Long-Term Outcomes in Survivors of Critical Illness and Interaction with Glucose Metabolism

Abstract

Even after surviving critical illness, many patients who are discharged from an Intensive Care Unit (ICU) die or suffer major morbidity. Diabetes is a risk factor for critical illness and its severity and critical illness also affects immediate glucose metabolism. In addition, substantial similarities exist between complications of critical illness and diabetes (nephropathy and neuropathy), such that critical illness may exacerbate complications of diabetes. Hitherto, the longer-term interaction between critical illness, glucose metabolism and outcomes had been scarcely evaluated. The aims of this thesis were to 1) determine long-term outcomes and evaluate follow-up care of ICU survivors with diabetes, 2) optimise glycaemic targets for critically patients with diabetes during the entire hospitalisation, 3) evaluate the long-term impact of stress hyperglycaemia, and 4) evaluate effects of critical illness on the autonomic nervous system, particularly with respect to nutrient intake and glucose metabolism. A large state-wide epidemiological study was conducted (Chapter 2.2) which identified that ICU survivors with diabetes experience greater loss of life-years after hospital discharge than those without diabetes. There is, therefore, a rationale to improve outcomes in this group. A novel intervention (an intensivist-endocrinologist ICU follow-up clinic) was created and a randomised controlled feasibility study was designed (Chapter 2.3). Although patients perceived that the intervention enhanced their recovery, the study did not meet feasibility criteria because of incomplete outcome data with high death rates. Survivors also experienced poor functional outcomes and high healthcare use (Chapter 2.4). In order to personalise glycaemic control during hospitalisation, it is necessary to rapidly determine premorbid glycaemia. A single centre observational study was undertaken to establish that point-of-care glycated haemoglobin testing is both accurate and feasible in ICU (Chapter 3.2). Despite substantial focus on glycaemia during ICU admission, data following ICU discharge are limited. Combined continuous glucose and electrocardiograph monitoring was utilised in a two-centre prospective cohort study to evaluate glycaemia and cardiac arrhythmias in patients prescribed insulin after ICU discharge (Chapter 3.3). Hypoglycaemia occurred frequently, was often nocturnal and asymptomatic, and increased the risk of bradycardia. These findings suggest that cautious prescription of insulin upon ICU discharge is necessary. Stress hyperglycaemia occurs frequently during critical illness in those without diabetes, but has traditionally been considered a transient phenomenon. A systematic review and metaanalysis was undertaken ascertaining that stress hyperglycaemia is associated with subsequent increased risk of both prediabetes and diabetes (Chapter 4.2). The potential mechanisms underlying this relationship and clinical implications were also addressed (Chapter 4.3). Autonomic neuropathy was evaluated in older patients at 3 and 12 months after ICU discharge using a longitudinal cohort study design (Chapter 5). Autonomic function was assessed specifically in the context of nutrient ingestion and glycaemia. Postprandial hypotension occurred frequently in this group and increased the risk of falls. In summary, this thesis has contributed to the understanding of long-term outcomes of ICU survivors with diabetes, as well as glycaemia and autonomic function after ICU discharge, and has evaluated a novel approach to follow-up care in survivors of critical illness with diabetes.