Skip to content

Differential Diagnosis of Persistent COVID-19 by Artificial Intelligence

Differential Diagnosis of Persistent COVID-19 by Artificial Intelligence

Status
UNKNOWN
Phases
NA
Study type
Interventional
Source
ClinicalTrials.gov
Registry ID
NCT05629793
Acronym
DICOPERIA
Enrollment
136
Registered
2022-11-29
Start date
2022-12-14
Completion date
2023-11-30
Last updated
2022-11-29

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Cognitive Dysfunction, COVID-19, COVID-19 Recurrent, Distress Respiratory Syndrome, Fatigue, SARS CoV 2 Infection

Keywords

Machine learning, Stress test, Cardiac variability, Voice recording, Skin conductance

Brief summary

The pandemic caused by SARS-CoV-2 infection has resulted, in addition to the well-known acute symptoms, in the emergence of persistent, diffuse and heterogeneous symptoms referred to as persistent COVID. Common symptoms include fatigue, shortness of breath, and cognitive dysfunction, among others, and result in an impact on daily functioning. Symptoms may be new onset, appear after initial recovery from an acute episode of COVID-19, or persist after the initial illness. Cardiac variability (HRV) was initially used in COVID-19 to predict mortality in the acute setting. Dysautonomia which partly evaluates HRV is frequent in patients with persistent COVID. Several groups have used voice or other respiratory noise analysis for the diagnosis of acute COVID. Patients in the persistent COVID cohort will be able to be differentiated from an age, sex and vaccination status matched cohort of recovered COVID patients without sequelae by means of a model created by Machine Learning that will be trained using cardiac variability (HRV), skin conductance and acoustic analysis data. The primary objetive will be to obtain a classification algorithm by Machine Learning to differentiate the group of patients with persistent COVID diagnosis from the paired group of recovered COVID patients without sequelae.

Detailed description

This is a validation study of a Machine Learning algorithm for the diagnosis of persistent COVID using clinical diagnosis as the gold standard. The sample will be composed of post-COVID patients, one group of which developed persistent COVID and another paired with the previous one with cured COVID without sequelae.

Interventions

OTHERExperimental tests

Walking for 6 minutes, sitting down and getting up from a chair for 1 minute and finally the cold test (Cold pressor) where the hand is introduced for 1 minute in water at 4ºC. The patient will be monitored by means of a Polar H10 chest strap, as used in sports, continuously and 02 saturation, TA and voice (exhalation while saying /a/ and dry cough) will be collected before and after the tests. Finally, skin conductance will be monitored by performing baseline tracing and then control while performing the cold test.

Sponsors

University of Vigo
CollaboratorOTHER
Galician South Health Research Institute
CollaboratorNETWORK
Fundacin Biomedica Galicia Sur
Lead SponsorOTHER

Study design

Allocation
NON_RANDOMIZED
Intervention model
PARALLEL
Primary purpose
TREATMENT
Masking
SINGLE (Investigator)

Masking description

The investigators performing the tests will be blinded to the diagnostic group.

Intervention model description

Persistent COVID group vs. Recovered COVID group

Eligibility

Sex/Gender
ALL
Age
18 Years to 70 Years
Healthy volunteers
No

Inclusion criteria

Persistent COVID group: Inclusion Criteria: * Age ≥18 and ≤70 years of age * Confirmed infection (PCR) with SARS- CoV-2 until 03/28/2022 and thereafter date. * Symptoms include: fatigue, respiratory distress or cognitive dysfunction, among others. * Symptoms persist or appear more than 3 months after onset of infection. * Symptoms last longer than 2 months and are not better explained by another diagnosis. * Symptoms appeared after initial recovery or persisted since disease debut. * Symptoms may fluctuate or remit over time. * Patients have capacity to consent and agree to participate in the study.

Exclusion criteria

* Active COVID-19 infection. * Cardiac arrhythmia, pacemaker carrier. * Other pathologies with dysautonomia. * Raynaud's phenomenon. * Other diseases that may affect exercise capacity or be aggravated by exercise shall also be excluded, such as: Uncontrolled heart failure, severe or symptomatic aortic stenosis, pulmonary edema, acute respiratory failure, recent pulmonary thromboembolism, lower limb thrombosis, infections, thyrotoxicosis, or orthopedic inability to walk. Recovery COVID group Inclusion Criteria: * Age ≥18 and ≤70 years of age * Confirmed infection (PCR) with SARS- CoV-2 until 03/28/2022 and thereafter date. * Full functional recovery. * Follow-up by Primary Care. * They have not presented three months after the onset of the disease: fatigue, respiratory distress or cognitive dysfunction, among others. * Patients have capacity to consent and agree to participate in the study.

Design outcomes

Primary

MeasureTime frameDescription
Differences of the group of patients with a persistent diagnosis of COVID from the age-matched group, sex and vaccination status of patients recovered from COVID without sequelae.8 weeksThrough an algorithm model created by Machine Learning that will be trained using cardic variability (HRV), skin conductance and acoustic analysis data.

Secondary

MeasureTime frameDescription
Voice recording8 weeksSounds produced by the patient at rest and after having performed the stress tests. The patient will be asked to take a deep breath and then pronounce the vowel /a/ in a sustained manner, in a comfortable tone and volume (3 times).
Skin conductance8 weeksmicro Siemens \[µS\]. By means of the Bitalino electrodermal activity recording system. It will be recorded at rest, during the Cold Pressor test and once it is finished, for at least two minutes, to assess the normalization of the conductivity curve.
Cardiac variability8 weeksNumber of times a contraction of the heart occurs in one minute, expressed in beats per minute, by means of a Polar chest strap, model H10. A baseline recording of 5 minutes duration will be taken, with the patient in a seated position. At the end of each test, recording is continued for 2 minutes to demonstrate the speed and degree of recovery after stress.
1minSTST8 weeksNumber of repetitions performed after sitting down and getting up from a chair without supporting the hands as many times as possible for 1 minute..
Cold Pressor test8 weeksOne hand is inserted into a container with water at 4-5ºC for 1 minute. Before and after the test, HRV, BP, and thermal conductance are recorded for 5 and 2 minutes -respectively- while lying supine.
6MWT8 weeksmetres/min. The patient will walk the maximum distance they can in 6 minutes.

Other

MeasureTime frameDescription
FEV18 weeksThe volume of air expelled during the first second of forced expiration in ml.
Age8 weeksYears
CO diffusion test8 weeksTo evaluate the transfer of oxygen from the alveolar space to the hemoglobin of the erythrocytes contained in the pulmonary capillaries. Effective alveolar-capillary area available for gas transfer in the lung. (%)
FEV1/ FVC8 weeksExpressed as a percentage (%), it indicates the proportion of the FVC that is expelled during the first second of the forced expiratory maneuver.
Sex8 weeksMale, Female
Current treatment8 weeksTreatment taken by the patient at the time of the study.
Date of PCR + SARS-CoV-28 weeksDD-MMM-YYYY
Epidemic wave8 weeksOf the 7 waves of COVID-19 that have occurred in Spain, a description will be given of the wave to which the infection of each patient included belonged. First, second, third, fourth, fifth, sixth, seventh, eighth, ninth or tenth wave.
Vaccination status at the time of infection8 weeksNumber of vaccines doses at the time of infection
FVC8 weeksIs the maximum volume of air exhaled, with the maximum possible effort, starting from a maximum inspiration in ml.

Countries

Spain

Contacts

Primary ContactAlejandro García Caballero, MD
alejandro.alberto.garcia.caballero@sergas.es988 38 55 00

Outcome results

None listed

Source: ClinicalTrials.gov · Data processed: Feb 4, 2026