Table 3. Characteristics of a cohort of postmenopausal Swedish women with a breast cancer diagnosis between 1970 and 2003, and SIR, with 95% CI, of gastric non-cardia adenocarcinoma.

	Gastric non-cardia adenocarcinoma
	Unexposed (1970–1987)	Exposed (1988–2003)	Total cohort (1970–2003)
No. of breast cancer cases	53 643	65 505	119 148
No. of person-years at risk	528 789	363 765	892 555
Mean years of follow-up	9.85	5.55	7.49
Mean age at entry (range)	69 (50–101)	68 (50–104)	68 (50–104)
No. of observed gastric non-cardia adenocarcinoma cases	257	84	341
No. of expected gastric non-cardia adeno-carcinoma cases	174.52	65.88	240.39
SIR (95% CI)	1.47 (1.30–1.66)	1.27 (1.03–1.57)	1.41 (1.27–1.57)
Latency interval after breast cancer diagnosis in years	Number of observed cases of gastric non-cardia adenocarcinoma SIR (95% CI)
1–4	105	46	151
	1.7 (1.4–2.06)	1.29 (0.96–1.72)	1.55 (1.32–1.82)
5–9	71	24	95
	1.4 (1.11–1.77)	1.06 (0.71–1.58)	1.3 (1.06–1.58)
10–14	39	14	53
	1.19 (0.87–1.63)	1.86 (1.1–3.14)	1.32 (1.008–1.72)
>15	42	0	42
	1.41 (1.04–1.91)		1.4 (1.04–1.9)

CI=confidence interval; SIR=standardised incidence ratio.

Patients diagnosed between 1970 and 1987 were considered unexposed to tamoxifen treatment, whereas those diagnosed between 1988 and 2003 were considered exposed. First year of follow-up was excluded. The earliest data available for the gastric cancer diagnosis were from 1961, but distinction between cardia and non-cardia diagnoses was only available from 1970.