Dyslipidaemia in rural Australia: prevalence, awareness, and adherence to treatment guidelines in the Greater Green Triangle Risk Factor Study

Cardiovascular disease (CVD) accounts for 18% of the total burden of disease in Australia,1 and the burden of CVD is higher in rural areas.2 Large clinical trials have demonstrated that treatment of hypercholesterolaemia is effective in both primary and secondary prevention of CVD,3,4 and target lipid goals have been progressively lowered, reflecting recent trial evidence. Treatment of hyperlipidaemia should be based on assessment of global CVD risk rather than plasma cholesterol levels alone, as high-risk groups benefit most.5

Little is known about the prevalence of risk factors or treatment patterns for CVD in rural areas.2 This is the first report of rural Australian population lipid levels, hyperlipidaemia awareness and assessment of adherence to more recent 20015 and 20056 Australian guidelines.

Three cross-sectional population surveys of chronic disease risk factors and related health behaviour were undertaken in the Greater Green Triangle region in South Australia and Victoria, in 2004–2006; these have been described previously.7,8 Each survey took a random sample from the electoral roll of people aged 25 to 74 years, stratified by 10-year age group and sex, with the exception of those aged 25–44 years, who were considered as one stratum. Strata sample sizes were 140 from the Limestone Coast (2004), 125 from Corangamite (2005) and 150 in Wimmera (2006). Analysis for this article was performed on data from participants who had all information required for 5-year CVD risk calculation. Aboriginal and Torres Strait Islander participants (n = 11) were not included in this analysis due to under-representation in the surveys and as the available risk calculators, based on Framingham data, have not been validated for use in this population.6

The survey methodology closely followed the World Health Organization MONICA (multinational MONItoring of trends and determinants in CArdiovascular disease) Project protocol and European Health Risk Monitoring recommendations.9,10 A postal questionnaire examined history of chronic disease, medication use and lifestyle habits. Health checks, including anthropometric measurements and venous blood sampling (after fasting for ≥ 10 hours), were carried out at survey sites by specially trained nurses. Blood samples were analysed at Flinders Medical Centre Clinical Trials Laboratory.7

Participants were divided into five mutually exclusive groups based on presence of established CVD or diabetes, lipid-lowering treatment, and CVD risk (Box 1). The 2001 guideline therapeutic target levels used were < 4 mmol/L for total cholesterol (TC), < 2.5 mmol/L for low-density lipoprotein cholesterol (LDL-C), > 1.0 mmol/L for high-density lipoprotein cholesterol (HDL-C) and < 2.0 mmol/L for triglycerides (TG) updated to < 1.5 mmol/L for TG in the 2005 guidelines.5,6 The extent to which a lower 2007 guideline LDL-C target (< 2.0 mmol/L) was achieved among patients with established CVD or diabetes, was also examined.11 These cut-off levels differ from those used to define metabolic syndrome.7

Ethics approval was received from the Flinders Clinical Research Ethics Committee.

Statistical analyses were undertaken using Stata, version 10.1 (StataCorp, College Station, Tex, USA). Age-specific means and prevalence data were calculated using the commands mean and CI, with 95% confidence intervals for prevalence data based on the Agresti–Coull interval. Analyses involving combined age groups were weighted to the local population and utilised survey-specific modules. Overall means and prevalence data were calculated using the commands mean and tabulate. P values for the differences between groups I–V and I–IV were calculated with the test command, following linear regression (continuous outcomes) or logistic regression (dichotomous outcomes) including age and sex in the models as covariates.

Overall, 3320 individuals were selected. After excluding 108 who had died or had left the area, there were 1563 participants (48.7%). Information on blood lipids, lipid-lowering medication, CVD risk, and history of CVD and diabetes was available for 1274 participants (40% of sample). Categorisation of participants is shown in Box 2. Overall population-weighted mean TC concentration was 5.38 mmol/L (95% CI, 5.30–5.45). Box 3 shows detailed results by age and sex. TC concentration exceeded 5.5 mmol/L for 39% of participants and 6.5 mmol/L for 13%. A wider definition of hypercholesterolaemia, constituting those with a TC concentration > 5.5 mmol/L or on cholesterol-lowering medication, gave a prevalence of 48%.

Overall population-weighted mean TG, LDL-C and HDL-C concentrations were 1.50 mmol/L (95% CI, 1.43–1.56), 3.23 mmol/L (95% CI, 3.16–3.30) and 1.46 mmol/L (95% CI, 1.44–1.49), respectively. Almost a third of the population had LDL-C concentration > 3.5 mmol/L, and prevalence of high TG and low HDL-C concentrations were 17% and 7%, respectively. In the population not on treatment, overall mean concentration of TC was 5.43 mmol/L (95% CI, 5.35–5.52), LDL-C was 3.30 mmol/L (95% CI, 3.22–3.38), HDL-C was 1.47 mmol/L (95% CI, 1.44–1.49), and TG was 1.45 mmol/L (95% CI, 1.38–1.52).

In total, 231 participants (12% population weighted) reported using lipid-lowering medication (Box 2). Of the 208 for whom lipid-lowering medication details were known, 199 (96%) were using statin mono-therapy. Of the 1043 participants who were not treated, 77 (7%) had established vascular disease or diabetes, 397 (38%) had high 5-year risk of CVD and 569 (55%) had low 5-year risk.

Untreated high-risk participants without CVD or diabetes (Group IV) had a mean TC level of 6.37 mmol/L and a mean LDL-C level of 4.14 mmol/L (Box 4). Between 19% and 30% of untreated high-risk individuals (groups II and IV, respectively) reported doctor-diagnosed high cholesterol. Almost all individuals being treated had their cholesterol measured in the previous year, compared with about half for untreated high-risk participants. Overall, three-quarters of participants aged 45 years or older had their cholesterol measured in the previous 5 years.

Three per cent of the high-risk study population had all lipids at target (Box 5). HDL-C and 2001 TG target levels were reached by most high-risk participants. TC and LDL-C levels were largely inadequate; treated patients with established CVD or diabetes (Group I) were most likely to be controlled.

Box 6 compares lipid levels from our study with those found in previous Australian studies.

In this rural population, a comprehensive examination of lipid levels, prevalence and assessment of adherence to guidelines for dyslipidaemia was completed, examining both the 2001 Australian guidelines and the 2005 guidelines, which became available shortly before the end of the study.

The National Health and Hospitals Reform Commission (NHHRC) has proposed tracking national progress towards tackling health inequities.16 Ample evidence exists to show that access to services is poorer in rural and remote areas than in metropolitan areas. Differences exist in statin prescribing in Australia after adjusting for age, sex and socioeconomic circumstance, with the highest prescribing rates seen in metropolitan areas.17 The NHHRC has proposed the establishment of a National Prevention Agency to build an evidence base, with the capacity and infrastructure to report on progress towards achieving better health outcomes.

This article represents an example of the baseline data a National Prevention Agency would require. Periodic follow-up to track progress would also be needed — without it, health inequality due to rurality will persist.

The participation rate in our study was low, so some caution is needed when interpreting the results. A comparison of the socioeconomic background — including primary occupation, rate of unemployment and total gross income of the survey participants — with population statistics available indicated that the participants closely represented the true populations of the areas surveyed.7 The participation rate was also comparable to those of other studies of this type.15,18,19 We cannot be sure of the effect that non-respondents would have on the results of this study. We suggest that as non-respondents typically have higher morbidity and mortality rates than respondents,20 their absence from our study would, if anything, result in a more conservative conclusion.

Analysing trends in blood lipid concentrations and prevalence of hyperlipidaemia in Australia over the past three decades is complex because of opposing changes in factors influencing blood lipid levels and methodological issues between the studies. The prevalence of overweight and obesity, dietary factors, exercise patterns, alcohol consumption and use of lipid-lowering medication all influence plasma lipid concentrations. Of particular note is the dramatic increase in prevalence of overweight and obesity, especially central obesity, over the past two decades.7 Increased body mass index is a causative factor of higher TG and lower HDL-C levels. The predominant non-genetic determinants of increased TC and LDL-C levels are dietary saturated fats, whereas TC and LDL-C levels are reduced by increasing dietary polyunsaturated fat and water-soluble fibre content.21

It is currently unknown whether there are differences in risk factor prevalence between metropolitan and rural populations. Although differences in HDL-C analysis methods exist between studies, results are still comparable. Similar methods were used for all studies for TC and TG analyses. Early studies by the National Heart Foundation of Australia (NHFA) did not estimate LDL-C (now the key target for modification), and were confined to metropolitan populations.12-14 AusDiab (the Australian Diabetes, Obesity and Lifestyle Study) did not report rural prevalence.15

A major concern is that in four large, primarily urban population studies there have been no improvements in mean TC levels since the early 1980s, and our study showed only a 0.2 mmol/L decline (Box 6). Although even a small improvement in the population mean favourably affects overall population CVD risk, the decline in the mean TC levels achieved is disappointing compared with what can be achieved.22 Mean HDL-C levels have shown minimal change (about 0.1 mmol/L). Mean TG levels appear to have deteriorated, especially among women (about 0.35 mmol/L), among whom increasing overweight and obesity is a major issue.7

The prevalence of hypercholesterolaemia (TC > 5.5 mmol/L) that we observed (39%) is lower than in previous studies (range, 47% to 51%),12-15 but the wider definition that includes both those with TC > 5.5 mmol/L or on cholesterol lowering medication reveals a prevalence of 48% in our study, unchanged from the 1980s.

Our study, completed just after the release of the 2005 Australian lipid guidelines in November 2005, provides unique opportunities to quantify adherence to guidelines in Australia. Previous Australian NHFA/CSANZ (Cardiac Society of Australia and New Zealand) lipid management guidelines (2001),5 which focus on absolute risk rather than lipid cut-offs, along with those from New Zealand, are considered among the most potentially effective in the world in reducing CVD-related deaths.23 This emphasis on absolute risk further enhances the effectiveness of treatment.6,11 A recent study of the AusDiab 1999–2000 population, primarily focusing on primary prevention, strongly argued this point and noted limitations in the corresponding Pharmaceutical Benefits Scheme criteria for treatment.24 Optimal implementation of current NHFA/CSANZ guidelines would further enhance the cost-effectiveness of lipid-lowering medication in Australia.

There has been a dramatic increase in lipid lowering medication use, from minimal in the 1980s, to 7.1% (women) and 8.2% (men) in 1999–2000,24 and 12% in our study. Lipid-lowering medication is almost solely comprised of statins, which have potent effects on TC and LDL-C levels, but little effect on HDL-C or TG levels. Dispensed prescriptions for lipid-lowering medications have trebled, from 6.1 to 18.0 million between 1997 and 2005 (Ms Debra Rowett, Director, Drug and Therapeutics Information Service, Repatriation General Hospital, Adelaide, personal communication, 2008). The total cost of lipid-lowering medications (including patient co-payments) in 2005 in Australia was $1.16 billion.

Results from our study using absolute risk show that current rates of detection and treatment are suboptimal. Many individuals requiring secondary or primary prevention were untreated. Even among those at high risk who were treated, a high proportion were not treated to targets. Elevated TG levels, especially according to the 2005 guidelines, were a major issue in all groups, and there was little use of fibrates, which is the most effective medication class for high TG levels.

Our results demonstrate the need for a national lipid guideline implementation program focused on primary care coupled with universal adoption of primary care chronic disease management systems such as the Australian National Primary Care Collaboratives.25 Strategies aimed at reducing mean TC levels in the whole population have a greater impact on overall burden of CVD than targeted treatment of high-risk individuals alone.5,26 The findings that mean Australian population TC has improved only a little over the past two decades reflect a lack of effective health promotion strategies, and an absence of population health monitoring and surveillance as recommended by the NHHRC. If better information becomes available about the impact of the strategies implemented, these initiatives can be modified or new ones implemented and the impact can be measured. It may be that some specific strategies will be required in rural areas and that not all rural areas will prove to be the same.

The balanced application of both targeted high-risk and population strategies is essential for the effective prevention of CVD. The proposed National Prevention Agency could undertake population-based strategies and also promote the implementation of evidence so that a high proportion of the population at risk of CVD is treated to the targets set out in national guidelines.

We hope that the National Prevention Agency will work to reduce the health inequality of high rates of CVD in rural and remote Australia.

2 Categorisation of participants by CVD risk status and lipid-lowering medication use

CVD = cardiovascular disease. NZ = New Zealand.