https://orcid.org/0000-0001-8992-8347
Abstract
Objectives. There is a paucity of data regarding the association between the use of high-sensitivity troponin (hs-cTn) compared with conventional troponin (cTn) and outcomes in chest pain patients in emergency departments (EDs). This study examined the impact of hs-cTnT on prognosis in chest pain patients in EDs. Design. In an observational cohort study, we included chest pain patients visiting the EDs of 14 hospitals in Sweden from 2011 to 2016. The study population was retrieved from each hospital, and information on characteristics and outcomes was collected from nationwide registries. Cox regression was used to estimate adjusted hazard ratios with 95% confidence intervals (HR, 95% CI) for (1) 1-year all-cause mortality, (2) missed acute coronary syndromes (ACSs), (3) use of coronary angiography, and (4) revascularizations within 30 days. Results. We included 170461 patients with chest pain where 62669 patients were tested with cTn while 107792 patients were tested with hs-cTnT. We found 4149 (4.6%) deaths in the cTn group and 6087 (3.7%) deaths in the hs-cTnT group. Patients in the hs-cTnT group had 9% lower mortality (0.91, 0.87–0.94), and were 14% more likely to undergo coronary angiography (1.14, 1.10–1.17), and 12% more likely to be revascularized (1.12, 1.08–1.17) than patients in the cTn group. Conclusions. Patients with chest pain visiting EDs using hs-cTnT had lower mortality and a higher likelihood of undergoing coronary angiographies and revascularizations than those using cTn. There may be a survival benefit of being tested with hs-cTnT compared with cTn in patients seeking medical attention for chest pain.
Introduction
Chest pain is the second most common reason for seeking medical attention in the emergency department (ED) and is therefore, associated with significant costs to healthcare systems [Citation1]. The introduction of high-sensitivity cardiac troponin (hs-cTn) assays has led to an earlier rule-out of myocardial infarction (MI), a shorter length of stay in the ED, a decline in admissions for chest pain, and consequently large substantial healthcare savings [Citation2,Citation3].
Only a few studies have investigated the prognostic importance of the transition to hs-cTn assays in patients with chest pain [Citation4–9]. In patients with chest pain admitted to cardiac care units no difference in the risk of mortality within 1 year was demonstrated [Citation4]. Similarly, in a study where patients with chest pain in the ED were randomized to either testing with hs-cTnI or cTn, no impact on prognosis was found, although a significant number of patients otherwise considered cTn-negative were reclassified to myocardial injury when tested with hs-cTnI [Citation7]. However, in one study an increased risk of all-cause mortality was found in patients with chest pain after the introduction of the hs-cTn assay when historical controls who were tested with cTn were used as a reference [Citation8]. However, in a recent study in which hs-cTnT is compared with cTn in the Emergency Department, where patients testing with hs-cTn were associated with a lower risk of MI and hospitalization but no differences in mortality at 1 year [Citation9].
There is a lack of “real-world” studies that have compared outcomes in patients with chest pain tested with hs-cTn versus cTn at different hospitals during the same time period. Therefore, we aimed to compare risks of all-cause mortality, missed acute coronary syndromes (ACSs), early revisits to the ED, readmissions, and the use of coronary angiography and revascularizations in all-comers with chest pain who were visiting the EDs of 14 hospitals in Sweden using either the hs-cTnT assay or a cTn assay during the years 2011 to 2016.
Methods
Study population
The study population consisted of all patients ≥35 years of age who sought medical attention for chest pain in the EDs of 14 hospitals in Sweden from 2011 to 2016. The selection of hospitals was done so there would be both hospitals included that used a cTn assay or the hs-cTnT assay during the study period.
Approval of the study was provided by the Regional Ethics Committee in Stockholm, Sweden. The study adhered to the principles in the Declaration of Helsinki.
Data sources
The study population was retrieved from each hospital’s administrative ED database. The data were then transferred to the Swedish National Board of Health and Welfare, who linked information about comorbidities, outcomes, and dispensed medication to the study population. Information about cTn assays and the locally applied cut-offs for the diagnosis of MI was obtained from the SWEDEHEART registry [Citation4]. The construction of the dataset and retrieval of the study population is described in Supplemental Figure 1.
Troponin assays
The only hs-cTn used was the hs-cTnT assay manufactured by Roche Diagnostics, Basel Switzerland. With this assay, the lowest concentration measurable with a 10% coefficient of variation (CV) is 13 ng/L, and the 99th percentile value among healthy controls is 14 ng/L [Citation10]. The hospitals and troponin assays that are used are shown in Supplemental Table 1.
Definitions
The index date was defined as the date when the patient visited the ED with a principal complaint of chest pain for the first time during the study period. Medication after the index date was defined as at least one filled prescription during the following 6 months. The ICD codes for comorbidities and outcomes are described in Supplemental Table 2.
Outcomes and follow-up
The primary outcome was all-cause mortality within 1 year. All patients, both admitted and discharged, were included in the study the day of arrival to the ED with chest pain and were followed for 1 year or until the time of death, whichever occurred first. If the patient came back with chest pain more than 1 year after the index date, a new follow-up period would be started.
Secondary outcomes were i) missed ACS, defined as readmission for MI within 30 days from ED discharge when follow-up started and ended at the admission date for MI, at the time of death, or at 30 days. Furthermore, we evaluated the likelihood of undergoing ii) coronary angiography and iii) revascularizations. For both of these outcomes, follow-up started at the index date and ended at the time of death, or after 30 days, whichever came first. For outcome iv) revisits to the ED, patients were followed from discharge from the ED until an ED revisit, death, or 7 days following the index visit, whichever came first.
The last day of follow-up for all-cause mortality was December 31, 2017. For all other outcomes, the end of follow-up was December 31, 2016.
Sensitivity analyses
Hospitals where the hs-cTnT assay was used had more visits, were more likely to be a university or a teaching hospital and located in a larger city, therefore, we evaluated whether the characteristics of the hospitals themselves, rather than the use of the hs-cTnT assay, were the determinant of the outcome by evaluating the association between seeking medical attention at a hospital using cTn or hs-cTnT and all-cause mortality at 1 year in patients with abdominal pain, and dyspnea.
Statistical analyses
Baseline characteristics are described as frequencies and percentages for categorical variables, and means and standard deviations for continuous variables. Cox proportional hazards models were used to estimate hazard ratios (HR) with 95% confidence intervals (CI) for the association between the use of a cTn assay and the use of the hs-cTnT assay for all-cause mortality, missed ACS, coronary angiography, revascularization, revisit to the ED, and readmission to hospital. Two models for each outcome were conducted: adjusted for age and sex (crude model), and adjusted for all comorbidities and socioeconomic variables described in . Date was used as the underlying time axis. Patients were followed until the outcome of death, emigration, or end of follow-up, whichever came first. The proportional hazards assumption was met. Absolute risks were calculated as the proportion of visits that came under risk and experienced the specific outcome during follow-up. Furthermore, we subgrouped the study population into those who were or were not diagnosed with MI during the index visit, and investigated the risk of all-cause mortality in relation to cTn assay used. In sensitivity analyses we evaluated the association between visiting the ED for i) abdominal pain, or ii) dyspnea at hospitals using the hs-cTnT assay, compared with hospitals using cTn assays for the outcome all-cause mortality. We also evaluated the association between cTn assay used and the outcomes all-cause mortality, coronary angiography, and revascularization restricted to hospitals with PCI available onsite. We also determined proportions of patients treated with cardiovascular medication during follow-up, and estimated odds ratios with 95% CIs for being treated with cardiovascular medication using logistic regression.
Table 1. Baseline characteristics for 170,461 patients with chest pain in 14 emergency departments in Sweden in 2011 to 2016 in relation to use of high-sensitivity troponin T or conventional troponin assays.
SAS software (Version 9.4 of the SAS System for MS Windows, SAS Institute, Cary, NC, USA), and R (Version 3.5.1, The R Foundation for Statistical Computing) were used for all statistical calculations.
Results
Study population
From 2997111 visits to 14 EDs in Sweden, we included 170461 patients ≥35 years of age with 256098 visits for chest pain, of which 89443 were visits to hospitals using cTn assays, and 166655 were visits to hospitals using the hs-cTnT assay .
All-cause mortality and missed acute coronary syndromes
There were 10236 deaths within 1 year. The yearly mortality rate was 4.6% in the cTn group and 3.7% in the hs-cTnT group. There was a 9% lower adjusted risk of death within 1 year in the hs-cTnT group, compared with the cTn group (HR 0.91, 95% CI 0.87–0.94) .
Table 2. All-cause mortality within 1 year and missed acute coronary syndrome within 30 days of a visit for chest pain in emergency departments using hs-cTnT compared with cTn.
Within 30 days of the index visit there were in total 2705 deaths, 1154 (1.3% per visit), and 1551 (0.93% per visit) in the cTn, and hs-cTnT groups, respectively. There was a 15% lower adjusted early mortality in the hs-cTnT group compared with the cTn group (HR 0.85, 95% CI 0.79–0.92).
Coronary angiographies and revascularizations
In total, 20028 coronary angiographies were performed within 30 days of the visit to the ED. Among patients in the cTn and hs-cTnT groups, 10.9%, and 12.2%, respectively, underwent coronary angiography within 30 days. Patients in the hs-cTnT group were 14% more likely to undergo a coronary angiography compared with patients in the cTn group (adjusted HR 1.14, 95% CI 1.10–1.17) .
Table 3. Coronary angiographies and revascularizations within 30 days in patients with chest pain in emergency departments using hs-cTnT compared with cTn.
There were in total 12899 revascularizations performed within 30 days of the visit for chest pain. Patients in the hs-cTnT group were 12% more likely to be revascularized than patients in the cTn group (adjusted HR 1.12, 95% CI 1.08′–1.17). In the hs-cTn group 7.8% were revascularized compared with 7,2% in the cTn group within 30 days from the visit . The proportion of coronary angiographies that led to revascularization was similar in both groups. Overall, men who underwent coronary angiographies were more likely to be revascularized than women: 71% vs. 54% in the cTn group, and 70% vs. 49% in the hs-cTnT group .
Emergency department revisits and hospital readmissions
Revisits to the ED within 7 days of being discharged home directly were unusual with 3.5%, and 4.6% events in the cTn, and hs-cTnT groups, respectively. Overall, there was no association between cTn assay used and the risk of a revisit Supplemental Table 7.
Mortality in patients with or without myocardial infarction at the index visit
We subgrouped patients into those with or without MI at the index visit Supplemental Table 2. The associations were similar to the whole study population with lower mortality in both patient groups with or without MI.
New cardiovascular medication
Overall there were small differences in cardiovascular medication started in conjunction with the index visit Supplemental Table 5, Supplemental Table 8. However, patients in the cTn group were treated with cardiovascular medication more frequently than patients in the hs-cTnT group. However, when patients were subgrouped into those with or without MI at the index visit we found that patients with MI in the hs-cTnT group were more likely to be treated with aspirin, beta-blockers, ACEi/ARBs, and statins at discharge Supplemental Table 8. After adjustment for confounders, including revascularization, the odds for being treated with cardiovascular medication was higher in the hs-cTnT group for all drugs except for treatment with P2Y12i and NOACs Supplemental Table 3, Supplemental Table 8.
Sensitivity analyses
When we restricted our analyses to hospitals that had PCI available onsite we found similar associations between troponin assay used and the all-cause mortality, coronary angiography, and revascularizations as in the whole study population. Among patients with abdominal pain there was a higher adjusted risk of death in patients in hospitals using hs-cTnT compared with cTn (HR 1.07, 95% CI 1.04–1.11). Similarly, among patients with dyspnea the risk of death within 1 year was higher in the hs-cTnT, compared with the cTn group (HR 1.03, 95% 1.01–1.06).
Discussion
In an observational cohort study, which included 170461 patients with chest pain in 14 EDs in Sweden from 2011 to 2016, we found that the use of hs-cTnT compared with cTn was associated with a 9% lower 1-year mortality, both in patients with and without MI at the index visit. There was a modest increase in both coronary angiographies, and revascularizations within 30 days in the hs-cTnT group.
The previous studies that evaluated the prognostic impact of being tested with hs-cTn, compared with cTn assays in chest pain patients, the majority found an unchanged mortality [Citation4–7,Citation9], except for one [Citation8]. Three of these studies, except for two [Citation4,Citation6,Citation8], used historical controls, which may have introduced bias. In the two studies [Citation5,Citation7] that were conducted prospectively, there was no association found between the troponin assay used and prognosis. A recent study partly supports our findings as there were reduced myocardial infarctions, hospital admission but no change in all-cause mortality after one year [Citation9]. We subgrouped patients into those with or without MI at the index visit. The association between troponin assay used and mortality was consistent in both patients with and without MI. Interestingly, in a nationwide Swedish study where patients with MI diagnosed with hs-cTnT were compared with those diagnosed with cTn no association between troponin assay used and all-cause mortality was found, although there was a small reduction in reinfarctions during 1 year of follow-up [Citation11]. That study included patients admitted during the years 2010 to 2013, while the present study included patients during the years 2011 to 2016. During the later period, new clinical guidelines had been introduced during the later period, which may have impacted outcomes in patients with MI [Citation12,Citation13]. For example, the use of ticagrelor in Sweden was increased among MI patients [Citation14].
The lower risk of all-cause mortality was paralleled by an increase in coronary angiographies and revascularizations after adjustment for differences between the two groups. A similar increase was found in a study that included patients cared for in cardiac care units in Sweden [Citation4]. However, in that study, the increase in coronary angiographies and revascularizations was not paralleled by lower mortality. In the present study, among patients without MI 3.5%, and 4.0% underwent coronary angiography, and 1.6%, and 1.7% underwent revascularization in the cTn, and hs-cTnT groups, respectively, which is higher than in other studies [Citation7]. After adjustment for differences between the groups, there was a significant increase in the use of both coronary angiographies and revascularizations. The increase in coronary angiographies and revascularizations was even larger among patients with MI, and the proportion of patients who underwent revascularization was higher than in previous studies [Citation5,Citation7]. These differences may be related to the differences in prognosis found between our study and others. In a previous study, it was found that early mortality in patients with MI in Sweden was lower than in the UK and that it may have been associated with less frequent use of revascularizations in the UK [Citation15]. In our study, there were marked differences in the use of coronary angiography and revascularizations in men and women, with men being twice as likely to undergo coronary angiography, and three times more likely to be revascularized. The absolute risk of death in women compared with men within 1 year of an MI was 50%, and 67% higher in the cTn and hs-cTnT groups, respectively. This may partly be explained by higher age and more prevalent comorbidities in women compared with men, but also by differences in treatment during the acute event, and in long-term secondary preventive measures [Citation16].
The present study confirmed findings from a previous study of reduced admission rates in patients with chest pain tested with hs-cTnT compared with cTn [Citation3]. The increase in discharges directly from the ED may raise concerns about safety. Therefore, we evaluated if the proportion of discharged patients who came back with an MI within 30 days increased with reduced admission rates: only 13, and 14 patients per 10000 discharges, respectively, would come back and be diagnosed with an MI within 30 days in the two groups. Overall, there was no association between the troponin assay used and the risk of missed ACS, although among women the risk was 30% lower when they were discharged from an ED where the hs-cTnT assay was used. This may be related to women being twice as likely to be reclassified to myocardial injury when tested with hs-cTn, which may have led to more aggressive secondary preventive measures in women in the hs-cTnT group [Citation7]. In addition, it is thought that women have more unspecific symptoms when they present with an MI, and may therefore more often be discharged accidentally with ongoing myocardial ischemia, and may therefore have more benefit from an assay with higher sensitivity [Citation16].
Among patients with MI at the index visit 89% of patients who were not treated with aspirin or beta-blockers, and 84% of those without statin therapy were started on these medications at discharge in the hs-cTnT group. The frequency of treatment with new cardiovascular medications was lower in patients with MI in the cTn group. In our previous study only treatment with statins differed between cTn groups, which may explain why no difference in the risk of mortality was found in that study [Citation11].
Strengths
The main strength of our study was the large number of patients included, which led to a high precision in our estimates and allowed us to perform analyses in relevant subgroups. In addition, we included patients from both large and small hospitals in different regions of Sweden. Furthermore, we were able to characterize patients on a very detailed level using nation-wide healthcare registers, and information from Statistics Sweden, which allowed us to adjust for differences between groups. Information about outcomes was collected from the National Patient Register, and the Cause-of-Death Register, which are complete for the whole country and are the reason there was no loss to follow-up. Furthermore, we carried out direct comparisons between hospitals using cTn assays and the hs-cTnT assay during the same time period, which avoided bias associated with unmeasured confounders, which were time-dependent and may have changed during the study period of 6 years. Thus, we believe that the external validity of our findings is high and may be generalized to other countries and healthcare systems and standards that are similar to those of Sweden.
Limitations
The decision limit for MI using the hs-cTnT assay, and the cTn assays is not comparable. The decision limit for MI was several times higher in hospitals using the cTn assay. The association between the use of hs-cTnT and improved outcomes may partly have been related to this. In the subgroup of patients tested with cTn without MI, likely there would have been several patients who would have been reclassified to MI if tested with a hs-cTn assay. A limitation is that none of the troponin assays used the 99th cutoff for decision-making. Hospitals that used the hs-cTnT assay were larger, with more visits to their EDs, and thus had larger volumes of MIs, which may have affected prognosis. However, the size of hospital is not necessarily associated with ranking in the SWEDEHEART registry on quality of care in patients with MI [Citation14]. Indeed, two of the hospitals where cTn assays were used were among the top-ranked hospitals in Sweden consistently during the years 2014 to 2016 [13]. Furthermore, in both the hs-cTnT and cTn groups two hospitals had PCI available 24/7. When we restricted our outcome analyses to hospitals with PCI available the association between troponin assay used and risks of death, coronary angiography, and revascularizations was similar to the whole cohort. To evaluate if the size of hospital may have affected outcomes through higher volumes we evaluated if patients with the first and the third most common principal complaint in the ED, abdominal pain and dyspnea, had a higher mortality in hospitals where cTn assays were used. We found a higher risk of death in patients with abdominal pain, and among patients with dyspnea in hospitals where hs-cTnT compared with cTnT was used. This indicates that it was not the characteristics of the hospital that led to the lower risk found at hospitals using the hs-cTnT assay, but more likely the risk assessment with the hs-cTnT assay itself. Still, patients in hospitals using cTn assays were older, had more comorbidities, and a shorter length of education than patients in hospitals using the hs-cTnT assay. However, we had abundant high-quality information from several registers that allowed us to adjust for these differences in multivariable analyses. Another limitation was that we did not have information about type of MI, which is why we could not distinguish type 2 from type 1 MI, two entities that have completely different prognoses [Citation17]. However, we only used the MI diagnosis in the primary position, and only 19% of patients with type 2 MI would have an MI diagnosis in the primary position in the Swedish Patient Register. Another major limitation was that we were not able to identify patients with myocardial injury, a group that may be as large or even larger than the group of type 1 MI in the era of hs-cTn assays [Citation7,Citation18]. This may partly limit the comparability with other studies in which patients with myocardial injury, be it type 1 MI, type 2 MI, or non-ischemic myocardial injury were identified. There was no external adjudication of ICD-10 codes, which is practically impossible in a study covering tens of thousands of admissions. Thus comorbidity among the patients is derived from diagnosis codes with its deficiencies. Kidney failure with the diagnosis code N18, is rarely used in clinical practice when it is a mild to moderate form. There is also a limitation for peripheral artery disease where only the diagnosis code I73 is used. Finally, like in every observational cohort study there may have been residual confounding present.
Conclusion
In a large cohort of patients with chest pain, we found a lower mortality in those seeking medical attention in EDs where the hs-cTnT assay was used compared with EDs where cTn assays were used. The lower mortality was paralleled with an increase in the use of coronary angiographies and revascularizations. In addition, patients who were diagnosed with MI in the hs-cTnT group were more frequently treated with secondary preventive medication compared with patients with MI in the cTn group. Our findings indicate that the use of hs-cTnT compared with cTn in patients with chest pain may lead to a survival benefit. The lower risk of death may be mediated by a better risk assessment.
Supplemental Material
Download MS Word (112.1 KB)Acknowledgments
The authors thank Tomas Andersson at Tomas Andersson Biostatistik AB for assistance with data management and statistical analyses. The construction of the database used in this study was sponsored by an unrestricted grant from Idorsia.
Disclosure statement
The corresponding author, on behalf of all authors, declares that all authors have completed the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr. Holzmann has received consultancy honoraria from Idorsia. The remaining authors declare no conflicts of interest.
Additional information
Funding
References
- Goodacre S, Thokala P. The economics of cardiac biomarker testing in suspected MI. Clin Biochem. 2015;48(4–5):213–217.
- Twerenbold C, Jaeger M, Rubini Gimenez K, et al. Impact of high-sensitivity cardiac troponin on use of coronary angiography, cardiac stress testing, and time to discharge in suspected acute myocardial infarction. Eur Heart J. 2016;37(44):3324–3332.
- Bandstein N, Ljung R, Lundbäck M, et al. Trends in admissions for chest pain after the introduction of high-sensitivity cardiac troponin. Int J Cardiol. 2017;240:1–7.
- Eggers KM, Lindahl B, Melki D, et al. Consequences of implementing a cardiac troponin assay with improved sensitivity at swedish coronary care units: an analysis from the SWEDHEART registry. Eur Heart J. 2016;37(30):2417–2424.
- Chew DP, Zeitz C, Worthley M, et al. Randomized comparison of high-sensitivity troponin reporting in undifferentiated chest pain assessment, circ. Cardiovasc. Qual. Outcomes. 2016;9:542–553.
- Corsini A, Vagnarelli F, Bugani G, et al. Impact of high-sensitivity troponin T on hospital admission, resources utilization, and outcomes. Eur. Heart J. Acute Cardiovasc. Care. 2015;4:148–157.
- Shah AS, Anand A, Strachan FE, High-STEACS Investigators, et al. High-sensitivity troponin in the evaluation of patients with suspected acute coronary syndrome: a stepped-wedge, cluster-randomised controlled trial. Lancet. 2018;392(10151):919–928.
- Bandstein N, Wikman A, Ljung R, et al. Survival and resource utilization in patients with chest pain evaluated with cardiac troponin T compared with high-sensitivity cardiac troponin T. Int J Cardiol. 2017;245:43–48.
- Lau G, Koh M, Kavsak PA, et al. Clinical outcomes for chest pain patients discharged home from emergency departments using high-sensitivity versus conventional cardiac troponin assays. Am Heart J. 2020; Mar; 221:84–94.
- Saenger AK, Beyrau R, Braun S, et al. Multicenter analytical evaluation of a high sensitivity troponin T assay. Clin Chim Acta. 2011;412(9–10):748–754.
- Odqvist M, Andersson PO, Tygesen H, et al. High-sensitivity troponins and outcomes after myocardial infarction. J Am Coll Card. 2018;71:2616–2624.
- Visseren FLJ, Mach F, Smulders YM, et al. 2021 ESC guidelines on cardiovascular disease prevention in clinical practice: developed by the task force for cardiovascular disease prevention in clinical practice with representatives of the european society of cardiology and 12 medical societies With the special contribution of the european association of preventive cardiology (EAPC). Eur Heart J. 2021;42:3227–3337.
- Collet JP, Thiele H, Barbato E, et al. 2020 ECS guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation of the european society of cardiology (ESC). Eur Heart J. 2021;42:1289–1367.
- https://www.ucr.uu.se/swedeheart/dokument-sh/arsrapporter-sh/swedeheart-annual-report-2017/viewdocument. 2019.
- Chung SC, Gedeborg R, Nicholas O, et al. Acute myocardial infarction: a comparison of short-term survival in national outcome registries in Sweden and the UK. Lancet. 2014;383:1305–1312.
- Mehta LS, Beckie TM, DeVon HA, American Heart Association Cardiovascular Disease in Women and Special Populations Committee of the Council on Clinical Cardiology, Council on Epidemiology and Prevention, Council on Cardiovascular and Stroke Nursing, and Council on Quality of Care and Outcomes Research, et al. American heart association cardiovascular disease in women and special populations committee of the council on clinical cardiology, council on epidemiology and prevention, council on cardiovascular and stroke nursing, and council on quality of care and outcomes research. Circulation. 2016; Mar 1133(9):916–947.
- Chapman AR, Shah ASV, Lee KK, et al. Long-term outcomes in patients with type 2 myocardial infarction and myocardial. Circulation. 2018;137:1236–1245.
- Roos A, Bandstein N, Lundbäck M, et al. Stable high-sensitivity cardiac troponin T levels and outcomes in patients with chest pain. J Am Coll Cardiol. 2017;70:2226–2236.