Hormonal contraception and colorectal cancer: a pooled null result may obscure site-specific associations

Agreement: I Agree Body: Dear Editor Lange and colleagues provide timely, reassuring evidence from nearly two million Danish women that contemporary hormonal contraception was not substantially associated with colorectal cancer diagnosed before age 50. Their nationwide design, daily updated exposure data, and extensive sensitivity analyses offer an important basis for counselling women using modern contraceptive products.1 Yet anatomy may be hiding inside the reassuring average. The study combined ICD-10 C18, C20, and C21, grouping colon, rectal, and anal cancers. International surveillance of early onset colorectal cancer generally uses C18-C20 and excludes C21.2 A pooled estimate close to unity therefore cannot establish that associations are absent across tumours with different anatomical origins and aetiologies. The authors could resolve this using existing data: report the numbers of C18, C20, and C21 cancers; repeat the primary analysis after excluding C21; and provide separate estimates for colon and rectal cancer, with a formal test for heterogeneity. If coding permits, proximal and distal colon estimates would be equally informative. Recent whole-genome data also caution against treating early onset colorectal cancer as biologically uniform.3 Clarifying whether linked pathology data permit exploratory analyses by histological or molecular subtype would help define the limits of the pooled result. The six-month exposure window creates a second, answerable uncertainty. Combining ongoing use with use stopped within six months may reduce reverse causation near diagnosis, but it is not a latency model for colorectal carcinogenesis. Two-year and five-year exposure lags, together with exclusion of cancers diagnosed during the corresponding early follow-up periods, would show whether the null estimate persists across more plausible exposure windows. Because follow-up ended at age 50, the inference should remain restricted to cancers diagnosed before age 50 rather than lifetime colorectal cancer risk. These are not requests for a new study, but focused analyses that the existing registry data should permit. If the null association remains after C21 is excluded, colon and rectal cancers are separated, and longer lags are applied, the reassuring conclusion would be considerably stronger. Until then, the evidence supports no substantial pooled association before age 50, rather than a uniform absence of association across tumour sites and time horizons. References 1 Lange IH, Hemmingsen CH, Meaidi A, etal. Contemporary hormonal contraception and colorectal cancer in premenopausal women: nationwide cohort study. BMJ 2026;394:e100065. doi: 10.1136/bmj-2026-100065. 2 Sung H, Siegel RL, Laversanne M, etal. Colorectal cancer incidence trends in younger versus older adults: an analysis of population-based cancer registry data. Lancet Oncol 2025;26:51-63. doi: 10.1016/S1470-2045(24)00600-4. 3 Díaz-Gay M, dos Santos W, Moody S, etal. Geographic and age variations in mutational processes in colorectal cancer. Nature 2025;643:230-40. doi: 10.1038/s41586-025-09025-8 No competing Interests: Yes The following competing Interests: Electronic Publication Date: Thursday, July 16, 2026 - 16:05 AI use: No, I have not used AI Highwire Comment Subject: Contemporary hormonal contraception and colorectal cancer in premenopausal women: nationwide cohort study Workflow State: Released Full Title: Hormonal contraception and colorectal cancer: a pooled null result may obscure site-specific associations Highwire Comment Response to: Contemporary hormonal contraception and colorectal cancer in premenopausal women: nationwide cohort study Check this box if you would like your letter to appear anonymously:: Last Name: Li First name and middle initial: Jiacheng Email: lijiacheng@njucm.edu.cn Address: No. 18 Yangsu Road, Gusu District, Suzhou 215000, Jiangsu Province, China Occupation: Dr Other Authors: Xiaopeng Wang Affiliation: Suzhou TCM Hospital Affiliated to Nanjing University of Chinese Medicine BMJ: Additional Article Info: Rapid response

Rapid Response: Hormonal contraception and colorectal cancer: a pooled null result may obscure site-specific associations Dear Editor Lange and colleagues provide timely, reassuring evidence from nearly two million Danish women that contemporary hormonal contraception was not substantially associated with colorectal cancer diagnosed before age 50. Their nationwide design, daily updated exposure data, and extensive sensitivity analyses offer an important basis for counselling women using modern contraceptive products.1 Yet anatomy may be hiding inside the reassuring average. The study combined ICD-10 C18, C20, and C21, grouping colon, rectal, and anal cancers.

International surveillance of early onset colorectal cancer generally uses C18-C20 and excludes C21.2 A pooled estimate close to unity therefore cannot establish that associations are absent across tumours with different anatomical origins and aetiologies. The