RU2838790C1 - Method for lethal outcome prediction in patients with covid-19 during first days of hospitalization according to integrated examination - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention relates to medicine, specifically to ultrasonic diagnostics, and can be used to predict the probability of a lethal outcome in patients with Covid-19 during the first days of admission to specialized hospital. Patient’s clinical state and blood lactate dehydrogenase (LDH) are determined. Ultrasonic examination of the thoracic and abdominal organs is performed; the derived data are used to calculate a prognostic index of a probability of the lethal outcome (p) by developed formula. at p ≥ 48% predict a high probability of the lethal outcome.

EFFECT: method enables predicting the lethal outcome in the patients suffering Covid-19 by taking a set of diagnostic measures within the first days of staying in hospital.

1 cl, 1 dwg, 3 ex

Description

Field of technology to which the invention relates

The invention relates to the field of ultrasound diagnostics and can be used to predict the likelihood of a fatal outcome in patients with COVID-19 hospitalized in a specialized hospital, based on an assessment of the clinical status, ultrasound examination of the lungs and pleural cavities, ultrasound assessment of liver hemodynamics, echocardiographic assessment of the right heart, as well as the level of lactate dehydrogenase in the serum of peripheral blood.

State of the art

One of the key features of the clinical course of COVID-19 is its ability to cause multisystem damage to various organs and tissues, which is due to the specific pathogenetic mechanisms of this infection. The SARS-CoV-2 virus uses angiotensin-converting enzyme 2 (ACE2) as a receptor to enter host cells. ACE2 is widely expressed in many tissues and organs of the human body, including the lungs, heart, kidneys, liver, vascular endothelial cells, and even some cells of the nervous system [Petrikov S.S., Ivannikov A.A., Vasilchenko M.K., Esaulenko A.N., Alidzhanova H.G. COVID-19 and the cardiovascular system. Part 1. Pathophysiology, pathomorphology, complications, long-term prognosis. Sklifosovsky Journal "Emergency Medical Care". 2021; 10 (1): 14-26. https://doi.org/10.23934/2223-9022-2021-10-1-14-26]. Broad expression of ACE2 explains the ability of the virus to infect a variety of cell types, resulting in multiple organ damage and systemic disorders. Thus, the clinical picture of COVID-19 may extend beyond purely respiratory symptoms and include involvement of the cardiovascular system, kidneys, liver, central and peripheral nervous system, as well as the hemostatic system. For example, lung involvement can manifest as acute respiratory distress syndrome (ARDS), which requires intensive care and can lead to a long-term decline in respiratory function. Cardiovascular involvement can manifest as myocarditis, arrhythmias, heart failure, and an increased risk of thromboembolic complications. Kidney involvement often leads to acute kidney injury, which aggravates the prognosis of the disease and may require dialysis. Liver dysfunction may manifest as elevated liver enzymes and jaundice, reflecting both direct viral injury and a systemic inflammatory response [Guarienti, F.A., Gonçalves, J.I.B., Gonçalves, J.B., Antônio Costa Xavier, F., Marinowic, D., & Machado, D.C. (2024). COVID-19: a multi-organ perspective. Frontiers in Cellular and Infection Microbiology, 2024; 14, 1425547].

In addition, endothelial dysfunction and hypercoagulation caused by the virus contribute to the development of thrombosis and microangiopathies, which can affect various organs and lead to additional complications [Podzolkov V.I., Volchkova E.V., Tarzimanova A.I., Bragina A.E., Ivannikov A.A., Bykova E.E., Shvedov I.I., Oganesyan K.A., Isaeva A.Yu. Predictors of thromboembolic complications in patients with severe SARS-CoV-2 coronavirus infection. Therapeutic archive. 2023; 95 (11): 907-912. doi: 10.26442/00403660.2023.11.202472]. The body's inflammatory response, known as the "cytokine storm," also plays a significant role in the pathogenesis of multisystem damage in COVID-19.

Ultrasound imaging has the potential to make a significant contribution to predicting adverse outcomes in COVID-19. In the context of the pandemic, the use of ultrasound imaging is particularly relevant due to its availability, safety, and the ability to perform examinations at the patient’s bedside without the need for patient transport. Lung ultrasound has proven to be an effective tool for assessing the extent of lung tissue damage in COVID-19. Characteristic ultrasound features such as multiple B-lines, consolidations, and pleural lines with discontinuous contours correlate with the severity of the disease. Thus, regular lung monitoring with ultrasound can help in early detection of patient deterioration and the need for intensive care. Cardiac ultrasound is also of significant importance given the frequency of cardiovascular complications in COVID-19. Echocardiography (EchoCG) can detect signs of myocarditis, left and right ventricular dysfunction, increased pulmonary artery pressure, and pericardial effusion. These changes may serve as prognostic markers for adverse outcomes, including the development of heart failure and increased risk of mortality. Ultrasound examination of abdominal organs, including the liver and kidneys, may reveal signs of their involvement, such as increased organ size, changes in parenchymal echogenicity, and the presence of free fluid. Involvement of these organs is often associated with severe COVID-19 and may serve as an additional prognostic factor [Varadarajan, V., Shabani, M., Ambale Venkatesh, B., & Lima, J.A. (2021). Role of imaging in diagnosis and management of COVID-19: a multiorgan multimodality imaging review. Frontiers in Medicine, 8, 765975].

In combination, the use of ultrasound diagnostic methods allows for a comprehensive assessment of the patient's condition, identification of multisystem lesions and monitoring of the dynamics of changes. This contributes to a more accurate prediction of the probability of unfavorable outcomes and individualization of treatment approaches. In addition, ultrasound diagnostics is a safe method without ionizing radiation, which is especially important when repeated studies are required.

A method for predicting the probability of death in hospitalized COVID-19 patients based on laboratory and instrumental studies is known [Patent RU 2806726 C1]. Risk factors obtained during hospitalization are determined. The ECG reveals the presence of: extrasystole, intraventricular blocks, atrioventricular block, atrial fibrillation, intraventricular conduction, pacemaker rhythm. Based on the results of chest computed tomography, the following is determined: 0 points - if there is no lesion, 1 point - lesion up to 25% of the lung volume, 2 points - lesion from 26 to 50% of the lung volume, 3 points - lesion from 51 to 75% of the lung volume, 4 points - lesion more than 76% of the lung volume. The pulse oximetry results are assessed: 0 points - above 95, 1 point - from 90 to 94, 2 points - from 75 to 89, 3 - below 75. The blood test determines the level of proBNP to IL, glucose, alkaline phosphatase, ESR, potassium, troponin and calcium, the number of leukocytes, neutrophils and monocytes. The patient's gender and age are determined. The probability of predicting the result (PPR) is calculated using the stated formula. If the PPR value is greater than 0.47, a high probability of death is predicted.

The disadvantages of this method include the large amount of data required to apply the method, including electrocardiography data, CT scan results, blood tests, and oxygen saturation values. In conditions of limited time and resources, this may make it difficult to use the method in real clinical practice.

A method for predicting a fatal outcome in SARS-CoV-2-associated pneumonia is known [Patent RU 2825705 C1]. The blood ferritin level on the day of hospitalization, the neutrophil coefficient, reflecting the ratio of band and segmented neutrophils, and blood oxygen saturation during the patient's stay in the anesthesiology and intensive care unit are determined. Taking into account the obtained data, the probability of a fatal outcome (p) is calculated using a given formula, and if the p value is more than 50%, a high probability of a fatal outcome is predicted.

The disadvantages of this method include limited flexibility, as it was developed for use in the anesthesiology and intensive care unit (ARI) and focuses on specific parameters, such as neutrophil coefficient and ferritin level, which makes it less universal. When using the method in other conditions, for example, when monitoring patients with mild to moderate COVID-19, the prognostic reliability of the result obtained may decrease. In addition, some of the parameters used, such as neutrophil coefficient (ratio of band and segmented neutrophils), are not standard in clinical practice and may be difficult for medical personnel to interpret. This may hinder the widespread use of the method, as it requires additional training or standardization of laboratory equipment.

A method for predicting the risk of death at the hospital stage in patients with COVID-19 is also known [Patent RU 2805928 C1]. The method includes determining the blood oxygen saturation in the air during the first two hours from the moment the patient is admitted to hospital and the glomerular filtration rate according to the CKD-EPI formula on the first day of hospitalization. The probability of death at the time of hospitalization is calculated using the logistic regression formula P = exp20.26 + (-0.0216 * X1) + (-0.2203 * X2) / 1 + exp20.26 + (-0.0216 * X1) + (-0.2203 * X2), where P is the probability of death during hospitalization for COVID-19; X1 is the glomerular filtration rate according to the CKD-EPI formula on the first day of hospitalization (ml / min / 1.73 m ² ); X2 is the oxygen saturation in air, %, determined during the first two hours of hospitalization using a pulse oximeter; exp is the exponent equal to 2.72.

The disadvantages of this method include the use of only two indicators - oxygen saturation and glomerular filtration rate according to the CKD-EPI formula, while it does not take into account the complex interactions of various factors affecting the prognosis of COVID-19, such as inflammation indicators, liver function and other parameters that can improve the accuracy of the prognosis.

The closest in essence to the claimed invention is a method for predicting the risk of death in patients with varying severity of COVID-19 infection [Patent RU 2820018 C1]. The patient's age, the degree of lung tissue damage according to chest computed tomography data, the ROX index and platelet level are determined. The obtained values of these indicators are substituted into the formula of the developed multivariate logistic regression model. If the P value is greater than 0.5, a high risk of death is predicted. If the P value is less than 0.5, a low risk of death is predicted.

The disadvantage of this method is the use of the ROX index, which is defined as the ratio of oxygen saturation normalized to FiO2 to the respiratory rate, which makes it extremely sensitive to changes in any of these parameters. For example, the slightest change in respiratory rate caused by patient anxiety or the use of sedatives can significantly affect the index value, which does not always correspond to the real severity of the condition. In addition, since the parameters used to calculate the ROX index depend on the respiratory support mode and can change during the day, the ROX index has limited stability. This reduces its reliability as a long-term prognostic indicator, since fluctuations in values can lead to erroneous predictions.

The technical problem solved in the present invention was the creation of a comprehensive, accessible and safe method for predicting the probability of a fatal outcome in patients with COVID-19.

Disclosure of invention

The technical result of the claimed invention is to predict the development of a fatal outcome in patients with COVID-19 during the first day of hospitalization. The advantage of the proposed method is a comprehensive approach to assessing the probability of a fatal outcome in patients with COVID-19 based on ultrasound data, the qSOFA scale and the level of LDH in the blood, which ensures high accuracy of the forecast. This method uses available ultrasound diagnostic methods, allowing for bedside assessment without the risk of radiation exposure. The model takes into account multisystem organ damage, which corresponds to the pathogenesis of COVID-19 and provides high predictive power.

The technical result of the claimed invention is achieved due to the fact that during the first day of hospitalization the patient undergoes a set of diagnostic measures, which include determining the scores on the qSOFA scale, ultrasound examination of the lungs and pleural cavities, ultrasound assessment of liver hemodynamics, echocardiographic assessment of the geometry of the right chambers of the heart, as well as determination of the LDH level in the serum of peripheral blood. Then the obtained results are substituted into the developed logistic regression equation, which has the form: p = 1 / (1 + e ^{- z} ),

where z=-9.235+1.33*X _qSOFA +2.26*X _liver +3.2*X _dilation +1.94*X _B-lines +2.04*X _hydrothorax +

+0.00935*X _LDH ,

p - the probability of a fatal outcome,

e is Euler's constant, approximately equal to 2.71,

X _qSOFA - qSOFA scores,

X _liver - ultrasound signs of liver hemodynamic disturbances (0 - no, 1 - yes),

X _dilation - EchoCG signs of dilation of the right heart chambers (0 - no, 1 - yes),

X _B-lines - B-lines merging on both sides polysegmentally (0 - no, 1 - yes),

X _hydrothorax - ultrasound signs of bilateral hydrothorax (0 - no, 1 - yes),

X _LDH - LDH concentration, U/l.

In the developed multivariate logistic regression model, the following parameters had a statistically significant impact on the prognosis: qSOFA scale; ultrasound signs of liver hemodynamic disturbances; echocardiographic signs of right heart dilation; B-lines merging on both sides polysegmentally; ultrasound signs of bilateral hydrothorax; LDH level.

The set of indicators used in the claimed method is interrelated, providing increased accuracy in predicting the probability of a fatal outcome in patients with COVID-19. Multiple merging B-lines detected during lung ultrasound indicate pronounced interstitial changes in the lung parenchyma, which cover multiple scanning zones. The obtained data are consistent with computed tomography data for severe and critical COVID-19. One of the enzymes whose level was assessed in patients with COVID-19 is LDH, an enzyme involved in the conversion of lactate to pyruvate in the cells of most body tissues and increasing after tissue destruction. Data from many studies indicate that determining LDH activity in serum is an important monitoring parameter for COVID-19 pneumonia, which allows monitoring the progression of the disease and determining timely treatment tactics for such patients. Considering that acute liver failure in patients with COVID-19 may be the result of direct damage to liver cells, impaired liver function can be regarded as a marker of disease progression, so the study of hemodynamic parameters is necessary. The qSOFA scale is an initial way to identify patients with a high risk of an unfavorable outcome of the infection. The development of acute respiratory distress syndrome (ARDS) is characteristic of lung damage in COVID-19, leads to respiratory failure and requires intensive care, and the fact of the presence of bilateral hydrothorax aggravates the impairment of respiratory function. The development of the pathological process in COVID-19 is due to lung damage, against this background, the pressure in the pulmonary circulation increases, which leads to dilation of the right chambers of the heart.

Thus, by measuring the values of the above-described indicators already on the first day of hospitalization, the claimed method allows for the prediction of a fatal outcome in patients with COVID-19.

The claimed method was developed based on the analysis of data from a prospective study of 100 patients who underwent examination and treatment in the infectious diseases department for patients with a new coronavirus infection of the N.V. Sklifosovsky Research Institute for Emergency Medicine of the Moscow Health Department in the period from 2020 to 2022.

Statistical analysis of the obtained data was performed using jamovi software version 3.6.2. At the first stage, the study sample was divided into two groups, depending on the development of a fatal outcome. The next stage was the development of a multivariate logistic regression model, which assessed the probability of a fatal outcome in patients with COVID-19.

To select predictors for the multivariate logistic regression model, the stepwise regression method with the exclusion of variables (backward elimination) based on the Wald criterion was used. The multivariate model included such variables as qSOFA scores, ultrasound signs of liver hemodynamic disturbances, diagnosed right heart dilatation according to echocardiography, diagnosed bilateral hydrothorax, as well as identified polysegmental B-lines merging on both sides, and LDH concentration.

The qSOFA score was assessed based on objective examination data, the condition of the lungs, pleura, pleural cavities, liver hemodynamics and the geometry of the right heart chambers were assessed using ultrasound, and the serum LDH concentration was assessed using a spectrophotometric method based on measuring enzyme activity.

All indicator values were determined on the first day of patients’ hospitalization.

To assess the quality of the binary classification of the developed model, ROC analysis (Receiver operating characteristic) was performed with the calculation of the AUC (area under curve) indicator, and the threshold value of the logistic function p was also determined, exceeding which predicted a high risk of developing a fatal outcome. The sensitivity of the developed multivariate logistic regression model at a given classification threshold was 88.6%, specificity - 90.8%, diagnostic efficiency - 90%. The area under the ROC curve is 0.970 (Fig. 1).

Implementation of the invention

During the first day from the moment of admission to the hospital, patients undergo a number of diagnostic measures, which include determining the qSOFA score, ultrasound examination of the lungs, pleura and pleural cavities, ultrasound assessment of liver hemodynamics, echocardiographic assessment of the size and volume of the right heart chambers. The data obtained from the examination are substituted into the logistic regression equation, thus, the indicators are multiplied by the coefficients calculated during the multivariate regression analysis, as a result of which the probability of a fatal outcome in patients with COVID-19 during hospitalization (up to 7 days) is calculated, expressed as a percentage:

p=1/(1+e ^-z ),

where z=-9.235+1.33*X _qSOFA +2.26*X _liver +3.2*X _dilation +1.94*X _B-lines +2.04*X _hydrothorax +

+0.00935*X _LDH ,

where p is the probability of a fatal outcome,

e is Euler's constant, approximately equal to 2.71,

X _qSOFA - qSOFA scores,

X _liver - ultrasound signs of liver hemodynamic disturbances (0 - no, 1 - yes),

X _dilation - EchoCG signs of dilation of the right heart chambers (0 - no, 1 - yes),

X _B-lines - B-lines merging on both sides polysegmentally (0 - no, 1 - yes),

X _hydrothorax - ultrasound signs of bilateral hydrothorax (0 - no, 1 - yes),

X _LDH - LDH concentration, U/l.

At p ≥ 48%, a high risk of developing a fatal outcome is predicted; at p < 48%, a low risk.

X _qSOFA - qSOFA (Quick Sequential Organ Failure Assessment) score, which is assessed by the presence of such parameters as respiratory rate ≥22/min, systolic blood pressure ≤100 mmHg, changes in mental status (GCS <15). Each feature is assigned 1 point, the total score can vary from 0 to 3.

X _liver - ultrasound signs of liver hemodynamic disturbances, which are assessed based on ultrasound examination data. Signs include portal vein dilation (>13 mm), presence of reverse blood flow in the portal vein, ascites or other markers of portal hypertension: 0 - ultrasound signs of disturbance are absent, 1 - present.

X _dilation - echocardiographic signs of right heart dilation, determined by echocardiography data. Criteria include an increase in the size of the right atrium (>60 ml) and/or right ventricle (end-diastolic size >40 mm). 0 - echocardiographic signs of right heart dilation are absent, 1 - present.

X _B-lines - B-lines merging on both sides polysegmentally, which are assessed during ultrasound examination of the lungs. Polysegmental fusion of B-lines characterizes pulmonary tissue edema. 0 - B-lines merging on both sides polysegmentally are absent, 1 - present.

X _hydrothorax - ultrasound signs of bilateral hydrothorax, which are determined during ultrasound examination of the pleural cavities. The sign is the presence of free fluid on both sides. 0 - ultrasound signs of bilateral hydrothorax are absent, 1 - present.

X _LDH - Lactate dehydrogenase (LDH) concentration, which is measured in the blood serum. LDH levels are expressed in units per liter (U/L). Values above the upper limit of normal (e.g., >250 U/L) indicate tissue damage.

The risk cutoff was established based on ROC (Receiver Operating Characteristic analysis). To evaluate the prognostic model, sensitivity and specificity were calculated at different probability thresholds (p). The optimal threshold (p = 48%) was defined as the point providing the best ratio of sensitivity and specificity of the model, which corresponds to the Youden’s J criterion (maximizing the sum of sensitivity and specificity minus 1).

The essence of the invention is illustrated by the following clinical examples.

Example #1

Patient B., 34 years old. Underlying disease: Coronavirus infection. Bilateral community-acquired polysegmental pneumonia. Condition after mechanical ventilation, resuscitation, central venous catheterization. Complication of the underlying disease: Septic shock. Sepsis. Multiple organ failure. Bilateral spontaneous pneumothorax on 11.04.20. Clinical death on 11.04.20, 12.04.20. Bedsores. Concomitant diseases: Erosive gastritis. Cardiac insufficiency. Erosive mucopurulent tracheobronchitis. Obesity stage 4. Pickwickian syndrome.

The qSOFA score was 3 points, the patient's lung ultrasound did not show bilateral merging B-lines and no bilateral hydrothorax was detected, the liver ultrasound showed hemodynamic disturbances, the echocardiography study showed dilation of the right heart, and the concentration of LDH in the peripheral blood serum was 809.

All parameters were measured during the first day of hospitalization.

When calculating the probability of a fatal outcome for this patient using the formula: p=1/(1+e ^-z ), where z=-9.235+1.33*3+2.26*1+3.2*1+1.94*0+2.04*0+0.00935*809, it was 99.96%, which corresponds to a high risk.

Upon further observation, the prognosis for this patient was confirmed. The patient died during hospitalization.

Example #2

Patient G., 47 years old. Primary diagnosis: Coronavirus infection caused by the COVID-19 virus, the virus has been identified (confirmed by laboratory testing regardless of the severity of clinical signs or symptoms). Complications: Bilateral polysegmental viral pneumonia. ODN. Mechanical ventilation from 26-29.04.20. Temporary tracheostomy installation from 26.03.2020. Limited erosive and ulcerative tracheitis. Minimal bilateral hydrothorax (resolved). Postpuncture hematoma of the soft tissues of the anterior and inner thighs, anterior abdominal wall in the organization stage. Concomitant: Thyroidectomy 2005. Hypothyroidism. Moderate anemia.

The qSOFA score was 1 point, the patient's lung ultrasound showed bilateral merging B-lines and ultrasound signs of bilateral hydrothorax, the liver ultrasound showed no hemodynamic disturbances, echocardiography showed no ultrasound signs of right-sided dilation, and the concentration of LDH in the peripheral blood serum was 352. All parameters were measured on the first day of hospitalization.

When calculating the probability of a fatal outcome for this patient using the formula: p=1/(1+e ^-z ), where z=-9.235+1.33*1+2.26*0+3.2*0+1.94*1+2.04*1+0.00935*352, it was 34.78%, which corresponds to a low risk.

During hospitalization, the patient's condition showed positive dynamics and he was discharged home for outpatient observation.

Example #3

Patient S., 81 years old. Underlying disease: Cerebral infarction with formation of an ischemic focus in the right frontal and left occipital lobes dated 03/14/2020. Cardioembolic pathogenetic variant according to TOAST. Underlying diseases: hypertension stage III, stage 3, risk of CCO4. Ischemic heart disease. Postinfarction cardiosclerosis. Paroxysmal form of atrial fibrillation. CHA2DS2-VASc scale 6 points, HAS-BLED scale 4 points. CHF2A. Complications: hemorrhagic transformation of the ischemic focus. Recurrent aspiration pneumonia localized in the lower sections on both sides, stage of regression. Pressure ulcer of the sacral region is in the resolution stage. Concurrent disease: Coronavirus infection caused by COVID-19 virus, virus identified (confirmed by laboratory testing regardless of severity of clinical signs or symptoms, convalescent. Comorbidities: CKD 4. Mild chronic anemia.

The qSOFA score was 1 point, according to the ultrasound examination of the lungs, the patient did not have bilateral B-lines and ultrasound signs of bilateral hydrothorax, according to the ultrasound examination of the liver, no hemodynamic disturbances were detected, according to the echocardiography examination, ultrasound signs of dilation of the right sections were also not detected, the concentration of LDH in the serum of peripheral blood was 243. All indicators were measured on the first day of hospitalization.

When calculating the probability of a fatal outcome for this patient using the formula: p=1/(1+e ^-z ), z=-9.235+1.33*1+2.26*0+3.2*0+1.94*0+2.04*0+0.00935*243, the P value was 0.36%, which corresponds to a low risk.

During hospitalization, the patient's condition showed positive dynamics and he was discharged home for outpatient observation.

Claims (11)

A method for predicting the probability of a fatal outcome for patients with COVID-19 during the first day of hospitalization, including determining the clinical condition of the patient and the level of lactate dehydrogenase (LDH) in the blood, conducting an ultrasound examination (ultrasound) of the chest and abdominal organs, based on the data obtained, the predictive indicator for the probability of a fatal outcome (p) is determined: p = 1 / (1 + e ^{- z} ), %, where z=-9.235+1.33*X _qSOFA +2.26*X _liver +3.2*X _dilation +1.94*X _B-lines +2.04*X _hydrothorax + +0.00935*X _LDH , e is Euler's constant, equal to 2.71, X _qSOFA - qSOFA scores, X _liver - ultrasound signs of liver hemodynamic disorders, X _liver is assigned "0 points" in the absence of disorders, "1 point" - in the presence of disorders, X _dilation - EchoCG signs of dilation of the right heart chambers, X _dilation is assigned "0 points" in the absence of dilation of the right heart chambers, "1 point" - in the presence of dilation of the right heart chambers, X _B-lines - detection of merging B-lines on both sides polysegmentally on ultrasound of the lungs, X _B-lines are assigned "0 points" in the absence of merging lines, "1 point" - if they are present, X _hydrothorax - ultrasound signs of bilateral hydrothorax, X _hydrothorax is assigned "0 points" in the absence of bilateral hydrothorax, "1 point" - in the presence of bilateral hydrothorax, X _LDH - LDH level in the blood, U/l, at p ≥ 48% a high probability of a fatal outcome is predicted.