The pharmaco-epidemiology of loop diuretic dispensing and its relationship to the diagnosis of heart failure and to prognosis

Friday, Jocelyn Marie (2023) The pharmaco-epidemiology of loop diuretic dispensing and its relationship to the diagnosis of heart failure and to prognosis. PhD thesis, University of Glasgow.

Abstract

Heart failure is a major and growing public health problem associated with poor patient outcomes, including reduced quality of life and high hospitalisation and mortality rates. It is a complex clinical syndrome rather than a single disease, which lacks a practical, universal, and standardised definition. Currently, the definition relies on the identification of symptoms and signs of cardiac dysfunction, such as ankle swelling and breathlessness, which are neither specific nor objective. Many patients are only diagnosed once their symptoms and signs are severe enough to require hospitalisation. Pathophysiologically, heart failure can be defined by the presence of salt and water retention, also known as congestion, associated with cardiac dysfunction. Within the United Kingdom, the pharmacological class of loop diuretics is used primarily for the treatment of congestion due to cardiac dysfunction. The aim of this thesis is to investigate the pharmacoepidemiology of loop diuretic dispensing and its relationship to the diagnosis of heart failure, with a particular focus on patient outcomes.

The first analysis describes the prevalence of repeated loop diuretic dispensing and/or diagnosis of heart failure within the NHS Greater Glasgow & Clyde Health Board population on 1st January 2012, including patient outcomes over the following five years. This research is thought to be the first population-level investigation into the prevalence of repeated loop diuretic dispensing and its prognostic significance in patients with and without a diagnosis of heart failure. The analysis found that an estimated 3.2% of the population received repeated loop diuretic dispensing, while only 1.3% of the population had a diagnosis of heart failure. Hospitalisation rates were higher in those with a loop diuretic (0.99 admissions per patient-year at risk for those with only repeated loop diuretic dispensing and 1.51 admissions per patient-year at risk for those with both) than those with only a diagnosis of heart failure (0.93 admissions patient-year at risk). All-cause mortality followed a similar pattern; adjusting for age, sex, socioeconomic deprivation and comorbidity status, the 5-year hazard ratio and (95% confidence interval) were 1.8 (1.8 - 1.9) for those with those only repeated loop diuretic dispensing and 2.3 (2.2 - 2.4) for those with both, while only 1.2 (2.2 - 2.4) for those with only a diagnosis of heart failure, implying that the presence of repeated loop diuretic dispensing is a marker of serious disease.

The second analysis stepped backwards in ‘patient-time’ to describe the pattern of hospitalisations in the year leading up to the initiation of loop diuretic dispensing or an incident diagnosis of heart failure using network graphs. While the precursors to heart failure are known, this research is thought to be the first to report the common patterns in events leading up to the initiation of loop diuretics. While there was little difference in comorbidity and medication levels 24 months prior, in the year leading up to the initiation, those who received a diagnosis of heart failure were more likely to be admitted for well-recognised contributors to the condition, including ischaemic heart disease in particular, but also atrial fibrillation/flutter and valve disease. In contrast, these patterns were not often seen in those who were only initiated on a loop diuretic, instead with a focus on admissions for non-specific symptoms and signs, most commonly unspecified chest pain.

The third analysis starts where the second leaves off. It assesses the prognostic relationship between the initiation of loop diuretic and diagnosis of heart failure on mortality and whether the sequence of these events matters using semi-Markov multi-state modes, a flexible model for use on longitudinal time data where there is an event-related dependence on outcomes. Those on repeated loop diuretic dispensing without a diagnosis of heart failure were majority women (62%). Many with evidence of left atrial dilation (53%), while those with a diagnosis of heart failure without a repeat loop diuretic were majority men (63%). Many had a history of myocardial infarction (51%). Hospitalisations and mortality were higher in those with a repeat loop diuretic (within the first year per patient-year at risk: hospitalisation, 1.44; mortality, 0.20) compared to those with a diagnosis of heart failure without a repeat loop diuretic (within the first year per patient-year at risk: hospitalisation, 1.47; mortality, 0.14). Rates were higher still in those with both loop diuretic and heart failure (where both events occurred together within the first year per patient-year at risk: hospitalisation, 1.74; mortality, 0.16; or where the diagnosis of HF preceded the initiation of loop diuretic, within the first year per patient-year at risk: hospitalisation, 1.68; mortality, 0.20), with the highest being in those who initiated the loop diuretic in advance of receiving a diagnosis of heart failure (within the first year per patient-year at risk: hospitalisation, 2.26; mortality, 0.28).

The fourth and final analysis subsets the population to investigate the mortality of the 24,921 patients with ischaemic heart disease according to whether or not they have had a repeat loop diuretic and/or diagnosis of heart failure; of whom, 3,806 had only repeat loop diuretic, 2,384 had only a diagnosis of heart failure, and 3,531 had both. This analysis found that after adjusting for age, sex, and other prognostic markers, mortality was associated with the repeat loop diuretic regardless of the patient’s heart failure status. Those with a repeat loop diuretic without a diagnosis of heart failure experienced substantially higher rates of cardiovascular (an estimated 15%) and all-cause mortality (47%) than those with a diagnosis of heart failure without a repeat loop diuretic (an estimated 8% cardiovascular and 19% all-cause mortality), while rates were highest for those with both (an estimated 25% cardiovascular and 57% all-cause mortality).

In conclusion, these analyses found that many more patients are repeatedly treated with loop diuretic than ever receive a diagnosis of heart failure. These patients are at a high risk of hospitalisation and death, and based on their characteristics, many probably have undiagnosed heart failure. From a public health and epidemiological perspective, the current definition of heart failure likely underestimates the true burden on the healthcare system. From the patient’s perspective, with the efficacy of angiotensin receptor-neprilysin inhibitor, sodium-glucose co-transporter-2 inhibitors, and mineralocorticoid receptor antagonistss, a missed diagnosis means a missed opportunity to improve the patient’s outcome and quality of life, regardless of their heart failure phenotype. Even more alarming, if these patients are receiving the loop diuretic inappropriately, the loop diuretic is likely causing these increased hospitalisation and mortality rates. If the loop diuretic can be safely withdrawn, other medications with diuretic properties exist which have good safety profiles. Ultimately, further research is required to determine the optimal strategy for managing these patients.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: R Medicine > RC Internal medicine; R Medicine > RM Therapeutics. Pharmacology
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
• Supervisor's Name: Cleland, Professor John, Wolters, Dr. Maria, Lewsey, Professor Jim and McAllister, Professor David
• Date of Award: 2023
• Depositing User: Theses Team
• Unique ID: glathesis:2023-83886
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 30 Oct 2023 15:02
• Last Modified: 30 Oct 2023 15:02
• Thesis DOI: 10.5525/gla.thesis.83886
• URI: https://theses.gla.ac.uk/id/eprint/83886