Comparison of the efficacy of static versus articular spacers in two-stage revision surgery for the treatment of infection following total knee arthroplasty: a meta-analysis

The prosthetic joint infection (PJI) is a devastating and complex complication after total knee arthroplasty (TKA). Although the incidence was only 1 to 2% [1‐3], the number of the joint arthroplasties being performed is increasing. It can lead to serious consequences such as bone defect and necrosis; thus, the diagnosis and effective treatment of postoperative infection are very important. The two-stage revision surgery is the “gold standard” for patients with infection following TKA, especially for patients with chronic infection. The implantation of antibiotic-containing bone cement spacers can be used to eradicate infected microbes before prosthesis implantation [4, 5]. In addition, spacers maintain length of involved limbs and prevent muscle and soft tissue contracture via simple bone cement blocks inserting into the joint space [6‐8]. Two types of antibiotic bone cement spacer are available, including articulating spacers and statics spacers. The articulating antibiotic-loaded cement spacer can be used to make femoral and tibial prosthesis via the artificial plastic method. This type of spacer is similar to the anatomical structure of the normal knee joint, and thus has a good match with the residual bone surface. Meanwhile, it can maintain joint space and joint activity, reduce soft tissue adhesion, atrophy and scar formation, and reduce the recurrence rate of infection [6, 9‐11]. The static spacers can not only effectively deliver high concentration of antibiotic to control infection, but also maintain the joint space and limb length. However, the flexion and extension of the knee joint can result in soft tissue contracture around the joint and difficulty with reimplantation [1, 12]. The use of articular bone cement spacer is becoming increasingly widespread. However, some researchers reported higher risks of complications and the infection rate for the articular bone cement spacers compared to static spacers [13, 14]. Currently, there exists controversy regarding which antibiotic spacers are superior in the treatment of infection following total knee arthroplasty. The aim of the present meta-analysis was to compare the clinical outcomes of static spacers with mobile spacers for the treatment of infection following TKA. This paper was conducted in strict accordance with the PICOS principle of formulating the inclusion criteria, and the related literatures were collected and retrieved at home and abroad.

TKA has been regarded as an effective way to treat knee joint osteoarthritis, rheumatoid arthritis, and other advanced knee disorders since 1968. With the continuous improvement of surgical techniques and prosthetic design, TKA has become one of the most successful techniques in orthopedics. With the aging of the population, the demand for total knee replacement is still increasing. The increase in the amount of surgery is associated with increasing postoperative complications, of which PJI is one of the greatest complications. The current “gold standard” for the treatment of periprosthetic infection is still the two-stage revision surgery, which can fundamentally eliminate the infection and create healthy and vigorous tissue for the preparation of new prosthesis implantation [23]. During the two-stage revision, the doctor’s main goal is to prevent soft tissue contracture around the joint, which may lead to second operation exposure and refurbished prosthetic implant difficulties, and thus increased the surgical difficulties [24]. Therefore, it is crucial to maintain the stability of the knee joint and the balance of the soft tissue around the joint during the exclusion period of antibiotic bone cement. Moreover, there was a certain concentration of antibiotics in the knee joint to eradicate the infection [25, 26]. Currently, the antibiotic bone cement spacer can be divided into two categories, articulating type (hinge type) and static type (fixed type), and there was a significant controversy about whether articulating or static spacers can provide better outcomes. In this meta-analysis, we analyzed the difference of the two spacers regarding postoperative effects, including the infection eradication rate, the soft tissue release rate during the two-stage revision surgery, the ROM after the revision surgery, and function scores.

There was no significant difference between the two spacers regarding eradicate rate of infection in this study. The majority of the included articles reported the similar antibiotic use time, antibiotics for at least 6 weeks in vivo during the period of spacers [6‐8, 16‐18, 20‐22]. There is no consistent standard for the best mix of methods for preparing high-dose antibiotic bone cement spacers. Most of the antibiotics associated with spacers mainly vancomycin (1~4 g of added 40 g bone cement) or gentamicin, tobramycin (2.4~4.8 g of added 40 g bone cement) [25]. Bone cement spacers should have a sufficiently high level of antibiotic to provide a relatively high concentration of local tissue release levels, while the dose of antibiotics should be low enough to prevent the mechanical properties of bone cement spacers weakened [27]. John F. Nettrour et al. [28] reported that there was no dose-dependent of antibiotic bone cement. Voleti et al. [29] compared the clinical effects between the articulating and static groups, 1526 patients were included in their systematic review (872 cases of articulating spacers and 654 cases of statics spacers). Their findings indicated no statistical significance difference regarding the infection eradication. The results of Voleti et al. study are consistent with Piver et al. [30] study. In contrast, Romano et al. [31] showed that the articulating spacers can achieve higher infection eradication than the static spacers (91.2 versus 87%). However, some authors suggested that the static spacer can provide greater release and space for the soft tissue around the joint infection or blood transfusion and can better eliminate the infection [32].

ROM and functional recovery are indicators of postoperative efficacy. In this meta-analysis, the results of ROM and KSS (function) score in articulating group significance were higher than in the static group. However, there was no significant difference regarding HSS and KSS (pain) score between the two groups. Dr. Javad Parvizi and Thorsten Gehrke concluded that there was an encouraging postoperative function in patients which were treated with articulating spacers in the two-stage revision of the knee surgery. In contrast, the increasing trend of ROM in statics spacers was higher than those patients with articulating spacers after follow-up for 2 years. Their results were consistent with some larger scale systematical reviews [29, 30, 33].

In this study, the soft tissue release rate during the two-stage revision surgery was used to assess whether there were differences regarding soft tissue contracture between the two different spacers. The results showed that there was no significant difference between either quadriceps or tibial tubercle osteotomy in two different spacers. However, Guild et al. [33] found that the patients with articulating spacers were less associated with the assistive technology than those with static spacers. The reason may be that knee activities during the exclusion period helps to maintain the length and flexibility of the extensor device and prevent the formation of scar tissue around the knee, quadriceps shortening, and joint capsule thickening and contracture [34]. The results of our study suggest no significant difference regarding the soft tissue release rate, which may be due to the relatively small number of participants and sample size. Randomized controlled studies with larger number of samples are still required.

Bone loss is also a common complication during the exclusion period. Guild et al. [33] reported that there was less bone loss in articulating group than the static group. A retrospective case-control study [8] reported that 15 (75%) patients had pre-existing tibial or femoral bone loss in the static group, 10 cases (65%) had pre-existing tibial bone loss, 13 (65%) patients had femoral bone loss, and 8 patients had tibial bone loss. However, there was no bone loss in the articulating group. Some researchers demonstrated that the static spacers did not restore the normal knee anatomical profile [35]. In the traditional “block” static spacers, the bone cement surface and the bone surface are in point-like contact; the uneven pressure distribution and the local high point pressure lead to a large amount of bone loss during the exclusion period. However, with the continuous updating of the static type of the design of the spacers, there has been “internal skeleton” type static spacer, and the stability of the knee joint will be further strengthened. Yoo et al. [36] reported that the four patients using an internal skeleton-type static spacer showed excellent clinical outcomes and no bone loss, and noted that this technique may be more suitable for patients with suspected bone loss after removal of the prosthesis. For these patients, the articulating spacers cannot provide sufficient stability. Most of the included literatures did not provide sufficient data for the bone loss in our research.

There are several limitations in this meta-analysis. (1) There were no RCTs in this article, and retrospective case-control studies inevitably lead to the recall bias and the confounding bias between subjects and will lead to objectivity influences. (2) The length of follow-up period in patients with static and articulating groups was different in some literatures. Each article had different follow-up time. Although the duration of follow-up was greater than 12 months, the deviation caused by different follow-up time may also affect the objectivity of the results. (3) All included studies did not specify any inclusion or exclusion criteria to determine which patients would receive the articulating spacers or the static spacers, resulting in considerable selection bias. (4) All the included patients in this study had larger age span. The differences in medical and health levels will lead to greater differences in postoperative efficacy. (5) During the data extraction of some continuous variable data, only the mean values were available in the literature; therefore, the standard deviation was obtained by averaging and the P value or 95% CI [37], which may lead to bias in the outcome and affect the objectivity. (6) Due to the limited number of literature, it was not possible to provide sufficient data for subgroup analysis, which may be one of the sources of heterogeneity.

In this meta-analysis, we found no significant difference between articulating and static spacers in terms of infection eradication, the soft tissue contracture, and the knee pain scores. The patients with articulating spacers were able to achieve better ROM and limb function, but there was no significant difference in postoperative pain between the static and articulating spacers. Further randomized controlled studies on these two different spacers are still required to be carried out to provide more accurate and objective data for a comprehensive and accurate analysis.