[Construction and evaluation of a prognostic model for severe acute pancreatitis based on CT scores and inflammatory factors]
Baohua Han,Wen Yang,Hui Wang,Shuchi Han,Zhibin Zhang,Huizhen Jiao,Lei Wang
DOI: https://doi.org/10.3760/cma.j.cn121430-20220411-00351
Abstract:Objective: To construct a prognostic model for severe acute pancreatitis (SAP) based on CT scores and inflammatory factors, and to evaluate its efficacy. Methods: 128 patients with SAP diagnosed admitted to the First Hospital Affiliated to Hebei North College from March 2019 to December 2021 were enrolled and given Ulinastatin combined with continuous blood purification therapy. The levels of C-reactive protein (CRP), procalcitonin (PCT), interleukins (IL-6, IL-8), tumor necrosis factor-α (TNF-α), and D-dimer were measured before and on the third day of treatment. An abdominal CT was performed on the third day of treatment to assess the modified CT severity index (MCTSI) and extra-pancreatic inflammatory CT score (EPIC). Patients were divided into the survival group (n = 94) and the death group (n = 34) according to the 28-day survival prognosis after admission. The risk factors for the SAP prognosis were analyzed using Logistic regression, which was then used to build nomogram regression models. The value of the model was evaluated using the concordance index (C-index), calibration curves and decision curve analysis (DCA). Results: Before treatment, the levels of CRP, PCT, IL-6, IL-8 and D-dimer in the death group were higher than those in the survival group. After treatment, the levels of IL-6, IL-8 and TNF-α in the death group were higher than those in the survival group. MCTSI and EPIC scores in the survival group were lower than those in the death group. Logistic regression analysis shows that, pre-treatment CRP > 140.70 mg/L, D-dimer > 2.00 mg/L, and post-treatment IL-6 > 31.28 ng/L, IL-8 > 31.04 ng/L, TNF-α > 31.04 ng/L, and MCTSI > 8 points were all independent risk factors for SAP prognosis [odds ratios (OR) and 95% confidence intervals (95%CI) were 8.939 (1.792-44.575), 6.369 (1.368-29.640), 8.546 (1.664-43.896), 5.239 (1.108-24.769), 4.808 (1.126-20.525), 18.569 (3.931-87.725), all P < 0.05]. Model 1 (consisting of pre-treatment CRP, D-dimer, and post-treatment IL-6, IL-8 and TNF-α) had a lower C-index than that model 2 (consisting of pre-treatment CRP, D-dimer, and post-treatment IL-6, IL-8 and TNF-α, and MCTSI; 0.988 vs. 0.995). The mean absolute error (MAE) and mean square error (MSE) of model 1 (0.034, 0.003) were higher than those of model 2 (0.017, 0.001). When the threshold probability was in the range of 0-0.66 or 0.72-1.00, the net benefit of model 1 was lower than that of model 2. When the threshold probability was in the range of 0.66-0.72, the net benefit of model 1 was higher than that of model 2. In addition, model 2 had a higher C-index than acute physiology and chronic health evaluation II (APACHE II) and bedside index of acute pancreatitis severity (BISAP, 0.995 vs. 0.833, 0.751). Model 2 had a lower MAE (0.017) and MSE (0.001) than APACHE II (0.041, 0.002). Model 2 had a lower MAE than BISAP (0.025). Model 2 had a higher net benefit than both APACHE II and BISAP. Conclusions: The prognostic assessment model of SAP consisting of pre-treatment CRP, D-dimer, and post-treatment IL-6, IL-8 and TNF-α, and MCTSI has high discrimination, precision and clinical application value, and is superior to APACHE II and BISAP.