The role of serum pyrrole-protein adduct in evaluating the severity and predicting the anticoagulant efficacy in patients with pyrroidine alkaloid-related hepatic sinusoidal obstruction syndrome / 中华消化杂志

ABSTRACT

Objective:To explore the role of serum pyrrole-protein-adduct (PPA) in evaluating the severity and predicting the anticoagulant efficacy in patients with pyrrolidine alkaloid-related hepatic sinusoidal obstruction syndrome (PA-HSOS).Methods:From April 2018 to December 2019, the data of 48 patients with PA-HSOS admitted and treated at Drum Tower Hospital, Affiliated Medical College of Nangjing University were collected, which included PPA level, portal vein velocity (PVV), ascites grading, PA-HSOS severity grading (according to the new severity grading criteria for suspected hepatic sinusoidal obstruction syndrome in adults by the European Society of Blood and Bone Marrow Transplantation and adjusted) and the outcome of anticoagulation. Patients with acute onset (onset of symptoms within 1 month after consuming pyrrolizidine alkaloid-containing plants) were taken as research subjects. The combination of PPA with PVV or with ascites classification of PA-HSOS severity assessment model was fitted by logistic regression, and the logit values of 2 combination models were calculated, the formula was logit 1=0.034×PPA(nmol/L)+ 0.055×PVV(cm/s)-3.287, logit 2=0.039×PPA(nmol/L)-2.712×ascites grade 2 (Yes=1, No=0)-0.388×ascites grade 3 (Yes=1, No=0)-0.899. The patients received initial anticoagulation therapy at Drum Tower Hospital, Affiliated Medical College of Nanjing University were selected as research subjects. The anticoagulant efficacy prediction model of combination of PPA with serum creatinine (SCR) and with hepatic venous pressure gradient (HVPG) was fitted by logistic regression, and the logit value was calculated, the formula was logit 3=0.013×PPA(nmol/L)+ 0.064×SCR (mol/L)+ 0.542×HVPG (mmHg, 1 mmHg=0.133 kPa)-16.005. The predictive value of PPA in evaluating the severity of PA-HSOS and anticoagulant efficacy was evaluated. Receiver operating characteristic curve analysis was performed for statistical analysis. Results:The serum PPA level of 48 patients was 10.81 nmol/L (3.91 nmol/L, 32.04 nmol/L). Among them, 33 cases (68.8%) were mild PA-HSOS, 3 cases (6.2%) were moderate PA-HSOS, no severe PA-HSOS case and 12 cases (25.0%) were very severe PA-HSOS. Among 23 patients received initial anticoagulant therapy at Drum Tower Hospital, Affiliated Medical College of Nanjing University and with complete data, 8 patients responded and survived, and 15 patients did not respond (5 patients died, 1 patient relieved after continue anticoagulant therapy, and 9 patients survived after switching to anticoagulant therapy and transjugular intrahepatic portosystemic shunt (TIPS) treatment). One patient without initial anticoagulant therapy, survived after TIPS treatment because of the progress of the disease. Area under the curve (AUC) of PPA to assess the severity of acute onset PA-HSOS was 0.75, 95% confidence interval ( CI) was 0.52 to 0.98 ( P=0.047). When PPA≥45.519 nmol/L, the specificity and sensitivity in evaluating severe and very severe PA-HSOS was 100.0% and 57.1%, respectively. AUC of combination of PPA and PVV to assess the severity of PA-HSOS was 0.77, 95% CI was 0.55 to 1.00 ( P=0.032). When the logit of combination model≥0.180, the specificity and sensitivity in evaluating severe and very severe PA-HSOS was 71.4% and 81.8%, respectively. AUC of combination of PPA and ascites grade (grade 1, 2 or 3) to assess the severity of PA-HSOS was 0.85, 95% CI was 0.63 to 1.00 ( P=0.005). When the logit of combination model≥0.347, the specificity and sensitivity in evaluating severe and very severe PA-HSOS was 85.7% and 92.0%, respectively. AUC of combination of PPA, SCR and HVPG to predict anticoagulation efficacy was 0.85, 95% CI was 0.69 to 1.00 ( P=0.009). When the logit≥0.393, the specificity and sensitivity in predicting anticoagulation efficacy was 62.5% and 91.7%, respectively. Conclusions:PPA can be used to assess the severity of acute onset PA-HSOS patients, and combined with ascites grading can significantly improve its efficiency. PPA combined with SCR and HVPG can better predict anticoagulant efficacy.