Outcomes of Debridement, Antibiotics and Implant Retention (DAIR) for Periprosthetic Joint Infection in a High-Volume Arthroplasty Centre

Abstract

Purpose

Prosthetic joint infection (PJI) following total knee arthroplasty (TKA) and total hip arthroplasty (THA) can be a devastating diagnosis. Debridement, antibiotics and implant retention (DAIR) is a preferred treatment modality for acute PJI. A retrospective analysis of infected primary arthroplasties to evaluate the success of DAIR and factors influencing its outcomes.

Methods

We retrospectively reviewed all patients who underwent DAIR for PJI at our unit between 2010 and 2018. Patients who underwent revision surgery as an index procedure, arthroscopic washout and those with less than two years of follow-up were excluded. Treatment failure was defined as revision arthroplasty for recurrence of infection within 2 years of the index procedure. Chi-square and Fischer’s exact test were used to compare between patient factors and DAIR outcomes. Kaplan–Meier survival curve and log-rank test were used to analyse implant survivorship following DAIR.

Results

Of the sixty patients (40 knees, 20 hips) who underwent DAIR, eighteen (13 knees, 5 hips) required revision arthroplasty within 2 years accounting for a success rate of 70%. Predictive factors for revision were American Society of Anaesthesiologist (ASA) score of greater than 2 (p = 0.021), BMI > 35 (p = 0.046), C Reactive protein (CRP) > 200 mg/L (p = 0.007) and Staphylococcus aureus growth (p = 0.012). The five-year survival rate for DAIR was 70%, which remained constant after two years from DAIR.

Conclusion

Success rate of DAIR in PJI was 70% which was comparable to similar studies in the literature. ASA > 2, BMI > 35, CRP > 200 and staphylococcus aureus growth were predictors for DAIR failure. Implant survival rate and duration were better following DAIR in early-onset PJI.

Introduction

Periprosthetic joint infection (PJI) is a serious complication with an incidence of 1–2% following primary arthroplasties, with infection identified as a reason for revision in upto a third of cases [1, 2]. With increasing life expectancy and procedural success, the number of arthroplasties performed continues to rise [1]. Management of PJI is complex and involvement of an interdisciplinary team is required for good functional outcomes [3], while intervention is often associated with significant patient morbidity and healthcare costs [4].

Debridement, antibiotics and implant retention (DAIR) can be a successful method for treating selected patients with PJI [5, 6]. It is quicker, cheaper and associated with less morbidity and fewer complications compared to revision arthroplasty; however, the success rate is variable within the literature [7–9]. The role of DAIR in acute hematogenous PJI is controversial with poorer results compared to acute postoperative PJI [7]. Multiple risk factors have been found to influence outcomes following DAIR, the commonest being microorganism isolated and duration of antibiotic cover [10].

The aim of this retrospective study was to evaluate the success rate of DAIR in acute PJI following primary hip and knee arthroplasties in our institution. A secondary objective was to identify the risk factors influencing outcomes and analyse implant survivorship following DAIR.

Materials and Methods

At our high volume, tertiary referral, elective arthroplasty centre, between 2010 and 2018, we identified 106 patients who underwent debridement for acute periprosthetic hip and knee joint infections from the institutional database. Diagnosis of PJI was based on clinical findings and laboratory inflammatory markers. Patients who underwent revision surgery as an index procedure, arthroscopic washout and those with less than two years of follow-up were excluded leaving 60 patients who underwent DAIR for final analysis. Patient details including demographic data, American Society of Anaesthesiologists (ASA) score, body mass index (BMI), comorbidities, time since index surgery and microbiology reports were collected from the institutional database.

Patients who presented within three months from index surgery were classified as early-onset PJI (n = 48), those presenting between three and twenty-four months from index procedure were classified as delayed onset PJI (n = 9) and patients presenting later than 2 years were grouped under late onset PJI (n = 3) [11].

DAIR was done in a standard operating theatre employing lateral decubitus position for THR and supine position with a tourniquet for TKR. Through the previous skin incision, arthrotomy and synovectomy were performed and a minimum of six samples were sent for culture. Implants were checked for stability and modular components (polyethylene, head and liner) were exchanged. Thorough wash was given with a minimum of 3 L of saline along with povidone iodine and H2O2. After 2016, the protocol was changed to include packing of the wound with antibiotic-loaded calcium sulphate beads (Stimulan, Biocomposites, Keele, UK) and we did not use H2O2. The kit includes 10 cc (20 g) of calcium sulphate hemihydrate powder, a pre-mixing solution bulb, pellet mould and spatula (Fig. 1). The arthrotomy wound was then closed and tourniquet released. All patients received antibiotic cover based on susceptibility testing for a mean duration of six months. The time duration for antibiotics was based on clinical findings and laboratory inflammatory markers on follow-up visits. Treatment failure was defined as the requirement of revision arthroplasty for recurrence of PJI within 2 years following DAIR.

Fig. 1.

Clinical picture of DAIR for primary total knee replacement showing stimulant kit (A), bead preparation using moulds (B) and prepared antibiotic beads being added to the knee wound during DAIR (C)

Statistical analysis was performed using the statistical package for social sciences software 23 (SPSS) and R. Chi-square and Fischer’s exact test (if number of observations < 5 in a cell) were used to compare patient factors and DAIR outcomes. Kaplan–Meier survival curve and log rank test were used to analyse implant survivorship following DAIR. Significance was set to p value of less than 0.05.

Results

Mean age of patients who underwent DAIR was 68.75 years (51–89 years) with male: female ratio of 1.6 (males = 37, females = 23). Index procedures included twenty hip (THR) and forty knee (TKR) arthroplasties. Mean duration of DAIR from index arthroplasty was 4 weeks (1–12) for early-onset PJI, 44.5 weeks for late-onset PJI (17–96) and 336 weeks for delayed onset PJI (173–500). Four patients underwent a second debridement for persistent infection. Of the sixty patients who underwent the DAIR procedure for PJI, eighteen patients (5 hips and 13 knees) required revision arthroplasty within 2 years accounting for 70% success rate for DAIR.

Predictive factors for failure of DAIR included ASA score greater than 2, CRP > 200, BMI > 35 and tissue growth of staphylococcus aureus. There was no significant association between diabetes mellitus, type of PJI (early, delayed, late onset), rifampicin treatment or stimulan use and DAIR failure (Table 1).

Table 1.

Risk factors influencing DAIR outcome

Characteristics	DAIR success (n = 42)	DAIR failure (n = 18)	p value
ASA grade > 2	2	5	0.021*
BMI > 35	6	12	0.046*
PJI type
Early onset (n = 48)	35	13
Delayed onset (n = 9)	6	3	0.339
Late onset (n = 3)	1	2
Modular component exchange	23	9	0.725
Rifampicin for Staph aureus	13	9	0.242
Staph aureus growth	15	13	0.012*
CRP > 200 mg/L	6	9	0.007*
Stimulan use	3	3	0.35

Bold values are statistically significant (p < 0.05)

*Fischer’s exact test

The polyethylene component was exchanged in 76.6% (46/60) of patients. Among the twenty hips, eleven had head and acetabular liner exchanged of which 8 were successful and 3 required revision. Data for modular component exchange was not available for 9 hips. Of the forty knees, data for modular component exchange was not available for 5 patients. The remaining thirty-five knees had polyethylene exchange, of which twenty-two had a successful outcome. Overall, there was no significant association between modular component exchange and DAIR outcome.

Staphylococcus aureus was the most common organism and was isolated in 46.6% of cases (Fig. 2). It was a predictive factor for DAIR failure with a failure rate of 46.4% compared to 15.6% in patients with no Staphylococcus aureus growth. All patients with Staphylococcus aureus growth received either flucloxacillin monotherapy or combination therapy. Rifampicin was added in 90.9% cases (20/22) with Staphylococcus aureus growth on culture. No cultures were methicillin resistant. Twelve patients had no organisms grown on culture and had a successful outcome following DAIR (100%).

Fig. 2.

Organisms isolated in patients with PJI undergoing DAIR procedure

Mean duration to follow-up in the DAIR success group was 5.52 years. Survival following DAIR was 70% at 2 years, remaining constant thereafter (Fig. 3). Within the “late” and “delayed” groups, there was a trend for lower survival than the “early” group, however, with low numbers in each group this failed to reach statistical significance (Fig. 4).

Fig. 3.

Kaplan–Meier curve showing overall survivorship of patients who underwent DAIR with 95% CI

Fig. 4.

Kaplan–Meier curve showing survival curves for early-onset vs late-onset vs delayed onset PJI treated by DAIR with 95% CI

Discussion

PJI remains a challenging complication following total joint replacement. Management options include suppressive antibiotic treatment, DAIR, one stage and two-stage exchange revision arthroplasties. DAIR remains the gold standard treatment for PJI presenting within 4 weeks from index surgery or hematogenous infection with symptoms less than 3 weeks [5]. It has the advantage of less patient morbidity and reduced healthcare costs compared to revision surgeries [12].

The major weakness of this study is the low number of patients and their heterogeneous nature. Despite analysing all patients undergoing DAIR at our high volume institution over an 8 year period, only 60 patients were found to be eligible. This small number of patients leads to difficulty in identifying statistically significant demographics factors, and survival trends (Fig. 3). Further, in this retrospective study conducted over an extended period of time, the surgical practice has changed with increasing emphasis on the exchange of components, and the use of antibiotic-loaded beads. However, our patient sample compares favourably with other studies and is a representative account of surgical practice.

In this retrospective single centre analysis, we obtained a success rate of 70% following DAIR in hip and knee PJI which is consistent with similar studies [13–16]. The success rate of DAIR in the literature varies from as low as 28–100% [9, 13–20]. The wide variability in results of DAIR may be due to different antibiotic regimes, organisms, debridement protocols and DAIR failure criteria. We defined treatment failure as those requiring revision within two years following index surgery which had been followed by studies with a larger cohort of patients [18, 19].

The 2018 musculoskeletal infection society (MSIS) consensus states that DAIR should be performed for acute PJI (< 4 weeks) [5, 8]. Ottesen et al. in their study of 67 patients reported lower success rates when DAIR was performed after 90 days [17]. A recent study by Wouthuyzen-Bakker M et al. involving 132 matched pairs of early and late acute PJI treated by DAIR reported lower success rates with late-onset PJI (46%) [20]. We also observed a lower success rate in DAIR performed for delayed onset PJI (33.3%).

Multiple risk factors such as patient comorbidities, BMI and ASA grade seem to influence the outcomes following DAIR. Azzam et al. showed that a higher ASA grade and intra-articular purulence was associated with increasing chances of failure [21]. Many authors have reported kidney disease, liver failure, smoking and steroid therapy as risk factors [22–24]. We also found a significant association between ASA grade of more than 2, BMI > 35 and DAIR failure. This is in contrast to the study by Ottesen et al. which reported no significant relationship between DAIR and ASA [17].

The role of inflammatory markers in PJI is controversial and serial interpretation is preferred for accurate analysis. We observed that a CRP of more than 200 mg/L at presentation was associated with DAIR failure. 60% of patients with CRP > 200 mg/L at initial presentation had failed after the DAIR procedure. A multicentric study on DAIR for PJI by Tamayo et al. also found CRP as an independent predictor for failure [25].

The bacterial load of polyethylene is usually higher due to its surface texture and this contributes to the persistence of biofilm even after good debridement. Tsang et al. in their systematic review had reported success rate of 73.9% with modular exchange compared to 60.1% with retainment of modular component [26]. We did not observe any significant association between modular component exchange and DAIR outcome (p > 0.05).

With respect to organisms, staphylococcus aureus has been proved to be an independent risk factor for DAIR failure [13, 21, 24]. Some studies have found no relation between organism and outcome following DAIR [17, 27]. In this study, the most common organism isolated was Staphylococcus aureus with a DAIR failure rate of 46.4% and a positive culture being predictive of DAIR failure (p < 0.05). There was a high incidence of polymicrobial growth (33.3%) in DAIR failure, which was similar to those reported in literature (5–39%) [13, 21, 25].

The duration of antibiotic cover after DAIR is a controversial topic with experts recommending a minimum of 6 weeks antibiotic cover [8]. Many studies regarding DAIR outcomes have used the antibiotic cover as proposed by Zimmerli et al. which consists of 2 weeks of intravenous antibiotics followed by 3–6 months of oral antibiotics [28]. In this study, the duration of antibiotics was based on susceptibility testing, clinical and inflammatory marker testing on follow-up visits. Patients received antibiotics for a mean duration of 6 months following the DAIR procedure. The role of rifampicin in treating staphylococcal PJI is clear with multiple studies showing a positive outcome [27]. However, we did not identify any association between rifampicin administration and DAIR outcome (p > 0.05).

Grammatopoulos et al. reviewed DAIR performed for 122 hips and reported 85% five-year implant survival rates [29]. We observed 70% five-year survival rates following DAIR with rates remaining constant after two years. Moreover, patients who underwent DAIR for early-onset PJI had better survival rates and duration compared to late and delayed onset PJI.

The role of antibiotic calcium sulphate beads in DAIR is controversial. We began using them after 2016. A total of six patients (5 knees and 1 hip) had stimulan beads applied at the time of DAIR with a success rate of 50% (p > 0.05) (Fig. 5). Abosala et al. in their systematic review concluded that adjunct antibiotic-loaded beads had a favorable outcome in revision procedures compared to DAIR. They also found increased complications with high-volume beads. The mean time to resolution of beads is reported to be from 3 to 12 weeks, however, we could not define the exact period as all our post DAIR radiographs were done at 3 months [30].

Fig. 5.

Plain radiographs of right total knee replacement after DAIR (A, B) and 3 months post DAIR (C, D) showing complete resolution of beads

This retrospective analysis adds to the existing literature on DAIR outcomes in PJI. DAIR performed for early onset PJI had better success compared to late and delayed onset PJI. However, prospective, multicentre series are required to identify large cohorts from which robust regression analysis can be performed to allow calculation of risk of failure, duration of antibiotics and optimum time frame for DAIR.

Conclusion

This study highlights that there is a role for DAIR in the treatment of acute hip and knee early-onset PJI. ASA score > 2, BMI > 35, CRP > 200 and staphylococcus aureus growth were predictors for DAIR failure. Implant survival rate and duration were better following DAIR in early-onset PJI compared to late and delayed onset PJI with survival rates remaining constant after two years.

Funding

No funds, grants, or other support was received.

Declarations

Conflict of Interest

The authors have no relevant financial or non-financial interests to disclose.

Ethical Approval

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008 (5). Informed consent was obtained from all patients for being included in the study.