Association of platelet distribution width with all-cause and cause-specific mortality in US adults

Platelet distribution width (PDW) indicates heterogeneity in circulating platelet sizes. Studies reporting PDW association with mortality were limited by small sample sizes. Therefore, we examined the relationship between PDW and all-cause and cause-specific mortality in a large representative cohort.

The NHANES III data were linked to mortality files to examine the association between PDW and mortality. We excluded participants <18 years old and had a history of myocardial infarction. Since the hazards violated the proportionality assumption, we used piece-wise spline with 5-year time intervals in Cox models without and with adjustment for age, gender, race, smoking history, diabetes mellitus, hypertension, eGFR and total cholesterol.

Of 15,688 participants, 53.2% were females, 36.2% had a history of hypertension, and 6368(40.6%) died during follow-up (range 0 to 31 years). The mean (SD) age of the participants was 47(20) years, platelet count was 275.0(71.7) 10⁹/L, and PDW 16.5(0.5). In multivariable analyses, PDW was associated with all-cause mortality at 0-5 years (HR = 1.44; 95%CI = 1.21, 1.72; P < 0.001) and at 5-10 years (HR = 1.23; 95%CI =1.03, 1.46; P = 0.02). Similarly, PDW association was significant for the first 0-5 years in cardiovascular mortality (HR = 1.58, 95%CI = 1.10, 2.25; P = 0.013) and for cancer mortality (HR = 1.48 (1.15, 95%CI = 1.15, 1.91, P = 0.003). For other-cause mortality, PDW remained significantly associated for 0-5 years (HR = 1.35, 95%CI =1.05, 1.74; P = 0.02) and for 5-10 years (HR = 1.38, 95%CI = 1.05, 1.83; P = 0.023).

PDW is an independent, but time-dependent, predictor of all-cause, cardiovascular, cancer and other-cause mortality up to 5 years. The mechanisms underlying this association need further study.