Baseline serum C-reactive protein and death from colorectal cancer in the NHANES III cohort.

ABSTRACT

Several prospective studies suggest that C-reactive protein (CRP), a nonspecific serologic marker of inflammation, might be linked to risk of colorectal cancer (CRC), whereas others have reported null or protective effects. We analyzed data from 7,072 participants (50-85 years) in the U.S. National Health and Nutrition Examination Survey III (1988-1994), a nationally representative cohort (n = 33,994; 2 months-85 years) with vital status follow-up to 2000. Hazard ratios (HRs) for mortality associated with baseline clinically raised (≥1.00 mg/dL) and intermediate (≥0.22-0.99 mg/dL) CRP levels were estimated using Cox proportional hazards regression controlling for CRC risk factors. There were 59 deaths from CRC, 106 from other obesity-related cancers (other-ORCs) and 1,130 from cardiovascular disease (CVD). Participants with clinically raised CRP at baseline were found to have a statistically significant greater risk of CRC death (HRs = 2.36-2.47) in comparison to persons with undetected levels. HRs were lower for death from other-ORC and CVD (1.82, 95% CI 1.05-3.15; 1.53, 95% CI 1.29-1.81, respectively). Intermediate CRP level was associated with a nonsignificant 10-21% increased risk for CRC death. HR for CRC death was higher among persons with a normal BMI (2.16, 95% 0.96-4.87, p = 0.06) compared to those who were overweight (1.22, 95% CI 0.53-2.78) or obese (1.23, 95% CI, 0.37-4.08). A similar pattern was observed for waist circumference. This effect modification suggests that the impact of chronic inflammation may be independent of excess body fat. Future research is recommended to confirm emerging data that elevated serologic CRP might reflect underlying colonic inflammation.