The British Journal of Diabetes & Vascular Disease http://dvd.sagepub.

com/

Metformin treatment for gestational diabetes

Steve L Hyer, Jyoti Balani, Antoinette Johnson and Hassan Shehata British Journal of Diabetes & Vascular Disease 2009 9: 220 DOI: 10.1177/1474651409346767 The online version of this article can be found at: http://dvd.sagepub.com/content/9/5/220

Published by:
http://www.sagepublications.com

Additional services and information for The British Journal of Diabetes & Vascular Disease can be found at: Email Alerts: http://dvd.sagepub.com/cgi/alerts Subscriptions: http://dvd.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://dvd.sagepub.com/content/9/5/220.refs.html

>> Version of Record - Oct 21, 2009 What is This?

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

Metformin treatment for gestational diabetes

Steve L Hyer,1 Jyoti Balani,1 Antoinette Johnson,2 Hassan Shehata2
Abstract
he prevalence of gestational diabetes mellitus (GDM) is rising as the pregnant population becomes older and more obese. This is concerning because GDM is associated with increased perinatal morbidity such as macrosomia and shoulder dystocia, and the need for instrumental delivery. In addition, the offspring of GDM women have increased long-term risks of obesity and type 2 diabetes. There is no doubt that treating women with GDM improves pregnancy outcomes. Conventionally this has been by diet and insulin. Although effective, insulin increases appetite leading to weight gain. It increases the risk of hypoglycaemia and needs to be given by injection. There is also a substantial cost in terms of time for teaching and educating patients. Metformin offers a logical alternative to insulin in GDM, by reducing insulin resistance. Recent trial evidence indicates it is safe and effective. We describe our experience with metformin in GDM and review the evidence. Br J Diabetes Vasc Dis 2009;9:220-225 Key words: gestational diabetes, macrosomia, metformin

Abbreviations and acronyms ACHOIS  Australian Carbohydrate Intolerance Study in pregnant women CI confidence interval FDA Food and Drug Administration FGR foetal growth restriction GDM gestational diabetes mellitus GI gastrointestinal HBGM home blood glucose monitoring HOMA Homeostasis Model Assessment OGTT oral glucose tolerance test MiG Metformin in Gestational Diabetes PCOS polycystic ovary syndrome RR risk ratio

utcomes is not yet established.4-5 Although insulin therapy is o highly effective, it has the disadvantages of needing to be given by injection and causing weight gain. Furthermore, insulin does not address the fundamental problem in these patients which is insulin resistance. For this reason, very large doses of insulin are often required. A safe and effective oral agent would be welcomed by patients and their healthcare professionals.

Introduction
Currently, the number of pregnancies complicated by type 2 diabetes and GDM exceed those affected by type 1 diabetes.1 Although women with GDM do not have an increased risk of congenital malformations,2 the risk of macrosomia is the same as for women with pre-existing diabetes. Macrosomia is associated with an increased risk of intrauterine death, birth trauma (especially shoulder dystocia), Caesarean section and neonatal hypoglycaemia. Furthermore, large for gestational age babies are at increased risk of obesity, metabolic syndrome and type 2 diabetes later in life.3 The best possible management of these patients to achieve optimal glucose control and consistently good neonatal
1

Conventional therapy
In the landmark ACHOIS study, women with GDM (glucose, fasting < 7.8 mmol/L; 2-h 7.811 mmol/L) during a 75g OGTT were randomised to intervention (diet and/or insulin) or routine antenatal care as for patients without diabetes.6 Serious perinatal morbidity was reduced in the intervention group (n5490) compared with controls (n5510) (1% vs. 4%; RR 0.32, 95% CI 0.140.73). No infants died in the intervention group, but there were five deaths in the control group (p50.06, twotailed). Infants born to the intervention group were less likely to be large for their gestational age (13% vs. 22%; adjusted treatment effect 0.62, 95% CI 0.470.81). Although anxiety scores were similar in both groups, those in the intervention group were less likely to experience depression (8% vs. 17% of controls). This study reinforces previous uncontrolled studies that treatment for GDM reduces serious perinatal outcomes. The decision to start insulin is based on maternal glycaemia taking into account foetal growth characteristics. Postprandial glucose testing is recommended to optimise glucose control and reduce macrosomia.7 However, there is controversy regarding the optimal timing of home glucose testing with some evidence to suggest that aiming for 1-h post-meals values of < 7.6 mmol/L

Department of Diabetes & Endocrinology, St Helier Hospital, Carshalton, Surrey, UK. 2 Department of Maternal Medicine, Epsom & St Helier Hospital University Hospitals NHS Trust, Carshalton, Surrey, UK. Correspondence to: Dr Hassan Shehata Department of Maternal Medicine, Epsom & St Helier Hospital University Hospitals NHS Trust, Wrythe Lane, Carshalton, Surrey, SM5 1AA, UK. Tel/Fax: +44 (0)20 8401 9928 E-mail: [email protected]

The Author(s), 2009. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav

10.1177/1474651409346767

220

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

results in less macrosomia than when the 2-h target (6.6mmol/L) is used, although the difference in macrosomia rates did not reach statistical significance.8 The most useful foetal growth parameter is foetal abdominal circumference; if the abdominal circumference is > 95th centile or is crossing centiles on serial scans, insulin therapy is considered. Insulin given with meals and at bedtime (basal-bolus insulin) is more effective than twice daily insulin regimens; if control is still not adequate, insulin by continuous infusion pump can be considered if a pump service with adequate support is available locally. Both Diabetes UK and the American Diabetes Association urge caution with the use of the new insulin analogues in pregnancy until more evidence becomes available on their safety. However, rapid acting insulin analogues are particularly suitable for pregnancy because they reduce postprandial glucose peaks and have a low risk of hypoglycaemia. In a trial of 322 pregnant women with diabetes, post-breakfast glucose peaks were significantly reduced by insulin aspart compared with conventional human insulin with similar pregnancy outcomes.9 In the same trial, insulin aspart reduced the risk of major maternal hypoglycaemia by 28% and major nocturnal hypoglycaemia by 52% In the USA, the FDA classifies both insulin aspart and insulin lispro as category B risk for pregnancy.10 By contrast, data on the use of long acting insulin analogues in pregnancy are limited and these should be used only if the potential benefit justifies the potential risk to the foetus (category C risk in pregnancy). The Detemir and Pregnancy Trial will complete in 2010. Of the 240 participants, 187 have delivered with one perinatal death and there appear to be no concerns.11

ceiving on metformin is not increased despite being older and heavier than controls.18 Long-term data are also reassuring. At 18 months, growth and motor social development of offspring of mothers conceiving and continuing on metformin is no different to control babies.19 Metformin offers the advantage of being in a tablet form and is usually well tolerated. The MiG study compared insulin treatment with metformin in a prospective randomised multicentre trial.20 Perinatal morbidity was compared in the metformin (n5373) and insulin (n5378) groups. In the metformin and insulin treatment arms, composite neonatal morbidity was 32.0% vs. 32.2%, RR (95% CI) 1.0 (0.801.23), p50.95 and rates of large for gestational age were 19.3% vs. 18.6%, p50.83. There was no increase in hypertensive complications. However, women with small for gestational age infants, i.e. with evidence of FGR, were excluded from the study. The trial investigators concluded that perinatal complication rates are similar and metformin was safe. Women using metformin gained an average 1.6kg less weight from enrolment to term compared with the insulin group, and a questionnaire assessing the acceptability of treatment indicated that most women would be happy to take metformin again in subsequent pregnancies. Long-term data in the offspring are being collected.

Experience in gestational diabetes

In 2007, the data safety monitoring committee for the MiG trial reported no safety issues in 550 patients enrolled in the trial and no changes in protocol were needed. Reassured by this information, we began using metformin in women with GDM whose glycaemic control was not adequate despite lifestyle and dietary changes. We defined suboptimal control as more than three HBGM readings outside of target range which at that time was fasting blood glucose > 6mmol/L (now > 5.5mmol/L), 1-h postprandial blood glucose > 8 mmol/L, 2-h postprandial blood glucose > 7 mmol/L. Patients tested before breakfast and again 1 or 2 h after meals. Patients were given a written information sheet which clearly stated that metformin was unlicensed for use in pregnancy. Those patients unhappy to take metformin in pregnancy were treated with insulin as previously. We initiated metformin at 500mg twice daily with food and the dose was titrated to a maximum of 2,500 mg/day according to HBGM results, consistent with the MiG trial protocol. If glucose control was still not obtained, insulin was added and the metformin continued. Patients were asked to return to clinic one week after any dose escalation and kept in close contact with the diabetes specialist nurse by telephone. Patients attended the joint antenatal diabetes clinic and we collected outcome data prospectively during the clinics. We subsequently audited the first 45 patients with GDM treated with metformin and compared the data with 35 patients attending the clinic in the previous year, matched for baseline characteristics (weight, age, ethnicity) whose records had been extracted from the diabetes register and who had been treated with insulin.

Why metformin?
Metformin improves insulin sensitivity and would be expected therefore to improve glucose tolerance in pregnancy by reducing the physiological rise in insulin resistance that occurs during pregnancy. Indeed clinical studies show a reduction in insulin resistance as assessed by HOMA of 46% when metformin is started before conception and continued during pregnancy.12 Metformin is widely used outside of pregnancy for type 2 diabetes as it has a low risk of causing hypoglycaemia and is not associated with weight gain. Data for metformin use in type 2 diabetes during pregnancy are beginning to emerge.13,14 In a retrospective review of 214 pregnancies in women with type 2 diabetes, metformin had been continued in 93 patients, 23 of whom had taken it for the duration of the pregnancy until term.15 Compared with controls, i.e. women in the same series who had not taken metformin, outcome measures including pre-eclampsia, perinatal loss and neonatal morbidity were similar between the groups. Theoretically, stopping metformin in order to plan a pregnancy may result in a greater teratogenic risk if there is deterioration of glycaemic control. There is a wealth of experience with metformin when used to treat infertility in PCOS from which there is no evidence that metformin is teratogenic.16,17 In these women, observational studies have shown reduced miscarriage rates in the first trimester.16 The rate of pre-eclampsia in women with PCOS con-

THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE

221

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

The mean age of the GDM patients was 31.4 years, 44% were Asian, 7% black, and the mean body mass index was 30.3. Mean baseline glycated haemoglobin A1C was 5.5% (4.57.2%). The responses to metformin are shown in figures 1 and 2. Some patients showed a rapid improvement within 1 week (e.g. M.K.; figure 1). Others required higher doses or additional basal insulin (figure 1). In a small number of patients there was a poor response usually indicating pre-existing type 2 diabetes
Figure 1. Illustrative case histories

or impaired glucose tolerance discovered during pregnancy and shown by a persistently abnormal OGTT post-partum. Mean responses to metformin are shown in figure 2. Baby growth and hence birth weight will be influenced both by the severity of hyperglycaemia and its duration. Mean baby weight increases with gestation at which metformin is commenced until 36 weeks with a small decrease after this time. It is important that intensive monitoring occurs as soon as

Patient M.K. 100% High blood glucose readings 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 33 34 35 Weeks of gestation Baby weight = 3,540 g
Key: MF 5 metformin

MF1500

36

37

Figure 2. Mean response to metformin (MF)

Patient S.S. 100% 90% High blood sugar readings 80% 70% 60% 50% 40% 30% 20% 10% 0% 24 25 26 27 28 29 30 31 32 33 34 35 36 MF 2000 MF1500

Weeks of gestation Baby weight = 3,196 g

Key: MF = metformin

222

VOLUME 9 ISSUE 5 . SEPTEMBER/OCTOBER 2009

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

Figure 3. Mean maternal weight gain in metformin (MF) only versus insulin-treated patients with gestational diabetes

100% 90% High blood glucose readings 80% 70% 60% 50% 40% 30% 20% 10% 0% 35 MF 1500

Patient Z.H.

MF 2000

MF 2500

36 37 Week of gestation

38 Baby Weight = 4,389 g Abnormal GTT

Key: MF = metformin

GDM is detected and drug treatment initiated immediately after dietary measures fail. During pregnancy there is enhanced elimination of metformin, which might suggest a need to adjust doses especially in late pregnancy.21 Metformin is stopped after delivery and OGTT repeated at 6 weeks post-partum. In fact, metformin can be used safely during breastfeeding.22

Comparison with insulin

In our early series, we found significantly less maternal weight gain in the group treated with metformin compared with

insulin-treated controls (p50.04; figure 3). Mean baby weight was reduced (3,491 g vs. 3,541 g) but this did not achieve statistical significance (figure 4). We have since studied 100 metformin-treated GDM patients compared with 100 insulintreated controls and confirmed the significant difference in maternal weight gain from enrolment to term.23,24 Although baby weights were not significantly different, birth weight centile for gestational age was significantly reduced in the metformin group using specific software that adjusts for maternal height, maternal weight at booking, parity, ethnic group and sex of baby.25

Figure 4. Mean baby weight (g) and week of gestation at which metformin (MF) treatment commenced

100% High blood glucose readings 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 25 26 27 28 29
MF 1000 MF 2000 MF + Insulin

Patient S.G.

30 31 32 33 34 Weeks of gestation

35

36

37

38

39

Baby weight = 2,952 g

Key: MF = metformin

THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE

223

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

Figure 5. Neonatal outcomes in metformin-treated and insulin-treated groups

100 Mean percentage of high blood sugar 90 80 70 60 50 40 30 20 10 0 12.6% 3.7% 5.5% 2.9% 3.7% 43%

nt

eks

tme

eks

eks

eks

we

we

we

we

trea

1-2

3-4

5-6

7-8

Pre

Data for neonatal outcomes are shown in figure 5. We found a significant reduction in the incidence of neonatal jaundice (defined as requiring phototherapy) in the metformin treated group. This is probably accounted for by the increased rate of prematurity in the insulin-treated group and their increased frequency of admissions to the special care baby unit. Our recent data from the larger series confirm this finding. Not all patients can tolerate metformin because of unpleasant GI side-effects even at the lowest doses, despite being advised to take the drug with food and up-titration of the dose. In addition some patients on reflection decide not to take it for fear of adverse effects on the baby. We found five of the 45 patients in our original series stopped metformin because of fear of harm to their baby and one patient stopped because of intolerable GI upset. Supplementary insulin was needed in addition to metformin in four patients to achieve glucose targets. In our larger series of 127 patients, 3.9% of women discontinued metformin because of GI side-effects and 10% required additional insulin. This compares with data from the MiG trial in which 1.9% of patients stopped because of adverse effects and 46% of patients required insulin added to the metformin. The difference in the need for insulin may reflect the tighter glycaemic targets used in MiG (fasting < 5.4 mmol/L, 1-h postprandial < 6.7 mmol/L).

uglycaemia (see figure 6 for treatment protocol). This has e resulted in a considerable easing of the workload burden for the diabetes specialist nurses who previously had large numbers of women at each clinic for initiation of insulin treatment. Our neonatologists have reported a significant decline in numbers of babies with hypoglycaemia and admissions to the special care baby unit. Our approach is in line with recent NICE guidance (2008)26 for diabetes in pregnancy which recommends that women with diabetes may be advised to use metformin as an adjunct or alternative to insulin when the likely benefits from improved glycaemic control outweigh the potential for harm.

Figure 6. Treatment protocol for gestation diabetes mellitus (GDM)

9-1

0w

eek

3,692 3,700 3,600 3,500 Baby weight 3,400 3,300 3,200 3,100 3,000 2,900 24-26 27-29 30-32 Week of gestation 33-35 36-38 3,196 3,395 3,412 3,540

Conclusions
On the basis of randomised controlled data and our own experience we now offer all GDM women metformin where lifestyle measures are inadequate to achieve or maintain

224

VOLUME 9 ISSUE 5 . SEPTEMBER/OCTOBER 2009

Downloaded from dvd.sagepub.com by guest on January 27, 2013

Achieving best Practice

Key messages
In our experience, compared with insulin, metformin

results in: less macrosomia fewer admissions to a special care baby unit  reduced incidence of neonatal jaundice and hypoglycaemia Metformin:  is a safe alternative to insulin when dietary and lifestyle measures are insufficient to achieve target glucose levels  offers several advantages over insulin in the management of gestational diabetes including less weight gain, no risk of hypoglycaemia, oral therapy, and reduced costs including specialist nurse time  can be combined with basal insulin when fasting glucose levels remain high should not be used if FGR is present is not licensed for use in pregnancy

The majority of our patients tolerate metformin and achieve good control, avoiding the need for injections. To date we have no experience with modified release metformin in pregnancy which may be better tolerated. There remain occasional patients who are unhappy about taking metformin in pregnancy and insulin alone (usually prandial insulin aspart and basal isophane insulin) is used for this group. For women with high fasting home glucose values (> 5.6 mmol/L) despite total daily doses of 2,500 mg metformin, we add in basal insulin such as isophane insulin. It is important to be vigilant for any evidence of FGR. We do not use metformin if FGR is present and if metformin has already been started, recommend discontinuing it. Vigilance is also required for maternal changes, such as development of a hypoxic state or reduced renal function, which would warrant discontinuation of metformin therapy.

References
1. Reece EA, Homko CJ. Pregnancy outcomes in gestations complicated by type 2 diabetes. Curr Diab Rep 2005;5:269-71. 2. Farrell T, Neale L, Cundy T. Congenital anomalies in the offspring of women with type 1, type 2 and gestational diabetes. Diabet Med 2002;19:322-6. 3. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in childhood: association with birth weight, maternal obesity, and gestational diabetes mellitus. Paediatrics 2005;115:e290-6. 4. Hyer SL, Shehata HA. Gestational diabetes mellitus. Curr Obstet Gynaecol 2005;15:368-74. 5. Fraser R, Heller SR. Gestational diabetes: aetiology and management. Obstet Gynaecol Reprod Med 2007;17:345-8.

6. Crowther CA, Hiller JE, Moss JR et al. The Australian Carbohydrate Intolerance Study in Pregnant Women (ACHOIS) Trial Group. Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 2005;352:2477-88. 7. de Veciana M, Major CA, Morgan MA et al. Postprandial versus preprandial blood glucose monitoring in women with gestational diabetes mellitus requiring insulin therapy. N Engl J Med 1995;333: 1237-41. 8. Weisz B, Shrim A, Homko CJ et al. One hour versus two hours postprandial glucose measurement in gestational diabetes: a prospective study. J Perinatol 2005;25:241-4. 9. Mathieson ER, Kinsley B, Amiel SA et al. On behalf of the Insulin Aspart Pregnancy Study Group. Maternal glycaemic control and hypoglycaemia in type 1 diabetic pregnancy. Diabetes Care 2007;30:771-6. 10. US Food and Drug Administration website: http://www.accessdata, fda,gov/drugsatfda_docs/label/2008/020357s031,021202s016lbl.pdf (Accessed 7th September 2009). 11. Duncan C. Highlights from the Fifth International Symposium on Diabetes and Pregnancy. Br J Diabetes Vasc Dis 2009;9:139-40. 12. Glueck CJ, Goldenberg N, Wang P et al. Metformin during pregnancy reduces insulin, insulin resistance, insulin secretion, weight, testosterone and development of gestational diabetes: prospective longitudinal assessment of women with polycystic ovary syndrome from preconception throughout pregnancy. Hum Reprod 2004; 19:510-21. 13. Hawthorne G. Metformin use and diabetic pregnancy-has its time come? Diab Med 2006;23:223-7. 14. Brown FM, Wyckoff J, Rowan JA et al. Metformin in pregnancy: Its time has not yet come. Diabetes Care 2006;29:485-6. 15. Hughes RCE, Rowan JA. Pregnancy in women with Type 2 diabetes: who takes metformin and what is the outcome? Diabet Med 2006;23:318-6. 16. Glueck CJ, Wang P, Goldenberg N, Seive-Smith L. Pregnancy outcomes among women with polycystic ovary syndrome treated with metformin. Hum Reprod 2002;17:2858-64. 17. Jakubowicz DJ, Iuorno MJ, Jakubowicz S. Effects of metformin on early pregnancy loss in the polycystic ovary syndrome. J Clin Endocrinol Metab 2002;87:524-9. 18. Glueck CJ, Bornovali S, Pranikoff J et al. Metformin, pre-eclampsia, and pregnancy outcomes in women with polycystic ovary syndrome. Diabet Med 2004;21:829-36. 19. Glueck CJ, Goldenberg N, Pranikoff J et al. Height, weight, and motorsocial development during the first 18 months of life in 126 infants born to 109 mothers with polycystic ovary syndrome who conceived on and continued metformin through pregnancy. Hum Reprod 2004;19:1323-30. 20. Rowan JA, Hague WM, Gao W et al. For the MiG Trial Investigators. Metformin versus insulin for the treatment of gestational diabetes. N Engl J Med 2008;358:2003-15. 21. Hughes RCE, Gardiner SJ, Begg EJ, Zhang M. Effect of pregnancy on the pharmacokinetics of metformin. Diabet Med 2006;23:323-6. 22. Glueck CJ, Salehi M, Sieve L, Wang P. Growth, motor, and social development in breast and formula-fed infants of metformin-treated women with polycystic ovary syndrome. J Pediatr 2006;148:628-32. 23. Balani J, Hyer SL, Rodin DA, Shehata H. Pregnancy outcomes in metformin-treated women with gestational diabetes. Diabetologia 2008;51(Suppl 1):S1-S588, 1147. 24. Balani J, Hyer SL, Rodin DA, Shehata H. Pregnancy outcomes in women with gestational diabetes treated with metformin or insulin: a case control study. Diabet Med 2009;26:798-802. 25. Gardosi J, Francis A. Customised Centile Calculator: GROW-Centile v 5.1. Gestation Network, www.gestation.net. Birmingham: Perinatal Institute, 2006. 26. National Institute for Health and Clinical Excellence Diabetes in pregnancy: full guideline. London: NICE, 2008.

THE BRITISH JOURNAL OF DIABETES AND VASCULAR DISEASE

225

Downloaded from dvd.sagepub.com by guest on January 27, 2013