Environment and Breast Cancer: Science Review

Evidence From Humans
Print this page
Childhood and Adolescent Pesticide Exposure and Breast Cancer Risk
Niehoff, N. M., Nichols, H. B., White, A. J., Parks, C. G., D'Aloisio, A. A., Sandler, D. P. Epidemiology. 2016. 27:3, 326-33.
Topic area
Environmental pollutant - Pesticides
Study design
Prospective cohort
Funding agency
NIEHS National Center for Advancing Translational
Study Participants
Menopausal Status
The menopausal status of women included in this study is listed here.
Stratified analysis based on menopausal status
Number in Cohort
Cohort: 50,756
Country where study was conducted
USA including Puerto Rico
Cohort participation rate
Initial response rate not reported. >95% for annua
Participant selection: Inclusion and exclusion criteria
Criteria used to select participants in the study.
50,884 participants were enrolled in the Sister Study, a prospective cohort study of women ages 35-74 with sisters who had been diagnosed with breast cancer. Participants were enrolled between 2003 and 2009 and were initially free of breast cancer. Women diagnosed with breast cancer prior to completing enrollment in the study or missing date of diagnosis were excluded from analysis (128 women).
Exposure Investigated
Exposures investigated
Self-reported residential history at baseline, including residence of longest duration before age 14. Detailed questions about each residence included whether it was used as a farm or orchard; whether it was in close proximity to orchards, golf courses, n
How exposure was measured
Questionnaire, by computer-assisted telephone interview
Exposure assessment comment
Exposure misclassification is a concern, given that participants were asked to recall childhood exposures. Exposure misclassification is expected to be non-differential in this prospective study, so effect estimates may be biased toward the null. For a subset of women, authors estimated potential exposure misclassification by asking participants' mothers to complete the questionnaire assessing their daughters' exposures. With mothers' recall treated as gold standard, exposure classification seemed somewhat better for farm-related exposures than for residential exposures. Details about specific products or chemicals used in pesticide application were not collected, and no metric of cumulative exposure was used for analysis, which may have led to substantial variability of exposure within the same analysis group. Direct exposure to spray of fogger trucks or airplanes before 1975 was used as a proxy for acute DDT exposure.
Early life exposures considered
Yes, residential exposures < 14 years of age and farm exposures < 18 years of age
Breast cancer outcome investigated
Primary incident breast cancer
Confounders considered
Other breast cancer risk factors, such as family history, age at first birth, and hormone replacement therapy use, that were taken into account in the study.
Based on causal diagram: Age at menarche, race, parity/breastfeeding, age at first birth. Authors report no difference in sensitivity analyses restricting to women with normal BMI or who never used HRT. They also state they considered the impact of adjust
Genetic characterization included
If the study analyzed relationships between environmental factors and inherited genetic variations, this field will be marked Yes. No, if not.
Strength of associations reported
For residential exposure ages 0-14 (reference = residence not treated with pest control chemicals):

Residence treated, no personal application: aHR 0.93 (95% CI 0.83, 1.0)
Residence treated, personal application at least some of the time: aHR 0.92 (95% CI 0.68, 1.2)

For farm exposure ages 0-18 (reference = never lived on a farm ≥ 12 months):

Lived on cotton or tobacco farm: aHR 1.3 (95% CI 1.0-1.6)
Lived on livestock farm: aHR 1.0 (95% CI 0.90-1.1)
Lived on farm growing field or orchard crops: aHR 1.0 (95% CI 0.89-1.1)
Lived on farm growing other cash crops: aHR 1.0 (95% CI 0.84-1.3)
Lived on a farm that used pesticides: aHR 0.95 (95% CI 0.82-1.1)

For farm exposure ages 0-18, among women who lived on a farm ≥ 12 months (reference = lived on a farm for ≥ 12 months that did not use pesticides):

Personally mixed, loaded, applied or helped others mix: aHR 0.82 (95% CI 0.57, 1.2)
Present in fields the same day applied to crops: aHR 0.82 (95% CI 0.60, 1.1)
Cleaned or helped clean equipment: aHR 0.97 (95% CI 0.60, 1.6)
Episode(s) with unusually high amount on skin/clothing: aHR 1.4 (95% CI 0.7, 2.5)

For ever vs never exposed to fog or spray and born between 1941-1958:
Total breast cancer: aHR 1.1 (95% CI 0.99, 1.3)
ER+/PR+ invasive: aHR 1.1 (95% CI 0.88, 1.3)
Premenopausal: aHR 1.3 (95% CI 0.92, 1.7)
Results Comments
All analyses were also performed restricted to ER+/PR+ tumor and stratifying by menopausal status. In general, no clear difference between these analyses compared to overall analysis of total breast cancer. Authors note because of the age distribution in their study, they were unable to replicate the analysis of Cohn 2007, which examined association between exposure for women born between 1931-1944 (< 14 years of age when DDT came into use and < 20 during peak use) and premenopausal breast cancer. However these authors do consider a critical window of pre-adolescent exposure by assessing residential exposure between 0-14 years of age and the sub-analysis of fog and spray exposure among women born between 1941 and 1958 captures early life (< 18 years old) exposure during the period of peak DDT use. No significant risks observed by residence type (urban, rural, suburban, small town) or for proximity to an orchard, greenhouse, or golf course.
Reviewers Comments
The average length of time subjects lived at a given residence between ages 0-14 was 10.8 years. Therefore, self-reports on residential exposure likely reflect exposures across childhood and adolescence.