New Test May Allow Early Discharge Of Chest Pain Patients

Each year in the US about six million people go to the emergency department with chest pain or other symptoms suggesting that they might be having a heart attack or other acute coronary syndrome (ACS). The vast majority of them do not have ACS, but because it is difficult to quickly rule out ACS many of them end up getting admitted to the hospital anyway.

Now a new study shows that a novel strategy using a high sensitivity cardiac troponin test can safely identify nearly two-thirds of people with chest pain in the emergency department who are at very low risk and who can safely be discharged.

Results from the High-STEACS study, which followed 6,304 patients in Scotland with suspected acute coronary syndrome, were published in the Lancet. In the first part of the study, investigators studied 4,870 patients, 19% of whom had a cardiac event within 30 days. 61% of the 3,799 patients who did not have a myocardial infarction had a high sensitivity troponin concentration below 5 ng/L. The negative predictive value was 99.6%. In the second phase of the study, the 5 ng/L threshold was validated in 1,061 patients. 56% of patients had levels below the threshold and the negative predictive value was 99.4%. The finding was consistent across all the major patient subgroups.

“Over the last two decades the number of hospital admissions due to chest pain has tripled,” said the first author of the paper, Anoop Shah (University of Edinburgh), in a press release. “The overwhelming majority of these patients do not have a heart attack. This study shows that low plasma cardiac troponin concentrations at presentation identify up to two-thirds of patients who are at very low risk of heart attack and could be safely discharged from the Emergency Department. Use of this approach is likely to have major benefits for both patients and healthcare providers.”

In an accompanying comment, Louise Cullenemail, William Parsonage, and Martin Than call the study “a huge advance.” They note, however, that the test may not be as sensitive in patients who present very early to the emergency department and that the results do not apply to “other assays, even other high-sensitivity assays.”

Asked to comment on the study, James De Lemos (University of Texas Southwestern Medical Center) said High-STEACS was the largest and best study to date:

“This is another important piece in the puzzle regarding applications of high sensitivity troponin testing for rapid disposition of ED patients with chest pain.  This study is the largest by far to test the ‘one and done’ strategy of excluding MI with a single troponin measurement, provided the value is very low or undetectable.  They are able to exclude MI in about 1/3 of individuals with this strategy, and demonstrate excellent NPV and very low event rates for those with an hs-cTn < 5.  This is an excellent study that incorporated two validation cohorts distinct from the derivation cohort.

“The problem here, as with all the other studies in this field, is that they didn’t actually discharge the patients.  Also, we do not know what ‘back end’ testing was done on these patients (i.e., stress testing, CTA etc).  Although it makes sense to discharge these low risk patients without additional evaluation, no one has explicitly studied this, and even the investigators are not doing this in their own centers.

“A few caveats need to be considered. First, this strategy will not have the same NPV in early presenters, as shown here as well. Second, although the authors say that there is no interaction by baseline GRACE score, the data do suggest better performance for low risk individuals. I believe that strategies that combines the hs-cTn with risk assessment and ECG findings, as has been done by Martin Than and his group from New Zealand, are the most intuitive and logical approaches to rapid ED evaluation.  Finally, and most importantly, the ‘one and done’ approach requires that the troponin assays are nearly perfectly reproducible. We need to be absolutely sure that 5 ng/L in one center is the same as 5 ng/L in another, and calibration problems or assay ‘drift’ will create major problems implementing this.”



Speak Your Mind