Abstract
BACKGROUND: The long-term effects of chronic low-dose radiation exposure remain a subject of debate, with studies suggesting a range of outcomes from negligible to significant health impacts. METHODS: This multi-center cohort study followed 210 newly hired healthcare workers in radiation-related occupations for 9 years, from January 2015 to December 2023. Health effects were evaluated using Repeated Measures Analysis of Variance (rMANOVA). Significant interactions were explored using Bonferroni's post hoc analysis and Bayesian methods. Repeated-measures regression models and Restricted Cubic Splines (RCS) models were employed to further assess the statistical significance and strength of evidence. Sensitivity analyses were performed employing Generalized Estimating Equations (GEE) to ensure robustness of the results. RESULTS: The per-protocol set included 111 subjects after accounting for exclusions and losses to follow-up. The rMANOVA models identified subtle but statistically significant differences across multiple follow-up intervals for various health indices. Parameters with Bayes factors (BF(10)) greater than 100, indicating extremely strong evidence for the alternative hypothesis, included RBC, MCV, MCH, MCHC, PH, SG, TP, ALB, GLB, and A/G. Per 1-mSv increment in instantaneous Hp(10), MCV decreased by 5.96 fL (95%CI -8.92 to -3.00, P < 0.001), MCH by 2.31 pg (95%CI -3.54 to -1.08, P < 0.001), and urine PH by 1.02 units (95%CI -1.65 to -0.40, P = 0.001). Conversely, each additional 1 mSv of cumulative Hp(10) was associated with an increase of 8.22 fL in MCV (95%CI 6.54-9.91, P < 0.001) and 2.37 pg in MCH (95%CI 1.67-3.07, P < 0.001), but also with reductions in hemoglobin (-3.48 g/L, 95%CI -6.36 to -0.59, P = 0.018), plateletcrit (-0.022%, 95%CI -0.037 to -0.008, P = 0.002), large-platelet ratio (-8.40%, 95%CI -14.35 to -0.25, P = 0.006), total protein (-2.10 g/L, 95%CI -3.67 to -0.52, P = 0.009), and globulin (-1.56 g/L, 95%CI -2.80 to -0.31, P = 0.014), together with a rise in the albumin/globulin ratio (0.123, 95%CI 0.01-0.23, P = 0.031). RCS models after adjusting for age, gender and profession revealed statistically significant dose-response relationships between Ln-transformed Hp(10) and the health indices of RBC, MCHC, and A/G. CONCLUSIONS: This pioneering longitudinal cohort study in western China demonstrates that chronic low-dose radiation exposure among healthcare workers can induce subtle yet significant health effects, underscoring the necessity of adhering to the ALARA principle for radiation exposure management.