Do comorbidity patterns predict outcomes in Asian heart failure?
Asian heart-failure patients fell into five natural comorbidity clusters, and a 'lean diabetic' group had the worst quality of life and nearly three times the one-year rate of death or hospitalisation of the youngest, healthiest group.
Source
Multimorbidity in patients with heart failure from 11 Asian regions: A prospective cohort study using the ASIAN-HF registry
Study at a glance
- Design
- Cohort — Prospective registry cohort (ASIAN-HF) across 11 Asian regions; latent class analysis on baseline comorbidities, then Cox models for one-year outcomes adjusted for demographics, NYHA class and medication.
- N
- N=6480 · 6,480 patients with chronic heart failure, of whom 1,204 had preserved ejection fraction; group sizes ranged from about one thousand to 1,759. The full Methods section is missing from the available text, so details of variables fed into the latent class model come only from the abstract.
- Population
- Adults with chronic heart failure (reduced or preserved ejection fraction) enrolled from inpatient and outpatient clinics in Hong Kong, Taiwan, China, Japan, Korea, India, Malaysia, Thailand, Singapore, Indonesia and the Philippines.
- Outcome
- Composite of all-cause death or heart-failure hospitalisation within one year (primary); death and heart-failure hospitalisation separately; quality of life (Kansas City Cardiomyopathy Questionnaire); echocardiographic cardiac structure.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The researchers analysed 6,480 patients with chronic heart failure enrolled prospectively in a registry spanning 11 Asian regions between 2012 and 2016. They used latent class analysis, a statistical method that finds hidden groups, to sort patients by their pattern of comorbidities. They then compared the groups on heart structure from standardised echocardiography, quality of life, and death or heart-failure hospitalisation over one year, adjusting for age, sex, symptom class and medication.
What they found
Five groups emerged: elderly/atrial fibrillation, metabolic (obese, diabetic, hypertensive), young (few comorbidities), ischemic, and lean diabetic (diabetic and hypertensive but not obese, with much kidney disease). Within one year 19.2% of all patients died or were hospitalised for heart failure, ranging from 11% in the young group to 29% in the lean diabetic group. After adjustment, compared with the young group the lean diabetic group had a hazard ratio of 1.79, elderly/AF 1.57, ischemic 1.51 and metabolic 1.28; the elderly/AF group had the highest risk of death alone (hazard ratio 1.71) and the lean diabetic group the highest risk of heart-failure hospitalisation (hazard ratio 1.99). The lean diabetic group also had the most concentric remodelling and was concentrated in Southeast Asia.
The limits
What it doesn't show
The groups describe associations, not causes: patients were enrolled with prevalent heart failure, so survival bias and reverse causation are possible (for example, lean diabetic patients may have lost weight because they were already frailer), and unmeasured confounders may remain. Sites and patients were not randomly selected and screening logs were incomplete, so the cohort may not represent all Asian heart-failure patients. Latent class analysis is hypothesis-generating, and the available text lacks the full Methods, so exactly which variables defined the classes cannot be checked here; the ischemic hazard ratio is also reported slightly differently in the abstract and results.
Key terms
- Latent class analysis
- A statistical method that finds hidden subgroups in a population from patterns in several variables, here which comorbidities tend to occur together.
- HFpEF vs HFrEF
- Heart failure with preserved ejection fraction (the heart pumps a normal fraction but fills poorly) versus reduced ejection fraction (the heart pumps weakly).
- Hazard ratio
- How much faster an event (such as death) occurs in one group than a reference group over follow-up; 1.79 means about 79% higher instantaneous risk.
- Concentric remodelling
- Thickening of the heart's left ventricular walls relative to chamber size, often linked to high blood pressure and metabolic disease.
- Survival bias
- Distortion that arises when a study includes only people who have survived long enough to be enrolled, missing those who died earlier.
Flashcards
0 of 11 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
In the ASIAN-HF registry, which multimorbidity group had the highest one-year rate of death or heart-failure hospitalisation?
Common questions
Why group patients by comorbidity pattern instead of counting comorbidities?
Specific combinations may matter beyond the total count. Even after adjusting for the number of comorbidities, group membership still predicted outcomes, with the ischemic group notably at higher risk.
Why is the 'lean diabetic' group surprising?
In Western cohorts diabetes in heart failure usually goes with obesity, but here a large group had diabetes and hypertension without obesity, mostly in Southeast Asia, and did worse than the obese metabolic group.
Does this mean treating diabetes would prevent these outcomes?
Not on its own evidence. The study is observational and shows associations; the authors suggest cardiometabolic drugs deserve testing in such groups, which would need trials.
More on Cardiovascular