Skip to main content
. 2015 Dec 16;7(2):2000–2008. doi: 10.18632/oncotarget.6626

Table 2. Incidences of oral cancer by metformin exposure in the original sample and the hazard ratios comparing metformin exposed to unexposed patients in the original sample and the matched sample.

Original sample Matched sample
Metformin use n/N Person-years Incidence rate(per 100,000 person-years) PS-adjusted model IPTW model n/N PS-adjusted model IPTW model
HR (95% CI) P HR (95% CI) P HR (95% CI) P HR (95% CI) P
Never users 119/16263 72760.7 163.6 1.000 1.000 119/16263 1.000 1.000
Ever users 1273/288198 1373325.0 92.7 0.584 (0.483-0.707) <0.0001 0.562 (0.465-0.678) <0.0001 77/16263 0.592 (0.444-0.789) 0.0003 0.602 (0.452-0.802) 0.0005
Tertiles of cumulative duration of metformin therapy (months)
Never users 119/16263 72760.7 163.6 1.000 1.000 119/16263 1.000 1.000
<21.5 715/95463 344951.9 207.3 1.403 (1.152-1.708) 0.0008 1.244 (1.024-1.511) 0.0281 35/5287 1.176 (0.806-1.717) 0.4000 1.127 (0.772-1.645) 0.5359
21.5-45.9 418/94666 471050.1 88.7 0.557 (0.453-0.684) <0.0001 0.526 (0.429-0.645) <0.0001 31/5307 0.693 (0.466-1.029) 0.0689 0.703 (0.473-1.043) 0.0802
>45.9 140/98069 557322.9 25.1 0.152 (0.119-0.194) <0.0001 0.138 (0.108-0.176) <0.0001 11/5669 0.197 (0.106-0.366) <0.0001 0.205 (0.110-0.380) <0.0001

n: incident cases of oral cancer, N: cases followed, HR: hazard ratio, CI: confidence intervals

PS: propensity score created from variables in Table 1

IPTW: Cox regression model incorporated with the inverse probability of treatment weighting (IPTW) using PS

The models shown in the table were created with covariates defined at the start of follow-up. Additional models created with covariates defined at censor were not much different from the respective models shown in the table.