The finding comes from a team led by Ali Azarbarzin of Brigham and Women’s Hospital and Harvard Medical School, working with colleagues at several institutions, including Atul Malhotra of the University of California, San Diego. The researchers pooled data from two of the most extensively characterized sleep cohorts in the world: the Sleep Heart Health Study, which enrolled community-dwelling adults with an average age of 64, and the Osteoporotic Fractures in Men Study, known as MrOS, which followed older men with an average age of 76. Together the analysis included 7,530 participants, roughly 47 percent of whom were men in the combined framing of the two cohorts, each of whom had undergone overnight polysomnography with continuous pulse oximetry.

The technical concept at the heart of the study is deceptively simple. Every apnea episode, whether central or obstructive, produces a fall in oxygen saturation followed by a recovery. Conventional metrics such as the apnea-hypopnea index count how often these events occur, and more recent work has quantified the hypoxic burden, meaning the cumulative depth and duration of the desaturations themselves. Oxygen overshoot burden flips the perspective: it measures the area under the curve of oxygen saturation above baseline, capturing how far and how long the blood oxygen climbs back — and sometimes beyond — after each event. The researchers computed this overshoot against two reference points, a stable baseline recorded during sleep and, in secondary analyses, a baseline recorded during quiet wakefulness, to confirm that the signal was not an artifact of how the baseline was chosen.

Over a median follow-up of nearly nine years — 8.92 years — the cohorts recorded 2,258 major adverse cardiovascular events, a composite that captures the kind of outcomes that matter most to patients: heart attacks, heart failure episodes, strokes, and cardiovascular death. When the team modeled oxygen overshoot burden as a continuous variable, they found that higher burden was associated with increased cardiovascular risk across the population, but the association was markedly stronger among individuals who exhibited any central apneas during their sleep studies. The interaction was statistically significant for both the sleep baseline and the wakefulness baseline calculations, with p values of 0.003 and 0.02 respectively, suggesting the finding is robust to how the reference point is defined.

The most striking results emerged when the researchers stratified participants by diagnosis and by overshoot burden. People with central sleep apnea whose oxygen overshoot burden was at or above the cohort median experienced a dramatically elevated rate of cardiovascular events: 54.6 percent of them suffered a major adverse cardiovascular event during follow-up, compared with 25.6 percent of controls. After adjustment for a comprehensive set of covariates, the hazard ratio was 1.45, with a 95 percent confidence interval of 1.16 to 1.82 and a p value of 0.001. In contrast, three other groups showed cardiovascular risk that was statistically comparable to the controls: patients with central sleep apnea but low overshoot burden, and patients with obstructive sleep apnea whether their overshoot burden was high or low.

That last contrast is what makes the study potentially paradigm-shifting. Obstructive sleep apnea, caused by the physical collapse of the upper airway, and central sleep apnea, in which the brain temporarily stops sending the signals that drive breathing, have long been lumped together in clinical scoring systems that count respiratory events without much regard to their mechanism. Yet the new data suggest that the cardiovascular consequences of the two disorders may travel along different physiological paths, and that the overshoot of oxygen after each pause in breathing may be a signature of the central form that carries particular danger. The result held up even after the researchers adjusted for the apnea-hypopnea index, the central apnea index, loop gain — a measure of the instability of the respiratory control system — and the hypoxic burden, indicating that oxygen overshoot captures risk information that these established metrics do not.

The biological rationale for the finding lies in oxidative stress. Repeated cycles of deoxygenation and reoxygenation, known as intermittent hypoxia followed by reoxygenation, are thought to generate reactive oxygen species in much the same way that reperfusion injury damages tissue after a blocked artery is reopened. Animal studies dating back to the early 1990s have shown that episodic hypoxia elevates blood pressure and drives inflammation and atherosclerosis, and human studies have documented oxidative stress, endothelial dysfunction, and vascular inflammation in sleep apnea patients. The overshoot may be the moment when this chemistry is most active: as oxygen floods back into blood that has just been depleted, the surge could catalyze the oxidative bursts that injure the vessel wall. Central sleep apnea, which is especially prevalent in patients with heart failure and is associated with unstable respiratory control and high loop gain, may produce particularly pronounced or repetitive overshoots, creating a vicious cycle in which the breathing disorder and the heart disease feed each other.

The clinical implications are substantial. Central sleep apnea has been a stubborn therapeutic problem: large randomized trials of adaptive servo-ventilation and continuous positive airway pressure in heart failure patients have produced mixed or sobering results, and the field has increasingly recognized that not all patients with the same diagnosis carry the same risk. By identifying a high-risk phenotype — central apnea with high oxygen overshoot — the study offers a potential tool for stratifying patients in future trials, and it raises the possibility that therapies aimed specifically at blunting the overshoot, whether through oxygen titration, ventilatory support, or phrenic nerve stimulation, could be tested in the patients most likely to benefit. Recent trials of nocturnal oxygen therapy and transvenous phrenic nerve stimulation are already probing this territory, and overshoot burden could become an endpoint or an enrollment criterion in the next generation of studies.

The authors are careful to note the limitations. The two cohorts are older and, in the case of MrOS, exclusively male, so the findings need validation in larger and more diverse populations. Pulse oximetry itself carries a known bias across skin pigmentation, a concern highlighted by prior research showing that oximeters can overestimate oxygen saturation in people with darker skin, which could affect the precision of overshoot measurements. The study is observational, so it demonstrates association rather than causation, and residual confounding cannot be fully excluded. Still, the size of the cohorts, the length of follow-up, the consistency of the signal across baseline definitions, and the specificity of the effect to central apnea make the result one of the most compelling pieces of evidence yet that the aftermath of each apnea — not just the event itself — shapes cardiovascular fate.

For patients and clinicians, the message is that the texture of sleep-disordered breathing matters, not merely its frequency. A night of breathing pauses that end in modest, controlled recoveries of oxygen appears very different from one in which each pause is followed by a dramatic rebound, and the difference may separate a benign pattern from a dangerous one. As the field moves toward physiological phenotyping of sleep apnea, oxygen overshoot burden joins hypoxic burden and loop gain in a growing toolkit of quantitative measures that promise to replace blunt event counts with a more precise picture of what is actually happening in the blood, the brainstem, and the heart. The next step, the researchers say, is validation — and, ultimately, interventional trials that test whether taming the overshoot can tame the risk.

Subject of Research: Oxygen overshoot burden in central sleep apnea and its association with major adverse cardiovascular events

Article Title: Oxygen overshoot burden of central sleep apnea and its association with cardiovascular outcomes

Article References: Azarbarzin, A., McKane, S., Stone, K. L., Germany, R., Redline, S., & Malhotra, A. (2026). Oxygen overshoot burden of central sleep apnea and its association with cardiovascular outcomes. Journal of Clinical Sleep Medicine, 22(1), Article 153. https://doi.org/10.1007/s44470-026-00161-0

Image Credits: AI Generated

DOI: 10.1007/s44470-026-00161-0

Keywords: central sleep apnea, oxygen overshoot burden, cardiovascular outcomes, sleep-disordered breathing, oxidative stress, hypoxic burden, polysomnography, pulse oximetry, heart failure, intermittent hypoxia, risk stratification, Sleep Heart Health Study

The study, led by Azarbarzin and colleagues and published in the Journal of Clinical Sleep Medicine, drew on 7,530 participants from two well-established cohort studies, the Sleep Heart Health Study and the Osteoporotic Fractures in Men Study, known as MrOS. Over nearly nine years of follow-up, 2,258 major adverse cardiovascular events occurred in this population. The investigators computed an oxygen overshoot burden for each participant: the cumulative area under the oxygen saturation curve that lies above an individual’s stable-sleep baseline, normalized for total sleep time. This is a deliberately individualized measure. Rather than comparing every patient against a fixed cutoff, the analysis anchors the signal to each person’s own resting saturation during uneventful sleep, so that even small relative rises above the personal baseline accumulate into a meaningful summary of nightly physiological stress.

The headline result is striking. Among participants with central sleep apnea and high overshoot burden, 54.6 percent experienced a major adverse cardiovascular event during follow-up, compared with 25.6 percent of control participants, corresponding to an adjusted hazard ratio of 1.45. Crucially, central sleep apnea with low overshoot burden carried no elevated risk relative to controls, and no corresponding association appeared in obstructive sleep apnea. The relationship survived statistical adjustment for the apnea-hypopnea index, the central apnea index, hypoxic burden, and, in the MrOS cohort, an estimate of loop gain, the control-system parameter that describes how vigorously breathing responds to disturbances in blood gases. That specificity is intriguing: it suggests the overshoot signal is not merely a proxy for how often someone stops breathing, but may capture something distinct about the instability of their ventilatory control.

The authors and commentators have proposed an intuitive biological explanation rooted in oxidative stress. The cycle of desaturation and reoxygenation that characterizes sleep apnea resembles ischemia-reperfusion injury, and repeated swings are thought to generate reactive oxygen species, inflammation, and vascular damage. Hypoxic burden, which integrates the depth and duration of event-related desaturation, already predicts cardiovascular outcomes better than a simple count of respiratory events, lending weight to the idea that the shape of the oxygen trace carries mechanistic information. Extending that logic above the baseline line is tempting: if falling oxygen is harmful, perhaps overshooting oxygen is harmful too, and the cumulative area of overshoot might quantify a dose of oxidative injury delivered night after night.

Yet a careful reading of the physiology counsels caution before accepting that interpretation. Oxygen overshoot as measured here is a relative oximetry signal, not demonstrated hyperoxia. A rise in peripheral oxygen saturation above a person’s stable-sleep baseline does not establish an elevated arterial partial pressure of oxygen, nor does it demonstrate increased tissue oxygen exposure. The study provides no mechanistic pathway, no oxidative biomarker measurements, and no bench or animal evidence showing that this specific above-baseline signal causes oxidative injury. Evidence from the intermittent hypoxia and reoxygenation literature cannot simply be transferred to this different signal. Indeed, previous work in obstructive sleep apnea found that greater post-event saturation overshoot was associated with lower nocturnal glucose, a pattern that argues against harm and even hints at benefit. Treating the oxygen rise itself as the causal exposure is, for now, premature.

A more plausible reading is that the oximetric overshoot is the visible tail of a much larger ventilatory and autonomic response. When an apnea terminates, the accumulated carbon dioxide and chemoreflex drive produce a vigorous recovery breath; the resulting hypocapnia is a direct signature of unstable ventilatory control, which is precisely the physiology that generates central sleep apnea and Cheyne-Stokes respiration in the first place. Human studies of apnea have documented marked sympathetic nerve activation and blood pressure surges around the termination of each event, followed by vagal modulation tied to lung inflation during the recovery phase. These autonomic oscillations, repeated hundreds of times a night, offer a credible route to myocardial infarction, arrhythmia, and stroke that does not require a modest rise in saturation to be intrinsically toxic. On this view, overshoot burden is a marker of the force of each recovery, not a poison in its own right.

This interpretation also exposes an important analytical gap in the new study. Adjusting for event frequency and hypoxic burden does not establish that overshoot is independent of the severity of each respiratory event. A longer or more severe apnea accumulates more hypercapnia, more chemoreflex stimulation, more arousal-related and sympathetic activation, and a longer loss of the vagal restraint normally provided by lung inflation. Those stimuli can then generate a larger recovery breath and a larger oxygen overshoot, and desaturation depth alone does not fully represent them. Consistent with this, work in obstructive sleep apnea has shown that ventilatory burden, a measure of the effort expended against collapsed airways, predicts cardiovascular outcomes and explains much of the variation in hypoxic burden. Oxygen overshoot may play an analogous role in central sleep apnea, summarizing a hidden physiological load that conventional indices leave unmeasured, without being the injurious agent itself.

Residual cardiac confounding remains another live possibility. In older community cohorts, central sleep apnea and Cheyne-Stokes respiration often reflect underlying cardiac dysfunction, elevated left-sided filling pressures, and prolonged circulation time, rather than mediating the cardiovascular consequences of those conditions. Heart failure identified through self-report and clinical records may miss subclinical disease, so some of the apparent association between overshoot burden and events could reflect undiagnosed cardiac impairment that both destabilizes breathing control and drives outcomes. Adjustment for loop gain in one cohort helps address the ventilatory-instability pathway, but detailed cardiac phenotyping, with objective measures of structure and function, is still needed before the direction of the arrow can be declared with confidence.

Two further cautions temper clinical translation. The central sleep apnea subgroup comprised only 303 participants, so confidence intervals around the headline event proportions deserve as much attention as the proportions themselves, and the findings require replication in larger and more diverse samples. The demonstration that a subject-specific baseline outperformed a fixed threshold of saturation at or above 96 percent confirms that the result depends on small relative differences in the oximetry signal, which places a premium on signal quality and measurement precision. The cohorts were 88 percent White, and known differential pulse-oximeter error by skin pigmentation is relevant to any saturation-based metric, although its effect on an above-baseline area calculation is currently unknown and warrants direct study.

The path forward is clear enough. The next study should measure event duration, airflow, respiratory effort, carbon dioxide, arousals, and event-level autonomic responses alongside oximetry, with detailed cardiac phenotyping and representative recruitment, and then test whether oxygen overshoot adds prognostic information once those features are accounted for. Until that work is done, oxygen overshoot burden should be regarded as a promising risk marker, a way of reading the recovery half of the apnea cycle that standard metrics ignore, rather than a demonstrated oxidative mechanism or a validated treatment target. Even so, the study is a reminder that the familiar oximetry trace still holds unexploited information, and that the line above the dips may matter as much as the dips themselves.

Subject of Research: Oxygen overshoot burden measured on overnight oximetry traces in central sleep apnea and its association with major adverse cardiovascular events.

Article Title: Above the line: oxygen overshoot in the oximetry trace

Article References: Manuel, A. R. G. (2026). Above the line: oxygen overshoot in the oximetry trace. Journal of Clinical Sleep Medicine, 22(1), Article 169. https://doi.org/10.1007/s44470-026-00183-8

Image Credits: AI Generated

DOI: 10.1007/s44470-026-00183-8

Keywords: central sleep apnea, oxygen overshoot, pulse oximetry, Cheyne-Stokes respiration, cardiovascular risk, hypoxic burden, ventilatory control, autonomic surges, oxidative stress, Sleep Heart Health Study, loop gain, sleep apnea