Table 1.

Association between handedness and incidence of breast cancer in study participants followed up at 16 years

			Hazard ratio (95% confidence interval)
Innate handedness	Cases	Estimated person years*	Crude	Adjusted†
Total
Non-left handed	361	153 422	1.00	1.00
Left handed	65	19 119	1.39 (1.09 to 1.81)	1.32 (0.99 to 1.76)
Premenopausal breast cancer
Non-left handed	57	32 113	1.00	1.00
Left handed	15	3329	2.41 (1.35 to 4.30)	2.20 (1.15 to 4.20)
Postmenopausal breast cancer
Non-left handed	257	127 426	1.00	1.00
Left handed	39	17 665	1.12 (0.80 to 1.57)	1.05 (0.75 to 1.48)
Body mass index ≤25
Non-left handed	217	95 964	1.00	1.00
Left handed	45	11 332	1.62 (1.17 to 2.24)	1.59 (1.15 to 2.20)
Body mass index >25
Non-left handed	144	57 458	1.00	1.00
Left handed	20	7787	1.05 (0.67 to 1.66)	1.04 (0.65 to 1.64)
Nulliparous
Non-left handed	61	16 486	1.00	1.00
Left handed	9	3759	0.68 (0.35 to 1.32)	0.70 (0.36 to 1.35)
Parous
Non-left handed	300	136 936	1.00	1.00
Left handed	56	15 360	1.58 (1.19 to 2.11)	1.59 (1.18 to 2.13)

The non-left handed group was the reference group.

74 random sample participants and 32 cases with missing data on covariates or prevalent cases were excluded from these analyses.

Fifty eight breast cancer cases were not analysed as premenopausal or postmenopausal breast cancer because menopausal information was not available and the age at diagnosis was 51-55 years.

^*The number of person years (lived in the total cohort) is extrapolated from the random sample.

^†Adjusted for socioeconomic status, age, height, body mass index (except in body mass index specific analysis), smoking status, history of breast cancer in mother or sister, age at menarche, parity status (except in parity specific analysis), all at baseline; adjusted for age at last known menstruation and menopausal status during follow-up (except for outcome of premenopausal or postmenopausal breast cancer).