Stroke

CLINICAL AND POPULATION SCIENCES

Baseline Cardiovascular Risk Factor Control in

Patients With Type 2 Diabetes and Coronary
Disease Versus Stroke: Secondary Analysis of
Cardiovascular Outcome Trials
Priyadarshini Balasubramanian , MD; Walter N. Kernan , MD; Kevin N. Sheth , MD; Anne Pernille Ofstad , MD;
Julio Rosenstock , MD; Christoph Wanner , MD; Bernard Zinman, MD; Michaela Mattheus , Dipl Biostat; Nikolaus Marx , MD;
Silvio E. Inzucchi , MD

BACKGROUND: Patients with type 2 diabetes (T2D) and cardiovascular disease are at increased risk for recurrent ischemic
events. Cardiovascular risk factor control is vital for secondary prevention, but how this compares among individuals with
different T2D macrovascular complications is unknown. We aimed to determine if there might be differences in risk factor
control in patients with T2D with previous stroke versus coronary artery disease (CAD).

METHODS: Cross-sectional analyses were performed on 12 856 patients with T2D with prior history of stroke with or without
CAD from 3 diabetes cardiovascular outcome trials: CARMELINA (The Cardiovascular and Renal Microvascular Outcome
Study With Linagliptin), EMPA-REG OUTCOME (Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes
Mellitus Patients), and CAROLINA (The Cardiovascular Outcome Study of Linagliptin vs Glimepiride in Type 2 Diabetes). Risk
factors at baseline assessed included dyslipidemia, hypertension, smoking, and current antiplatelet/anticoagulant therapy.
Downloaded from http://ahajournals.org by on August 27, 2023

Control, respectively, was defined as LDL (low-density lipoprotein)-C <100 mg/dL or statin use, systolic blood pressure
<140 and diastolic blood pressure <90 mm Hg, not currently smoking, and use of an antiplatelet/anticoagulant medication.
The odds ratio of 3 to 4 (or good) versus 0 to 2 (or suboptimal) risk factors controlled was analyzed by logistic regression
models.

RESULTS: The odds for good versus suboptimal risk factor control in patients with CAD alone was higher than in those with
stroke alone across all 3 trials odds ratios (95% CI): CARMELINA, 2.05 (1.67–2.51), EMPA-REG OUTCOME, 2.50 (2.10–
2.99), and CAROLINA, 1.63 (1.21–2.20). The respective odds ratios were lower (and rendered nonsignificant in CAROLINA)
when cardiovascular risk factor control in patients with both CAD and stroke were compared with those with stroke alone:
CARMELINA, 1.45 (1.13–1.87); EMPA-REG OUTCOME, 1.62 (1.25–2.08); and CAROLINA, 1.16 (0.74–1.83).

CONCLUSIONS: In contemporary populations of patients with T2D, there was significant discordance in control of
cardiovascular risk factors between patients with stroke versus CAD, with the former having less optimal control. The
intermediate results in patients with both CAD and stroke suggest that these differences could be related at least in part
to clinician factors.

REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifiers: NCT01243424, NCT01131676, NCT01897532.

GRAPHIC ABSTRACT: A graphic abstract is available for this article.

Key Words: anticoagulant ◼ cardiovascular disease ◼ hypertension ◼ odds ratio ◼ risk factors

Correspondence to: Priyadarshini Balasubramanian, MD, Section of Endocrinology, Department of Medicine, 333 Cedar St, Yale School of Medicine, New Haven, CT
06510. Email priyadarshini.balasubramanian@yale.edu
This manuscript was sent to Erica Jones, Editor-in-Training, and Seemant Chaturvedi, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.122.042053.
For Sources of Funding and Disclosures, see page 2021.
© 2023 The Authors. Stroke is published on behalf of the American Heart Association, Inc., by Wolters Kluwer Health, Inc. This is an open access article under the
terms of the Creative Commons Attribution License, which permits use, distribution, and reproduction in any medium, provided that the original work is properly cited.
Stroke is available at www.ahajournals.org/journal/str

Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053 August 2023   2013

 Balasubramanian et al
CLINICAL AND POPULATION
Nonstandard Abbreviations and Acronyms
SCIENCES
CAD coronary artery disease
CARMELINA The Cardiovascular and
Renal Microvascular
Outcome Study With
Linagliptin
CAROLINA The Cardiovascular Out-
come Study of Linagliptin
vs Glimepiride in Type 2
Diabetes
DPP-4 dipeptidyl peptidase 4
EMPA-REG OUTCOME Empagliflozin Cardiovas-
cular Outcome Event Trial
in Type 2 Diabetes Mellitus
Patients
LDL low-density lipoprotein
SGLT2  sodium-glucose
cotransporter-2
T2D type 2 diabetes
UACR  urine albumin/creatinine ratio
I
schemic stroke is a leading cause of morbidity and
mortality world wide and contributes significantly to
health care expenditures. Type 2 diabetes (T2D) is
common among patients with stroke and is a major risk
factor for recurrent stroke events. During the last few
Downloaded from http://ahajournals.org by on August 27, 2023
decades, research has shown that the recurrence risk
may be reduced by careful attention to specific interven-
tions including lipid lowering, management of hyperten-
sion, smoking cessation, and antiplatelet therapy. Despite
these advances the risk of recurrent stroke remains as
high as 15% to 30% in the first 5 years after an initial
event.1 About 1 in 2 survivors remain with some form of
disability and about 1 in 7 require institutionalized care.2
Preventing recurrent stroke as well as other atheroscle-
rotic cardiovascular disease events is therefore a major
goal of care for patients with cerebrovascular disease.
There are few data, however, on the relative control of
these risk factors during the outpatient care of patients
after stroke versus myocardial infarction (MI) or, more
broadly, coronary artery disease (CAD).
Over the past decade, large cardiovascular outcome
trials in patients with T2D have demonstrated cardio-
vascular safety and efficacy of newer glucose-lowering
medications. The regulatory mandate to assess the car-
diovascular safety of new glucose-lowering medications
stemmed from concerns related to older drugs used
for T2D that may have actually increased cardiovascu-
lar risk.3 In the outcome trials conducted with the newer
T2D medications, extensive information was obtained
at baseline regarding prior cardiovascular diseases as
well as control of atherosclerotic cardiovascular disease
risk factors. This provided an opportunity to compare
2014   August 2023
CV Risk Factor Control in T2D Patients With Stroke
the degree of baseline cardiovascular risk factor control
among patients with T2D with prior stroke and CAD. We
specifically analyzed the results from 3, large cardio-
vascular outcome trials in patients with T2D, seeking to
determine if there might be significant differences in the
extent to control cardiovascular risk factors in patients
with T2D with previous stroke versus CAD.
METHODS
The data, analytic methods, and study materials will be made
available to other researchers for purposes of reproducing the
results or replicating the procedure. To ensure independent
interpretation of clinical study results and enable authors to fulfil
their role and obligations under the ICMJE criteria, Boehringer
Ingelheim grants all external authors access to relevant clinical
study data. In adherence with the Boehringer Ingelheim Policy
on Transparency and Publication of Clinical Study Data, sci-
entific and medical researchers can request access to clinical
study data after publication of the primary article and second-
ary analyses in peer-reviewed journals and regulatory and reim-
bursement activities are completed, normally within 1 year after
the marketing application has been granted by major Regulatory
Authorities. Researchers should use the https://vivli.org/ link to
request access to study data and visit https://www.mystudywin-
dow.com/msw/datasharing for further information.
The article follows the STROBE (Strengthening the
Reporting of Observational Studies in Epidemiology) reporting
guidelines (Supplemental Material). Written informed consent
was obtained from all participants in the original trials, per Good
Clinical Practice guidelines. Given the nature of this study, no
additional consent was warranted.
Study Design
We performed a series of cross-sectional analyses of patient
level data from 3 recent cardiovascular outcome trials in
T2D: EMPA-REG OUTCOME (Empagliflozin Cardiovascular
Outcome Event Trial in Type 2 Diabetes Mellitus Patients),
CAROLINA (The Cardiovascular Outcome Study of Linagliptin
vs Glimepiride in Type 2 Diabetes), and CARMELINA (The
Cardiovascular and Renal Microvascular Outcome Study With
Linagliptin), where we had complete access to participant-level
baseline data. EMPA-REG OUTCOME was a multicenter ran-
domized placebo-controlled trial that evaluated the efficacy and
safety of the SGLT2 (sodium-glucose cotransporter-2) inhibitor
empagliflozin in addition to standard of care in 7020 patients
with T2D and established cardiovascular disease.4 CAROLINA
was a multicenter, active-controlled, randomized clinical trial
that studied cardiovascular outcomes in 6033 patients with
T2D and high cardiovascular risk randomized 1:1 to linagliptin,
a DPP-4 (dipeptidyl peptidase 4) inhibitor, or the sulfonylurea
glimepiride when added to usual therapy.5 High cardiovascular
risk was defined as documented atherosclerotic cardiovascu-
lar disease, multiple cardiovascular risk factors (at least 2 of
the following: T2D duration >10 years; systolic blood pressure
>140 mm Hg or receiving at least 1 blood pressure–lowering
treatment; current smoker; low-density lipoprotein cholesterol
≥135 mg/dL [3.5 mmol/L] or receiving lipid-lowering treat-
ment), aged at least 70 years, or evidence of microvascular
Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053
 Balasubramanian et al CV Risk Factor Control in T2D Patients With Stroke

Table 1. Major Design Characteristics of the 3 Trials was defined as baseline (1) LDL (low-density lipoprotein) cho-

CLINICAL AND POPULATION

EMPA-REG
lesterol <100 mg/dL or statin use, (2) systolic blood pressure
TRIAL CARMELINA ­OUTCOME CAROLINA <140 mm Hg and diastolic blood pressure <90 mm Hg, (3) not
smoking, and (4) use of antiplatelet or anticoagulant drugs,

SCIENCES
Randomized Linagliptin vs Empagliflozin vs Linagliptin vs
study placebo placebo Glimepiride respectively. In a sensitivity analysis, we additionally used LDL
­medication <70 mg/dL or statin use to define control of the lipid risk factor.
Number of 6979 7020 6033 For all 3 studies, the information on characteristics to define
patients the cardiovascular disease groups and cardiovascular risk fac-
Major • Type 2 • Type 2 diabetes • Type 2 tor control were mandatory to be collected as per each clini-
­inclusion ­diabetes • BMI ≤45 kg/m2 ­diabetes cal trial protocol. Across all 3 studies, there was 1 patient in
criteria • HbA1c • Established CVD • HbA1c of EMPA-REG OUTCOME with missing information on stroke,
6.5–10% • eGFR≥30 mL/ 6.5%–8.5% while all patients provided data to assess CAD. Among patients
• High CV risk min/1.73 m2 AND
and renal risk* • HbA1c 7.0–9.0% • High CV risk*
with CAD and stroke, 99.8% provided data for the assessment
if no blood of cardiovascular risk factor control in EMPA-REG OUTCOME,
glucose-lowering 97.7% in CAROLINA, and 99.0% in CARMELINA.
treatment, 7.0–
10% if on blood
glucose-lowering Statistical Analyses
treatment
Analyses were performed in each trial separately, based on
Median 2.2 y 3.1 y 6.3 y patients randomized and treated with available data for cardio-
follow-up
vascular disease groups and cardiovascular risk factor control
BMI indicates body mass index; CARMELINA, The Cardiovascular and Renal and on pooled treatment arms within each trial. Baseline char-
Microvascular Outcome Study With Linagliptin; CAROLINA, The Cardiovascular
Outcome Study of Linagliptin vs Glimepiride in Type 2 Diabetes; CV, cardiovas-
acteristics for cardiovascular disease groups were reported
cular; CVD, cardiovascular disease; EMPA-REG OUTCOME, Empagliflozin Car- as means (continuous variables) or numbers and proportions
diovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients; eGFR, (categorical variables) based on patients with data available for
estimated glomerular filtration rate; and HbA1c, hemoglobin A1c. assessment of cardiovascular risk factor control.
*Definitions included in the text.
The odds ratio (OD) of having all or most (3 to 4, and there-
fore good) versus less (0 to 2, and therefore suboptimal) car-
complications (impaired kidney function [eGFR of 30–59 mL/ diovascular risk factors controlled was analyzed by a logistic
min per 1.73 m2 or urine albumin/creatinine ratio ≥30 mg/g]) or regression model that included a term for the cardiovascular
proliferative retinopathy.5 CARMELINA was a third multicenter disease group. We performed 2 comparisons: (1) patients with
Downloaded from http://ahajournals.org by on August 27, 2023

randomized placebo-controlled trial enriched with patients with
evidence of chronic kidney disease that evaluated the effect of
linagliptin versus placebo, when added to usual care, on cardio-
vascular and prespecified kidney outcomes in 6979 patients
with T2D and high cardiovascular and renal risk. High cardio-
vascular risk and renal risk in CARMELINA was defined as
history of macrovascular disease with urine albumin/creatinine
ratio ≥30 mg/g; or reduced eGFR with or without urine albu-
min/creatinine ratio >200 mg/g.6
The major design features of the 3 trials have been shown
in Table 1.
For the current exploratory analyses, we included partici-
pants from these 3 trials who had a history of stroke with or
without CAD and available data on baseline cardiovascular risk
factor control.
Definition of Cardiovascular Disease Groups
Patients were defined as having a history of stroke or CAD
based on investigator reports. CAD was defined as history of
MI or coronary artery bypass graft or otherwise documented
CAD. We defined 3 mutually exclusive groups for subsequent
comparison: one that had a history of stroke only (ie, without
CAD), 1 with a history of CAD only (ie, without stroke), and 1
that had a history of both stroke and CAD.
Definition of Cardiovascular Risk Factor Control
Data on risk factors were captured at baseline. Risk factors
assessed were dyslipidemia, hypertension, smoking, and use
of antiplatelet and anticoagulant drugs. Control of a risk factor
CAD alone versus stroke alone and (2) patients with both CAD
and stroke versus patients with stroke alone. These analyses
were repeated in subgroups by age, sex, and geographic region
in separate logistic regression models including terms for sub-
group and the interaction term of subgroup with cardiovascu-
lar disease group. In sensitivity analyses, we also assessed the
control of 4 versus less (0 to 3) and 2 to 4 versus less (0 to 1)
cardiovascular risk factors.
In further sensitivity analyses we performed adjustment for
age, sex and region. All analyses were exploratory on a nominal
2-sided α=0.05 without adjustment for multiplicity. All statisti-
cal analyses were performed using SAS software, version 9·4
(Cary, NC.)
RESULTS
The demographics and selected baseline characteristics
by cardiovascular disease category for each of the 3 tri-
als are shown in Table 2.
Overall, 80.7% participants in the CARMELINA trial,
86.6% participants in EMPA-REG OUTCOME and
79.4% participants in the CAROLINA trial had 3 or 4
cardiovascular risk factors controlled at baseline. Among
participants with CAD alone, 83.9% in the CARMELINA
trial, 89.1% in the EMPA-REG OUTCOME trial and
83.1% in the CAROLINA trial had 3 or 4 cardiovascu-
lar risk factors controlled and the corresponding rates
among participants with stroke alone were consistently
lower at 71.7%, 76.6%, and 75.1%, respectively. When

Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053 August 2023   2015

 Balasubramanian et al CV Risk Factor Control in T2D Patients With Stroke

Table 2. Baseline Characteristics by CV Disease Categories in the 3 Trials

CLINICAL AND POPULATION

CARMELINA TRIAL EMPA-REG OUTCOME TRIAL CAROLINA TRIAL

CAD and CAD and CAD and

SCIENCES

Stroke alone CAD alone stroke Stroke alone CAD alone stroke Stroke alone CAD alone stroke

n=576 n=3339 n=735 n=932 n=4776 n=698 n=297 n=1341 n=162

Male, n (%) 339 (58.9) 2268 (67.9) 468 (63.7) 513 (55.0) 3651 (76.4) 486 (69.6) 176 (59.3) 1021 (76.1) 112 (69.1)

Age, y; mean±SD 64.9±8.7 65.9±9.0 66.1±8.2 62.1±8.8 63.2±8.6 65.7±7.5 63.5±9.2 65.1±8.7 66.0±8.5

T2D duration, y, n (%)

 >10 358 (62.2) 2093 (62.7) 468 (63.7) 451 (48.4) 2768 (58.0) 423 (60.6) 77 (25.9) 362 (27.0) 56 (34.6)

 >5 to ≤10 112 (19.4) 697 (20.9) 153 (20.8) 243 (26.1) 1197 (25.1) 169 (24.2) 92 (31.0) 407 (30.4) 45 (27.8)

≤5 106 (18.4) 549 (16.4) 114 (15.5) 238 (25.5) 811 (17.0) 106 (15.2) 128 (43.1) 572 (42.7) 61 (37.7)

Obstructive sleep apnea* 20 (3.5) 177 (5.3) 27 (3.7) 26 (2.8) 303 (6.3) 44 (6.3) 8 (2.7) 77 (5.7) 19 (11.7)

Region, n (%)

 Europe 250 (43.4) 1462 (43.8) 458 (62.3) 432 (46.4) 1846 (38.7) 346 (49.6) 138 (46.5) 627 (46.8) 73 (45.1)

 Asia 62 (10.8) 263 (7.9) 55 (7.5) 235 (25.2) 933 (19.5) 109 (15.6) 78 (26.3) 260 (19.4) 31 (19.1)

 North America 67 (11.6) 609 (18.2) 94 (12.8) 92 (9.9) 1105 (23.1) 144 (20.6) 31 (10.4) 283 (21.1) 33 (20.4)

 Latin America 197 (34.2) 1005 (30.1) 128 (17.4) 150 (16.1) 649 (13.6) 74 (10.6) 44 (14.8) 142 (10.6) 23 (14.2)

 Africa - - - 23 (2.5) 243 (5.1) 25 (3.6) 6 (2.0) 29 (2.2) 2 (1.2)

BMI, kg/m2; mean±SD 30.8±5.4 31.4±5.2 31.3±4.9 29.9±5.4 30.7±5.2 30.9±5.28 29.05±4.93 29.7±5.0 30.3±5.1

HbA1c, %, mean±SD 7.95±1.00 7.96±0.99 8.01±1.04 8.04±0.89 8.07±0.83 8.04±0.86 7.13±0.63 7.14±0.55 7.11±0.60

SBP, mm Hg, mean±SD 141.2±18.4 139.4±17.4 140.3±16.4 137.1±16.8 134.4±16.7 137.8±17.5 135.9±17.4 135.5±16.5 138.1±15.9

DBP, mm Hg, mean±SD 79.3±11.1 77.3±10.4 78.7±9.8 78.5±9.5 76.2±9.8 76.9±9.9 79.6±9.2 78.3±9.4 79.5±10.3

LDL-C, mg/dL, mean±SD 94.4±39.7 86.8±38.8 94.0±41.5 97.3±41.1 81.6±33.6 86.5±35.3 92.5±35.8 87.8±32.8 91.9±37.5

eGFR,† mL/min/1.73 m2, 62.1±25.8 58.6±25.3 62.1±24.1 76.4±22.7 73.6±20.7 68.9±19.5 77.2±21.3 73.9±18.7 73.3±18.7
mean±SD

eGFR<60 mL/min/1.73 m2 269 (46.7) 1848 (55.3) 353 (48.0) 221 (23.7) 1232 (25.8) 226 (32.4) 59 (19.9) 307 (22.8) 41 (25.3)

Insulin use, n (%) 311 (54.0) 1853 (55.5) 396 (53.9) 369 (39.6) 2333 (48.8) 363 (52.0) 0 (0.0) 2 (0.1) 0 (0.0)

Metformin use, n (%) 369 (64.1) 1966 (58.9) 459 (62.4) 692 (74.2) 3572 (74.8) 481 (68.9) 178 (59.9) 744 (55.5) 81 (50.0)
Downloaded from http://ahajournals.org by on August 27, 2023

Statin use, n (%) 402 (69.8) 2686 (80.4) 547 (74.4) 612 (65.7) 3914 (82.0) 544 (77.9) 199 (67.0) 1099 (82.0) 113 (69.8)

Antihypertensive use, n (%) 545 (94.6) 3264 (97.8) 714 (97.1) 850 (91.2) 4608 (96.5) 688 (98.6) 269 (90.6) 1284 (95.7) 158 (97.5)

Antiplatelet use, n (%)

 Aspirin 322 (55.9) 2520 (75.5) 515 (70.1) 639 (68.6) 4191 (87.8) 563 (80.7) 173 (58.2) 1009 (75.2) 115 (71.0)

 Clopidogrel 101 (17.5) 813 (24.3) 170 (23.1) 86 (9.2) 550 (11.5) 90 (12.9) 58 (19.5) 298 (22.2) 31 (19.1)

 Dipyridamole 5 (0.9) 3 (0.1) 0 (0.0) 23 (2.5) 11 (0.2) 12 (1.7) 14 (4.7) 5 (0.4) 12 (7.4)

Anticoagulant use, n (%)

 Vitamin K antagonists 34 (5.9) 264 (7.9) 77 (10.5) 40 (4.3) 269 (5.6) 79 (11.3) 24 (8.1) 94 (7.0) 19 (11.7)

 Direct thrombin inhibitors 11 (1.9) 25 (0.7) 9 (1.2) 3 (0.3) 14 (0.3) 2 (0.3) 1 (0.3) 4 (0.3) 0 (0.0)

 Direct Factor Xa inhibitors 4 (0.7) 46 (1.4) 20 (2.7) 1 (0.1) 2 (<0.1) 2 (0.3) 1 (0.3) 0 (0.0) 0 (0.0)

 Heparin 2 (0.3) 28 (0.8) 9 (1.2) 5 (0.5) 11 (0.2) 1 (0.1) 4 (1.3) 3 (0.2) 1 (0.6)

BMI indicates body mass index; CAD, coronary artery disease; CARMELINA, The Cardiovascular and Renal Microvascular Outcome Study With Linagliptin; CARO-
LINA, The Cardiovascular Outcome Study of Linagliptin vs Glimepiride in Type 2 Diabetes; CV, cardiovascular; DBP, diastolic blood pressure; EMPA-REG OUTCOME,
Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c; LDL-C, low
density lipoprotein - cholesterol; MDRD, Modification of Diet in Renal Disease Study; SBP, systolic blood pressure; and Xa, coagulation factor Xa.
*By patient report only. Likely represents an underestimate of true prevalence.
†MDRD equation.

participants with both CAD and stroke were analyzed, the
results appeared to be intermediate, at 78.6%, 84.1%,
and 77.8% across the 3 trials, respectively. About 16.1%
participants in CARMELINA trial, 10.9% participants
in EMPA-REG OUTCOME, and 16.9% participants in
CAROLINA trial had 0 to 2 risk factors controlled in the
CAD alone group and the corresponding rates among
participants with stroke alone were 28.3%, 23.4%, and
24.9%, respectively. Among patients with both CAD and
stroke, 21.4%, 15.9%, and 22.2% had 0 to 2 risk factors
controlled. The percentage of patients with good versus
suboptimal cardiovascular risk factor control by cardio-
vascular disease groups is shown in Figure 1. Cardio-
vascular risk factor control by cardiovascular groups is
shown in Table 2.
The relative odds for good versus suboptimal risk fac-
tor control (ie, 3–4 risk factors versus ≤2 risk factors) in
patients with CAD alone was higher than in those with
stroke alone across all 3 trials (ORs [95% CI]: CAR-
MELINA 2.05 [1.67–2.51]; EMPA-REG OUTCOME,

2016   August 2023 Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053

 Balasubramanian et al
Downloaded from http://ahajournals.org by on August 27, 2023
2.50 [2.10–2.99]; and CAROLINA, 1.63 [1.21–2.20];
Figure 2). The respective ORs (95% CIs) were lower
(and rendered nonsignificant in CAROLINA) when car-
diovascular risk factor control in patients with both CAD
and stroke were compared with those with stroke alone:
CARMELINA, 1.45 (1.13–1.87), EMPA-REG OUT-
COME, 1.62 (1.25–2.08), and CAROLINA, 1.16 (0.74–
1.83). Using the more stringent criterion of LDL <70
mg/dL or statin use, the results were consistent, show-
ing that the odds for good versus suboptimal risk fac-
tor control in patients with CAD alone were higher than
those in stroke alone across the 3 trials (ORs [95% CIs]:
CARMELINA 2.10 [1.73–2.55]; EMPA-REG OUTCOME
2.44 [2.07–2.89]; and CAROLINA 1.74 [1.31–2.31]).
Also, in patients with both CAD and stroke the respec-
tive odds for good control were higher than in patients
with stroke alone (ORs [95% CIs]: CARMELINA, 1.56
[1.23–1.99], EMPA-REG OUTCOME, 1.65 [1.29–2.10];
and CAROLINA, 1.06 [0.69–1.62]; data not shown). In
subgroup analyses by age, sex, and geographic region,
the above differences in cardiovascular risk factor con-
trol across cardiovascular disease categories were gen-
erally consistent in all 3 trials (interaction P<0.05), while
Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053
CV Risk Factor Control in T2D Patients With Stroke
CLINICAL AND POPULATION
SCIENCES
Figure 1. The percentage of patients
with good (3–4 risk factors) vs
suboptimal (0–2 risk factors)
controlled by cardiovascular disease
subgroups at baseline in the 3 trials.
CAD indicates coronary artery disease;
CARMELINA, The Cardiovascular and
Renal Microvascular Outcome Study With
Linagliptin; CAROLINA, The Cardiovascular
Outcome Study of Linagliptin vs
Glimepiride in Type 2 Diabetes; and
EMPA-REG OUTCOME, Empagliflozin
Cardiovascular Outcome Event Trial in Type
2 Diabetes Mellitus Patients.
across regions in EMPA-REG OUTCOME, there was
some difference in the magnitude of the effect.
Results for the control of 4 versus less (ie, 0–3) and 2
to 4 versus less (ie, 0–1) cardiovascular risk factors were
generally consistent in that the odds of good versus sub-
optimal control was generally higher in patients with CAD
alone versus stroke alone while the differences (odds
ratios) were smaller when patients with both CAD and
stroke were compared to those with stroke alone (but
nonsignificant in CAROLINA; full data shown in Supple-
mental Material).
Following adjustment for age, sex, and region, the
results for good versus suboptimal risk factor control (ie,
3–4 risk factors versus ≤2 risk factors) in patients with
CAD alone versus stroke alone and of those with both
CAD and stroke versus stroke alone were consistent
across all 3 studies (data not shown).
DISCUSSION
Among 12 856 patients with diabetes in 3 recent car-
diovascular outcome trials of glucose-lowering therapy,
achievement of goals for preventive therapies appeared
August 2023   2017
 Balasubramanian et al CV Risk Factor Control in T2D Patients With Stroke
CLINICAL AND POPULATION
SCIENCES
Downloaded from http://ahajournals.org by on August 27, 2023

Figure 2. Forest plot showing cardiovascular (CV) risk factor control by CV disease groups in the 3 trials.
CAD indicates coronary artery disease; CARMELINA, The Cardiovascular and Renal Microvascular Outcome Study With Linagliptin; CAROLINA,
The Cardiovascular Outcome Study of Linagliptin vs Glimepiride in Type 2 Diabetes; and EMPA-REG OUTCOME, Empagliflozin Cardiovascular
Outcome Event Trial in Type 2 Diabetes Mellitus Patients.

better in patients with history of CAD compared with
stroke. The largest relative difference in risk factor control
was observed between patients with CAD alone versus
stroke alone, that is, where clinical management is being
influenced by the presence of disease solely in that vas-
cular bed. Our findings were intermediate for those with
both CAD and stroke. This may provide a clue for the
underlying reasons for the differences we detected. The
presence of coronary disease itself, irrespective of stroke,
appears to increase the likelihood of better risk factor
control. Interestingly, the major driver for the overall dif-
ference in risk factor control across the CAD and stroke
groups were blood pressure control, LDL-C control, or
statin use and use of antiplatelet/anticoagulants which
are predominantly clinician-driven. In contrast, there was
similar control of smoking across these groups, which is
predominantly a patient decision. Although patient adher-
ence plays a role in control of both blood pressure and
LDL-C and smoking cessation can indeed be influenced
by practitioners, clinician-related factors are likely the
major influence in the first 2. The intermediate results in
patients with both CAD and stroke support this point. The
magnitude of differences was variable across the 3 tri-
als, with the largest differences in the EMPA-REG OUT-
COME trial and smallest in the CAROLINA trial. This is
likely related to the fact that the latter study’s population
generally had more recently diagnosed T2D and fewer
underlying comorbidities. According to some studies, the
guidelines-adherent management of modifiable cardio-
vascular risk factors including dyslipidemia, hypertension,

2018   August 2023 Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053

 Balasubramanian et al
and smoking cessation can decrease the risk of recur-
rence of stroke by about 80%.7 Therefore, irrespective of
the reasons of these disparities, our findings underscore
the need for improved care in patients with stroke.
Ours is the first study conducted in a T2D popula-
tion enrolled in modern outcome trials, where cardio-
vascular risk factor control was assiduously recorded.
These data are consistent with research in the general
population. A multicenter observational case-control
study that included 5458 patients from 1444 primary
health centers in Spain compared the rates of risk
factor control in patients with ischemic stroke versus
patients with CAD. The therapeutic targets analyzed
in that study were significantly lower in patients after
ischemic stroke for hypertension (23.0% versus 27.2%)
and dyslipidemia (13.6% versus 20.3%) compared with
patients with CAD.8 Another retrospective study com-
pared the risk factor control 1 year after discharge in
patients with ischemic stroke versus MI and reported
higher odds for optimal blood pressure management
in patients who suffered MI in the past year com-
pared to those with stroke. However, the study found
no differences in control rates for hyperlipidemia and,
if anything, higher odds of better glycemic control in
patients with ischemic stroke who had T2D.9 A third
study, involving a nationwide cross-sectional investiga-
tion in France, compared the management of hyperten-
sion in stroke versus CAD patients and reported poorer
blood pressure control in the former versus the latter
Downloaded from http://ahajournals.org by on August 27, 2023
(24.5% versus 34.2%; P<0.01). The study also noted
that antihypertensive monotherapy was more common
in patients with stroke than in those with CAD (43.2%
versus 31.4%; P<0.0001), which could explain the dif-
ferences in blood pressure achievement.10
Stroke and CAD are primarily the result of athero-
sclerosis and treatments for secondary prevention are
substantially similar. So why might there be differences
in the quality of preventive care for patients with these
2 conditions? One possible explanation is that patients
with CAD are more commonly cared for by cardiolo-
gists (in addition to primary care providers), who tend
to follow their patients regularly and, in many circum-
stances, indefinitely. This is particularly the case after
acute cardiovascular events (eg, MI or invasive coro-
nary interventions). Moreover, this specialty has been
particularly focused on risk factor modification. In con-
trast, however, most patients with stroke will see a neu-
rologist during the hospitalization and later in the office
for a limited number of visits. Most patients with stroke
may also be seen by physical medicine and rehabili-
tation providers for several months, but their focus is
mainly on the recuperative process with, traditionally,
less attention to risk factor assessment or manage-
ment. The hypothesis that intensity of subspecialty care
may explain differences in risk factor control is sup-
ported by research from the Veterans Administration
Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053
CV Risk Factor Control in T2D Patients With Stroke
showing that patients with history of MI have more
CLINICAL AND POPULATION
outpatient visits (mean of 7.9±6.1 visits [median 7])
compared with patients with history of stroke (mean
SCIENCES
of 6.0±4.5 visits [median 5]; P<0.0001) within the first
year after discharge.9 The difference was attributable
to an increase in cardiology visits.
Other potential explanations for the differences
observed relate to underlying disease status (CAD ver-
sus stroke). For example, there tends to be an increased
prevalence of uncontrolled hypertension in patients with
stroke. Data from REGARDS, a population-based cohort
study, suggests that the prevalence of apparent resis-
tant hypertension defined as uncontrolled blood pres-
sure with ≥3 antihypertensive medication classes (or use
of ≥4 antihypertensive medication classes regardless of
blood pressure level) was 24.9% in patients with stroke
compared to 17.0% in those without.11 Even though the
direction of causation between stroke and treatment
resistant hypertension, (ie, if the stroke was due to previ-
ous resistant hypertension or the resistant hypertension
is a result of vascular changes secondary to atheroscle-
rosis) is difficult to establish, a higher percentage of
patients with stroke seemed to have difficult to control
hypertension based on this study. Also, from these tri-
als’ databases, we were not able to distinguish 2 causes
of stroke (ischemic and cardioembolic) which may not
respond similarly to modification of all the cardiovascular
risk factors analyzed. Finally, guideline-directed thera-
pies in patients with CAD with concurrent heart failure
includes renin-angiotensin system blockers, beta block-
ers, and mineralocorticoid receptor antagonists—each of
which have blood pressure–lowering effects. This may
contribute to less severe hypertension in at least some
patients with CAD who would then be more likely to
achieve target blood pressure than in those with stroke.
The percentage of patients with control of each cardio-
vascular risk factor by cardiovascular disease group is
shown in Table 3.
Patients with more severe strokes may also be vul-
nerable to clinician implicit bias due to disability and
dependency on caregivers for aspects of daily living.
This may influence decision-making or perceived treat-
ment goals and contribute to less optimal risk factor
management in stroke patients. In the most profoundly
affected patients, institutionalization may further limit
access to aggressive outpatient care.12 In addition,
patient-centered communication that is critical to trust
and improved treatment adherence could be compro-
mised in patients with residual neurological deficits.
Other patient-related factors could also account for the
observed differences. Some patients with a history of
stroke could have accompanying cognitive impairment
either secondary to the stroke itself or prior damage
from coexisting cerebrovascular disease, making com-
pliance with medications and follow-up with physician
appointments more challenging.
August 2023   2019
 Balasubramanian et al CV Risk Factor Control in T2D Patients With Stroke

Table 3. CV Risk Factor Control by CV Disease Groups in achieve LDL cholesterol level <100 mg/dL, or adher-
CLINICAL AND POPULATION

the 3 Trials ence in the trial participants. These might have provided
CARMELINA TRIAL more insights into the severity of risk factor and provider
SCIENCES

Stroke alone, CAD alone, CAD and prescribing patterns. We did not categorize the stroke
N=576 N=3339 stroke, N=735 events as atherosclerotic/lacunar/hemorrhagic strokes.
N with BP controlled* 269 (46.7%) 1712 (51.3%) 342 (46.5%) This is a limitation of our study, but we suspect that it
(% of total)
would not make a major difference to the results of the
N with LDL-C 474 (82.3%) 2984 (89.4%) 613 (83.4%) study as ischemic stroke is much more common than
­controlled† or statin
use (% of total)
hemorrhagic strokes in patients with diabetes with only
N with nonsmoking (% 520 (90.3%) 2982 (89.3%) 653 (88.8%)
a small fraction having the latter. Finally, our definition of
of total) stroke included only those with cerebrovascular events,
N on antiplatelet or 424 (73.6%) 2914 (87.3%) 635 (86.4%) and not those with cerebrovascular disease, per se. This
anticoagulant (% of is in contrast to the definition of CAD in our study, which
total) incorporated both those with prior ischemic events and
EMPA-REG OUTCOME TRIAL interventions, but also those with more milder forms of
Stroke alone, CAD alone, CAD and obstructive CAD. Given the systemic nature of athero-
N=932 N=4776 stroke, N=698 sclerosis, it is very likely that those individuals assigned
N with BP controlled* 546 (58.6%) 3043 (63.7%) 389 (55.7%) to the CAD-only and stroke-only groups actually did
(% of total)
have disease in other vascular beds.
N with LDL-C 731 (78.4%) 4364 (91.4%) 601 (86.1%)
­controlled† or statin
use (% of total)
Conclusions
N with nonsmoking (% 831 (89.2%) 4140 (86.7%) 615 (88.1%)
of total) In summary, we found significant discordance in the man-
N on antiplatelet or 736 (79.0%) 4441 (93.0%) 629 (90.1%)
agement of cardiovascular risk factors between patients
anticoagulant (% of with stroke versus those with CAD who were enrolled in
total) clinical trials of diabetes therapy. Patients with a history
CAROLINA TRIAL of stroke, compared with CAD, had less optimal risk fac-
Stroke alone, CAD alone, CAD and tor control. The intermediate results in patients with both
N=297 N=1341 stroke, N=162 CAD and stroke suggest that these differences could
Downloaded from http://ahajournals.org by on August 27, 2023

N with BP controlled* 184 (62.0%) 795 (59.3%) 80 (49.4%) be related at least in part to clinician factors. Improving
(% of total)
clinical outcomes after stroke, particularly as regards to
N with LDL-C 233 (78.5%) 1183 (88.2%) 134 (82.7%) reducing recurrent vascular events, will require a better
­controlled† or statin
use (% of total) understanding of the reasons behind these differences.
N with nonsmoking (% 256 (86.2%) 1127 (84.0%) 142 (87.7%)
More broadly, addressing them is likely to have a benefi-
of total) cial impact on the health of patients with stroke.
N on antiplatelet or 236 (79.5%) 1179 (87.9%) 143 (88.3%)
anticoagulant (% of
total) ARTICLE INFORMATION
BP indicates blood pressure; CAD, coronary artery disease; CARMELINA, The Received November 20, 2022; final revision received June 2, 2023; accepted
Cardiovascular and Renal Microvascular Outcome Study With Linagliptin; CARO- June 12, 2023.
LINA, The Cardiovascular Outcome Study of Linagliptin vs Glimepiride in Type 2 Presented in part at the American Diabetes Association Scientific Sessions,
Diabetes; EMPA-REG OUTCOME, Empagliflozin Cardiovascular Outcome Event New Orleans, LA, June 3–7, 2022.
Trial in Type 2 Diabetes Mellitus Patients; LDL-C, low density lipoprotein -choles-
terol; and SBP, systolic blood pressure. Affiliations
*BP controlled defined as SBP <140 and diastolic blood pressure <90 mm Hg. Section of Endocrinology, Department of Medicine (P.B.), Section of General Inter-
†LDL-C controlled defined as LDL-C<100 mg/dL. nal Medicine, Department of Medicine (W.N.K), Department of Neurology (K.N.S),
and Section of Endocrinology, Department of Medicine (S.E.I), Yale School of Medi-
cine, New Haven, CT. Boehringer Ingelheim Norway KS, Asker (A.P.O.). Oslo Dia-
Limitations betes Research Center, Norway (A.P.O.). Velocity Clinical Research at Medical City,
Dallas, TX (J.R.). Würzburg University Clinic, Germany (C.W.). Lunenfeld-Tanenbaum
It may be difficult to fully generalize the findings from Research Institute, Mount Sinai Hospital, University of Toronto, ON, Canada (B.Z.).
Boehringer Ingelheim Pharma GmbH & Co KG, Germany (M.M.). Department of
these 3 cardiovascular outcome trials as their patient Internal Medicine, University Hospital; RWTH Aachen University, Germany (N.M.).
populations were dictated by the specific inclusion/
exclusion criteria in each. Accordingly, our findings only Acknowledgments
The authors meet criteria for authorship as recommended by the International
pertain to the trial populations, and it would be of inter- Committee of Medical Journal Editors (ICMJE) and were fully responsible for all
est to see how our observations fare in large-scale content and editorial decisions and were involved at all stages of the manuscript
observational datasets in T2D. We also did not analyze development. The authors thank the investigators, coordinators, and patients who
participated in the trials. Editorial assistance with preparation of figures was sup-
medication data, such as number and class of antihy- ported financially by Boehringer Ingelheim and provided by Paul Lidbury of El-
pertensives, use of high/moderate intensity statins to evate Scientific Solutions.

2020   August 2023 Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053

 Balasubramanian et al
Sources of Funding
The EMPA-REG OUTCOME (Empagliflozin Cardiovascular Outcome Event Trial
in Type 2 Diabetes Mellitus Patients), CAROLINA (The Cardiovascular Outcome
Study of Linagliptin vs Glimepiride in Type 2 Diabetes), and CARMELINA (The
Cardiovascular and Renal Microvascular Outcome Study With Linagliptin) trials
were sponsored by the Boehringer Ingelheim and Eli Lilly and Company Diabetes
Alliance.
Disclosures
Dr Sheth reports compensation from Astrocyte for consultant services; service
as President for Advanced Innovation in Medicine; grants from Hyperfine; em-
ployment by Yale School of Medicine; grants from Biogen; compensation from
Cerevasc for consultant services; a patent pending for Stroke wearables licensed
to Alva Health; compensation from CSL Behring for consultant services; com-
pensation from Sense for data and safety monitoring services; compensation
from Certus for consultant services; grants from Novartis; grants from BARD;
compensation from ZOLL Medical Corporation for data and safety monitoring
services; and compensation from Rhaeos for consultant services. Dr Rosenstock
reports grants from Novo Nordisk; compensation from Intarcia for consultant ser-
vices; travel support from Novo Nordisk; travel support from Applied Therapeu-
tics; grants from Boehringer Ingelheim; travel support from Boehringer Ingelheim;
grants from Pfizer; compensation from Oramed for consultant services; grants
from Hanmi Pharmaceutical Co, Ltd; compensation from Boehringer Ingelheim
for consultant services; compensation from Zealand for consultant services; com-
pensation from Novo Nordisk for consultant services; grants from Intarcia; travel
support from Intarcia; compensation from Eli Lilly and Company for consultant
services; travel support from Sanofi US Services, Inc; compensation from Regor
Pharmaceuticals, Inc for consultant services; compensation from Applied Thera-
peutics for consultant services; grants from Sanofi US Services, Inc; compensa-
tion from Hanmi Pharmaceutical, Co Ltd, for consultant services; grants from
Eli Lilly and Company; grants from Novartis; grants from Applied Therapeutics;
compensation from Sanofi US Services, Inc for consultant services; grants from
Oramed; and travel support from Oramed. Dr Wanner reports compensation from
Astellas Pharma for consultant services; compensation from Sanofi US Services,
Inc for consultant services; compensation from Fresenius USA, Inc for consul-
tant services; compensation from Idorsia for consultant services; compensation
from Vifor Fresenius Medical Care Renal Pharma, Ltd, for consultant services;
compensation from Eli Lilly and Company for consultant services; compensa-
Downloaded from http://ahajournals.org by on August 27, 2023
tion from Amicus Therapeutics Inc. for consultant services; compensation from
Takeda California Inc. for consultant services; compensation from Boehringer
Ingelheim for consultant services; compensation from Chiesi Farmaceutici for
consultant services; compensation from Amgen for consultant services; compen-
sation from GlaxoSmithKline for consultant services; compensation from Merck
for consultant services; compensation from AstraZeneca for consultant services;
compensation from Novo Nordisk AS for consultant services; compensation from
Bayer for consultant services; compensation from Novartis for consultant ser-
vices; and compensation from Gilead Sciences for consultant services. Dr Zin-
man has served as a consultant for Boehringer Ingelheim, Novo Nordisk and
Eli Lilly. Dr Ofstad and M. Mattheus are employees of Boehringer Ingelheim. Dr
Marx reports compensation from AstraZeneca for other services; compensation
from Sanofi Aventis for other services; compensation from Boehringer Ingelheim
for other services; compensation from Bristol-Myers Squibb for other services;
compensation from Bristol-Myers Squibb for other services; compensation from
AstraZeneca for consultant services; compensation from Novo Nordisk for other
services; compensation from Kowa Research Institute for consultant services;
compensation from Novo Nordisk for other services; compensation from Merck
Sharp and Dohme for other services; compensation from Genfit for other servic-
es; grants from Boehringer Ingelheim; compensation from Boehringer Ingelheim
for other services; compensation from Lilly Deutschland for other services; and
compensation from Amgen for other services. Dr Inzucchi has served as a con-
Stroke. 2023;54:2013–2021. DOI: 10.1161/STROKEAHA.122.042053
CV Risk Factor Control in T2D Patients With Stroke
sultant and member of clinical trial steering committees for Boehringer Ingelheim,
CLINICAL AND POPULATION
AstraZeneca, Novo Nordisk, Merck, Pfizer, and Bayer. He has delivered lectures
sponsored by Boehringer Ingelheim, AstraZeneca, and Merck. The other authors
report no conflicts.
SCIENCES
Supplemental Material
Table S1
Figures S1–S2
STROBE checklist
REFERENCES
1. Elnady HM, Mohammed GF, Elhewag HK, Mohamed MK, Borai A. Risk fac-
tors for early and late recurrent ischemic strokes. Egypt J Neurol Psychiatr
Neurosurg. 2020;56:56.
2. Hardie K, Hankey GJ, Jamrozik K, Broadhurst RJ, Anderson C. Ten-
year risk of first recurrent stroke and disability after first-ever stroke
in the Perth Community Stroke Study. Stroke. 2004;35:731–735. doi:
10.1161/01.STR.0000116183.50167.D9
3. Di Angelantonio E, Kaptoge S, Wormser D, Willeit P, Butterworth AS,
Bansal N, O’Keefe LM, Gao P, Wood AM, Burgess S, et al; Emerging Risk
Factors Collaboration. Association of cardiometabolic multimorbidity with
mortality. JAMA. 2015;314:52–60. doi: 10.1001/jama.2015.7008
4. Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, Mattheus M,
Devins T, Johansen OE, Woerle HJ, et al; EMPA-REG OUTCOME Investiga-
tors. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabe-
tes. N Engl J Med. 2015;373:2117–2128. doi: 10.1056/NEJMoa1504720
5. Rosenstock J, Kahn SE, Johansen OE, Zinman B, Espeland MA, Woerle HJ,
Pfarr E, Keller A, Mattheus M, Baanstra D, et al; CAROLINA Investigators.
Effect of linagliptin vs glimepiride on major adverse cardiovascular outcomes
in patients with type 2 diabetes: the CAROLINA randomized clinical trial:
the CAROLINA randomized clinical trial. JAMA. 2019;322:1155–1166. doi:
10.1001/jama.2019.13772
6. Hackam DG, Spence JD. Combining multiple approaches for the secondary
prevention of vascular events after stroke: a quantitative modeling study.
Stroke. 2007;38:1881–1885. doi: 10.1161/STROKEAHA.106.475525
7. Alvarez-Sabin J, Quintana M, Hernandez-Presa MA, Alvarez C, Chaves J,
Ribo M. Therapeutic interventions and success in risk factor control for sec-
ondary prevention of stroke. J Stroke Cerebrovasc Dis. 2009;18:460–465.
doi: 10.1016/j.jstrokecerebrovasdis.2009.01.014
8. Bravata DM, Daggy J, Brosch J, Sico JJ, Baye F, Myers LJ, Roumie CL,
Cheng E, Coffing J, Arling G. Comparison of risk factor control in the year
after discharge for ischemic stroke versus acute myocardial infarction.
Stroke. 2018;49:296–303. doi: 10.1161/STROKEAHA.117.017142
9. Amar J, Cambou JP, Touzé E, Bongard V, Jullien G, Vahanian A, Coppé G,
Mas J. Comparison of hypertension management after stroke and myocar-
dial infarction: results from ECLAT1--a French nationwide study: results
from ECLAT1-a French nationwide study. Stroke. 2004;35:1579–1583. doi:
10.1161/01.STR.0000131547.71502.81
10. Howard VJ, Tanner RM, Anderson A, Irvin MR, Calhoun DA, Lackland DT,
Oparil S, Muntner P. Apparent treatment-resistant hypertension among
individuals with history of stroke or transient ischemic attack. Am J Med.
2015;128:707–14.e2. doi: 10.1016/j.amjmed.2015.02.008
11. Kleina T, Horn A, Suhr R, Schaeffer D. Current status of medical care for
nursing home residents in Germany - results of an empirical study. Gesund-
heitswesen. 2017;79:382–387. doi: 10.1055/s-0035-1549971
12. Arif H, Aijaz B, Islam M, Aftab U, Kumar S, Shafqat S. Drug compliance
after stroke and myocardial infarction: a comparative study. Neurol India.
2007;55:130–135. doi: 10.4103/0028-3886.32783
August 2023   2021