Abstract： Objective To evaluate the incidence and prognosis of postoperative acute kidney injury （AKI） in patients after cardiovascular surgery， and analyse the value of AKI criteria and classification using the Acute Kidney Injury Network （AKIN） definition to predict their in-hospital mortality. Methods A total of 1 056 adult patients undergoing cardiovascular surgery in Renji Hospital of School of Medicine， Shanghai Jiaotong University from Jan. 2004 to Jun. 2007 were included in this study. AKI criteria and classification under AKIN definition were used to evaluate the incidence and in-hospital mortality of AKI patients. Univariate and multivariate analyses were used to evaluate preoperative， intraoperative， and postoperative risk factors related to AKI. Results Among the 1 056 patients， 328 patients（31.06%） had AKI. In-hospital mortality of AKI patients was significantly higher than that of non-AKI patients （11.59% vs. 0.69%， P＜0.05）. Multivariate logistic regression analysis suggested that advanced age （OR=1.40 per decade）， preoperative hyperuricemia（OR=1.97）， preoperative left ventricular failure （OR=2.53）， combined CABG and valvular surgery （OR=2.79）， prolonged operation time （OR=1.43 per hour）， postoperative hypovolemia （OR=11.08） were independent risk factors of AKI after cardiovascular surgery. The area under the ROC curve of AKIN classification to predict in-hospital mortality was 0.865 （95% CI 0.801-0.929）. Conclusion Higher AKIN classification is related to higher in-hospital mortality after cardiovascular surgery. Advanced age， preoperative hyperuricemia， preoperative left ventricular failure， combined CABG and valvular surgery， prolonged operation time， postoperative hypovolemia are independent risk factors of AKI after cardiovascular surgery. AKIN classification can effectively predict in-hospital mortality in patients after cardiovascular surgery， which provides evidence to take effective preventive and interventive measures for high-risk patients as early as possible.
Objective To evaluate the predicted value of APACHEⅡ score at admission for deep fungal infection（DFI） in patients with severe acute pancreatitis （SAP）.Methods The clinical data of 132 patients with SAP from January 2006 to June 2011 in our hospital were analyzed retrospectively. The receiver operating characteristic curve （ROC） was used for evaluating the predicted value.Results Thirty-nine patients with SAP infected DFI （29.5%）,of which 36 patients （92.3%） infected with Candida albicans,2 patients （5.1%） with Candida tropicalis,1 patient （2.6%） with pearl bacteria.And,among these 39 patients,27 patients （69.2%） infected at single site,12 patients （30.8%） infected at multi-site. The APACHEⅡ score in 39 patients with DFI was higher than that of 93 patients without DFI （17.1±3.8 versus 9.7±2.1, t=14.316,P=0.000）.The ROC for APACHEⅡ score predicting DFI was 0.745（P=0.000）, 95%CI was 0.641－0.849.When the cut off point was 15,it showed the best forecast performance,with specificity 0.81, sensitivity 0.72,Youden index 0.53. Conclusions The APACHEⅡ score at admission can preferably predict DFI in patients with SAP； when the APACHEⅡ score is greater than 15,it prompts highly possible of DFI,so preventive anti-fungal treatment may be necessary.
ObjectiveTo investigate the predictive factors of portal vein thrombosis (PVT) before and after splenectomy and gastroesophageal devascularization for liver cirrhosis with portal hypertension. MethodsSixty-one cases of liver cirrhosis with portal hypertension who underwent splenectomy and gastroesophageal devascularization were enrolled retrospectively. The patients were divided into PVT group and non-PVT group based on the presence or absence of postoperative PVT on day 7. The clinical factors related with PVT were analyzed. ResultsThere were 25 cases in the DVT group and 36 cases in the non-DVT group. The results of univariate analysis showed that the preoperative platelet (P=0.006), activated partial thromboplastin time (P=0.048), prothrombin time (P=0.028), and international normalized ratio (P=0.029), postoperative fibrin degradation product (P=0.002) and D-dimer (P=0.014) on day 1, portal venous diameter (P=0.050) had significant differences between the DVT group and non-DVT group. The results of logistic multivariate regression analysis showed that the preoperative platelet (OR=0.966, 95% CI 0.934-1.000, P=0.048) and postoperative fibrin degradation product on day 1(OR=1.055, 95% CI 1.011-1.103, P=0.017) were correlated with the PVT. The PVT might happen when preoperative platelet was less than 34.5×109/L (sensitibity 80.6%, specificity 60.0%) or postoperative fibrin degradation product on day 1 was more than 64.75 mg/L (sensitibity 48.0%, specificity 91.7%). ConclusionPreoperative platelet and postoperative fibrin degradation product on day 1 might predict PVT after splenectomy and gastroesophageal devascularization for liver cirrhosis with portal hypertension.
ObjectiveTo explore the early predictive value of Wells score and D-dimer for acute pulmonary embolism. MethodsEighty-two cases with acute pulmonary embolism comfirmed by computed tomography pulmonary angiography and (or) lung ventilation/perfusion scan were retrospectively studied from October 2013 to October 2014 in our hospital. Another 82 cases without acute pulmonary embolism in the chest pain center simultaneously were selected as control group. The data on admission were analyzed including Wells score, D-dimer, pH, PCO2, PO2, P(A-a)O2, brain natriuretic peptide, troponin I of two groups of patients. Relevant variables were selected by multivariate logistic regression analysis. The receiver operating characteristic (ROC) curve was made by sensitivity as the ordinate and 1 minus specificity as abscissa. The area under ROC curve (AUC) for relevant variables was calculated and the variable with higher AUC was selected. The best threshold, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were achieved from the ROC curves. ResultsThe multivariate logistic regression analysis showed that Wells score (OR=8.114, 95%CI 1.894-34.761, P=0.005) and D-dimer (OR=1.009, 95%CI 1.001-1.017, P=0.021) could predict APE early. The AUC, sensitivity, specificity, PPV, NPV of Wells score for the early prediction of patients with acute pulmonary embolism were 0.990, 50.0%, 100.0%, 100.0%, 66.7%, respectively. The AUC, sensitivity, specificity, PPV, NPV of D-dimer for the early prediction of patients with acute pulmonary embolism were 0.986, 95.1%, 97.6%, 97.5%, 95.2%, respectively. ConclusionWells score and D-dimer have high predictive value in patients with acute pulmonary embolism, and can be used in preliminary screening of acute pulmonary embolism in the emergency department.
ObjectiveTo evaluate clinical value of colon leakage score (CLS), a preoperative predictive scoring system, for risk of anastomotic leakage after left-sided colorectal cancer surgery. MethodsThe clinical data of 310 patients who underwent left-sided colorectal cancer surgery from January 2010 to December 2014 were studied retrospectively. Risk factors for postoperative anastomotic leakage were analyzed by univariate analysis. The sensitivity and specificity of CLS system were determined by receiver operating characteristic (ROC) curve analysis. Resultsa total of 14 patients were diagnosed as anastomotic leakage. The point of CLS for the patients with anastomotic leakage was significantly higher than that for the patients without anastomotic leakage (14.21±5.76 versus 4.43±3.36, t=9.474, P=0.000). The results of ROC curve analysis showed that the sensitivity and specificity of the CLS system were 92.9% and 88.6%, respectively. The area under the curve was 0.957 (95% CI 0.924-0.991). The best cut off value of CLS was 10 (The Youden index was 0.867). The results of univariate analysis showed that the age, preoperative hemoglobin level, status of intestinal obstruction, and blood loss were associated with postoperative anastomotic leakage (P<0.05). ConclusionThe preoperative predictive score system CLS could accurately predict occurrence of anastomotic leakage. While large, multicenter prospective randomized controlled trial is still needed to further confirm it.
Objective To explore the predicted precision of discharged patients number using curve estimation combined with trend-season model. Methods Curve estimation and trend-season model were both applied, and the quarterly number of discharged patients of 363 hospital from 2009 to 2015 was collected and analyzed in order to predict discharged patients in 2016. Relative error between predicted value and actual number was also calculated. Results An optimal quadratic regression equation Yt=3 006.050 1+202.350 8×t–3.544 4×t2 was established (Coefficient of determination R2=0.927, P<0.001), and a total of 23 462 discharged patients were predicted based on this equation combined with trend-season model, with a relative error of 1.79% compared to the actual number. Conclusion The curve estimation combined with trend-season model is a convenient and visual tool for predicting analysis. It has a high predicted accuracy in predicting the number of hospital discharged patients or outpatients, which can provide a reference basis for hospital operation and management.
ObjectivesTo estimate the latest burden of disability adjusted life years (DALYs) for liver cancer in China and the long-term trend, and to make future prediction.MethodsBased on the visualization platform of Global Burden of Disease 2016, data on the DALYs for liver cancer in China was extracted. The very recent status in 2016 and the previous trend from 1990 to 2016 were described, using annualized rate of change (ARC). The burden from 2017 to 2050 was further predicted by combining the ARC and the Chinese population data projected by the United Nation.ResultsIn 2016, the total DALYs for liver cancer in China was estimated as 11 539 000 person years (accounting for 54.6% of the global burden), and years of life lost (YLLs) and years lived with disability (YLDs) contributed 98.9% and 1.1%, respectively. The age-standardized DALY rate was 844.1 per 100 000 (3.0 times of the global average) and the male-to-female ratio was 3.4. The DALY rate continuously increased from 1990–2016 (ARC=0.57%), particularly in recent 5 years (ARC=1.75%). Among the DALYs for all cancers, liver cancer contributed approximately 20% and constantly remained as the top 2 (ranking as the number one before year 2005). There were inverse trends in gender, with increasing in males and decreasing in females (ARC was 0.77% and –0.11%, respectively). Hepatitis B infection continually kept the leading cause of DALYs for liver cancer (accounting for nearly 57%), and the DALY rate was gradually increasing (ARC=0.43%). Although the peak age of DALY rate was stable at 65to 69 years, the peak age of the DALYs changed from 55 to 59 years in 1990 to 60 ~ 64 years in 2016. In 2050, the estimated DALYs for liver cancer in China will reach 14.37 million person years, 20.0% more than that in 2017.ConclusionsThe DALYs caused by liver cancer in China exceeds the overall burden of all other countries in the world, and accounts for 1/5 of DALYs for all cancers in local population. The burden in males has been continuously rising, and the leading cause remained unchanged as hepatitis B infection. With population aging, the DALYs for liver cancer in China will be incessant to increase, suggesting the necessity to implement continuous effort in risk factors prevention (e.g. hepatitis B infection), and efficient management in high risk population of liver cancer.
ObjectivesTo explore the construction method of prediction model of absolute risk for breast cancer and provide personalized breast cancer management strategies based on the results.MethodsA case-control design was conducted with 2 747 individuals diagnosed as primary breast cancer by pathology in West China Hospital of Sichuan University from 2000 to 2017 and 6 307 healthy controls from Breast Cancer Screening Cohort in Sichuan Women and Children Center and Chengdu Shuangliu District Maternal and Child Health Hospital. Standardized questionnaires and information management systems in hospital were used to collect information. Decision trees, logistic regression, the formula in Gail model and registration data in China were used to estimate the probability of 5-year risk of breast cancer. Eventually a ROC (receiver operating characteristics) curve was drawn to identify optimal cut-off value, and the power was evaluated.ResultsThe decision tree exported 4 variables, which were urban or rural sources, number of live birth, age and age at menarche. The median 5-year risk and interquartile range of the controls was 0.027% and 0.137%, while the median 5-year risk and interquartile range of the cases was 0.219% and 0.256%. The ROC curve showed the cut-off value was 0.100%. Through verification, the sensitivity was 0.79, the specificity was 0.73, the accuracy was 0.75, and the AUC (area under the curve) was 0.79.ConclusionsThe methods used in our study based on 9 054 female individuals in Sichuan province could be used to predict the 5-year risk for breast cancer. Predictor variables include urban or rural sources, number of live birth, age, and age at menarche. If the 5-year risk is more than 0.100%, the person will be judged as a high risk individual.
ObjectiveTo establish a forecasting model for inpatient cases of pediatric limb fractures and predict the trend of its variation.MethodsAccording to inpatient cases of pediatric limb fractures from January 2013 to December 2018, this paper analyzed its characteristics and established the seasonal auto-regressive integrated moving average (SARIMA) model to make a short-term quantitative forecast.ResultsA total of 4 451 patients, involving 2 861 males and 1 590 females were included. The ratio of males to females was 1.8 to 1, and the average age was 5.655. There was a significant difference in age distribution between males and females (χ2=44.363, P<0.001). The inpatient cases of pediatric limb fractures were recorded monthly, with predominant peak annually, from April to June and September to October, respectively. Using the data of the training set from January 2013 to May 2018, a SARIMA model of SARIMA (0,1,1)(0,1,1)12 model (white noise test, P>0.05) was identified to make short-term forecast for the prediction set from June 2018 to November 2018, with RMSE=8.110, MAPE=9.386, and the relative error between the predicted value and the actual value ranged from 1.61% to 8.06%.ConclusionsCompared with the actual cases, the SARIMA model fits well with good short-term prediction accuracy, and it can help provide reliable data support for a scientific forecast for the inpatient cases of pediatric limb fractures.
ObjectiveTo establish a predictive model of surgical site infection (SSI) following colorectal surgery using machine learning.MethodsMachine learning algorithm was used to analyze and model with the colorectal data set from Duke Infection Control Outreach Network Surveillance Network. The whole data set was divided into two parts, with 80% as the training data set and 20% as the testing data set. In order to improve the training effect, the whole data set was divided into two parts again, with 90% as the training data set and 10% as the testing data set. The predictive result of the model was compared with the actual infected cases, and the sensitivity, specificity, positive predictive value, and negative predictive value of the model were calculated, the area under receiver operating characteristic (ROC) curve was used to evaluate the predictive capacity of the model, odds ratio (OR) was calculated to tested the validity of evaluation with a significance level of 0.05.ResultsThere were 7 285 patients in the whole data set registered from January 15th, 2015 to June 16th, 2016, among whom 234 were SSI cases, with an incidence of SSI of 3.21%. The predictive model was established by random forest algorithm, which was trained by 90% of the whole data set and tested by 10% of that. The sensitivity, specificity, positive predictive value, and negative predictive value of the model were 76.9%, 59.2%, 3.3%, and 99.3%, respectively, and the area under ROC curve was 0.767 [OR=4.84, 95% confidence interval (1.32, 17.74), P=0.02].ConclusionThe predictive model of SSI following colorectal surgery established by random forest algorithm has the potential to realize semi-automatic monitoring of SSIs, but more data training should be needed to improve the predictive capacity of the model before clinical application.