Background

As a terminal means for treating osteoarthritis, joint arthroplasty can effectively reduce pain and improve quality of life. However, periprosthetic joint infection (PJI) is a severe complication of joint arthroplasty that causes significant pain and economic loss. It is expected that 10,000 patients with PJI will require revision each year by 2030 [1]. In recent years, improvements in surgical techniques and surgical conditions have led to a decline in the incidence of PJI from 1%–23% to 1%–2% [2]. With improvements in economic conditions and people's quality of life requirements, the total number of arthroplasty procedures have increased rapidly, and the number of PJIs has increased accordingly. Two-stage revision is considered the gold standard for treating PJI [3, 4]. However, the ideal time interval between surgical treatments, optimal antimicrobial agent, and duration of treatment remains controversial and the reported postoperative infection recurrence rate varies widely. In addition, some patients are in poor physical condition and may not be able to tolerate a second surgery. In recent years, the single-stage revision technique has received widespread attention and its application is increasing worldwide. Compared with two-stage revision, single-stage revision is more conducive to the functional recovery of the affected limb, reduces the occurrence of complications, reduces the overall treatment cost, reduces the surgical trauma, and improves patient satisfaction [5,6,7]. Moreover, several studies have reported comparable success with single-stage revision versus two-stage revision [6, 8, 9]. However, the evidence regarding single- and two-stage revision for PJI is inconsistent. This meta-analysis aimed to determine whether the reinfection and reoperation rates differ between the two treatment modalities and to ultimately reduce uncertainty in clinical decision-making for PJI treatment.

Methods

According to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement, this meta-analysis was performed in agreement [10]. The protocol for this meta-analysis was registered on PROSPERO (Registration No: CRD 42022369943).

Inclusion criteria

Study type: randomized controlled trial, cohort study, or retrospective study (Level I to III evidence). Study population: patients undergoing PJI. Intervention and control: single-stage in the treatment group, two-stage in the control group. Outcome index: clear reinfection or reoperation reported. Reinfection can be defined as the recurrence of clinical, serologic, or radiographic signs of infection during the follow-up period after the initial infection has been controlled. Reoperation can be defined as the patients need for further revision surgery.

Exclusion criteria

Letters, case reports, reviews, animal trials, or republished studies; Studies lacking a control group; Patients with septic arthritis or tuberculous arthritis.

Search strategy

Two of the authors (YZ and HL) performed the search in PubMed, EBSCO, Medline, and the Cochrane Central Register of Controlled Trials from the inception dates to May 30, 2023, using the keywords “(Two-stage or 2-stage or two stage or second-stage or double-stage) and (Single-stage or one-stage or 1-stage) and (arthroplasty or replacement) and (unhealed or infection or reoperate* or revise)”. No language restrictions were applied during the search.

Study selection

Two researchers (YZ and ZFW) screened the retrieved literature strictly and individually against inclusion and exclusion criteria. If two researchers do not agree during the literature screening process, it will be left to the senior researcher (HL).

Data collection process

Data on relevant outcome measures were extracted from the literature that met the inclusion criteria, including first author, year of publication, number of patients included, population characteristics (age, gender, comorbidities, etc), study design, PJI definition criteria, used, joint, surgical strategy, definition of failure (reoperation for infection, DAIR (debridement, antibiotic and implant retention), suppressive antibiotics), reason for reoperation other than infection and timming, follow-up by two researchers (SHF and HL) individually.

Outcomes

The primary outcome was the incidence of reinfection. A secondary outcome was the incidence of reoperation.

Assessment of risk of bias and quality of evidence

Two researchers (HL and SHF) independently assessed the quality of all included trials based on Cochrane risk-of-bias criteria [11]. The Newcastle–Ottawa scale (NOS) was used to evaluate the literature quality of the retrospective studies [12].

Data synthesis

The Meta-analysis was performed using Stata (version 17; StataCorp, 2021) software. The heterogeneity was assessed by using the Q test and I2 value calculation. Suppose the heterogeneity was not present (P>0.1 and I2<50%), the data was combined with a fixed effect model. The random effects model was used if heterogeneity was present (P<0.1 or I2 >50%). The odds ratio (OR) and their associated 95% confidence interval (CI) were used to assess outcomes, and a P value less than 0.05 suggested that the difference was statistically significant.

Subgroup analyses

We performed subgroup analyses for different surgical areas and periods of surgery.

Sensitivity analyses

We performed a sensitivity analysis on a case-by-case exclusion basis using random effect models.

Results

A total of 1663 documents were retrieved, 1012 duplicate documents were eliminated, the remaining 651 documents were read for abstracts and titles, 594 irrelevant documents were excluded, and 1 document failed to obtain the full text. The remaining 56 articles were read in full text. Fifteen studies were excluded, of which one review study, five case reports, three outcomes were no recurrence of infection, and six participants were without PJI. A total of 41 studies were included in the systematic review [5,6,7,8,9, 13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48], of which one study was excluded from the meta-analysis as the reinfection or reoperation outcomes could not be extracted [26]. A total of 40 articles were included in the meta-analysis (Fig. 1). The characteristics of the included studies are detailed in Table 1.

Fig. 1
figure 1

Flow diagram for search and selection of included studies

Full size image
Table 1 Characteristics of included studies
Full size table

A total of 41 retrospective studies were included in our systematic review. We used the NOS to assess the methodological quality and risk of bias. The quality scores were 6 to 9, indicating an overall low risk of bias (Table 1).

Reinfection

A total of 37 studies reported the recurrence of infection [6, 8, 9, 13,14,15,16,17,18,19,20, 22,23,24,25, 27,28,29,30,31,32,33,34,35,36, 38,39,40,41,42,43,44,45,46,47,48,49]. Van den Kieboom et al. [44] included both superficial and deep infections. We did not exclude superficial infections as these may result in deep infections. Among the cohort evaluated by Larsson et al., [30] we excluded one patient in the single-stage group who experienced treatment failure because the appropriate criteria were not met. There was mild heterogeneity between studies (I2=24.3%, P=0.106), and a fixed-effect model was used. There was no difference in the reinfection rate after single- versus two-stage revision for PJI (OR: 0.88; 95% CI: 0.73–1.07; P=0.209; Fig. S1A). As different surgical sites and surgery periods may have been a source of heterogeneity, subgroup analyses were performed. There was no difference in the reinfection rate between the single- and two-stage groups among the subgroups with PJI of the hip (OR: 1.35; 95% CI: 0.66–2.76; P=0.410; I2=53.5%; Fig. S1B), knee (OR: 0.76; 95% CI: 0.58–1.00; P=0.052; I2=0%; Fig. S1B), or shoulder (OR: 0.55; 95% CI: 0.16–1.88; P=0.338; I2=0%; Fig. S1B). Subgroup analysis based on the surgery period showed no significant difference in the reinfection rate between the single- and two-stage groups that underwent surgery after 2005 (OR: 0.79; 95% CI: 0.58–1.08; P=0.142; I2=0%; Fig. 2C), during both surgery periods (OR: 0.78; 95% CI: 0.58–1.07; P=0.125; I2=0%; Fig. S1C), or before 2005 (OR: 2.69; 95% CI: 0.58–12.37; P=0.204; I2=60.6 %; Fig. S1C).

Reoperation

A total of 18 studies reported the number of reoperations [5, 7,8,9, 16, 21,22,23, 33, 35, 36, 38,39,40,41,42,43,44]. There was no significant difference in the reoperation rate between the single- and two-stage groups (OR: 1.04; 95% CI: 0.79–1.37; P=0.792; I2=52.2 %; Fig. S2A). Considering the heterogeneity of the results, subgroup analyses were performed for different surgical sites and surgery periods. Subgroup analyses showed no difference in the reoperation rate after single-stage revision versus two-stage revision for PJI of the hip (OR: 1.49; 95% CI: 0.77–2.89; P=0.239; I2=76.6%; Fig. S2B), knee (OR: 0.93; 95% CI: 0.74–1.16; P=0.509; I2=2.6%; Fig. S2B), or shoulder (OR: 1.10; 95% CI: 0.31–3.75; P=0.880; Fig. S2B). Subgroup analysis based on the surgery period showed no difference in the reoperation rate between the single- and two-stage groups that underwent surgery after 2005 (OR: 0.77; 95% CI: 0.55–1.08; P=0.129; I2=0%; Fig. S2C) or during both surgery periods (OR: 1.23 95% CI: 0.82–1.83; P=0.316; I2=69.2%; Fig. S2C). Only one study that reported reoperation data was performed before 2005, and statistical calculations could not be performed because the number of events in both groups was 0.

Sensitivity analysis

A sensitivity analysis of the included studies was performed on a case-by-case exclusion basis. The remaining studies were combined using the OR values if any study was excluded. No individual study had a significant impact on the results (Fig. S3A and B).

Risk of bias

As shown in Fig. 2, the funnel plots showed some asymmetry, but the Harbord test showed no evidence of publication bias regarding reinfection (P=0.537) and reoperation (P=0.322).

Fig. 2
figure 2

Funnel plot of the included studies in this meta-analysis for the incidence of reinfection (A) and reoperation (B)

Full size image

Discussion

Several systematic reviews and meta-analyses have explored reinfection rates after single- or two-stage revision, but the two treatment protocols were not compared and the studies were limited to a single surgical site analysis [49,50,51,52,53,54,55,56]. Nagra et al. [57] published a meta-analysis of the two treatment options in 2016, but included only five retrospective studies comparing 796 patients with PJI of the knee. Since the publication of the meta-analysis by Nagra et al., [57] there has been a significant increase in studies reporting reinfection rates after single-stage revision for PJI of the knee. Therefore, we searched the literature for relevant studies and included studies evaluating the treatment of knee, hip, and shoulder PJI to determine whether the reinfection and reoperation rates differed between the single- and two-stage revision groups.

Our study found no difference in the reinfection and reoperation rates between the single-and two-stage groups. The decision whether to perform a single or two-stage revision is made at the discretion of the surgeon after considering all the details of the patient and the surgical site; therefore, there was some allocation bias that led to this result. In addition, for patients with hip and shoulder PJI, only part of the prosthesis may be revised [21, 40], leading to incomplete debridement and an increased reinfection rate after single-stage revision. The reoperation rate did not differ between the single-stage and two-stage groups for different surgical sites. As the data collected for the current analysis spanned a long period from 1969 to 2019, which may constitute a potential confounding source for our analysis, we performed subgroup analyses of surgery periods. These subgroup analyses showed no significant differences in the rates of reinfection and reoperation between the two groups. Svensson et al. [41] defined reinfection as the need for reoperation due to reinfection. However, some patients may have had secondary infections that resolved with medication and did not require a second revision surgery. This may have led to increased reporting bias.

Although two-stage revision has traditionally been considered the gold standard for treating PJI [58], it significantly reduces patient activity time to a total of approximately twice as long as single-stage revision. Removal of a well-immobilized prosthesis may also lead to degeneration of bone stock and perioperative fractures [21]. Single-stage revision offers advantages such as a similar failure rate as a two-stage replacement, reduced hospitalization and costs, and improved cost-effectiveness [58, 59]. Our findings suggest that there was no difference between single- and two-stage revision in the rates of reinfection and reoperation. Studies have found that the risk factors for failure of single-stage revision may be related to prior joint infection with Enterococcus or Streptococcus species [60], so it may be interesting to compare the outcomes of single-stage revision to treat PJI caused by these two bacterial species. Moreover, different studies have used different surgical procedures and methods for the criteria for single- versus two-stage revision, and there is no regulation of the use of antibiotics. The present study focused on whether single-stage revision can achieve the same treatment effect as two-stage revision while reducing the surgery time, pain, and cost. Therefore, more comparisons of antibiotics and optimization of surgical procedures need to be performed to provide a basis for formulating relevant guidelines.

Strengths

This is the first comprehensive comparison of the efficacy of single- and two-stage revision for PJI. This meta-analysis pooled 40 published studies involving 8711 patients with PJI, which may improve the statistical power of the data analysis and thus provide more reliable estimates. Sources of heterogeneity were analyzed, and subgroup analyses were performed for different surgical sites and periods of surgery. Our results showed that the success rate of single-stage revision was comparable to that of two-stage revision, challenging the assumption that two-stage revision is the gold standard for PJI. Clinicians are encouraged to consider single-stage revision for eligible patients with PJI. Compared with studies within a single country, our study pooled relative data from multiple countries worldwide, enhancing the universal applicability of the findings. Based on the Harbord tests and funnel plots, there was no significant publication bias in the included studies. Therefore, the results based on the available evidence are compelling.

Limitations

This study has several limitations. First, the most significant limitation of our article is that the included studies were all non-randomized controlled studies. The allocation of patients was not based on randomization but rather on the surgeons' experience, resulting in a preference for two-stage revision in patients with contraindications to single-stage revision or those with severe joint infection [9, 28], leading to allocation bias. Therefore, the confidence of the results needs to be further confirmed by randomized controlled trials. Second, the definition of reinfection after revision differed between studies. Castellani et al. [15] defined the outcome as a failure without stating the rates of reinfection or revision. Thus, we could only judge whether patients had reinfection based on the description of the definition in the complete text, and discussed each patient to decide whether to include them in the group with reinfection, which may have deviated from the authors’ original definition [15]. Third, Kheir, [27] Mahieu et al., [35] and Van den Kieboom et al. [44] studied patients with specific bacterial infections or those with negative bacterial cultures, which increased the bias of the results. Fourth, the present review included studies with follow-up periods ranging from 6 months to 22 years. Some studies had a very long follow-up, and the reason for reoperation was independent of the surgical modalities; in other studies, the follow-up time needed to be longer, resulting in missing outcome measures. Fifth, when we performed statistical calculations, we did not adjust the original data in accordance with confounding factors but simply combined the original data statistically, which increased the bias of the article. Sixth, in the studies we included, both partial and complete implant removal were incorporated, to some extent, increasing the heterogeneity of the article.

Conclusions

To our knowledge, this is the first meta-analysis to summarize the current evidence about the differences between single- and two-stage revision in treating PJI. We found that there was no difference between single- and two-stage revision in the reinfection and reoperation rates. Recognizing constraints in our inclusion/exclusion criteria and the observed diversity, we acknowledge the challenge of making definitive conclusions. Hence, we recommend a thoughtful, case-by-case consideration of the choice between single- and two-stage revision, considering patient-specific factors and staying attuned to ongoing research advancements.