Data from this large sample of asthma patients show that symptoms indicative of high OSA risk or diagnosed and untreated OSA are each associated with persistent asthma symptoms during the daytime. These associations were at least as strong as those found between OSA symptoms or history and persistent asthma symptoms at night, and were most robust among subjects who had both persistent daytime and nighttime symptoms. These relationships emerged independently of obesity and other contributors to poor asthma control. Furthermore, CPAP use attenuated the likelihood of persistent daytime symptoms. Although this cross-sectional study cannot prove causal relationships, our data strengthen available evidence for the relationship between OSA and asthma – specifically implicating a role for OSA in daytime asthma control – in a manner analogous to previously reported effects of this same nocturnal sleep disorder on daytime hypertension [20
]. These results suggest that OSA may have carry-over effects during the day, possibly through inflammatory pathways that exacerbate asthma, and that CPAP treatment for OSA may improve asthma around the clock.
By testing associations of OSA separately with daytime and nighttime symptoms that could not be separated previously [13
] and in a large sample of patients, the current findings represent a key advance which builds on knowledge generated by earlier reports [13
] that used the Asthma Control Questionnaire (ACQ).The ACQ was validated to assess overall asthma control [21
], and not to discriminate between daytime and nighttime asthma control, as most of its symptom-items do not specify timing. In this large sample of patients, we found significant independent associations of high OSA risk or diagnosed OSA with persistent daytime asthma symptoms that were at least as strong as those found with persistent nighttime symptoms ( and ).
Additionally, we observed a reduced likelihood for persistent daytime asthma symptoms () with CPAP use among previously diagnosed OSA patients. Previous interventional studies of CPAP therapy for OSA in asthma report improved outcomes, such as asthma symptoms [8
], rescue bronchodilator use [8
], PEFR [8
], and disease-specific quality of life measured on validated questionnaires [11
]; however, these studies have focused on patients with predominant nocturnal symptoms. In the only study which examined daytime and nighttime asthma symptoms separately in relation to OSA, Chan et al found that two weeks of CPAP in 9 subjects with OSA significantly improved asthma not only during the night but also during the day, and also improved morning and evening pre/postbronchodilator PEFR [8
]. Cessation of CPAP returned the PEFR to pre-CPAP levels [8
]. Our findings () now add to this evidence and together suggest that treatment for OSA in patients with asthma could improve asthma symptoms day and night
The mechanisms through which OSA worsens asthma still remain largely unstudied. Shared features or common comorbid conditions, such as obesity [22
] and GERD [23
], may link OSA with asthma. However, when we controlled for these variables in the multivariate models ( and ), the associations of interest persisted, as another recent study has also found [24
], suggesting that other links exist. Several mechanisms by which the events of airway obstruction in OSA might lead to immediate, but short lived effects on asthma symptoms have also been proposed. Increased nocturnal bronchial reactivity [25
] could be caused by vagal neural receptor activation that accompany the Muller maneuvers of obstructive events [26
]. Hypoxia stimulates the carotid body, and enhances vagal activity and bronchial reactivity [27
]. Additionally, short exposure to sustained hypoxia impairs defenses particularly important to asthmatics at night, such as cough [28
] and arousal thresholds to resistive loading [29
], and suppresses perception of asthma symptoms [30
]. These effects are short lived in most cases and thus would be expected to affect primarily nighttime asthma.
In contrast, our observed associations of high OSA risk and diagnosed OSA with persistent daytime symptoms ( and ) and the “protective” association with CPAP use (), do suggest that nocturnal OSA events may have carryover effects on asthma during the daytime. Inflammatory changes present in the UAW of OSA patients (reviewed in [31
]) may also be present in the lower airways. Compared with controls, OSA patients have elevated numbers of neutrophils, and higher concentrations of bradykinin and vasoactive intestinal peptide in nasal lavage fluid. Subepithelial edema, mucous gland hypertrophy, reduction of connective tissue and inflammatory cell infiltration were present on histology of surgical specimens from uvulopalatopharyngoplasty and tonsillectomy. In the uvula for example, although with a different cellular profile in each structural layer, this leukocyte-predominant inflammation is present at levels as deep as the musculature and may contribute to the UAW dysfunction. The snoring-related mechanical stress, and local and systemic effects of intermittent hypoxia have been implicated in the occurrence of UAW inflammation [31
]. These OSA features may have similar effects on the lower airways. Induced sputum, an established method to evaluate lower airway inflammation [33
] confirms a neutrophilic profile in OSA patients [34
] and polymorphonuclear neutrophil cell numbers correlate with AHI [35
]. Moreover, preliminary work in animal models shows neutrophilic lung inflammation arising from both apnea-related strenuous respiratory efforts and intermittent hypoxia [36
]. Additionally, one established feature of OSA is a sustained state of systemic inflammation, implicated in the associated cardiovascular morbidity [38
], which also shares similarities with asthma-related systemic inflammation [39
]. Thus OSA, through direct airway or systemic effects, may exacerbate a neutrophilic phenotype, one that is poorly characterized but increasingly recognized among subsets of asthmatics [40
]. CPAP may have multiple beneficial effects on the mechanical and neuromechanical properties of the lower airway, it may ameliorate gastroesophageal reflux and also the systemic and local inflammation [41
]. It is also possible that CPAP, through sleep restoration, may help improve asthma control around the clock, as recent studies indicate that proper sleep is an important factor in the control of disease and quality of life of asthma patients [43
]. A robust univariate relationship of diagnosed and untreated OSA with persistent nocturnal asthma symptoms was noted (). Attenuation of this association in the final multivariate model () does not negate its existence. Our subset of diagnosed but untreated OSA subjects was small (n=60), relative to much larger subsets of the other covariates included in the model, ie rhinitis (n=674), GERD (n=351), obesity (n=285). The small sample size limited statistical power to test for this specific association. Nevertheless, as our questionnaire suggests, there is likely a substantial burden of unrecognized and untreated OSA among asthmatics, a population which, as our data also put forth, could realize considerable benefit [8
] by having their sleep disorder addressed.
Limitations of our study include use of a questionnaire-based approach (SA-SDQ) to evaluate OSA symptoms prospectively, in a large number of patients, and also reliance on established clinical diagnoses in only a small number of subjects. However, prospective polysomnography on all subjects would have been prohibitively expensive for this study. Additionally, in a large sample of sleep patients, the SA-SDQ demonstrated high internal validity, and good sensitivity and specificity [18
]. This instrument has high diagnostic value in comparison to other sleep apnea screening instruments [45
]. Although the SA-SDQ has not been validated specifically in asthma patients, the scale does predict PSG-diagnosed OSA well in other samples [46
]. Our observation that the subset of subjects with diagnosed OSA had SA-SDQ scores nearly 10 points higher on average (37±6 vs. 28±7, p<0.0001) suggests that this instrument does have good utility in these patients; however, further studies aimed at its validation in asthmatics are needed.
The possibility of an overlap between OSA symptoms and nighttime asthma may also be raised. However, the symptoms included in the SA-SDQ primarily concern loud snoring and witnessed apneas specifically during sleep, and their worsening when supine or with alcohol. These sleep-related symptoms should be specific to upper rather than lower airway compromise. Further studies are necessary to validate the SA-SDQ specifically in patients with asthma.
Although asthma may contribute to OSA pathogenesis [7
], the cross-sectional design of this study precludes conclusions about causality. In the context of data from interventional studies [8
] and readily conceived underlying mechanisms as discussed above, our findings are suggestive of an effect of OSA on both daytime and nighttime asthma symptom control. Clearly, research on OSA and asthma is often complicated by shared comorbidities and potential for multidirectional causal pathways. For example, we cannot be certain whether obesity, GERD and psychopathology are more closely tied, etiologically, to asthma or OSA, or some may be intermediary variables in causal pathways between asthma and OSA. Finally, conducting this study in specialty-based clinic populations may have underestimated the true relationships of interest. The main breathing route during sleep is through the nose, and nasal congestion is a recognized risk factor for OSA [48
]. The high prevalence of rhinitis corroborated with its “protective” association with asthma symptoms (most likely a treatment effect), may have attenuated the true OSA risk, because nasal steroid treatment improves nasal breathing and reduces AHI in patients with rhinitis [49
]. This in turn may have attenuated the association of high OSA risk with asthma symptoms.