hypertension, intrauterine growth restriction, fetal growth restriction, small for gestational age, preterm birth, preterm delivery, premature birth, low birth weight, or birth weight. ◦ Studies reporting at least one predefined maternal or neonatal outcome, including pre-eclampsia, eclampsia, gestational diabetes mellitus, pregnancy-induced hypertension, intrauterine growth restriction, fetal growth restriction, small for gestational age, preterm birth, preterm delivery, premature birth, low birth weight, or birth weight. ◦ Study designs: cross-sectional, case–control, cohort studies (prospective or retrospective), and randomised controlled trials. ◦ Study designs: cross-sectional, case–control, cohort studies (prospective or retrospective), and randomised controlled trials. ◦ Peer-reviewed journal articles and postgraduate theses with sufficient methodological and outcome data. ◦ Peer-reviewed journal articles and postgraduate theses with sufficient methodological and outcome data. Exclusion criteria ◦ Case reports, narrative reviews, systematic reviews, editorials, or commentaries. ◦ Case reports, narrative reviews, systematic reviews, editorials, or commentaries. ◦ Conference abstracts without full-text data. ◦ Conference abstracts without full-text data. ◦ Non-human studies. ◦ Non-human studies. ◦ Studies lacking extractable data on magnesium levels or pregnancy outcomes. ◦ Studies lacking extractable data on magnesium levels or pregnancy outcomes. Information sources and search strategy A comprehensive electronic search will be conducted in PubMed, Embase, Scopus, Cochrane Library, DOAJ, LILACS, and African Index Medicus. Grey literature will be searched via OpenGrey and ProQuest Dissertations and Theses. Reference lists of included studies will be hand-searched to identify eligible articles. The search strategy was developed in consultation with an expert in the field and peer reviewed using the PRESS 2015 checklist [21]. No language or date limits will be applied. Non-English titles and abstracts will initially be translated using Google Translate or equivalent translation software to determine potential eligibility. For studies that appear eligible at full-text stage, translation will be undertaken using a combination of machine translation and, where necessary, assistance from a fluent speaker or professional translation support to ensure accurate extraction of methods, exposure definitions, outcome definitions, and numerical data. Articles that cannot be translated with sufficient accuracy for eligibility assessment or data extraction will be listed in an appendix with reasons for exclusion. Full search strategies for each database will be reported and included as an appendix. Boolean operators and keywords will incorporate variations in terminology, such as: ◦ Pregnancy terms: (“pregnancy” OR “gestation” OR “maternal”). ◦ Pregnancy terms: (“pregnancy” OR “gestation” OR “maternal”). ◦ Magnesium terms: (“magnesium” OR “MgSO4” OR “hypomagnesemia” OR “hypermagnesemia”). ◦ Magnesium terms: (“magnesium” OR “MgSO4” OR “hypomagnesemia” OR “hypermagnesemia”). ◦ Outcome terms: (“pre-eclampsia” OR “preeclampsia” OR “eclampsia” OR “gestational diabetes” OR “pregnancy-induced hypertension” OR “intrauterine growth restriction” OR “fetal growth restriction” OR “small for gestational age” OR “preterm birth” OR “preterm delivery” OR “premature birth” OR “low birth weight” OR “LBW” OR “birth weight” OR “pregnancy outcome”). ◦ Outcome terms: (“pre-eclampsia” OR “preeclampsia” OR “eclampsia” OR “gestational diabetes” OR “pregnancy-induced hypertension” OR “intrauterine growth restriction” OR “fetal growth restriction” OR “small for gestational age” OR “preterm birth” OR “preterm delivery” OR “premature birth” OR “low birth weight” OR “LBW” OR “birth weight” OR “pregnancy outcome”). ◦ Comorbidity terms: (“HIV” OR “anaemia” OR “obesity”). ◦ Comorbidity terms: (“HIV” OR “anaemia” OR “obesity”). The search strategy was piloted on the 23rd of May 2025 to determine the appropriateness of the selected keywords and electronic databases (Table 2). Table 2Potential search strategy on PubMedSearchQuery#1((((("pregnant"[All Fields]) OR ("pregnancy"[All Fields])) OR ("maternal"[All Fields])) OR ("mother"[All Fields])) OR ("gestation"[All Fields])) OR ("gestational"[All Fields])#2("preeclampsia"[All Fields] OR "pre-eclampsia"[All Fields] OR "eclampsia"[All Fields] OR "hypertension, pregnancy-induced"[MeSH Terms] OR "pregnancy-induced hypertension"[All Fields] OR "gestational hypertension"[All Fields] OR "hypertensive disorder"[All Fields] OR "hypertensive disorders of pregnancy"[All Fields] OR "gestational diabetes"[All Fields] OR "diabetes, gestational"[MeSH Terms] OR "gestational diabetes mellitus"[All Fields] OR "GDM"[All Fields] OR "intrauterine growth restriction"[All Fields] OR "IUGR"[All Fields] OR "fetal growth restriction"[All Fields] OR "FGR"[All Fields] OR "small for gestational age"[All Fields] OR "SGA"[All Fields] OR "infant, small for gestational age"[MeSH Terms] OR "preterm birth"[All Fields] OR "premature birth"[All Fields] OR "preterm delivery"[All Fields] OR "preterm labour"[All Fields] OR "preterm labor"[All Fields] OR "premature birth"[MeSH Terms] OR "low birth weight"[All Fields] OR "LBW"[All Fields] OR "birth weight"[All Fields] OR "birth weight"[MeSH Terms] OR "pregnancy outcome"[All Fields] OR "pregnancy outcome"[MeSH Terms] OR "pregnancy complications"[MeSH Terms] OR "pregnancy complication"[All Fields])#3((((((((("magnesium"[All Fields]) OR ("magnesium"[MeSH Terms])) OR ("magnesium sulphate"[All Fields])) OR ("magnesium sulfate"[All Fields])) OR ("magnesium sulfate"[MeSH Terms])) OR ("mgso4"[All Fields])) OR ("hypermagnesemia"[All Fields])) OR ("hypomagnesemia"[All Fields])) OR ("low magnesium"[All Fields])) OR ("high magnesium"[All Fields])#4#1 AND #2 AND #3 Potential search strategy on PubMed Study selection and data management All identified records will be imported into Zotero (Version 6.0.30) for reference management and duplicate removal. After deduplication, two reviewers (NA and VD) will independently screen titles and abstracts, followed by full-text screening of potentially eligible studies. Discrepancies will be resolved through discussion, with a third reviewer (CB) acting as arbitrator when necessary. Calibration exercises will be conducted prior to screening using a random sample of studies to ensure consistency. Inter-rater agreement will be quantified using Cohen’s kappa statistic. All screening decisions will be documented, and the study selection process will be summarized using a PRISMA 2020 flow diagram, including explicit reporting of duplicate records removed (Fig. 1) [20, 22]. PRISMA 2020 flow diagram describing selection of studies for systematic review of magnesium levels in pregnancies with adverse outcomes [23] Data extraction A standardised data extraction form will be developed in Microsoft® Excel and piloted on a subset of 10 studies prior to full extraction [24]. Two reviewers (NA and VD) will independently extract data, with disagreements resolved by consensus or by a third reviewer (CB). The following data will be extracted: ◦ Study characteristics: author(s), year, country, study design, setting, sample size. ◦ Study characteristics: author(s), year, country, study design, setting, sample size. ◦ Population characteristics: maternal age, gestational age, parity, BMI, HIV status, antiretroviral therapy use, anaemia status, obesity, socioeconomic indicators. ◦ Population characteristics: maternal age, gestational age, parity, BMI, HIV status, antiretroviral therapy use, anaemia status, obesity, socioeconomic indicators. ◦ Exposure assessment: specimen type (serum/plasma/erythrocyte/urine), timing of sampling by trimester, laboratory assay methods, units, reference ranges. ◦ Exposure assessment: specimen type (serum/plasma/erythrocyte/urine), timing of sampling by trimester, laboratory assay methods, units, reference ranges. ◦ Exposure format: continuous measures (mean/SD), categorical measures (quartiles, deficiency thresholds), z-scores. ◦ Exposure format: continuous measures (mean/SD), categorical measures (quartiles, deficiency thresholds), z-scores. ◦ Nutritional co-exposures: magnesium supplementation (dose, formulation, timing, adherence), dietary magnesium intake (assessment method where available), other supplements (e.g. calcium, vitamin D, iron). ◦ Nutritional co-exposures: magnesium supplementation (dose, formulation, timing, adherence), dietary magnesium intake (assessment method where available), other supplements (e.g. calcium, vitamin D, iron). ◦ Outcome definitions and ascertainment: operational definitions, diagnostic criteria, data source (clinical diagnosis, chart review, registry). ◦ Outcome definitions and ascertainment: operational definitions, diagnostic criteria, data source (clinical diagnosis, chart review, registry). ◦ Effect estimates: adjusted and unadjusted estimates (OR, RR, HR, mean difference) with confidence intervals. ◦ Effect estimates: adjusted and unadjusted estimates (OR, RR, HR, mean difference) with confidence intervals. ◦ Statistical modelling: covariates included, model type, handling of missing data. ◦ Statistical modelling: covariates included, model type, handling of missing data. ◦ Additional notes: author-reported limitations and reviewer comments. ◦ Additional notes: author-reported limitations and reviewer comments. Risk of bias assessment Risk of bias will be assessed independently by two reviewers (NA and VD) using tools appropriate to study design: ◦ Hoy’s risk of bias tool will be used exclusively for prevalence and cross-sectional studies reporting magnesium status distributions [25]. ◦ Hoy’s risk of bias tool will be used exclusively for prevalence and cross-sectional studies reporting magnesium status distributions [25]. ◦ ROBINS-I will be applied to non-randomised observational studies assessing associations between magnesium levels and pregnancy outcomes [26]. ◦ ROBINS-I will be applied to non-randomised observational studies assessing associations between magnesium levels and pregnancy outcomes [26]. ◦ Cochrane Risk of Bias 2 (RoB 2) tool will be used for randomised controlled trials evaluating magnesium supplementation [27]. ◦ Cochrane Risk of Bias 2 (RoB 2) tool will be used for randomised controlled trials evaluating magnesium supplementation [27]. Disagreements will be resolved by consensus or consultation with a third reviewer (CB). Risk of bias assessments will inform sensitivity analyses and GRADE certainty ratings [28]. GRADE assessment The certainty of evidence for each outcome will be assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework [28].This approach evaluates the following five domains: Risk of bias–based on assessments using Hoy’s tool, ROBINS-I and RoB 2 for individual studies [25–27]. Risk of bias–based on assessments using Hoy’s tool, ROBINS-I and RoB 2 for individual studies [25–27]. Inconsistency—assessed through heterogeneity in results across studies (e.g. I2 > 50%). Inconsistency—assessed through heterogeneity in results across studies (e.g. I2 > 50%). Indirectness—evaluates how directly the evidence applies to the review question in terms of population, exposure, and outcomes. Indirectness—evaluates how directly the evidence applies to the review question in terms of population, exposure, and outcomes. Imprecision—determined by the size of the confidence intervals and the overall sample size contributing to the estimate. Imprecision—determined by the size of the confidence intervals and the overall sample size contributing to the estimate. Publication bias—evaluated through funnel plots and Egger’s regression test when 10 or more studies are available for a pooled estimate [29]. Publication bias—evaluated through funnel plots and Egger’s regression test when 10 or more studies are available for a pooled estimate [29]. Two reviewers (NA and VD) will perform GRADE assessments independently. Any disagreements will be resolved through consensus or consultation with a third reviewer (CB). GRADEpro GDT software will be used to generate Summary of Findings tables [30]. These tables will summarise effect estimates, number of studies and participants, and the overall certainty of the evidence (rated as high, moderate, low, or very low). Data synthesis and analysis Observational studies and randomised controlled trials will be analysed in separate analytical streams. No pooled estimates will combine interventional and observational evidence. Meta-analysis will be considered when at least three studies with sufficient clinical and methodological homogeneity are available for a given outcome and exposure definition. This threshold was selected to avoid reliance on a single study or a pairwise comparison while still allowing synthesis in areas where the evidence base may be limited. When only three to four studies contribute to a pooled estimate, between-study variance will be interpreted cautiously, confidence intervals will be emphasised, and findings will be supported by narrative interpretation rather than treated as definitive. Statistical analyses will be performed using MetaXL and/or Stata 19. Random-effects models will be used as the primary approach where studies are sufficiently comparable but clinical or methodological heterogeneity is expected. Quality-effects models will be used as a sensitivity analysis to examine the influence of study quality on pooled estimates. Where heterogeneity is substantial or studies are not sufficiently comparable, results will not be pooled and will instead be synthesised narratively [31, 32]. Effect measures will be synthesised by study design as follows: Cross-sectional and prevalence studies: pooled means or proportions of magnesium levels. Cross-sectional and prevalence studies: pooled means or proportions of magnesium levels. Case–control studies: pooled odds ratios. Case–control studies: pooled odds ratios. Cohort studies: pooled relative risks or hazard ratios. Cohort studies: pooled relative risks or hazard ratios. Randomised controlled trials: pooled intervention effects analysed separately. Randomised controlled trials: pooled intervention effects analysed separately. Heterogeneity will be quantified using the I2 statistic and interpreted alongside clinical and methodological differences between studies. Prespecified subgroup analyses will be conducted by trimester, geographic region, country income classification, laboratory measurement method, comorbidity status, and study design. Meta-regression will be undertaken only if sufficient studies are available, generally at least ten studies per covariate, to reduce the risk of unstable estimates. Where quantitative synthesis is not appropriate, findings will be synthesised narratively using structured narrative synthesis principles. Discussion This systematic review and meta-analysis will synthesise the available evidence on maternal magnesium status and its association with adverse pregnancy outcomes, addressing a long-standing gap in the literature characterised by variable results, heterogeneous measurement approaches, and limited contextual interpretation. By integrating observational and interventional evidence within pre-specified analytic streams, this review aims to clarify the extent to which maternal magnesium levels are associated with outcomes such as pre-eclampsia, gestational diabetes mellitus, preterm birth, and low birth weight, while explicitly acknowledging the methodological and biological complexity underpinning this field. Magnesium plays a well-established role in vascular regulation, inflammatory modulation, and cellular metabolism, processes that are central to placental development and maternal adaptation to pregnancy. The widespread use of magnesium sulphate in the management of severe pre-eclampsia and eclampsia further highlights its clinical relevance. However, translating biological plausibility into population-level evidence has proven challenging. Existing studies differ substantially in terms of exposure assessment, timing of measurement across gestation, laboratory reference ranges, and outcome definitions, limiting direct comparability and contributing to conflicting conclusions. This review is designed to address these challenges by systematically mapping patterns of maternal magnesium status across pregnancy and examining associations within a transparent, design-stratified synthesis framework. A key strength of the proposed review lies in its explicit separation of observational and randomised controlled trial evidence. Observational studies will primarily inform associations between maternal magnesium levels and pregnancy outcomes, supporting hypothesis generation and identification of vulnerable subgroups. In contrast, randomised controlled trials will be synthesised separately to explore the effects of magnesium supplementation, recognising their distinct role in causal inference. This hierarchical approach reflects established principles of evidence synthesis and avoids inappropriate pooling of fundamentally different study designs. The review will place particular emphasis on understanding how comorbid conditions that are common in many low- and middle-income countries, such as HIV infection, anaemia, and obesity, may modify observed associations. These conditions may influence magnesium metabolism, inflammatory pathways, and pregnancy risk independently, and their co-occurrence may amplify susceptibility to adverse outcomes. Rather than restricting the evidence base geographically, this review adopts a global inclusion strategy with prespecified subgroup analyses by region, country income classification, and comorbidity status. This will allow LMIC relevance to be examined analytically rather than assumed a priori, thereby strengthening both internal validity and external applicability. Importantly, while this review will describe trimester-specific distributions and patterns of maternal magnesium status, it does not aim to define clinical diagnostic cut-offs or establish normative reference ranges. Instead, findings will be interpreted in terms of relative patterns, associations, and variability across populations and study designs. Such an approach is appropriate given current heterogeneity in laboratory methods and reference standards, and it avoids overextension of conclusions beyond the available evidence. The findings of this review are expected to inform future research priorities by identifying critical gaps, including under-represented populations, inconsistencies in exposure measurement, and the need for well-designed trials that account for baseline magnesium status and relevant comorbidities. In doing so, the review may contribute to a more nuanced evidence base to support context-sensitive antenatal care strategies, particularly in settings where nutritional deficiencies and pregnancy-related morbidity remain highly prevalent. Limitations Several limitations inherent to the existing evidence base are anticipated and will be considered when interpreting the findings of this review. First, a substantial proportion of included studies are expected to be observational in nature, rendering them susceptible to residual confounding. Maternal magnesium status may act as a proxy for broader dietary quality, socioeconomic status, health-seeking behaviour, or access to antenatal care, factors that are not uniformly measured or adequately adjusted for across studies. As a result, observed associations should not be interpreted as causal in the absence of supporting interventional evidence. Second, heterogeneity in magnesium assessment represents a significant methodological challenge. Studies vary in specimen type (serum, plasma, erythrocyte), assay methodology, timing of sampling across gestation, and reference ranges used to define adequacy or deficiency. These differences may reflect both biological and analytical variability, limiting the comparability of results and constraining the extent to which pooled estimates can be generated. Prespecified subgroup and sensitivity analyses are intended to mitigate, but not eliminate, this limitation. Third, outcome definitions and ascertainment methods may differ across studies and settings, particularly for conditions such as pre-eclampsia and gestational diabetes mellitus, which are subject to evolving diagnostic criteria. Such variability may contribute to heterogeneity and affect the precision of pooled estimates. Fourth, despite the inclusion of grey literature and the absence of language restrictions, publication bias remains a potential concern, particularly if smaller or null studies are underreported. Formal assessments of publication bias will be conducted where sufficient studies are available, but their interpretability may be limited in the presence of substantial heterogeneity. Finally, while subgroup analyses will explore regional and comorbidity-specific patterns, data availability may limit the strength of inferences in certain subpopulations, particularly in low-resource settings where laboratory measurement of magnesium is less routinely performed. Ethics and dissemination No ethical approval is required as this review involves secondary analysis of published data. Findings will be disseminated via peer-reviewed publications and academic conferences. Registration and amendments This protocol has been developed in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) 2015 guidelines. The completed PRISMA-P checklist is provided as a supplementary file. The final systematic review and meta-analysis will be reported in accordance with the PRISMA 2020 statement to ensure transparency and completeness of reporting. The protocol is registered with PROSPERO (CRD42024518427). The PROSPERO record has been updated to reflect the methodological refinements made during revision, including the PICOS framing, separated analytic streams for observational and interventional evidence, design-specific risk-of-bias tools, and planned subgroup analyses. Any further amendments to the protocol will be documented with dates, rationale, and description of changes in both the PROSPERO record and an appendix to the final publication. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements The authors would like to extend thanks to Dr Anushka Ajith for assistance with the search strategy. Abbreviations Human immunodeficiency virus Deoxyribonucleic acid Gestational diabetes mellitus Pregnancy-induced hypertension Intrauterine growth restriction Pre-eclampsia Low-to-middle income countries Preferred Reporting Items for Systematic Review and Meta-analysis for Protocols Directory of Open Access Journals Authors’ contributions NA and VD conceptualised the study and developed the protocol. NA drafted the manuscript. VD and CB critically revised the manuscript. All authors read and approved the final manuscript. Funding This protocol forms part of a Master’s thesis, which received funding for the empirical portion of the study from the National Research Foundation’s Thuthuka Grant (TTK170508230162) in collaboration with the University of KwaZulu-Natal and Medical Research Council of South Africa (SIR Grant UNS14197). The institutions listed have no vested interest in the study. Furthermore, there was no input from these organisations in the interpretation of the result. Declarations Competing interests The authors declare that they have no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3.