While December's editorial on non-cardiac chest pain is thoughtful and thorough,(1) there is a strange lack of emphasis on skeletal chest pain. I am not sure if this is due to selection of patients, but I wonder if it is the lack of a diagnostic test for skeletal pain. Since this may involve up to 73% of patients referred with chest pain to cardiac clinics,(2) it would seem to be of paramount importance. The authors comment on the possibility that "abnormalities" (namely oesophageal and respiratory) may be "coincidental rather than causative" and that 60% of patients with normal coronary angiograms and non-cardiac chest pain have a psychiatric disorder. We all occasionally recommend angiography in patients who cannot adapt to the pidgeonhole we offer to give more strength to our arm, which suggests that a higher percentage of patients with non-cardiac chest pain and psychiatric disorders will get angiography than those who are mentally more balanced; in addition Mayou and his colleagues have themselves demonstrated that patients who are told that their hearts are normal but are not given a clear diagnosis remain in trouble,(3) so it is perhaps not surprising that psychiatrists find a high level of psychiatric disorder in these people.

In a letter to Heart,(4) I reported the reproduction of chest pain by testing the passive range of movements and found that of 27 patients with clinically non-cardiac chest pain one had oesophageal reflux (and responded to treatment for this), one had costochondral joint pain and 22 had their symptoms reproduced by spinal movements. Incidentally at one year some still had symptoms but none had been off work because of their presenting symptoms or admitted to hospital with chest pain. This pain can sometimes be both reproduced and relieved by intercostal nerve block (C Davidson, personal communication). We are doing a larger study at present but it does seem that manipulating the spine as described by Chambers et al is a sound way of demonstrating to patient and doctor that the the pain is mechanical. Only 7 out of 27 responded to non-steroidal anti-inflammatory drugs; I think it is likely that postural training or (pace the unbelievers) manipulation will do better. The hardest patients to manage seem to be those with cardiac and skeletal pain.

The authors suggest a psychological cause if there is a situational or phobic component to the somatic symptoms; but anxiety appears to increase mechanical symptoms such as sciatica due to a lumbar disc, perhaps by increased muscle tone; it is not the cause of the symptoms.

Exercise testing is a two edged sword (did they ever make single edged swords?). The authors mention the value of a negative test, but of equal importance is the danger of a positive in someone with coronary disease but non-cardiac chest pain, and the risk of intervention with no relief of symptoms, further angiography etc.

Finally the authors suggest that for patients with "continuing symptoms---with coexisting psychological problems" a specialist nurse who has received additional training should be employed in the cardiac clinic to help them. Surely if the patient has been told that they have no cardiac disease the last place they should be followed is the cardiac clinic whatever the actual cause of their symptoms.

I know the north-south divide involves more than house prices, but I find it hard to believe that chest pain in Oxford or London is too different from that in Lancashire, assuming that referral is not biased. To demonstrate a physical cause is of tremendous importance as we are looking at thousands of referrals per year who, if managed by their general practitioner, would contribute to a significant drop in our waiting lists. I suggest that some of those thousands have demonstrable skeletal pain, and that we are missing them.

References

1 Chambers J, Bass C, Mayou R. Non-cardiac chest pain: assessment and management. Heart 1999;82:656-7.
2 Jain D, Fluck D, Sayer JW, et al. Ability of a one-stop chest pain clinic to identify high risk patients... J R Coll Phys Lond 1997;31:401-4.
3 Mayou R, Bryant B, Sanders D, et al. A controlled trial of cognitive behavioural therapy for non-cardiac chest pain. Psychol Med 1997;27:1021-31.
4 Best RA. Non-cardiac chest pain: a useful physical sign? [letter] Heart 1999;81:450.

In my review on risk stratification in acute coronary syndromes, "diagnostic value" was used conventionally to refer to the ability of predischarge tests to predict future coronary events, particularly death and myocardial infarction.

In response to the 3 additional points:

1. Cost-effectiveness Underwood et al are correct to caution me on statements of cost-effectiveness. My contention was (and remains) that predischarge stress testing is usually as effective as perfusion imaging for risk stratification and costs less. The superior quality of SPECT compared with conventional perfusion imaging may confer some advantage (see below) but the considerable capital costs involved will inevitably limit its application.

2. Exercise ECG versus Perfusion Imaging Underwood et al are incorrect to infer superiority of predischarge perfusion imaging from the data presented in the metaanalysis of Shaw et al (1). Cardiac event rates were higher in the studies of myocardial perfusion imaging, readily accounting for the apparent difference between positive predictive values for these diagnostic tests. It is for the same reason, incidentally, that both tests appear to perform better in patients who have not received thrombolytic therapy. Shaw emphasises in her metaanalysis that both the exercise ECG and the radionuclide perfusion scan are blunt tools for risk stratification and my algorithm recommends use of either test, without expressing a preference. Underwood et al draw attention to 3 papers comparing the exercise ECG with SPECT imaging that post-dated ShawÆs metaanalysis (2-4). Two showed some advantage for SPECT imaging although positive predictive values for events after hospital discharge appeared low (specific data not provided), confirming previous reports (2-3). The other was a small (n=71) retrospective series, in which the only multivariate predictor of death and recurrent infarction was LV ejection fraction (4). Only when the endpoint was extended to incorporate unstable angina and heart failure did perfusion imaging provide independent prognostic information. Set against the 36 studies (16,960 patients) included in ShawÆs metaanalysis, the additional 3 selected by Underwood et al are not overly persuasive but suggest that SPECT technology may have an edge over the exercise ECG for risk stratification.

3. Incremental value of perfusion imaging Although one of the studies quoted by Underwood et al found that SPECT perfusion imaging was not independently predictive of death and recurrent infarction if LV ejection fraction was included in the multivariate model (4), they quote another in which incremental value was demonstrated (5). Data for stress testing are similarly contradictory. For this reason I recommended application of these tests (exercise ECG or perfusion scan) only in patients with advanced LV dysfunction (LVEF <_40 in="in" whom="whom" risk="risk" is="is" greatest.="greatest." unpublished="unpublished" data="data" for="for" our="our" own="own" low="low" patients="patients" discharged="discharged" with="with" lvef="lvef"/>40% by GUSTO criteria (6)) show the estimated risk of death in the first year is only 3.8% (2.1-5.4%) and of death and recurrent infarction 9.8% (7.1-12.4%). The idea that these relatively low risk survivors of acute myocardial infarction would benefit from further risk stratification using perfusion imaging (or stress tesing) seems barely credible. This does not of course mean that the tests should not be done but it does mean that incautious predictions of "cost-effectiveness" may be incorrect.

In summary, SPECT imaging may offer a small advantage over the exercise ECG for detection of residual ischaemia in patients with acute coronary syndromes, although it remains a blunt tool for risk stratification. Speaking from the perspective of the coronary care unit, there is no doubt that the availability and low cost of exercise testing (as opposed to the relative nonavailabiltiy and high cost of SPECT imaging) will make it a more practical solution for predischarge risk stratification in most centres. However, perfusion imaging is a reasonable, perhaps better, alternative in those units able to provide a predischarge service for upward of 700 coronary patients per year, and my algorithm allows for this.

Adam D Timmis MD, FRCP

References

1. Shaw LJ, Peterson ED, Kesler K, Hasselblad V, Califf RM. A metaanalysis of predischarge risk stratification after acute myocardial infarction with stress electrocardiographic, myocardial perfusion, and ventricular function imaging. Am J Cardiol 1996;78:1327-37.

2. Travin MI, Dessouki A, Cameron T, Heller GV. Use of exercise technetium-99m sestamibi SPECT imaging to detect residual ischemia and for risk stratification after acute myocardial infarction Am J Cardiol 1995;75:665-9.

3. Brown KA, Heller GV, Landin RS, Shaw LJ, Beller GA, Pasquale MJ, Haber SB. Early dipyridamole (99m)Tc-sestamibi single photon emission computed tomographic imaging 2 to 4 days after acute myocardial infarction predicts in-hospital and postdischarge cardiac events: comparison with submaximal exercise imaging. Circulation 1999;100:2060-6.

4. Dakik HA, Mahmarian JJ, Kimball KT, Koutelou MG, Medrano R, Verani MS. Prognostic value of exercise 201Tl tomography in patients treated with thrombolytic therapy during acute myocardial infarction. Circulation. 1996;94:2735-42.

5. Mahmarian JJ, Mahmarian AC, Marks GF, Pratt CM, Verani MS. Role of adenosine thallium-201 tomography for defining long-term risk in patients after acute myocardial infarction. J Am Coll Cardiol 1995;25:1333-40.