Gestational Diabetes Mellitus and the Maternal Heart

Pregnancy is associated with major adaptations of physiology, including of the glucoregulatory and cardiovascular systems. Often considered a stress test of the glucoregulatory system, pregnancy can uncover a predilection to later diabetes development in mothers if gestational diabetes mellitus (GDM) is diagnosed (1). GDM also predicts an increased risk of cardiovascular diseases (CVD) later in the lives of affected mothers (2). In normal pregnancy, cardiac output increases by approximately 30–50%, which is achieved through a reduction in systemic vascular resistance (20–30% by midpregnancy) and increases in maternal blood volume (about 1.1 to 1.6 L), heart rate (15–25%), stroke volume (20–30%), and left ventricular mass (24–34%) (3). These cardiac adaptations place women with preexisting cardiac conditions (e.g., congenital heart disease, valvular heart disease, and cardiomyopathies) at high risk of heart failure during pregnancy (3). Unknown is whether assessment of the response of the heart in women with GDM to the cardiac stress test of pregnancy can be used to more accurately predict those at greatest risk of future CVD.

In this issue of Diabetes Care, Thirunavukarasu et al. (4) explore the adaptation of the heart to pregnancy in a prospective study of 30 women with GDM compared with 38 healthy pregnant (HP) women. They used ³¹P magnetic resonance spectroscopy (³¹P-MRS) and cardiovascular magnetic resonance imaging (CMR) to assess myocardial energetics, cardiac structure, and function at ∼30 weeks of gestation (4). The women were well matched for age, gestational age, and ethnicity, but the group with GDM had higher BMI (31 ± 5 kg/m² vs. 25 ± 5 kg/m²; P = 0.0001) and systolic (120 ± 9 mmHg vs. 113 ± 10 mmHg; P = 0.004) and diastolic (74 ± 8 mmHg vs. 72 ± 7 mmHg; P = 0.01) blood pressures in the 1st trimester. Systolic and diastolic blood pressure differences persisted into the 3rd trimester. Of the 30 GDM women, 12 (40%) were treated with metformin, 4 of whom required additional insulin, and 18 were treated with lifestyle only (4).

The key findings of the cardiac investigations in the women with GDM compared with the HP women were a lower phosphocreatine-to-ATP (PCr/ATP) ratio (1.9 ± 0.4 vs. 2.2 ± 0.3; P < 0.0,001), a higher left ventricular (LV) mass (103 ± 18 g vs. 90 ± 13 g; P = 0.001), a lower LV end-diastolic volume (67 ± 11 mL/m² vs. 76 ± 12 mL/m²; P = 0.002), and lower global longitudinal shortening (−18 ± 3% vs. −20 ± 3%; P = 0.008) (4). These differences remained statistically significant when adjusted for BMI and gestational weight gain. LV ejection fraction and mitral inflow E/A ratio and deceleration time were similar in the GDM and HP women (4).

While the LV structural changes are consistent with concentric hypertrophy, a likely consequence of the higher blood pressure in these women with GDM and obesity, the lower PCr/ATP ratio and global longitudinal shortening are suggestive of some decompensation in myocardial function in these women, which is clearly of concern. Unanswered questions relate to 1) the mechanisms causing the observed reductions in myocardial energy status and contractile function in women with GDM and obesity, 2) the relative roles of GDM and obesity in these findings, and 3) how these findings of cardiac maladaptation to pregnancy can be used to predict future maternal cardiac health.

Phosphorylation of creatine to PCr occurs when ATP is abundant in the cardiomyocyte, enabling the buildup of an energy supply that can be rapidly accessed for ATP generation at times of increased demand (5). A fall in the myocardial PCr/ATP ratio is an indicator of cardiac energy deficiency and dysfunction (5). For example, the PCr/ATP ratio is reduced in myocardial ischemia and is reported to be low in the failing heart (5,6). Relevant to the findings of this article, low ratios have also been reported in subjects with type 2 diabetes and obesity (5,7,8) as well as in those with hypertrophic cardiomyopathy (9). Low PCr/ATP ratios in the heart can be worsened in cardiac stress testing either by exercise or pharmacological means (e.g., dobutamine) (9,10). In subjects with obesity, a lower PCr/ATP ratio can be improved through weight loss (11).

It is therefore not entirely surprising that lower PCr/ATP ratios were found in the myocardium of women with GDM and obesity at 30 weeks of gestation (Fig. 1). The combination of the increased cardiac work requirements in pregnancy with the effects of higher blood pressures in the women with GDM and obesity will increase demand for cardiac energy. Oxygen supply to meet this energy demand may be compromised if vascularization of the rapidly hypertrophing myocardium is inadequate. An additional factor in pregnancy is that the heart relies more on free fatty acids (FFA) for energy, as the insulin resistance of pregnancy diverts glucose away from the heart for use by the fetus and maternal brain. Mitochondrial ATP production from fatty acids is less efficient than from glucose with respect to O₂ consumption and will be further compromised if there is a degree of hypoxia (12). FFA levels have also been reported to be inversely related to the PCr/ATP ratio in individuals with type 2 diabetes (8). Furthermore, the women with GDM in this study had higher 3rd trimester plasma triglyceride, FFA, and ketone body concentrations (4).

With 14 of the 30 women with GDM having been treated with metformin (4), the possibility that metformin is contributing to the lower PCr/ATP ratio warrants consideration. Metformin is known to reduce the efficiency of mitochondrial ATP production through inhibitory effects on both complex Ι of the respiratory chain and mitochondrial glycerophosphate dehydrogenase activity (13) (Fig. 1). Metformin has also been shown to be cardioprotective in ischemia-reperfusion injury (14). The cardiac protective effect of metformin may therefore come with the cost of a lower energy state. This raises the possibility that the lower PCr/ATP in the women with GDM and obesity is a consequence of cardioprotective mechanisms in which insulin resistance limits nutrient-induced stress to the myocardium (15). A point may be reached by which such cardioprotective mechanisms compromise cardiac function by leading to a myocardial energy deficit.

A limitation in the study of Thirunavukarasu et al. (4) is that it is difficult to disentangle the effects of obesity and GDM on the cardiac findings, even though GDM changes persisted after adjustment for BMI. Women with obesity only or GDM only also need to be studied. From the known effects of obesity on cardiac energetics and function outside pregnancy (7), it is likely that obesity is a significant contributor to the findings of this study. Obesity and GDM, however, could have synergistic effects resulting in the cardiac maladaptation to pregnancy.

The importance of this study of the hearts of women with obesity and GDM is that changes of real concern have been demonstrated (4). It adds weight to the findings of an earlier study that showed impaired cardiac function indices in women with GDM assessed by an ultrasonic cardiac output monitor (16). Further studies are required to determine the significance of the findings. We need to first understand the serial changes in ³¹P-MRS and CMR that occur during normal cardiac adaptation to pregnancy through all trimesters and postpartum. Only then can studies in pregnant women with obesity, GDM, and preexisting diabetes be fully interpreted. Investigation of effects of therapies such as diet, exercise, insulin, and metformin on heart function in pregnancy of women with GDM and obesity is also important. Long-term follow-up studies are necessary to determine whether the maladaptation of these cardiac indices in pregnancies of women and GDM and obesity are fully reversible or result in some permanent cardiac injury. Long-term studies are also needed to determine whether measurement of cardiac indices by ³¹P-MRS and CMR during pregnancy can be used to develop prediction models for future CVD risk in affected women.