The burden of illness of non-cardiac chest pain

Authors


Correspondence to: Professor N. J. Talley, Department of Medicine, The University of Sydney, Nepean Hospital, Level 5, South Block, PO Box 63, Penrith, NSW 2751, Australia. E-mail: ntalley@med.usyd.edu.au

SUMMARY

Background : Non-cardiac chest pain is a common condition affecting approximately one-quarter of the population during their lifetime, but the long-term economic costs of non-cardiac chest pain are poorly defined.

Methods : A MEDLINE and Current Contents search was performed from 1991 to 2002 using specific keywords. All major articles on the subject of non-cardiac chest pain in this period were reviewed and their reference lists searched.

Results : Limited studies suggest that the majority of those with non-cardiac chest pain do not consult a doctor regarding their symptoms; the drivers of health care seeking are not known. The impact on the quality of life in consulters can be severe, with as many as 36% reporting much lower quality of life levels. The diagnosis of non-cardiac chest pain can be difficult due to the heterogeneous nature of the condition, with significant overlap of gastro-oesophageal reflux disease, chest wall syndromes and psychiatric disease, which may drive up the costs of management. The prognosis appears to be good, but there are conflicting results in long-term studies.

Conclusions : The costs of non-cardiac chest pain to the health care system are likely to be large and represent a significant proportion of each Western country's health care budget. Further studies are required to determine methods of reducing health care costs.

Introduction

Early presentation to hospital after the onset of acute chest pain has been promoted by organizations such as the American Heart Association, British Heart Foundation and the Heart Foundation of Australia, so that patients with acute myocardial infarction can benefit from early thrombolytic therapy.1, 2 Despite this, over the last decade, most hospitals have observed a decline in the number of presentations with acute myocardial infarction, while there has been a proportionately greater increase in admissions with unstable angina or non-cardiac chest pain (defined as chest pain not due to underlying detectable ischaemic heart disease).3

These observations raise a number of clinically important questions. How many people suffer from recurrent non-cardiac chest pain in the community? How many of these individuals seek medical attention repeatedly for their non-cardiac chest pain? What are the costs to the health care system and society? Can changes to health care seeking behaviour for recurrent unexplained chest pain limit the rising costs to the health care system? We sought to address these issues by a review of the available literature.

Methods

This article is based on a search strategy using the electronic databases MEDLINE (Index Medicus) and Current Contents (1991–2002). The keywords used for the search included combinations of non-cardiac chest pain, noncardiac chest pain, chest pain unexplained undetermined, economic, costs, decision analysis, consulting, health care seeking and absenteeism. The search revealed 659 abstracts from both databases. Additional manual searches were made using the reference lists from the selected articles to retrieve other papers relevant to the topic. No language restriction was placed on any of the literature searches. All articles that included information on cost (n = 5) were included in this review.

How big is the problem?

The causes of non-cardiac chest pain are diverse and overlap, with the condition accounting for approximately 2–5% of all emergency presentations.2 Despite this significant burden, the prevalence and natural history of non-cardiac chest pain remain poorly understood because of a lack of population-based studies.

The prevalence of non-cardiac chest pain has been ascertained primarily from hospital-based studies or case studies, which have concentrated on chest pain in relation to ischaemic heart disease.4, 5 There are very few studies which have estimated the population prevalence of non-cardiac chest pain. Locke et al., in a community survey in Olmsted County, MN, USA (n = 1511), reported that 23% had non-cardiac chest pain (which was defined as ‘those who reported chest pain but did not have a history of cardiac disease’).6 This was based on a semi-rural US population sample. In a study of a non-Hispanic white (n = 1848) and Mexican American (n = 3272) population in Texas, 3% had chest pain thought to be angina (based on the Rose Questionnaire), and 30% had pain thought not to be angina.7 In another population-based study of 7735 randomly selected men, aged 40–59 years, the prevalence of ‘other chest pain’ was reported to be 24%; the selection criteria were based on the Rose Angina Questionnaire, examination by a nurse and the exclusion of previous cardiovascular diagnoses.8 Brattberg et al. reported that the prevalence of ‘chest pain’ in a population-based longitudinal study of 321 Swedish individuals, aged 53–87 years, was 28%, but cardiac and non-cardiac cases were not distinguished.9

In Australia, we have obtained initial population-based data on the epidemiology of non-cardiac chest pain.10 Our results suggest that non-cardiac chest pain is a prevalent condition in the community; amongst our population-based sample of 672 randomly selected individuals in Sydney, Australia, 39% reported having chest pain at some time in their lives; 7% reported a history of myocardial infarction and 8% of angina. A total of 219 cases (33%) were classified as non-cardiac chest pain.10

Non-cardiac chest pain is a heterogeneous condition (Table 1). Moreover, the possible causes often overlap. However, gastro-oesophageal reflux disease is considered to be the most important explanation, followed by psychiatric disease and other causes.4, 10–18

Table 1.  Hospital- or population-based studies that have classified non-cardiac chest pain into clinical groups
StudyDesignIschaemic
heart disease
GERDPsychiatric
disorders
Musculoskeletal
disorders
Microvascular
disease
Others
  • GERD, gastro-oesophageal reflux disease.

  • *

     Not considered.

  •  Already excluded.

Cannon et al.17 (n = 87)Hospital-based***73%27%
Janssens et al.15 (n = 60)Hospital-based88%***12%
Rouan et al.18 (n = 772)Out-patients clinic 8% 9% 1%23%*59%
Katon et al.14 (n = 74)Hospital-based*79%**21%
Wise et al.16 (n = 100)Hospital-based**16%*84%
Fruergaard et al.4 (n = 204)Hospital-based31%42%*28%*12%
Eslick et al.10 (n = 672)Population-based 7%54%24%11%*12%

What drives people to seek help?

How an individual perceives, evaluates and acts upon an illness is defined as illness behaviour; abnormal illness behaviour characterizes individuals who are frequent seekers of health care and who are over-concerned with minor illnesses. Very little is known about health care utilization amongst people experiencing non-cardiac chest pain. However, if one-quarter of the population experiences chest pain, it is very unlikely that most of these subjects seek health care.6 We do not know how many individuals with non-cardiac chest pain in the community actually present for medical care, and what drives people to seek help. The perception and verbal report of non-cardiac chest pain is a complex process, probably mediated by a number of factors, including symptom severity, anxiety, depression, individual cultural values and secondary gain.19

Studies have suggested that only a minority of patients with heartburn or dyspepsia seek medical advice in the general population, but whether these data can be extrapolated to non-cardiac chest pain is unknown.20 However, as gastro-oesophageal reflux disease explains some non-cardiac chest pain in the community, it seems reasonable to learn from this disease.19 Among patients with heartburn, psychological factors are strongly associated with health care seeking behaviour.20 One study observed that, in those with heartburn, presenters and non-presenters for health care differed significantly from each other with regard to symptom severity and psychological distress (e.g. phobia, obsessionality and somatization).20 The study concluded that the interaction of heartburn severity and psychosocial factors influences health care seeking behaviour.20 This is similar to presenters and non-presenters amongst patients with irritable bowel syndrome.21

How does recurrent chest pain affect apatient's life?

Studies in patients with non-cardiac chest pain followed for 5 years suggest a persistence of symptoms in the majority of patients, and a continued pattern of impaired functional status and re-presentations.22, 23 Tew et al., in a small hospital-based prospective study, observed that patients with non-specific chest pain reported more frequent visits for medical care than did ischaemic heart disease patients.22 The impact of non-cardiac chest pain on those who have consulted includes greater impairment of daily activities, emotional distress and higher levels of anxiety than in those with ischaemic heart disease.24 The impact of non-cardiac chest pain on the quality of life in non-consulters is unknown.

Studies have produced conflicting findings regarding the long-term outcome of patients with acute chest pain.25, 26 Ockene et al. followed 57 patients with normal coronary arteries, who had been referred for coronary angiography because of chest pain.25 Sixteen months later, 58% still had symptoms as severe as those prior to cardiac catheterization and 29% were still unable to work. Despite reassurances to the contrary, 44% still believed themselves to have cardiac disease. Other studies have confirmed this picture;27 the prognosis is good in terms of survival, but there is considerable continued morbidity and consumption of medical resources.

However, Wilhelmsen et al., in a 16-year cohort study, found a high cardiovascular and non-cardiovascular mortality among male patients with ‘non-specific’ chest pain who had negative investigations for cardiac ischaemia.26 They stated that it was important to be suspicious of early coronary artery disease, particularly in men with apparent non-cardiac chest pain. This study population was biased due to inadequate stratification by gender and age, with all the study participants being male and aged 51–59 years at baseline. Although generally the prognosis for patients with non-cardiac chest pain is considered to be excellent, this has not been adequately assessed in the general population and deserves further attention.

What does non-cardiac chest pain cost?

Direct costs

It is now recognized that non-cardiac chest pain accounts for a significant proportion (2–5%) of chest pain cases presenting to hospital accident and emergency departments. Therefore, the levels of health care utilization and the costs associated with non-cardiac chest pain are high. Less than 20% of chest pain admissions are actually found to have coronary artery disease.28 In the USA, it has been estimated that it costs approximately $8 billion dollars for the initial care of patients suspected to have an acute coronary syndrome, but who are subsequently found not to have coronary artery disease.28 This high cost is driven by the high incidence of missed myocardial infarction, the associated litigation and the subsequent mortality of those sent home.29 In Australia, non-cardiac chest pain has been shown to be a major contributor to the total public health cost of chest pain.2

The long-term economic cost of undiagnosed chest pain is unknown. Faxon et al. demonstrated that the finding of a normal coronary angiogram decreased with the number of days spent in hospital in the 2 years after the angiogram, compared to the previous 2 years.30 However, in functional terms, the patients were little changed. Although hospital room costsweresignificantly reduced ($2967 to $572), other costs were not considered in this study. It was calculated that, in 1989, a patient with chest pain but normal coronary arteries cost approximately US $3500 in hospital costs alone in the year following exclusion of coronary artery disease.31

The actual costs of care are variable, because the investigations considered often reflect the speciality of the investigators concerned. Usually, more than one cause for chest pain is not considered. For example, Katon et al. investigated a population undergoing cardiac catheterization from a psychiatric view point.14 They found that a very high proportion (43%) of the patients with normal coronary arteries had panic disorder. However, Janssenset al., primarily using gastroenterological investigations, found that over 90% of their patients had either ‘likely’, ‘probable’ or ‘suspected’ oesophageal causes for their chest pain.15 Evidently, some overlap must occur, yet neither set of investigators considered this. Different again was the study of Wise et al., who investigated patients for the presence of musculoskeletal pathology.16 They found this to be present and the likely cause of symptoms in 16% of their patients. If investigations are performed to uncover the presence of reduced coronary flow reserve, 30–80% of such patients would appear to suffer from this condition.17, 32 It is evident that the apparent prevalence of a particular disorder is related to the degree of enthusiasm with which it is sought. The causes may act synergistically to produce symptoms. Cannon et al. found that 82% of patients with angina-like chest pain and normal epicardial vessels had either reduced coronary artery flow reserve and/or abnormal oesophageal motility disorders.33 However, the exact connection between these abnormalities remains obscure.

Recently, evidence has been accumulating that non-invasive testing is cost-effective for patients who present with chest pain.34 This study involved the comparison of the following interventions using a cost-effectiveness analysis and sensitivity analysis: no testing, exercise electrocardiography, exercise echocardiography, exercise single photon emission computerized tomography and coronary angiography. Both exercise electrocardiography and echocardiography were cost-effective for patients with mild to moderate risk of coronary heart disease. However, this study did not specifically report on patients with non-cardiac chest pain, and the diagnostic procedures referred to in the study were used to determine coronary artery disease (i.e. exercise electrocardiography, coronary angiography). Another study compared the costs of an Emergency Department-based accelerated diagnostic protocol, based on tests that detect acute myocardial infarction during its early stages, against hospitalization in patients with chest pain.35 The primary outcomes measured were the length of stay and the total cost of treatment. The findings suggested that, by using the accelerated diagnostic protocol, $567 in total hospital costs were saved per patient treated, and the hospitalization rates, length of stay and total costs for low risk patients with chest pain were also reduced. This study did not differentiate between patients with cardiac and non-cardiac chest pain.

Indirect costs

No studies have specifically assessed the indirect costs (e.g. work days lost, value of ‘unpaid’ work) of non-cardiac chest pain. However, based on studies in functional gastrointestinal disorders, such as irritable bowel syndrome and non-ulcer dyspepsia, the indirect costs are likely to be very substantial.36, 37 Indirect costs may also accrue annually because of the chronicity of symptoms.

Other gastrointestinal disorders

Due to the limited economic data for non-cardiac chest pain, we sought to review the costs in irritable bowel syndrome and non-ulcer dyspepsia. Irritable bowel syndrome, like non-cardiac chest pain, is costly because the drugs used to treat the condition are required long term, and there may be a tendency for many patients to re-present due to the inadequacies of current therapy.38, 39 On the contrary, the indirect and intangible costs of irritable bowel syndrome appear to be much greater, but these burdens obviously do not have such an impact on those responsible for purchasing and providing health care for irritable bowel syndrome sufferers.36

Functional dyspepsia, like irritable bowel syndrome, is costly (US $55 000 per 1000 population), largely due to work absenteeism and other indirect costs.40 The economic implications of functional dyspepsia are constrained by the fact that it is only possible to make this diagnosis in a minority of individuals with the disorder, because many do not seek medical attention and investigation is not appropriate for all that do.37 Studies which attempt to assess the cost-effectiveness of management may be further constrained by the fact that there is now good evidence that many patients with dyspepsia seek medical attention not so much because of the severity of symptoms, but because they fear that the symptoms signal the presence of some serious underlying disease.37

Chest pain clinics: the answer?

Chest pain clinics are common in the USA and are gaining in popularity in Australia and the UK. The purpose of these clinics is to systematically evaluate patients with low to intermediate risk chest pain (assumed to be non-cardiac after an appropriate diagnostic work-up in the Emergency Department). However, recently, the value of such clinics has been questioned.41 Chest pain clinics can reassure patients and reduce re-presentation rates, but only if they can offer a comprehensive diagnostic strategy.42 In a large number of patients (up to 39%), symptoms persist, functional status remains impaired and re-presentations are common.43, 44

Future studies

The role of empirical acid suppression is promising, but is as yet unproven. There have been a few trials conducted among patients with non-cardiac chest pain using the proton pump inhibitor omeprazole.45, 46 Fass et al. conducted a double-blind, randomized, controlled trial on 39 patients with suspected non-cardiac chest pain using either omeprazole (40 mg am/20 mg pm for 7 days) or placebo.45 They found that this omeprazole test was sensitive (78%) and specific (87%) for the identification of gastro-oesophageal reflux disease patients, which resulted in significant cost savings and decreased the number of diagnostic tests subsequently undertaken. In addition, a second trial was conducted among 36 patients randomized to receive omeprazole, 20 mg twice daily (n = 17), or placebo (n = 19) for 8 weeks.46 Patients taking omeprazole improved in terms of the number of chest pain free days (81% vs. 44%, P = 0.03) and chest pain severity (81% vs. 50%, P = 0.057) compared to placebo.46

There remains a need for an outcome-based management trial of proton pump inhibitors vs. standard care for all patients presenting to the Emergency Department with presumed non-cardiac chest pain. Acid suppression may reduce re-presentation rates and improve symptoms in up to 75% of these patients. Tricyclic antidepressants have been shown to be effective in treating patients with non-cardiac chest pain.47 It would be worthwhile testing empirical tricyclic therapy in those failing acid suppression at low risk of cardiac disease in a management trial. New pharmacological treatments employing visceral analgesics may have utility in the future; octreotide, a somatostatin analogue, has shown promise in treating selected groups of patients with non-cardiac chest pain.48 A preliminary placebo-controlled trial of octreotide in patients with non-cardiac chest pain showed a significant decrease in pain perception after intra-oesophageal balloon distension with a 30-mm long latex balloon.48

Conclusions

The answers to the list of questions raised in the ‘Introduction’ section remain largely unknown. Non-cardiac chest pain appears to be common in the community, although the epidemiology of this problem is ill-defined. Most studies have shown that extensive investigation does not adequately reduce re-presentations, despite trials of several different diagnostic algorithms in the Emergency Department. The prognosis for patients with non-cardiac chest pain remains controversial. The socio-economic consequences are substantial, placing a significant drain on hospital resources and the health care system, but need better quantification. In order to constrain the rising costs of investigation, whilst improving long-term outcomes, more precise data about the long-term clinical outcomes and economic consequences of non-cardiac chest pain are urgently required.

Acknowledgements

Guy D. Eslick is supported by a Biomedical Scholarship from the Gastroenterological Society of Australia (GESA).

We would like to thank Natasha A. Koloski for her assistance with references and advice.

Ancillary