Risk Factors for Periprosthetic Infection Following Limb Salvage Surgery in Bone Sarcomas

Simple Summary

Advancements in the treatment of bone sarcomas have improved survival and enabled limb-saving surgeries for most patients, but complications like periprosthetic joint infections (PJIs) remain a significant challenge. This study explores factors that may increase the risk of PJI in patients who receive endoprostheses after bone sarcoma removal. By analyzing data from 66 patients treated between 2014 and 2021, we found that factors such as higher BMI, older age, surgical site, and specific health conditions (measured by ASA score and the Charlson Comorbidity Index) were linked to a greater likelihood of infection. Polymicrobial infections and resistant bacteria, like MRSA, were associated with recurring infections. These findings will help surgeons better understand and manage infection risks in these complex cases. Future research with larger patient groups is needed to confirm these results and improve outcomes for patients undergoing limb salvage surgery.

Abstract

Background: Multimodal treatment of bone sarcomas has improved survival and allowed limb salvage surgery in the majority of these patients. Periprosthetic joint infection (PJI) constitutes a challenging complication. Controversy remains regarding the risk factors for PJI. Here, we aim to identify them. We also discuss pathogens and treatments. Methods: The authors reviewed the institutional database to retrieve endoprostheses implanted after bone sarcoma resection from 2014 to 2021. In total, 66 eligible patients were identified. Results: A total of 14 (21.21%) periprosthetic infections were diagnosed. Of these, 10 occurred in men (71.43%, p = 0.143). Mean BMI, age at the time of surgery, and ASA score were significantly higher among patients who developed PJI (p = 0.003, 0.044, and 0.033, respectively). Site was an important factor as well (p = 0.029). The number of comorbidities and the Charlson Comorbidity Index were also higher among these patients (p = 0.264, 0.060, respectively). Histology did not play a role in PJI (p = 0.385). Conclusions: Our data allow surgeons to better understand and control risk factors for PJI. We identified BMI, age, ASA score, site, and the Charlson Comorbidity Index as the main risk factors. Polymicrobial infections and methicillin-resistant Staphylococcus aureus are associated with recurrent infections. A multicentric study with a larger cohort is needed.

1. Introduction

Overall management of bone sarcomas is a complex matter, involving a multidisciplinary team. Multimodal treatment has improved survival and allowed limb salvage surgery among these patients [1,2,3,4], with amputation, once the mainstay of treatment, being performed infrequently nowadays [3,5]. In these limb-sparing procedures, tumor prostheses are the standard of care for limb reconstruction after long-bone resection, granting stability, early function, and good long-term outcomes [1,6,7]. Bearing all this in mind, currently, more than 95% of osteosarcomas of the lower extremity are potentially treatable with a limb-sparing surgery [8,9,10].

Several factors may lead to periprosthetic complications, though, including infection, aseptic loosening, mechanical failure, and fracture. In 2014, Henderson et al. [11] classified the failure of limb salvage after reconstructive surgery for bone tumors into three general categories further divided into six types of complications. Mechanical complications include soft-tissue failure (type 1), aseptic loosening (type 2), and structural failure (type 3) [11]. Non-mechanical complications comprise infection (type 4) and tumor progression (type 5) [11]. Finally, pediatric complications are included in type 6 complications of the Henderson classification [11].

Periprosthetic joint infection (PJI) following limb salvage surgery in bone sarcomas (type 4 in the Henderson classification) constitutes a major and challenging complication, and its incidence (8–15%, with some studies reporting values as high as 43%) greatly exceeds that of primary conventional prostheses (1–2%) [1,2,4,5,12,13,14,15], due to patients’ immunosuppression, extensive soft-tissue dissection, long operating times, and neoplastic diseases [2,4,14,16,17].

Infection may lead to a delay in chemotherapy treatment, and the need for further surgeries [8] and long-term antibiotic treatment, causing further poor joint function and quality of life [4,18]. Early and accurate diagnosis and identification of pathogens are of great importance for selecting adequate treatment and improving outcomes [4]. Nevertheless, according to previous observations, PJI in the upper limb seems to be much lower (0–2.8%) than in the lower limb (7–25%) [1,12], the most frequent place for musculoskeletal tumors [1].

Several treatments, in association with antibiotics, are indicated in the case of PJI: surgical debridement, one-stage revision, and two-stage revision. Two-stage revision seems to be the most efficient strategy to eradicate a PJI, even though it also results in low bone stock and less soft-tissue coverage [1,5].

Given the high incidence and devastating impact of PJI on sarcoma patients, a variety of risk factors have been proposed, yet controversy remains regarding their true involvement. In this study, the authors aim to identify risk factors for PJI in the specific setting of bone sarcoma patients, while also clarifying the implications of site and kind of tumor for PJI, as well as the responsible pathogens.

2. Materials and Methods

2.1. Data Source

In this single-center retrospective observational study, following institutional board approval, the authors reviewed the institutional database to retrieve endoprostheses implanted after bone sarcoma resection from 2014 to 2021.

Benign tumors and metastases were excluded, as well as patients with malignant tumors operated on during this period whose first surgery occurred before 2014. Therefore, the authors identified 66 eligible patients. We focused exclusively on sarcoma patients to maintain a homogeneous study population, given that the oncologic and surgical characteristics in these patients differ markedly from those in non-cancer patients. Although a comparison with non-cancer patients might have been of interest, our institutional database did not provide a sufficient non-cancer cohort for robust analysis.

The tumor prostheses included proximal humerus, proximal femur, distal femur, total femur, and proximal tibia, using implantcast^TM (Buxtehude, Germany) prostheses in all cases.

2.2. Outcome Measures

Patient files were reviewed to obtain demographic data, including sex, age at diagnosis, body mass index (BMI), alcohol and tobacco use, ASA score, comorbidities, site and side, resection length, operative time, estimated blood loss, histology of the specimen, stage at diagnosis, chemotherapy and radiotherapy, and final outcome. The Charlson Comorbidity Index (CCI) was calculated for each patient based on their documented comorbid conditions using standard criteria to quantify the overall comorbidity burden and predict 10-year survival.

Patients with PJI were further analyzed and data on the time of infection, pathogen, antibiotic treatment, surgical treatment, and outcome were retrieved.

2.3. Statistical Analysis

Data analysis was performed using Microsoft Office Excel 2016^TM (Microsoft Corp, Redmond, WA, USA) and SPSS Statistics V29^TM (IBM Corp., Armonk, NY, USA). Comparisons between groups were performed using an independent sample t test, unpaired two-tailed t test, the Mann–Whitney U test, one-way ANOVA, the Kruskal–Wallis test, Fischer’s exact test, and the Kaplan–Meier method. A p value < 0.05 was considered significant. We assessed the normality of our data using the Shapiro–Wilk test. When normality assumptions were not met—particularly given the small number of infection cases (14 vs. 52 non-infections)—non-parametric tests (e.g., Mann–Whitney U test) were employed.

Due to the limited sample size and the risk of overfitting, multivariable regression analysis was not performed.

3. Results

3.1. Patient Characteristics

Thirty-five patients were male (53%) and thirty-one were female (47%). The average patient age at the time of the surgery was 38.8 ± 21.0 years (range 8–81 years), with a median clinical follow-up of 35.3 ± 33.2 months (range 1–188 months).

Mean BMI was 23.9 ± 4.0 (range 16–37). Six (9.1%) patients reported tobacco use and two patients reported alcohol abuse (3%). Mean ASA score was 2.3 ± 0.6 (range 1–3), and 12 (18.2%) patients reported other comorbidities (namely, heart, respiratory, vascular, and endocrine disorders). The Charlson Comorbidity Index predicts the 10-year survival in patients with multiple comorbidities; the mean score in this study was 3.5 ± 2.1 (range 2–9).

The distal femur was the most common site of bone sarcoma (18 cases, 27.3%), followed by the proximal femur (16 cases, 24.2%) and the proximal humerus (12 cases, 18.2%). Nine cases affected the proximal tibia (13.6%), eight the pelvis (12.1%), two the total femur (3%), and one occurred in the distal humerus (1.5%). Thirty-one cases occurred on the right side (47%).

In terms of medical treatment, 6 patients received neoadjuvant and 8 adjuvant chemotherapy alone, whereas 27 received both and 25 received none. There are no data related to chemotherapy in two patients. On the other hand, neoadjuvant radiotherapy was performed in 2 patients, adjuvant in 5, and 57 patients received none. No information regarding radiotherapy was available for two patients. We used STIMULAN^TM (Biocomposites Ltd., Wilmington, NC, USA), an absorbable calcium sulfate antibiotic carrier that allows the management of dead space, in 22 (33.3%) of our patients.

Considering surgical data, on average, operative time was 276.4 ± 109.5 min (range 51–665), resection length was 180.2 ± 60.0 mm (range 80–388), and estimated blood loss was 527.0 ± 449.2 mL (range 150–2500; no data in 39 cases).

There were 35 cases of osteosarcoma (53%), 17 of chondrosarcoma (25.8%), 9 of Ewing sarcoma (13.6%), 2 cases of fibrosarcoma (3%), 1 case of undifferentiated pleomorphic sarcoma (1.5%), 1 case of leiomyosarcoma (1.5%), and 1 case of small round cell sarcoma (1.5%).

In terms of classification, we used the TNM G classification for sarcomas, which groups the stages into IA and B; IIA, B, and C; III; and IV. Thirty patients (45.4%) were in stage IIB (T1a or b, N0, M0, and G3 or 4), and only one infection occurred in a patient with a lower stage than this.

Twenty-eight patients died during follow-up (42.4%)—nine of the osteosarcoma patients (9/35; 25.7%), twelve of the chondrosarcoma patients (12/17; 70.6%), five of the Ewing sarcoma patients (5/9; 55.6%), one of the fibrosarcoma ones (1/2; 50%), and the solo cases of leiomyosarcoma and undifferentiated pleomorphic sarcoma (100%).

One case was lost to follow-up. This patient was referred to our institution by a hospital in another country, and returned to his country of origin after treatment.

The previous stated demographics are summarized in

Table 1. Summary of baseline characteristics.

Baseline Characteristics
Sex (female; %)	31 (46.97)
Age at time of surgery (in years)	38.79 ± 21.03 (range 8–81)
Follow-up (in months)	35.33 ± 33.25 (range 1–188)
BMI (kg/m2)	23.88 ± 4.00 (range 16–37)
ASA score	2.29 ± 0.60 (range 1–3)
Comorbidities (no.; %)	12 (18.18)
Charlson Comorbidity Index	3.54 ± 2.11 (range 2–9)
Site of bone sarcoma (no.; %)
- Proximal humerus - Distal humerus - Pelvis - Proximal femur - Distal femur - Total femur - Proximal tibia	- 12 (12.12) - 1 (1.51) - 8 (12.12) - 16 (24.24) - 18 (27.27) - 2 (3.03) - 9 (13.63)
Chemotherapy (yes; %)	27 (51.92)
Radiotherapy (yes; %)	7 (10.94)
STIMULANTM (yes; %)	22 (33.33)
Operative time (min)	276.35 ± 109.46 (range 51–665)
Resection length (mm)	180.24 ± 60.02 (range 80–388)
Blood loss (mL)	527.04 ± 449.22 (range 150–2500)
Histology (no.; %)
- Osteosarcoma - Chondrosarcoma - Ewing sarcoma - Fibrosarcoma - Undifferentiated pleomorphic sarcoma - Leiomyosarcoma - Small round cell sarcoma	- 35 (53.03) - 17 (25.76) - 9 (13.63) - 2 (3.03) - 1 (1.51) - 1 (1.51) - 1 (1.51)
Periprosthetic infection (no.; %)	14 (21.21)

for a better understanding of the baseline characteristics.

3.2. Periprosthetic Infection—Risk Factors

A total of 14 (21.2%) periprosthetic infections were diagnosed during follow-up (Figure 1, Table 1). The mean time for infection development was 12.1 ± 15.5 months, with nine cases occurring in the 1st year following surgery, three between the 1st and 2nd years, and only two cases after 2 years of surgery (Figure 1,

Table 2. Risk factors for periprosthetic infection after tumor resection.

Risk Factor	Periprosthetic Joint Infection Patients
	No.	p
Sex (men)	10 (71.43%)	0.143
BMI (kg/m2)	26.00 ± 4.00	0.003
Age at time of surgery (in years)	50.00 ± 18.73	0.044
Operative time (min)	309.36 ± 153.33	0.345
Resection length (mm)	190.23 ± 58.75	0.653
ASA score (ASA 3)	9 (64.28%)	0.033
Comorbidities (% of patients with comorbidities)	4 (33.33%)	0.264
Charlson Comorbidity Index	4.64 ± 2.79	0.060
Site of bone sarcoma
- Proximal humerus - Distal humerus - Pelvis - Proximal femur - Distal femur - Total femur - Proximal tibia	- 0 (0) - 0 (0) - 3 (60%) - 2 (14.28%) - 4 (28.57%) - 2 (100%) - 1 (33.33%)	0.029
Chemotherapy	-	No statistical association
Radiotherapy	-	No statistical association
STIMULANTM	4 (18.18%)	0.759
Histology
- Osteosarcoma - Chondrosarcoma - Ewing sarcoma - Fibrosarcoma - Undifferentiated pleomorphic sarcoma - Leiomyosarcoma - Small round cell sarcoma	- 8 - 2 - 2 - 1 - 1 - 0 - 0	No statistical association

). Besides PJI, we had 12 (18.2%) more cases of local complications: 2 cases of osteosarcoma and 5 cases of chondrosarcoma recurrence, 4 cases of aseptic loosening, and 1 case of dislocation.

Ten of the periprosthetic infections occurred in men (71.4%, p = 0.143). The mean BMI of patients who developed PJI was 26.0 ± 4.0 (vs. 22.3 ± 3.6, p = 0.003), mean age at the time of surgery was 50.0 ± 18.7 (vs. 34.6 ± 19.2, p = 0.044), mean operative time was 309.4 ± 153.3 min (vs. 267.5 ± 94.3 min, p = 0.345), and mean resection length was 190.2 ± 58.8 mm (vs. 182.2 ± 55.9 mm, p = 0.653). None of the six smokers developed an infection. Regarding ASA score, one ASA 1 patient developed infection in contrast to four ASA 2 and nine ASA 3 patients (p = 0.033). In terms of comorbidities, 33.3% of the patients with comorbidities developed infection, as opposed to 18.5% of the patients without (p = 0.264). Patients who eventually developed PJI had a mean Charlson Comorbidity Index of 4.6 ± 2.8 vs. 3.2 ± 2.1 of the ones who did not (p = 0.060).

No patient with a humeral prosthesis developed PJI, whereas PJI was observed in 100% of patients with total femur replacements, 60% of those with pelvic reconstructions, 14.3% of those with proximal femur prostheses, 28.6% of those with distal femur prostheses, and 33.3% of those with proximal tibia prostheses (p = 0.029). Regarding laterality, eight patients with right-sided implants and six with left-sided implants developed PJI (p = 0.549).

There was no statistical relation in terms of going through chemotherapy and/or radiotherapy and infection in our sample. We had 4 infections out of 22 patients (18.2%) treated with STIMULAN^TM and 10 out of 44 patients (22.7%) that did not receive it (p = 0.759).

PJI occurred in eight osteosarcomas, two chondrosarcomas, two Ewing sarcomas, one fibrosarcoma, and the one pleomorphic sarcoma (p = 0.385) (Figure 2).

Out of the 28 patients that died during follow-up, 7 had had PJI vs. 7 out of the 37 that did not die (p = 0.762).

Table 2 summarizes these data.

3.3. Periprosthetic Infection—Pathogens and Treatment

As mentioned above, a total of 14 patients developed PJI. Regarding Gram-positive organisms, methicillin-susceptible Staphylococcus aureus (MSSA) was identified in one patient, and methicillin-susceptible Staphylococcus lugdunensis was detected in two cases (one in association with a multisusceptible Aerococcus viridans). Additionally, two patients were diagnosed with methicillin-resistant Staphylococcus aureus (MRSA) and two with Staphylococcus epidermidis (one of these cases also had Escherichia coli and Enterobacter cloacae).

Gram-negative pathogens were also isolated. In one case, multisusceptible Escherichia coli and Klebsiella pneumoniae (only resistant to ampicillin) were identified. Serratia grimesii, resistant to cefuroxime, ampicillin, and amoxicillin/clavulanic acid, was found in two cases. In another case, multisusceptible Proteus mirabilis, Escherichia coli, and Klebsiella pneumoniae (only susceptible to carbapenems) were isolated, while multiresistant Klebsiella pneumoniae was detected in one case (in association with Candida albicans). Finally, one case yielded a multiresistant Enterobacter cloacae complex and Citrobacter braakli.

Fungal involvement was noted in the case with multiresistant Klebsiella pneumoniae, where Candida albicans was isolated (already mentioned above).

In terms of antibiotic therapy, and in accordance with the commonly observed germens and their sensitivities in our hospital for infections in primary arthroplasties, the patients were started on intravenous piperacillin–tazobactam and vancomycin, which was afterwards adjusted consonantly with the antibiotic sensitivity test (AST). We added an antibiofilm antibiotic (usually rifampicin for Gram-positive and ciprofloxacin for Gram-negative bacteria, if resistance was not verified in AST).

Out of the 14 observed infections, 12 were detected early and considered acute and were, therefore, submitted to debridement, antibiotics, and implant retention (DAIR). The infection was successfully treated without relapses in seven cases, but five cases (the one with Staphylococcus epidermidis, Escherichia coli, and Enterobacter cloacae; the one with Proteus mirabilis, Escherichia coli, and Klebsiella pneumoiniae; the two patients with MRSA; and the one with Klebsiella pneumoniae and Candida albicans) evolved to chronic infection. Four of them resulted in amputation after multiple infections, and one was successfully treated with a two-stage revision. Two of these acute infections occurred in patients with relapse of the disease that died soon after.

Two chronic infections were diagnosed ad initium. One was submitted to a two-stage revision, but eventually resulted in transfemoral amputation. The other chronic infection was due to MSSA and was successfully treated with a one-stage revision, without relapses.

Table 3. Demographics of patients with periprosthetic infection.

Sex	Age	Site	Histology	Time to Infection (in Months)	Organism
M	69	DF	Osteosarcoma	1	MSSA E. faecalis
M	32	PT	Osteosarcoma	32	S. epidermidis
M	33	PT	Ewing sarcoma	8	E. coli K. pneumoniae
F	51	PT	Osteosarcoma	3	S. grimesii
M	17	DF	Osteosarcoma	6	MRSA
M	52	DF	Osteosarcoma	2	MRSA
M	22	Pelvis	Ewing sarcoma	4	S. marcescens
M	49	DF	Fibrosarcoma	55	S. epidermidis E. coli E. cloacae
M	64	PF	Pleomorphic sarcoma	13	S. lugdunensis A. viridans
M	22	DF	Osteosarcoma	23	S. lugdunensis
M	69	PF	Chondrosarcoma	17	P. mirabilis E. coli K. pneumoniae
F	61	Pelvis	Osteosarcoma	1	E. cloacae C. braakli
F	51	Pelvis	Osteosarcoma	1	K. pneumoniae C. albicans
F	51	TP	Osteosarcoma	3	S. grimesii

M, male; F, female. PF, proximal femur; DF, distal femur; PT, proximal tibia. MSSA—methicillin-susceptible Staphylococcus aureus; MRSA—methicillin-resistant Staphylococcus aureus.

summarizes these data related to demographics and pathogens in PJI after bone sarcoma resection.

4. Discussion

Advances in both local and systemic treatments now allow limb preservation in over 90% of bone cancer patients, with megaprostheses serving as the primary reconstruction method despite potential complications [1,2,3,4,8,9,10,19]. In our study of 66 bone sarcoma cases (2014–2021), osteosarcoma was the most frequent subtype (53.0%) and the distal femur the most common site (27.3%), which is consistent with previous reports [20].

Out of the 66 patients, 25 received no form of chemotherapy, largely because most of these cases were chondrosarcomas—a tumor type that generally responds poorly to chemotherapy [19].

Moreover, during follow-up, 28 patients died due to disease progression. When excluding the isolated cases of undifferentiated pleomorphic sarcoma and fibrosarcoma, chondrosarcoma was associated with the highest mortality rate (12/17; 70.6%), and seven patients required reintervention for local relapse. These findings may indicate both the aggressive nature of chondrosarcomas and their poor response to chemotherapy.

In addition to systemic disease progression, these tumors contribute significantly to local complications. In our cohort, 26 cases (39.4%) experienced local complications—whether periprosthetic infection or other issues—that necessitated further treatment.

Infection remains a major concern in these patients, as complications related to their implants can be devastating. Although many risk factors for periprosthetic joint infection (PJI) have been described, controversy persists. Thus, we retrospectively investigated risk factors associated with PJI.

We identified 14 periprosthetic infections (21.2% of our population), with the majority occurring within the first two years after prosthesis implantation (12/14, 85.7%) (Figure 1, Table 2). As reported in the literature, the incidence of PJI in megaprostheses is higher than that seen in primary conventional prostheses (1–2%), and some studies report rates as high as 43% [2]. Although our incidence was lower than some reports, it remains slightly above the 8–15% reported by others [1]. This wide range across studies underscores the current lack of consensus, making definitive conclusions difficult.

Regarding sex, although 10 of the 14 infections occurred in men (71.4%), this difference did not reach statistical significance (p = 0.143). Nonetheless, other studies have suggested that males may have a higher risk of PJI in both primary and revision surgeries [1,20,21], possibly due to differences in the hormonal environment and skin microbiota [20,21].

Some studies have associated a higher BMI with an increased risk of PJI following primary arthroplasty of the hip or knee [22], but there are not many reports on PJI and BMI following megaprosthesis implantation. For instance, Fujiwara et al. [1] analyzed 121 patients who underwent tumor prosthesis of the lower limb after resection of musculoskeletal tumors and found no association or trend between higher BMIs and PJI. In our cohort, though, PJI tended to occur in patients with a higher BMI (26.0 ± 4.0 vs. 22.3 ± 3.6), and the results were statistically different (p = 0.003). Considering the results in primary arthroplasties and the tendency we verified here, we believe a bigger cohort is mandatory to definitively state this association.

There seems to be a clear trend in older people developing infection when compared to their younger counterparts (50.0 ± 18.7 vs. 34.6 ± 19.2, p = 0.044), and the differences were statistically significant (p = 0.044), despite our small sample.

Fujiwara et al. [1] observed in their study that a longer operative time was associated with a higher risk of PJI. Many other reports have also verified this association [23,24]. In our study, the mean operative time was larger in patients who eventually developed PJI (309.3 ± 153.3 vs. 267.5 ± 94.3), but the results were not statistically different (p = 0.345).

Resection length was tendentially longer in patients who developed PJI (190.2 ± 58.75 vs. 182.2 ± 55.9, p = 0.653), and again we believe a bigger sample would allow us to obtain statistically different results. This factor has not been widely reported in the literature, and, in fact, it is something that the surgeon cannot usually control, having to extirpate the tumor with wide margins to prevent recurrence. Even so, we believe it is an important factor to report in future works so that we can establish its impact on PJI and therefore implement additional measures to prevent it in patients at increased risk.

As expected, we observed an increased number of infections in patients with higher ASA scores (one ASA 1 patient vs. four ASA 2 vs. nine ASA 3), and the results reached statistical significance (p = 0.033). One should note that 9 out of the 14 infections occurred in ASA 3 patients (64.3%), and 9 of the 24 ASA 3 patients developed infection (37.5%). Likewise, patients who reported comorbidities were more prone to infection (33.3% vs. 18.5%, p = 0.3264), as well as patients who had a higher Charlson Comorbidity Index (4.6 ± 2.8 vs. 3.2 ± 2.1), which was again really close to statistical significance (p = 0.060). Many studies have related PJI and comorbidities [2], but few have found a direct correlation between ASA and PJI and even less between the Charlson Comorbidity Index and PJI (let alone PJI in the context of tumor resection), as we seem to have found here.

Previous studies show a lesser risk of PJI in the proximal humerus when compared to the lower limb [1,24]. Other studies seem to show that, among megaprostheses in the lower limb following tumor resection, infection tends to be higher in the distal femur and proximal tibia, whereas the proximal femur tends to work as a protective factor [2,25]. Resection of proximal tibia tumors often leads to difficulty in terms of soft-tissue coverage, explaining the higher risk of infection, as shown by the drastic decrease in PJI following the introduction of gastrocnemius flaps as part of the reconstruction in this site [26]. In our sample, we did find correlation between site and PJI (p = 0.029), with the proximal femur being the least infected site in the lower limb, and we had no PJI in the proximal humerus (or in the only case of distal humerus, for instance). As anticipated, side played no role in the risk of infection (p = 0.549).

From a theoretical point of view, we could expect a higher PJI rate in patients undergoing chemotherapy and/or radiotherapy, as an overall depression in the immune system could set the optimal environment for infection. Our data did not show that, though. Other works in the past corroborate our findings [1,2,27], but there are also some studies reporting an association between infection and this kind of treatment [28]. Given the high rate of infection in these megaprostheses as opposed to what is verified in primary arthroplasty, we have recently added STIMULAN^TM, a way of delivering local antibiotic therapy, intending to decrease our infection rate. Considering the results of this study, there is no difference in terms of risk of infection between those who received STIMULAN^TM and those who did not (p = 0.759), but we need a bigger sample in order to draw further conclusions (Table 3).

We found no association between histology and PJI (p = 0.385) (Figure 2, Table 3). Most of our infections occurred in osteosarcomas (8 out of 14; 57.1%), but this was also the most common tumor in our cohort (35 out of 66; 53.0%).

No relation between infection and eventual death was found in our study (p = 0.762). However, infection-related mortality remains an important clinical issue. Several reports have suggested that PJI may contribute to increased mortality in patients undergoing megaprosthesis reconstruction by delaying chemotherapy, necessitating multiple surgical interventions, and exacerbating underlying comorbidities. Although the literature on the direct association between PJI and overall survival is limited, the adverse impact of infection on patient outcomes is well recognized. In our study, the absence of a statistically significant relationship may be due to the small sample size and limited statistical power. Future studies with larger cohorts are warranted to clarify whether infection-related complications independently predict increased mortality in this patient population.

The DAIR procedure is associated with a high rate of failure (from over 20% to over 40%) [29,30,31]. Despite this, DAIR remains a widely accepted first-line strategy for acute infections, particularly when early intervention is feasible and the infecting organisms are amenable to targeted antibiotic therapy [31]. In our study, 12 out of our 14 infections were acute, and 7 out of these 12 (58,3%) were successfully treated with the DAIR procedure, with no reinfections observed during follow-up. This success rate is comparable to those reported in the literature, supporting the use of DAIR in selected acute cases. Notably, the failures in our cohort occurred in patients with infections that were polymicrobial and involved multiresistant organisms—factors that have been associated with poorer outcomes following DAIR [28,29,30]. On the other hand, we diagnosed two chronic infections, and one of them was then submitted to a two-stage revision and one to a one-stage revision. The first one, a polymicrobial infection, eventually recurred, and ended in the worst-case scenario, an amputation. The one submitted to a one-stage revision was due to MSSA and never relapsed, so this seems to be a good option in selected cases, even though we only have this one case to report.

We recognize several limitations to our study. To begin with, it is a retrospective analysis of a small sample of only 66 patients at a single institution. Moreover, there is a lot of heterogeneity in terms of diagnosis and site. Ultimately, these aspects limited the associations between the potential risk factors we considered and infection. Additionally, we acknowledge that the observed associations with a higher BMI and increased Charlson Comorbidity Index (CCI) may be partly coincidental, due to the limited sample size and potential residual confounding factors. These findings should be interpreted with caution and validated in larger, prospective studies that can better adjust for these variables.

5. Conclusions

Megaprostheses are currently the main means of reconstruction after bone tumor resection, allowing more than 90% of limb salvage nowadays. Infection is a major concern in these patients, though.

In this study, we identified BMI, age, ASA score, and site as the main risk factors for PJI, with BMI being the only modifiable one, with population measures to prevent overweight. Men, the presence of comorbidities, and a higher Charlson Comorbidity Index seem to be important factors, as well. Longer resection length and operative time have also shown a tendency towards PJI.

The DAIR procedure is a good option in acute PJI, so we believe it is still a valuable weapon if applied to selected patients. Polymicrobial infections and MRSA are associated with recurrent infections.

Given our results, it is clear that a multicentric study with a much larger cohort is needed to definitively establish an association between these factors and PJI and validate our results.

Our data, however, further contribute to the understanding of this complex and important matter, allowing surgeons to better control some of these factors in the future in order to diminish the PJI rate.