We have conducted a study to assess fetal LAS in relation to maternal diabetes. We successfully evaluated the impact of maternal DM on fetal atrial size (LAA) and multi chamber strain. The fetal echocardiographic changes included thicker IVS, a trend towards larger LAA and impairment of fetal LAS and RAS in fetuses exposed to maternal DM.

The multivariable modelling demonstrated that maternal DM is significantly associated with lower fetal left atrial function. Furthermore, asides from an expected association between fetal size and LA size, there is a trend towards maternal diabetes increasing LA size, albeit not reaching statistical significance. As expected, mothers who were diabetic had higher BMIs, weight and their fetuses had higher FHR.

Although RA strain was lower in maternal diabetes and RV FWS trended to be lower on multivariable adjustment determinants of RAS were EFW and RV FWS.

Our study adds to previous literature on fetal strain analysis. It reproduces findings by Miranda et al. showing no significant difference in LV GLS [5], which, is different to the observation of impaired LV GLS shown by Kurlarni et al. and Aguilera et al. [4, 19]. Furthermore, our study shows a trend to reduction in RV strain, which was also demonstrated by Miranda et al. The current study evaluated LAS and RAS and added a multivariable analysis demonstrating the independent association between fetal LAS and maternal diabetes. The results suggest that in utero exposure to maternal DM may cause early functional impairment of fetal left atrial function- independent to left ventricular function. Further, our research supports the relationship between reduced fetal left atrial shortening fractions in diabetic pregnancies presented by Zielinsky et al. [20]

DM is known to result in increased systemic inflammation and oxidation, and this may impact fetal LA strain. However, we did not collect the necessary biomarkers to assess this in this study, and future research would be important to investigate this.

We have demonstrated the impact of maternal diabetes on the developing fetus in utero using a sensitive marker of cardiac function (atrial stain).

The clinical implications of these findings would suggest that maternal DM impacts LA strain, which may be an effect of atrial impairment or underlying diastolic dysfunction. These results need to be replicated and contextualized in a future study that incorporates diastology to see if the observed differences are a result of fetal diastology .

Previous assessment of fetal LAS, due to placental dysfunction did not show significant impairment of atrial function [11]. A previous study using tissue Doppler imaging and pulse wave Doppler imaging has demonstrated changes in flow patterns in maternal obesity and diabetes [21].

Maternal diabetes and resulting fetal hyperglycaemia has been shown to impact fetal diastolic function as early as the first trimester [22], with myocardial hypertrophy and an independently reduced mitral E/A ratio as illustrated by Mohsin et al [24].

The clinical relevance of this myocardial hypertrophy is unclear as it has been shown to resolve within the first few months are resolution of fetal hyperinsulinemia [24].  The resultant fetal hyperglycaemia may cause atrial and ventricular dysfunction earlier than the time necessary for overt structural changes. This parallels findings in adults where atrial function measured by LAS is impaired in diabetic patients despite normal left atrial size [23].

This impairment in cardiac function may explain the higher fetal heart rate demonstrated in our study and previously using cardiotocography [25]. The association between fetal heart rate and atrial strain suggests the mechanism by which maternal diabetes increases fetal heart rate is related to cardiac dysfunction and the need to maintain cardiac output by increasing heart rate.

The effect of maternal DM was non significant and the primary determinants on multivariable modelling were EFW and RV FWS. This may be related to the relatively modest sample size of the current study.

There are several limitations to this study, which include the modest sample size, moderate reproducibility, modest inter-observer variability, the lack of additional maternal baseline data and the need for alternative Doppler studies to better demonstrate changes in cardiac function. Further measures of fetal diastolic function (eg. inflow Doppler, umbilical artery, MPI) were not measured in this study, and may be of significance in future studies.

We did not find the expected difference in fetal birth weight between the two groups. However, this may be due to the observed reduced fetal gestation time in mothers with DM, as the earlier delivery may explain the lack of expected difference in fetal birth weight.

Due to the sample size being modest, the trends, which were seen in LV GLS, LA size and RV FWS and maternal diabetes, may be unmasked in a larger study with higher statistical power. Indeed, the current study can help inform sample size calculations for future studies.

Modest inter-observer agreement was noted demonstrating limited reproducibility. This may indicate that further training between the operators is required to streamline measurement techniques. We did, however, perform the cardiac functionality assessment offline, which allowed the operators FP and SS to be blinded to the maternal health status, reducing bias. Post processing image analysis also allowed the operator time to accurately review the cine loop and select the clearest segment of each clip. Although our findings demonstrate a good association with DM and cardiac function, further exploration with additional data on maternal glycated haemoglobin and blood sugar levels at the time of the scan would strengthen the results.

The study may have also benefited from an addition of tissue Doppler imaging, as there are observed significant difference in flow patterns of foetuses in NDC versus diabetes mothers [21].

The high heart rate and interplay with frame rate limited our ability to assess phasic changes in LA function as outlined by Rato et al. [10]. A limitation of this study is the lack of data regarding the angle of insonation when obtaining images, which has been shown by Semmler et al. [26], to influence left ventricular global longitudinal strain significantly, and hence can be assumed to similarly effect left atrial parameters.

We did not investigate the relationship between, or stratify the results base on, different levels of control of maternal DM. Future studies with higher sample sizes would be required to further investigate this relationship.

Further measures of fetal diastolic function (eg. inflow Doppler, umbilical artery, MPI) were not measured in this study, and may be of significance in future studies.