30 / 48
72
VOLUME 12 NUMBER 2 • NOVEMBER 2015
RESEARCH ARTICLE
SA JOURNAL OF DIABETES & VASCULAR DISEASE
Prevalence and covariates of electrocardiographic left
ventricular hypertrophy in diabetic patients in Tanzania
JJK LUTALE, H THORDARSON, Z GULAM-ABBAS, K VETVIK, E GERDTS
Correspondence to: JJK Lutale
Institute of Medicine and Centre for International Health, University of
Bergen, Bergen, Norway
Muhimbili University College of Health Sciences, Dar es Salaam, Tanzania
e-mail:
kente315@yahoo.co.ukH Thordarson
Department of Medicine, Haukeland University Hospital, Bergen, Norway
Z Gulam-Abbas
Muhimbili University College of Health Sciences and Abbas Medical Centre,
Dar es Salaam, Tanzania
K Vetvik
Institute of Medicine, University of Bergen, Norway
E Gerdts
Institute of Medicine, University of Bergen, Norway
Previously published in
Cardiovasc J Afr
2008;
19
: 8–14
S Afr J Diabetes Vasc Dis
2015;
12
: 72–78
Summary
Background:
Left ventricular hypertrophy (LVH) has been
demonstrated to be a powerful predictor of cardiovascular
(CV)morbidityandmortalityindiabeticaswellashypertensive
patients. However, less is known about the prevalence of
electrocardiographic LVH (ECG-LVH) and its relation to other
CV risk factors in diabetic patients in sub-Saharan Africa.
Therefore, the aim was to assess the prevalence of ECG-
LVH in diabetic patients in Dar es Salaam, Tanzania, and its
relation to other cardiovascular risk factors.
Methods:
Two hundred and thirty-seven consecutive patients
attending theMuhimbili diabetic clinic were studied. ECG-LVH
was diagnosed by Sokolow-Lyon voltage and Cornell voltage-
duration product criteria. Q waves, ST-segment deviation,
T-wave abnormalities and intraventricular conduction defects
were classified by the Minnesota codes. Blood pressure (BP),
serum creatinine, cholesterol and triglyceride levels, and
HbA
1c
and urinary albumin and creatinine concentrations
were determined.
Results:
The prevalence of LVH in patients was 16% by either
ECG criteria; 12.2% by Sokolow-Lyon and 5.1% by Cornell
product criteria. Patients with LVH had significantly higher
systolic and mean BP and pulse pressure, and a higher
prevalence of ST-segment abnormalities, T-wave inversion
and albuminuria than those without LVH (all
p
< 0.05). In
multivariate logistic regression analysis, systolic BP was the
only independent predictor of ECG-LVH. The prevalence of
ECG-LVH increased by 15% per 10 mmHg higher systolic BP
[OR 1.151 (95% CI 1.009–21.314),
p
< 0.05]. Clustering of
cardiovascular risk factors differed significantly between type
1 and type 2 diabetes patients. On average, type 1 patients
had 0.8 and type 2 had 2.2 additional CV risk factors.
Conclusion:
ECG-LVH was present in 16% of diabetic patients
in Tanzania. Systolic BP was the most important predictor of
ECG-LVH. Clustering of CV risks was significantly higher in
type 2 than in type 1 diabetics, demonstrating the need for
systematic multiple risk-factor assessment in these patients.
Left ventricular hypertrophy (LVH), whether diagnosed by
electrocardiography or echocardiography, is a manifestation of
cardiac target-organ damage and has been demonstrated to be
a powerful predictor of cardiovascular morbidity and mortality in
diabetic
1,2
as well as hypertensive patients.
3,4
Physiologically, LVH is
a structural and functional adaptation of the left ventricle chamber
to increased afterload. In previous studies, main determinants
of ECG-LVH, including advanced age,
5
male gender,
6,7
obesity,
8,9
glucose intolerance, diabetes mellitus, lipid abnormalities, cigarette
smoking and microalbuminuria
10
have been identified.
ECG-LVH has particularly been associated with hypertension in
African patients.
11-13
However, less is known about prevalence of
electrocardiographic left ventricular hypertrophy (ECG-LVH) and its
relation to other cardiovascular (CV) risk factors in diabetic patients
in sub-Saharan Africa. Therefore, the aim of the present study was
to assess the prevalence of ECG-LVH and its relation to other CV risk
factors in diabetic patients attending the diabetes outpatient clinic
at the Muhimbili National University Hospital in Dar es Salaam.
Methods
All 290 patients attending the diabetes outpatient clinic at Muhimbili
National Hospital between 1 August 2003 and 1 February 2004
were invited to participate in the study. All 271 patients without
cardiac or renal failure, cerebral vascular disease or advanced liver
disease were invited; 263 accepted participation in the present
study. Muhimbili Hospital is the national referral and a university
teaching hospital. All patients gave written informed consent
before enrolment in the study.
The study was approved by the Scientific and Publication
Committee of Muhimbili University College of Health Sciences
and the Regional Ethical Committee III in Norway. The study was
conducted in accordance with the Helsinki declaration. Patients
were classified as type 1 or type 2 diabetics according to the 1997
World Health Organisation (WHO)
14
clinical criteria, based on age
at diagnosis, mode of onset (acute versus insidious presentation),
duration of disease, current treatment, body mass index (BMI),
waist-to-hip ratio, blood pressure, random or fasting glucose, HbA
1c
and urine ketone levels.
Patients aged 30 years or younger at onset of diabetes, with acute
presentation of classical symptoms, who required insulin therapy
to control hyperglycaemia were classified as type 1. Patients older
than 30 years who needed insulin treatment, even if they had had
diabetes for a short duration, or were metabolically uncontrolled
and/or underweight, were also classified as type 1.
Patients over 30 years at diagnosis and not needing insulin
for metabolic control were classified as type 2 diabetics. Patients
younger than 30 years were also classified as type 2 if they were
obese and/or had diabetes duration of more than 10 years without