Practice pattern of use of high sensitivity troponin in the outpatient settings

Abstract Background High‐sensitivity troponin assays (hs‐Tn) detect lower serum concentrations than prior‐generation assays and help guide acute coronary syndrome (ACS) evaluation in emergency departments. Outpatient hs‐Tn utilization is not well described. Hypothesis Outpatient providers use hs‐TnT to triage patients with suspected ACS. Methods We compared the volume of outpatient prior‐generation troponin tests in the pre‐hsTn implementation period (January 2015‐March 2018) with outpatient hs‐TnT volume in the post‐implementation period (April 2018‐January 2020). Triage patterns were compared between patients with hs‐TnT≥99th vs <99th percentile, using two‐sample t tests. In patients triaged home, adverse events were compared between patients with hs‐TnT≥99th vs <99th percentile, using log‐rank tests. Results Across a large tertiary healthcare system, a mean of 80 prior‐generation tests/month were ordered during the pre‐hsTn implementation period compared with 12 hs‐TnT tests/month in the post‐implementation period. Prior‐generation orders rose by 1.72 tests/month during pre‐implementation, vs a decline of 2.74 hs‐TnT tests/month during post‐implementation (P < .001). Among 129 hs‐TnT orders, most were placed by cardiologists (54%) and primary care providers (32%). Patient symptoms at the time of troponin ordering included dyspnea (34%) and chest pain (33%), although 25% were asymptomatic. Among symptomatic patients (n = 74), those with hs‐TnT > 99th percentile were more likely to be sent to the ED (RR, 3.36; 95% CI, 1.22‐9.25; P = .002). Among patients sent home (n = 66), those with hs‐TnT > 99th percentile had more adverse events by 6 months (3.3% vs 22.2% RR, 6.67; 95% CI, 1.04‐42.9; P = .026). Conclusions In this healthcare system, outpatient troponin utilization significantly declined since hs‐TnT implementation. Some providers use hs‐TnT to triage patients with suspected ACS to the ED; others test asymptomatic patients and some send patients home despite high hs‐TnT values.


| INTRODUCTION
Compared to prior generation troponin assays, high-sensitivity assays for troponin (hs-Tn) can detect biochemical evidence of myocardial injury at lower concentrations, thus improving the sensitivity for patients with suspected acute coronary syndromes (ACS), and potentially accelerating diagnosis and early therapeutic decisions. 1,2 In 2017, the Food and Drug Administration approved the first hs-Tn assay in the United States, intended for the evaluation of patients with suspected ACS in inpatient and emergency department settings. As a result, evidence-based algorithms have been generated for using hs-Tn in the ACS evaluation of patients who present to the emergency department (ED) or in inpatient settings. 2,3 In contrast, there are few data describing the utilization of hs-Tn measurements in non-ED outpatient settings, 4,5 and in patients without ACS symptoms. We describe early practice patterns in outpatient utilization of a high sensitivity assay for troponin T (hs-TnT) in an academic tertiary care center.

| METHODS
We retrospectively identified all outpatient hs-TnT ordered since assay implementation (post-implementation period, 4 January 2018-31 January 2020) and all the prior fourth generation TnT ordered in the 3 years before hs-TnT implementation (pre-implementation period, 1 January 2015-31 March 2018) at the Brigham Health (which includes one academic tertiary care hospital, one community hospital and 28 affiliated clinics in Massachusetts). The rate of change in outpatient troponin orders between pre-implementation and postimplementation periods was compared to determine any statistically significant change in trend, using a two-sided alpha value of 0.05.
Electronic medical records were reviewed independently by two physicians to identify primary reason for ordering hs-TnT, as well as patient and ordering provider characteristics, and patient outcomes.
The prior generation Tn assay was the Roche fourth generation troponin T assay, which was used uniformly by all sites during the preimplementation period. The clinical cutpoint was 0.01 ng/mL, which corresponds with the 99th percentile reference limit. Hs-TnT were categorized as: undetectable (reported as <6 ng/L), detectable below the institutional 99th% sex-specific cutoff (≥15 ng/L in men, ≥10 ng/L in women), and detectable ≥99th% sex-specific cutoff. Unadjusted two-sample t tests were used to compare the proportion of symptomatic patients referred to the ED between patients with hs-TnT ≥99th% and < 99th%/undetectable (reference group). In patients sent home after outpatient encounters, log-rank tests were used to compare adverse cardiovascular event rates (defined as time to death, first ACS or hospitalization for heart failure) for patients with hs-TnT ≥99th% and < 99th% (reference group). Analyses were performed using the STATA software (College Station, TX). This study was approved by the institutional review board of Mass General Brigham.

| RESULTS
During the pre-implementation period, outpatient providers ordered an average of 80 Tn tests/month in total across the healthcare F I G U R E 1 Time trend in outpatient troponin ordering. The y-axis indicates the number of prior generation troponin assays ordered per month (blue line) and the number of hs-TnT assays ordered per month (orange line) during the study period (x-axis). The dotted vertical line represents the implementation of the hs-TnT assay, and divides the study period into a pre-implementation and post-implementation phase. The change in slope between the two periods was compared to determine any statistically significant change in trend (as shown by the P-value). hs-TnT, high-sensitivity troponin T system, with a mean increase of +1.72 tests per each month during this period ( Figure 1A). During the post-implementation period, providers ordered a total of 129 hs-TnT tests at an average of 12 hs-TnT tests/month across Brigham Health, with a testing rate that decreased by 1.01 tests per month during this period. The post-implementation period had a net change of 2.74 (95% confidence interval [CI]: 1.82-3.64, P < .001). fewer tests per month compared to the pre-implementation period.
Outpatient hs-TnT tests were most commonly ordered for dyspnea (34.1%) and chest pain (33.3%). 10.7% patients were in outpatient heart failure clinics that could provide intravenous diuresis (Table 1). Mean age of patients in which outpatient hs-TnT was ordered was 66.2 years, 51.9% were male, 31.8% had coronary artery disease and 46.5% had history of heart failure. The mean detectable hs-TnT value was 40.8 ng/L. A total of 31% of symptomatic and 29% of asymptomatic patients had undetectable hs-TnT (Table 1). Ordering providers were primarily cardiologists (54.3%), followed by primary care (31.8%) and ambulatory urgent care providers (7.0%) ( Table 1).
Overall, 52.7% of patients were sent home, 14.7% were scheduled for outpatient ischemic evaluation, and 16.4% were transferred to the ED ( Figure 2). One patient declined referral to the ED.
Lastly, almost a quarter of patients (n = 35, 24.0%) were asymptomatic at the time of hs-TnT ordering: the majority of them (66.7%) had elevated hs-TnT ≥99th%, while 23.3% and 10% had undetectable and < 99th% hs-TnT, respectively. The majority of them (83.3%) were sent home, while 10% and 6.7% were referred for outpatient ischemic evaluation and ED evaluation, respectively. Overall, 10% of asymptomatic patients experienced an adverse event in the subsequent 6 months (data not shown).

| DISCUSSION
We describe the outpatient utilization pattern of troponin assays in a large academic medical center in the United States. While there was an increase in outpatient ordering of prior generation troponin assays in the 3 years prior to implementation of the high-sensitivity assay, the launch of hs-TnT was associated with a subsequent significant decline in outpatient troponin utilization. Almost one quarter of patients were asymptomatic at the time of hs-TnT ordering. Higher outpatient hs-TnT levels were associated with a significantly higher likelihood of ED referral. Among patients sent home from clinic, higher hs-TnT levels were associated with significantly higher shortterm cardiovascular event rates.
Citing incomplete evidence, current consensus recommendations discourage the use of this assay outside of ACS evaluation in traditional ED and inpatient care settings. Several recent studies have tried to apply potential outpatient strategies to identify patients at low risk of coronary events. A recent meta-analysis found that a single undetectable hs-TnT, combined with a non-ischemic electrocardiogram, has a > 99% negative predictive value for ACS for patients Abbreviations: ED, emergency department; Hs-TnT, high-sensitivity troponin T assay; OP, outpatient. a The Hs-TnT <99th % group was used as reference for statistical analysis. b CV Event, cardiovascular event, defined as time to death, first acute coronary syndrome or hospitalization for heart failure, since index outpatient encounter.
declining outpatient troponin utilization since implementation of the high-sensitivity assay seems supportive of these guidance recommendations, and may be due to clinician unfamiliarity with the assay and appropriate concerns about interpretation of its results.
At the same time, however, our finding that 25% of hs-TnT assays was ordered for asymptomatic patients suggests that clinicians may

| CONCLUSIONS
Taken together, our findings support the need to develop guidance from professional societies on the use of hs-Tn in the outpatient settings. 4

DATA AVAILABILITY STATEMENT
The first and corresponding author had full access to data, which was analyzed at Brigham and Women's Hospital. We encourage investigators interested in data sharing and collaboration to contact the corresponding author.