24–25 Nov 2026
IARC
Europe/Paris timezone

Longitudinal BMI Patterns to identify Sensitive Periods for Cancer Etiology and Mortality

T1-3
24 Nov 2026, 10:05
10m
RC 150 - Auditorium (IARC)

RC 150 - Auditorium

IARC

25 avenue Tony Garnier, Lyon, France
Short oral Cancer Etiology, Epidemiology & Risk Factors Cancer Etiology, Epidemiology & Risk Factors

Speaker

Leonie Courcoul (IARC)

Description

Background: Repeated measurements are increasingly available in epidemiological cohorts, creating opportunities to investigate whether exposures at different ages contribute differently to later health outcomes. However, exposure histories are often reduced to a single summary measure, limiting the ability to identify sensitive periods across the life course.

Methods: We extended a landmark-based weighted cumulative index of exposure (WCIE) framework to accommodate time-to-event outcomes. The proposed approach combines trajectory reconstruction, multiple imputation, WCIE modelling, and Cox regression to estimate age-specific associations from sparse longitudinal exposure histories. The framework was applied to body mass index (BMI) histories (BMI assessments at three timepoints throughout adulthood) from the EPIC cohort in relation to postmenopausal breast cancer and all-cause mortality after age 55 years.

Results: BMI associations varied substantially according to age at exposure. For postmenopausal breast cancer, BMI was inversely associated with risk during early adulthood (up to about age 40 years) and positively associated with risk later in life. For all-cause mortality, the strongest positive associations were observed during mid-adulthood (between 30 and 45 years old). The analyses also highlighted that the flexibility of age-specific association functions depends on the amount of longitudinal information available for trajectory reconstruction.

Conclusion: Making full use of repeated measurements can provide insights into how the association between exposure and disease risk varies across the life c course, which are not accessible with single-exposure measurements or simple exposure summaries. The proposed framework offers a practical approach for identifying sensitive periods from sparse longitudinal data and survival outcomes.

Author

Co-authors

Presentation materials

There are no materials yet.