Reply to WB Grant

Dear Editor:

We agree that total cancers are heterogeneous conditions and analyzing them as a group may obscure possible associations of individual cancers with 25-hydroxyvitamin D [25(OH)D] concentrations (1). However, space constraints did not allow us to examine the associations with individual cancers in detail and to discuss this issue adequately. We note Grant’s comments about the duration of follow-up. Table 1 below shows the HRs per increase of 20 nmol/L in serum 25(OH)D concentration for total cancer mortality and incidence according to follow-up time.

TABLE 1.

HRs (95% CIs) for cancer mortality and cancer incidence by follow-up time per 20-nmol/L increase in serum 25(OH)D in 14,614 men and women in EPIC Norfolk, 1997–2012¹

	Deaths			Cancer incidence
Follow-up time	No. of events	HR (95% CI)	P	No. of events	HR (95% CI)	P
2 y
Age, sex, and month adjusted	86	0.80 (0.65, 0.99)	0.04	387	0.93 (0.84, 1.02)	0.11
Multivariable adjusted2		0.84 (0.65, 1.08)	0.016		0.92 (0.83, 1.03)	0.16
4 y
Age, sex, and month adjusted	234	0.82 (0.72, 0.93)	0.002	788	0.96 (0.90, 1.02)	0.18
Multivariable adjusted2		0.82 (0.71, 0.96)	0.01		0.95 (0.88, 1.02)	0.14
6 y
Age, sex, and month adjusted	364	0.91 (0.82, 1.00)	0.06	1174	0.98 (0.93, 1.03)	0.35
Multivariable adjusted2		0.93 (0.83, 1.04)	0.18		0.98 (0.93, 1.05)	0.61
8 y
Age, sex, and month adjusted	560	0.93 (0.86, 1.00)	0.06	1594	0.97 (0.93, 1.03)	0.19
Multivariable adjusted2		0.94 (0.86, 1.03)	0.18		0.98 (0.93, 1.03)	0.42
10 y
Age, sex, and month adjusted	744	0.93 (0.86, 0.99)	0.03	1974	0.99 (0.95, 1.03)	0.48
Multivariable adjusted2		0.93 (0.86, 1.01)	0.08		0.99 (0.95, 1.04)	0.71
Final
Age, sex, and month adjusted	1086	0.93 (0.88, 0.99)	0.017	3121	1.01 (0.97, 1.04)	0.77
Multivariable adjusted2		0.94 (0.89, 1.00)	0.07		1.02 (0.99, 1.06)	0.21

¹EPIC, European Prospective Investigation into Cancer and Nutrition; 25(OH)D, 25-hydroxyvitamin D.

²Adjusted for age, sex, month, BMI, cigarette smoking, alcohol intake, plasma vitamin C, physical activity, diabetes, history of cardiovascular disease, history of cancer, social class, and educational level.

There is indeed some support for somewhat lower HRs for mortality due to cancer with shorter follow-up time, such that the lowest significant HRs were observed for the 2- and 4-y follow-up, as hypothesized by Grant, and a trend at the 10- and 13-y follow-up. The cancer incidence data showed no statistical significance or trends with follow-up time.

We agree that cancer is heterogeneous, and we will indeed be examining in more detail the associations between the incidence of individual cancers with 25(OH)D.