The SA Journal Diabetes & Vascular Disease Vol 12 No 1 (July 2015)

VOLUME 12 NUMBER 1 • JULY 2015

SA JOURNAL OF DIABETES & VASCULAR DISEASE

RESEARCH ARTICLE

Statistically significant factors associated with high LDL-C levels

included ethnicity, hypertension, DM, and the presence of coronary

and cerebrovascular heart disease. Factors associated with low

HDL-C levels were a high waist circumference, DM and being

treated by a specialist. Elevated TG was associated with female

gender, obesity, DM and peripheral artery disease. However, the

only statistically significant factors independently associated with

the presence of all three lipid abnormalities were obesity and Asian

as well as mixed-ancestry ethnicity.

Based on the current data, it is unclear whether the findings in

regard to ethnicity are biologically or sociologically determined. Even

though this study was conducted exclusively in the private healthcare

sector in South Africa, Asian or mixed-ancestry ethnicity most likely

still correlates partially with social deprivation, which has been shown

to be a risk factor for cardiovascular disease. Social deprivation may

also affect access to medical care, with less access to specialist care

and a bias towards less aggressive treatment. Studies from other

countries have shown that ethnic minorities or immigrants often

receive less aggressive cardiovascular care,

as also observed in this

study, with black patients receiving lower-dose potency of statins,

despite the majority of patients being at high risk.

Socio-economic status has also been associated with statin

adherence,

as has ethnicity.

In the South African context, lower

socio-economic status would, for instance, often correlate with

membership of a medical scheme option that restricts lipid-lowering

treatment to less-potent (and less-costly) options. Lower income

may also influence the willingness and ability to pay ‘co-payments’

that are often required to access more potent lipid-lowering

therapy. However, factors such as provider bias, access to treatment

and differential adherence do not completely explain the observed

ethnic differences, as black patients generally still experience the

highest level of socio-economic deprivation as a legacy of South

Africa’s past history.

Lesser goal attainment may also in part be due to differences in

baseline lipids. In the Heart of Soweto study, there were significant

differences in untreated lipid profiles by ethnicity in patients

presenting for cardiovascular care

at a tertiary referral centre. The

odds ratio (compared to black patients) for elevated LDL-C levels

in Asian and mixed-ancestry patients was 4.66 and 2.44 mmol/l,

respectively. Indian and mixed-ancestry patients also had higher

median TG levels (1.8 and 1.4 mmol/l, respectively) than black

patients (1.1 mmol/l).

In addition to identifying factors that are associated with

dyslipidaemia in statin-treated patients, DYSIS in South Africa

(along with previous DYSIS studies) also highlights the deficiencies

of lipid-lowering therapy in clinical practice. Other researchers

analysing the efficacy of lipid-lowering therapies have supported

this conclusion,

35,36

including another recent study analysing statin-

treated South African patients.

Together, these findings suggest

that there is a need to improve upon existing treatment strategies

(e.g. combination of current therapies for optimal patient efficacy,

utilisation of more-potent statins, improving adherence) while also

developing novel therapeutic approaches.

Combination therapieswere evaluated in theAustrianCholesterol

screening and Treatment (ACT) II study, which evaluated the effect

of lipid-lowering therapies in high-risk, statin-treated patients with

elevated LDL-C levels. Interestingly, combination therapy consisting

of simvastatin and ezetimibe (used for 73% of patients in the ACT

II study) resulted in 40.3% of patients meeting their LDL-C goals,

with a decline in LDL-C levels from a baseline of 31.3% following

12 months of intensified therapy.

High-dose statins are another option to achieve LDL-C targets

in high-risk patients.

38,39

Improving adherence is a challenge that

physicians face every day, and some strategies that have shown

promise include regular phone calls by a practice nurse, regular

review by a community pharmacist and providing a medication

Table 4.

Factors independently associated with LDL-C, HDL-C and TG abnormalities: results from multiple regression analyses (or, 95% CI)

LDL-C not at target*

†

(≥ 1.8/2.5/3.0 mmol/l)

Low HDL-C*

[< 1.0 (m)/1.2 (w) mmol/l]

Elevated TG*

(>1.7 mmol/l)

LDL-C not at target, low

HDL-C, elevated TG*

Age ≥ 70 years

0.57 (0.43–0.77)

Female

0.43 (0.32–0.58)

1.33 (1.02–1.74)

Asian vs Caucasian

2.48 (1.19–5.16)

Black vs Caucasian

Mixed ancestry vs Caucasian

2.12 (1.36–3.32)

2.78 (1.50–5.19)

Alcohol consumption > 2 units/week

0.50 (0.31–0.79)

BMI ≥ 30 kg/m² (obesity)

1.74 (1.33–2.29)

2.11 (1.27–3.50)

WC > 102 (m)/> 88 cm (w)

1.71 (1.26–2.32)

Hypertension

1.55 (1.12–2.13)

Diabetes mellitus

1.36 (1.01–1.82)

1.58 (1.17–2.15)

1.49 (1.12–1.98)

Cerebrovascular disease

1.89 (1.39–2.57)

Peripheral artery disease

2.35 (1.09–5.07)

Specialist (Card/Endo/Dia/Int/Oth)

2.01 (1.46–2.76)

*Models contained the following variables: age, gender, ethnicity, 1st-grade family history of premature CVD, current smoker, sedentary lifestyle, alcohol

consumption > 2 units/week, BMI ≥ 30 kg/m² (obesity), waist circumference > 102 cm in men/> 88 cm in women, hypertension, diabetes mellitus, coronary heart

disease, cerebrovascular disease, heart failure, peripheral artery disease, RR ≥ 140/90 mmHg (systolic/diastolic), 20–40 vs 10 mg/day simvastatin equivalent, ≥ 80

vs 10 mg/day simvastatin equivalent, ezetimibe.

Backward selection (alpha = 0.05) was done.

†

Patients with SCORE risk 1–4%: LDL-C ≥ 3.0 mmol/l; patients with SCORE risk 5–9%: LDL-C ≥ 2.5 mmol/l; patients with CVD, DM, and/or SCORE risk ≥ 10%:

LDL-C ≥ 1.8 mmol/l

Card = cardiologist, Endo = endocrinologist, Dia = diabetologist, Int = internist, Oth = other speciality, ns = not significant (

> 0.05), OR = odds ratio, CI =

confidence interval.