From dysbiosis to prediction: a novel gut microbiota-derived index for spontaneous bacterial peritonitis in HBV-related cirrhosis.

SBP in HBV-related cirrhosis was characterized by enrichment of pathogenic taxa and depletion of short-chain fatty acid producers.

• Pathogenic taxa enriched in SBP included Escherichia-Shigella, Klebsiella, Veillonella, and Streptococcus.
• Depleted taxa included short-chain fatty acid producers: Prevotella, Roseburia, Faecalibacterium, and Bacteroides.
• These microbial shifts formed the basis of the SBP Microbiota Index (SBP-MI).
• Fecal 16S rRNA sequencing was used to characterize microbial composition across 135 participants.

The study enrolled 135 participants across four groups representing a spectrum from health to SBP.

• Groups included healthy controls (n=40), compensated cirrhosis (n=30), cirrhosis with ascites but without SBP (n=40), and SBP (n=25).
• Fecal 16S rRNA sequencing and clinical data were obtained from all 135 participants.
• An additional 140 cirrhotic patients with ascites were prospectively followed for 6 months with SBP occurrence as the endpoint.
• Among the prospective cohort, 40 patients provided paired fecal samples for longitudinal analysis.

During follow-up, microbial changes differed between patients who improved versus those who progressed to SBP.

• Improved patients had greater microbial diversity and higher levels of beneficial commensals.
• Progression to SBP was linked to expansion of Haemophilus.
• SBP-MI effectively tracked these longitudinal microbial changes.
• SBP-MI outperformed the Hepatitis B Cirrhosis Dysbiosis Index in tracking disease-related microbial changes.

Multivariable Firth logistic regression identified INR, previous history of SBP, and SBP-MI as independent predictors of SBP.

• Firth logistic regression was used to handle potential issues with sparse data in multivariable modeling.
• Three independent predictors were identified: INR, previous history of SBP, and SBP-MI.
• These three variables were combined to construct the SBP risk prediction model (SBP-RP).

The SBP risk prediction model (SBP-RP) achieved an AUC of 0.91 for predicting SBP occurrence.

• The AUC was 0.91 (95% CI: 0.82–0.99).
• Calibration of the model 'appeared acceptable in this cohort.'
• The model integrated both microbial (SBP-MI) and clinical factors (INR and previous SBP history).
• The model was evaluated in the prospective cohort of 140 cirrhotic patients with ascites followed for 6 months.

SBP-MI outperformed the existing Hepatitis B Cirrhosis Dysbiosis Index in characterizing microbial imbalance related to SBP.

• SBP-MI was constructed specifically from key microbial shifts identified in HBV-related cirrhosis patients with SBP.
• The Hepatitis B Cirrhosis Dysbiosis Index was used as a comparator index.
• SBP-MI demonstrated superior performance in tracking microbial changes during longitudinal follow-up.