Electrocardiography for cardiac risk stratification in haemodialysis patients

Abstract

Cardiovascular disease is common in end stage renal disease. Emerging evidence suggests that repolarisation abnormalities play an important part in the pathogenesis of cardiac mortality. Risk stratification methods have traditionally relied on echocardiography. The ECG is in our opinion under-utilised for risk stratification in haemodialysis. The purpose of this thesis was to test the hypothesis that traditional and novel ECG parameters are independently predictive of cardiac outcomes in haemodialysis. Were this case, the ECG would have the potential to be used as a screening tool in haemodialysis patients. In the first study (chapter 4) we examined the prognostic potential of QRS - T angle, an indicator of repolarisation heterogeneity that has been shown to be associated with cardiac outcomes in other populations. In our cohort of haemodialysis patients (n=171) we demonstrated that QRS -T angle carries independent prognostic value for MACE (follow up 2.3 ± 1.1 years): hazard ratio 2.215, (1.188 - 4.131), p=0.012. In the next study (chapter 5) we looked specifically at the presence of ECG strain and demonstrated that is associated with increased risk of MACE (HR 2.961, CI: 1.254 - 6.990, p= 0.013), but not all cause mortality, independently of LVH. For the following study, we aimed to evaluate these biomarkers already used as diagnostic tests and assess whether an ECG based model can be reliably used as a diagnostic test for pending MACE in haemodialysis patients. We found that ECG biomarkers were overall very poor as a screening tool for cardiovascular outcomes and all cause mortality. This was evidenced by the very low sensitivity and very low AUC values on the ROC curves of all of them. In chapter 6 we evaluated the diagnostic accuracy of electrocardiographic methods of calculating LVH compared to RT3DE and demonstrated that ECG methods for assessment of LVH that rely on voltage criteria have very low sensitivity and unreliable specificity compared to RT3DE and also conventional M- Mode echocardiography. In the last study (chapter 7) we explored whether KIM-1 (a novel biomarker of cardiovascular risk in ESRD) exhibits any association or correlation with ECG recorded parameters of abnormal conduction and arrhythmia. Based on our data, we could not identify a correlation between KIM-1 and the above biomarkers. Our results do not support the above hypotheses. Individual ECG markers demonstrate variable degrees of association with cardiac outcomes in particular group of patients, but this association was not reliably reproduced and often ceased to exist with the addition of other risk factors. A limitation of the findings is the possibility that the small sample size in some studies prevented conclusive results.

Date of Award	1 Aug 2021
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Dimitrios Poulikakos (Supervisor), Philip Kalra (Supervisor) & Darren Green (Supervisor)