The main result of this study was the impairment of BR responses in patients with OSA compared to controls. Our results showed in these patients a prolonged latency of R1 response after stimulation of both supraorbital nerves and a delayed ipsilateral and contralateral latency of R2 response after stimulation of the right supraorbital nerve. The alteration of BR reported in this study may be considered as the electrophysiological evidence of brainstem dysfunction associated with the OSA condition. Although our findings demonstrated a BR response impairment in patients with moderate-severe OSA, the mechanisms at the basis of these results can only be hypothesised. Delayed BR responses may, in fact, reflect an altered brainstem neurotransmission due to the impaired myelination of cranial nerves and dysregulation of brainstem synaptic pathways, both related to hypoxia and sleep impairment [21]. Considering that R1 is thought to represent the disynaptic reflex connecting the main trigeminal sensory nucleus to the ipsilateral facial nucleus, the delayed R1 response recorded in patients with OSA seems to suggest the involvement of synaptic relay stations sited in the mid pons and lower pontine tegmentum. The lesser implication of R2 than R1 responses may indicate a greater preservation of the multisynaptic pathway linking the spinal tract of the trigeminal nucleus to the facial nucleus, as specifically investigated by the R2 response, which could be interpreted as a minor involvement of medulla oblongata structures in OSA. Therefore, the BR pattern abnormalities found in patients with moderate-severe OSA allow us to hypothesise that this sleep disorder may be associated with impairment in structures mainly present in the pons, which is a crucial region for the sleep–wake cycle regulation and is also a region that holds the main nuclei involved in the sleep and respiratory control, namely, the Medial parabrachial/Kolliker Fuse, Lateral parabrachial and GABAergic and glutamatergic ventral pons regions [22,23,24]. Conversely, the mainly respiratory nuclei located in the medulla oblongata, comprising the ventral respiratory neurons associated with the nucleus ambiguous and the dorsal respiratory neurons (DRN) associated with the nucleus tractus solitarius [25], seem to be less affected in OSA due to the partial preservation of the conduction of R2 responses. On these bases, it could be speculated that pontine structures are more vulnerable to a possible OSA-related intermittent nocturnal hypoxemia effect than the medulla centres, with a further dysfunction of pontine reticular pathways in patients with OSA.

Few studies have investigated the effects of OSA on BR responses with mixed results [26,27,28,29]. In line with the present findings, the study by Urban et al. (1996), although not showing significant changes in electrophysiological brainstem responses in a group of 18 patients with OSA, documented abnormalities in the left BR R1 component in one patient, suggesting a left pontine lesion [10]. However, more recently, Tavsanli and colleagues found no significant differences in BR response latencies between patients with OSA and controls [30]. Considering the paucity of studies investigating the BR responses in patients with OSA, further studies are needed for reaching a unequivocal conclusion, that currently cannot be postulated. However, the present data, coupled with previous studies based on evoked potentials and peripheral nerve conduction recordings, may suggest the impairment in electrophysiological conduction in patients with OSA, and that it can be possibly attributed to a hypoxic mechanism inducing myelin damage and axonal alteration [31,32,33,34,35,36,37]. In particular, visual and auditory evoked potentials studies showed prolonged latency and reduced amplitude of the evoked responses [36, 37], hypothesizing a myelin damage due to intermittent hypoxia and neuro-inflammation based on delayed latency, and the occurrence of microvascular events to explain the reduced amplitude.

Considering that CPAP is the gold standard treatment for patients with moderate-severe OSA [38, 39], the present study evaluated the effects of beneficial CPAP therapy on the BR documenting a partial improvement of responses with the reduction of the delayed ipsilateral and contralateral R2 latency after the right supraorbital nerve stimulation. No other significant modifications of the BR responses were observed, in particular in the R1 latencies. These data, showing the persistent alteration of the R1 responses despite the effective CPAP therapy, seem to confirm the greater vulnerability of the pontine structures compared to those of the medulla towards a possible hypoxic effect, which remains stable although the significant improvement of the AHI following CPAP treatment. Nonetheless, it is important to note that the oxygen saturation parameters were evaluated exclusively at the polygraphic recording performed at baseline, while at the 6-month follow-up, the AHI was monitored through the software report. Therefore, although the CPAP effect on the AHI, it is not possible to disentangle whether the lack of improvement in the R1 response can be related to the susceptibility of the pontine structure to the chronic suboptimal oxygen levels. To our knowledge, no previous study had explored the effects of CPAP treatment on BR responses in OSA, and thus, the comparison of the present results with previous investigations cannot be performed. However, a study performed on severe OSA and evaluating the effect of CPAP therapy on the auditory evoked potentials did not document an improvement of electrophysiological responses after treatment, suggesting an un-modifiable effect of OSA on the pontine regions [26] and not replicating previous results [29]. Taking into account the whole literature about the effects of OSA on the electrophysiological tests and the possible beneficial effect of CPAP treatment, there are studies documenting the significant improvement of optic nerve function and neuropsychological event-related potentials in patients with OSA treated by CPAP, thus highlighting the importance of performing further studies with a larger group of patients to test this effect [26,27,28,29].

This study presents some limitations that need to be addressed. The diagnosis of OSA was performed by using polygraphic cardiorespiratory monitoring, which did not allow the evaluation of sleep structure. Furthermore, the inclusion and exclusion criteria were restricted, which allowed the inclusion of a homogeneous sample of patients not affected by comorbidities, however on the other hand limited the sample size, possibly affecting the BR statistical power.

In conclusion, considering the detrimental effects of OSA on brain functions, as reported by previous electrophysiological studies, this BR study demonstrated that OSA can impair brainstem functioning, mostly involving the mid-pons region. Hence, the present study underlines the importance of early recognition of OSA and thus starting CPAP therapy in order to preserve and recover brain and brainstem network functions. The application of electrophysiologic tools exploring brainstem functions, such as BR, may be proposed for the assessment of OSA as an instrument that may improve the comprehension of pathophysiologic mechanisms and clinical outcomes of this common sleep disorder.