RU2788285C1 - Method for determination of probability of lethal outcome at hospital stage after coronary bypass surgery - Google Patents

Abstract

FIELD: medicine; cardiovascular surgery.

SUBSTANCE: after coronary bypass surgery (hereinafter – CBS), a body mass index (BMI, kg/m²), a glomerular filtration rate (GFR, ml/min/1.73 m²), the total ejection fraction of the left ventricle (TEF LV, %), the presence of critical atherosclerosis damage to main arteries of lower extremities (MALE, yes/now qualitative indicator), and previous CBS (yes/no qualitative indicator) are determined, based on which a probability of the presence of lethal outcome (P) is determined according to the formula. If P is equal to or more than 1.5%, a conclusion is made about a high risk of lethal outcome.

EFFECT: method provides an increase in the accuracy of preoperative prediction of lethal outcome after CBS, taking into account modern region adaptiveness.

1 cl, 2 ex, 3 dwg, 2 tbl

Description

The invention relates to medicine, namely to cardiovascular surgery, and is a method for predicting the risk of death at the hospital stage after coronary bypass surgery.

Ischemic heart disease (CHD) is the cause of death for more than half of all patients who die from diseases of the circulatory system. One of the most effective methods of treating IHD in patients with multivessel coronary artery disease is coronary bypass grafting (CABG), the widespread use of which leads to a significant increase in the duration and quality of life of patients. Mortality after CABG, along with other adverse cardiovascular events, is an indicator of the quality of cardiac surgery.

In clinical practice, physicians regularly face difficulties in making decisions regarding the choice of the optimal method of myocardial revascularization in patients with coronary artery disease, despite the available risk scores (for example, EuroSCORE I, II, STS Score, etc.) (ESC / EACTS Guidelines for myocardial revascularization 2018 Russian Journal of Cardiology 2019; 24 (8)). The reason for the difficulties that arise is caused by the low accuracy of predicting cardiovascular events using the available risk scales, which is due to the lack of regional adaptation of existing scales, incomplete consideration of initial clinical data and risk factors, errors and limitations of mathematical methods for calculating risks, "obsolescence" of clinical databases used to create scales, as a result, they may not reflect modern reality. So, in the work of Shikhverdiev N.N. et al. (2019) the value of the EuroSCORE II scale showed insufficient prognostic significance for assessing the risk of developing 30-day mortality and associated complications (AUC=0.532; p=0.676) (Shikhverdiev N.N., Ushakov D.I., Peleshok A.S. ., Krivopalov V.A., Sizenko V.V. New approaches to risk stratification of elderly cardiac surgery patients. Clinical and experimental surgery. 2019; 7(1): 23-27. doi: 10.24411/2308-1198-2019-11003 ).

Other researchers also state the insufficient effectiveness of the available risk scales (Chicherina E.N., Pinegina Yu.V. Factors affecting the outcome of coronary artery bypass surgery. A systematic review of Russian literature data. Vyatka Medical Bulletin. 2017; 2 (54): 12- 18 etc.). Thus, the creation of a modern effective method for predicting the risk of death after CABG, specific for the conditions of the Russian Federation, is currently a very urgent task.

A known method for predicting the risk of mortality within 30 days after isolated coronary artery bypass grafting (Reid C., Billah B., Dinh D, Smith J, Skillington P., Yii M., Seevanayagam S., Mohajeri M, Shardey G. An Australian risk prediction model for 30-day mortality after isolated coronary artery bypass: the AusSCORE J Thorac Cardiovasc Surg 2009; 138(4):904-10 DOI: 10.1016/j.jtcvs.2009.03.020). Based on the analysis of the results of CABG in 11,823 patients with coronary artery disease, of which 7709 (65.9%) underwent isolated CABG, a method was developed for preoperative prediction of the risk of 30-day mortality, taking into account such predictors as age, functional class of angina pectoris, total ejection fraction of the left ventricle , urgency of surgery, previous open heart surgery, hypercholesterolemia, peripheral vascular disease, and cardiogenic shock. Declared effectiveness of the AusSCORE method: ROC AUC ¹ - 0.834, p=0.241 (Receiver Operating Characteristic Area Under Curve (ROC AUC). Interpretation of the quality results of the ROC AUC model: 0.9-1.0 - excellent; 0.8-0.9 - very good, 0.7-0.8 - good, 0.6-0.7 - average, 0.5-0.6 - unsatisfactory).

The main disadvantage of this method is regional non-adaptation, taking into account differences in health care models and medical punctures used in practice. In addition, it is important to take into account the "obsolescence" of the clinical databases used (2009), and, therefore, the method cannot take into account the evolutionary and technical changes of recent years in relation to approaches to the criteria for selecting patients for surgery, technical support for operations, approaches to drug therapy. accompanying patients with coronary artery disease, both before and after the intervention, and much more.

The technical result of the proposed method is to increase the accuracy of the possibility of preoperative prediction of death after CABG, taking into account modern regional adaptation to healthcare models and medical protocols used in practice).

The calculation of the probability of a patient's death is based on clinical diagnostic characteristics determined by multivariate logistic regression. Predictors that have a statistically significant effect on the development of death after CABG include:

1) body mass index (BMI, kg / m ² ) - a quantitative indicator calculated by the formula: BMI \u003d weight (kg) / height (m ² );

2) glomerular filtration rate (GFR, ml / min / 1.73 m ² ) - a quantitative indicator determined by the MDRD formula (Modification of Diet in Renal Disease Study - the recommended formula for assessing GFR (Levey AS, Greene T., Kusek J ., Beck G. A simplified equation to predict glomerular filtration rate from serum creatinine (abstract) J. Am. Soc. Nephrol. 2000;11: 155A): GFR=186 × (serum creatinine, mg/dL) - 1.154 × (age, years) - 0.203 (for women, the result is multiplied by 0.742, for blacks - by 1.210);

3) total left ventricular ejection fraction (LV FEV, %) - a quantitative indicator determined by the formula: LV FEV \u003d EDV - EDV / EDV × 100%, where EDV - end diastolic LV volume (ml), ESD - end systolic LV volume (ml), by transthoracic echocardiography;

4) the presence of critical atherosclerotic lesions of the main arteries of the lower extremities (MALE) - a qualitative indicator (yes / no), determined on the basis of the detection of atherosclerosis NALE (stenoses ≥ 70%) by triplex ultrasonic dopplerography of vessels (USDG).

5) previous CABG - a qualitative indicator (yes / no), determined according to the anamnesis data (based on a previous CABG operation).

The technical result is achieved by calculating the probability of death after CABG operation according to the formula:

P=1/(1+e- ^z ) × 100%,

where P is the probability of a fatal outcome; e - Euler number - a mathematical constant equal to 2.71828; z=4.517 - 0.102X _BMI - 0.022X _GFR - 0.092X _{LV EF} + 1.494X _{critical NAUT} + 2.226X CV _history , where X _BMI - body mass index (kg/m ² ), X _GFR - glomerular filtration rate ( ml / min / 1.73 m ² ), X _{LV FEV} - total left ventricular ejection fraction (%), X _{critical NAAT lesion} (0 - no critical NAAT lesion (<70%), 1 - presence of critical NAAT lesion (≥70 %)), X _{history of CABG} (0 - no history of CABG, 1 - history of CABG).

The resulting regression model is statistically significant (p<0.001). Based on the value of the Nigelkirk coefficient of determination, the model explains 23.1% of the observed variance in mortality during hospitalization.

Quality metrics: area under the ROC curve was ROC AUC 0.838±0.046 (95% CI 0.747-0.929; p<0.001).

The proposed method was implemented on the basis of clinical data of 2239 patients with stable coronary artery disease who underwent CABG for the period from 2018 to 2021.

The formation of the database for the implementation of the method was based on the criteria for data sampling. Inclusion criteria: stable course of coronary artery disease before surgery; critical atherosclerotic lesion of the coronary arteries, confirmed by the results of coronary angiography and requiring CABG. Exclusion criteria: LV aneurysm resection and/or surgery to reconstruct heart valves (plasty, prosthetics, etc.) and/or simultaneous revascularization of other vascular beds (brachycephalic arteries, arteries of the lower extremities, etc.) simultaneously with myocardial revascularization.

The analyzed target features included preoperative data: history data; risk factors (gender, age, weight, smoking, arterial hypertension, diabetes mellitus, obesity, severity of angina pectoris, history of myocardial infarction, history of myocardial revascularization (CS or PCI), multifocal atherosclerosis, chronic kidney disease); examination results (ECG rhythm, total LV EF according to echocardiography, the presence of critical atherosclerotic lesions of the brachiocephalic arteries and arteries of the lower extremities, damage to the trunk of the left coronary artery, damage to the anterior interventricular artery (ALAD), the number of affected main coronary arteries according to coronary angiography); laboratory data (total cholesterol, low density lipoproteins, creatinine); the level of operational risk of death according to the EuroSCORE II scale (%); volumes and conditions of surgical interventions (carrying out cardiopulmonary bypass (EC), LAD bypass, use of the left internal mammary artery, the number of formed bypasses, EC time) and postoperative outcomes (hospital mortality after CABG).

Quantitative indicators were assessed for compliance with the normal distribution using the Shapiro-Wilk test (if the number of subjects was less than 50) or the Kolmogorov-Smirnov test (if the number of subjects was more than 50). In the absence of a normal distribution, quantitative data were described using the median (Me) and the lower and upper quartiles (Q1-Q3). Categorical data were described with absolute values and percentages. Comparison of two groups in terms of a quantitative indicator, the distribution of which differed from the normal one, was performed using the Mann-Whitney U-test. Comparison of percentages in the analysis of four-field contingency tables was performed using Pearson's chi-square (χ ² ) test (with expected phenomena values of more than 10), comparison of percentages in the analysis of multi-field contingency tables was performed using Pearson's χ ² test. The construction of a predictive model of the probability of a certain outcome was performed using the method of logistic regression. Nigelkirk's R ² coefficient served as a measure of certainty indicating that part of the variance that can be explained by logistic regression. To assess the diagnostic significance of quantitative signs in predicting a certain outcome, the method of analyzing ROC curves was used. The separating value of the quantitative trait at the cut-off point was determined by the highest value of the Youden index.

In the considered cohort of patients (n=2239), hospital mortality was 1.3% (n=30). Using the binary logistic regression method, a predictive model was developed to determine the probability of a fatal outcome during hospitalization, depending on the indicators that demonstrate statistical significance during multivariate regression. Predictors such as BMI, GFR, LV FEV had an inverse relationship with the probability of death. Predictors such as the presence of a critical NAUT lesion according to ultrasound and CABG in the anamnesis were accompanied by an increase in the likelihood of a lethal outcome (Table 1).

Figure 1 presents an estimate of the odds ratio with 95% CI for the studied predictors of death during hospitalization.

When evaluating the dependence of the probability of a fatal outcome on the value of the logistic function P using ROC analysis, the following curve was obtained.

The figure 2 shows the ROC-curve that characterizes the dependence of the probability of death during hospitalization on the value of the logistic function P.

The area under the ROC curve was 0.838±0.046 with a 95% CI of 0.747-0.929. The resulting model was statistically significant (p<0.001).

The figure 3 shows the analysis of the sensitivity and specificity of the model depending on the threshold values of the logistic function P.

Thus, when assessing the dependence of the probability of a fatal outcome on the value of the logistic function P using ROC analysis, the curve shown in Figure 3 was obtained. The area under the ROC curve was 0.838±0.046 with 95% CI 0.747-0.929. The resulting model was statistically significant (p<0.001). The threshold value of the logistic function P at the cut-off point, which corresponded to the highest value of the Youden index, was 0.015. The presence was predicted at the value of the logistic function Р≥1.5%. The sensitivity and specificity of the model were 72.4% and 84.0%, respectively.

Examples of the proposed method

Example 1

Patient U., 73 years old, was admitted for a planned CABG with a diagnosis of multifocal atherosclerosis. Coronary artery disease. Functional class IV angina (FC) (CCS). Postinfarction cardiosclerosis (2015). CHF 2A st., 2 st. by NYHA. Mitral insufficiency 1-2 tbsp. Arterial hypertension (grade 1, very high risk of CVD). Bicarotid stenosis (ICA on the left - up to 70%, ICA on the right - up to 60%). Atherosclerosis of the arteries of the lower extremities 2 tbsp. (STA on the right is narrowed to 75%, SFA on the left - 60%, left common femoral artery - up to 75%). Type 2 diabetes mellitus (moderate severity, compensated). Diabetic polyneuropathy. Chronic kidney disease 3b st. Glaucoma. Gastroduodenitis. Peptic ulcer of the 12th duodenum (remission).

Survey results:

Electrocardiography (ECG) - sinus rhythm, correct, heart rate (HR) 55 bpm. Incomplete blockade of PVNPBG. Cicatricial changes on the anterior wall of the left ventricle.

Echocardiography (EchoCG) - myocardial contractility is reduced. The cavities of the heart are moderately dilated. FEV LV - 38%. Mitral insufficiency 1-2 tbsp. Hypertrophy of the left heart. Hypokinesis of the anterior anterior septal wall at the middle and apical levels.

UZDG - signs of systemic atherosclerosis: stenosis of both internal carotid arteries (ICA): ICA on the left - up to 70%, ICA on the right - up to 60%. Stenosis of the vertebral artery up to 45%. Atherosclerosis of the arteries of the lower extremities with signs of calcification: the superficial femoral artery is narrowed to 80% on the right, 60% on the left, and up to 75% on the left common femoral artery. Calcification and diffuse atherosclerosis of the arteries of both legs.

Coronary angiography: right type of blood flow. Multiple lesions of the coronary bed. LCA trunk - stenosis up to 60%, LAD - up to 95% in the middle third, RCA - critically stenotic in the proximal section, circumflex branch (OB) - stenosis up to 80% in the proximal section. Signs of diffuse changes in the distal bed. Syntax SCORE - 28. There are no contraindications to surgical treatment.

BMI - 24 kg / m ² .

Serum creatinine - 139 µmol/l; GFR - 42 ml / min / 1.73 m ² .

The risk of death according to EuroSCORE is 2-7.6%.

Calculation of the risk of death based on the proposed formula:

P \u003d 1 / (1 + e ^-z ) × 100%,

where P is the probability of an outcome in %; e (Euler number) - a mathematical constant equal to 2.71828; z - 4.517 - 0.102X _BMI (24 kg / m ² ) - 0.022X _GFR (42 ml / min / 1.73 m ² ) - 0.092X _{LV FEV} (38%) + 1.494X _{crich. atherosclerosis of LE arteries} (1 - the presence of a critical lesion of NAUT) + 2.226x _{CABG in history} (0 - no history of CABG).

So for this patient

z=4.517 - 0.102 × 24 - 0.022 × 42 - 0.092 × 38 + 1.494 × 1 + 2.226 × 0=-0.857

P=1/(1+2.71828 ^-(-0.857) ) × 100%=29.8%.

The risk of death when using the proposed method is 29.8%.

Due to severe angina pectoris and damage to the LCA trunk, a decision was made to perform coronary revascularization.

LAD shunting with LVHA, coronary artery bypass grafting of the VTK and OB with autogenous vein were performed under EC conditions. The duration of the operation was 5 hours 49 minutes. Interoperative blood loss in 1 day - 700 ml. According to the control shuntography, all shunts are passable.

The postoperative period was complicated by the phenomena of cardiovascular insufficiency, respiratory failure, the clinic of cerebral edema, multiple organ failure, as well as infectious complications (right-sided lower lobe pneumonia). Bed-day in the ICU: 12. The death of the patient - on the 12th day from multiple organ failure.

Example 2

Patient E., 61 years old, was admitted for planned coronary bypass surgery with a diagnosis of coronary heart disease. Angina IV FC (CCS). Postinfarction cardiosclerosis (2019). Arterial hypertension (grade 1, very high risk of CVD). Diabetes mellitus type 2 (moderate severity, remission). Obesity 1 tbsp. Condition after mastectomy on the right in 2012

Survey results:

ECG - sinus rhythm, correct with a heart rate of 55 bpm. Violation of intraventricular conduction.

EchoCG - myocardial contractility is preserved. The cavities of the heart are not dilated. FEV LV 62%. Symmetrical LV hypertrophy. Moderate aortic stenosis 1 degree with signs of calcification.

UZDG - in the internal carotid arteries on both sides, stenoses up to 40% were noted.

On the lower extremities - the main type of blood flow; no significant arterial lesions were found.

Coronary angiography is the right type of blood flow. Multiple lesions of the coronary bed. LCA trunk - stenosis 55%, LAD - 80% and 90% in the proximal and middle third, VTC OV - critically narrowed from the mouth; RCA - 70% stenosis in the proximal third. Syntax SCORE - 26. There are no contraindications to surgical treatment.

The risk of death according to EuroSCORE is 2-0.9%.

BMI - 32 kg / m ² .

Creatinine 67 µmol / l, GFR - 85 ml / min / 1.73 m ² .

Calculation of the risk of death based on the formula we proposed:

P \u003d 1 / (1 + e ^-z ) × 100%,

where P is the probability of an outcome in %; e (Euler number) - a mathematical constant equal to 2.71828; z \u003d 4.517 - 0.102X _BMI (32 kg / m ² ) - 0.022X _SKf (85 ml / mi / 1.73m ² ) - 0.092X _{LV FEV} (64%) + 1.494X _{critical. atherosclerosis of LE arteries} (1 - the presence of a critical lesion of NAUT) + 2.226x _{CABG in history} (0 - no history of CABG).

So in this case;

z=4.517 - 0.102 × 32-0.022 × 85 - 0.092 × 64 + 1.494 × 0 + 2.226 × 0=-4.881;

P=1/(1+2.71828 ^-(-4.881) ) × 100%=0.7%.

The risk of death when using the proposed method is 0.7%.

LAD shunting with the help of LVHA, coronary artery bypass grafting of the OB and RCA with autoveins were performed under EC conditions. The duration of the operation is 4 hours 10 minutes. Interoperative blood loss in 1 day - 300 ml. According to the control shuntography, all shunts are passable.

Postoperative period without complications. Bed-day in the ICU - 1. Bed-day in the department - 8. Discharged in a satisfactory condition.

The given examples confirm the achievement of a technical result, namely, an increase in the accuracy of the possibility of preoperative prediction of death after CABG, taking into account modern regional adaptation to healthcare models and medical protocols used in practice.

Claims (3)

Method for determining the probability of death at the hospital stage after coronary bypass surgery (CS), including the determination of body mass index (BMI, kg/m ² ), glomerular filtration rate (GFR, ml/min/1.73 m ² ), total ejection fraction of the left ventricle (LV FEV, %), the presence of critical atherosclerotic lesions of the main arteries of the lower extremities - a qualitative indicator (yes / no), previously undergone CABG - a qualitative indicator (yes / no), based on which, according to the formula P \u003d 1 / (1 + e ^-z ) × 100%, where P is the probability of a fatal outcome; e - Euler number - a mathematical constant equal to 2.71828; z=4.517 - 0.102X _BMI - 0.022X _GFR - 0.092X _{LV EF} + 1.494X _{critical NAUT} + 2.226X _{KS in history} , where X _BMI - body mass index (kg/m ² ), X _GFR - glomerular filtration rate ( ml / min / 1.73 m ² ), X _{LV EF} - total left ventricular ejection fraction (%), X _{critical NAAT lesion} : 0 - no critical NAAT lesion <70%, 1 - presence of a critical NAAT ≥70%, X _{History of CABG} : 0 - no history of CABG, 1 - history of CABG, the probability of death is calculated, with a P value equal to or more than 1.5%, a conclusion is made about a high risk of death.

Images