Obstructive Sleep Apnea and Biomarkers of Alzheimer's Disease

Randomized, parallel group, non-blinded, controlled clinical trial compared with conventional treatment.

1. Protocol and intervention Patients with an Apnea-Hypopnea index (AHI)>30 h-1 will be assigned, using a 1:1 randomization table, to lifestyle recommendations treatment or to lifestyle recommendations plus nasal CPAP, for a period of 4 months. CPAP pressure will be titled with automatic using an AutoSet II device, ResMed.
2. Sample size For the estimation of the sample size, previous data were used (PLoS ONE 2019;14(9): e0221255). In this case, in order to compare the effect of CPAP it would be necessary to randomize a total of 100 patients with OSA and 30 control subjects.
3. Ethical considerations

• Indication of CPAP treatment for the prevention of Alzheimer's disease is not yet accepted.
• Those patients with an urgent study indication for the diagnosis and treatment of respiratory sleep disorders (professional drivers, respiratory failure or risk professions) will be excluded from the project. In the other cases, the delay in healthcare for the performance of Polysomnography and CPAP titration exceeds the duration of the study, therefore that patients assigned to the control arm (conventional treatment) will not be exposed to a higher risk than the general population.

1. Methods.

. Polysomnography. It will be used as a screening test previous the randomization of the included patients. The exploration and its interpretation will be carried out following the recommendations of the American Academy of Sleep Medicine. Based on the results of this test, only those patients with an AHI will continue in the study when AHI >30 h-1. In these patients the determinations listed below will be carried out immediately before randomization (visit 1), at 4 weeks (visit 2), 16 weeks (visit 4) of allocation to the corresponding treatment group

• Clinical data: The medication used and other associated diseases already diagnosed will be recorded.
• Vital signs: recording of blood pressure, by means of a conventional triple taking, and heart rate.
• Anthropometric characteristics: age and weight in basal conditions, the perimeter of the neck and waist / hip index will be measured. The lean mass index will be determined using a Bodystat 1500 impedance system (Bodystat Ltd, United Kingdom).
• Questionnaires: Epworth Sleepiness Scale, generic health-related quality of life (SF-12 and EuroQoL) and daily physical activity (International Physical Activity Questionnaire or iPAQ).
• Laboratory findings: blood count, coagulation, creatinine, serum sodium and potassium; glycated hemoglobin (HbA1c) and blood glucose and insulin levels in fasting, from these values will determine the resistance and sensitivity to insulin using the homeostatic model assessment (HOMA) and Quantitative insulin sensitivity check index (QUICKI) indices; total cholesterol, HDL-cholesterol, LDL-cholesterol and triglycerides; and troponin I, homocysteine, N-terminal pro-brain natriuretic peptide (NT-pro BNP) and highly sensitive C-reactive protein.
• Blood venous samples. 20 mL of venous blood will be collected, storing the plasma obtained at -80 ºC for subsequent determination of β amyloid biomarkers (Aβ40, Aβ42, tau y tau 181 (P-tau 181).
• Night pulse oximetry. To assess the existence of residual nocturnal hypoxemia.
• In the patients assigned to the CPAP treatment arm, a count of the hours of use of the CPAP will be carried out at each visit by directly reading the automatic counter of each equipment.

Statistic analysis The data will be expressed as mean ± standard deviation, median (interquartile range) or percentage, depending on their type and distribution. For comparison between groups Student's t-test, the U-Mann-Whitney or the chi-square test will be used, as appropriate. The relations between variables will be analyzed using Pearson's correlation and multiple linear regression analysis. The effect of treatment will be evaluated using general linear models and repeated measures analysis of variance, with multiple comparisons post-hoc using the Bonferroni test. A multiple logistic regression model will be applied to determine the related variables with a response to treatment. Values of p <0.05 will be considered statistically significant. The statistical study will be performed with the SPSS program version 22.0.