Dose-response relationship of serum iron, ferritin, and transferrin saturation with cancer-related mortality: An analysis of population-based data.

Abstract

11043 Background: Altered iron homeostasis has been linked to cancer outcomes, with epidemiologic evidence suggesting that both reduced iron availability (low transferrin saturation or anemia) and elevated iron markers, particularly ferritin, may correspond to higher cancer mortality. However, population-level data quantifying dose–response relationships between continuous iron indices and cancer-related mortality remain limited. This study evaluates the dose–response association of serum iron, transferrin saturation (TSAT), and ferritin with cancer mortality using nationally representative NHANES data. Methods: NHANES laboratory data were linked to National Death Index mortality files (1999–2019). Survey-weighted Cox proportional hazards models and Fine–Gray competing-risk models were used to assess linear, quintile-based, and restricted cubic spline–modeled non-linear relationships between iron biomarkers and cancer mortality. Models were further adjusted to evaluate independent effects of each biomarker. Results: A total of 52,593 participants were included, with 6,875 recorded deaths. In survey-weighted log₂ Cox models, serum iron (HR 1.41; 95% CI 0.17–11.68) and ferritin (HR 1.14; 95% CI 0.84–1.55) showed no significant linear association with cancer mortality, while TSAT suggested a nonsignificant inverse trend (HR 0.30; 95% CI 0.02–5.43). Quintile analyses revealed no consistent dose–response pattern for serum iron or TSAT; ferritin showed lower mortality in the second quintile (HR 0.33; 95% CI 0.12–0.92) without a graded pattern across higher quintiles. In mutually adjusted models including all three biomarkers, none demonstrated independent linear associations. Fine–Gray analyses identified a statistically significant but clinically small association for ferritin, whereas serum iron and TSAT remained nonsignificant. Restricted cubic spline models showed no clear non-linear dose–response for any biomarker. Conclusions: Among the iron indices evaluated, ferritin demonstrated the strongest and most consistent association with cancer-related mortality, although no clear linear or non-linear dose–response pattern emerged. Ferritin may serve as a more informative marker for risk stratification, but further mechanistic and longitudinal studies are warranted.