Short term efficacy of a lifestyle intervention programme on cardiovascular health outcome in overweight Indigenous Australians with and without type 2 diabetes mellitus: The healthy lifestyle programme (HELP)

1. Introduction 

Indigenous Australians are at higher risk for cardiovascular disease (CVD) and type 2 diabetes mellitus (T2DM) compared to the general Australian population [1], [2]. Combined with psychosocial factors, this greater disease burden results in a dramatic 16–20 years decrease in Indigenous life expectancy [3]. Several longitudinal studies have shown that certain risk factors strongly associated with mortality and morbidity are highly modifiable [4], [5], [6]. The strongest of these modifiable risk factors include systolic blood pressure [4] and LDL cholesterol [7], [8]. Other cardiovascular risk factors discovered through several studies include glycaemic control in diabetics [5], homocysteine [9], microalbuminuria [10] and obesity [11].

The Multiple Risk Factor Intervention Trial and Steno-2 Study have shown that lifestyle intervention has a significant benefit on mortality and morbidity outcomes for both CVD [6] and T2DM [12] which are correlated primarily with favorable modifications of CVD risk factors such as systolic blood pressure and LDL cholesterol. Recent longitudinal data from the INTERHEART [13] and Dubbo [14] studies have re-emphasized the impact of three eminently modifiable risk factors: smoking, diabetes control and hypertension on survival time.

Although costs involved in delivering and maintaining a lifestyle intervention programme are prohibitive on a national level, targeting high risk groups such as Indigenous Australians could yield a net cost benefit. This is due largely to decreases in cost of managing the complications of T2DM and CVD [15]. There are likely to be many strata of cardiovascular risk within the Indigenous population and identification of very high risk individuals is difficult, with evidence that conventional risk stratification algorithms based on Caucasian population studies tending to underestimate cardiovascular risk [16].

Few studies have examined the effects of lifestyle intervention on Indigenous Australians with T2DM [17], [18]. The healthy lifestyle intervention project (HELP) will increase the pool of knowledge in Indigenous health; in particular the response of CVD and T2DM control measures to a lifestyle intervention programme. To our knowledge, the HELP is a novel study examining the effects of a lifestyle intervention programme on cardiovascular and diabetic health outcomes in urban Indigenous Australians. The primary aim of this study was to determine the effectiveness of lifestyle intervention on improving diabetes and cardiovascular risk factors.

2. Methods 

2.2. Clinical and biochemical data 

Baseline measurements included a detailed medical history, physical examination and laboratory testing. Presence of T2DM, CVD, smoking and socioeconomic measures were recorded. Physical parameters recorded were weight, height, waist circumference, hip circumference and blood pressure. Waist circumference was taken at the midpoint between the iliac crest and the lower ribs. Hip circumference was measured at the point of maximum protuberance of the buttocks. Blood pressure was measured using an Omron sphygmomanometer (Omron, Illinois, USA) on rested subjects in the seated position.

Subjects were required to abstain from smoking and all oral intake, excepting plain water, for 8h prior to venesection. Laboratory testing was carried out in Queensland Health Pathology and Scientific Services laboratories in The Prince Charles Hospital and the Princess Alexandra Hospital, Brisbane, Australia. Measures assayed for were plasma glucose, HbA1c, lipid profile (LDL cholesterol, HDL cholesterol, triglycerides), homocysteine, c-peptide and serum creatinine. Plasma glucose was measured using a hexokinase method, cholesterol and triglycerides with a peroxidase mediated oxidative reaction in a Dade Behring RxL automated clinical chemistry analyser (Dade Behring Inc., Illinois, USA). HbA1c was assayed by ion exchange high performance liquid chromatography in a Biorad Variant 2 analyser (Biorad, Sydney, Australia). Lastly, homocysteine was measured by fluorescence polarisation immunoassay on the AxSYM analyser (Abbott, Illinois, USA). Urine samples were collected for determination of urine albumin to creatinine ratio.

Insulin resistance was calculated from fasting serum c-peptide and fasting glucose concentrations using the homeostasis model assessment of insulin resistance (HOMA-IR) index [21]. The HOMA-IR index has been validated in normal range weight and moderately obese type 2 diabetic patients [22]. Creatinine clearance (mL/min) was estimated with the Cockcroft–Gault formula:

3. Results 

There were 101 study participants, 44 (44%) had T2DM, 11 (11%) had impaired fasting glucose (IFG) and the remainder had neither T2DM nor IFG.

3.1. Baseline 

Predictably, participants with T2DM were significantly older (57 years versus 43 years, P<0.001), had higher FPG (8.0mmol/L versus 5.0mmol/L, P<0.001), HbA1c (7.3% versus 5.0%, P<0.001), triglycerides (2.1mmol/L versus 1.5mmol/L, P=0.002) and homocysteine (11.1μmol/L versus 10.0μmol/L, P=0.045) than non-diabetic participants (Table 1). These findings are consistent with an age dependent increase of T2DM prevalence and cardiovascular disease risk found in longitudinal studies of T2DM. Notably, indices of renal function were significantly poorer in the diabetic group, as reflected by significantly lower creatinine clearances (73mL/min versus 83mL/min, P=0.003) and higher urinary albumin to creatinine ratios (10.8 versus 1.2, P=0.002). Non-diabetic subjects exhibited a significantly higher LDL cholesterol (3.2mmol/L versus 2.8mmol/L, P=0.03), which is likely due to significantly higher rates of HMGCoA reductase (statin) use in the diabetic group (47.5% versus 4.8%, P<0.001). Non-diabetic subjects had lower insulin resistance (2.0 versus 3.2, P<0.001) compared to the diabetic group. The diabetic group had significantly higher waist circumference and body mass index compared to the non-diabetic group, however, there were no significant differences in waist hip ratio between the two groups.

Table 1. Baseline differences between diabetic vs. non-diabetic vs. IFG groups

	Variable	Diabetic, n=44	Non-diabetic, n=46	P (diabetic vs. non-diabetic)	IFG, n=11	P (diabetic vs. IFG)
	Female (%)	77.3	75.6	0.99	54.5	0.13
	Clinical CVD (%)	18.6	2.5	0.03a	18.2	0.99
	Current smoker (%)	26.2	42.5	0.17	30.0	0.21
	AMb use (%)	60.0	6.5	<0.001a	50.0	0.57
	Statin use (%)	47.5	4.8	<0.001a	40.0	0.74
	Metformin use (%)	62.5	4.8	<0.001a	0.0	0.001a
	Age (years)	56.5 (52.8–60.1)	43 (39 to 47)	<0.001a	49.8 (40 to 60)	0.12
	Waist circumference (cm)	108 (103–112)	98 (91 to 104)	0.01a	109 (93 to 126)	0.91
	Waist hip ratio	0.92 (0.88–0.95)	0.88 (0.82–0.93)	0.19	0.90 (0.79–1.00)	0.61
	Body mass index (kg/m2)	34 (32–36)	30 (28–33)	0.03a	36 (27–45)	0.99
	Systolic BP (mmHg)	132 (127–139)	125 (120–130)	0.04a	132 (123–142)	0.94
	Diastolic BP (mmHg)	85 (81–89)	84 (80–87)	0.56	84 (79–89)	0.94
	MABP (mmHg)	101 (97–104)	97 (93–101)	0.16	100 (96–105)
	Diet Scorec	16.3 (15.3–17.3)	16.0 (14.7–17.3)	0.70	16.5 (10.2–22.9)	0.92
	Steps	6148 (4620–7675)	7163 (5503–8824)	0.39	12430 (8179–16680)	0.03a
	Fasting plasma glucose (mmol/L)	8.0 (7.2–8.7)	5.0 (4.8–5.2)	<0.001a	6.1 (5.9–6.3)	0.01a
	HbA1c (%)	7.3 (6.9–7.7)	5.5 (5.4–5.7)	<0.001a	6.1 (5.8–6.4)	0.001a
	HOMA IRd	3.2 (2.7–3.6)	2.0 (1.7–2.3)	<0.001a	2.7 (2.0–3.4)	0.53
	Total cholesterol (mmol/L)	4.9 (4.6–5.2)	5.1 (4.8–5.4)	0.23	4.8 (4.4–5.2)	0.82
	LDL cholesterol (mmol/L)	2.8 (2.6–3.0)	3.2 (2.9–3.5)	0.03a	2.8 (2.3–3.2)	0.94
	HDL cholesterol (mmol/L)	1.1 (1.0–1.2)	1.3 (1.2–1.4)	0.04a	1.1 (0.9–1.3)	0.75
	Triglycerides (mmol/L)	2.1 (1.8–2.4)	1.5 (1.2–1.7)	0.002a	2.5 (1.1–4.0)	0.34
	Fasting homocysteine (μmol/L)	11.1 (10.3–12.0)	10.0 (9.2–10.8)	0.045a	11.4 (10.0–12.7)	0.80
	Calculated creatinine clearance (mL/min)	73 (65–80)	83 (75–91)	0.003a	78 (45–111)	0.04a
	Urinary albumin–creatinine ratio	10.8 (3.6–18.0)	1.2 (0.7–1.7)	0.002a	0.8 (0.5–1.1)	0.058

Results for continuous variables are reported as mean (95% CI), results of categorical variables are reported as percentages.

a Statistical significance P<0.05 (two-tailed). b Angiotensin system modulators (includes ACE inhibitors and AII receptor antagonists). c Higher scores indicate healthier diets. d Homeostasis model assessment of insulin resistance.

Marked similarities between the T2DM and IFG groups (Table 1) can still be appreciated despite the small numbers present in the IFG group causing a reduction in power of discriminatory tests. Besides definitional glycaemic differences, the IFG group shared a similar cardiovascular risk profile with the diabetic group, except for the significantly better renal function found in the IFG group as marked by a calculated creatinine clearance close to that of the non-diabetic group. Other cardiovascular risk factors were not statistically significant between the two groups. Insulin resistance was also not significantly different between diabetic and IFG groups (HOMA IR 3.2 and 2.7, respectively, P=0.53).

3.2. Six-month follow up 

A summary of the short-term efficacy of the HELP intervention programme is shown in Table 2. Eighty (79%) were available for assessment during the follow up period. There were no significant differences in smoking pattern, ACEM use or statin use recorded in follow up diaries. In general, there were small but significant improvements in waist circumference, diastolic blood pressure, mean arterial blood pressure, total cholesterol and triglyceride levels. There were small but statistically significant disimprovements in HbA1c and HDL levels. However, due to the small magnitude of these changes, it is likely that only the improvements in waist circumference (reduced by 3.1cm from baseline) and diastolic blood pressure (reduced by 4.6mmHg from baseline) are of clinical relevance at this time point.

Table 2. Change from baseline at 6-month follow up for participants (n=80)

	Variable	Mean difference from baseline (95% CI)	P
	Waist circumference (cm)	−3.1 (−0.7 to −5.4)	0.01a
	Waist hip ratio	−0.03 (−0.06 to 0.01)	0.14
	Body mass index (kg/m2)	−0.4 (−0.80 to 0.05)	0.09
	Systolic BP (mmHg)	−3.3 (−9.0 to 2.3)	0.25
	Diastolic BP (mmHg)	−4.6 (−1.2 to −8.0)	0.01a
	MABP (mmHg)	−4.2 (−0.6 to −7.7)	0.02a
	Diet Score	0.1 (−0.9 to 1.1)	0.77
	Steps	261 (−642 to 1166)	0.552
	Fasting plasma glucose (mmol/L)	0.14 (−0.09 to 0.38)	0.234
	HbA1c (%)	0.31 (0.18 to 0.44)	<0.001a
	HOMA IR	−0.13 (−0.09 to 0.38)	0.113
	Total cholesterol (mmol/L)	−0.26 (−0.10 to −0.42)	0.002a
	LDL cholesterol (mmol/L)	−0.09 (−0.25 to 0.07)	0.26
	HDL cholesterol (mmol/L)	−0.09 (−0.05 to −0.13)	<0.001a
	Triglycerides (mmol/L)	−0.18 (−0.02 to −0.33)	0.03a
	Fasting homocysteine (μmol/L)	0.24 (−0.29 to 0.76)	0.37
	Creatinine clearance (mL/min)	1.24 (−3.2 to 5.7)	0.58
	Urine ACR	−1.3 (−4.3 to 1.6)	0.37

a Statistical significance P<0.05 (two-tailed).

4. Conclusions 

The diabetic group of Indigenous Australians had, in general, a more unfavorable cardiovascular risk profile compared to the non-diabetic group. The exception was LDL cholesterol being lower in the T2DM group, this is probably due to the institution of a more aggressive lipid management plan for these diagnosed diabetics by their primary care physicians through the use of statins. Comparisons at baseline were consistent with an age dependent increase of T2DM prevalence and cardiovascular risk found in longitudinal studies of T2DM.

With regards to cardiovascular risk management targets for diabetes in Australia [24], management for our T2DM group remains suboptimal. However, the mean CVD risk parameters of the T2DM group were very close to set targets, and the likelihood of achieving them in the near future is high.

Renal function is a strong independent cardiovascular risk factor in people with and without T2DM [25], [26]. In our study participants, we found a significantly reduced level of renal function as reflected by calculated creatinine clearance and urinary albumin and creatinine ratio. Urinary albumin to creatinine ratio is eight times greater in the diabetic group compared to the non-diabetic group. This emphasizes the urgent need to adequately manage glycaemic control and hypertension in T2DM to prevent deterioration in renal function.

Preservation of renal function and definitional differences in fasting plasma glucose levels are the main differences between the T2DM and IFG groups. Insulin resistance in both groups are markedly elevated and are not significantly different between these two groups. Measures of obesity and age are non-significant between these groups, it is probable that this high degree of insulin resistance in the IFG group will be translated to the development of T2DM in the future. This temporal difference in development of T2DM may have a genetic cause and warrants further investigation [27].

The cardiovascular risk profile of our small IFG group as reflected by blood pressure and lipid profile is not significantly different to that of the T2DM group. Our findings support a more aggressive preventative measures in the IFG group. The Australian Diabetes Society draft summary recommendations for management of prediabetes recommends oral hypoglycaemics for overweight Australian with prediabetes and lifestyle modifications. Angiotensin converting inhibitors or angiotensin receptor antagonists should be first line antihypertensive medications in this population, as they have been shown to be particularly efficacious in preserving renal function and improving health outcome in people with T2DM [28].

In terms of the short-term efficacy endpoints, we found general trends towards improvement of cardiovascular risk factors. Most of the changes are small in magnitude and are unlikely to be clinically relevant at this stage with the exception of the detected reductions in waist hip ratio and diastolic blood pressure.

Waist hip ratio has been proposed to be a superior index of cardiovascular risk in Australians [29]. In our Indigenous Australian population with high prevalence of central adiposity and T2DM, it may be a more subtle and accurate means of assessing follow up and clinically relevant effects of our intervention programme. We recorded a 4mmHg reduction in diastolic blood pressure from baseline at our 6-month follow up time point despite no significant changes in recorded antihypertensive use during the 6 months. MacMahon et al. established in 1990 after correcting for regression dilutional bias, that a 5mmHg reduction in diastolic blood pressure would correlate to 34% less stroke and 21% less coronary heart disease [30].

Our findings would have been strengthened with the establishment of a control group. Unfortunately, the Indigenous communities approached during establishment of this study declined to be involved as measurement only control groups. Without this group, it was difficult to assess the significance of the 0.3% disimprovement in HbA1c found after 6 months. As diabetes medication use did not change significantly during this time, this disimprovement may reflect a temporal variation in HbA1c due to differences in food intake in the holiday season immediately preceding follow up. This trend will be investigated further at the next time point.

In summary, our baseline findings indicate that our urban Indigenous Australian participants with T2DM are close to meeting management targets but have poorer renal status compared to participants without T2DM. Indigenous Australians with IFG had a similar cardiovascular risk profile as Indigenous Australians with T2DM with the preservation of renal function although this observation is based on a very small sample size. It may be beneficial to institute more aggressive preventative strategies in Indigenous Australians with IFG aimed at preserving renal status and minimization of cardiovascular risk. In the short term, our 6-month follow up results show a general trend towards improvement of cardiovascular end points, including clinically relevant changes in waist circumference and diastolic blood pressure.