Clinical evaluation of a new technique for custom-made spacers in septic two-stage revision of total hip arthroplasties

86 patients with a septic two-stage revision operation with implantation of the custom-made spacer as previously described were included in the study [26]. The periprosthetic joint infection was diagnosed preoperatively in all cases according to the criteria of the Musculoskeletal Infection Society (MSIS) and the International Consensus on Musculoskeletal Infection (ICM) 2018 [27, 28].

45 cases were explanted endofemorally and 41 cases were explanted transfemorally. The transfemoral approach was performed with a modified Wagner technique as previously published [13, 15]. In brief, a posterolateral incision to the femur was applied and the posterolateral region of the femur was exposed via the septum intermusculare laterale with ligation of the perforating vessels. Following the preparation of the lateral circumference of the femur in the area where the estimated end of the osteotomy flap was going to be positioned, two 3.2 mm holes were drilled under cooling. One hole was positioned above the linea aspera and one hole was positioned 180 degrees ventromedial from the first hole. The osteotomy at the proximal ventromedial trochanter region was applied with a chisel at the vasto-gluteal border. In the next step the dorsolateral osteotomy was performed using a water-cooled oscillating saw, followed by the transverse osteotomy connecting the two drill holes. Then an osteotomy of the distal ventromedial at 3 cm was performed. The ventromedial osteotomy was completed in the proximal direction with a chisel driven blind under the vastus lateralis muscle to the proximal end of the osteotomy. The vastus lateralis muscle thus remains attached to the mobile bone flap. After implantation of the cement-covered interim stem, the bone flap was closed with two double cerclage wires (1.5 mm diameter) (Fig. 1). In the second-stage surgery, the bone flap was reopened and closed again with cerclages after the implantation of the new prosthetic components. All surgeries were performed in Orthopaedic Clinic Markgröningen.

Post-explantation a radical debridement followed and the custom-made interim prosthesis was implanted. The stem spacer component consisted of a regular cemented prosthesis stem. Prior to encasing the stem with bone cement supplemented with individually mixed antibiotics, it was coated with the patient’s own blood. Furthermore, implantation was performed with the cement being 6 min old to reduce the quality of interdigitation of cement. Both steps were carried out to facilitate later removal. The acetabular spacer consisted of a support ring [Müller cup, Ganz ring, or a Burch–Schneider-acetabular reinforcement ring depending on the acetabular bone defect (ZimmerBiomet, Winterthur, Switzerland)] and a polyethylene cup cemented herein. The support rings were fixed with two to maximal four screws and with an individual mixture of antibiotics in the cement according to the susceptibility of the microorganism.

The femur and acetabulum component were articulated with a metal head on the polyethylene cup (Figs. 2, 3).

Copal cement (Heraeus, Darmstadt, Germany) was used as industrially prepared cement. An individual and specific mixture of antibiotic substances was applied to the bone cement according to the antibiotic susceptibility profile of the microorganisms detected preoperatively. A maximum of 10% of the total cement powder weight was added as antibiotic substance to avoid mechanical problems. The cement of the spacer contained two antibiotic substances in 37 cases, three in 45 cases, and four in 3 cases (Table 1).

The individual parenteral antibiotic treatment started perioperatively. After 2 weeks of parenteral antibiotic therapy, the antibiotic treatment was changed to oral administration. The high bioavailability of rifampicin and ciprofloxacin allowed their oral administration already from the second day following surgery.

After 2 weeks of antibiotic treatment and mobilization with partial weight bearing on the operated leg, the patients were discharged from hospital. There was no restriction in range of motion for the operated joint except movements hazardous for dislocation.

During the re-implantation procedure, at least five biopsies were taken for bacteriological examination. Antibiotic treatment followed the same protocol as after the first-stage surgery.

After re-implantation of the new prosthesis, mobilization with partial weight bearing of 10 kg for a period of 6 weeks was allowed. Full weight bearing was gradually reached 3 months postoperatively as previously described by other authors for other cementless revision stems [13‐17].

Examination of all patients took place before the operation, 3, 6, 9, 12, 18, and 24 months postoperatively. Later follow-up examinations were based on a regular interval of every 2 years. Patients who no longer attended regular follow-up examinations, as described, were classified as lost to follow-up. Patients were judged definitely infection free according to Diaz-Ledezma et al. [9] if he or she was free from mortality related to periprosthetic joint infection, free from subsequent surgical intervention for periprosthetic joint infection, and if there was a microbiological and clinical absence of the infection for at least 2 years. If there was suspicion of a periprosthetic joint infection at a follow-up, the diagnosis was again ruled out or confirmed with the MSIS criteria 2014 the ICM criteria 2018.

The Staffelstein-Score was used to record the clinical and functional results. This is a score validated in German to record objective and subjective criteria for hip arthroplasty. It strongly resembles the Harris Hip Score [21], which does not exist in a validated form in German for which reason the Staffelstein-Score was applied. Items which are assessed similarly in both scores are “Pain”, “Support”, “Walking distance”, “Limp”, “Activities” (shoes, socks), “Stairs”, “Public transportation” and “Range of motion”.

Statistical analyses were conducted using IBM SPSS Version 20 (IBM, Armonk, NY, USA) and Microsoft Excel (Microsoft, Redmond, WA, USA). Distributions of variables within the groups were assessed by histograms and a non-parametric approach was chosen. Continuous variables are presented as medians and ranges, and categorical variables as frequencies. Comparison between groups was performed by Mann–Whitney U test, Wilcoxon test or Chi-square test as appropriate. All reported p values are two-sided, with an alpha level of 0.05 and have not been adjusted for multiple testing. Survival is presented with a Kaplan–Meier curve.

The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the local ethics board of the University Hospital of Tübingen (registration number 418/2021BO2).

Septic two-stage prosthesis revision surgery was performed in 86 patients with late periprosthetic infection of the hip endoprosthesis between May 2013 and September 2017. 29 women and 57 men formed the study collective. Median age was 70 (27–92) years. The average body mass index was 29.3 ± 6.1 kg/m². Diabetes mellitus and rheumatoid disease were known in 12 (11%) and 7 patients (6%), respectively. ASA classification was as follows: ASA 1: 2 patients, ASA 2: 38 patients, ASA 3: 45 patients and ASA 4: 1 patient. With 71%, the majority of explanted prostheses were primary implants (69% cementless, 18% hybrid, 10% cemented, 1.5% bipolar and 1.5% surface replacement prostheses). 29% of explanted prostheses were revision implants.

21 patients had already undergone one septic revision and three patients had already undergone multiple revision operations.

In 45 patients, an endofemoral approach and in 41 patients a transfemoral approach were performed. There was no statistical difference regarding sex (p = 0.175), body mass index (p = 0.935), diabetes mellitus (p = 0.244), rheumatoid diseases (p = 0.470), ASA classification (p = 0.890), type of explanted prosthesis (p = 0.161), previous septic revisions (p = 0.239), and follow-up (p = 0.384) between both groups. However, patients with endofemoral approach were of significantly younger age (p = 0.023) (Table 2).

Most frequently detected were Staphylococcus epidermidis and Cutibacterium acnes (Supplementary material, Table 1). In two cases, there was no causative microorganism detected. However, because of a fistula communicating with the joint and elevated CRP in one case and elevated CRP, a positive histopathological result and elevated polymorphonuclear neutrophil percentage in the joint fluid in the other case, they both met the criteria for a PJI. In 17 cases two different and in 1 case three different causative organisms were identified. In one case, even four microorganisms were found positive.

Median follow-up until dropout was 50 (2–87) months. Dropout reasons were reinfection (8 cases), death (6 cases), end of follow-up (53 cases) or lost to follow-up (19 cases). The 19 patients were lost to follow-up at 2–55 months (median: 24 months) postoperatively. No perioperative deaths were observed in the collective. The reported dropouts due to death were not directly linked to septicemia or treatment-specific complications associated with the infected prosthesis.

At the end of the follow-up, 78 cases out of the whole collective (90%) had no reinfection and 8 cases (10%) had to be classified as “reinfected”. Of the eight cases with reinfection, four were in the “endofemoral” (10%) and four in the “transfemoral” (12%) group. This distribution was not statistically different (p = 0.590). We additionally performed a worst case survival analysis, classifying the lost to follow-up cases as reinfected. Considering the worst case scenario, the reinfection rate was 33%. Survivorship (reinfection-free) at 36 months postoperatively was 94.6% (95% CI 91.6–100%) for the endofemoral group and 92.9% (95% CI 83.8–100%) for the transfemoral group. The difference for the function of survival between the groups was not statistically significant (log-rank test, p = 0.742). Calculating the worst case scenario, the survivorship at 36 months postoperatively was 83.6% (95% CI 73.1–95.5%) and 67.0% (95% CI 53.8–83.4%) for the endofemoral and the transfemoral collective, respectively (Fig. 4). However, the difference in the function of survival is also not statistically significant in the worst case scenario (log rank test, p = 0.184).

In three cases (7%: one case with Staphylococcus epidermidis, two cases with Staphylococcus capitis) in the endofemoral collective and in one case (3%: Staphylococcus epidermidis) in the transfemoral collective the samples taken during the reimplantation operation were positive for bacterial infection. However, in all three cases there was only just one positive culture out of at least five samples. Therefore, these samples were classified as contamination. None of these cases with a positive detection of microorganisms at the time of the reimplantation operation had a reinfection during follow-up.

The whole collective presented a median Staffelstein-Score with the spacer at the end of the interim phase of 60 (range 31–81) out of 120 possible maximum points. Comparing the clinical status of both procedures, with a median of 62 (range 31–81) for the endofemoral and of 60 (range 32–80) for transfemoral surgery, there was no statistically significant difference (p = 0.410) (Fig. 5).