Researchers at the Indian Institute of Technology Bombay in India and the QIMR Berghofer Medical Research Institute in Australia have created a rapid method to determine whether a COVID-19 patient is likely to show severe symptoms.

WHAT IT'S ABOUT

The classification algorithm that was developed is based on infrared spectra of blood plasma acquired using the Agilent Cary 630 FTIR Spectrometer by California-based biotechnology firm Agilent Technologies. 

In their study, whose findings were published in the journal Analytical Chemistry, the researchers collected infrared spectra of blood plasma from 160 COVID-positive patients from Mumbai – 130 as a training set for the development of the multivariate statistical model and another 30 as a blind test set for the model's validation. 

The Agilent spectrometer showed "slight but observable" differences in the sample blood plasma spectra between severe and non-severe COVID-19 patients. 

"In particular, there were differences in two infrared regions that correspond to sugar and phosphate chemical groups, as well as primary amines, which occur in specific types of proteins," said Michelle Hill, head of the Precision and Systems Biomedicine Research Group at QIMR Berghofer.

The study also found that diabetes is a "key predictor" of severe COVID-19, according to Sanjeeva Srivastava, professor at the Indian Institute of Technology Bombay. 

Following this, the algorithm was fed with other clinical parameters, such as age, sex, diabetes mellitus, and hypertension and then tested on 30 samples for a blind test. It was later revealed that it got a 69.2% specificity and 94.1% sensitivity in predicting who among COVID-19 patients would become severely ill.

However, it resulted in more false positives than predictions, Prof. Srivastava noted. "We hope that with more testing, we can reduce these false positives," he said.

WHY IT MATTERS

Healthcare systems across the globe have been overwhelmed by the continuing COVID-19 outbreaks, leading to shortages in hospital resources like beds and ventilators. 

The World Health Organization has emphasised the importance of early identification and triaging of patients based on severity to help free up resources. 

Agilent said in a statement that the latest research can potentially lend support to healthcare institutions making critical decisions on hospital resources. 

THE LARGER TREND

An AI tool was recently developed that can tell a COVID-19 patient's likelihood of survival from hospitalisation. The web-based COVID Risk Calculator by Hong-Kong based AI systems developer Deep Longevity provides a patient's COVID-19 risk score, expected time to death and survival probability curve. The company pointed out that assigning risks to admitted patients is still an "essential, albeit grim, necessity" as hospitals around the globe continue to be overwhelmed with new COVID-19 cases.

ON THE RECORD

"We are very excited about this study, and happily supported the researchers in their fight against COVID-19 by placing the Cary 630 FTIR spectrometer for this study. Their work highlights the potential of ATR-FTIR spectroscopy for COVID-19 and infectious disease research, and we will continue to support research in this field," said Andrew Hind, AVP of Research and Development under the Molecular Spectroscopy Division at Agilent.