2017/10/04 by Claudio Liguori, Agostino Chiaravalloti, Francesca Izzi +5 · 1 citation
Medicine · Neuroscience · Psychology · #Amyloid (mycology) #Computer science #Dementia and Cognitive Impairment Research #Medicine #Neuroscience #Obstructive Sleep Apnea Research #Pathology #Psychology #Risk factor #Sleep (system call) #Sleep and related disorders
paper · pdf · doi:10.1093/brain/awx281
openalex publication_date 2017/10/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04
Sir, Obstructive sleep apnoea (OSA) is a condition with increasing prevalence and frequently diagnosed in middle-age and elderly subjects (Heinzer et al., 2015). OSA has been widely associated with the risk of cognitive impairment (Vaessen et al., 2015), and recently with the pathological alteration of cerebral amyloid-β42 dynamics (Osorio et al., 2014; Ju et al., 2016; Liguori et al., 2017). Therefore, OSA may represent a risk factor for Alzheimer’s disease neurodegeneration (Osorio et al., 2014; Ju et al., 2016; Liguori et al., 2017). However, it is a condition easily treated by continuous positive airway pressure (CPAP), which usually restores brain structure changes and cognitive impairment (Canessa et al., 2011). A 57-year-old male complaining of snoring, subjective sleep impairment, daytime sleepiness, and nocturia was visited at our Sleep Medicine Centre. The patient also reported memory and attention deficits, although he had already performed the neuropsychological tests, the results of which were normal. Therefore, once admitted at our Sleep Medicine Centre he underwent: Polysomnography (PSG), performed as previously reported (Pierantozzi et al., 2016). Briefly, the montage consisted of two electroculographic channels, three electromyographic channels (chin and anterior tibialis muscles) and eight EEG channels (F4, C4, O2, A2, F3, C3, O1, A1). Oronasal flow, thoracic and abdominal movements (plethysmography), pulsoximetry and electrocardiography measured the cardiorespiratory parameters. PSG was scored according to the international standard criteria of the American Academy of Sleep Medicine (Iber et al., 2007). Brain MRI, performed with a 1.5 T superconductive system (OptimaTM MR450w, GE Medical System). Head-coil was used and standard sequences were obtained in axial planes with 3-mm slice thickness. Neuropsychological tests, counting the Mental Deterioration Battery, which is a standardized and validated neuropsychological battery including cognitive tests pertaining to the elaboration of verbal and visuospatial materials (Carlesimo et al., 1996). Lumbar puncture, performed in the decubitus position with an atraumatic needle, between 8:00 and 9:30 am, 2 h after morning awakening. Blood specimens were also obtained at the same time as the lumbar puncture procedure. CSF samples were collected in polypropylene tubes using standard sterile techniques and, immediately after collection, were centrifuged to eliminate cells and cellular debris and immediately frozen at −80°C until analysis. CSF amyloid-β42, amyloid-β40, total tau (t-tau), phosphorylated tau (p-tau) and orexin levels were determined according to previously published standard procedures, using commercially available assays (Innotest β-Amyloid 1-42, Innotest β-Amyloid 1-40, Innotest h-T-tau, Innotest Phospho-T-tau 181; Orexin A/Hypocretin-1 RIA Kit; Phoenix Pharmaceuticals) (Sancesario et al., 2010; Liguori et al., 2014). CSF samples were run in duplicate on the same plate. Moreover, researchers (M.N., S.B.) who were completely blinded to the time of CSF sample (baseline or after 1 year CPAP treatment) performed the analyses. Previously established cut-off values for amyloid-β42, t-tau and p-tau were considered: <500 pg/ml for amyloid-β42, >375 pg/ml for t-tau, and >52 pg/ml for p-tau (Fagan et al., 2006; Mulder et al., 2010; Duits et al., 2014). The ratios t-tau/amyloid-β42 and amyloid-β42/amyloid-β40 were also calculated, and the cut-off of >0.52 and ≤0.06, respectively were considered suggestive of Alzheimer’s disease pathology (Duits et al., 2014; Dorey et al., 2015). Both amyloid-PET and 2-deoxy-2-(18F) fluoro-d-glucose (18F-FDG) PET (18F-FDG-PET). 18F-FDG-PET was performed as previously reported (Liguori et al., 2016). Briefly, 185–210 MBq of 18F-FDG were hydrated with 500 ml of NaCl 0.9% and then injected intravenously with lights off. The patient remained in a resting condition with eyes closed for 30 min prior to PET scan. For amyloid-PET 350 MBq were injected and PET scan started 90 min after injection. A low-ampere head CT was performed before PET image acquisition for attenuation correction (40; 120 Kv). The duration of brain PET scans was set at 15 min. The PET/CT system Discovery VCT (GE Medical Systems) was used to assess 18F-FDG and amyloid brain distribution. All the images were acquired in a 3D-mode standard technique in a 256 × 256 matrix. Reconstruction was performed using the 3D reconstruction method of ordered subsets expectation maximization (OSEM) with 20 subsets and four iterations. Taking all the exams into account, PSG showed the impairment of sleep efficiency coupled with the mild reduction of stage 3 of non-REM and REM sleep (Fig. 1). Notably, the apnoea-hypopnoea index (AHI) was 39.7/h. Brain MRI appeared unremarkable (Fig. 2). Lumbar puncture showed pathological CSF levels of amyloid-β42 and normal CSF concentrations of t-tau, p-tau and orexin (Fig. 1). Notably, at that time both t-tau/amyloid-β42 (0.57) and amyloid-β42/amyloid-β40 (0.03) ratios were suggestive of Alzheimer’s disease pathology. However, amyloid-PET resulted normal and 18F-FDG-PET did not show focal hypometabolisms (Fig. 2). Polysomnographic and CSF data at baseline and after 1-year CPAP treatment. Hypnograms (A) at baseline and (B) after 1-year CPAP treatment. (C) The evolution of CSF biomarkers at baseline and after 1-year CPAP treatment. Principal axis is for CSF levels of orexin, t-tau, p-tau, amyloid-β42; secondary axis is for CSF levels of amyloid-β40. Brain MRI, amyloid-PET and 18F-FDG-PET. (A) The hybrid PET/T1-weighted MRI showing a normal 2-deoxy-2-(18F) fluoro-d-glucose uptake in frontal and parietal regions. (B) The amyloid PET/T1-weighted MRI scan at the same level showing no pathological amyloid deposits in the same cortical regions. (C and D) Axial cuts of the same imaging modalities showing a normal biodistribution of both radiolabelled compounds in the bilateral temporal lobes. (E and F) Axial PET/T1-weighted MRI scan in temporal and fronto-parietal regions. At the end of the diagnostic work-up, the patient started CPAP treatment and was admitted at follow-up. Every 6 months he repeated the neuropsychological tests, which remained normal. At 1-year follow-up, considering the recovery of subjective cognitive deficits and the resolution of daytime sleepiness and subjective sleep impairment, the patient repeated PSG, which showed the improvement of sleep efficiency and continuity and the increase of stage 3 non-REM and REM sleep (Fig. 1) associated with the reduction of AHI to 2.7/h. The patient also repeated lumbar puncture, which documented normal CSF t-tau and p-tau levels, the increase of CSF amyloid-β40 concentrations and, unexpectedly, the recovery to normal CSF concentrations of amyloid-β42 (Fig. 1). Moreover, both t-tau/amyloid-β42 (0.17) and amyloid-β42/amyloid-β40 (0.09) ratios recovered to values not suggestive of Alzheimer’s disease pathology. Finally, CSF orexin levels decreased, although they remained in the normal range (Fig. 1). This emblematic case shows the normalization of cerebral amyloid-β dynamics after CPAP therapy in a patient affected by OSA and subjective cognitive impairment (SCI). Consistently, both the CSF indices suggestive of Alzheimer’s disease pathology (t-tau/amyloid-β42 and amyloid-β42/amyloid-β40) before starting CPAP, recovered to normal values 1 year after CPAP treatment. Both OSA and SCI actually represent conditions considered at risk for the development of Alzheimer’s disease (Osorio et al., 2014; Ju et al., 2016; Liguori et al., 2017; Rabin et al., 2017). On the one hand, OSA has been related to the alteration of cerebral amyloid-β dynamics, since both the reduction of CSF amyloid-β42 levels and the increase of amyloid-β brain deposition have been documented in OSA patients (Osorio et al., 2014; Ju et al., 2016; Liguori et al., 2017). On the other hand, SCI is a condition associated with both CSF amyloid-β42 pathological levels and increased cerebral β-amyloid deposits in crucial brain regions for Alzheimer’s disease pathology (Colijn and Grossberg, 2015). Therefore, it is possible to hypothesize that OSA and SCI may represent preclinical stages of Alzheimer’s disease, in which biomarker changes occur in a very early stage of neurodegeneration. Nowadays, substantial efforts are being spent trying to identify and possibly treat Alzheimer’s disease at the preclinical or mild stages of neurodegeneration. It was recently hypothesized that OSA may be considered as a preclinical Alzheimer’s disease condition in which CPAP treatment may stop or, in an optimistic suggestion, reverse amyloid-β pathology (Liguori et al., 2017). The present report showed the pathological reduction of CSF amyloid-β42 concentrations and the alteration of ratios suggestive of Alzheimer’s disease pathology in a patient affected by severe OSA. This finding is not unexpected since the negative effect of OSA on amyloid-β metabolism and clearance, related to night-time intermittent hypoxia and sleep fragmentation, has already been documented (Osorio et al., 2014; Ju et al., 2016; Liguori et al., 2017). Accordingly, OSA has been associated with reduction of both amyloid-β40 and amyloid-β42 CSF levels and pathological deposition of amyloid-β in relevant brain areas (Ju et al., 2016; Yun et al., 2017). However, it has been supposed that CSF amyloid-β42 levels reduce earlier than amyloid-β deposition, and isolated pathological reduction of CSF amyloid-β42 concentrations may represent the first sign of amyloid pathology in preclinical Alzheimer’s disease (Palmqvist et al., 2016). In the present OSA patient, we documented pathological CSF amyloid-β42 levels without amyloid deposition at PET examination. The discordance between CSF amyloid-β42 measurement and amyloid-PET deposition may indicate a very preclinical stage of Alzheimer’s disease pathology, which can be considered suitable for starting disease-modifying treatment targeting amyloid-β pathology or interventions on modifiable risk factors for Alzheimer’s disease neurodegeneration (Palmqvist et al., 2016). In keeping with this observation, the patient started CPAP treatment, considered the gold standard intervention for OSA, and was admitted at follow-up. One year later, since he reported the resolution of cognitive complaints, a second lumbar puncture was performed to investigate the effect of CPAP treatment on CSF amyloid-β42 levels. CSF analysis showed the unexpected normalization of CSF amyloid-β42 concentrations; moreover, both ratios suggestive of Alzheimer’s disease pathology recovered to normal values. Therefore, we hypothesized that CPAP treatment took the patient back to a healthy condition. CPAP therapy may allow the normalization of CSF amyloid-β42 concentrations, since it recovers sleep fragmentation, and may reverse the glymphatic system dysfunction caused by OSA (Xie et al., 2014; Ju et al., 2016). Moreover, CPAP prevents both the nocturnal involuntary Valsalva manoeuvres related to the recurrent sleep apnoeas (which can reduce the brain glymphatic drainage by increasing veins pressure) and the night-time intermittent hypoxia (demonstrated affecting amyloid-β metabolism and promoting β-secretases activity) re-establishing physiological nocturnal breathing (Shiota et al., 2013; Wostyn et al., 2017). We also observed reduction of CSF orexin levels after 1 year CPAP treatment, although CSF orexin concentrations remained in a normal range. Considering that OSA alters sleep quality and continuity, this finding opens a new possibility, where the orexinergic system may change its regulation in relation to sleep impairment. In keeping with this hypothesis, the reduction of CSF orexin levels after CPAP treatment may be explained by the improvement of sleep continuity and the reduction of the wakefulness after sleep onset. The positive effect of CPAP on cognitive deficits in OSA patients has already been demonstrated (Canessa et al., 2011); however, no reports are present in the literature assessing CSF biomarkers in OSA patients before and after CPAP treatment. Hence, this case report firstly showed the positive effect of CPAP therapy on CSF biomarkers, and in particular in normalizing CSF amyloid-β42 levels, reversing CSF ratios of Alzheimer’s disease pathology and reducing CSF orexin levels. We are aware that the effect of CPAP therapy on the CSF biomarkers in our OSA patient may be explained only by suppositions and needs to be confirmed by clinical trials assessing CSF Alzheimer’s disease biomarkers in OSA patients before and after CPAP treatment. However, the present observation invites further attention and research regarding the triggering effect of OSA on Alzheimer’s disease pathology, since it can actually represent the first treatable risk factor for Alzheimer’s disease neurodegeneration. No funding was received towards this work.