Background:  The perception remains that HIV⁺ patients would have a worse surgical outcome than HIV^- patients because of suppressed immune status and poor viral control. In a matched design we compared HIV⁺ patients to HIV^- patients with respect to surgical outcomes in the HAART era. 

Methods:  All selected procedures (appendectomy, cardiac surgery, mastectomy, cholecystectomy, hip and knee replacement/arthroplasty, and hernia repair) among more than 5000 HIV⁺ members of the Kaiser Permanente Northern California (KP) HMO (cases) in the period July 1997 to June 2002, were identified for review. Cases were matched 1:1 to randomly selected HIV^- KP patients (controls) on type, year and location of surgery, and gender and age at surgery. Patient characteristics, pre-surgical conditions (CD4, VL, Hgb, WBC, use of antibiotics, and prior DM, HTN, CAD, smoking, OI), as well as complications and mortality through 12 months post-op were collected from paper medical records and electronic databases. Concordance between pairs with respect to presence/absence of complications was analyzed using McNemar’s test. 

Results:  To date, 641 charts have been reviewed and 256 of 395 (65%) case-control pairs have been analyzed. Mean age of cases was 47 years; 92% were male; 95% of cases and 93% of controls were followed through 12 months post-op or death. Cases and controls were similar with respect to Hgb, WBC, pre-surgical co-morbidities, mean LOS, and number of post-op surgical visits. Among cases, mean CD4 was 429 cells/mL, 51% had VL BLQ, 64% were taking 3 or more ARV, 67% previously diagnosed AIDS, and mean time since HIV diagnosis was 8 years. Complications were various but uncommon (including infection) and were not more frequent among cases than controls (10.9% vs 9.8%; McNemar’s exact, p = 0.77).  Cases with VL >10,000 vs <10,000 at surgery were more likely to have complications (23% vs 8%, Fisher’s exact, p = 0.005), but not with regard to CD4 <200 (p = 0.58). There were more (NS) deaths in 1 year post-op among cases than controls (7/256 vs 2/256) but they could be attributed to preexisting diagnoses.  

Conclusions:  To our knowledge, this is the largest study of surgical outcomes in HIV. Outcome for a variety of surgeries in the HAART era is no different for HIV⁺ than for HIV^- patients. HIV status should not by itself be a criterion for surgical consideration. Pending multivariate analysis, viral suppression to below 10,000 copies might improve surgical outcomes among HIV⁺ patients.

Keywords: Surgery; HIV/AIDS; Outcome