HbA1c levels in non-diabetic older adults - No J-shaped associations with primary cardiovascular events, cardiovascular and all-cause mortality after adjustment for confounders in a meta-analysis of individual participant data from six cohort studies.
Level 3 - non-randomized controlled study
Individual participant data meta-analysis of prospective cohort studies
PubMed 26867584 · doi:10.1186/s12916-016-0570-1
What was done
Individual participant data meta-analysis of six population-based cohort studies from Europe and the United States including 28,681 non-diabetic adults aged 50 years or older. Participants were followed for a mean of 10.7 years to assess associations between baseline HbA1c categories (<5.0%, 5.0 to <5.5% [reference], 5.5 to <6.0%, and 6.0 to <6.5%) and all-cause mortality, cardiovascular mortality, and primary cardiovascular events (myocardial infarction or stroke) using multivariable Cox proportional hazards models.
What was found
Over follow-up, 6,769 participants died (2,648 from cardiovascular causes) and 2,493 experienced a primary cardiovascular event. A linear association was observed for primary cardiovascular events across HbA1c categories. Adjusting for cardiovascular risk factors reduced excess risk by approximately 50%, resulting in hazard ratios (95% CI) for the highest category (6.0 to <6.5%) of 1.14 (1.03-1.27) for all-cause mortality, 1.17 (1.00-1.37) for cardiovascular mortality, and 1.19 (1.04-1.37) for cardiovascular events. Associations for very low HbA1c (<5.0%) were inconsistent across cohorts and statistically non-significant in pooled analyses. In NHANES, where an initial J-shaped curve appeared, the elevated risk at very low HbA1c lost significance after adjusting for race/ethnicity, alcohol intake, BMI, iron deficiency anemia, and liver function biomarkers.
Why it matters
Apparent J-shaped risk curves linking low HbA1c to increased mortality in non-diabetic older adults are likely artifacts of confounding from underlying conditions like anemia or liver dysfunction rather than true glycemic toxicity.
Limits
Observational design precludes causal inference, and residual confounding cannot be ruled out. HbA1c was measured only at baseline. Biomarker adjustments for liver function and anemia could only be evaluated within specific cohorts that had those measurements available.
Cited by
- contradicts All-cause mortality data show better outcomes for individuals with a hemoglobin A1c of 5.0% compared to 5.5%.