Central Sleep Apnea

Central sleep apnea (CSA) is recurrent cessation or marked reduction of airflow during sleep caused by loss or instability of the brainstem's own respiratory drive, rather than by a mechanically obstructed airway. On polysomnography this is defined, in broad terms, by a central apnea-hypopnea index of at least 5 events per hour in which more than half of the scored respiratory events are central.^[1] CSA is far less common overall than obstructive sleep apnea, but it clusters heavily in 2 clinical populations that a pharmacist encounters routinely: patients on chronic opioid therapy, where central and ataxic breathing patterns are a genuine, dose-related complication of long-term use rather than a rare idiosyncratic reaction, and patients with heart failure with reduced ejection fraction (HFrEF), where CSA takes the specific crescendo-decrescendo form known as Cheyne-Stokes respiration. High-altitude periodic breathing, recent stroke or other structural brainstem/CNS disease, and rare idiopathic CSA make up most of the remainder.

The single most important safety concept in this chapter is that CSA does not behave like OSA, either diagnostically or therapeutically. Continuous positive airway pressure (CPAP), the mainstay of OSA treatment, works by mechanically splinting an obstructed airway open — a mechanism that simply does not apply when the airway was never obstructed in the first place, so CPAP is a reasonable first trial but a meaningfully less consistent fix in CSA than in OSA. And adaptive servo-ventilation (ASV), a device purpose-built to smooth out the Cheyne-Stokes breathing pattern and highly effective at doing exactly that, was shown in a large randomized trial to increase both all-cause and cardiovascular mortality when used in patients with HFrEF (left ventricular ejection fraction 45% or less) and predominant CSA — despite effectively normalizing the breathing pattern it was measured against.^[2] Effectively treating the sleep study is not the goal; keeping the patient alive is, and in this one specific population those 2 things pulled in opposite directions. This is now a hard contraindication, not a nuance.^[3]

Central versus obstructive — 1 mechanism, 2 completely different diseases

Both CSA and OSA produce the same end result on a pulse oximeter — a period of reduced or absent airflow followed by a desaturation and often an arousal — which is exactly why the distinction cannot be made from symptoms or oximetry alone. The mechanisms are opposite. In obstructive sleep apnea, the pharyngeal airway collapses or narrows despite the brain continuing to signal the chest and diaphragm to breathe — respiratory effort persists, and often visibly increases, against a blocked airway. In central sleep apnea, the brainstem's respiratory control centres simply stop, or destabilize, the signal to breathe — both airflow and respiratory effort go flat together, because nothing downstream is even being asked to move.^[1] Polysomnography distinguishes the two using dedicated airflow and thoracoabdominal effort belts recorded side by side; a home sleep apnea test, which typically lacks a true effort channel, cannot reliably make this distinction and is not an adequate test to diagnose or exclude CSA — see Assessment & Diagnosis.

Central apnea on polysomnography — airflow and respiratory effort both go flat together, because the brainstem has stopped signalling the chest to move at all. (NascarEd, Wikimedia Commons, CC BY-SA 3.0)

Obstructive apnea on polysomnography, for direct comparison — airflow goes flat while respiratory effort continues, and often intensifies, against the obstructed airway. (NascarEd, Wikimedia Commons, CC BY-SA 3.0)

Opioid-induced central and ataxic breathing

Chronic opioid therapy produces sleep-disordered breathing through several overlapping mechanisms: direct mu-opioid-receptor-mediated suppression of the brainstem's rhythm-generating circuitry (the pre-Bötzinger complex), blunted chemoreceptor sensitivity to rising CO₂, and destabilized upper-airway dilator muscle activity.^[4] The resulting pattern is often genuinely irregular rather than the smooth waxing-and-waning of Cheyne-Stokes respiration — clusters of breaths of varying depth separated by pauses, sometimes described as an ataxic or Biot's pattern — and can include central apneas, central hypopneas, and, less predictably than in Cheyne-Stokes respiration, prolonged periods of shallow, irregular breathing that do not meet a formal apnea definition but still fragment sleep and reduce oxygenation. This is a dose-related phenomenon: the risk and severity of opioid-associated CSA rise with the daily morphine-equivalent dose, and some degree of sleep-disordered breathing with a central component is found in a substantial minority — commonly cited at roughly 1 in 4 — of patients on chronic opioid therapy who undergo polysomnography.^[4] Patients themselves are frequently unaware anything is wrong: unlike OSA, opioid-induced CSA is not reliably accompanied by loud snoring, so the bed partner's report of "no snoring" should never be used to reassure a patient on chronic high-dose opioids who reports unrefreshing sleep, daytime sedation beyond what the opioid itself would explain, or witnessed breathing pauses.

Heart-failure-associated Cheyne-Stokes respiration

In HFrEF, a prolonged circulation time between the lungs and the brainstem chemoreceptors, combined with an abnormally sensitive ventilatory response to small changes in CO₂ (a high "loop gain"), destabilizes the normal breathing rhythm into a smooth, repeating crescendo-decrescendo pattern: tidal volume gradually builds to a peak, tapers back down, and is followed by a central apnea or hypopnea before the cycle begins again — Cheyne-Stokes respiration. It is common in more severe HFrEF, tends to track with reduced ejection fraction and NT-proBNP/BNP elevation, and both causes and is worsened by poor sleep quality, elevated sympathetic tone, and nocturnal hypoxemia — a cycle that plausibly contributes to the excess cardiovascular risk seen in heart failure patients with untreated Cheyne-Stokes respiration, even though, as detailed in Choose Initial Treatment, correcting the breathing pattern itself has not reliably translated into a mortality benefit.^[5]

Other causes

• High-altitude periodic breathing — a Cheyne-Stokes-like pattern triggered by hypoxia-driven hyperventilation and the resulting hypocapnia at altitudes above roughly 2,500 m; it resolves with descent or acclimatization and does not, on its own, indicate an underlying cardiac or opioid cause.
• Stroke and other structural CNS disease — CSA is common in the acute period after a stroke, particularly with brainstem involvement, and often improves over the following weeks to months as the acute injury resolves; a tumour, demyelinating disease, or other structural brainstem lesion can also present with CSA.
• Idiopathic (primary) CSA — a diagnosis of exclusion once opioid use, heart failure, altitude, and a structural CNS cause have been ruled out; genuinely uncommon.
• Treatment-emergent (complex) CSA — central apneas that appear or persist when a patient with predominantly obstructive sleep apnea is started on CPAP; usually transient and often resolves with continued CPAP use over several weeks, but worth recognizing as a distinct pattern rather than assuming CPAP titration has simply failed (see Common Errors to Avoid).

• Chronic opioid therapy, particularly at a higher daily morphine-equivalent dose — the single most important modifiable risk factor a pharmacist will directly influence.^[4]
• Heart failure with reduced ejection fraction — risk and severity of Cheyne-Stokes respiration rise as ejection fraction falls and heart failure severity worsens.^[5]
• Atrial fibrillation, which frequently coexists with both HFrEF and Cheyne-Stokes respiration and can independently worsen the underlying loop-gain instability.
• Male sex — central sleep apnea, like obstructive sleep apnea, is substantially more common in men.
• Older age.
• Recent stroke, particularly with brainstem involvement, or another acute or structural CNS lesion.
• Ascent to high altitude (generally above roughly 2,500 m) without adequate acclimatization.
• Concurrent use of other CNS/respiratory depressants (benzodiazepines, gabapentinoids, sedating antihistamines, alcohol) alongside chronic opioid therapy, which can compound opioid-related ventilatory instability even though these agents do not independently cause a central pattern to the same degree.
• Atrial septal defect or another right-to-left cardiac shunt, and end-stage renal disease — both recognized, less common associations.

History

Ask specifically about witnessed breathing pauses, gasping or choking arousals, unrefreshing sleep, morning headache, and daytime sleepiness beyond what an underlying opioid dose or heart-failure fatigue would explain on its own. Snoring is frequently absent or minimal in pure CSA, unlike OSA, so its absence should never be used to rule the condition out. Always take a complete, current medication history including the total daily opioid dose in morphine-equivalents, and a focused cardiac history (known HFrEF, most recent ejection fraction, NYHA class, recent decompensation).

Polysomnography — the only way to actually make this diagnosis

Full, attended, in-laboratory polysomnography (PSG) with dedicated airflow, thoracoabdominal respiratory-effort, and oximetry channels is required both to diagnose CSA and to distinguish it from OSA — see the side-by-side tracings in Recognize the Condition. A central apnea-hypopnea index of at least 5 events per hour, with more than half of the total scored respiratory events being central, is the general threshold used to define clinically significant CSA.^[1] Home sleep apnea testing (HSAT) most consumer and clinical HSAT devices do not include a true respiratory-effort channel and cannot reliably distinguish a central event from an obstructive one — an HSAT reported as showing a high apnea-hypopnea index in a patient with a high pretest suspicion for CSA (chronic high-dose opioids, known HFrEF) should prompt referral for full PSG rather than being treated as diagnostic on its own.^[1]

What the sleep study should characterize, specifically

• The central apnea-hypopnea index, and the proportion of total events that are central versus obstructive versus mixed.
• The pattern of the central events — a smooth crescendo-decrescendo pattern supports Cheyne-Stokes respiration (look specifically for coexisting HFrEF); an irregular, ataxic pattern with variable-depth clusters supports an opioid mechanism.
• Overnight oxygen saturation nadir and total time spent below 90% saturation.
• Presence of a coexisting obstructive component — mixed OSA/CSA is common, particularly in opioid-associated disease, and changes the treatment approach (see Choose Initial Treatment).

Other investigations directed at the underlying cause

• Echocardiography (or review of the most recent one) to confirm and characterize ejection fraction in any patient with newly diagnosed CSA and no clearly identified alternative cause, and in every patient with known heart failure being worked up for Cheyne-Stokes respiration.
• BNP/NT-proBNP as a supportive marker of heart failure severity, not as a stand-alone diagnostic test for Cheyne-Stokes respiration.
• A current, itemized opioid medication list with total daily morphine-equivalent dose for any patient on chronic opioid therapy being evaluated for CSA.
• Neuroimaging when a stroke or another structural CNS cause is suspected clinically, particularly with a new focal deficit or a recent stroke.
• Altitude and travel history when the presentation is temporally linked to recent ascent.

• Confirm the diagnosis with real polysomnography before committing a patient to long-term device therapy — do not treat CSA based on a home sleep apnea test alone.
• Identify which of the major etiologies (opioid, heart failure, altitude, stroke/CNS, idiopathic) is actually driving the pattern, since management differs meaningfully between them.
• In opioid-associated CSA, achieve the lowest effective opioid dose for the patient's pain-management goals, in genuine collaboration with the prescriber — this is the single highest-yield intervention available, not an afterthought to device therapy.
• In heart-failure-associated Cheyne-Stokes respiration, optimize guideline-directed medical therapy for the underlying heart failure itself before or alongside any device trial.
• Reduce the central apnea-hypopnea index and improve oxygenation using CPAP, supplemental oxygen, or, where appropriate, adaptive servo-ventilation — while respecting the hard safety boundary around ASV in HFrEF.
• Never prescribe or continue adaptive servo-ventilation in a patient with LVEF 45% or less and moderate-to-severe CSA-predominant sleep-disordered breathing.
• Set realistic expectations: unlike OSA, full normalization of the breathing pattern is not always achievable, and treating the sleep study is not itself proven to extend life in every population covered by this chapter.
• Recognize and escalate red flags — worsening heart failure, severe hypoxemia, or signs of opioid toxicity — promptly rather than managing them as routine sleep-clinic findings.

Heart-failure-associated Cheyne-Stokes respiration — treatment ladder

Opioid-associated central sleep apnea — treatment approach

Other causes

High-altitude periodic breathing generally resolves with descent or with time for acclimatization and does not usually require device therapy unless the exposure will be prolonged. Stroke-associated CSA is frequently transient and is reassessed with repeat testing over the following weeks to months rather than treated as a fixed, lifelong diagnosis at the time of the acute event. Idiopathic CSA follows the same acetazolamide/theophylline/PAP framework as heart-failure-associated CSA once heart failure itself has been excluded.

• Optimize the underlying heart failure — guideline-directed medical therapy is a genuine, foundational treatment for Cheyne-Stokes respiration, not merely a parallel task; see the Heart Failure chapter and Choose Initial Treatment above.
• CPAP — first-line device trial in both major etiologies; effective at reducing the central apnea-hypopnea index and improving oxygenation in Cheyne-Stokes respiration, though meaningfully less consistent than in OSA, and even less predictably effective in opioid-associated CSA.^[5][4]
• Nocturnal supplemental oxygen — reduces the severity of Cheyne-Stokes respiration by lowering the exaggerated chemoreceptor sensitivity that drives the crescendo-decrescendo pattern; a simpler, often better-tolerated option than PAP therapy for some patients, and does not carry the same HFrEF-specific safety concern as ASV.
• Adaptive servo-ventilation (ASV) — highly effective at normalizing the Cheyne-Stokes pattern, but contraindicated in patients with LVEF 45% or less and moderate-to-severe CSA-predominant sleep-disordered breathing; reasonable in patients with preserved or mildly reduced ejection fraction, or non-cardiac CSA, when CPAP/oxygen are inadequate.^[2][3]
• Bilevel PAP with a backup rate (BPAP-ST) — an option under sleep-medicine supervision, particularly for opioid-associated CSA not adequately controlled with simple CPAP.
• Phrenic nerve stimulation — an implantable device (placed and managed by a specialist centre, not initiated in routine sleep clinic) that stimulates the phrenic nerve to restore a more regular breathing pattern; a genuine option for selected patients with moderate-to-severe CSA who have not responded to or cannot tolerate PAP therapy, though patient selection and long-term outcome data continue to evolve — a pharmacist's role here is largely recognizing the device exists and confirming it is not confused with a PAP device on medication/device reconciliation.
• Repeat polysomnography after any significant opioid dose change, heart-failure therapy optimization, or device initiation — confirming the breathing pattern has actually changed, not just assuming it from symptoms alone, is itself a core part of non-pharmacologic management.
• Altitude descent or acclimatization for high-altitude periodic breathing, and time plus repeat testing for stroke-associated CSA in the weeks to months after the acute event.

Pregnancy

Central sleep apnea specifically related to opioid therapy or HFrEF is uncommon in pregnancy, but any pregnant patient on chronic opioid therapy or with peripartum cardiomyopathy/heart failure who reports witnessed breathing pauses or unrefreshing sleep should still be worked up rather than dismissed as routine pregnancy-related fatigue. Acetazolamide should generally be avoided in pregnancy given animal data showing fetal harm at supratherapeutic doses, and used only if a specialist judges the benefit clearly outweighs the risk; theophylline requires individualized specialist input. Any opioid-dose discussion in pregnancy needs to weigh withdrawal risk to the fetus/neonate against the benefits of dose reduction — this is a genuinely specialist-level discussion, not one to manage from a general framework, and should involve obstetric/maternal-fetal medicine, addiction medicine, and pain management together.

Breastfeeding

Acetazolamide's passage into breast milk is not well characterized, so caution and individualized specialist input is reasonable. Opioid selection and dosing in a breastfeeding patient should follow the specific guidance in each opioid's own Drug Database entry and the Opioid-Related Disorders chapter rather than being decided from this chapter alone.

Renal impairment

Acetazolamide is contraindicated in significant renal impairment given reduced clearance and the risk of worsening metabolic derangement; avoid or use with specialist input. Opioid clearance is reduced in renal impairment for several agents (notably morphine, via active metabolite accumulation) — see the Chronic Kidney Disease chapter and each opioid's own monograph for dose-adjustment guidance, which becomes doubly important when the same opioid is also contributing to CSA.

Hepatic impairment

Theophylline clearance is reduced in significant hepatic impairment, raising toxicity risk at a given dose — use lower doses with closer monitoring. Several opioids, including methadone, have altered and less predictable clearance in hepatic impairment; involve a pharmacist or specialist experienced with opioid dosing in this setting explicitly rather than adjusting empirically.

Older adults

Both major etiologies in this chapter are more common with advancing age, and older adults are more susceptible to the sedative, cognitive, and fall-risk effects of any CNS-depressant medication (including opioids) that may be layered onto CSA-related sleep fragmentation. Apply an even more deliberate opioid-minimization approach where clinically feasible, and consider a lower threshold for cardiology/sleep-medicine referral given the more blunted, easier-to-miss symptom reporting common in this group.

Coexisting obstructive sleep apnea

Mixed OSA/CSA is common, particularly with opioid-associated disease, and treatment-emergent (complex) CSA can appear when a predominantly obstructive patient is started on CPAP. Do not assume a CPAP titration has simply failed the first time central events appear — many cases of treatment-emergent CSA resolve with continued CPAP use over several weeks; escalate to BPAP-ST or ASV (respecting the LVEF-based safety criterion) only if central events persist on repeat assessment.

Confirm a repeat, objective reassessment — repeat polysomnography where feasible, or at minimum a genuine respiratory-event download from a compliant PAP or ASV device, not just usage hours — after any significant change: an opioid dose reduction or rotation, initiation or up-titration of heart-failure guideline-directed medical therapy, or new device therapy. Symptom improvement alone is not a reliable substitute for confirming the breathing pattern has actually changed.

Opioid-associated CSA

Track the total daily morphine-equivalent dose over time as its own monitoring parameter, alongside standard pain and function measures — a rising dose in a patient with known or suspected CSA is itself a signal to revisit the treatment plan, not only a pain-management data point. Screen for daytime hypoventilation (morning headache, confusion, daytime hypercapnia if measured) at follow-up, and reinforce naloxone access and overdose-recognition counselling for the patient and household as part of standard chronic-opioid-therapy safety practice, independent of the CSA diagnosis itself — see the Opioid-Related Disorders chapter.

Heart-failure-associated Cheyne-Stokes respiration

Monitor heart failure status (weight, volume status, NYHA class, and guideline-directed medical therapy titration) at every follow-up alongside the sleep-apnea-specific plan — the two are managed together, not in sequence. Confirm any patient on ASV genuinely has LVEF above 45%, and re-confirm this at any point ejection fraction is reassessed, since a falling ejection fraction over time can convert a previously appropriate ASV prescription into a contraindicated one.^[2][3]

Pharmacologic adjuncts

Monitor electrolytes and renal function with acetazolamide, and obtain theophylline levels with dose changes or any new interacting medication, given its narrow therapeutic index — review for arrhythmia symptoms specifically in the HFrEF population.

Transitions of care

Confirm the specific PAP/ASV device type — not just "on a breathing machine at night" — is captured accurately at every medication and device reconciliation, particularly at a hospital admission or a new prescriber; an ASV device continued by default in a patient whose ejection fraction has since fallen to 45% or less is a genuine, preventable safety failure.

1. Opioid-associated CSA with no improvement despite a genuine, adequately sized dose reduction or rotation: escalate to sleep-medicine-supervised device therapy (CPAP, then BPAP-ST or ASV per the LVEF-based safety criterion) rather than repeatedly adjusting the opioid regimen alone.
2. Heart-failure-associated Cheyne-Stokes respiration not improving despite optimized guideline-directed medical therapy and a CPAP or oxygen trial: reassess LVEF before considering ASV — proceed only if LVEF is above 45%; if LVEF is 45% or less, ASV remains contraindicated regardless of how poorly controlled the sleep-disordered breathing is on other measures.^[2][3]
3. Persistent central events on CPAP in a patient newly started for suspected OSA: reassess for treatment-emergent (complex) CSA — continue CPAP and reassess in several weeks before assuming escalation is required, unless the patient has a clear alternative driver (opioids, HFrEF) that changes the picture.
4. Worsening symptoms or a rising central apnea-hypopnea index despite device therapy that was previously effective: reassess for a new or worsening driver — declining ejection fraction, an increased opioid dose, new atrial fibrillation, or a new CNS event — rather than simply intensifying the existing device settings.
5. A patient who cannot tolerate PAP therapy in any form: revisit acetazolamide or theophylline as an adjunct (see Pharmacologic Treatment Options), reinforce the opioid-reduction or heart-failure-optimization plan as applicable, and consider phrenic nerve stimulation referral for selected, appropriately worked-up patients.
6. Ongoing diagnostic uncertainty about the underlying cause despite the standard workup: refer to sleep medicine for a specialist-level reassessment rather than continuing empiric therapy indefinitely without a confirmed driver.

• Diagnosing or excluding central sleep apnea from a home sleep apnea test alone — most HSAT devices lack a true effort channel and cannot reliably distinguish central from obstructive events.^[1]
• Prescribing or continuing adaptive servo-ventilation in a patient with LVEF 45% or less and moderate-to-severe CSA-predominant sleep-disordered breathing — this is a hard contraindication based on a documented mortality signal, not a relative preference.^[2][3]
• Treating a normalized apnea-hypopnea index on a device download as proof the patient's outcome has improved — CANPAP showed CPAP can normalize physiology in HFrEF-associated CSA without improving transplant-free survival for the population as a whole.^[5]
• Reassuring a patient or dismissing a CSA workup because "there's no snoring" — snoring is frequently absent or minimal in CSA, unlike OSA, and its absence should never be used to lower suspicion.
• Treating opioid-associated CSA as primarily a device-therapy problem rather than starting with a genuine, collaborative opioid dose-reduction or rotation discussion.^[4]
• Assuming a CPAP titration has simply "failed" the first time central events appear in a patient started for suspected OSA, rather than recognizing possible treatment-emergent (complex) CSA and reassessing after continued use.
• Continuing a fixed ASV or opioid regimen indefinitely without periodically reconfirming the safety criteria that justified it in the first place — a falling ejection fraction or a rising opioid dose can convert a previously appropriate plan into a contraindicated or unsafe one.
• Using theophylline in a patient with HFrEF without explicitly weighing its proarrhythmic risk against the population's already-elevated baseline arrhythmia risk.
• Managing witnessed apneic pauses with cyanosis in a patient on chronic opioids as routine CSA follow-up rather than urgently evaluating for opioid toxicity/overdose as a distinct, more time-critical problem.

How to use this table: Doses require patient-specific confirmation, including renal/hepatic adjustment where relevant. Select any drug name to open its complete Pharmasist Drug Database page. This table is intentionally short — central sleep apnea is primarily a device-therapy and underlying-cause-treatment condition, not a pharmacotherapy-first one; see Non-Pharmacologic Treatment and Choose Initial Treatment for the CPAP/oxygen/ASV framework and its critical HFrEF-specific safety limit.

Adjunctive Pharmacotherapy — Heart-Failure-Associated or Idiopathic Central Sleep Apnea

Reserve for patients in whom PAP therapy and/or supplemental oxygen are inadequate or not tolerated, after guideline-directed heart-failure therapy has been optimized. Neither agent is a first-line or stand-alone treatment.

Chronic Opioid Therapy — Dose Reduction or Rotation Is the Primary Intervention for Opioid-Associated CSA

These rows exist to flag the agents most relevant to the dose-reduction/rotation discussion, not to recommend initiating or increasing any of them. See the Opioid-Related Disorders and Acute Pain chapters for the full deprescribing/rotation framework.