Among women with diabetes, the worst pregnancy outcome is seen in the subgroup of women with diabetic nephropathy. Development of severe preeclampsia that leads to early preterm delivery is frequent. Predictors and pathophysiological mechanisms for the development of preeclampsia among women with diabetes and observational studies that support antihypertension treatment for pregnant women with microalbuminuria or diabetic nephropathy preventing preeclampsia and early preterm delivery are presented here. Obtaining and maintaining strict glycemic control before and during pregnancy is paramount to prevent preterm delivery. The cornerstones of diabetes management are appropriate diet and insulin, although the risk of severe hypoglycemia always needs to be taken into account when tailoring a diabetes treatment plan. Pathophysiological mechanisms of the increased risk of hypoglycemia during pregnancy are explored, and studies evaluating the use of insulin analogs, insulin pumps, and continuous glucose monitoring to improve pregnancy outcomes and to reduce the risk of severe hypoglycemia in pregnant women with type 1 diabetes are reported. In addition to strict glycemic control, other factors involved in fetal overgrowth are explored, and restricting maternal gestational weight gain is a promising treatment area. The optimal carbohydrate content of the diet is discussed. In summary, the lessons learned from this clinical research are that glycemic control, gestational weight gain, and antihypertension treatment all are of importance for improving pregnancy outcomes in pregnant women with preexisting diabetes. An example of how to use app technology to share the recent evidence-based clinical recommendations for women with diabetes who are pregnant or planning pregnancy is given.
Introduction
Among pregnant women with preexisting diabetes, the worst pregnancy outcome occurs in the subgroup of women with diabetic nephropathy (1,2). In the late 1990s, the prevalence of microalbuminuria and diabetic nephropathy among pregnant women with diabetes was 10% and 5%, respectively. The development of preeclampsia with severe hypertension and nephrotic proteinuria was common in these women and often led to early preterm delivery (1–3). Outside of pregnancy, strict antihypertension treatment was proven to be efficient in controlling hypertension and proteinuria. However, whether antihypertension treatment during pregnancy in women with diabetes and kidney involvement could improve pregnancy outcome was not known.
Preterm delivery, before 37 weeks, occurred in 40% of women with preexisting diabetes, whether or not they had kidney involvement (3). This prevalence was approximately four times higher compared with the background population (3,4).
Strict glycemic control with appropriate diet and human insulin was the cornerstone of treatment to prevent preterm delivery in women with diabetes. However, strict glycemic control in women with type 1 diabetes was associated with a high risk of severe hypoglycemia (5).
At present, fetal overgrowth is the most significant fetal complication. Approximately half of women with preexisting diabetes deliver infants that are large for gestational age (LGA) (4,6). Poor glycemic control is a well-known predisposing factor for LGA, but other factors predisposing to fetal overgrowth need to be considered and explored.
At the Center for Pregnant Women with Diabetes in Copenhagen, we have a long tradition of clinical research that leads to optimized clinical care (7). Each year, approximately 100 women with type 1 diabetes and 50 women with type 2 diabetes are followed during pregnancy. Our clinical care and research involves a fruitful collaboration in a team involving endocrine and obstetric expertise, including doctors, nurses, midwives, laboratory technicians, dietitians, and medical students.
This overview covers the clinical research performed in Copenhagen over the past two decades in order to improve the pregnancy outcome in all women with preexisting diabetes and, in particular, women suffering from microalbuminuria and/or diabetic nephropathy.
Microalbuminuria and Diabetic Nephropathy in Pregnant Women With Diabetes
Twenty years ago development of preeclampsia with severe hypertension and nephrotic proteinuria leading to early preterm delivery was frequently observed among women with type 1 diabetes complicated by diabetic nephropathy (1,2). In nonpregnant subjects, antihypertension treatment with blockers of the renin-angiotensin system has for years been a well-established cornerstone in the treatment of diabetic nephropathy and microalbuminuria to control hypertension and proteinuria and thereby prevent the progression of the kidney disease (8–10). However, the possible benefit of antihypertension treatment during pregnancy was still a matter of debate. Although antihypertension treatment might prevent severe hypertensive complications in pregnant women without diabetes, it might also, in theory, impair fetal blood flow and fetal growth (11). Therefore, in the late 1990s, our treatment strategy was only to initiate or intensify antihypertension treatment if blood pressure exceeded 140/95 mmHg in pregnant women with type 1 diabetes.
Pathophysiological studies have shown that preeclampsia is often preceded by higher blood pressure (12), elevated urinary albumin excretion (12), endothelial dysfunction (12), impaired maximal vasodilatation of the arterial vessels (12), and increased levels of components of the renin-angiotensin system (13), markers of cardiac overload (14), and antiangiogenic factors (15,16) in women with diabetes in early pregnancy. The majority of these factors can be modulated by antihypertension treatment. To prevent the rise in blood pressure and urinary albumin excretion leading to preeclampsia, tight antihypertension treatment during pregnancy is therefore, theoretically, beneficial in these women.
In line with this, a case of a normotensive woman with type 1 diabetes and microalbuminuria touched my heart. She had been treated with antihypertension treatment for 8 years up to pregnancy without progression from microalbuminuria to diabetic nephropathy. According to our antihypertensive strategy at that time, antihypertensive treatment was discontinued in early pregnancy. Shortly after, she developed hypertension and severe proteinuria and met the criteria for severe preeclampsia and had to be delivered preterm at 28 weeks. She gave birth to a girl born with cerebral bleeding. This case in combination with several similar cases with poor outcomes in women with preexisting microalbuminuria or diabetic nephropathy led us to investigate the prevalence of preeclampsia in relation to the occurrence of kidney involvement. In a cohort of 240 women, the prevalence of preeclampsia was 6% in women with normal urinary albumin excretion, 42% in women with microalbuminuria (30–300 mg/24 h), and 64% in women with diabetic nephropathy (albumin excretion >300 mg/24 h) in early pregnancy (3). Preeclampsia in women with microalbuminuria or diabetic nephropathy was often severe, with early development leading to preterm delivery before 34 weeks. In the year 2000, we changed our antihypertension strategy in women with microalbuminuria or diabetic nephropathy to initiate or increase antihypertension treatment if office blood pressure exceeded 140/90 mmHg or albumin excretion exceeded 2,000 mg/24 h. If the women already were on antihypertension treatment, this was changed to antihypertension agents well tolerated in pregnancy such as methyldopa, labetalol, or nifedipine (11).
In a subsequent unselected cohort of 20 normotensive pregnant women with type 1 diabetes and microalbuminuria treated with this strategy, a significant reduction in early preterm delivery before 34 weeks was seen compared with the previous cohort where antihypertension treatment was given less rigorously (Table 1) (17,18). However, the prevalence of preeclampsia and preterm delivery was still high. Therefore, in 2004, we decided to intensify the strategy by initiating or increasing antihypertensive treatment when blood pressure exceeded 135/85 mmHg or urinary albumin excretion exceeded 300 mg/24 h (Table 2). This strategy seemed to be associated with further improvement as fewer women with type 1 diabetes and microalbuminuria developed preeclampsia or delivered preterm (Table 1) (16–19). We have found a similar effect in women with type 2 diabetes complicated with microalbuminuria or diabetic nephropathy who were treated according to the same antihypertension strategy (20). Since 2004, this strategy for antihypertension treatment has been implemented for women with type 1 and type 2 diabetes complicated with microalbuminuria or diabetic nephropathy at our center. We now also use this strategy in women with preexisting diabetes and essential hypertension or pregnancy-induced hypertension, including preeclampsia.
Pregnancy outcome in three cohorts of women with type 1 diabetes from the same center with different antihypertension strategies
| Years | |||
|---|---|---|---|
| 1995–1999 | 2000–2003 | 2004–2006 | |
| Antihypertension strategy | BP >140/95 mmHg | U-alb >2,000 BP >140/90 AH-shift | U-alb >300 BP >135/85 AH-shift |
| Preeclampsia (%) | 42 | 20 | 0 |
| Early preterm delivery <34 weeks (%) | 23 | 0 | 0 |
| Years | |||
|---|---|---|---|
| 1995–1999 | 2000–2003 | 2004–2006 | |
| Antihypertension strategy | BP >140/95 mmHg | U-alb >2,000 BP >140/90 AH-shift | U-alb >300 BP >135/85 AH-shift |
| Preeclampsia (%) | 42 | 20 | 0 |
| Early preterm delivery <34 weeks (%) | 23 | 0 | 0 |
Presence of at least one of the mentioned criteria prompted initiation and optimizing of antihypertension treatment. AH-shift, shift of ongoing antihypertension treatment prepregnancy to treatment allowed in pregnancy (methyldopa, labetalol, and nifedipine); BP, blood pressure (mmHg); U-alb, urinary excretion of albumin (mg/24 h).
Copenhagen guideline for initiation and optimization of antihypertension treatment in pregnant women with preexisting diabetes with diabetic nephropathy, including microalbuminuria
Initiation and optimization of antihypertension treatment is required if at least one of the following criteria is present:
|
Initiation and optimization of antihypertension treatment is required if at least one of the following criteria is present:
|
Alongside optimized antihypertension therapy, the occurrence of preeclampsia has declined and the prevalence of early preterm delivery before 34 weeks has declined from 10% (3) to 1% in an unselected cohort of 260 pregnant women with preexisting diabetes delivering at our center in the period 2012–2014 (S.K. Nørgaard, N.C. Do, B. Asbjørnsdottir, L. Ringholm, P. Damm, E.R.M., unpublished observations).
Preterm Delivery and Glycemic Control
The high prevalence of preterm delivery in the late 1990s (21) was tightly associated with HbA1c in third trimester (22). It is well known that maternal glucose freely passes through the placenta to the fetal circulation and maternal hyperglycemia thereby induces fetal hyperinsulinemia, fetal overgrowth, relative fetal hypoxia, and increased prevalence of fetal morbidity and mortality (7,23). This emphasizes the importance of strict glycemic control in pregnant women with diabetes. However, a high incidence of severe hypoglycemia with the women needing help from a third person was observed among women with type 1 diabetes (5,24). In a prospective study including 108 women, 45% of women experienced at least one episode of severe hypoglycemia in pregnancy with the highest incidence between 8 and 16 weeks’ gestation (24). In a small cohort of pregnant women with type 2 diabetes who were treated with insulin, 17% experienced at least one episode of severe hypoglycemia during pregnancy (25).
A history of severe hypoglycemia the year preceding pregnancy (5,24), self-estimated impaired hypoglycemia awareness, and particularly the combination of these two factors (25) were significant risk factors for severe hypoglycemia during pregnancy (25). Pregnancy-induced nausea and vomiting have been proposed to be contributing factors for severe hypoglycemia in early pregnancy (5) due to fluctuations in food intake. However, a large prospective study evaluating the occurrence of nausea and vomiting among pregnant women with type 1 diabetes with or without episodes of severe hypoglycemia found that nausea and vomiting were not more prevalent in women with episodes of severe hypoglycemia. Therefore, nausea and vomiting are not major contributing factors for severe hypoglycemia in diabetic pregnancy (25).
In women with type 1 diabetes, pregnancy per se may cause impaired counterregulatory hormonal responses (26,27). Clamped hypoglycemia failed to elicit an increase in glucagon levels, whereas the adrenaline response was markedly lower and the threshold for adrenaline release was decreased compared with those in a nonpregnant control group without diabetes (26). In our prospective study, women with type 1 diabetes and at least one episode of severe hypoglycemia during pregnancy had a lower adrenaline response to hypoglycemia than the remaining women (28). This may render pregnant women with type 1 diabetes more dependent on the second-line defense hormones, such as growth hormone and cortisol (29). Growth hormone has well-established hyperglycemic and insulin antagonistic properties and may exert its glycemic effects directly or indirectly by IGF-I (30,31). In our prospective study, those women experiencing repeated severe hypoglycemic events (five or more events) during pregnancy had IGF-I levels that were 25% lower throughout pregnancy compared with women without severe hypoglycemia (32).
The renin-angiotensin system may also play a role for the risk of severe hypoglycemia during pregnancy in women with type 1 diabetes (33). A significant association between severe hypoglycemia the year before pregnancy and high ACE activity (33) was demonstrated in our prospective cohort in accordance with previous reports in studies of nonpregnant adults (34,35) with type 1 diabetes. However, during pregnancy there was only a tendency toward an association between the ACE activity and the incidence of severe hypoglycemia.
The use of insulin analogs, insulin pumps, and continuous glucose monitoring from before pregnancy might be of value to improve glycemic control with a reduction in the risk of severe hypoglycemia during pregnancy. The short-acting insulin analog insulin aspart has been tested in a large randomized trial including 322 pregnant women with type 1 diabetes. The study demonstrated that insulin aspart was at least as well tolerated and effective as human insulin when used as meal-time insulin in a basal-bolus regimen together with NPH insulin in pregnant women with type 1 diabetes. Additionally, this insulin analog may offer new clinical benefits in terms of better postprandial glucose control and a trend toward reduction of preterm delivery compared with treatment with regular human insulin (36,37).
The long-acting insulin analog detemir has been tested in a large randomized trial in pregnant women with type 1 diabetes, demonstrating that insulin detemir was at least as well tolerated and effective as human NPH insulin. Detemir may offer new clinical benefits in terms of better fasting glucose values (38,39).
A tendency toward fewer preterm deliveries was observed both in the study with insulin aspart (36,37) and insulin detemir (38,39), and no unexpected side effects were observed.
An observational study in pregnant women with type 1 diabetes showed that the incidence of severe hypoglycemia in pregnancy was reduced by 36% while maintaining good glycemic control (40). This was achieved in a routine care setting by applying a multifactorial educational approach with a focus on prevention of hypoglycemia and concomitantly more frequent use of insulin analogs and insulin pumps. Insulin analogs are now the most commonly used insulin in pregnancy in Europe (41).
Insulin pump treatment may be of value during pregnancy, but appropriately powered randomized trials have not been performed to evaluate their efficacy. Improved pregnancy outcomes were not demonstrated in an observational study among women with type 1 diabetes using insulin pumps compared with women using multiple daily injection therapy in our center (42). Nevertheless, insulin pump treatment during pregnancy has become more prevalent (40). With insulin pump treatment, the average carbohydrate-to-insulin ratio declined gradually from 12 to 3 g carbohydrate per unit insulin from early to late pregnancy (42).
The use of continuous glucose monitoring in addition to intermittent finger-prick blood glucose monitoring has been investigated in two clinical trials with conflicting results. The first study demonstrated a positive effect on HbA1c and fetal overgrowth (43), but the second study could not confirm these positive results with a real-time continuous glucose monitoring device (44). Glucose control and pregnancy outcome were not improved in women using the continuous glucose monitoring device before and during labor (45). Use of real-time continuous glucose monitoring with alarms did not reduce the risk of severe hypoglycemia in women at high risk of this condition (46). In fact, a combination of patient education and frequent clinical visits was equally effective as the use of continuous glucose monitoring (46). The technology is constantly improving, with more precision of the glucose monitoring at a lower cost. Despite the poor evidence at present, our pragmatic strategy is to offer real-time continuous glucose monitoring with alarms in early pregnancy to women who have experienced severe hypoglycemia within the past year prior to pregnancy or in early pregnancy.
Fetal Overgrowth: The Role of Maternal Gestational Weight Gain
Fetal growth can be evaluated as the birth weight SD score from the mean birth weight of a local cohort of infants adjusted for sex and gestational age. During the past 20 years, the average SD score of the newborns of women with type 1 diabetes has declined from 1.78 (21,48) in Denmark to a recent value of 1.0 (47) at our center. Despite this improvement in the average fetal size, the prevalence of LGA infants is still 40–50% in our and other populations (44,49).
Fetuses of women with diabetes often present with disproportionate overgrowth as indicated by increased abdominal circumference in relation to the head circumference. In practice, the largest circumference is often the area across the shoulders (50). Fetal overgrowth is therefore associated with increased risk of complications during labor, such as delayed progression of labor and shoulder dystocia (50). In many countries, the prevalence of planned and acute caesarean sections among women with pregestational diabetes is therefore high.
Shoulder dystocia occurs during vaginal delivery when the head is born and the remaining part of the infant is stuck and not immediately born. Special obstetric maneuvers by the birth assistants are needed to deliver the shoulders and thereby complete the delivery. In our population, the risk of shoulder dystocia among vaginal deliveries is 6% among women with type 1 diabetes (50). One-quarter of these infants need life resuscitation at birth, and some of these infants suffer from injuries of bones and nerves (50).
In addition to the short-term complications of delivery, the intrauterine exposure to hyperglycemia is also is associated with increased long-term risk of obesity, prediabetes, and risk factors for cardiovascular disease during adult life (51). Lower cognitive function in adolescent offspring of women with type 1 diabetes has also been reported (48).
The prevalence of LGA infants is positively associated with a higher HbA1c in late pregnancy (47,52), but other mechanisms are also present. It is well known that there is a positive association between maternal and fetal weight gain during normal pregnancy. The Institute of Medicine (IOM) (53) therefore in 2009 published recommendations for appropriate gestational weight gain in women according to BMI classes prior to pregnancy (Table 3).
Copenhagen guidelines for maternal gestational weight gain according to the different BMI categories in women with diabetes in relation to the IOM guidelines for healthy women
| Total weight gain (kg) | ||
|---|---|---|
| Copenhagen guidelines for women with diabetes | IOM guidelines for healthy women | |
| Normal weight | 10–15 | 11.5–16 |
| Overweight | 5–8 | 7–11.5 |
| Obese | 0–5 | 5–9 |
| Total weight gain (kg) | ||
|---|---|---|
| Copenhagen guidelines for women with diabetes | IOM guidelines for healthy women | |
| Normal weight | 10–15 | 11.5–16 |
| Overweight | 5–8 | 7–11.5 |
| Obese | 0–5 | 5–9 |
A large observational cohort study in obese Danish women without diabetes showed that a gestational weight gain of 0–5 kg resulted in the most optimal pregnancy outcome (54). On the basis of these local observations, we decided to recommend a gestational weight gain of 0–5 kg among obese women with diabetes at our center in the year 2008. For normal-weight and overweight women with diabetes, the national Danish recommendations of 10–15 kg were used until 2013, when the recommendation for overweight women was lowered to 5–8 kg (Table 3).
In early pregnancy, the weight gain includes increase in maternal fat depots, whereas in late pregnancy, the weight gain is mainly related to the growth of the fetus, placenta, and uterus. For pragmatic practical reasons, the Copenhagen recommendations for gestational weight gain in women with diabetes were for obese, overweight, and normal-weight women 0, 100, and 100 g per week before gestational week 20 and 200, 300, and 400 g per week in the remaining part of pregnancy, respectively. With these recommendations the maternal gestational weight gain can be followed at each clinical visit and more focus on gestational weight gain can be recommended to the woman, if necessary.
Because of the discrepancies between our Copenhagen recommendations for gestational weight gain for women with diabetes and the IOM guidelines for women in general (Table 3), an observational study including obese women with diabetes was performed (55). The cohort of obese women with type 2 diabetes was divided into those with excessive gestational weight gain (>5 kg) and those obtaining appropriate weight gain or no gestational weight gain at all. The prevalence of LGA infants was 39% in the women with excessive gestational weight gain and 12% in the remaining women gaining <5 kg (55). The infant morbidity, including neonatal hypoglycemia and jaundice, was also significantly higher among the women with excessive weight gain (55). The prevalence of small-for-gestational-age infants (<10% percentile) was comparable between the two groups.
The study was expanded to include overweight and normal-weight women with type 2 diabetes with excessive gestational weight gain defined according to the IOM definitions for the different BMI classes. The infants of mothers with excessive gestational weight gain were approximately 500 g heavier at birth compared with the infants of the remaining women. The prevalence of LGA infants was 48% in women with excessive gestational weight gain and 20% in the remaining women, whereas the prevalence of small-for-gestational-age infants was similar between the groups (56).
In a cohort of women with type 1 diabetes, an independent association between maternal gestational weight gain and fetal overgrowth after adjustment for HbA1c in late pregnancy was demonstrated (57). Similar associations between maternal excessive gestational weight gain and fetal overgrowth in women with diabetes have also been described by others (58). Avoiding excessive gestational weight gain thus seems to be a promising way of reducing fetal overgrowth, and this topic needs further investigation.
Diet Treatment to Control Glucose and Gestational Weight Gain
The diet of pregnant women with diabetes should include sufficient nutrients, but it should not induce excessive gestational weight gain. There is some evidence that intake of low–glycemic index carbohydrates is associated with better intake of micronutrients and improved pregnancy outcome in comparison with high–glycemic index diet in women without diabetes and in women with gestational diabetes mellitus (59). One small study suggested that it is safe to recommend 40% of the diet as carbohydrate in pregnant women with diabetes. However, a minimum daily intake of 175 g carbohydrates during pregnancy is recommended by IOM in order to secure sufficient intake of carbohydrates for the fetal growth and fetal brain function (60), which both depend on carbohydrates. Insufficient carbohydrate intake induces ketone production from degradation of fat. High amounts of ketone bodies in the circulation may increase the risk of ketoacidosis in women with type 1 diabetes and might also be related to impaired brain function in the offspring (48,61).
It is well known that low-carbohydrate diets may improve glycemic control in nonpregnant patients with diabetes, and low-carbohydrate diets are widely used in our diabetes community. However, the recommendations of a daily intake of at least 175 g carbohydrates during pregnancy may necessitate the daily use of carbohydrate counting when aiming for a relatively low amount of carbohydrate in the diet. In addition, carbohydrate counting is useful to adjust the meal-related insulin dose (62). Recommendation of carbohydrate counting thus seems logical to all pregnant women with diabetes on insulin.
In our clinic, we now recommend a daily intake of 175 g of low–glycemic index carbohydrates to pregnant women with diabetes. For practical, pragmatic reasons, this is judged to consist of 150 g from the six major contributors of dietary carbohydrates (bread, potatoes, rice, pasta, dairy products, and fruit). At present we recommend 20, 40, and 40 g of carbohydrates for breakfast, lunch, and dinner, respectively, and 10–20 g for three to four snacks. Breakfast often induces a high rise in postprandial glucose level, and therefore the amount of carbohydrates is restricted more at this meal compared with the remaining meals.
An App as a Tool to Involve Patients
Modern diabetes treatment involves the patient, especially during pregnancy. It is therefore of uppermost importance that the women have access to free information about the recommendations for diet and diabetes care when planning pregnancy and during pregnancy. Initiated by and in collaboration with one of our pregnant women with diabetes (Hellen Marguerite Juul), we have developed an app called Pregnant with Diabetes (63) (Fig. 1). The app is intended for the women with diabetes who are pregnant or planning pregnancy and includes the treatment recommendations at Center for Pregnant Women with Diabetes in Copenhagen. The research summarized in this overview is now implemented as clinical recommendations in this app. It is available in several languages and has been downloaded from almost every country in the world, in total more than 10,000 times.
The app Pregnant with Diabetes was designed for women with diabetes who are pregnant or planning pregnancy.
The app Pregnant with Diabetes was designed for women with diabetes who are pregnant or planning pregnancy.
Future Perspectives
The prevalence of preeclampsia also is high in women with diabetes without kidney involvement in early pregnancy. Evaluation of the efficacy and safety of the present antihypertension strategy used for women with diabetes with pregnancy-induced and essential hypertension are needed. Observational studies of home blood pressure measurements to evaluate the magnitude of white coat hypertension in women with diabetes are needed before the initiation of randomized clinical trials to investigate the effect of antihypertension treatment on the prevention of preeclampsia and preterm delivery.
To improve glycemic control during pregnancy, and thereby pregnancy outcome, more focus on the diet the women are actually eating is mandatory. Pragmatic dietary recommendations to pregnant women with diabetes that are easy to follow, that focus on the quality and quantity of carbohydrates, and that aim for optimal glycemic control and gestational weight gain need to be developed and evaluated.
One of the best documented methods to restrict gestational weight gain among women without diabetes is motivational interviewing (64). The cost-effectiveness of motivational interviewing with a focus on improving the diet and thereby preventing hyperglycemia and excessive gestational weight gain in pregnant women with diabetes deserves to be investigated.
Conclusions
The lessons learned from our clinical research over the past 20 years is that glycemic control, gestational weight gain, and antihypertension treatment all are of importance for improving pregnancy outcome in pregnant women with preexisting diabetes.
Article Information
Acknowledgments. The author would like to thank Peter Damm (Center for Pregnant Women with Diabetes, University of Copenhagen) and Dr. Lene Ringholm (Steno Diabetes Center, Gentofte, Denmark) for help in the preparation of the manuscript and Hellen Marguerite Juul for help with preparation of the app.
Funding. E.R.M. is funded by the Novo Nordisk Foundation.
Duality of Interest. E.R.M. received funds for lectures for Eli Lilly, Sanofi, AstraZeneca, and Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.
