Developing and Validating a Prognostic Model for Mortality in Patients with Heart Failure
Publication Date
October 15, 2021
Creator
Abstract
Objectives: Develop and validate a prognostic model for all-cause mortality in patients diagnosed with Heart Failure. Also, to stratify mortality rates of patients diagnosed with Heart Failure, by BMI.
Study design: This was a retrospective open cohort study.
Setting: General practices in England providing data for the Clinical Practice Research Datalink.
Participants: All patients who were 18 years and over at the study start date and had been diagnosed with Heart Failure.
Sample size: The study analysed information on 109,577 patients diagnosed with Heart Failure over a 12-month follow up failure from 1 January 2000 to 31 December 2017 from the Clinical Practice Research Datalink.
Predictors: Predictors considered included age, gender, BMI, NYHA class, systolic blood pressure, prior history of left ventricular hypertrophy, prior history of hypertension, prior history of cerebrovascular disease, prior history of chronic obstructive pulmonary disease, use of diuretics, use of digoxin, use of a calcium-channel blocker, use of lipid lowering agents, use of an angiotensin-converting enzyme (ACE) inhibitor, use of a β-blocker, creatinine, bilirubin, CRP, prior history of anaemia, prior history of cardiomyopathy, dyspnoea, prior history of type-2 diabetes mellitus, albumin, race, prior history of coronary heart disease, prior history of cancer, prior history of alcohol use, smoking status, systolic blood pressure, eGFR, serum albumin, total cholesterol levels, HDL, LDL, VLDL jugular venous pressure, prior history of pulmonary congestion, prior history of atrial fibrillation, and prior history of chronic kidney disease.
Main outcome measure: 12-month mortality.
Statistical analysis: Multivariable cox regression modelling was used to determine predictors of Heart Failure and consequently, develop a 1-year prognostic model for all-cause mortality in patients with Heart Failure, at any point during the study period.
Results: 93,140 patients did not die during the follow up period and 16,437 died during the follow up period. Out of the 31 predictors assessed 29 were significant. BMI proved to be a significant predictor of heart failure mortality, with the general trend of lower BMI reducing mortality risk. The resulting prognostic model obtained a model validation score (C-statistic) of 0.79. The agreement between the observed and predicted proportion of events showed excellent apparent calibration. The D statistic of 1.65 suggests that the model provides a reasonably high amount of prognostic separation.
Conclusions: Heart failure mortality has several significantly associated comorbidities, drugs, and biomarkers. Lower BMI has been found to generally decrease mortality risk, with an exception being Obesity Class I patients having a lower hazard risk ratio than Pre-Obesity patients. Further research must be completed to better understand the complete system of predictors and confounders responsible for HF mortality.
Item Type
ethesis
Thesis Type
MRes
Supervisors
Subjects (LC)
Associated Schools / Departments
School of Medicine
eprints ID
66549
UoN Repository URI
Except where otherwise noted, this item's license is described as
File(s)![Thumbnail Image]()
Name
Dissertation final 2.pdf
Type
Full-text
Description
Examined
Size
3.3 MB
Format
Adobe PDF
Checksum (MD5)
d506400e832fb5c04cf49157102d454a