Objective To construct a risk prediction model for post-stroke dysphagia(PSD)based on clinical and laboratory indicators of elderly stroke patients with explainable machine learning.Methods A retrospective analysis was conducted on 3994 stroke patients hospitalized in Depart-ment of Neurology of First Medical Center of Chinese PLA General Hospital from October 2010 to December 2021.Among them,the 1390 cases admitted during January 2019 and December 2021 were assigned into an external validation set,and the 2604 cases admitted during October 2010 to January 2019 were into a training group.Those from the training group were further divided into a training set(1823 cases)and an internal validation set(781 cases)in a 7∶3 ratio,and also grouped into a PSD subgroup(773 cases)and a non-PSD group(1831 cases).With occurrence of swallowing difficulties as an endpoint,risk prediction models were constructed using random for-est(RF),eXtreme Gradient Boosting(XGBoost),Support Vector Machine(SVM),and logistic regression.ROC curve analysis was employed to evaluate the performance of our models.After the optimal model was selected,SHAP was employed to interpret feature contributions.Results There were significant differences in muscle strength,right/left-sided stroke,and area of brain in-jury between the PSD and the non-PSD groups(P＜0.01).The PSD group had obviously larger proportions of hypertension,diabetes,and drinking history,increased neutrophil counts,and de-creased levels of potassium and albumin when compared with the non-PSD group(P＜0.05,P＜0.01).Multivariate logistic regression analysis showed that age,drinking history,diabetes,hyper-tension,muscle strength grade,area of brain injury,hemispheric stroke,neutrophil count,and al-bumin and potassium levels were risk factors for PSD(P＜0.05,P＜0.01).The external validation results showed that the area under curve value of the RF model,XGBoost model,SVM model,and our logistic model was 0.883,0.902,0.877,and 0.868,respectively.The distribution of SHAP value showed that drinking history,hypertension and diabetes were positively correlated with PSD risk;Muscle strength was negatively correlated with the risk;Age growth was positively correlated with the risk;Subtentorial lesions showed stronger predictive efficacy than supratentorial lesions and entire lesions;The bilateral and right-sided stroke had higher risk for PSD than the left-sided stroke.Conclusion The model based on the XGBoost model shows best performance in predicting the risk for swallowing disorders in elderly patients after stroke.

Objective To construct a risk prediction model for post-stroke dysphagia(PSD)based on clinical and laboratory indicators of elderly stroke patients with explainable machine learning.Methods A retrospective analysis was conducted on 3994 stroke patients hospitalized in Depart-ment of Neurology of First Medical Center of Chinese PLA General Hospital from October 2010 to December 2021.Among them,the 1390 cases admitted during January 2019 and December 2021 were assigned into an external validation set,and the 2604 cases admitted during October 2010 to January 2019 were into a training group.Those from the training group were further divided into a training set(1823 cases)and an internal validation set(781 cases)in a 7∶3 ratio,and also grouped into a PSD subgroup(773 cases)and a non-PSD group(1831 cases).With occurrence of swallowing difficulties as an endpoint,risk prediction models were constructed using random for-est(RF),eXtreme Gradient Boosting(XGBoost),Support Vector Machine(SVM),and logistic regression.ROC curve analysis was employed to evaluate the performance of our models.After the optimal model was selected,SHAP was employed to interpret feature contributions.Results There were significant differences in muscle strength,right/left-sided stroke,and area of brain in-jury between the PSD and the non-PSD groups(P＜0.01).The PSD group had obviously larger proportions of hypertension,diabetes,and drinking history,increased neutrophil counts,and de-creased levels of potassium and albumin when compared with the non-PSD group(P＜0.05,P＜0.01).Multivariate logistic regression analysis showed that age,drinking history,diabetes,hyper-tension,muscle strength grade,area of brain injury,hemispheric stroke,neutrophil count,and al-bumin and potassium levels were risk factors for PSD(P＜0.05,P＜0.01).The external validation results showed that the area under curve value of the RF model,XGBoost model,SVM model,and our logistic model was 0.883,0.902,0.877,and 0.868,respectively.The distribution of SHAP value showed that drinking history,hypertension and diabetes were positively correlated with PSD risk;Muscle strength was negatively correlated with the risk;Age growth was positively correlated with the risk;Subtentorial lesions showed stronger predictive efficacy than supratentorial lesions and entire lesions;The bilateral and right-sided stroke had higher risk for PSD than the left-sided stroke.Conclusion The model based on the XGBoost model shows best performance in predicting the risk for swallowing disorders in elderly patients after stroke.

BACKGROUND:Scaphoid fractures are common wrist fractures.The percutaneous screw is used to treat scaphoid fractures.The closed reduction of the fracture site,the precise placement of the hollow compression screw in the scaphoid axis and the compression fixation of the fracture site can promote fracture healing and achieve the better function of the wrist. OBJECTIVE:To evaluate the efficacy of percutaneous hollow compression screw fixation of scaphoid fractures using a novel Kirschner wire guide. METHODS:Between January 2015 and December 2020,clinical data from 15 patients with scaphoid fractures selected at the Department of Hand Surgery,Third Hospital of Hebei Medical University were retrospectively analyzed.All patients underwent percutaneous hollow compression screw fixation by a novel Kirschner wire guide.Fracture healing time,operation time,number of intraoperative fluoroscopies,time to return to work and complications were collected.According to the modified Mayo wrist scoring system,the functional outcomes of wrists were assessed 12 months after surgery.Wrist flexion,extension,ulnar deviation,radial deviation and grip strength were measured. RESULTS AND CONCLUSION:All patients were followed up for 12 months.The compression screw was located in the axial position of the scaphoid by routine immediate postoperative radiographs.All scaphoid fractures united at an average of 10.0 weeks.The average operation time was 55.7 minutes.The number of intraoperative fluoroscopies was 10.9.The average time to return to work was 10.3 weeks.The results of wrist joint function were excellent in 9 cases,good in 5 cases,and average in 1 case,with an excellent and good rate of 93.3%.No significant differences in wrist flexion,extension,ulnar deviation,radial deviation and grip strength were found between the affected and healthy sides(P＞0.05).None of the patients had wound infection,malunion,screw displacement,or screw breakage.These findings indicate that the application of a novel guide can shorten the operation time,reduce the number of intraoperative fluoroscopies,improve the accuracy of screw insertion,and shorten the time to return to work.The function of the wrist was satisfactory after the operation.The novel Kirschner wire guide made percutaneous hollow compression screw fixation easier for scaphoid fractures.

Objective To investigate the early resuscitation effect of hemoglobin-based oxygen carriers (HBOC) in rats with uncontrolled hemorrhagic shock.Methods 170 Sprague-Dawley (SD) rats were randomly divided into five groups:lactate Ringer solution (LR) control group,whole blood control group,and 0.5％,2.0％,5.0％ HBOC groups,with 34 rats in each group.The uncontrolled hemorrhagic shock model in SD rats was reproduced by cutting off the splenic artery branch,and induced mean arterial pressure (MAP) reducing to 40 mmHg (1 mmHg =0.133 kPa).The corresponding solution was infused after model reproduction in each group,maintaining MAP at 50 mmHg for 1 hour,then completely ligating and hemostasis,and maintaining MAP at 70 mmHg for 1 hour and 80 mmHg for 1 hour respectively,after maintaining MAP 80 mmHg,all were supplemented with LR to 2 times blood loss volume.The survival rate and blood loss rate were observed in 16 rats in each group,hemodynamics parameters including MAP,left ventricular systolic pressure (LVSP) and the maximum rate of left ventricular pressure rise (+dp/dtmax) were determined in another 10 rats,and cardiac output (CO) and tissue oxygen supply (DO2) were observed in the rest 8 rats.Results ① When resuscitation by LR alone,the blood loss rate of animals was as high as 60％ to 70％.Compared with the LR control group,whole blood recovery could significantly reduce the blood loss rate before hemostasis in uncontrolled hemorrhagic shock rats [(46.6 ± 4.5)％ vs.(62.3 ± 4.0)％,P ＜ 0.01];0.5％,2.0％,5.0％ HBOC could significantly decrease the blood loss rate,especially in 5.0％ HBOC group with significant difference as compared with that in the LR control group [(45.6±4.1)％ vs.(62.3±4.0)％,P ＜ 0.01].② When LR was used alone for resuscitation,the rats died quickly and survived for a short time.Only one rat survived for 12 hours,and no rat survived for more than 24 hours.Compared with the LR control group,whole blood resuscitation could improve the survival rate of uncontrolled hemorrhagic shock rats,and the survival time was significantly prolonged (hours:20.4± 4.6 vs.3.5 ± 1.1,P ＜ 0.01);0.5％,2.0％ and 5.0％ HBOC also significantly prolonged the survival time of rats.The 5.0％ HBOC group had the best effect,4 rats survived in 24 hours,and the survival time was significantly longer than that of the LR control group (hours:18.4 ± 4.0 vs.3.5 ± 1.1,P ＜ 0.01),and it was the same as the whole blood control group.③ Compared with pre-shock,CO,DO2 and hemodynamic parameters of uncontrolled hemorrhagic shock rats were significantly decreased,and the above parameters were gradually increased with the prolongation of rehydration time.Compared with the LR control group,whole blood resuscitation could significantly increase CO and DO2,and improve hemodynamics in rats with uncontrolled hemorrhagic shock at different time points.Three concentrations of HBOC could also increase CO,DO2 and other hemodynamic parameters of rats at 1 hour of maintaining MAP of 80 mmHg after hemostasis and 1 hour and 2 hours after resuscitation.The effect of 5.0％ HBOC group was more significant than that of the LR control group with statistically significant difference [CO (× 10-3,L/min):72.84±2.84 vs.63.11±2.38 at 1 hour of maintaining MAP of 80 mmHg,70.25±4.55 vs.59.88 ± 9.31 at 1 hour after resuscitation,71.51 ± 2.90 vs.53.24 ± 6.32 at 2 hours after resuscitation;DO2 (L· min-1 · m-2):271.9± 13.5 vs.159.1 ±25.4 at 1 hour of maintaining MAP of 80 mmHg,261.0± 15.0 vs.145.7±20.1 at 1 hour after resuscitation,249.6± 12.0 vs.107.4± 18.2 at 2 hours after resuscitation;MAP (mmHg):82.1±2.1 vs.74.0±2.8 at 1 hour of maintaining MAP of 80 mmHg,107.5±9.3 vs.64.0±5.7 at 1 hour after resuscitation,104.0±9.7 vs.73.0±4.2 at 2 hours after resuscitation;LVSP (mmHg):128.6±7.9 vs.103.8±0.8 at 1 hour of maintaining MAP of 80 mmHg,129.3±± 15.0 vs.99.4±0.0 at 1 hour after resuscitation,127.5± 11.3 vs.97.4±0.0 at 2 hours after resuscitation;+dp/dt max (mmHg/s):6 534.2±± 787.6 vs.5 074.0± 71.7 at 1 hour of maintaining MAP of 80 mmHg,5 961.5 ±± 545.4 vs.4 934.5 ± 510.2 at 1 hour after resuscitation,5 897.4± 350.5 vs.4 534.7 ±489.2 at 2 hours after resuscitation,all P ＜ 0.05].Conclusions HBOC infusion prolonged the survival time,increased survival rate,and improved hemodynamics,cardiac function and tissue oxygen supply in a dose-dependent manner in the early stage of uncontrolled hemorrhagic shock.The recovery effect of 5.0％ HBOC was similar to that of the whole blood.

BACKGROUND:To improve local microenvironment and reduce local scars is conducive to peripheral nerve regeneration that promotes nerve function recovery.OBJECTIVE:To evaluate the effect of fresh amniotic membrane on the regeneration of tinjured peripheral nerve.METHODS:Sixty healthy Sprague-Dawley rats were randomly divided into three groups (n=20 per group) after constructing a model of sciatic nerve injury of the unilateral leg. In group A, the nerve was wrapped with fresh human amnion at the anastomosis end after the repair of nerve. In group B, the nerve was wrapped with biofilm at the anastomosis end after the repair of nerve. In group C, no treatment was conducted after the repair of nerve (blank control). The effects were evaluated by anatomical observation, light microscope observation, immunohistochemical detection (2, 4, 8, 12 weeks after surgery), transmission electron microscope observation, axon imaging analysis, action potential detection, and sciatic nerve function index (4, 8, 12 weeks after surgery).RESULTS AND CONCLUSION: (1) Gross observation. The amniotic membrane and biofilm were absorbed partialy at postoperative 2 weeks, mostly at postoperative 4 weeks and completely at postoperative 8 weeks. In the groups A and B, the nerve was adhered slightly and loosely to the surrounding tissues, with a fair range of motion. In the group C, the nerve was tightly adhered to the surrounding tissues, with a poor range of motion. (2) Observation under light microscope. The nerve regeneration was better in the groups A and B than group C at 2, 4, 8, 12 weeks postoperatively. (3) Observation under electron microscope. Regenerated nerve fibers were rarely seen and lamelar structures were unclear in the three groups at 4 weeks postoperatively. Then, increased regenerated nerve fibers, thickened myelin sheath, clear lamelar structure and enlarged axon diameter were found in the groups A and B compared with the group C at 8 and 12 weeks postoperatively. (4) Immunohistochemical detection. The expression and distribution of S-100 protein in the groups A and B were better than those in the group C. (5) Axon image analysis. Groups A and B were superior to the group C in the diameter of myelinated nerve fibers, thickness of myelin sheath and number of regenerated nerve fibers. There was a significant difference by statistical analysis (P < 0.05). (6) Electrophysiological examination. Shorter latency period, higher amplitude and faster nerve conduction velocities were observed in the groups A and B compared with the group C (P < 0.05). (7) The sciatic function index. The sciatic function index in group A or B was significantly higher than that in group C (P < 0.05). To conclude, the human amniotic membrane can reduce adhesion between the damaged nerve and surrounding tissues, and prevent scarring at the anastomosis end. In addition, it promotes the regeneration of nerve fibers, increase axon diameter and myelin sheath thickness, and ease inflammatory and immune responses at the neural incision.

Objective To investigate the effects of hydrochloride fasudil on inflammation and secondary brain injury in rats after intracerebral hemorrhage (ICH).Methods One hundred and seventeen male Wistar rats were randomly divided into normal group (n=9),sham-operated group (n=36),ICH group (n=36) and hydrochloride fasudil treatment group (ICH/HF,n=36).ICH was caused by injecting non-anticoagulant autologous arterial blood into the right caudate nucleus,3 mg HF was dissolved in 2 mL 0.9％ saline solution and administrated intraperitioneally once daily (12 mg/kg/d) since 12 h ofICH onset.Rats in the sham-operated group and ICH group received an equal volume of normal saline.On one,3,7 and 14 d ofICH induction,Bederson method was used to score neurological deficits in rats;interleukin (IL)-6 and tumor necrosis factor (TNF)-α levels in serum were detected by enzyme-linked immunosorbent assay (ELISA).Brain edema was measured by comparing wet and dry brain weights.Inflammatory cell infiltration around the hematoma was visualized using hematoxylin and eosin,and changes in neuronal morphology were detected by Nissl staining.Immunohistochemistry method was used to detect the expressions of Caspase-3 and Bcl-2.Results (1) The neurological deficit scale scores on one and 3 d ofICH in the ICH/HF group were significantly decreased as compared with those in the ICH group (P＜0.05).(2) Serum IL-6 (except 14 d of ICH) and TNF-α levels in ICH/HF group were significantly decreased as compared with those in the ICH group (P＜0.05).(3) Brain water contents at each time point in ICH/HF group were significantly decreased as compared with those in the ICH group (P＜0.05).(4) HE staining showed the sizes of hematoma on one and 3 d ofICH in the ICH/HF group and ICH group were consistent;but ICH/HF group had smaller hematomas on 7 d of ICH,and disappeared hematoma and glial cell hyperplasia on 14 d of ICH.Mass neutrophils infiltration developed in the perihematomal areas of the ICH and ICH/HF groups;however,neutrophils infiltration in the ICH/HF group ([20.56±6.37] cells/field) was significantly decreased as compared with that in the ICH group ([41.50±8.91] cells/field) on one d of ICH (P＜0.05).(5) Nissl staining showed neuron death and loss in the perihematomal areas on one,3 and 7 d of ICH in the ICH group;but ICH/HF group had significantly increased neuron number as compared with the ICH group (P＜0.05).(6) The Caspase-3 expression on 3 and 7 d ofICH in the ICH/HF group was significantly decreased as compared with that in the ICH group (P＜0.05);the Bcl-2 expressions at each time point in the ICH/HF group was significantly increased as compared with that in the ICH group (P＜0.05).Conclusion HF exerts neuroprotective effects in ICH rat models by decreasing serum IL-6 and TNF-α levels,reducing neutrophils infiltration and neuronal necrosis around the hematoma,restraining the cell apoptosis through down-regulating Caspase-3 and up-regulating expression of Bcl-2,alleviating the brain edema,and improving the neurological outcome.

BACKGROUND:Experiments have demonstrated that biological membranes can be usedtorecon struct thetendon she athandin hibit exogenou shealing of thetendon.Therefore,the semembrane sprovide a good bed for tendon gliding and reduce tendon adhesion.
OBJECTIVE:To compare the effectsof acelular amniotic membrane and medical membraneagainst tendon adhesion during the repair oftendon sheath defects.
METHODS:ToesIIIfrom the bipeds of 66 leghorns were chosen to prepare tendon injury and tendon sheath defect models, which were randomly divided into three groups (n=22 per group). Amnion group were repaired with acelular amniotic membrane, medical membrane group with absorbable membrane, and control group had no treatment on tendon sheath defects. Gross, histological and biomechanical tests of each group were performed at 2, 4, 8, 12 weeks after surgery.
RESULTS AND CONCLUSION:At 12 weeks after surgery, in the amniotic membrane and medical membrane groups, the tendon sheath formed completely, and the tendon healed well, with no adhesion, but in the control group, there was serious tendon adhesion. At 8 weeks after surgery, the number of synovial cells in the false sheath was highest in the amniotic membrane group sequentially followed by the medical membrane group and control group. In the amniotic membrane group, the rough endoplasmic reticulum expanded highly and secreted exuberantly in the matrix, while in the control group, the synovial cells presented with messy arrangement, and expanded vacuoles in the matrix were weaker than those in the other two groups. At 12 weeks after surgery, fibroblasts were arrayedtidily in layerwith dense structure in the medical membrane and amniotic membrane groups;but in the control group, fibroblasts were distributed disorderly with loose structure. Tendon sliding distance and total flexor toe angle in the amniotic membrane and medical filmgroups were significantly larger than those in the control group (P < 0.05),butthere was no significant difference between the medical membrane and amniotic membrane groups. Additionally, the maximum tensile fracture strength had no significant difference among three groups at 12 weeks after surgery. These results indicate that both amniotic membrane and medical membrane can markedlyprotect the tendon from exogenous healing and adhesion.

Objective To explore renal transplantation model in non-human primate cynomolgus monkeys.Methods 50 non-human primates' kidneys were transplanted into the lower part of the abdomen with end-to-side anastomosis of renal artery to aorta and renal vein to inferior vena eava,and with end-to-end anastomosis of ureter to bladder.Results In the 50 cases,1 case death as accident of anesthesia;7 cases with postoperative complications,and all with creatinine sudden rise,after ultrasonic examinations showed that 2 cases with renal vein thrombosis,and 5 cases appeared urinary leakage.All animal models were without surgical infections,and with normal serum creatinine,urine output.Conclusion Non-human primate animal kidney transplantation model establishment method is reliable,but should pay attention to the the surgical technique training,complications prevention.The model is valuable for application in the research of immune tolerance,heterogeneous transplant.

Objective To evaluate the effect of debridement hepatectomy with selective hepatopetal blood occlusion in the treatment of severe hepatic trauma.Methods The clinical data of 55 patients with severe hepatic trauma treated by debridement hepatectomy with selective hepatopetal blood occlusion were retrospectively analyzed.20,20 and 15 patients were with grade Ⅲ,Ⅳ and Ⅴ hepatic trauma respectively,combined with major peripheral hepatic vascular injury in 14 cases and with other trauma in 35 cases.Additional procedures including liver suture repair in 7 cases,perihepatic gauze packing in 3 cases,inferior vena cava repair in 5 cases,hepatic vein repair in 4 cases,hepatic vein ligation in 3 cases and hepatic artery ligation in 2 cases were performed.Other operations such as craniotomy debridement in 3 cases,cholecystectomy in 6 cases,T tube drainage of common bile duct in 4 cases,splenectomy in 5 cases,pancreatic tail resection in 2 cases,left kidney resection in 1 case,thoracic cavity closed drainage in 9 cases,partial small bowel resection or repair in 4 cases and stomach repair in 1 case were performed as needed.Results The operations were successful in 47 patients.Postoperative complications were observed in 19 cases (34.5％) including coagulation disorders in 1 case,postoperative abdominal bleeding in 2 cases,intestinal obstruction in 1 case,liver and renal dysfunction in 4 cases,abdominal infection in 3 cases,incision infection in 2 cases,pulmonary infection in 4 cases,pleural effuion in 10 cases.Death occurred in 8 patients (14.5％),the cause of death were hemorrhagic shock in 3 cases,combined with severe craniocerebral injury in 2 cases,septic shock in one case,and multiple organ failure in 2 cases.Conclusions Debridement hepatectomy with slective hepatopetal blood occlusion is an effective treatment for severe hepatic trauma.