CPAP data is readily available. Increasingly, so are metrics from other therapy modalities. So why are some clinicians positing that therapeutic device-supplied numbers aren’t enough?
By Sree Roy
Sleep physician Ernesto Eusebio, MD, faced a conundrum with one of his obstructive sleep apnea (OSA) patients. Despite undergoing an in-lab CPAP titration and then trialing autoPAP when the titrated fixed pressure fell short, she remained symptomatic—even though the autoPAP reported a residual AHI below 5.
Perplexed, Eusebio, the founder and medical director of Sleep & Apnea Institute of Florida, sent her home with an EnsoData ensoHST. The photoplethysmogram (PPG)-based home sleep test (HST), designed for multinight diagnostic testing and long-term therapy monitoring and titration, identified many residual sleep-breathing events that the positive airway pressure device did not report.
Once Eusebio titrated her therapy using the HST data, lo and behold, her symptoms subsided.
Eusebio has since seen this scenario repeat. “We learned this because of the eyes that RPM [remote patient monitoring] gives us at home,” he says.
Eusebio readily concedes that his CPAP-adherent-but-symptomatic patients who are helped with HST for monitoring are rare. “It helped these patients, but it’s a small sample,” he notes.
However, there are other scenarios where he uses HST—generally, ensoHST or Compumedics’ Somfit—for therapy monitoring more often: 1) while patients are waiting for their scheduled in-lab study so they can start treatment earlier, and 2) for Inspire neurostimulator titrations, where patients often subsequently decline to return to the sleep lab because they feel so much better and have multinight HST data showing low residual AHI.
Eusebio’s experience reflects the insights gleaned by sleep clinicians across the United States and Canada who have begun folding HST into their OSA therapy management protocols. These clinicians are finding discrepancies between CPAP and HST data, addressing physiological concerns (such as unusual heart rate patterns) that would otherwise have stayed hidden, comparing standardized data when patients switch therapy modalities, and in some cases, discovering that what a therapeutic device indicates is “well-controlled” sleep apnea may not be controlled very well at all.
Sleep Health Analysis
“It’s changed everything,” says pulmonologist-sleep physician Sahil Chopra, MD, co-founder of Empower Sleep, where one- to two-week baseline home sleep studies are the norm, followed by nightly therapy monitoring, typically using either a SleepImage (cardiopulmonary coupling-based software) or EnsoData’s software.
For instance, CPAP’s effect on sleep quality is not universally positive. About a third of CPAP-adherent patients experience stabler sleep, more deep sleep, and less sleep fragmentation—the positive outcomes everyone hopes for. In another third, their sleep architecture is basically unchanged—neither positive nor negative. But in about a third, “their sleep goes down the drain,” says Chopra, citing the HST data.
The HST data allow clinicians to identify patients whose sleep quality deteriorates on CPAP—often due to endotypic factors such as a low arousal threshold—and adjust therapy (via options that include reducing pressures, adding a hypnotic, or switching modalities).
Heart rate, another metric unavailable from CPAP, has also become key to Chopra’s assessments. “When there is the absence of heart rate dipping, that is a predictor for cardiovascular mortality risk,” he says. Heart rate tracking allows him to gauge whether a therapy is normalizing a patient’s physiology or if perhaps a referral to a cardiologist is warranted.
CPAP devices effectively give physicians three “very limited” metrics: device usage, mask leak, and residual AHI, Chopra says, while a home sleep study reveals the health of a person’s sleep when they’re using the therapy.
Standardizing the Discussion
In Calgary, Alberta, Canada, limited in-lab sleep center capacity means almost all people undergo HST for an OSA diagnosis. To Elaina Zebroff, RRT, co-owner of My Sleep Services in Calgary, it just makes sense to continue to use HST to track therapy efficacy, whether the patient gets a CPAP from My Sleep Services or is referred elsewhere first, such as to a dentist for an oral appliance. “When patients are diagnosed through a home sleep study, we use those metrics when we’re talking to patients about what this all means,” Zebroff says. “It only makes sense to have that continuum of information when they go on treatment.”
Zebroff’s clinic puts every new CPAP patient on 30 days of ensoHST monitoring, then a round of either 7- or 30-day monitoring at their three, six, or 12-month check-in (depending on their baseline and progress). “We use most metrics available on the home sleep testing, in combination with consultation with the patient. So their subjective feedback, and the objective feedback, guide the next steps or clinical intervention,” Zebroff says.
Many clinicians who use HST for therapy monitoring emphasize that having an “apples-to-apples” comparison is critical for decision making. Comparing CPAP data to the recordings from, say, neurostimulator software, has little value because the algorithms aren’t identical, they argue.
Even using the same HST for monitoring that was used for diagnosis could prove important.
“The algorithms for scoring and detection of sleep-related breathing events and other related metrics can vary among the devices,” says sleep physician Dimi Barot, MD, founder and CEO of Arima Health Technology. “Once you’re comfortable with one or two different types of devices…try to standardize, so the quality of your data can also be standardized.”
At his practice, Barot finds that HST efficacy tests are less crucial for patients on CPAP (though a “fair amount” still benefit, he says). But for patients on non-positive airway pressure modalities, “we rely a lot more on the efficacy home sleep test and make decisions specifically based upon that,” which can be “fundamentally different” than decisions based on device data alone, he says.
AHI Discrepancies
Clinicians also share a surprising discovery about using HST and CPAP data concurrently: discrepancies in the residual AHI. In general, when they witness a difference, it’s the HST that reports a higher apnea-hypopnea burden.
“It’s more than a few points difference,” says Chopra, who is also boarded in internal medicine and critical care. There is no pattern based on baseline disease burden either. “It’s not a linear relationship,” he says.
It means that select patients that the CPAP classifies as having “well-controlled” sleep apnea may still be having mild, moderate, or even severe levels of untreated OSA (in some cases, along with significant hypoxia, according to Chopra’s HST data).
In Canada, Zebroff has arrived at the same conclusion. “We don’t use the AHI coming off the CPAP to guide clinical decisions,” she says. “It does not correlate, in our expertise and our clinical discretion, with home sleep testing or PSG….If we take the CPAP SD card, we’ll look at waveforms, flow analysis, and the leak—that absolutely guides clinical decisions. But just looking at the AHI off the CPAP machine doesn’t hold a lot of weight.”
Overlooked Sleep Health Factors
Sleep clinicians are also uncovering influences on sleep quality and OSA expression that have rarely been accounted for in clinical practice, including those stemming from behavior and environment.
Sam Rusk, co-founder and chief AI officer at EnsoData, has used ensoHST himself to glean how routine choices change his sleep metrics. Shifts in bedtime, eating certain foods, and alcohol intake all register in his sleep. Longitudinal HST can create a “patient feedback loop to understand what patients are doing at a deeper level, so they can make themselves healthier,” Rusk says.
At Empower Sleep, for patients exploring CPAP alternatives, nightly data helps determine which adjunctive options—such as nasal strips, mouth tape, or side-sleeping aids—move the needle. “They can track what’s working and what’s not,” Chopra says. “The more data you have, the more precise the recommendations can be.”
Managing Extra Logistics
However, the sheer volume of HST mailing, tracking, and data review can become overwhelming.
Reviewing the raw data is essential, sleep physician Barot emphasizes. So to save time elsewhere, Arima Health Technology partners with third-party HST logistics vendors, including VirtuOx Sleepifi, to handle HST drop-shipping. “Play to your strengths,” Barot says. “Our strength is clinical care, and—assuming we have access to the appropriate technology to support exceptional clinical care—there are other companies that can do the logistics very well.”
A solution at My Sleep Services is to give each patient 45 minutes when they get their HST to ensure it is set up properly—including logging in to the app and making Bluetooth connections. Also, Zebroff serves as the point person to manage the extra data and address patient concerns.
Empower Sleep has gone a step further, building an AI-powered software platform to parse the data and flag patients who need attention. Chopra now sells the software by request to other clinicians embarking on longitudinal HST.
Validation Concerns
A stickler for American Academy of Sleep Medicine guidelines, Eusebio still sends many patients for in-lab sleep studies, despite the usefulness he has found in HST for therapy monitoring. He wishes there were a greater body of evidence for HST for repeat therapy efficacy and fine-tuning studies. “I always have that lab in the back of my head, thinking ‘I need to get this verified,’” Eusebio says. His advice? “Try to make up for its weaknesses by testing as many nights as possible before making a clinical decision,” he says, adding, “If you can, back it up with the lab.”
When asked, EnsoData’s Rusk clarifies that ensoHST is “validated to include a representative sample of the clinical population, but there are always going to be nuanced disease expressions that may require fallback on an in-lab study. There could continue to be more validation over time on specific patient populations with potentially confounding health conditions.” The company continues to work with investigators to publish more validation papers in peer-reviewed journals.
For his part, Eusebio also sees weaknesses with in-lab studies—weaknesses that can be overcome by the strengths of HST for therapy monitoring.
“We trust these lab tests—a CPAP titration or an Inspire fine-tune—where the patient slept for only 20 minutes. What we’re now finding is these patients go home on those settings, and they may still not be the right settings. They may benefit from additional adjustments…You can see the desaturations, the events happening, and you can confirm that the device was active at the time.
“I’m not saying [HST for therapy monitoring] will replace the status quo. I would never consider that, but it certainly will enhance it.”
Learn More About Newer Uses for HST:
ID 71138923 © Skypixel | Dreamstime.com
