Abstract
Background:
Periprosthetic ankle joint infection is a feared complication of total ankle arthroplasty because the implant fails in the majority of cases. However, risk factors for developing these infections are unknown.
Methods:
We aimed to determine risk factors for infection in a matched case-control study that included twenty-six patients with periprosthetic ankle joint infection and two control groups, each consisting of fifty-two patients.
Results:
The prevalence of periprosthetic ankle joint infection within our cohort was 4.7%. Four infections (15%) had a hematogenous origin and twenty-two (85%), an exogenous origin. Staphylococcus aureus was the most common pathogen, followed by coagulase-negative staphylococci. Preoperative predisposing factors associated with infection included prior surgery at the site of infection (odds ratio [OR] = 4.56, 95% confidence interval [CI] = 0.98 to 21.35, and OR = 4.78, 95% CI = 1.53 to 14.91, in comparison with the two control groups) and a low American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot score (35.8 versus 49.8 and 47.6 in the two control groups, p ≤ 0.02). The mean duration of the index surgery was significantly longer in the case group than in both control groups (119 versus eighty-four and ninety-three minutes, p ≤ 0.02). After surgery, persistent wound dehiscence (OR = 15.38, 95% CI = 2.91 to 81.34, p = 0.01, in comparison with both control groups) and secondary wound drainage (OR = 7.00, 95% CI = 1.45 to 33.70, and OR = 5.31, 95% CI = 1.01 to 26.78, in comparison with the two control groups, p ≤ 0.04) were associated with the development of a periprosthetic ankle joint infection.
Conclusions:
Patients at risk for periprosthetic ankle joint infection following total ankle arthroplasty include those with a history of surgery on the ankle, a low preoperative AOFAS hindfoot score, and a long operative time. Postoperatively, patients with a prolonged wound dehiscence or a secondary wound-healing problem are also at risk for infection.
Level of Evidence:
Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
End-stage ankle osteoarthritis has traditionally been treated with arthrodesis1. During the last thirty years, however, total ankle arthroplasty has been used with increasing frequency. First and second-generation arthroplasties were associated with a high rate of mechanical failure, mainly because of poor implant design2. The improvement of both surgical technique and implant design has led to a significant decrease in failure rates2,3. Currently, implant-associated infections have become an important issue in ankle surgery4-12. In a recent systematic literature review, Glazebrook et al.3 classified the complications of ankle arthroplasty, and they considered infections to be high-grade complications because the implant fails in the majority of cases. We analyzed a cohort of patients with total ankle arthroplasty to evaluate the prevalence of these infections and to identify both patient-related and surgery-related risk factors for infection.
Patient Population
Patients who had been treated for periprosthetic joint infection were identified from a cohort of 408 patients who had undergone total ankle arthroplasty at our institution between August 1, 2006, and June 30, 2010. Data for all patients were obtained by using a computerized cohort database. All patients within this cohort are followed at regular intervals after arthroplasty with clinical and radiographic examination. The study was approved by the Ethics Committee of the University of Basel, Switzerland.
Cases and Controls
Patients who were treated for periprosthetic joint infection during the given period were the cases for this study. Only the first episode of a periprosthetic joint infection in a given patient, not the relapses, was included in the analysis. The diagnosis of periprosthetic joint infection included clinical signs of pain, effusion, erythema, and induration and at least one of the following criteria: (1) growth of the same microorganism in at least two cultures of synovial fluid and/or periprosthetic tissue, (2) visible pus surrounding the joint, (3) acute inflammation on histopathological examination (≥10 neutrophils/high-power field)13, or (4) the ability to probe the base of a wound to the implant14. After detailed review of the patient history, infection acquisition was categorized as “exogenous” or “hematogenous.” Exogenous cases were locally acquired through the wound. The case was considered hematogenous when the patient had had an uneventful postoperative course for at least three months after the index arthroplasty and/or there was a distant infection source. All cases that did not fulfill the criteria for a hematogenous infection were categorized as exogenous.
Controls were chosen from the cohort of patients who had undergone total ankle arthroplasty during the study period but had not developed a periprosthetic joint infection. The case group was compared with two control groups, each in a ratio of 1:2 patients. The first control group was obtained by matching each case with two controls on the basis of two constitutional variables (age and sex). To obtain the second group, each case was matched according to the date of the total ankle arthroplasty (time-matched controls). These controls underwent total ankle arthroplasty directly prior to and directly after the case (one each). The same method was applied to patients referred for the treatment of periprosthetic joint infection, although the controls were selected from our cohort.
Potential Risk Factors and Statistical Analysis
The medical records of the cases and controls were reviewed for potential risk factors. Patient-related variables included body mass index (BMI), diabetes mellitus, malignant tumor, corticosteroid use, chronic renal failure (estimated glomerular filtration rate of <30 mL/min), chronic heart failure, chronic skin disease, polyneuropathy, peripheral arterial occlusive disease, and smoking.
Potential risk factors for the development of periprosthetic joint infection were divided into four categories: (1) the indication for total ankle arthroplasty, (2) clinical characteristics prior to the index surgery, (3) characteristics of the index surgery itself, and (4) the postoperative course after the index surgery.
Indication-associated variables included primary osteoarthritis, rheumatoid arthritis, posttraumatic arthritis, and a preexisting ankle arthrodesis.
Factors related to the clinical presentation prior to the index surgery included previous surgery at the site of infection, the visual analog scale (VAS) score for pain (0 = “no pain” and 10 = “pain as bad as it could be”), the range of ankle joint motion, and the American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot score15,16. This score is a clinician-based 100-point score, with the highest number of points reflecting the best clinical state. The score attributes 40 points to pain, 50 points to function, and 10 points to hindfoot alignment.
The duration and type of index surgery (primary or revision total ankle arthroplasty, with or without subtalar arthrodesis) were considered to be the potential risk factors related to the index surgery.
The variables associated with the postoperative course were poor wound-healing and secondary wound drainage. Poor wound-healing was defined as persistent wound dehiscence along the surgical incision for fourteen days or more. The cutoff of fourteen days was selected empirically. Secondary wound drainage was defined as a wound that became wet after it was initially judged to be dry. Both variables regarding the postoperative course were obtained from detailed written and/or photographic documentation of the patient’s surgical wounds.
Statistical Methods
If not otherwise indicated, the odds ratio was calculated for each potential risk factor after matching of cases and controls. The Student t test, Fisher exact test, or Pearson chi-square test was applied where appropriate. A two-tailed p value of <0.05 was considered significant.
Source of Funding
This work was supported by grants from the Swiss Society of Internal Medicine and the Clinic of Orthopaedic Surgery, Kantonsspital, Liestal, Switzerland (fellowship for B.K.). The funding source did not play a role in the investigation.
Patient Population and Microbiological Findings
Over the four-year period, we identified twenty-six infections after the index total ankle arthroplasty. The median interval between the joint arthroplasty and the diagnosis of periprosthetic joint infection was 193 days (range, ten to 3762 days). Six (23%) of the twenty-six cases occurred within ninety days after implantation; sixteen (62%), between ninety days and two years; and four (15%), after two years. Seven cases (27%) were referred from other hospitals. The type of infection was exogenous in twenty-two cases (85%) and hematogenous in four (15%). The sources for the hematogenous cases were skin and soft-tissue infections (two cases) and dental and perianal abscesses (one case each). After exclusion of the seven referred cases, the prevalence of periprosthetic joint infections in our total ankle arthroplasty cohort was nineteen (4.7%) of 401.
The demographic data and comorbid conditions of the cases and the controls are presented in Table I. The characteristics were similar in the three groups. The patients with periprosthetic joint infection used corticosteroids more frequently than the time-matched control group did. However, this result was not observed in the age/sex-matched control group and therefore not conclusive.
The microbiological findings are presented in Table II. Staphylococcus aureus was the most common pathogen, followed by coagulase-negative staphylococci. The infection was polymicrobial in four cases (15%).
Risk Factors for the Development of Infection
The results of the risk factor analysis are presented in Table III.
Among the variables categorized as an “indication for total ankle arthroplasty,” only “revision of a preexisting ankle arthrodesis” was more frequent in the group with periprosthetic joint infection (four cases) than in the control groups (none in the age/sex-matched group and one in the time-matched group, p ≤ 0.03).
Characteristics before the index arthroplasty that were associated with periprosthetic joint infection included prior surgery at the site of infection (odds ratio [OR] = 4.56, 95% confidence interval [CI] = 0.98 to 21.35 in comparison with the age/sex-matched group and OR = 4.78, 95% CI = 1.53 to 14.91 in comparison with the time-matched group) and a low AOFAS hindfoot score (35.8 versus 49.8 in the age/sex-matched group and 47.6 in the time-matched group, p ≤ 0.02).
The mean duration of the index operation was significantly longer in the case group (119 minutes) than in both control groups (eighty-four minutes in the age/sex-matched group and ninety-three minutes in the time-matched group, p ≤ 0.02). A total ankle arthroplasty without subtalar arthrodesis was associated with a lower risk of infection (OR = 0.24, 95% CI = 0.08 to 0.69, and OR = 0.14, 95% CI = 0.04 to 0.51, p ≤ 0.01). In contrast, revision arthroplasties were more frequent in the case group (23%) than in the control groups (2% in the age/sex-matched group and 0% in the time-matched group, p < 0.01).
In the postsurgical phase, poor wound-healing was associated with the development of periprosthetic joint infection (OR = 15.38, 95% CI = 2.91 to 81.34, p = 0.01, in comparison with both control groups), as was secondary wound drainage (OR = 7.00, 95% CI = 1.45 to 33.70 in comparison with the age/sex-matched group and OR = 5.31, 95% CI = 1.01 to 26.78 in comparison with the time-matched group, p ≤ 0.04).
Total ankle arthroplasty is an increasingly used yet challenging procedure, and periprosthetic ankle joint infection is considered a high-grade complication3. Published series of fifty or more cases of total ankle arthroplasty have had infection rates ranging from 2% to 8.6%4,6-8,11,17-19. This wide range may be due to variability associated with the small numbers of infections in the case series. The prevalence of infection in our cohort was 4.7%. Our institution serves as a tertiary care center for patients needing arthroplasty13, and we have treated a large cohort of patients with total ankle arthroplasty20,21. Another reason for the variability of the reported prevalences of infection may be the lack of an unequivocal definition of periprosthetic joint infection following ankle arthroplasty. The authors of most studies of periprosthetic joint infection distinguished between deep and superficial wound infections5,18,19. Although there are standard definitions for surgical site infections22,23, they are not readily applied to implant-associated infections. Considering that all wound complications following hip and knee arthroplasties are significantly associated with implant-associated infections24, we believe that the distinction between superficial and deep wound infections in patients with arthroplasty should be reconsidered. This issue might be particularly relevant to patients with total ankle arthroplasty, since only sparse soft tissues surround the ankle joint. However, our results and those in the literature indicate that the infection rate following total ankle arthroplasty is higher than that observed after total hip and knee arthroplasties25.
To the best of our knowledge, no study has evaluated the risk factors for developing infection following primary total ankle arthroplasty. Several comorbid conditions, such as rheumatoid arthritis, diseases requiring long-term corticosteroid therapy, diabetes mellitus, obesity, smoking, and malignant tumors, have been described as risk factors for prosthetic hip and knee joint-associated infections24-26. In our cases, none of these conditions were significantly more prevalent than in either control group except for corticosteroid use, and the results of the analyses of that factor were inconsistent between the two comparisons and hence not conclusive. It is possible that the failure to identify significant associations was due to the small sample sizes. It should be noted that van der Heide et al.7 reported a high prevalence of periprosthetic joint infection (8.6%) after total ankle arthroplasty in patients with rheumatoid arthritis, but they did not analyze a control group for comparison. Rheumatoid arthritis has also been identified as a risk factor for periprosthetic knee joint infection24,26,27.
In our study, the only indication for total ankle arthroplasty that was associated with an increased risk for periprosthetic joint infection was revision of a preexisting ankle arthrodesis (Table III). The total number of infections in this category was small. Also, a preexisting ankle arthrodesis often increases both the complexity of the total ankle arthroplasty and the surgical time. Hence, it is possible that this variable is linked to those surgery-related risk factors.
When we considered clinical characteristics at the time of the index arthroplasty, we found that the majority of the patients with periprosthetic joint infection had had at least one prior surgical intervention at the site where the infection later developed. The significance of this risk factor was found in the comparisons with both control groups. The finding is in agreement with those of previous studies demonstrating that the number of operations is a risk factor for surgical site infections in orthopaedic patients28. In patients undergoing total ankle arthroplasty, this risk factor must be considered critically because the surrounding soft tissue is unfavorable for repeated surgery. Moreover, posttraumatic arthritis is the most common indication for a total ankle arthroplasty. Consequently, most of these patients have had previous surgery and the surrounding soft-tissue mantle is commonly impaired. It is conceivable that such soft-tissue conditions make the patient susceptible to exogenous infection, the type observed in the majority of our cases. For the same reason, exogenous infections can be missed for several months.
Our patients with periprosthetic joint infection had significantly lower preoperative AOFAS hindfoot scores than did the control groups. We cannot explain this observation because the preoperative VAS pain scores and ankle motion did not differ between the groups. However, as observed in patients with rheumatoid arthritis7, it is possible that ankle joints with poor hindfoot alignment are more predisposed to infection.
The risk factors for periprosthetic joint infection related to the index surgery were in agreement with the results from previously published studies, although in a different patient population. A prolonged operative time is a recognized risk factor28. The total ankle arthroplasties without an accompanying arthrodesis of the subtalar joint had shorter operating times, and the patients were less prone to infection. Similarly, revision arthroplasties often require a longer operative time and the patients are at higher risk for infection24,29. It makes sense that the operative time is often associated with the complexity of the intervention and that these risk factors are not independent of each other.
In the postoperative course, a wound dehiscence persisting for fourteen days or more and the presence of secondary wound drainage were associated with periprosthetic joint infection. Up to 15% of patients have delayed wound-healing after total ankle arthroplasty3,6,30. Wound dehiscence increases the risk of postoperative exogenous infection. As a consequence, any factor leading to foot swelling, and perhaps to wound dehiscence, should be prevented. There is no precise cutoff for the duration of dehiscence at which the risk of infection increases significantly. Retrospectively, our data suggest that the risk increases after eight to ten days of dehiscence, although because the cutoff (fourteen days) was set prior to data collection and our study was not designed to subcategorize this variable, we cannot draw any firm conclusions. However, a wet wound developing after initially adequate wound-healing in the postoperative course should raise the suspicion of periprosthetic joint infection.
As with every case-control study, our analysis is susceptible to sampling and differential measurement bias. Nevertheless, except for the referred cases, all patients were followed prospectively after the index total ankle arthroplasty. Hence, cases and controls were selected in the same way and before the outcome (periprosthetic joint infection) occurred. A limitation of our study design is the matching of controls for different variables, which produced two separate case-control analyses. Another limitation is the small absolute number of periprosthetic joint infections, making the interpretation of statistical results difficult. These factors together may explain the inconsistent results. Finally, limitations of the AOFAS hindfoot score should be considered15. The score is used to assess outcomes after foot and ankle surgery, but it has not yet been validated. Nevertheless, it is the most commonly used measure in ankle surgery15, reflecting its importance in clinical practice.
In conclusion, our study revealed several risk factors for periprosthetic ankle joint infection. These include a low AOFAS hindfoot score, previous ankle surgery prior to the index arthroplasty, prolonged index surgery time, prolonged wound dehiscence, and the development of a wet wound after wound-healing has occurred postoperatively.
Note: The authors thank Barbara Every, ELS, of BioMedical Editor, for editing assistance. They also thank Sven Trelle, MD, Associate Director of Clinical Trial Unit Bern at the Institute of Social and Preventive Medicine of the University of Bern, Switzerland, for valuable advice on epidemiological and statistical issues.
SooHoo
NF;
Zingmond
DS;
Ko
CY. Comparison of reoperation rates following ankle arthrodesis and total ankle arthroplasty. J Bone Joint Surg Am.
2007;89:2143-9.[PubMed][CrossRef]
Gougoulias
NE;
Khanna
A;
Maffulli
N. History and evolution in total ankle arthroplasty. Br Med Bull.
2009;89:111-51.[PubMed][CrossRef]
Glazebrook
MA;
Arsenault
K;
Dunbar
M. Evidence-based classification of complications in total ankle arthroplasty. Foot Ankle Int.
2009;30:945-9.[PubMed][CrossRef]
Henricson
A;
Knutson
K;
Lindahl
J;
Rydholm
U. The AES total ankle replacement: a mid-term analysis of 93 cases. Foot Ankle Surg.
2010;16:61-4.[PubMed][CrossRef]
Claridge
RJ;
Sagherian
BH. Intermediate term outcome of the agility total ankle arthroplasty. Foot Ankle Int.
2009;30:824-35.[PubMed][CrossRef]
Lee
KB;
Cho
SG;
Hur
CI;
Yoon
TR. Perioperative complications of HINTEGRA total ankle replacement: our initial 50 cases. Foot Ankle Int.
2008;29:979-84.
van der Heide
HJ;
Schutte
B;
Louwerens
JW;
van den Hoogen
FH;
Malefijt
MC. Total ankle prostheses in rheumatoid arthropathy: Outcome in 52 patients followed for 1-9 years. Acta Orthop.
2009;80:440-4.[PubMed][CrossRef]
Knecht
SI;
Estin
M;
Callaghan
JJ;
Zimmerman
MB;
Alliman
KJ;
Alvine
FG;
Saltzman
CL. The Agility total ankle arthroplasty. Seven to sixteen-year follow-up. J Bone Joint Surg Am.
2004;86:1161-71.[PubMed]
Spirt
AA;
Assal
M;
Hansen
ST
Jr. Complications and failure after total ankle arthroplasty. J Bone Joint Surg Am.
2004;86:1172-8.[PubMed]
Saltzman
CL;
Amendola
A;
Anderson
R;
Coetzee
JC;
Gall
RJ;
Haddad
SL;
Herbst
S;
Lian
G;
Sanders
RW;
Scioli
M;
Younger
AS. Surgeon training and complications in total ankle arthroplasty. Foot Ankle Int.
2003;24:514-8.[PubMed]
Henricson
A;
Skoog
A;
Carlsson
A. The Swedish Ankle Arthroplasty Register: an analysis of 531 arthroplasties between 1993 and 2005. Acta Orthop.
2007;78:569-74.[PubMed][CrossRef]
Besse
JL;
Colombier
JA;
Asencio
J;
Bonnin
M;
Gaudot
F;
Jarde
O;
Judet
T;
Maestro
M;
Lemrijse
T;
Leonardi
C;
Toullec
E; l'AFCP. Total ankle arthroplasty in France. Orthop Traumatol Surg Res.
2010;96:291-303.[PubMed][CrossRef]
Giulieri
SG;
Graber
P;
Ochsner
PE;
Zimmerli
W. Management of infection associated with total hip arthroplasty according to a treatment algorithm. Infection.
2004;32:222-8.[PubMed][CrossRef]
Grayson
ML;
Gibbons
GW;
Balogh
K;
Levin
E;
Karchmer
AW. Probing to bone in infected pedal ulcers. A clinical sign of underlying osteomyelitis in diabetic patients. JAMA.
1995;273:721-3.[PubMed][CrossRef]
Naal
FD;
Impellizzeri
FM;
Rippstein
PF. Which are the most frequently used outcome instruments in studies on total ankle arthroplasty?Clin Orthop Relat Res.
2010;468:815-26.[PubMed][CrossRef]
Kitaoka
HB;
Alexander
IJ;
Adelaar
RS;
Nunley
JA;
Myerson
MS;
Sanders
M. Clinical rating systems for the ankle-hindfoot, midfoot, hallux, and lesser toes. Foot Ankle Int.
1994;15:349-53.[PubMed]
Hurowitz
EJ;
Gould
JS;
Fleisig
GS;
Fowler
R. Outcome analysis of agility total ankle replacement with prior adjunctive procedures: two to six year followup. Foot Ankle Int.
2007;28:308-12.[PubMed][CrossRef]
Doets
HC;
Brand
R;
Nelissen
RG. Total ankle arthroplasty in inflammatory joint disease with use of two mobile-bearing designs. J Bone Joint Surg Am.
2006;88:1272-84.[PubMed][CrossRef]
Kitaoka
HB;
Patzer
GL. Clinical results of the Mayo total ankle arthroplasty. J Bone Joint Surg Am.
1996;78:1658-64.[PubMed]
Hintermann
B;
Barg
A;
Knupp
M;
Valderrabano
V. Conversion of painful ankle arthrodesis to total ankle arthroplasty. Surgical technique. J Bone Joint Surg Am.
2010;1:55-66.[CrossRef]
Hintermann
B;
Barg
A;
Knupp
M;
Valderrabano
V. Conversion of painful ankle arthrodesis to total ankle arthroplasty. J Bone Joint Surg Am.
2009;91:850-8.[PubMed][CrossRef]
Mangram
AJ;
Horan
TC;
Pearson
ML;
Silver
LC;
Jarvis
WR. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol.
1999;20:250-78.[PubMed][CrossRef]
Wilson
AP;
Treasure
T;
Sturridge
MF;
Grüneberg
RN. A scoring method (ASEPSIS) for postoperative wound infections for use in clinical trials of antibiotic prophylaxis. Lancet.
1986;1:311-3.[PubMed][CrossRef]
Berbari
EF;
Hanssen
AD;
Duffy
MC;
Steckelberg
JM;
Ilstrup
DM;
Harmsen
WS;
Osmon
DR. Risk factors for prosthetic joint infection: case-control study. Clin Infect Dis.
1998;27:1247-54.[PubMed][CrossRef]
Del Pozo
JL;
Patel
R. Clinical practice. Infection associated with prosthetic joints. N Engl J Med.
2009;361:787-94.[PubMed][CrossRef]
Pulido
L;
Ghanem
E;
Joshi
A;
Purtill
JJ;
Parvizi
J. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res.
2008;466:1710-5.[PubMed][CrossRef]
Jämsen
E;
Huhtala
H;
Puolakka
T;
Moilanen
T. Risk factors for infection after knee arthroplasty. A register-based analysis of 43,149 cases. J Bone Joint Surg Am.
2009;91:38-47.[CrossRef]
de Boer
AS;
Mintjes-de Groot
AJ;
Severijnen
AJ;
van den Berg
JM;
van Pelt
W. Risk assessment for surgical-site infections in orthopedic patients. Infect Control Hosp Epidemiol.
1999;20:402-7.[PubMed][CrossRef]
Peersman
G;
Laskin
R;
Davis
J;
Peterson
M. Infection in total knee replacement: a retrospective review of 6489 total knee replacements. Clin Orthop Relat Res.
2001;392:15-23.[PubMed][CrossRef]
Gougoulias
N;
Khanna
A;
Maffulli
N. How successful are current ankle replacements?: a systematic review of the literature. Clin Orthop Relat Res.
2010;468:199-208.[PubMed][CrossRef]
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.