How Sleep Apnea is Diagnosed
Polysomnography (in-laboratory sleep study) is the gold standard for diagnosing obstructive sleep apnea, though home sleep apnea testing with a technically adequate device is an acceptable alternative for uncomplicated patients with high pretest probability of moderate-to-severe OSA. 1
Diagnostic Pathway
Step 1: Clinical Evaluation and Risk Assessment
Before any testing, diagnosis requires a comprehensive sleep evaluation that includes:
Symptoms to assess: habitual snoring, witnessed apneas, gasping/choking at night, excessive daytime sleepiness, morning headaches, dry mouth upon waking, non-restorative sleep, nocturia, impaired concentration, memory problems 2
Physical examination findings: obesity (particularly BMI >30), increased neck circumference, retrognathia/micrognathia, macroglossia, enlarged uvula/soft palate, tonsillar hypertrophy, modified Mallampati score, Friedman tongue position 2
High-risk comorbidities: hypertension, cardiovascular disease, type 2 diabetes, stroke, atrial fibrillation, heart failure, metabolic syndrome 1, 2
Critical caveat: Questionnaires alone (Berlin, STOP-Bang, Epworth Sleepiness Scale) cannot diagnose OSA and must be combined with objective sleep testing. They are screening tools only. 1
Step 2: Selecting the Appropriate Diagnostic Test
Polysomnography (PSG) is Required For:
- Patients with significant cardiorespiratory disease
- Suspected respiratory muscle weakness from neuromuscular conditions
- Awake hypoventilation or suspected sleep-related hypoventilation
- Chronic opioid medication use
- History of stroke
- Severe insomnia
- When home sleep testing is negative, inconclusive, or technically inadequate
- Suspected coexisting sleep disorders (central sleep apnea, periodic limb movements, parasomnias) 1
Home Sleep Apnea Testing (HSAT) is Acceptable For:
- Uncomplicated adult patients with signs/symptoms indicating increased risk of moderate-to-severe OSA
- No significant comorbidities as listed above
- High pretest probability based on clinical evaluation 1
Important limitation: HSAT should not be used to rule out OSA. If a single HSAT is negative, inconclusive, or technically inadequate, proceed to in-laboratory PSG. 1
Step 3: Diagnostic Criteria
OSA is diagnosed when the apnea-hypopnea index (AHI) meets threshold criteria:
- AHI ≥5 events/hour with typical symptoms (snoring, witnessed apneas, daytime sleepiness)
- AHI ≥15 events/hour regardless of symptoms 2
Severity classification based on AHI:
- Mild: 5-14 events/hour
- Moderate: 15-29 events/hour
- Severe: ≥30 events/hour 2
Hypopneas are typically defined as ≥30% reduction in airflow with either ≥3-4% oxygen desaturation or an arousal from sleep. 3
Step 4: When Initial Testing is Negative
If PSG is negative but clinical suspicion remains high, consider a second polysomnogram, as night-to-night variability in AHI can be substantial due to sleep position, alcohol use, and sleep-stage distribution. 1, 4
Common Pitfalls to Avoid
Do not diagnose OSA based on questionnaires or clinical prediction tools alone without objective sleep testing 1
Do not use HSAT in complicated patients with the comorbidities listed above—these require full PSG 1
Do not accept a negative HSAT as ruling out OSA—proceed to PSG if clinical suspicion persists 1
Be aware of pulse oximetry racial bias in newer wearable devices, which may underestimate hypoxemia in patients with darker skin pigmentation 5
Consumer wearable devices (smartwatches, rings) may raise awareness but should not replace medical-grade diagnostic testing 5, 2
Split-Night Protocol Option
A split-night study (diagnostic portion followed by CPAP titration in the same night) may be considered if clinically appropriate, though this is a weaker recommendation based on resource considerations rather than superior diagnostic accuracy. 1
Emerging Considerations
Multi-night home testing may improve diagnostic accuracy by accounting for night-to-night AHI variability, particularly around diagnostic thresholds, though single-night testing remains standard practice. 4 Advanced metrics beyond AHI (hypoxic burden, T90, event duration) may better capture cardiovascular risk but are not yet incorporated into standard diagnostic criteria. 6