关键词: 肝脓肿, 隐源性感染, 临床特征, 影像学表现, 病原学, 宏基因组二代测序

Abstract:

Objective This study aims to investigate the clinical characteristics， associated risk factors， and pathogen profiles of cryptogenic and non-cryptogenic pyogenic liver abscess （PLA）， and to evaluate the diagnostic performance of metagenomic next-generation sequencing （mNGS） for pathogen identification in PLA. Methods A retrospective analysis was performed on 167 patients diagnosed with pyogenic liver abscess admitted to our hospital between January 2020 and December 2025. Patients were stratified into the cryptogenic group （n = 105） and the non-cryptogenic group （n = 62） based on whether a definitive infectious source was identified. Baseline characteristics， laboratory results， and imaging features were compared between the two groups. Among 41 patients who underwent both examinations， we further evaluated the diagnostic performance of mNGS versus conventional bacterial culture. Results A total of 167 patients diagnosed with PLA were enrolled in this study. Compared with the non-cryptogenic group， patients in the cryptogenic group had a higher prevalence of diabetes mellitus， a higher neutrophil percentage， and elevated procalcitonin levels （all P < 0.05）. In contrast， biliary diseases， intra-abdominal malignancies， and multiple hepatic lesions were significantly more prevalent in the non-cryptogenic group （all P < 0.05）. Multivariate logistic regression analysis identified fever （OR = 3.778， 95%CI： 1.688 ⁃8.456）， diabetes mellitus （OR = 2.212， 95%CI： 1.043 ⁃ 4.691）， and elevated hemoglobin level （OR = 1.019， 95%CI： 1.002 ⁃ 1.036） as independent risk factors for cryptogenic PLA. Klebsiella pneumoniae was the most common predominant pathogen isolated in this cohort. Among the 41 patients who received both mNGS testing and conventional pathogen culture， mNGS achieved a significantly higher positive detection rate than conventional culture ［95.12% （39/41） vs. 73.17% （30/41）， P = 0.022］. Additionally， mNGS had a shorter turnaround time and superior detection performance for anaerobic bacteria and uncommon pathogens. Conclusions Cryptogenic and non-cryptogenic pyogenic liver abscesses demonstrate distinct clinical heterogeneity in underlying etiologies， clinical manifestations， inflammatory profiles， and lesion distribution. Etiology-based classification can facilitate early risk stratification and individualized clinical management. For patients with complicated pyogenic liver abscesses or unclear infectious sources， mNGS acts as a valuable adjunct to conventional microbiological testing， enabling accurate pathogen identification and targeted optimization of antimicrobial therapy.

Key words: liver abscess, cryptogenic infection, clinical features, imaging findings, etiology, metagenomic next-generation sequencing