Prognostic Tool May Improve COVID-19 Management

A score developed in China to quantify COVID-19 pneumonia severity appeared to predict mortality in U.S. patients, and may help to eventually guide treatment decisions, a researcher said.

When applied to a small cohort of U.S. patients at the time of hospitalization, the MuLBSTA score was associated with in-hospital death with an area under the receiver operating curve of 0.813, reported Jurgena Tusha, MD, of Wayne State University School of Medicine in Detroit.

“An influx of SARS-CoV-2 infection has led to unanswered questions. One such question raised was how to risk stratify these patients in order to direct further management,” Tusha said at a presentation at the annual meeting of the American College of Chest Physicians.

She explained the components of the MuLBSTA score, which are weighted according to importance:

• Multilobe infiltrate (Yes +5)
• Absolute lymphocyte count less than 0.8 (Yes +4)
• Bacterial coinfection detected by sputum or blood culture (Yes +4)
• Smoking history (Active smoker +3, Prior smoker +2)
• History of hypertension (Yes +2)
• Age older than 60 (Yes +2)

Earlier work with the MuLBSTA score indicated it could predict 90-day mortality in patients, at rates of 0.47% for patients with a score of 0 to around 69% with a score of 20.

“I thought this was very interesting, because I thought the score takes into [account] some of the factors that actually make COVID a unique disease,” such as age and hypertension, said Marc Feinstein, MD, of Memorial Sloan Kettering Cancer Center in New York City, who moderated the session at which Tusha presented, but was not involved with the research.

The researchers reviewed charts for 163 hospitalized patients with COVID-19 pneumonia from March 15 to April 10, using those data to develop MuLBSTA scores. Mortality was the primary outcome, with secondary outcomes of length of stay and ventilator support.

Overall, 29.4% of patients died, with ICU mortality of about 51% and ventilator-associated mortality of about 63%. Feinstein said the latter figure seemed high — at his institution it was 25%, “similar to other ICU patients.”

Tusha said her data “might be biased” because it came from Michigan, where the population is older, and was from the beginning of the pandemic in March and April.

Patients in the ICU were a mean age of 68, 55% were women, 60% were white. Mean length of stay for ICU patients was about 15 days, and for non-ICU patients a little over 6 days. Mean MuLBSTA score was 8.67 for patients who survived and 13.6 for patients who died.

Tusha’s group also found a positive correlation between high MuLBSTA score and ventilator support (OR 1.30, 95% CI 1.17-1.44, P=0.0001) and length of stay.

Further studies are required to validate the study in larger patient cohorts with 90 day follow-up after discharge, the authors wrote, as Tusha noted this was a single retrospective chart review.

One clinician in the virtual audience asked how this score “helps me in patient management in terms of preventing decline in terms of mechanical ventilation?”

Tusha said her team was looking into that question: for example, how to devise treatment strategies “based on [patients’] clinical presentation on day 1.”

“When they first reach the ED if the score is at a certain level” it could help determine “whether or not they should be getting a different type of treatment,” she said.