Академический Документы
Профессиональный Документы
Культура Документы
DOI 10.1007/s00404-014-3566-0
GENERAL GYNECOLOGY
123
Arch Gynecol Obstet
123
Arch Gynecol Obstet
The characteristics of included studies are shown in The fixed-effects model was used for all secondary
Table 1 and the criteria used for diagnosis and criteria for maternal outcomes, and the results showed that there was
starting medical treatment are shown in Table 2. A total of no significant difference between the two treatments in
cesarean delivery rate (Pheterogeneity = 0.09, I2 = 44 %,
n = 1,594, RR = 0.94, 95 % CI 0.82–1.08, P = 0.36),
whereas significant decreased results were observed for
weight gain after enrollment (Pheterogeneity = 0.10,
I2 = 49 %, n = 1,530, MD = -1.28, 95 % CI -1.54 to
-1.01, P \ 0.0001) and gestational age at delivery
(Pheterogeneity = 0.26, I2 = 20 %, n = 2,072, MD = 0.94,
95 % CI -0.21 to -0.01, P = 0.03) in the metformin
treatment, compared with insulin treatment.
123
123
Table 1 Characteristics of 11 included studies in this meta-analysis
References Area Age, metformin Age, insulin n1 n2 Dose (mg/ Dose (u) BMI, BMI, insulin Gestation weeks at Maternal Neonatal
d) metformin randomization outcomes outcomes
Moore et al. USA 27.1 (4.7) 27.7 (6.7) 32 31 1,000–2,000 NA NA NA 27.8 28.9 (5.0) d, e a, b, f, h, i
[40] (6.5)
Rowan et al. Multi- 33.5 (5.4) 33.0 (5.1) 363 370 1,750–2,500 50 35.1 (8.3) 34.6 (7.2) 30.2 30.1 (3.2) b, c, d, e, a, b, c, d,
[29] center (30–90) (3.3) f, g e, f, g,
i, j
Tertti et al. Finland 31.9 (5.0) 32.1 (5.4) 110 107 500–2,000 2–42 29.4 (5.9) 28.9 (4.7) 30.3 30.4 (1.8) a, c, d, e, a, b, c, e, f,
[39] (2.0) f, g g, i, j
Niromanesh Iran 30.7 (5.5) 31.8 (5.1) 80 80 1,000–2,500 NA 28.1 (4.0) 27.1 (2.1) 28.7 28.6 (3.6) a, b, c, d, a, b, c, d,
et al. [24] (3.7) e, f, g e, f, g, h,
i
Hasan et al. Pakistan 30.29 (3.06) 30.88 (3.6) 75 75 500–3,000 NA 29.17 (1.94) 28.74 (2.69) 29.53 29.2 (1.48) a, d, e a, b, d, h, i,
[30] (1.33) j
Ijäs et al. Finland 32.3 (5.6) 31.7 (6.1) 47 50 750–2,250 30 31.5 (6.5) 30.8 (5.4) 30.0 30.0 (4.0) d, e a, b, c, h, i,
[25] (4.9) j
Spaulonci Brazil 31.93 (6.02) 32.76 (4.66) 46 46 1,700–2,550 NA 31.96 (4.75) 31.39 (5.71) 32.18 32.05 (3.50) b, c, d, a, b, c, d,
et al. [38] (3.70) e, g e, h
Ibrahim et al. Egypt 29.84 (5.37) 29.84 (5.37) 46 44 1,500–2,000 1.14A 31.83 (3.23) 28.7 (3.71) d, e a, b, g, i
[36]
Barrett et al. Multi- 33.3 (32.6–34.0) 32.9 (32.2–33.5) 219 213 1,750–2,500 NA 34.5 33.75 30.1 30.3 a, c, g –
[34] center (33.5–35.5) (32.73–34.77) (29.7 (29.9–30.7)
–30.6)
Barrett et al. Multi- 33.1 (32.4–33.8) 32.5 (31.9–33.2) 236 242 1,750–2,500 NA 35.2 34.6 20–33 20–33 e, b, c, d
[35] center (34.2–36.2) (33.7–35.5)
Mesdaghinia Iran 29.6 (5.3) 30.2 (5.9) 100 100 500–2,500 0.5/kg/d 27.6 28.46 27.9 28.9 (3.8) – a, b, c, e,
et al. [37] (3.22) f, g
Multi-center: New Zealand and Australia; n1: case of Metformin group; n2: case of Insulin group
Maternal outcomes: a HbA1c at 36–37 weeks gestation, b fasting blood glucose, c weight gain after enrollment, d Cesarean delivery, e gestational age at delivery, f incidence of pregnancy-
induced hypertension, g incidence of preeclampsia
Neonatal outcomes: a incidence of hypoglycemia, b birth weight, c prevalence of large for gestational age infants, d prevalence of small for gestational age infants, e premature birth, f shoulder
dystocia, g birth defect, h incidence of hyperbilirubinemia, i admission to neonatal intensive care unit, j birth trauma
BMI body mass index
A
IU/kg
Arch Gynecol Obstet
Arch Gynecol Obstet
Table 2 Criteria for diagnosis Study Criteria for diagnosis of GDM Criteria for starting medical
and starting medical treatment treatment
of GDM
Loading Fasting 1h 2h 3h Fasting Postprandial
sugar (g) (mg/dl) (mg/dl) (mg/dl) (mg/dl) (mg/dl) (mg/dl)a
Moore et al. [40] 100 105 190 165 145 105 120
Rowan et al. [29] 75 99 126 97.2 120.6
Tertti et al. [39] 75 95.4 180 154.8 99 140.4 (1 h)
Niromanesh et al. [24] 50 95 130 100 95 120
Hasan et al. [30] 75 95 180 155 100 126
Ijäs et al. [25] 75 95.4 198 172.8 95.4 120.6
Spaulonci et al. [38] 75/100 95 180 155 140 95 120
GDM gestational diabetes
mellitus Ibrahim et al. [36] 50 95 130 100 95 120
a
The postprandial levels (2 h) Barrett et al. [34] 50 95 130 100 95 120
offered in most included studies, Barrett et al. [35] 50 95 130 100 95 120
and postprandial levels (1 h) Mesdaghinia et al. [37] 100 95 180 155 140 95 120
offered in two included studies
provided the data on LGA infants; meta-analysis of these (Pheterogeneity = 0.49, I2 = 0, n = 1,501, RR = 0.82,
data failed to demonstrate a significant difference between 95 % CI 0.67–0.99, P = 0.04).
the two treatments (Pheterogeneity = 0.23, I2 = 26 %,
n = 1,977, RR = 0.91, 0.75–1.10, P = 0.33, Fig. 4b).
Publication bias
Similarly, no significant difference was indicated according
to the pooled analysis on the rate of SGA infants between
Publication bias was assessed for the outcome of birth
the two treatments (Pheterogeneity = 0.49, I2 = 0,
weight which was reported in most included studies. As
n = 1,710, RR = 0.88, 95 % CI 0.66–1.20, P = 0.43,
shown in Fig. 5, no obvious publication bias was observed
Fig. 4c). Hypoglycemia was reported in nine studies, and
with the relatively symmetrical funnel plots.
the combined result showed a significantly lower rate of
hypoglycemia in the metformin group as compared with
the insulin group (Pheterogeneity = 0.28, I2 = 18 %,
n = 1,793, RR = 0.69, 95 % CI 0.55–0.87, P = 0.001, Discussion
Fig. 4d). As shown in the pooled result of six studies, there
was no significant difference between the two treatments A total of 11 studies comprising 2,712 patients were
with respect to the incidence of hyperbilirubinemia included for present meta-analysis. With respect to the
(Pheterogeneity = 0.36, I2 = 8 %, n = 778, RR = 0.88, primary maternal outcomes, there was no significant dif-
95 % CI 0.63–1.22, P = 0.43, Fig. 4e). ference between the two groups except that PIH rate was
significantly lower in the metformin group. The pooled
Secondary neonatal outcomes analysis regarding secondary outcomes of pregnant women
revealed that when compared with the insulin group, sig-
The fixed-effects model was used for all secondary nificantly lower average weight gains after enrollment and
neonatal outcomes, and meta-analysis revealed that there average gestational ages at delivery existed in the metfor-
was no significant difference between the two treatments min group. This finding is consistent with the previous
in premature birth (Pheterogeneity = 0.17, I2 = 38 %, meta-analysis [26], and suggests a few benefits and risks of
n = 1,402, RR = 1.32, 95 % CI 0.92–1.88, P = 0.13), metformin treatment for GDM.
shoulder dystocia (Pheterogeneity = 0.58, I2 = 0, Although the cellular and molecular mechanisms of
n = 1,373, RR = 0.73, 95 % CI 0.36–1.47, P = 0.37), metformin remain unclear, Viollet et al. [41] considered
birth defect (Pheterogeneity = 0.52, I2 = 0, n = 1,382, that the benefit from metformin in PIH rate may be
RR = 0.76, 95 % CI 0.44–1.31, P = 0.32), and birth attributed to its effect on endothelial functions and reactive
trauma (Pheterogeneity = 0.39, I2 = 0, n = 1,197, oxygen species production. In fact, meta-analyses have
RR = 0.79, 95 % CI 0.42–1.46, P = 0.39). However, confirmed a reduced effect on cardiovascular mortality
compared with the insulin group, significantly decreased with metformin, as compared with any other oral diabetes
rate of NICU was observed in the metformin group agents [42, 43]. In the current study, similar HbA1c and
123
Arch Gynecol Obstet
Fig. 2 Assessment of the risk of bias in the included studies. judgment about each methodological quality item for each included
a Methodological quality graph: authors’ judgment about each study. ?: low risk of bias; ‘‘?’’: unclear risk of bias; ‘‘-’’: high risk of
methodological quality item presented as percentages across all bias
included studies; b methodological quality summary: authors’
FBG levels were observed between the two treatments, could not confirm significant difference between the two
which suggested that metformin was comparable with groups in NICU or birth weights; whereas significant lower
insulin in glycemic control. Though the anti-hyperglycemic incidence of hypoglycemia was presented for the metfor-
mechanism of metformin required to be confirmed, it was min group [27]. It is noteworthy that all 10 included studies
found that the liver was the main site of action for met- regarding birth weights did not find a significant difference
formin, and suppression of hepatic gluconeogenesis was between the two groups. We considered that the pooled
regarded as the most widely accepted model of the anti- result of our study regarding birth weights may be influ-
hyperglycemic action of metformin [44]. enced by some potential bias. The more recent RCT of
On the other hand, our results indicated that the average Ibrahim et al. [36] demonstrated a significantly reduced
birth weights, as well as the incidence of hypoglycemia and rate of NICU and hypoglycemia in the metformin group.
NICU were significantly lower in the metformin group as Similar results were obtained in the study of Hasan et al.
compared with the insulin group. This result, however, [30], and a rational interpretation suggested that signifi-
differs from the meta-analysis by Gui et al. [26] which cantly lower neonatal NICU admissions were associated
showed there was no significant difference between the two with lesser hypoglycemia and hyperbilirubinemia at birth
groups in the rate of hypoglycemia or NICU. In addition, in metformin group. However, in the study of Rowan et al.
they failed to demonstrate significant lower average birth [29] similar rates of neonatal hypoglycemia were found
weights in the metformin group, although slightly lower between metformin and insulin groups, although the result
tendency was found. Similarly, the other meta-analysis also observed a higher proportion of severe neonatal
123
Arch Gynecol Obstet
Fig. 3 Forest plots of primary maternal outcomes comparing metformin with insulin. a HbA1c, b FBG, c incidence of PIH, d incidence of
preeclampsia. MD mean differences, CI confidence intervals, RR risk ratio, FBG fasting blood glucose, PIH pregnancy-induced hypertension
hypoglycemia (glucose less than 28.8 mg/dL) in the insulin validated because of limited available information in this
group (8.1 %) compared with the metformin group (3.3 %, regard to date, meanwhile, the exact mechanism by which
P = 0.008). The authors considered that metformin could metformin affect the fetal physiology remains to be
cross the placenta to cause neonatal hypoglycemia, but was elucidated.
less likely to cause severe neonatal hypoglycemia [29]. All 11 included studies for present meta-analysis are
Actually, metformin crosses the placenta with a maternal- RCTs with relatively high quality. In addition, there was
to-fetal transfer rate estimated of 10–16 %, which may no significant heterogeneity across studies for all study
naturally affect fetal physiology directly [45]. In the cur- outcomes except for HbA1c, and no obvious publication
rent meta-analysis, the finding that a significantly reduced bias was observed. On the other hand, potential limita-
rate of hypoglycemia in the metformin group requires to be tions should be considered for our result. First, limited
123
Arch Gynecol Obstet
123
Arch Gynecol Obstet
123
Arch Gynecol Obstet
in the treatment of gestational diabetes: a prospective randomised 36. Ibrahim MI, Hamdy A, Shafik A, Taha S, Anwar M, Faris M
study. BJOG 118(7):880–885 (2014) The role of adding metformin in insulin-resistant diabetic
26. Gui J, Liu Q, Feng L (2013) Metformin vs insulin in the man- pregnant women: a randomized controlled trial. Arch Gynecol
agement of gestational diabetes: a meta-analysis. PLOS One Obstet 289(5):959–965
8(5):e64585 37. Mesdaghinia E, Samimi M, Homaei Z, Saberi F, Moosavi SGA,
27. Su D, Wang X (2014) Metformin vs insulin in the management of Yaribakht M (2013) Comparison of newborn outcomes in women
gestational diabetes: a systematic review and meta-analysis. with gestational diabetes mellitus treated with metformin or
Diabetes Res Clin Pract 104(3):353-357 insulin: A randomised blinded trial. Int J Prev Med 4(3):327
28. Dhulkotia JS, Ola B, Fraser R, Farrell T (2010) Oral hypogly- 38. Spaulonci CP, Bernardes LS, Trindade TC, Zugaib M, Francisco
cemic agents vs insulin in management of gestational diabetes: a RPV (2013) Randomized trial of metformin vs insulin in the
systematic review and metaanalysis. Am J Obstet Gynecol management of gestational diabetes. Am J Obstet Gynecol
203(5):457–459 209(1):34. e31–34. e37
29. Rowan JA, Hague WM, Gao W, Battin MR, Moore MP (2008) 39. Tertti K, Ekblad U, Koskinen P, Vahlberg T, Rönnemaa T (2013)
Metformin versus insulin for the treatment of gestational diabe- Metformin vs. insulin in gestational diabetes. A randomized study
tes. N Engl J Med 358(19):2003–2015 characterizing metformin patients needing additional insulin.
30. Hasan JA, Karim N, Sheikh Z (2012) Metformin prevents mac- Diabetes Obes Metab 15(3):246–251
rosomia and neonatal morbidity in gestational diabetes. Pak J 40. Moore LE, Briery CM, Clokey D, Martin RW, Williford NJ,
Med Sci 28(3):384–389 Bofill JA, Morrison JC (2007) Metformin and insulin in the
31. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, management of gestational diabetes mellitus: preliminary results
Ioannidis JP, Clarke M, Devereaux P, Kleijnen J, Moher D (2009) of a comparison. J Reprod Med 52(11):1011–1015
The PRISMA statement for reporting systematic reviews and 41. Viollet B, Guigas B, Garcia NS, Leclerc J, Foretz M, Andreelli F
meta-analyses of studies that evaluate health care interventions: (2012) Cellular and molecular mechanisms of metformin: an
explanation and elaboration. Ann Intern Med 151(4):W-65–W-94 overview. Clin Sci 122(6):253–270
32. Higgins JP, Green S, Collaboration C (2008) Cochrane handbook 42. Lamanna C, Monami M, Marchionni N, Mannucci E (2011)
for systematic reviews of interventions, vol 5. Wiley Online Effect of metformin on cardiovascular events and mortality: a
Library, Europe meta-analysis of randomized clinical trials. Diabetes Obes Metab
33. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Mea- 13(3):221–228
suring inconsistency in meta-analyses. BMJ 327(7414):557 43. Selvin E, Bolen S, Yeh H-C, Wiley C, Wilson LM, Marinopoulos
34. Barrett HL, Gatford KL, Houda CM, De Blasio MJ, McIntyre SS, Feldman L, Vassy J, Wilson R, Bass EB (2008) Cardiovas-
HD, Callaway LK, Nitert MD, Coat S, Owens JA, Hague WM cular outcomes in trials of oral diabetes medications: a systematic
(2013) Maternal and neonatal circulating markers of metabolic review. Arch Inter Med 168(19):2070–2080
and cardiovascular risk in the metformin in gestational diabetes 44. Rena G, Pearson ER, Sakamoto K (2013) Molecular mechanism
(MiG) trial responses to maternal metformin versus insulin of action of metformin: old or new insights? Diabetologia
treatment. Diabetes Care 36(3):529–536 56(9):1898–1906
35. Barrett HL, Nitert MD, Jones L, O’Rourke P, Lust K, Gatford 45. Nanovskaya TN, Nekhayeva IA, Patrikeeva SL, Hankins GD,
KL, De Blasio MJ, Coat S, Owens JA, Hague WM (2013) Ahmed MS (2006) Transfer of metformin across the dually per-
Determinants of maternal triglycerides in women with gestational fused human placental lobule. Am J Obstet Gynecol 195(4):
diabetes mellitus in the metformin in gestational diabetes (MiG) 1081–1085
study. Diabetes Care 36(7):1941–1946
123