RU2536278C1 - Method for prediction of early complications following coronary artery bypass graft surgery - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention can be used for the purpose of predicting early postoperative (for the first 6 postoperative days) complications in the form of organ dysfunctions in the patients following coronary artery bypass graft surgery. To this effect, within 25-35 minutes upon the beginning of the bypass procedure, the following central haemodynamics values are measured: blood pressure (BP) and central venous pressure (CVP), homeostasis including the blood lactate concentration (Lact), oxygen partial pressure in venous (pvO2) and arterial blood (paO2), and the packed cell volume (Ht). Besides, the time of the bypass (t1), the time of the aortic compression (t2) and the patient's age are determined, and a discriminator is derived by formula: D=3.321+(0.006)*AP+(0.011)*CVP+(1.030)*Lact+(-0.009)*paO2+(0.005)*pvO2+(-0.124)*Ht+(-0.023)*age+(0.014)*t1+(-0.019)*t2. If D<-0.6, the low risk of early postoperative organ dysfunctions is stated. The values D>0.6 enable stating the high risk of the organ dysfunctions. The value falling within the range of -0.6<D<0.6 provides characterising an uncertainty interval, and if the derived value D falls within the range, the patient's state is monitored including more frequent measurement of the standard parameters.

EFFECT: more rapid achievement of the prediction results with maintaining the high prediction accuracy, availability of the results during the bypass graft surgery directly, once the bypass procedure has been completed.

2 tbl, 4 dwg, 2 ex

Description

The invention relates to medicine, namely to cardiovascular surgery, anesthesiology and resuscitation, and can be used to predict the development of complications in the early postoperative period (in the first 6 days after surgery) in patients performing coronary artery bypass surgery (CABG) in cardiopulmonary bypass, which can be expressed as organ dysfunctions, in particular acute renal failure, respiratory failure, acute cardiovascular failure and acute myocardial infarction, atrial fibrillation, encephalopathy and cerebrovascular events. In this case, the forecast of the onset of adverse events, according to the claimed method, is made during the operation or after completion of the cardiopulmonary bypass according to risk criteria - parameters of homeostasis and hemodynamics, taken directly during the operation.

Currently, coronary artery bypass surgery is one of the most common types of cardiac surgery. In the Russian Federation over the past 5 years, the number of operations performed annually has increased by almost 1.5 times, amounting to more than 27.5 thousand in 2010 [Bokeria L.A., Gudkova R.G., 2010]. The development of surgical techniques, methods of anesthesiological aid and organ protection under conditions of cardiopulmonary bypass (IC) allowed to significantly improve the results of such operations and reduce the mortality rate to 1-2.5% [Bokeria L.A., Gudkova R.G., 2010; Miyata, H., et al., 2012]. However, despite the obvious successes, the problem of postoperative complications remains unresolved, most of which are associated with the aggressive effects of cardiopulmonary bypass.

The development in the early postoperative period of initially isolated organ dysfunctions then leads to decompensation of preexisting disorders in patients with initially complex concomitant pathology and the formation of multiple organ failure, accompanied by a ten-fold increase in hospital mortality.

According to various authors, the incidence of postoperative acute renal failure after surgery under IR conditions is from 1% to 30%, while hemodialysis is required in 23-25% of these patients, and mortality in this group increases to 40% [Rodrigues A.J. et al., 2009; Shahzad D Raja, 2005; Stall wood M.I. et al., 2004]. The development of left ventricular failure after shunt interventions may result from reperfusion injury to the myocardium [Hammon J.W. et al., 2003] and require the use of inotropic support or circulatory support methods. According to Wongcharoen W. et al. (2012), the incidence of acute myocardial infarction after coronary bypass surgery can reach 30%. Atrial fibrillation is the most common complication of coronary artery bypass surgery and occurs in 20–40% of patients who have undergone this type of intervention [Zangrillo A. et al., 2004]. The leading role in the pathogenesis of postoperative atrial fibrillation is presumably played by the effects of IR and cardioplegia [Archbold A., et al., 2003]. Cerebral infarction develops in the postoperative period in about 3% of cases, and cerebral artery microembolism in the process of IC plays the main role in their pathogenesis. Knipp S.C., et al., 2004.

The development of postoperative complications even with isolated dysfunction significantly increases the likelihood of a fatal outcome, and in a favorable case leads to an increase in the length of stay in the intensive care unit, the length of hospitalization, complicates the postoperative rehabilitation of the patient and worsens the subsequent quality of life. While the restoration of the quality of life of patients with chronic coronary artery disease is a priority for myocardial revascularization operations.

The work of many researchers over the past few decades has been devoted to solving the problem of risk assessment and predicting the onset of adverse outcomes of coronary artery bypass surgery. The results of these studies, based on database analyzes, have led to the development of integrated scales for assessing the likelihood of complications and a fatal outcome. Examples of the most common risk assessment systems are Euroscore, QMMI, etc. At the same time, their high predictive ability was shown [Giedrius V., et al., 2004; Nashef S.A. et al., 1999].

However, it should be noted that these systems are intended exclusively for preoperative use and are based on the assessment of medical history data and the results of laboratory tests conducted immediately before the intervention. With this approach, the task of intraoperative monitoring of vital functions and predicting the development of adverse events remains unresolved, taking into account changes that occur in the patient’s body directly during surgery in an IR environment. This is especially true in groups of patients with the same initial level of risk of adverse events, that is, for example, an equal number of Euroscore scores.

The prior art method for predicting the results of cardiac surgery in the early postoperative period in patients with isolated, combined and combined lesions of the valvular apparatus of the heart - patent RU 2251964 C1. To predict the risk of postoperative mortality, an analysis of the presence of several risk criteria from the above list is used: the presence of a functional class of circulatory failure above the third, the initial fraction of the left ventricle is less than 45%, the initial pulmonary hypertension is at least 60 mm Hg, and the concentration of serum creatinine in the preoperative a period of more than 102 μm / l, cardiopulmonary bypass time more than 230 minutes. If there are three or more of the above risk criteria, the prognosis is considered unfavorable.

However, this method is informative only in relation to the onset of a fatal outcome in the postoperative period and does not allow predicting the development of non-lethal complications in the form of organ dysfunctions, which, when developed in the postoperative period, lead to an extension of the length of stay of patients in the intensive care unit, in the hospital, and increase the cost of treatment and lower quality of life in the post-hospital phase. In the known method, preoperative risk criteria are used without taking into account changes in the parameters of homeostasis that occur during surgery in conditions of cardiopulmonary bypass. The only risk criterion associated with the use of cardiopulmonary bypass (IR) is an IR duration of more than 230 minutes, which is not applicable for coronary bypass operations, the duration of which is much shorter.

Also known from the prior art is a method for assessing the risk of adverse events after coronary artery bypass grafting in cardiopulmonary bypass, disclosed in the article "Stratification of adverse outcomes by preoperative risk factors in coronary artery bypass graft patients: an artificial neural network prediction model" Chee-Fah Chong, Yu-Chuan Li, Tzong-Luen Wang, Hang Chang AMIA Annu Symp Proc. 2003; 2003: 160-164. The authors of this article based on a retrospective analysis of data from 563 patients who underwent coronary artery bypass grafting under cardiopulmonary bypass, proposed a methodology for assessing the risk of adverse events: death, cardiac arrest, coma, renal failure requiring dialysis, and the need for mechanical ventilation for more than 14 days. To form a forecast, the neural network method was used. As predictors in the method used more than 20 variables characterizing preoperative signs. The authors did not use intraoperative parameters for prediction.

However, in this method, only preoperative indicators were used to assess the risk, which also does not allow to assess the impact of surgical intervention and IR procedures on the risk of adverse events after operations. Only extremely severe complications were also predicted, the frequency of occurrence of which after coronary artery bypass surgery does not currently exceed 1-5%.

Closest to the claimed is a method for predicting adult respiratory distress syndrome and multiple organ failure during coronary artery bypass grafting using cardiopulmonary bypass according to patent RU 2138049 C1, according to which the degree of operational risk is calculated on a scale of clinical, anamnestic and laboratory risk factors (clinical scale) with additional definition of risk degree on the scale of immunological and metabolic risk factors (pathogenetic scale). When comparing risk on both scales, a conclusion is made about the forecast. If the degree of risk coincides on both scales, the accuracy of the forecast increases. At high risk on any of the scales, a conclusion is made about an unfavorable prognosis.

The disadvantage of this method is the large number of predictor indicators, the determination of which requires complex immunological tests using expensive reagents and special laboratory equipment, time-consuming, which limits the applicability of the proposed method for predicting the development of complications in the early postoperative period.

Another disadvantage of this method is the difficulty of its application in everyday clinical practice due to the need to carry out calculations on several scales and the ambiguity in the interpretation of the forecast results with multidirectional predictions of the scales used. The determination of most predictor parameters is performed in the preoperative period, which does not allow fully taking into account the effect of cardiopulmonary bypass on homeostasis. In the description of the method there are no clear criteria for the diagnosis of respiratory distress syndrome and multiple organ failure. In addition, the method was developed on a sample of patients whose surgical interventions were performed under conditions of cardiopulmonary bypass and moderate hypothermia, and the possibility of using this method and the accuracy of the prognosis for coronary artery bypass surgery under normothermia is unknown.

The objective of the invention is the creation for clinical practice of a new method for predicting the development of complications in the early period after bypass surgery of coronary arteries in cardiopulmonary bypass based on intraoperative monitoring data.

The technical result to which the claimed invention is directed is to increase the speed of obtaining forecast results while maintaining high prognostic accuracy, as well as the ability to obtain results online - directly during the coronary artery bypass surgery at the end of the cardiopulmonary bypass.

The problem is solved in that the method of intraoperative determination of the risk of developing organ dysfunctions in the early period after the operation of isolated coronary artery bypass grafting under cardiopulmonary bypass involves measuring in the range of 25-35 minutes from the start of the cardiopulmonary bypass procedure of central hemodynamics, including the level of arterial ( AR) and central venous pressure (CVP), indicators of homeostasis, including the concentration of lactate in the blood (Lact) , partial pressure of oxygen in venous (pvO2) and arterial blood (paO2), hematocrit (Ht), as well as determination of the duration of cardiopulmonary bypass (t1) and clamping of the aorta (t2), and patient age (age), followed by determination of the discriminant value functions according to the formula:

D = 3.321 + (0.006) * AP + (0.011) * CVP + (1.030) * Lact + (- 0.009) * paO2 + (0.005) * pvO2 + (- 0.124) * Ht + (- 0.023) * Age + (0.014) * t1 + (- 0.019) * t2,

and when receiving values of D <-0.6, they conclude that there is a low risk of developing organ dysfunctions in the early period after the operation, when receiving values of D> 0.6 they conclude that there is a high risk of developing organ dysfunctions, while the interval of values is -0.6 <D <0.6 is characterized as an uncertainty interval; upon receipt of the D value from this interval, the state of the patient is monitored, including a more frequent measurement of standard parameters.

The invention is illustrated by the drawings shown in figures 1-4.

Figure 1 presents a diagram of the grouping of values of the discriminant function (D-function) for patients without postoperative complications, figure 2 - for patients with postoperative complications; 1 and 2 along the X axis indicate the value of the function D, along the Y axis the number of patients with the corresponding value of D; figure 3 presents the ROC curve for the developed model (AUC = 0.847; p = 0.016; confidence interval 0.656-1.0) for the control sample; figure 4 presents a graphical distribution of the D-function values of patients of the control sample with postoperative complications (point 1 along the X axis) and without them (point 0 along the X axis), the values of function D are plotted along the Y axis. D values for each patient are indicated squares.

The inventive method was developed on the basis of an analysis of a retrospectively selected sample of 33 patients who underwent isolated coronary artery bypass grafting under conditions of cardiopulmonary bypass at the Regional Cardiac Surgery Center of Saratov. The sample was divided into two groups:

- Group 0 - without postoperative complications, which included 33 patients: men - 26 (80%), women - 7 (20%), the average age of patients in this group was 58.4 ± 8.5 years (42-76 years );

- Group 1 with the complicated course of the postoperative period (the structure of complications is presented in Table 1), which included 24 patients: men - 19 (79.3%), women - 6 (20.7%), the average age of patients in the group was 59.6 ± 7.9 years (42-73 years).

Both groups according to the results of statistical comparison using nonparametric methods were comparable by sex, age, nature of coronary artery damage and the structure of concomitant pathology (see Table 2), which allowed us to use this sample to develop a model for predicting postoperative complications

Complicated was considered the early (1-6 days from the time of surgery) postoperative period in the presence of at least one of the following conditions:

- the development of renal failure - an increase in blood creatinine level of more than 200 μmol / l or more than 50% of the preoperative level;

- the development of respiratory failure, requiring the use of prolonged mechanical ventilation;

- development of paroxysm of atrial fibrillation;

- hemodynamic instability, requiring the use of inotropic or vasopressor support and / or methods of auxiliary circulation;

- Signs of acute myocardial infarction;

- Symptoms of encephalopathy and signs of acute cerebrovascular accident (stroke).

The structure of postoperative complications, the development of which was noted in the first 6 days after surgery, is presented in table 1.

Table 1 Group 1, N = 24 The number of patients who have noted the development of this complication, abs. Share in the group,% Acute myocardial infarction 3 12.5 Acute renal failure eleven 45.8 Respiratory failure requiring extended ventilation 6 25.0 Paroxysm of atrial fibrillation 8 33.3 Acute cardiovascular failure 8 33.3 Symptoms of encephalopathy and signs of stroke 6 25.0

The development of multiple organ failure (involving 3 or more systems) was observed in 11 (45.8%) patients in this group. Mortality in group 1 was 37.5% - 9 patients died.

Brief preoperative characteristics of the groups and parameters of surgical interventions are presented in table 2.

table 2 Euroscore, points The probability of death according to Euroscore,% Ejection fraction,% The number of bypass arteries IR time, min Aortic clamping time, min Group 1.88 ± 1.1 1.37 ± 0.51 59.7 ± 11.9 2.42 ± 0.76 147.4 ± 99.2 64.7 ± 34.2

one Group 0 2.55 ± 2.1 2.3 ± 1.8 58.8 ± 10.3 2.43 ± 0.57 79.2 ± 24.2 41.5 ± 15.0 R 0.439 0.174 0.952 0.774 0.0011 * 0.0118 * * - differences between groups are significant

As follows from the table, between the groups there are statistically significant differences in the duration of IR and time of clamping of the aorta. The results obtained are completely consistent with the data of other authors [Wesselink R.M., 1997; Babaev MA, 2011] that with an increase in IC time, the likelihood of developing organ failure and the severity of its clinical manifestations increases.

Before surgery, both groups evaluated the risk of adverse outcomes using the Euroscore scale. All patients underwent isolated coronary artery bypass surgery under IR normothermia. Anesthesia and IR were performed according to the unified protocol adopted at the clinic, pharmacological cold cardioplegia was used. Conventional hemodynamic parameters were monitored.

The following indicators were selected for further analysis: blood pressure (AR, mmHg), central venous pressure (CVP, mmHg), blood lactate level (Lact, mmol / l), hematocrit (Ht,% ), the partial pressure of oxygen in arterial and venous blood (paO2 and pvO2, mmHg). For the forecast, predictor values were used, determined within 25-35 minutes from the beginning of cardiopulmonary bypass. It should be noted that throughout the entire period of intraoperative monitoring, the values of these parameters were within the reference range of normal values in patients of both groups. The following indicators were also used for the prognosis: patient age (age), IR duration (t1) and aortic clamping time (t2).

To determine the prognostic significance of changes in the selected parameters in relation to the development of postoperative complications, the method of discriminant analysis was used. To carry out discriminant analysis, the IBM SPSS 19.0 software package was used, and MedCalc 7.4 was used to graphically display the results of the ROC analysis characterizing the predictive accuracy of the obtained model.

Discriminant analysis is a statistical method that allows you to solve the problem of assigning a studied object to one of several available groups based on the determination of a certain number of features. The technology of automated discriminant analysis to obtain the discriminant function is presented in detail in the book Nasledov A.D. SPSS: professional statistical data analysis / A.D. Nasledov, St. Petersburg, 2011. The prognostic model obtained as a result of discriminant analysis is a linear discriminant function. The assignment of patients to the group with or without postoperative complications using the proposed diagnostic model is based on comparing the values of the obtained linear discriminant function of a particular patient with the average values of the corresponding functions for patients of the studied groups. To calculate the discriminant function in the program SPSSS 19.0. the values of intraoperative predictors selected for the prediction were entered, as well as data on known outcomes. As a result, a linear discriminant function of the form was obtained:

D = 3.321 + (0.006) * AP + (0.011) * CVP + (1.030) * Lact + (- 0.009) * paO2 + (0.005) * pvO2 + (- 0.124) * Ht + (- 0.023) * Age + (0.014) * t + (- 0.019) * t2,

where D is the value of the discriminant function.

The resulting model is characterized by a high degree of predictive accuracy, which is confirmed by the value of the canonical correlation coefficient, which for this function is 0.833. For this coefficient, the confidence level is p <0.001, which indicates a high predictive accuracy of the obtained model.

The differences between the groups are also characterized by different average values of the D-function for each of the studied groups. Figure 1 graphically shows the distribution of patients of each of the groups of the training sample on the plane in accordance with the values of the function D calculated for each of the patients. It can be seen that the largest number of patients of group 1 with postoperative complications is grouped in the positive half-plane near the average value of D for this group, which is 1.93. At the same time, most patients without postoperative complications (group 0) are grouped by the value of function D in the negative half-plane, near the average for this group, which is -1.10. Thus, in FIGS. 1 and 2, it can be seen that the D values for groups with and without postoperative complications are shifted relative to the 0 coordinate plane in the positive and negative parts, respectively.

The most accurate assignment of a particular patient to one of the groups will occur in cases where the linear discriminant function will take values close to the average group, or when the values of the discriminant function are located, respectively, in areas of less negative and more positive values of the average group values. In the case of “getting” the values of the discriminant function of the studied patient into the zone located “between the groups”, especially near the value conditionally dividing the distance between the middle groups in two, the accuracy of this mathematical model will obviously decrease, thus forming a certain “range of uncertainty ".

The accuracy of the developed model was tested on an additional control sample of 20 patients who underwent coronary artery bypass surgery with known outcomes. To verify the predictive accuracy of the developed model in the control sample, the ROC analysis procedure was used (the results are graphically presented in Figs. 3 and 4). The clinical significance of the studied parameter was numerically evaluated using the AUC indicator (area under curve) in accordance with the generally accepted expert scale for AUC values.

In the course of the ROC analysis, the value of D was compared with the known outcome of the presence or absence of early postoperative complications. Figure 3 presents the actual ROC curve, along the axes, respectively, indicators of sensitivity and 100 specificity, presented in percent (%). The area under the ROC curve obtained for our model, and, accordingly, the forecasting accuracy is 0.847. Figure 4 presents the assessment of the threshold value of the D-function, namely, a schematic grouping of the values of the D-function corresponding to patients with and without postoperative complications, the value of the function D is also indicated with the optimal combination of sensitivity and specificity, which makes it possible to distinguish between the studied groups.

The obtained value AUC = 0.847 corresponds to an expert assessment of the model quality “very good”, while the value p = 0.016 indicates the statistical reliability of the calculated parameters. The value D = -0.6 is recognized as “critical”, which allows one to distinguish between groups with and without complications. Thus, when using this model, when values are less than -0.6, a smooth course of the postoperative period will be most likely; greater than -0.6, the development of postoperative complications is predicted; when a value of = -0.6 is obtained, the prognosis should be considered uncertain, in clinical practice, based on the greater risk of unrecognizing the development of complications than their overdiagnosis, these patients should be assigned to the group with predicted development of complications. When using this model in clinical practice, the need for a differentiated therapeutic and tactical approaches is obvious depending on the results of assigning the patient to a particular group.

During surgery, upon completion of the cardiopulmonary bypass procedure, the prognosis is calculated by substituting the values of the parameters of the studied patient into the formula. Upon receipt of a value of D <-0.6, a conclusion is made that there is a low risk of developing organ dysfunctions in the early period after an operation; when a value of D> 0.6 is obtained, a conclusion is made that there is a high risk of developing organ dysfunctions, and if a value of D = - 0.6 the prognosis is characterized as doubtful, and when the D value is obtained from this interval, the patient is monitored, monitoring the state of the internal environment of the body, including more frequent measurement of standard parameters (hemodynamics, blood chemistry, urine output, and etc.). This category of patients is subject to closer monitoring in the intensive care unit.

Example 1. Patient K., 58 years old. A mammary coronary artery bypass grafting was performed and 3 coronary artery bypass grafting was performed. The duration of the IR is 79 minutes, the duration of clamping the aorta is 30 minutes. At the 30th minute of the IR, the determined parameters had the following values: HELL = 49 mm Hg, CVP = 70 mm Hg, blood lactate level - 2.3 mmol / L; pvO2 = 39 mmHg, paO2 = 204 mmHg, hematocrit = 25%. When calculating the values of the discriminant function by substituting the indicated values in the formula D = 1.215, the value is more critical "-0.6", close to the average for the group with complications. In the postoperative period, this patient had hypotension, which required inotropic support.

Example 2. Patient S., 59 years old. A mammary coronary artery bypass grafting was performed and 3 coronary artery bypass grafting was performed. The duration of the IR is 55 minutes, the duration of clamping the aorta is 34 minutes. At the 30th minute of the IR, the determined parameters had the following values: HELL = 41 mm Hg, CVP = 0 mm Hg, blood lactate level -1.1 mmol / L; pvO2 = 29.5 mmHg, paO2 = 293 mmHg, hematocrit = 21.9%. When calculating the values of the discriminant function by substituting the indicated values in the formula D = -1.738, the value is less than the “critical” one. A smooth course of the postoperative period was noted. It should be noted that in both cases the values of the monitored indicators did not go beyond the limits of normal values, however, they were classified by the model differently.

Thus, the use of the claimed invention - prognostic models in the dynamic monitoring of clinical and laboratory parameters allows you to adjust the treatment tactics at the initial stages of the formation of pathological processes.

Claims (1)

The method of intraoperative determination of the risk of developing organ dysfunctions in the early period after the operation of isolated coronary artery bypass grafting under cardiopulmonary bypass, including measuring in the range of 25-35 minutes from the start of the cardiopulmonary bypass procedure of central hemodynamics, including the level of arterial (AP) and central venous pressures (CVP), indicators of homeostasis, including the concentration of lactate in the blood (Lact), the partial pressure of oxygen in noznoy (pvO2) and arterial blood (paO2), hematocrit (Ht), as well as determination indexes duration artificial circulation (t1), the time of aortic cross-clamping (t2), and age of the patient (age) with subsequent determination of the value of the discriminant formula features:
D = 3.321 + (0.006) * AP + (0.011) * CVP + (1.030) * Lact + (- 0.009) * paO2 + (0.005) * pvO2 + (- 0.124) * Ht + (- 0.023) * age + (0.014) * t1 + (- 0.019) * t2,
and when receiving values of D <-0.6, they conclude that there is a low risk of developing organ dysfunctions in the early period after the operation, when receiving values of D> 0.6 they conclude that there is a high risk of developing organ dysfunctions, while the interval of values is -0.6 <D <0.6 is characterized as an uncertainty interval; upon receipt of the D value from this interval, the patient's condition is monitored, including a more frequent measurement of standard parameters.

Images