A meta-analysis of metformin and insulin on maternal outcome and neonatal outcome in patients with gestational diabetes mellitus

Abstract

Introduction

The use of metformin for treating gestational diabetes mellitus (GDM) remains controversial because it can pass through the placenta. This meta-analysis aimed to compare the effects of metformin and insulin on maternal and neonatal outcomes in patients with GDM.

Methods

We conducted a comprehensive search of the PubMed, Embase, and Cochrane Library databases, focusing on randomized controlled trials (RCTs) that evaluated the impacts of metformin and insulin on both maternal and neonatal outcomes in patients with GDM.

Results

Twenty-four RCTs involving 4934 patients with GDM were included in this meta-analysis. Compared with insulin, metformin demonstrated a significant reduction in the risks of preeclampsia (RR 0.61, 95% CI 0.48 to 0.78, p < .0001), induction of labor (RR 0.90, 95% CI 0.82 to 0.98, p = .02), cesarean delivery (RR 0.91, 95% CI 0.85 to 0.98, p = .01), macrosomia (RR 0.67, 95% CI 0.53 to 0.83, p = .0004), neonatal intensive care unit (NICU) admission (RR 0.75, 95% CI 0.66 to 0.86, p < .0001), neonatal hypoglycemia (RR 0.55, 95% CI 0.48 to 0.63, p < .00001), and large for gestational age (LGA) (RR 0.80, 95% CI 0.68 to 0.94, p = .007). Conversely, metformin showed no significant impact on gestational hypertension (RR 0.84, 95% CI 0.67 to 1.06, p = .15), spontaneous vaginal delivery (RR 1.13, 95% CI 1.00 to 1.08, p = .05), emergency cesarean section (RR 0.94, 95% CI 0.77 to 1.16, p = .58), shoulder dystocia (RR 0.65, 95% CI 0.31 to 1.39, p = .27), premature birth (RR 0. 92, 95% CI 0.61 to 1.39, p = .69), polyhydramnios (RR 1.11, 95% CI 0.54 to 2.30, p = .77), birth trauma (RR 0.87, 95% CI 0.54 to 1.39, p = .56), 5-min Apgar score < 7 (RR 1.13, 95% CI 0.76 to 1.68, p = .55), small for gestational age (SGA) (RR 0.93, 95% CI 0.71 to 1.22, p = .62), respiratory distress syndrome (RDS) (RR 0.74, 95% CI 0.50 to 1.08, p = .11), jaundice (RR 1.09, 95% CI 0.95 to 1.25, p = .24) or birth defects (RR 0.80, 95% CI 0.37 to 1.74, p = .57).

Conclusions

The findings suggest that metformin can reduce the risk of certain maternal and neonatal outcomes compared with insulin therapy for GDM. However, long-term follow-up studies of patients with GDM taking metformin and their offspring are warranted to provide further evidence.

Keywords:

• Metformin
• insulin
• gestational diabetes
• meta-analysis
• randomized controlled trials

Introduction

Gestational diabetes mellitus (GDM) is the predominant metabolic complication during pregnancy and affects 12%–18% of all pregnancies [1]. Various determinants, including genetic, epigenetic, and environmental factors, have been implicated in the onset of GDM [2]. Suboptimal glycemic management during this critical period can culminate in a spectrum of unfavorable outcomes for both mothers and newborns, including maternal preeclampsia, hypertension, neonatal macrosomia, and hypoglycemia [3].

Therefore, the cornerstone of GDM management is glycemic control. Lifestyle interventions, including medical nutrition therapy and daily exercise, serve as the primary approach for GDM treatment. When lifestyle modifications alone do not achieve glycemic targets, the introduction of medications is considered [4].

Although the degree of insulin resistance is increasing throughout the course of pregnancy for patients with GDM, insulin remains a widely accepted and safe option [5] because it does not cross the placental barrier, thereby ensuring fetal safety [6]. However, there may be barriers to insulin treatment during pregnancy, such as cost, availability, fear of needles and hypoglycemia, and patient preferences. In these instances, metformin can serve as an alternative solution [7].

Metformin is a drug that plays a role in inhibiting hepatic gluconeogenesis and increasing peripheral insulin sensitivity. A landmark randomized controlled trial (RCT) in 2008 [8] demonstrated that metformin did not cause more perinatal complications. Notably, nearly half of the patients with GDM required additional insulin therapy to meet pregnancy targets. Therefore, its utilization remains a subject of debate [9]. Although the Society for Maternal-Fetal Medicine (SMFM) supports the application of metformin in the management of blood glucose levels in GDM [10], some studies have emphasized the caution of its use and the necessity of further research [11].

As interest in metformin as a GDM treatment grows, a comprehensive analysis comparing metformin and insulin outcomes in patients with GDM is essential. While prior meta-analyses have compared metformin and insulin use during pregnancy [12], these analyses were constrained by the limited number of RCTs and mixed results for patients with type 2 diabetes. In light of the publication of numerous new RCTs, we performed a meta-analysis using the most recent data to compare outcomes of metformin versus insulin in patients with GDM.

Materials and methods

This meta-analysis, adhering to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [13], was registered at the International Prospective Register of Systematic Reviews (number CRD42022341314).

Literature search

We searched PubMed, Embase, and the Cochrane Library from inception to June 2022. The following keywords were used: “gestational diabetes mellitus,” “insulin,” “metformin,” and “randomized controlled trials.” The detailed search strategy is provided in Appendix 1.

Inclusion criteria

This meta-analysis included the following studies: (1) they were RCTs published in full in English; (2) the criteria for diagnosing GDM included the American Diabetes Association, the World Health Organization, the Carpenter and Coustan guidelines, and the Australian Pregnancy Association consensus guidelines; and (3) the interventions included metformin alone (with supplemental insulin when required) or solely insulin.

Exclusion criteria

This meta-analysis excluded the following studies: (1) ongoing studies; (2) duplicate publications; (3) review articles, case reports, conference abstracts and guidelines; and (4) studies that included patients with type 2 diabetes.

Data extraction and outcome measures

The following information was extracted for the included RCTs: first author, year, country, sample size, number of individuals requiring insulin supplementation, and gestational weeks. Maternal outcomes included preeclampsia, induction of labor, cesarean delivery, gestational hypertension, spontaneous vaginal delivery, emergency cesarean section, shoulder dystocia, premature birth, polyhydramnios, and birth trauma. On the other hand, neonatal outcomes included macrosomia, neonatal hypoglycemia, neonatal intensive care unit (NICU) admission, respiratory distress syndrome (RDS), macrosomia, large for gestational age (LGA), small for gestational age (SGA), 5-min Apgar score <7, birth defects and jaundice.

Quality assessment of individual studies

Two independent investigators critically assessed the risk of bias for the included studies using the Cochrane Collaboration’s risk of bias tool [14]. In conjunction with the guidelines from the Cochrane Handbook for Systematic Reviews of Interventions version 6.3, the quality evaluation covered aspects such as selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases. Based on the evaluation, results were categorized into low risk, high risk, or unclear risk.

Statistical analysis

Dichotomous data were expressed as relative risk ratios (RRs) and 95% confidence intervals (CIs). An I² value greater than 50% indicated significant heterogeneity. Fixed-effects models were used for I² < 50%, and random-effects models were used for I² > 50% [15]. Subgroup analysis was performed on whether insulin was supplemented in the metformin group. Publication bias in outcome evaluation was reported only when the number of included studies was adequate (≥10). A p value of < .05 in a two-tailed test was considered statistically significant, and all the statistical calculations were performed using STATA version 16.0 (StataCorp, College Station, TX, USA) and Review Manager version 5.4 (Cochrane Collaboration, Software Update, Oxford, UK).

Results

Literature search and study characteristics

The initial search included 274 records. After removing duplicates, there were 219 records. After screening the titles and abstracts, 184 studies were excluded. We read the full texts of these articles, and 24 RCTs [8,16–38] were ultimately included in the quantitative analysis.

Figure 1 depicts the search process, and Table 1 lists the basic characteristics of the eligible RCTs. In total, 4934 patients with GDM were included in our quantitative analyses, with 2409 in the metformin group (experimental arm) and 2525 in the insulin group (control arm). The 24 trials included in this review were conducted across twelve countries from 2008 to 2022. Specifically, they were distributed as follows: Egypt (4 trials), Pakistan (5 trials), Iran (3 trials), Finland (4 trials), and one trial each in Ireland, Macedonia, Spain, Brazil, Eldorgua, the United States, New Zealand, and India. Notably, in 14 of these studies [8,16–21,23–25,31,33,36,37], insulin was supplemented in the metformin group due to inadequate glycemic control.

Figure 1. Flow chart for the selection of relevant studies.

Table 1. Basic characteristics of the eligible studies.

Study	Year	Country	Gestational weeks	Metformin group	Insulin required	Insulin group
Rowan JA	2008	New Zealand	20–33	363	168	370
Ijäs H	2011	Finland	12–34	47	15	50
Ainuddin J	2012	Pakistan	20–35	75	18	75
Niromanesh S	2012	Iran	20–34	80	11	80
Mesdaghinia E	2013	Iran	24–34	100	22	100
Spaulonci CP	2013	Brazil	NA	46	12	46
Tertti K	2013	Finland	22–34	110	23	107
Ruholamin S	2014	Iran	24–33	50	0	50
Tertti K	2014	Finland	22–34	58	23	107
Ainuddin J	2015	Pakistan	20–36	75	32	75
Ashoush S	2016	Egypt	26–32	48	11	47
Saleh HS	2016	America	26–34	67	0	70
Arshad R	2017	Pakistan	NA	25	0	25
Khan RMA	2017	Pakistan	24–32	385	0	385
Eid SR	2018	Egypt	22–30	113	0	116
Simeonova-Krstevska S	2018	Macedonia	23–35	48	0	101
Galal M	2019	Egypt	28–34	52	4	54
Ghomian N	2019	Ireland	24–28	143	0	143
Huhtala M	2020	Finland	30–36	110	23	107
Mansour A	2021	Egypt	24–28	70	0	70
Mir S	2021	Pakistan	20–28	120	0	120
Picón-César	2021	Spain	14–35	94	4	97
Dasari P	2022	India	NA	80	13	80
Saraswat D	2022	Ecuador	24–28	50	0	50

NA: not applicable.

Maternal outcomes

Preeclampsia

Twelve studies [8,18,20,21,24,26,28–30,33,34,38] involving 3093 patients with GDM included preeclampsia (Figure 2). The pooled results from the included studies revealed a reduction in the incidence of preeclampsia in the metformin group compared to the insulin group (RR 0.61, 95% CI 0.48 to 0.78, p < .0001), with some heterogeneity among the studies (I² = 48%, heterogeneity p = .03).

Figure 2. Pooled analysis of preeclampsia.

Induction of labor

Data on the induction of labor were obtained from 12 studies [8,16,17,21,24,26,27,29,33,36–38] with 2430 patients with GDM (Figure 3). With an I² of 23% (heterogeneity p = .21) indicating some heterogeneity, our analysis suggested a decreased probability of induction of labor in the metformin group (RR 0.90, 95% CI 0.82 to 0.98, p = .02).

Figure 3. Pooled analysis of induction of labor.

Cesarean delivery

A total of 20 studies [8,16–18,20–22,24–31,33,34,36–38] included 4039 patients with GDM and provided data on cesarean delivery (Figure 4). The heterogeneity was low (I² = 42%, heterogeneity p = .03), and the risk of cesarean delivery with metformin was lower than with insulin (RR 0.91, 95% CI 0.85 to 0.98, p = .01).

Figure 4. Pooled analysis of cesarean delivery.

Other maternal outcomes

Additionally, other maternal outcomes were collected, as shown in Table 2. Eleven studies [8,18,20,21,24,25,28,29,33,36,37] reported gestational hypertension, fifteen studies [16,17,21,24,26–31,33,34,36–38] reported spontaneous vaginal delivery, five studies [8,18,26,29,37] reported emergency cesarean section, four studies [8,18,19,22] reported shoulder dystocia, fifteen studies [8,18–21,25,26,28–32,36–38] reported premature birth, two studies [26,29] reported polyhydramnios, and four studies [8,24,32,36] reported birth trauma.

Table 2. Summary of other outcomes.

Outcomes	Metformin group (n/N)	Control group (n/N)	RR (95% CI)	p
Maternal outcomes
Gestational hypertension	118/1502	141/1511	0.84(0.67, 1.06)	.15
Spontaneous vaginal delivery	750/1398	714/1461	1.13(1.00, 1.28)	.05
Emergency cesarean section	142/703	153/716	0.94(0.77, 1.16)	.58
Shoulder dystocia	10/593	16/600	0.65(0.31, 1.39)	.27
Premature birth	152/1807	188/1877	0.92(0.61, 1.39)	.69
Polyhydramnios	14/180	13/186	1.11(0.54, 2.30)	.77
Birth trauma	30/675	35/685	0.87(0.54, 1.39)	.56
Neonatal outcomes
5-min Apgar score < 7	45/812	40/825	1.13(0.76, 1.68)	.55
SGA	91/1196	105/1267	0.93(0.71, 1.22)	.62
RDS	43/1133	59/1149	0.74(0.50, 1.08)	.11
Jaundice	284/1299	267/1358	1.09(0.95, 1.25)	.24
Birth defects	11/311	14/317	0.80(0.37, 1.74)	.57

SGA: small for gestational age; RDS: respiratory distress syndrome.

Neonatal outcomes

Macrosomia

Thirteen studies [16–21,25,28,29,33,34,36,38] reported on the risk of macrosomia (Figure 5), and the results showed a lower risk of macrosomia in the metformin group (RR 0.67, 95% CI 0.53 to 0.83, p = .0004), with no heterogeneity (I² = 0%, heterogeneity p = .53).

Figure 5. Pooled analysis of macrosomia.

NICU

Seventeen studies [8,16–19,21–26,29,32–34,36,37], focused on NICU outcomes (Figure 6). This analysis revealed a lower risk of NICU admission in the metformin group (RR 0.75, 95% CI 0.66 to 0.86, p < .0001), and the heterogeneity was low (I² = 6%, heterogeneity p = .39).

Figure 6. Pooled analysis of NICU.

Neonatal hypoglycemia

In total, 19 studies [16–26,28,29,32,34–38] on neonatal hypoglycemia were conducted (Figure 7). A lower risk of neonatal hypoglycemia was determined in the metformin group (RR 0.55, 95% CI 0.48 to 0.63, p < .00001) based on the low heterogeneity (I² = 6%, heterogeneity p = .38).

Figure 7. Pooled analysis of neonatal hypoglycemia.

LGA

LGA outcomes were explored in 13 studies [8,16,18–21,24,26,29,30,33,36,37] (Figure 8). The pooled results from the included studies suggested a decreased risk of LGA in the metformin group (RR 0.80, 95% CI 0.68 to 0.94, p = .007), with no significant heterogeneity among the studies (I² = 0%, heterogeneity p = .81).

Figure 8. Pooled analysis of LGA.

Other neonatal outcomes

Further examination revealed additional neonatal outcomes, as detailed in Table 2. Six studies [8,18,20,26,29,32] reported a 5-min Apgar score < 7, twelve studies [8,17,18,20,22,24,26,29,30,33,36,37] reported SGA, eleven studies [8,17–20,22,24,26,29,35,36] reported RDS, 13 studies [16–24,26,28,29,36,] reported jaundice, and four studies [18,26,34,36] reported birth defects.

Subgroup analysis

According to our subgroup analysis (Table 3), several outcomes demonstrated statistical significance, suggesting that metformin, either alone or in combination with insulin, may have positive effects on certain maternal and neonatal outcomes. Notably, the risk of preeclampsia was significantly lower in the metformin-alone group than in the insulin group, with an RR of 0.58 (95% CI 0.42 to 0.80, p = .001). In addition, the incidence of neonatal hypoglycemia was significantly lower in both metformin groups, with RRs of 0.68 (95% CI 0.53 to 0.88, p = .004) for the combined group and 0.59 (95% CI 0.50 to 0.69, p < .001) for the metformin alone group. The risk of macrosomia also significantly decreased in the metformin-alone group, with an RR of 0.65 (95% CI 0.48 to 0.88, p = .006). These statistically significant findings highlight the potential value of metformin in managing maternal complications and improving neonatal outcomes, particularly in reducing the risks of preeclampsia, neonatal hypoglycemia, and macrosomia. However, these insights necessitate further validation and understanding through additional prospective studies and long-term follow-ups to comprehend their enduring health impacts.

Table 3. Subgroup analysis.

Outcomes	Subgroup	Number of studies	RR (95% CI)	p
Maternal outcomes
Preeclampsia	Metformin plus insulin vs insulin	6	0.70 (0.49, 1.02)	.06
	Metformin alone vs insulin	6	0.58 (0.42, 0.80)	.001
Induction of labor	Metformin plus insulin vs insulin	8	0.93 (0.83, 1.04)	.23
	Metformin alone vs insulin	4	0.91 (0.69, 1.20)	.53
Cesarean delivery	Metformin plus insulin vs insulin	13	0.90 (0.81, 1.00)	.06
	Metformin alone vs insulin	7	0.99 (0.87, 1.13)	.91
Gestational hypertension	Metformin plus insulin vs insulin	9	0.83 (0.63, 1.09)	.18
	Metformin alone vs insulin	2	0.89 (0.57, 1.37)	.6
Spontaneous vaginal delivery	Metformin plus insulin vs insulin	8	1.18 (0.95, 1.47)	.13
	Metformin alone vs insulin	7	1.04 (0.92, 1.18)	.5
Emergency cesarean section	Metformin plus insulin vs insulin	3	0.95 (0.74, 1.22)	.7
	Metformin alone vs insulin	2	0.95 (0.65, 1.38)	.8
Shoulder dystocia	Metformin plus insulin vs insulin	3	0.69 (0.31, 1.50)	.35
	Metformin alone vs insulin	1	0.34 (0.01, 8.15)	.5
Premature birth	Metformin plus insulin vs insulin	9	1.23 (0.88, 1.72)	.22
	Metformin alone vs insulin	6	0.68 (0.31, 1.48)	.33
Birth trauma	Metformin plus insulin vs insulin	3	0.76 (0.41, 1.40)	.38
	Metformin alone vs insulin	1	1.07 (0.50, 2.28)	.84
Neonatal outcomes
Macrosomia	Metformin plus insulin vs insulin	9	0.76 (0.53, 1.09)	.13
	Metformin alone vs insulin	4	0.65 (0.48, 0.88)	.006
NICU	Metformin plus insulin vs insulin	12	0.78 (0.66, 0.90)	.002
	Metformin alone vs insulin	5	0.83 (0.61, 1.14)	.25
Neonatal hypoglycemia	Metformin plus insulin vs insulin	11	0.68 (0.53, 0.88)	.004
	Metformin alone vs insulin	8	0.59 (0.50, 0.69)	0
LGA	Metformin plus insulin vs insulin	10	0.84 (0.70, 1.00)	.06
	Metformin alone vs insulin	3	0.76 (0.49, 1.19)	.24
5-min Apgar score < 7	Metformin plus insulin vs insulin	3	1.09 (0.63, 1.90)	.74
	Metformin alone vs insulin	3	1.14 (0.63, 2.05)	.65
SGA	Metformin plus insulin vs insulin	8	0.98 (0.72, 1.32)	.89
	Metformin alone vs insulin	4	0.77 (0.39, 1.54)	.47
RDS	Metformin plus insulin vs insulin	7	0.79 (0.51, 1.23)	.3
	Metformin alone vs insulin	4	0.62 (0.27, 1.39)	0.25
Jaundice	Metformin plus insulin vs insulin	9	0.76 (0.55, 1.03)	.08
	Metformin alone vs insulin	4	1.16 (0.98, 1.37)	.07
Birth defects	Metformin plus insulin vs insulin	2	0.92 (0.38, 2.23)	.85
	Metformin alone vs insulin	2	0.52 (0.09, 2.78)	.44

NICU: neonatal intensive care unit admission; SGA: small for gestational age; LGA: large for gestational age; RDS: respiratory distress syndrome.

Publication bias

Egger’s test was performed to evaluate publication bias in this meta-analysis, and the results are presented in Table 4.

Table 4. The Egger's tests of the results in this meta-analysis.

Outcomes	Egger’s tests (p value)	Publication bias
Gestational hypertension	.149	(−)
Preeclampsia	.415	(−)
Induced labor	.764	(−)
Spontaneous vaginal delivery	.060	(−)
Cesarean section	.355	(−)
Preterm birth	.444	(−)
Macrosomia	.315	(−)
NICU	.330	(−)
Neonatal hypoglycemia	.701	(−)
SGA	.261	(−)
LGA	.150	(−)
RDS	.720	(−)
Jaundice	.006	(+)

NICU: neonatal intensive care unit admission; SGA: small for gestational age; LGA: large for gestational age; RDS: respiratory distress syndrome.

Assessment of study quality

For selection bias, 12 studies that provided clear details about the generation of random sequences were assessed as low risk. The remaining 12 studies did not specify their methodology, resulting in an unclear risk of bias. Concerning allocation concealment, seven studies explicitly described the process and were rated as low risk, while 17 studies did not provide this information, leading to an unclear risk of bias. In terms of performance bias, seven studies were open-label experiments, indicating a high risk of bias. Conversely, three studies mentioned participant blinding and were categorized as having a low risk of bias. Thirteen studies did not address performance bias and were considered to have unclear risks. Regarding detection bias, three studies reported that assessors were blinded, indicating a lower risk of bias, while 21 studies did not explicitly mention detection bias, resulting in an unclear risk. For attrition bias and reporting bias, all studies were rated as low risk, while for other biases, they were categorized as having an unclear risk. The bias risk graph (Figure 9) and bias risk summary (Figure 10) provide visual representations. In conclusion, the quality of the included studies met the criteria set out in the Cochrane Handbook for Systematic Reviews of Interventions Version 6.3.

Figure 9. Risk of bias graph.

Figure 10. Risk of bias summary.

Discussion

In the meta-analysis, we identified important maternal and neonatal outcomes associated with metformin treatment for GDM. Regarding maternal outcomes, metformin significantly reduced the risk of preeclampsia, induction of labor, and cesarean delivery. However, it had no significant impact on the risk of gestational hypertension, spontaneous vaginal delivery, emergency cesarean section, shoulder dystocia, premature birth, polyhydramnios, or birth trauma. Concerning neonatal outcomes, neonates of mothers treated with metformin significantly reduced the risk of macrosomia, NICU admission, neonatal hypoglycemia and LGA. However, there were no significant differences in 5-min Apgar score <7, SGA, RDS, jaundice or birth defects.

Compared with the previous literature, we found that some of the outcomes were different. In their 2015 review, Kitwitee and colleagues [39] presented the main findings that metformin has a favorable effect on neonatal hypoglycemia and NICU admission, and these outcomes were similar to ours. However, they found that metformin did not reduce the outcomes of preeclampsia, cesarean delivery, LGA, or macrosomia, which is completely contrary to our findings. A salient point from their work is that 14-46% of those on metformin still required insulin supplementation, underscoring that while metformin has its advantages, clinicians must be astute in monitoring glucose levels and supplementing with insulin when necessary. The 2019 review by Lanlan Guo et al. [40] yielded findings that aligned partially with our findings. Likewise, metformin significantly reduced the risk of preeclampsia, NICU admission, and neonatal hypoglycemia. However, outcomes such as cesarean delivery and LGA are contrary to our findings. They also highlighted that while metformin exhibits efficacy, there is still a pressing need to monitor the long-term outcomes in the offspring of GDM patients treated with this drug, indicating a call for further research. Similarly, a recent analysis by Bo Sheng [41] supported the positive effects of metformin, focusing only on short-term neonatal outcomes. Metformin reduced the risk of macrosomia, NICU admission, and neonatal hypoglycemia. However, the LGA outcomes differed from our findings. In summary, while metformin shows promise for short-term GDM treatment, future studies should prioritize investigating its long-term effects.

In our meta-analysis, we focused exclusively on patients with GDM, in contrast to the meta-analysis by He et al. [12], which included patients with both GDM and type 2 diabetes. While their meta-analysis also indicated that metformin significantly reduced the incidence of neonatal hypoglycemia and NICU admission, the results related to gestational age were significantly different from our findings. Specifically, their results showed that metformin decreased the risk of LGA but increased the risk of SGA. This is most likely due to the inclusion of patients with type 2 diabetes. In addition, our meta-analysis found no statistical significance for gestational hypertension or neonatal jaundice. In contrast, Li and colleagues [42] found that the risk of gestational hypertension and neonatal jaundice was significantly increased in the insulin group. Given these outcomes, metformin has emerged as a promising alternative to insulin in clinical scenarios.

Furthermore, while our analysis focused primarily on GDM outcomes, it is crucial to consider the effects of metformin in various contexts. For example, the PregMet study [43], primarily for polycystic ovary syndrome (PCOS), offers relevant long-term insights. This study tracked children exposed to metformin over 5–10 years and found significantly higher body mass index (BMI) compared to those exposed to a placebo. These findings emphasize the complexity of the long-term effects of metformin on offspring. It is essential to cautiously approach metformin prescriptions during pregnancy due to potential long-term implications. As we gain more understanding and long-term data, clinical guidelines may need adjustment.

About recent developments in GDM treatment, the EMERGE study [44] presented new perspectives. These findings contribute significantly to the ongoing discussion about GDM management strategies. This study evaluated the early treatment of metformin, contrasting its effects with a placebo. Although there was no significant difference in the primary composite outcome of insulin initiation or fasting glucose levels at specified gestational weeks, these findings offer valuable insight into the broader spectrum of metformin’s application in GDM management.

In the context of managing GDM, one paramount concern is the potential impact of medications on neonatal outcomes, including birth weight. The SGA represents a critical neonatal outcome that has been associated with various perinatal complications. Our meta-analysis explored the association between metformin treatment and the occurrence of SGA. Based on the pooled data, the risk of SGA with metformin was comparable to that with insulin. These findings suggest that metformin, whether used alone or in combination with insulin, does not significantly alter the risk of delivering an SGA infant in comparison to insulin treatment alone. It is also worth noting the broader context. The use of metformin in GDM management is often favored due to its oral administration, reduced hypoglycemic risk compared to insulin, and potential benefits in preventing the progression to type 2 diabetes postpartum. Our current data indicating no significant increase in SGA risk provide further evidence supporting the safety of metformin in the context of GDM. However, it is crucial to approach this issue with caution. While our meta-analysis did not reveal a statistically significant difference in SGA risk, individual patient factors, severity of GDM, or timing of metformin initiation can all play roles. Clinicians should consider these findings in the broader context of comprehensive GDM management, taking into account both maternal and neonatal outcomes.

We recognize that there is a distinct difference in disease severity between patients with GDM treated with metformin alone and those requiring insulin supplementation. This distinction is critical, as it reflects the varying treatment needs based on disease severity. In interpreting our findings, it is important to recognize that comparing these two groups may not provide direct equivalence in risk levels.

We adopted a rigorous search strategy for the included RCTs involving strict screening, quality assessment, and data extraction to ensure the authenticity of the study results. We sought to evaluate the effects of metformin and insulin in the treatment of patients with GDM from an objective and comprehensive perspective. Our meta-analysis suggested that metformin may be a safer and more effective agent for treating GDM and has the potential to replace insulin in future clinical practice. It can reduce some risk events for maternal and neonatal outcomes.

However, there are several limitations in our meta-analysis. First, our meta-analysis was limited by the number and quality of the included studies. Second, the time span of the included studies was extensive, and there was a lack of uniformity in the diagnostic criteria for GDM. Third, our subgroup analysis is somewhat limited in scope. In the future, subgroups could be analyzed in terms of country, gestational week, and other maternal baseline characteristics. Fourth, primary cesarean cannot be distinguished in the original datasets. This distinction is crucial because primary cesarean rates can provide valuable insights into the clinical management and decision-making processes in GDM. Finally, long-term maternal and neonatal outcomes have not been studied or evaluated. The lack of relevant evaluations will make it difficult to judge the long-term effects of drugs, which reminds us that purposeful studies may be able to evaluate long-term outcomes in mothers and infants over time.

In conclusion, our meta-analysis, based on short-term data, suggested that metformin may curtail the risk of select adverse outcomes in patients with GDM and their newborns compared to insulin. These findings suggested that metformin could be a viable and efficient medication for managing GDM. However, given that metformin permeates the placental barrier, further research is needed to fully understand its long-term effects on patients with GDM and their offspring.

Conclusion

Overall, these findings suggest that metformin can provide some desirable benefits compared to insulin therapy, reducing the risk of certain maternal and neonatal outcomes in patients with GDM.

Statement of ethics

All analyses were based on previously published studies. Thus, neither ethical approval nor patient consent was needed.

Disclaimers

We declare that the views expressed in the articles submitted are our own and not the official position of the institution or funder.

Disclosure statement

No potential conflicts of interest were reported by the author(s).

Data availability statement

The data generated for this study are available from the corresponding authors upon reasonable request.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Appendix 1

(Specific retrieval strategy in PubMed)

Specific retrieval strategy in PubMed:

(“Metformin”[Mesh] OR Dimethylbiguanidine[Title/Abstract] OR Dimethylguanylguanidine[Title/Abstract] OR Glucophage[Title/Abstract] OR Metformin Hydrochloride[Title/Abstract] OR Hydrochloride, Metformin[Title/Abstract] OR Metformin HCl[Title/Abstract] OR HCl, Metformin[Title/Abstract]) AND ("Insulin"[Mesh] OR Insulin, Regular[Title/Abstract] OR Regular Insulin[Title/Abstract] OR Soluble Insulin[Title/Abstract] OR Insulin, Soluble[Title/Abstract] OR Insulin A Chain[Title/Abstract] OR Sodium Insulin[Title/Abstract] OR Insulin, Sodium[Title/Abstract] OR Novolin[Title/Abstract] OR Iletin[Title/Abstract] OR Insulin B Chain[Title/Abstract] OR Chain, Insulin B[Title/Abstract]) AND (“Diabetes, Gestational”[Mesh] OR Diabetes, Pregnancy-Induced[Title/Abstract] OR Diabetes, Pregnancy Induced[Title/Abstract] OR Pregnancy-Induced Diabetes[Title/Abstract] OR Gestational Diabetes[Title/Abstract] OR Diabetes Mellitus, Gestational[Title/Abstract] OR Gestational Diabetes Mellitus[Title/Abstract]) AND (randomized controlled trial[Publication Type] OR randomized[Title/Abstract] OR placebo[Title/Abstract])