[Association between novel high-density lipoprotein cholesterol metabolic phenotypes and all-cause and cardiovascular disease mortality in older adults in communities in Beijing].
Ye X, Du S, et al. • Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi • 2026
Novel HDL-C metabolic phenotypes mHDL-C and mHDL-C△ were significantly associated with all-cause and CVD mortality risks in community-dwelling older adults in Beijing, revealing metabolic heterogeneity and residual mortality risk undetectable by traditional HDL-C concentration.
Key Findings
Results
Both traditional HDL-C and mHDL-C were negatively associated with all-cause mortality in community-dwelling older adults.
HDL-C: HR=0.84, 95%CI: 0.79-0.91 for all-cause mortality after adjusting for confounders
mHDL-C: HR=0.84, 95%CI: 0.76-0.91 for all-cause mortality after adjusting for confounders
Study included 8,578 older adults (≥60 years) from the Beijing Healthy Aging Cohort Study with complete baseline data
Follow-up continued until March 31, 2021, with a median survival of 2,257.50 days
A total of 1,162 all-cause deaths occurred during follow-up
Results
Both traditional HDL-C and mHDL-C were negatively associated with CVD mortality in community-dwelling older adults.
HDL-C: HR=0.73, 95%CI: 0.65-0.81 for CVD mortality after adjusting for confounders
mHDL-C: HR=0.70, 95%CI: 0.61-0.80 for CVD mortality after adjusting for confounders
558 CVD deaths occurred among the 1,162 total all-cause deaths during follow-up
Cox proportional hazards regression models were used to analyze associations
Results
mHDL-C△ (the residual of mHDL-C) was positively associated with both all-cause and CVD mortality, in contrast to HDL-C and mHDL-C.
mHDL-C△ represents the difference between observed HDL-C and mHDL-C (the metabolically predicted HDL-C value)
A random forest regression model was constructed to derive mHDL-C and mHDL-C△
Results
mHDL-C△ demonstrated the best reclassification performance for predicting all-cause mortality among the HDL-C metabolic phenotypes examined.
mHDL-C△ achieved an NRI (net reclassification improvement) of 0.122 for all-cause mortality
Predictive performance was evaluated using NRI, integrated discrimination improvement, and time-dependent area under the receiver operating characteristic curve
mHDL-C△ outperformed traditional HDL-C in reclassification of all-cause mortality risk
Conclusions
The novel metabolic phenotypes mHDL-C and mHDL-C△ can reveal metabolic heterogeneity and residual mortality risk not detectable by traditional HDL-C concentration.
Traditional HDL-C concentration alone does not capture the full metabolic heterogeneity relevant to mortality risk
mHDL-C△ captures residual mortality risk beyond what is explained by traditional HDL-C levels
These indicators are suggested as potential auxiliary predictors for mortality risk and tools for cardiovascular risk stratification and precision prevention in community-dwelling older adults
What This Means
This research suggests that a standard blood test measuring HDL cholesterol (often called 'good cholesterol') may not tell the whole story about an older person's risk of dying from heart disease or other causes. Scientists developed two new ways to measure HDL cholesterol using a computer model: one called mHDL-C (which estimates what HDL-C level would be expected based on a person's overall metabolic profile) and another called mHDL-C△ (the difference between their actual HDL-C and this expected value). The study followed over 8,500 community-dwelling adults aged 60 and older in Beijing for several years, during which 1,162 died, including 558 from cardiovascular disease.
The study found that both the traditional HDL-C and the new mHDL-C measure were linked to lower risk of death — higher levels meant lower risk. However, mHDL-C△ showed the opposite pattern: people with higher mHDL-C△ had a 28% higher risk of dying from any cause and a 42% higher risk of dying from cardiovascular disease. This means that if someone's actual HDL-C is higher than expected based on their metabolism, that excess may actually signal hidden health problems rather than extra protection. The mHDL-C△ measure also did the best job of correctly reclassifying people into higher or lower risk groups compared to using traditional HDL-C alone.
This research suggests that looking beyond simple HDL cholesterol levels — by also considering how that level fits with the rest of a person's metabolism — could help identify older adults who are at higher risk of dying, even when their standard cholesterol numbers look normal. These new measures could potentially improve how doctors identify and monitor cardiovascular risk in elderly populations, though further research would be needed to confirm their clinical utility.
Ye X, Du S, Wang C, Wan Y, Chen S, Shi Y, et al.. (2026). [Association between novel high-density lipoprotein cholesterol metabolic phenotypes and all-cause and cardiovascular disease mortality in older adults in communities in Beijing].. Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi. https://doi.org/10.3760/cma.j.cn112338-20260108-00022