It’s not often that a journal paper serves as an inflection point for a field of medicine. “Metrics of sleep apnea severity: beyond the apnea-hypopnea index” is one of those articles.
“In a nutshell, the paper¹ that was published by Malhotra and a panel of experts suggests clinically significant OSA and severity should not be measured by AHI alone,” says Hitendra Patel, MD, FACP, FCCP, Medical Director, Wellstar Pulmonary and Critical Care Medicine. “We know there are other risk factors, comorbidities for example, but what else should we be measuring? How can machine learning and AI help us sift through the data and make sense of it?”
In the study by Malhotra et al, the authors reviewed the history of the apnea-hypopnea index (AHI), calling it the “best studied metric of OSA, albeit imperfect.” They then explored alternative metrics and future directions to gather clinically meaningful OSA endophenotypes.
As researchers explore these options, technology can help answer an underlying question: “If AHI isn’t the optimal metric, then what metrics should we instead be factoring into the equation for defining OSA severity?”
AHI Imperfections
Since Block et al. first described hypopnea in 1979, definitions have evolved and eventually led to the publication of the AASM scoring manual. This evolution, however, has led to inconsistency and confusion. According to the study authors, “Just as the evolution and variation in respiratory monitoring technology and inconsistency in event definition has caused confusion and complicated comparisons across studies, the terminology for frequency of respiratory events has also been inconsistent.”
The authors note that while OSA is thought to affect approximately 1 billion people worldwide, the lack of clear event definition makes it difficult to understand if most are asymptomatic, minimally symptomatic, or lack candidacy for therapy. The historical use of AHI to define OSA is now being challenged, “motivated by a growing recognition of the limitations of the AHI to predict adverse effects of OSA and to predict responsiveness to treatment,” according to the paper.
There is concern that AHI falls short of capturing physiological abnormalities that underlie its neurocognitive, metabolic, and cardiovascular effects, according to the authors. AHI as a metric does not fully explain variance in these symptoms or diseases outcomes. The authors point out that AHI is limited in predictive ability and precision in reflecting the true OSA-related exposure that is the cause of adverse outcomes, individual differences in response to OSA, and competing (non-OSA) causes of outcomes of interest.
Additional Risk Factors & Metrics
Additional risk factors explored by the authors included daytime sleepiness, quality of life, motor vehicle crashes, hypertension, coronary artery disease, stroke, and mortality.
“For the last 20 years, we’ve defined sleep apnea with AHI index and focused treatment this way, but clinically AHI has not always been a very reliable metric,” says Patel. “When I treat a patient with an AHI of 8 who has hypertension and excessive sleepiness, and they feel better with improved quality of life, it is meaningful. On the flip side, I’ve got a patient with an AHI of 33 with no symptoms or comorbidities and who is not sleepy. Well, what have you treated? You’ve treated a number,” he says.
While one metric is unlikely to characterize OSA and its risks, the authors propose considering additional metrics including hypoxic burden, arousal intensity, odds ratio product, and cardiopulmonary coupling.
Novel Strategies
Sleep apnea severity can be quantified using any number of approaches. The authors propose five, including evaluation of symptom subtypes, genetics, blood biomarkers, machine learning, and wearable technologies. Advancement of these approaches have one thing in common—they all require vast amounts of data.
“Somnoware aggregates millions of sleep and respiratory datapoints and also pulls in patient data from electronic health record systems,” says Subath Kamalasan, Somnoware CEO. “The ability to aggregate data from testing devices, patient-reported surveys, wearables, etc can help pave the way for research to study novel metrics to characterize OSA and its risks. The platform is being utilized for original research by some of the leading research institutions in the US.”
Kamalasan continues, “The platform’s rules engine can also use logic to define criteria for sleep apnea severity. Rules can be created to flag patients that meet defined criteria for identification of risk factors. Based on those factors, patients can be evaluated as needed. The possibilities are endless. We look forward to utilizing our data to help researchers move beyond AHI for diagnosing OSA and classifying disease severity.”
Learn more about Somnoware. Schedule a demo at somnoware.com.
Source: 1. Malhotra A, et al. Sleep. 2021 Jul 9;44(7):zsab030
Franklin Holman is the Director of Marketing at Somnoware. He can be reached at franklinh@somnoware.com or on his LinkedIn profile.