Sleep Disorders in Mitochondrial Diseases

The most frequent sleep disorders described in the context of MDs belong to the sleep-disordered breathing (SDB) group. SDB includes a constellation of disturbances classified in four major categories: obstructive sleep apnea (OSA), central sleep apnea (CSA) syndrome, sleep-related hypoxemia disorders, and sleep-related hypoventilation disorders [22]. Each of these disorders has been associated with MDs.

The physiological changes that occur during sleep make this a critical time for the process of breathing. First, the effect of gravity in the supine position determinates a reduction of the total lung capacity and a narrowing of the velopharynx [23], both contributing to the increased upper airway resistance, as well as the reduced tonic drive of pharyngeal dilator muscles [24]. Furthermore, physiological modifications observed in breathing during sleep include a reduction of respiratory rate [25], a diminished sensitivity of chemoreceptor of the respiratory center [26], and the absence of stimulus of wakefulness drive to respiration [27]. Finally, the reduction of the muscular tone involving the accessory respiratory muscles is associated with normal diaphragmatic activity in order to guarantee an adequate ventilation. This aspect is particularly important for REM sleep characterized by a complete muscular atonia and, therefore, represents the most critical sleep stage for respiration [28]. On these bases, it is expected that patients affected by neuromuscular disorders are particularly vulnerable to develop SDB [29]. OSA is the most common subtype of SDB, and it is characterized by intermittent and repetitive episodes of partial (hypopnea) or complete (apnea) obstruction of the upper airways causing falls in blood oxygen hemoglobin saturation and disruption of sleep. Daytime symptoms include excessive daytime sleepiness (EDS), fatigue, morning headache, and cognitive or mood alterations (e.g., memory loss, irritability, and depression). In our previous paper, we described a large population of adult patients affected by MDs investigated by a polysomnographic study [19••], revealing a high prevalence of OSA (35/103, 34%). Particularly, the prevalence of this specific SDB was significantly higher in the phenotypes of MDs associated with the higher grade of muscular involvement. Interestingly, the classical risk factors for OSA described in general population, such as obesity, were not significantly associated with OSA in our MD patients. Příhodová and colleagues observed a high prevalence of OSA (22%) in patients affected by Leber hereditary optic neuropathy (LHON) and dominant optic atrophy (DOA). Interestingly, the prevalence of OSA in these patients did not differ between symptomatic and asymptomatic groups [21••]. At the same time, the burden of OSA in MDs has been also revealed in pediatric populations. In particular, Jeyakumar et al. [15] observed a higher prevalence of OSA (9.8%) in pediatric patients affected by MDs when compared with that described in the general pediatric population (2%). These findings were further confirmed by Mosquera and colleagues [17••], reporting a pediatric group of patients affected by MDs by means of video-polysomnography. In their study, the authors revealed a high incidence of SDB (10/18, 56%) with a clear prevalence of OSA (6/18, 33.3%). Interestingly, the prevalence of SDB was prominent in patients with an abnormal muscular tone, while the classical risk factors of OSA, such as adenotonsillar hypertrophy, allergic rhinitis history, and post-tonsillectomy status, were not associated with SDB, suggesting that the genetic neuromuscular disease contributes to sleep respiratory disturbances. These data seem to confirm the results reported in our research paper, documenting the key role of skeletal muscle involvement in the occurrence of specific SDB [19••]. Moreover, the presence of SBD in pediatric MD population suggests that respiratory sleep disorders are an early manifestation of disease, and therefore, polysomnography should be performed as promptly as possible in these fragile patients.

Conversely, Smits et al. [20••] found a low prevalence of OSA (1/20, 5%) in a population of 20 adult patients with chronic progressive external ophthalmoplegia (PEO), with slightly increased AHI. On the other hand, in this population, CSA was the prominent SDB (4/20, 20%). A possible explanation for the discrepancy of data between the two studies is probably to be found in the different phenotypes associated with mutations in POLG gene. In our recent published article [19••], the patients associated with pathological variants in POLG presented a PMM with PEO phenotype, while Smits and colleagues investigated the presence of SDB in subjects with POLG mutations and ataxia neuropathy spectrum (ANS) phenotype. In light of these considerations, it is possible to hypothesize that patients with pathogenic variants in the same gene present OSA or CSA depending on a phenotype with predominant muscular or nervous system involvement. Another possible explanation of this conflicting findings is that, at least in part, some of the observed respiratory events could be classified as “pseudo-central” [30]. These events appear mainly during REM sleep characterized by a reduction of the oro-nasal flow due the diminished intercostal muscle activity and, in turn, to a reduction of the excursion of the rib cage. However, OSA has been described in association with other phenotypes of MDs, such as Leigh syndrome (LS) [17••, 31‐33], Kearns–Sayre syndrome (KSS) [34], and neuropathy, ataxia, and retinitis pigmentosa (NARP) syndrome [35]. In conclusion, literature data document that OSA is a frequent sleep disorder in MDs, especially in phenotypes with predominant skeletal muscle involvement.

As aforementioned, sleep represents a state of vulnerability for patients affected by neuromuscular disorders, including MDs characterized by different degrees of involvement of the respiratory muscles. For this reason, specific subgroups of MDs have an increased risk of developing sleep-related hypoventilation/hypoxemia [34, 36]. As consequence of muscular atonia observed during REM, hypoventilation/hypoxemia usually first appears during this sleep stage [16, 19••] and, subsequently, progresses towards NREM sleep. This is mainly due to two mechanisms: the progression of the muscular weakness due to the underline disease and the decreased ventilatory drive due to a diminished sensitivity of chemoreceptor to chronic hypercarbia. In particular, the metabolic derangement, as observed in MDs, could concur to lower the chemosensitivity to hypoxia and hypercarbia. In fact, a depressed ventilatory drive response to hypoxia/hypercapnia has been described in patients affected by PEO [16, 37, 38] and mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (MELAS) syndrome [39]. In particular, sleep-related hypoventilation has been associated with primary mitochondrial dysfunction with a prominent involvement of CNS [32, 34, 35, 39‐42], suggesting that also a dysfunction of the brainstem respiratory center participates to the development of the sleep-related respiratory disturbance. In light of these considerations, it is not surprising that MD phenotypes with predominant CNS involvement, such as LHON “plus” [41] and LS [32], have been associated with the central hypoventilation syndrome, a peculiar life-threatening sleep disorder characterized by ineffective breathing till respiratory arrest during the sleep.

Therefore, it is possible to hypothesize that hypoventilation observed during sleep in MDs recognizes a combined etiology involving CNS dysfunction and muscular weakness [32, 43]. Moreover, sleep-related hypoventilation/hypoxemia have been described also in mitochondrial pediatric population [17••, 33], underlining that these disorders could manifest even in the early stages of the disease.

As previously mentioned, CSA has also been described in association with MDs [16, 20••, 33, 35, 42, 44]. This specific subgroup of SDB is characterized by air-flow cessation in consequence of diminished or absent respiratory effort, due to the lack of drive to breath. CSA can be associated with wakefulness hypercapnia (hypercapnic CSA) or normocapnia (non-hypercapnic CSA) [45]. Hypercapnic CSA is the result of an impaired central drive due to a dysfunction of brainstem respiratory centers (e.g., CNS diseases, specific medications) or an impaired respiratory efference (e.g., muscular disorders, motor neuron diseases). Conversely, non-hypercapnic CSA is idiopathic or frequently related to congestive heart failure. In the latter case, a peculiar pattern of crescendo/decrescendo ventilatory pattern is observed, mostly during NREM sleep, and defined as Cheyne–Stokes breathing (CSB). MDs encompass a large variety of phenotypes with different grades of muscular, neurological, and cardiac involvement. On these bases, it is reasonable to presume that they are widely associated to CSA, and therefore, the clinical suspicion of disordered breathing of central origin should be raised in those patients with genetic mitochondrial dysfunction who manifest cardiomyopathy, as observed in a LHON patient with a severe cardiac involvement [21••] or the involvement of the CNS. In fact, most of the cases described in literature of CSA in MDs are related to clinical phenotypes such as LS [33, 46], NARP [35], and KSS [34]. CSA and a diminished ventilatory response to inhaled CO2 have been described by Manni and colleagues [16] in a cohort of patients affected by PEO, without molecular characterization but defined as “ophthalmoplegia plus” for the involvement of the central and/or peripheral nervous system. Similarly, Smits et al. [20••] revealed high prevalence of CSA (4/20; 20%) in PEO patients associated with POLG mutations and ANS phenotype. Although the absence of genetic data in the article of Manni et al. does not allow us to draw definitive conclusions, it is possible to speculate that a combined mechanism of impaired CNS control of respiration and respiratory muscle weakness concur to determinate sleep-disordered breathing of central origin.

Subjective nocturnal sleep dysfunction, evaluated by the Pittsburgh sleep quality index (PSQI), has been reported in about 75% of patients with MDs [20••] and in 70% of patients with mitochondrial optic neuropathies [21••]. Similarly, EDS, evaluated by means of Epworth sleepiness scale (ESS), has been described in adult patients affected by primary mitochondrial disorders with a prevalence ranging from 27 [14] to 33% [18] and up to 66% in pediatric population [17••]. In both studies, the prevalence of EDS appears significantly higher than in general population. EDS appears to be prevalent (4/36, 11.1%) also in mitochondrial optic neuropathies, regardless of the presence of ocular symptoms [21••]. A single study by Guilleminault and colleagues [47] objectively evaluated EDS by means of multiple sleep latency test (MSLT) in patients with neuromuscular disorders, including two patients with MDs. Both subjects presented subjective EDS (ESS > 10) objectively confirmed by a mean sleep latency of about 8 min at MSLT. Interestingly, after correcting the underlying sleep respiratory disorder, these patients presented a normalization of both subjective and objective EDS. Other experiences indicate that a treatment of an underlying sleep disorder could ameliorate daytime symptoms correlated to a chronic sleep deprivation promoted by sleep disruption [31, 35, 43]. Conversely, subjective sleep complaints were not present in a study group of patients with concomitant sleep apnea and REM-related hypoventilation [16]. Moreover, in large cross-sectional studies the prevalence of perceived sleep dysfunction [20••] and EDS [17••, 20••] is higher than the prevalence of a concomitant sleep disorder or abnormal findings on polysomnography. Similarly, in LHON and DOA [21••], no significant correlation was observed between polysomnographic parameters and poor subjective sleep quality. Therefore, it is difficult to establish whereby EDS and subjective sleep dysfunction are a reflection of underlying sleep disorder or a direct manifestation of MDs.

Other sleep disorders, in addition to those belonging to the SDB category, have been rarely described in association with MDs. Sleep-onset and maintenance insomnia have been reported in a high prevalence of patients (15/36, 41.7%) with mitochondrial optic neuropathies [21••]. Smits et al. [20••] reported a high prevalence (7/20, 35%) of restless legs syndrome (RLS) in PEO patients: two carrying POLG mutations, whereas the other five are associated with single mtDNA deletion or other mutations. Surprisingly, the presence of RLS was not associated to increased sleep latency neither to worst subjective sleep quality. Regarding polysomnographic findings, they found a high prevalence of periodic limb movements (PLM) in their population (mean PLM index 25.5 events/h), and in nine cases (45%), the PLM index was higher than 15 events/h which is considered the pathological cut-off [48]. PLM were more common in those patients who complained poor subjective nocturnal sleep quality, suggesting a high prevalence of periodic limb movement disorder (PLMD) in their cohort of patients affected by genetic mitochondrial dysfunction. RLS and nocturnal leg cramps seems to be prevalent (4/36, 11.1%) in patients affected by LHON and DOA, while PLM detected on PSG were present in few patients of the same cohort (2/36, 5.6%) [21••]. A case of RLS in a patient affected by PEO with a POLG pathological variants has been reported by Aitken et al. [49]. In this case, the patient presented an asymmetric uptake of tracer in the putamen at DaTscan, suggesting a dysfunction of the dopaminergic system, as usually observed in Parkinson disease and idiopathic RLS [50]. Interestingly, Haschka et al. [51] reported an association between RLS and the mitochondrial iron deficiency in peripheral monocytes, suggesting that mitochondrial dysfunction can concur to aggravate RLS symptoms. Increased PLM index (pediatric cut-off > 5 events/h) has been described also by Mosquera et al. [17••] in two pediatric patients, without subjective sleep movement complaints. Finally, a peculiar sleep-related movement disorder defined as excessive fragmentary hypnic myoclonus has been reported by Pincherle and colleges [52]. In this case report, the authors describe a patient with brainstem lesions on MRI, who underwent to V-PSG, presenting sub-continuous and arrhythmic myoclonic jerks occurring during both NREM and REM sleep and associated to sleep-onset insomnia, reduced sleep efficiency, and increased wake after sleep onset (WASO). Finally, a delayed sleep–wake phase disorder, a circadian rhythm sleep disorder (CRSD), has been described in a family of diabetes mellitus associated with m.3243A>G mutation [53]. Interestingly, the circadian rhythm disorder dramatically improved after the administration of coenzyme Q10, suggesting that circadian rhythm disorder can be a direct manifestation of MD, as indicated by recent evidence of a crosstalk between the mitochondria and the circadian clock [54]. Conversely, no CRSD have been observed in a large cohort of patients with mitochondrial optic neuropathies [21••], reinforcing the hypothesis that the retinohypothalamic tract, essential for light-dependent regulation of the circadian rhythm, is sufficiently preserved in these pathologies.

SDB are a frequent comorbidity of MDs, and an early recognition is crucial in order to prevent further clinical deterioration of these fragile patients. Data from the literature support that mechanical ventilation is effective to improve nocturnal breathing, to restore a normal sleep architecture [35] and to ameliorate daytime symptoms [31, 47]. Particularly, in cases associated with severe central nocturnal hypoventilation [32, 41, 44, 46], treatment is lifesaving, because this condition can lead to sudden death during sleep. An early recognition and treatment could reduce the risk of further respiratory deterioration and the need for invasive mechanical ventilation or tracheotomy [32, 34, 35, 41, 46]. Moreover, a pharmacological treatment has been proposed [39, 40] to treat hypoventilation in these patients with drugs that stimulate hypoxic ventilatory response (e.g., aminophylline, theophylline, almitrine) with inconsistent results. The treatment of the underlying metabolic disorder associated with MD could also concur to ameliorate the SDB [31, 42, 43]. For pediatric population, adenotonsillectomy has been proposed for the treatment of OSA, in the absence of sufficient data to support this approach [15]. Therefore, a combined treatment with ventilatory support and pharmacological therapy appears desirable to treat SDB in MDs.