Received: 22 February 2018 Revised: 10 May 2018 Accepted: 22 May 2018

DOI: 10.1111/dom.13377

ORIGINAL ARTICLE

Sex differences in management and outcomes of patients

with type 2 diabetes and cardiovascular disease: A report
from TECOS
Joakim Alfredsson MD1,2 | Jennifer B. Green MD1 | Susanna R. Stevens MS1 |
Shelby D. Reed PhD1 | Paul W. Armstrong MD3 | M. Angelyn Bethel MD4 |
Samuel S. Engel MD5 | Darren K. McGuire MD6 | Frans Van de Werf MD7 |
Irene Hramiak MD8 | Harvey D. White DSc9 | Eric D. Peterson MD1 |
Rury R. Holman FRCP4 | on behalf of the TECOS Study Group

1
Duke Clinical Research Institute, Duke University
School of Medicine, Durham, North Carolina Aim: To examine sex differences in baseline characteristics and outcomes in patients with type
2
Department of Cardiology and Department of 2 diabetes and atherosclerotic vascular disease.
Medicine and Health, Linköping University, Materials and methods: Cox models were used to analyse the association between sex and out-
Linköping, Sweden
comes in the Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), a randomized,
3
Canadian VIGOUR Centre, University of
placebo-controlled trial assessing the impact of sitagliptin on cardiovascular (CV) outcomes in
Alberta, Edmonton, Canada
4
patients with type 2 diabetes and atherosclerotic vascular disease.
Diabetes Trials Unit, Oxford Centre for
Diabetes, Endocrinology and Metabolism, Results: A total of 4297 women and 10 374 men were followed for a median of 3.0 years.
University of Oxford, Oxford, UK Women were slightly older and more often had cerebrovascular disease and peripheral arterial
5
Merck & Co., Inc, Kenilworth, New Jersey disease but less often coronary heart disease than men. At baseline, women were less likely to
6
Division of Cardiology, Department of use aspirin or statins. The primary composite outcome of CV death, myocardial infarction,
Medicine, University of Texas Southwestern stroke, or hospitalization for unstable angina occurred in 418 women (9.7%) and 1272 men
Medical Center, Dallas, Texas
7
(12.3%; 3.48 vs 4.38 events/100 participant-years, crude hazard ratio [HR] 0.79, 95% confi-
Department of Cardiovascular Sciences,
dence interval [CI] 0.71-0.89, adjusted HR 0.64, 95% CI 0.55-0.74; P < .0001). Women also had
University of Leuven, Leuven, Belgium
8 a significantly lower risk of secondary CV outcomes and all-cause death.
Division of Endocrinology and Metabolism,
University of Western Ontario, London, Canada Conclusions: In this large prospective study of people with type 2 diabetes and CV disease,
9
Auckland City Hospital, Green Lane women had different CV disease burden, worse CV risk factor profiles, and less use of indicated
Cardiovascular Service, Auckland, medications than men. Despite this, women had significantly lower risk of CV events, suggesting
New Zealand
that the cardioprotective effects of female sex extend to populations with type 2 diabetes.
Correspondence
Jennifer B. Green, MD, Duke Clinical Research
Institute, P.O. Box 17969, Durham, NC 27715. KEYWORDS
Email: jennifer.green@duke.edu cardiovascular, diabetes, sex difference, sitagliptin
Funding information
TECOS was funded by Merck Sharp & Dohme
Corp., a subsidiary of Merck & Co., Inc.,
Kenilworth, New Jersey.

1 | I N T RO D UC T I O N
Diabetes is a major health concern, affecting close to 9% of the adult
population worldwide. It has been projected that the number of
Joakim Alfredsson and Jennifer B. Green contributed equally.
people with diabetes will increase from 425 million in 2017 to nearly
630 million in 2045.1 Diabetes prevalence varies by region, being
highest in North America and the Caribbean, the Middle East and
North Africa, and Southeast Asia. The vast majority of people with
diabetes currently live in low- and middle-income countries, with the
greatest increases expected to occur in areas with economies

Diabetes Obes Metab. 2018;1–10. wileyonlinelibrary.com/journal/dom © 2018 John Wiley & Sons Ltd 1
2 ALFREDSSON ET AL.
improving to the middle range.1 The prevalence of type 2 diabetes has
expanded as a result of factors such as urbanization, obesity, physical
inactivity and ageing, with comparable disease prevalence in women
and men.1,2 Diabetes is a strong risk factor for cardiovascular disease
(CVD), and among people with type 2 diabetes, CVD is the leading
cause of death.3
Although women without diabetes have a lower risk of cardiovas-
cular (CV) events than men, this advantage is reportedly lost in women
4,5 6
with type 2 diabetes. The INTERHEART case–control study of par-
ticipants with vs without myocardial infarction (MI), comprising
>27 000 individuals from 52 countries, identified diabetes as one of
nine independent CVD risk factors, and found that a diagnosis of dia-
betes was more strongly associated with MI in women than in men.
The odds ratios for diabetes as a risk factor for a first MI were 4.26
(95% CI 3.68-4.94) in women and 2.67 (95% CI 2.43-2.94) in men,
with a statistically significant interaction of the association by sex
(P < .0001). A Framingham Heart Study analysis found that estab-
lished coronary heart disease conveyed a higher risk of CV mortality
than did diabetes in men, but that diabetes posed the greater risk in
7 first occurrence of a primary composite outcome comprising CV
women. Several studies have confirmed a higher relative risk of CVD
in women with vs without diabetes compared to men with vs without
diabetes; however, the absolute risk of CVD in women with diabetes
8 occurrence of CV death, non-fatal MI, or non-fatal stroke. Other sec-
compared to men with diabetes has reportedly been higher,
9–12 5,13,14
similar, or lower. Reported disparities in CV outcomes by
sex are incompletely understood but may be related to differences in
CV risk factor burden and management, timing of disease diagnosis,
2,15–17
or other factors. Analyses of more contemporary populations
are needed to explore these sex-related differences in the context of
current care.
The primary aim of the present analysis was to characterize over-
all differences by sex in baseline characteristics, CV risk factor man-
agement, and important CV and non-CV outcomes in patients with
type 2 diabetes and established atherosclerotic CVD (ASCVD)
enrolled in the Trial Evaluating Cardiovascular Outcomes with Sitaglip-
tin (TECOS). Whether there are sex-specific effects of treatment with
the dipeptidyl peptidase-4 (DPP-4) inhibitor sitagliptin was also
analysed.
2 | MATERIALS AND METHODS
2.1 | Study population and study medication
TECOS was an international, multicentre, randomized, placebo-con-
trolled, double-blind trial evaluating the impact of sitagliptin on CV
outcomes. The study design18 and major outcomes of the trial have
19
been published. Briefly, 14 735 people with type 2 diabetes and
established ASCVD, defined as a history of major coronary artery dis-
ease, ischaemic cerebrovascular disease, or atherosclerotic peripheral
arterial disease (PAD), were enrolled at 673 sites in 38 countries
between December 2008 and July 2012. The protocol was approved
by the ethics committees at each participating trial site, and all partici-
pants provided written informed consent to trial participation. Eligible
participants were aged ≥50 years, with a glycated haemoglobin
(HbA1c) level of 6.5% to 8.0% (48 mmol/mol to 64 mmol/mol) when
treated with stable doses of one or two oral antihyperglycaemic
agents (metformin, pioglitazone or sulphonylureas), or treated with
insulin with or without metformin. Participants were excluded if their
estimated glomerular filtration rate (eGFR) was <30 mL/min/1.73 m2
at baseline, or if they had experienced ≥2 severe hypoglycaemic epi-
sodes during the previous 12 months. Participants were assigned ran-
domly to an oral dose of sitagliptin 100 mg/d (or 50 mg/d if their
baseline eGFR was <50 mL/min/1.73 m2) or matching placebo, given
in addition to their baseline antihyperglycaemic regimen. Assigned
study medication was continued throughout follow-up, with prede-
fined dose adjustments for changes in eGFR. Glycaemia was managed
as per local/regional standards of care. Uptitration of existing antihy-
perglycaemic medications and the addition of new antihyperglycaemic
medications was permitted; however, use of open-label DPP-4 inhibi-
tors or glucagon-like peptide-1 receptor agonists was prohibited.
2.2 | Study outcomes
The primary composite outcome in TECOS was defined as the time to
death, non-fatal MI, non-fatal stroke, or hospitalization for unstable
angina (UA). The secondary composite outcome was time to the first
ondary outcomes included the occurrence of the individual compo-
nents of the composite outcomes and, in addition, hospitalization for
heart failure (HF), initiation of additional antihyperglycaemic agents,
severe hypoglycaemia, renal failure, acute pancreatitis and cancer. In
our analyses, we assessed sex differences for all of these outcomes as
well as initiation of insulin or other additional antihyperglycaemic
therapy.
An independent clinical events classification committee, blinded
to randomized treatment, adjudicated all events of death, MI, stroke,
hospitalization for UA or HF, acute pancreatitis, and cancers. The defi-
nitions used to adjudicate events have previously been listed in
detail.18 In the main trial, no significant differences in these outcomes
were found with sitagliptin treatment compared with placebo.
2.3 | Statistical analysis
Baseline characteristics are presented as median (25th, 75th percen-
tile) for continuous variables and as number (percentage) for categori-
cal variables. Women and men were compared using chi-squared tests
for categorical variables and Wilcoxon rank-sum tests for continuous
variables. Changes in laboratory and vital status measures over the
course of the study were analysed with repeated-measures analysis of
variance and presented as overall least-squares mean differences
between men and women. Models for blood pressure used the
unstructured covariance structure, while HbA1c and LDL cholesterol
models used a spatial power covariance structure.
Cox proportional hazards regression models stratified by region
were used to analyse the association between sex and primary and sec-
ondary outcomes. Data were censored at the last day the participant
was known to be free of the event in question. SAS PROC MI (SAS
Institute, Inc., Cary, North Carolina) was used to impute missing base-
line data, and this imputed dataset was used in modelling. No more
ALFREDSSON ET AL.
than 3% of data were imputed for any particular variable. Event rates
are presented as total number of events and as events per 100 partici-
pant-years of follow-up. Associations between sex and outcomes are
shown with adjusted hazard ratios (HRs) and 95% confidence intervals
(CIs). Variables were chosen for inclusion in the multivariate adjustment
model if they were selected for the CV death, MI, or stroke composite
endpoint using Cox models with backward elimination, with α = 0.05
required for retention; these baseline characteristics included age, race,
ethnicity, HbA1c, New York Heart Association class, smoking, MI,
chronic obstructive pulmonary disease, coronary artery disease, stroke,
carotid artery stenosis, atrial fibrillation or flutter, insulin use, amputa-
tions, diabetic neuropathy, foot ulcers, blood pressure, heart rate,
height, body mass index (BMI) and eGFR. Randomized treatment and
diabetes duration were forced in the model because of clinical rele-
vance, and region was included as a stratification variable for consis-
tency with the primary manuscript (Table S1). The linearity assumption
of Cox proportional hazards regression models was checked in all end-
points for the continuous variables. Where piecewise linear splines for
continuous variables were necessary, a cutpoint was selected that
would work reasonably well for all endpoints. The proportional hazards
assumption was checked for the adjustment model and no major
violations were identified. HRs for sitagliptin treatment vs placebo were
computed, including the P value for the interaction of sex and
treatment. Unadjusted Kaplan–Meier curves are provided by sex
for the primary and all-cause death outcomes. Data were analysed
with SAS version 9.4. P values <.05 were considered statistically
significant.
The trial was registered at http://www.clinicaltrials.gov:
NCT00790205.
3 | RESULTS
3.1 | Baseline characteristics
A total of 4297 women and 10 374 men were enrolled in the TECOS
intention-to-treat population. The median (interquartile range) dura-
tion of follow-up was 3.0 (2.3-3.8) years. On average, women were
slightly older (66 vs 65 years; P < .0001) but with a similar duration of
diabetes (10 years) at randomization. Men more often had prior coro-
nary heart disease (79.3% vs 61.3%; P < .0001) including prior MI,
percutaneous coronary intervention, or coronary artery bypass graft-
ing, while women more often had prior cerebrovascular disease (31.2
vs 21.7%; P < .0001), PAD (20.6% vs 14.9%; P < .0001), and HF
(21.8% vs 16.4%; P < .0001). Women more often had diabetic neu-
ropathy (27.2% vs 21.1%; P < .0001), but less often had an amputa-
tion (1.7% vs 2.9%; P < .0001). Women had higher BMI (30.5 vs
2
29.2 kg/m ), systolic blood pressure (BP; 135 vs 132 mm Hg), and
LDL cholesterol levels (2.4 vs 2.1 mmol/L), as well as worse kidney
2
function (eGFR 68 vs 74 mL/min/1.73m ), but lower prevalence of
smoking (8.5% vs 12.6%; P < .0001 for all). At baseline, women were
less likely to be treated with aspirin (72.5% vs 81.0%) or a statin
(73.0% vs 82.7%) than men (both P < .0001). Hormone replacement
therapy (HRT) was reported in 4.7% of women. The proportions of
women and men enrolled varied by region. The greatest proportion of
3
women were enrolled in Eastern Europe (36.7% of all women
enrolled), while the greatest proportion of men were enrolled in the
region categorized as Asia Pacific and other (33.5% of all men
enrolled). Detailed baseline characteristics by sex are presented in
Table 1. Over the follow-up period, women had slightly higher mean
HbA1c (estimated overall mean difference 0.64 mmol/mol), systolic
BP (estimated overall mean difference 1.57 mm Hg), diastolic BP (esti-
mated overall mean difference 0.87 mm Hg), and LDL cholesterol
(estimated overall mean difference 0.35 mmol/L; all unadjusted
P values from repeated-measures analysis of variance <.0001). The
between-sex differences in aspirin and statin use noted at baseline
persisted throughout the trial.
3.2 | Association between sex and outcomes,
independent of randomized treatment
The primary composite outcome occurred in 418 of 4297 women
(9.7%) and 1272 of 10 374 men (12.3%), corresponding to 3.48 vs
4.38 events per 100 participant-years (crude HR 0.79, 95% CI
0.71-0.89; P < .0001). The association of lower risk for women vs
men for this outcome was strengthened after adjustment for differ-
ences in baseline characteristics (HR 0.64, 95% CI 0.55-0.74;
P < .0001 [Figure 1]). Differences between women and men for fatal
outcomes were similar, with women having a lower incidence of CV
death (1.55 vs 1.76 events/100 participant-years [adjusted HR 0.56,
95% CI 0.45-0.70; P < 0.0001]) and all-cause death (2.17 vs 2.58
events/100 participant-years [adjusted HR 0.56, 95% CI 0.47-0.67;
P < .0001]; Figure 1). The unadjusted Kaplan–Meier rates over
48 months for the primary and mortality endpoints are shown in
Figure 2. Among the secondary CV outcomes, women had lower
event rates for the composite endpoint of CV death, MI and stroke
(P < .0001), as well as the individual outcomes of MI (P = .0047),
stroke (P < .0001), or hospitalization for HF (P = .0164). Rates of hos-
pitalization for UA did not differ significantly between men and
women (Figure 1).
Post hoc sensitivity analyses restricted to the 10 863 patients
with prior coronary heart disease identified persistent differences in
the incidence of the primary composite CV outcome in women vs
men, corresponding to 4.09 vs 4.52 events per 100 participant-years
(adjusted HR 0.71, 95% CI 0.60-0.84; P < .0001), and the composite
endpoint of CV death, MI and stroke: 3.48 vs 3.88 events per 100 par-
ticipant-years (adjusted HR 0.67, 95% CI 0.56-0.81; P < .0001). In
women compared with men in this cohort, there was a lower inci-
dence of CV death (1.58 vs 1.74 events/100 participant-years
[adjusted HR 0.55, 95% CI 0.42-0.71; P < .0001]) and all-cause death
(2.18 vs 2.56 events/100 participant-years [adjusted HR 0.54, 95% CI
0.43-0.67; P < .0001]).
There was no statistical difference in initiation of insulin or any
co-interventional anti-hyperglycaemic therapy between women and
men, or in the overall incidence of severe hypoglycaemia. Finally, can-
cer (adjusted HR 0.74, 95% CI 0.57-0.97; P = .0280) and renal failure
(adjusted HR 0.49; 95% CI 0.32-0.74; P = .0008) occurred less often
in women (Figure 1).
4 ALFREDSSON ET AL.

TABLE 1 Baseline characteristics according to sex

Characteristic Female (n = 4297) Male (n = 10 374) Pa

Age, y 66 (60, 72) 65 (59, 71) <.0001
Race, n (%) <.0001
White 2896 (67.4) 7061 (68.1)
Black 224 (5.2) 223 (2.1)
Asian 824 (19.2) 2441 (23.5)
Other 353 (8.2) 649 (6.3)
Hispanic or Latino, n (%) 648 (15.1) 1150 (11.1) <.0001
Region, n (%) <.0001
Eastern Europe 1578 (36.7) 2387 (23.0)
Latin America 557 (13.0) 914 (8.8)
North America 667 (15.5) 1927 (18.6)
Western Europe 405 (9.4) 1671 (16.1)
Asia Pacific and other 1090 (25.4) 3475 (33.5)
BMI, kg/m2 30.5 (26.7, 34.8) 29.2 (26.1, 32.7) <.0001
Height, cm 158 (153, 163) 171 (166, 177) <.0001
Systolic BP, mm hg 135 (125, 147) 132 (122, 144) <.0001
Systolic BP < 140 mm Hg, n (%) 2400 (56.0) 6415 (62.0) <.0001
Diastolic BP, mm Hg 80 (70, 85) 78 (70, 83) <.0001
Heart rate, bpm 72 (66, 80) 72 (64, 78) <.0001
Baseline eGFR, mL/min/1.73m2 68 (57, 84) 74 (60, 90) <.0001
eGFR <60 mL/min/1.73m2, n (%) 1255 (29.5) 2069 (20.1) <.0001
Urinary albumin: creatinine ratio, g/Mol 1.11 (0.38, 3.69) 1.20 (0.41, 3.85) .0068
Total cholesterol, mmol/L 4.51 (3.80, 5.40) 3.96 (3.39, 4.70) <.0001
LDL cholesterol, mmol/L 2.39 (1.81, 3.16) 2.10 (1.63, 2.69) <.0001
HDL cholesterol, mmol/L 1.20 (1.01, 1.43) 1.04 (0.88, 1.22) <.0001
Triglycerides, mmol/L 1.66 (1.21, 2.31) 1.59 (1.13, 2.23) <.0001
Qualifying HbA1c, mmol/Mol 55.2 (50.8, 60.7) 55.2 (50.8, 59.6) .0078
Duration of diabetes, y 10.0 (5.5, 16.0) 10.0 (5.0, 16.0) .5805
Prior CVD (coronary), n (%) 2634 (61.3) 8229 (79.3) <.0001
≥50% coronary stenosis 1714 (39.9) 5973 (57.6) <.0001
Myocardial infarction 1513 (35.2) 4742 (45.7) <.0001
Prior PCI 1350 (31.8) 4364 (42.7) <.0001
Prior CABG 611 (14.2) 3053 (29.4) <.0001
Prior cerebrovascular disease, n (%) 1342 (31.2) 2246 (21.7) <.0001
Prior peripheral arterial disease, n (%) 885 (20.6) 1548 (14.9) <.0001
Prior congestive HF, n (%) 937 (21.8) 1706 (16.4) <.0001
History of hypertension, n (%) 3881 (90.3) 8767 (84.5) <.0001
Hyperlipidaemia/dyslipidaemia, n (%) 3168 (73.7) 8072 (77.8) <.0001
Atrial fibrillation/atrial flutter, n (%) 300 (7.0) 867 (8.4) .0051
COPD, n (%) 360 (8.4) 757 (7.3) .0247
NYHA class 3 or higher, n (%) 153 (3.7) 220 (2.2) <.0001
Cigarette smoking, n (%) <.0001
Never smoked 3135 (73.0) 4014 (38.7)
Current smoker 367 (8.5) 1311 (12.6)
Prior smoker 795 (18.5) 5049 (48.7)
Amputation, n (%) 71 (1.7) 306 (2.9) <.0001
Diabetic neuropathy, n (%) 1169 (27.2) 2185 (21.1) <.0001
Foot ulcer, n (%) 97 (2.3) 296 (2.9) .0419
Retinopathy, n (%) 588 (13.7) 1276 (12.3) .0221
Blindness, n (%) 65 (1.5) 170 (1.6) .5800
Cancer within past 5 y (excluding non-melanoma skin 59 (1.4) 268 (2.6) <.0001
cancer), n (%)
ALFREDSSON ET AL. 5

TABLE 1 (Continued)

Characteristic Female (n = 4297) Male (n = 10 374) Pa

Antidiabetic medication, n (%)
Metformin 3426 (79.7) 8540 (82.3) .0002
Sulphonylurea 1885 (43.9) 4760 (45.9) .0256
Thiazolidinedione 97 (2.3) 299 (2.9) .0336
Insulin 1047 (24.4) 2361 (22.8) .0359
Antihypertensive, n (%)
β blocker 2642 (61.5) 6680 (64.4) .0009
ACE inhibitor or ARB 3397 (79.1) 8158 (78.6) .5748
Calcium channel blocker 1627 (37.9) 3334 (32.1) <.0001
Diuretic 2018 (47.0) 4002 (38.6) <.0001
Antiplatelet therapy, n (%)
Aspirin 3114 (72.5) 8404 (81.0) <.0001
Other antiplatelet 770 (17.9) 2417 (23.3) <.0001
Baseline vitamin K antagonist (eg, warfarin/coumarol), 244 (5.7) 756 (7.3) .0004
n (%)
Lipid lowering therapy, n (%)
Statin 3136 (73.0) 8583 (82.7) <.0001
Fibrate 214 (5.0) 729 (7.0) <.0001
Niacin 45 (1.0) 242 (2.3) <.0001
Ezetimibe 205 (4.8) 556 (5.4) .1433

Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; BP, blood pressure; CABG, coronary artery
bypass grafting; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; HbA1c, glycated
haemoglobin; HF, heart failure; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; TIA, transient ischaemic attack. Data are
median (25th, 75th percentile) unless otherwise indicated.
a
P values from Wilcoxon rank sum or chi-squared tests.

Event Events per 100 participant-years Hazard Ratio (95% CI) P-value
Female Male

CV death, MI, stroke, or hospitalization for UA 3.48 4.38 0.64 (0.55, 0.74) <.0001
CV death, MI, or stroke 3.06 3.82 0.62 (0.53, 0.72) <.0001
CV Death 1.55 1.76 0.56 (0.45, 0.70) <.0001
MI 1.11 1.61 0.70 (0.55, 0.90) 0.0047
Stroke 0.78 0.88 0.52 (0.38, 0.71) <.0001
Hospitalization for UA 0.47 0.62 0.82 (0.55, 1.21) 0.3177
All Cause Death 2.17 2.58 0.56 (0.47, 0.67) <.0001
Hospitalization for HF 0.93 1.14 0.71 (0.53, 0.94) 0.0164
Severe Hypoglycaemia 0.92 0.66 1.05 (0.76, 1.45) 0.7751
Pancreatitis 0.08 0.08 1.03 (0.39, 2.73) 0.9544
Cancer 0.96 1.45 0.74 (0.57, 0.97) 0.0280
Initiate Co-interventional Therapy 9.59 10.19 0.94 (0.85, 1.04) 0.2346
Initiate Insulin 4.24 4.09 0.96 (0.81, 1.13) 0.5970
Renal failure 0.38 0.56 0.49 (0.32, 0.74) 0.0008

0.25 0.5 1 2
Favours Female Favours Male

FIGURE 1 Association of sex and outcomes. Hazard ratios were calculated from adjusted Cox proportional hazards regression models, as
described in the Methods section. CV, cardiovascular; HF, heart failure; MI, myocardial infarction; UA, unstable angina

3.3 | Association between sex and effects of
randomized treatment
There was no statistical difference in the primary composite outcome
with sitagliptin treatment compared with placebo in men (4.35 vs 4.41
events/100 participant-years [HR 1.00, 95% CI 0.90-1.12]) or women
(3.37 vs 3.58 events/100 participant-years [HR 0.95, 95% CI 0.78-1.15;
interaction P value = .64]). Similarly, there were no statistical treatment
differences for the combined outcome of CV death, MI or stroke in
men (3.83 vs 3.81 events/100 participant-years [HR 1.02, 95% CI
0.90-1.14]) or women (2.97 vs 3.16 events/100 participant-years
[HR 0.94, 95% CI 0.77-1.16; interaction P value = .53]) or the individual
outcomes of all-cause mortality, CV death, MI, stroke, or hospitalization
for UA or HF for participants treated with sitagliptin compared with
placebo (all interaction P values >.10). Sitagliptin treatment was not
associated with severe hypoglycaemia, cancer, or kidney failure in men
6 ALFREDSSON ET AL.

(A)
20
Male
Female

Primary cardiovascular endpoint (%)

16

12

4 Unadjusted HR (95% CI)

for female vs male =
0.79 (0.71, 0.89)
P < 0.0001
0
0 12 24 36 48
Months
At risk
Male 10374 9551 8958 4702 1752
Female 4297 3977 3720 1916 730

(B)
20
Male
Female

16
All-cause mortality (%)

12

4 Unadjusted HR (95% CI)

for female vs male =
0.80 (0.69, 0.91)
P = 0.0012
0
0 12 24 36 48
Months
At risk
Male 10374 10036 9732 5267 2013
Female 4297 4165 4036 2145 834

FIGURE 2 Kaplan–Meier curves by sex are for the A, primary composite cardiovascular and B, mortality endpoints. HR, hazard ratio

or women. The incidence of acute pancreatitis was low in both sexes
but was associated with sitagliptin treatment compared with placebo in
men (0.13 vs 0.03 events/100 participant-years [HR 4.03, 95% CI
1.51-10.76]) but not in women (0.05 vs 0.11 events/100 “patient-years”
[HR 0.46, 95% CI 0.12-1.79; interaction P value = .011]). Both sexes
had reduced initiation of additional antihyperglycaemic therapy with
sitagliptin treatment compared with placebo (Figure S1).
4 | DISCUSSION
In the large TECOS clinical trial of people with type 2 diabetes and
established ASCVD, no significant differences in CV outcomes were
found with sitagliptin therapy compared with placebo in the setting of
usual diabetes care. However, in this predefined subgroup analysis com-
paring men and women, we found significantly lower rates of incident
major CV events in women compared with men, despite women having
worse baseline CV risk factor profiles and less use of statins and antith-
rombotic agents. Women had lower rates of all CV outcomes assessed,
other than hospitalization for UA; importantly, significantly lower rates
for the key outcomes of CV and all-cause death were found. Moreover,
the magnitude of the estimated associations between female sex and
lower CV event rates was substantially increased by adjustment for dif-
ferences in baseline characteristics. No heterogeneity of efficacy
between sex and effects of sitagliptin treatment was identified.
ALFREDSSON ET AL.
Women enrolled in TECOS were slightly older and more often
had a history of cerebrovascular disease, PAD and congestive HF than
men, while men more often had a history of coronary disease. Women
had higher BMI, systolic BP and LDL cholesterol levels, and worse
kidney function. Women had lower rates of smoking compared with
men and a similar duration of diabetes. Women were less likely than
men to be prescribed aspirin or statin therapy at baseline. These base-
line differences are in accordance with earlier reports.20
4.1 | Diabetes as a CV risk factor in men and women
There is a general notion that diabetes is more deleterious as a CV risk
factor for women than for men, and that the female advantage regard-
ing CV risk among patients without diabetes is lost in the context of
type 2 diabetes.2,9–12 Higher risk of fatal CV events in women with
diabetes compared with those without diabetes is supported by two
previous meta-analyses.8,12 In addition, the Euro Heart Survey of
Acute Coronary Syndromes reported that, after multivariable adjust-
ment for baseline differences, women with MI and diabetes had
higher in-hospital mortality compared with women without diabetes,
in contrast to the findings in men with diabetes vs those without
diabetes;20 however, these reports do not address between-sex
differences in people with diabetes or, most relevant to the present
analyses, people with diabetes plus established ASCVD, instead focus-
ing on risk differences between people with and without diabetes in
analyses stratified by sex.
4.2 | Between-sex differences in CVD
Diabetes increases the risk of CVD, but the more pronounced
increase in relative risk associated with diabetes for women compared
with men appears in part to be reflective of the lower risk in women
compared with men among those without diabetes. This was illus-
trated by a previous study from the First National Health and Nutri-
tion Examination Survey (NHANES), reporting changes in heart
disease and coronary heart disease mortality rates in individuals with
and without diabetes in 1982 to 1984 compared with 1971 to 1975.
While mortality rates declined in men with and without diabetes, age-
adjusted mortality rates increased in women with diabetes; however,
absolute mortality rates were higher in men at both time points and
all ages.13 A large study of 158 426 people with diabetes and
314 115 people without diabetes found that diabetes was associated
with significantly higher CV risk, with an almost doubled incidence of
coronary heart disease, stroke, HF and all-cause mortality found in
people with vs without diabetes of both sexes. 14
After adjustment for
age, number of concomitant medications and prior CVD, the differ-
ences were attenuated but women still had a 10% higher risk of coro-
nary heart disease and a 19% higher risk of all-cause mortality
associated with diabetes. Compatible with the present findings, the
incidence of coronary heart disease in those with diabetes was twice
as high in men compared to women.
In the TECOS population of people with type 2 diabetes and con-
comitant ASCVD we found a 44% significantly lower risk of CV death
in women compared with men. In line with our findings, a study of
>50 000 men and women (1108 participants with diabetes) found
7
that the risk of coronary heart disease mortality in patients with both
diabetes and a history of MI was substantially lower in women than in
men, with women who had neither diabetes nor prior MI as compara-
tor (adjusted HR 8.96, 95% CI 5.52-14.55, and 18.35, 95% CI
13.31-25.29, respectively).5 The TECOS data also suggest a possible
cardioprotective effect of female sex in a contemporary population
with type 2 diabetes and established CVD.
4.3 | ASCVD complications in TECOS
Types of established ASCVD at baseline differed between sexes in
TECOS, with stroke and PAD being more frequent in women and cor-
onary heart disease more frequent in men, which may have contrib-
uted to the subsequently lower risk of CV events in women.
However, the magnitude of association between diabetes and stroke
risk is greater in women compared with men,21 and no apparent sex
difference in long-term survival in patients with both stroke and
diabetes was observed in the Swedish MONICA study22 or in a cohort
where outcomes were adjusted for multiple potential confounders.23
Of note, the incidence of new stroke was also lower in women than
men in TECOS. It is therefore unlikely that prior stroke is a major
explanatory factor for the observed differences between men and
women in our study.
Generally, PAD is regarded as a coronary heart disease equivalent
with regard to its prognostic importance for CV outcomes. In the
Dutch Bypass Oral anticoagulants or Aspirin Study of 313 men and
169 women with PAD, 44% of the women and 31% of the men had a
CV event (vascular death, MI, stroke or major amputation) during
5 years of follow-up.24 A recent Danish study that included 4945
women and 6289 men undergoing peripheral vascular surgery, with
optimal medical therapy over >5 years' follow-up, found that the inci-
dence of the composite of death/MI/stroke was marginally higher in
men (adjusted HR 1.10, 95% CI 1.05-1.16).25 Taking these results
together, we do not believe that differences in the prevalence of PAD
explains the substantially lower CV event rate in women in TECOS. In
addition, the sensitivity analyses performed exclusively in the partici-
pants with prior coronary heart disease also found significantly lower
risks of major CV outcomes and death in women compared with men.
4.4 | Regional differences
The percentages of participants enrolled from various regions around
the world did vary between women and men; therefore, regional dif-
ferences in participant characteristics, risk factor management, or
prognosis may have contributed to the between-sex differences in
outcomes. Cox proportional hazards regression models were stratified
by region in the present analyses; however, this approach may not
have fully compensated for important regional disparities in care. A
previous report from the TECOS trial found significant differences in
achievement of overall secondary prevention treatment targets
between regions.26 The disparities in risk factor management between
women and men in TECOS, however, do not suggest that superior
care of women in a particular region would explain the overall reduced
risk of serious CV outcomes.
8 ALFREDSSON ET AL.

4.5 | CV risk factor burden delivery as well as by the menopausal stage in which therapy is used.
The information collected in TECOS does not permit an in-depth anal-
Higher burden and less optimal management of modifiable CV risk
ysis of the effects of HRT on CV risk, or otherwise provide a biological
factors in women with diabetes compared with their male counter-
explanation for the between-sex differences which were found.
parts have repeatedly been shown in registry data from Europe and
In summary, the present analysis of the large contemporary popu-
the United States, and may have an impact on subsequent CV event
lation with type 2 diabetes and established CVD enrolled in TECOS
rates.16,27 The majority of these datasets, however, are >10 years old
identified significant differences in baseline CV risk factor manage-
and do not reflect the implementation of more recent treatment rec-
ment between women and men. Despite this, women had significantly
ommendations in women and the elderly. In the present analyses, we
lower rates of important new CV events compared with men during a
observed unfavourable differences in BMI, blood pressure, LDL cho-
median follow-up of 3 years. These findings suggest that the cardio-
lesterol, and kidney function in women compared with men; however,
protective effect of female sex may be maintained despite a diagnosis
despite these differences in characteristics and lower rates of statin
of diabetes and the presence of ASCVD. Analyses of sex-related out-
and antithrombotic use, women experienced lower rates of major CV
comes in similar recently completed or ongoing trials will be critical to
events than men. As all enrolled participants had established ASCVD,
confirm these findings, and to further explore this clinical paradox in
it is unlikely that women had significantly different indications for
secondary prevention of CV complications. Treatment with sitagliptin
these therapies than men. This suggests that with equal risk factor
in addition to usual care had a similar safety profile in women and
management, CV outcomes may have been even lower in women.
men. It is important to recognize and address continued sex-related
Initiatives to promote adoption of guideline-based CV care in women
disparities in care and outcomes by sex, as women with diabetes are
with diabetes, and to enroll such patients in CV outcomes trials, are
likely to derive additional benefit from further optimization of CV risk
needed.
factor management. Enhanced efforts to enroll women in CV out-
In the publication of the primary TECOS results, no difference in
comes trials are needed in order to effectively individualize and apply
the risk of major adverse CV events with sitagliptin compared with
CV risk reduction strategies.
placebo was reported.19 In the present, more comprehensive analyses
of outcomes by sex, we have again found no suggestion of heteroge-
neity of the effects of sitagliptin therapy vs placebo on major CV out- ACKNOWLEDGMENTS
comes, including CV death, MI, stroke, all-cause death, or
Peter Hoffmann, an employee of the Duke Clinical Research Institute
hospitalization for HF or UA. The observation that sitagliptin vs pla-
(DCRI), provided editorial support.
cebo had no significant effect on major adverse CV outcomes is con-
R.R.H. is a National Institute for Health Research Senior Investiga-
sistent with findings from trials of other DPP-4 inhibitors, including
tor. The TECOS trial was funded by Merck Sharp & Dohme Corp., a
saxagliptin and alogliptin.28,29 However, no detailed sex-specific data
subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA. The study was
have been published from those trials. The present study provides
designed and run independently by the DCRI and the University of
important evidence that the DPP-4 inhibitor sitagliptin has a similar
Oxford Diabetes Trials Unit (DTU) in an academic collaboration with
safety profile in women and men. Assessment for sex differences in
the sponsor. All analyses were performed by DCRI and DTU indepen-
effects of therapies is advocated by many groups, including the US
dent of the sponsor. The authors are solely responsible for the design
Food and Drug Administration.30
and conduct of this study, all analyses, the drafting and editing of the
There are a number of limitations to these analyses. First, the
paper, and its final contents. All authors agreed to submit the report
TECOS study population is from a randomized clinical trial population
for publication; the funder had no role in this decision.
with inclusion and exclusion criteria contributing to differences rela-
tive to patients in routine clinical practice. In particular, the eligibility
criteria resulted in enrolment of patients with established CV disease
Conflicts of interest
and only modest hyperglycaemia and relatively preserved kidney func- J.A. has received consulting fees from Bristol-Myers Squibb, Sanofi-
tion at baseline. Our results are therefore addressing secondary CV Aventis, Eli Lilly and Bayer, lecture fees from AstraZeneca, Novartis,
risk and may not be applicable to all routinely encountered types of Merck, Sharp & Dohme and Sanofi-Aventis, and research funding
patients with diabetes; however, the TECOS trial has enhanced clinical from AstraZeneca. J.B.G. has received grants from Merck, AstraZe-
applicability as a result of the trial's pragmatic design, including inte- neca, and GlaxoSmithKline, grants and personal fees from Merck,
gration into usual care practices in 38 countries worldwide. In addi- other support from Boehringer-Ingelheim, and personal fees from
tion, far fewer women than men were enrolled, which somewhat Bioscientifica and The Endocrine Society. S.R.S. has no disclosures to
compromises the ability to make between-sex comparisons. Limited report. S.D.R. has received research support from Merck, AstraZeneca,
information pertaining to treatment for acute CVD events is available, Sanofi US, and Janssen Pharmaceuticals, Inc. P.W.A. has received
and differences in such care may not have been comparable between grants, personal fees and non-financial support from Merck and grants
men and women. Disparities in acute CVD event management, how- from AstraZeneca. M.A.B. has received grants, personal fees, and
ever, seem unlikely to explain the observed reductions in incidence of other support from Merck, other support from Boehringer-Ingelheim,
multiple different CV event types in women. Some women reported Novo Nordisk, Glaxo and Smith Kline, and non-financial support from
using HRT at baseline. The impact of such therapy on CV event rates Bayer. S.S.E. is an employee of Merck & Co., Inc. D.K.M. has received
is not well understood, and may vary by HRT type and route of personal fees from Boehringer-Ingelheim, Janssen Research and
ALFREDSSON ET AL.
Development LLC, Sanofi-Aventis Group, Merck, Lilly USA, Novo
Nordisk, GlaxoSmithKline, AstraZeneca, Merck, Eisai Pharmaceu-
ticals, Lexicon, University of Oxford, Duke Clinical Research
Institute, Partners Healthcare, and the Cleveland Clinic Founda-
tion. F.V.d.W. has received study grants and personal fees from
Merck, AstraZeneca and Boehringer-Ingelheim. I.H. has received
personal fees from Amgen, AstraZeneca/Bristol-Myers Squibb,
Boehringer Ingelheim, Eli Lilly, GlaxoSmithKline, Hoffmann-La
Roche, Insulet, Janssen-Ortho (JNJ), Merck Frosst, Novo Nordisk,
Sanofi and Takeda, and research support from AstraZeneca/Bristol-
Myers Squibb, Eli Lilly, GlaxoSmithKline, Janssen-Ortho (JNJ), Lexicon,
Medtronic, Merck Frosst and Novo Nordisk. H.D.W. reports grants
from Sanofi Aventis, Eli Lilly, National Institutes of Health, Merck
Sharp & Dohme, AstraZeneca, Omthera Pharmaceuticals, Pfiser,
Elsai Inc., DalGen Products and Services, and personal fees from
AstraZeneca. E.D.P. has received grants and personal fees from
Janssen, grants from Eli Lilly, and personal fees from AstraZeneca,
Bayer and Sanofi. R.R.H. has received grants and personal fees from
Merck, grants from Bayer, AstraZeneca and Bristol-Myers Squibb,
personal fees from Amgen, Bayer, Intarcia, Novartis, Novo Nordisk,
and other support from GlaxoSmithKline, Janssen, and Takeda.
Author contributions
J.A. and J.B.G contributed to the study design and data analysis
and interpretation, and drafted and edited the manuscript.
S.R.S. performed the statistical analysis, provided data tables and
figures, and edited the manuscript. S.D.R. contributed to the study
design and data analysis and interpretation, and edited the manu-
script. P.W.A., M.A.B., S.S.E., D.K.M., F.V.d.W., I.H. and H.D.W. con-
tributed to the data analysis and interpretation and edited the
manuscript. E.D.P. and R.R.H. contributed to the study design and
data analysis and interpretation and edited the manuscript. All authors
reviewed and approved the final submitted report. E.D.P. and
R.R.H. are the guarantors of this work and, as such, had full access
to all of the data in the study and take full responsibility for the
work as a whole, including the study design and integrity of
the data.
Prior presentation
Parts of this work were presented at the 2016 meeting of the
European Association for the Study of Diabetes.
ORCID
Jennifer B. Green http://orcid.org/0000-0002-9967-5352
Shelby D. Reed http://orcid.org/0000-0002-7654-4464
Samuel S. Engel http://orcid.org/0000-0002-4439-6356
Rury R. Holman http://orcid.org/0000-0002-1256-874X
RE FE R ENC E S
1. International Diabetes Federation. IDF Diabetes Atlas. 8th ed.; 2017
http://www.diabetesatlas.org. Accessed April 26, 2018.
2. Regensteiner JG, Golden S, Huebschmann AG, et al. Sex differences in
the cardiovascular consequences of diabetes mellitus: a scientific
9
statement from the American Heart Association. Circulation. 2015;
132:2424-2447.
3. Morrish NJ, Wang SL, Stevens LK, Fuller JH, Keen H. Mortality and
causes of death in the WHO multinational study of vascular disease in
diabetes. Diabetologia. 2001;44(suppl 2):S14-S21.
4. Kannel WB, Wilson PW. Risk factors that attenuate the female coro-
nary disease advantage. Arch Intern Med. 1995;155:57-61.
5. Hu G, Jousilahti P, Qiao Q, Katoh S, Tuomilehto J. Sex differences in
cardiovascular and total mortality among diabetic and non-diabetic
individuals with or without history of myocardial infarction. Diabetolo-
gia. 2005;48:856-861.
6. Anand SS, Islam S, Rosengren A, et al. Risk factors for myocardial
infarction in women and men: insights from the INTERHEART study.
Eur Heart J. 2008;29:932-940.
7. Natarajan S, Liao Y, Cao G, Lipsitz SR, McGee DL. Sex differences in
risk for coronary heart disease mortality associated with diabetes and
established coronary heart disease. Arch Intern Med. 2003;163:
1735-1740.
8. Huxley R, Barzi F, Woodward M. Excess risk of fatal coronary heart
disease associated with diabetes in men and women: meta-analysis of
37 prospective cohort studies. BMJ. 2006;332:73-78.
9. Roche MM, Wang PP. Sex differences in all-cause and cardiovascular
mortality, hospitalization for individuals with and without diabetes,
and patients with diabetes diagnosed early and late. Diabetes Care.
2013;36:2582-2590.
10. Kalyani RR, Lazo M, Ouyang P, et al. Sex differences in diabetes and
risk of incident coronary artery disease in healthy young and
middle-aged adults. Diabetes Care. 2014;37:830-838.
11. Juutilainen A, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Similarity of
the impact of type 1 and type 2 diabetes on cardiovascular mortality
in middle-aged subjects. Diabetes Care. 2008;31:714-719.
12. Peters SA, Huxley RR, Woodward M. Diabetes as risk factor for inci-
dent coronary heart disease in women compared with men: a system-
atic review and meta-analysis of 64 cohorts including 858,507
individuals and 28,203 coronary events. Diabetologia. 2014;57:
1542-1551.
13. Gu K, Cowie CC, Harris MI. Diabetes and decline in heart disease mor-
tality in US adults. JAMA. 1999;281:1291-1297.
14. Baviera M, Santalucia P, Cortesi L, et al. Sex differences in cardiovas-
cular outcomes, pharmacological treatments and indicators of care in
patients with newly diagnosed diabetes: analyses on administrative
database. Eur J Intern Med. 2014;25:270-275.
15. Wexler DJ, Grant RW, Meigs JB, Nathan DM, Cagliero E. Sex dispar-
ities in treatment of cardiac risk factors in patients with type 2 diabe-
tes. Diabetes Care. 2005;28:514-520.
16. Norhammar A, Stenestrand U, Lindback J, Wallentin L. Women youn-
ger than 65 years with diabetes mellitus are a high-risk group after
myocardial infarction: a report from the Swedish Register of Informa-
tion and Knowledge about Swedish Heart Intensive Care Admission
(RIKS-HIA). Heart. 2008;94:1565-1570.
17. Strom Williams JL, Lynch CP, Winchester R, Thomas L, Keith B,
Egede LE. Gender differences in composite control of cardiovascular
risk factors among patients with type 2 diabetes. Diabetes Technol
Ther. 2014;16:421-427.
18. Green JB, Bethel MA, Paul SK, et al. Rationale, design, and organiza-
tion of a randomized, controlled Trial Evaluating Cardiovascular Out-
comes with Sitagliptin (TECOS) in patients with type 2 diabetes and
established cardiovascular disease. Am Heart J. 2013;166:983-989.e7.
19. Green JB, Bethel MA, Armstrong PW, et al. Effect of sitagliptin on car-
diovascular outcomes in type 2 diabetes. N Engl J Med. 2015;373:
232-242.
20. Dotevall A, Hasdai D, Wallentin L, Battler A, Rosengren A. Diabetes
mellitus: clinical presentation and outcome in men and women with
acute coronary syndromes. Data from the euro heart survey ACS.
Diabet Med. 2005;22:1542-1550.
21. Peters SA, Huxley RR, Woodward M. Diabetes as a risk factor for
stroke in women compared with men: a systematic review and
meta-analysis of 64 cohorts, including 775,385 individuals and 12,539
strokes. Lancet. 2014;383:1973-1980.
22. Eriksson M, Carlberg B, Eliasson M. The disparity in long-term sur-
vival after a first stroke in patients with and without diabetes
10 ALFREDSSON ET AL.

persists: the northern Sweden MONICA study. Cerebrovasc Dis. 29. White WB, Cannon CP, Heller SR, et al. Alogliptin after acute coronary
2012;34:153-160. syndrome in patients with type 2 diabetes. N Engl J Med. 2013;369:
23. Zhao W, An Z, Hong Y, et al. Sex differences in long-term outcomes 1327-1335.
among acute ischemic stroke patients with diabetes in China. Biol Sex 30. U.S. Food and Drug Administration. FDA research, policy, and work-
Differ. 2015;6:29. shops on women in clinical trials. www.fda.gov/ScienceResearch/
24. Wisman PP, Tangelder MJ, van Hattum ES, de Borst GJ, Moll FL. SpecialTopics/WomensHealthResearch/ucm131731.htm. Accessed
Young women with PAD are at high risk of cardiovascular complica- August 22, 2017.
tions. Eur J Vasc Endovasc Surg. 2012;43:441-445.
25. Budtz-Lilly JW, Petersen CN, Pedersen TF, Eldrup N. Male sex associ-
SUP P O R T I NG I NFO R M A T I O N
ated with increased long-term cardiovascular mortality after periph-
eral vascular surgery for atherosclerosis despite optimal medical
treatment. Eur J Vasc Endovasc Surg. 2015;50:767-773. Additional supporting information may be found online in the Sup-
26. Bethel MA, Green JB, Milton J, et al. Regional, age and sex differences porting Information section at the end of the article.
in baseline characteristics of patients enrolled in the Trial Evaluating
Cardiovascular Outcomes with Sitagliptin (TECOS). Diabetes Obes
Metab. 2015;17:395-402.
How to cite this article: Alfredsson J, Green JB, Stevens SR,
27. Ferrara A, Mangione CM, Kim C, et al. Sex disparities in control and
treatment of modifiable cardiovascular disease risk factors among et al. Sex differences in management and outcomes of patients
patients with diabetes: Translating Research Into Action for Diabetes with type 2 diabetes and cardiovascular disease: A report from
(TRIAD) study. Diabetes Care. 2008;31:69-74. TECOS. Diabetes Obes Metab. 2018;1–10. https://doi.org/10.
28. Scirica BM, Bhatt DL, Braunwald E, et al. Saxagliptin and cardiovascular
1111/dom.13377
outcomes in patients with type 2 diabetes mellitus. N Engl J Med. 2013;
369:1317-1326.