Background: Reports of HIV-related inpatient hospital admissions have typically been based on local, cross-sectional data. Using comprehensive hospital discharge data from 12 states (including CA, FL, NJ, and NY), we evaluated trends in HIV-related inpatient admission rates and lengths of stay in 1996, 1998, and 2000 as a function of selected discharge diagnoses.

Methods: Using ICD-9 diagnoses from discharge abstracts in the Health Care Utilization Project-State Inpatient Database (SID), we identified HIV-related admissions and classified them as involving opportunistic illness (OI), complications of injection drug use (IDU), and liver related complications (LRC). IDU complications were defined as abscess, cellulitis, osteomyelitis, bacteremia, endocarditis, and poisoning by analgesics. LRC were identified as acute and subacute necrosis of the liver, chronic liver disease and cirrhosis, liver abscess, hepatic coma, portal hypertension, hepatorenal syndrome, hepatocellular carcinoma, gastrointestinal bleed, Mallory-Weiss Tear, and viral hepatitis.

Results: We evaluated 327,306 HIV hospitalizations in the 12 states under study. Between 1996–2000, HIV hospitalizations for OIs decreased significantly from 41%–29% of all HIV hospitalizations (p < 0.001); hospitalizations for complications of IDU remained relatively constant between 5.9%–6.9%; and hospitalizations for LRC increased significantly from 13%–18% of all HIV hospitalizations (p < 0.001). Mean length-of-stay was significantly longer for LRCs than non-LRCs (10.4 vs 8.9 days, p < 0.0001). The proportion of all HIV admissions covered by Medicare increased from 17%–23%. By contrast, the proportion of LRC Medicare admissions rose from 17%–25% (p < 0.0001). Conclusions: Our results show, in multiple states, declines in OI-related hospital admission rates, but an increase in hospitalization rates for LRC as well as longer LOS for LRC admissions between 1996–2000. If this trend continues, LRC may become one of the principal comorbidities in HIV infected patients. This could have a dramatic impact on the burden of disease, the costs of care, and publicly funded insurance.