The apnea number surgeons use may be the wrong one for predicting complications

Obstructive sleep apnea is a highly prevalent and heterogeneous condition that predisposes to postoperative complications, and adequate metrics of its severity to stratify that risk in clinical settings are still needed — which is how the authors of a new study frame the problem they set out to address [s1].

The metric in standard use is the apnea-hypopnea index, or AHI: a count of how many times an hour breathing stops or shallows. Being a count, it carries no information about how long each event lasted or how far oxygen fell, which means two patients with identical indexes can accumulate very different amounts of hypoxia over a night.

A study published in JAMA Network Open on February 24 tested whether a measure that captures that difference does better at predicting what happens to these patients after major surgery [s1].

The measure

Sleep apnea-specific hypoxic burden (SASHB) is defined as the area under the desaturation curve associated with sleep-related obstructive respiratory events [s1]. In plain terms: not how many times oxygen dropped, but how much oxygen was lost, integrated over the night. It is expressed in percent-minutes per hour.

The study

The authors assembled a multicentre clinic-based cohort linked with a health administrative database, comprising adults diagnosed with obstructive sleep apnea between May 2007 and December 2018 who subsequently underwent major noncardiothoracic surgery between diagnosis and December 2024 [s1]. Analysis ran from January to December 2025.

The primary outcome was a 30-day composite of stroke, atrial fibrillation, heart failure, myocardial infarction, venous thromboembolism, and all-cause mortality [s1].

What was found

Among 2,286 patients with obstructive sleep apnea — median age 58 years (IQR 49–66), 1,472 (64.4%) men — who underwent major noncardiothoracic surgery a median of 4.5 years (IQR 1.9–7.5) after their apnea diagnosis, the composite outcome occurred in 80 patients (3.5%) within 30 days [s1].

The event rate rose across hypoxic burden strata [s1]:

• Low SASHB (<32% min/h): 1.6% — 12 events
• High SASHB (≥80% min/h): 5.8% — 44 events

In adjusted analysis, compared with low SASHB [s1]:

• SASHB 32 to <80% min/h: odds ratio 1.76 (95% CI, 0.86–3.59)
• SASHB ≥80% min/h: odds ratio 2.79 (95% CI, 1.42–5.49)

Only the highest stratum reached statistical significance; the middle stratum's confidence interval crosses 1. The gradient is monotonic, but the evidence supports a distinction between the severely hypoxic and everyone else rather than a fine-grained three-tier stratification.

A risk score combining age, emergency admission before surgery, and SASHB was associated with the outcome, with an area under the receiver operating characteristic curve of 0.73 (95% CI, 0.68–0.77) [s1]. That is moderate discrimination — better than chance, well short of a decisive test.

The finding with the most practical reach

The authors report that similar findings were obtained using a simplified version of SASHB derived automatically from the single oximetry signal extracted from diagnostic sleep studies [s1].

This is the part that determines whether any of it could be used. Full hypoxic burden calculation requires event-by-event annotation of a polysomnogram. A version computable from the oximetry channel alone can be derived retrospectively from studies already sitting in sleep-lab archives, and prospectively from far simpler recordings.

What this does not establish

The interval is long. Surgery occurred a median of 4.5 years after apnea diagnosis, with an interquartile range of 1.9 to 7.5 years [s1]. The hypoxic burden being used as a predictor was measured years before the operation. Weight, apnea severity, and treatment status can all change substantially over that span, and the abstract does not report whether patients were on CPAP at the time of surgery.

It is observational. SASHB may be a marker of cardiovascular vulnerability rather than a cause of postoperative events. Severe nocturnal hypoxia and a fragile cardiovascular system travel together.

The outcome count is small. Eighty events across 2,286 patients constrains how many covariates can be adjusted for and how precisely subgroups can be estimated.

It is a single cohort. The study was conducted by investigators at French respiratory and sleep research centres [s1], and it includes only patients who were referred, diagnosed, and had a diagnostic sleep study on record. People with undiagnosed apnea, who are the majority, are not represented.

The authors state the necessary caveat themselves: further research is needed to determine whether interventions guided by SASHB scores can modify postoperative risk [s1]. Showing that a number predicts an outcome is not the same as showing that acting on the number changes it.

What to watch

External validation in an independent cohort; whether the oximetry-derived simplified SASHB performs comparably when computed prospectively; and whether any trial tests a SASHB-guided perioperative protocol against usual care.

This article describes observational research and is not medical advice.

Sources

• [s1] Bailly S, Sabil A, Blanchard M, et al. Sleep Apnea-Specific Hypoxic Burden and Postoperative Outcomes of Major Noncardiothoracic Surgery. JAMA Network Open, published online 2026-02-24.