Skip to main content
. 2016 May 23;6(5):e010889. doi: 10.1136/bmjopen-2015-010889

Table 3.

HRs (95% CI) for prediction of type 2 diabetes with respect to FPG change quartiles; Tehran Lipid and Glucose Study (TLGS) (2001–2010)

Quartile of interest 1 (lowest) 2 3 4 (highest) p trend Harrell's C AIC
FPG change, mmol/L* <−0.29 −0.29–0.06 0.07–0.39 ≥0.40
 Diabetes cases 64 71 57 96
 Person-years 9335 9981 9162 8253
 Incidence (per 1000 person-year) 6.85 7.11 6.22 11.63
Model I 1.00 1.08 (0.77 to 1.52) 0.96 (0.67 to 1.37) 1.71 (1.24 to 2.33) 0.001 0.65 3700.65
Model II 1.00 1.11 (0.79 to 1.56) 0.98 (0.68 to 1.40) 1.65 (1.2 to 2.27) 0.004 0.72 3636.42
Model III 1.00 1.06 (0.76 to 1.5) 0.9 (0.63 to 1.29) 1.43 (1.03 to 1.98) 0.06 0.73 3619.2

Significant values (p<0.05) are shown in bold typeface. Model I: FPG change categories (crude model). Model II: model I+baseline measurements of current smoking, family history of diabetes, education, BMI, WC, TG/HDL-C ratio and its change, TC/HDL-C ratio and its change, lipid drug and hypertension change categories. Model III: model II+2-hPCG change.

We defined the hypertension change category based on the first (1999–2001) and second (2002–2005) examination (yes/yes, yes/no, no/yes and no/no). AIC, Akaike information criterion.

*1st quartile (<−0.29 mmol/L); 2nd quartile (−0.29–0.06 mmol/L); 3rd quartile (0.07–0.39 mmol/L); 4th quartile (≥0.4 mmol/L); multivariate Cox analyses were conducted using age as timescale.

2-hPCG, 2 h post-challenge plasma glucose, mmol/L; BMI, body mass index; FPG, fasting plasma glucose, mmol/L; HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides; WC, waist circumference.