J. R. Coll. Surg. Edinb., 43, February 1998, 43-44
Does early wound infection after elective orthopaedic surgery lead on to chronic sepsis?
J. R. BENGER, A. J. KELLY AND I. G. WINSON
Department of Orthopaedic Surgery, Southmead Hospital,
Westbury on Trym, Bristol, UK
Infection is an uncommon, but occasionally devastating, complication of orthopaedic surgery. The definition of post-operative infection remains problematic. A high rate of early post-operative sepsis has previously been reported using a clinical definition of wound infection as recommended by the Surgical Infection Study Group. The purposes of this study is to determine the rate of ongoing wound problems and deep sepsis 1 year after these early wound infections. Of 1131 consecutive orthopaedic procedures, there were 70 wound infections occurring within 30 days of surgery. Adequate follow-up data were obtained in 67(97%) of the 69 patients alive at 1 year. Of these 67, three had definite evidence and two possible evidence of ongoing wound problems and/or deep sepsis. It is concluded that early post-operative wound infection as defined by the Surgical Infection Study Group is a poor predictor (4-10%) of ongoing wound problems and deep sepsis at 1 year. All of the confirmed cases of late sepsis were found to be associated with revision arthroplasty and/or pin tract sepsis.
Keywords: infection, orthopaedic surgery.
We have previously reported an apparently high rate of early post-operative wound sepsis after elective orthopaedic surgery in a single centre,1 using a clinical definition of wound infection as recommended by the Surgical Infection Study Group.2 These definitions are described as 'memorable, easy to apply and based on clinical evidence'. A substantially higher audited rate of wound infection is produced by this clinical definition than by one which requires bacteriological confirmation, and this was clearly shown in our earlier study.1
After 1131 consecutive procedures, there were 53 'minor' and 17 'major' wound infections occurring within 30 days of surgery. The purpose of this study was to determine the rate of ongoing wound problems and deep sepsis in these patients at 1 year.
Seventy patients who were defined as having had early wound infection were assessed by clinical review or postal questionnaire at least 18 months after their original orthopaedic surgery. The following criteria were assessed: pain at rest, persistent pyrexia, continuing wound erythema or discharge, increased erythrocyte sedimentation rate, increased white blood cell count, positive deep cultures and radiological changes suggestive of infection.
Fifty of the 70 patients were reviewed in an outpatient clinic, and the remaining 20 were sent a short questionnaire designed to detect any ongoing problems. Non-responders, and those in whom the questionnaire suggested any possibility of late infection (i.e. a single positive response to any of a list of screening questions) were contacted personally by telephone, as was their GP, for further information. Those few patients suspected of having evidence of late-infection were then examined in more detail.
From the original study group of 1131 patients, 70 were known to have had clinical evidence of early wound infection, with 53 of these being defined as minor infections and 17 major.
Clinical review was available for 50 patients and questionnaires were sent to the remaining 20, of which 14 were returned. Three of the six outstanding patients were contacted by telephone and adequate information was obtained by contacting the GP of the fourth. Two patients were lost to follow-up. One patient died 2 months after her original surgery of an unrelated cause. Adequate follow-up information was therefore available for 67 patients (i.e. 97% if those alive at 1 year).
Of these 67 patients, 62 (93%) had no evidence of ongoing wound problems or deep sepsis. In the remaining five, three had definite deep sepsis related to their original procedure, while in the other two there was a clinical suspicion of deep sepsis without objective confirmation. In the early post-operative phase two of these five patients were judged to have 'minor' infections (patients, 1 and 5), while the remainder had infections classified as 'major'. These five patients are now considered in more detail.
Case 1
A 69-year-old man underwent a revision total knee replacement after supracondylar fracture of the femur. A Denham pin was used for skeletal traction, but he developed a proven pin tract infection which was, however, thought to have resolved by the time of his revision. Subsequently the implant became infected, requiring a two-stage revision.
Case 2
A 58-year-old man underwent open ankle arthrodesis using a compression technique. An infection developed in the pin sites of his external fixator, which then led on to sinus formation with continuing pain and non-union at the arthrodesis.
Case 3
A 73-year-old woman developed deep infection after a two-stage revision of an infected total hip prosthesis.
Case 4
A 68-year-old man required revision of a total hip replacement at 5 months for persistent pain and early periprosthetic lysis suggesting infection. Biopsy samples taken at the time of the revision did not confirm infection and the patient had none of the above symptoms and signs of infection 1 year after his one-stage revision.
Case 5
A 51-year-old woman had pain and delayed union after a first metatarsophalangeal arthrodesis. The possibility of infected delayed union was considered, particularly in view of her early post-operative wound infection, but at exploration there was no evidence of infection and the arthrodesis united after compression screw fixation.
The ideal definition of post-operative wound infection remains problematic.1,2 A substantially higher audited rate of wound infection is produced by applying the clinical definition proposed by the Surgical Infection Study Group than by one which requires microbiological confirmation. After apparent early post-operative wound infection in 70 patients, only three (4%) had definite ongoing wound problems or deep sepsis at 1 year. The worst case scenario for late infections, including the two cases which were not confirmed objectively and the two lost to follow-up, would be 10%, and this represents an overall late infection rate of 0.6% in the original series of 1131 patients. In contrast with this use of clinical definitions, Johnson and Bannister showed that early microbiologically proved superficial wound infection preceded 30% of infected total knee replacements.3
The purpose of this study was to determine the rate of ongoing wound problems and deep sepsis after early wound infection and not the overall prevalence of infection at 1 year. We have therefore not assessed the 1061 patients from the study who did not have any evidence of early wound sepsis, and the prevalence of deep sepsis in this group remains unknown.
Two cases of late sepsis occurred in association with revision arthroplasty and two with the use of an external fixator or skeletal traction. Revision joint surgery,3a open ankle arthrodesis4 and external fixation5 have all been previously associated with a higher than average infection rate.
The finding of early post-operative wound infection in 70 of 1131 patients (6%) has important implications for patient counselling, general practice and community services in the early postoperative period. However, early wound infection, as defined in this way, is a poor predictor (4-10%) of ongoing problems and deep sepsis at 1 year. It appears that the chance of deep sepsis is primarily determined by the procedure itself, with revision arthroplasty and external fixation representing the highest risk.
Paper accepted 10 December 1996
Correspondence: J. R. Benger, 14 Lakewood Road, Henleaze, Bristol BS1O 5HH, UK.
© 1998 The Royal College of Surgeons of Edinburgh, J. R. Coll. Surg. Edinb., 43, February, 43-44