Background: Incidence rates of colorectal cancer (CRC) have increased among persons born in and after the 1960s, implicating risk factors in early life. We examined the impact of antibiotic use during childhood on risk of CRC in adult offspring of the Child Health and Development Studies (CHDS), a multi-generational cohort followed prospectively for more than 60 years.

Methods: Established in 1959, the CHDS enrolled nearly all pregnant women receiving care from the Kaiser Foundation Health Plan (Oakland, CA and surrounding East Bay) between June 1959 and September 1966, with deliveries through June 1967. Prescribed antibiotics (oral or injection) were abstracted from offspring’s medical records from birth through age five years, including the date and indicating condition. Incident diagnoses of CRC in adult offspring were ascertained through 2021 by linkage with the California Cancer Registry; we identified matched controls for each case (3:1) by randomly selecting offspring without CRC from strata of sex and birth year. We used conditional logistic regression models to estimate odds ratios (OR) for associations with antibiotic use (none vs. any) and number of prescribed antibiotics (< vs. > 90^th percentile).

Results: We identified 51 adult offspring diagnosed with CRC (diagnosed at ages 23 – 59 years; 33.3% non-Hispanic Black) and 153 sex- and birth year-matched controls. By age two years, 90.4% of cases were prescribed at least one antibiotic (mean 3.7 0, range 0 – 19) compared to 73.2% of controls (mean 3.1, range 0 – 16). This difference narrowed by age five years: 94.1% of cases were prescribed at least one antibiotic (mean 6.4, range 0 – 27) compared to 80.4% of controls (mean 6.2, range 0 – 31). The most commonly prescribed antibiotics were Bicillin CR (penicillin G benzathine / penicillin G procaine, 14.2%), Gantrisin (acetyl sulfisoxazole, 13.6%), and tetracycline (7.7%). Any antibiotic use by age two years was associated with CRC in adulthood (OR 3.32, 95% CI 1.24, 8.93), and the magnitude of the association increased with the number of prescribed antibiotics (> 90^th percentile: OR 4.43, 95% CI 1.25, 15.73). Associations persisted through age five years and were robust to adjustment for maternal race, maternal body mass index, and birth weight.

Conclusions: Our results suggest that antibiotic use during childhood is a risk factor for CRC, likely mediated by alterations in gut microbiota. Interventions addressing antibiotic stewardship in early life may be important to reduce the future burden of CRC, particularly among younger adults.

Barbara Cohn

Appearances