Prevalence and outcome of patients referred for chest pain with high‐sensitivity troponin elevation and no diagnosis at discharge

Abstract

Introduction

Specialized chest pain units appear to increase the proportion of patients with acute chest pain who are properly evaluated, but some of them remain doubtful.

Hypothesis

The aim of this study was to evaluate the survival and occurrence of cardiovascular events in patients without diagnosis at the end of management of chest pain with high‐sensitivity troponin (Tn) elevation.

Method

All consecutive patients who came to the cardiac emergency room of Poitiers University Hospital between January 1, 2014, and August 7, 2015, for chest pain and Tn elevation were included. The primary endpoint was the number of undiagnosed patients; secondary endpoints included survival and major adverse cardiac events.

Results

A total of 1001 patients (695 male; mean age, 68 ±16 years) who had chest pain and Tn elevation were included. Median follow‐up was 24.5 (IQR, 14.7–29.5) months. Forty‐seven (4.7%) patients remained without diagnosis. Compared with patients with diagnosis, these patients were younger (53.6 ±19.7 years; P < 0.0001) and had less hypertension (29.8%; P < 0.0001), diabetes (4.3%; P = 0.0016), and history of coronary artery disease (6.4%; P < 0.0001). No patients died or experienced MACE in 6‐month follow‐up. Survival curves showed the probability of survival was excellent, not only at 6 months, but also at 36 months (P = 0.0025).

Conclusions

Less than 5% of patients referred for chest pain and with high‐sensitivity Tn elevation remained without diagnosis after adapted care in the chest pain unit. Their 6‐month prognosis was excellent.

1. INTRODUCTION

In the emergency department (ED), the most severe and most frequent outcome of chest pain is acute coronary syndrome (ACS). However, this symptom combines a wide variety of etiologies, including cardiac but not ischemic pain, as well as extracardiac pain. In the end, <25% of the chest pain admitted to the ED proves to be ACS,1 and mean costs of admission and lengths of stay for patients with nonspecific chest pain are significant.2 In the late 1990s in the United States, specialized chest pain units (CPU) were created. They appeared to increase the proportion of patients with chest pain who are thoroughly evaluated, with improved quality of care and lower costs.3 The effectiveness of these units in the management of myocardial infarction (MI) has been demonstrated; there was a 50% decrease in mortality and cardiovascular events.4, 5, 6

Anginal pain in ACS may have several presentations.7 Typical chest pain is characterized by a retrosternal sensation of pressure or heaviness radiating to the left arm, neck, or jaw, which may be intermittent or persistent. Atypical complaints are more often observed in the elderly, in women, and in patients with diabetes mellitus (DM), chronic renal disease, or dementia. Measurement of a biomarker of cardiomyocyte injury, preferably high‐sensitivity cardiac troponin (cTn), is mandatory in all patients with suspected non–ST‐segment elevation ACS.8, 9, 10 High‐sensitivity assays are recommended over less sensitive ones.11, 12, 13 Data from large multicenter studies have consistently shown that sensitive and high‐sensitivity cTn assays increase diagnostic accuracy for MI at the time of presentation compared with conventional assays.14, 15, 16 Cardiac Tn must be integrated into a compatible clinical scenario of ACS. Indeed, cTn and TnT and TnI are specific to myocardium; increases indicate a cellular injury, but there could be multiple causes17, 18, 19, 20, 21, 22, 23 that do not always indicate an ischemic cause.

The aim of this study was to evaluate the incidence of chest pain associated with high‐sensitivity Tn elevation without diagnosis at discharge in a CPU and to assess the survival and occurrence of cardiovascular events in these patients.

2. METHODS

2.1. Study design and patient population

This retrospective study was carried out at Poitiers University Hospital, France. A total of 1001 consecutive patients who came to the cardiac ED of Poitiers University Hospital between January 1, 2014, and August 7, 2015, and had cTn elevation were included. There were no exclusion criteria. Patients could be initially seen in the ED or directly in the CPU.

2.2. Data collection

Data were collected using hospital archives and hospital software archiving medical data. All the hospitalization and consultation reports were examined. If necessary, the general practitioners were contacted by phone and the reports of the referring cardiologists were obtained. Sociodemographic and clinical characteristics, medical history, laboratory results, electrocardiogram (ECG) description, and transthoracic echocardiography (TTE) reports were collected. Global Registry of Acute Coronary Events (GRACE) score was systematically recalculated. Cardiology examinations reports and their completion time, final diagnosis, and date of last follow‐up were also collected.

2.3. Diagnosis

The diagnosis was the one adopted at the end of the medical care, sometimes after several hospitalizations or consultations. Patients were considered without diagnosis if all cardiological hypotheses were invalidated by medical exams. Except for life‐threatening cases, decision‐making for each examination for every patient was discussed by the intensive care unit's practitioners at twice‐daily meetings, according to the clinical probability of ACS. ACS diagnosis was ruled out if there were no obstructive coronary arteries at angiography or cardiac computed tomography (CT), or if another diagnosis explaining the clinical picture was confirmed.

2.4. Electrocardiography

ECGs were classified into the following categories: normal, abnormal (ischemic T‐wave or ST‐segment change), abnormal but stable compared with prior ECG, and noncontributory (left bundle branch block or paced rhythm).

2.5. High‐sensitivity Tn

Tn elevation was defined by an increase beyond the 99th percentile,7 which was >0.013 μg/mL in our laboratory. We collected the peak level during hospitalization.

2.6. Survival

The last follow‐up of patients was collected in their medical records. All patients without diagnosis were systematically recalled (or their general practitioner, if necessary) and were asked if any major adverse cardiac events (MACE) had occurred; MACE was defined as a composite of cardiac death, MI, unstable angina, acute heart failure requiring hospitalization, ventricular arrhythmias, or if the patient needed a new hospitalization in a cardiology department.

2.7. Study endpoints

The primary endpoint was the number of undiagnosed patients, and secondary endpoints included mortality and occurrence of MACE in this population at 6 months of follow‐up.

2.8. Statistical analysis

Categorical variables are expressed as number and percentage and continuous variables as median and interquartile range (IQR) or mean ±SD. Groups were compared using the χ² test for categorical variables and the Mann–Whitney test for quantitative variables. Survival curves were built according to these cutoff values using the Kaplan–Meier method and compared using a log‐rank test. Analyses were performed using SAS version 9.4 (SAS Institute, Inc., Cary, NC). Two‐sided P values <0.05 were considered statistically significant.

3. RESULTS

3.1. Population characteristics

A total of 6062 consecutive patients came to the cardiac emergency room of Poitiers University Hospital between January 1, 2014, and August 7, 2015. Among them, 1001 patients (695 male; mean age, 68 ±16 years) had chest pain and Tn elevation and were included (Figure 1). All of them were admitted to the hospital. Baseline characteristics are summarized in Table 1.

Figure 1.

Flowchart of the cohort. A total of 6062 consecutive patients came to the cardiac ED of Poitiers University Hospital between January 1, 2014, and August 7, 2015. Among them, 1001 (695 males; mean age, 68 ±16 years) had chest pain and Tn elevation and were included. Median follow‐up was 24.5 months (IQR, 14.7–29.5 months). Abbreviations: ED, emergency department; IQR, interquartile range; Tn, troponin

Table 1.

Baseline characteristics of the patients, according to diagnosis

	With Diagnosis, n = 954	Without Diagnosis, n = 47	P Value
General
Female sex	289 (30.3)	17 (36.2)	0.42
Mean age, y	69.1 ±15.7	53.6 ±19.7	<0.0001
BMI, kg/m2	26.9 ±4.9	27.2 ±5.0	0.52
GRACE score	148.3 ±36.5	109.3 ±33.3	<0.0001
Medical history
HTN	575 (60.3)	14 (29.8)	<0.0001
DM	217 (22.7)	2 (4.3)	0.0016
Current smoking	231 (24.2)	14 (29.8)	0.39
Hyperlipidemia	463 (48.5)	18 (38.3)	0.18
Family history of CAD	161 (16.9)	7 (14.9)	0.84
AF	131 (13.7)	3 (6.4)	0.19
CAD	333 (34.9)	3 (6.4)	<0.0001
CABG	66 (6.9)	0 (0.0)	0.07
PAD	83 (8.7)	0 (0.0)	0.03
CKD ≥ stage 3	117 (12.3)	2 (4.3)	0.11
VTE	60 (6.3)	0 (0.0)	0.11
Neoplasia	99 (10.4)	3 (6.4)	0.62
Biology
Peak Tn, μg/L	0.30 (0.05–1.72)	0.18 (0.03–0.77)	0.0059
NT‐proBNP, ng/L	869 (221–3027)	320 (123–714)	<0.0001
Cr, μmol/L	86 (73–110)	76 (64–86)	0.0003
GFR, mL/min/1.73 m2	70 (51–86)	90 (74–109)	<0.0001
Leukocytes, ×109/L	9.0 (7.2–11.5)	8.7 (6.4–10.1)	0.0271
CRP, mg/L	5 (2–18)	6 (2–24)	0.54
18‐Lead ECG
Normal	232 (24.3)	28 (59.6)	<0.0001
Abnormal	506 (53.0)	14 (29.8)	0.0024
Abnormal but stable	123 (12.9)	4 (8.5)	0.50
Uninterpretable	90 (9.4)	1 (2.1)	0.12
TTE
LVEF, % (Simpson biplane)	54.2 ± 12.4	58.8 ± 11.5	0.0015
Segmentary kinetic abnormality	520 (54.5)	8 (17.0)	<0.0001
Severe valvular disease	54 (5.7)	0 (0.0)	0.10
Pericardial effusion	29 (3.0)	1 (2.1)	1.0000
RV overload	44 (4.6)	0 (0.0)	0.26

Abbreviations: AF, atrial fibrillation; BMI, body mass index; CABG, coronary artery bypass grafting; CAD, coronary artery disease; CKD, chronic kidney disease; Cr, creatinine; CRP, C‐reactive protein; DM, diabetes mellitus; ECG, electrocardiogram; GFR, glomerular filtration rate; GRACE, Global Registry of Acute Coronary Events; HTN, hypertension; IQR, interquartile range; LVEF, left ventricular ejection fraction; PAD, peripheral arterial disease; RV, right ventricular; SD, standard deviation; TTE, transthoracic echocardiography; VTE, venous thromboembolism.

Data are expressed as n (%), mean ± SD, or median (IQR).

3.2. Cardiac testing

A total of 977 patients (97.6%) underwent TTE. The next most commonly used cardiac examination was coronary angiography (688 patients). It was carried out in first intention in 81.1% of patients and permitted diagnosis in 588 patients (85.4%). Fifty (5.0%) had coronary CT in first intention, 28 (2.8%) had myocardial scintigraphy, 16 (1.6%) had cardiovascular magnetic resonance, and 3 (0.3%) had stress echocardiography. A total of 199 patients had no cardiac examination other than echocardiography, because of their echocardiographic or clinical diagnosis, refusal of cardiac examination, or decision to receive noninvasive treatment in view of advanced age or significant morbidity.

3.3. Diagnosis and follow‐up

A total of 735 patients (73.4%) had an ischemic cause; 193 patients (19.3%) had a cardiac but nonischemic cause; and 26 patients (2.6%) had an extracardiac pathology. Forty‐seven patients (4.7%) remained undiagnosed. The diagnoses of all patients are summarized in Table 2 , with the Tn peak for each one. Median follow‐up was 24.5 months (IQR, 14.7–29.5 months). Only 26 patients (2.6%) were lost to follow‐up at 6 months.

Table 2.

Final diagnosis (N = 1001)

		Peak Tn, μg/L
Ischemic
STEMI	235 (23.5)	3.50 (1.41–6.26)
NSTEMI	437 (43.7)	0.28 (0.07–1.06)
Chronic CAD	57 (5.7)	0.03 (0.02–0.06)
Coronary artery spasm	6 (0.6)	0.08 (0.04–0.14)
Cardiac nonischemic
Rhythm disorder	46 (4.6)	0.04 (0.02–0.09)
PE	35 (3.5)	0.07 (0.03–0.15)
HF	26 (2.6)	0.06 (0.03–0.15)
Severe valvular disease	21 (2.1)	0.05 (0.04–0.06)
Myocarditis	18 (1.8)	1.00 (0.23–1.43)
Takotsubo syndrome	13 (1.3)	0.31 (0.26–0.53)
Cardiac tamponade	8 (0.8)	0.05 (0.04–0.08)
Acute aortic syndrome	5 (0.5)	0.13 (0.04–0.23)
Hypertensive crisis	5 (0.5)	0.02 (0.02–0.03)
HCM	5 (0.5)	0.03 (0.02–0.07)
Others	11 (1.1)
Extracardiac disease
Infectious pneumonia	9 (0.9)	0.03 (0.02–0.05)
COPD	2 (0.2)	0.03 (0.03–0.03)
Others	15 (1.5)
Unknown etiology
	47 (4.7)	0.18 (0.03–0.77)

Abbreviations: CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; HCM, hypertrophic cardiomyopathy; HF, heart failure; IQR, interquartile range; NSTEMI, non–ST‐segment elevated myocardial infarction; PE, pulmonary embolism; STEMI, ST‐segment elevated myocardial infarction; Tn, troponin.

Data are expressed as n (%) or median (IQR).

3.4. Unknown etiology

Forty‐seven (4.7%) patients remained without diagnosis. Two of them (4.3%) were lost to follow‐up before 6 months. Compared with patients with diagnosis, these patients were younger (age 53.6 ±19.7 years; P < 0.0001), had less high blood pressure (29.8%; P < 0.0001), less DM (4.3%; P = 0.0016), and less history of coronary artery disease (6.4%; P < 0.0001). Their GRACE score was lower (109.3 ±33.3; P = 0.002). They had more normal ECG (59.6%; P < 0.0001), higher left ventricular ejection fraction (58.8% ±11.5%, P = 0.015), and fewer segmentary kinetic left wall abnormalities (17.0%; P < 0.0001). Median length of stay for these patients was 3 days (IQR, 2–5 days). None of the followed patients died or had MACE in 6 months of follow‐up. Only 1 of them (2.3%) was rehospitalized in the cardiac department, 4 months later, with a diagnostic of pulmonary embolism. Survival curves showed that the probability of survival was excellent not only at 6 months, but also at 36 months (P = 0.0025; Figure 2).

Figure 2.

Survival probability curves for patients with and without diagnosis. Kaplan–Meier survival curves and log‐rank test results comparing survival rates between patients with diagnosis (blue line) and without diagnosis (red line)

4. DISCUSSION

The fundamental purpose of a CPU is to differentiate ACS from other causes of chest pain. Data from the Germany⁶ and the United Kingdom³ demonstrated the superiority of CPUs compared with standard emergency service units (EDs), with a better prognosis for treated ACS patients.

In our study, the diagnostic sensibility of patients admitted to the CPU for chest pain and elevated Tn was 95.3% (95% confidence interval: 91.8–98.8, P 0.0001). In our population, coronary angiography was the cardiac examination most often used in first intention (81.1% of the cases). This rate is close to the 88.3% found in the Illman study.24 The difference is explained by the use of coronary CT as the first‐intention examination in 5% of our cases. This is a local feature due to the easy access to a dedicated CT scanner. Two diagnoses were underrepresented in our population. The first was spontaneous coronary artery dissection (SCAD). According to the literature, SCAD is observed in ≥10% of women age < 50 years presenting with ACS,25 and up to 35% in a recent Japanese study.26 We had 28 (2.8%) women age < 50 years with a diagnosis of ACS, so we could expect 2 or 3 cases of SCAD. The second was coronary spasm with only 6 (0.6%) patients, which was far less than that in previous literature.

Regarding the 47 patients (4.7%) remaining without diagnosis despite adapted medical care, their prognosis was excellent, because none had died or had MACE at 6 months of follow‐up. Significant characteristics of this group compared with patients with diagnosis, such as their lower age, lower GRACE score, lower cTn peak, lower N‐terminal pro brain natriuretic peptide, more normal ECG, and higher left ventricular ejection fraction, are expected compared with a population with a large proportion of ischemic etiologies and explain part of the excellent prognosis in these patients. Their lower prevalence of high blood pressure and DM and better glomerular filtration rate are major confounding factors because of the younger population. However, among these 47 patients, if all of them had TTE, 23 (48.9%) had only 1 other cardiac test, 22 (46.8%) had 2, and 2 (4.3%) had 3. We realize that TTE leads to a main hypothesis, but if this hypothesis is rejected by the first test, investigations stop in half of the cases. In these patients, coronary angiography is still the examination most often used in first intention. The number of undiagnosed patients could be lower with further tests, but given the excellent prognosis of these patients, there might be no expected benefit from these additional examinations.

Thus, cTn elevations are common in many disease states and do not necessarily indicate the presence of MI. Indeed, lowering the decision cutoff allows earlier detection of MI, increases the number of cases with MI, and decreases the number of cases with unstable angina.27 However, these undiagnosed patients question the specificity of cTn.

4.1. Study limitations

This study has some limitations. First, it is a retrospective work, based on the analysis of hospitalization and consultation reports. Second, despite a large number of patients, it remains a single‐center study, with local features. Some baseline characteristics between the 2 groups were different and may be confounding factors on morbidity and mortality.

5. CONCLUSION

In this large single‐center study of patients referred for chest pain and with high‐sensitivity Tn elevation, <5% of patients remained with an unknown etiology for their chest pain, despite adapted care in our CPU. Their 6‐month prognosis was excellent.

Conflicts of interest

The authors declare no potential conflicts of interest.

Lordet V, Lesbordes M, Garcia R, et al. Prevalence and outcome of patients referred for chest pain with high‐sensitivity troponin elevation and no diagnosis at discharge. Clin Cardiol. 2018;41:953–958. 10.1002/clc.22984