Dear Editor,

We thank Underwood for his kind comment about our study[1] which demonstrated an absence of gender bias in the investigation and management of patients referred to our open access chest pain clinic. We can reassure him that this study relied on routinely collected data and clinical staff were not aware that they would be under scrutiny with regard to gender bias. Also, primary physicians had guidelines as to which patients could be referred and an exclusion was limited mobility. Furthermore, the supervising cardiology specialist registrars made management decisions based on individual clinical judgement although, as stated, these decisions were checked by consultants. As can be deduced from our results, only patients who actually had an exercise test were included in the study and therefore, patients who were not exercised because of abnormal resting electrocardiograms were excluded.

His comments about the role of treadmill exercise testing, myocardial perfusion imaging (MPI) and coronary arteriography are also timely as the establishment of open access clinics in the United Kingdom has become a priority since the publication of the National Service Framework for coronary heart disease.[2] In summary, Underwood contends that Southampton had a low use of MPI, that a high proportion of women referred for coronary arteriography had normal coronary arteries and that if MPI were used more frequently, "inappropriate" coronary arteriography could be avoided and large cost savings could be made.

An ideal diagnostic test would be one which provides the most information for the least cost and at low risk to the patient. There is no doubt that MPI provides valuable prognostic information which may be superior to that obtained by coronary arteriography. However, it has a high radiation burden and requires expensive equipment and in the United Kingdom, is not as widely available as coronary arteriography.[3] The serious complication rate for coronary arteriography has been estimated at 1% with a mortality rate of 1 in 2000.[3] However, it uniquely provides the anatomical data required for revascularisation by coronary artery bypass surgery or percutaneous transluminal coronary angioplasty.

In our study, we showed that for men, treadmill exercise testing had a high positive predictive value of 95.2%. At the end of our study period, 286 men had had coronary arteriography of which 170 (59.4%) were referred for revascularisation for either symptomatic or prognostic reasons. It could be argued therefore, that if MPI had been carried out instead of arteriography, then approximately 40% of men may have been spared arteriography. However, 60% of men would have had an additional investigation and because of this, they would have waited longer for revascularisation. In our centre, the waiting time for MPI is x months.

In contrast to men, the positive predictive value for treadmill exercise testing in women was only 72% and as it is known that false positive rates are higher in women, the British Cardiac Society guidelines do suggest that MPI should be carried out as a first line investigation for chest pain in women.[3] However, from personal experience, we know that this is not common practice in the UK and that this is partly because of limited availability. Also, the open access chest pain clinics being established as part of the National Service Framework will all use treadmill testing for initial diagnosis and none are restricted to men. In our study, all women with 2 mm or more of ST depression on their electrocardiogram during their exercise test who were referred for further investigation were referred for arteriography rather than MPI. Of these, 56% were referred for revascularisation. Therefore, the argument as to whether these women should have had MPI first is the same as that for all men who were referred for arteriography. Similarly, for women who had ST depression of less than 2mm and who were referred for arteriography, the intervention rate was 42.3%. For those women who had no electrocardiogram changes and who were referred for arteriography, the intervention rate was only 13%. For this group, it would seem that MPI would perhaps have been more appropriate. Of all women referred for arteriography in our series, we found that 56.2% had normal coronary arteries. This is similar to the 50% of women undergoing diagnostic arteriography in the CASS study[4] but greater than the 30.7% found in the RITA study.[5] However, this latter study included patients with myocardial infarction and unstable angina. These high rates of normal arteriographic findings are likely to reflect the low positive predictive value of exercise testing in women rather than indiscriminate referral for arteriography. For instance, of the women who had MPI in our study, only 3.9% were reported to have findings compatible with coronary artery disease (data not originally reported). This suggests that women suspected of having a low probability for coronary artery disease were referred for MPI while women thought to have a greater probability were referred for arteriography. Underwood reports in the EMPIRE study that centres which used MPI had a normal angiogram rate of 26% while those that did not had a rate of 43%.[6] However, these figures are not comparable to ours since they are not sex specific. If we were to combine our results for men and women, then 29.3% of all patients undergoing arteriography had normal results!

Underwood estimates that based on the EMPIRE results, Southampton could save £65,000 a year if MPI were to be used more frequently. We do not agree with this claim. The main deficiencies of the EMPIRE study are that it was a retrospective notes trawl; it was relatively small with 8 centres in 4 countries recruiting an average of only 49 patients each and hypothetical rather than actual costs were used in financial calculations. We are told that there were "discrepant results" when each centre supplied information on costs and charges. They estimated that exercise tests, MPI and coronary arteriography cost £70, £220 and £1,100 respectively. However, the assumed cost of angiography is excessive, now being as low as £700 in most high volume centres. If we applied these figures to our female population, then the cost for the investigations which actually occurred is £149,190 (601 exercise tests, 137 arteriograms, 51 MPI studies). If we then consider a situation where every woman had an exercise test, where women who were actually referred for arteriography had MPI instead and where only those women who were referred for revascularisation had arteriography, then the cost would be £114,930 (601 exercise tests, 45 arteriograms, 188 MPI studies). This figure is likely to be higher in real life because it does not take into account false positive MPI results e.g. those due to breast shadows, which could lead to further investigation by arteriography. Also, it does not take into account situations where arteriography rather than MPI is indicated for clinical reasons. It can be seen that cost savings can be made but not on the scale claimed.

In the EMPIRE study, Underwood concedes that there are no randomised trials assessing the cost-effectiveness of strategies of investigation in patients with symptoms suggestive of coronary artery disease. This situation is still true. Even without such studies, common sense tells us that there are situations where the use of MPI is valuable. For instance, in patients thought to have a low probability of coronary artery disease, a normal MPI result would indicate an extremely good prognosis. On the other hand, in patients thought to be at high risk from coronary events, then coronary arteriography would be more appropriate regardless of sex, poor mobility or abnormal resting electrocardiogram. The question as to whether all women who have a history suggestive of angina should have MPI as a first investigation rather than an exercise test has never been asked in any study and we stand by our assertion that further research is needed.

References
(1) Wong Y, Rodwell A, Dawkins S, Livesey SA, Simpson IA. Sex differences in investigation results and treatment in subjects referred for investigation of chest pain. Heart 2001;85:149-52.

(2) Anon. National Service Framework for Coronary Heart Disease. Modern standards and service models. Department of Health, March, 2000.

(3) de Bono D. Investigation and management of stable angina: revised guidelines 1998. Heart 1999;81:546-55.

(4) Kennedy JW, Killip T, Fisher LD, Alderman EL, Gillespie MJ, Mock MB. The clinical spectrum of coronary artery disease and its surgical and medical management. Circulation 1982;66(Suppl 3):16-23.

(5) Henderson RA, Raskino CL, Hampton JR. Variations in the use of coronary arteriography in the UK: the RITA trial coronary arteriogram register. Q J Med 1995;88:167-73.

(6) Underwood SR, Goodman B, Salyani S, Ogle JR, Ell PJ. Economics of Myocardial Perfusion Imaging in Europe - the EMPIRE study. Eur Heart J 1999;20:157-65.