Clinical and non-clinical markers of prognosis in heart failure

Dewan, Pooja (2021) Clinical and non-clinical markers of prognosis in heart failure. PhD thesis, University of Glasgow.

Abstract

Heart failure (HF) is a major cause of morbidity and mortality, and the prevalence of HF is only increasing globally. The rise in prevalence is primarily attributed to a combination of increasing survival especially in patients in industrialized countries and increasing incidence in low- and middle-income countries (mostly in a younger population). The clinical course of HF varies from patient to patient. For some, an initial diagnosis of HF is soon followed by multiple hospitalisations deeply impacting their quality of life, others have a fairly indolent course and some die soon after a diagnosis of HF is made. The treatment for many also depends on various factors including the phenotype of HF, the aetiology of HF and other co-existent chronic conditions to name a few. There are patients with HF who may not be candidates for intensive invasive procedures but would on the other hand benefit from supportive care and palliative care advice with treatment being directed towards preservation of quality of life. Physicians are therefore often faced with the question of the prognosis their patients with HF face. Accurate assessment of prognosis is therefore important in shared decision making for patients with HF. However, assessment of prognosis is not straightforward. Reliance on a clinician’s acumen or single prognostic markers such as left ventricular ejection fraction (LVEF) and New York heart association (NYHA) class can be inaccurate and is not advised. Therefore, multivariable models were turned to in order to paint a more accurate picture of a patient’s prognosis by incorporating different individual markers known to be associated with clinical outcomes in HF. Multiple prognostic models have consequently been developed for assessment of prognosis in HF. However, uptake of these in clinical practice remain low. Many factors contribute including issues with reproducibility of prognostic ability in different populations, unavailability of variables and complexity of statistical methodologies. The evolving risk of different outcomes due to pharmacological and non-pharmacological advances in HF is another influencing factor. I consequently conducted a systemic analysis of the literature of prognostic models in HF – focusing primarily on a single phenotype of HF – HF with reduced ejection fraction (HFrEF). I identified several variables common to most models, with LVEF, sex, age, NYHA class being some of the most frequently featured. Inclusion of more contemporary prognostic markers such as NT-proBNP and non-clinical markers such as region, race/ethnicity and socioeconomic status was however very less frequent or absent altogether. Given this background, the aim of this thesis was to explore a select set of clinical and non-clinical markers, some of which have featured in previous models to review their prognostic importance along with a few which have not been featured in risk models in the past. The analyses presented were conducted in three contemporary clinical trial datasets in HFrEF – ATMOSPHERE, PARADIGM-HF and DAPA-HF. I used a variety of statistical measures to assess the association between 3 commonly used markers – LVEF, sex & chronic obstructive pulmonary disease (COPD) and 4 uncommonly/previously unused markers – geography & ethnicity, income inequality and frailty – and common clinical outcomes examined in HF. Different outcomes were tested – including cardiovascular, non-cardiovascular & all-cause death and first & recurrent HF, cardiovascular & all-cause hospitalisations. Cox regression was used to study the association between LVEF and COPD with various clinical outcomes. I used competing risk regression to study the other markers of prognosis and their association with clinical outcomes. In the DAPA-HF cohort, each 5% decrease in LVEF was associated with a 20% higher risk of HF hospitalisation (95% CI 1.13 – 1.27) and a 20% higher risk of cardiovascular death (95% CI 1.13 – 1.28). The risks of the same outcomes in those with COPD was 78% (95% CI 1.44 – 2.20) and 28% (95% CI 1.00 – 1.63) respectively. The rest of the analyses were carried out in a pooled cohort of the ATMOSPHERE and PARADIGM-HF trials. Women had a 19% lower risk of HF hospitalisation (95% CI 0.74 – 0.90) and 26% lower risk of cardiovascular death (95% CI 0.67 – 0.81). Among the Asian countries, the highest and lowest risk of hospitalisation for HF was seen in patients belonging to Taiwan (1.88; 95% CI 1.46 – 2.42) and India (0.44; 95% CI 0.36 – 0.54) respectively. In the same patients living in the Philippines had the highest risk of cardiovascular death (sHR 1.87; 95% CI 1.36 – 2.57) and the lowest risk of the same outcome was seen in those living in Japan (subdistribution hazard ratio (sHR) 0.68; 95% CI 0.46 – 0.98). When levels of income inequality were examined, patients lining in countries with the greatest inequality had a 57% higher risk of hospitalisation for HF (95% CI 1.36 – 1.81) and the risk of cardiovascular death was 50% greater (95% CI 1.29 – 1.74) compared to patients living in countries with the lowest income inequality. Using an acceptable method, I found that 69% of the population in ATMOSPHERE and PARADIGM-HF were frail. In the same population, the frailest patients carried a 89% higher risk of HF hospitalisation (95% CI 1.69 – 2.11) and the sHR for cardiovascular death was 2.14 (95% CI 1.92 – 2.38). All the above listed associations were statistically significant. In conclusion, I found that a select set of traditionally featured markers in prognostic models in HF remained strong predictors of hospitalisation and mortality in contemporary set of HF populations. In addition, several non-clinical and clinical markers that have infrequently featured in previous prognostic markers also carry significant value in measuring risk of clinical outcomes in HF. The inclusion of such markers may improve the predictive ability and clinical applicability of prognostic models in HF in the future.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
• Supervisor's Name: McMurray, Prof. John and Jhund, Prof. Pardeep
• Date of Award: 2021
• Depositing User: Theses Team
• Unique ID: glathesis:2021-82670
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 03 Feb 2022 16:32
• Last Modified: 08 Apr 2022 16:53
• Thesis DOI: 10.5525/gla.thesis.82670
• URI: https://theses.gla.ac.uk/id/eprint/82670

Related URLs

• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI