Clinical outcomes and survivorship of two-stage total hip or knee arthroplasty in septic arthritis: a retrospective analysis with a minimum five-year follow-up

Septic arthritis (SA) is a rare but potentially devastating condition, causing significant pain and disability to affected patients [1, 2]. Its management is challenging, and a multidisciplinary approach is always required. Recently, the incidence of SA has increased because of the aging of the population and the numerous intra-articular procedures performed. Generally, the extent of infection in the setting of SA is limited compared with periprosthetic joint infection (PJI) because of the absence of previously implanted devices. However, since early onset SA can be treated with open arthroscopic debridement, whereas chronic infections need more radical resection and debridement and are associated with wider joint degeneration, the timing of the diagnosis is fundamental [3]. Over the years, several surgical strategies have been proposed to treat evolutive SA. In 1943, Girdlestone described resection arthroplasty, which was demonstrated to efficiently relieve pain and morbidity related to infection; however, leg length discrepancies and poor functional outcomes were common complications of the procedure [4, 5]. More recently, total joint arthroplasty (TJA) has been described as a reliable solution to treat SA with wide joint degeneration to control symptoms and guarantee acceptable restoration of function [6‐9]. One-stage TJA has demonstrated good functional outcomes, but the results are still unsatisfactory in terms of eradication in cases of active infection [7, 10, 11].

The use of cement spacers in staged procedures is well established, showing excellent outcomes in joint function and infection eradication [12‐15].

The aim of this study was to assess the clinical outcomes of patients treated with two-stage TJA in the setting of SA of the hip and knee at the medium-term follow-up. Endpoints of this analysis were clinical and functional outcomes, expressed as clinical scores of validated questionnaires, rate of eradication of the infection, and survivorship of the implants in this cohort of patients.

The local institutional review board approved this single-center study. Informed consent was obtained from all individual participants included in the study.

Forty-seven consecutive patients (25 male, 22 female) met the inclusion criteria and were enrolled in the study. Of these, 25 (53.2%) suffered from SA of the hip and 22 (46.8%) had SA of the knee. The overall mean age at stage one was 55.9 ± 14.3 years. The overall mean BMI was 26.4 ± 4.4 kg/m². All the patients had been reimplanted at a mean of 14.7 (range, 10.7 to 22.9) weeks after the first stage. The mean follow-up was 85.2 ± 15.4 months. The most common comorbidities were chronic cardiovascular diseases (29.8%) and diabetes mellitus (25.5%). A history of intravenous drug abuse (21.3%), HIV (17.7%), and HCV (17.0%) were other frequently encountered conditions in this series. Seventeen (36.2%) patients had previous surgery to the affected joint and 6 (12.8%) had previous infiltrative therapies; in 24 (51.1%) cases, the infection was classified as primary SA. Details on demographics, ASA scores, comorbidities, and general characteristics of the study population are presented in Table 1.

The appropriate surgical approach to evolutive SA associated with extensive joint degeneration continues to be debated. One-stage replacement was proposed as a viable technique to manage patients with SA of the hip and knee, but the risk of increased PJI is still concerning [7, 11]. Moreover, one-stage procedures are not indicated in cases of active infection. In these cases, two-stage replacement may be considered the treatment of choice to recover function and avoid infection recurrences [14, 23]. Papanna et al. [10] applied a differentiated protocol based on active versus quiescent SA, which was treated using two-stage versus one-stage replacement, respectively. In this cohort of 18 patients, no recurrence of infection or implant failures were registered at a mean 70-month follow-up. Anagnostakos et al. [24] treated 16 patients with two-stage replacement, achieving 88% control of infection. In general, the available literature demonstrated that two-stage procedures after SA provided good to excellent clinical results, with success rates ranging from 84 to 100% [6, 12, 15, 25, 26]. The results of this study demonstrated overall primary infection control in 92% of hips and 90.9% of knees and an overall implant survival of 93.6% after the second stage. Notably, patients developing SA are often affected by specific comorbidities (diabetes mellitus, intravenous drug abuse) that are well-known risk factors for PJI and generalized sepsis. From this point of view, a staged approach is safer and more advisable for chronic SA management.

Scores obtained at final follow-up from HHS, KSS, and KSS-F were classified as good to excellent. Two poor results were registered in patients with prior post-surgical sequelae, which affected the final scores, as mentioned in the “Results” section. These results are comparable with those available in the literature.

Different types of spacers have been used in the setting of two-stage surgery; some authors have used prefabricated spacers, while others have used handmade cement spacers or cement beads [10, 26‐28]. However, cement spacers seem to guarantee better functional outcomes during the period between stages when compared with cement beads [9, 14, 29, 30]. In 2019, Li et al. [31] conducted a comparative analysis on patients who received two-stage exchange arthroplasty following SA; patients in group I underwent the Girdlestone procedure and subsequent TJR, whereas patients in group II were implanted with a cement spacer at the first stage. The presence of the spacer effectively maintained leg length and simplified the second surgery, decreasing blood loss and operative time. Based on the specific bone defect encountered in the first stage, we used a preformed antibiotic-loaded cement spacer loaded with vancomycin and gentamicin. A handmade acetabular spacer was always coupled to the preformed spacer in the SA of the hips [13]. In knees, reinforced stems were incorporated into the preformed spacer when needed to guarantee more stability. In our cohort of patients, such parameters as LLD and offset did not demonstrate any significant difference from pre- to post-operative values, corroborating the assumption that spacers can efficiently preserve the length and appropriate muscular tension of the treated limb. However, two cases (4.3%) of hip spacer dislocation were registered.

Although two-stage arthroplasty is a reliable procedure, surgeons should consider related complications, such as mortality, which is higher compared to that associated with other treatments (e.g., one-stage surgery), and mechanical or septic complications that the patient could undergo during the interstage period [32‐34]. In a recent study, Xu et al. [30] found that older age, infection by a resistant pathogen, and high pre-operative levels of CRP were independent predictors of treatment failure after two-stage exchange arthroplasty for SA. At the same time, they found that serum ESR, CRP, and interleukin 6 (IL-6) had no benefits in predicting infection persistency before second-stage prosthesis implantation. In our series, in four cases (8.6%), patients had complications related to the spacer, which were successfully managed with spacer exchange in two (4.3%) cases and conversion to TJA in the other two (4.3%) cases, with optimal final outcomes.

Chronic SA often massively impairs bone stock, especially on the acetabular side (Fig. 3). The subsequent necessary debridement leads to massive bone loss in hips and in knees. The surgeon should be aware of this feature of SA and should always be prepared to manage complex bone defects (Fig. 4). In such surgical situations, an extremely accurate spacer implantation tailored to a patient’s bone defects should prevent further bone loss and interstage complications.

It is important to highlight that several limitations affect this study. The retrospective nature of the analysis contains inherent limitations that must be considered when evaluating results. Although we applied our institutional two-stage surgery protocol to all the patients included, the type of spacer and antibiotic therapy were individualized, and this could have resulted in a bias in the analysis. The absence of control groups made any considerations on different treatment options impossible, and the small sample size limits the statistical power of this analysis. Different types of SA (post-infiltrative, post-surgery, primary) were all analyzed together, representing a diagnostic bias of our analysis. Finally, no further investigations (e.g., PET) have been performed during follow-up to exclude late-onset chronic infections and confirm a disease-free condition. However, the strengths of this study were the prospective collection of data and relatively long follow-up (>5 years), as well as that all the patients underwent a standardized protocol of treatment and follow-up and diagnoses and surgery were performed in a standard manner by the same surgeon.

Two-stage arthroplasty provides good to excellent clinical outcomes in cases of active SA of the hip and the knee, as well as high percentages of infection control and implant survival. During the interstage period, cement spacers maintained adequate biomechanical parameters of the affected limbs. More high-quality case-control studies are needed to clarify which is the best treatment in the setting of SA.