Chest pain: is the history useful?

Almost everyone has chest pain at some time, maybe after an extra glass of port, or after sitting awkwardly, or during a panic attack.1 Although non-cardiac chest pain has a benign prognosis,2 any cardiologist will remind you that coronary disease is the commonest cause of death in the West. So we tend to do an angiogram ‘just to make sure’. However, overinvestigation of chest pain may fail to reassure the patient,3 may actually entrench the idea of a serious illness4 and may interfere with the basic relationship between doctor and patient.5 It would therefore be an advantage to improve the clinical differentiation of cardiac and non-cardiac chest pain. Students usually remember that cardiac pain is central in distribution and crushing in quality. However, the most reliable features of chronic stable angina are its precipitation by exercise and relief with rest. We usually classify chest discomfort as either ‘typical’ or ‘atypical’ of a cardiac origin, depending how closely related to exercise it is. However, these terms are subjective and mean different things to different people. Thus the vast majority of patients - about 95% - undergoing angiography for suspected chronic stable angina have pain related at some time to exertion,6, 7 but the number in whom the relationship is tight is small, about 15-38%. The incidence of ‘typical’ pain could therefore be taken as 95% or 15-38% depending on definition. Can ‘typical’ pain be defined more objectively? A logistic regression analysis has shown that three out of 50 questions differentiated patients with normal anatomy from those with coronary disease (Table 1). The questions conformed to clinical experience and were related to precipitation on exercise, occurrence at rest and the duration of the pain. The answers were classified as either ‘typical’ or ‘atypical’ so that the patient could have between 0 and 3 ‘typical’ symptoms. The number of ‘typical’ symptoms was directly related to the likelihood of coronary disease. We have confirmed this result in a prospective study of unselected patients. Those aged under 55 years with no ‘typical’ symptoms had an 11% chance of coronary artery disease, while those aged 55 or more with three ‘typical’ responses had an 85% chance of coronary artery disease (unpublished data). In neither study, however, was there any important difference in the quality, site or radiation of the pain. Glyceryl trinitrate (GTN) is often used as a therapeutic trial, but most patients in both studies claimed relief with GTN irrespective of their coronary anatomy. Patients with normal anatomy tended to report that relief occurred after many minutes, which is out of keeping with the pharmacodynamics of the drug. A number of symptoms thought to suggest non-cardiac pain were also found not to be helpful. These included localisation of the pain under a finger or chest wall tenderness. On the other hand, associated symptoms such as pins and needles, palpitations and breathlessness were statistically more frequent in the patients with normal coronary anatomy, although these symptoms did not add predictive power to the model. This bears out some of the characteristics suggested by Potokar and Nutt in their review article on pages 110-114 of this issue.1 The rate of normal coronary anatomy at angiography can vary between 5% and 53% between different centres,8, 9 and from 13% to 24% between different operators at the same centre.10 These variations are likely to be related partly to the clinician's confidence in making a diagnosis of non-cardiac pain based solely on non-invasive information. Guidelines, including chest pain models similar to the one described above, may help to standardise rates of normal anatomy used in conjunction with the coronary risk profile and, in selected cases, non-invasive investigation.