No female excess in peripheral arterial disease after age 60

Background. Global estimates report peripheral arterial disease (PAD) to be more prevalent in women than in men, with the difference widening across age, despite women carrying a lower burden of every major associated risk factor. These comparisons are crude rather than age-standardized. Methods. We analyzed 7140 US adults aged 40–85 years with measured ankle–brachial index (ABI) in NHANES 1999–2004, using the survey design throughout (Taylor lineariza- tion; 43 strata, 87 primary sampling units). We decomposed the crude sex difference in PAD prevalence into rate and age composition components, estimated age-standardized differences, and—because ABI is a ratio—decomposed the sex difference in log ABI exactly into ankle (numerator) and brachial (denominator) contributions. Prognostic consequences were evaluated against 2780 deaths through 2019. Results. Crudeprevalencewas6.32%inwomenand5.01%inmen(difference1.31percentage points, 95% CI 0.11 to 2.51). Kitagawa decomposition attributed 46.7% of this to age composition: women were 5.9% of their own sex at age 80 and older versus 3.9% of men, in a stratum where prevalence approaches 24%. After age standardization the difference fell to 0.70 percentage points and was no longer distinguishable from zero, consistently across three reference populations (p= 0.22 to 0.24). Stratified by age, there was no detectable female excess at 60 years and older (adjusted OR 0.99, 95% CI 0.81 to 1.22; 504 cases), unchanged excluding the top-coded band (60–79 years: OR 1.00, 95% CI 0.77 to 1.29). This agrees with the modeled high-income estimates themselves, which imply odds ratios of 0.83 to 0.98 across ages 60–79 and exceed unity only in strata above 80 years. The residual rate component lay below 60, but a formal age interaction was not significant (p= 0.11). Nonetheless women’s ABI was lower throughout (1.080 versus 1.123; difference−0.044, 95% CI−0.051 to−0.036). This deficit was not explained by risk factors, by effect modifica- tion (F = 1.56, p= 0.21), or by stature (height coefficient 0.00080 per cm versus the 0.00286 required; z=−7.39, p= 1.5 × 10−13). It was confined to the bulk of the distribution: the sex gap was−0.05 at the median but zero at the 2nd percentile, and the prevalence difference was monotone in the cutpoint, vanishing at ABI ≤ 0.80. The deficit changed origin with age—an ankle deficit before 50 (−0.042), and after 60 an entirely brachial excess with ankle pressures equal or higher in women (interaction p= 3.2 × 10−5)—accompanied by a widening pulse pressure difference (p= 1.5 × 10−13) with diastolic pressure lower in women at every age. Non-compressibility, which could artificially raise ankle pressure, was three- to sixfold more common in men after 60, and excluding it strengthened the finding. Conditional on ABI entered flexibly, brachial pressure retained independent prognostic information (β = 0.65, p= 0.022). Interpretation. The reported female excess in PAD prevalence is substantially an artifact of crude comparison across differing age structures, and is absent after age 60—the very range in which modeling studies report it to be largest. Women nonetheless have systematically lower ABI after 60, but this arises from the brachial denominator and is confined to the middle of the distribution, so it depresses the index without producing additional cases at the diagnostic threshold.