Metformin Safety in the Management of

Gestational Diabetes
Pedro Marques, MD, Maria Raquel Carvalho, MD, Luísa Pinto, MD, Sílvia Guerra, MD
Disclosures
Endocr Pract. 2014;20(10):1022-1031. 

Abstract and Introduction

Abstract

Objective The use of metformin in pregnant women is still controversial, despite the increasing
reports on metformin's safety and effectiveness. We aimed to evaluate the maternal and neonatal
safety of metformin in subjects with gestational diabetes mellitus (GDM).

Methods We retrospectively reviewed the clinical records of 186 pregnancies complicated with
GDM surveilled at Hospital de Santa Maria, Lisboa, between 2011 and 2012. The maternal and
neonatal outcomes of 32 females who took metformin during pregnancy were compared with 121
females controlled with diet and 33 insulin-treated females.

Results Of the 186 GDM subjects, 32 (17.2%) received metformin during pregnancy. No statistical
differences between the diet and metformin groups were found with regard to the rates of abortion,
prematurity, preeclampsia, macrosomy, small-for-gestational-age (SGA) or largefor-gestational-age
(LGA) newborns, cesarean deliveries, neonatal intensive care unit (NICU) admissions, and birth
malformations or neonatal injuries. Similarly, there were no differences between the metformin and
insulin groups with regard to the referred outcomes. No abortions or perinatal deaths were recorded
in the metformin group. Ten out of 32 metformin patients required additional insulin.

Conclusion This retrospective study suggests that metformin is a safe alternative or additional

treatment to insulin in females with GDM. Metformin was not associated with a higher risk of
maternal or neonatal complications when compared to the insulin or diet groups.

Introduction

Gestational diabetes mellitus (GDM) is defined as any degree of glucose intolerance with onset or
initially identified during pregnancy.[1] The overall incidence of GDM has been increasing as
pregnant women become older and more obese,[2,3] and it varies widely depending on the diagnostic
criteria and characteristics of the studied population, such as ethnicity.[4,5] The prevalence rates of
GDM are higher in Asian/Pacific Islander, Non-Hispanic Black and Hispanic populations than in
Non-Hispanic White populations.[6] The prevalence of GDM ranges from 1.1 to 25.5% of all
pregnancies in the U.S.,[7] and it is estimated to affect 2.2% and 15% of South American and Indian
pregnancies, respectively.[8,9]
GDM is associated with an increased risk of maternal and perinatal complications including
preeclampsia, cesarean delivery, macrosomia, birth injuries and trauma, prematurity, hypoglycemia,
and neonatal respiratory distress.[10] Several studies have demonstrated that the risk of adverse
pregnancy outcomes increases continuously with elevated glucose levels and can be reduced with
effective treatment.[10,11]Interventions to modify lifestyle have been shown to improve perinatal
outcomes, and if hyperglycemia persists, additional treatments, traditionally insulin, are often
administered.[12]

While effective, insulin therapy has several disadvantages for pregnant women, including multiple
daily injections, risk of hypoglycemia, increased risk of maternal weight gain, and higher cost.
Moreover, it requires adjustments based on maternal weight, glucose levels, and diet and physical
activity, necessitating frequent assessments. Therefore, a safe and effective oral agent would offer
advantages over insulin.[13,14] Metformin, as the firstline medication for type 2 diabetes mellitus
(T2DM) in nonpregnant patients is an oral agent candidate for GDM. It decreases hepatic
gluconeogenesis and improves peripheral glucose uptake and is not associated with weight gain or
hypoglycemia.[15] Metformin crosses the placenta with a maternal-to-fetal transfer rate estimated at
10 to 16%, so it may directly affect fetal physiology.[16] The long-term effects of metformin on future
metabolic disorders in the offspring are currently unknown and constitute an important barrier for its
use.[13] However, it is speculated that metformin induces a more favorable intrauterine environment
and may reduce the offspring's long-term metabolic complications (e.g., obesity or diabetes mellitus
[DM]).[13,17]

Although several studies have reported that metformin is safe and suitable for GDM,[5,11,13,14,18–24] its use
in pregnant women remains controversial. Furthermore, inconsistencies in clinical outcomes across
studies and limited data on the safety, benefits, and risks of metformin in GDM have made it
difficult to establish definitive conclusions.[14]

In this retrospective study, we aimed to evaluate the maternal and neonatal safety of metformin in
GDM by comparing several maternal and neonatal outcomes in patients with GDM treated with
metformin, insulin, or dietary regimens.

Discussion 

The appropriate management of GDM can reduce adverse

pregnancy outcomes.[12,27] As insulin is proven to be safe and
effective, it is considered the gold standard therapy for GDM.
[28]
 Metformin use in GDM remains controversial. If metformin
proves to be effective and safe, it will be highly cost effective
and more convenient for patients.[13] Important studies,
summarized in Table 5 and Table 6, were conducted to clarify
the role of metformin in GDM, and the majority reported results
indicating that its use is favorable. Furthermore, some relevant
medical societies are allowing its use in selected cases.[29]
Globally, our retrospective study reinforces the conclusions of
relevant studies such as the MiG trial[11] and reports by
Balani[13] and Tertti;[23] we did not observe increases in adverse
maternal or perinatal outcomes in patients with GDM exposed
to metformin. When we compared the metformin group with the
diet and insulin groups, we found no significant increases in the
rates of preeclampsia, abortion, cesareans, prematurity,
macrosomia, SGA or LGA newborns, NICU admissions,
malformations, or neonatal injuries.
Baseline characteristics were analyzed to assure group
comparability. The metformin and insulin groups were
comparable, but the metformin group had a higher BMI than the
diet group. Nevertheless, similar outcomes were obtained in
both groups, reinforcing the safety of metformin. Moreover, our
results emphasize the safety profile of metformin because of the
comparison of different risk populations; the metformin group
included patients with more difficult-to-treat GDM whereas the
diet group included patients with more manageable GDM.
In contrast with the MiG trial,[11] we found a nonsignificantly
lower proportion of preterm births in GDM exposed to metformin
and fewer NICU admissions. The frequency of preterm birth in
the MiG trial was higher in the metformin group and was
associated with a greater frequency of spontaneous births,
raising the concern of an unrecognized effect of metformin on
the labor process. These findings were not confirmed in
subsequent studies.[5,13,19,22,20,24] Nevertheless, the increased rate
of preterm births in the MiG trial was not associated with higher
rates of other complications.
Macrosomia and LGA newborns are associated with perinatal
complications.[10,30] In our study, no differences were observed
between the metformin and diet or insulin groups regarding birth
weight, macrosomia, or LGA newborns. Tertti,[5] Balani,
[13]
 Mesdaghinia,[20] and Niromanesh[20] reported lower rates of
macrosomia in their metformin groups.
The rate of caesarean section was not significantly higher in the
metformin group, and no perinatal deaths or abortions occurred
in this group. The metformin group had rates of 3.1% and 6.3%
for neonatal malformations and injuries, respectively.
Mesdaghinia,[20] Balani,[13] and Tertti[5] concluded that metformin
does not increase the risk of fetal anomalies.
Surprisingly, we observed a higher mean weight gain during
pregnancy in the metformin group compared to the insulin
group. This result was unexpected considering that metformin
inhibits hepatic gluconeogenesis and stimulates glucose uptake
in peripheral tissues and that insulin is linked to weight gain. We
hypothesize that the higher BMI mean in the metformin group
may be a possible explanation: these women could be more
susceptible to weight gain or be careless about implementing
lifestyle modifications. Moreover, those subjects were
diagnosed mostly at 24 to 28 weeks, meaning that they had a
shorter period for nutritional and pharmacological intervention,
while the majority of subjects in the insulin group were
diagnosed before week 13. Most studies have reported lower
weight gain with metformin.[11,13,18,21]
he percentage of GDM that requires pharmacological treatment
varies from 20 to 60%.[5] Pharmacological treatment was
needed in 34.9% of our subjects. Nearly one-third of our GDM
patients receiving metformin required additional insulin. Some
series have reported lower rates, but Rowan found a higher
number of GDM patients treated with metformin that required
insulin (Table 5). This heterogeneity depends on the study
population and the objectives used to titrate the
pharmacological treatment.[13] Comparing the baseline features
of these 2 groups, it appears that a familial history of DM, older
age, and higher BMI may predict failure of the isolated use of
metformin. However, no statistical significance was detected for
these parameters, possibly due to the small number of subjects
in each group.
In the metformin plus insulin group, no significant maternal or
neonatal outcome differences were observed when compared
to metformin only except for prematurity and NICU admissions.
This difference could be explained by the maternal features
rather than treatment modality (the premature births occurred in
older mothers with ages of 37, 37, and 44).
Serious adverse effects from metformin were not found, and no
subjects had to stop treatment. Conversely, Rowan [11] and
Balany[13] reported interruption rates of 1.9% and 5%,
respectively.
An increased risk for developing T2DM after pregnancy has
been reported.[31,32] We did not identify any cases of T2DM on
postpartum OGTT, but IFG/IGT was verified in 13.6% of
subjects. Postpartum OGTT is important for patient
reclassification and should be requested in this setting.
Conclusion 

In conclusion, our findings suggest that metformin, alone or with

additional insulin, is a safe and effective treatment option in
GDM. We did not identify any significant risks or poorer
outcomes associated with metformin use. Its safety was
reinforced when compared with the diet group that had better
baseline features (lower mean BMI and HbA1C) and
theoretically harmless GDM cases. The evidence from this
retrospective study encourages the use of metformin in GDM
women.