Introduction: The World Cancer Research Foundation and the American Institute for Cancer Research have reported the role of calcium intake in reducing the risk of colorectal cancer (CRC). Nevertheless, there remains a critical need to explore whether this beneficial association varies depending on source of calcium or tumor anatomic subsites. Moreover, there are disparities in calcium intake by race and ethnicity, but the impact of low calcium consumption on the relationship with CRC in these populations is unclear.

Methods: Calcium intake was estimated from foods only (dietary calcium, mg/1000 kcal/day), including both dairy and non-dairy calcium, from supplements only (supplemental calcium, mg/day), and from both sources (total calcium, mg/day). Among 481,645 participants who were cancer-free at baseline, 10,566 incident cases of CRC were identified during 7,339,156 person-years of follow-up, with a median of 18.3 years of follow-up. Multivariable-adjusted Cox proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) with the first sex-specific quintile of calcium intake as the reference group.

Results: Median total calcium intake for the lowest (Q1) and highest quintile (Q5) was 421.5 mg/day and 1,794.48 mg/day, respectively (Q1 and Q5 was 424.9 mg/day and 1,653.8 mg/day for males and 415.5 mg/day and 1,954.4 mg/day for females). On average, dairy, non-dairy, and supplemental sources accounted for 42.1%, 34.2%, and 23.7% of total calcium, respectively. Total calcium intake was associated with lower risk of CRC (HR_Q5 vs. _Q1=0.72, 95% CI: 0.66, 0.78, P_trend<0.01). Similar results were observed by different sources of calcium (Table 1) and by tumor location (Table 2). Among non-Hispanic Black participants (N = 18,682, no. of CRC = 406), median total calcium intake was 394.5 mg/day (Q1) and 1,804.8 mg/day (Q5); Q1 and Q5 were 385.6 mg/day and 1,628.6 mg/day for males and 400.5 mg/day and 1,917.1 mg/day for females. On average, dairy, non-dairy, and supplemental sources accounted for 35.5%, 45.4% and 19.1% of total calcium in non-Hispanic Black participants, respectively. The association between total calcium and CRC was also inverse, albeit not statistically significant (HR_Q5 vs. _Q1=0.57, 95% CI: 0.30, 1.06, P_trend=0.09), among non-Hispanic Black participants, with no evidence of effect modification by race (P_{heterogeneity}=0.47).

Conclusions: A strong dose-response relationship between higher calcium intake and lower CRC risk was observed overall, regardless of source of calcium or tumor location. While calcium intake may vary by race and ethnicity, the potential for calcium to play a role in CRC prevention was consistent in non-Hispanic Black participants. Increasing calcium intake, particularly among population subgroups with lower intakes, may reduce avoidable differences in cancer risk.

Linda Liao

Appearances