The SA Journal Diabetes & Vascular Disease Vol 12 No 2 (December 2015)

27 / 48

SA JOURNAL OF DIABETES & VASCULAR DISEASE REVIEW
VOLUME 12 NUMBER 2 • NOVEMBER 2015 69
total number of study participants recruited, study sites (e.g.
single or multicentre), gender ratios, ethnicity ratios, timelines of
results presented (e.g. single or longitudinal data) and periods of
enrollment.
To compare results, we standardised (or converted or conformed)
certain measurement units in order to maintain consistency (e.g.
LDL-C in mmol/l instead of mg/dl). The control or baseline results of
studies were reported instead of interventional group data. Only the
latest data were selected from studies with multiple time periods.
Studies excluded from the review had one or more of the
following characteristics: non-English language, studies conducted
before 2009, participating patients younger than 18 years of age,
participants reported to have had any diabetes other than T2DM
(e.g. gestational, type 1 or steroid induced), studies that reported
insufficient data or less than two of the three major risk factors
being compared, and studies that consisted of large HMO claims
databases. The latter was chosen as an exclusion criterion as larger-
sized cohort studies would have biased the results of this review.
Data presented in this article were collected from the results of
other studies and are limited to the authors’ definitions of control.
This review did not allow for the access of patient-level data of
different studies included in the review to be accessed. It was
assumed that all data extracted for this study were collected from
the medical records of patients who willingly participated in the
studies included in this review. The relevant data were captured
into a secure database using Microsoft Excel 2010. Ethical approval
was obtained from the University of the Witwatersrand Human
Research Ethics Committee (Medical).
Results
The authors of this study set out to determine how diabetes care
compared across different settings, given the healthcare challenges
faced especially by under-resourced areas. Of the 511 (154 from
Ovid MEDLINE + 32 Pubmed + 325 Sabinet) titles initially identified
between 2009 and 2014, 14 studies fulfilled the inclusion criteria.
These 14 studies originated from 19 different countries (some
studies included more than a single country) and we enrolled a
total of 25 629 patients.
There were 17 high-income, one upper-middle- (South Africa)
and one low-income (Uganda) country included in the review
(grouped according to the United Nations’ economies by per-capita
country classification).7 Cohort sizes ranged from 50 to 4 926
patients. Twelve studies contained results for all major risk factors
(HbA1c, BP and LDL-C), while the rest included at least two-thirds
of the measured risk factors. There were eight studies (57.1%) that
defined treatment targets as per the AmericanDiabetes Association.8
The characteristics of each study are outlined in Table 1.
In 12 studies (25 354 patients) that used an HbA1c level of 7.0%
or less to define control, 44.5% (range 19.2–70.5%) of patients
achieved target.9-20 In two studies (275 patients) where HbA1c level
was defined as < 6.5 and < 8.0%, respectively, 56.6 and 60.0% of
patients reached their targets, respectively.21,22
In eight studies (18 089 patients), which had the definition of
target BP of 130/80 mmHg or less (systolic and diastolic combined),
35.2% (range 7.4–66.3%) of patients achieved target.9-11,16-19,21
In four studies (7 240 patients) where systolic BP targets of 130
mmHg or less (alone) defined control, 32.7% (range 21.3–50.0%)
of the subjects achieved target.12,13,15,20 In two studies (300 patients)
with a BP target of either < 140/90 or < 140/80 mmHg, 24.0 and
56.0% of patients achieved goal, respectively.14,22
In the 11 studies (24 702 patients) that used LDL-C levels of
either 2.5 or 2.6 mmol/l (100 mg/dl or less) to define control, 51.4%
(range 20.0–82.9%) of patients achieved goal.10-20 One study (225
patients) with a total cholesterol target of < 200 mg/dl (5 mmol/l)
had 49.3% of patients at goal.21 Two studies (702 patients) did not
measure lipid levels.9,22
In general, more patients reached target for LDL-C than for
HbA1c levels, with the poorest achievement of targets being BP.
The widest variability of target achievement was LDL-C (variation
of 62.9%), followed by BP and then HbA1c (least variability). The
highest and lowest achieved targets were those by an American
(LDL-C, 82.9%) and a German study (BP, 7.4%), respectively.18,19
Table 1. Study characteristics
First author
(reference)
Year of
publi-
cation Location
Cohort
size
(n)
Achievement of target
HbA1c BP (< 130/ LDL-C
(< 7%) 80 mmHg) (< 2.6 mmol/l)
Al-Taweel9 2013 Kuwait 652 19.2 46.0 –
Braga10 2012 Canada 3002 52.6 53.6 64.2||
Casagrande11 2013 USA 4926 52.5 51.1 56.2
Goderis12 2009 Belgium 2495 54.0 50.0‡ 42
Hermans13 2013 Belgium,
Greece,
Luxem-
bourg,
Portu-
gal,
Spain,
UK
3996 49.2 27.3‡ 40.8
Kibirige14 2014 Uganda 250 20.8 56.0§ 20.0
Klisiewicz15 2009 South
Africa
150 30.7 21.3‡ 50.7||
Lee16 2009 Korea 926 49.2 66.3 51.0
Morren21 2010 Trinidad 225 56.6† 53.6 49.3**
Pinchevsky17 2013 South
Africa
666 26.2 45.8 53.8||
Sease18 2013 USA 95 35.8 62.1 82.9
Stone19 2013 Belgium 1044 59.7 27.6 49.7
Stone19 2013 France 1056 65.3 14.9 52.4
Stone19 2013 Germany 959 48.6 7.4 30.7
Stone19 2013 Ireland 950 53.4 24.9 76.9
Stone19 2013 Italy 984 35.7 20.8 40.4
Stone19 2013 Nether-
lands
1021 70.5 20.3 58.9
Stone19 2013 Sweden 550 56.5 27.1 47.3
Stone19 2013 UK 1033 39.1 25.0 74.5
Umar-
Kamara22
2011 USA 50 60.0* 24.0¶ –
Webb 20 2014 South
Africa
599 27.0 32.0‡ 33.0||
Exceptions to the above targets are indicated by the following:
†HbA1c < 6.5%; *HbA1c < 8.0%;
‡systolic blood pressure only < 130 mmHg;
§systolic/diastolic blood pressure < 140/80 mmHg; ¶systolic/diastolic blood
pressure < 140/90 mmHg;
||low-density lipoprotein cholesterol < 2.5 mmol/l;
**total cholesterol < 5.18 mmol/l.