Endocrine Abstracts

Background: The association between hypoglycaemia and CVEM is controversial. It is not ethical or practical to design RCTs to study this association as the primary clinical outcome measure. Therefore, utilization of epidemiological principles to study this association is essential.

Objective: To systematically explore the data from MA, RCTs and OBS on hypoglycaemia and CVEM applying Bradford Hill criteria [strength, consistency, dose response or biological gradient, specificity, temporality, biological plausibility, coherence, experimental evidence, and analogy].

Methods: A systematic literature search was performed using PubMed database, with several combinations of MeSH terms. Bibliography mining was also done on relevant articles to be as inclusive as possible.

Results: Strength: Three MAs demonstrated significant increase in CVEM; data driven from RCTs and OBS illustrated Hazard Ratio (HR) of 2.0 and 4.0, respectively. Consistency: Four RCTs and two OBS with entirely different populations illustrated increase in CVEM related to hypoglycaemia. Dose response or biological gradient: One MA demonstrated HR of 1.6 for mild hypoglycaemia versus 2.3 for severe hypoglycaemia. Specificity, temporality, biological plausibility and coherence: Not possible to assess directly through studies with CVEM measures; these criteria can be fulfilled by biological and physiological studies; hypoglycemia leads to abnormalities in myocardial perfusion, thrombotic process, cardiac repolarization; these abnormalities could potentially lead to increased risk of CVEM. Temporality has been shown with studies that utilized continuous glucose monitoring and ECG, simultaneously. Experimental evidence: Evidence exists from both interventional and observational studies with hard and surrogate outcomes. Analogy: Collectively the data from biological, physiological and outcome studies for heart disease such as heart failure illustrated increase in CVEM in population without and with diabetes through same pathways that have been observed for hypoglycaemia.

Conclusion: Available data from literature collectively illustrates that the association between hypoglycaemia and cardiovascular events fulfil Bradford Hill criteria for association.