Participants
We enrolled 296 community-dwelling probable AD from the visitors to the Dementia Clinic of the Seoul National University Bundang Hospital (SNUBH) from 2011 to 2020. Among them, 104 underwent a 18F-florbetaben amyloid brain positron emission tomography (PET) scan, and 93 were found to be amyloid beta (Aβ)-positive.
We excluded the participants with following conditions: possible or probable DLB or Parkinson’s disease dementia (PDD); any major psychiatric and/or neurological disorders that could affect cognitive function other than AD; any history of brain tumors, substance abuse or dependence, and use of medications such as clonazepam or exogenous melatonin that may influence RBD symptom; any serious medical conditions that could affect the structure and/or function of the pineal gland or abnormalities in pineal gland morphology such as neoplastic lesions or extremely large cystic gland (diameter greater than 15.0 mm) (26); and those with high risk of restless legs syndrome (positive on Cambridge-Hopkins Restless Legs Syndrome questionnaire) (27) and obstructive sleep apnea (STOP-BANG questionnaire score of ≥ 5 points) (28), all of which could mimic symptoms of RBD(29, 30).
All participants were fully informed with the protocol of this study, and provided written informed consents signed by themselves or their legal guardians. This study was approved by the Institutional Review Board of the SNUBH.
Diagnostic assessments
Geriatric psychiatrists with expertise in dementia research conducted in person standardized diagnostic interviews, detailed medical histories, and physical/neurological examinations using the Korean version of the Consortium to Establish a Registry for Alzheimer’s Disease Assessment Packet Clinical Assessment Battery (CERAD-K)(31) and the Korean version of the Mini-International Neuropsychiatric Interview(32). Additionally, research neuropsychologists administered the CERAD-K Neuropsychological Assessment Battery(CERAD-K-N)(31, 33), Digit Span Test(34), Frontal Assessment Battery(35), and Geriatric Depression Scale(36).
Trained research nurses collected data on age, sex, years of education, duration of AD (months), intracranial volume (ICV), history of head injury, amount of smoking (packs/day) and alcohol drinking (standard units/week) over the past 12-month period, and use of drugs influencing sleep or motor activity, including cholinesterase inhibitors (donepezil, rivastigmine, and galantamine), antidepressants (selective serotonin reuptake inhibitor, serotonin norepinephrine reuptake inhibitor, and others), carbamazepine, triazolam, zopiclone, quetiapine, clozapine, and sodium oxybate to each participant. We diagnosed dementia according to the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders Text Revision criteria(37). Global severity of dementia was determined according to the Clinical Dementia Rating(38). We determined probable AD according to the National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer’s Disease and Related Disorders Association diagnostic criteria(39). We diagnosed probable or possible DLB and PDD according to the diagnostic criteria proposed by McKeith et al(25), in which the presence of RBD features was ignored in the current study.
Assessment of brain amyloid deposition
We performed 18F-florbetaben amyloid brain PET imaging using a Discovery VCT scanner (General Electric Medical Systems; Milwaukee, WI, USA) in three-dimensional (3D) acquisition mode. The participants were injected with 8.1 mCi (300 MBq) of 18F-florbetaben (Neuraceq) as a slow single intravenous bolus (6 sec/mL) in a total volume of up to 10 mL. After a 90-minute uptake period, we obtained a 20-minute PET images comprising four 5-minute dynamic frames. The determination was based on the visual interpretation of tracer uptake in the gray matter of the following four brain regions: the temporal lobes, frontal lobes, posterior cingulate cortex/precuneus, and parietal lobes. Participants were considered Aβ positive if smaller areas of tracer uptake were equal to or higher than those present in the white matter extending beyond the white matter rim to the outer cortical margin involving the majority of the slices within at least one of the four brain regions (“moderate” Aβ deposition) or a large confluent area of tracer uptake (i.e., signal intensity) was equal to or higher than that present in the white matter extending beyond the white matter rim to the outer cortical margin and involving the entire region including the majority of slices within at least one of the four brain regions (“pronounced” Aβ deposition). Participants were considered Aβ negative if tracer uptake in the gray matter is lower than that in the white matter in all four brain regions (no β-amyloid deposition).
Assessment of rapid eye movement sleep behavior disorder symptoms
We evaluated behavioral features of RBD using the REM Sleep Behavior Disorder Screening Questionnaire (RBDSQ)(40). The RBDSQ is a self-reported screening instrument used to diagnose RBD and comprises 10 items assessing the most prominent clinical features of RBD: items 1 to 4, the frequency and content of dreams and their relationship to nocturnal movements and behaviors; item 5, self-injuries and injuries to the bed partner; item 6, four subsections specifically assessing nocturnal motor behavior (e.g., questions about nocturnal vocalization (6.1), sudden limb movements (6.2), complex movements (6.3), or bedside items that fall down (6.4)); items 7 and 8, nocturnal awakenings; item 9, disturbed sleep in general; and item 10, the presence of any neurological disorder. Each item could be answered as ‘‘yes’’ or ‘‘no.’’ The RBDSQ score ranges from 0 to 13 points, with higher scores indicating more features associated with RBD. We defined probable RBD (pRBD) as having a total score of 5 or higher on the RBDSQ(40). The questionnaire was completed by the participants with the corroboration from their partners.
Assessment of pineal gland volume
We obtained 3D structural T1-weighted spoiled gradient echo magnetic resonance (MR) images using a Philips 3.0 Tesla Achieva scanner (Philips Medical Systems; Eindhoven, the Netherlands) within 3 months of clinical assessments with the following parameters: acquisition voxel size = 1.0 × 0.5 × 0.5 mm; sagittal slice thickness = 1.0 mm; repetition time = 4.61 ms; echo time = 8.15 ms; number of excitations = 1; flip angle = 8°; field of view = 240 × 240 mm; and acquisition matrix size = 175 × 256 × 256 mm in the x-, y-, and z-dimensions. We implemented bias field correction to remove the signal intensity inhomogeneity artifacts of MR images using Statistical Parametric Mapping software (version 12, SPM12; Wellcome Trust Centre for Neuroimaging, London; http://www.fil.ion.ucl.ac.uk/spm). We resliced the MR images into an isotropic voxel size of 1.0 × 1.0 × 1.0 mm3. We measured ICV using FreeSurfer software (version 5.3.0; http://surfer.nmr.mgh.harvard.edu) to adjust for interindividual variabilities in brain volume. We assessed pineal gland volume as described in our previous work(16). In brief, trained researchers constructed a 3D mask of each pineal gland by manually segmenting the pineal gland slice-by-slice on the resliced T1-weighted MR images at 1.0 × 1.0 × 1.0 mm3 using the ITK-SNAP software (version 3.4.0; http://www.itksnap.org). We measured pineal gland volume and pineal cysts volume and estimated the volume of pineal parenchyma (VPP) by subtracting the pineal cysts volume from the pineal gland volume (Figure 1). We defined a pineal cyst as an area of homogenous intensity that was isointense to the cerebrospinal fluid in T1 sequence images with a diameter of 2.0 mm or greater(41).
The intra-rater and inter-rater intraclass correlation coefficient were 0.983 (95% confidence interval [CI] = 0.956–0.993, p < 0.001) and 0.934 (CI = 0.828–0.974, p < 0.001), respectively.
Statistical analyses
We compared the continuous variables using the independent samples t-tests and categorical variables using the chi-squared tests between groups. We compared VPP between the participants with pRBD and those without pRBD using analysis of covariance that adjusted for age, sex, years of education, ICV, head injury, smoking, alcohol drinking, and use of drugs influencing sleep or motor activity as covariates. We examined the association between VPP and the risk of pRBD using binary logistic regression analysis that was adjusted for the same covariates. We examined the diagnostic performance of the VPP for pRBD using the receiver operating characteristic (ROC) analysis. We calculated the optimal cutoff value and area under the curve (AUC) using Youden index maximum (sensitivity + specificity − 1). We examined the association of VPP with RBDSQ total score (RBDSQ-T) and the item-6 score of the RBDSQ (RBDSQ-6) using multiple linear regression model adjusted for the covariates stated above. The RBDSQ item 6 comprises four subitems on the core symptoms of RBD; nocturnal vocalization (6.1), sudden limb movements (6.2), complex movements (6.3) or bedside items that fall down (6.4). We conducted the same analyses only on the Aβ-positive AD patients.
For all analyses, we considered a two-tailed p-value less than 0.05 as statistically significant, and employed Bonferroni corrections to reduce type I error when multiple comparisons were conducted. We performed ROC analyses using MedCalc for Windows version 18.11.3 (MedCalc Software, Mariakerke, Belgium), and all the other statistical analyses using the Statistical Package for the Social Sciences for Windows version 20.0 (International Business Machines Corporation, Armonk, NY).