Research, Development, and Application of Intelligent Diagnostic System for Orthostatic Hypotension

Research, Development, and Application of Intelligent Diagnostic System for Orthostatic Hypotension

Orthostatic hypotension (OH) has a high incidence rate of 30%-50% in the elderly and populations with neurodegenerative diseases. The resulting cerebral hypoperfusion significantly increases the risk of cerebral ischemia, falls, and cognitive decline. Traditional OH diagnosis primarily relies on intermittent cuff blood pressure measurements, leading to low detection rates and an inability to provide scientifically effective OH classification. Furthermore, existing research often overlooks cerebral hemodynamic mechanisms, particularly the assessment of dynamic cerebral autoregulation (dCA), making it difficult to study the mechanisms behind OH and its associated symptoms.

To address these issues, the research team has preliminarily developed an "Intelligent Diagnostic System for Orthostatic Hypotension". This system innovatively integrates synchronous and continuous monitoring of multiple parameters, including non-invasive beat-to-beat blood pressure, transcranial Doppler (TCD) cerebral blood flow velocity, and electrocardiogram (ECG). It also enables the quantitative assessment of dynamic cerebral autoregulation function. The project will collaborate with fifteen high-level clinical centers in China to collect data from 2000 patients with orthostatic hypotension. The aim is to establish and externally validate a risk stratification model for OH. By integrating multimodal clinical and hemodynamic data, the investigators intend to construct an automated, precise intelligent system for the classification, subtyping, and risk stratification of OH. This initiative will establish a standardized diagnostic and management pathway covering early screening, precise classification, early warning, and stratified intervention. The goal is to provide key technological support for enhancing the early identification and standardized management of OH, thereby reducing its associated disability and mortality rates.

This prospective, multicenter, observational cohort study aims to develop and validate an intelligent diagnostic and risk stratification system for orthostatic hypotension (OH). The study plans to enroll approximately 2000 participants from 15 tertiary clinical centers in China between March 2026 and February 2029. The target population comprises adult patients (≥18 years) with Parkinson's disease (PD) or multiple system atrophy (MSA), and patients aged ≥50 years with diabetes mellitus who are suspected or diagnosed with OH. A key technical inclusion criterion is the presence of adequate bilateral temporal bone windows for reliable transcranial Doppler (TCD) monitoring.

The core methodology involves synchronous, continuous, and non-invasive monitoring of beat-to-beat blood pressure (BP), bilateral cerebral blood flow velocity (CBFv) in the middle cerebral arteries, electrocardiogram (ECG), and end-tidal carbon dioxide (PetCO₂) during a standardized active standing test. Following a 10-minute supine rest, participants rapidly stand and remain upright for up to 10 minutes. Using this integrated data stream, OH is classified as Initial, Classic, or Delayed per consensus hemodynamic thresholds. Dynamic cerebral autoregulation (dCA) is quantitatively assessed offline via transfer function analysis (TFA) of the BP and CBFv signals, deriving phase, gain (absolute and normalized), and coherence parameters in very low frequency (VLF) and low frequency (LF) bands.

Participants are followed for 24 months, with a telephone follow-up at 12 months and an in-person visit at 24 months that includes a repeat stand test and cognitive assessment. The primary technical endpoints are the algorithm-based classification of OH subtype/etiology and the quantitative dCA parameters. Secondary endpoints include the performance (sensitivity, specificity, area under the curve [AUC]) of the derived multimodal risk model in predicting clinical events such as falls, syncope, cognitive decline, and all-cause mortality.

Data analysis will involve machine learning/statistical modeling on a development cohort to generate the risk stratification model, followed by external validation on a separate cohort to assess generalizability and clinical utility.

Inclusion Criteria:

1. Adult patients (≥18 years old).
2. Clinical diagnosis of Parkinson's disease (PD) OR multiple system atrophy (MSA) OR diabetes mellitus (if diabetic, must be aged ≥50 years).
3. Suspected or diagnosed with orthostatic hypotension (OH).
4. Presence of adequate acoustic temporal bone windows for Transcranial Doppler (TCD) monitoring.
5. Willing and able to provide informed consent.

Exclusion Criteria:

1. Significant intracranial or extracranial arterial stenosis (≥70% confirmed by ultrasound).
2. Recent stroke or intracerebral hemorrhage (confirmed by CT/MRI).
3. Severe cardiac arrhythmias (e.g., atrial fibrillation) or severe valvular heart disease.
4. Bilateral temporal bone windows insufficient for TCD monitoring.
5. Pregnancy or lactation.
6. Inability to cooperate with the testing procedures.
7. Other systemic diseases that significantly affect cerebral blood flow regulation (e.g., severe thyroid or renal dysfunction).