Phase angle and transthyretin in relation to all-cause mortality in hemodialysis patients: a multicenter prospective cohort study

Malnutrition is a major determinant of mortality in patients receiving maintenance hemodialysis (HD), yet its multidimensional nature is not fully captured by individual biomarkers. Phase angle (PhA), a marker related to cellular integrity and body composition, and serum transthyretin (TTR), an inflammation-sensitive biochemical marker related to visceral protein status, represent complementary domains of patient assessment. We investigated their associations with all-cause mortality and whether the PhA–mortality association differed across TTR levels. This multicenter prospective cohort included 307 adults receiving maintenance HD and followed them for a median of 36.9 months (IQR, 27.5–39.3). Baseline PhA and serum TTR were assessed concurrently. Cox proportional-hazards models adjusted for age, sex, Kt/V, dialysis session duration and frequency, dialysis vintage, and cause of kidney failure, with diabetes mellitus additionally included in Model 2. TTR was modeled continuously for formal PhA×TTR interaction testing, while its median (40.38 mg/dL) was used for exploratory stratification. Thirty-six-month discriminatory performance was assessed using time-dependent receiver operating characteristic analysis. During follow-up, 98 participants (31.9%) died. Higher PhA was associated with lower mortality in the crude model (HR, 0.850; 95% CI, 0.776–0.931; P < 0.001) and Model 1 (HR, 0.853; 95% CI, 0.771–0.944; P = 0.002), but not after additional adjustment for diabetes (HR, 0.879; 95% CI, 0.758–1.018; P = 0.085). TTR was not significantly associated with mortality in any model. In exploratory analyses, the inverse PhA–mortality association remained significant in the lower-TTR group before diabetes adjustment, whereas the higher-TTR group showed a borderline crude association (HR, 0.782; 95% CI, 0.611–1.001). Formal interaction testing showed no evidence that TTR modified the PhA–mortality association. At 36 months, both PhA and TTR showed modest discriminatory performance for all-cause mortality. Higher PhA was associated with lower all-cause mortality, although this association was attenuated after adjustment for diabetes. TTR was not significantly associated with mortality, and no evidence of PhA×TTR effect modification was observed. These findings do not support a differential PhA–mortality association across TTR levels.