. 2013 Jan 22;15(1):R7. doi: 10.1186/bcr3378

Table 2.

Survival and incidence of new breast cancer related to breast density

Outcome	Model	Label	Per 10% density	Percent density^a		Percent density categories

				Low	High	< 25%	25 to < 50%	≥ 50%
Breast cancer death	All	Number	607	304	303	215	209	183

		Deaths	27	12	15	9	7	11

		HR^b	0.95	1	0.83	1	0.60	0.89

		95% CI	0.79-1.15		0.36-1.92		0.22-1.66	0.34-2.33

		P value	0.61		0.67		0.33	0.80

	All plus interaction (density × radiation)	Number	607	304	303	215	209	183

		Deaths	27	12	15	9	7	11

		HR^b	1.38	1	5.31	1	0.96	2.54

		95% CI	0.99-1.92		0.66-43.1		0.33-2.79	0.65-9.88

		P value	0.06		0.12		0.94	0.18

	With radiation	Number	361	175	170	128	131	102

		Deaths	16	11	5	9	2	5

		HR^b	0.77	1	0.33	1	0.19	0.42

		95% CI	0.60-0.99		0.11-1.06		0.04-0.91	0.12-1.40

		P value	0.04		0.06		0.04	0.16

	Without radiation	Number	246	118	128	87	78	81

		Deaths	11	1	10	0	5	6

		HR	1.46	1	5.57		1	1.34

		95% CI	1.00-2.14		0.65-48.0			0.34-5.34

		P value	0.05		0.12			0.68

Other causes of death	All	Number	607	304	303	215	209	183

		Deaths	100	55	45	44	28	28

		HR^b	1.08		1.37		1.06	1.32

		95% CI	0.98-1.20		0.88-2.13		0.64-1.75	0.77-2.25

		P value	0.13		0.16		0.83	0.32

Second breast cancer	All plus interaction (density × radiation)	Number	607	304	303	215	209	183

		Cases	71	32	39	22	23	26

		HR^b	1.13	1	1.72	1	1.31	2.11

		95% CI	0.96-1.33		0.88-2.55		0.69-2.50	0.98-4.50

		P value	0.14		0.15		0.41	0.06

^aLow density = < 35%; high density ≥ 35%. ^bHazard ratios (HRs) and 95% confidence intervals (CIs) for the effect of density estimated by Cox regression with adjustment for age at diagnosis, ethnicity, overweight/obesity, stage of disease at diagnosis (in situ, localized, or advanced), and radiation treatment when applicable.