Objective To explore the predictive value of two models based on machine learning algorithms in predicting the initial and subsequent doses of tacrolimus in kidney transplant recipients. Methods A retrospective analysis was conducted on the medical records of 1 013 Chinese kidney transplant recipients at the First Affiliated Hospital of Sun Yat-sen University from January 2015 to April 2019, focusing on the initial and subsequent doses in kidney transplant recipients. Thirty-three variables were collected for the initial dose, and twenty-six variables for the subsequent dose. A genetic algorithm combined with a random-restart hill-climbing algorithm was used to determine a small number of key clinical variables through majority voting, and variables with Lasso regression coefficients less than the optimal variable coefficient threshold were further eliminated. The selected clinical variables were input into a cascaded deep forest (CDF) and TabNet deep neural network for analysis and comparison based on structured tabular data, and the leave-one-subject-out method was used for validation. Results A total of 613 recipients were included in the training set, and 116 recipients were in the external validation set. In the initial dose algorithm of tacrolimus, the clinical variables ultimately included target concentration, time from surgery to target concentration, body weight, gender, type of surgery, time from surgery to first dose, WuZhi capsule, calcium channel blocker, creatinine, hemoglobin and CYP3A5. In the subsequent dose algorithm, the clinical variables ultimately included target concentration, time from surgery to target concentration, WuZhi capsule, creatinine, alanine aminotransferase, aspartate aminotransferase, previous dose, previous dose concentration and time from surgery to previous concentration. Based on the above variables, the TabNet model showed better predictive performance than the CDF model: in the initial dose prediction, the accuracy of the predicted dose within ±20% of the actual dose was 0.801, and the fitting index R² was 0.436; in the subsequent dose prediction, the corresponding accuracy and R² were 0.939 and 0.902, respectively. The results of feature contribution showed that CYP3A5 and target concentration contributed the most to the prediction of initial dose, while previous dose and its corresponding concentration had the greatest impact on subsequent dose prediction. In addition, the results of independent external validation were also satisfactory. Conclusions The optimized TabNet predictive model may provide important reference for drug dose prediction based on machine learning algorithms in clinical practice.

Objective To analyze the clinical features of patients with coronavirus disease 2019 (COVID-19) in Shanghai and to investigate the risk factors for disease progression to severe cases. Methods The clinical data of 292 adult patients with COVID-19 hospitalized in Shanghai Public Health Clinical Center from January 20, 2020 to February 10, 2020 were retrospectively analyzed, including 21 severe patients and 271 mild patients. The demographic characteristics, epidemiological history, history of underlying diseases and laboratory examinations were compared between the two groups. Measurement data were compared using t test or Mann-Whitney U test. The count data were compared using hi-square test. The binary logistic regression equation was used to analyze the risk factors for the progression of patients to severe cases. Results Among the 292 patients, 21 were severe cases with the rate of 7.2% (21/292). One patient died, and the mortality rate was 4.8% in severe patients. The severe patients aged (65.0±15.7) years old, 19 (90.5%) were male, 11 (52.4%) had underlying diseases, 7 (33.3%) had close relatives diagnosed with COVID-19. The mild patients aged (48.7±15.7) years old, 135 (49.8%) were male, 74 (27.3%) had underlying diseases, 36 (13.3%) had close relatives diagnosed with COVID-19. The differences between two groups were all significant statistically ( t =-4.730, χ 2 =12.930, 5.938 and 4.744, respectively, all P <0.05). Compared with the mild patients, the levels of absolute numbers of neutrophils, alanine aminotransferase, aspartate aminotransferase, lactate dehydrogenase, creatinine, serum cystatin C, C reactive protein (CRP), procalcitonin , D -dimer, pro-B-type natriuretic peptide (proBNP), serum myoglobin, creatine kinase (CK), creatine kinase isoenzyme (CK-MB), serum troponin I (cTnI) in severe patients were all significantly higher ( U =2 091.5, 1 928.0, 1 215.5, 729.0, 1 580.5, 1 375.5, 917.5, 789.5, 1 209.0, 1 434.0, 638.0, 964.5, 1 258.0 and 1 747.5, respectively, all P <0.05), while the levels of lymphocyte count, albumin, transferrin, CD3 + T lymphocyte count, CD8 + T lymphocyte count and CD4 + T lymphocyte count in severe patients were all significantly lower ( U =1 263.5, t =4.716, U =1 214.0, 962.0, 1 167.5 and 988.0, respectively, all P <0.05). Further logistic regression analysis showed that the albumin (odds ratio ( OR )=0.806, 95% CI 0.675-0.961), CRP ( OR =1.016, 95% CI 1.000-1.032), serum myoglobin ( OR =1.010, 95% CI 1.004-1.016), CD3 + T lymphocyte count ( OR =0.996, 95% CI 0.991-1.000) and CD8 + T lymphocyte count ( OR =1.006, 95% CI 1.001-1.010) at admission were independent risk factors for the progression of COVID-19 patients to severe illness (all P <0.05). Conclusions Severe cases of patients with COVID-19 in Shanghai are predominantly elderly men with underlying diseases. Albumin, CRP, serum myoglobin, CD3 + T lymphocyte count and CD8 + T lymphocyte count could be used as early warning indicators for severe cases, which deserve more clinical attention.

Objective To evaluate the effect of deep and moderate neuromuscular blockade on surgical conditions during minor laparoscopic gynecologic surgery.Methods Sixty-five patients,with expected surgery time ＜ 3 h,aged 18-60 yr,with body mass index＜30 kg/m2,of American Society of Anesthesiologists physical status Ⅰ orⅡ,scheduled for elective laparoscopic gynecological surgery,were allocated into deep neuromuscular blockade group (group D,n =33) and moderate neuromuscular blockade group (group M,n=32) using a random number table.After induction of anesthesia,the patients were tracheally intubated and mechanically ventilated.Cisatracurium was continuously infused to maintain the degree of neuromuscular blockade in both groups to achieve the target degree post-tetanic count of 1 or 2 in group D and train-of-four (TOF) count of 1 or 2 in group M.Surgical conditions were assessed and scored after surgery.The recovery index,time for TOF ratio returning to 0.7 and 0.9,surgery time,mean intra-abdominal pressure,extubation time and TOF ratio at extubation were recorded.Results Compared with group M,the mean intra-abdominal pressure was significantly decreased,and the extubation time and time for TOF ratio returning to 0.7 and 0.9 were prolonged in group D (P＜0.05).There was no significant difference in the other parameters between the two groups (P＞0.05).Conclusion Moderate neuromuscular blockade can provide better surgical conditions for minor laparoscopic gynecological surgery with shorter recovery time.

We designed a locator for puncture under CT examination.Thirty-three patients of both sexesaged 16-76 yr weighing 46-80 kg undergoing bilateral thoracic or lumber sympathetic nerve block under CT examination were enrolled in this study.One side was punctured under the guidance of the locator,while the other side was punctured by conventional technique.The time consumed during puncture,CT scan adjusting time,the deviation of the needle from the targets and the incidence of puncture complications were recorded.The results showed that compared with the conventional technique,with the locator the time consumed during puncture was significantly shorter,the number of attempts and the deviation of the needle from the target were significantly reduced.The procedure was successfully performed on both side in all patients without serious complications such as hemothorax and pneumothorax.

Objective To compare the clinical efficacy and safety of isotretinoin erythromycin gel, a gel containing isotretinoin (0.05%) and erythromycin (2%), versus adapalene gel in the treatment of mild to moderate acne vulgaris. Methods A multicenter, randomized, open, parallel-controlled clinical study was conducted. A total of 192 patients with mild to moderate (Grade Ⅰ -Ⅲ ) acne vulgaris were enrolled in this study according to the grading criteria for acne severity in guidelines for the treatment of acne in China. Efficacy analysis was carried out in 169 patients and safety analysis in 190 patients. The patients were classified into trial group (n = 86) and control group (n = 83 ) to be treated with isotretinoin erythromycin gel or adapalene gel once a night for 6 weeks. Patients were evaluated at the baseline, on week 2, 4 and 6 during the treatment for the count of comedones (both open and closed), inflammatory papules and pustules, severity of acne and local or general adverse effects. Results After the start of treatment, the response rate gradually increased and severity of acne decreased in both groups. On week 6, the total response rate was 51.16% in the trial group and 40.96% in the control group (P ＞ 0.05), while a greater reduction in the count of pustules and inflammatory lesions was observed on week 4 and 6 in the trial group with a lower severity grade of acne compared with the control group (P ＜ 0.05 or 0.01 ). Adverse reactions were similar in both groups and manifested as tolerable local irritation. Conclusions The efficacy of isotretinoin erythromycin gel is similar to that of adapalene gel in the treatment of mild to moderate acne vulgaris, however, isotretinoin erythromycin gel seems superior to adapalene gel in reducing inflammatory lesions and rapidly improving severity of acne vulgaris.

Objective To determine the relationship between serum levels of MCAF/MCP-1 (monocyte chemotactic and activating factor/monocyte chemoattractant protein-1), RANTES(regulated on activation, normal T-cell expressed and secreted) and the disease activity of systemic lupus erythematosus(SLE). Methods Serum levels of MCAF and RANTES were measured by ELISA. Results ①Serum level of MCAF but not RANTES, was significantly increased in patients with SLE as compared with controls. ②Serum level of MCAF but not RANTES, was markedly higher in patients with active disease than those with inactive disease. ③No significant differences were found in the serum levels of MCAF and RANTES between patients with renal damage and those without. Conclusions These results suggest that MCAF may be involved in the pathogenesis of SLE, and serum MCAF levels could be an indicator for the disease activity of SLE.