Non-cardiac surgery perioperative period large-dose blood transfusion line diagram prediction method

Abstract

The invention provides a non-cardiac surgery perioperative period large-dose blood transfusion line diagram prediction method, and relates to the technical field of medical auxiliary prediction model construction, and the method comprises the following steps: S1, determining a research object: employing a retrospective case contrast research method, and collecting suitable cases of non-cardiac surgery treatment and perioperative period blood transfusion treatment for research; the model is internally verified by utilizing R software, the actually observed and predicted perioperative period large-dose blood transfusion has good consistency, the calibration degree of the model is high, the prediction model can be used for predicting perioperative period large-dose blood transfusion of various different surgical patients, the sample size is the largest in the existing research, and the method is suitable for large-dose blood transfusion prediction in the perioperative period. Through research on influence factors of large-dose blood transfusion, a prediction model of large-dose blood transfusion in the perioperative period is established, a clinician is helped to recognize blood transfusion risk factors of an operative patient in an early stage before an operation, and the probability that large-dose blood transfusion possibly occurs in the perioperative period of the patient is effectively predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of medical auxiliary prediction model construction, and in particular to a nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery. Background Art

[0002] Massive intraoperative bleeding is a common complication of surgical operations. When patients suffer from severe perioperative anemia or hemorrhagic shock due to intraoperative blood loss, timely and reasonable blood transfusion can save the patient's life, but blood transfusion therapy also has certain risks and costs, and blood products are very scarce and precious. Therefore, prudent and reasonable blood transfusion is very important. Patient-centered precise individualized management can improve patients' prognosis and quality of life, and can also effectively save precious blood resources. Therefore, predicting the perioperative red blood cell transfusion risk for surgical patients, optimizing the blood preparation process, and rationally allocating blood resources are important links in individualized perioperative blood management.

[0003] At present, there have been research reports at home and abroad on the factors affecting the risk of massive intraoperative bleeding or massive perioperative blood transfusion, such as patient age, body mass index (BMI), shock index, preoperative hemoglobin, left ventricular ejection fraction, and preoperative international normalized ratio (INR). These factors have been clearly proven to be related to large-dose blood transfusion, but it is impossible to accurately predict the patient's transfusion risk by evaluating only a single independent influencing factor.

[0004] In recent years, nomograms have been widely used in clinical research and prediction models because of their advantages of integrating different univariate variables and visualizing the predicted probability results of multivariate regression models. Nomogram prediction models for blood transfusion have been partially reported in some clinical subspecialty surgeries such as liver surgery, neurosurgery, and hip replacement surgery, but the number of relevant samples is small, only 200 to 1,000 cases, and most of the current studies are focused on a single discipline or a single type of surgical patient, and the predicted outcomes are basically focused on whether the surgical patient has a perioperative red blood cell transfusion outcome. There are few studies on the perioperative red blood cell transfusion volume of surgical patients, and there is a lack of interdisciplinary, large-sample related research. The perioperative red blood cell transfusion risk prediction model that can be applied to different types of surgical patients at the same time needs further in-depth exploration. Summary of the invention

[0005] The invention provided by the present invention aims to provide a prediction method for perioperative high-dose blood transfusion nomograms for non-cardiac surgery. The model is internally verified by using R software, and there is a good consistency between the actual observed and predicted perioperative high-dose blood transfusions, indicating that the calibration of the model is high. The prediction model can be used for the prediction of perioperative high-dose blood transfusions for various types of surgical patients, and it has the largest sample size in existing studies. Through the study of the influencing factors of high-dose blood transfusions, a prediction model for perioperative high-dose blood transfusions is established to help clinicians identify the risk factors of blood transfusions for surgical patients early before surgery and effectively predict the probability of high-dose blood transfusions that may occur in patients during the perioperative period, thereby strengthening perioperative blood management in an individualized manner, guiding surgical doctors to implement preoperative blood combination more accurately, providing blood combination approval basis for the blood transfusion department, optimizing the blood combination application process, and evaluating the patient's physical health and surgical risk before anesthesia and grading them, so that the six prediction factors of the prediction model are all easy to obtain, and have good operability in clinical application.

[0006] In order to achieve the above-mentioned effects, the present invention provides the following technical solution: a nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery, comprising the following steps:

[0007] Step 1. Determine the research subjects: A retrospective case-control study method was used to collect appropriate cases of patients undergoing non-cardiac surgery and perioperative blood transfusion therapy for inclusion in the study.

[0008] Step 2: Search observation indicators: Retrospectively collect the patient's medical records during hospitalization, including the patient's general information, surgical anesthesia-related information, and preoperative blood test results.

[0009] Step 3: Univariate analysis between groups: Conduct univariate analysis on the collected observation indicators for inter-group comparison.

[0010] Step 4: Multivariate Logistic regression analysis: The statistically significant influencing factors among the single factors between the groups were included in the binary multivariate Logistic regression model. The results of the multivariate regression analysis were used to screen out independent risk factors related to perioperative large-dose blood transfusion to construct a nomogram model.

[0011] Step 5. Use R software to establish a nomogram: import the independent influencing factors obtained from the multivariate regression analysis into the function, use R software to draw a nomogram, and construct a nomogram prediction model that can predict the risk of large-dose blood transfusion during non-cardiac surgery.

[0012] Step 6. Evaluation and verification of the prediction model: Use R software to draw the receiver operating characteristic curve, determine the area under the curve, combine the C-index to measure the discrimination of the nomogram model, draw the calibration graph, observe the relationship between the original curve of the model and the calibration curve and the ideal curve, and evaluate the calibration of the model.

[0013] Furthermore, the collection criteria in step one include inclusion criteria and exclusion criteria. The inclusion criteria include patients undergoing non-cardiac surgery during hospitalization and patients undergoing allogeneic red blood cell transfusion during hospitalization. The exclusion criteria include age less than 15 years old and incomplete clinical case data.

[0014] Furthermore, in step one, the total transfusion volume of allogeneic red blood cell transfusion therapy for the patient during hospitalization is divided into a non-large-dose transfusion group and a large-dose transfusion group, the total transfusion volume of the non-large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume <8u, and the total transfusion volume of the large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume ≥8u.

[0015] Furthermore, in step 2, the general information includes gender, age, weight, height, BMI, diabetes, hypertension, heart disease, blood transfusion history, allergy history, surgical history, smoking and drinking habits.

[0016] Furthermore, in step 2, the surgical anesthesia-related data include intraoperative blood transfusion volume, surgical department, surgery name, whether it is an emergency surgery, whether it is a day surgery, surgery classification, surgery type, surgery duration, intraoperative blood loss, surgical incision description, ASA anesthesia classification and anesthesia time.

[0017] Furthermore, in step 2, the preoperative blood test results include preoperative hemoglobin, albumin, white blood cell count, platelet count, INR, PT and APTT.

[0018] Furthermore, the step of univariate analysis among the three groups includes:

[0019] S301. Normal distribution test of continuous variables: All continuous variables are tested for normal distribution using the functions in R language. The test for normal distribution is expressed as ks.test(x, "pnorm")#P>0.05, and the test for non-normal distribution is expressed as P≤0.05;

[0020] S302. Nonparametric test of continuous variables: Use the function in R language to perform nonparametric test on the continuous variables of the samples. If the difference between the groups is not significant, it is indicated as wilcox.test(tested variable ~ grouping variable, paired = F) #P ≥ 0.05. If the difference between the groups is significant, it is indicated as P < 0.05.

[0021] S303, chi-square test for categorical variables: chi-square test was performed on all categorical variables using the function in R language. The identification field of sample classification was represented as tab1<-table(x, md$whether or not a large-dose blood transfusion)#"whether or not a large-dose blood transfusion was performed", insignificant differences between groups were represented as CrossTable(tab1, chisq=T, expected=T, fisher=T)#P≥0.05, and significant differences between groups were represented as P<0.05;

[0022] S304. Single factor analysis results: Based on the above single factor analysis, single factor analysis results are obtained.

[0023] Furthermore, in step 4, the condition for the existence of statistical significance in the univariate analysis between the groups is that the variable has a P value less than 0.05 in the univariate analysis.

[0024] Furthermore, in step five, the score is marked on the line segment in the Nomogram to represent the degree of influence of the factor, and the sum of each individual indicator is added to obtain the total score of the prediction indicator, and finally the probability of perioperative large-dose blood transfusion can be obtained.

[0025] Furthermore, in step six, the model was internally validated using R software, a validation set was established using the Bootstrap method of resampling the data set 1000 times with replacement, and the consistency index was calculated using the Harrell-C-statistic function.

[0026] The present invention provides a nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery, which has the following beneficial effects: the model is internally verified by using R software, and there is a good consistency between the actually observed and predicted perioperative high-dose blood transfusions, indicating that the model has a high degree of calibration, and the prediction model can be used to predict perioperative high-dose blood transfusions in patients with various types of operations, and has the largest sample size in existing studies. By studying the influencing factors of high-dose blood transfusions, a prediction model for perioperative high-dose blood transfusions is established to help clinicians identify the risk factors for blood transfusion in surgical patients early before surgery and effectively predict the probability of high-dose blood transfusions that may occur in patients during the perioperative period, thereby strengthening perioperative blood management in an individualized manner, guiding surgical doctors to implement preoperative blood combination more accurately, providing a basis for blood combination approval for the blood transfusion department, optimizing the blood combination application process, and evaluating and grading the patient's physical health and surgical risk before anesthesia, so that the six prediction factors of the prediction model are easier to obtain and have good operability in clinical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of the method for predicting large-dose blood transfusion nomogram during perioperative period of non-cardiac surgery of the present invention;

[0028] Figure 2 A nomogram for predicting perioperative high-dose blood transfusion in surgical patients according to the method for predicting perioperative high-dose blood transfusion in non-cardiac surgery of the present invention;

[0029] Figure 3 It is a schematic diagram of an example of using a perioperative high-dose blood transfusion nomogram prediction model of a perioperative high-dose blood transfusion nomogram prediction method for non-cardiac surgery of the present invention;

[0030] Figure 4 A ROC curve diagram of the nomogram model of the nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery of the present invention for predicting perioperative high-dose blood transfusion in surgical patients;

[0031] Figure 5 The present invention is a calibration curve diagram for predicting perioperative high-dose blood transfusion in surgical patients using a nomogram model of the perioperative high-dose blood transfusion prediction method for non-cardiac surgery. DETAILED DESCRIPTION

[0032] See also Figure 1-5 The present invention provides a technical solution: a nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery, comprising the following steps:

[0033] Step 1. Determine the research subjects: A retrospective case-control study method was used to collect appropriate cases of patients undergoing non-cardiac surgery and perioperative blood transfusion therapy for inclusion in the study.

[0034] Step 2: Search for observation indicators: Retrospectively collect the patient's medical records during hospitalization, including the patient's general information, surgical anesthesia-related information, and preoperative blood test results.

[0035] Step 3: Univariate analysis between groups: Conduct univariate analysis on the collected observation indicators for inter-group comparison.

[0036] Step 4: Multivariate Logistic regression analysis: The statistically significant influencing factors among the single factors between the groups were included in the binary multivariate Logistic regression model. The results of the multivariate regression analysis were used to screen out the independent risk factors related to perioperative large-dose blood transfusion in order to construct a nomogram model.

[0037] Step 5. Use R software to establish a nomogram: import the independent influencing factors obtained from the multivariate regression analysis into the function, use R software to draw a nomogram, and construct a nomogram prediction model that can predict the risk of large-dose blood transfusion during non-cardiac surgery.

[0038] Step 6. Evaluation and verification of the prediction model: Use R software to draw the receiver operating characteristic curve, determine the area under the curve, combine the C-index to measure the discrimination of the nomogram model, draw the calibration graph, observe the relationship between the original curve of the model and the calibration curve and the ideal curve, and evaluate the calibration of the model.

[0039] Specifically, the collection criteria in step one include inclusion criteria and exclusion criteria. The inclusion criteria include patients who underwent non-cardiac surgery during hospitalization and patients who underwent allogeneic red blood cell transfusion during hospitalization. The exclusion criteria include age less than 15 years and incomplete clinical data of the cases.

[0040] Specifically, in step one, the total transfusion volume of allogeneic red blood cell transfusion therapy for patients during hospitalization is divided into a non-large-dose transfusion group and a large-dose transfusion group. The total transfusion volume of the non-large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume <8u, and the total transfusion volume of the large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume ≥8u.

[0041] Specifically, in step 2, the general information includes gender, age, weight, height, BMI, diabetes, hypertension, heart disease, blood transfusion history, allergy history, surgical history, smoking and drinking habits.

[0042] Specifically, in step 2, the surgical anesthesia-related data include the intraoperative blood transfusion volume, surgical department, surgery name, whether it is an emergency surgery, whether it is a day surgery, surgery classification, surgery type, surgery duration, intraoperative blood loss, surgical incision description, ASA anesthesia classification and anesthesia time.

[0043] Specifically, in step 2, the preoperative blood test results include preoperative hemoglobin, albumin, white blood cell count, platelet count, INR, PT and APTT.

[0044] Specifically, the steps of univariate analysis between the three groups include:

[0045] S301. Normal distribution test of continuous variables: All continuous variables are tested for normal distribution using the functions in R language. The test for normal distribution is expressed as ks.test(x, "pnorm")#P>0.05, and the test for non-normal distribution is expressed as P≤0.05;

[0046] S302. Nonparametric test of continuous variables: Use the function in R language to perform nonparametric test on the continuous variables of the samples. If the difference between the groups is not significant, it is indicated as wilcox.test(tested variable ~ grouping variable, paired = F) #P ≥ 0.05. If the difference between the groups is significant, it is indicated as P < 0.05.

[0047] S303, chi-square test for categorical variables: chi-square test was performed on all categorical variables using the function in R language. The identification field of sample classification was represented as tab1<-table(x, md$whether or not a large-dose blood transfusion)#"whether or not a large-dose blood transfusion was performed", insignificant differences between groups were represented as CrossTable(tab1, chisq=T, expected=T, fisher=T)#P≥0.05, and significant differences between groups were represented as P<0.05;

[0048] S304. Single factor analysis results: Based on the above single factor analysis, single factor analysis results are obtained.

[0049] Specifically, in step 4, the condition for statistical significance in the univariate analysis between groups is the variable with a P value less than 0.05 in the univariate analysis.

[0050] Specifically, in step five, the score is marked on the line segment in the Nomogram to represent the degree of influence of the factor, and the sum of each individual indicator is added to obtain the total score of the prediction indicator, and finally the probability of perioperative large-dose blood transfusion can be obtained.

[0051] Specifically, in step six, the model was internally validated using R software, a validation set was established using the Bootstrap method of 1000 resamplings with replacement, and the consistency index was calculated using the Harrell-C-statistic function.

[0052] The method of the embodiment is used for detection and analysis, and compared with the prior art. It can be concluded that when the embodiment is used, the model is internally verified by using R software, and there is a good consistency between the actual observed and predicted perioperative large-dose blood transfusions, indicating that the model has a high degree of calibration. The prediction model can be used to predict perioperative large-dose blood transfusions for various types of surgical patients, and has the largest sample size in existing studies. By studying the influencing factors of large-dose blood transfusions, a prediction model for perioperative large-dose blood transfusions is established to help clinicians identify transfusion risk factors for surgical patients early before surgery and effectively predict the probability of large-dose blood transfusions that may occur in patients during the perioperative period, thereby strengthening perioperative blood management in an individualized manner, guiding surgeons to implement preoperative blood combination more accurately, providing a basis for blood combination approval for the blood transfusion department, optimizing the blood combination application process, and evaluating the patient's physical health and surgical risk before anesthesia and grading them, so that the six prediction factors of the prediction model are easier to obtain and have good operability in clinical applications.

[0053] The present invention provides a nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery:

[0054] 1. Materials and methods for model building

[0055] 1.1 Establish modeling objects and groups

[0056] 1.1.1 Research subjects

[0057] This study adopted a retrospective case-control study method. Appropriate cases of non-cardiac surgery and perioperative blood transfusion were collected for inclusion in the study.

[0058] Inclusion criteria: patients who underwent non-cardiac surgery during hospitalization; patients who received allogeneic red blood cell transfusion during this hospitalization.

[0059] Exclusion criteria: age less than 15 years old; incomplete clinical data of cases.

[0060] 1.1.2 Experimental Grouping

[0061] According to the total amount of allogeneic red blood cell suspension transfused during hospitalization, the patients were divided into a non-large-dose transfusion group (allogeneic red blood cell suspension transfusion volume <8u) and a large-dose transfusion group (allogeneic red blood cell suspension transfusion volume ≥8u). According to the quality requirements of standard whole blood and component blood, each unit of whole blood contains 200 ml of red blood cells.

[0062] 1.2 Research Indicators

[0063] 1.2.1 Outcome indicators:

[0064] Whether the total amount of allogeneic red blood cell suspension transfused during the operation and within 24 hours after the operation is greater than or equal to 8 U is used as the outcome indicator of the study observation. This indicator is also used as the basis for the study grouping and the predicted outcome of the model.

[0065] 1.2.2 Prediction indicators:

[0066] The medical records of patients during hospitalization were collected retrospectively, including the following three parts:

[0067] (1) Patient's general information: gender, age, weight, height, BMI [(kg / m 2 ) = weight (kg) / height 2 m 2 ], etc.; past medical history includes: diabetes, hypertension, heart disease (coronary heart disease, heart failure, organic heart disease, etc.), blood transfusion history, allergy history, surgical history, smoking and drinking, etc.

[0068] (2) Information related to surgical anesthesia: intraoperative blood transfusion volume, surgical department, surgery name, whether it is an emergency surgery, whether it is a day surgery, surgery classification, surgery type, surgery duration, intraoperative blood loss, surgical incision description, ASA anesthesia classification, anesthesia time, etc.

[0069] (3) The patient's preoperative blood test results: preoperative hemoglobin, albumin, white blood cell count, platelet count, INR, PT, APTT, etc.

[0070] 1.3 Statistical analysis

[0071] The main statistical work of the study was completed using R software (Version 4.3.2; ).

[0072] 1.3.1 Univariate analysis:

[0073] The Kolmogorov-Smirnov test was used to check whether the data of each group met the normal distribution. The measurement data that met the normal distribution were described as "mean ± standard deviation", and the t test was used for inter-group comparison; the measurement data that did not meet the normal distribution were described as "median [upper quartile to lower quartile]", and the non-parametric test (Wilcoxon test) was used for inter-group comparison. For count data, the "frequency (percentage)" method was used to describe, and the chi-square test was used for inter-group comparison.

[0074] 1.3.2 Binary multivariate logistics regression analysis:

[0075] Variables with P values ​​less than 0.05 in univariate analysis were considered statistically significant in the inter-group comparison. These variables were further introduced into the binary multivariate logistic regression model. The results of multivariate regression analysis were used to screen out independent risk factors associated with perioperative high-dose blood transfusion to construct a nomogram model.

[0076] 1.3.3 Constructing the Nomogram Model

[0077] The independent influencing factors obtained from multivariate regression analysis were imported into the function using the lrm function in the rms package of R software, and the Nomogram was drawn using R software to construct a nomogram prediction model that can predict the risk of large-dose blood transfusion during non-cardiac surgery. In the Nomogram, the score is marked on the line segment to represent the degree of influence of the factor. The sum of each single indicator is added to obtain the total score of the prediction indicator, and finally the probability of perioperative large-dose blood transfusion can be obtained.

[0078] 1.3.4 Validation and evaluation of prediction models

[0079] The model was internally validated using R software. The validation set was established by using the Bootstrap method with 1000 resamplings with replacement, and the consistency index (C-index) was calculated using the Harrell-C-statistic function. At the same time, the receiver operating characteristic (ROC) curve was drawn using R software to determine the area under the curve (AUC), which was combined with the C-index to measure the discrimination of the nomogram model. Finally, the calibration function of R software was used to draw the calibration graph, observe the relationship between the original curve of the model and the calibration curve and the ideal curve, and evaluate the calibration of the model.

[0080] 2. Specific implementation steps and results of model establishment

[0081] The following further describes the establishment of the patented technology of this prediction model in combination with specific implementation steps and corresponding results:

[0082] 1. Data collection and preprocessing

[0083] 1.1 Collection and preprocessing of research subjects:

[0084] 1.1.1 Collect the list of research subjects: collect information on patients who received perioperative red blood cell suspension transfusion therapy within a period of time through the blood transfusion patient information system: According to the blood transfusion department information system recording plan, the removal and use of each bag of blood products will generate a record in the information system. The key information collected initially includes the hospitalization number of such patients, the amount of blood drawn each time, and the date of blood drawing. List the hospitalization numbers of perioperative blood transfusion patients, which is the list of research subjects.

[0085] 1.1.2 Summarize and summarize the outcome indicators: Use the data statistics function of Excel to summarize and merge the multiple blood collection records of each patient during a single surgery and the same perioperative period. Count the total red blood cell doses taken multiple times during the perioperative period for each patient. The total amount of red blood cell suspension transfused during the perioperative period for each patient (U) is the outcome indicator for establishing the model training set.

[0086] 1.1.3 Sample Grouping:

[0087] According to the outcome index, that is, the total amount of allogeneic red blood cell suspension transfused during the perioperative period, the subjects were divided into a non-large-dose transfusion group (allogeneic red blood cell suspension transfusion volume <8u) and a large-dose transfusion group (allogeneic red blood cell suspension transfusion volume ≥8u).

[0088] 1.2. Collection and preprocessing of predictors

[0089] 1.2.1 Collect observation indicators of perioperative massive blood transfusion: First, by studying the research reports on the risk of massive intraoperative bleeding or perioperative massive blood transfusion at home and abroad, it was found that factors such as patient age, body mass index (BMI), shock index, preoperative hemoglobin, left ventricular ejection fraction, and preoperative international normalized ratio (INR) have been clearly proven to be related to massive blood transfusion. Secondly, combined with clinical experience, the factors that may be related to perioperative blood transfusion include general information such as patient gender, height, and weight; surgical related information such as type of surgery, duration of surgery, and anesthesia classification, as well as preoperative tests such as white blood cell count and platelet count. Using the hospital big data platform and His inpatient information system, the hospitalization number of the patients in the "research subject list" was imported, and the predictive factors that may be related to massive blood transfusion for each patient were exported from the system to obtain a database of observation indicators.

[0090] 1.2.2 Preprocessing of predictors:

[0091] (1) Clean the database and discard relevant data that cannot be analyzed for correlation:

[0092] Some predictive factors in the preliminary observation indicator database could not be analyzed for correlation due to missing key information. For example, the derived "imaging examination" could not view the specific examination time due to data platform record reasons, and could not correspond to the perioperative blood transfusion time, so it was discarded; for example, some blood test items such as "arterial blood gas analysis results" were not routine perioperative examination items, and only some patients had this data, so statistical analysis could not be performed and were therefore discarded.

[0093] (2) Processing of missing values:

[0094] Missing values ​​refer to data missing or truncated due to lack of information in rough data, resulting in incomplete values ​​of one or some attributes in the existing data set. The missing value processing solution is as follows:

[0095] The summary() function in the base package of R language is used to calculate the number of missing values, and the which(is.na()) function is used to find the location of missing values.

[0096] Variable data with missing values ​​greater than 80% were directly deleted.

[0097] For variable data with missing values ​​between 30% and 80%, the missing values ​​of the variable were treated as a separate category for univariate analysis between groups.

[0098] For variable data with missing values ​​less than 30%, missing values ​​will be filled. Continuous variables that follow a normal distribution will be filled with the mean; continuous variables that do not conform to a normal distribution will be filled with the median; and categorical variables will be filled with the mode.

[0099] (3) Numerical transformation of continuous variables:

[0100] BMI (body mass index): A custom function in Excel software was used to convert the patient's weight / height squared (kg / m 2) in the database to obtain the BMI value of each patient.

[0101] Surgery / anesthesia duration: Use the time difference calculation function of Excel software to add the patient's surgery / anesthesia end time and surgery / anesthesia start time obtained in the database to obtain the surgery / anesthesia duration (min).

[0102] 2. Univariate analysis between groups

[0103] 2.1 Normal distribution test of continuous variables:

[0104] All continuous variables: operation time (min), anesthesia time (min), height, weight, BMI, intraoperative blood loss, preoperative hemoglobin level, preoperative albumin level, preoperative white blood cell count, preoperative platelet count, preoperative INR, preoperative PT, preoperative APTT were tested for normal distribution using the Kolmogorov-Smirnov Tests function in the stats package of the R language:

[0105] ks.test(x,"pnorm")#P>0.05 means it obeys the normal distribution; P≤0.05 means it does not obey the normal distribution.

[0106] The final test results indicated that all continuous variables did not conform to the normal distribution.

[0107] 2.2 Nonparametric tests for continuous variables:

[0108] The Wilcoxon Rank Sum and Signed Rank Tests function in the R language stats package is used to perform non-parametric tests on the continuous variables of the sample:

[0109] wilcox.test(tested variable ~ grouping variable, paired = F) #P ≥ 0.05 indicates that the difference between the groups is not significant; P < 0.05 indicates that the difference between the groups is significant.

[0110] 2.3 Chi-square test was used for categorical variables:

[0111] All categorical variables: gender, preoperative anemia, day surgery, emergency surgery, surgery category, surgery grade, ASA grade, surgical incision description, history of hypertension, history of diabetes, history of heart disease, history of blood transfusion, history of allergies, history of surgery, history of smoking, and history of drinking were tested using the Cross Tabulation with Tests for Factor Independence function in the R language gmodels package:

[0112] tab1<-table(x,md$Whether to give a large dose of blood transfusion)#"md$Whether to give a large dose of blood transfusion" is the identification field for sample classification

[0113] CrossTable(tab1,chisq=T,expected=T,fisher=T)#P≥0.05 indicates that there is no significant difference between the groups; P<0.05 indicates that there is a significant difference between the groups.

[0114] 2.4 Results of univariate analysis:

[0115] According to the above univariate analysis, the results of univariate analysis were obtained (see Table 1 for details): age, weight, BMI, preoperative albumin level, preoperative APTT, operation time, anesthesia time, operation grade, ASA grade, surgical incision description, history of hypertension, operation history, and intraoperative blood loss were related factors for perioperative large-dose blood transfusion (p value < 0.05, the difference between the groups was statistically significant). Height, preoperative hemoglobin level, preoperative white blood cell count, preoperative INR, preoperative PT, preoperative platelet count, gender, whether it was day surgery, whether it was emergency surgery, type of surgery, diabetes, heart disease, previous history of blood transfusion, previous history of allergy, smoking history, and drinking history were not significantly correlated with perioperative large-dose blood transfusion (p> 0.05, the difference between the groups was not statistically significant).

[0116] Table 1. Univariate analysis results of factors affecting perioperative high-dose blood transfusion

[0117]

[0118]

[0119]

[0120]

[0121] 2. Binary Multivariate Logistic Regression Analysis

[0122] Before multivariate analysis, considering the convenience of using regression models, the univariate variables were assigned categorical variables according to their respective clinical significance or clinical usage habits:

[0123] Table 2. Value table for factors related to perioperative high-dose blood transfusion

[0124]

[0125] Whether the patient received perioperative high-dose blood transfusion was taken as the dependent variable. The variables with statistically significant inter-group comparison (P value < 0.05) in the univariate analysis results were selected: age, operation time, BMI, intraoperative blood loss, preoperative albumin level, APTT, preoperative anemia, surgical grade, ASA anesthesia grade, surgical incision description, history of hypertension, and previous surgical history (variable assignment see Table 2). A total of 12 factors related to perioperative high-dose blood transfusion were included in the multivariate logistic regression analysis.

[0126] The Fitting Generalized Linear Models function in the stats package in R software was used to perform regression analysis on the relationship between the dependent variable and the independent variable. The code is as follows, where "mydata$Whether to give a large dose of blood transfusion" is the basis for the binary classification of the sample, and "family=binomial" is the binary logistic regression analysis method selected in the regression analysis family of the program package.

[0127] R language code:

[0128] glm(mydata$Whether to give a large dose of blood transfusion~

[0129] mydata$operation duration+mydata$BMI+mydata$bleeding volume+mydata$APTT+mydata$albumin+mydata$operation history+mydata$anemia+mydata$asa grade+mydata$operation level, family=binomial,

[0130] data = mydata)

[0131] The OR value and 95% confidence interval (CI) were calculated, and the results of multivariate regression analysis were summarized:

[0132] Table 3 Multivariate Logistic regression analysis of factors affecting perioperative high-dose blood transfusion

[0133]

[0134]

[0135]

[0136] 4. Draw a nomogram prediction model

[0137] According to the results of multivariate logistic regression analysis, five predictor variables were included: operation duration, intraoperative blood loss, preoperative albumin level, surgical grade, and anesthesia grade. "Preoperative anemia" was also included as a predictor variable because it has important clinical significance for evaluating perioperative blood transfusion and blood combination. The six predictor variables were assigned values ​​according to clinical significance and model practicality and set as categorical variables.

[0138] The Logistic Regression Model function and Draw a Nomogram Representing a RegressionFit function in the R software Package rms version 6.7-1 and Package base version 4.3.1 were used to establish a multivariate regression relationship between the above six predictor variables and the independent variables, and then the multivariate regression relationship was transformed into a nomogram model. The specific code and steps are as follows:

[0139]

[0140]

[0141]

[0142] Among them, lrm is the function of establishing multi-factor regression relationship, nomogram is the function of converting regression relationship into nomogram drawing; "md" (mydata) is the data table of 5252 research samples and six variables included in the regression relationship.

[0143] Run the program and get Figure 2 Shown is a nomogram for predicting perioperative high-dose blood transfusion in surgical patients.

[0144] Each single indicator has a corresponding score (Points), and the sum of all single indicator scores predicts the summary score of the indicator (Total Points). The summary score of the predictive indicator corresponds to the value of the probability on the large-dose transfusion line segment. The probability of perioperative large-dose transfusion can be obtained by reading the graph.

[0145] The following is an example of how to use the nomogram prediction model:

[0146] For example, a patient with a pelvic fracture is to undergo pelvic osteotomy and internal fixation. The estimated duration of the operation is 5 hours, and the estimated intraoperative blood loss is 1200 ml. The preoperative blood test results show that the albumin level is 28 g / L, the hemoglobin (Hb) level is 81 g / L, the surgical grade is 4, and the ASA grade is III. During preoperative preparation, the surgeon uses the nomogram to predict the patient's perioperative transfusion risk. The individual scores of each predictive indicator are as follows: the surgery duration score is 14, the intraoperative blood loss score is 32, the preoperative hemoglobin level score is 26, the preoperative anemia score is 6, the surgical grade score is 40, the ASA grade score is 10, and the predictive indicator summary score is 128. The predicted probability of perioperative large-dose transfusion (red blood cell transfusion exceeding 8 U) for this patient is about 0.2, indicating that the patient has a low risk of perioperative large-dose transfusion.

[0147] 1. Evaluation of the predictive performance of the nomogram model

[0148] The model was internally validated using R software. 1000 resamplings with replacement were performed, and the model was internally cross-validated using the Bootstrap method. The Harrell C statisti function calculated the resampling average C-index to be 0.859, 95% CI [0.857, 0.861].

[0149] In addition, the ROC curve of this model was drawn for further analysis, and its area under ROC (AUC) was calculated in R software = 0.864 (see Figure 4 ).

[0150] Finally, the calibration graph was drawn, and it was observed that the original curve of the model (apparent) and the correction curve after internal verification of the model (Bias-corrected) in the calibration graph were close to the ideal curve (Ideal) (see Figure 5 ), indicating good agreement between the observed and predicted perioperative high-dose blood transfusions, indicating a high degree of calibration of the model.

[0151] 2. Innovation of the nomogram prediction model

[0152] 2.1 Wide range of application

[0153] At present, most of the research on the nomogram prediction model of blood transfusion in China focuses on a single discipline or a single type of surgical patient. There is a lack of interdisciplinary nomogram prediction models for perioperative red blood cell transfusion risk that can be applied to different types of surgical patients at the same time. This study collected patients of all surgical types except cardiac surgery to screen for transfusion risk and establish a model. This prediction model can be used to predict large-dose blood transfusions in perioperative patients of various different surgical types.

[0154] 2.2 Large number of research samples

[0155] After reviewing relevant literature, the current nomogram prediction model for perioperative transfusion risk has been partially reported in some subspecialty surgeries such as liver surgery, neurosurgery, and hip replacement surgery, and the number of relevant samples used to establish the regression model of the nomogram is small (only 200 to 1,000 cases). This study collected a total of 5,252 non-cardiac surgery patients who received transfusion therapy in the past five years, which is the largest sample size in the study of establishing a transfusion risk nomogram prediction model in China.

[0156] 2.3 Establishing a prediction model for “large-dose” blood transfusion in the blood preparation process for the first time

[0157] At present, the research on perioperative transfusion risks at home and abroad basically focuses on predicting the outcomes of whether surgical patients will receive red blood cell transfusion during the perioperative period, and there are few studies on the amount of red blood cell transfusion in surgical patients during the perioperative period.

[0158] According to the corresponding core medical system "Clinical Blood Use Review System", when the total amount of blood used by a patient on the same day reaches or exceeds 1600ml (i.e. 8U red blood cell suspension) per day, a professional doctor with an intermediate or higher professional title must submit a request, and after approval by the department director, a report must be submitted to the medical department to obtain final permission before blood can be prepared. After reviewing relevant literature on the risk of perioperative blood loss in non-cardiac surgery at home and abroad, there is no report on the risk factors for predicting perioperative bleeding greater than 1600ml, and there is no relevant guideline recommendation for which type of surgery can exceed 8U of preoperative blood. Therefore, whether the preoperative blood volume exceeds 8U is currently mostly based on the subjective judgment of surgeons, lacking objective evidence-based basis and guideline recommendations; and in the approval process of the medical department, there is also a lack of judgment basis for whether to approve the application for 8U of preoperative blood volume. Therefore, in clinical work, for surgical cases with a high expected bleeding risk, it is difficult to effectively avoid the adverse effects of insufficient blood preparation on surgery and patients or the waste of blood resources caused by unused blood.

[0159] According to the above-mentioned "Clinical Blood Use Review System" regulations, this application innovatively defines the amount of blood transfusion (red blood cell suspension) greater than 8U during the perioperative period as "large-dose blood transfusion". By studying the influencing factors of large-dose blood transfusion, a prediction model for perioperative large-dose blood transfusion is established to help clinicians identify surgical patients' transfusion risk factors early before surgery and effectively predict the probability of large-dose blood transfusion in patients during the perioperative period. This will strengthen perioperative blood management in an individualized manner, guide surgeons to implement preoperative blood combination more accurately, provide the blood transfusion department with a basis for blood combination approval, and optimize the blood combination application process.

[0160] 2.4 Good operability in clinical application

[0161] Before using this experimental model to predict large-dose blood transfusion, it is necessary to collect the above six prediction indicators: "operation duration, intraoperative blood loss, preoperative albumin level, preoperative anemia, surgical grade, and ASA anesthesia grade". Among them, "operation duration" and "intraoperative blood loss" are modeled using relevant data in the patient's medical records and surgical records when establishing the model. When using the model, for different surgical subspecialties and different surgical methods, it is necessary to combine the patient's surgical site, lesion range, surgical difficulty, and the surgeon's surgical proficiency to make estimates. In the nomogram model, "operation duration" is represented by "<2h, 2-4h, 4-6h, 6-8h, >8h" as different levels of variables and corresponding prediction scores, and "intraoperative blood loss" is represented by "<1000ml, 1000-2000ml, 2000-3000ml, 3000-4000ml, >4000ml" as different levels of variables and corresponding prediction scores, which provides feasibility and practicality for the use of "estimated operation duration" and "estimated intraoperative blood loss" in clinical use. At the same time, the final nomogram model corresponding score indicator subtitles are also presented in "estimated operation duration" and "estimated blood loss". "Preoperative albumin level" and "preoperative anemia" can directly and objectively obtain the classification results of preoperative albumin level and preoperative anemia through preoperative blood test results. This system is used to evaluate and grade the patient's physical health and surgical risk before anesthesia. These two evaluation indicators are currently the surgical risk assessment and grading indicators commonly used by clinicians in preoperative evaluation. The six predictive factors of this prediction model are easy to obtain and have good operability in clinical application.

[0162] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A nomogram prediction method for perioperative high-dose blood transfusion in non-cardiac surgery, characterized in that: The following steps are involved: S1. Determine the research subjects: adopt a retrospective case-control study method to collect suitable cases of non-cardiac surgery and perioperative blood transfusion therapy for inclusion in the study; S2. Search observation indicators: retrospectively collect the medical records of patients during hospitalization, including the general information of the patients, surgical anesthesia-related information and preoperative blood test results; S3. Univariate analysis between groups: The collected observation indicators were subjected to univariate analysis for inter-group comparison; S4. Multivariate Logistic regression analysis: The statistically significant influencing factors among the single factors between the groups were included in the binary multivariate Logistic regression model. The results of the multivariate regression analysis were used to screen out the independent risk factors associated with perioperative large-dose blood transfusion to construct a nomogram model. S5. Use R software to establish a nomogram: import the independent influencing factors obtained from the multivariate regression analysis into the function, use R software to draw a nomogram, and construct a nomogram prediction model that can predict the risk of large-dose blood transfusion during non-cardiac surgery; S6. Evaluation and validation of the prediction model: The receiver operating characteristic curve was plotted using R software to determine the area under the curve. The discrimination of the nomogram model was measured using the C-index. The calibration graph was plotted to observe the relationship between the original model curve and the calibration curve and ideal curve to evaluate the calibration of the model.

2. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: The collection criteria in step S1 include inclusion criteria and exclusion criteria. The inclusion criteria include patients who underwent non-cardiac surgery during hospitalization and patients who underwent allogeneic red blood cell transfusion during hospitalization. The exclusion criteria include patients who were younger than 15 years old and had incomplete clinical data.

3. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S1, the total transfusion volume of allogeneic red blood cell transfusion therapy for the patient during hospitalization is divided into a non-large-dose transfusion group and a large-dose transfusion group, the total transfusion volume of the non-large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume <8u, and the total transfusion volume of the large-dose transfusion group is an allogeneic red blood cell suspension transfusion volume ≥8u.

4. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S2, the general information includes gender, age, weight, height, BMI, diabetes, hypertension, heart disease, blood transfusion history, allergy history, surgical history, smoking and drinking habits.

5. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S2, the surgical anesthesia-related data include intraoperative blood transfusion volume, surgical department, surgery name, whether it is an emergency surgery, whether it is a day surgery, surgery classification, surgery type, surgery duration, intraoperative blood loss, surgical incision description, ASA anesthesia classification and anesthesia time.

6. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S2, the preoperative blood test results include preoperative hemoglobin, albumin, white blood cell count, platelet count, INR, PT and APTT.

7. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: The steps of univariate analysis between groups in step S3 include: S301. Normal distribution test of continuous variables: All continuous variables are tested for normal distribution using the functions in R language. The test for normal distribution is expressed as ks.test(x, "pnorm")#P>0.05, and the test for non-normal distribution is expressed as P≤0.05; S302. Nonparametric test of continuous variables: Use the function in R language to perform nonparametric test on the continuous variables of the samples. If the difference between the groups is not significant, it is indicated as wilcox.test(tested variable ~ grouping variable, paired = F) #P ≥ 0.

05. If the difference between the groups is significant, it is indicated as P < 0.

05. S303, chi-square test for categorical variables: chi-square test was performed on all categorical variables using the function in R language. The identification field of sample classification was represented as tab1<-table(x, md$whether or not a large-dose blood transfusion)#"whether or not a large-dose blood transfusion was performed", and the insignificant difference between the groups was represented as CrossTable(tab1, chisq=T, expected=T, fisher=T)#P≥0.05, and the significant difference between the groups was represented as P<0.05; S304. Single factor analysis results: Based on the above single factor analysis, single factor analysis results are obtained.

8. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S4, the condition for the existence of statistical significance in the single factor between the groups is the variable with a P value less than 0.05 in the single factor analysis.

9. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S5, the score is marked on the line segment in the Nomogram to represent the influence of the factor, and the sum of each single index is added to obtain the total score of the prediction index, and finally the probability of perioperative large-dose blood transfusion can be obtained.

10. The method for predicting perioperative high-dose blood transfusion in non-cardiac surgery according to claim 1, characterized in that: In step S6, the model was internally validated using R software, a validation set was established using the Bootstrap method of 1000 resamplings with replacement, and the consistency index was calculated using the Harrel-C-statistic function.

Images