{
  "abstract": "Background The optimal timing of diagnostic paracentesis in hospitalized patients with decompensated cirrhosis and ascites remains controversial. While guidelines recommend early paracentesis, observational studies show conflicting results due to confounding by indication.Methods We performed a retrospective cohort study using the 2022 National Readmissions Database. We identified 41,254 hospitalized adults with decompensated cirrhosis who underwent paracentesis. We used double machine learning (DML) with cross-fitting to estimate the causal effect of early paracentesis (performed on hospital day 1-2) versus delayed paracentesis (day 3+) on in-hospital mortality, adjusting for 25 baseline confounders. We conducted pre-specified subgroup analyses based on illness severity and complication profiles.Results Early paracentesis was performed in 90.7% (37,437/41,254) of patients. Overall mortality was 5.1% (2,116 deaths) ( IDDF2026-ABS-0238 Table 1). After adjusting for confounding using DML, early paracentesis showed a non-significant trend toward mortality reduction (adjusted risk difference: -1.20 percentage points [95% CI: -2.47 to 0.06]; p=0.061). However, pre-specified subgroup analyses revealed significant mortality benefits in specific populations: patients without hepatic encephalopathy (-1.94 pp [95% CI: -3.21 to -0.67]; p=0.003; number needed to treat [NNT]=51), patients without renal dysfunction (-1.52 pp [95% CI: -2.83 to -0.20]; p=0.024; NNT=66), and patients with high severity illness (-1.39 pp [95% CI: -2.74 to -0.03]; p=0.045; NNT=72). Early paracentesis was associated with increased length of stay (+2.45 days [95% CI: 1.85 to 3.05]; p<0.001) and reduced days alive and out of hospital at 30 days (-1.20 days [95% CI: -1.55 to -0.86]; p<0.001) (FIDDF2026-ABS-0238 Table 2)Conclusions Early diagnostic paracentesis reduces mortality in hospitalized cirrhosis patients without hepatic encephalopathy or renal dysfunction, though at the cost of increased hospital length of stay. These findings support a precision medicine approach to paracentesis timing based on patient-specific complication profiles.Abstract IDDF2026-ABS-0238 Table 1Baseline Characteristics by Paracentesis TimingCharacteristicEarly Paracentesis (n=37,437)Delayed Paracentesis(n=3,817)P-valueSMDDemographicsAge, mean (SD)58.2 (12.4)59.1 (13.1)<0.0010.07Female, %38.636.40.0130.05Elderly (≥75y), %12.314.8<0.0010.07Disease SeverityAPRDRG Mortality Risk 3-4, %61.869.5<0.0010.16APRDRG Severity 3-4, %73.183.2<0.0010.24Decompensation count, mean (SD)1.6 (0.8)1.8 (1.0)<0.0010.22ComplicationsHepatic encephalopathy, %4.77.9<0.0010.13Variceal bleeding, %8.911.2<0.0010.08Spontaneous bacterial peritonitis, %12.39.8<0.0010.08Renal dysfunction, %11.617.8<0.0010.18Acute kidney injury, %22.428.9<0.0010.15ComorbiditiesComorbidity count, mean (SD)3.2 (1.8)3.8 (2.0)<0.0010.32Alcoholic etiology, %42.338.7<0.0010.07Diabetes mellitus, %38.942.3<0.0010.07Hypertension, %45.247.80.0030.05Cancer history, %15.618.9<0.0010.09Hospital FactorsEmergency admission, %78.982.3<0.0010.09Weekend emergency, %19.221.40.0020.05Diagnosis count, mean (SD)18.3 (8.4)22.1 (9.7)<0.0010.42Procedure count, mean (SD)4.2 (3.1)6.8 (4.5)<0.0010.67SocioeconomicLow income quartile, %28.931.20.0050.05Rural location, %8.77.90.0950.03Medicaid insurance, %33.938.2<0.0010.09Outcomes (Unadjusted)In-hospital mortality, %4.0515.67<0.0010.4Length of stay, median (IQR)5 (3-8)13 (8-20)<0.0011.12SMD = standardized mean difference. SMD >0.10 indicates meaningful imbalance.Abstract IDDF2026-ABS-0238 Table 2Treatment effects by subgroupSubgroupNBaseline Mortality (%)ATE (pp)95% CIP-valueNNTOverall41,25415.67-1.2-2.47 to 0.060.06183By Encephalopathy StatusNo encephalopathy39,05315.2-1.94-3.21 to -0.670.00351Has encephalopathy2,20122.612.07-4.91 to 9.050.562—By Renal FunctionNo renal dysfunction36,07013.13-1.52-2.83 to -0.020.02466Renal dysfunction5,18429.63-1.88-7.22 to 3.460.49—By SeverityHigh severity (3-4)31,42815.24-1.39-2.74 to -0.030.04572Low severity (1-2)9,82602.67-4.10 to 9.440.439—By Mortality RiskHigh risk (3-4)26,48919.46-1.17-2.65 to 0.310.12—Low risk (1-2)14.76501.63-1.05 to 4.320.234—By DecompensationSingle event14,19616.62-1.42-3.68 to 0.830.216—Multiple events2.84820.71-2.75-8.56 to 3.050.352—",
  "authors": [
    {
      "affiliations": [
        "Rutgers University, United States"
      ],
      "name": "Lawin Steve Mathew Lasington"
    },
    {
      "affiliations": [
        "New York Medical College, United States"
      ],
      "name": "Jeril Lasington"
    },
    {
      "affiliations": [
        "Boston University, United States"
      ],
      "name": "Swamynathan Umamaheshwaran"
    },
    {
      "affiliations": [
        "New York Medical College, United States"
      ],
      "name": "Harika Manohar"
    }
  ],
  "title": "IDDF2026-ABS-0238 Early paracentesis and mortality in cirrhosis without encephalopathy or renal dysfunction - a causal analysis",
  "uid": "fbf1b85a-ae40-5515-87ce-49459ed8d875"
}