Risk factors for superficial wound complications in hip and knee arthroplasty
Superficial wound complications have been consistently implicated in the development of prosthetic joint infection. This cohort study aimed to determine perioperative risk factors associated with superficial wound complications. The study was performed over an 18-month period (January 2011 to June 2012) and included 964 patients undergoing prosthetic hip or knee replacement surgery. The factors associated with superficial wound complication differed according to arthroplasty site. In the combined cohort the following factors were associated with superficial wound complications: the use of 0.5% chlorhexidine in 70% alcohol for surgical skin preparation compared with 1% iodine in 70% alcohol (odds ratio (OR) 4.75; 95% confidence interval (CI) 1.42, 15.92; p = 0.012); increasing age (OR, 1.13; 95% CI, 1.06,1.19; p 0.18); increasing body mass index (BMI) (OR, 1.08; 95% CI, 1.05,1.12; p < 0.001); rheumatoid arthritis (OR, 2.56; 95% CI, 1.17, 5.58; p 0.018); and increasing blood transfusions (OR, 1.26; 95% CI, 1.06,1.49; p 0.008). In the hip arthroplasty cohort, the use of 0.5% chlorhexidine in 70% alcohol for surgical skin preparation (OR, 13.35; 95% CI, 2.11, 84.29; p 0.006), increasing BMI (OR, 1.13; 95% CI, 1.06, 1.19; p < 0.001) and increasing blood transfusions (OR, 1.26; 95% CI, 1.06, 1.49; p 0.008) were associated with superficial wound complications. In the knee arthroplasty cohort rheumatoid arthritis (OR, 2.75; 95% CI, 1.03, 7.33; p 0.043) and increasing tourniquet time (OR, 1.01; 95% CI, 1.00, 1.02; p = 0.029) were independent predictors of superficial wound complications. Further research is warranted to assess the impact of modification of these factors on the subsequent development of wound complications and prosthetic joint infection.
Prosthetic joint surgery has led to improvement in symptoms of osteoarthritis and patient's quality of life. The popularity of this surgery continues to increase worldwide, including in Australia . Data from epidemiological studies estimate that the number of knee and hip arthroplasties will increase to 4.5 million procedures per year by 2030 in the United States of America . Infective complications of the prosthesis occur in 1–3% of patients undergoing hip and knee arthroplasty; however, they can have catastrophic consequences and often result in patient morbidity and significant cost to the public healthcare system . In addition to the direct complications of infection, the need for treatment with antibiotics contributes to emerging antimicrobial resistance . Superficial wound complications, such as surgical site infection (SSI) and prolonged wound discharge, have been consistently implicated in the development of prosthetic joint infection and may increase the risk of subsequent deep infection by up to 35-fold [5, 6]. While there are a number of epidemiological studies examining risk factors for deep prosthetic infections in hip and knee arthroplasty, risk factors for superficial wound complications are not as well established.
The aims of this study were (i) to describe the incidence and severity of superficial wound complications in a cohort of patients undergoing hip and knee arthroplasty and (ii) to elucidate perioperative factors associated with the development of superficial wound complications.
Patients and Methods
Study setting and population
This retrospective cohort study was performed at St Vincent's Hospital Melbourne (SVHM), Victoria, Australia, an 848-bed tertiary public hospital. This centre comprises 16 orthopaedic surgeons performing over 800 prosthetic hip and knee replacements per year. All hospital postoperative care is performed according to the SVHM Hip or Knee Replacement Clinical Pathway, which has been described previously . All patients received antibiotic prophylaxis, in keeping with national guidelines, prior to skin incision and >5 min prior to tourniquet inflation . The study was performed over an 18-month period (January 2011 to June 2012) and included all patients undergoing elective primary and revision hip and knee arthroplasty during this period, as systematically recorded in a database prospectively compiled by the Department of Orthopaedics . Patients were excluded if they underwent arthroplasty for fractured neck of femur, or revision arthroplasty with prior history of septic arthritis affecting the index joint, or if they were lost to follow-up. Patients were followed-up from the date of index arthroplasty until review in the orthopaedic clinic 6 weeks post-surgery. The study design was reviewed and approved by the SVHM Human Research Ethics Committee.
Superficial wound complication was recorded if the patient developed either a superficial incisional SSI or prolonged wound ooze in the 30 days following index arthroplasty surgery.
Superficial incisional SSI was defined as per the US Centers for Diseases Control and Prevention (CDC) Criteria [10, 11]. In keeping with the US CDC criteria, a superficial incisional SSI was not recorded in the case of stitch abscess, navigation pin site infection, cellulitis not in communication with the wound, or in the case that a general practitioner commenced antibiotics for ‘swelling’, ‘erythema’ or ‘increased pain’, which was not subsequently recognized as a superficial incisional SSI by the orthopaedic surgeon or treating physician .
Prolonged wound ooze was defined if there was documented drainage from the surgical incision that required intervention, such as superficial surgical debridement, or if it led to deviation from normal care as per the SVHM arthroplasty clinical pathway, such as delayed discharge from hospital .
A data collection spreadsheet was designed to document potential risk factors for superficial wound complications. Risk factors were drawn from the current literature as well as novel, biologically plausible risk factors and included patient co-morbidities and factors surrounding operative and postoperative care. Information was extracted from the database and from careful medical chart review by a single researcher (KC). Information regarding postoperative complications was obtained primarily from orthopaedic clinic notes, as well as emergency department or medical record entries if readmission occurred. All management strategies and decisions were not randomized and were based on the treating surgeon's preference at the time of surgery. There was no telephone or recall contact with patients for the purposes of this study.
Descriptive statistics were used to summarize and report the data. Descriptive analyses were based on percentages and frequencies for categorical variables and for continuous variables, mean and standard deviation (SD) or medians and interquartile range (IQR) if the data were skewed. Logistic regression was performed to produce odds ratios (ORs) with 95% confidence intervals (CIs) for the association between each variable and the presence or absence of superficial wound complication. Multivariable logistic regression techniques were used in the assessment of risk factors by adding in forward substitution factors identified as significant in the univariate analysis (p value <0.1) or risk factors previously identified in the published literature. Fisher's exact test was performed to compare rates of superficial wound complication and prosthetic joint infection. All reported p-values were two-tailed and for each analysis p < 0.05 was considered statistically significant. All analyses were performed using Stata 11.2 (StataCorp, College Station, TX, USA, 2009).
Over the 18-month study period, 1006 patients underwent prosthetic hip or knee replacement surgery at SVHM. Forty-two patients were excluded from the study (eight with fractured neck of femur, 12 with septic revision arthroplasty and 22 who were lost to follow-up). Therefore, 964 patients were included in this current study (453 hip and 511 knee arthroplasties). The demographic characteristics of the cohorts are outlined in Table 1.
Table 1. Demographic characteristics of the hip and knee arthroplasty cohorts
|Median age, years (IQR)||69 (60, 75)||70 (64, 76)|
|Median BMI, kg/m2 (IQR)||29.7 (26.2, 33.7)||33.7 (29.6, 38.4)|
|Female gender||272 (60%)||335 (66%)|
|Rheumatoid arthritis||19 (4%)||39 (8%)|
|Immunosuppressant medications||10 (2%)||32 (6%)|
|Systemic corticosteroids||13 (3%)||25 (5%)|
|Diabetes mellitus||58 (13%)||118 (23%)|
|Preoperative prescription of warfarin||28 (6%)||47 (9%)|
|American Society of Anaesthesiologists (ASA) score|
|1||15 (3%)||12 (2%)|
|2||25512 (2%)||249 (49%)|
|3||171 (38%)||233 (46%)|
|4||12 (3%)||17 (3%)|
|Surgical antibiotic prophylaxis|
|Cefazolin||437 (96%)||483 (96%)|
|Vancomycin||12 (3%)||22 (4%)|
|Gentamycin||345 (76%)||395 (77%)|
|Image guided surgery (IGS)||–||94 (18%)|
|Median tourniquet time, minutes (IQR)||–||75 (60, 90)|
|Median operation time, minutes (IQR)||100 (85, 115)||95 (85, 115)|
|Staples||324 (72%)||397 (81%)|
|Subcuticular sutures||123 (28%)||94 (19%)|
|Presence of drain tube||165 (36%)||262 (51%)|
|Median drain output in mL (IQR)||250 (140, 400)||380 (140, 700)|
|Median transfusion, units (range)||0 (0, 8)||0 (0, 7)|
|Surgical skin preparation|
|1% iodine in 70% alcohol||447 (99%)||501 (98%)12|
|0.5% chlorhexidine in 70% alcohol||6 (1%)||9 (2%)|
Hip and knee cohorts were analysed as a combined cohort and then separately. Overall, 88 (9%) patients developed a superficial wound complication (42 had superficial incisional SSI, 40 had prolonged wound discharge and six had both).
The combined cohort results of the univariate and multivariate logistic regression analysis are outlined in Table 2. On univariate analysis, the following factors were associated with the development of a superficial wound complication: increasing age, increasing body mass index (BMI), rheumatoid arthritis, warfarin, the use of 0.5% chlorhexidine in 70% alcohol for surgical skin preparation prior to surgical incision compared with 1% iodine in 70% alcohol, blood transfusion and skin closure with staples rather than sutures. On multivariate analysis, the following factors were independently associated with the development of superficial wound complications: increasing age, increasing BMI, rheumatoid arthritis, skin preparation with 0.5% chlorhexidine in 70% alcohol and increasing number of units of blood transfused in the postoperative period.
Table 2. Univariate and multivariate analysis of risk factors for superficial wound complications in the combined cohort
|Age, years||1.02 (1.00, 1.05)||0.044||1.03 (1.00, 1.06)||0.018|
|Female gender||0.98 (0.62, 1.54)||0.9|| || |
|Prosthetic hip replacement||Reference||–|| || |
|Prosthetic knee replacement||0.92 (0.59, 1.43)||0.7|| || |
|BMI, kg/m2||1.06 (1.02, 1.09)||0.001||1.08 (1.05, 1.12)||<0.001|
|Warfarin||2.04 (1.05, 3.94)||0.035|| || |
|Rheumatoid arthritis||1.97 (0.93, 4.15)||0.077||2.56 (1.17, 5.58)||0.018|
|Immunosuppressant medication||2.08 (0.89, 4.82)||0.089|| || |
|Systemic corticosteroids||1.93 (0.78, 4.75)||0.2|| || |
|Diabetes mellitus||1.36 (0.80, 2.30)||0.3|| || |
|Surgical skin preparation|
|1% iodine in 70% alcohol||Reference||–|| || |
|0.5% chlorhexidine in 70% alcohol||3.74 (1.17, 12.0)||0.027||4.75 (1.42, 15.92)||0.012|
|Staples||Reference||–|| || |
|Subcuticular sutures||0.50 (0.27, 0.95)||0.033||0.54 (0.29, 1.03)||0.063|
|Operation time (min)||1.00 (0.99, 1.01)||0.8|| || |
|Presence of drain tube||0.95 (0.61, 1.48)||0.8|| || |
|Drain loss (mL)||0.99 (0.99, 1.00)||0.8|| || |
|Blood transfusion (units of red blood cells)||1.24 (1.06, 1.44)||0.006||1.26 (1.06, 1.49)||0.008|
In the cohort of 453 patients who underwent elective hip arthroplasty, a wound complication occurred in 43 patients at a rate of 9%. Readmission to hospital for management of the wound complication was necessary in 11 of these cases (26%). In the univariate analysis of hip arthroplasty patients, the following factors were associated with superficial wound complication: increasing BMI, use of 0.5% chlorhexidine in 70% alcohol for surgical skin preparation, blood transfusion and warfarin therapy. On multivariate analysis increasing BMI, skin preparation with 0.5% chlorhexidine in 70% alcohol and increasing blood transfusion requirements predicted superficial wound complications (Table 3).
Table 3. Univariate and multivariate analysis of risk factors for superficial wound complications in the hip arthroplasty cohort
|Age, years||1.02 (0.99, 1.06)||0.1|| || |
|Female gender||1.26 (0.64, 2.49)||0.5|| || |
|BMI||1.12 (1.06, 1.18)||<0.001||1.13 (1.06, 1.19)||<0.001|
|Warfarin||3.96 (1.57, 10.00)||0.004||2.93 (0.99, 8.63)||0.050|
|Rheumatoid arthritis||1.31 (0.29, 5.89)||0.7|| || |
|Immunosuppressant medications||2.66 (0.55, 12.99)||0.2|| || |
|Systemic corticosteroids||1.92 (0.41, 9.00)||0.4|| || |
|Diabetes mellitus||0.74 (0.25, 2.16)||0.6|| || |
|Surgical skin preparation|
|1% iodine in 70% alcohol||Reference||–||–||–|
|0.5% chlorhexidine in 70% alcohol||11.08 (2.16, 56.86)||0.004||13.35 (2.11, 84.29)||0.006|
|Staples||Reference||–|| || |
|Subcuticular sutures||0.44 (0.18, 1.07)||0.070|| || |
|Operation time (min)||1.00 (0.99, 1.01)||0.8|| || |
|Presence of drain tube||1.85 (0.96, 3.55)||0.07|| || |
|Drain loss (mL)||0.99 (0.99, 1.00)||0.7|| || |
|Blood transfusion (units of red blood cells)||1.32 (1.09, 1.60)||0.005||1.37 (1.10, 1.70)||0.005|
In the cohort of 511 knee arthroplasty patients, a wound complication occurred in 45 patients at a rate of 9%. Readmission to hospital for management of the wound complication was necessary in 15 of these cases (33%). The following factors were identified on univariate analysis: rheumatoid arthritis, diabetes mellitus and increasing tourniquet time. On multivariate analysis, rheumatoid arthritis and increasing tourniquet time predicted superficial wound complications (Table 4).
Table 4. Univariate and multivariate analysis of risk factors for superficial wound complications in the knee arthroplasty cohort
|Age, years||1.02 (0.98, 1.05)||0.4|| || |
|Female gender||0.85 (0.45, 1.61)||0.6|| || |
|BMI||1.03 (0.98, 1.08)||0.2|| || |
|Warfarin||1.26 (0.47, 3.37)||0.6|| || |
|Rheumatoid arthritis||2.50 (1.03, 6.04)||0.042||2.75 (1.03, 7.33)||0.043|
|Immunosuppressant medication||2.03 (0.74, 5.57)||0.2|| || |
|Systemic corticosteroids||2.07 (0.68, 6.31)||0.2|| || |
|Diabetes mellitus||1.57 (0.80, 3.07)||0.1||2.03 (0.99, 4.16)||0.054|
|Surgical skin preparation|
|1% iodine in 70% alcohol||Reference||–|| || |
|0.5% chlorhexidine in 70% alcohol||1.29 (0.16, 10.6)||0.8|| || |
|Staples||Reference||–|| || |
|Subcuticular sutures||0.64 (0.26, 1.57)||0.3|| || |
|IGS||1.30 (0.62, 2.73)||0.5|| || |
|Tourniquet||2.94 (0.33, 18.85)||0.6|| || |
|Tourniquet time (min)||1.01 (1.00, 1.02)||0.029||1.01 (1.00, 1.02)||0.029|
|Operation time (min)||1.00 (0.99, 1.01)||0.9|| || |
|Presence of drain tube||0.60 (0.32, 1.13)||0.1|| || |
|Drain loss (mL)||1.00 (0.99, 1.00)||0.5|| || |
|Blood transfusion (units of red blood cells)||1.08 (0.82, 1.43)||0.6|| || |
Prosthetic joint infection
Over the course of the study, 14 patients presented with prosthetic joint infection (seven prosthetic hip and seven prosthetic knee infections), with an overall infection rate of 1.45%. In prosthetic hip infections, six patients had preceding superficial wound complications (Fisher's exact test p < 0.001). In prosthetic knee infections, four patients had preceding superficial wound complications (Fisher's exact test p 0.006).
Epidemiological studies have consistently implicated superficial wound complications in the development of deep prosthetic joint infection [5, 13]. Indeed, in this study 71% of patients subsequently diagnosed with prosthetic joint infection had preceding superficial wound complications. We postulate that identification and modification risk factors for superficial wound complications will lead to a decrease in the incidence of subsequent deep prosthetic joint infection. In addition, superficial wound complications themselves are associated with patient morbidity and cost to the healthcare system independent of the development of deep infection, such as prolonged hospital stay, readmission, ongoing treatments and reduced patient satisfaction [5, 6, 14, 15]. In this study, 9% of patients developed a superficial wound complication and of these, 30% required readmission to hospital for management. This study has identified a number of novel and clinically relevant risk factors for superficial wound complications. In addition, this study has highlighted the risk factors that differ according to arthroplasty site.
The optimal agent for surgical skin preparation is a controversial issue. In a large, multicentre randomized clinical trial in clean-contaminated surgery, skin preparation with 2% chlorhexidine gluconate and 70% isopropyl alcohol was associated with a reduced number of superficial and deep SSIs when compared with 10% povidone-iodine . This study did not include a third comparator arm with povidone-iodine combined with alcohol; therefore it is difficult to attribute the reduction in SSIs to the chlorhexidine, the alcohol or to the combination of both agents . At SVHM either chlorhexidine with alcohol or iodine with alcohol are recommended for skin antisepsis as per the hospital infection control recommendations and the decision about which specific agent is used is based on the surgeon's preference. In this current study, surgical skin preparation with chlorhexidine 0.5% in alcohol 70% was associated with a five-fold increased risk of superficial wound complications compared with iodine 1% in alcohol 70%. The association was particularly marked in the hip arthroplasty cohort, with a 13-fold increased risk. This association may reflect differing anti-infective properties of the agents. Counter to this, it may be argued that the difference observed reflects the surgeon's preference and surgical experience. The agents chosen in this study differed from the randomized control trial; however, this study raises interesting questions regarding surgical preparation choice that requires further evaluation.
Obesity has been previously implicated in prosthetic joint infections, particularly in hip arthroplasty patients [9, 18, 19]. In addition, obesity is a risk factor for prolonged postoperative wound drainage . There are a number of potential explanations for the association between increasing BMI and superficial wound complications, including the need for larger surgical incision, increased incidence of fat necrosis and prolonged or more complicated arthroplasty surgery [9, 20].
Rheumatoid arthritis leads to impaired immune function and previous studies have suggested interplay between the underlying disease process as well as use of immunosuppressant medications [21, 22]. In this study, the use of systemic corticosteroids or other immunosuppressant medications was not associated with the development of superficial wound complications, suggesting the importance of impaired immunity secondary to the underlying inflammatory process itself. The association between rheumatoid arthritis and superficial wound complications varied with the joint replaced. Rheumatoid arthritis was implicated in knee but not hip arthroplasty. Patients with rheumatoid arthritis have an increased rate of carriage of Staphylococcus aureus and this organism is isolated in a higher proportion of prosthetic joint infections in patients with rheumatoid arthritis compared with other patients [21, 23]. We postulate that the association between rheumatoid arthritis and superficial wound complications may represent interplay between increased Staphylococcus aureus carriage and the overall vulnerability of the prosthetic knee joint to infective complications [21, 24].
Blood transfusion has been previously identified as a risk factor for wound infection in surgical patients, including those having orthopaedic surgery [25, 26]. This association may reflect more prolonged or complicated surgery with increased intraoperative blood loss, or development of haematoma. Haematoma formation impairs healing by increasing wound tension and reducing tissue perfusion, as well as acting as a culture medium for pathogens [6, 27]. Overall, this finding argues for strategies to minimize intraoperative blood loss, including attention to haemostasis and drainage of the surgical field.
In this knee arthroplasty cohort, an increase in tourniquet time was associated with the development of superficial wound complications. This finding is in keeping with other published studies; increased tourniquet time has been identified as a risk factor for deep infection and for impaired wound healing and prolonged wound discharge after total knee arthroplasty . The reason for this effect of increased tourniquet time may relate to a number of factors. Firstly, the prolonged tourniquet application may lead to local inflammation and tissue hypoxia and subsequent compromised wound healing. In addition, a number of studies have implicated tourniquet use in increased total postoperative blood loss [29, 30]. Finally, the use of the tourniquet reduces the tissue concentrations of prophylactic antibiotics, such as cephazolin, which may increase the risk of surgical site infections; however, in this current study, antibiotic prophylaxis was given prior to tourniquet inflation .
The benefit of subcuticular soluble sutures or staples remains a contentious issue, with a recent meta-analysis suggesting an increased rate of superficial SSI when wounds were closed with staples compared with sutures . In this study, wound closure with staples was associated with an increase in superficial wound complications on univariate analysis in the combined and hip arthroplasty cohorts, but this difference was not observed upon multivariate analysis. Therefore the question of optimal wound closure requires further investigation.
Strengths of this study include an analysis of consecutively operated patients over an 18-month period and a relatively large sample size. A single researcher was responsible for all data collection, classification and coding, which promoted consistency in the analysis. In addition, hospital care is standardized for all arthroplasty patients at SVHM through the protocol-driven clinical care pathway.
Limitations of the study are related to the retrospective study design and the possible inaccuracy or misinterpretation of the information contained in the medical records. However, we attempted to address this by setting clear definitions of wound complication prior to data collection. Secondly, not all factors related to wound healing could be accounted for, such as nutrition, hygiene and wound care, particularly after patients were discharged home from hospital. Finally, we included data regarding prosthetic joint infection; however, given the contemporaneous setting of this study, 12- and 24-month follow-up data for some patients are incomplete, therefore some delayed and late infections may not be captured. Superficial wound complications, however, are a more typical feature of early prosthetic joint infection rather than delayed or late infection (where pain is the predominant clinical feature) .
This study has identified unique factors in the evolution of superficial wound complications. These data provide clinicians with evidence to support pre-emptive strategies to identify patients at risk of superficial wound complications, particularly obese patients or patients with rheumatoid arthritis. This study has highlighted the importance of perioperative factors, including skin preparation, tourniquet time and blood transfusion, in the development of superficial wound complications. This study provides an impetus for further clinical research, particularly assessing the impact of modification of these identified factors on the subsequent development of both superficial wound complications and subsequent prosthetic joint infections.
Dr Trisha Peel is supported by a National Health and Medical Research Council Medical and Dental Postgraduate Research Scholarship (APP1017038). Dr Michelle Dowsey is supported by a National Health and Medical Research Council Early Career Australian Clinical Fellowship (APP1035810).
PFMC has received consultancy fee and funds for research (not related to the current work) from De Puy and funds for research from Allergan and royalties from Zimmer.
All other authors: no conflict of interest to declare.