Risk Factors for WUS

Generally, ischemic stroke is most likely to occur in the early morning.Reference Tsai and Albers⁵ A meta-analysis of 31 studies of 11,816 stroke patients with known times of symptom onset found a circadian pattern of stroke onset with 55% of ischemic strokes occurring between 12:00AM and 6:00AM.Reference Elliott⁷ The increased risk of ischemic stroke in the early morning is attributable to many physiologic and disease factors. Much of this work follows the cardiology literature, as it is well known that myocardial infarction (MI) is more likely to occur in the early morning hours with about 40% of MI occurring in the morning.

In the morning, as persons begin to awaken and REM sleep becomes predominant, there is increased sympathetic nervous activity, Reference Somers, Dyken, Mark and Abboud⁸ which in turn stimulates the Renin-Angiotensin-Aldosterone system leading to elevations in blood pressure and heart rate. Further increasing the risk of WUS in the early morning is increased platelet aggregation and elevated levels of prothrombotic factors in the early morning.Reference Andrews, Gralnick, Merryman, Vail and Quyyumi^9–Reference Wouters, Lemmens, Dupont and Thijs¹² Platelet aggregation increases by as much as 27%–71% in the morning, which in part can be explained by increases in whole-blood platelet count and hematocrit as well as a doubling to tripling of catecholamines levels in the early morning.Reference Andrews, Gralnick, Merryman, Vail and Quyyumi⁹ These physiologic changes, which occur in sleep and in the early morning in normal individuals may explain at least in part, why ischemic stroke has a tendency to occur during sleep as platelet aggregation favors the formation of clots, specifically in the early morning. Physiologic changes favoring stroke during sleep may be enhanced in those with stroke risk factors. In hypertensive elderly patients, there are significant elevations in von Willebrand factor (vWF) early in the morning that may contribute to enhanced morning platelet aggregation.Reference Kario, Yano, Matsuo, Hoshide, Asada and Shimada^10, Reference Goto, Sakai and Goto^13, Reference Ikeda, Handa and Kawano¹⁴ Atrial fibrillation may also be associated with WUS, as the odds of a new atrial fibrillation diagnosis is nearly four-fold higher in patients with WUS compared to those having stroke in wakefulness (non-WUS).Reference Riccio, Klein and Pagani Cassara^15, Reference Kim, Kim, Kwon, Kim and Kang¹⁶

Disordered sleep may further predispose to WUS. There is some evidence that WUS is associated with severe obstructive sleep apnea (OSA) in men. Analysis of the Sleep Heart Health Study data of over 5,000 community-dwelling individuals showed that severe OSA with at least 20 apneas or hypopneas per hour of sleep is associated with an almost 3-fold ischemic stroke incidence over 8 years in men.Reference Redline, Yenokyan and Gottlieb¹⁷ No such increased stroke incidence was found in women. Consistent with this gender differentiation, another study showed that among 54 men with ischemic stroke who underwent polysomnography, nearly half with WUS compared to under 20% with non-WUS had severe OSA and frequent oxygen desaturation. Still in this study, the majority of both those with WUS and without WUS had OSA as defined by an apnea-hypopnea index (AHI) of ≥ 5, 81.5% with WUS versus 79.6% with non-WUS. The average AHI in WUS patients was reported to be 22.5 (26.9 in men, 9.8 in women), while the mean AHI in patients with non-WUS was 19.7 (17.9 in men, 22.8 in women).Reference Koo, Bravata and Tobias¹⁸ On the other hand, Hsieh et al. defined severe sleep-disordered breathing (SDB) as AHI ≥ 30, which was more prevalent in WUS patients compared to non-WUS patients (38.5% of WUS patients, 8.9% of non-WUS patients).Reference Hsieh, Lai, Liu, Hsieh and Hsu¹⁹ In OSA, increased cerebral blood flow occurs in the immediate post-apnea period and after about 15 seconds an approximately 20% decrease in cerebral blood flow occurs. At the same time oxygen desaturation following an apnea decreases partial pressure of oxygen in this decreased amount of blood going to the brain. The end result may be an ischemic stroke.Reference Zirak, Gregori-Pla and Blanco²⁰ Additionally, intermittent hypoxia in OSA is also associated with blood hypercoagulability as both blood viscosity and platelet aggregation increase, while fibrinolytic activity decreases.Reference Bradley and Floras²¹ Indeed, other studies have shown that frequent nocturnal oxygen desaturation is more frequent among those with WUS compared to non-WUS. A study in a South Korean stroke unit showed that nearly 30% of persons with WUS compared to just 12% of those with non-WUS had frequent nocturnal desaturation.Reference Kim, Ko and Jeong²² Patients with WUS appear to have similar age and gender distribution as those with non-WUS.Reference Moradiya and Janjua⁶

Mechanisms of WUS

Analysis of International Stroke Trial data of over 17,000 acute ischemic strokes showed that WUS was more likely to occur by a lacunar mechanism and less likely to occur through a large vessel anterior circulation mechanism.Reference Moradiya and Janjua⁶ Our data also showed that WUS was more likely to occur through a small vessel mechanism, as 42.9% of 28 persons with WUS and just 14.0% of 44 persons with non-WUS had stroke occurring by small vessel mechanism.Reference Tanimoto , Mehndiratta and Koo²³ In two independent stroke cohorts, we also found that persons with WUS had significantly higher low-density lipoprotein (LDL) than persons with non-WUS with large differences in each study of about 20 mg/dL.Reference Koo, Bravata and Tobias^18, Reference Tanimoto , Mehndiratta and Koo²³ In addition, to diabetes and hypertension, hyperlipidemia and specifically high LDL cholesterol is an independent risk factor for lacunar stroke.Reference Bezerra, Sharrett and Matsushita²⁴ Patients with blockages of small arteries may be especially vulnerable to WUS in the setting of OSA, when near ischemic brain regions are further challenged with the severe hypoxia of OSA.

Ozdemir et al. reported in a case series that the presence of patent foramen ovale might increase the risk of cryptogenic strokes on awakening due to paradoxical emboli leading to systemic embolism.Reference Ozdemir , Beletsky, Hachinski and Spence²⁵ Patients with severe OSA are about 3 times more likely to have a large patent foramen ovale than persons without OSA.Reference Shaikh, Jaye and Ward²⁶ During an obstructive apnea, changes in intrathoracic pressure result in a transient increase in right-sided heart pressure, predisposing to paradoxical embolism.Reference Beelke, Angeli and Del Sette²⁷ Additionally, persons with WUS are more likely than those with non-WUS to have right-to-left shunting.Reference Ciccone, Proserpio, Roccatagliata, Nichelatti, Gigli and Parati²⁸

Neuroimaging in WUS

In WUS, the exact time of stroke onset is unknown. Patient with WUS presenting to an acute care facility beyond 4.5 hours of sleep onset are frequently considered ineligible for IV thrombolytic therapy, even though the time of onset may in fact have been within 4.5 hours.Reference Jauch, Saver and Adams^29–32 Because of the unknown time of stroke onset in WUS, many studies have targeted neuroimaging as a way to approximate time of onset and potentially select patients for thrombolytic treatment.Reference Tsai and Albers⁵

Computed Tomography in WUS

Computed Tomography (CT) is uniformly used in clinical practice to exclude cerebral hemorrhage in patients with acute neurological symptoms, and also to guide the use of thrombolytic therapy in patients with hyperacute ischemic stroke.Reference Adams, Adams and Brott^33, Reference Tomura, Uemura, Inugami, Fujita, Higano and Shishido³⁴ Early changes seen on CT result from cerebral ischemia affecting water diffusion regulation in brain tissue leading to cytotoxic edema within a few minutes of an ischemic insult.Reference Alegiani, MacLean and Braass³⁵ This edema occurs due to increased microvascular permeability, as the ischemic insult causes significant disruption of microvascular barriers.Reference Zoppo, Kummer and Hamann³⁶ A few CT-based studies reported similarity in early ischemic changes observed in both WUS and non-WUS.Reference Huisa, Raman and Ernstrom^37–Reference Todo, Moriwaki, Saito, Tanaka, Oe and Naritomi⁴⁰ Roveri et al. compared early ischemic changes on CT imaging in WUS and non-WUS patients with similar clinical and demographic characteristics. There was a comparable degree of early ischemic change as evidenced by similar amounts of parenchymal hypoattenuation, blurring of the gray-white junction, and swelling as judged by the Alberta Stroke Program Early CT Score (ASPECTS), in the baseline CT scan in both groups. Follow-up CT scans in both groups also showed similarity in the extent of infarction. Based on these findings, it was presumed that the onset of WUS was shortly before, or even upon awakening.Reference Roveri, La Gioia, Ghidinelli, Anzalone, De Filippis and Comi³⁸

The ability of CT-perfusion (CTP) to detect the extent of the ischemic penumbra following an ischemic event was evaluated in a few studies. For this purpose, different CTP parameters were used including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT) and time-to-peak (TTP). CBV is the amount of blood present per brain tissue volume, while MTT is the average time the red blood cells spend in the capillary circulation between the arterial inflow and the venous outflow. TTP is defined as the time taken by the contrast material to achieve the best possible enhancement in the area of interest.Reference Khandelwal⁴¹ If the CBF is decreased, as in an ischemic event, but the CBV is stable or even increased, this indicates reversible ischemia. In contrast, if both CBF and CBV are decreased, this is considered a sign of irreversible ischemia.Reference Tomandl, Klotz and Handschu^42, Reference Srinivasan, Goyal, Al Azri and Lum⁴³ Using these parameters, Wintermark et al. stated that CTP could be used to draw the line between the area of ischemic infarction and the salvageable penumbra.Reference Wintermark, Flanders and Velthuis⁴⁴ Campbell et al. used a different CTP approach in which they stated that the relative CBF is the optimal parameter in appreciating the size of the infarcted core of the ischemic tissue.Reference Campbell, Christensen and Levi⁴⁵

Promising results were obtained in EXTEND, a multicentric randomized clinical trial, which randomized both WUS patients (if within 9 hours from the midpoint of sleep) and patients presenting 4.5–9.0 hours after the last known well time to IV alteplase or placebo. The investigators used both CTP imaging and perfusion-diffusion magnetic resonance imaging (MRI) techniques to determine the extension of the hypoperfused, yet salvageable brain areas. A favorable outcome was defined as a modified Rankin score (mRS) of 0–1, which was reported in 40 patients (35.4%) in the alteplase group versus 33 patients (29.5%) in the placebo group (adjusted risk ratio, 1.44; 95% CI 1.01–2.06; p = 0.04). However, symptomatic intracerebral hemorrhage occurred in 7 patients (6.2%) in the alteplase group versus 1 patient (0.9%) in the placebo group (adjusted risk ratio, 7.22; 95% CI 0.97–53.5; p = 0.05).Reference Ma, Campbell and Parsons⁴⁶ Furthermore, a meta-analysis was conducted looking at the pooled data from EXTEND, Reference Ma, Campbell and Parsons⁴⁶ ECASS-4: EXTENDReference Ringleb, Bendszus, Bluhmki, Donnan, Eschenfelder and Fatar⁴⁷ and EPITHETReference Davis, Donnan and Parsons⁴⁸ (EXTEND trial used CTP and MRI to judge the onset of stroke, ECASS-4: EXTEND and EPITHET trials used MRI only) to evaluate the efficacy of IV alteplase in acute ischemic stroke when used beyond the 4.5-hour window. Excellent functional outcome, defined as mRS of 0–1 at 90 days, occurred in 36% of the alteplase group and 29% of the placebo group, while functional independence, defined as mRS of 0–2 at 90 days, occurred in 49% of the alteplase group and 44% of the placebo group. Adverse events of symptomatic intracerebral hemorrhage occurred more in the alteplase group (5% alteplase group and <1% placebo group, odds ratio 9.7; 95% CI 1.23–76.55, p = 0.031).Reference Campbell, Ma and Ringleb⁴⁹

Currently, TWIST is an ongoing open-label randomized controlled clinical trial that evaluates the efficacy of tenecteplase in patients presenting with WUS.⁵⁰ Tenecteplase is a genetically modified alteplase molecule, which has higher specificity to fibrin, a longer half-life, and more rapid onset of action than alteplase.Reference Tsikouris and Tsikouris^51, Reference Dunn and Goa⁵² This study is also assessing the ability of plain CT and CT angiography in recognizing WUS patients who are eligible to benefit from thrombolytic treatment with tenecteplase.⁵⁰

As compared to CT, the use of MRI in ischemic strokes is more complicated and takes longer time to be done.Reference Vymazal, Rulseh, Keller and Janouskova⁵³ Despite the broader availability of CT imaging as compared to MRI, the ability of CT to detect early ischemic changes in the brain is limited as the sensitivity of CT in recognizing early ischemic changes depends on the duration and severity of focal cerebral ischemia.Reference Kummer, Nolte, Schnittger, Thron and Ringelstein⁵⁴ Another disadvantage of CT imaging carries the risk of radiational exposure.Reference Vymazal, Rulseh, Keller and Janouskova⁵³

Magnetic Resonance Imaging (MRI) in WUS

Due to the limitations of CT in estimating stroke onset and evaluating the degree of damage, MRI has been used in different studies to perform these tasks.Reference Chalela, Kidwell and Nentwich^55–Reference Schellinger, Thomalla and Fiehler⁵⁷ MRI has the advantage of better tissue delineation than CT and facilitates the characterization of tissue areas at risk of infarction in the immediate time after an ischemic event (ischemic penumbra).Reference Albers, Thijs and Wechsler^58–Reference Thomalla, Schwark and Sobesky⁶⁰ Two MRI-based mismatch principles are currently followed to detect an ischemic penumbra: (a) perfusion weighted imaging (PWI)–diffusion-weighted imaging (DWI) mismatch and (b) DWI–fluid attenuated inversion recovery (FLAIR) mismatch.Reference Kurz, Advani, Behzadi, Eldoen, Farbu and Kurz⁶¹

PWI–DWI mismatch is based on the sequence of events taking place in the brain directly following an ischemic insult. In the few minutes after vessel occlusion, loss of blood supply causes irreversible damage to a core of infarcted tissue. There is also an ischemic penumbra surrounding this infarcted core, which is at risk for infarction but can still be salvaged.Reference Bang⁶² The infarction core is characterized by impaired diffusion of water molecules resulting from cytotoxic injury, edema and depleted cellular energy due to hypoperfusion, and this diffusion defect can be detected by DWI.Reference Bang⁶² On the other hand, PWI provides information about microcirculation in the capillary network.Reference Kim, Kang and Kim^63, Reference Tatlisumak, Strbian, Abo Ramadan and Li⁶⁴ Thus, it can capture the hemodynamic changes taking place in the ischemic tissue and the surrounding brain regions.Reference Neumann-Haefelin , Wittsack and Wenserski⁶⁵

PWI-DWI mismatch refers to the difference between the area of infarcted brain (diffusion defect) and the area of the brain where perfusion is compromised (perfusion defect).Reference Jovin, Saver and Ribo^66, Reference Lansberg, Straka and Kemp⁶⁷ Heiss et al. stated that MRI-based PWI-DWI mismatch could be used to differentiate the area with irreversible damage from the surrounding area with reversible hypofunctionality with a comparable reliability to positron-emission tomography (PET).Reference Heiss and Sobesky⁶⁸ Therefore, patients with a PWI-DWI mismatch are thought to potentially benefit from thrombolytic therapy in acute stroke settings. Fink et al. reported a similarity of PWI-DWI mismatch within 3 hours in both WUS and non-WUS. PWI-DWI mismatch was found in 82% of known-onset stroke patients and 73% of WUS patients.Reference Fink, Kumar and Horkan¹ The same mismatch concept was used to guide mechanical and chemical thrombolytic therapy in a case report of two WUS patients, both of whom had severe deficits at onset, yet favorable ultimate outcomes, one being completely free of symptoms and the other having only mild language deficit.Reference Iosif, Oppenheim, Trystram, Domigo and Meder⁶⁹

Several studies looked at the reliability of PWI–DWI mismatch in combination with DWI–FLAIR mismatch in identifying eligible patients with unknown-onset strokes to guide thrombolytic therapy.Reference Breuer, Schellinger and Huttner^70–Reference Kang, Sohn and Hong⁷³ Like DWI, MRI FLAIR sequencing leverages abnormalities in the properties of water to detect changes in cerebral tissue exposed to ischemia. DWI detects defects in diffusion of water within a few minutes after the ischemic event, while FLAIR detects cerebral edema over 3-6 hours. Based on these principles, the presence of DWI defect, without a matching FLAIR lesion, indicates that the ischemic infarction is of a recent onset, most likely < 4.5 hours.Reference Rubin and Barrett^39, Reference Aoki, Kimura, Iguchi, Shibazaki, Sakai and Iwanaga^74–Reference Thomalla, Rossbach and Rosenkranz⁷⁶

Using these imaging concepts, Petkova et al. retrospectively studied FLAIR and DWI records of 130 patients with acute stroke of known onset. Sixty-three patients had their imaging done within the first 3 hours after the onset of stroke symptoms, while the remaining 67 patients underwent imaging more than 3 hours after the symptom onset. The sensitivity and specificity of DWI–FLAIR mismatch in identifying strokes, which occurred within 0–3 hours were 90% and 93%, respectively.Reference Petkova, Rodrigo and Lamy⁷⁷ Other studies have shown that DWI-FLAIR mismatch has 62% sensitivity and 78% specificity in identifying stroke patients within 4.5 hours of symptom onset.

Several investigators have used this concept of DWI–FLAIR mismatch to guide thrombolytic therapy in WUS.Reference Aoki, Kimura and Iguchi^78–Reference Mourand, Milhaud and Arquizan⁸⁰ Outcomes in these patients in general were favorable and indicated that persons with WUS and DWI–FLAIR mismatch without contraindication to thrombolytic therapy may indeed benefit from this therapy without having excess risk of symptomatic hemorrhage.

MRI-Guided Clinical Trials

Recognizing the knowledge gap in acute therapy for WUS, investigators have carried out clinical trials to guide IV t-PA therapy in strokes with unknown symptom onset including WUS. Table 2 outlines details of these trials which are also commented on in the below section.

Table 1.

Studies assessing correlation between OSA and WUS.

BMI indicates body mass index.

Table 2.

Trials and observational studies assessing efficacy of MRI-guided thrombolytic therapy in WUS, Unknown-onset stroke (UOS) and/or 4.5–9 hours post stroke onset.

Sample size refers to UOS patients who underwent MRI/CT scanning and received IV thrombolysis. mRs indicates modified Rankin scale.

N/A, not applicable; WUS, wake-up strokes.

The MR WITNESS trial is a single arm safety trial, which included 183 patients who had acute ischemic stroke with unwitnessed symptom onset and were last known to be well 4.5–24 hours earlier. Safety outcomes included the primary outcome of symptomatic hemorrhage and a secondary outcome of brain edema risk. DWI–FLAIR mismatch was used to guide potential therapy. Patients were excluded if there was severe stroke with NIH Stroke Scale of > 25, a large infarction demonstrated on DWI (>100 cm³), or more than minimal FLAIR hyperintensity (> 1.15 signal intensity ratio). Eighty patients, including 57 WUS patients received IV t-PA. Symptomatic hemorrhage occurred in only one patient, and symptomatic cerebral edema occurred in only three subjects, both rates not different from symptomatic hemorrhage and edema rates observed in the ECASS-3 study. Ultimately, the trial successfully demonstrated that IV t-PA could safely be administered to acute ischemic stroke patients with unwitnessed symptom onset and last known well of 4.5–24 hours when quantitative DWI–FLAIR mismatch was present.Reference Schwamm, Wu and Song⁸¹

The MRI-Based Thrombolysis in Wake-Up Stroke (WAKE-UP) trial is a multicenter, randomized, double blinded, placebo-controlled clinical trial that involved patients with strokes of unknown onset, predominantly (about 90%) those patients with WUS. Eligible patients had ischemic stroke with unknown onset, a last known well time of greater than 4.5 hours, and DWI-FLAIR mismatch showing acute stroke on DWI with no hyperintensity shown on FLAIR imaging. Two hundred and fifty-four patients were randomized to receive IV t-PA and 249 patients were randomized to receive placebo. The alteplase group were more likely than the placebo group to have a favorable outcome (modified Rankin Scale of 0 or 1), 53.3% versus 41.8% (adjusted odds ratio of favorable outcome of 1.61; 95% CI 1.09–2.36 with a p = 0.02. Deaths (odds ratio 3.38; 95% CI 0.92–12.52, p = 0.07) and symptomatic hemorrhage (odds ratio 4.95; 95% CI 0.57–42.87; p = 0.15) occurred more commonly in the alteplase group, but not significantly so. The investigators concluded that IV t-PA is associated with favorable clinical 90-day outcome in patients with unknown-onset strokes when administration is guided by DWI-FLAIR mismatch. Unfortunately, the study had to be terminated early due to discontinuation of funding, thus limiting the interpretation of safety results which showed trends toward increases in symptomatic hemorrhage and mortality in the alteplase compared to placebo group.Reference Thomalla, Simonsen and Boutitie⁸²

ECASS-4: EXTEND is a double-blinded randomized placebo-controlled clinical trial that used both PWI and DWI techniques to determine the presence of salvageable ischemic brain tissues in patients who presented 4.5–9 hours after the onset of the stroke or those who had the stroke symptoms upon waking up. Patients were randomized to either IV alteplase or placebo and the mRS distribution at 90 days did not show a significant difference between the two groups (odds ratio 1.20; 95% CI 0.63–2.27, p = 0.58). Also, the mortality rate at 90 days was not remarkably different between the two groups (11.5% alteplase and 6.8 placebo, p = 0.53) and single event of symptomatic intracerebral hemorrhage occurred in the alteplase group. However, it should not escape from notice that this trial has been terminated prematurely with a limited sample size of 119 patients (61 alteplase and 58 placebo).Reference Ringleb, Bendszus, Bluhmki, Donnan, Eschenfelder and Fatar⁴⁷

The THAWS trial is a randomized controlled trial, which, similar to the WAKE-UP trial, used MRI-based thrombolysis in Japanese patients with stroke of unknown onset, and compared its efficacy to standard treatment.Reference Koga, Toyoda and Kimura⁸³ The alteplase and standard care groups had similar rates for favorable outcome, defined as mRS of 0–1 at 90 days (relative risk 0.97; 95% CI 0.68–1.41, p = 0.892). However, it should be noted that there was comparable safety of alteplase use in stroke with unknown onset to that of standard care.Reference Koga, Yamamoto, Inoue, Asakura, Aoki and Kanzawa⁸⁴

Therapy for OSA in WUS

There is a recognized high prevalence of OSA in ischemic stroke between 60% and 80%.Reference Johnson and Johnson⁸⁵ Evidence further suggests that severe OSA and severe hypoxemia occur with increased frequency in patients with WUS compared non-WUS.Reference Koo, Bravata and Tobias^18, Reference Kim, Ko and Jeong²² In WUS, OSA may be directly related to the mechanism of stroke and it may be particularly important in these patients to conduct polysomnography to rule out OSA. Furthermore, severe hypoxemia in OSA can only be detrimental in the setting of a new stroke and may in part account for neurologic worsening and poor functional and cognitive outcomes, which are known to occur at higher rates when OSA co-occurs with stroke.Reference Iranzo, Santamaria, Berenguer, Sanchez and Chamorro^86–Reference Aaronson, Bennekom and Hofman⁸⁸ Aaronson et al. reported in the results of their case-control study that stroke patients with OSA have a significant impairment of their attention, executive functions, visuoperception, psychomotor abilities, and intellectual functions when compared to stroke patients who did not have OSA.Reference Aaronson, Bennekom and Hofman⁸⁸ Based on this, a randomized controlled trial evaluated the functional and cognitive outcomes of CPAP use for 4 weeks in stroke patients and compared these outcomes to those in stroke patients who received rehabilitation treatment but not CPAP. Taking into consideration that this trial did not include patients with OSA, the CPAP use in stroke patients was associated with a better cognitive outcome in attention and executive functioning.Reference Aaronson, Hofman and Bennekom⁸⁹ For these reasons it is important to conduct polysomnography in all patients with ischemic stroke but perhaps particularly in those with WUS. Recent data shows that in persons with ischemic stroke and OSA, continuous positive airway pressure (CPAP) when used compliantly is associated with improvement in the one year modified Rankin Scale of nearly one point.Reference Bravata, Sico and Vaz Fragoso⁹⁰ These data are for patients with ischemic stroke regardless of stroke timing. In WUS, treatment of OSA may be especially beneficial if the mechanism of stroke directly involved OSA.

The SAVE trial is an open-label randomized controlled trial comparing the use of CPAP combined with usual care to usual care alone in patients with moderate to severe OSA and either coronary artery disease or a cerebrovascular disease over a span of 3.7 years. The primary outcome of the study was death from cardiovascular events, stroke, MI or hospitalization due to unstable angina, heart failure or transient ischemic attacks. CPAP use did not decrease the risk of cardiovascular or cerebrovascular events including stroke (hazard ratio with CPAP 1.10, 95% CI 0.91–1.32, p = 0.34). The effectiveness of CPAP in affecting outcomes could have been reduced by the overall low CPAP usage in the intervention group of 3.3 hours per night.Reference McEvoy, Antic and Heeley⁹¹