Iron Deficiency Anemia and Hematochezia Are Red Flags for Colorectal Cancer in Young Patients

Demb J, Liu L, Murphy CC, et al. Young-onset colorectal cancer risk among individuals with iron-deficiency anaemia and haematochezia. Gut 2021;70:1529–1537. The incidence of colorectal cancer (CRC) diagnosed in individuals aged <50 years, or young-onset CRC (YCRC), is increasing. Iron deficiency anemia (IDA) and hematochezia are potential red flag symptoms of YCRC; however, there are few data on the magnitude of risk associated with these symptoms; therefore, mixed guidance on appropriate evaluation of these symptoms in patients aged <50 years. Demb et al (Gut 2020;70:1529–1537) conducted a retrospective matched cohort study of US veterans aged 18–49 from 1999 to 2016 to examine the association between IDA/hematochezia and YCRC. The study population was identified from validated Veterans Health Administration databases, allowing investigators to link person-level data to cause-specific mortality. For the IDA cohort, they included veterans who met the World Health Organization criteria for IDA (hemoglobin of <130 g/L in men, <120 g/L in women; a ferratin level of ≤15 ng/mL; or a transferrin saturation of ≤16%). For the hematochezia cohort, veterans were identified by International Classification of Diseases 9th or 10th edition codes. For each veteran with IDA and hematochezia, 4 controls matched by birth year, sex, and first visit in the Veterans Health Administration system were included. All patients had ≥5 years of available follow-up; those with YCRC, inflammatory bowel disease, or previous diagnostic codes for IDA were excluded. The primary outcome was YCRC within 5 years of start of follow-up. The 5-year YCRC incidence was derived using Kaplan–Meier estimation. Hazard ratios were estimated using mixed Cox regression models adjusted for race/ethnicity, body mass index, diabetes, smoking status, and aspirin use. Subgroup analyses were conducted based on age and sex. The number needed to scope (NNS) to identify 1 YCRC case was estimated by computing the inverse of YCRC prevalence among exposed individuals over the postulated baseline CRC prevalence. In the IDA cohort, 47,800 veterans were matched with 191,200 controls with 0.8 million person-years of follow-up. Among those with IDA, there were 184 YCRC cases (cumulative incidence 0.45%; 95% confidence interval [CI], 0.38%–0.51%) versus 73 YCRC cases in the controls (cumulative incidence, 0.05%; 95% CI, 0.04%–0.07%), with an adjusted hazard ratio of 10.81 (95% CI, 8.15–14.33). A sensitivity analysis yielded similar results when women with prior menorrhagia or hysterectomy were excluded. The 5-year cumulative incidence in those with IDA increased with age (0.14% in those aged <30 years; 0.14% vs 0.61% in those aged 40–49 years), with similar results when further stratified by sex. Overall NNS for those with IDA was 259.8 (95% CI, 226.8–301.6). In the hematochezia cohort, 130,748 veterans were matched with 522,992 controls with 2.62 million person-years of follow-up. Among those with hematochezia, there were 406 YCRC cases (cumulative incidence, 0.33%; 95% CI, 0.30%–0.36%) versus 150 YCRC cases in the controls (cumulative incidence, 0.03%; 95% CI, 0.03%–0.04%), with an adjusted hazard ratio of 10.66 (95% CI, 8.76–12.97). The 5-year cumulative incidence in those with hematochezia increased with age (0.05% in those aged <30 years; 0.14% s 0.50% in those aged 40–49 years), with similar results when further stratified by sex. The overall NNS for those with hematochezia was 322.0 (95% CI, 293.4–359.1). Although the exact etiology to the rising incidence of CRC in young patients is unknown, there are multiple immediate strategies to decreasing the burden of CRC in this population. One approach is to optimize prompt evaluation of, and follow-up of signs and symptoms associated with CRC. Unfortunately, those under the age of 50 have a 152- to 217-day delay from time of symptom onset to CRC diagnosis compared with 29.5–87.0 days in those aged >50 years (Adv Cancer Res 2021;151:1–37). Although IDA and hematochezia are established signs and symptoms of CRC, they have alternate benign causes in young individuals, such as hemorrhoids or menorrhagia. Thus, the magnitude of CRC risk associated with these clinical presentations in young patients is unclear. Accordingly, medical providers may delay or forgo endoscopic evaluation assuming benign etiologies, leading to delays in YCRC diagnosis. Establishing the magnitude of CRC risk associated with IDA and hematochezia can inform the priority of endoscopic evaluation of these clinical presentations in young patients. Demb et al found a 10-fold increased risk of YCRC for those with either IDA or hematochezia. As expected, the risk of YCRC increases with age approaching 50 years. These results reinforce the critical importance of endoscopic evaluation at a younger age in all patients with these signs or symptoms. It is important to acknowledge some of the limitations in generalizing the results from this study. Although the sample size was large and data source was comprehensive, the veteran population may not be generalizable to the non-veteran public. Although the study accounted for available risk factors that may be disproportionately present in veterans (male sex, smoking status), there may be other service-related factors that influence cancer risk. Furthermore, the authors were unable to adjust for family history of colorectal neoplasia or presence of hereditary cancer syndromes. The severity of signs and symptoms could not be adjusted for. Although the authors were able to collect stage of CRC at diagnosis, it is unclear if colonoscopy at onset of signs or symptoms could have downgraded the stage of diagnosis. Despite these limitations, this study has several important strengths. The authors leveraged a comprehensive database linking patient-specific data to mortality-cause data within the largest health care system in the United States. They were able to adjust for CRC risk factors such as race/ethnicity, body mass index, smoking status, aspirin use, and diabetes. Multiple sensitivity analyses were completed to explore potential unknown confounders, which all yielded similar results. There are several points of interest in the data presented. Of those diagnosed with YCRC, 44% and 14% of cancers were located in the proximal colon in the IDA and hematochezia cohorts, respectively. This finding suggests that colonoscopy, as opposed to flexible sigmoidoscopy, is the optimal endoscopic diagnostic tool for evaluation of IDA and hematochezia. Furthermore, the authors estimated that approximately 260 and 322 colonoscopies, respectively, need to be completed to diagnose 1 case of YCRC in those presenting with IDA or hematochezia, respectively. They point out this is NNS is lower than the NNS to detect 1 CRC case in average-risk individuals aged 50–75 years. Thus, although there is an overwhelming shortage of colonoscopy access within the VA system (J Gen Intern Med 2020;35:1776–1782), this study reinforces that colonoscopy resources should be secured for those with IDA and hematochezia. The subgroup analyses completed suggest that the increased risk remains when stratified by age and sex, but the highest risk is in males aged 40–49 years. Thus, in resource restricted environments, it may be reasonable to prioritize colonoscopy accordingly. This study reinforces that IDA and hematochezia are important red flags for YCRC and should not be dismissed. These results strengthen the body of evidence supporting timely colonoscopy in individuals under age 50 presenting with IDA or hematochezia and all efforts should be made to ensure colonoscopy access for this population.