Background: Gestational diabetes mellitus (GDM) is the most common complication in pregnancy, representing a serious risk for the mother and fetus. Identifying new biomarkers to ameliorate the screening and improving GDM diagnosis and treatment is crucial. During the last decade, a few studies have used speckle tracking echocardiography (STE) for assessing the myocardial deformation properties of fetuses (FGDM) and infants (IGDM) of GDM women, providing not univocal results. Accordingly, we performed a meta-analysis to examine the overall influence of GDM on left ventricular (LV) and right ventricular (RV) global longitudinal strain (GLS) in both FGDM and IGDM.
Methods: All echocardiographic studies assessing conventional echoDoppler parameters and biventricular strain indices in FGDM and IGDM vs. infants born to healthy pregnant women, selected from PubMed and EMBASE databases, were included. The studies performed on FGDM and IGDM were separately analyzed. The subtotal and overall standardized mean differences (SMDs) in LV-GLS and RV-GLS in FGDM and IGDM studies were calculated using the random-effect model.
Results: The full texts of 18 studies with 1046 babies (72.5% fetuses) born to GDM women and 1573 babies of women with uncomplicated pregnancy (84.5% fetuses) were analyzed. Compared to controls, FGDM/IGDM were found with a significant reduction in both LV-GLS [average value -18.8% (range -11.6, -24.2%) vs. -21.5% (range -11.8, -28%), p < 0.05)] and RV-GLS [average value -19.7% (range -13.7, -26.6%) vs. -22.4% (range -15.5, -32.6%), p <0.05)]. Large SMDs were obtained for both LV-GLS and RV-GLS studies, with an overall SMD of -0.91 (95%CI -1.23, -0.60, p < 0.001) and -0.82 (95%CI -1.13, -0.51, p < 0.001), respectively. Substantial heterogeneity was detected for both LV-GLS and RV-GLS studies, with an overall I2 statistic value of 92.0% and 89.3%, respectively (both p < 0.001). Egger's test gave a p-value of 0.10 for LV-GLS studies and 0.78 for RV-GLS studies, indicating no publication bias. In the meta-regression analysis, none of the moderators (gestational age, maternal age, maternal body mass index, maternal glycosylated hemoglobin, white ethnicity, GDM criteria, ultrasound system, frame rate, FGDM/IGDM heart rate, and anti-diabetic treatment) were significantly associated with effect modification in both groups of studies (all p > 0.05). The sensitivity analysis supported the robustness of the results.
Conclusions: GDM is independently associated with biventricular strain impairment in fetuses and infants of gestational diabetic mothers. STE analysis may allow for the early detection of subclinical myocardial dysfunction in FGDM/IGDM.
Keywords: biventricular; fetuses; gestational diabetes mellitus; global longitudinal strain; infants.