番茄社区

PARIS -- Using different cutoffs for high-sensitivity cardiac troponin I (hs-cTnI) testing based on risk accurately ruled out MI and sent patients home from the emergency department sooner without missing adverse cardiac events, the HISTORIC trial found.

In the stepped-wedge trial of over 30,000 consecutive patients, introduction of the risk-based approach reduced length of stay at the emergency department by over 3 hours compared with standard care (6.8 vs 10.1 hours, P<0.001), reported Nicholas Mills, MD, PhD, of the University of Edinburgh in Scotland.

And 74% of patients under the new pathway were discharged without requiring hospital admission versus 53% under standard protocols (adjusted risk ratio 1.57, 95% CI 1.34-1.83, P<0.001).

For the primary safety endpoint, 2.5% of patients in the standard group died from cardiac causes or had an MI at 12 months post-discharge versus 1.8% of those in the early rule-out group (adjusted OR 1.02, 95% CI 0.74-1.40).

"Adoption of this approach will have major benefit for both patients and healthcare providers," said Mills during a late-breaking press briefing at the 2019 European Society of Cardiology (ESC) congress.

For example, many patients will need only a single troponin test under the algorithm to lead to a decision on admission, he noted, which could have "absolutely enormous" cost savings.

After a baseline hs-cTnI test, those identified as low risk (<5 ng/L) and with symptoms beyond 2 hours are discharged (or retested if symptom-onset was within 2 hours). High-risk patients -- those with a troponin level greater than the sex-specific 99th percentile cutoff, the threshold recommended in guidelines -- are admitted to hospital.

Intermediate-risk patients (5 ng/L to 99th percentile) were re-tested at 3 hours, after which those with a level <3 ng/L were discharged and the rest were admitted to hospital.

"There's an argument that it might be superior to using 99th percentile [for all patients], because it helps clinicians to focus on the highest-risk patients," Mills said. "There's lots of practical reasons why this should be a way to triage patients with acute chest pain."

Roxana Mehran, MD, of Mount Sinai Hospital in New York City and one of the press briefing moderators, called HISTORIC a landmark study but told 番茄社区 that it's likely the FDA would want a study conducted in U.S. emergency departments, with particular attention to the timing of blood draws, before a similar risk-based approach could be implemented.

"I entirely agree," said James Januzzi, MD, of Massachusetts General Hospital in Boston.

"One cannot translate results from different healthcare systems to the United States for one very specific reason: In the United States, troponin testing is done with a much lower threshold," he said, which skews the proportion of intermediate risk -- and medically more complex -- patients among those tested to be higher in U.S. emergency departments.

But another obstacle exists. The FDA has restricted reporting results of these high sensitivity troponin assays to concentrations of 6 ng/L or above. The first of these assays was cleared for use by the agency in 2017.

Januzzi told 番茄社区 that the agency has focused on the fact that in these very low concentrations the assays become less precise, with concerns about incorrect assignment of either a true or false positive, or a true or false negative.

"To be very clear, however, that's based on analytical assumptions and less on the results of prospective clinical trials," he said. "What this study adds is that despite the fact that the imprecision of the assay at this very low concentration rises quite substantially, the risk of misattribution of either the presence of an MI or the absence of an MI is extremely small, and it's reflected in the safety data."

Mills said the results have already changed practice in Scotland and that other countries are moving in this direction, too.

"The guidelines have been wedded to using a single threshold because they didn't have any evidence to base the guidelines on, that's why they used expert consensus," said Mills. "It's really since the adoption of high sensitivity assays that the evidence has started to emerge."

A simultaneously-released paper in from the same group provided backing for the cutoff of <5 ng/L, which identified two-thirds of patients as low-risk at presentation and offered a negative predictive value of ≥99.5% for cardiac death or MI at 30 days.

"The key insight is that the diagnostic performance of that risk-stratification threshold is identical across all age groups," Mills told 番茄社区. "Whether you're 20 or 90, it has exactly the same performance. That is completely different to the diagnostic threshold where the positive predictive value plummets as you get older."

HISTORIC included 31,492 consecutive patients with suspected acute coronary syndrome who presented at seven hospitals in Scotland. During the 6- to 9-month validation phase, the standard approach was used. This was followed by a 6-month randomization phase where the standard approach and risk-based pathway were used, and then another equivalent period using the new pathway alone. Adherence to the risk-based approach was noted in up to 92% of participants. Testing was done via the Abbott Diagnostics ARCHITECT_STAT hs-cTnI assay.

Follow-up for the primary safety outcome of MI or cardiac death occurred at 30 days and 1 year. Adverse cardiac events at 30 days, however, were too few to determine noninferiority between the two groups -- 57 events among the 14,700 patients (0.4%) treated during the standard phase period compared with 56 events among 16,792 patients (0.3%) during the early rule-out phase of the study.

Comment