Cardiac hypertrophy in fetuses of diabetic women is a well-known phenomenon that might also cause changes in cardiac function.1,2 There are changes in intracardiac blood flow in fetuses of diabetic mothers throughout gestation.3 The main tool used to study the human fetal intracardiac blood flow and cardiac function is Doppler velocimetry of atrioventricular blood flow. This important aspect of human fetal circulation also has been characterized in normal fetuses.4–6.
The purposes of the present study were to examine possible changes in cardiac function in fetuses of diabetic mothers by comparing the atrioventricular blood flow patterns of fetuses of diabetic and nondiabetic women and to determine more specifically whether cardiac compliance is reduced in fetuses of diabetic women.
Materials and Methods
Between January 1995 and January 1997, we conducted a prospective longitudinal study of women with pregestational diabetes (class B or C) who were between 22 weeks' gestation and term. All diabetic women with glycosylated hemoglobin level lower than 6.5% and mean blood glucose lower than 105 mg/dL were asked to participate in this study. All patients included in the study had an early ultrasound (before 16 weeks' gestation) that confirmed gestational age. When fetal anomalies were detected or when the ultrasound showed a discrepancy in gestational age of more than 1 week from the last menstrual period, the patients were disqualified. The hospital research committee approved the study, and informed consent was obtained from all participants. Thirty-one of 35 pregnant diabetic women (class B or C) who visited our diabetes clinics agreed to participate in this study.
Doppler studies of the blood flow through the mitral and tricuspid valves were done every 4 weeks using a pulsed-wave Doppler ultrasound device (Acuson 128 XP10, Mountain View, CA) with a 3.5- or 5-MHz transducer. We used a high pulse-repetition frequency in the cardiac Doppler studies. Temporal-average intensity was below 50 mW/cm2. The total scanning time was limited to 25 minutes, and all scans were done by one of two examining doctors. The following indices were calculated from the flow velocity waveforms: the peak velocity during rapid ventricular filling (E wave) and during the atrial systole (A wave), and the ratio between these velocities (E/A ratio); and the velocity time integral of the atrioventricular blood flow, which is the sum of the atrioventricular flow velocities during the cardiac cycle. The velocity time integral correlates with volume flow. The velocity time integral multiplied by the heart rate, which is the sum of the flow velocities per minute and correlates with volume flow per minute, was also calculated. To improve the accuracy of these calculations we accepted only measurements obtained with a beam angle less than 20 degrees. We performed the study after obtaining an apical or basal four-chamber view of the fetal heart. Septal thickness was measured by M-mode fetal echocardiography using the lateral four-chamber view of the fetal heart.
The statistical program SPSSPC (SPSS Inc., Chicago, IL) was used to analyze the data. The mean ± two standard deviations (SD) was calculated at 22, 26, 30, 34, and 38 weeks' gestation for the various Doppler indices. The different Doppler indices obtained in fetuses of diabetic women were compared with the Doppler indices in normal fetuses that we previously published.4 The control group was randomly selected and included 25 low-risk pregnant women who had an early ultrasound confirming gestational age.4 The Mann-Whitney test was used to compare fetuses of diabetic and nondiabetic women at different stages of pregnancy. We also calculated the interobserver variability of our measurements.
Each patient had four or five fetal echocardiographic exams at 22, 26, 30, 34, and 38 weeks' gestation. The success rate in obtaining the measurements from the mitral valve was 93% and from the tricuspid valve, 91%. The coefficient of correlation between the two observers was 0.96 (P < .001, slope = 1.011, constant = −0.023) and 0.93 (P < .001, slope = 1.015, constant = −0.026) in measuring E/A ratio and velocity time integral multiplied by heart rate, respectively. The mean maternal age of diabetic women was 32 ± 7 years and of nondiabetic women 29 ± 6 years (P < .05). The mean gestational age at delivery of diabetic women was 38.1 ± 1.8 weeks' gestation and of nondiabetic women 40.3 ± 0.9 weeks' gestation (P < .05). The mean birth weight of infants of diabetic women was 3780 ± 756 g and of infants of nondiabetic women 3405 ± 541 g (P < .05). Ten diabetic women (38%) and three nondiabetic women (10%) had cesarean deliveries (P < .05). There were no adverse outcomes in either group (all Apgar scores greater than 8 and no admissions to the neonatal intensive care unit).
As shown in Table 1 and Figures 1 and 2, the E/A ratio of the mitral and tricuspid valves did not increase in fetuses of diabetic women during the third trimester of pregnancy and was significantly higher in fetuses of nondiabetic women compared with fetuses of diabetic women at 34 and 38 weeks' gestation. The velocity time integral multiplied by heart rate of the mitral and tricuspid valves was higher, but not significantly, in fetuses of nondiabetic women compared with fetuses of diabetic women at 34 and 38 weeks' gestation. Mitral and tricuspid E-wave increased in fetuses of diabetic mothers as in normal fetuses between 22–38 weeks' gestation. Mitral and tricuspid A-wave increased only in fetuses of diabetic women between 30 and 38 weeks' gestation and was significantly higher at 34 and 38 weeks' gestation compared with fetuses of nondiabetic women (Figure 2). The interventricular septum was significantly thicker in fetuses of diabetic women compared with normal fetuses between 26 and 38 weeks' gestation (Table 2).
Two important aspects of cardiac function can be assessed by studying the blood flow patterns across the atrioventricular valves. First, cardiac output or factors that correlate with cardiac output can be measured at the level of the atrioventricular valves when no cardiac malformation is present. Second, the relationship between the two components of atrioventricular blood flow, rapid ventricular filling (E-wave) and atrial contraction (A-wave), could provide information about cardiovascular function and might correlate with cardiac compliance. In fact, improvement of cardiac compliance will increase the relative contribution of the rapid ventricular filling (E-wave) to the total ventricular filling. Other conditions might influence the atrioventricular blood flow, such as loading conditions and afterload.
In normal pregnancies, the E/A ratio of the mitral and tricuspid valves increases throughout gestation.4–7 These changes are caused by an increase of the rapid ventricular filling (E-wave) throughout gestation, because the ventricular filling associated with atrial contraction (A-wave) does not change significantly during pregnancy.4 The increase of stroke volume and cardiac output throughout gestation probably results from increased size of the valve orifices because velocity time integral changes only moderately during pregnancy.
Previous studies8,9 suggested that different changes occur in cardiac function in fetuses of pregestational diabetic women; most prominently, the E/A ratio of the mitral and tricuspid valves is lower in fetuses of women with pregestational diabetes than in fetuses of nondiabetic women. Theoretically, this difference might indicate poor cardiac compliance in fetuses of diabetic women. We examined in detail the atrioventricular blood flow patterns in fetuses of diabetic women and made the following observations. First, the E/A ratio of the mitral and tricuspid valves did not increase in fetuses of diabetic women as it did in fetuses of nondiabetic women throughout gestation. It becomes significantly lower in fetuses of diabetic women compared with fetuses of nondiabetic women only during the second half of the third trimester. Second, the relatively low E/A ratio observed in fetuses of diabetic women probably is not caused by poor cardiac compliance. This conclusion comes from our observation that the rapid ventricular filling (E-wave) increases in fetuses of diabetic women as in fetuses of nondiabetic women throughout gestation. The low E/A ratio resulted from a higher A-wave detected in these fetuses compared with fetuses of nondiabetic women. Although the A-wave remains constant throughout the second half of pregnancy in normal fetuses, the present study found that, in fetuses of diabetic women, the A-wave increases during the second half of pregnancy. Third, the high A-wave might indicate increased cardiac contractility and increased ventricular filling due to changes in cardiac output in fetuses of diabetic women. In fetal lambs infused with exogenous insulin, cardiac output increased.10 In the present study, mitral and tricuspid velocity time integral multiplied by heart rate, which correlates with cardiac output, was not significantly higher in fetuses of diabetic women compared with fetuses of nondiabetic women during the second half of the third trimester.
Our study population included only fetuses of women whose pregestational diabetes was well controlled. Rizzo et al9 found that, as early as 12 weeks' gestation, fetuses of poorly controlled insulin-dependent diabetic mothers had a lower E/A ratio in the mitral and tricuspid valves compared with normal fetuses. However, in their study, fetuses of well-controlled insulin-dependent diabetic women had significant changes in cardiac function compared with normal fetuses during this early stage of the second trimester. We did not find these changes in fetuses of well-controlled diabetic women until the second half of the third trimester. These differences might result from differences in study groups, eg, the cutoff value of glycosylated hemoglobin we used (6.5%) compared with higher values used by others (8.5%). More studies are needed to examine the effect of maternal control of diabetes on fetal cardiac function.
1. Weber HS, Copel JS, Reece EA, Green J, Kleinman CS. Cardiac growth in fetuses of diabetic mothers with good metabolic control. J Pediatr 1991;118:103–7.
2. Gandhi JA, Zhang XY, Maidman JE. Fetal cardiac hypertrophy and cardiac function in diabetic pregnancies. Am J Obstet Gynecol 1995;173:1132–6.
3. Rizzo G, Arduini D, Romanini C. Accelerated cardiac growth and abnormal cardiac flow in fetuses of type I diabetic mothers. Obstet Gynecol 1992;80:369–76.
4. Weiner Z, Efrat Z, Zimmer E, Itskovitz-Eldor J. Fetal atrioventricular blood flow throughout gestation. Am J Cardiol 1997;80:658–62.
5. Reed KL, Meijboom EJ, Sahn DJ, Scagnelli SA, Valdes-Cruz LM, Shenker L. Cardiac Doppler flow velocities in human fetuses. Circulation 1986;73:41–6.
6. Kenny JF, Plappert T, Doubilet P, Doubilet P, Saltzman DH, Cartier M, et al. Changes in intracardiac blood flow velocities and right and left ventricular stroke volume with gestational age in the normal human fetus: A prospective Doppler echocardiographic study. Circulation 1986;74:1208–16.
7. De Smedt MCH, Visser GHA, Meijboom EJ. Fetal cardiac output estimated by Doppler echocardiography during mid- and late gestation. Am J Cardiol 1987;60:338–42.
8. Rizzo G, Arduini D, Romanini C. Fetal cardiac function in fetuses of type I diabetic mothers. Am J Obstet Gynecol 1991;164:837–43.
9. Rizzo G, Arduini D, Capponi A, Romanini C. Cardiac and venous blood flow in fetuses of insulin-dependent diabetic mothers: Evidence of abnormal hemodynamics in early gestation. Am J Obstet Gynecol 1995;173:1775–81.
10. Milley JR, Rosenberg AA, Philipps AF, Molteni RA, Jones MD Jr, Simmons MA. The effect of insulin on ovine fetal oxygen extraction. Am J Obstet Gynecol 1984;149:673–8.