Innate left handedness and risk of breast cancer: case-cohort study

Introduction

Among the proposed origins of breast cancer are intrauterine influences, such as exposure to sex hormones.1 Such exposure may also influence cerebral lateralisation, with hand preference being one of its manifestations. We know only of case-control studies on a putative common origin of left handedness and breast cancer, some of which show an association.2 We assessed the association between handedness and incidence of breast cancer in a population based prospective cohort of healthy, middle aged women followed up for 16 years.

Participants, methods, and results

In a breast cancer screening study in Utrecht, the Netherlands, 12 178 women born between 1932 and 1941 and recruited between 1982 and 1985 (participation rate 40%) had baseline questionnaire data recorded about reproductive history, demography, lifestyle, and innate hand preference and had anthropometric measures taken. Linkage with the regional cancer registry provided data on all new cases of invasive breast cancer that occurred until 1 January 2000. Follow-up for adequate information about the person years lived for all 12 178 women would have been costly and time consuming, so we ascertained vital status until 1 January 2000 in a random selection of 1500 women and used their follow-up data to calculate person years lived in the 12 178 women. Exclusion of women with incomplete data left 1426 women; 55 women from the random sample and 371 from outside the random sample had breast cancer (incidence of breast cancer 2.5 per 1000 person years). We used Prentice's method for Cox regression as recommended by Barlow and colleagues3 for analysis with the statistical software SAS (version 8.2, SAS Institute, NC, USA) and a dedicated macro (http://lib.stat.cmu.edu/general/robphreg).

The random sample comprised 165 (11.6%) left handed women. Mean age at baseline was similar for left and non-left handed women (47.4 (range 41.6-53.1) and 47.0 (41.0-53.1) years respectively). These groups did not differ in anthropometry, socioeconomic status, smoking habits, family history of breast cancer, or reproductive history (except for parity—21.8% of left handed v 10.9% of non-left handed women were nulliparous).

The relative risk for breast cancer in left handed women compared with non-left handed women was 1.39 overall (table). The risk was 2.41 when the cancer was premenopausal (diagnosis before reported onset of menopause or, if menopausal data were unavailable, diagnosis at age < 51 years), but there was no excess risk for postmenopausal cancers (1)). Adjustment for risk factors hardly affected the overall association between left handedness and incidence of breast cancer. We found an excess risk for breast cancer in left handed women with a body mass index of border=0> 25 but not in those whose index was > 25 (P interaction 0.07), and in parous but not nulliparous women (P interaction 0.02).

Comment

We found that left handed women are more than twice as likely to develop premenopausal breast cancer as non-left handed women. This risk is compatible with left handedness being a marker of constitutional risk rather than of environmental risk as with postmenopausal breast cancer.

Our findings among premenopausal women may be compatible with a stronger association in women with a normal body mass index, as high body mass index is a particular risk factor for postmenopausal breast cancer. However, our data did not allow for a robust analysis of this issue.

The origin of the association may lie in intrauterine exposure to steroid hormones, as supported by data showing a higher prevalence of left handedness in women with intrauterine exposure to diethylstilbestrol.4 Although the underlying mechanisms remain elusive, our results support the hypothesis that left handedness is related to increased risk of breast cancer.

This article was posted on bmj.com on 26 September 2005: http://bmj.com/cgi/doi/10.1136/bmj.38572.440359.AE