Association between time-dependent postoperative vitamin D levels and survival in early-onset colorectal cancer patients undergoing curative surgery: A real-world database analysis 

Background:  Previous studies have found high vitamin D levels are associated with improved colorectal cancer (CRC) survival. However, no reports have focused on early-onset CRC (EOCRC, CRC diagnosed before age 50), whose incidence has increased dramatically in the United States and many other countries. Moreover, previous studies rarely considered the timing of vitamin D measurements, the dynamic change in vitamin D levels, and patients' comorbidity conditions, which may confound the association.  Methods:  Using the University of California Health Data Warehouse (UCHDW), an electronic health record (EHR) database encompassing data from six UC medical centers from 2012 through 2025, we initiated an EOCRC patient cohort. EOCRC patients who had no prior history of malignancy, underwent curative surgery, and had at least one vitamin D measurement after surgery were included in the study. Vitamin D level was treated as both continuous variable and categorical variable (Vitamin D deficiency: 25(OH)D < 20ng/mL; Vitamin D insufficiency: 20ng/mL ≤ 25(OH)D <30ng/mL; Vitamin D sufficiency: 25(OH)D ≥ 30ng/mL). To account for dynamic changes in vitamin D levels, postoperative vitamin D levels were treated as time-dependent variables. A time-dependent Cox regression model was performed and hazard ratio (HR) and 95% confidence intervals (CI) were estimated to measure the association between Vitamin D and EOCRC survival. Results:  We identified 424 eligible EOCRC patients, with an average follow-up time of 5.42 years. Each patient had an average of 2.68 postoperative vitamin D measurements, and 83 patients died during the follow-up period. Data was organized as person-time structure for the time-dependent cox regression model and were adjusted for age, gender, race, tumor locations, body mass index (BMI), socioeconomic status, comorbidity index, and seasons of vitamin D measurement. Patients with vitamin D sufficiency (HR 0.36, 95%CI 0.21-0.65) and insufficiency (HR 0.34, 95%CI 0.18-0.64) were associated with a better survival compared with patients with Vitamin D deficiency.  When vitamin D level was treated as a continuous variable in the model, higher vitamin D levels were still associated with better overall survival (HR 0.96, 95%CI 0.94-0.99).  Conclusion:  This is the first study to investigate the association between vitamin D levels and EOCRC survival using a time-dependent Cox regression model. Our results show that postoperative vitamin D non-deficiency status is associated with better survival compared to vitamin D deficiency. Our findings may provide evidence for clinical management of vitamin D levels in EOCRC patients.