Outcomes are very important. They help us decide if we are doing our job properly. They help patients choose who should carry out their operation and where it should be done. Increasingly these outcomes will be monitored and reported by others and this may lead to great variance in the reported outcome, particularly if case mix varies.
Therefore it is crucial that any outcome that is measured has a standard definition that is relatively easy to measure . Death of course is the easiest outcome to measure but in isolation of other contributing factors, is not a helpful comparator. Despite this, hospital mortality rates remain a popular means of comparing institutions.1
Surgical site infection (SSI) is important as it can significantly affect outcome. The rate of SSI within a category of surgical procedures varies considerably between hospitals. Some of this may be due to variation in risk factors in the patient population, and furthermore when the rate is based on small numbers of procedures, the estimate will not be accurate.2
The article by Willis-Owen et al is to be commended for using a clear definition of infection when looking at factors affecting the incidence of infection after hip and knee replacement.3
Infection following any operation can reflect many factors including the comorbidity of the patient, the cleanliness of the hospital, the skills of the operative team, the appropriate use of antibiotics and the standard of nursing so such comparisons are not straightforward.4,5 Deep infection following an arthroplasty is a disaster for the patient and expensive to manage.6
The trouble is that the definition of infection used in many papers in this and other journals seems remarkably woolly and it is timely that an agreed definition should be used. The problem of course is that any large series of arthroplasties reported will have been collected over a prolonged time period and the definition of infection may have changed. Is superficial redness around the skin wound at ten days which is culture negative a SSI? What about the wound swab taken routinely at the time of revision which grows a coagulase negative staph, when five others and all the inflammatory markers are negative?
The easiest time to identify infections should be during the immediate postoperative period when the patient is in hospital; although with better anaesthetics and advances in physiotherapy, the average length of stay in orthopaedics has decreased from 7 days in 2004 to 5 days in 2007/8 meaning that infections can be missed.7
The Hospital Infection Society (HIS) definition simply required there to be a break in the epithelial surface with pus, fever, erythema or a physician diagnosis8 as had previously been suggested by Glenister.9 Specific definitions were produced by the CDC in 1992 and this was refined in 1998.10,11 For all prospective monitoring a clear definition as used in this paper or that available from the Health Protection Agency (HPA) should be used.12 The current HPA guidance definitions are:
A superficial incisional infection arises within 30 days of the operation and has one of the following:
1. Purulent drainage from the superficial incision
2. Positive culture from the sinus or tissue and pus cells are present
3. At least two of the following symptoms and signs are present:
a) pain or tenderness
b) localised swelling
and i) the superficial incision is deliberately reopened by surgeon to treat the infection unless the cultures are negative or ii) the clinican diagnoses a superficial incisional infection.
A deep incisional infection is diagnosed if the deep tissues around an implant are infected within one year in association with any of the following:
1. Purulent discharge from the organ space
2. Positive culture from the organ space with pus cells present
3. Spontaneous dehiscence or opening of a wound in association with a temperature >38°C or localized pain and tenderness
4. Abscess or other proof of deep infection confirmed by surgery, histology or radiology
5. Diagnosis of deep infection by attending surgeon or physician
This guidance recognises that with decreasing lengths of stay, many of these events will be happening in the community where very different perceptions of the possibility of infection and the use of antibiotics may be prevalent. Although recording of events after discharge is recommended, it still remains optional and this means that a hospital rigorously following up all its patients could still have a much higher rate of infection than one just relying on the current inpatient surveillance.13
These discrepancies have also been identified by NICE.5 Other series have used the definition of a positive culture from the deep tissues at any stage to define a deep infection and have continued surveillance indefinitely. This has shown that of all deep infections, 29% will arise in the first 3 months (usual definition of early), 35% will arise in the first year and 36% will arise after that.14
Whilst these definitions have been used by many authors to report their infection rates, others have been less impressed with the reliability of clinical features and have relied on histology also with the presence of 5 neutrophils/hpf indicating infection even if cultures were negative.15,16
Patients perceptions of infection may also vary. Blom and his colleagues investigated this by asking patients if they recalled having an infection following their hip or knee replacement, asking them specifically if they recalled a hospital readmission, further surgery or being put on antibiotics. Their results showed a 1.08% rate of infection for primary hip replacements and similar for knees.17
Long established registries have used more pragmatic definitions of infection such as the need for revision where infection is identified. The Finnish registry records an infection rate of 0.33% for unicompartmental knee replacements; 0.52% for primary knee replacements and 1.91% for revision knee replacements and believes that it probably underestimates the true incidence.18 They point out that this is because of the difficulty of capturing all subsequent events. The Norwegian register also points out that a proportion of revisions done for aseptic loosening will in fact be due to infection and that the diagnosis of infection will be proportional to the efforts made to detect it.19
Every year this Journal publishes numerous articles which mention infection rates after either hip or knee arthroplasty. Virtually none of them define how the diagnosis of infection was made and usually the figure reported is the number of revisions done for infection. To be meaningful, a clear definition of infection should be agreed and adhered to when reporting infection rates. How else can we truly compare our outcomes?
1. Jarman B, Bottle A, Aylin P. Monitoring Changes in Hospital Standardised Mortality Ratios. BMJ 2005;330:339.
2. Health Protection Agency. Surveillance of Healthcare Associated Infections Report, 2008. London: Health Protection Agency.
3. Willis-Owen CA, Konyves A, Martin DK. Factors affecting the incidence of infection in hip and knee replacement. J Bone Joint Surg [Br] 2010;92-B:1128-33.
4. Bosco J, Slover J, Haas J. Perioperative Strategies for Decreasing Infection: A Comprehensive Evidence-Based Approach. J Bone Joint Surg [Am] 2010;92-A:232-239.
5. National Institute for Health and Clinical Excellence. Surgical site infection 2008(clinical guideline 74) www.nice.org.uk/CG74 (accessed 18 August 2010).
6. Mangram A, Horan T, Pearson M, Silver L, Jarvis W. Guideline for Prevention of Surgical Site Infection. Am J Infect Control 1999;27:97-134.
7. Health Protection Agency 2009. Trends in Rates of Healthcare Associated Infection in England 2004 – 2008. London: Health Protection Agency.
8. Spencer R. National Prevalence Survey of Hospital Acquired Infection: Definitions. A Preliminary Report of the Steering Group of the Second National Prevalence Survey. J Hosp Infect 1993;24:69-76.
9. Glenister HM, Taylor J, Cooke EM, Bartlett CL. A Study of Surveillance Methods for Detecting Hospital Infection. Public Health Laboratory Service 1992.
10. Horan T, Gaynes R, Martone W, Jarvis W, Emori T. CDC definition of Nosocomial Surgical Site Infection : a modification of the CDC definitions of surgical wound infection. Infect Control Hosp Epidem 1992;13:606-8.
11. Crowe M, Cooke E. Review of Case Definitions for Nosocomial Infections – Towards a Consensus. J Hosp Infect 1998;39:3-11.
12. Protocol for surveillance of surgical site infection. Health Protection Agency. http://www.hpa.org.uk/web/HPAwebFile/HPAweb_C/1194947388966 (accessed 18 August 2010).
13. Wilson J, Charlett A, Leong G, McDougall C, Duckworth G. Rates of surgical site infection after hip replacement as a hospital performance indicator: analysis of data from the English mandatory surveillance system April 2004-March 2005. Infect Control Hosp Epid 2008;29:219-226.
14. Phillips J, Crane T, Noy M, Elliott T, Grimer R. The incidence of deep prosthetic infection in a specialist orthopaedic hospital: a 15 year prospective survey. J Bone Joint Surg [Br] 2006:88-B;943-948.
15. Saleh K, Olson M, Resig S, et al. Predictors of wound infection in hip and knee replacement, results from a 20 year surveillance program. J Orth Research 2002;20:506-515.
16. Atkins B, Athanasou N, Deeks JJ, et al. Prospective evaluation of criteria for microbiological diagnosis of prosthetic joint infection at revision arthroplasty. J Clin Microbiology 1998;36:2932-9.
17. Blom A, Taylor A, Pattison G. Whitehouse, S. Bannister G. Infection after total hip arthroplasty, the Avon experience. J Bone Joint Surg [Br] 2003;85-B:956-959.
18. Jamsen E, Huotari K, Huhtala H, Nevalainen J, Konttinen Y. Low rates of infected knee replacements in a nationwide series – is it an underestimate? Acta Orth 2009;80:205-12.
19. Engesaeter L, Espehaug B, Lie S, Furnes O, Havelin I. Does cement increase the risk of infection in primary total hip arthroplasty? Acta Orth 2006;77:351-8.
Grimer R, FRCS
Ludman L, BA, SRN
The Royal Orthopaedic Hospital, Birmingham, United Kingdom