From the Journals

Complete blood count scoring can predict COVID-19 severity 


 

A scoring system based on 10 parameters in a complete blood count with differential within 3 days of hospital presentation predict those with COVID-19 who are most likely to progress to critical illness, new evidence shows.

Advantages include prognosis based on a common and inexpensive clinical measure, as well as automatic generation of the score along with CBC results, noted investigators in the observational study conducted throughout 11 European hospitals.

“COVID-19 comes along with specific alterations in circulating blood cells that can be detected by a routine hematology analyzer, especially when that hematology analyzer is also capable to recognize activated immune cells and early circulating blood cells, such as erythroblast and immature granulocytes,” senior author Andre van der Ven, MD, PhD, infectious diseases specialist and professor of international health at Radboud University Medical Center’s Center for Infectious Diseases in Nijmegen, the Netherlands, said in an interview.

Furthermore, Dr. van der Ven said, “these specific changes are also seen in the early course of COVID-19 disease, and more in those that will develop serious disease compared to those with mild disease.”

The study was published online Dec. 21 in the journal eLife.

The study is “almost instinctively correct. It’s basically what clinicians do informally with complete blood count … looking at a combination of results to get the gestalt of what patients are going through,” Samuel Reichberg, MD, PhD, associate medical director of the Northwell Health Core Laboratory in Lake Success, N.Y., said in an interview.

“This is something that begs to be done for COVID-19. I’m surprised no one has done this before,” he added.

Dr. Van der Ven and colleagues created an algorithm based on 1,587 CBC assays from 923 adults. They also validated the scoring system in a second cohort of 217 CBC measurements in 202 people. The findings were concordant – the score accurately predicted the need for critical care within 14 days in 70.5% of the development cohort and 72% of the validation group.

The scoring system was superior to any of the 10 parameters alone. Over 14 days, the majority of those classified as noncritical within the first 3 days remained clinically stable, whereas the “clinical illness” group progressed. Clinical severity peaked on day 6.

Most previous COVID-19 prognosis research was geographically limited, carried a high risk for bias and/or did not validate the findings, Dr. Van der Ven and colleagues noted.

Early identification, early intervention

The aim of the score is “to assist with objective risk stratification to support patient management decision-making early on, and thus facilitate timely interventions, such as need for ICU or not, before symptoms of severe illness become clinically overt, with the intention to improve patient outcomes, and not to predict mortality,” the investigators noted.

Dr. Van der Ven and colleagues developed the score based on adults presenting from Feb. 21 to April 6, with outcomes followed until June 9. Median age of the 982 patients was 71 years and approximately two-thirds were men. They used a Sysmex Europe XN-1000 (Hamburg, Germany) hemocytometric analyzer in the study.

Only 7% of this cohort was not admitted to a hospital. Another 74% were admitted to a general ward and the remaining 19% were transferred directly to the ICU.

The scoring system includes parameters for neutrophils, monocytes, red blood cells and immature granulocytes, and when available, reticulocyte and iron bioavailability measures.

The researchers report significant differences over time in the neutrophil-to-lymphocyte ratio between the critical illness and noncritical groups (P < .001), for example. They also found significant differences in hemoglobin levels between cohorts after day 5.

The system generates a score from 0 to 28. Sensitivity for correctly predicting the need for critical care increased from 62% on day 1 to 93% on day 6.

Pages

Next Article:

   Comments ()