Polymerase chain reaction can detect bacterial DNA in aseptically loose total hip arthroplasties

Identifying low-grade infection in failed total hip arthroplasties is an important but difficult task. This study investigated the ability of the polymerase chain reaction to identify low-grade infection during revision of total hip arthroplasties that failed from aseptic causes. One hundred thirteen specimens from 31 total hip arthroplasties revised for aseptic loosening were compared with 105 control specimens from 28 primary total hip arthroplasties. All surgeries were done in laminar flow operating rooms. No primary or revision specimen had positive microbiologic cultures. No revision specimen had histologic evidence suggestive of infection. Using the polymerase chain reaction with a detection threshold of 10 organisms per cubic centimeter of specimen, bacterial DNA was identified in 39 of 85 revision tissue specimens (46%) compared with 18 of 84 primary tissue specimens (21.4%). Bacterial DNA was identified in the synovial fluid of three specimens taken from 28 revision total hip arthroplasties (10.7%) and in two specimens taken from 21 primary total hip arthroplasties (9.5%). As multiple specimens were sent for each hip, a maximum of 16 of 31 revision total hip arthroplasties (52%) and eight of 28 primary total hip arthroplasties (29%) were considered to be infected. Bacterial DNA can be found in many specimens obtained from revised total hip arthroplasties considered to be aseptically loose. Because bacterial DNA identified at primary total hip arthroplasty was assumed to be attributable to contamination rather than present in healthy tissues, the overall specimen contamination rate of 19% and case contamination rate of 29% indicate that the polymerase chain reaction has poor specificity at this sensitivity level for diagnosing infection in revision total hip arthroplasty.