Skip to main content
. Author manuscript; available in PMC: 2014 Jan 29.
Published in final edited form as: Int J Cancer. 2012 Feb 10;131(7):1667–1677. doi: 10.1002/ijc.27394

Table 3.

Multivariable-adjusted cancer mortality rates and hazard ratios of cancer mortality and major site-specific cancer mortality by glycated hemoglobin value in non-diabetic participants, and by diabetes in the Atherosclerosis Risk in Communities Study (ARIC), 1990–2006.


Glycated Hemoglobin, %
in Non-Diabetic Participants
Women Total <5.0 5.0–5.6 ≥5.7 Diabetes*
All Cancer
No. of cases 372 33 180 110 49

Person yrs of follow-up 101,246 7,502 64,548 21,349 7,847

Mortality rate 391.5 501.9 297.4 547.2 601.9

HR (95% CI) 1 1.77
(1.22,2.58)
1.00
(Ref)
1.81
(1.41,2.32)
2.25
(1.63,3.13)

HR (95% CI) 2 1.82
(1.25,2.64)
1.00
(Ref)
1.58
(1.23,2.05)
1.96
(1.40,2.76)

Lung Cancer
No. of cases 105 7 55 31 12

Person yrs of follow-up 101,246 7,502 64,548 21,349 7,847

Mortality rate 110.9 119.0 90.7 148.4 146.6

HR (95% CI) 1 1.29
(0.59,2.85)
1.00
(Ref)
1.66
(1.04,2.63)
1.77
(0.94,3.40)

HR (95% CI) 2 1.18
(0.53,2.60)
1.00
(Ref)
1.35
(0.83,2.21)
1.73
(0.88,3.40)

Post-menopausal Breast
Cancer
No. of cases 40 4 19 8 9

Person yrs of follow-up 86,236 5,604 53,988 19,423 7,221

Mortality rate 48.1 63.4 38.2 38.6 85.7

HR (95% CI) 1 2.12
(0.72,6.27)
1.00
(Ref)
1.03
(0.44,2.44)
3.22
(1.40,7.43)

HR (95% CI) 2 2.28
(0.77,6.75)
1.00
(Ref)
0.88
(0.37,2.13)
2.34
(0.97,5.62)

Colorectal Cancer
No. of cases 29 5 14 6 4

Person yrs of follow-up 101,246 7,502 64,548 21,349 7,847

Mortality rate 30.6 65.2 24.6 36.3 53.2

HR (95% CI) 1 3.23
(1.15,9.04)
1.00
(Ref)
1.31
(0.48,3.57)
2.49
(0.79,7.90)

HR (95% CI) 2 3.30
(1.17,9.24)
1.00
(Ref)
1.18
(0.43,3.27)
2.25
(0.68,7.46)

Men
All Cancer
No. of cases 515 32 300 143 40

Person yrs of follow-up 80,271 6,513 49,472 18,121 6,165

Mortality rate 641.3 602.7 649.9 728.0 595.3

HR (95% CI)1 0.88
(0.61,1.27)
1.00
(Ref)
1.16
(0.94,1.42)
0.94
(0.67,1.31)

HR (95% CI) 2 0.97
(0.67,1.40)
1.00
(Ref)
1.08
(0.87,1.33)
0.92
(0.66,1.30)

All Cancer minus Prostate
Cancer
No. of cases 471 27 280 125 39

Person yrs of follow-up 80,271 6,513 49,472 18,121 6,165

Mortality rate 584.4 515.2 601.8 647.4 583.9

HR (95% CI) 1 0.79
(0.53,1.18)
1.00
(Ref)
1.11
(0.89,1.37)
1.00
(0.71,1.40)

HR (95% CI) 2 0.89
(0.60,1.32)
1.00
(Ref)
1.01
(0.80,1.26)
0.92
(0.65,1.30)

Lung Cancer
No. of cases 182 7 104 60 11

Person yrs of follow-up 80,271 6,513 49,472 18,121 6,165

Mortality rate 225.7 132.8 221.1 315.1 172.3

HR (95% CI) 1 0.56
(0.26,1.20)
1.00
(Ref)
1.41
(1.02,1.96)
0.74
(0.39,1.39)

HR (95% CI) 2 0.65
(0.30,1.40)
1.00
(Ref)
1.31
(0.94,1.84)
0.81
(0.43,1.52)

Colorectal Cancer
No. of cases 46 5 25 10 6

Person yrs of follow-up 80,271 6,513 49,472 18,121 6,165

Mortality rate 57.2 89.0 60.4 51.7 91.6

HR (95% CI) 1 1.66
(0.63,4.34)
1.00
(Ref)
0.97
(0.46,2.06)
1.69
(0.68,4.20)

HR (95% CI) 2 1.73
(0.66,4.53)
1.00
(Ref)
0.86
(0.40,1.85)
1.51
(0.59,3.81)

Prostate Cancer
No. of cases 44 5 20 18 1

Person yrs of follow-up 80,271 6,513 49,472 18,121 6,165

Mortality rate 56.9 87.5 48.0 80.6 11.4

HR (95% CI) 1 2.15
(0.80,5.74)
1.00
(Ref)
1.81
(0.93,3.51)
0.29
(0.04,2.18)

HR (95% CI) 2 2.06
(0.77,5.52)
1.00
(Ref)
1.65
(0.84,3.25)
0.25
(0.03,1.91)

Abbreviations: No., number; yrs, years; HR, Hazard Ratio; CI, Confidence Interval

*

A participant was classified as having diabetes if they self-reported a diagnosis of diabetes and/or were taking medication to treat diabetes at the first or second visits.

The rate per 100,000 women/men is given, standardized to the age and race distribution in the analytic cohort.

The Cox proportional hazards models were adjusted for:

(1)

age (continuous) at the second visit, and race/ethnicity (white, black);

(2)

covariates in (1) and ARIC study site, education level (less than high school, high school/equivalent, college or above), cigarette smoking status (never, former, current), cigarette-years smoked (continuous), body mass index (continuous), and waist circumference (continuous).

The Cox proportional hazards models for women were also adjusted for post-menopausal hormone use (no, yes).