Urine test may predict COPD exacerbations 1 week in advance
Paired with mobile app, dipstick test could help to prevent flares: Study
A urine test developed in a dipstick format by Global Access Diagnostics helped not only distinguish between stable and exacerbation states among chronic obstructive pulmonary disease (COPD) patients but also was able to predict exacerbations a week in advance.
That’s according to the results of a new study in which patients tested their urine daily at home for inflammatory biomarkers, with data sent to a mobile app.
With these data, the researchers developed an artificial intelligence (AI) tool to predict the time to COPD exacerbation, a period when symptoms suddenly worsen. The team noted that the use of this new testing tool could offer doctors a chance to intervene early to prevent COPD flares or make them less severe.
“We found the AI tool could reliably predict a flare up in symptoms seven days prior to a diagnosis,” Chris Brightling, PhD, a consulting physician at Glenfield Hospital in the U.K. and the study’s senior author from the University of Leicester, said in a university press release. “The advantage of sampling urine is that it’s relatively quick and easy for patients to do at home on a daily basis.”
The study, “Artificial neural network risk prediction of chronic obstructive pulmonary disease (COPD) exacerbations using urine biomarkers,” was published in ERJ Open Research. Four of its authors work at Global Access Diagnostics, previously Mologic, which develops rapid diagnostic tests and partly funded the study.
Developing a test like ‘a personal weather forecast of an impending flare-up’
COPD occurs when long-lasting inflammation damages the airways, causing symptoms such as shortness of breath, persistent cough, and recurrent infections. Exacerbations, or flare-ups, are the “principal driver of poor health status, [hospitalizations], mortality and healthcare costs,” the researchers wrote.
Several clinical features and biomarkers have been proposed as predictors of COPD exacerbations, but none of them provide “a reliable risk prediction of the timing of an onset of an exacerbation,” the researchers wrote.
Brightling noted that what’s needed is a better predictor.
“It would be better if we could predict an attack before it happens and then [personalize] treatment to either prevent the attack or reduce its impact,” Brightling said. “We wanted to develop a predictive test that would act like a personal weather forecast of an impending flare-up.”
The researchers first looked for urine biomarkers that could distinguish between stable and exacerbation states in COPD. They drew on data from 55 adults with COPD who had at least one exacerbation in the previous year that required corticosteroids or antibiotics. This included patients admitted to the hospital to treat the flare-up. The patients had a median age of 69, and about two-thirds were men.
By comparing urine samples collected during periods of stable disease versus periods of COPD exacerbations, the researchers identified a set of 10 inflammatory biomarkers — NGAL, TIMP1, CRP, fibrinogen, CC16, fMLP, TIMP2, A1AT, B2M, and MMP8 — that could discriminate between disease states with an accuracy of 84%.
To validate their findings, the researchers asked a group of 105 additional COPD patients to test their urine daily for six months, sending the data back via an app on their mobile phones. Paired urine samples from 26 patients — who had a median age of 70 and experienced a total of 33 exacerbations — were used to validate the set of biomarkers.
The set of 10 biomarkers could discriminate urine samples taken during exacerbation from those collected during a stable disease period with an accuracy of 81%, the team found.
Test checks levels of 5 biomarkers in urine to predict COPD exacerbations
An artificial neural network (ANN) was used to identify the best panel of urine biomarkers to predict COPD exacerbations. This was then used to predict the risk of a future exacerbation. An AAN is an artificial intelligence (AI) method that teaches computers to process data in a way that is inspired by the human brain.
Based on data from 85 patients, five of the biomarkers — NGAL, TIMP1, CRP, fibrinogen, and CC16 — were selected to develop a prototype urine test called Headstart.
In this test, a strip of paper is dipped into a sample of urine to check for the levels of the five biomarkers. According to the team, the results can be read quickly using an optic reader, a device that translates visual data into digital data.
Using the ANN to recognize patterns in data, researchers were able to predict a COPD exacerbation within a 13-day time window with an accuracy of 89%, a sensitivity of 95%, and a specificity of 85%. Here, the test’s sensitivity refers to its ability to correctly identify those experiencing an exacerbation within 13 days, while specificity refers to correctly identifying those with stable disease.
This research is promising because it suggests we can use AI [artificial intelligence] analysis of urine samples to predict a flare up before it starts. … If it proves successful in the longer term, this testing could make sure patients get the treatment and care they need to reduce symptom flare-ups as quickly as possible.
Apostolos Bossios, MD, PhD, a respiratory physician-scientist at the Karolinska Institutet and Karolinska University Hospital, in Sweden, who was not involved in the study, said a test for predicting COPD flares would be helpful for both patients and clinicians.
“There is no cure for COPD, so monitoring and treatment is crucial for helping patients stay well enough to carry out their normal day-to-day activities,” said Bossios, also the head of the European Respiratory Society’s airway disease assembly.
“This research is promising because it suggests we can use AI analysis of urine samples to predict a flare up before it starts,” Bossios added. “If it proves successful in the longer term, this testing could make sure patients get the treatment and care they need to reduce symptom flare-ups as quickly as possible.”
About the Author
