Sleep Apnea Diagnosis: Why 'Mild' Can Be Misleading
A sleep physician argues the term 'mild OSA' is often misinterpreted, leading to under-treatment. The apnea-hypopnea index (AHI) alone fails to capture key factors like symptom severity and health risks.
The label 'mild obstructive sleep apnea' can be dangerously misleading, according to sleep physician David E. McCarty, MD. He argues that relying solely on the apnea-hypopnea index (AHI) to gauge severity often overlooks a patient's actual symptoms and health risks.
McCarty, writing for Sleep Review, states the term 'mild' is typically attached to an AHI between 5 and 14.9 events per hour. He observes that in clinical practice, this descriptor often migrates from describing a test result to characterizing the entire problem. "'An AHI in the mild range' becomes 'mild sleep apnea,' which naturally becomes 'a mild problem,' which morphs into 'nothing important,'" he notes. The danger, he says, is that the term can be "cognitively sedating," making uncertainty feel settled before important clinical questions are asked.
The Limits of the AHI Metric
Major medical bodies acknowledge the AHI's limitations. McCarty cites the American Heart Association's position that the standard AHI thresholds are empirical conventions. Event frequency alone does not capture hypoxic burden, event duration, sleep fragmentation, or foundational symptoms like daytime sleepiness. Published literature adds to the confusion, with research statements finding inconsistent definitions and evidence for the consequences of so-called mild OSA.
These criticisms reveal a category whose apparent precision exceeds its clinical usefulness. McCarty illustrates this with a hypothetical example. Two people can have an identical AHI of 9 events per hour. One may feel well with few issues, while the other could be falling asleep at work, waking with headaches, or have substantial cardiometabolic vulnerability. Their stories, and their need for treatment, are completely different despite the same arithmetic.
Handle Treatment Decisions
The critique is not that 'mild' always means 'dangerous,' which would simply replace one label-based approach with another. McCarty points out that sometimes physiology in the mild AHI range does correspond to a minimal problem. Evidence supporting treatments like CPAP for less symptomatic individuals is limited. Large randomized CPAP trials have not demonstrated hoped-for reductions in cardiovascular events in selected populations, and a US task force found insufficient evidence for screening asymptomatic adults.
These findings should discipline any reflex to treat the AHI as an automatic mandate to intervene. However, neutral trial results do not mean nobody with a low AHI has a treatable problem. Newer analyses suggest treatment benefit may be concentrated in phenotypes with greater physiological risk. The lesson is that the AHI tier is a poor stand-in for a patient's actual 'reasons to treat.'
A Framework for Patient-Centered Care
In the clinical framework at Rebis Health, where McCarty serves as chief medical officer, the 'reasons to treat' are organized into five domains: risk, snoring, sleep, wake, and comorbidities. The goal is to force the question of 'why' treat back into the open. What makes this physiology consequential for this person? Sometimes the honest answer is 'nothing, yet.' Patient-centered care, he argues, is not a vehicle for more treatment but for better reasons for treatment.
The next question is how to measure success. If treatment was started because a patient couldn't stay awake in afternoon meetings, success cannot be defined solely by a lower AHI. If it was chosen because a couple hasn't slept together peacefully in years, the bed partner's experience is a core part of the outcome.
Moving Beyond a Single Number
This is not an argument to discard the AHI but to restore it to its proper role. Hypoxic burden, event duration, symptoms, blood pressure, and patient priorities are all pieces of a story too complex for any single number to narrate alone. Emerging research on endophenotypes is moving toward the same conclusion, motivated by the fact that approximately half of patients do not achieve satisfactory outcomes under conventional AHI-centered treatment pathways.
Patient-centered and lifestyle-metabolic models suggest chronic OSA care should incorporate patient values and modifiable drivers like posture, adiposity, and diet. McCarty is candid about the evidence boundary, stating no randomized trial has shown that replacing severity labels with a narrative framework improves hard outcomes like cardiovascular events. Endophenotyping and patient-centered care remain developing research programs.
The strongest claim is epistemic and ethical: the categorical vocabulary of AHI grades does not map reliably onto biology or lived experience. A framework that keeps diagnosis, alternatives, reasons for treatment, and outcomes in active tension is more faithful to reality. The danger begins when clinical shorthand becomes thought-when 'MILD' stops describing where a number landed and starts dictating how much the patient's problem matters.
