Background Pyrrolidine alkaloid‑induced hepatic sinusoidal obstruction syndrome (PA-HSOS) is a drug-induced liver injury characterized by sudden and rapidly progressive portal hypertension. Hepatic venous pressure gradient (HVPG) is the gold standard for the diagnosis of sinus hypertension, at least in viral and alcoholic liver disease, while its role in evaluating disease severity and treatment choice and predicting prognosis in patients with PA-HSOS needs further verification. The aim of this study was to investigate the clinical value of HVPG in patients with PA-HSOS.
Methods Patients who were diagnosed with PA-HSOS according to the “Nanjing criteria” and received HVPG measurements in our hospital between January 2016 and April 2020 were reviewed retrospectively. Next, the predictive efficacy of HVPG in initial anticoagulant therapy, the prognostic survival of patients in different HVPG groups, the relationship between HVPG and the Drum Tower Severity Scoring (DTSS), the correlation between HVPG and pathological data were all evaluated.
Results A total of 76 PA-HSOS patients were included in the study. Multivariate logistic regression analysis identified HVPG as independently associated with nonresponse to initial anticoagulation (95% CI: 1.006-1.413, P=0.043). The receiver operating characteristic curve (ROC) identified a cutoff value of 20.165 mmHg (sensitivity: 0.744, specificity: 0.697) for predicting nonresponse to initial anticoagulation, and the area under the curve (AUC) was 0.741 (95% CI: 0.626-0.857, P<0.001). When HVPG>20.165 mmHg was combined with serum total bilirubin (STB), heart rate (HR) and blood urea nitrogen (BUN) for modeling, the AUC was 0.881 (P<0.001). Seventy-six patients were divided into two groups according to whether HVPG was greater than 20.165 mmHg, and the patients in the high HVPG group (HVPG>20.165 mmHg) had a significantly worse prognostic survival than the patients in the low HVPG group (HVPG<20.165 mmHg) (P=0.022, χ2=5.285). Additionally, there was a linear correlation between HVPG and the area of sinusoidal bleeding in 76 patients (P=0.008, R=0.343). Furthermore, after excluding the patients with an onset time of more than 1 month, HVPG improved the efficacy in predicting nonresponse to anticoagulant therapy (AUC=0.789, 95% CI: 0.654-0.924, P=0.001), and the linear relationship between HVPG and bleeding area in the sinusoids was enhanced (P=0.001, R=0.499). In addition, a linear relationship between HVPG and DTSS emerged (P<0.001, R=0.522).
Conclusions HVPG can assist in predicting the outcome of anticoagulant therapy and prognosis in patients with PA-HSOS, and this prediction is more accurate for patients within one month of disease onset.