Diabetes Epidemiology and Management 12 (2023) 100160
Contents lists available at ScienceDirect
Diabetes Epidemiology and Management
journal homepage: www.elsevier.comOriginal articleRelationships of blood pressure and control with microvascular
dysfunction in type 2 diabetesCharles F. Hayfron-Benjamina,b,c,*, Theresa Ruby Quartey-Papa cfio , Tracy Amo-Nyarkob,
Ewuradwoa A Antwid, Patience Vormatue, Melody Kwatemah Agyei-Fedieleyb,f,
Kwaku Amponsah Obengb,c
a Departments of Internal Medicine, Vascular Medicine, and Respiratory Medicine, Amsterdam UMC, University of Amsterdam, Cardiovascular Sciences, Amster-
dam, the Netherlands
b Department of Physiology, University of Ghana Medical School, Ghana
c Department of Anesthesia and Critical Care, Korle Bu Teaching Hospital and University of Ghana Medical School, Ghana
d Department of Medicine, 37 Military Hospital, Ghana
e Department of Biomedical Sciences, University of Cape Coast, Ghana
f Department of Anaesthesia, Greater Accra Regional Hospital, Accra, GhanaA R T I C L E I N F O
Article History:
Received 18 June 2023
Accepted 29 June 2023
Available online 5 July 2023Abbreviations: BMI, body mass index; BP, blood press
DNS, Diabetic neuropathy symptom; HBA1c, glycated h
lipoprotein; LDL, low-density lipoprotein; NPDR, Non-p
thy; OR, odds ratio; PDR, Proliferative diabetic retinopa
sure; T2D, type 2 diabetes mellitus; VPT, Vibration perce
* Corresponding author at: Departments of Internal
and Respiratory Medicine, Amsterdam UMC, Universit
605, Dansoman, Accra, Ghana.
E-mail addresses: c.hayfronbenjamin@amsterdamum
(C.F. Hayfron-Benjamin).
https://doi.org/10.1016/j.deman.2023.100160
2666-9706/© 2023 The Authors. Published by Elsevier M
(http://creativecommons.org/licenses/by-nc-nd/4.0/)A B S T R A C T
Background: In type 2 diabetes mellitus (T2D), cardiovascular risk factors including glycemic control differen-
tially affect various microcirculatory beds. To date, studies comparing the impact of blood pressure (BP) on
various microvascular beds in T2D are limited. We assessed the associations of BP and its control with neural,
renal, and retinal microvascular dysfunction.
Methods: This was a cross-sectional study among 403 adults with T2D. Microvascular dysfunction was based
on nephropathy (albumin-creatinine ratio ≥ 30 mg/g), neuropathy (vibration perception threshold ≥ 25 V
and/or Diabetic Neuropathy Symptom score > 1), and retinopathy (based on retinal photography). Logistic
regression was used to examine the associations of hypertension, systolic BP, and diastolic BP with microvas-
cular dysfunction with adjustments for age, sex, diabetes duration, smoking pack years, HbA1c concentration,
total cholesterol concentration, and BMI.
Results: The mean age (§ SD), proportion of females, and proportion of hypertensives were 56.35 (§ 9.91)
years, 75.7%, and 49.1%, respectively. In a fully adjusted model, hypertension was significantly associated
with neuropathy [odds ratio 3.44, 95% confidence interval 1.96−6.04, P < 0.001] and nephropathy [2.05 (1.09
−3.85), 0.026] but not for retinopathy [0.98 (0.42−2.31), 0.970]. Increasing Z-score systolic BP was signifi-
cantly associated with nephropathy [1.43 (1.05−1.97), 0.025] but not for neuropathy [1.28 (0.98−1.67),
0.075] or retinopathy [1.27 (0.84−1.91), 0.261]. Increasing Z-score diastolic BP was significantly associated
with nephropathy [1.81 (1.32 − 2.49), < 0.001] but not retinopathy [1.38 (0.92−2.05), 0.120] or neuropathy
[0.86 (0.67−1.10), 0.230].
Conclusion: Our study shows varying strengths of associations of hypertension, systolic BP, and diastolic
BP with microvascular dysfunction in different microcirculatory beds. Hypertension prevention and/or
control may be valuable in the prevention/treatment of microvascular disease, especially nephropathy,
and neuropathy.
© 2023 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)Keywoards:
Blood pressure control
Hypertension
Microvascular complications
Nephropathy
Neuropathy
Retinopathy
Type 2 diabetesure.; CI, confidence interval.;
emoglobin; HDL, high-density
roliferative diabetic retinopa-
thy; SBP, systolic blood pres-
ption threshold
Medicine, Vascular Medicine,
y of Amsterdam, P O Box DC
c.nl, charlesfhb1@gmail.com
asson SAS. This is an open access article under the CC BY-NC-ND licenseIntroduction
Diabetes mellitus remains a global public health problem, affect-
ing over ten percent of the world’s adult population [1]. A character-
istic complication of diabetes mellitus is microvascular disease
including retinopathy, neuropathy, and nephropathy [2]. These
microvascular complications are important causes of disability. For
example, diabetic retinopathy is the commonest cause of blindness
in adults [3], diabetic nephropathy is the commonest cause of
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160end-stage renal failure [4], and diabetic neuropathy is a leading cause
of leg ulcers, gangrene, and lower limb amputations [5]. Type 2
diabetes mellitus (T2D), the commonest type of diabetes, is associ-
ated with other vascular risk factors including hypertension, obesity,
and dyslipidemia [1].
There are important gaps in the treatment of diabetes-related
microvascular complications. For example, there is currently no effec-
tive treatment for diabetic neuropathy [6]. Given this, a multifactorial
approach targeting glycemia and other conventional risk factors has
been recommended for the prevention of diabetic microvascular dis-
ease [6]. Targeting optimum glucose control may not always be effec-
tive in the prevention of microvascular complications in T2D. For
example, while enhanced glucose control limits the progression of
neuropathy in type 1 diabetes, there is limited evidence for the same
in T2D [7]. This has warranted the need for identifying alternative
strategies for preventing the development and progression of micro-
vascular disease in T2D. Based on work done by our team among
sub-Saharan Africans [8−11], and those of other researchers in other
populations [2,12,13], non-conventional candidate risk factors
including low-grade inflammation, hyperuricemia, and macrovascu-
lar dysfunction also do not fully explain the burden of vascular dys-
function. One potential factor is the impact of blood pressure (BP)
control on systemic microvascular injury. It has previously been
reported that patients with T2D have a loss of myogenic tone and
vascular hypertrophy in resistance vessels [14]. There is also evidence
that deranged myogenic tone and vascular hypertrophy in T2D may
be partially ameliorated after treatment with the angiotensin-recep-
tor blocker candesartan [15]. Given the above, BP and its control
could play an important role in the development and/or progressions
of microvascular complications. However, the prognostic potency of
the individual vascular risk factors for vascular injury is known to
vary in the different vascular beds [16]. We hypothesize that the
associations of BP and its control with microvascular dysfunction
differ by microvascular beds. Therefore, we assessed the associations
of BP and its control with neural, renal, and retinal microvascular
dysfunction.
Methods
Study design
This cross-sectional study among adult Ghanaians with T2D man-
aged at a National Diabetes Management and Research center in
Ghana has been previously described [17]. Between 2019 and 2022, a
total of 500 adults with established diagnoses of T2D and who did
not have primary heart disease and/or previous/current heart failure
were recruited for pulmonary, cardiac, and vascular functional
assessment. The patients were systematically sampled from patients
who reported for clinic visits. The diagnosis of T2D was based on fast-
ing plasma glucose (FPG) ≥ 7.0 mmol/l and/or 2−h plasma glucose ≥
11.1 mmol/l and/or on hypoglycemic agents who reported the start
of their diabetes age > 30 years, and whose diabetes initially did not
require insulin for management. The current analyses included 403
participants with data on at least one measure of microvascular func-
tion. Before the start of data collection, ethical approval was obtained
from the Ethics Committees of the University of Ghana College of
Health Sciences (CHS-Et/M6-P20.14/2017−2018) and the Korle Bu
Teaching Hospital (KBTH-IRB/000,124/2019). All participants pro-
vided written informed consent before enrolment in the study.
Measurements
Assessment of baseline sociodemographic/clinical characteristics
including smoking and diabetes duration, measurement of height,
weight, and concentrations of fasting plasma glucose, HbA1c, choles-
terol, and serum creatinine has also been previously described [17].2
BP measurement and control
BP was measured thrice using the Omron BP Monitor HEM-907XL
device, with appropriate-sized cuffs after at least 5 minutes of rest
while seated. The mean of the last two BP measurements was used
for the analyses. Hypertension was based on a clinical diagnosis
code/documentation in the medical records, evidenced by docu-
mented systolic BP (SBP) ≥ 140mmHg and/or diastolic BP (DBP) ≥
90mmHg, and/or being on antihypertensive medication treatment
[18]. Suboptimal BP control was defined per the 2017 American Col-
lege of Cardiology/American Heart Association guidelines criteria and
European Society of Cardiology/European Society of Hypertension
guidelines (for individuals with hypertension and diabetes) as sys-
tolic BP >/= 130 mmHg and/or diastolic BP >/= 80 mmHg [19]. These
cutoff values are consistent with the American Diabetes Association’s
recommendation for individuals with diabetes with higher cardiovas-
cular risk [20].
Assessment of neural microvascular dysfunction
Neural microvascular dysfunction was based on the presence of
neuropathy. Vibration perception threshold (VPT) was measured using
the Horwell Neurothesiometer (Scientific Laboratory Supplies Ltd, Not-
tingham) at the metatarsophalangeal joint of both feet in a two-step
manner (ascending and descending methods). The VPT procedure was
demonstrated to each participant before the VPT testing was done. To
explain the nature of the vibration to be perceived, the vibration was
demonstrated at the planter surface of the participant’s thumb. VPT
testing was done while the patient was in the supine position and the
neurothesiometer was placed in a position where the patient could
not see the vibration settings. Using light pressure, the plastic tip of
the neurothesiometer was applied against the plantar surface of the
great toe of the foot. For the ascending method, the induced vibrations
were gradually increased by increasing the device voltage at a rate of
0.5 Vs/second until the participant first detected the vibration. The
voltage at which the participant first detected the vibration was docu-
mented as the ascending VPT. Two other ascending vibrating tests
were performed, and the arithmetic mean of consistent values was
used as the final ascending VPT used in the analyses. The consistent
voltages were voltages whose values do not differ by > 0.5 V. The
technique was repeated using a descending approach by decreasing
the voltage from 50 V at a rate of 0.5 Vs/second and identifying the
lowest voltage at which no vibration was perceived; this voltage was
documented as the descending VPT. The average of consistent
descending voltages was used in the final descending VPT used in the
analyses. Symptoms of diabetic neuropathy were scored with the Dia-
betic Neuropathy Symptom (DNS) score [21]. Neuropathy was diag-
nosed if the VPT was ≥ 25 V [22] and/or a DNS score ≥ 1 [21].
Assessment of renal microvascular dysfunction
Renal microvascular dysfunction was based on the presence of
albuminuria. Direct analyses of urinary albumin and creatinine con-
centration were performed on an early morning urine sample. Uri-
nary albumin concentration was measured by an immunochemical
turbidimetric method (Roche Diagnostics). Urinary creatinine con-
centration was measured by a kinetic spectrophotometric method
(Roche Diagnostics). Albuminuria was based on the urinary albumin-
creatinine ratio (ACR) ≥ 30 mg/g [category ≥ A2]) according to the
2012 Kidney Disease: Improving Global Outcomes (KDIGO) guide-
lines [23].
Assessment of retinal microvascular dysfunction
Retinal microvascular dysfunction was based on the presence of
retinopathy. The ZEISS 500 Fundus Camera (ZEISS Inc. JENA) was
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160used for retinal photography after dilatation with tropicamide (1%)
and phenylephrine (2.5%) ophthalmic solutions. Retinal images were
analyzed and graded by a certified ophthalmologist according to the
Early Treatment Diabetic Retinopathy Study (ETDRS) criteria [24].
Diabetic retinopathy when present was classified as either prolifer-
ative diabetic retinopathy (PDR) or non-proliferative diabetic
retinopathy (NPDR). NPDR was graded as mild, moderate, severe, or
very severe. Mild NPDR was based on the presence of at least one
microaneurysm in the absence of any criteria for moderate, severe, or
very severe NPDR. Moderate NPDR was based on the presence of
hemorrhage/microaneurysm ≥ standard photograph #2A or soft exu-
dates (cotton wool spots), venous beading, and intraretinal microvas-
cular abnormalities present, in the absence of any criteria for severe
or very severe NPDR or PDR. Severe NPDR was based on the presence
of hemorrhage/microaneurysm ≥ standard photograph #2A in all 4
quadrants or venous beading in at least 2 quadrants or intraretinal
microvascular abnormalities ≥ standard photograph #8A in at least 1
quadrant. Very severe NPDR was based on the presence of any two
criteria for severe NPDR in the absence of PDR. PDR was graded as
either early PDR, high-risk PDR, or severe PDR. Because of the smaller
number of individuals in the moderate group (N = 1) for the NPDR
group, we merged moderate and severe NPDR groups. Further, only
two individuals were in the high-risk or severe PDR groups. There-
fore, we graded diabetic retinopathy into either mild NPDR/early PDR
or moderate or severe NPDR/high risk or severe PDR.
Statistical analysis
Data were analyzed using the IBM SPSS version 25 for Windows.
Differences in clinical characteristics between individuals with and
without hypertension or sub-optimal BP control were assessed by
the chi
square test or Fisher’s exact test for categorical variables,
t
test for continuous variables, or the Mann
Whitney U
test for
variables not normally distributed. Multivariate logistic regression
analyses were used to examine the associations of hypertension, sub-
optimal BP control, and increasing z-score systolic BP or diastolic BP
(independent variables) with measures of microvascular dysfunction
(dependent variables), with adjustment for potential covariates. The
minimal sufficient adjustment sets for estimating the direct effect of
impaired spirometry on microvascular dysfunction were determinedFig. 1. Directed acyclic graph for estimating the direct effect of hyperte
3
by a directed acyclic graph (DAG) (Fig. 1). Based on the DAG, the min-
imal sufficient adjustment sets were age, sex, diabetes duration,
smoking pack years, HbA1c concentration, total cholesterol concen-
tration, and BMI. Three models were used to examine the data. Model
1 was unadjusted. Model 2 was adjusted for age and sex. Model 3 was
adjusted for age, sex, diabetes duration, smoking pack years, HbA1c
concentration, total cholesterol concentration, and BMI. Odds ratios
(ORs) and their corresponding 95% CI were estimated. A statistical
test of significance was set at a P-value < 0.05.
Results
General characteristics of the study participants
The baseline characteristics of the study population are summa-
rized in Table 1. Compared with individuals with T2D without hyper-
tension, individuals with T2D with hypertension were much older,
had a longer mean duration of diabetes, and were more frequently
users of statins and insulin. The mean BMI, waist circumference, and
WHR were higher in T2D individuals with hypertension compared
with their non-hypertensive comparative group. There were no sig-
nificant differences in the mean HbA1C and cholesterol concentra-
tions between the two groups. Similar observations were made when
the two groups were compared in terms of hypertensive control,
except for the observations that the T2D with suboptimal BP control
had a higher female-to-male ratio and a similar mean WHR com-
pared with the T2D with optimal BP control group.
Microvascular dysfunction in the study population
Table 2 compares the measures of neural, renal, and retinal micro-
vascular function in the study population. Respectively 31.0%, 25.4%,
and 28.8% had markers of neural, renal, and retinal microvascular
dysfunction. The mean VPT in the left and right legs were respectively
38% higher (12.30 Vs vs. 8.94 Vs, P < 0.001) and 34% higher (12.04 Vs
vs. 8.96 Vs, P < 0.001) in the T2D with hypertension group compared
to the T2D without hypertension group (P < 0.001 for both). The
prevalence rate of neuropathy was nearly thrice as high in the T2D
with hypertension group compared with the T2D without hyperten-
sion group (46.5% vs 16.1, P < 0.001). These observed disparities werension or blood pressure measures on microvascular dysfunction.
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160
Table 1
General characteristics of the study participants.
All Participants (N = 403) Hypertension Status * BP Control *
Normotensive (N = 205) Hypertensive (N = 198) p-value Optimum (N = 111) Suboptimum (N = 293) P-value
Sex (%) 0.165 0.027
Female 305 (75.7%) 149 (72.7%) 156 (78.8%) 75 (67.6%) 230 (78.8%)
Male 98 (24.3%) 56 (27.3%) 42 (21.1%) 36 (32.4%) 62 (21.2%)
Age (years) 56.35 (§9.91) 53.74 (§11.11) 59.05(§7.63) <0.001 53.05 (§10.58) 57.61 (§9.36) <0.001
Higher education (%) 149 (43.2%) 68 (42.2%) 81 (44.0%) 0.745 49 (55.1%) 100 (39.1%) 0.009
Smoking status (%) 0.747 0.711
Never smoker 394 (97.8%) 201 (98.0%) 193 (97.5%) 108 (97.3%) 286 (97.9%)
Current/previous smoker 9 (2.2%) 4 (2.0%) 5 (2.5%) 3 (2.7%) 6 (2.1%)
Smoking pack years* 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.687 0.0 (0.0) 0.0 (0.0) 0.703
Duration of diabetes (years) 11.15 (§7.17) 9.44(§7.48) 12.73 (§6.48) <0.001 8.71 (§6.65) 11.99 (§7.15) <0.001
Insulin use (%) 138 (34.2%) 47 (22.9%) 91 (46.0%) <0.001 28 (25.2%) 110 (37.7%) 0.019
Statin use (%) 191 (47.4%) 58 (28.3%) 133 (67.2%) <0.001 42 (37.8%) 149 (51.1%) 0.019
Systolic BP, mmHg 136.51 (§16.23) 128.60 (§32) 144.70 (§16.50) <0.001 119.48 (§7.29) 142.98 (§13.82) <0.001
Diastolic BP, mmHg 79.47 (§8.40) 77.61 (§7.38) 81.40 (§8.95) <0.001 72.36 (§5.21) 82.18 (§7.78) <0.001
Pulse rate, beats per minute 79.62 (§11.15) 78.70 (§11.40) 80.57 (§10.83) 0.093 78.60 (§10.82) 80.00 (§11.26) 0.262
Anthropometry
BMI, kg/m2 29.55 (§6.07) 28.48 (§5.84) 30.65 (§6.12) <0.001 28.4 (§5.47) 29.98 (§6.24) 0.019
Waist circumference, cm 96.16 (§12.18) 93.83 (§11.34) 98.56 (§12.57) <0.001 93.70 (§10.82) 97.09 (§12.55) 0.012
WHR 0.90 (§0.07) 0.89 (§0.08) 0.91 (§0.74) 0.047 0.89 (§0.08) 0.90 (§0.07) 0.799
Biochemical Measurements
HbA1c,% 7.98 (§1.78) 8.03 (§1.85) 7.93 (§1.70) 0.596 8.00 (§1.87) 8.00 (§1.74) 0.903
Total cholesterol, mmol/l 4.88 (§1.27) 4.91 (§1.23) 4.85 (§1.31) 0.712 4.6 (§1.13) 4.95 (§1.32) 0.110
Triglyceride, mmol/l 1.23 (§0.51) 1.20 (§0.49) 1.26 (§0.52) 0.374 1.18 (§0.44) 1.25 (§0.36) 0.228
HDL- cholesterol, mmol/l 1.34 (§0.35) 1.37 (§0.36) 1.32 (§0.33) 0.171 1.37 (§0.36) 1.33 (§0.34) 0.339
LDL-cholesterol, mmol/l 2.97 (§1.15) 2.98 (§1.14) 2.95 (§1.16) 0.783 2.78 (§1.05) 3.04 (§1.18) 0.078
Obesity is defined as BMI ≥ 30 kg/m2; Central obesity defined as waist to hip ratio ≥ 0.90 for males and ≥ 0.85 for females.
Higher education is defined as secondary school education or higher.
Abbreviations: BMI = body mass index, HBA1c = glycated hemoglobin; WC = waist circumference.less masked when individuals with T2D with and without suboptimal
BP control were compared.
The median ACR was 63% higher in the in the T2D with hyperten-
sion group compared to the T2D without hypertension group
(19.50 mg/g vs. 12.00 mg/g, P < 0.001). Similarly, the proportion of
individuals with albuminuria was nearly 80% higher in individuals
with T2D with hypertension compared with those with T2D without
hypertension (32.6% vs. 18.2%, P = 0.003). Similar observations were
made when individuals with T2D with and without suboptimal BP
control were compared.
The rates, types, and grades of retinopathy were similar in the T2D
with hypertension and T2D without hypertension groups. However,Table 2
Microvascular Dysfunction in the Study Population.
Microvascular circulation Hypertensio
A. Neural microvascular All Participants (N = 403) Normotensive (N = 205) Hype
VPT left leg, volts 10.59 (§7.93) 8.94 (§5.77) 12.3
VPT right leg, volts 10.47 (§8.06) 8.96 (§6.22) 12.0
Neuropathy (%) 125 (31.0%) 33 (16.1%) 92 (4
B. Renal microvascular All Participants (N = 342) Normotensive (N = 170) Hype
ACR, mg/g 12.00 (24.00) 12.00 (18.00) 19.5
Nephropathy 87 (25.4%) 31 (18.2%) 56 (3
C. Retinal microvascular All Participants (N = 177) Normotensive (N = 77) Hype
Retinopathy 51 (28.8%) 19 (24.7%) 32 (3
Retinopathy type
PDR 4 (2.3%) 1 (1.3%) 3 (3.
NPDR 47 (26.6%) 18 (23.4%) 29 (2
Retinopathy Grade
Mild NPDR / early PDR 45 (25.4%) 18 (23.4%) 27 (2
Moderate to severe NPDR / 6 (3.4%) 1 (1.3%) 5 (5.
high risk or severe PDR
ACR expressed as median 9interquartile range).
Abbreviations: ACR = urinary albumin to creatinine ratio; DNS = Diabetic Neuropathy Symp
nopathy; VPT = vibration perception threshold.
4
differences were observed when individuals with T2D with and with-
out suboptimal BP control were compared. For example, the prevalence
of retinopathy was over twice as high in individuals with T2D with
suboptimal BP control compared with those with optimal BP control
(33.6% vs. 14.0%, P = 0.013). Also, individuals with T2D with suboptimal
BP control were less likely to have mild NPDR or early PDR compared
with those with optimal BP control (29.1% vs. 14.0%, P = 0.036).
Relationship between measures of BP and microvascular dysfunction
In the unadjusted and age and sex-adjusted models, hypertension
was positively associated with higher odds of neuropathy andn Status * BP Control *
rtensive (N = 198) p-value Optimum (N = 111) Suboptimum (N = 292) P-value
0 (§9.38) <0.001 9.24 (§6.67) 11.10 (§8.31) 0.021
4 (§9.36) <0.001 9.77 (§7.47) 10.74 (§8.27) 0.291
6.5%) <0.001 27 (24.3%) 98 (33.6%) 0.091
rtensive (N = 172) p-value Optimum (N = 92) Suboptimum (N = 250) P-value
0 (34.00) <0.001 11.00 (18.00) 17.00 (29.00) 0.001
2.6%) 0.003 14 (15.2%) 73 (29.2%) 0.008
rtensive (N = 100) p-value Optimum (N = 43) Suboptimum (N = 134) P-value
2.0%) 0.318 6 (14.0%) 45 (33.6%) 0.013
0.496 0.040
0%) 0 (0.0%) 4 (3.0%)
9.0%) 6 (14.0%) 41 (30.6%)
0.315 0.036
7.0%) 6 (14.0%) 39 (29.1%)
0%) 0 (0.0%) 6 (4.4%)
tom; NPDR = non-proliferative diabetic retinopathy; PDR = proliferative diabetic reti-
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160
Table 3
Logistic regression models for the between measures of obesity (independent variable) and pulmonary dysfunction (dependent variable).
Neuropathy (N = 125) Nephropathy (N = 87) Retinopathy (N = 51)
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Hypertension 4.52 (2.84 − 7.21), 4.74 (2.91 − 7.74), 3.44 (1.96 − 6.04), 2.17 (1.31 − 3.58), 1.88 (1.11 − 3.19), 2.05 (1.09 − 3.85), 1.44 (0.74 − 2.80), 1.38 (0.69 − 2.75), 0.98 (0.42 − 2.31),
<0.001 <0.001 <0.001 0.003 0.018 0.026 0.287 0.360 0.970
Suboptimal BP control 1.57 (0.96 − 2.58), 1.51 (0.90 − 2.51), 1.11 (0.61 − 2.04), 2.30 (1.22 − 4.32), 1.94 (1.01 − 3.72), 1.63 (0.77 − 3.47), 3.12 (1.23 − 7.94), 3.02 (1.16 − 7.88), 2.18 (0.72 − 6.61),
0.075 0.116 0.732 0.010 0.046 0.205 0.017 0.024 0.170
Z-score systolic BP 1.44 (1.16 − 1.79), 1.44 (1.14 − 1.81), 1.28 (0.98 − 1.67), 1.62 (1.26 − 2.07), 1.49 (1.15 − 1.94), 1.43 (1.05 − 1.97), 1.58 (1.13 − 2.21), 1.56 (1.10 − 2.22), 1.27 (0.84 − 1.91),
0.001 0.002 0.075 <0.001 0.003 0.025 0.007 0.013 0.261
Z-score diastolic BP 0.97 (0.78 − 1.19), 0.97 (0.78 − 1.20), 0.86 (0.67 − 1.10), 1.65 (1.26 − 2.16), 1.69 (1.29 − 2.22), 1.81 (1.32 − 2.49), 1.45 (1.05 − 2.01), 1.47 (1.06 − 2.04), 1.38 (0.92 − 2.05),
0.749 0.770 0.230 <0.001 <0.001 <0.001 0.025 0.022 0.120
Model 1 = unadjusted.
Model 2 = adjusted for age and sex.
Model 3 = Adjusted for age, sex, diabetes duration, smoking pack years, HbA1c concentration, total cholesterol concentration, and BMI.
Abbreviations: BMI = body mass index; BP = BP.nephropathy but not retinopathy (Table 3). The association remained
statistically significant in the fully adjusted model for neuropathy
[odds ratio 3.44, 95% confidence interval 1.96−6.04, P < 0.001] and
nephropathy [2.05 (1.09−3.85), 0.026] but not for retinopathy [0.98
(0.42−2.31), 0.970]. While suboptimal BP control was positively
associated with nephropathy and retinopathy in both unadjusted
and age-sex-adjusted models, the strength of associations was not
statistically significant in the fully adjusted model for each of the
measures of microvascular dysfunction: neuropathy [1.11 (0.61
−2.04), 0.732]; nephropathy [1.63 (0.77−3.47), 0.205]; retinopathy
[2.18 (0.72−6.61), 0.170].
Increasing Z-score systolic BP was positively associated with
higher odds of neuropathy, nephropathy, and retinopathy in the
unadjusted and age and sex-adjusted models. In the fully adjusted
model, the positive association between Increasing Z-score systolic
BP and microvascular dysfunction was significant for nephropathy
[1.43 (1.05−1.97), 0.025] but not for neuropathy [1.28 (0.98−1.67),
0.075] or retinopathy [1.27 (0.84−1.91), 0.261].
Increasing Z-score diastolic BP was not significantly associated
with neuropathy in all three models including the fully adjusted
model [0.86 (0.67−1.10), 0.230]. Increasing Z-score diastolic BP was
positively associated with higher odds of nephropathy and retinopa-
thy in the unadjusted and age and sex-adjusted models. In the fully
adjusted model, the strength of association was significant in the
fully adjusted model for nephropathy [1.81 (1.32−2.49), < 0.001) but
not for retinopathy [1.38 (0.92−2.05), 0.120].
Discussion
Summary of key findings
Our study has four key findings. First, neuropathy was the most
common microvascular complication and nephropathy was the least
common microvascular complication in our study population of West
Africans with T2D. Secondly, coexisting hypertension and/or subopti-
mal BP control was associated with an increased likelihood of micro-
vascular dysfunction in individuals with T2D; hypertension diagnosis
was significantly associated with neuropathy and nephropathy, while
suboptimal BP was significantly associated with nephropathy and
retinopathy. Thirdly, the positive associations of hypertension with
neuropathy and nephropathy were independent of the conventional
vascular risk factors. Finally, there were differential associations of
measures of BP and control with microvascular dysfunctions in differ-
ent microcirculation.
Discussion of key findings
In our study population of West Africans with T2D managed at a
national diabetes management and referral center, neuropathy was
the commonest microvascular complication, while nephropathy was
the least common. To the best of our knowledge, no study among
West Africans has compared the burden of microvascular dysfunc-
tions in the three key microcirculations affected by diabetes. Our
finding of neuropathy as the commonest complication is consistent
with prior studies among individuals of European ethnic origins as
well as those living in high-income settings [6,25]. Similar reports
exist for individuals of Asian ethnic origin living in low-resource set-
tings [26]. While about 50% of individuals with T2D may eventually
develop diabetic neuropathy [27], prevalence rates from cross-sec-
tional studies have varied based on the population studied. Our
reported diabetic neuropathy prevalence rate of 31.0% is slightly
higher than that reported in high-income settings such as a cross-
sectional multicenter hospital clinic-based study in the UK that
reported a prevalence rate of 28.5 [28]. We have previously reported
on how the prevalence of retinopathy and nephropathy in our study
population compared with other populations [17]. The biological5
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160basis of increased susceptibility of the neural microcirculation to
dysfunction, compared with other microcirculations, is not fully
understood. Proposed explanations include enhanced hyperglyce-
mia-induced oxidative stress in diabetic neurons [25] or the fact that
some nondiabetic neuropathies may be misdiagnosed as diabetic
neuropathy [6].
The current study shows that coexisting hypertension and/or sub-
optimal BP control is associated with an increased likelihood of
microvascular dysfunction in individuals with T2D. Our findings
agree with a limited study that assessed the impact of hypertension
on diabetes-related complications based on questionnaires [29].
Kesavamoorthy et al. reported that the proportion of individuals
reporting diabetes neuropathy and retinopathy was higher among
individuals with diabetes and hypertension compared with those
without hypertension [29]. Our findings and that of Kesavamoorthy
et al. are supported by evidence that hypertension or elevated BP
alone can cause derangements in the microvascular circulation
[30−33]. Mechanistically, hypertension alters the mechanisms regu-
lating vasomotor tone resulting in enhanced vasoconstriction or
reduced vasodilator responses to various vasodilators [34]. In addi-
tion, the loss of myogenic tone and vascular hypertrophy in resis-
tance vessels of hypertensive patients with T2D results in increased
wall thickness to radius ratio, resulting in microcirculatory dysfunc-
tion [14]. Regardless of the above mechanistic model, some studies
have not found hypertension to substantially enhance the risk of
microvascular injury in T2D. For example in the Wisconsin Epidemio-
logic Study of Diabetic Retinopathy, a diagnosis of hypertension, sys-
tolic BP, and diastolic BP were not associated with the incidence and
progression of retinopathy in individuals with T2D [35]. Our findings
suggest that in individuals with T2D of West African ethnic origin,
coexisting hypertension and/or suboptimal BP control are associated
with an increased likelihood of microvascular injury. BP control in
West Africans with T2D may thus have protective benefits including
enhanced microvascular function.
A novel contribution of our study to the existing literature is our
comparison of the differential effects of BP and control measures on
different microcirculation. While prior studies have compared the
differential effects of hypertension on different microcirculatory
beds, none has done so in the setting of T2D. Our results show that
measures of abnormal BP are more strongly associated with renal
microvascular dysfunction than neural and retinal microvascular dys-
function. In the fully adjusted model, both increasing systolic BP and
diastolic BP were independently associated with an increased likeli-
hood of nephropathy, but not neuropathy or retinopathy. Our find-
ings are consistent with previous observations that show that high-
flow, low-impedance organs including the brain and kidney are par-
ticularly vulnerable to high BP pulsatility since the pulsatile load in
these organs penetrates deeply into their microvascular beds [36].
T2D is independently associated with stiffness of the elastic proximal
arteries including the aorta [37,38]. In the setting of aortic stiffness,
there is a loss in the ability to dampen the pulsatility of ventricular
ejection [36]. However, the distal muscular arteries may not stiffen.
Therefore, there is a decrease in the difference or gradient of stiffness
between the proximal elastic arteries and distal muscular arteries,
resulting in a reduction in the backward reflection of pulsatile energy.
The pulsatile energy that is not reflected backward may be transmit-
ted toward the small arteries of target organs [36]. Because the renal
microvasculature is more susceptible to high BP pulsatility compared
with the neural and retinal microcirculation, measures of increasing
BP are more likely to be associated with renal microvascular dysfunc-
tion, compared with neural and retinal microvascular dysfunction,
Like systolic and diastolic BP, a diagnosis of hypertension was
independently associated with nephropathy. However, a diagnosis of
hypertension was also independently associated with neuropathy.
The strength of the association between hypertension diagnosis and
neuropathy was more pronounced than that of nephropathy. For6
example, in a fully adjusted model, hypertension diagnosis was asso-
ciated with 244% increased odds of neuropathy (P < 0.001) and 105%
increased odds of nephropathy (P = 0.026); there was no significant
association with retinopathy. Given the evidence of no effective treat-
ment for diabetic neuropathy [6], interventions aimed at preventing
hypertension in individuals with T2D may be valuable in reducing
the risk of diabetic neuropathy. It remains unclear why a diagnosis of
hypertension had different strengths of associations with microvas-
cular dysfunction, compared with measures of BP. It may well be the
case that hypertension per se is more strongly associated with neu-
ropathy, but poor BP control more adversely affects high-flow, low-
impedance organs including the kidney [36]. Future longitudinal
studies may confirm or refute this claim.Strengths and limitations
To the best of our knowledge, our study is the first to compare the
impact of BP and its control on complementary measures of systemic
microvascular dysfunction in individuals with T2D. The strengths of
our study include an objective assessment of three complementary
measures of microvascular dysfunction and the adjustments for a
wide range of confounders. Our study has limitations. First, the cross-
sectional design limits us from precluding reverse caus. Secondly, we
did not assess microvascular dysfunctions of the coronary and cere-
bral microcirculation due to the current technical challenges associ-
ated with microvascular functional testing in these circulations. For
example, coronary microcirculation is beyond the resolution of inva-
sive or noninvasive coronary angiography [39].Conclusions
The current study demonstrates differential strengths of associa-
tions of hypertension and increasing BP with dysfunction in the renal,
neural, and retinal microcirculation in T2D. The positive associations
of hypertension with neuropathy and nephropathy persisted after
adjustment for a wide range of conventional risk factors. Increasing
systolic and diastolic BP were independently associated with an
increased likelihood of nephropathy but not neuropathy or retinopa-
thy. Based on our findings, individuals with T2D may benefit from
interventions aimed at hypertension prevention, to prevent the like-
lihood of developing neuropathy or nephropathy. In the setting of
T2D, BP control aimed at reducing the likelihood of microvascular
dysfunction may most benefit the kidneys. Our findings provide use-
ful insights into the possible role of BP control in systemic microvas-
cular dysfunction and provide opportunities for future research
aimed at determining the mechanisms linking differential impacts of
hypertension and/or BP control to microvascular dysfunction in dif-
ferent microcirculatory beds, as well as studies targeting strategies
for microvascular dysfunction prevention and/or treatment.Authorship contributions
All authors contributed substantially to this article and approved
the submission. C.F.H-B and T.R.Q-P conceived the idea. C.F.H-B and
P.V. performed the experiment. C.F.H-B, T.A-N., and K.A.O. were
responsible for statistical analysis and data interpretation. Each
author contributed important intellectual content during article
drafting or revision and accepts accountability for the overall work
by ensuring that questions about the accuracy or integrity of any por-
tion of the work are appropriately investigated and resolved. C.F.H-B.
takes responsibility for the fact that this study has been reported
honestly, accurately, and transparently, that no important aspects of
the study have been omitted, and that any discrepancies from the
study as planned have been explained.
C.F. Hayfron-Benjamin, T.R. Quartey-Papafio, T. Amo-Nyarko et al. Diabetes Epidemiology and Management 12 (2023) 100160Funding
This work was funded by the Faculty Development Grant of the
University of Ghana College of Health Science.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Acknowledgments
We are very grateful to the study participants for taking part in
the study. Patience Vormatu deserves special commendation for coor-
dinating the recruitment and study procedures. We thank the
research assistants (Abraham Ablor, Alexander Danquah, Maame
Boatemaa Ansong, Michael Adjei, and Nii Adjetey Kwaw Wills) for
assisting with interviews and anthropometric measurements. We also
Edem Ahiabor for assisting with the retinal photography. We grate-
fully acknowledge the leadership and staff of the NDMRC for helping
with various aspects of the study. We also acknowledge Elikem Jasper
Butsey, Faustina Aba Amoah, Odelia Tamakloe, and Seth Agyemang,
and the staff of the National Cardiothoracic center Laboratory and the
KBTH Accident and Emergency Laboratory for running the biochemi-
cal tests. We are grateful to Dr. Charles Antwi-Boasiako, Dr. Yvonne
Dei-Adomako, and Mr. Enoch Mensah for their support in biobank
management and the high-quality storage of collected samples.
References
[1] Sun H, Saeedi P, Karuranga S, Pinkepank M, Ogurtsova K, Duncan BB, et al. IDF
Diabetes Atlas: global, regional and country-level diabetes prevalence estimates
for 2021 and projections for 2045. Diabetes Res Clin Pract 2022;183:109119. doi:
10.1016/j.diabres.2021.109119.
[2] Fowler MJ. Microvascular and macrovascular complications of diabetes. Clin Dia-
betes 2008;26:77–82. doi: 10.2337/diaclin.26.2.77.
[3] Fong DS, Aiello LP, Ferris FL, Klein R. Diabetic retinopathy. Diabetes Care
2004;27:2540–53. doi: 10.2337/diacare.27.10.2540.
[4] Association AD. Diabetic nephropathy. Diabetes Care 2003;26(1):s94–8 suppl.
doi: 10.2337/diacare.26.2007.S94.
[5] Haneda M, Utsunomiya K, Koya D, Babazono T, Moriya T, Makino H, et al. A new
classification of diabetic nephropathy 2014: a report from joint committee on dia-
betic nephropathy. J Diabetes Investig 2015;6:242–6. doi: 10.1111/jdi.12319.
[6] Pop-Busui R, Boulton AJM, Feldman EL, Bril V, Freeman R, Malik RA, et al. Diabetic
neuropathy: a position statement by the American Diabetes Association. Diabetes
Care 2017;40:136–54. doi: 10.2337/dc16-2042.
[7] Callaghan BC, Little AA, Feldman EL, Hughes RAC. Enhanced glucose control for
preventing and treating diabetic neuropathy. Cochrane Database Syst Rev
2012;6:CD007543. doi: 10.1002/14651858.CD007543.pub2.
[8] Hayfron-Benjamin CF, der Zee AHM van, den Born BJ van, Amoah AGB, Meeks
KAC, Klipstein-Grobusch K, et al. Association between C reactive protein and
microvascular and macrovascular dysfunction in sub-Saharan Africans with and
without diabetes: the RODAM study. BMJ Open Diabetes Res Care 2020;8:
e001235. doi: 10.1136/bmjdrc-2020-001235.
[9] Hayfron-Benjamin CF, Amoah AGB, AH Maitland-van der Zee, van Charante EPM,
Galenkamp H, den Born BJ van, et al. Associations between macrovascular and
renal microvascular dysfunction in type 2 diabetes and non-diabetes. Microvasc
Res 2021;136:104162. doi: 10.1016/j.mvr.2021.104162.
[10] Hayfron-Benjamin C, van den Born BJ, Maitland - van der Zee AH, Amaoh AGB,
Meeks KAC, Klipstein-Grobusch K, et al. Microvascular and macrovascular complica-
tions in type 2 diabetes Ghanaian residents in Ghana and Europe: the RODAM study.
J Diabetes Complications 2019;33:572–8. doi: 10.1016/j.jdiacomp.2019.04.016.
[11] Hayfron-Benjamin CF, Mosterd C, Maitland - van der Zee AH, van Raalte DH,
Amoah AGB, Agyemang C, et al. Inflammation and its associations with aortic
stiffness, coronary artery disease and peripheral artery disease in different ethnic
groups: the HELIUS Study. EClinicalMedicine 2021;38:101012. doi: 10.1016/j.
eclinm.2021.101012.
[12] Climie Rachel E, van Sloten Thomas T, Rosa-Maria Bruno, Taddei S, Empana JP,
Stehouwer CDA, et al. Macrovasculature and microvasculature at the crossroads
between type 2 diabetes mellitus and hypertension. Hypertension 2019;73:
1138–49. doi: 10.1161/HYPERTENSIONAHA.118.11769.
[13] Muiesan ML, Agabiti-Rosei C, Paini A, Salvetti M. Uric acid and cardiovascular dis-
ease: an update. Eur Cardiol Rev 2016;11:54–9. doi: 10.15420/ecr.2016:4:2.7
[14] Schofield I, Malik R, Izzard A, Austin C, Heagerty A. Vascular structural and func-
tional changes in type 2 diabetes mellitus: evidence for the roles of abnormal
myogenic responsiveness and dyslipidemia. Circulation 2002;106:3037–43. doi:
10.1161/01.cir.0000041432.80615.a5.
[15] Malik RA, Schofield IJ, Izzard A, Austin C, Bermann G, Heagerty AM. Effects of
angiotensin type-1 receptor antagonism on small artery function in patients with
type 2 diabetes mellitus. Hypertens Dallas Tex 2005;45:264–9 1979. doi:
10.1161/01.HYP.0000153305.50128.a1.
[16] Gregorio Brevetti, Giuseppe Giugliano, Linda Brevetti. Hiatt William R. Inflamma-
tion in peripheral artery disease. Circulation 2010;122:1862–75. doi: 10.1161/
CIRCULATIONAHA.109.918417.
[17] Antwi-Boasiako C, Obeng KA, Amissah-Arthur KN, Musah L, Abaidoo B, Awuviri
HW, et al. Association between albuminuria and retinal microvascular dysfunc-
tion in type 2 diabetes with and without hypertension. Diabetes Epidemiol
Manag 2023;11:100139. doi: 10.1016/j.deman.2023.100139.
[18] Stergiou GS, Palatini P, Parati G, O’Brien E, Januszewicz A, Lurbe E, et al. 2021
European Society of Hypertension practice guidelines for office and out-of-office
blood pressure measurement. J Hypertens 2021;39:1293. doi: 10.1097/
HJH.0000000000002843.
[19] Whelton PK, Carey RM, Mancia G, Kreutz R, Bundy JD, Williams B. Harmonization
of the American College of Cardiology/American Heart Association and European
Society of Cardiology/European Society of Hypertension Blood Pressure/Hyper-
tension Guidelines: comparisons, Reflections, and Recommendations. Eur Heart J
2022;43:3302–11. doi: 10.1093/eurheartj/ehac432.
[20] Passarella P, Kiseleva TA, Valeeva FV, Gosmanov AR. Hypertension management in
diabetes: 2018 Update. Diabetes Spectr 2018;31:218–24. doi: 10.2337/ds17-0085.
[21] Meijer JWG, Smit AJ, Sonderen EV, Groothoff JW, Eisma WH, Links TP. Symptom
scoring systems to diagnose distal polyneuropathy in diabetes: the Diabetic Neu-
ropathy Symptom score. Diabet Med J Br Diabet Assoc 2002;19:962–5. doi:
10.1046/j.1464-5491.2002.00819.x.
[22] Young MJ, Breddy JL, Veves A, Boulton AJ. The prediction of diabetic neuropathic
foot ulceration using vibration perception thresholds. A prospective study. Diabe-
tes Care 1994;17:557–60. doi: 10.2337/diacare.17.6.557.
[23] Lamb EJ, Levey AS, Stevens PE. The Kidney Disease Improving Global Outcomes
(KDIGO) guideline update for chronic kidney disease: evolution not revolution.
Clin Chem 2013;59:462–5. doi: 10.1373/clinchem.2012.184259.
[24] Solomon SD, Goldberg MF. ETDRS Grading of Diabetic Retinopathy: still the Gold
Standard? Ophthalmic Res 2019;62:190–5. doi: 10.1159/000501372.
[25] Edwards JL, Vincent AM, Cheng HT, Feldman EL. Diabetic neuropathy: mecha-
nisms to management. Pharmacol Ther 2008;120:1–34. doi: 10.1016/j.pharm-
thera.2008.05.005.
[26] Ali A, Iqbal F, Taj A, Iqbal Z, Amin MJ, Iqbal QZ. Prevalence of microvascular com-
plications in newly diagnosed patients with type 2 diabetes. Pak J Med Sci
2013;29:899–902. doi: 10.12669/pjms.294.3704.
[27] Pirart J. [Diabetes mellitus and its degenerative complications: a prospective
study of 4,400 patients observed between 1947 and 1973 (3rd and last part)
(author’s transl)]. Diabete Metab 1977;3:245–56.
[28] Young MJ, Boulton AJ, MacLeod AF, Williams DR, Sonksen PH. A multicentre study
of the prevalence of diabetic peripheral neuropathy in the United Kingdom hospi-
tal clinic population. Diabetologia 1993;36:150–4. doi: 10.1007/BF00400697.
[29] Kesavamoorthy G, Singh AK, Sharma S, Kasav JB, Mohan SK, Joshi A. Burden of
diabetes related complications among hypertensive and non hypertensive dia-
betics: a comparative study. J Clin Diagn Res JCDR 2015;9:LC10–4. doi: 10.7860/
JCDR/2015/14461.6452.
[30] Wong TY, Mitchell P. Hypertensive retinopathy. N Engl J Med 2004;351:2310–7.
doi: 10.1056/NEJMra032865.
[31] Levy BI, Schiffrin EL, Mourad JJ, Agostini D, Vicaut E, Safar ME, et al. Impaired tis-
sue perfusion: a pathology common to hypertension, obesity, and diabetes melli-
tus. Circulation 2008;118:968–76. doi: 10.1161/CIRCULATIONAHA.107.763730.
[32] Yang CP, Lin CC, Li CI, Liu CS, Lin WY, Hwang KL, et al. Cardiovascular risk factors
increase the risks of diabetic peripheral neuropathy in patients with type 2 diabe-
tes mellitus: the Taiwan Diabetes Study. Medicine (Baltimore) 2015;94:e1783.
doi: 10.1097/MD.0000000000001783.
[33] Ponirakis G, Petropoulos IN, Alam U, Ferdousi M, Asghar O, Marshall A, et al.
Hypertension contributes to neuropathy in patients with type 1 diabetes. Am J
Hypertens 2019;32:796–803. doi: 10.1093/ajh/hpz058.
[34] De Boer MP, Meijer RI, Wijnstok NJ, Jonk AM, Houbert AJ, Stehouwer CD, et al.
Microvascular dysfunction: a potential mechanism in the pathogenesis of obe-
sity-associated insulin resistance and hypertension. Microcirc N Y N 2012;19:5–
18 1994. doi: 10.1111/j.1549-8719.2011.00130.x.
[35] Klein R, Klein BE, Moss SE, Davis MD, DeMets DL. Is blood pressure a predictor of
the incidence or progression of diabetic retinopathy? Arch Intern Med
1989;149:2427–32.
[36] Laurent S, Agabiti-Rosei C, Bruno RM, Rizzoni D. Microcirculation and macrocircu-
lation in hypertension: a dangerous cross-link? Hypertension 2022;79:479–90.
doi: 10.1161/HYPERTENSIONAHA.121.17962.
[37] Zheng M, Zhang X, Chen S, Song Y, Zhao Q, Gao X, et al. Arterial stiffness preceding
diabetes. Circ Res 2020;127:1491–8. doi: 10.1161/CIRCRESAHA.120.317950.
[38] Elias MF, Crichton GE, Dearborn PJ, Robbins MA, Abhayaratna WP. Associations
between type 2 diabetes mellitus and arterial stiffness: a prospective analysis
based on the Maine-Syracuse Study. Pulse 2017;5:88–98. doi: 10.1159/000479560.
[39] Taqueti VR, Di Carli MF. Coronary microvascular disease pathogenic mechanisms
and therapeutic options: JACC State-of-the-Art Review. J Am Coll Cardiol
2018;72:2625–41. doi: 10.1016/j.jacc.2018.09.042.