. 2014 Jun 4;9(6):e98508. doi: 10.1371/journal.pone.0098508

Table 3. Joint associations of weight status and cardiorespiratory fitness with incident type 2 diabetes mellitus.

	BMI <25 kg/m²			BMI ≥25 kg/m²
	Fit	Unfit		Fit	Unfit
	(upper two-thirds)	(lower one-third)	P value^*	(upper two-thirds)	(lower one-third)	P value
Estimated VO_2max
median (IQR), ml/min/kg	40.7 (37.8–44.5)	32.8 (30.8–34.0)		38.4 (36.4–41.4)	31.5 (29.5–33.3)
No. of subjects	2,096	785		245	397
No. of cases	89	43		35	32
Person-years	12,785	4,646		1,364	2,393
Age-, sex-, and baseline year-adjusted	1.00 (reference)	1.30 (0.90, 1.87)	0.16	3.26 (2.20, 4.81)	1.93 (1.29, 2.89)	0.033
Multivariable-adjusted model 1^†	1.00 (reference)	1.32 (0.91, 1.90)	0.15	2.94 (1.98, 4.37)	1.83 (1.21, 2.76)	0.053

Abbreviations: BMI, body mass index; IQR, interquartile range.

*P value between fit and unfit in each category of BMI.

^†

Adjusted for age, baseline year (2003, 2004, or 2005), smoking status (non-smoker or current smoker consuming 1 to 10, 11 to 20, or ≥21 cigarettes per day), alcohol consumption (non-drinker or drinker consuming <1 unit, 1 to <2 units, or ≥2 units per day, 1 unit of alcohol corresponds to 1 go of Japanese sake which approximately contains 23 g of ethanol), sleep duration (<7, 7 to <9, or ≥9 hours per day), hypertension (presence or absence), and family history of diabetes (presence or absence).