Understanding the Respiratory Report Without Worshipping the AHI

Opening Clinical Case

Two patients each have a respiratory event index or apnea-hypopnea index of 18 events per hour. The first reports mild snoring, occasional unrefreshing sleep, and minimal daytime impairment; her tracing shows limited desaturation and little else that looks dramatic. The second has prolonged REM-clustered events, deep desaturation, marked daytime impairment, and morning headaches. If you stop at the headline number, they look the same. If you read the report like a sleep physician, they are not the same at all.

That is the point of this chapter. The respiratory report is meant to be interpreted clinically, not worshipped like a sacred tablet. The AHI matters, but it is only one part of the story, and the job of the fellow is to figure out what kind of breathing problem is actually being described.pmc.ncbi.nlm.nih+1

What You Should Be Able to Do After This Chapter

By the end of this chapter, you should be able to:

• Distinguish AHI, REI, and RDI and know why they are not interchangeable.

• Recognize when event type changes the diagnosis and treatment.

• Interpret sleep-stage and positional patterns without overcalling certainty.

• Read oxygen data as more than a single nadir.

• Identify clues to RERAs, periodic breathing, and hypoventilation.

• Write a useful clinical impression rather than a machine-generated summary.

The Core Concept

The respiratory report is a summary of a physiology problem, not the physiology itself. The index tells you how often breathing went wrong. It does not automatically tell you why it went wrong, how badly it affected oxygenation, whether the patient was asleep the whole time, or whether the report hides important stage or positional clustering.pmc.ncbi.nlm.nih+1

Medical explanation: the reported indices are ratios. The numerator counts respiratory disturbances, but the denominator may be total sleep time, monitoring time, or sleep time plus additional event types depending on the measure. That is why AHI, REI, and RDI are related but not automatically interchangeable.pmc.ncbi.nlm.nih+1 Explain It Like I’m 10: the numerator counts breathing disturbances. The denominator asks, “Out of how much time?” Change the time measurement and the final number may change. Back to clinical medicine: once you understand the denominator and the type of event being counted, you can stop pretending that one index solves every sleep question.

A practical comparison helps anchor the definitions before we go deeper.

Note: the exact report format can vary by lab and device type, so always verify what was actually used to generate the reported number.

AHI, REI, and RDI

AHI is the number of apneas plus hypopneas per hour of sleep. REI is the number of respiratory events per hour of monitoring time, which is why it is commonly used in HSAT reports. RDI generally includes apneas, hypopneas, and respiratory effort-related arousals, though labs and device reports may vary in how they present the value. The key point is not merely that the numbers differ; it is that they are describing related but distinct things.pmc.ncbi.nlm.nih+2

The denominator is where people get in trouble. AHI requires true sleep time, so it is more precise when sleep staging is available. REI uses monitoring time, so it can dilute the event rate when the patient is awake for long periods. That is why a severely insomniac patient can have a deceptively low REI if the study included lots of wakefulness. RDI adds arousal-related respiratory disturbance, which can better capture sleep fragmentation in some patients but still does not solve every interpretive problem.pmc.ncbi.nlm.nih+1

Explain It Like I’m 10: the count may be the same, but if you divide by a bigger chunk of time, the final number gets smaller. Back to clinical medicine: when comparing reports, confirm what was counted and what time base was used before you compare severity across studies.

Obstructive, Central, and Mixed Events

The total AHI can hide mechanism. A patient with a total AHI of 18 could have mostly obstructive events, mostly central events, or a mixed picture with meaningful periodic breathing or treatment-emergent central instability. The report should tell you not only how often events occurred, but also what kind of events they were and whether respiratory effort was present.aasm+1

Obstructive events preserve effort. Central events do not. Mixed events have features of both. That distinction changes diagnosis and treatment. If you miss the event type, you may treat obstruction while the problem is actually central instability or hypoventilation. The total AHI is useful, but it is not the whole crime scene.

Explain It Like I’m 10: the AHI tells you how often breathing went wrong. It does not automatically tell you why it went wrong. Back to clinical medicine: event type matters because the treatment logic is different for obstruction, central apnea, and mixed physiology.

Event-type table

Sleep-Stage Patterns

Stage-specific patterns are often where the story becomes clinically interesting. REM-related worsening is common because muscle tone decreases and obstruction can become more severe in vulnerable airways. NREM disease may be more evenly distributed. If the report shows limited REM time, then the confidence about REM-predominant disease drops because the patient may simply not have spent enough time in REM to reveal the phenotype.pmc.ncbi.nlm.nih+1

The same logic applies in the opposite direction. If almost all the events occur in REM, the overall AHI can look modest while the REM burden is substantial. This is exactly why a clinician should read the stage distribution rather than stop at the summary line. If the stage data are sparse, say so in the impression. Quietly pretending the data are robust does not make them robust.

Explain It Like I’m 10: if you only watched one scene in the movie, you cannot claim you saw the whole plot. Back to clinical medicine: limited REM time reduces confidence in ruling out REM-predominant disease and can make the overall AHI look more reassuring than it should.

When limited REM matters

• The overall AHI may underestimate clinically relevant REM disease.

• Deep desaturations can cluster in REM and drive symptoms.

• A report with minimal REM time should be interpreted cautiously.

• If the clinical suspicion remains high, the absence of REM on one night is not a force field.

Positional Patterns

Supine and nonsupine event distribution can be highly informative. Some patients have clear positional phenotype, with much worse obstruction while supine and relatively mild disease off their back. Others have disease that is more generalized. Limited nonsupine sleep creates the same interpretive problem as limited REM: sparse data should not be overinterpreted as proof that disease is absent in that position.pmc.ncbi.nlm.nih+1

This matters because positional disease can affect management. If a patient has strong supine worsening, positional therapy may be a useful adjunct, especially when the disease is otherwise mild or moderate. But if the nonsupine sample is tiny, do not overstate how mild the off-back disease truly is. Sparse data are a poor foundation for overconfidence.

Explain It Like I’m 10: if the patient barely spent any time on their side, you do not really know what the side position would have shown. Back to clinical medicine: positional phenotype is real, but it must be interpreted with adequate sampling.

Oxygen Metrics

Oxygen data deserve more respect than the single nadir most reports highlight. Baseline saturation, oxygen nadir, desaturation frequency, and time below a threshold all help describe physiologic burden. A patient with brief dips to a low nadir is not the same as a patient who spends a long time hovering near that nadir. That difference is clinically important and too often hidden by report summaries.pmc.ncbi.nlm.nih+1

Intermittent desaturation suggests event-driven pathology such as OSA, whereas sustained desaturation should prompt consideration of hypoventilation, pulmonary disease, or another gas-exchange problem. Artifact also matters. A bad oximetry signal can create fake drama or falsely minimize the burden, and either outcome is bad medicine. Oxygen numbers are helpful, but they need context.

Explain It Like I’m 10: the nadir tells you the lowest point. It does not tell you whether the patient touched that number for two seconds or stayed near it for an hour. Back to clinical medicine: interpret saturation trends, not just the single lowest number.

A visual comparison makes that distinction much easier to remember.

Patterns suggesting additional disease

A baseline saturation that is low before events begin, sustained desaturation unrelated to discrete apneas, or a mismatch between the event count and the oxygen burden should make you think beyond routine OSA. That is where hypoventilation, lung disease, cardiac disease, and measurement artifact come back into the conversation. The respiratory report should not let the saturation curve quietly tell a different story while the summary page waves its AHI around like a tiny flag.

Arousals, RERAs, and Flow Limitation

Respiratory effort-related arousals and flow limitation matter because they explain why some patients have marked sleep fragmentation and symptoms despite a modest AHI. A patient may have repeated inspiratory flow flattening, increased effort, and frequent arousals without enough complete apnea or hypopnea events to drive the AHI very high. That does not make the problem trivial. It means the report may be capturing a different expression of sleep-disordered breathing.

RERAs can be especially relevant when the clinical story is “I sleep all night but I feel awful,” or when insomnia, nonrestorative sleep, and awakenings dominate. The diagnostic label should be used carefully because not every lab reports RERAs identically and not every report with a modest AHI is truly benign. If the patient is fragmented all night, the body does not care that the index looks polite.

Explain It Like I’m 10: the airway is not fully blocking, but breathing still gets harder and the brain keeps waking up to deal with it. Back to clinical medicine: symptoms can be substantial even when the AHI is not dramatic, especially when RERAs and flow limitation are prominent.

Periodic Breathing and Cheyne–Stokes Breathing

Periodic breathing is a waxing-and-waning pattern of ventilation often accompanied by central apneas or hypopneas. Cheyne–Stokes breathing is the classic form, characterized by crescendo-decrescendo ventilation with central pauses and often seen in heart failure, stroke, and other neurologic or cardiopulmonary settings. It is a pattern recognition diagnosis as much as a metric-based one.pubmed.ncbi.nlm.nih+2

Why does it matter? Because the total AHI may conceal the pattern. A report can list a total event frequency while the actual physiology is telling you about ventilatory control instability and central cycling. That pattern has implications for differential diagnosis and next-step evaluation. The attending is often less interested in the number than in whether you noticed the crescendo-decrescendo story hiding inside it.

Explain It Like I’m 10: the breathing keeps rising and falling like a wave, and the pauses happen at the low point of the wave. Back to clinical medicine: periodic breathing should trigger a search for clinical context such as heart failure, stroke, or other causes of ventilatory instability.

A waveform-style example helps the pattern stick.

Clues Suggesting Hypoventilation

Hypoventilation should be suspected when the report shows sustained desaturation, especially if it is paired with elevated carbon dioxide, morning headache, obesity, neuromuscular weakness, chest-wall disease, or medication effects. Oxygen saturation alone is insufficient because hypoventilation is a ventilation problem, not just an oxygenation problem. The patient may be hypoventilating even when the report does not look like repetitive apnea.pmc.ncbi.nlm.nih+1

Elevated serum bicarbonate is a valuable clue because it can reflect chronic CO₂ retention. If the patient has obesity hypoventilation syndrome, neuromuscular weakness, chest-wall restriction, or chronic opioid exposure, you should read the respiratory report with a more skeptical eye. A modest AHI does not rule out clinically important hypoventilation. The lungs and the drive system sometimes fail in ways that make the AHI look less dramatic than the physiology deserves.

Explain It Like I’m 10: oxygen saturation alone is not enough because the patient can be breathing too shallowly and still not show a simple apnea pattern. Back to clinical medicine: sustained desaturation plus the right clinical context should raise suspicion for hypoventilation even if the AHI is not extreme.

Writing the Clinical Impression

The clinical impression should tell the reader what kind of breathing disorder is present, how severe it appears, what patterns matter, and what the next step should be. Do not simply rewrite the machine output in slightly more elegant English. Say whether the predominant event type is obstructive, central, or mixed. Summarize severity, stage effect, positional effect, oxygenation, and any concern for hypoventilation or periodic breathing. Mention technical limitations. Then give the clinical bottom line and next step.

A good respiratory impression is a physician’s interpretation, not a data dump. It should include the mechanism, the phenotype, and the caveats. If a study has limited REM sleep, sparse nonsupine sleep, artifact-prone oximetry, or signals suggestive of hypoventilation, say so. If the report suggests periodic breathing, say that too. The attending should be able to read your impression and understand what to do without feeling like they need a second coffee and a decoder ring.

Practical impression template

• Predominant event type.

• Overall severity.

• REM and positional effects.

• Oxygen burden and pattern.

• Concern for hypoventilation or periodic breathing.

• Technical limitations.

• Clinical correlation and next step.

Ten questions to ask before signing the respiratory impression

1. What is the predominant event type?

1. Is the denominator sleep time or monitoring time?

1. Does the report show REM clustering?

1. Is there positional dependence?

1. Is oxygen burden clinically meaningful beyond the nadir?

1. Are there RERAs or flow limitation?

1. Is there periodic or Cheyne–Stokes breathing?

1. Are there clues to hypoventilation?

1. Are there technical limitations or artifact concerns?

1. What is the next clinical step?

Return to the Opening Case

The two patients with AHI 18 are not equivalent because the report is not just about the index. The patient with mild desaturation and minimal symptoms likely has a lower physiologic burden than the patient with prolonged REM events, severe desaturation, and marked daytime impairment. The decisive differences are event distribution, oxygen burden, symptom burden, and likely mechanism, not the headline number alone.

The correct impression is not “mild-moderate OSA, same as the other one.” It is something more like: “OSA with clinically meaningful REM-clustered disease and significant oxygen burden” or “OSA with limited physiologic impact,” depending on the tracing. The AHI got you into the room; the rest of the report tells you what kind of room it is.

What the Attending Will Ask

1. What is the difference between AHI and REI? AHI uses total sleep time, while REI uses monitoring time. They are not interchangeable.pmc.ncbi.nlm.nih+1

1. What does RDI add? RDI includes respiratory effort-related arousals in addition to apneas and hypopneas, depending on the report format.pmc.ncbi.nlm.nih+1

1. Why can the same AHI mean different clinical severity? Because AHI does not capture oxygen burden, event duration, REM clustering, positional burden, arousal burden, or symptoms.pmc.ncbi.nlm.nih+1

1. Why is limited REM time important? Because it lowers confidence in ruling out REM-predominant disease and may understate the burden.

1. Why does a low REI sometimes underestimate severity? Because REI uses monitoring time rather than true sleep time, which can dilute the index in insomnia or wakefulness-heavy studies.

1. What is the significance of periodic breathing? It suggests ventilatory control instability and should prompt attention to underlying heart failure, stroke, or other cardiopulmonary contexts.pubmed.ncbi.nlm.nih+1

1. How do you think about desaturation nadir? It is only one part of oxygen burden; duration and baseline saturation matter too.

1. When should you suspect hypoventilation? When desaturation is sustained, CO₂ is elevated, bicarbonate is high, or the clinical context suggests ventilatory impairment.pmc.ncbi.nlm.nih+1

Mistakes Smart Fellows Still Make

1. Treating AHI and REI as identical. This happens because both are “event rates.” It matters because the denominator is different. Avoid it by checking whether the study used sleep time or monitoring time.

1. Ignoring limited sleep time. This happens because the report summary looks complete. It matters because sparse REM or nonsupine sleep weakens confidence. Avoid it by reading the architecture of the study, not just the top line.

1. Ignoring REM or positional clustering. This happens because the overall AHI is easier to remember. It matters because phenotype can drive symptoms and treatment. Avoid it by looking at stage and position breakdowns.

1. Diagnosing hypoventilation from desaturation alone. This happens because low oxygen is dramatic. It matters because hypoventilation is about ventilation and CO₂. Avoid it by looking for hypercapnia and clinical clues.

1. Missing central or periodic breathing hidden inside the total AHI. This happens because total event frequency can obscure mechanism. It matters because treatment logic changes. Avoid it by checking event type and pattern, not just the number.

1. Writing the impression as a machine summary. This happens because the report template invites passivity. It matters because the referring clinician needs a clinical synthesis. Avoid it by answering the actual question posed by the study.

The Board Exam Is Trying to Trick You

1. The question gives two patients with the same AHI and asks who is more severe. The trap is to say they are equivalent. The correct answer is the patient with deeper desaturations, REM clustering, symptoms, or larger oxygen burden.

1. The stem says the REI is low but the patient has severe insomnia. The trap is to dismiss disease. The correct reasoning is that monitoring time may have diluted the index.

1. The report shows limited REM sleep. The trap is to call REM-predominant disease absent. The correct answer is that confidence is limited.

1. The tracing shows waxing and waning ventilation. The trap is to call it ordinary OSA. The correct answer is periodic breathing or Cheyne–Stokes pattern, depending on the context.pubmed.ncbi.nlm.nih+1

1. The report has sustained desaturation and high bicarbonate. The trap is to call it simple OSA. The correct answer is to think hypoventilation.

How to Explain This to a Patient

“Your sleep study gives us more than one number. The event count is helpful, but we also look at how long the events lasted, how low the oxygen dropped, whether the events happened more in REM sleep or when you were on your back, and whether the pattern suggests another breathing problem. Two patients can have the same index and still have different amounts of real-world impact. That is why we interpret the whole report, not just one number.”

Practical Pearls

• AHI, REI, and RDI are related but not interchangeable.

• The denominator matters as much as the numerator.

• Event type changes diagnosis and treatment.

• REM and positional clustering can hide behind a modest overall AHI.

• Oxygen nadir is not the same as oxygen burden.

• Sustained desaturation is a red flag for more than simple repetitive obstruction.

• RERAs and flow limitation can explain symptoms when AHI looks modest.

• Periodic breathing should prompt attention to central control and clinical context.

• Hypoventilation requires thinking about CO₂, not just saturation.

• A good respiratory impression is a synthesis, not a transcript.

The Bottom Line

• The respiratory report is a clinical tool, not a scoreboard.

• AHI, REI, and RDI describe related but different things.pmc.ncbi.nlm.nih+2

• Sleep time and monitoring time are not the same denominator.

• Event type matters because obstruction, central events, and mixed physiology are not treated the same way.

• REM and positional clustering can make disease look milder than it is.

• Oxygen nadir is useful but incomplete.

• Baseline saturation, time below threshold, and desaturation pattern matter.

• RERAs and flow limitation can explain symptoms despite a modest AHI.

• Periodic breathing and Cheyne–Stokes patterns should not be missed.pubmed.ncbi.nlm.nih+1

• Sustained desaturation plus the right clinical context should trigger concern for hypoventilation.

• The clinical impression should say what the report means, not just what it measured.

Question 1

A PSG report lists an AHI of 14 events per hour and an RDI of 29 events per hour. What is the most likely reason the RDI is higher?

A. RDI excludes hypopneas B. RDI includes respiratory effort-related arousals C. RDI is always calculated using monitoring time D. RDI only counts central apneas E. RDI cannot be reported on PSG

Question 2

A patient with a low REI on HSAT had severe insomnia and was awake for much of the recording. Which interpretation is most accurate?

A. The low REI reliably excludes OSA B. The REI may underestimate severity because monitoring time was used C. HSAT always overestimates disease severity D. The patient must have central sleep apnea E. The REI is identical to the AHI

Question 3

A report shows a total AHI of 18 with events clustered almost entirely in REM sleep, but there were only 25 minutes of REM. What is the best interpretation?

A. REM-predominant disease is confidently ruled out B. REM findings are more meaningful because REM time was long C. Confidence in stage-specific conclusions is limited because REM time was sparse D. The overall AHI completely overrides stage distribution E. This pattern indicates hypoventilation by itself

Question 4

Which report feature most strongly suggests periodic breathing?

A. Repetitive snoring only B. Waxing-and-waning ventilation with central pauses C. Isolated desaturations after limb movements D. Flow limitation without arousal E. Frequent awakenings without respiratory changes

Question 5

Which combination most strongly suggests possible hypoventilation rather than simple repetitive obstructive apneas?

A. Brief intermittent desaturations and normal bicarbonate B. Sustained desaturation, elevated bicarbonate, and obesity or opioid exposure C. Loud snoring and normal oxygen saturation D. Frequent arousals with no oxygen change E. Supine worsening only

Question 1

Correct answer: B. RDI includes respiratory effort-related arousals.

Why it is correct: RDI commonly adds RERAs to apneas and hypopneas, which can raise the index.pmc.ncbi.nlm.nih+1 Why the others are wrong: A is false; C describes a denominator issue but not the reason for the difference; D is false; E is false. Learning point: RDI can capture arousal-based respiratory disturbance that AHI misses.

Question 2

Correct answer: B. The REI may underestimate severity because monitoring time was used.

Why it is correct: HSAT commonly uses monitoring time rather than true sleep time, which can dilute the index in insomnia or wakefulness-heavy studies.pmc.ncbi.nlm.nih+1 Why the others are wrong: A overstates the value of a single index; C is opposite of the common limitation; D is not implied; E is false. Learning point: denominator problems matter in insomnia.

Question 3

Correct answer: C. Confidence in stage-specific conclusions is limited because REM time was sparse.

Why it is correct: limited REM exposure makes REM-predominant inference less secure. Why the others are wrong: A and D are overconfident; B is the opposite of sparse REM; E is not supported by the information alone. Learning point: stage-specific patterns require adequate stage sampling.

Question 4

Correct answer: B. Waxing-and-waning ventilation with central pauses.

Why it is correct: that is the hallmark of periodic or Cheyne–Stokes breathing.pubmed.ncbi.nlm.nih+1 Why the others are wrong: A is snoring, not periodic breathing; C suggests other causes; D and E are nonspecific. Learning point: pattern recognition is essential.

Question 5

Correct answer: B. Sustained desaturation, elevated bicarbonate, and obesity or opioid exposure.

Why it is correct: this combination suggests impaired ventilation and possible chronic CO₂ retention.pmc.ncbi.nlm.nih+1 Why the others are wrong: A, C, D, and E do not specifically suggest hypoventilation. Learning point: hypoventilation is a CO₂ question, not just an oxygen question.