Table 2. Hazard ratios for risk of cardiovascular disease (CVD) associated with use of RAS inhibitors during follow-up among patients with type 2 diabetes who did not use RAS inhibitors for at least 2.5 years before enrollment.

Use vs nonuse of RAS inhibitors	Hazard ratio	95% CI	P
Reported in meta-analysis of RCTs	0.92	0.84–1.00
Time-fixed Cox model including immortal time
Univariable analysis	1.16	0.92–1.47	0.213
Adjusted for covariables at enrollmenta	0.62	0.48–0.80	<0.001
Further adjusted for other drug useb	0.66	0.51–0.86	0.002
Time-dependent Cox model including immortal time
Univariable analysis	2.45	1.35–2.08	<0.001
Adjusted for covariables at enrollment	1.43	1.10–1.88	0.009
Adjusted for covariables at enrollment among non-RAS inhibitor users  or time-dependent covariables in usersc	1.30	0.99–1.70	0.062
Further adjusted for other drug useb	1.42	1.08–1.88	0.013
Time-fixed Cox model excluding immortal time among RAS inhibitor users
Univariable analysis	1.85	1.46–2.34	<0.001
Adjusted for covariables at enrollment	0.96	0.74–1.25	0.757
Adjusted for covariables at enrollment in non-RAS inhibitor users  or at the time of use of RAS inhibitors in usersc	0.86	0.66–1.12	0.251
Further adjusted for other drug useb	0.89	0.68–1.17	0.408
Time-fixed Cox model excluding immortal time among users and adding it among nonusers
Univariable analysis	2.57	2.02–3.28	<0.001
Adjusted for covariables at enrollment	1.36	1.04–1.78	0.027
Adjusted for covariables at enrollment in non-RAS inhibitor users  or at the time of use of RAS inhibitors in usersc	1.22	0.93–1.60	0.151
Further adjusted for other drug useb	1.28	0.97–1.68	0.084

Abbreviations: RAS, renin–angiotensin system; RCT, randomized controlled trial.

^aCovariables at enrollment included age, sex, occupation, duration of diabetes, body mass index, smoking status, alcohol use, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglyceride, systolic blood pressure, HbA_1c, natural log-transformed spot urinary albumin: creatinine ratio, and estimated glomerular filtration rate.

^bDrug use included use of insulin, statins, gliclazide, glibenclamide, and thiazolidinediones from enrollment (or initiation of RAS inhibitors in time-fixed models with exclusion of immortal time) to CVD, death, or July 30, 2005, whichever came first.

^cCovariables at the time of initiation of RAS inhibitors during follow-up among RAS inhibitor users were estimated using partial coefficients of age and duration of diabetes that were obtained from all other covariables at enrollment.