Individual-ecological triangulation of risk factors for peripheral artery disease: concordant evidence for smoking and diabetes, discordant evidence for kidney disease
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Peripheral artery disease (PAD) is studied either through individual-level cohorts or national mortality statistics; the two evidence streams carry different biases and have not been formally compared within one design.
Methods
We triangulated two pre-specified streams. Individually, we analyzed 3,268 NHANES 1999-2004 participants aged 40 years and older with ankle-brachial index measurements (235 with PAD) linked to mortality through 2019, using survey-weighted logistic and Cox models, a Fine-Gray model, and attributable fractions. Ecologically, we used national Multiple Cause of Death records for 1999-2024 (390,174 PAD deaths) to estimate five-year-band trends in the share of PAD deaths mentioning tobacco (F17.0-F17.9), diabetes (E00-E89), or renal failure (N17-N19). Per a pre-specified protocol, we compared directions of association, each bridge quantified by a bootstrap direction-agreement probability.
Results
Weighted PAD prevalence was 5.05% (95% CI, 4.18 to 5.92). PAD was associated with cardiovascular death (hazard ratio, 2.65; 95% CI, 1.87 to 3.76). Smoking showed the strongest PAD association (odds ratio, 3.37) and the steepest rise in its mortality share (2.20% to 18.50%; direction-agreement probability, 1.000). Diabetes was associated with PAD (1.68) and rose after a 2014 trough (7.35% to 10.92%; probability, 1.000). Renal failure was associated with PAD (1.78) but its mortality share was flat for 26 years (14.55% to 14.49%; probability, 0.445).
Conclusions
The two streams agreed in direction for smoking and diabetes and diverged for kidney disease, reflecting different estimands. Smoking and diabetes are the factors supported by both streams for population-level PAD prevention.
