Cardiac troponin (cTn) is instrumental in screening and diagnosing myocardial ischemia in adults. However, the role of cTn screening in the pediatric population is less clear. The purpose of this study was to evaluate the current clinical practice, diagnostic and prognostic value, and resource utilization associated with cTn assays in the pediatric population. A multicenter, retrospective review of all cTn assays performed on patients aged ≤18 years from January 2003 to December 2010 in the Intermountain Healthcare system was conducted. Data collected included patient demographics, location, presenting symptoms, provisional and discharge diagnoses, additional tests, clinical outcomes (hospitalization days, ventilation, and death), and patient charges. During the study period, cTn assays were performed on 3,497 pediatric patients. The most common presenting diagnoses were chest pain (40%), trauma (11%), and poisoning or drug overdose (9%). Irrespective of diagnosis, elevated cTn was associated with an increased rate of hospitalization, ventilation, and death. Overall, 12% of patients had elevated cTn. Of the patients with chest pain, 4% had elevated cTn, 53% of whom were diagnosed with myopericarditis. In the myopericarditis group, 66% presented with fever, and 98% had abnormal electrocardiographic findings. For patients presenting with chest pain, approximately $162,000 was spent per positive result. In conclusion, cTn screening has strong prognostic value in pediatric patients, even in noncardiac diagnoses such as trauma or drug overdose. However, cTn screening in pediatric patients with chest pain provides minimal benefits and is associated with increased resource utilization, unless patients have constitutional symptoms, such as fever and/or electrocardiographic abnormalities.
Serum cardiac troponin (cTn) screening is often obtained in pediatric patients in a variety of clinical settings, especially in patients presenting with chest pain to emergency rooms. This practice may result in excessive medical testing as well as clinical decision making based on data of unknown significance. The objective of the present study was to examine the current clinical practice, contribution to patient diagnosis and management, prognosis, and resource utilization associated with using cardiac cTn screening in a pediatric population.
Methods
This study was a multicenter, retrospective review using the Enterprise Data Warehouse database and associated medical records in the Intermountain Healthcare electronic system from January 2003 to December 2010. Intermountain Healthcare is an internationally recognized system of 23 hospitals and a full range of medical services that provides most care for all pediatric patients in Utah and the surrounding Intermountain region. All pediatric patients aged 0 to 18 years who had serum cTnI levels drawn during this time frame were included in this study. The study was approved by the University of Utah and Intermountain Healthcare institutional review boards.
Medical records were reviewed for the following data: (1) date of birth; (2) age; (3) gender; (4) location of the patient encounter (emergency room, inpatient, or outpatient), as well as the medical center where the cTnI assay was ordered; (5) presenting complaints; (6) electrocardiographic (ECG) data; (7) echocardiographic data; (8) hospitalization days; (9) mechanical ventilation (yes or no); (10) death; (11) discharge diagnosis; and (12) charges associated with the encounter during which cTnI was drawn. ECG results were categorized as normal or abnormal on the basis of the American Heart Association’s 2009 recommendations for standardization of interpretation. For example, findings such as J point elevation or interventricular conduction delay were considered normal, while findings such as ST-segment elevation or ectopy were regarded as abnormal. Hospitalization, ventilation, and death were considered adverse outcomes. The presence of any of these 3 adverse outcomes was considered the “composite adverse outcome.”
Cardiac troponin I isoform, a highly sensitive and cardiac specific biomarker, was used for cTn screening at Intermountain Healthcare hospitals during the entire study period. During the 8 years of this study, the normal reference range for serum cTnI assays in the Intermountain Healthcare hospitals varied from 0 to 2.0 ng/ml in the earlier years to <0.05 ng/ml in the later years. Between these time points, 0.4 ng/ml was considered the upper limit of normal. These changes reflect the overall improvement in the sensitivity and specificity of commercially available cTnI assays over the past decade. From 2007 to 2010, most assays were performed using the i-STAT point-of-care assay (Abbott Laboratories, Abbott, Park, Illinois). For the purpose of this study, cTnI levels were dichotomized as normal or elevated using a threshold of 0.4 ng/ml. Additionally, because some patients had multiple cTnI levels drawn per encounter and others only had 1 level drawn, only the highest level for each patient was used for this analysis.
Dichotomous and other categorical variables were compared using Pearson’s chi-square test without continuity correction. Distributions of continuous variables between groups were compared using Wilcoxon’s rank-sum tests. Associations between pairs of continuous variables were assessed using Spearman’s correlation coefficients. Continuous variables were compared using Wilcoxon’s rank-sum test. Where cTnI levels are considered as continuous variables, values recorded as inequalities were assigned the value of the end point of the interval. For example, “<0.02” was assigned a value of 0.02. All p values were unadjusted and assume 2 tails. Analysis was carried out using R language and environment (The R Project for Statistical Computing, Vienna, Austria).
Results
During the study period, 3,564 cTnI assays were performed in 3,497 pediatric patients. Table 1 lists the patient demographics, locations of cTnI assays, presenting diagnoses, additional cardiac tests ordered, and patient dispositions of the entire study cohort. There were 1,941 male patients (56%), and the median age of the cohort was 16 years (range 0 to 18). Most patients (1,998 [58%]) were screened in emergency rooms. A total of 403 patients (12%) had elevated cTnI levels, 159 of whom (40%) were seen in a tertiary care pediatric hospital. The most common reason for obtaining a cTnI assay was chest pain (1,409 [40%]), followed by trauma (397 [11%]) and overdose or poisoning (314 [9%]). Electrocardiography was performed in 2,688 patients (77%), and 791 (23%) underwent echocardiography. Most patients (1,831 [66%]) were discharged after their encounters, 958 (27%) needed hospital admission, 392 (11%) were mechanically ventilated, and 139 (4%) died. None of the patients were diagnosed with ischemic coronary artery disease or coronary artery abnormalities.
| Variable | Value |
|---|---|
| Patient characteristics | |
| Age (years), median (interquartile range) | 16 (13–18) |
| Male | 1,941 (56%) |
| Female | 1,556 (44%) |
| High cTnI | 403 (12%) |
| Normal cTnI | 3,094 (88%) |
| Location | |
| Tertiary care pediatric hospital | 603 (17%) |
| Other hospitals | 2,894 (83%) |
| Emergency room | 1,998 (58%) |
| Inpatients | 973 (28%) |
| Outpatients | 480 (14%) |
| Presenting symptoms | |
| Chest pain | 1,409 (40%) |
| Trauma | 397 (11%) |
| Overdose/poisoning | 314 (9%) |
| Syncope | 192 (6%) |
| Palpitations/arrhythmia | 183 (5%) |
| Miscellaneous | |
| Congenital heart disease with miscellaneous symptoms | 114 (3%) |
| Newborn with miscellaneous symptoms | 257 (7%) |
| Other | 631 (18%) |
| Additional tests | |
| Electrocardiography | 2,688 (77%) |
| Echocardiography | 791 (23%) |
| Patient disposition | |
| Discharged | 1,831 (66%) |
| Adverse outcomes | |
| Hospitalization | 958 (27%) |
| Ventilation | 392 (11%) |
| Death | 139 (4%) |
| Composite adverse outcomes | 1,028 (29%) |
Compared to patients with normal cTnI, patients with elevated cTnI were younger (median age 13 vs 16 years, p <0.0001), more often male (62% vs 55%, p = 0.01), more often inpatients (65% vs 23%, relative risk [RR] 10, p <0.0001), and were more often seen at a pediatric tertiary care hospital (40%, p <0.0001) ( Table 2 ). Compared to patients with trauma (23%) or poisoning or overdose (12%), a very small percentage of patients with syncope (3%) or chest pain (4%) had elevated cTnI levels. Irrespective of the presenting diagnosis, patients with elevated cTnI had increased risks for hospitalization (RR 3, p <0.0001) for longer durations (median 8.5 vs 3 days, p <0.0001), mechanical ventilation (RR 9, p <0.0001), and death (RR 25, p <0.0001).
| Variable | High cTnI | Normal cTnI | p Value | RR |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age (years), median (interquartile range) | 13 (0–17) | 16 (13–18) | <0.0001 | |
| Male | 248 (62%) | 1,693 (55%) | 0.01 | 1.3 |
| Female | 155 (38%) | 1,401 (45%) | ||
| Total | 399 | 3,052 | ||
| Patient location | ||||
| Pediatric tertiary care hospital | 159 (40%) | 444 (14%) | <0.0001 | 3.1 |
| Other hospitals | 244 (60%) | 2,650 (86%) | ||
| Visit type | ||||
| Emergency room | 50 (13%) | 1,948 (64%) | ||
| Inpatient | 261 (65%) | 712 (23%) | <0.0001 | 10.7 |
| Outpatient | 88 (22%) | 392 (13%) | <0.0001 | 7.3 |
| Presenting diagnoses | ||||
| Chest pain | 63 (4%) | 1,346 (96%) | ||
| Trauma | 91 (23%) | 306 (77%) | <0.0001 | 5.1 |
| Overdose/poisoning | 39 (12%) | 275 (88%) | <0.0001 | 2.8 |
| Syncope | 6 (3%) | 186 (97%) | 0.39 | 0.7 |
| Palpitations/arrhythmias | 9 (5%) | 174 (95%) | 0.78 | 1.1 |
| Miscellaneous | ||||
| Congenital heart disease | 35 (31%) | 79 (69%) | <0.0001 | 6.9 |
| Newborn | 66 (26%) | 191 (74%) | <0.0001 | 5.7 |
| Other | 94 (15%) | 537 (85%) | <0.0001 | 3.3 |
| Patient outcomes | ||||
| Hospitalization | 290 (89%) | 668 (27%) | <0.0001 | 3.3 |
| Hospitalization days, median (interquartile range) | 8.5 (4–22) | 3 (1–7) | <0.0001 | |
| Required ventilation | 217 (63%) | 175 (7%) | <0.0001 | 8.9 |
| Death | 96 (43%) | 43 (2%) | <0.0001 | 25 |
| Composite risk | 338 (93%) | 690 (28%) | <0.0001 | 3.3 |
| Additional tests | ||||
| Electrocardiography | 283 (11%) | 2,405 (89%) | ||
| Echocardiography | 269 (34%) | 522 (66%) |
There was no difference in age between patients with chest pain with elevated cTnI and those with normal levels ( Table 3 ). Male patients, however, accounted for 75% (RR 2.6, p = 0.004) of the elevated group. The RR of having elevated serum cTnI in patients presenting with chest pain and either fever or abnormal ECG results was 13 times higher (p <0.0001 for both) than in the rest of the patients with chest pain. The 2 most common discharge diagnoses for patients presenting with chest pain and elevated cTnI were myopericarditis (53%, RR 82, p <0.0001) and overdose or poisoning (7%, RR 27, p <0.0001) ( Table 3 ). The 13 patients who were discharged with nonspecific chest pain despite having elevated cTnI levels by our analysis did have levels >0.4 ng/ml but at the time of their encounters were within the upper limit of normal, and the results were interpreted as such by the managing physicians.
| Variable | High cTnI | Normal cTnI | p Value | RR |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age (years), median (interquartile range) | 16 (14–18) | 17 (15–18) | 0.42 | |
| Male | 47 (75%) | 698 (52%) | 0.004 | 2.6 |
| Female | 16 (25%) | 648 (48%) | ||
| Presenting characteristics | ||||
| Febrile | 24 (40%) | 40 (3%) | <0.0001 | 13 |
| Afebrile | 36 (60%) | 1,200 (97%) | ||
| Abnormal ECG results | 47 (78%) | 236 (19%) | <0.0001 | 13 |
| Normal ECG results | 13 (22%) | 1,027 (81%) | ||
| Discharge diagnoses | ||||
| Chest pain | 13 (21%) | 1,117 (89%) | ||
| Myopericarditis | 32 (52%) | 2 (0.2%) | <0.0001 | 82 |
| Pericarditis | 1 (2%) | 9 (0.7%) | 0.011 | 9 |
| Overdose/poisoning | 4 (7%) | 9 (0.7%) | <0.0001 | 27 |
| Palpitations/arrhythmia | 1 (2%) | 17 (1%) | 0.091 | 4.8 |
| Shortness of breath | 0 (0%) | 14 (1%) | 0.69 | 0 |
| Abdominal pain | 0 (0%) | 19 (1.5%) | 0.64 | 0 |
| Pulmonary embolus | 0 (0%) | 10 (0.8%) | 0.73 | 0 |
| Other | 10 (16%) | 58 (5%) | <0.0001 | 13 |
Most patients diagnosed with myopericarditis presented with chest pain (72%) ( Table 2 ). Compared to the overall cohort, they were similar in age (median 16 years) but showed a male preponderance (n = 34 [72%], RR 2.1, p <0.023; Table 4 ) and were more commonly inpatients (n = 35 [75%], RR 8, p <0.0001). With regard to fever and abnormal ECG results, 97% of patients with fever and 93% of patients with abnormal ECG results had elevated cTnI, but because of the small number of patients in the groups without fever and with normal ECG findings, these percentages were not statistically significant. The data for the myopericarditis cohort, as well as for other diagnoses, are listed in Table 4 .
| High cTnI | Normal cTnI | p Value | RR | |
|---|---|---|---|---|
| Myopericarditis | ||||
| Age (years), median (IQR) | 15 (10.5–17) | 17.5 (16.8–18) | 0.079 | |
| Male | 31 (91%) | 3 (9%) | 0.90 | 1 |
| Female | 12 (92%) | 1 (8%) | ||
| Fever | 28 (97%) | 1 (3%) | 0.22 | 1.1 |
| Abnormal ECG results | 40 (93%) | 3 (7%) | 0.12 | 1.4 |
| Trauma | ||||
| Age (years), median (IQR) | 16 (11–17) | 16 (14–18) | 0.016 | |
| Male | 66 (24%) | 210 (76%) | 0.48 | 1.2 |
| Female | 25 (21%) | 96 (79%) | ||
| Fever | 1 (9%) | 10 (91%) | 0.18 | 0.33 |
| Abnormal ECG results | 27 (28%) | 70 (72%) | 0.023 | 1.8 |
| Overdose/poisoning | ||||
| Age (years), median (IQR) | 18 (16–18) | 17 (16–18) | 0.17 | |
| Male | 25 (13%) | 169 (87%) | 0.75 | 1.1 |
| Female | 14 (12%) | 106 (88%) | ||
| Fever | 0 (0%) | 10 (100%) | 0.19 | 0 |
| Abnormal ECG results | 19 (24%) | 60 (76%) | 0.0008 | 2.7 |
| Sepsis | ||||
| Age (years), median (IQR) | 15 (2.25–17) | 15 (8–17) | 0.96 | |
| Male | 9 (31%) | 20 (69%) | 0.78 | 0.91 |
| Female | 13 (34%) | 25 (66%) | ||
| Fever | 10 (26%) | 28 (74%) | 0.087 | 0.55 |
| Abnormal ECG results | 9 (39%) | 14 (61%) | 0.31 | 1.5 |
| Palpitations/arrhythmia | ||||
| Age (years), median (IQR) | 10 (0–15) | 16 (12–18) | 0.005 | |
| Male | 5 (5%) | 96 (95%) | 0.98 | 1 |
| Female | 4 (5%) | 78 (95%) | ||
| Fever | 1 (6%) | 17 (94%) | 0.99 | 1 |
| Abnormal ECG results | 6 (9%) | 64 (91%) | 0.088 | 3.1 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
