Gestational diabetes mellitus and risk of neonatal respiratory distress syndrome: a systematic review and meta-analysis

Abstract

Aim

Gestational Diabetes Mellitus (GDM), a common pregnancy complication characterized by glucose intolerance, is increasingly recognized as a risk factor for Neonatal Respiratory Distress Syndrome (NRDS). This study aimed to systematically review and quantify the association between GDM and NRDS.

Methods

A comprehensive search was conducted in PubMed, Scopus, Embase, and Web of Science from their inception through July 30, 2024, to identify relevant studies. A total of 44 studies, including 50 datasets and over 6.2 million participants, were included in the analysis. Meta-analyses were performed using random-effects models to estimate pooled odds ratios (ORs) and assess heterogeneity among studies. Subgroup analyses were conducted based on study design, gestational age, diagnostic methods, and geographical regions.

Results

Our meta-analysis demonstrated a statistically significant association between GDM and an increased risk of NRDS in newborns (OR 1.9; 95%CI 1.5–2.3). A sub-group analysis based on studies participants showed significant association in both GDM-based (OR, 2.0; 95%CI, 1.5–2.7) and NRSD-based studies (OR, 1.7; 95%CI, 1.3–2.3). This association was consistent across other various subgroups, including both term and preterm pregnancies and across different continents. Sensitivity analysis confirmed the robustness of these findings, and cumulative meta-analysis showed a steady increase in the strength of the association over time.

Conclusion

Our findings highlight GDM as a significant risk factor for NRDS, underscoring the need for early detection and effective management of GDM to reduce adverse neonatal outcomes. However, limitations such as residual confounding, high heterogeneity among studies, and evidence of publication bias should be considered when interpreting these results. Future research should address these issues by including diverse populations and accounting for key confounders to better understand the GDM-NRDS relationship and explore targeted interventions to mitigate the risk in infants born to mothers with GDM.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13098-024-01539-x.

Introduction

Neonatal Respiratory Distress Syndrome (NRDS) is a serious respiratory condition that primarily affects preterm infants; however, term infants are also at risk [1]. It is characterized by insufficient production of surfactant, a substance that prevents the collapse of the lungs’ alveoli, leading to impaired gas exchange and respiratory failure [2]. Globally, NRDS accounts for a substantial proportion of neonatal deaths, particularly in low- and middle-income countries, particularly in Middle East, Sub-Saharan Africa and Asia, where access to advanced neonatal care is limited [1, 3]. The condition’s epidemiological burden is considerable, with an estimated incidence of 1–7% among live births, and it accounts for 30 to 40% of newborn hospital admissions [4, 5]. Risk factors for NRDS include preterm birth, male gender, cesarean delivery, maternal diabetes, and perinatal asphyxia [6]. Among these, maternal diabetes mellitus (gestational or pre-gestational) has been increasingly recognized as a critical risk factor for the development of NRDS in newborns, even in those born at term [7].

Gestational Diabetes Mellitus (GDM) is a form of diabetes first diagnosed during pregnancy, characterized by glucose intolerance and hyperglycemia [8]. It affects a significant percentage of pregnancies worldwide, with prevalence rates ranging from 1% to > 30%, depending on the population studied and the diagnostic criteria used [8]. Regarding WHO regions, the highest prevalence of GDM was estimated for the Eastern Mediterranean and South-East Asia regions (median ∼ 15%), and the lowest prevalence was estimated for European and North America/Caribbean regions (median 6–7%) [8]. GDM poses a considerable health burden, contributing to both maternal and neonatal morbidity [9]. For the mother, GDM increases the risk of hypertensive disorders, cesarean delivery, and the future development of type 2 diabetes [10]. For the infant, GDM is associated with a range of adverse outcomes, including macrosomia, birth trauma, preterm birth, and metabolic complications [11]. Of particular concern is the increased risk of NRDS, which is thought to be due to delayed lung maturation associated with the hyperglycemic environment in utero [7].

The purpose of this systematic review and meta-analysis is to evaluate and synthesize the existing evidence on the association between GDM and the risk of NRDS. By pooling data from multiple studies, this analysis aims to provide a clearer understanding of the extent to which GDM contributes to the development of NRDS, which could inform clinical practices and guidelines for managing pregnancies complicated by GDM.

Methods

This meta-analysis study was done based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines [12]. A Population, Exposure, Comparator, Outcome (PECO) framework was established as follows: The population included pregnant women diagnosed with gestational diabetes mellitus (GDM). The exposure was GDM. The comparison was pregnant women without GDM. The outcome was the incidence of neonatal respiratory distress syndrome (RDS). Based on this framework, the research question was proposed: What is the risk of neonatal respiratory distress syndrome in infants born to mothers with gestational diabetes mellitus compared to those born to mothers without GDM?

Search strategy

At first, we developed a search strategy for PubMed with the assistance of a medical information specialist. The strategy included a combination of Medical Subject Heading (MeSH) terms and keywords related to GDM and RDS. This search strategy was then adapted for use in other databases as needed to ensure comprehensive coverage of relevant studies (Table S1). We used following keywords and linked them using the Boolean operators OR and/or AND: “maternal diabetes”, “gestational diabetes”, “pregnancy diabetes”, “infants of diabetic mothers”, “GDM”, “diabetes”, “respiratory distress syndrome”, “hyaline membrane disease”, “RDS”, “respiratory distress”, “acute respiratory distress syndrome”. Then, we did comprehensive searches on four scientific databases including PubMed, Scopus, Embase, and Web of Science to identify potential eligible studies published from inception through July 30, 2024 without any language, time or geographical restrictions. We further searched the bibliographies of articles that met inclusion criteria as well as relevant reviews or systematic reviews articles and Google Scholar (20 first pages) to identify further eligible studies and gray literature. Google Translate (https://translate.google.com/) was used to assess non-English articles for potential inclusion. EndNote X20.0 software (Thomson Reuters, California, USA) was employed for record management, deduplication, and the initial screening of abstracts and titles.

Study selection criteria

Two researchers (FY and HL) independently assessed all retrieved studies for inclusion, with any discrepancies resolved by a third investigator (CD). The inclusion criteria were as follows: (a) comparative cross-sectional, cohort, or case-control studies; (b) studies where the exposure group consisted of pregnant women with GDM or infants born to women with GDM; (c) studies that included appropriate and representative healthy participants as a control group; (d) studies where the outcome of interest was neonatal RDS; and (e) studies that reported relative risk (RR), odds ratio (OR), or hazard ratio (HR) with 95% confidence intervals (CI) or adequate data to calculate these. We excluded studies if they: (a) had a sample size of fewer than 30 participants; (b) included pregnant women with pre-existing diabetes mellitus (pre-GDM); (c) used datasets duplicated in other studies; (d) focused solely on neonatal RDS without providing extractable data on maternal characteristics; (e) lacked a control group; or (f) did not present original data, such as case reports, case series, literature reviews, or systematic reviews.

Data extraction and quality assessment

Two investigators (FY and HL) independently extracted the following data from each study: the first author’s name, the year of publication, the time period during which the study was conducted, the country in which the study took place, the WHO region, the country’s income level, the study design, the method used to diagnose GDM, the gestational age of pregnant women, the number of cases screened, the number of infants who developed RDS, the adjusted OR, RR or HR with a 95% confidence interval (CI) for RDS risk (if available), and the variables used to adjust the OR. The studies were categorized into two sub-groups based on the type of participants: (a) GDM-based studies, where pregnant women with and without GDM were followed to assess the incidence of RDS in their infants, and (b) NRDS-based studies, where infants with and without RDS were retrospectively evaluated in relation to the health status of their mothers, specifically considering GDM. The quality of the studies included in the analysis was rigorously assessed using the Newcastle–Ottawa Scale (NOS), a tool specifically designed to evaluate the methodological soundness of non-randomized studies. This scale examines various aspects of study design, including the selection of study groups, the comparability of the groups, and the ascertainment of outcomes or exposures. Each study was assigned a score based on these criteria, with a maximum possible score of 9 points. A score of 7 or higher was considered indicative of high methodological quality, reflecting robust design and reliable findings.

Statistical analysis

All analyses were conducted using Stata software (version 17.0; Stata Corporation, College Station, TX, USA). Statistical significance was determined with two-sided P-values, where P ≤ 0.05 was considered significant. The effect size was represented as odds ratios (ORs), and pooled measures were synthesized using inverse-variance random-effects models (REM) [13]. Heterogeneity among studies was evaluated using the I² statistic and χ² test [14]. Meta-regression was performed to explore potential sources of heterogeneity, including publication year, sample size, continent, study design (cohort or case-control), study quality assessment score, and participant type [15]. A separate REM analysis was conducted using only adjusted ORs. Additionally, subgroup analyses were carried out based on continent, study design, WHO region, country income level, and study quality. Influence analysis was used to identify any individual studies that significantly impacted the pooled results. A cumulative meta-analysis was performed to illustrate the trend of results over time by sequentially adding studies based on publication year. Funnel plot asymmetry and publication bias were assessed using Begg’s test [16].

Results

Our comprehensive search across four major databases initially identified 24,021 studies, with an additional 102 studies found through Google Scholar and manual reference list searches. Following preliminary screening and the application of inclusion and exclusion criteria, 44 studies comprising 50 datasets published between 2002 and 2024 were deemed eligible for inclusion in the meta-analysis [17–60]. The study selection process, along with the reasons for excluding certain studies, is illustrated in Fig. 1 using a PRISMA flowchart. The included studies comprised 13 case-control, 9 prospective cohort, and 28 retrospective studies, with a combined total of 220,122 cases and 4,081,397 healthy controls. These studies were conducted across 23 countries, including Australia, Bangladesh, Brazil, China, France, India, Indonesia, Iran, Israel, Italy, Mexico, Poland, Portugal, North Macedonia, Romania, Saudi Arabia, Slovenia, South Korea, Spain, Sweden, Taiwan, the United Kingdom, and the USA. In terms of geographic distribution, the studies spanned five WHO regions, with the exception of the African region. Among the included datasets, 34 were GDM-based and 16 were NRDS-based studies. Based on quality assessment, 44 datasets were classified as high quality and six as moderate quality. The key characteristics of the eligible studies are summarized in Table 1.

Fig. 1.

PRISMA flow chart showing study selection process for studies evaluation the association between gestational diabetes mellitus and risk of neonatal respiratory distress syndrome

Table 1.

Main characteristics of studies included assessing the relationship between gestational diabetes mellitus and neonatal respiratory distress syndrome

Authors	Study period	Country	Study design	Gestational age	Diagnosis method of DM	Number of cases	Number of outcomes	Number of controls	Number of outcomes	Quality assessment
Eastern Mediterranean
Mohammadbeigi et al. (2008)* [17]	2006–2006	Iran	RC	NS	None	70	15	350	18	6
Gasim et al. (2012)* [18]	2001–2008	Saudi Arabia	CC	NS	OGTT	220	3	220	2	7
Almarzouki et al. (2013)* [19]	2008–2008	Saudi Arabia	CC	NS	MR + OGTT	69	15	80	5	8
Kouhkan et al. (2018)* [20]	2014–2017	Iran	PC	NS	OGTT	287	27	287	22	9
Osman et al. (2024)* [21]	2023–2023	Saudi Arabia	CC	NS	MR	113	18	369	31	8
European				NS
Wróblewska-Seniuk et al. (2004)† [22]	1994–2000	Poland	RC	NS	MR	352	19	221	1	7
Fadl et al. (2010)* [23]	1991–2003	Sweden	PC	NS	OGTT	10,525	32	1,249,772	2500	8
Simo˜es et al. (2011)† [24]	1999–2010	Portugal	CC	≥ 24	OGTT	105	30	315	43	8
Guillen et al. (2014)* [25]	1999–2011	Spain	CC	> 24	OGTT	106	20	166	54	8
Kovo et al. (2015)* [26]	2008–2013	Israel	RC	≥ 37	OGTT	41	2	41	3	7
Mortier et al. (2017)* [27]	2011–2012	France	PC	≥ 34	FBS + OGTT	60	12	384	20	8
Billionnet et al. (2017)† [28]	2012–2012	France	RC	≥ 22	MR	57,629	2075	735,519	21,330	8
Billionnet et al. (2017)† [28]	2012–2012	France	RC	≥ 28	MR	57,383	1951	729,105	19,686	8
Billionnet et al. (2017)† [28]	2012–2012	France	RC	≥ 37	MR	52,780	1056	684,398	10,950	8
Bricelj et al. (2017)† [29]	2003–2012	Slovenian	RC	34–37	OGTT	363	11	7400	317	7
Rosen et al. (2017)* [30]	2007–2014	Israel	RC	37–42	OGTT	2236	2	43,876	11	9
Capobianco et al. (2020)* [31]	2017–2018	Italy	CC	NS	OGTT	183	19	207	11	8
Krstevska et al. (2020)* [32]	None	Macedonia	CC	NS	OGTT	50	15	50	4	7
Preda et al., (2021)* [33]	2018–2021	Romania	PC	≥ 24	OGTT	51	16	46	0	7
Monteiro et al. (2022)* [34]	2011–2018	Portugal	RC	NS	OGTT	84	24	166	54	9
Myszkowski et al. (2023)* [35]	2014–2018	Poland	RC	33–37	MR	33	11	177	57	7
Karkia et al. (2023)* [36]	2010–2022	UK	PC	NS	FBS + OGTT	2089	136	49,122	1618	9
Latin America				NS
Freitas et al. (2019)* [37]	2018–2018	Brazil	RC	NS	OGTT	47	1	93	0	6
Violante-Ortíz et al. (2022)* [38]	2009–2020	México	RC	NS	OGCT	342	32	453	20	7
North America				NS
Rauh-hain et al. (2009)* [39]	1998–2006	USA	PC	NS	OGTT	565	16	23,044	206	9
Boghossian et al. (2014)* [40]	2002–2010	USA	RC	≥ 20	MR + ICD-9	2529	114	58,478	1873	9
Kawakita et al. (2017)† [41]	2002–2008	USA	RC	24–42	MR + ICD-9	11,327	453	208,355	6251	9
Werner et al. (2019)* [42]	2010–2015	USA	CC	34–37	MR	306	30	2525	253	10
Battarbee et al. (2020)* [43]	1999–2002	USA	RC	24–42	MR	10,549	491	182,464	6990	8
South-East Asian
Wahi et al. (2010)* [44]	2007–2008	India	PC	24–28	OGTT	132	2	140	0	6
Mannan et al. (2012)* [45]	None	Bangladesh	PC	24–28	OGTT	72	8	72	3	7
Naher et al. (2015)* [46]	2006–2007	Bangladesh	RC	NS	MR	50	29	50	7	5
Kumari et al. (2018)* [47]	2011–2016	India	RC	NS	GTT	170	8	191	3	8
Naik et al. (2019)* [48]	2014–2016	India	PC	< 38	OGTT	424	18	424	7	7
Modak et al. (2023)* [49]	2020–2021	India	RC	24–34	OGTT	34	4	266	0	6
Negara et al. (2024)* [50]	None	Indonesia	RC	NS	MR	39	38	56	45	3
Westren Pacific			RC
Stone et al. (2002)* [51]	1996–1996	Australia	RC	NS	MR	2169	54	58,231	874	9
Liu et al. (2014)† [52]	2008–2010	China	CC	≥ 37	MR	40	20	578	185	8
Feng et al. (2018)* [53]	2011–2015	China	CC	NS	OGTT	996	53	996	20	9
Ye et al. (2018)† [54]	2015–2017	China	CC	NS	MR	320	45	320	15	8
kim et al. (2018)† [55]	2012–2016	South Korea	CC	≥ 37	None	80	2	116	1	7
Xi et al. (2020)† [56]	2017–2019	China	CC	≥ 34	OGTT	101	16	101	13	7
Lin et al. (2021)* [57]	2000–2012	Taiwan	RC	NS	MR + ICD-9	1946	71	27,217	430	8
Harrison et al. (2022)† [58]	2018–2020	Australia	RC	≥ 24	OGTT	2491	240	12,474	995	8
Kim et al. (2022)* [59]	1995–2005	South Korea	RC	≥ 24	OGTT	25	10	1027	121	8
Kim et al. (2022)* [59]	2005–2018	South Korea	RC	≥ 24	OGTT	105	20	607	275	8
Zhang et al. (2023)† [60]	2019–2020	China	RC	25–37	OGTT	217	96	423	131	8
Zhang et al. (2023)† [60]	2019–2020	China	RC	< 32	OGTT	69	61	121	92	8
Zhang et al. (2023)† [60]	2019–2020	China	RC	32–34	OGTT	68	26	125	21	8
Zhang et al. (2023)† [60]	2019–2020	China	RC	34–37	OGTT	80	9	179	18	8

* GDB-based studies; † N-RSD-based studies

Abbreviations: MR, medical redords; FBS, Fasting blood sugar; NS, not specified; CC, case-control; PC, prospective cohort; RC, retrospective cohort; USA, United States; UK, United Kingdom; ICD-9, International Classification of Diseases codes; OGTT, Oral glucose tolerance tests

Results of overall meta-analysis

As illustrated in Fig. 2, the overall meta-analysis identified a statistically significant positive association between GDM and the incidence of NRDS (OR, 1.9; 95% CI, 1.5–2.3; I² = 97.6; p < 0.001). Additional analysis using a REM on adjusted ORs further supported this significant positive association (OR, 1.3; 95% CI, 1.1–1.5; I² = 90.6%; Fig. 3). The assessment of potential “small-study effects” through funnel plots, as shown in Figure S1, revealed asymmetry, suggesting the presence of publication bias. However, Begg’s test did not provide evidence of publication bias (p = 0.13).”

Fig. 2.

Forest plot of overall pooled ORs, pooled with random effects, regarding the association between gestational diabetes mellitus and risk of neonatal respiratory distress syndrome

Fig. 3.

Forest plot of adjusted ORs, pooled with random effects, regarding the association between gestational diabetes mellitus and risk of neonatal respiratory distress syndrome

Sub-group meta-analysis based on GDM- or NRDS-based datasets

With respect to GDM-based studies, our subgroup meta-analysis, which included 34 eligible datasets covering 47,704 pregnant women with GDM and 1,909,261 without GDM, found NRDS incidence of 9.2% (95% CI, 8.1–10.4%; 1786/47,704) and 3.2% (95% CI, 2.8–3.5%; 21,763/1,909,261), respectively (Table 2). The REM showed a significantly higher incidence of NRDS among pregnant women with GDM (OR, 2.0; 95% CI, 1.5–2.7; I² = 94.6%), suggesting a strong positive association between GDM during pregnancy and the occurrence of NRDS.

Table 2.

Sub-group analyses of the pooled odds ratios for the association between gestational diabetes mellitus and neonatal respiratory distress syndrome

Variables	Datasets (n)	Pooled prevalence of outcome in cases % (95% CI)	Pooled prevalence of outcome in controls % (95% CI)	Odds ratios (95% CI)	Heterogeneity (I2%)
					I2
General	50	9.63 (8.76–10.51)	4.69 (4.27–5.11)	1.9 (1.54–2.35)	97.6
Type of studies
GDM-based	34	9.24 (8.06–10.42)	3.2 (2.8–3.5)	2.01 (1.49–2.71)	94.64
NRDS-based	16	10.11 (8.85–11.37)	5.63 (5.01–6.25)	1.74 (1.34–2.27)	97.81
Study design
Case-control	13	11.42 (8.21–14.63)	7.66 (6.01–9.31)	1.7 (1.2–2.4)	83.84
Prospective cohort	9	6.83 (4.09–9.58)	2.32 (1.19–3.46)	2.7 (1.6–4.4)	48.92
Retrospective cohort	28	11.56 (10.36–12.77)	6.34 (5.61–7.08)	1.9 (1.4–2.5)	98.82
Gestational age
Term	19	6.21 (5.09–7.32)	6.24 (5.37–7.11)	1.6 (1.1–2.3)	99.15
Preterm	10	20.71 (14.19–27.24)	12.85 (9.81–15.89)	1.8 (1.4–2.2)	51.46
Not specified	21	14.07 (11.66–16.49)	4.36 (3.48–5.25)	2.38 (1.67–3.41)	89.55
Diagnostic method
OGTT	31	8.86 (7.77–9.95)	2.58 (2.33–2.83)	1.9 (1.4–2.6)	86.81
MR or ICD	19	9.32 (8.14–10.49)	3.90 (3.41–4.39)	1.9 (1.4–2.5)	98.6
Study quality
High	44	7.49 (6.68–8.29)	4.96 (4.51–5.40)	1.71 (1.41–2.08)	97.17
Moderate	6	32.41 (-3.44-68.26)	10.28 (5.91–14.66)	6.80 (3.96–11.69)	0
Publish year
2002–2010	6	9.63 (8.76–10.51)	1.82 (0.64–3.01)	2.92 (1.18–7.22)	78.65
2011–2019	24	6.21 (5.24–7.18)	3.93 (3.37–4.48)	1.87 (1.45–2.40)	97.8
2020–2024	20	21.23 (17.64–24.81)	8.50 (7.51–9.48)	1.85 (1.24–2.75)	95.31
Income levels
High	30	7.05 (6.11–7.98)	5.64 (4.95–6.34)	1.79 (1.31–2.44)	98.99
Lower middle	9	20.23 (8.02–32.45)	5.96 (3.33–8.60)	2.96 (1.98–4.43)	17.61
Upper middle	11	20.59 (15.20-25.99)	12.61 (9.78–15.45)	1.97 (1.49–2.60)	53.55
WHO regions
Western Pacific	5	9.65 (3.75–15.55)	9.01 (3.03–14.99)	1.37 (0.57–3.34)	76.73
South-East Asian	17	5.68 (4.52–6.84)	4.13 (3.24–5.01)	2.07 (1.45–2.97)	98.76
Latin America	2	2.19 (0.68–3.70)	2.17 (0.98–3.36)	2.43 (1.46–4.04)	0
European	5	6.00 (4.57–7.43)	3.58 (3.14–4.03)	1.76 (1.10–2.81)	95.59
Eastern Mediterranean	7	25.72 (8.32–43.11)	10.46 (6.26–14.65)	3.12 (1.81–5.39)	43.78
North America	14	16.81 (14.15–19.47)	9.98 (8.78–11.18)	1.66 (1.07–2.58)	93.68
Type of RDS
RDS	2	6.66 (5.60–7.71)	3.30 (3.15–3.46)	2.78 (1.29–5.95)	74.15
ARDS	48	9.65 (8.76–10.54)	4.71 (4.28–5.14)	1.9 (1.5–2.3)	97.67

Abbreviations: MR, medical redords; GDB, gestational diabetes mellitus; NRDS, neonatal respiratory distress syndrome; ICD-9, International Classification of Diseases codes; OGTT, Oral glucose tolerance tests

In a separate analysis of 16 eligible studies involving 172,418 infants with NRDS and 2,172,136 without NRDS, the pooled prevalence of GDM among their mothers was 10.1% (95% CI, 8.8–11.4%; 5690/172,418) and 5.6% (95% CI, 5.0–6.2%; 53,823/2,172,136), respectively (Table 2). The REM indicated a significantly higher prevalence of GDM in mothers of infants with NRDS (OR, 1.7; 95% CI, 1.3–2.3; I² = 97.8%), further supporting a significant positive association between GDM during pregnancy and the occurrence of NRDS.

Sub-group meta-analyses based on characteristics of studies

Subgroup analyses by study design revealed a significant link between GDM and NRDS in prospective case-control studies (OR, 1.7; 95% CI, 1.2–2.4), prospective cohort studies (OR, 2.7; 95% CI, 1.6–4.4), and retrospective cohort studies (OR, 1.9; 95% CI, 1.4–2.5). When considering gestational age, this association was evident in both term (OR, 1.6; 95% CI, 1.1–2.3) and preterm (OR, 1.8; 95% CI, 1.4–2.2) pregnancies. In terms of diagnostic methods, studies utilizing oral glucose tolerance test (OGTT) (OR, 1.9; 95% CI, 1.4–2.6) and those relying on medical records or ICD codes (OR, 1.9; 95% CI, 1.4–2.5) both demonstrated significant associations. This association persisted across all subgroups when stratified by study quality, publication year, country income levels, and WHO regions, with the exception of the Western Pacific region (refer to Table 2). Additionally, two studies assessed the relationship between GDM and the risk of acute respiratory distress syndrome (ARDS) in neonates. A REM applied to these studies indicated a significant association between GDM and ARDS (OR, 1.9; 95% CI, 1.5–2.3).

Sensitivity and cumulative analysis

We performed a sensitivity analysis to determine whether the overall pooled effect sizes were influenced by any single study. The analysis revealed that the removal of any individual study did not significantly impact the overall pooled odds ratio (OR), confirming the robustness of our results (Figure S2). Additionally, a cumulative meta-analysis indicated a steadily increasing association that maintained statistical significance as more studies were added over time (Figure S3).

Discussion

This comprehensive systematic review and meta-analysis, provides robust evidence supporting a significant positive association between GDM and the development of NRDS. The pooled analysis of 44 studies from 23 countries, encompassing over 2.2 million participants, consistently demonstrated a heightened risk of NRDS in infants born to mothers with GDM, regardless of study design, gestational age, diagnostic method, or geographic location. This finding is consistent with previous evidence and reinforces the importance of recognizing GDM as a critical risk factor for NRDS. Our findings are consistent with a previous meta-analysis [7] which indicated that both GDM (OR, 1.57) and pre-existing diabetes mellitus (pre-GDM) (OR, 2.6) are significantly associated with an increased risk of NRDS. However, our meta-analysis, encompassing a significantly larger number of eligible studies and a sample size approximately three times greater than the previous study, yielded a pooled odds ratio for GDM that was approximately 26% stranger (ORs, 1.9 vs. 1.5). Furthermore, our findings are in line with previous evidence indicating that GDM had adverse effects on pregnancy and neonatal outcomes [61, 62]. Previous research has consistently demonstrated that GDM is associated with a range of adverse maternal and neonatal complications beyond NRDS [61, 63]. For instance, GDM increases the risk of hypertensive disorders, such as preeclampsia, which can lead to premature birth and other complications for both mother and baby [61]. GDM is also linked to an increased risk of cesarean delivery and the future development of type 2 diabetes in the mother [61, 64]. In infants, GDM is associated with macrosomia (large birth weight), birth trauma, preterm birth, and metabolic complications, such as hypoglycemia and hyperbilirubinemia [61, 64].

The underlying mechanisms by which GDM contributes to NRDS are complex and multifaceted, involving a combination of factors related to the hyperglycemic environment in utero [65]. One key mechanism is the delay in lung maturation associated with maternal hyperglycemia [66]. Elevated glucose levels in the fetal circulation can disrupt the normal developmental processes of the lungs, leading to reduced surfactant production. Surfactant, a lipoprotein complex essential for maintaining alveolar stability and preventing collapse, is crucial for proper gas exchange [65]. Insufficient surfactant production in infants born to mothers with GDM can lead to alveolar collapse and the development of NRDS. Additionally, hyperglycemia can also induce oxidative stress and inflammation in the fetal lungs, further contributing to impaired lung development and an increased risk of NRDS [67]. Furthermore, GDM can lead to placental insufficiency, which can restrict oxygen and nutrient delivery to the fetus, further compromising lung development and increasing the susceptibility to NRDS [68, 69]. A deeper understanding of these intricate mechanisms is crucial for developing targeted interventions to mitigate the risk of NRDS in infants born to mothers with GDM.

The strength of this meta-analysis lies in its extensive number of studies, large sample size, and comprehensive search strategy, which includes multiple databases and gray literature, ensuring a thorough and unbiased evaluation of the available evidence. The inclusion of a diverse range of studies, including both GDM-based and NRDS-based designs, strengthens the generalizability of the findings. Furthermore, the rigorous quality assessment using the Newcastle-Ottawa Scale, with the majority of studies classified as high quality, enhances the reliability of the results. However, this meta-analysis also has notable limitations. Firstly, the reliance on observational studies inherently limits the ability to establish causal relationships. Despite efforts to adjust for potential confounders, residual confounding cannot be entirely ruled out. Key confounding variables, such as maternal BMI, socioeconomic status, and concurrent pregnancy complications, were not consistently accounted for across all included studies, which could influence both the risk of GDM and NRDS. Secondly, substantial heterogeneity was observed among the included studies, likely reflecting variations in study populations, diagnostic criteria, outcome definitions, and regional practices. Although meta-regression and subgroup analyses were performed to explore the sources of heterogeneity, these methods may not fully resolve the issue, which could impact the reliability and generalizability of the pooled estimates. Thirdly, evidence of publication bias, as suggested by asymmetry in the funnel plots, indicates the possibility that studies reporting stronger associations may have been more likely to be published. This bias could potentially skew the overall results toward a stronger association between GDM and NRDS. Finally, the geographic scope of the analysis was somewhat limited, as no studies from the African region were included, and some WHO regions were underrepresented. This restricts the global applicability of the findings. Given these limitations, the results of this meta-analysis should be interpreted with caution. Future research should aim to include more diverse populations and ensure consistent adjustment for confounders such as maternal BMI and socioeconomic status to strengthen the evidence base.

This comprehensive systematic review and meta-analysis provides compelling evidence for a significant positive association between GDM and the risk of NRDS. The robust findings, consistent across diverse study designs and geographic locations, underscore the importance of GDM as a critical risk factor for NRDS development. This highlights the need for early detection and effective management of GDM during pregnancy to reduce adverse neonatal outcomes, particularly NRDS. Our findings support targeted management strategies for GDM during pregnancy to potentially mitigate NRDS risk, especially in high-risk populations. While the meta-analysis provides valuable insights, future research is warranted to further elucidate the mechanisms linking GDM to NRDS. Well-controlled longitudinal cohort studies focusing on diverse populations are particularly needed to address residual confounding and better evaluate causality. Additionally, studies from underrepresented regions, such as Africa, should be prioritized, and strategies to mitigate the impact of heterogeneity and publication bias in future meta-analyses should be explored. By deepening our understanding of the GDM-NRDS relationship and addressing these research gaps, we can implement evidence-based strategies to improve neonatal health and reduce the burden of this serious respiratory condition globally.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Cuixia Ding conceived and designed this study. Fang Yang and Hua Liu searched databases and collected data from included studies. Fang Yang and Hua Liu performed the statistical analyses. Fang Yang and Cuixia Ding wrote, revised and confirmed the original draft of the manuscript. All authors reviewed and approved the final version before submission. Two first authors are equally contributed in this study.

Funding

There is no special funding for. The corresponding author had full access to all study data and final responsibility for the decision to submit for publication.

Data availability

We included all data in main manuscript and supplementary files. Further data that supports the findings of this study are available from the corresponding authors upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.