Keeping Score on the STS Risk Score

The Society of Thoracic Surgeons has a formula for estimating a patient’s risk of dying following cardiac surgery, the STS Predicted Risk of Mortality score. Some new data reported during the past 2 months suggest that the score doesn’t predict death as well as some experts thought. It looks like when it’s applied to very sick patients, it may produce an overly pessimistic estimate and predict more deaths than patients will actually experience.

A patient’s mortality risk score derives from a list of 30 clinical and demographic inputs, factors like type of surgery, age, sex, hypertension, diabetes, cardiac history, vascular health, hemodynamics, etc. Based on all this, the formula spits out a patient’s probability of dying during the 30 days following the proposed surgery.

The PARTNER trial enrolled patients with severe aortic stenosis to compare a new technique of percutaneous aortic valve replacement with standard open surgical replacement. An initial report on results from the randomized portion of the study occurred last month at the annual meeting of the American College of Cardiology, and some more details on strokes and other neurologic outcomes got reported a few days ago at the American Association for Thoracic Surgery’s annual meeting in Philadelphia.

The patients entered into PARTNER were very, very sick. Their average age was about 83 years, and about 95% had the two highest grades of heart failure, New York Heart Association class III or IV. Their STS risk scores were also high, averaging about 12, which meant these patients had a 12% predicted risk of dying during the 30 days following open surgical replacement of their dysfunctional aortic valve.

These patients “were probably the highest 10% of risk on the STS score,” said Dr. D. Craig Miller, the cardiac surgeon who presented the neurologic data at the AATS meeting. “STS scores have never been validated at this extreme. Never before were enough patients [with scores this high] operated on to validate the STS score ” at this level, Dr. Miller said.

The reality was that the scores broke down. Instead of having a 12% 30-day mortality rate, the patients who underwent open surgery had about an 8% death rate. “We were pleasantly surprised by the low death rates,” at least in comparison to what the STS scores predicted, he said.

The only caveat to this good news was that the less-than-dire outcomes of some patients might have been very specific for the high-level treatment that patients received at the 26 centers that participated in this carefully structured trial. “What would be the results in the real world? That remains unanswered,” Dr. Miller said.

—Mitchel Zoler (on Twitter @mitchelzoler)