Background
Intrauterine exposure to a diabetic environment is associated with adverse fetal myocardial remodeling. The aim of this study was to assess the biventricular systolic and diastolic function of fetuses exposed to maternal diabetes (MD) compared with control subjects, using a comprehensive cardiac functional assessment and exploring the role of speckle-tracking to assess myocardial deformation. The authors hypothesized that fetuses exposed to MD present signs of biventricular dysfunction, which can be detected by deformation analysis.
Methods
A cross-sectional study was conducted in 129 fetuses with structurally normal hearts, including 76 fetuses of mothers with diabetes and 53 of mothers without diabetes. Maternal baseline characteristics, standard fetoplacental Doppler indices, and conventional echocardiographic and myocardial deformation parameters were prospectively collected at 30 to 33 weeks of gestation.
Results
Fetuses of mothers with diabetes had a significantly thicker interventricular septum compared with control subjects (median, 4.25 mm [interquartile range (IQR), 3.87–4.50 mm] vs 3.67 mm [IQR, 3.40–3.93 mm), P < .001), but no effect modification was demonstrated on myocardial deformation analysis. No significant differences were found in conventional systolic and diastolic functional parameters for the left ventricle and right ventricle, except for lower left ventricular cardiac output in the MD group (median, 320 mL/min [IQR, 269–377 mL/min] vs 365 mL/min [IQR, 311–422 mL/min], P < .05]. Deformation analysis demonstrated a significantly lower early diastolic strain rate (SRe) and late diastolic strain rate (SRa) for both ventricles in the MD group (left ventricle: SRe 1.85 ± 0.72 vs 2.26 ± 0.68 sec −1 , SRa 1.50 ± 0.52 vs 1.78 ± 0.57 sec −1 ; right ventricle: SRe 1.57 ± 0.73 vs 1.97 ± 0.73 sec −1 , SRa 2 ± 0.77 vs 1.68 ± 0.79 sec −1 ; P < .05), suggesting biventricular diastolic impairment. Additionally, the right ventricle presented a lower global longitudinal strain in the study group (−13.67 ± 4.18% vs −15.52 ± 3.86%, P < .05). Multivariate analysis revealed that maternal age is an independent predictor of left and right ventricular global longitudinal strain ( P < .05), with a significant effect only in MD after group stratification.
Conclusions
Fetuses of mothers with diabetes present signs of biventricular diastolic dysfunction and right ventricular systolic dysfunction by deformation analysis in the third trimester of pregnancy. They may represent a special indication group for functional cardiac assessment, independently of septal hypertrophy. Two-dimensional speckle-tracking could offer an additional benefit over conventional echocardiography to detect subclinical unfavorable changes in myocardial function in this population.
Highlights
- •
Intrauterine exposure to a diabetic environment is a risk factor for fetal development.
- •
A comprehensive echocardiographic evaluation with myocardial deformation is performed.
- •
The study confirms the feasibility and reproducibility of fetal 2D speckle-tracking.
- •
Fetuses of diabetic mothers have biventricular diastolic dysfunction.
- •
Fetuses of diabetic mothers have right ventricular systolic dysfunction.
Exposure to a diabetic intrauterine environment has long been recognized as a risk to the fetus, with a double impact on the fetal heart. During embryogenesis it has a teratogenic effect, increasing the incidence of structural cardiac anomalies. More commonly, infants of mothers with diabetes are at higher risk for developing fetal hypertrophic cardiomyopathy, characterized by asymmetric myocardial hypertrophy, more evident in the third trimester of pregnancy and affecting mainly the interventricular septum.
Advances in high-resolution echocardiography and Doppler interrogation of the fetal heart have improved our understanding of fetal cardiac function and response to adverse intrauterine conditions. It has been suggested that diastolic dysfunction may represent the earliest change on fetal hypertrophic cardiomyopathy in fetuses of mothers with diabetes, preceding systolic dysfunction, and in some cases progression to symptomatic heart failure. However, the commonly applied evaluation methods provide only indirect information about systolic and diastolic function and do not measure deformation of the myocardium itself. Speckle-tracking echocardiography is a recent development based on the tracking of unique speckle patterns or “kernels” in the myocardium throughout the cardiac cycle. It has the potential to provide new insights into fetal origins of disease and into the understanding of fetal heart function in specific disease states.
The aim of this study was to assess biventricular systolic and diastolic function in fetuses of mothers with diabetes compared with control subjects, using conventional fetal echocardiography and exploring the potential role of deformation analysis by speckle-tracking on cardiac functional assessment. We hypothesized that fetuses exposed to maternal diabetes (MD) present signs of biventricular dysfunction and that deformation analysis could be a sensitive and reproducible technique to assess global and segmental cardiac function of fetuses in this condition.
Methods
Study Population
We conducted an observational cross-sectional study at a tertiary referral center for fetal cardiology. Patients were selected at the time of the examination appointment, following a convenience sampling strategy, between July 2016 and June 2017. Echocardiographic data were acquired in a prospective manner specifically for the purpose of the study. The study protocol was approved by the local ethics committee, and written informed consent was obtained from all study participants.
A total of 129 subjects were recruited. The MD group comprised 76 pregnant women with gestational diabetes diagnosed by current international recommendations (fasting plasma glucose at first trimester or glucose tolerance test using 75 g glucose between 24 and 28 weeks of gestation) or with pregestational diabetes. The control group comprised 53 pregnant women without diabetes with normal results on fetal echocardiographic evaluation. Referral reasons for control group were family history of congenital heart disease ( n = 13), unconfirmed suspicion of heart disease on obstetric anomaly scan ( n = 8), polyhydramnios ( n = 8), increased nuchal translucency thickness in the first trimester ( n = 8), intracardiac echogenic focus ( n = 7), exposure to teratogens in the first trimester ( n = 4), advanced maternal age ( n = 3), and mild tricuspid regurgitation on obstetric scan ( n = 2). Exclusion criteria applied to both groups were structural or chromosomal fetal anomalies, fetal arrhythmias, fetal growth restriction, evidence of fetal infection, multiple pregnancies, pregnancies conceived by assisted reproductive technology, and maternal chronic disease other than diabetes mellitus. All control group subjects had normal findings on postnatal clinical examinations, which included medical history and physical examination.
For a subgroup analysis, the MD group was divided into subjects receiving no pharmacologic treatment ( n = 42) and those receiving pharmacologic treatment (insulin and/or metformin; n = 34).
Fetal examinations were performed between 30 and 33 weeks of gestation in both groups, with gestational age calculated on the basis of crown-rump length measurement on first-trimester ultrasound.
Maternal baseline characteristics including age, height, and weight (for body mass index calculation), smoking habit, multiparity, and diabetes treatment were collected at the time of the ultrasound examination. Detailed fetal echocardiography was performed by an experienced cardiologist to exclude congenital heart disease. A standard fetoplacental Doppler evaluation and comprehensive fetal echocardiography were added to routine fetal echocardiography.
Imaging Protocol
All ultrasound studies were performed using a single high-resolution ultrasound platform (Vivid E95; GE Healthcare, Little Chalfont, United Kingdom) equipped with a C1-6 curved-array fetal transducer. Two-dimensional (2D) echocardiographic data were acquired at a high frame rate (60–150 frames/sec). Pulsed-wave Doppler parameters were recorded from three or more waveforms, with insonation angle < 20°. For offline postprocessing, EchoPAC version 201.71 (GE Healthcare) was used for both conventional and deformation analysis.
Fetoplacental Doppler Evaluation
Estimated fetal weight was calculated using the formula of Hadlock et al. Standard obstetric Doppler evaluation comprised measurement of the pulsatility index for the umbilical artery, middle cerebral artery, ductus venosus, and aortic isthmus. The cerebroplacental ratio was calculated by dividing middle cerebral artery and umbilical artery pulsatility indices.
Fetal Echocardiography
Cardiovascular evaluation comprised morphometric measurements and functional assessment of the heart, the latter by both conventional echocardiography and deformation analysis. The following parameters, measured as described previously, were included.
Cardiac morphometry: cardiothoracic ratio (by the ellipse method in a cross-sectional view of the fetal thorax at end-diastole), ventricular sphericity index (base-to-apex length/basal diameter in a four-chamber view of the heart at end-diastole), interventricular septal wall thickness (by M-mode echocardiography in a transverse four-chamber view at end-diastole), and left and right atrial areas (on 2D images of a four-chamber view at end-ventricular systole, the maximum point of atrial distension)
Conventional systolic function: left and right cardiac output (cross-sectional aortic or pulmonary annular area × aortic or pulmonary flow velocity integral × heart rate), left ventricular shortening fraction (by M-mode echocardiography from the transverse four-chamber view using the Teichholz formula), and mitral and tricuspid annular plane systolic excursion (by M-mode echocardiography in an apical or basal four-chamber view)
Conventional diastolic function: mitral and tricuspid early (E) and late (A) diastolic filling ratio (E/A ratio) and isovolumic relaxation time
Conventional global function: modified left myocardial performance index assessed by pulsed-wave Doppler (the ratio between isovolumic times [contraction + relaxation] and ejection time)
Deformation Analysis: Two-dimensional speckle-tracking analysis was performed to assess left and right myocardial deformation ( Figure 1 ). Global longitudinal strain and strain rate (systolic, early diastolic, and late diastolic) were analyzed and compared between groups. Segmental deformation was studied as well.
Figure 1
Example of left ventricular myocardial deformation assessment in a four-chamber view of a fetal heart. (A) Global longitudinal strain with a graphic representation of average strain curves displayed as function of time for each of six segments. (B) Systolic and early and late diastole peak strain rates with a graphic representation of strain rate curves at the top and a table recording the peak strain rate values for each segment at the bottom. S% , Peak systolic strain; SRa , late diastolic strain rate; SRe , early diastolic strain rate; SRs , peak systolic strain rate.
A 2D video clip of a four-chamber view of the fetal heart was used for offline analysis. Special attention was taken to clear delineation of the ventricular walls, and frame rate optimization was performed during image acquisition. At least three cine loops of a four-chamber view were acquired. Raw, uncompressed data were then sent to the workstation for analysis. A single cardiac cycle was defined by the selection of two consecutive end-diastolic frames (corresponding with mitral valve closure). Left and right ventricular endocardial borders were defined manually in end-diastole and tracking curves of endocardial and epicardial borders created automatically by the software in subsequent frames. Adequacy of tracking was assessed visually. The software segmented automatically the ventricles into six equidistant segments (basal septal, mid septal, apical septal, apical lateral, mid lateral, and basal lateral) and provided segmental and global analyses of deformation parameters. The same technique was used for both ventricles. For the right ventricle only the free wall data (basal lateral, mid lateral, and apical lateral segments) were analyzed to provide global and segmental deformation parameters.
Statistical Analysis
Descriptive data are summarized as mean ± SD for normally distributed variables and as median and interquartile range for skewed distributions. Data were analyzed using Stata version 14 (StataCorp, College Station, TX). For continuous variables, unpaired t tests and analysis of variance were used for two-group and three-group comparisons, respectively. P values < .05 were considered to indicate statistical significance.
Univariate linear regressions were used to interrogate the effect of maternal baseline characteristics (age, body mass index, smoking habit, and multiparity) and of interventricular septal thickness (as a continuous variable) on left and right ventricular deformation parameters. To adjust for possible confounding, a multivariate linear regression model including the covariates maternal age, body mass index, smoking habit, and multiparity was built.
To assess intraobserver and interobserver variability, 16 random subjects (eight from each group) were analyzed on two separate occasions by the same operator and two different operators, respectively. Intraclass correlation coefficients (ICC) were assessed using a two-way mixed-effects model for intraobserver variability and a two-way random-effects model for interobserver variability. ICCs > 0.80 were considered excellent and 0.60 to 0.80 good.
Results
Baseline Maternal and Fetal Characteristics
Characteristics of the study populations are reported in Table 1 . The MD group included seven pregnant women with pregestational diabetes and 69 with gestational diabetes. Maternal baseline characteristics were similar for the MD and control groups, except for maternal weight and body mass index, which were higher in the MD group. Gestational age at ultrasound examination was similar between groups, as expected. Estimated fetal weight and fetoplacental Doppler parameters did not differ significantly between groups.
| Control group ( n = 53) | MD group ( n = 76) | P | |
|---|---|---|---|
| Maternal characteristics | |||
| Maternal age (y) | 32 ± 5.61 | 33 ± 4.91 | .75 |
| Height (cm) | 161.79 ± 5.93 | 161.41 ± 6.50 | .74 |
| Weight (kg) | 59 (55–65) | 66 (57–81) | .003 |
| BMI (kg/m 2 ) | 23.31 (20.61–25.22) | 25.07 (22.04–29.76) | .001 |
| Smoking | 6 (11%) | 10 (13%) | .79 |
| Multiparity | 30 (57%) | 45 (59%) | .72 |
| Fetoplacental ultrasound evaluation | |||
| Gestational age (wk) | 31.25 ± 1.28 | 31.19 ± 0.13 | .74 |
| Estimated fetal weight (g) | 1,680.95 ± 291.70 | 1,661.66 ± 236.07 | .68 |
| Umbilical artery PI | 1.04 ± 0.17 | 1.04 ± 0.18 | .93 |
| Middle cerebral artery PI | 2.05 ± 0.34 | 2.06 ± 0.34 | .84 |
| Cerebroplacental ratio | 1.94 (1.73–2.17) | 1.99 (1.73–2.28) | .71 |
| Ductus venosus PI | 0.53 (0.41–0.69) | 0.51 (0.40–0.67) | .63 |
| Aortic isthmus PI | 2.56 ± 0.31 | 2.67 ± 0.31 | .081 |
Morphometric and Functional Echocardiographic Indices
Results of conventional echocardiography are displayed in Table 2 . Interventricular septal thickness was significantly higher in the MD group. No significant differences were present between the two groups in conventional echocardiographic data for systolic and diastolic function for both ventricles, except for left-sided cardiac output, which was lower in the MD group.
| Control group ( n = 53) | MD group ( n = 76) | P | |
|---|---|---|---|
| Cardiac morphometry | |||
| Cardiothoracic ratio | 0.29 ± 0.03 | 0.28 ± 0.03 | .35 |
| Left atrial area (cm 2 ) | 1.96 (1.75–2.20) | 1.82 (1.64–2.02) | .034 |
| Right atrial area (cm 2 ) | 1.96 (1.72–2.16) | 1.78 (1.66–2.13) | .062 |
| Global sphericity index | 1.24 ± 0.11 | 1.23 ± 0.09 | .69 |
| IVS thickness (mm) | 3.67 (3.40–3.93) | 4.25 (3.87–4.50) | <.001 |
| Systolic function | |||
| Left CO (mL/min) | 365 (311–422) | 320 (269–377) | .005 |
| Right CO (mL/min) | 569 (466–725) | 590 (482–667) | .72 |
| Left SF (%) | 36 (29–39) | 35 (32–43) | .28 |
| MAPSE (mm) | 6.99 ± 1.09 | 6.67 ± 1.12 | .16 |
| TAPSE (mm) | 9.19 ± 1.60 | 8.98 ± 1.28 | .45 |
| Diastolic function | |||
| Tricuspid E/A ratio | 0.78 (0.74–0.84) | 0.78 (0.72–0.82) | .37 |
| Mitral E/A ratio | 0.80 (0.74–0.85) | 0.79 (0.75–0.85) | .81 |
| IRT (msec) | 54 (50–57) | 54 (49–58) | .68 |
| Global function | |||
| Mod MPI | 0.56 ± 0.07 | 0.55 ± 0.07 | .69 |
Deformation Analysis
Deformation analysis for the left and right ventricles is reported in Table 3 . There were no significant differences in the acquisition frame rate and fetal heart rate between groups.
| Control group ( n = 53) | MD group ( n = 76) | P | |
|---|---|---|---|
| Frame rate (frames/sec) | 110 (96–126) | 110 (91–124) | .73 |
| Fetal heart rate (beats/min) | 134 ± 12 | 137 ± 12 | .14 |
| Left ventricle | |||
| Global longitudinal strain | |||
| Systolic peak (%) | −16.99 ± 2.42 | −16.01 ± 3.45 | .076 |
| Time to peak (msec) | 215 ± 27 | 207 ± 27 | .096 |
| FW systolic peak (%) | −16.68 ± 2.80 | −15.54 ± 3.88 | .071 |
| IVS systolic peak (%) | −18.03 ± 2.99 | −17.16 ± 4.20 | .20 |
| Systolic strain rate | |||
| Systolic peak (1/sec) | −1.73 ± 0.34 | −1.62 ± 0.39 | .10 |
| Time to peak (msec) | 96 ± 37 | 95 ± 30 | .82 |
| Diastolic strain rate | |||
| Early diastolic (1/sec) | 2.26 ± 0.68 | 1.85 ± 0.72 | .001 |
| Late diastolic (1/sec) | 1.78 ± 0.57 | 1.50 ± 0.52 | .005 |
| Right ventricle | |||
| Global longitudinal strain | |||
| Systolic peak (%) | −15.52 ± 3.86 | −13.67 ± 4.18 | .012 |
| Time to peak (msec) | 200 ± 35 | 199 ± 41 | .90 |
| Systolic strain rate | |||
| Systolic peak (1/sec) | −1.49 ± 0.39 | −1.42 ± 0.45 | .35 |
| Time to peak (msec) | 89 ± 41 | 99 ± 39 | .19 |
| Diastolic strain rate | |||
| Early diastolic (1/sec) | 1.97 ± 0.73 | 1.57 ± 0.73 | .002 |
| Late diastolic (1/sec) | 2.00 ± 0.77 | 1.68 ± 0.79 | .021 |
Left Ventricle
There were no differences in systolic deformation of the left ventricle between groups, namely for global longitudinal strain and systolic strain rate. No systolic segmental differences were found between groups when comparing the left ventricular free wall and interventricular septum ( Table 3 ), nor for each individual segment (data not shown).
Regarding diastolic function, the MD group had more negative longitudinal early and late diastolic strain rates than control subjects, indicative of worse diastolic function ( Table 3 ). Segmental analysis showed that the diastolic impairment was evident for all six left ventricular segments, except for late diastolic strain rate in the basal septal segment (data not shown).
A subgroup analysis was performed to compare (1) control subjects, (2) subjects with MD receiving no pharmacologic treatment, and (3) subjects with MD receiving pharmacologic treatment. Left ventricular global longitudinal strain was significantly more impaired in subjects with MD receiving no pharmacologic treatment compared with the other two subgroups ( P = .01; Figure 2 ). Also, subjects with MD receiving no pharmacologic treatment had lower early ( P = .013) and late ( P = .01) diastolic strain rate values than control subjects ( Figure 2 ).
Right Ventricle
Right ventricular global longitudinal strain was significantly lower in the MD group compared with control subjects ( Table 3 ). Analysis of the right ventricular segments showed that basal lateral and mid lateral were the segments most affected in the MD group (basal lateral, −15.1% vs −17.1%, P = .039; mid lateral, −14.1% vs −15.9%, P = .026). Diastolic function seemed also to be impaired in the right ventricle, with the MD group presenting more negative longitudinal early and late diastolic strain rates ( Table 3 ). Segmental analysis revealed that the diastolic impairment was evident for all RV segments (data not shown).
Multivariate regression analysis showed that maternal age is an independent predictor of left ventricular and right ventricular global longitudinal strain, independently of maternal body mass index, smoking habit, and multiparity ( Table 4 ). Further analysis demonstrated a significant interaction between study group and maternal age. After stratification, it was possible to observe a statistically significant effect of maternal age on ventricular deformation for the MD group only. For each additional year of maternal age in the MD group, there were decreases in left ventricular global longitudinal strain and right ventricular global longitudinal strain of −0.24% (95% CI, −0.40% to −0.008%) and −0.24% (95% CI, −0.43% to −0.045%), respectively ( Figure 3 ).
