Efficacy Heterogeneity of Hemoperfusion Plus Continuous Venovenous Hemofiltration for Hypertriglyceridemia‐Induced Acute Pancreatitis Stratified by Severity of Inflammatory Response

BACKGROUND: This study aimed to investigate the heterogeneity in the efficacy of hemoperfusion (HP) plus continuous venovenous hemofiltration (CVVH) versus CVVH alone in reducing the incidence of severe acute pancreatitis (SAP) among patients with hypertriglyceridemia-induced acute pancreatitis (HTG-AP), stratified by the inflammatory marker C-reactive protein (CRP). In addition, we assessed secondary outcomes encompassing laboratory response, disease severity and organ function, healthcare resource utilization, and treatment-related indicators, so as to delineate the broader efficacy profile of the combination therapy and its relationship with the primary endpoint. Furthermore, we sought to identify a CRP-based therapeutic window for the combination therapy. METHODS: This single-center retrospective study included 204 patients who were assigned to the HP + CVVH group (n = 103) or the CVVH group (n = 101). Multivariable logistic regression was used to assess the overall difference in efficacy between the two groups. Restricted cubic splines (RCS) were employed to analyze the nonlinear relationship between treatment efficacy and CRP levels and to identify a CRP-based therapeutic window. Secondary outcomes were analyzed using multivariable logistic regression for categorical variables and generalized linear models for quantitative variables. RESULTS: After multivariable adjustment, the incidence of SAP did not differ significantly between the HP + CVVH and CVVH groups (p > 0.05). RCS analysis revealed a significant interaction between treatment group and CRP levels. The conditional odds ratio for SAP (HP + CVVH vs. CVVH) was significantly below 1 only within a specific CRP range: 194.96-296.91 mg/L in the primary analysis and 200.77-285.81 mg/L after propensity score adjustment (p < 0.05). Regarding secondary outcomes, HP + CVVH was associated with significantly lower CRP, procalcitonin, and triglyceride levels on day 3, more SIRS-free days within 14 days, lower peak intra-abdominal pressure, shorter ICU length of stay, fewer CVVH sessions, earlier initiation of enteral nutrition, and higher blood purification-related costs (all p < 0.05). No significant differences were observed in organ failure-free days, incidence of pancreatic necrosis, hospital length of stay, total hospitalization cost, total morphine equivalent dose, carbapenem use, or duration of mechanical ventilation (all p > 0.05). CONCLUSION: HP + CVVH showed no overall superiority to CVVH alone in preventing SAP. Nevertheless, secondary outcome analyses demonstrated improved inflammatory control, faster resolution of systemic inflammatory response, reduced intra-abdominal pressure, and better recovery in several resource-related and treatment-related indicators, indicating measurable biological and clinical process benefits. These benefits, however, did not translate into a uniform reduction in SAP. The effect of HP + CVVH on the hard endpoint was conditional, nonlinear, and dependent on baseline inflammatory burden, with significant benefit confined to a specific CRP therapeutic window. Therefore, improvements in secondary endpoints should not be equated with hard-endpoint benefit, and the decision to initiate HP + CVVH should be individualized according to the intensity of the inflammatory response.