Conservative two-stage revision with primary components of infected total hip arthroplasty: An analysis of survival, clinical and radiographic outcomes

Few studies provide an analysis of conservative two-stage revision of hip periprosthetic joint infection (PJI) and its impact on final outcome. A conservative revision is defined when soft tissues and bone quality enable the use of primary prosthetic components. Data of patients treated for chronic hip PJI who underwent two-stage revision between 2009 and 2016 and had a minimum of 2 years of follow-up were collected. Oxford Hip Score (OHS), Harris Hip Score (HHS) and radiological and microbiological data were retrieved and analysed. Clinical and functional outcome, survival, mortality, eradication, reinfection and re-revision rates within subgroups of patients with primary components and revision components are reported herein. A total of 148 patients underwent two-stage hip exchange with a mean follow-up of 55.6 ± 23.1 months and a mean age at surgery of 64.3 ± 12.7 years. Forty-four percent of patients underwent conservative revision. The mean HHS significantly improved from 40.6 ± 9.4 points to the final value of 87.8 ± 10.5 points (p = .002), and the mean OHS went from 20.3 ± 3.8 points to 40.3 ± 5. points (p< .001). Patients who were treated with primary components or isolated revision stems in the second stage had a significant reduction in surgical times (p< .001). The mortality rate for all causes of death was 6.8%, the eradication rate was 89.9%, the reinfection rate was 4.7% and the reoperation rate was 7.4% without differences between conservative and non-conservative revisions. Two-stage exchange arthroplasty for total hip arthroplasty (THA) PJI is a good strategy that provides satisfactory results, high eradication rates and no further need for revision. The conservative two-stage revision in patients with adequate bone stock represents a feasible option with good results and survival rates.

Few studies provide an analysis of conservative two-stage revision of hip periprosthetic joint infection (PJI) and its impact on final outcome. A conservative revision is defined when soft tissues and bone quality enable the use of primary prosthetic components. Data of patients treated for chronic hip PJI who underwent two-stage revision between 2009 and 2016 and had a minimum of 2 years of follow-up were collected. Oxford Hip Score (OHS), Harris Hip Score (HHS) and radiological and microbiological data were retrieved and analysed. Clinical and functional outcome, survival, mortality, eradication, reinfection and re-revision rates within subgroups of patients with primary components and revision components are reported herein. A total of 148 patients underwent two-stage hip exchange with a mean follow-up of 55.6 ± 23.1 months and a mean age at surgery of 64.3 ± 12.7 years. Forty-four percent of patients underwent conservative revision. The mean HHS significantly improved from 40.6 ± 9.4 points to the final value of 87.8 ± 10.5 points (p = .002), and the mean OHS went from 20.3 ± 3.8 points to 40.3 ± 5. points (p< .001). Patients who were treated with primary components or isolated revision stems in the second stage had a significant reduction in surgical times (p< .001). The mortality rate for all causes of death was 6.8%, the eradication rate was 89.9%, the reinfection rate was 4.7% and the reoperation rate was 7.4% without differences between conservative and non-conservative revisions. Two-stage exchange arthroplasty for total hip arthroplasty (THA) PJI is a good strategy that provides satisfactory results, high eradication rates and no further need for revision. The conservative two-stage revision in patients with adequate bone stock represents a feasible option with good results and survival rates. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Introduction
Periprosthetic joint infection (PJI) is one of the most common and devastating complications after total hip arthroplasty (THA), occurring in approximately in 1% of cases [1,2]. The two-stage exchange have high success rate ranging from 85 to 100% [3][4][5][6]. Nevertheless, the explantation during the first-stage represents a technical challenge especially with regard to femoral and acetabular bone stock. A conservative revision is defined when soft tissues and bone quality enable the use of primary prosthetic components. The main advantages of standard implants in revision surgery are the reduced invasiveness, the shorter operative times and preserved bone stock. Nonetheless, in case of severe bone deficit, a second-stage reconstruction could be extremely difficult to manage with standard implants and revision components are required [7][8][9].
Several studies have been conducted on the management of PJI [10,11] but few provide any data on conservative two-stage revisions and detailed aspects of the type of components that were used for the second stage revision of hip PJI.
The aim of the present study is to provide data on the clinical, radiological and microbiological outcomes of patients treated for PJI after total hip replacement with conservative twostage exchange with primary components. Moreover, we will analyse the complication rates, reoperations, re-revisions, survivorship and features of the components that were used during the second stage and their relative impact on the clinical outcome of patients with a minimum two-year follow-up.

Materials and methods
Data for this retrospective, single-centre study were collected using the prospective institutional arthroplasty registry.
This study was approved by the Institutional Review Board, and all included patients gave written informed consent to analysis of their medical records and data collection in the setting of this study.
Three to six intra-operative biopsies for microbiological analysis were obtained by default. Lateral femoral window osteotomy [16] was performed in patients with well-fixed stems and/or cemented implants. Sonication of the infected implant was routinely performed. An intra-articular surgical drain was used until the second postoperative day.
Specific intravenous antibiotics were routinely administered for at least 4 weeks, followed by oral administration lasting at least 2 weeks. When intraoperative cultures were negative, intravenous administration of glycopeptides and fluoroquinolones was protracted for 3 weeks and was followed by oral fluoroquinolones for 4 more weeks. All patients underwent a washout period without any antibiotic therapy of at least 2 weeks before undergoing preoperative second stage screening for infection (Serum CRP and ESR, Synovial White Blood Cell count and percentage, Synovial Leukocyte esterase, synovial fluid culture and Alpha defensing in doubtful cases).
At the time of re-implantation, new surgical debridement and spacer sonication were carried out. Furthermore, three to six culture samples and 1 specimen for definitive histological examination and frozen section were collected.
All the reimplantation procedures were performed through the posterolateral approach. A surgical drain was left in place until the second postoperative day.
The prosthetic components that were used during the second stage are reported in detail and analysed in the setting of the present study.
Conservative revision was defined as a two-stage procedure using primary acetabular cups and stems. Indications for conservative revision were acetabular Paprosky type I or IIA-IIB and femoral Paprosky type I or II-IIIA bone defects. The Trabecular Metal Monoblock Acetabular Cup System (Zimmer Inc, Warsaw, IN) with highly cross-linked polyethylene liner (Longevity; Zimmer) and the CLS Spotorno Stem (Zimmer, GmbH, Winterthur, Switzerland) were used in all patients underwent conservative two stage revision.
Indication for non-conservative revision were severe bone defects (acetabular Paprosky > IIB and femoral Paprosky > IIIA). Acetabular revision components included jumbo cups, augments, modular dual mobility cups and custom-made cups, while long monoblock cementless stems (Wagner; Zimmer Inc, Warsaw, IN) were used as revision femoral components. No modular or cemented stems were used. Fig 1 shows an example of a two-stage revision that was carried out with primary components.
Patients began ambulation either with partial (50%) or toe-touch weight-bearing and a walker or crutches starting the second postoperative day regardless of the type of prosthetic components that were used.
Supervised physical therapy and continuous passive motion were started the day after surgery and went on for 6 to 8 weeks.

Follow-up
All patients were assessed clinically and radiographically for a mean of 55.6 ± 23.1 months and none of the patients were lost to follow-up.
Patients without recent follow-up were contacted for the present study.
Patients underwent complications were re-scheduled to be examined with shorter follow up depending on specific condition.
The standard anteroposterior pelvis, frontal hip, and true lateral radiographs were taken one day post-operatively and at 3, 6, and 12 months post-surgery, and annually thereafter.

Outcomes
The clinical outcome was evaluated by Harris Hip Score (HHS), Oxford Hip Score (OHS) and range of motion at 3, 6, and 12 months after the procedure, and annually thereafter.
The radiographic outcome was assessed evaluating the presence of radiological alterations (loosening, osteolysis, migration, subsidence, cortical hypertrophy or malposition and leg length discrepancy) reviewed by 2 trained orthopaedic fellows (AC, SL). Unclear cases were solved by consensus. Stem alignment was assessed by the axial alignment of the proximal femur on the AP and lateral radiographs.
Malposition was defined when the main axis of the stem differed >5 mm from the anatomic femoral axis. The lateral inclination of the acetabular component was evaluated by defining neutral as an inclination between 30 and 45˚ [17]. Osteolysis was defined as new, expansive radiotransparent lesions that were not present in the immediate postoperative period. A stem was considered to be "subsided" when the gap between the apex of the great trochanter and the most lateral side of the stem shoulder increased by > 2mm between the immediate postoperative period and the last follow-up evaluation. The pre-and post-operative lateral hip offsets were assessed by digital measurement [18].
Minor complications (wound dehiscence, superficial wound problems) as well as major ones (deep infection, aseptic loosening, intra-operative or post-operative fractures, revision, reoperation) were investigated and fully reported. Septic recurrence was defined as a secondary infection following the two-stage exchange that was caused by the same pathogens with the same antibiogram. New infection was defined as a secondary infection after the two-stage exchange that was caused by other pathogens or different antibiogram.
Revision was defined as any kind of surgery after the second stage requiring the removal of fixed implant components. Reoperation was defined as any kind of surgical procedure that involved the specific hip joint following the reimplantation, with or without the removal of implant components. Implant survival was defined as the time from two stage to any revision or latest follow-up.
The primary outcome of interest was the difference between conservative and non-conservative revision in mortality, clinical and functional outcome and radiological results in patients underwent staged septic rTHA.
The secondary outcome measures were defined as revision rate for septic recurrence, eradication rate and differences between patients with primary components and patients with revision components.
According to the Diaz-Ledezma criteria [19], successful reimplantation was defined as control of the infection (healed wound without fistula, drainage, or pain), no secondary surgical procedures due to infection after reimplantation surgery (second stage), and no occurrence of PJI-related mortality (i.e., sepsis or necrotizing fasciitis).

Statistical analysis
Categorical variables were expressed as the absolute number of cases and/or percentage.
The Shapiro-Wilk Test was used to identify normally distributed parameters. Differences between means were calculated with the T-test for continuous variables or with the Mann-Whitney U test if not normally distributed. The non-parametric Wilcoxon Signed Rank test was used to compare continuous matched pre-operative and final data. Categorical variables were calculated using the Chi-square test or Fisher's exact test.
The analysis of variance (ANOVA) was used to compare means of continuous normally distributed variables in two or more independent comparison groups. The Kruskal-Wallis test was used for not normally distributed variables.
The non-parametric Spearman's rho coefficient was used to assess correlation between continuous or ordinal values.
Variables achieving the p value < 0.1 in univariate analysis were examined using multivariate logistic regression analysis and backward selection process. The significance threshold for tests was set at p < .05.
The inter-observer reliability for radiological analysis was evaluated by Cohen's Kappa coefficient. Kaplan-Meier survival function curves were created using all parameters to analyse survivorship free of revision for any reason for all the implants. A p-value of <0.05 was considered statistically significant.

Results
A total of 253 patients were screened for study eligibility related to hip chronic PJI between 2009 and 2016. Twenty-three patients did not undergo the two-stage procedure due to comorbidities, complications or death, 34 had undergone the first stage procedure at another institution, 40 had had a previously failed two-stage exchange and 8 were lost to follow-up. A total of 148 patients completed the two-stage hip exchange and their data was retrieved for study evaluation. The mean inter-stage period was 12.7 ± 4.8 weeks. The mean follow-up was 55. 6  Eighty patients (54.1%) were female and 68 (45.9%) were male. Table 1 shows the results of isolated bacteria. With regard to co-morbidities, 21 patients were diabetic (14.2%), 13 (8.8%) were affected by heart failure and 4 (2.7%) had rheumatoid arthritis. Ninety-seven (65.6%) were non-smokers, 24 (16.2%) were smokers and 27 (18.2%) were former smokers [20].
Sixty-six (44.6%) patients underwent a conservative revision and 82 (55.4%) a non-conservative revision. Details are listed within the Table 2. Table 2 shows the overall features of patients who underwent conservative and non-conservative two-stage rTHA.
Five (3.4%) reoperations for spacer exchanges were performed during the inter-stage phase due to 3 recurrent dislocations, 1 displaced femur fracture and 1 infection persistence. No significant differences between pre-operative HHS (p = .229), pre-operative OHS (p = .265) and final HHS (p = .097) scores within conservative and non-conservative groups were observed. However, patients that underwent conservative revisions had had significantly fewer previous surgeries (p< .003) and lower Paprosky bone loss grade at second stage (Table 2). Patients with isolated acetabular revision components had lower final OHS scores than patients with primary components (p = .009), revision stems (p = .016) and acetabular and femoral revision components (p = .036). (Fig 2).

Clinical and functional outcome
The type of prosthetic components that were used during the second stage procedures significantly influenced the operative time (p< .001). Both subgroups of patients who underwent conservative second stage revision or second stage surgery with revision stems and primary cups had significantly shorter surgical times (p< .05) compared to other subgroups. Details concerning the duration of second stage surgeries are shown in Fig 3. The type of prosthetic components (revision or primary) that were used for second stage surgeries did not influence mortality, eradication, reinfection or re-operation rates (p> .05).

Radiographic evaluation
At the patients' most recent follow-up evaluations, X-rays demonstrated a 97.3% rate of good anteroposterior and lateral positioning, while 4 outliers were observed, two of which were in varus stem positioning (both primary stems) and two in horizontal cup positioning (dual mobility cups). No differences were found between conservative and non-conservative rTHA. One patient of conservative group had femoral diaphyseal cortical hypertrophy and occasional tight pain with mild functional limitation.
Eleven patients showed radiolucent lines of less than 1 mm after the implantation with no progression over time. Six of them were on the femoral side (5 in Gruen's zone 6 on the anteroposterior view and in zone 13 on the lateral view, and 1 in zones 2-6 on the anteroposterior view and zones 9-13 on the lateral view) and seven of them were on the acetabular side (6 in DeLee-Charnley zones 1-2, and 1 in zone 1). Heterotopic ossifications were observed in 17 patients (12 Brooker grade II and 5 Brooker grade III).
No statistically significant associations were found when we compared radiolucency, ossification, leg length discrepancy and malalignment between primary and revision components.
The overall model test was found to be significant by multivariate analysis (p< .001). Age at surgery (p = .001) and non-eradication of infection (p = .020) were both significantly independent factors correlated to mortality.

Eradication of infection
The eradication rate in the present series was 89.9% with 15 patients receiving suppressive therapy, with no significant differences between conservative and non-conservative rTHAs (p> .05) ( Table 3). Thirteen patients had clinically controlled infection after three months of antibiotic therapy and 2 patients required a second two-stage procedure due to septic recurrence. When microbiological findings were analysed by univariate analysis, we found that Gram negative (odds ratio: 11.7; 95% CI: 2.6 to 53.4; p< .001) and polymicrobial infections (odds ratio: 5.7; 95% CI: 1.8 to 18.3; p = .002) had a significant influence on the eradication rate.
Multivariate analysis confirmed that Gram negative (p< .001) and polymicrobial infections (p< .001) were both significantly associated with non-eradication of the infection.
Of the 7 patients with reinfection, 3 died during follow-up (2 of sepsis and 1 of a metastatic pancreatic tumour).
Five patients underwent a re-revision for reinfection (3 underwent a second two-stage procedure and 2 had Girdlestone arthroplasty for relevant comorbidities thus precluding a second two-stage procedure).
On the contrary, the type of prosthetic components (p = .321), the presence of comorbidities (p = .054), and the operative time of the second stage procedure (p = .133) did not significantly influence the reinfection rate.
After multivariate analysis was applied, the regression model showed that infections sustained by Gram negative (p< .001) and polymicrobial flora (p = .003) were both significantly correlated with reinfection.
Age at surgery (p = .090) and type of implant used in the second stage (p = .104) were not significant risk factors correlated to reoperation.
Multivariate analysis highlighted that polymicrobial infections (p< .001) and Gram negative infections (p< .001) were significantly associated with reoperation, while regression analysis did not confirm that dislocation and loosening were significantly associated.

Discussion
The main finding of the present study is that conservative two-stage revision of infected THR shows good clinical and radiological outcome and high rates of eradication and implant survival at midterm follow-up.
Due to the progressively increasing trend of revisions in ever-younger patients [21], nowadays greater attention is being paid to component selection and "de-escalation" surgery [22]. Casella et al. [23] presented their results regarding twenty-one consecutive revision procedures involving conservative hip arthroplasty using cementless primary components. Nevertheless, the series included a limited number of septic cases and it focused on the revision of conservative primary arthroplasties.
To our knowledge, no previous studies have reported the clinical and radiological outcomes of large series of conservative two-stage rTHAs.
In their systematic review of the literature, Cavagnaro et al. [24,25] highlighted that femoral cementless revision is a feasible option in Paprosky type I and II defects, and reported a survival rate of 95.6% after 4.7 years of follow-up. However, few data on conservative primary acetabular component selection are available in the current literature.

PLOS ONE
The present study showed that 44.6% of patients who completed the two-stage exchange for hip PJI were treated with conservative revision and primary components. The advantage of a primary implant is the decreased operative time without any significant increase in the failure rates. In spite of these findings, the present study highlighted more dislocations in patients treated with conservative revisions (4.5% vs 0%) without any significant increase in reoperations and revisions. The clinical outcomes of patients with conservative and non-conservative revision were also somewhat comparable, with the exception of OHS results. Patients with isolated acetabular revision components had lower OHS scores at final follow-up. It is not surprising that the poor subjective clinical outcome might be due to the severity of the bone loss at the acetabular side, to the multiple previous surgeries and to the altered hip biomechanics of these patients [26,27]. Data confirmed that patients with revision components had significantly more previous surgeries than patients with primary components (p = .003).
Several outcome measures and risk factors have been analysed in the present study. Gram negative and polymicrobial infections have been identified as factors of worse outcome. These pathogens are especially challenging in septic revisions due to the small number of effective antibiotics, the high frequency of Multidrug Resistance (MDR), Gram-negative bacilli especially in southern European regions and low drug bioavailability within the bony tissue [28,29].
The presence of comorbidities has been identified as a significant risk factor for worse outcome and increased mortality in two-stage revisions by several authors [28,29]. However, univariate and multivariate analysis of the present study did not confirm this aspect, although the data nearly reached significance. This can probably be explained by the high prevalence of comorbidities in our population (64.2%).
The mortality rate in the present study was 6.8% which is comparable to rates that have been reported in other studies ranging from 2.9 to 19% [30,31], thus confirming that the twostage procedure for infected THR has a remarkable mortality rate. Moreover, we have to consider that not all patients with infected THRs concluded the two-stage procedure due to related comorbidities.
This study has several limitations: first, it is a retrospective study with relatively short follow-up. Second, the exclusion of patients who did not complete the two-stage procedure represents a selection bias that influences the final results. Third, the lack of anaesthesiologic data precluded the stratification of patients, and the scoring or indexing of related comorbidities could have skewed the results and the relative impact of comorbidities on the final outcome.

Conclusion
Conservative two-stage exchange arthroplasty for THA PJI represents a good strategy with satisfactory results and high eradication rate and survival. The use of primary components in two-stage revision in patients with adequate bone stock represents a feasible option providing good results and survival rates.