A new machine learning algorithm seems to be able to use pulse oximeter waveforms to estimate volume status in patients with cirrhosis.
Researchers from the Northwestern University Feinberg School of Medicine, in Chicago, found that analyzing fingertip photoplethysmography waveforms during breath-holding estimates intravascular volume overload more accurately than measuring serum brain natriuretic peptide (BNP).
Nikhilesh R. Mazumder, MD, MPH, a transplant hepatology fellow at Northwestern, said inaccurate estimation of the intravascular space can lead to adverse events in patients with cirrhosis who are taking diuretics for volume overload.
“Seeing all these patients in clinic can be challenging, so I was thinking about a way that we can quantify their physiology noninvasively from home,” Mazumder said. Although decreased pulse pressure at the end of breath holding represents normal physiology, its absence is a sign of volume overload, he added. His group presented the findings at the 2020 Liver Meeting (abstract 126).