WASHINGTON – Oxygen desaturation index (ODI) scores showed significant variation across two software systems, a study showed.
The researchers assessed the ODI scores of 106 patients using the ResMed ApneaLink Plus system (AL) and the Compumedics Grael Profusion PSG3 system (Comp). “AL ODI values tended to be higher than Comp ODI values, but with significant variability,” they said.
AL showed a bias of an additional 4.4 events per hour (95% limits of agreement, –5.8 to 14.6 events per hour) for ODI scores at 4% desaturation and a bias of an additional 7.1 events per hour (95% limits of agreement, –6.4 to 20.6 events per hour) at 3% desaturation (J Clin Sleep Med. 2017;13[4]:599-605).
This may be problematic for physicians evaluating patients during sleep studies who rely on ODI scores at 3% and 4% desaturations to create accurate apnea severity assessments, the investigators said.
“[The] wide limits of agreement in our study highlight that clinicians cannot be confident that an ODI4% recorded in the AL is the same as that recorded in the Comp,” wrote Yvonne Ng, MBBS, of the department of lung and sleep medicine at Monash Health, Victoria, Australia, and her colleagues. “The differences are large enough to significantly affect diagnostic thresholds for OSA [obstructive sleep apnea] and, in particular, moderate-severe OSA.”
The researchers gathered data from patients undergoing sleep analysis at the Monash Medical Centre, who were, on average, 47 years of age, had a body mass index score of 32 kg/m2, and had an apnea hypopnea index (AHI) of 23.2.
ODI3% scores analyzed through Comp diagnosed 66 patients with OSA (ODI3% greater than or equal to 5 events per hour), while desaturation events analyzed through the AL system diagnosed 90 patients, a 36% increase over Comp (P = .0002).
When researchers tested for moderate to severe OSA (ODI3% greater than or equal to 15 events per hour), 32 patients were diagnosed using the Comp system, compared with 59 patients using the AL system.
Disparities in these measurements create uncertainty among clinicians, who rely on ODI measurements for scores that are accurate and can be easily replicated using an algorithm, the researchers said.
“The current work demonstrates that significantly more patients would receive a diagnosis of OSA, or more particularly, moderate-severe OSA, with the AL ODI, compared to the Comp ODI,” Dr. Ng and her colleagues wrote.
When sensitivity scores for Comp and AL were compared, AL ODI3% scores were significantly more sensitive than Comp, with sensitivity scores of 96% vs. 58%.
Using different fingers for measuring desaturation during the test or differences in algorithms used to assess ODI scores were possible sources of the disparities, the researchers noted.
Differences in internal processing between the two systems were the most likely causes of the discrepancies between the data collected using each system, they added.
Because there is no universal standard for ODI measurements, the researchers were unable to determine which system was more accurate.
Several of the researchers reported receiving financial support, research equipment, or consultancy fees from various entities.
ezimmerman@frontlinemedcom.com
On Twitter @eaztweets