13.4.10.6 Diabetes management in pregnancy

Although the outlook for the woman with diabetes has greatly improved since the discovery of insulin, the goal of the St. Vincent Declaration (1989) that the outcome of diabetic pregnancy should approximate that of nondiabetic pregnancy has still not been realized. In the mid 1990s, a number of regional UK centres reported a four-fold to ten-fold increase in congenital malformations and three- to five-fold increase in perinatal mortality, compared with the background population. A general increase in the prevalence of type 2 diabetes is being translated into the pregnancy context and outcomes appear similar to those of type 1 diabetes. The problem of pregnancy planning and other key demographic and pregnancy-related features were highlighted in a major UK Confidential Enquiry into Maternal and Child Health (CEMACH) during 2002–2003, which has provided a largely unrivalled source of reference (1). While the relevance of overt hyperglycaemia to maternal and perinatal outcomes is now clearly established, the significance of minor degrees of hyperglycaemia for maternal/fetal outcome has been the subject of much controversy and dogma. The lack of a robust evidence base is reflected in the lack of consensus among published guidelines (2).

Despite these limitations, the outcome of pregnancy for most women with diabetes is good, and this undoubtedly reflects improved obstetric surveillance and better management of maternal hyperglycaemia over the last several decades. The aim is, through education and maternal empowerment, to optimize blood glucose control both before and during pregnancy, so that pregnancy may proceed as normally as possible and result in the birth of a normal baby at near term.

The last few years have seen the publication of a number of landmark observational studies and randomized trials (3–8), which have the potential to alter the diagnostic and therapeutic landscape considerably. Some guidance for the management of diabetes in pregnancy has recently been published (9, 10).

The UK CEMACH survey estimated the frequency of type 1 diabetes to be 1 in 364 (0.27%) and type 2 to be 1 in 955 births (0.10%) (1). Type 1 diabetes dominates in northern European populations, but the prevalence of type 2 diabetes and gestational diabetes mellitus in pregnancy is increasing (up to 20% of pregnancies in certain populations), and varies significantly between ethnic groups and locations. This has resulted in a predominance of patients with type 2 diabetes over patients with type 1 in some diabetes clinics. The increase in gestational diabetes mellitus appears to be compounded by obesity, which now affects about 1 in 5 women who give birth. Pregnancy may cause or worsen obesity through excessive weight gain, and obesity may complicate pregnancy by increasing the risk of fertility problems, excess fetal growth, and maternal hypertensive and diabetic disorders.

Classical risk factors for adverse pregnancy outcome in diabetic mothers are well recognized, and will be modified to some extent by the type and duration of diabetes, glycaemic control, and diabetes-related vascular complications (2). General factors include age, parity, weight, hypertension, smoking, and drug abuse. Relevant obstetric factors include previous miscarriage, multiple pregnancy, nutritional deficiency, late booking, and poor obstetric history. Risks to the mother include progression of pre-existing diabetic complications, spontaneous abortion, and, in later pregnancy, pre-eclampsia, polyhydramnios, macrosomia, operative delivery, and stillbirth—all reported to be more common in diabetic women. Iatrogenic risks relate to more intensive blood glucose control. Specific risks to the baby include both intrauterine growth retardation (small for dates) and fetal overgrowth (macrosomia). The associated risks of prematurity, operative delivery, and neonatal hypoglycaemia all require expert neonatal supervision. Long-term implications relate to the risk of recurrent gestational diabetes mellitus, future diabetes in the mother and intrauterine programming of disease in the offspring. Education of women at risk and regular antenatal clinic attendance are important potential modifiers of most of these factors.

There is an adaptation of maternal metabolism during pregnancy to ensure growth and development of the fetus. This involves a greater fall in plasma glucose and amino acids, and a greater rise in free fatty acids to overnight fasting than in the nonpregnant state (‘accelerated starvation’), associated with hepatic insulin resistance. In later pregnancy, a progressive rise in postprandial glucose and its associated insulin response, associated with decreased insulin sensitivity, parallels the growth of the fetal placental unit and rapidly reverses after delivery. This ‘facilitated anabolism’ brings about appropriate changes in carbohydrate, amino acid, and lipid metabolism, and ensures adequate nutrients for the developing fetus.

Women who lack the necessary β cell reserve—either absolutely, as in type 1 diabetes, or relatively, as in type 2 diabetes or gestational diabetes mellitus—will have abnormal adaptation of carbohydrate, protein, and fat metabolism. The pregnant woman with type 1 diabetes requires sufficient insulin to compensate for increasing caloric needs, increasing adiposity, decreasing exercise, and increasing anti-insulin hormones. The insulin dose to maintain normoglycaemia and prevent maternal ketosis may increase up to threefold in the course of pregnancy in type 1 diabetes, and women with type 2 diabetes will usually require insulin treatment in pregnancy, often at high doses because of their obesity and physical inactivity. An understanding of these underlying pathophysiological mechanisms is necessary for the successful management of these women during pregnancy.

The hypothesis that maternal hyperglycaemia accelerates fetal growth through provision of excessive glucose to the fetus at a time when the fetal pancreas can respond by increasing its production of insulin—an important fetal growth factor—was first enunciated by Jorgen Petersen some 50 years ago (11). This hypothesis, which is supported by animal and epidemiological data, has provided a basis for the concept of fetal programming. Other maternal fuels are likely also to be implicated.

Internationally agreed definitions of type 1 and type 2 diabetes apply during pregnancy. The fasting and 2-h cut-points employed in these criteria indicate unequivocal hyperglycaemia that may or may not have been recognized before pregnancy. The metabolic adaptations of pregnancy, however, affect both the fasting and 2-h plasma glucose values and suggest the need for more specific diagnostic criteria pertaining to pregnancy. The term ‘gestational diabetes’ is defined as carbohydrate intolerance with onset or first recognition during pregnancy. This definition provides little insight into underlying pathophysiology, the spectrum of associated hyperglycaemia, and the impact of such a diagnosis for maternal fetal outcome.

To date, a number of differing approaches in the UK, Europe, and the USA have caused confusion (12, 13, 15). The essence of the debate surrounds the relevance of minor degrees of glucose intolerance to maternal and fetal outcome. Traditionally, two different schemes have been employed. Both involve an oral glucose tolerance test (OGTT), but differ in the glucose load and interpretation. Much of the controversy derives from the fact that the initial criteria for the diagnosis of gestational diabetes mellitus, established more than 40 years ago, were chosen to identify women at high risk for the development of diabetes outside pregnancy, or were derived from adaptation of criteria used from nonpregnant persons, not to identify pregnancies with increased risk of adverse perinatal outcome (13). A fundamental problem is that the relation between hyperglycaemia during pregnancy and adverse outcome appears to be continuous.

The Hyperglycaemia and Pregnancy Outcome (HAPO) trial was a multicentre, multicultural, observational study designed to examine whether maternal hyperglycaemia, short of diabetes, is associated with adverse maternal fetal outcome (3). The study comprised 25 000 women who underwent a 75 g oral glucose tolerance test (OGTT) at an average of 28 weeks gestation with care-givers blinded to the results, unless the fasting venous plasma glucose was above 5.8 mmol/l or the 2-h was greater than 11.1 mmol/l. The results demonstrated a continuum of risk, without clear thresholds, between each of the OGTT glucose measures (fasting, 1-h, and 2-h post-glucose load) and each of the four primary outcomes: macrosomia, primary Caesarean section, neonatal hypoglycaemia, and fetal hyperinsulinism (Fig. 13.4.10.6.1) (3). The associations were modestly attenuated, but persisted, after controlling for multiple confounding variables, including field centre, gestational age at delivery, and maternal body mass index (BMI). Positive associations were also found with increasing glucose levels, and each of the five secondary outcomes: premature delivery, shoulder dystocia or birth injury, intensive neonatal care, hyperbilirubinaemia, and pre-eclampsia. The large study size, broad inclusion criteria, and similarity across centres in the associations between maternal glycaemia and outcomes support the development of outcome-based criteria for classifying maternal glucose metabolism in pregnancy.

A consensus statement for the diagnosis of gestational diabetes mellitus based on the HAPO data has been published (14). The cut-points from that concensus are given in Table 13.4.10.6.1., together with earlier recommendations (12,15). The glycaemic threshold values chosen from the HAPO data are consistent with an odds ratio of 1.75 higher than mean glucose values (i.e. 1.0). Thirteen per cent of the total group had one or more values greater than or equal to the threshold. It seems likely that these new criteria, and the facility to diagnose gestational diabetes mellitus with a single cut-off value, will result in a large increase in the number of patients with gestational diabetes mellitus. While the cost implications are obviously relevant, the recommendations need to be viewed in the context of the increasing prevalence of overweight, gestational diabetes mellitus and type 2 diabetes in this population.

There is again a lack of international agreement in this area. Several differing approaches to screening are detailed in position statements and society recommendations, reflecting the lack of a gold standard diagnostic test and limited evidence showing the benefit of treatment. Some authorities have recommended universal or selective screening, while others have indicated the practice should be abandoned. In North America, screening traditionally has been based on a 50 g 1-h non-fasting glucose challenge test. In the UK (16), until recently, screening comprised:

The recent NICE guidelines (9) have reverted to traditional risk factors:

even though these have been shown to lack sensitivity and specificity.

It is unclear as to whether the HAPO investigators will make any recommendations in relation to screening. Pragmatically, a single screening and diagnostic test at 28 weeks, such as a fasting glucose, has considerable appeal, but concerns about post-meal hyperglycaemia as the main abnormality in some ethnic groups remain. Screening strategies will have to be evaluated in relation to new HAPO diagnostic criteria. Routine screening will be also necessary at the first visit, but debate continues as to whether blood testing should be reserved only for high-risk groups. For some populations, this will translate, essentially, into universal screening.

These mothers may have had either type 1 or type 2 diabetes before conception. For both types of diabetes, the aim is to optimize control before conception combined with active screening for, and treatment of, diabetic complications.

A multidisciplinary team operating in a secondary or tertiary care setting is a commonly adopted model for the provision of pregnancy care for women with diabetes. Our clinic in Belfast started in the late 1950s and was among the first such clinics in the UK. Essential members of the team include an obstetrician and a diabetes physician supported by a diabetes specialist nurse, dedicated dietitian, and a diabetes-trained midwife. Review, which is usually fortnightly, is initially centred on diabetic rather than obstetric issues. Patients are often seen weekly as term approaches.

Recognition that congenital malformations were increased in the infants of mothers with diabetes first occurred more than 40 years ago, and was quickly seen to be linked with maternal hyperglycaemia. The concept of pre-pregnancy care was developed after Pedersen observed that the occurrence of hypoglycaemic reactions during the first trimester and insulin coma was low in pregnancies resulting in malformed infants, indicating poor diabetes regulation at that time. Judith Steel established a prepregnancy clinic in Edinburgh in 1976, and this model has now become accepted practice (17). A recent meta-analysis of 14 studies of pre-pregnancy care showed a threefold reduction in the risk of major congenital malformations among 1192 offspring who received such care, compared with 1459 offspring of mothers who did not (18). It is important for the mother to realise that the risks are reduced with any improvement in HbA_1c (9).

Despite these data, the UK CEMACH study showed that 62% of women with type 1 diabetes and 75% of women with type 2 diabetes had no evidence of pre-pregnancy counselling (1). Suboptimal pre-conception care was associated with a fivefold increased risk of poor pregnancy outcome (defined as death after 20 weeks or a major congenital malformation). Women with type 2 diabetes are less likely to receive formal pre-pregnancy care.

This can be defined as the education of, and discussion with, women of reproductive age about pregnancy and contraception. It is an essential component of every consultation in primary and/or specialist care and includes education and discussion on:

This is the additional care needed to prepare a woman with diabetes for pregnancy, and involves a close partnership between the woman and health care professionals. Ideally, it should begin at least 6 months before she embarks on a pregnancy and includes advice and discussion on the following:

The factors contributing to poor outcome in type 2 diabetes are complex, and include suboptimal diabetes control, use of potentially teratogenic drugs, older age, obesity, and greater socioeconomic deprivation and ethnic diversity. Many of these can be addressed by pre-conception care. Obesity should be addressed with intensive lifestyle support before pregnancy. Unless the women has polycystic ovarian syndrome (PCOS) and the benefits of metformin outweigh the potential disadvantages, women with type 2 diabetes should be transferred to insulin before pregnancy and oral hypoglycaemic drugs discontinued. Following the use of metformin in PCOS women, there is growing interest in the continuing use of metformin in Type 2 diabetic women in pregnancy, with insulin as needed to achieve glucose targets, but there good outcome studies are still required.

Few evidence-based dietary guidelines exist for women with diabetes, with or without obesity. A healthy lifestyle consisting of a well-balanced diet and moderate physical activity should be encouraged for all women during pregnancy. The aim is to limit blood glucose values post meals and to prevent hypoglycaemia between meals. Low glycaemic-index foods help to achieve this. As with a healthy, nonpregnant diet, approximately 50% of the total energy is provided by carbohydrate and less than 35% from fat. Individual advice around appropriate carbohydrate intake, and the adjustment of insulin prior to exercise to avoid hypoglycaemia, is necessary, especially early in pregnancy. For women with type 2 and gestational diabetes mellitus, 30 minutes of walking once or twice a day after meals is realistic, easily achievable, and can lower postprandial blood glucose values.

Appropriate weight targets should be based on the woman’s pre-pregnancy weight. For normal weight women, a 10–12.5 kg pregnancy weight gain is considered optimal, as it is associated with fewer pregnancy-related complications. By contrast, for underweight women (BMI <19.8 kg/m²), a pregnancy weight gain of 12.5–18 kg is recommended. Minimizing unnecessary weight gain in obese subjects with type 2 diabetes and women with gestational diabetes mellitus can improve maternal glycaemic control, reduce the risk of macrosomia, and improve pregnancy outcomes. Current UK guidelines recommend women with a prepregnancy BMI greater than 27 kg/m² to restrict their calorie intake to around 25 kcal/kg per day in the second trimester. Official guidance for the management of significantly obese women remains limited, and the most recent (2009) American Institute of Medicine recommendations simply grouped all women with a prepregnancy BMI equal to or greater than 30 kg/m², in whom they recommended a total weight gain of 11–20 lbs (5–9.1 kg) during pregnancy.

Despite modern technology, optimizing glycaemic control remains demanding for both the patient and clinician. In type 1 diabetes, blood glucose should be measured up to 8 times daily (before and after each meal, at bedtime, and, intermittently, in the middle of the night). In type 2 diabetes or gestational diabetes mellitus, the frequency of monitoring is, to some extent, dictated by the treatment strategy, but more frequent measurements identify more frequent hyperglycaemia, leading to a higher insulin usage and a lower incidence of macrosomia. A recent UK guideline (9) has recommended target values of HbA_1c less than 6.1% (43 mmol/mol), if feasible, and capillary measurements of 3.5–5.9 mmol/l pre prandially and less than 7.8 mmol/l 1-h post-prandially. HbA_1c provides an important index of peri-conceptional hyperglycaemia, but changes too slowly to be used to inform treatment adjustments throughout pregnancy, where its main use is for retrospective analysis of control in each trimester, and to support home glucose-monitoring data. Any strategy for intensive diabetes control must constantly be balanced against the risk of maternal hypoglycaemia.

Most patients are now using a multiple dose injection (MDI) insulin regimen, although there is little evidence to support the use of one particular regimen over another. MDI usually comprises a short-acting insulin taken before meals and an intermediate-acting insulin at bedtime (often given with a pen device). Some advocate the prescription of a basal isophane insulin more than once daily, and there may be justification for this approach, with the increasing use of rapid-acting insulin analogues for meal coverage, if the gap between meals is greater than 3 h. Those patients with type 2 diabetes who are controlled on a twice-daily fixed-mixture insulin regimen prepregnancy are often changed to an MDI regimen for the duration of pregnancy.

There has been increasing interest in the role of insulin analogues in pregnancy. Conceptually, their advantages include fewer episodes of hypoglycaemia, a reduction in postprandial glucose excursions, and increased patient satisfaction. A randomized trial in type 1 diabetes, comparing insulin aspart with regular soluble insulin, showed similar efficacy, with a tendency to lower rates of hypoglycaemia and without apparent toxicity (19). Currently, insulin aspart is licensed for use in pregnancy. More data on the role and safety of long-acting analogues are needed, although retrospective data to date have reported no evidence of toxicity.

There are no convincing data that continuous subcutaneous insulin infusion is superior to MDI for the majority of women, but its selected use in experienced centres and with motivated patients with difficult-to-control diabetes may be appropriate. Care must be taken to avoid ketoacidosis. It seems likely that capillary glucose monitoring reveals only a minor fraction of glucose variability, and this may be relevant to adverse fetal outcome. A recent randomized trial showed an improvement in glycaemic control among subjects with type 1 and type 2 diabetes with the use of continuous glucose monitoring in each trimester, compared with regular monitoring (20). Given the cost of each sensor, it is unlikely that this technology will be routinely available, but its selected use in problematic patients may supplement home blood glucose monitoring data.

A significant proportion of these women will have previously unrecognized type 2 diabetes, as suggested by the persistence of glucose intolerance post partum. The validity of gestational diabetes mellitus as a diagnostic entity has been disputed, but support for the concept has come from the HAPO study (7) and two recent randomized trials (4, 5) (Table 13.4.10.6.2). The ACHOIS Trial (4) included women diagnosed by WHO diagnostic criteria (fasting glucose < 7 mmol/l (range 3.4–6.2; mean 4.8 mmol/l) and 2-h glucose between 7.8–11.1 mmol/l (median 8.6 mmol/l)). In the MFMU Network trial (5), gestational diabetes mellitus was diagnosed by a fasting value of less than 5.3 mmol/l and any two of 1-h, 2-h, and 3-h values after an 100 g OGTT of ≥10.0, 8.6, and 7.8 mmol/l, respectively. In both trials, average birth weight, frequency of babies being large for gestational age (LGA), and pre-eclampsia were reduced by treatment. The MFMU trial showed a reduction in Caesarean section rate, even though women were identified as having gestational diabetes mellitus—an important point because of concerns that simply identifying women as having gestational diabetes mellitus increases the risk of intervention in delivery because of increasing professional anxieties. The frequencies of adverse outcomes in the untreated arms of these studies were similar to those found with maternal glycaemia above a putative threshold in the HAPO study.

Women with gestational diabetes mellitus are usually diagnosed during screening or by a diagnostic OGTT around the middle of pregnancy. Hyperglycaemia first recognized in early pregnancy is more often previously undiagnosed type 1 or type 2 diabetes and HbA_1c is frequently elevated, which may not be the case in gestational diabetes mellitus. If type 1 diabetes is considered most likely, insulin should be started immediately to prevent the unexpected development of ketoacidosis. Medical nutritional therapy with regular review of blood glucose monitoring results and fetal growth is successful in controlling hyperglycaemia in 80% to 90% of patients with gestational diabetes mellitus (4, 5). Predictors of the need for additional therapy are the degree of hyperglycaemia, gestational age at diagnosis, and use of insulin in a previous pregnancy. Insulin therapy is indicated for persistent maternal hyperglycaemia in order to prevent fetal complications, especially those related to compensatory hyperinsulinaemia and ensuing macrosomia. There is good evidence that such an approach is safe and effective. The goals of treatment are similar to type 1 diabetes and MDI regimens are frequently used. The starting dose of insulin is 1.0 U/kg/day and can easily be commenced as an outpatient in a diabetes centre, with rapid upward titration by subsequent daily telephone call.

Two recent randomized controlled trials examined the role of oral hypoglycaemic drugs in the management of gestational diabetes mellitus. In one study (6), 404 women with gestational diabetes mellitus with fasting hyperglycaemia were randomized to either insulin or glibenclamide (glyburide) starting at 11 to 33 weeks gestation. After treatment, both groups had similar rates of LGA infants (≥ 90th centile) (12% versus 13%), macrosomia (≥ 4 kg) (7% versus 4%), birth weight, congenital anomalies (both 2%), and hypoglycaemia (9% versus 6%). Satisfactory control was achieved in 88% of the sulphonylurea group and 84% of the insulin group. A second trial (7) randomized 751 women with gestational diabetes mellitus aged 18–45 years to metformin (up to 2500 g daily) or insulin, from 20 to 33 weeks’ gestation. The rate of the primary outcome (a composite of neonatal hypoglycaemia, respiratory distress, need for phototherapy, birth trauma, 5-min Agpar score less than 7, or prematurity) was 32% in the metformin group and 32.2% in the insulin group (relative risk, 0.99; 95% confidence interval, 0.80–1.23). The rate of other secondary outcomes (neonatal anthropometric measurements, maternal glycaemic control, maternal hypertensive complications, postpartum glucose tolerance) did not differ significantly between the groups. There were no significant adverse events associated with the use of metformin. Supplemental insulin was required in 46.3% of women taking metformin.

These two trials suggest that metformin (alone or with supplemental insulin) and glibenclamide may be effective and safe treatment options for women with gestational diabetes mellitus who meet the usual criteria for starting insulin (9). Data on safety of these agents in the first trimester are more limited, but the results of small retrospective studies and the increasing use of metformin in the polycystic ovarian syndrome are reassuring. Metformin, unlike glibenclamide, crosses the placenta and caution is necessary until safety in the first trimester has been evaluated more fully. Follow-up data focusing on both short- and long-term neonatal outcomes are required. Oral hypoglycaemic therapy for gestational diabetes mellitus is appealing, particularly in locations where insulin is not readily available and in the context of obesity. UK NICE guidelines now recommend metformin, and suggest that glibenclamide may be considered for the management of gestational diabetes mellitus if hypoglycaemic therapy is required (9).

When counselling a woman with advanced diabetic complications about pregnancy, the risks both to herself and to the baby must be discussed. The development or progression of retinopathy, nephropathy, or autonomic neuropathy during pregnancy may present a major management dilemma and requires the closest of collaboration between relevant colleagues. Diabetic ketoacidosis is a serious problem to the health and even the viability of the fetus, and every mother should be instructed in monitoring urinary ketones and on how to seek urgent advice if needed. Advanced microvascular disease may be associated with intrauterine growth retardation.

Pregnancy is a risk factor for the progression of diabetic retinopathy. The exact mechanism remains unknown, but relevant factors include rapid improvement in glycaemic control, altered haemodynamic properties of pregnancy, and immune-inflammatory components. Ideally, all diabetic patients should have a detailed eye examination prior to pregnancy. This permits relevant treatment (including photocoagulation) before instituting aggressive glycaemic management during pregnancy and is essential to reduce the risk of retinopathy progression during pregnancy. The UK NICE guideline recommends that a retinal assessment should be offered following the first antenatal appointment (if not performed in previous 12 months), at 28 weeks if the first assessment was normal, and, additionally, at 16–20 weeks if any retinopathy was present. It is our practice to screen at booking and then during each trimester, with more frequent screening if there has been poor control and retinopathy at booking, referring many of these patients to our ophthalmology service.

During pregnancy, the term is used to describe an heterogeneous group of patients with either microalbuminuria or varying degrees of proteinuria, with or without maternal hypertension or significant impairment in renal function. While recent retrospective data have reported perinatal survival rates of 95%, these vary with the stage of nephropathy and are accompanied by very high rates of pre-eclampsia (32–65%), preterm delivery (57–91%), and fetal growth restriction (12–45%). Tight control of blood glucose and blood pressure, before and during pregnancy, along with close fetal surveillance and timely delivery are needed to optimize pregnancy outcome. Drugs acting on the angiotensin system are teratogenic and should be discontinued before or possibly at conception: in patients with more severe degrees of nephropathy, pre-pregnancy. It is also necessary to balance the risk of a protracted period off ACE inhibitor therapy if conception is delayed. Methyldopa and labetolol are alternative agents. There is a need for trials that examine the optimal blood pressure during pregnancy, but the consensus is to aim for levels below 130/ 80 mm/Hg in subjects with diabetic nephropathy. Frequently, this requires more than one drug.

Screening for microalbuminuria should take place in all women with type 1 and type 2 diabetes at booking (if not performed in previous 12 months), and referral to a nephrologist should be considered if serum creatinine is 120 µmol/l or higher, or total protein excretion is more than 2 g/day. The literature would suggest that proteinuria levels increase from the first trimester to term across all stages of nephropathy, returning to pre-pregnant levels postpartum. Recent evidence has also shown that microalbuminuria in early pregnancy is associated with a fourfold increased risk of pre-eclampsia in pregnant women with type 1 diabetes.

Accurate dating of the pregnancy is an imperative and is best achieved by ultrasound examination at 8–10 weeks. The mother with type 1 diabetes will have already made contact with her diabetes team by this time and peri-conceptual control should be reviewed.

When fetal death occurs, it is usually in the final weeks of pregnancy in the context of poor glycaemic control, polyhydramnios, and/or accelerated fetal growth. In contrast, diabetic women with vasculopathy and/or pre-eclampsia may develop intrauterine growth restriction and fetal demise as early as the second trimester, probably related to placental vascular disease. The occurrence of fetal compromise or stillbirth when the fetus is normally grown or macrosomic is most likely to result from chronic fetal hypoxia and/or fetal acidaemia secondary to maternal/fetal hyperglycaemia and fetal hyperinsulinaemia. The UK CEMACH enquiry found that fetal surveillance was poor in up to 45% of cases (1).

The goal of obstetric surveillance is to identify fetuses at risk, in order to intervene in a timely and appropriate fashion to reduce perinatal morbidity and mortality. Given the limitations of the available tests and lack of rigorous scientific trials, all protocols used for fetal surveillance are empirical and all have limitations. The UK NICE guideline recommends ultrasound monitoring of fetal growth and amniotic fluid volume every 4 weeks from 28 to 36 weeks, and individualized monitoring of fetal wellbeing for women at risk of intrauterine growth restriction (IUGR) who are those with macrovascular disease or nephropathy (9). Tests of fetal wellbeing before 38 weeks are not recommended unless there is a risk of IUGR. In the absence of prospective trials, a similar schedule of fetal surveillance for women with insulin-requiring gestational diabetes mellitus associated with macrosomia or other risk factors would seem reasonable. Ultrasound measurement of abdominal circumference may also guide the clinician as to the need for insulin therapy in conjunction with the results from home blood glucose monitoring.

The primary objectives are to avoid the fetus dying in utero and the hazards of obstructed labour or shoulder dystocia associated with fetal macrosomia. As a consequence, Caesarean section rates for women with pregestational diabetes in most parts of the world are more than 50% (67% in the UK Survey, compared with the overall population rate of 24%; see Table 13.4.10.6.3). Iatrogenic prematurity has resulted in high rates of admission to neonatal intensive care in type 1 diabetes.

The indications for Caesarean section are often multiple and vary considerably with individual hospital policy. The rate tends to be lower in women with type 2, many of whom have had previous pregnancies at a time when glucose tolerance was normal. For the obstetrician, the major consideration influencing the mode of delivery remains the risk of birth injury. Current UK guidelines advise that women with diabetes should be offered elective delivery after 38 weeks, assuming no other significant factors have developed before this time (9), and it may be that this recommendation should also apply to women with gestational diabetes mellitus who are treated with insulin. An individualized approach to the timing and mode of delivery is essential. This is particularly so in pre-gestational and insulin-requiring gestational diabetes mellitus, where many factors need to be taken into consideration, including glycaemic control, diabetes complications, past obstetric history, fetal growth (macrosomia or IUGR), and the availability of health care resources in labour.

Pre-term labour can be particularly hazardous for the infant of the diabetic mother. Beta-sympathomimetic agents used to suppress uterine contractions, and corticosteroids used to accelerate fetal lung maturation, may result in significant and prolonged maternal hyperglycaemia, and even ketoacidosis, and the need for supplementary insulin must be anticipated. We have successfully developed an inpatient algorithm for steroid treatment in which an additional 30% of long-acting insulin is given on the evening of the first steroid dose, followed by a 50% increment of each insulin dose on days 2 and 3, a 30% increment on day 4, and a 20% increment on day 5, administered under close supervision (21).

The management of labour should follow standard practice as for the nondiabetic woman. Given the desire not to prolong pregnancy unduly, induction of labour is widely utilized and usually involves a combination of first prostaglandins followed, frequently, by oxytocin. Careful monitoring of progress is facilitated by the use of a partograph and continuous electronic fetal monitoring by cardiotocography. Management of diabetes during labour should follow an established protocol in a dedicated centre with a neonatal care unit equipped and staffed to deliver the most sophisticated level of care.

The literature would suggest that the maintenance of maternal blood glucose between 4–7 mmol/l during labour and delivery reduces the incidence of both neonatal hypoglycaemia and ‘fetal distress’ (9). Hourly capillary glucose measurements provide a ready guide to the success of management and the need for insulin adjustment.

There is no consensus over how best to achieve optimal maternal glucose control during labour and delivery. The first stage of labour is associated with a decrease in the need for insulin and a constant glucose requirement (22). In Belfast, for many years, we have successfully used an intravenous glucose/insulin infusion supplemented by additional insulin doses. Alternatively, some centres use a constant glucose infusion with insulin being infused separately by an infusion pump.

As soon as the cord is cut, the rate of the insulin infusion should be approximately halved, as insulin sensitivity returns to normal within minutes of the shutdown of the uteroplacental circulation. Regular capillary blood glucose readings and intravenous fluids are continued until the mother is able to eat normally.

Breastfeeding should be fully supported and dietetic advice at this stage is essential. The recommendation is that postpartum calorie requirements are increased from 25 kcal/kg per day for non-breastfeeding women to 27 kcal/kg per day for those women who wish to breastfeed (based on postpartum weight). Because maternal hypoglycaemia is most likely to occur within an hour of breastfeeding, this is an important time to measure blood glucose. In most cases, hypoglycaemia can be avoided by eating a small snack before breastfeeding, rather than making excessive adjustment of insulin doses.

For mothers with type 2 diabetes, insulin should usually be withdrawn at delivery with regular monitoring of capillary blood glucose. Those mothers who were previously on oral hypoglycaemic agents can resume these postpartum if they do not intend to breastfeed. If breastfeeding is desired, the question of whether or not to prescribe oral hypoglycaemic agents is controversial and will depend on individual circumstances and assessment of the risks and benefits. Currently, the UK NICE guidelines recommend that these drugs can be used during breastfeeding (9). If blood glucose levels, after careful monitoring, are not satisfactory with dietary measures alone, insulin should be reinstituted for a period. The mother may be able to return to oral hypoglycaemic therapy once breastfeeding has ceased.

For patients with gestational diabetes mellitus, insulin is usually discontinued following delivery. Capillary glucose monitoring should continue for several days to ensure a return of both fasting and postprandial values to the normal range. If these are satisfactory, blood testing can cease and the patient is booked for an assessment of glucose tolerance and review approximately six weeks after delivery.

The major complications include neonatal hypoglycaemia, macrosomia, and the risk of congenital anomalies and perinatal mortality. Other complications include the respiratory distress syndrome, hypocalcaemia, and polycythaemia.

Patient preference and health status are the two main factors that determine the choice of contraception for women with diabetes (22). Intrauterine contraceptive methods are particularly suited to women who do not wish to become pregnant within the next year. In women without vascular disease who wish to conceive sooner, combined (oestrogen and progesterone) hormonal contraception is considered safe. In general, the lowest dose (oestrogen ≤ 35 mcg) and potency formulation should be used as here the absolute increase in arterial thromobembolism is very low (1/12 000) and comparable to that among healthy users and nonusers. Women with longstanding diabetes, hypertension, microvascular, or cardiovascular complications, those who are less than 6 weeks postpartum, and probably also those who smoke and who have a BMI of 35 kg/m² or more, should not use oestrogen-containing contraceptives; progesterone-only methods (injections, implants, or tablets) may be used. Women with previous gestational diabetes mellitus share many risk factors for type 2 diabetes, but short-term prospective studies have not shown any adverse effects of low-dose/potency combined preparations on glucose or lipid metabolism. Barrier and ‘natural’ family planning methods are not ideal because of high failure rates. Following completion of childbearing, partner vasectomy and female sterilization are available. When faced with an unintended pregnancy, women with diabetes must receive additional guidance reflecting their increased risk for major congenital anomalies. Clinicians must understand the range of contraceptive options available and promote effective methods.

All women should be reviewed at 6–7 weeks after delivery, when they should be counselled regarding contraception and future pregnancy planning. Women with pre-existing diabetes (type 1 or type2) are referred back to their pre-pregnancy care providers. Women with gestational diabetes mellitus are offered a 75 g oral glucose tolerance test (OGTT) between 6 and 12 weeks, and non-attenders are followed up. Gestational diabetes mellitus is a recognized factor for the future development of gestational diabetes mellitus (recurrence rates range from 30% to 84%, related to body weight and ethnicity), type 2 diabetes, and cardiovascular disease. In addition, risk factors for the development of gestational diabetes mellitus are similar to those of type 2 diabetes and the metabolic syndrome. The postnatal visit, therefore, provides a unique opportunity to offer specific lifestyle advice, to screen for cardiovascular factors, and remind women of the need for early referral in the eventuality of future pregnancy. Those with normal or impaired glucose tolerance should have annual diabetes screening by fasting glucose, and, ideally, repeat OGTT 3-yearly in the community.

There is now increasing recognition that offspring of women with diabetes mellitus (whether type 1, type 2, gestational diabetes mellitus, or maturity-onset diabetes of the young) during pregnancy are at increased future risk of diabetes, obesity, and cardiovascular disease. Since these risk factors develop early in life, they place the offspring of women with diabetes at risk throughout adulthood, and at high risk for becoming obese during childhood and for developing diabetes or gestational diabetes mellitus by the time they reach child-bearing age. The major public health challenge for the future is to break this vicious cycle by the prevention of diabetes or impaired glucose tolerance until after child-bearing age, and, possibly, also by better control of diabetes during pregnancy (23). Recent studies have shown that weight management and physical activity with/without pharmacological measures are effective in delaying type 2 diabetes in women with previous gestational diabetes mellitus, and are cost effective (24).