Table 4.

Associations between pre-diagnostic blood glucose level¹ and glioma by age at glioma diagnosis, cohort and gender.

Age at diagnosis	Hazard ratios2 (95% confidence intervals)				P trend 3
	≤50 years	>50–≤60 years	>60–≤70 years	>70 years
AMORIS 4 cohort
Men	1.01 (0.87 to 1.79)	0.79 (0.68 to 0.92)	0.80 (0.69 to 0.93)	0.84 (0.64 to 1.11)	0.08
Cases, n	126	104	126	34
Women	0.94 (0.69 to 1.28)	0.94 (0.73 to 1.19)	0.78 (0.60 to 0.98)	1.07 (0.82 to 1.42)	0.93
Cases, n	56	68	64	26
Total	0.98 (0.85 to 1.13)	0.81 (0.71 to 0.92)	0.81 (0.72 to 0.92)	0.93 (0.76 to 1.15)	0.10
Cases, n	182	172	190	60
Person-years at risk	2,211,354	1,995,967	1,816,781	1,481,712
Me-Can 5 cohort
Men	0.97 (0.61 to 1.54)	1.08 (0.82 to 1.42)	0.72 (0.47 to 1.10)	0.506 (0.29 to 0.85)	0.06
Cases, n	20	50	22	7
Women	0.78 (0.43 to 1.42)	0.77 (0.51 to 1.17)	0.87 (0.61 to 1.24)	0.72 (0.41 to 1.25)	0.94
Cases, n	18	29	33	10
Total	0.87 (0.60 to 1.27)	0.96 (0.76 to 1.22)	0.81 (0.62 to 1.07)	0.60 (0.40 to 0.90)	0.26
Cases, n	38	79	55	17
Person-years at risk	900,761	662,854	663,638	440,112

¹Glucose was transformed to its standardised log (i.e. mean = 0, standard deviation = 1); one unit of this transformed value equals one standard deviation.

²Hazard ratios adjusted for age at lab test, date of lab test, fasting, triglycerides, cholesterol, and gender (totals only); with separate baseline hazards for age at glioma diagnosis and for sub-cohort in the Me-Can data set.

³Linear contrasts (ANOVA, F-test).

⁴Apolipoprotein and MOrtality RISk (AMORIS).

⁵Metabolic syndrome and Cancer project (Me-Can); number of cases differs from Table 1 due to missing values of triglycerides, cholesterol, and glucose.

⁶A hazard ratio of 0.50 among men whose blood was drawn when they were older than 70 years indicates that the glioma rate decreases 50% per unit increase of the standardised natural log of blood glucose.