Incidence of peripheral vestibular disorders in individuals with obstructive sleep apnea

1 Introduction

Obstructive sleep apnea (OSA) syndrome is one of the most prevalent chronic respiratory disorders, which is characterized by recurrent pharyngeal collapse during sleep. Repetitive episodes of airflow limitation causing oxygen desaturation and hypercapnia lead to increased respiratory drive followed by frequent arousal [19]. Intermittent hypoxia, hypercapnia, and disturbed sleep architecture are associated with various pathological phenomenon resulting from OSA [19]. Previous literature showed that OSA increased the risk of systemic diseases including hypertension, stroke, myocardial infarction, heart failure, diabetes, headache, etc [10].

Several laboratory studies have shown that peripheral vestibular end-organs and the vestibular nucleus can be damaged by hypoxia or reactive oxygen species [4, 35]. Previous cross-sectional studies reported that caloric responses were decreased in subjects with OSA, especially in cases with worse apnea-hypopnea index [8, 13]. In a population study, the risk of vertigo was significantly higher in individuals with OSA [29]. In addition, the incidence of Meniere’s disease was reported to be higher in female and in middle-aged subjects with OSA [14]. Thus, not only may subjective dizziness be related to daytime somnolence and OSA, but peripheral vestibular disorders may also be associated with OSA [8, 29]. So far, however, there has been a lack of studies assessing the incidence of specific vestibular diseases in individuals with OSA.

In the present study, we hypothesized that the incidence of peripheral vestibular disorders would be affected by preexisting OSA. A population-based cohort study using the Korean National Health Insurance Service (KNHIS) Sample Cohort database was designed to compare the incidence of peripheral vestibular disorders in individuals with OSA and matched non-OSA subjects. Disorders causing peripheral vertigo included BPPV, Meniere’s disease and vestibular neuritis, which are known to be due to pathologies of the otolith organs, endolymphatic system, and vestibular pathway, respectively. In short, this study aimed to evaluate the risks of peripheral vestibular disorders in individuals with OSA using a nationwide claims data.

2 Methods

2.2 Study design and selection of cohorts

A total of 3,037 individuals diagnosed and managed with OSA were identified from 1 Jan 2002 to 31 Dec 2015 (Fig. 1). Participants who had been diagnosed with dizziness-related disorders (H81, H82, and H83, peripheral vertigo; R42, subjective dizziness; I63 and I61, cerebral infarction and hemorrhage) before the study was initiated were excluded.

OPEN IN VIEWER

Fig. 1

Flowchart of the construction of the cohorts.

The OSA patients were matched 1:4 with comparison participants who had never been diagnosed with OSA [1, 26]. The matching process was conducted for age, sex, income, dwelling place, and Charlson Comorbidity Index (CCI) scores. The latter is calculated based on the presence of seventeen selected medical conditions including congestive heart failure, dementia, diabetes, liver disease etc [6]. The score ranges from 0 (no comorbidities) to 33 after weighting each item according to the validated method [6]. To prevent selection bias during the matching process, the participants were sorted using a random number order, then selected from top to bottom. It was assumed that the comparison participants were involved at the same time as the corresponding matched OSA patient. Propensity score matching was performed.

In a total of 2,082 OSA patients and 8,328 matched non-OSA subjects, development of peripheral dizziness was evaluated from the time of enrollment to December 31, 2015.

3 Results

This study evaluated the incidence of peripheral vestibular disorders in individuals with OSA using nationwide medical claim data. Individuals diagnosed with OSA showed meaningfully higher incidence of peripheral vertigo than subjects without OSA (IRR = 6.28, adjusted HR = 6.64). The incidences of BPPV, Meniere’s disease, and vestibular neuritis in the OSA group were more than 6, 3 and 9 times higher, respectively, than in the non-OSA group. To the best of our knowledge, the present study is the first to assess the risk of specific peripheral vestibular disorders in individuals with OSA using a population database. The results support a positive association between OSA and risk of peripheral vertigo.

OSA is a very common sleep-related breathing disorder characterized by recurrent episodes of hypoxia and reoxygenation. Chronic exposure to intermittent hypoxia and the resulting activation of the sympathetic nervous system contribute to damage to the systemic cardiovascular and endocrine systems [5]. The peripheral vestibular system may also be affected by hypoxia and the endothelial changes [4, 35]. Previous literature showed that poor sleep quality in OSA was also associated with vertigo [15, 33]. In a population study using the National Health Insurance Research Database of Taiwan, people with sleep apnea had a significantly higher cumulative incidence of vertiginous syndromes (code 386 in the ICD-9-CM system; includes peripheral vertigo, Meniere’s disease, labyrinthitis, labyrinthine fistula, vertigo due to central origin) than those without sleep apnea (HR = 1.71) [29]. In our study, the diagnostic codes were narrowed to specific peripheral vestibular disorders including BPPV, Meniere’s disease, vestibular neuritis, and other peripheral vestibulopathy. As a result, the hazard of developing peripheral vertigo was notably higher in people with OSA (HR = 6.64) and in females (HR = 1.92).

One of the hypotheses explaining the general effects of OSA on peripheral vestibular function is nocturnal cerebral ischemia. In OSA, cerebral blood flow increases and decreases along with concomitant changes of arterial pressure during repetitive apnea [3, 30]. Due to the rapid changes of blood pressure and flow, it is difficult for the cerebral autoregulation system to avoid cerebral ischemia. As the inner ear structure is supplied by terminal branches of the anterior inferior cerebellar artery without collateral circulations, the vestibular end-organs would be particularly vulnerable to circulatory alternations causing tissue ischemia. With regard to BPPV, previous literatures have been suggested that vascular insufficiency of the inner ear is a possible cause [11, 34]. Periodic experimental hypoxic stress in marine vertebrates has been reported to significantly alter the chemical composition of the otolith [23]. Since otolith homeostasis is maintained by the balanced production of reactive oxygen species by NADPH oxidase-3 (NOX3), it is possible that dysregulation of NOX3 in response to ischemia-reperfusion leads to defective otoconial homeostasis which may cause otolith to be easily dislodged [16, 21]. In Meniere’s disease, a histopathologic temporal bone study showed that a circulatory disturbance may be the cause of abnormal endolymph homeostasis [12, 27]. As the volume of endolymphatic fluid increases, the cochlear lateral wall containing capillary networks is compressed against the bony labyrinth, and the increased cochlear outflow resistance may lead to ischemia-reperfusion injury. For this reason, subjects with preexisting problems of microvasculature could be susceptible to inner ear damage caused by endolymphatic hydrops [7]. In addition, a clinical study revealed that OSA patients with Meniere’s disease benefited from solitary continuous positive airway pressure therapy for managing vertigo and hearing loss [24]. In vestibular neuritis, although the etiology of acute unilateral peripheral vestibular deficit still remains unclear, it could be suggested that the chronically repeated hypoxic episodes would affect the neuronal activities of the vestibular nucleus [35] as well as immune responses to herpes virus [36].

The prevalence of peripheral vestibular disorders is known to be higher in women than in men [9]. In our study, as expected, the incidence was higher in females than in males for BPPV, Meniere’s disease, and vestibular neuritis (Tables 2, 4). The multivariable analysis showed that women had 1.92 times the hazards of developing peripheral vestibular disorders compared to men. The ratio of incidence rate between the OSA and non-OSA cohorts, however, did not meaningfully differ according to gender, meaning that effect of OSA on the incidence of peripheral vestibular disorder would be similar for both women and men.

The strength of this study is that it is a nationwide population-based cohort study that careful excluded possible confounding risk factors for peripheral vestibular disorders and built a control cohort strictly matched for socio-economic factors and comorbidity index. However, this study also has several limitations. First, we used diagnostic codes in administrative claims data. Although medical accessibility in South Korea is excellent [25], people who were unaware of their sleep problems or not willing to visit hospitals could not be included in the analysis. Also, vestibular migraine was not evaluated not only because it tends to be the result of central rather than peripheral pathology but also because the ICD-10 system did not designate a separate diagnostic code for it. In addition, issue of diagnostic accuracy inevitably follows, which is thought to have little impact on the reliability of the comparison results, as this can occur with similar trends in both the OSA and matched comparison groups. Second, the claims data did not include any relevant information about the severity of OSA, polysomnography results and whether subjects had been treated properly. Finally, individuals with OSA who were followed by otorhinolaryngologists or neurologists were more likely to be diagnosed when they developed dizziness. Although there are limitations, the higher incidence of peripheral vertigo in individuals with OSA was consistently observed, with a more than six-fold increased risk, and further studies are needed to follow. We expect that large-scale prospective clinical studies will provide further insights into the causal relationship between OSA and peripheral vertigo in the future.

5 Conclusion

In this population-based cohort study using nationwide administrative claims data, people with a diagnosis of OSA had a higher incidence of peripheral vestibular disorders, including BPPV, Meniere’s disease, and vestibular neuritis, than socio-demographically matched people who had never been diagnosed with OSA. It is recommended to consider peripheral vertigo when counseling and managing OSA.