Table 5

 Evidence on medical versus surgical drainage of the septic joint

StudyStudytypeNo of patientsDuration of studyTreatment/interventionOutcome measuresConclusions from studyLevel of evidence providedCaveats
Vispo Seara et al35,Retrospective886-Year collection period, mean follow- up 2.5 yearsAll patients had arthroscopy, but depending on grade, also had shaving, limited synovectomy, resection of adhesions, or debridement of loose bone and cartilage. Divided into four groups (Gachter score)Number of procedures, Lysholm functional score, extension, flexion and painCorrelated outcome to presence of premorbid degenerative change, age of patient, and delay from onset to first surgery. Number of procedures was related to organism, but final result was not related to organismBOutcome may be related to several factors. Study power not sufficient to control for them all. Surgery not same for all groups - outcome may be determined by surgery rather than severity
Wirtz et al36Retrospective5112-Year collection period, mean follow-up 2.2 years24 in arthrotomy and synovectomy group, 27 in arthroscopy group. Both groups drainedLarson score, range of motion (total). Correlated with time to treatmentArthroscopic treatment was better than arthrotomy and synovectomy on all measures overall, but significant difference only for patients treated within 5 days when analysed relative to time from onsetBStatistical analysis limited, and the study is underpowered
Stutz et al,37Retrospective7610 Year collection period, mean follow-up not statedArthroscopy, irrigation and debridement where necessary, all groups. Divided into three groupsNumber of procedures, presence of infection at 1 yearThe number of procedures required to eradicate infection is related to the stage of the infection at presentationBStaging may have led to different (more interventional) operations for higher grades, so more than one variable. Underpowered
Richard et al38Retrospective875 Year collection, mean follow-up not statedMixed: “conservative” (open lavage/synovectomy, debridement/splintage/oral antibiotics) or “joint sacrifice” - arthrodesis, joint excision or amputationPresence or absence of infection Description of functional results50% cure of infection with “conservative” treatment. 31/87 amputated. Articular resection remained painful and stiff. There were 8 arthrodeses Regional infection led to poor resultsBConclusions specific to finger surgery. No generalisations can be made. No detailed statistical analysis
Thiery39Multicentre, retrospective, questionnaire46Average follow-up 7.1 monthsAll cases had arthroscopic drainageResults according to aetiology, pathological agent, delay prior to arthroscopy, analysis of “failures of treatment”80% cure rate, 10% failure, 10% relapse. Better if done early. No cure if >3 monthsBNo comparator group
Goldenberg et al40Retrospective review598 Years42 Needle aspirations, 17 open lavageComplete recovery, poor result indicated by reduction of movement, ankylosis, secondary osteomyelitis, persistent effusion, deathBetter function in needle aspiration group (67% vs 42% good result). 12% vs 5% death rate for aspiration group vs surgicalBHeterogeneous groups, hips mostly surgical, wrists mostly aspirated. Study underpowered to answer the question