Participants

Participants were recruited from Orange County communities through outreach events, mailing lists, word-of-mouth, online portals, a research volunteer registry, and through the University of California Irvine (UCI) Alzheimer’s Disease Research Center (ADRC), and all procedures were conducted as part of the Vascular Senescence and Cognition (VaSC) Study at UCI. Older adults aged 55 to 89 years who were living independently were included (Table 1). Study exclusions were a prior diagnosis of dementia, history of clinical stroke, family history of dominantly inherited neurodegenerative disorders, current neurological or major psychiatric disorders that may impact cognitive function, history of moderate-to-severe traumatic brain injury, current use of medications impairing the central nervous system, current organ failure or other uncontrolled systemic illness, and contraindications for brain MRI. Eligibility for the study was verified by a structured clinical health interview and review of current medications with the participant and, when available, a knowledgeable informant study partner. All participants underwent neurological and neuropsychological evaluations performed using the Uniform Data Set (UDS), and additional neuropsychological tests, as described in the neuropsychological testing section. This study was approved by the UCI Institutional Review Board, and all participants gave informed consent. The anonymous data that support the findings of this study are available upon reasonable request from the corresponding author, DAN, through appropriate data-sharing protocols.

Table 1 Participant characteristics and demographicsContinuous BP data acquisition

Participants were asked to take medications as normally prescribed and abstain from caffeine on the morning of data collection. Beat-to-beat BP measurements were obtained continuously during supine rest in a 3 T Siemens MRI scanner, using an MRI-compatible non-invasive continuous BP finger cuff device (Biopac®). First, the participant rests for 3 min in the supine position prior to the calibration period. During calibration, BP waveforms are acquired by the continuous monitoring device and 2 static pressures are simultaneously acquired using a calibrated, MRI-compatible automatic BP device with an inflatable brachial artery cuff (TeslaDUO). These static pressures are used to calibrate the continuous BP monitor using the Caretaker® system (Biopac®). After calibration, continuous BP was monitored during MRI for 7 min and further data processing was performed as previously described [23].

Systolic blood pressure average real variability

Beat-to-beat BPV is quantified as systolic blood pressure (SBP) average real variability (ARV), a measure of systolic beat-to-beat BPV which has been demonstrated as reliable in older adults [23, 24] regardless of antihypertensive medication use [23]. ARV calculates the average of absolute changes between consecutive blood pressure readings and is calculated as:

$$\text=\frac \sum_^|_-_|$$

where n represents the number of blood pressure readings obtained during continuous blood pressure monitoring and k represents the beat index of the readings as previously described [23, 25]. Systolic ARV was chosen as the measure of beat-to-beat BPV in the present study due to its increased reliability compared to other BPV measures such as standard deviation and coefficient of variation [23], and its decreased susceptibility to outliers [24]. Additionally, ARV has the advantage of considering the temporal ordering of systolic BP measurements and is therefore a more specific measure of beat-to-beat fluctuations in blood pressure [23, 26].

Vascular risk factors (VRF)

Vascular risk factor (VRF) burden was determined through clinical interviews with the participant and informant (when available), and review of current medications and medical history. The assessed VRFs included a history of cardiovascular disease (e.g., heart failure, angina, stent placement, coronary artery bypass graft, intermittent claudication), hypertension, hyperlipidemia, type 2 diabetes, atrial fibrillation, left ventricular hypertrophy, and transient ischemic attack. Total VRFs were summed for each participant and elevated VRF burden was defined as ≥ 2 VRFs (vs. 0–1) as described previously [27, 28].

APOE genotyping

Fasted blood samples were obtained by venipuncture and used to determine the participant’s APOE genotype. Genomic DNA was extracted using the PureLink Genomic DNA Mini Kit (Thermo). The isolated DNA concentration was determined using a NanoDrop One (Thermo). DNA was then stored at − 80 °C for long-term storage. Isolated DNA was first diluted to a concentration of 10 mg/μL. PCR reactions were performed in a final volume of 25 μL containing 25 ng DNA, 0.5 μM of both forward and reverse primers (forward: ACGGCTGTCCAAGGAGCTG; reverse: CCCCGGCCTGGTACACTG), and 1 × SYBR Green Master Mix (Qiagen) diluted in H2O. For the amplification, a T100 Thermal Cycler (BioRad) was used with the following settings: 95 °C for 10 min; 32 cycles of 94 °C for 20 s, 64 °C for 20 s, and 72 °C for 40 s; followed by 72 °C for 3 min. Fifteen microliters of the DNA PCR product was digested with Hhal-fast enzyme at 37 °C for 15 min. The digested PRC product was added to a 3% agarose gel in 1 × borax buffer for gel electrophoresis. The gel was run at 175 V for 25 min and visualized on ChemiDoc (BioRad) with a GelRed 10,000 × gel dye. APOE4 carrier status was defined as APOE4 carriers (at least one copy of the ε4 allele) or APOE4 non-carriers (no copies of the ε4 allele), as previously described [29]. All analyses were performed at the same lab at the University of Arizona (KER).

Brain volumes

All participants underwent brain MRI scans conducted on a 3 T Siemens Prisma scanner with 20-channel head coil. High-resolution 3D T1-weighted anatomical (Scan parameters: TR = 2300 ms; TE = 2.98 ms; TI = 900 ms; flip angle = 9 deg; FOV = 256 mm; resolution = 1.0 × 1.0 × 1.2 mm3; Scan time = 9 min) images were acquired, using 3-dimensional magnetization-prepared rapid gradient-echo (MPRAGE) sequences.

For region-of-interest (ROI) analysis, post-processing of T1 scans was accomplished in FreeSurfer 7.4.1 [30] using an automated segmentation algorithm that is robust to anatomical variability including ventricular enlargement associated with neurological diseases and aging for quantification of bilateral hippocampal volumes [31]. After automated segmentation, each individual subject was checked for any inaccuracies or misclassifications; manual corrections were made as needed with FreeSurfer’s built-in editing tools, cases were then re-processed, and resulting volumes were used for analyses.

Plasma biomarkers

Blood plasma from fasted blood samples was separated by centrifugation and stored at -80 °C until AD biomarker assays. All plasma Aβ40 and Aβ42 concentrations were obtained using the digital immunoassay, Simoa Neurology 3-Plex A (N3PA) Advantage Kit (Quanterix). Plasma total tau was also obtained but not analyzed due to questions regarding its relationship with brain AD pathological changes [32]. Plasma levels of GFAP and NfL were determined using a single molecule array, (Simoa®) Neurology 2-Plex B (N2PB) Kit (Quanterix), following the manufacturer’s protocol on the HD-X machine. Accepted ranges were as follows: NfL = 0– ~ 2000 pg/mL and GFAP = 0– ~ 40,000 pg/mL. All biomarker assays were conducted in the same lab at UCI (EH).

Neuropsychological testing

All participants underwent a clinical interview and comprehensive neuropsychological assessment by a trained technician or doctoral student under the supervision of a licensed clinical neuropsychologist. The assessment included multiple tests of memory, attention/executive function, and language. All neuropsychological testing and diagnostic assessments were conducted blinded to all clinical, biomarker, and imaging findings. Composite scores were created for memory, attention/executive function, and language. A memory composite score was created by averaging the demographically corrected (age, sex, and education) z-scores from three memory tests which included story memory delayed recall (either Weschler Memory Scale–Revised [WMS-R] Logical Memory-II [33] or Craft story delayed recall [34, 35]), word list delayed recall (Rey Auditory Verbal Learning Test [RAVLT] Trial 7 [36] or CERAD word list 30-min delayed recall [37]), and word list delayed recognition (RAVLT Recognition [36] or CERAD word list 30-min delayed recognition [37]). An attention/executive function composite was created by averaging the demographically corrected z-scores from the attention/executive function tests which included the trail-making test A [38], trail-making test B [38], and one other attention/executive test which included either the Golden Stroop color and word test [39], the D-KEFS color-word interference test [40], or the digit span backward test. Lastly, a language composite score was created by averaging the demographically corrected z-scores of the three language tests which included semantic verbal fluency (Animals) [41], confrontational naming (Boston Naming Test [42] or multilingual naming test [43]), and phonemic verbal fluency (FAS) [41].

Data analysis

One hundred five participants underwent continuous BP monitoring and brain MRI. One participant was excluded after 3 SD outlier screens (+ 4.35 SD left hippocampal volume, + 4.31 SD right hippocampal volume) resulting in a total analyzed sample size of 104 for volumetric analyses. A subset of these participants also had neuropsychological testing characterization (n = 103) available for analysis. Demographically adjusted neuropsychological component scores were screened for outliers and averaged into memory (n = 93), attention/executive (n = 101), and language (n = 101) composite scores as previously described. The relationship between beat-to-beat BPV and demographically adjusted neuropsychological domain composite scores was investigated. A subset of 56 participants with plasma GFAP and NfL assays were also analyzed.

Linear regression models assessed whether beat-to-beat BPV was associated with hippocampal atrophy, neuropsychological domain composite z-scores, and plasma markers of neuroaxonal/neuroglial injury with and without adjustment for age and sex where applicable (volumetric and plasma biomarker analyses), and for average beat-to-beat SBP and VRF burden. Volumetric analyses also included total intracranial volume (TIV) as a covariate. All analyses were performed in R [44].

Additional sensitivity analyses were performed for the significant age, sex, VRF burden, and average beat-to-beat SBP-adjusted findings. These analyses included additional correction for APOE4 carrier status, plasma Aβ42/40, and where appropriate, neuropsychological testing site. Benjamini–Hochberg false discovery rate (FDR) correction [45] was also applied for the primary analyses (L/R hippocampal volumes, plasma GFAP, plasma NfL, and memory composite z-score).