To the Editor: In a recent article, Thompson and colleagues provided useful information on the prevalence of cardiovascular risk factors in urban Aboriginal people.1 Using the Sheffield table of absolute risk,2 the authors estimated that “15% men and 6% women had an absolute risk > 15% of a cardiovascular event within 10 years”.

The Sheffield risk table was developed for assessing the risk of coronary deaths rather than the risk of cardiovascular events.2 Moreover, the validity of applying the Sheffield table and other risk assessment tools based on the Framingham risk functions to Aboriginal people is yet to be assessed. The lower risk estimate in women reported by Thompson and colleagues may simply reflect the higher cholesterol concentration cut-offs for women in the Sheffield table.

The true risk difference between sexes in Aboriginal people may not be as dramatic as Thompson and colleagues suggest.

Firstly, data in Box 1 of their article show that there was little difference between men and women as regards past history of cardiovascular disease.

Secondly, Aboriginal women experience a higher prevalence than men of some cardiovascular risk factors such as diabetes,1,3 abnormal HDL cholesterol level and overweight.3

Thirdly, our own research suggests that there may be a substantial difference between estimated and observed risks. Using data from a cross-sectional study of 681 Australian Aboriginal people in a remote community,3 we performed a similar analysis to that of Thompson et al. Based on the Framingham functions,4 we estimated that 10-year risks of coronary heart disease for women were much lower than those for men in all age groups (a finding similar to that of Thompson and colleagues). However, in a related study of the same Aboriginal community (as yet unpublished), when we analysed cohort data from 838 participants with 13 years of follow-up, the observed coronary disease rates for women were as high as those for men (Box).

The discrepancy we found between estimated and observed risks is a warning that researchers and clinicians need to be cautious when applying existing risk assessment tools to Aboriginal people.

In reply: We appreciate the commentary by Wang and Hoy on the problems of the use of risk scores for assessing cardiovascular risk in Aboriginal people.

In general, we agree that caution is essential in using tables that predict absolute risk of cardiovascular events. However, despite their limitations, absolute risk estimates are being encouraged by Australian, European, New Zealand and US authorities as a practical aid to targeting coronary disease preventive measures.1 An estimated risk of > 15% of a fatal cardiovascular event within 10 years, based on the Sheffield or Framingham scores, is now recommended as an indication for active treatment. Our prime purpose in providing an estimate of absolute risk in the Perth urban Aboriginal population was to demonstrate that a program of cardiovascular risk assessment with strong Aboriginal community support is capable of detecting high-risk people who will benefit from intensive risk-lowering strategies.

Wang and Hoy’s caution about applying absolute risk estimates based on the Framingham population to unrelated populations is of particular importance in the case of Australian Indigenous people, in whom diabetes and the related metabolic syndrome may be the predominant risk factors.

We have recently completed an analysis of the determinants of carotid atherosclerosis in the same population described in our earlier study.2 Our results confirm that, while the Framingham estimates (based on sex, age, LDL cholesterol and blood pressure) are indeed predictors of carotid atherosclerosis, their predictive value is significantly enhanced by the addition of markers of diabetes status and obesity.

The 13-year follow-up study of the Aboriginal cohort referred to by Wang and Hoy will provide unique data to help identify reliable risk predictors specific to Aboriginal people, and we look forward to its publication.