Risk prediction method for formation of postoperative blood hypercoagulable state of total knee arthroplasty patient

Abstract

The invention discloses a risk prediction method for postoperative blood hypercoagulable state formation of total knee arthroplasty patients, which comprises the following steps of: obtaining a factor screening table conforming to clinical practical work based on hospital medical record data of total knee arthroplasty patients caused by osteoarthritis in the orthopedics department through literature evidence-based query of previous research data; a nomogram risk prediction model with good distinction degree, accuracy and practicability is constructed by using a statistical means; according to the nomogram risk prediction method, the total score of the patient can be obtained by adding the scores of the risk factors, the corresponding risk probability is finally found on the nomogram, and the nomogram risk prediction method can be used for estimating the occurrence rate of the postoperative blood hypercoagulable state of the patient, so that efficient prediction and advanced intervention are realized, the process is easy to understand, the chart operation is simple, the generalizability is high, and the cost is low. And in addition, an electronic storage product can be imported, so that patients and medical personnel can conveniently call out and evaluate at any time.

Description

Technical Field

[0001] The present invention relates to a risk prediction method, and in particular to a risk prediction method for the formation of a hypercoagulable state in blood after total knee replacement surgery. Background Art

[0002] Total knee replacement surgery has become the main surgical method for treating end-stage knee osteoarthritis. This technology can effectively improve patients' symptoms and improve their quality of life, and the number of surgeries has continued to grow. However, total knee replacement surgery is a common major orthopedic surgery. During total knee replacement, patients' physical activities are limited, and the time they spend in bed is increased. In addition, their own physical function is impaired, which can easily lead to hypercoagulation of the blood and further form venous thrombosis, resulting in delayed postoperative recovery, prolonged hospitalization, increased economic and physical and mental burdens on patients, and even severe cases may develop into pulmonary embolism, endangering life safety.

[0003] The Caprini score is currently the most widely used scale for orthopedic thrombosis risk assessment. However, it assigns a large weight to the type of surgery and underestimates the scores of other important risk factors. It cannot provide accurate clinical risk grading for total knee replacement patients and is not good at distinguishing high-risk individuals, making it difficult to focus on identifying truly high-risk patients.

[0004] Timely identification of platelet activation and accurate prediction of hypercoagulable state are the key to thrombosis prevention. Clinical prediction model tools have been widely used in the field of clinical medicine in recent years. By building a prediction model, the risk of individual adverse events can be predicted. However, there is still a lack of accurate and effective assessment tools for the risk of hypercoagulable state after total knee replacement. Summary of the invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a risk prediction method for the formation of hypercoagulable state after total knee replacement surgery, to establish a reliable and easy-to-use prediction model for hypercoagulable state after total knee replacement surgery, to provide an assessment tool for early identification of high-risk patients, and to intervene in advance, thereby reducing the incidence of hypercoagulable state and reducing the occurrence of adverse postoperative complications.

[0006] Technical solution: The method for predicting the risk of hypercoagulable state in patients after total knee replacement surgery of the present invention comprises the following steps:

[0007] (1) Construct a potential risk factor screening table through literature-based evidence-based methods combined with clinical practice;

[0008] (2) Based on the hospital medical record system, qualified medical records were screened according to the sample inclusion and exclusion criteria to construct the original data set;

[0009] (3) The original data set was pre-standardized and randomly split into a training set and a validation set using R language software, with the randomization ratio of the training set to the validation set set being 7:3;

[0010] (4) The data in the training set were imported into SPSS software, and items that met the statistical significance requirement (p < 0.05) were preliminarily screened out through univariate analysis, and then independent risk factors were finally obtained through binary logistic regression;

[0011] (5) The independent risk factors were imported into the R language software. The R package was loaded and the nomogram function was used to construct the nomogram risk prediction model. The model was evaluated by drawing ROC curves, calibration curves, and decision curves, and the validation set data was used for verification.

[0012] Preferably, the items collected in the qualified medical records in step 2 include:

[0013] (1) General demographic information: gender, age, height, weight, drinking history, hypertension, diabetes, and coronary heart disease;

[0014] (2) Preoperative individual assessment: preoperative systolic blood pressure, preoperative diastolic blood pressure, American Society of Anesthesiologists score, and self-care ability score;

[0015] (3) Intraoperative data: operation time, intraoperative blood loss, and anesthesia method;

[0016] (4) Laboratory examination indicators: glucose, creatinine, C-reactive protein, red blood cells, hemoglobin, hematocrit, platelet count, prothrombin time, PT international normalized ratio, activated partial thromboplastin time, thrombin time, fibrinogen, D-dimer, erythrocyte sedimentation rate, and platelet width.

[0017] Preferably, step 2 further comprises calculating the patient's body mass index (BMI) according to the height and weight, and classifying the patient into underweight, normal weight, overweight and obesity according to the standard.

[0018] Preferably, the sample selection criteria in step 2 are as follows:

[0019] (1) Age ≥ 40 years;

[0020] (2) patients undergoing total knee replacement surgery due to osteoarthritis;

[0021] (3) Complete medical records;

[0022] (4) The operation was performed by the same surgical team.

[0023] Preferably, the sample rejection criteria in step 2 are as follows:

[0024] (1) Patients transferred to other hospitals or departments during the course of treatment;

[0025] (2) thrombosis existed before admission;

[0026] (3) Patients with coagulation disorders.

[0027] Preferably, the pre-standardization processing method in step 3 is to check for inevitable missing values, supplement the missing data with the average, and perform numbering processing on the categorical data.

[0028] Preferably, the independent risk factors in step 4 include preoperative hospitalization time, hypertension, diabetes, coronary heart disease, platelet count, intraoperative blood loss, and fibrinogen.

[0029] Preferably, the specific steps of univariate analysis in step 4 are as follows:

[0030] (41) The training set data were imported into SPSS software, and the measurement data were expressed as mean ± standard deviation. The independent sample t test was used for comparison between groups;

[0031] (42) Enumeration data were expressed as percentages, and inter-group comparisons were performed using the chi-square test or Fishers exact test;

[0032] (43) Items that meet the statistical significance requirement of p < 0.05 were screened out from the measurement data and count data;

[0033] The measurement data included age, BMI, self-care ability score, preoperative hospital stay, operation time, intraoperative blood loss, preoperative systolic blood pressure, preoperative diastolic blood pressure, glucose, creatinine, C-reactive protein, red blood cells, hemoglobin, hematocrit, platelet count, prothrombin time, PT international normalized ratio, activated partial thromboplastin time, thrombin time, fibrinogen, D-dimer, and erythrocyte sedimentation rate;

[0034] The counting data included gender, drinking history, hypertension, diabetes, coronary heart disease, American Society of Anesthesiologists score, and anesthesia method.

[0035] Preferably, the nomogram risk prediction model formula in step 5 is: -1.297+(-0.589×preoperative hospitalization time)+(-1.095×presence of hypertension)+1.131×presence of diabetes+1.648×presence of coronary heart disease+0.010×intraoperative bleeding volume+(-0.290×platelet count)+0.644×fibrinogen.

[0036] Preferably, the R package in step 5 includes a pRoc package for analyzing and drawing ROC curves, an rmda package for drawing decision curves, and an rms package and an Hmisc package for building models and drawing calibration curves.

[0037] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: 1. By establishing a risk prediction model, it can accurately assess the risk of postoperative hypercoagulation in patients early; 2. Early identification of high-risk patients, early prevention and treatment can effectively avoid the transformation of hypercoagulation to thrombosis, reduce the occurrence of adverse postoperative concomitant diseases, and facilitate patient recovery; 3. It provides orthopedic surgeons with more practical and specific risk assessment tools to help them develop personalized prevention and treatment plans based on the individual risk conditions of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the nomogram prediction model of the present invention;

[0039] Figure 2 It is a schematic diagram of the ROC curve of the present invention;

[0040] Figure 3 It is a schematic diagram of the calibration curve of the present invention;

[0041] Figure 4 It is a schematic diagram of the decision curve of the present invention. DETAILED DESCRIPTION

[0042] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0043] 1. Through CNKI, VIP, Wanfang, PubMed, Web of Science and other databases, based on the search terms formed by the search formula, comprehensively review the published literature, use evidence-based methods to summarize the risk factors, and form a potential risk factor screening table;

[0044] The electronic medical records of the patients were accessed through the hospital medical record system. The sample selection criteria were as follows: (1) age ≥ 40 years; (2) patients undergoing total knee replacement surgery due to osteoarthritis; (3) complete medical records; (4) the surgery was completed by the same surgical team, and the sample exclusion criteria were as follows: (1) patients transferred to other hospitals or departments during the operation; (2) patients with thrombosis before admission; (3) patients with coagulation disorders for screening. The medical records of eligible patients undergoing total knee replacement treatment for osteoarthritis were collected and the data were entered on the spot using an Excel spreadsheet. Afterwards, two people checked the data for accuracy and for missing and abnormal values.

[0045] The general demographic data, preoperative individual assessment, intraoperative data, and laboratory test indicators of the patients were collected for retrospective analysis. A total of 490 patients were included, and 30 items were collected, including: gender, age, height, weight, drinking history, hypertension, diabetes, coronary heart disease, preoperative systolic blood pressure, preoperative diastolic blood pressure, American Society of Anesthesiologists score, self-care ability score, operation time, intraoperative blood loss, anesthesia method, glucose, creatinine, C-reactive protein, red blood cells, hemoglobin, hematocrit, platelet count, prothrombin time, PT international normalized ratio, activated partial thromboplastin time, thrombin time, fibrinogen, D-dimer, erythrocyte sedimentation rate, platelet width, to construct the original data set;

[0046] 2. The original data set was pre-standardized, missing values ​​were supplemented with the mean, categorical variables were numbered, and the original data set was randomly split into training set and validation set using R language software. The randomization ratio of training set and validation set was set to 7:3. The training set was used for modeling and the validation set for verification. The training set was divided into hypercoagulable group and non-hypercoagulable group.

[0047] 3. SPSS27.0 statistical software was used for data analysis. The measurement data were expressed as mean ± standard deviation, and the independent sample t-test was used for inter-group comparison. The count data were expressed as percentages, and the chi-square test or Fishers exact test was used for inter-group comparison. Items that met the statistical significance requirement, i.e. p < 0.05, were entered into binary logistic regression analysis to obtain the independent risk factors for the formation of hypercoagulable state after total knee replacement.

[0048] The statistical results showed that 11 items including age, preoperative hospital stay, intraoperative blood loss, platelet count, activated partial thromboplastin time, fibrinogen, preoperative systolic blood pressure, gender, hypertension, diabetes, and coronary heart disease met the statistical significance requirement, i.e., p < 0.05, as shown in Table 1 for details.

[0049] Binary logistic regression analysis showed that preoperative hospitalization time, hypertension, diabetes, coronary heart disease, platelet count, intraoperative blood loss, and fibrinogen were independent risk factors for the development of hypercoagulable state in patients after total knee replacement (see Table 2 for details).

[0050]

[0051]

[0052]

[0053] Table 1

[0054]

[0055] Table 2

[0056] 4. Import the independent risk factor data into R language software to construct a risk prediction model for the formation of hypercoagulable state after total knee replacement, and visualize the results through nomograms. Figure 1 ; Draw the ROC curve, calibration curve and decision curve of the nomogram risk prediction model to evaluate the discrimination, accuracy and clinical effectiveness of the prediction model.

[0057] The results showed that the AUC value of the training set was 0.898, the AUC value of the validation set was 0.896, and the average AUC value was greater than 0.7, which proved that the nomogram risk prediction model had good discrimination. Figure 2 ; Draw the calibration curve of the nomogram. The calibration curve and the ideal curve have good consistency, indicating that the model has good accuracy. For details, see Figure 3 ; Decision curve analysis is a simple method for evaluating the application value of various researches such as clinical prediction models and diagnostic tests. DCA analysis was performed on the data in the modeled nomogram. From the DCA curve, it can be seen that in the threshold range of 0 to 1, the curve is higher than the reference line, indicating that the net benefit value of the model is high, suggesting that the nomogram prediction model has good clinical effectiveness. For details, see Figure 4 .

[0058] The clinical test values ​​of the patients are collected and input, and imported into the nomogram risk prediction model. In the nomogram risk prediction model, according to the specific conditions of the patients after total knee replacement, the clinical test values ​​include: preoperative hospitalization time, hypertension, diabetes, coronary heart disease, intraoperative blood loss, platelet count PLT, fibrinogen FIB, and each item is selected to have a score value. The scores of the score scale corresponding to all the selected score values ​​are added to obtain a total score, and the risk value of the patient's blood hypercoagulable state after total knee replacement corresponding to the total score is determined as the risk probability of the patient's blood hypercoagulable state after total knee replacement;

[0059] Preoperative hospitalization time module is used to determine the patient's preoperative hospitalization time, with an optional range of 1-8;

[0060] The hypertension module is used to determine whether the patient has hypertension, if not, it is 0, if yes, it is 1;

[0061] The diabetes module is used to determine whether the patient has diabetes, if not, it is 0, if yes, it is 1;

[0062] The coronary heart disease module is used to determine whether the patient has coronary heart disease, if not, it is 0, if yes, it is 1;

[0063] The intraoperative bleeding volume module is used to determine the patient's intraoperative bleeding volume, with an optional range of 20-800;

[0064] Platelet count PLT module, used to determine the patient's platelet count, with an optional range of 50-500;

[0065] Fibrinogen FIB module, used to determine the patient's fibrinogen, with an optional range of 2-8.5;

[0066] The total score module is used to add up the scores of all score scales, with an optional range of 0-220;

[0067] The outcome module is used to determine the risk of patients developing a hypercoagulable state, with an optional range of 0.001 to 0.999;

[0068] For example, a patient who underwent total knee replacement had been hospitalized for 3 days before surgery and suffered from diabetes. During the surgery, he lost 200 ml of blood and his platelet count was 150×10 9 / L, the fibrinogen test value was 8g / L. According to the nomogram prediction model: when the preoperative hospitalization time = 3, the corresponding score was about 23 points; when there was diabetes, the corresponding score was about 10 points; when the intraoperative blood loss = 200, the corresponding score was about 14 points; when the platelet count = 150, the corresponding score was about 78 points; when the fibrinogen = 8, the corresponding score was about 28 points.

[0069] Therefore, the total score can be calculated as: 23+10+14+78+28=153 points. According to the model, the probability of hypercoagulable risk outcome corresponding to the total score of 153 is close to 0.9 (90%), indicating that the postoperative risk is high.

Claims

1. A method for predicting the risk of hypercoagulable state in patients after total knee replacement surgery, characterized in that: The following steps are involved: (1) Construct a potential risk factor screening table through literature-based evidence-based methods combined with clinical practice; (2) Based on the hospital medical record system, qualified medical records were screened according to the sample inclusion and exclusion criteria to construct the original data set; (3) The original data set was pre-standardized and randomly split into a training set and a validation set using R language software, with the randomization ratio of the training set to the validation set set being 7:3; (4) The data in the training set were imported into SPSS software, and items that met the statistical significance requirement (p < 0.05) were preliminarily screened out through univariate analysis, and then independent risk factors were finally obtained through binary logistic regression; (5) The independent risk factors were imported into the R language software. The R package was loaded and the nomogram function was used to construct the nomogram risk prediction model. The model was evaluated by drawing ROC curves, calibration curves, and decision curves, and the validation set data was used for verification.

2. The risk prediction method according to claim 1, characterized in that: The items collected in the qualified medical records in step 2 include: (1) General demographic information: gender, age, height, weight, drinking history, hypertension, diabetes, and coronary heart disease; (2) Preoperative individual assessment: preoperative systolic blood pressure, preoperative diastolic blood pressure, American Society of Anesthesiologists score, and self-care ability score; (3) Intraoperative data: operation time, intraoperative blood loss, and anesthesia method; (4) Laboratory examination indicators: glucose, creatinine, C-reactive protein, red blood cells, hemoglobin, hematocrit, platelet count, prothrombin time, PT international normalized ratio, activated partial thromboplastin time, thrombin time, fibrinogen, D-dimer, erythrocyte sedimentation rate, and platelet width.

3. The risk prediction method according to claim 1, characterized in that: Step 2 also includes calculating the patient's body mass index (BMI) based on height and weight, and categorizing it into underweight, normal weight, overweight, and obesity according to the standard.

4. The risk prediction method according to claim 1, characterized in that: The sample selection criteria described in step 2 are as follows: (1) Age ≥ 40 years; (2) patients undergoing total knee replacement surgery due to osteoarthritis; (3) Complete medical records; (4) The operation was performed by the same surgical team.

5. The risk prediction method according to claim 1, characterized in that: The sample elimination criteria in step 2 are as follows: (1) Patients transferred to other hospitals or departments during the course of treatment; (2) thrombosis existed before admission; (3) Patients with coagulation disorders.

6. The risk prediction method according to claim 1, characterized in that: The pre-standardization method described in step 3 is to check for unavoidable missing values, use the average to supplement the missing data, and number the categorical data.

7. The risk prediction method according to claim 1, characterized in that: The independent risk factors described in step 4 include preoperative hospital stay, hypertension, diabetes, coronary heart disease, platelet count, intraoperative blood loss, and fibrinogen.

8. The risk prediction method according to claim 1, characterized in that: The specific steps of univariate analysis in step 4 are as follows: (41) The training set data were imported into SPSS software, and the measurement data were expressed as mean ± standard deviation. The independent sample t test was used for comparison between groups; (42) Enumeration data were expressed as percentages, and inter-group comparisons were performed using the chi-square test or Fishers exact test; (43) Items that meet the statistical significance requirement of p < 0.05 were screened out from the measurement data and count data; The measurement data included age, BMI, self-care ability score, preoperative hospital stay, operation time, intraoperative blood loss, preoperative systolic blood pressure, preoperative diastolic blood pressure, glucose, creatinine, C-reactive protein, red blood cells, hemoglobin, hematocrit, platelet count, prothrombin time, PT international normalized ratio, activated partial thromboplastin time, thrombin time, fibrinogen, D-dimer, and erythrocyte sedimentation rate; The counting data included gender, drinking history, hypertension, diabetes, coronary heart disease, American Society of Anesthesiologists score, and anesthesia method.

9. The risk prediction method according to claim 1, characterized in that: The formula of the nomogram risk prediction model described in step 5 is: -1.297 + (-0.589 × preoperative hospitalization time) + (-1.095 × presence or absence of hypertension) + 1.131 × presence or absence of diabetes + 1.648 × presence or absence of coronary heart disease + 0.010 × intraoperative blood loss + (-0.290 × platelet count) + 0.644 × fibrinogen.

10. The risk prediction method according to claim 1, characterized in that: The R packages described in step 5 include the pRoc package for analyzing and drawing ROC curves, the rmda package for drawing decision curves, the rms package and the Hmisc package for building models and drawing calibration curves.

Images