14.10 Diabetes in pregnancy 2627

14.10 Diabetes in pregnancy 2627

ESSENTIALS Diabetes in pregnancy is predominantly either pre-​existing type 1 or type 2 diabetes mellitus, or gestational diabetes, the latter defined as diabetes or glucose intolerance first diagnosed during the preg- nancy. Gestational diabetes usually arises in the late second trimester and is common, affecting from 2–​6% to 15–​20% of pregnant women depending on diagnostic criteria and country of origin. Gestational diabetes is most commonly diagnosed on the basis of an oral glucose tolerance test performed at 24–​28 weeks’ gesta- tion by a plasma glucose at 0 minutes of more than 5.1 (or >5.6, depending on the authority) mmol/​litre, or at 120 minutes of more than 8.5 (or >7.8) mmol/​litre. Diabetes affects fertilization, implantation, embryogenesis, or- ganogenesis, fetal growth and development, and neonatal and perinatal morbidity and mortality. Preconception counselling is associated with improved pregnancy outcomes. Key aspects of periconceptional and pregnancy management include optimization of glycaemic control, stopping of medications contra-​indicated in pregnancy, avoidance of hypoglycaemia and diabetic ketoacidosis, and screening and management of diabetic complications. Risks to the fetus of maternal diabetes include congenital malfor- mations, fetal macrosomia, intrauterine growth restriction, and those from the increased incidence of maternal pre-​eclampsia. Long-​term adverse effects such as increased susceptibility to metabolic disease later in life are also recognized. Once pregnancy is confirmed, women with pre-​existing diabetes should be encouraged to book early in the pregnancy for manage- ment by a hospital-​based multidisciplinary team. Optimal blood glu- cose targets are a fasting capillary glucose of 5.3 mmol/​litre, a one hour after meal value below 7.8 mmol/​litre, or a two-​hours after meal value below 6.4 mmol/​litre. In some women with gestational diabetes this may be achieved with diet and exercise, but oral hypo- glycaemic agents (typically metformin or glibenclamide) and/​or in- sulin are often required. Women with diabetes should give birth in hospitals where 24 hour-​a-​day advanced neonatal resuscitation skills are available. Induction of labour or an elective caesarean section before 38 + 6 weeks of pregnancy is recommended if spontaneous labour has not occurred before then because of the increased risk of stillbirth. The effect of pregnancy on maternal glycaemic control ceases very quickly post-​partum, hence women with pre-​existing diabetes taking insulin should immediately revert to their pre-​pregnancy regimen after birth, but with a lower insulin dose. Introduction Diabetes mellitus (DM), whether pre-​existing, or new onset in preg- nancy has major long-​term implications for both maternal and fetal health. The overall incidence of diabetes in women of child- bearing age has increased over the last decade and is anticipated to continue to rise. One of the underlying causes of this is the rising prevalence of obesity, and this is impacting on the type of diabetes as well as the demographic factors of pregnant women with diabetes. Understanding the type of diabetes a woman has and the associated risk factors is key to the management of diabetes in pregnancy, as well as the implication for success or failure of potential interven- tion strategies. In addition, as the number of pregnant women with diabetes rises, there is a substantial impact on healthcare costs and resource utilization, emphasizing the need for the implementation of the best evidence-​based cost-​efficient treatment strategies. The physiology of glucose homeostasis during pregnancy In normal early pregnancy there is an increase in insulin sensitivity and a fall in fasting plasma glucose in non​obese women. This increase in insulin sensitivity accounts for the fall in insulin requirements seen in women with type 1 DM and their susceptibility to hypogly- caemia during the first trimester. Insulin sensitivity continues to in- crease at the beginning of the second trimester, after which there is a progressive reduction in insulin sensitivity until term, and by late pregnancy insulin sensitivity will have fallen by 40–​60%. The underlying mechanisms for these changes are complex and due in part to maternal and placental adipokines and cytokines, as well as to the upregulation of the growth hormone/​insulin-​like growth factor axis by high circulating concentration of placental 14.10 Diabetes in pregnancy Bryony Jones and Anne Dornhorst

Section 14  Medical disorders in pregnancy 2628 growth hormone. The contributions of the early increase in corpus luteal and then placental progesterone and 17β-​oestrogen by 10-​ and 30-​fold, respectively, on insulin resistance is unclear. The physiological reduction in insulin sensitivity in later preg- nancy facilitates the maternal-​fetal transfer of glucose and other nutrients across the placenta. Decreased insulin sensitivity de- creases the glucose uptake in maternal muscle, the principal site of whole-​body glucose disposal, thereby redirecting glucose to the fetus where it is the principal fetal substrate. This transfer of glu- cose is facilitated by GLUT-​1, a specific placental glucose trans- porter protein, and by the maternal-​fetal glucose concentration gradient. By late pregnancy the degree of maternal hepatic and peripheral insulin resistance results in an increased hepatic glu- cose output and free fatty acid release from maternal adipose stores that increase during early pregnancy due to the relative insulin resistance. This normal physiological fall in insulin sensitivity in pregnancy is greater in obese women and those with gestational diabetes mellitus (GDM). Maternal insulin secretion increases in response to the fall in insulin sensitivity, such that post-​prandial insulin secretion is in- creased 200–​250% from baseline levels by late pregnancy. This is ac- companied by an adaptive increase in the numbers of β cells in small islets, implying β-​cell neogenesis rather than duplication of β cells in existing islets. The increase in maternal insulin secretion helps promote maternal lipogenesis in early pregnancy and the deposition of adipose stores. If the maternal insulin response to pregnancy is inadequate, ma- ternal hyperglycaemia will develop, leading to the development of gestational diabetes mellitus. While most women who become hyperglycaemic are at risk of future type 2 DM, and start pregnancy with increased insulin resistance such that their insulin demands to maintain euglycaemia in pregnancy are extremely high, a smaller proportion of women are in the early preclinical phases of type 1 DM. The increased insulin demands of pregnancy account for the observation that there is a threefold increase in the new presentation of type 1 DM in pregnancy. Identifying this small group of women is important as early insulin treatment and close antenatal surveillance is important. Epidemiology and classification Diabetes in pregnancy is predominantly either pre-​existing type 1 or type 2 diabetes mellitus, or gestational diabetes, the latter de- fined as diabetes or glucose intolerance first diagnosed during the pregnancy. There are other rarer forms of pre-​existing diabetes, including monogenetic diabetes, recognition of which in preg- nancy will increase as awareness and the ability to diagnose them increases. While the overall incidence of women with diabetes of any type is increasing, the largest proportional rise is seen in women with pre-​existing type 2 diabetes mellitus and gestational diabetes. In 2012, overall prevalence of gestational diabetes across Europe was reported as between 2 to 6%, but this is dependent on the diag- nostic criteria used and the ethnic and demographic mix of the antenatal population, and higher values will be reported as more universal screening and testing for gestational diabetes is intro- duced (see later). Maternal gestational diabetes and obesity are both independently associated with adverse pregnancy outcomes, and their combination has a greater impact than either one alone. Obesity and gestational diabetes are also associated with adverse long-​term cardiovascular health in both the mother and the child in later life. In 2015 it was es- timated that in England and Wales around 35 000 pregnant women annually would have either pre-​existing diabetes (20%) or gesta- tional diabetes mellitus (80%). Gestational diabetes usually arises in the late second trimester as a consequence of the pregnancy-​induced changes in maternal carbo- hydrate metabolism. However, undiagnosed type 2 diabetes mellitus is increasingly seen, and it is essential to identify these women as their clinical management needs to be similar to those with pre-​ existing recognized type 2 and optimized as early in pregnancy as possible. The diagnosis of diabetes and type of diabetes in pregnancy The criteria for the diagnosis of diabetes outside of pregnancy can be made by any of the following; • a fasting plasma glucose of 7.0 mmol/​litre or above; • a two-​hour plasma glucose of 11.1 mmol/​litre or above during a 75 g oral glucose tolerance test taken as anhydrous glucose dis- solved in water; • a glycosylated haemoglobin (HbA1c) of 6.5% (48  mmol/​mol) or above; • a random plasma glucose of 11.1 mmol/​litre or above in the pres- ence of hyperglycaemic symptoms. The distinction between type 1 and type 2 diabetes mellitus can usually be made on the basis of clinical risk factors. Both types 1 and 2 are heterogeneous diseases in which the clinical presenta- tion and disease progression in adults may vary considerably and can appear to overlap. The diagnosis of type 1 diabetes mellitus, an autoimmune disease characterized by β-​cell destruction, can if necessary usually be confirmed by the presence of islet specific autoantibodies and low circulating or undetectable concentration of serum C-​peptide, a measure of endogenous insulin secretion. These islet cell autoantibodies include those directed against in- sulin, glutamic acid decarboxylase (GAD) (GAD65), the tyro- sine phosphatases IA-​2 and IA-​2b, and ZnT8. The diagnosis of type 2 diabetes mellitus remains a clinical diagnosis. Other than hyperglycaemia, including elevated HbA1c levels, there are no specific clinical diagnostic tests for type 2 diabetes mellitus. The lack of β-​cell autoimmunity, other autoimmune diseases, or other known causes for hyperglycaemia all favour a diagnosis of type 2. Most, but not all, patients with type 2 are overweight or obese, and many will have a family history of type 2 diabetes mellitus or other metabolic risk factors. See Chapter  13.9.1 for further discussion. In pregnancy, if the diagnosis of pre-​existing type 1 diabetes mellitus is uncertain, the presence of GAD antibodies would be ex- pected to confirm the diagnosis in around 80% of cases, and the de- cision to undertake further testing for antibodies will depend on risk assessment (Fig. 14.10.1).

14.10  Diabetes in pregnancy 2629 Gestational diabetes mellitus Definition and diagnosis Gestational diabetes mellitus is defined as a new diagnosis of dia- betes during pregnancy. There remains some controversy as how best to diagnose it. The Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study was designed to establish internationally agreed diagnostic criteria. In this landmark observational study more than 25 000 non​diabetic pregnant women were recruited between 2000 and 2006 from nine different countries. All women underwent a 75 gram, two-​hour OGTT between 24 and 32 weeks’ gestation, with healthcare providers blinded to the results unless sug- gestive of overt diabetes (fasting plasma glucose above 5.8 mmol/​l (105 mg/​dl), or the 2-​hour value above 11.1 mmol/​l (200 mg/​dl), or a random plasma glucose at 34–​37 weeks above 8.9 mmol/​l (160 mg/​ dl)). The fasting, 1-​ and 2-​hour plasma glucose results of the OGTT were analysed from 23 316 mother-​newborn pairs for their asso- ciation with four primary outcomes: macrosomia (corrected birth weight >90th centile), primary caesarean delivery, clinical neonatal hypoglycaemia, and hyperinsulinemia (as assessed by cord serum c-​peptide >90th centile for the study group as a whole). The analysis of the data showed the associations between the OGTT results at dif- ferent time points and the four primary outcomes were continuous, with no clear inflection points. The results of the HAPO analysis formed the basis of the International Association of the Diabetes and Pregnancy Study Groups (IAPDSG) 2014 criteria for gestational diabetes. The new criteria for gestational diabetes were based on the threshold of glu- cose value at each time point that gave a 1.75 likelihood of an adverse outcome when compared with the mean reference value at the three separate time points. The American Diabetes Association (ADA), World Health Organization, and International Diabetes Federation all subsequently endorsed the IAPDSG diagnostic criteria for gesta- tional diabetes, which included a 75 gm OGTT performed between 24 to 28 weeks. In contrast to the HAPO-​derived criteria, following a cost-​ effectiveness analysis of published intervention studies the National Institute for Health and Care Excellence (NICE) 2015 Diabetes and Pregnancy update proposed a higher fasting value and a lower two-​ hour oral glucose tolerance test value, with no one-​hour value for the diagnosis of gestational diabetes (Table 14.10.1). The NICE criteria also advocate the use of universal screening of five risk factors iden- tified at the initial antenatal visit, with a 75 gm OGTT between 26 to 28 weeks performed on only those women with one or more risk fac- tors, these being a body mass index (BMI) above 30 kg/​m2, a previous macrosomic baby weighing 4.5 kg or above, previous gestational dia- betes, a first-​degree relative with diabetes, and belonging to an ethnic family origin with a high background prevalence of diabetes. An earlier diagnostic OGTT (at 16 weeks) is recommended for women with a history of previous gestational diabetes, with repeat at 24–​28 weeks if the first test was normal. The five risk factors recom- mended by NICE is not exhaustive and leaves out other recognized risk factors, including increasing age and a history of polycystic ovary syndrome. The NICE-​recommended screening at booking potentially may miss up to 25% of women with gestational diabetes. Applying the new IAPDSG criteria to the original HAPO cohort diagnosed between 15 to 20% of parturients as having gestational diabetes, depending on country of origin. Gestational diabetes mellitus in early pregnancy There is currently no agreed strategy for identifying women not known to have diabetes who have significant hyperglycaemia in early pregnancy. Glycosylated haemoglobin (HbA1c) is less reliable No need to test for type 1 DM Test for Type 1 DM using GAD* antibodies Autoimmune thyroid disease, coeliac disease, Addison’s vitiligo, myasthenia gravis, pernicious anaemia. autoimmune hepatitis, Two or more risk factors for type 2 DM- BMI > 27kg/m2, family history of type 2 DM, age > 40 yrs., nonwhite, previous GDM 1st or 2nd degree relative with type 1 DM or other autoimmune diseases No risk factors for type 2 DM Strong family history of type 2 DM Fig. 14.10.1  Factors affecting the decision to test for GAD antibodies in a patient with diabetes in pregnancy.

Section 14  Medical disorders in pregnancy 2630 in pregnancy due to the higher turnover of erythrocytes, and fasting blood glucose falls in early pregnancy. However, an HbA1c meas- urement of ≥5.9% (≥41 mmol/​mol) at the initial antenatal clinic ap- pears to be a good discriminator for women who have significant glucose intolerance on a subsequent OGTT performed at around 20 weeks’ gestation, and with an increased risk of adverse pregnancy outcomes. This HbA1c value is lower than the diagnostic value for diabetes outside pregnancy (HbA1c 6.5% (48 mmol/​mol). Aetiology As described, pregnancy induces a state of relative decreased insulin sensitivity due to both maternal and placental factors. High circu- lating levels of maternal free fatty acids, adipokines, and cytokines may be factors contributing to this. Women who are obese and those who have gestational diabetes have increased levels of many of the adipokines and inflammatory cytokines associated with de- creased insulin sensitivity, including leptin, tumour necrosis factor-​α (TNF​α) and interleukin-​6, while the insulin sensitizing adipocyte cytokine adiponectin is reduced. The placenta is a major source of these cytokines, hormones, and inflammatory mediators, as well as of human placental lactogen and placental growth hormone. Pregnancy is a physiologically leptin-​resistant state and results in a significant upregulation of leptin, which in turn leads to hyperglycaemia. The human growth hormone (GH) gene cluster contains a single gene expressed solely in the anterior pituitary (GH1), plus four genes expressed in the placenta (SHL1, CSH1, GH2, CSH2). GH1 encodes solely pituitary GH, which is also a ligand for the maternal GH re- ceptor and stimulates the insulin-​like growth factor (IGF)-​1 axis. In pregnancy the 6–​8-​fold increase in placental growth hormone in- creases and replaces normal circulating pituitary growth hormone by 20 weeks’ gestation, and this is thought to decrease insulin sensitivity by inhibiting key components of the insulin-​signalling pathway. The other members of the GH gene cluster, CSH1 and CSH2, en- code for human placental lactogen which is a ligand for the maternal prolactin receptors. This can be detected in the maternal circulation by six weeks’ gestation and has been implemented in augmenting maternal insulin resistance and promoting the supply of fatty acids and glucose to the placenta. Changes in the placental microcirculation and placental dysfunc- tion leading to increased fetal complications have been attributed to the adverse diabetic metabolic milieu resulting from hypergly- caemia, hyperinsulinism, dyslipidaemia, and secretion of adipokines and inflammatory cytokines, causing increased oxidative stress and protein glycosylation. Rarer forms of diabetes Monogenic β-​cell diabetes is thought to be responsible for approxi- mately 2% of all diabetes diagnosed before the age of 45 years, about 80% of such patients being misdiagnosed as either type 1 or type 2 diabetes, reflecting lack of clinical awareness and limited clinical access to genetic testing. Clues to the diagnosis of monogenic forms of diabetes include lack of typical characteristics of type 1 diabetes (no islet cell auto- antibodies, low or no insulin requirement five years after diagnosis, persistence of measurable stimulated C-​peptide, absence of dia- betic ketoacidosis), or type 2 diabetes (lack of obesity, hypertension, dyslipidaemia), in the presence of a strong family history of diabetes. An autosomal dominant family history of diabetes may be caused by monogenic disruption of glucokinase (GCK), hepatic nuclear factor (HNF)-​1A, HNF-​4A, or HNF-​1B gene expression. The mo- lecular diagnosis of such forms of diabetes is important because it enables genetic counselling, predictive genetic testing in affected families, personalized tailoring of medication, and provides patient information regarding prognosis. GCK-​related diabetes Making the diagnosis of GCK-​related diabetes is important as it al- lows patients to be reassured that they have an inherited tendency to mild hyperglycaemia, which requires no specific glucose-​lowering treatment or self-​monitoring of blood glucose (outside of preg- nancy), because the condition is not associated with microvascular or macrovascular complications. It also has important implications for management of pregnancy, because in parturients with GCK-​ related diabetes the risk of fetal macrosomia depends upon the fetal GCK genotype. If the fetus has not inherited the GCK mutation it will respond to maternal hyperglycaemia with increased insulin produc- tion, leading to excess fetal growth (adding an additional approxi- mately 550–​700 g by term). Alternatively, if the fetus does inherit the GCK abnormality, it will sense the maternal hyperglycaemia as normal, produce normal amounts of insulin and have normal growth, and in this context the aggressive lowering of maternal glu- cose into the normal range may adversely affect fetal growth. As it is not currently possible to determine the fetal genotype non​invasively during pregnancy, the decision on whether to treat maternal hyperglycaemia in parturients with GCK-​related diabetes is made on the basis of monitoring fetal abdominal circumference, with a reading above the 75th percentile being the recommended threshold to start insulin and normalize maternal glycaemia. In the future, fetal genotyping using cell-​free fetal DNA from maternal plasma sampling during early pregnancy may assist in this context. Table 14.10.1  Recommendations for screening and oral glucose tolerance testing in pregnancy IAPDSG criteria NICE 2015 criteria Screening None Risk factors at 1st antenatal clinic

1. BMI >30 kg/​m2
2. Previous history of GDM
3. Previous macrosomic birth, ≥ 4.5 kg
4. Family history
5. Minority ethnic family origin with a high prevalence of diabetes 75 gm OGTT Universal at
   24–​28 weeks Selective testing 26–​28 weeks Based on the presence of risk factors Plasma glucose IAPDSG criteria one or more of the following thresholds
   be met or exceeded: NICE 2015 criteria one or more of the following thresholds be met or exceeded 0 minute 5.1 mmol/​litre (92 mg/​dl) 5.6 mmol/litre (100 mg/​dl) 60 minute 10 mmol/​litre (180 mg/​dl) 120 minute 8.5 mmol/​litre (153 mg/​dl) 7.8 mmol/​litre (140 mg/​dl) GDM, gestational diabetes mellitus.

14.10  Diabetes in pregnancy 2631 The impact of diabetes on the mother and fetus during pregnancy Diabetes affects all aspects of a woman’s reproductive life, including pregnancy. Diabetes, specifically maternal hypergly- caemia, affects fertilization, implantation, embryogenesis, or- ganogenesis, fetal growth, and development and neonatal and perinatal morbidity and mortality. Women with pre-​existing diabetes or those who develop diabetes during the pregnancy are at added risk of maternal and fetal complications during preg- nancy. Maternal diabetes and obesity also influence the risk of future obesity and diabetes in the child, through epigenetic fac- tors and fetal programming in utero. Preconception planning, surveillance, and screening throughout pregnancy by a multi- disciplinary specialist team can lessen the risks for both mother and fetus. Preconception counselling for women with diabetes Preconception counselling is associated with improved pregnancy outcomes. Uptake is highly dependent on sociodemographic fac- tors, and women with the greatest social deprivation scores, those with type 2 diabetes mellitus, and those from ethnic minority groups are less likely to access these services. As half of all pregnancies are unplanned, counselling women about pregnancy should form part of the ongoing care pathway for all women of childbearing age with diabetes. When any potentially teratogenic medication is prescribed as part of a woman’s routine diabetes care (such as an angiotensin converting enzyme (ACE)-​ inhibitor or a statin (HMG-​CoA reductase inhibitor)) women need to be counselled on the need to stop this prior to or as soon as preg- nancy is confirmed in an unplanned pregnancy. In addition, ad- vice should be given on the need to take high dose folic acid (5 mg daily) three months prior to pregnancy and continued for the first 12 weeks in pregnancy to reduce the risk of fetal neural tube defects. The key elements of preconception counselling and assessment are shown in Table 14.10.2. Management of diabetes peri-​conceptionally
and in pregnancy Optimizing glycaemic control Women need to be supported to achieve optimal glycaemic control peri-​conceptionally and throughout pregnancy to reduce the risk of miscarriage, congenital malformation, stillbirth, and neonatal death. Improved control prior to pregnancy also reduces the risk of deterioration of diabetic retinopathy and nephropathy during the pregnancy. Providing women with structured education around in- sulin management and dose adjustment of their rapid acting insulin to match their carbohydrate intake improves glycaemic control and lessens the risk of hypoglycaemia. Meta-​analysis of large data sets shows that congenital malfor- mations, preterm delivery and maternal hyperglycaemia in the first trimester of pregnancy are all reduced in women who receive preconception counselling. There is a strong positive association of fasting glucose or glycated haemoglobin in women with type 1 diabetes mellitus and type 2 diabetes mellitus and major fetal anom- alies, with multiple organ anomalies associated with the poorest glycaemic control. This association is so strong that 2015 NICE guidance is that women with an HbA1c above 86 mmol/​mol (10%) be strongly advised not to get pregnant. 2015 NICE guidance is to aim for an HbA1c below 48 mmol/​mol (6.5%) if this is achievable without causing problematic hypogly- caemia. Plasma glucose targets are fasting levels of 5–​7 mmol/​l on waking for women with type 1 diabetes mellitus and a plasma glu- cose level of 4–​7 mmol/​l before meals at other times of the day. To achieve this level of glycaemic control women with type 1 diabetes mellitus will require the use of a multiple dose insulin regime con- sisting of 4 or 5 insulin injections a day given as 1 or 2 basal insulin injections and 3 bolus rapid acting insulin injections to cover the main meals, or a continuous subcutaneous insulin infusion (CSII) pump. Recent technological advances around insulin pumps, con- tinuous glucose monitoring, and automated bolus insulin calcula- tors that help calculate pre-​meal insulin dosing depending on the amount of carbohydrate eaten, are all available to help women to achieve these targets. In the future, closed-​loop insulin delivery technologies, combining real-​time continuous glucose monitoring with CSII using a computer algorithm, will also be available, and early work in this field holds great promise that lower HbA1c levels will be achievable with less risk of hypoglycaemia. For women with type 2 diabetes mellitus, achieving an HbA1c below 48 mmol/​mol (6.5%) prior to pregnancy is easier, as these women will be at less risk of hypoglycaemia due to their relatively short history of diabetes and less dependency on insulin therapies. However, many such women will need to be started on an insulin prior to a planned pregnancy as all oral agents other than metformin will need to be stopped. Medication review Some antihypertensive and lipid lowering agents routinely used in the management of diabetes have been associated with an increased risk of congenital malformations when used in the first trimester. These include ACE-​inhibitors, angiotensin II receptor blockers, and the HMG-​CoA reductase inhibitors known collectively as statins. NICE guidance cautions against the use of thiazides and related Table 14.10.2  Key elements of preconception counselling Optimizing glycaemic control Medication review Screening and management of diabetes complications Advice regarding diet and weight Information on pregnancy risk to the mother, including particular risks of hypoglycaemia and DKA Increased risk of hypertensive disease in pregnancy Information on pregnancy risk to the baby—​risk of congenital abnormalities, preterm birth, and risk of macrosomia and growth restriction Importance of pregnancy planning and accessing multidisciplinary care in early pregnancy DKA, diabetic ketoacidosis.

Section 14  Medical disorders in pregnancy 2632 diuretics in pregnancy. Daily 5 mg of folic acid is recommended three months prior to any planned pregnancy. Screening and management of diabetic complications During preconception counselling women need to be informed of how a pregnancy may affect their own health, including the impact on pre-​existing micro and macrovascular disease. It is important to ensure that a retinal and renal assessment has been performed within the previous year, and to seek information on symptoms of peripheral and autonomic neuropathy, and all cardiovascular risk factors including hypertension and hypercholesterolaemia. Retinopathy Pregnancy can lead to new onset diabetic retinopathy or worsening of pre-​existing disease. Sight-​threatening diabetic retinopathy in pregnancy is rare, but proliferative diabetic retinopathy which accel- erates during pregnancy may not regress post-​partum. Screening for retinopathy is therefore recommended prior to pregnancy, during early pregnancy, and again at 28 weeks’ gestation. Any retinopathy detected requires treatment as it may deteriorate during pregnancy, and ongoing follow-​up during pregnancy and postnatally will be re- quired. Diabetic retinopathy is not a contraindication to rapid opti- mization of blood glucose control in women who present with a high HbA1c, nor is it a contraindication to pregnancy or vaginal birth. Renal function Physiological changes occur in renal function in pregnancy. Glomerular filtration rate, creatinine clearance, and protein excre- tion all increase. Renal function is usually preserved in women with diabetes who start pregnancy with normal renal function. Women with diabetic nephropathy (serum creatinine >1.5  mg/​dl or pro- teinuria of more than 3 g protein/​24 hours) have an increased risk of preeclampsia, preterm delivery, fetal growth restriction, peri- natal death, and a permanent deterioration in renal function (see Chapter 14.5). Assessment of renal function at the outset of preg- nancy is important as optimization of blood glucose as well as blood pressure before pregnancy may improve maternal and fetal out- comes. Furthermore, knowing baseline renal function (serum cre- atinine and urinary microalbumin excretion or urinary albumin/​ protein creatinine ratio) is essential for later comparison in case preeclampsia is suspected. Autonomic neuropathy Pre-​existing autonomic neuropathy is an important risk factor for poor glucose control, increased glucose variability and hypogly- caemia. It is also associated with gastroparesis. Although pregnancy does not worsen autonomic neuropathy, it can worsen the symp- toms of gastroparesis due to the hormonal and mechanical effects of pregnancy that independently slow intestinal motility. Patients with gastroparesis may develop severe nausea and vomiting, as well as malabsorption, that complicate the timing of insulin administration and contribute to high levels of glucose variability and post-​prandial hypoglycaemia. Women with autonomic neuropathy should have an anaesthetic assessment in the third trimester of pregnancy due to the associated increased anaesthetic risk. Maternal complications of diabetes in pregnancy Hypoglycaemia Hypoglycaemia in pregnancy is a significant problem, affecting up to 70% of women with pre-​existing diabetes and associated with ex- cess mortality. Undoubtedly the strict glycaemic targets women are expected to achieve before and during pregnancy are an important factor. Awareness of hypoglycaemic symptoms is reduced in preg- nancy and women require assessment and education around hypo- glycaemia avoidance. This is particularly important for those with initial poor glycaemic control who undergo rapid intensification of their insulin management. Hypoglycaemia is particularly prevalent in the first half of preg- nancy, when insulin requirements actually fall, and the first few weeks post-​partum (Fig. 14.10.2). Women who have frequent 16 14 12 10 8 6 4 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 Gestational week Number of events Fig. 14.10.2  Hypoglycaemic episodes during pregnancy in women with type 1 diabetes. Reproduced from Nielsen LR et al. (2008). Hypoglycemia in pregnant women with type I diabetes. Diabetes Care, 31(1), 9–​14. Copyright © 2008, American Diabetes Association.

14.10  Diabetes in pregnancy 2633 hypoglycaemic events may benefit from “smart” diabetes tech- nologies, including insulin pumps with continuous glucose moni- toring that allows the glucose monitor to temporarily switch off
the insulin pump. Diabetic ketoacidosis Diabetic ketoacidosis is a rare but serious complication in preg- nancy and has an associated fetal mortality of up to 10%. All women with type 1 diabetes mellitus should be provided with blood ketone testing strips and a meter, instructed on their use, and advised to test for ketonaemia if they become hyperglycaemic or unwell, and be aware of the indications to seek urgent medical advice. Diabetic ketoacidosis can occur at near normal blood glucose levels during pregnancy, hence it should be excluded in any preg- nant woman with pre-​gestational diabetes and persistent nausea and vomiting. The risk of diabetic ketoacidosis increases in late preg- nancy due to increased insulin demands and enhanced lipolysis. The treatment of diabetic ketoacidosis in pregnancy is the same as for non​pregnant patients. Fetal assessment should occur following ma- ternal stabilization, noting that fetal heart rate abnormalities usu- ally correct with maternal treatment. Pre-​existing diabetic complications Pre-​existing complications differ in women with type 1 or type 2 diabetes mellitus, or gestational diabetes. Women with type 1 dia- betes mellitus are likely to have had diabetes for longer than those with type 2 diabetes mellitus, and established diabetic microvascular disease increases with diabetes duration. Retinopathy, nephropathy, and neuropathy, both peripheral and autonomic, are all more common in women with type 1 diabetes mellitus. Women with type 2 diabetes mellitus are more likely to have a duration of diabetes of less than 10 years and usually less than 5 years, but this may change in the future due to the increasing prevalence of type 2 diabetes mellitus among younger women. All women with diabetes, particularly those with longstanding type 1diabetes mellitus and older obese women with type 2 diabetes mellitus and gestational diabetes, are at an increased risk of diabetic macrovascular disease, including coronary heart disease. Diabetic macrovascular disease remains uncommon in pregnancy, but when it precedes or occurs for the first time in pregnancy, both maternal and fetal outcomes are poor. As the prevalence of macrovascular complications increases with duration of diabetes mellitus, and the numbers of women choosing to delay childbirth into their thirties and forties increases, the prevalence of macrovascular complica- tions in the pregnant population are likely to increase. Other maternal complications The potential maternal complications from conception to the post-​ partum period are listed in Table 14.10.3. Risks to the fetus Although maternal glucose crosses the placental barrier, insulin does not and, consequently, increases in maternal glucose stimulate fetal insulin production and result in hyperplasia of fetal insulin Table 14.10.3  Maternal complications of diabetes from conception to the post-​partum period Time Issue Type 1 DM Type 2 DM GDM Conception Fertility Subfertility due to poor control Delayed menarche Early menopause Autoimmune ovarian failure Increased prevalence of weight related amenorrhoea (anorexia) Subfertility if associated with polycystic ovarian syndrome 1st trimester 0–​12 wks Miscarriage Increased Poor glycaemic control Increased if due to PCO Hypoglycaemia High risk of severe recurrent hypoglycaemia Loss of hypoglycaemic awareness 2nd trimester 13–​28 wks Retinopathy Worsening retinopathy Hypoglycaemia High risk of severe recurrent hypoglycaemia Autonomic neuropathy Gastroparesis Nephropathy Worsening renal function Increased proteinuria 3rd trimester 29–​40 wks Preeclampsia Pregnancy-​induced hypertension Operative birth Birth trauma Failed induction of labour Operative birth including caesarean section Birth trauma Post-​partum Post-​partum haemorrhage Increased risk Long term, 1–​10 years Expected improvement of any pregnancy related deterioration in retinopathy and nephropathy Long-​term risk of coronary vascular disease, and diabetic microvascular complications Recurrent GDM Future type 2 DM Increased risk of CVD CVD, cardiovascular disease; DM, diabetes mellitus; GDM, gestational diabetes mellitus.

Section 14  Medical disorders in pregnancy 2634 sensitive tissues. Maternal hyperglycaemia also enhances produc- tion of hPG, fetal IGF, and TNF​α. Like insulin, these act as fetal growth factors and result in accelerated fetal growth, macrosomia, and enlargement of the heart and liver. Congenital malformations As previously discussed, diabetes mellitus is also associated with an increased risk of congenital malformation which is dependent upon glycaemic control (Fig. 14.10.3). When women with diabetes have a normal HbA1c in early pregnancy, congenital malformation rates approach that of the background population (risk of around 2%), but as HbA1c values increase from two to eight standard deviations above the normal range the malformation risk rises from 3 to 10% respectively. Diabetes mellitus is associated with a wide variety of malformations involving the cardiovascular system (e.g. transposition of the great vessels, ventricular septal defect), the central nervous system (e.g. anencephaly, spina bifida, hydrocephaly and holoprosencephaly), the genitourinary system, and the skeleton. The exact mechanism of the embryopathy remains uncertain, but there is evidence for contri- butions from increased oxidative stress due to elevated superoxide dismutase activity, reduced levels of myoinositol and arachidonic acid, and inhibition of the pentose phosphate shunt pathway. Fetal macrosomia Maternal weight gain, even without diabetes, predisposes to fetal macrosomia, hence obese pregnant women are advised to avoid un- necessary weight gain during pregnancy. The causes of macrosomia are not fully understood. The Freinkel hypothesis attributes excessive fetal insulin to increased transport of maternal fuel to the conceptus as the predominant cause of fetal macrosomia. However, fetal growth is complex and influenced by maternal, paternal, and fetal genes and factors, uterine environment, and maternal and fetal hormonal status. Infants of diabetic mothers have an additional influence of maternal fluctuations in glycaemia, as a hyperinsulinized fetus can cause a glu- cose steal phenomenon in late pregnancy, reducing maternal insulin requirements and precipitating maternal hypoglycaemia. Macrosomia is associated with risks for birth for both the fetus and the mother. It is a risk factor for shoulder dystocia which can be difficult to predict but may occur more commonly in fetuses of diabetic mothers with an estimated weight of more than 4 kg, and up to 50% in fetuses weighing 4.5 kg or more. Operative delivery and caesarean section are increased in both preexisting and gesta- tional diabetes, independent of the effect of birth weight, potentially related to placental dysfunction leading to fetal distress in labour. Intrauterine growth restriction Pre-​existing diabetes is also associated with intrauterine growth re- striction (IUGR), which is diagnosed when ultrasound-​estimated fetal weight is below the 10th percentile for gestational age, implying a possible pathological process causing the low fetal weight. IUGR in pregnancy complicated by pre-​existing diabetes is usually caused by placental dysfunction related to maternal vasculopathy. Prevention of IUGR should ideally start before pregnancy. Strict gly- caemic control and intensive treatment of nephropathy and hyper- tension are essential. Low-​dose aspirin initiated before 16 gestational weeks may also reduce IUGR risk in women with vasculopathy. Umbilical and other fetal Doppler studies can guide diagnosis and surveillance of fetuses with IUGR as these may reflect placental re- sistance and function. Pre-​eclampsia There is an increased risk of pre-​eclampsia in women with pre-​ existing diabetes, especially in those with diabetic nephropathy. Preterm delivery may be required, balancing the risks of delivery and neonatal short term and long-​term effects of prematurity. Delivery management and the timing of delivery is made according to maternal well-​being, the degree of glycaemic control, the pres- ence of diabetic complications, growth of the fetus, evidence of uteroplacental insufficiency, and the results of fetal surveillance. Perinatal mortality remains high among infants of mothers with type 1 and type 2 diabetes mellitus. Although high glucose levels have been implicated, the exact mechanism remains unclear. Neonatal and longer-​term effects The impact of maternal hyperglycaemia persists into the neonatal period with an increased risk of neonatal hypoglycaemia, respiratory distress, polycythaemia, hypocalcaemia, and hyperbilirubinaemia. The severity of these fetal metabolic problems correlates with levels of C-​peptide, insulin, and erthyropoietin in amniotic fluid and fetal cord blood, but these are not measured routinely. There is also a growing body of evidence suggesting a longer-​term effect of poor glycaemic control on neonatal outcome. Fetal pancreatic islet cells subjected to hyperglycaemia in utero appear to contribute to an increased susceptibility to metabolic disease later in life, such as obesity, type 2 diabetes mellitus, and the metabolic syndrome. The risks of developing type 2 diabetes mellitus are complex and poorly understood, but infants with both decreased and increased birth weight are at increased risk of developing type 2 diabetes mellitus when compared to those born with a normal birth weight. Management of diabetes in pregnancy Pre-​existing diabetes Once pregnancy is confirmed, women with pre-​existing diabetes should be encouraged to book early in the pregnancy with an early dating scan at 11–13 weeks. Multidisciplinary, consultant-​led, 30 25 20 15 10 5 0 Women without diabetes 6.1–7.7 7.8–10.0

10.0 HbA1c (%) Congenital malformation (%) Fig. 14.10.3  Risk of congenital malformation dependent upon glycaemic control. Reproduced from BMJ, Taylor and Davison, 334: 742–​745, copyright © 2007, with permission from BMJ Publishing Group Ltd.

14.10  Diabetes in pregnancy 2635 hospital-​based care is usually the most appropriate level of care, with the multidisciplinary team (MDT) comprising of endocrinologists, diabetes nurses, dieticians, specialist midwives, and obstetricians. In addition to routine antenatal care and screening tests, regular sur- veillance (usually every two weeks) is suggested to assess blood glu- cose levels, screen for fetal anomaly, and assess fetal growth (Table 14.10.4). All women should have contact details and telephone ac- cess between clinic visits with a member of the MDT team. The optimal blood glucose targets remain the same throughout pregnancy and are similar regardless of the type of diabetes, namely a fasting capillary glucose of 5.3 mmol/​litre, a one hour after meal value below 7.8 mmol/​litre, or a two-​hours after meal value below 6.4 mmol/​litre. However, these targets may need to be individual- ized and increased for women at risk of hypoglycaemia. Women with type 1, type 2, or gestational diabetes on multiple dose insulin regimen or pump therapy are advised to perform home glucose monitoring fasting, one hour pre-​meals, one to two hours post-​meal, and before bed. Women with gestational diabetes man- aged with diet and exercise alone, or oral therapy, only need to test fasting and one to two hours post-​meals. Gestational diabetes Women with gestational diabetes should be taught home glu- cose monitoring to ensure that their glycaemic targets are met throughout the duration of pregnancy. The aim of treatment is to achieve euglycaemia (while avoiding hypoglycaemia), as this has been shown to benefit both maternal and fetal outcomes. The primary intervention for women diagnosed with gesta- tional diabetes but with a fasting blood glucose below 6 mmol/​l is dietary counselling in combination with physical activity and self-​ monitoring of blood glucose. In women who have a fasting blood glucose on OGTT above 7.0 mmol/​l, diet and exercise alone would not be expected to lower this value to 5.3 mmol/​litre within an ac- ceptable time period, and therefore the prompt initiation of insulin therapy, with or without metformin, is advised. For women with a fasting blood glucose on OGTT between 6.0 to 7.0 mmol/​l, imme- diate treatment with insulin therapy, with or without metformin, should be considered if there is clinical evidence of macrosomia or polyhydramnios. Dietary advice All patients with gestational diabetes should be referred to specialist dieticians, the key elements of dietary advice being to substitute complex for simple carbohydrates and increase dietary fibre. The need for pharmacological therapy for gestational diabetes depends entirely on the severity of the glucose intolerance, in addition to fetal growth as assessed on prenatal ultrasound. Diet alone will maintain the fasting and postprandial blood glucose values within the target range in approximately half of women with gestational diabetes. Exercise The role of exercise in gestational diabetes may be even more im- portant than in women with pre-​existing diabetes, given that ex- ercise in some women may lessen the need for medical therapy. Moderate exercise in women with gestational diabetes is well tol- erated and has been shown to lower maternal glucose levels. Using exercise after a meal in the form of a brisk walk may blunt the post- prandial glucose excursions sufficiently to avoid the need of medical therapy. Establishing a regular routine of modest exercise during pregnancy may also have long-​lasting benefits for women with ges- tational diabetes due to their appreciable risk of developing type 2 diabetes in the future. Pharmacological treatments If diet and exercise are insufficient, oral hypoglycaemic agents and/​ or insulin will be required. Insulin does not cross the placenta, but is more difficult to administer, and the risk of hypoglycaemia and weight gain is greater than with oral hypoglycaemic agents. Oral hypoglycaemics, particularly metformin and glibenclamide, have been demonstrated to be safe and efficacious in pregnancy. The advantages of metformin include ease of use, low risk of hypo- glycaemia, and limitation of maternal weight gain and weight re- tention post-​partum. Although there is a theoretical concern that metformin crosses the placenta, recent evidence has demonstrated acceptable short-​term outcomes, with longer term outcomes yet to be clearly defined. Use of metformin was recommended by NICE in 2015, but it does not currently have a UK licence for use in diabetes in pregnancy. Glibenclamide is the only sulfonylurea that has been studied in a large randomized control trial performed in women with gestational diabetes. Maternal glycaemic control, macrosomia, neonatal hypoglycaemia, and neonatal outcomes were not different between women managed with glibenclamide or insulin. There is either very limited or no safety or outcome data on women with gestational diabetes being treated with other oral agents, including pioglitazone, metglitinides, acarbose, and incretins. If blood glucose control remains above target following treatment with the oral agents, then a recombinant human insulin should be considered. It is common to add a short and medium duration acting insulin to achieve better 24-​hour control while continuing the Table 14.10.4  Management plan for diabetes in pregnancy Early pregnancy 8–​9 weeks Booking visit Dating scan at 11–13 weeks HbA1C, renal profile, assessment of proteinuria Baseline BP Referral for retinopathy screen 12 weeks Diabetes review First trimester ultrasound scan 14–​16 weeks Routine antenatal care 18–​20 weeks Anomaly scan including fetal echocardiography 22 weeks Diabetes review 24 weeks Ultrasound fetal growth 26 weeks 28 weeks Ultrasound fetal growth 30 weeks 32 weeks Ultrasound fetal growth 34 weeks 36 weeks Ultrasound fetal growth 37 weeks 38 weeks Aim for delivery of women with pre-​existing diabetes Consider delivery of women with complex gestational diabetes 40 weeks Offer induction of labour in women with gestational diabetes

Section 14  Medical disorders in pregnancy 2636 metformin therapy. Glibenclamide is usually discontinued once in- sulin is required. There is evidence suggesting an association between vitamin D deficiency/​insufficiency and gestational diabetes, although the mo- lecular or cellular mechanisms for this association is unclear. It is currently not known whether vitamin D supplementation can re- duce the risk of developing gestational diabetes and/​or improve glycaemic control in diabetic pregnant women with vitamin D defi- ciency/​insufficiency. Obstetric matters Timing, mode, and place of birth Women with diabetes should give birth in hospitals where 24 hours-​ a-​day advanced neonatal resuscitation skills are available. Women with type 1 or type 2 diabetes mellitus and no other complications should be advised that birth between 37 + 0 weeks and 38 + 6 weeks of pregnancy lessens their heightened risk of stillbirth. Induction of labour or an elective caesarean section before 38 + 6 weeks of pregnancy is recommended if spontaneous labour has not occurred before then. Birth before 37 + 0 weeks may be advisable if there are metabolic or maternal or fetal indications. Antenatal steroids for fetal lung maturation can be used if a preterm delivery is anticipated, but close glycaemic monitoring is required alongside careful insulin titration. Management of diabetes in labour During labour and delivery, most women with pre-​existing diabetes should be managed with a sliding scale of intravenous insulin and dextrose infusion to maintain capillary plasma glucose between 4 and 7 mmol/​litre to lessen the risk of neonatal hypoglycaemia. Hourly capillary plasma glucose should be performed. Women with gestational diabetes using diet, oral hypoglycaemic agents, or small doses of insulin can cease the medications in labour, with continued regular assessment of maternal capillary glucose. Women with gestational diabetes requiring large doses of insulin may require an insulin sliding scale in labour. Post-​partum care The effect of pregnancy on maternal glycaemic control ceases very quickly post-​partum, hence women with pre-​existing diabetes taking insulin should immediately revert to their pre-​pregnancy regimen after birth, but with a lower insulin dose. It is important for women and healthcare staff to be prepared for the potential increased risk of hypoglycaemia in the immediate post-​partum period. Breastfeeding, lack of sleep, and the need for maternal adjustment to not requiring ‘tight glycaemic’ control as during the pregnancy, all contribute to this increased risk, which can be reduced if post-​partum women take 25–​30% less than their pre-​pregnancy insulin dose. Women taking insulin who breastfeed should test their blood glucose before and after breastfeeding and be encouraged to have a snack before and have one available when feeding. Due to the heightened risk of hypoglycaemia when breastfeeding, women need to be warned of the potential dangers to the newborn when breast- feeding in bed. Healthcare staff and patients should be trained to monitor for both neonatal and maternal hypoglycaemia. Women with type 2 diabetes mellitus on metformin prior to and during pregnancy can safely continue this post-​partum and during breastfeeding. The concentration of metformin in breast milk is low, and infant exposure to metformin has been reported to range be- tween 0.3 to 1.1% of the weight-​normalized maternal dose. Breastfeeding appears to confer an advantage to all women, redu- cing incidence of developing type 2 diabetes mellitus in those with and those without gestational diabetes. Breastfeeding has also been linked to a decreased risk of the infant developing obesity and im- paired glucose tolerance in later life. Women with gestational diabetes can usually cease all diabetes medication post-​partum, with most reverting to normal glucose levels soon after birth, although 5–​10% will fulfil the criteria for type 2 diabetes. Although screening women with prior gestational dia- betes for subsequent type 2 is accepted, there is a lack of consensus as to how and when this should be done. The ADA recommends using a 75 gm OGTT to screen for persistent diabetes 6–​12 weeks’ post-​partum, and lifelong screening for development of diabetes or pre-​diabetes at least every three years. In the United Kingdom, the 2015 NICE guidelines recommend a laboratory glucose prior to dis- charge, and a fasting glucose 6–​13 weeks post-​partum or a HbA1C after 13 weeks, with an annual fasting plasma glucose or HbA1C thereafter. All women with prior gestational diabetes remain at risk of diabetes in the future, with up to half developing type 2 diabetes mellitus within 10 years. A programme of diet and exercise can im- prove this risk, hence dietary and lifestyle modifications, including regular exercise, should be advocated in the long term. Post-​partum care should include advise on family planning and the need for contraception that allows women time to optimize their health prior to any future pregnancy. Areas of uncertainty, controversy, and future developments The maternal and neonatal obstetric outcomes in women with pre-​ existing diabetes have the potential to improve with continuing advances in diabetes management, fetal surveillance, and neonatal care. However, a significant improvement in pregnancy outcomes will require more women with pre-​existing diabetes to plan their pregnancy and participate in preconception counselling to optimize their glycaemic control and identify and manage proactively any dia- betic complications prior to pregnancy. Preconception counselling improves both maternal and fetal short-​ and long-​term outcomes. It will become increasingly important to identify women with gestational diabetes, not only to minimize the impact of gesta- tional diabetes on the pregnancy, but also for the future health of the child. Women with gestational diabetes represent the future population of parous women with diabetes, mostly type 2. This po- tential provides an opportunity for targeted prevention strategies to curb the current rise in obesity and type 2 diabetes. The association

14.10  Diabetes in pregnancy 2637 of obesity and glucose intolerance in the offspring of women with pre-​existing diabetes mellitus or gestational diabetes provides an opportunity for intergenerational risk prevention for future obesity and diabetes. FURTHER READING Bellamy L, et al. (2009). Type 2 diabetes mellitus after gestational dia- betes: a systematic review and meta-​analysis. Lancet, 373, 1773–​9. Confidential Enquiry into Maternal and Child Health (2005). Pregnancy in women with type 1 and type 2 diabetes in 2002–​03, England, Wales and Northern Ireland. London, CEMACH. Crowther CA, et al. (2005). Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med, 352, 2477–​86. Garcia-​Patterson A, et  al. (2010). Insulin requirements throughout pregnancy in women with type 1 diabetes mellitus: three changes of direction. Diabetologia, 53, 446–​51. Guerin A, Nisenbaum R, Ray JG (2007). Use of maternal GHb concen- tration to estimate the risk of congenital anomalies in the offspring of women with prepregnancy diabetes. Diabetes Care, 30, 1920–​5. Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study Group (2008). Hyperglycemia and adverse pregnancy outcomes.
N Engl J Med, 358, 1991–​2002. International association of diabetes and pregnancy study groups (2010). International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care, 33, 676–​82. Macintosh MC, et  al. (2006). Perinatal mortality and congenital anomalies in babies of women with type 1 or type 2 diabetes in England, Wales, and Northern Ireland:  population-​based study. BMJ, 333, 177. National Institute for Health and Care Excellence (NICE) (2015). Diabetes in Pregnancy:  Management from Preconception to the Postnatal Period. https://​www.nice.org.uk/​guidance/​ng3 Rowan JA, et al. (2008). Metformin versus insulin for the treatment of gestational diabetes. N Engl J Med, 358, 2003–​15. Secher AL, et al. (2013). The effect of real-​time continuous glucose monitoring in pregnant women with diabetes: a randomized con- trolled trial. Diabetes Care, 36, 1877–​83. Stewart ZA, et al. (2016). Closed-​loop insulin delivery during preg- nancy in women with type 1 diabetes. N Engl J Med, 375, 644–​54. Tennant PW, et  al. (2014). Pre-​existing diabetes, maternal glycated haemoglobin, and the risks of fetal and infant death: a population-​ based study. Diabetologia, 57, 285–​94. Tripathi, A., et al. (2010). Preconception counseling in women with diabetes: a population-​based study in the north of England. Diabetes Care, 33, 586–​8. Zhao P, et  al. (2016). Maternal gestational diabetes and childhood obesity at age 9–​11: results of a multinational study. Diabetologia, 59, 2339–​48.