Summary

Family doctors are faced with the difficult task of reassuring the vast majority of patients with chest pain while selecting the few with significant coronary artery disease for investigation. Investigating those without disease entrenches the idea of disease and increases the patient's anxiety. Recent studies demonstrate correlation between chest pain duration, reproducibility of chest pain and the absence of rest pain to coronary artery disease on angiography. Together with Framingham risk analysis this chest pain score can help in risk-stratifying patients with chest pain, preventing unnecessary investigations. Early diagnosis of other causes of chest pain and appropriate treatment is also helpful. Oesophageal and psychological causes of chest pain remain the most common causes of atypical chest pain. Adequate treatment of coronary risk factors is important to prevent development of coronary artery disease. Reassurance is the primary treatment for atypical chest pain and its effectiveness is related to a trusting doctor patient relationship.

摘要

治 療 胸 痛 是 家 庭 醫 生 面 對 的 一 項 難 題 ， 必 須 從 大 量 病 人 中 間 找 出 很 少 數 有 顯 著 冠 心 病 可 能 的 病 人 做 進 一 步 檢 查 ， 其 他 絕 大 部 分 則 需 要 耐 心 解 釋 安 慰 使 其 安 心 。 因 為 對 無 冠 心 病 者 進 行 檢 查 ， 會 加 重 其生 病 的 感 覺 、 增加 焦 慮 。 近 期 的 研 究 顯 示 胸 痛 的 持 續 時 間 ， 反 覆 發 作 和 靜 止 痛 與 冠 心 狀 動 脈 造 影 之 間 有 相 關 性 。如 果 連 同 Framingham 危 險 因 素分 析 一 起 使 用 ， 得 出 的 胸 痛 評 分 ， 就 可 以 按 胸 痛 的 危 險 性 分 組， 避 免 不 必 要的 檢 查 。 盡 早 診 斷 和 治 療 其 他 原 因 的 胸 痛 也 很 重 要 ， 心 理 性 和 食管 性 胸 痛 是 非 典 型 胸 痛 最 常 見 的 兩 種 情況 。 正 確 控 制 危 險 因 素 對 預 防 冠 心 病 很 重 要 。 基 層 醫 生 治 療 非 典 型 胸 痛 主要 不 是 通 過 於 知 識 和 檢查 ， 而 是相 互 信 賴 的 醫 療 關 係 ， 通 過 安 慰 解 釋 令 病 人 解 除 擔 心 。

---

Background

Chest pain is one of the most common clinical problems that present to the doctor. While the management of ischaemic heart disease is familiar territory for the majority of family doctors, the majority of patients who present to the family practitioner have non-cardiac chest pain. Despite the benign prognosis of non-cardiac chest pain, both the diagnosis and the management of non-cardiac chest pain remains a most difficult task. In this article we will review the progress made in the diagnosis and management of non-cardiac chest pain.

Although chest pain may be due to life threatening ischaemic heart disease, the majority of patients with chest pain have a benign aetiology.¹ Family doctors have the unenviable task of picking out the few with serious heart disease among the many who need nothing more than simple reassurance. With the increasing fear of litigation from missing coronary artery disease, this task has become more and more difficult. It is important not to succumb to legal threat and adopt the practice of defensive medicine as over exposure of the patient to unnecessary investigations is costly and ethically doubtful.² As we have argued elsewhere,³ normal investigations often fail to reassure⁴ the patient and often entrenches⁵ the idea of disease. It is harmful to the doctor patient relationship and has a small but real morbidity and mortality. Therefore, unnecessary investigations motivated by defensive medical practice are harmful to patients. In this review, we offer an alternative to defensive medicine in facing the difficult task of selecting the few with serious coronary artery disease among the many with non-cardiac chest pain.

The Framingham database⁶ has offered substantial help for the frontline medical practitioner who is struggling with the task of risk stratifying patients who present with chest pain. Young females are extremely unlikely to have coronary artery disease, but their risk of coronary artery disease increases rapidly after the menopause. Cholesterol, hypertension, smoking and diabetes all have a significant effect upon the 10 year risk of coronary artery disease. A shorthand application of the Framingham data would be to investigate all those who are older than 55 years of age and consider investigation in those males over 40 with risk factors. However, this is a substantial group of patients and so we depend on history taking to determine who has genuine "typical" chest pain. Unfortunately, there has been little research correlating "typical" chest pain with findings of coronary artery disease on angiography. Instead, "typical" and "atypical" chest pain have been defined subjectively. Traditional ideas of "typical" chest pain have been shown to be poor differentiators between those with and without coronary artery disease on coronary angiography. Radiation to the throat, jaw, neck; location of the pain, and the quality of the chest pain all failed to differentiate coronary artery disease from non-cardiac chest pain.^7,8 Cooke et al⁷ found that only three out of 50 chest pain characteristics could differentiate between 65 patients with normal coronary anatomy and 65 controls with coronary artery disease. These were reproducibility, duration, and absence of rest pain. For reproducibility, typical is defined as chest pain being reproducible in 10 out of 10 episodes of strenous exertion. For rest pain, typical is defined as pain occurring at rest in less than 10% of episodes. For duration, typical is defined as less than or equal to 5 minutes. The reproducibility score (0 or 1), the rest pain score (0 or 1) and the duration score (0 or 1) can be added together to give a score between 0-3, where 0 is very atypical chest pain, and three typical angina pain (Table 1). This chest pain score has subsequently been validated in 250 patients presenting for coronary angiography⁸ as well as in 363 patients attending a tertiary referral center with chest pain.⁹ Used in conjunction with the Framingham risk calculations, the chest pain score allows further differentiation of patients into low and high risk. In our patient population, chest pain score allowed us to identify half of the patient's referred for exercise testing as low risk with a subsequent event rate of less than 8% in two years. Therefore, even among those already considered at high risk and referred from their general practitioners, the chest pain score was able to identify a lower risk subgroup who might not need further investigation. We suggest using the rest score to differentiate patients initially. Young patients with rest pain occurring at more than 10% of the time need no further investigations. Older patients or patients at high Framingham risks require investigation unless the chest pain score is 0 indicating very atypical pain. This system will greatly reduce the number of normal investigations we perform and be beneficial to the patients.

Table 1: The 'typical' chest pain questionnaire
Question		Typical	Atypical
1.	"If you go up a hill on ten separate occasions, on how many of these do you experience chest pain?"	10	<10
2.	"If you have the pain ten times in a row, how many happen when you are resting or sitting quietly?"	0 or 1	>1
3.	"How long does the pain usually last?"	5 mins	>5 mins

Treatment of disease often involves the balance of risks. Anticoagulation for atrial fibrillation is a good example. However, when the physician is exposed to unilateral risks, there is a danger that he or she might expose the patient to opposing risks excessively. For example, a transplant surgeon might see only acute rejection cases and have a limited exposure to the lymphomas that develop after 10 years of overly aggressive anti rejection therapy. Consequently, he or she might be tempted to be overly aggressive with immunosuppression as his or her practice is biased by one sided observation. A similar situation occurs with chest pain. Patients with missed coronary artery disease inevitably get seen by a cardiologist when they are having their myocardial infarction or coronary angiography. However, patients who are investigated and found to have non-cardiac chest pain rarely return to see the cardiologist thus biasing the cardiologist's experience. Fortunately, we have excellent long term studies that inform us of the risks of over investigating non-cardiac chest pain. Although the mortality for patients with non-cardiac chest pain is similar to an age sex matched control population, they suffer from substantial morbidity. Half of the patients continue to have worsening chest pain. This results in job loss or job change in 50% of patients.^10-14 This group of patients suffers greatly and is burdened by the continuing belief of having heart disease. This is often entrenched by investigation. If we minimise the risk of missing coronary artery disease, we increase the risk of entrenching the idea of coronary artery disease in those with non-cardiac chest pain. We also put these patients at risk of mortality and morbidity associated with investigations, which is about one per 100 patients for angiogram. Therefore we suggest a risk balancing strategy, accepting that there will occasionally be patients with coronary artery disease that we miss, whilst being aware of the dangers of entrenching the idea of coronary artery disease. In our experience in the exercise test laboratory, we often see patients who have had only one or two weeks of chest pain several months ago. Our studies above^7-9 all included the entry criteria of a history of chest pain of longer than 1 month. In our busy and fast paced clinics, it is often easier to book an investigation or refer a patient than to see the patient again for reassessment. In cases where the chest pain score is intermediate and the history is short, we recommend applying the physician's friend, time. Chronic stable angina is very likely to persist and the risks of re-evaluating a month later is minimal.

Therefore, in evaluating chest pain, we must be wary of blindly practicing defensive medicine. Instead, we should be aware of the harm of unnecessary investigations. We recommend the careful use of the Framingham risk analysis as well as the chest pain score for risk stratification of patients with chest pain. For difficult cases, reviewing the patient in a month's time is more valuable than ordering investigations "just in case" of coronary artery disease.

Since non-cardiac chest pain is so much more common than ischaemic heart disease, most family practitioners will see many patients with non-cardiac chest pain. Some will have had coronary angiography, others negative exercise test or other stress tests. A few will have very atypical pain and belong to low risks group and not had further investigations. Successful management of these patients is essential to prevent them from progressing onto the next stage of investigation and the further entrenching of the idea of disease. Therefore we need a clear and thorough understanding of the causes and treatments of non-cardiac chest pain.

Although there are many causes of atypical chest pain, we will cover only the common and diagnostically challenging ones here. Serious atypical chest pain caused by pulmonary embolism, aortic dissection, and chest trauma are usually readily identified by medical practitioners and will not be further covered here. Musculoskeletal chest pain is also easily diagnosed by the demonstration of tenderness and will also not be further discussed. The remaining causes of atypical chest pain are discussed below.

Oesophageal pain

About a third of the patients with chest pain despite normal coronary anatomy have an oesophageal cause for their pain.¹⁵ This high prevalence has prompted some authors to recommend a trial of proton pump inhibitor in all patients with chest pain despite normal coronary anatomy.¹⁶ However, as oesophageal disease rates vary between different countries and reflux is closely related to body mass, we suggest a trial of proton pump inhibitors in those with clinical suspicion of oesophageal disease. Those patients with high body mass index, acid burning in the mouth, pain on lying flat, hoarse voice in the mornings, food regurgitation should initially undergo a trial of proton pump inhibitor therapy. Although the clinical features of oesophageal pain is seldom seen, oesophageal abnormalities that correlate well to symptoms occur frequently in many populations worldwide and is likely to be more prominent than our clinical impression would suggest in Hong Kong.

Psychological

Psychological abnormalities exists in 40-60% of those with normal coronary anatomy as oppose to 20% of controls.^17-19 Panic disorder, anxiety neurosis and depression are the commonest abnormalities reported.^17-19 Cognitive behavioural therapy have been shown to be beneficial to patients with chest pain despite normal coronary anatomy.²⁰ Once again, we recommend using clinical sense to differentiate those with anxiety, or panic attack symptoms and recommending them for cognitive behavioural therapy. In our own practice, as well as reports from others, we have found that patients are reluctant to enter into psychological therapy. Consequently, we tend to give this group of patients a course of imipramine, as several investigators have shown a reduction in chest pain episodes with imipramine treatment.^21,22 We recognise that this is not universally accepted practice, but urge the individual practitioner to try out this suggested therapy in patients who are at low risk of coronary artery disease and have a psychological component to their chest pain.

Oestrogen deficiency

There are two small double blinded randomised controlled cross over studies of hormone replacement therapy in post menopausal women with chest pain despite normal coronary anatomy.^23,24 Both demonstrate a benefit in number of chest pain episodes, but only one showed improved exercise time. However, since hormone replacement therapy is relatively safe and recommendable in many postmenopausal patients, we routinely start postmenopausal women on hormonal therapy once we have diagnosed non-cardiac chest pain. Although the pros and cons of hormonal replacement therapy have been long debated, it seems that its risks and benefits are balanced and in the presence of atypical chest pain that is bothersome, it may well be worth a try.

Risk factors

Many patients with non-cardiac chest pain have risk factors for coronary artery disease. The presence of these risk factors may bias the patient to seek medical attention. Despite presenting with atypical chest pain, these patients should have their risk factors treated to reduce their chance of developing coronary artery disease. Hypertension with left ventricular hypertrophy is thought to produce chest pain. Treatment may improve their chest pain as well as reduce their risk of developing ischaemic heart disease.

Reassurance

Reassurance is the largest therapeutic component for most patients presenting with chest pain. Yet there is little research into reassurance. However, we know that investigations have little role in reassurance. Macdonald et al²⁵ demonstrated reduction in anxiety from echocardiography in only 15% of the 300 patients referred for echocardiography. Six percent had increase anxiety after normal echocardiography. The benefit of normal investigation in reassurance is minimal. In another study²⁶ he demonstrated worsening anxiety in all 10 patients with symptoms after normal echocardiography. Lucock et al²⁷ found that physicians' rating of success of reassurance corresponded well to patients' ratings in patients who have undergone normal endoscopy. Those with anxiety beforehand had little benefit of reassurance from endoscopy, with their anxiety level returning to baseline within 24 hours and sustained at 1 year. A normal endoscopy was reassuring only in those with low initial anxiety. These three studies demonstrate that investigations have no benefit for those who are anxious of having disease. Yet in our clinical practice we tend to investigate only those with substantial anxiety. Those who have symptoms, like chest pain, will tend to do worse with investigations. Therefore, we should reassure the patient strongly with detailed history taking and clinical examination, repeated statement of the lack of serious illness and suggestion that the symptoms will resolve.²⁸ It seems that extensive knowledge of the disease makes little difference in the impact of the doctor's reassurance,²⁹ a trusting doctor patient relationship is more important.³⁰ Reassurance often fails in those with underlying psychological illness and these patients may benefit from psychological counselling.

Conclusions

Atypical chest pain is a diagnostic and therapeutic challenge in our current age of defensive medicine. The physician's fear of missing disease must be balanced by the realisation of the dangers of entrenching the idea of disease with unnecessary investigations. Patients often have considerable morbidity and experience very poor health despite a normal coronary anatomy. Recent studies have documented three chest pain characteristics that help predict coronary artery disease on angiography. Using these factors together with Framingham risk analysis can effectively risk stratify patients and avoid unnecessary investigations. A keen awareness of alternative causes of chest pain and psychological abnormalities can also help to reduce the number of patients with atypical chest pain sent for investigations. In intermediate risk patients, a timely review may be more cost effective and less harmful to the patient than early investigation. Success in the difficult task of reassuring the patient is not dependent on investigations or knowledge but on the trusting doctor patient relationship.

Key messages

1. Normal coronary angiography entails a small but real mortality risk as well as considerable morbidity after angiography.
2. Cardiovascular risk analysis and chest pain score can help with risk stratification of patients with chest pain.
3. Determination of other causes of chest pain and appropriate treatment should be given to patients with non cardiac chest pain.
4. Reassurance is the most important component of the treatment of non cardiac chest pain.

---

---

References