Application and validation of a surgical risk calculator in a liver surgery cohort of patients in Veracruz, Mexico.

OBJECTIVE: Multiple models have tried to predict the morbidity and mortality of liver resections (HR). This study aims to determine the efficacy and validity of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) surgical risk calculator in a cohort of patients undergoing HR in Veracruz, Mexico. MATERIAL AND METHODS: Retrospective analysis of patients undergoing HR between 2005 and 2019. To estimate the performance of the calculator, the observed results were compared with the average risk predicted by the calculator, using ROC curve, Brier score and Z test. RESULTS: 67 patients were evaluated, mean age 51.9 years of age, 50.7% female. The majority of resections (56.7%) were for malignancy and 62.1% were partial hepatectomies (up to 3 liver segments). The morbidity was 25.4% and the mortality 4.5%. There was a good prediction in the complications (serious complication: C = 0.725 statistic, Brier score 0.26, p = 0.006 and any statistical complication C = 0.731, Brier score 0.33, p = 0.005) and mortality (C = 0.922 statistic, Brier score 0.005, p = 0.014). CONCLUSIONS: The application of the ACS-NSQIP calculator in patients undergoing HR has good discrimination power and good predictive ability. Prediction of postoperative risks achieves a preoperative planning of the appropriate procedure.

OBJETIVO: Múltiples modelos han intentado predecir la morbilidad y mortalidad de las resecciones hepáticas (RH). Este estudio tiene por objetivo determinar la eficacia y validez de la calculadora de riesgo quirúrgico del American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) en una cohorte de pacientes sometidos a RH en Veracruz, México. MATERIAL Y MÉTODOS: Análisis retrospectivo de pacientes sometidos a RH entre 2005 y 2019. Se compararon los resultados observados con la media del riesgo previsto por la calculadora, mediante Curva ROC, score de Brier y prueba Z. RESULTADOS: Se evaluaron 67 pacientes, con 51.9 años de edad media, un 50.7% de sexo femenino. La mayoría de las resecciones (56.7%) fueron por patología maligna y el 62.1% fueron hepatectomías parciales. La morbilidad fue del 25.4% y la mortalidad del 4.5%. Hubo una predicción buena en las complicaciones (complicación seria, estadística: C: 0.725, score Brier: 0.26, p = 0.006 y cualquier complicación, estadística: C: 0.731, score Brier: 0.33, p = 0.005) y la mortalidad (estadística C: 0.922, score Brier: 0.005, p = 0.014). CONCLUSIONES: La aplicación de la calculadora ACS-NSQIP en pacientes sometidos a RH tiene buen poder de discriminación y buena habilidad predictiva. Predecir riesgos postoperatorios logra una adecuada planeación preoperatoria del procedimiento.

Performance of the American College of Surgeons National Surgical Quality Improvement calculator for early postoperative outcomes following bile duct repair hepaticojejunostomy.

PURPOSE: A Roux-en-Y hepaticojejunostomy (HJ) is the preferred method for repairing bile duct injuries (BDIs). The American College of Surgeons (ACS) established the National Surgical Quality Improvement Program (NSQIP) online risk calculator to predict risks for morbidity and mortality. The objective of the study is to assess the use of the ACS NSQIP calculator for patients undergoing HJ for BDI repairs outside the NSQIP cohort. METHODS: An IRB-approved retrospective study of Bismuth-Strasberg type E BDI HJ (2008-2020) was performed. Clinical data was introduced in the NSQIP calculator, and morbidity and mortality were determined. Perioperative risk factors were obtained. Comparisons of postoperative complications and NSQIP-predicted complication rate were carried out. RESULTS: Eighty-two patients were included (age: 42.2 ± 15.7 years; 81% female; BMI 27.1 ± 4.4 kg/m2). The most common injury was E4 (36.3%). A total of 40.2% patients had a 30-day complication. Mortality was 2.4%. Preoperative sepsis and high body mass index (p = 0.01) were significantly related to complications (p = 0.01) (univariate analysis). Older age (p = 0.01) and higher ASA class (p = 0.02) were significantly related to mortality (univariate analysis). None was statistically significant in multivariate analysis. Comparison between morbidity and mortality and the calculated NSQIP was not statistically significant. Postoperative mortality had a statistically significant trend (C-value = 0.72, p = 0.055). CONCLUSION: Preoperative sepsis, high body mass index, age, and ASA classification were associated with worse outcomes in HJ BDI repair. The ACS NSQIP calculator did not have a good performance in a population outside the NSQIP data. Further larger studies will need to corroborate these results.

Low-dose Thymoglobulin vs Basiliximab Induction Therapy in Low-Risk Living Related Kidney Transplant Recipients: A Prospective Randomized Trial.

CONTEXT: Thymoglobulin is used effectively as induction agent in kidney transplantation but the optimal dose is not well established. OBJECTIVE: Demonstrate that low-dose thymoglobulin (3 mg/kg) has similar efficacy and safety compared to basiliximab induction in low-risk kidney transplantation under standard maintenance immunosuppression DESIGN, SETTING, PARTICIPANTS: Prospective randomized study in kidney transplant patients (12/2016-05/2018). INCLUSION CRITERIA: Recipients > 18 years, first living donor transplant. EXCLUSION CRITERIA: Second and multiorgan transplant, ABO incompatibility, positive cross-match, panel reactive antibodies (PRA) > 30%, positive donor-specific antibody, human immunodeficiency virus, hepatitis B surface antigen, hepatitis C virus positive, white blood cells < 2000 cells/mm3, platelets < 75,000 cells/mm3 and malignancy. INTERVENTION: Group A: basiliximab (20 mg D0 and D4). Group B: thymoglobulin (3 mg/kg total). Maintenance immunosuppression: tacrolimus, mycophenolate mofetil, and steroids. MAIN OUTCOME MEASURES: Biopsy-proven acute rejection (BPAR), delayed graft function, slow graft function, leukopenia, infections, adverse events, graft loss, estimated glomerular filtration rate, and death within 12 months. RESULTS: 100 patients (basiliximab, n = 53) (thymoglobulin, n = 47) were included. Donor and recipient characteristics were similar except for longer dialysis (basiliximab), PRA class I (1.2% basiliximab, 4.5% thymoglobulin), HLA match (basiliximab 2.8, thymoglobulin 2.2), and cytomegalovirus status. BPAR rate was basiliximab 3.8% and thymoglobulin 6.4% (P = ns). Delayed graft function (basiliximab 3.8%; thymoglobulin 4.3%), slow graft function, and 12-month leukopenia (basiliximab 11.3%, thymoglobulin 21.3%) were similar between groups (P = ns). There was no difference in infections and adverse events between groups. Patient and graft survival were as follows: basiliximab 98.1% and 92.5%, thymoglobulin 100% and 93.6% (P = ns). CONCLUSION: Low-dose thymoglobulin induction (3 mg/kg) can be used effectively and safely in low-risk kidney transplantation with good results during the first year post-transplant.

Factors associated with patency loss and actuarial patency rate following post-cholecystectomy bile duct injury repair: long-term follow-up.

PURPOSE: Several factors may contribute to bile duct injury (BDI) repair failure. The objective of our study was to evaluate factors that contribute to the loss of patency and influence the actuarial patency rate of BDI repairs in an 11-year period. METHODS: Retrospective review study of patients who underwent a hepaticojejunostomy for a type E Bismuth-Strasberg BDI (2008-2019). The outcomes are the following: primary patency attained, loss of primary patency, and actuarial primary patency rate. Logistic regression for loss of patency and Cox regression for actuarial patency rate were used. RESULTS: Seventy-nine patients (age 42.3 ± 15.8 years, 81% female) were studied. Most common index operation was open cholecystectomy (60.8%). Most common Bismuth-Strasberg lesion was E4 (38%). Primary patency was 93.4%. Mean follow-up was 36 ± 34 months. Ten-year actuarial patency was 53.9%. Factors associated with loss of patency were vasculobiliary injury, biliary stents, and 90-day biliary complications (univariate); number of surgeries before repair and postoperative cholangitis (univariate and multivariate) (p < 0.05). Factors that impacted actuarial patency rate were (univariate analysis) 90-day biliary complications; postoperative cholangitis and index treatment period stenosis (p < 0.05). No factors impacted actuarial patency rate in multivariate analysis. CONCLUSION: Postoperative cholangitis is associated with loss of patency and had potentially detrimental effect on the actuarial patency rate in BDI repair.

Retroactive Application of a New Risk Index for Living Donor Kidney Transplantation to Renal Transplants in Veracruz, Mexico.

BACKGROUND: The Living Kidney Donor Profile Index (LKDPI) was recently created. This model predicts recipient risk of graft loss after living donor transplant. Herein, we applied the LDKPI to our population to analyze its performance. METHODS: A retrospective analysis of all living donor kidney transplants from 2003 to 2018 from 2 transplant centers in Veracruz, Mexico, was used. LKDPI was calculated in a webpage (www.transplantmodels.com). Donor and recipient demographics and transplant data included in the model were registered. Pearson correlation between the LKDPI percentage and death-censored graft survival was performed. Kaplan-Meier survival (log-rank) and Cox regression analysis were compared between the LKPDI quartiles. P < .05 was considered statistically significant. RESULTS: In total, 821 transplants were included (mean age 31.7 ± 10.5 years, 62.5% male, n = 513). Mean follow-up was 64.7 ± 46.2 months. Mean estimated survival (Kaplan-Meier) was 128.9 ± 3 months (95% confidence interval [CI], 123-134). Ten-year death-censored graft survival was 61.4%. Median LKPDI was -2%, and mean LKDPI was -2.6% ± 14.6% (range, -50% to 42%). Pearson coefficient correlation between the LKDPI and death-censored graft survival was 0.024 (P = .4). Area under the curve (receiver operating characteristic [ROC]) for the LKDPI and death-censored graft loss was 0.54 (95% CI, 0.505-0.591) (P = .04). Recipients with the lowest LKDPI had lower risk of death-censored graft loss than other quartiles (P = .014 log-rank). Cox regression analysis was significant for the lower LKDPI quartile (<20%) (Exp B = 0.35; 95% CI, 0.14-0.9; P = .03). CONCLUSION: The LKDPI applies with moderate discrimination predictive power in our population. The best LKDPI patient has better death-censored graft survival. Further studies might continue to validate the LKDPI in other cohorts.

Annual Analysis (2018) of the Kidney Transplant Waiting List of a Social Security Hospital in Veracruz, Mexico.

BACKGROUND: In Mexico during 2018, 15,072 patients were waiting for a deceased donor kidney transplant, and 969 deceased donor kidney transplants were performed. There is no annual data report of the waiting list activity in Mexico. Herein, we analyzed our kidney transplant waiting list activity in 2018. METHODS: We performed a waiting list analysis in our unit during 2018. Patient and status characteristics (active, deceased, inactive, or transplant) were registered. Differences between status were determined. A P < .05 was considered statistically significant. RESULTS: In total, 467 patients were waiting, and 74 patients were included on the list (57.7% male, mean age 38.5 ± 11.3 years and mean BMI 24.9 ± 4.7 kg/m2); 92.8% were state residents. The most common end-stage renal disease diagnosis was unknown (40.9%). In total, 94.9% were on dialysis (mean time 5.1 ± 3.14 years), and for 90.9%, this was the first transplant. PRA class I and class II were 19.9% ± 30.6% and 12.9% ± 27.1%, respectively. Mean EPTS was 19.8% ± 9.4%. Mean waiting time was 2.88 ± 2.3 years. In total, 21 deceased donor patients (3.9%) were transplanted; 57 (10.5%) patients had an inactive status, and 3 (0.6%) received a living donor kidney transplant with a proven mortality of 1.8% (n = 10). Patients who underwent deceased donor transplant were younger and had more time on dialysis, lower PRA class I, and more time on the waiting list (P < .05 by analysis of variance). CONCLUSION: There are more patients included on the list than patients off the list. There are significant differences between patients who received a transplant and inactive and active patients that needs to be shortened.

Effect of Panel-Reactive Antibody on Graft Survival in Living Kidney Donor Transplantation: Analysis of 10 Years in a Transplant Center in Veracruz, Mexico.

BACKGROUND: Pretransplant anti-HLA antibodies are a risk factor for graft rejection and loss, and its percentage estimate is known as panel-reactive antibody (PRA). Our objective was to evaluate the influence of PRA on the survival of renal grafts from living donors over a period of 10 years. METHODS: Retrospective analysis was completed in all living donor transplants with PRA class I and class II from October 2008 to December 2018 with follow-up until June 2019. The methods used for the PRA were flow cytometry and Luminex. Graft survival (not censored) was evaluated by Kaplan-Meier (log-rank) and Cox regression. P < .05 was considered significant. RESULTS: The study included 393 patients. PRA class I mean was 9.8 ± 20% (0%-98%) and class II mean was 8.6 ± 17.8% (0%-97.8%). Of the patients, 81.9% had a PRA <20% for any class. Uncensored graft survival at 1, 5, and 10 years was 90.3%, 76.2%, and 69.3%, respectively. Mean estimated uncensored graft survival in PRA <20% patients (103.9 ± 2.7, 95% confidence interval [CI] 96.6-11.2) was higher than that of PRA >20% patients (61.5 ± 5.7, 95% CI 50.3-72.8) (P = .005 log-rank). Cox regression (univariate) was statistically significant for PRA class I (Exp [B] 1.01, 95% CI 1.003-1.02, P = .009) and for PRA >20% any class (Exp [B] 2.074, 95% CI 1.222-3.520, P = .007). CONCLUSION: PRA class I and PRA >20% any class are associated with lower graft survival. PRA must be considered to determine immunologic risk and to choose an immunosuppressive regimen in kidney transplantation.

Acute Rejection Is a Strong Negative Predictor of Graft Survival in Living-Donor Pediatric Renal Transplant: 10-Year Follow-Up in a Single Mexican Center.

OBJECTIVES: Kidney transplant is the optimal treatment for children with end-stage renal disease. Multiple factors affect patient and graft survival. We assessed determinants of long-term patient/graft survival in our center by a retrospective review of pediatric living donor (< 18 years) kidney transplants from February 2003 to December 2016. MATERIALS AND METHODS: Donor and recipient demo-graphic data and immunosuppression use were gathered for analyses. Transplant outcomes included patient/graft survival, acute rejection, and 1-year estimated glomerular filtration rate. Patient/graft survival results were analyzed by Kaplan-Meier, and Cox proportional hazards regression model was used for risk factors (univariate/multivariate). P ≤ .05 was statistically significant. RESULTS: Ninety-nine patients were included. Age was 13.4 ± 3.08 years, 64.6% were male, and 88.9% were on dialysis with time of 17.1 ± 12.6 months. Mean donor age was 36.6 ± 7.7 years, and most were females (63.6%). Donor estimated glomerular filtration rate was 89.4 ± 16.9 mL/min/1.73 m2. HLA match was 3.2 ± 1.05. Panel reactive antibody showed 8.6 ± 20.5%. Of total patients, 47.5% used induction, 88.9% used cyclo-sporine, and 100% used mycophenolate mofetil. Five- and 10-year patient survival rates were 93.2% and 93.2%. One-year acute rejection was 14.1%, with rate of 24.2% throughout follow-up. One-year estimated glomerular filtration rate was 76.4 ± 25.6 mL/min/1.73 m2. Five- and 10-year graft survival rates were 62.6% and 43.3%. Multivariate analysis confirmed donor age and acute rejection episodes throughout follow-up as risk factors for graft survival (P < .05). CONCLUSIONS: Acute rejection and donor age are important risk factors for 10-year graft survival in living-donor pediatric kidney transplant in our program.