· Web vieweTable 13 Subgroup analysis of Risk of Ischemic Stroke Among Patients Using ARBs and ACEIs eFigure 1 Study flow diagram eFigure 2. Distribution of propensity scores for

Supplementary Online Content

eMethods

eTable 1. Propensity Score Model Results of Probability of Using Angiotensin II

receptor blockers

eTable 2. Baseline Characteristics of Diabetes Patients

eTable 3. Drug persistence of ACEI and ARB during follow-up period

eTable 4. Prescriptions of ACEI and ARB during follow-up period

eTable 5. Incidence and Risk of MACEs and adverse effects Among Patients Using

ARBs and ACEIs before Propensity Score Matching.


ARBs and ACEIs With Different Enrollment Period Before Propensity Score

Matching (As Treated Analysis)


ARBs and ACEIs Before Propensity Score Matching, by 30-, 60-, 90-Day Medication

Gap (As Treated Analysis)

eTable 8. Risk of MACEs and adverse effects Among Patients Using ARBs and

ACEIs, stratified by Hospital Levels.

eTable 9. Sensitivity Test for Risk of MACEs and adverse effects Among Patients

Using ARBs and ACEIs Before Propensity Score Matching with Adjusting Hospital

Levels

eTable 10. Effect of Continuing Use of ARBs on MACEs and adverse effects

eTable 11. Subgroup analysis of Risk of All-cause Mortality Among Patients Using

ARBs and ACEIs

eTable 12. Subgroup analysis of Risk of Myocardial Infarction Among Patients Using

ARBs and ACEIs

eTable 13 Subgroup analysis of Risk of Ischemic Stroke Among Patients Using ARBs

and ACEIs

eFigure 1 Study flow diagram

eFigure 2. Distribution of propensity scores for using ARB before and after matching

eFigure 3. Overlap in distribution of propensity scores for using ARB among ACEI

and ARB cohort before and after matching

eFigure 4. Cumulative Incidence of Adverse Outcomes in ACEI and ARB Users with

Type 2 Diabetes

(A) hospitalization for acute kidney injury, (B) hospitalization for hyperkalemia.

All patients were ACEI/ARB-naïve and newly diagnosed with type 2 diabetes.

ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor

blocker, MACE, major adverse cardiac event

eFigure 5. Forest Plots of ACEI and ARB Effects

Subgroup analysis measured the risks of (A) myocardial infarction, (B) ischemic

stroke, and (C) all-cause mortality.


blocker; RAAS, Renin-angiotensin-aldosterone system

eMethods

Statistical analysis

Marginal structural models

Marginal structural models aim to control for the effects of time-dependent

confounders by prior treatment by estimating each subject’s probability of having

their observed treatment yielding inverse-probability-of-treatment weights (IPTW).

The outcome of MACE associated with time-dependent ACEI and ARB treatment is

accessed in a regression model that is weighted using the IPTWs. The inverse

probability weights were estimated by fitting, separately for each arm. A participant

contributed as many observational periods at baseline and for each 3-month interval in

the logistic regression models to calculate IPTWs during the follow-up, i.e., from

baseline to the occurrence of using opposite drugs, primary outcomes of interest,

death, or end of study, whichever occurred first. Patients were censored when he/she

used opposite drugs or death when outcomes of interest was not death. The model

included not only all static baseline variables (age, sex, year of index date,

urbanization, monthly income, hospital level of first prescription of ACEI/ARB,

outpatient visits of metabolism &endocrinology, in the past one year, CCI, aDCSI

scores, medication use and comorbidities), but also time-varying covariates (day since

index date, CCI, aDCSI scores and comorbidities). For each subject, the weight of

probability of remaining uncensored was also calculated. Finally, we fitted a pooled

logistic regression including all covariates in eTable 1 except month of index date and

weighting by IPTW combined with weight of probability of remaining uncensored to

estimate the causal hazard ratio for the effect of treatment on our outcome.

Propensity score analysis

The propensity score method was used to estimate the differences in the clinical

characteristics of newly diagnosed patients initiating ARBs or ACEIs (as reference) in

the logistic regression models. eTable 1 shows predefined baseline covariates and

odds ratios associated with each covariate, indicating the likelihood for using ARBs.

We used the standardized difference to compare the 2 groups after propensity score;

all values of standardized difference in each covariate in Table 1 were statistically

insignificant, representing balancing between ACEI and ARB users adequately after

controlling for the propensity score. A propensity score-matched cohort was generated

by the process that each ARB user was matched to one ACEI user according to

propensity scores closest in value, based on nearest-neighbor matching without

replacement and using calipers of width equal to 0.1 standard deviation of the logit of

the propensity score. The distribution of propensity score and overlap between cohorts

after propensity matching was nearly similar and complete and shown in eFigure 2 &

3.

eTable 1. Propensity Score Model Results of Probability of Using Angiotensin II receptor blockers.

95% CIParameter Estimate Odds Ratios Lower Upper P valueAge, per year 0.0038 1.004 1.002 1.006 <.0001Year of Index Date2000 12001 0.0522 1.054 0.941 1.179 0.36412002 0.5682 1.765 1.584 1.967 <.00012003 1.0438 2.84 2.552 3.161 <.00012004 1.256 3.511 3.161 3.901 <.00012005 1.2088 3.349 3.004 3.735 <.00012006 1.4381 4.213 3.781 4.694 <.00012007 1.5996 4.951 4.444 5.517 <.00012008 1.8005 6.052 5.429 6.748 <.00012009 1.8829 6.572 5.896 7.327 <.00012010 2.2174 9.184 8.218 10.262 <.00012011 2.4854 12.006 10.200 14.132 <.0001

Month of Index DateJanuary 1February 0.1152 1.122 1.024 1.229 0.0132March 0.0636 1.066 0.974 1.166 0.1679April 0.1121 1.119 1.017 1.23 0.021May 0.1352 1.145 1.041 1.258 0.0051June 0.0648 1.067 0.973 1.17 0.1697July 0.1964 1.217 1.109 1.336 <.0001August 0.1559 1.169 1.064 1.283 0.0011September 0.1845 1.203 1.094 1.322 0.0001October 0.2083 1.232 1.123 1.351 <.0001November 0.2168 1.242 1.133 1.362 <.0001December 0.1866 1.205 1.1 1.321 <.0001

Male -0.0456 0.955 0.917 0.995 0.0286Monthly income, NT dollars

Dependent 10–19,100 -0.0671 0.935 0.882 0.991 0.023619,100–42,000 -0.0688 0.934 0.889 0.981 0.0061> 42,000 0.1338 1.143 1.061 1.232 0.0005

Urbanization level*

1 12 -0.0189 0.981 0.943 1.022 0.35643 0.1381 1.148 1.046 1.26 0.00374 -0.1316 0.877 0.711 1.081 0.2172

Hospital level of first prescription of ARB or ACEI

Level I (Medical center) 1.7211 5.591 5.257 5.946 <.0001Level II 1.4181 4.129 3.903 4.368 <.0001Level III 1.3517 3.864 3.662 4.077 <.0001Level IV (Local medical clinics) 1

Outpatient Visits of Metabolism &Endocrinology, in the past one year 05 visits 1 610 visits 0.1772 1.194 1.127 1.265 <.0001 1115 visits 0.3 1.35 1.144 1.592 0.0004 >15 visits 0.3988 1.49 1.042 2.131 0.029Charlson comorbidity index score

1 12 -0.00554 0.994 0.942 1.05 0.8423 -0.0898 0.914 0.856 0.977 0.008 4 -0.0819 0.921 0.84 1.01 0.0808

Adapted Diabetes Complications Severity Index scorec 0.00379 1.004 0.98 1.028 0.7542

Anti-diabetic drugs Acarbose inhibits enzymes 0.0257 1.026 0.921 1.143 0.6394 Sulfonylurea -0.0381 0.963 0.918 1.009 0.1123 Insulin -0.1601 0.852 0.715 1.015 0.0724 Metformin -0.2454 0.782 0.748 0.818 <.0001 Thiazolidinediones 0.5254 1.691 1.52 1.882 <.0001 Glinide 0.011 1.011 0.907 1.127 0.8422 Dipeptidyl peptidase-4 inhibitor 0.441 1.554 1.165 2.074 0.0027Anti-hypertensive drug Alpha-blocker 0.0579 1.06 0.918 1.223 0.4294 Beta-blocker 0.00407 1.004 0.948 1.063 0.8894 Calcium channel blocker 0.0379 1.039 0.992 1.087 0.1038 Diuretics -0.1591 0.853 0.799 0.91 <.0001

Other anti-hypertensive drug -0.2145 0.807 0.672 0.969 0.0218Other concomitant medications

Antiplatelet agent d 0.0233 1.024 0.961 1.09 0.4711Steroid -0.0847 0.919 0.838 1.007 0.0698Nitrate -0.0957 0.909 0.802 1.03 0.1332NSAID -0.0594 0.942 0.892 0.996 0.0348PPI 0.1358 1.145 0.989 1.327 0.0699

Statin 0.0189 1.019 0.961 1.08 0.5267 SSRI -0.284 0.753 0.59 0.96 0.0223Comorbidities Coronary artery disease 0.034 1.035 0.976 1.096 0.2497 Receiving PCI 0.0675 1.07 0.832 1.376 0.5997 Receiving CABG 0.2975 1.346 0.834 2.175 0.224 Hypertension 0.3753 1.455 1.368 1.548 <.0001 Heart failure -0.0949 0.909 0.821 1.008 0.0693 Peripheral vascular disease -0.0742 0.928 0.807 1.068 0.2995 Peptic ulcer disease -0.0291 0.971 0.921 1.024 0.2807 Liver disease -0.0155 0.985 0.934 1.038 0.5682 Chronic kidney disease 0.1268 1.135 1.041 1.238 0.004 Atrial fibrillation -0.2143 0.807 0.666 0.979 0.0292 Dyslipidemia 0.0675 1.07 1.027 1.115 0.0014 Valvular heart disease 0.1239 1.132 1.024 1.251 0.0149 Cancer 0.084 1.088 0.989 1.197 0.0841 Autoimmune disease -0.0684 0.934 0.807 1.08 0.3575a Urbanization levels in Taiwan are divided into four strata according to the Taiwan National Health Research Institute publications. Level 1 designates the most urbanized areas, and level 4 designates the least urbanized areas.b Charlson Comorbidity Index score is used to determine overall systemic health. With each increased level of CCI score, there are stepwise increases in the cumulative mortality.

CAdapted Diabetes Complications Severity Index is a 13-point scale from 7 complication categories: retinopathy, nephropathy, neuropathy, cerebrovascular, cardiovascular, peripheral vascular disease, and metabolic, ranging from each complication. Each complication produced a numeric score ranging from 0 to 2 (0 = no abnormality, 1 = some abnormality, 2 = severe abnormality)d Including aspirin, clopidogrel, ticlopidine and cilostazol.Abbreviations: NT dollars, new Taiwan dollars; NSAID, Non-steroidal anti-inflammatory drugs; PPI, proton pump inhibitors; SSRI, Selective serotonin re-uptake inhibitors; PCI,

Percutaneous coronary intervention; CABG, Coronary artery bypass grafting.

eTable 2. Baseline Characteristics of Diabetes PatientsBefore Propensity Score-Matching

Characteristics ARB cohort ACEI cohortAdjustedP Value†

Patient (no.) 21,436 30,777Mean age (SD), year 55.2 (12.4) 55.4 (12.3) 0.988Male 12,528 (58.4) 17,498 (56.8) 0.933Year of index date 2000 685 (3.2) 3,020 (9.8) 0.147 2001 971 (4.5) 3,970 (12.9) 0.146 2002 1,389 (6.5) 3,502 (11.4) 0.741 2003 1,789 (8.3) 2,984 (9.7) 0.997 2004 2,243 (10.5) 3,112 (10.1) 0.991 2005 1,721 (8.0) 2,758 (9.0) 0.993 2006 2,014 (9.4) 2,604 (8.5) 0.966 2007 2,238 (10.4) 2,323 (7.5) 0.808 2008 2,388 (11.1) 2,114 (6.9) 0.624 2009 2,549 (11.9) 2,165 (7.0) 0.559 2010 2,698 (12.6) 1,799 (5.8) 0.224 2011 751 (3.5) 426 (1.4) 0.352Monthly income, NT dollars Dependent 5,589 (26) 8,421 (27.3) 0.985 <19,100 4,510 (21) 6,059 (19.6) 0.972 19,10041,999 8,850 (41.2) 13,650 (44.3) 0.965 42,000 2,487 (11.6) 2,647 (8.6) 0.817Urbanizationa

Level 1 8,693 (40.6) 12,369 (40.2) 0.995Level 2 11,532 (53.8) 16,724 (54.3) 0.993Level 3 1,037 (4.8) 1,396 (4.5) 0.987Level 4 (rural area) 174 (0.8) 288 (0.9) 0.999

Hospital level of first prescription of ARB or ACEILevel I (Medical center) 5,001 (23.3) 4,174 (13.6) 0.405Level II 6,081 (28.4) 6,307 (20.5) 0.711Level III 6,967 (32.5) 7,562 (24.6) 0.762Level IV (Local medical clinics) 3,387 (15.8) 12,734 (41.4) 0.016

Outpatient Visits of Metabolism &Endocrinology, in the past one year

05 visits 17,895 (83.4) 27,050 (87.8) 0.780 610 visits 3,155 (14.7) 3,257 (10.5) 0.770 1115 visits 322 (1.5) 376 (1.2) 0.963 >15 visits 64 (0.2) 94 (0.3) 0.999Charlson Comorbidity Index Scoreb

1 8,066 (37.6) 11,952 (38.8) 0.9852 5,421 (25.2) 7,823 (25.4) 0.9983 3,933 (18.3) 5,607 (18.2) 0.998 4 4,016 (18.7) 5,395 (17.5) 0.973

Adapted Diabetes Complication Severity Index scoreC 0.86 (1.21) 0.79 (1.15)

0 11,316 (52.8) 16,847 (54.7) 0.598 1 5,527 (25.8) 7,893 (25.6) 0.656 2 2,375 (11.1) 3,271 (10.6) 0.159 3 2,218 (10.3) 2,766 (9.0) 0.244Drugs for diabetes Acarbose inhibits enzymes 835 (3.8) 786 (2.5) 0.815 Sulfonylurea 5,274 (24.6) 8,053 (26.1) 0.983 Insulin 299 (1.3) 324 (1.0) 0.943 Metformin 8,450 (39.4) 11,323 (36.7) 0.959 Thiazolidinediones 1,062 (4.9) 657 (2.1) 0.361 Glinide 819 (3.8) 766 (2.4) 0.812 Dipeptidyl peptidase-4 inhibitor 215 (1) 68 (0.2) 0.176Anti-hypertensive drug Alpha-blocker 400 (1.8) 547 (1.7) 0.994 Beta-blocker 3,028 (14.1) 4,181 (13.5) 0.988 Calcium channel blocker 6,056 (28.2) 7,535 (24.4) 0.912 Diuretics 2,191 (10.2) 3,476 (11.2) 0.991 Other anti-hypertensive drug 197 (0.9) 495 (1.6) 0.899Other concomitant medications

Antiplatelet agent ‡ 2,840 (13.2) 3,554 (11.5) 0.941Steroid 954 (4.4) 1,639 (5.3) 0.997Nitrate 588 (2.7) 863 (2.8) 0.998NSAID 2,956 (13.7) 5,061 (16.4) 0.997

PPI 476 (2.2) 413 (1.3) 0.811 Statin 3,624 (16.9) 3,787 (12.3) 0.769 SSRI 129 (0.6) 190 (0.6) 0.999Comorbidities

Coronary artery disease 4,315 (20.1) 5,974 (19.4) 0.986 Hypertension 19,152 (89.3) 26,962 (87.6) 0.995 Heart failure 1,161 (5.4) 1,694 (5.5) 0.998 Peripheral vascular disease 444 (2) 652 (2.1) 0.998 Peptic ulcer disease 5,717 (26.6) 7,696 (25) 0.969 Chronic kidney disease 1,845 (8.6) 2,442 (7.9) 0.977 Liver disease 4,951 (23) 7,301 (23.7) 0.992 Atrial fibrillation 218 (1) 333 (1) 0.998 Dyslipidemia 10,296 (48) 13,189 (42.8) 0.914 Valvular heart disease 960 (4.4) 1,082 (3.5) 0.920 Cancer 1,272 (5.9) 1,553 (5) 0.950 Autoimmune disease 394 (1.8) 526 (1.7) 0.991

* All data were descripted as number (%), except mean age and propensity score.aUrbanization levels in Taiwan are divided into four strata according to the Taiwan National Health Research Institute publications. Level 1 designates the most urbanized areas, and level 4 designates the least urbanized areas.bCharlson Comorbidity Index (CCI) score is used to determine overall systemic health. With each increased level of CCI score, there are stepwise increases in the cumulative mortality.

C Adapted Diabetes Complications Severity Index is a 13-point scale from 7 complication categories: retinopathy, nephropathy, neuropathy, cerebrovascular, cardiovascular, peripheral vascular disease, and metabolic, ranging from each complication. Each complication produced a numeric score ranging from 0 to 2 (0 = no abnormality, 1 = some abnormality, 2 = severe abnormality).Abbreviations: SD, standard deviation; StD, Standardized difference; NT dollars, new Taiwan dollars; ACEI, angiotensin-converting-enzyme inhibitors; ARB, Angiotensin II receptor blockers; NSAIDs, Non-steroidal anti-inflammatory drugs; SSRI, Selective serotonin re-uptake inhibitors.† Adjuted for propensity score.‡ Including aspirin, clopidogrel, ticlopidine and cilostazol.

eTable 3. Drug persistence of ACEI and ARB during follow-up periodContinuing use, patient No. (%)

Medication gap All cohortOne year

periodTwo year

periodThree year

periodFour year

periodFive year

periodTen year

period90 days

ACEI cohort 30,777 17,879 (58.1) 12,209 (39.7) 9,250 (30.1) 7,269 (23.6) 5,992 (19.5) 1,839 (6.0%)ARB cohort 21,436 15,481 (72.2) 11,868 (55.4) 9,763 (45.5) 8,544 (39.9) 7,650 (35.7) 2,783 (13.0%)

60 daysACEI cohort 30,777 16,161 (52.5) 10,252 (33.3) 7,397 (24.0) 5,624 (18.3) 4,507 (14.6) 1,451 (4.7%)ARB cohort 21,436 13,412 (62.6) 8,853 (41.3) 6,572 (30.7) 5,385 (25.1) 4,576 (21.3) 1,690 (7.9%)

30 daysACEI cohort 30,777 13,859 (45.0) 8,186 (26.6) 5,805 (18.9) 4,410 (14.3) 3,525 (11.5) 1,129 (3.7%)ARB cohort 21,436 10,168 (47.4) 5,816 (27.1) 4,122 (19.2) 3,262 (15.2) 2,676 (12.5) 950 (4.4%)

Abbreviations: ACEI, angiotensin-converting-enzyme inhibitors; ARB, Angiotensin II receptor blockers.

eTable 4. Prescriptions of ACEI and ARB during follow-up period

All cohortAdd-on opposite

drug†

Continuous treatment

Discontinuing treatment

Switching to the opposite drug

Medication gap Patient No. Patient No. (%) Patient No. (%) Patient No. (%) Patient No. (%)90 days

ACEI cohort 30,777 4,699 (15.3%) 1,839 (6.0%) 27,115 (88.1%) 1,823 (5.9%)ARB cohort 21,436 3,155 (14.7%) 2,783 (13.0%) 18,435 (86.0%) 218 (1.0%)

60 daysACEI cohort 30,777 4,227 (13.7%) 1,451 (4.7%) 27,642 (89.8%) 1,684 (5.5%)ARB cohort 21,436 2,777 (13.0%) 1,690 (7.9%) 19,530 (91.1%) 216 (1.0%)

30 daysACEI cohort 30,777 3,869 (12.6%) 1,129 (3.7%) 28,187 (91.6%) 1,461 (4.7%)ARB cohort 21,436 2,440 (11.4%) 950 (4.4%) 20,284 (94.6%) 202 (0.9%)

Abbreviations: ACEI, angiotensin-converting-enzyme inhibitors; ARB, Angiotensin II receptor blockers.

eTable 5. Incidence and Risk of MACEs and adverse effects Among Patients Using ARBs and ACEIs before Propensity Score Matching.ARB cohort ACEI cohort (As Reference) Crude Adjusted† Competing Risk ‡

No. of Event

Person-years

Incidence rate*

No. of Event

Person-years

Incidence rate*

Hazard Ratio(95% CI)

p ValueHazard Ratio

(95% CI)p Value


p Value

Myocardial infarction 413 125,982 3.28 848 222,970 3.80 0.91 (0.81-1.03) 0.130 0.90 (0.79-1.02) 0.110 0.93 (0.80-1.07) 0.308Ischemic Stroke 1,033 123,751 8.35 2,145 217,561 9.86 0.87 (0.81-0.94) <0.001 0.95 (0.88-1.03) 0.233 0.96 (0.87-1.05) 0.327All-cause mortality 2,506 127,086 19.72 4,868 225,344 21.60 0.96 (0.92-1.01) 0.141 0.95 (0.90-1.00) 0.063 -Hospitalization for acute kidney injury

702 125,248 5.60 1,380 222,154 6.21 0.98 (0.90-1.08) 0.725 0.92 (0.84-1.02) 0.119 0.92 (0.82-1.03) 0.133

Hospitalization for hyperkalemia

272 126534 2.15 495 224,057 2.21 1.06 (0.91-1.23) 0.450 1.05 (0.89-1.23) 0.564 1.02 (0.85-1.23) 0.813

* per 103 person-years.† Adjusted for propensity score.‡ In this model, death was considered as a competing event.Abbreviations: CI, confidence interval, ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors.

eTable 6. Incidence and Risk of MACEs and adverse effects Among Patients Using ARBs and ACEIs With Different Enrollment Period Before Propensity Score Matching (As Treated Analysis)

ARB cohort ACEI cohort (As Reference) Crude Adjusted†

No. of Event

Person-years

Incidence rate*

No. of Event

Person-years

Incidence rate*


P ValueHazard Ratio

(95% CI)P Value

2000-2006Myocardial infarction 103 36,024 2.86 152 49,335 3.08 0.87 (0.68-1.12) 0.295 0.81 (0.62-1.07) 0.140Ischemic Stroke 284 35,744 7.95 429 48,836 8.78 0.89 (0.77-1.04) 0.130 0.91 (0.77-1.08) 0.283All-cause mortality 186 36,254 5.13 268 49,544 5.41 1.02 (0.84-1.23) 0.869 0.90 (0.73-1.11) 0.317Hospitalization for acute kidney injury 108 35,944 3.00 137 49,259 2.78 1.01 (0.78-1.30) 0.947 0.88 (0.66-1.16) 0.350Hospitalization for hyperkalemia 48 36,184 1.33 71 49,441 1.44 0.86 (0.60-1.25) 0.433 0.76 (0.51-1.14) 0.181

2007-2011Myocardial infarction 46 19,947 2.31 23 10,363 2.22 1.01 (0.61-1.67) 0.974 1.02 (0.60-1.73) 0.943Ischemic Stroke 109 19,854 5.49 60 10,345 5.80 0.94 (0.69-1.29) 0.713 0.91 (0.65-1.28) 0.595All-cause mortality 154 20,034 7.69 63 10,392 6.06 1.37 (1.02-1.84) 0.037 1.11 (0.82-1.51) 0.511Hospitalization for acute kidney injury 81 19,772 4.10 30 10,321 2.91 1.38 (0.91-2.11) 0.130 1.22 (0.95-11.30) 0.379Hospitalization for hyperkalemia 31 19,959 1.55 9 10,368 8.68 1.84 (0.87-3.87) 0.109 1.54 (0.71-3.33) 0.276

* per 103 person-years.† Adjusted for propensity score.Abbreviations: CI, confidence interval, ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors.

eTable 7. Incidence and Risk of MACEs and adverse effects Among Patients Using ARBs and ACEIs Before Propensity Score Matching by 30-, 60-, 90-day medication gap (As Treated Analysis)

ARB cohort ACEI cohort (As Reference) Crude Adjusted†

No. of Event

Person-years

Incidence rate*

No. of Event

Person-years

Incidence rate*


P ValueHazard Ratio

(95% CI)P Value

90-day medication gapMyocardial infarction 149 55,972 2.66 175 59,698 2.93 0.89 (0.72-1.11) 0.303 0.86 (0.67-1.09) 0.210Ischemic Stroke 393 55,598 7.07 489 59,181 8.26 0.85 (0.74-0.97) 0.016 0.91 (0.78-1.05) 0.200All-cause mortality 340 56,216 6.05 331 59,907 5.53 1.16 (1.00-1.35) 0.058 0.96 (0.81-1.13) 0.631Hospitalization for acute kidney injury

189 55,717 3.39 167 59,580 2.80 1.19 (0.96-1.46) 0.109 0.96 (0.76-1.21) 0.741


79 56,143 1.41 80 59,809 1.34 1.04 (0.76-1.41) 0.827 0.88 (0.63-1.25) 0.479

60-day medication gapMyocardial infarction 114 41,353 2.76 156 50,061 3.12 0.88 (0.69-1.13) 0.320 0.81 (0.62-1.05) 0.117Ischemic Stroke 383 41,060 9.33 524 49,616 10.56 0.88 (0.77-1.01) 0.067 0.89 (0.77-1.02) 0.103All-cause mortality 256 41,490 6.17 284 50,229 5.65 1.13 (0.95-1.34) 0.159 0.92 (0.77-1.11) 0.409Hospitalization for acute kidney injury

56 41,447 1.35 66 50,151 1.32 1.03 (0.72-1.47) 0.888 0.86 (0.58-1.27) 0.438


135 41,137 3.28 144 49,929 2.88 1.14 (0.90-1.44) 0.283 0.89 (0.69-1.16) 0.397

30-day medication gap

Myocardial infarction 94 28,471 3.30 130 40,237 3.23 1.04 (0.80-1.36) 0.772 0.87 (0.65-1.16) 0.343Ischemic Stroke 299 28,283 10.57 462 39,844 11.60 0.92 (0.80-1.07) 0.284 0.88 (0.75-1.03) 0.123All-cause mortality 188 28,553 6.58 241 40,354 5.97 1.09 (0.90-1.32) 0.388 0.89 (0.72-1.10) 0.282Hospitalization for acute kidney injury

94 28,312 3.32 110 40,122 2.74 1.23 (0.93-1.62) 0.139 0.93 (0.69-1.25) 0.628


36 28,534 1.26 51 40,301 1.27 1.01 (0.66-1.54) 0.975 0.84 (0.53-1.33) 0.452

* per 103 person-years.† Adjusted for propensity score.Abbreviations: CI, confidence interval, ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors.

eTable 8. Risk of MACEs and adverse effects Among Patients Using ARBs and ACEIs, by Hospital Levels.

Patient No (%).Myocardial infarction

Ischemic Stroke All-cause mortalityHospitalization for

acute kidney injury


Hospital Level ARB cohort ACEI cohort HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI)Level I (Medical center) 5,001 (23.3) 4,174 (13.6) 0.97 (0.76-1.26) 0.88 (0.74-1.05) 0.90 (0.81-1.01) 0.85 (0.69-1.04) 1.08 (0.77-1.53)Level II 6,081 (28.4) 6,307 (20.5) 1.01 (0.79-1.30) 0.99 (0.85-1.16) 0.96 (0.87-1.06) 1.00 (0.84-1.19) 1.02 (0.76-1.36)Level III 6,967 (32.5) 7,562 (24.6) 0.94 (0.67-1.04) 0.94 (0.82-1.08) 0.92 (0.83-1.01) 0.99 (0.82-1.19) 1.06 (0.80-1.40)Level IV (Local medical

clinics)3,387 (15.8) 12,734 (41.4) 0.73 (0.52-1.04) 1.01 (0.83-1.23) 1.10 (0.97-1.25) 0.78 (0.60-1.02) 1.07 (0.70-1.64)

Abbreviations: HR, hazard ratio; CI, confidence interval; ARB, Angiotensin II receptor blockers; ACEI, angiotensin-converting-enzyme inhibitors.

eTable 9. Risk of MACEs and adverse effects Among Patients Using ARBs and ACEIs Before Propensity Score Matching

Crude Model 1a Model 2 b Model 3 c

HR (95% CI)

PHR

(95% CI)P

HR(95% CI)

PHR

(95% CI)P

Myocardial infarction 0.91 (0.81-1.03) 0.130 0.90 (0.79-1.02) 0.103 0.90 (0.79-1.02) 0.101 0.94 (0.81-1.08) 0.381Ischemic Stroke 0.87 (0.81-0.94) <0.001 0.95 (0.87-1.03) 0.191 0.95 (0.88-1.03) 0.232 0.96 (0.87-1.05) 0.335All-cause mortality 0.96 (0.92-1.01) 0.141 0.95 (0.90-1.00) 0.065 0.95 (0.90-1.00) 0.054 0.96 (0.90-1.02) 0.156Hospitalization for acute kidney injury

0.98 (0.90-1.08) 0.725 0.93 (0.84-1.03) 0.151 0.92 (0.84-1.02) 0.113 0.92 (0.82-1.03) 0.144


1.06 (0.91-1.23) 0.450 1.06 (0.90-1.24) 0.504 1.05 (0.89-1.23) 0.571 1.04 (0.86-1.25) 0.695

a Adjusted for decile of propensity score.b Cox models with sandwich estimator of hospital-level (Four level: medical center, metropolitan hospital, local hospital, and medical clinic) and adjusted for propensity score.c Cox models with shared-frailty model of the individual hospital identifier and adjusted for propensity score.Abbreviations: HR, hazards ratio; CI, confidence interval.

eTable 10. Effect of Continuing Use of ARBs on Risk of MACEs and adverse effects Marginal Structural Models with Inverse-Probability-

of-Treatment WeightsHazard Ratio

(95% CI)p Value

Myocardial infarction 0.78 (0.42-1.46) 0.433Ischemic Stroke 1.25 (0.85-1.83) 0.261All-cause mortality 0.95 (0.88-1.01) 0.104Hospitalization for acute kidney injury 0.96 (0.53-1.75) 0.902Hospitalization for hyperkalemia 1.00 (0.34-2.89) 0.997

Abbreviations: CI, confidence interval, ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors.

eTable 11. Subgroup analysis of Risk of All-cause Mortality Among Patients Using ARBs and ACEIs

CharacteristicHazard Ratio*

(95% CI)P Value

InteractionP Value

GenderMale 0.948 (0.880-1.022) 0.164 0.834Female 0.961 (0.872-1.060) 0.425

Age20-40 years 0.993 (0.762-1.295) 0.961 0.96441-60 years 0.961 (0.869-1.062) 0.43661-80 years 0.939 (0.863-1.023) 0.149≧81 years 0.968 (0.807-1.161) 0.727

Charlson Comorbidity Index score1 0.971 (0.871-1.082) 0.593 0.3722 0.862 (0.759-0.979) 0.0223 0.986 (0.859-1.131) 0.837≧4 0.988 (0.886-1.102) 0.832

HypertensionYes 0.936 (0.880-1.000) 0.050 0.145No 1.073 (0.908-1.267) 0.410

Chronic kidney diseaseYes 0.955 (0.820-1.113) 0.556 0.959No 0.953 (0.894-1.016) 0.144

Heart failureYes 0.814 (0.693-0.955) 0.012 0.034No 0.970 (0.910-1.034) 0.351

Coronary artery diseaseYes 0.925 (0.824-1.039) 0.188 0.565No 0.963 (0.899-1.032) 0.286

RAAS mono-therapyYes 0.970 (0.893-1.054) 0.477 0.543No 0.935 (0.858-1.018) 0.120

* Adjusted for propensity score; ACEI cohort: reference group Abbreviations: CI, confidence interval; ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors; RAAS, Renin-angiotensin-aldosterone system.

eTable 12. Subgroup analysis of Risk of Myocardial Infarction Among Patients Using ARBs and ACEIs


(95% CI)P Value

InteractionP Value









* Adjusted for propensity score; ACEI cohort: reference group Abbreviations: CI, confidence interval; ARBs, Angiotensin II receptor blockers; ACEIs, angiotensin-converting-enzyme inhibitors, ; RAAS, Renin-angiotensin-aldosterone system.

eTable 13. Subgroup analysis of Risk of Ischemic Stroke Among Patients Using ARBs and ACEIs


(95% CI)P Value

InteractionP Value









* Adjusted for propensity score; ACEI cohort: reference group Abbreviations: CI, confidence interval; ARB, Angiotensin II receptor blockers; ACEI, angiotensin-converting-enzyme inhibitors, ; RAAS, Renin-angiotensin-aldosterone system.

eFigure 1. Study flow diagram

eFigure 2. Distribution of propensity scores for using ARB before and after matching

eFigure 3. Overlap in distribution of propensity scores for using ARB among ACEI

and ARB cohort before and after matching

eFigure 4. Cumulative Incidence of Adverse Outcomes in ACEI and ARB Users with

Type 2 Diabetes

(A) hospitalization for acute kidney injury, (B) hospitalization for hyperkalemia.

All patients were ACEI/ARB-naïve and newly diagnosed with type 2 diabetes.


blocker.

(4A) hospitalization for acute kidney injury

(4B) hospitalization for hyperkalemia

eFigure 5. Forest Plots of ACEI and ARB Effects

Subgroup analysis measured the risks of (A) myocardial infarction, (B) ischemic

stroke, and (C) all-cause mortality.


blocker; RAAS, Renin-angiotensin-aldosterone system

(5A) myocardial infarction

(5B) ischemic stroke

(5C) all-cause mortality

Documents

· Web vieweTable 13 Subgroup analysis of Risk of Ischemic Stroke Among Patients Using ARBs and ACEIs eFigure 1 Study flow diagram eFigure 2. Distribution of propensity scores for