Abstract
Objective
To systematically synthesize clinical, structural, biomarker, and safety outcomes of knee joint distraction (KJD) and implantable medial compartment shock absorbers (ISA) for tibiofemoral knee osteoarthritis (KOA), and to summarize comparative evidence versus arthroplasty, high tibial osteotomy (HTO), orthoses, and non-operative care.
Design
A PRISMA-based systematic review of PubMed, EMBASE, Scopus, and Cochrane Library from inception to 1 October 2025 identified peer-reviewed clinical studies of KJD or ISA for tibiofemoral KOA. Two reviewers independently screened records, extracted data, and assessed risk of bias. Owing to substantial clinical and methodological heterogeneity and overlapping cohorts, a narrative synthesis was prespecified and no quantitative meta-analysis was performed.
Results
Seventeen studies (13 KJD, 4 ISA) reporting on approximately 400 patients met the inclusion criteria. KJD yielded clinically important 1- to 2-year improvements in WOMAC/KOOS and VAS pain, with arthroplasty-free survival of roughly 75% to 85% at 5 to 9 years in selected series, accompanied by increases in radiographic joint-space width and MRI-derived cartilage thickness in the most affected compartment. ISA consistently improved WOMAC pain and function and showed higher 2-year arthroplasty-free survival than HTO or non-operative comparators.
Conclusions
Current evidence, based on small heterogeneous cohorts at low-to-moderate certainty, suggests that KJD and ISA can provide meaningful short- to mid-term symptom relief and delay arthroplasty in carefully selected patients. KJD and ISA address different indications and mechanisms and should be considered complementary rather than interchangeable joint-preserving strategies. Larger, independently replicated randomized trials with standardized structural and clinical endpoints are needed before widespread adoption.
Introduction
Knee osteoarthritis (KOA) is a chronic degenerative disease with a high incidence and one of the main causes of musculoskeletal disability and functional limitation.1 -3 Recently, the Global Burden of Disease projected a substantial growth in OA in the next 25 years, underscoring the need for durable joint-preserving strategies that can delay arthroplasty in working-age patients.3 -5
Knee replacement is certainly a safe and effective intervention in reducing pain and improving function, still suffers from complications both biological (i.e., infections, allergies) and mechanical (i.e., wearing, loosening), that underscore the need to try with joint preserving measures first.6 -10
Recently, building on ankle data, unloading via distraction has been explored for the as a joint-preserving intervention.11,12 Knee joint distraction (KJD) is a joint-preserving procedure that temporarily unloads the tibiofemoral (TF) surfaces by increasing the separation between tibia and femur during joint loading through a metallic frame, thereby decreasing the inflammatory action on the affected cartilage and potentially promoting the restoring of joint homeostasis.13,14 Prospective cohorts and randomized trials have shown clinically meaningful improvements in pain and function at various follow-up time points reinforcing biological plausibility for structure modification.15
-17 Across the reports, the most common regimen is approximately 5 mm of distraction for ~6 weeks, yet the optimal “dose” has not been unanimously defined yet18,19 (
Schematic representation of an external knee distraction system (KJD).
Schematic representation of an internal shock absorber system (ISA), medial side of the knee.
Previous systematic reviews and meta-analyses have already examined knee joint distraction and related unloading strategies. Earlier work by Li et al. 22 and others focused on the KineSpring system and early ISA constructs, while subsequent reviews and meta-analyses evaluated KJD outcomes, with an emphasis on clinical benefit and radiographic joint space changes.23 -27 Building on this literature, the present review incorporates the latest randomized and phase III comparative trials, systematically considers both external-frame KJD and ISA devices, and pairs clinical outcomes with radiographic, MRI, and biomarker findings, thereby providing an updated joint-preserving perspective across the tibiofemoral OA spectrum.
Hence, this systematic review aims to synthesize contemporary clinical evidence on knee joint distraction and implantable shock absorbers as joint-preserving strategies for tibiofemoral KOA, summarizing patient selection, clinical outcomes, structural and biomarker changes, complications, and comparative results versus arthroplasty, high tibial osteotomy, orthoses, and non-operative care. It was hypothesized that both KJD and ISA would provide clinically meaningful short- to mid-term improvements in pain and function in carefully selected patients, but that the overall certainty of evidence would remain limited due to small sample sizes, heterogeneous protocols, and non-randomized study designs.
Material and Methods
The present systematic review was conducted in accordance with PRISMA 2020 guidelines. The protocol was registered in PROSPERO (ID: CRD420251175841).
PubMed, EMBASE, Scopus, and the Cochrane Library were searched from inception to October 1, 2025, using the following terms using the following terms: ((“knee joint distraction” OR (“joint distraction” AND knee) OR “arthrodiastasis” OR “distraction arthroplasty”) OR (knee AND (“implantable shock absorber” OR “shock absorber” OR “load absorber” OR “unloading device” OR “KineSpring” OR “Atlas system” OR “Atlas” OR “Misha” OR KneeReviver OR ArthroSave))) AND (osteoarthritis OR osteoarthritis OR gonarthrosis).
Screening was conducted by 2 independent reviewers (SP and PQ). Eventual discrepancies were solved with a third senior reviewer (GA). Conference abstracts and other forms of gray literature were not systematically searched. Only peer-reviewed clinical studies reporting outcomes of knee joint distraction and/or implantable shock absorbers for tibiofemoral KOA, with at least 12 months of follow-up and written in English, were included. Records in other languages were excluded at title/abstract level; no full-text articles were excluded solely because of language. After title and abstract screening, full-text was obtained and the reference list of each articles was checked to retrieve potential missing reports. Preclinical studies, expert opinions, reviews, and case reports were excluded from the screening process. For KJD studies, we included external-fixator–based knee joint distraction procedures performed for tibiofemoral knee osteoarthritis with a minimum of 12 months’ follow-up. No a priori restrictions on Kellgren–Lawrence grade or limb alignment were imposed at the review level. For ISA studies, studies evaluating the implantable partial unloading devices used to treat medial compartment–dominant tibiofemoral OA with a minimum of 12 months’ follow-up were included. No a priori restrictions were imposed on Kellgren–Lawrence grade or alignment at the review level.
Risk of bias was assessed through the RoB2 tool for randomized controlled trials and with the ROBINS-I tool for non-randomized studies.28,29
The primary outcome of interest was change in pain and function as measured by validated patient-reported outcome measures (PROMs) such as the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Knee injury and Osteoarthritis Outcome Score (KOOS). Secondary outcomes included other PROMs (e.g., VAS/NRS pain, Oxford Knee Score, SF-36), structural imaging measures (radiographic joint-space width and MRI-derived cartilage thickness or denuded bone area), synovial fluid or serum biomarkers, and adverse events. Given the expected clinical and methodological heterogeneity (differences in patient selection, frame or implant design, distraction protocols, comparators, and outcome definitions), as well as overlapping publications from the same underlying cohorts at multiple follow-up time points, a narrative synthesis was prespecified and no quantitative meta-analysis was performed. When multiple articles reported outcomes from the same cohort, we summarized the number of reporting articles and the approximate number of patient-level outcome observations. Where precise de-duplication was not feasible, the counts are explicitly described as total observations and may include some overlap between publications.
Results
Study Selection
A total of 17 studies fulfilled the inclusion criteria and were finally included. A PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) flowchart of the literature review is reported in
PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) flowchart of the literature review.
Thirteen studies investigated the outcomes of the KJD, whereas 4 studies were focused on the outcomes of the ISA. A synopsis of the main characteristics of the included studies is reported in
List and Characteristics of the Included Studies (External Distractors).
List and Characteristics of the Included Studies (Shock Absorbers).
Several KJD publications report different follow-up time points from the same underlying cohorts, and some ISA reports by the same group share overlapping populations. Therefore, for outcome-specific summaries, patient numbers refer to total patient-level outcome observations across articles, rather than strictly unique individuals, unless otherwise specified.
Quality Assessment
Randomized controlled trials
The 5 different types of bias analyzed in ROB-2 were classified into “low risk,” “high risk,” or “unclear risk with some concerns.
28
Based on this, 2 studies18,40 were found to be associated with a medium risk of bias whether one with a high risk of bias
41
(
RoB2 risk-of-bias assessment for randomized controlled trials of knee joint distraction. Three RCTs were assessed. None of the trials were at low risk of bias in all domains; most were judged at some concerns or high risk in at least one domain.
Non-randomized clinical studies
The ROBINS-I tool assesses 7 different types of bias across 7 domains: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported results. For each domain, the risk of bias is classified as “low risk,” “moderate risk,” “serious risk,” or “critical risk” based on signaling questions. Afterwards, these domain-level risk judgments are synthesized into an overall risk of bias judgment for the study. The key criteria are: “Low risk of bias” if all domains are low risk, “Moderate risk of bias” if at least one domain is moderate risk, “Serious risk of bias” if at least one domain is serious risk but none are critical risk, and “Critical risk of bias” if any domain is critical risk. This ultimately categorizes the non-randomized interventional studies into 4 tiers of quality: “Low risk of bias,” “Moderate risk of bias,” “Serious risk of bias,” and “Critical risk of bias.”
29
Among the studies assessed, 5 study demonstrated critical risk of bias, 7 studies had a serious risk of bias, 6 a moderate risk of bias and none had a low risk of bias (
ROBINS-I risk-of-bias assessment for non-randomized knee joint distraction studies. The majority were judged at serious or moderate risk of bias, with no study at overall low risk.
ROBINS-I risk-of-bias assessment for non-randomized implantable shock absorber studies (n = 4). All studies were at least at serious risk of bias.
External frame—clinical outcomes
Ten KJD articles reported WOMAC outcomes, corresponding to 277 patient-level observations across partly overlapping cohorts. Early prospective Dutch series by Intema et al. 42 and their extensions by Wiegant et al. 15 and van der Woude et al. 16 showed large, clinically relevant improvements in total WOMAC and in pain, stiffness, and function subscales at 1 to 2 years, with mean gains well above the minimal clinically important difference and a high proportion of OARSI-OMERACT responders. These improvements were partially maintained at 3 to 5 years, although some attenuation was observed in a subset of patients.
More recent work by Struik et al. 19 reported a >15-point WOMAC improvement post-distraction, with 72% of patients classified as responders at 1 year (declining to 55% at 2 years). Struik et al. 33 found comparable WOMAC improvements between Knee Reviver and Dynamic Monotube devices, despite differences in baseline KL grades.
Analysis from Jansen et al. 30 linked WOMAC pain score improvements to reductions in Kellgren-Lawrence grade and biomarker changes (e.g., MCP-1 increases correlated with pain reduction, R2 = 0.41).
Across approximately 135 pain observations, VAS or NRS scores improved by about 40% to 60% at 1 to 2 years after KJD, mirroring the WOMAC pain and KOOS pain results.15,16,34,43,44 In randomized comparisons, KJD achieved similar improvements in pain and function to HTO at 2 years and comparable short-term pain relief to an unloading orthosis, while TKA generally produced the largest absolute gains in general health and quality-of-life metrics.34,35,43,44
Other clinical scales showed parallel trends, with meaningful improvements in KOOS, JOA, and functional performance tests after distraction.16,19,31,35,37 KOOS subscales improved in both prospective series and comparative trials, and in the cohort combining KJD with microfracture, most patients reached clinically important gains in symptoms, activities of daily living, sport, and quality-of-life domains. 37 Improvements in JOA score, EQ-5D, SF-36 physical components, and timed up-and-go performance were also observed, with KJD generally providing similar functional gains to HTO and somewhat smaller but still large gains compared with TKA.31,35,43,44 In the only study applying OARSI-OMERACT responder criteria, 75% of patients met the combined pain and function response definition at 2 years, in line with the large WOMAC and VAS improvements reported in that cohort. 15
Range of motion decreased during the distraction phase but returned to baseline or slightly improved thereafter in cohorts that reported ROM, with flexion commonly recovering to >110° and extension deficits improving or normalizing by 6 to 12 months after frame removal.16,19,35,38
External frame—radiological outcomes
Ten KJD articles (≈218 patient-level observations) reported radiographic joint-space width (JSW) as a structural outcome and generally demonstrated gains in minimum and/or mean JSW in the most affected compartment at 1 to 2 years after distraction15,16,18,19,33,34,42,44,45 Early Dutch cohorts showed mean increases in minimum JSW of roughly 0.5 to 1.0 mm at 1 to 2 years, with partial persistence of these gains up to 5 years in extended follow-up, although mean JSW differences were smaller and sometimes no longer statistically significant in the overall cohort.15,16,18,42
In comparative randomized and controlled trials, KJD produced larger minimum JSW increases than HTO in the most affected compartment and greater JSW gains than an unloading orthosis, despite similar or slightly greater short-term symptom relief in the TKA and orthosis groups.34,35,44 Later distraction cohorts with purpose-built devices showed JSW increases of comparable magnitude to earlier external-fixator constructs, supporting the reproducibility of the structural response across different frame designs.19,33
MRI-based structural outcomes were reported in smaller subsets of KJD cohorts and consistently showed increased cartilage thickness and reduced denuded bone area in the distracted compartment at 1 to 2 years, which paralleled improvements in WOMAC pain and function.15,17,42 Longer-term follow-up suggested that these cartilage gains may attenuate over time, with non-significant trends toward partial loss of the early benefit at 5 years, but without complete return to baseline in many patients.16,18 In one series, higher baseline damage and certain distraction-related parameters were associated with greater increases in cartilage thickness, linking structural response to both disease severity and biomechanical loading changes. 17 Lineham et al., 36 in a hinged arthrodiastasis cohort, also reported modest increases in cartilage volume and surface area and reductions in denuded bone area and bone marrow lesions at 12 months, again in parallel with symptomatic improvement.
Kellgren–Lawrence grade improved in a subset of KJD patients, with downgrading of radiographic severity in the most affected compartment in some studies and with such improvement correlating with better pain outcomes.16,18 Subchondral bone density, which was elevated at baseline in the most affected compartment, partially normalized after distraction, supporting a remodeling response of the subchondral plate and trabecular bone. 17 The modified MOCART score was assessed in a single long-term series and showed substantial increases at 6 years, consistent with durable cartilage repair in a subset of patients. 37
External frame—biological outcomes
Three studies (2 of them from the same cohort at different points in time) assessed collagen type II breakdown (CTXII) and synthesis (PIIANP) biomarkers (patients n = 75). Intema et al., 42 in their 1-year follow-up of knee joint distraction (KJD) for moderate osteoarthritis (OA), reported a trend toward reduced collagen type II breakdown (CTXII: −11 ± 39%; P = 0.078) and increased synthesis (PIIANP: +103 ± 298%; P = 0.060) between 6 and 12 months, suggesting early cartilage remodeling. Wiegant et al., 15 extending follow-up to 2 years, observed a numerical decline in CTXII (329-229 ng/mmol) and rise in PIIANP (1811-1856 ng/mL), though these changes lacked statistical significance. Conversely, Jansen et al. 41 demonstrated divergent dynamics: a transient but significant increase in CTXII (indicating elevated degradation) at 3 and 12 months (P = 0.020), followed by sustained increases in PIIANP at 12 (P = 0.008) and 24 months (P < 0.001), reflecting a delayed shift toward net collagen synthesis.
Jansen et al. 30 only evaluated dynamic changes in SF biomarkers during and KJD. At 3 to 4 weeks of distraction, TGFβ-1, TIMP-1, and IL-6 levels increased significantly, while activin A decreased (P < 0.05). By 6 to 7 weeks post-distraction, further changes were observed: LTBP2, TGFβ-1, IL-6, and MCP-1 increased, with activin A remaining reduced (P < 0.05). Notably, MCP-1 elevations during early distraction (3 weeks) were positively associated with 1-year improvements in minimum joint-space width (JSW) (R2 = 0.46, B = 0.008, 95% CI: 0.001-0.015) and standardized JSW (R2 = 0.46, B = 0.007, 95% CI: 0.001-0.013). Conversely, TGFβ-1 increases during this period correlated negatively with 1-year JSW changes (R2 = 0.64, B = −0.002, 95% CI: −0.003 to 0.001). Clinically, greater MCP-1 increases at 3 and 6 weeks predicted larger reductions in WOMAC pain at 1 year (R2 = 0.60, B = 0.094; R2 = 0.40, B = 0.033, respectively).
External frame—adverse events
Across KJD cohorts, pin tract infection was the most frequently reported adverse event, occurring in a substantial proportion of patients (often 50%-80%) and typically managed with local care and oral antibiotics.15,16,19,33,42,44,46 Although mostly superficial, these infections represent a considerable treatment burden and are inherent to the use of an external fixator.15,16,42
Serious complications were less common but included cases of osteomyelitis, pin-related femoral fracture, and persistent pain requiring repeated arthroscopic procedures or frame removal.16,18,19 Knee stiffness requiring manipulation under anesthesia at the time of frame removal was also reported in some series, particularly in regular-care cohorts.19,33,45 Over long-term follow-up, a proportion of patients progressed to total knee arthroplasty because of recurrent symptoms or osteoarthritis progression, with higher conversion rates in regular-care cohorts than in the earlier prospective trials.16,18,19,33 In randomized comparisons, the incidence of pin-site infections and related complications was substantially higher after KJD than after HTO, orthosis, or TKA, although most KJD complications were low grade and resolved without lasting sequelae.34,35,44,46
ISA—clinical outcomes
Four studies assessed WOMAC outcomes in approximately 140 patient-level observations . WOMAC scores demonstrated significant and sustained improvements across studies evaluating knee joint distraction (KJD) and implantable devices. Hayes et al. 47 reported a 60% statistically significant reduction in WOMAC pain (37.9 ± 13.4 to 15.2 ± 15.2) and function scores (36.5 ± 16.5 to 10.9 ± 14.7) over 5 years post-KineSpring® implantation. Similarly, Slynarski et al. 48 observed a 72% decrease in WOMAC pain (53.5 ± 8.6 to 15.0 ± 10.8) and 61% improvement in function (48.4 ± 17.2 to 18.8 ± 14.8) over 2 years with the Atlas system. Comparative studies highlighted superior outcomes for implantable devices versus traditional interventions. Diduch et al. 20 found greater WOMAC pain reduction in the ISA group (76% improvement) compared to high tibial osteotomy (HTO) (65% improvement), with 95.8% of ISA patients classified as responders versus 87.9% for HTO. Pareek et al. 49 further reinforced the efficacy of ISA in a randomized trial: 19% of all study subjects progressed to unicompartmental or total knee arthroplasty (UKA or TKA) during the 2-year follow-up, with 100% of these cases occurring in the Control group versus 0% of ISA subjects (P = 0.001). Only one study analyzed KSS: Hayes et al. 47 reported significant improvements in both the total KSS and its function subscale following implantation of the KineSpring® Knee Implant System. The mean total KSS improved from 63.6 ± 12.2 at baseline to 91.6 ± 12.9 at 60 months post-procedure (P < 0.05), reflecting a 44% enhancement in overall knee performance. The KSS function subscale demonstrated similar gains, rising from 72.6 ± 22.0 at baseline to 98.4 ± 3.6 at final follow-up (P < 0.05 at all timepoints except 6 weeks, where P = 0.105). These improvements were sustained over the 5-year study period.
Range of motion decreased transiently after ISA implantation but returned to baseline or improved slightly by 6 to 12 months and remained stable over 2 to 5 years, indicating that partial unloading did not compromise knee mobility.20,47,48
ISA—radiological outcomes
Collectively, these studies lacked quantitative or qualitative evaluations of joint structure, cartilage integrity, or bone remodeling via imaging modalities (e.g., radiography, MRI), limiting insights into the structural impact of the interventions. Future research should integrate radiographic metrics to correlate clinical improvements with anatomical changes.
ISA—adverse events
In the study conducted by Slynarski et al. 48 one patient reported knee stiffness 1.5 years after surgery that ultimately resolved, and one patient had a deep infection that was resolved with antibiotics and no further sequelae. Diduch et al., 20 had one instance (1/81) of a screw partially backed out (ISA subject) and one instance (1/81) of a screw breakage (HTO subject).
Discussion
The main finding of the present systematic review is that KJD offers clinically meaningful pain relief and functional improvement for carefully selected relatively young patients with tibiofemoral OA, with parallel signals of structural response on radiography and MRI in a subset of cases. Early prospective cohorts and subsequent randomized and non-inferiority trials consistently reported improvements in clinical scores and increases in JSW, with MRI-derived cartilage thickness gains most evident in the most affected compartment. These benefits can persist beyond the first postoperative year, and in some series up to 5 years, although attenuation is observed in part of the cohort over time.15,16,18,41,44,50
The present findings align with, and extend, the existing high-level evidence on knee joint distraction and unloading. Li et al. 22 emphasized the potential role of KineSpring-type implantable shock absorbers as an intermediate option between non-operative care and arthroplasty in medial compartment osteoarthritis, primarily from an economic and early clinical perspective. Takahashi et al. 23 and Goh et al. reported that KJD can yield substantial improvements in pain and function together with increased joint-space width, but highlighted that their meta-analyses were constrained by small sample sizes, heterogeneous distraction protocols, and limited long-term follow-up. Jansen et al. 25 confirmed, in a systematic review and meta-analysis, that clinical benefits and structural changes after KJD may persist at mid- to long-term follow-up in selected cohorts, whereas Lineham et al. 26 concluded that hinged arthrodiastasis, although conceptually attractive, currently lacks convincing evidence of superiority over standard static distraction constructs. More recently, Fernandes et al. 27 framed KJD, ISA devices, orthoses, and other strategies within a comprehensive biomechanical overview of invasive and non-invasive unloading methods.
Building on these prior reviews, the current systematic review combines external-frame KJD and modern ISA within a single joint-preserving framework, incorporates the most recent randomized and phase III comparative trials, and integrates clinical outcomes with radiographic, MRI, and biomarker data. This approach helps to clarify how KJD and ISA occupy different niches along the osteoarthritis spectrum, how their benefits compare with HTO and arthroplasty in the short to mid-term, and where the overall certainty of evidence remains limited by small cohorts, overlapping populations, and moderate-to-serious risk of bias.
An important nuance emerging from this review is that KJD and ISA do not target identical clinical phenotypes. KJD cohorts predominantly comprised relatively young patients (often in their fifth or sixth decade) with advanced, frequently bicompartmental tibiofemoral osteoarthritis in whom total knee arthroplasty would ordinarily be considered but was deemed too early or undesirable. In contrast, ISA studies enrolled patients with medial compartment–dominant disease, preserved lateral and patellofemoral joint surfaces, and mild-to-moderate radiographic severity, for whom high tibial osteotomy or extended non-operative management would more typically be alternatives.
Mechanistically, KJD produces temporary global joint distraction using an external fixator, aiming to reduce mechanical load and synovial inflammation while stimulating a reparative response in cartilage and subchondral bone. ISA devices, by comparison, provide permanent partial unloading of the medial compartment via an extra-articular implant, functioning more akin to a dynamic medial unloader than to true distraction. These distinct indications and mechanisms suggest that KJD and ISA occupy different niches along the OA spectrum and should be considered complementary, rather than directly interchangeable, joint-preserving options. A pragmatic summary of typical indications and contraindications for KJD versus ISA, distilled from the inclusion criteria and baseline populations of the included trials, is provided in
Typical Indications and Contraindications for Knee Joint Distraction (KJD) Versus Implantable Shock Absorbers (ISA) Based on Included Studies.
The comparative evidence reviewed here suggests that, in carefully selected patients, KJD can achieve improvements in pain and function of similar magnitude to HTO at 2 years, whereas TKA generally provides the largest absolute gains in PROMs.18,38,41 For ISA, randomized and controlled studies indicate superior symptom relief and arthroplasty-free survival compared with HTO or non-operative care in the short to mid-term.20,49 However, these comparisons are based on relatively small samples, limited follow-up, and few independent replications. Arthroplasty and HTO therefore remain the established standards for many patients with advanced or malaligned disease, and joint-preserving distraction or ISA should be considered adjunct or alternative options in well-counseled individuals rather than default replacements for conventional surgery.
A central question is whether outcomes are primarily driven by the frame configuration and device choice or by patient selection and execution of the surgical protocol. Historically, distraction has been performed with general-purpose monotube external fixators; more recently, a purpose-built, biplanar device has been introduced. Comparative clinical and bench data indicate no clear superiority of one device over the other at 2 years, with stiffness and gap maintenance influenced more by construct geometry (e.g., pin length/planes) than brand per se.19,33 Mechanical studies also suggest that under physiological loading, some loss of the distracted gap can occur, reinforcing the importance of meticulous intraoperative alignment of the hinge/axis and parallel tibiofemoral distraction, as well as vigilant postoperative pin care.51,52 Within this context, standardizing the construct parameters (pin planes, hinge location relative to the transepicondylar axis, target distraction and ramp-up schedule) would substantially increase inter-study comparability.
Despite the role in tibiofemoral OA has been deeply investigated and proven, less evidence is available in the presence of relevant patellofemoral (PF) osteoarthritis. Most distraction studies prioritize tibiofemoral outcomes and global PROMs; dedicated PF imaging and PF-specific scores are rarely reported. Consequently, there is insufficient evidence that distraction alone addresses PF-dominant symptoms. In those cases, adjunct procedures such as arthroscopic lateral patellar facetectomy with/without lateral release or even patellofemoral arthroplasty may be indicated.32,39,53 -55 Future KJD trials should prospectively phenotype the PF joint, stratify analyses by PF involvement, and include PF-specific outcomes, both radiological and clinical, to clarify whether combined strategies outperform KJD alone.56,57
Another unresolved issue is the “dose,” the amount of distraction. The most used protocol used of 5 mm maintained for 6 weeks—appears effective and is the best studied, but it is not necessarily the optimum. A dedicated study suggested 6 weeks are sufficient for meaningful clinical benefit and cartilaginous repair, while other cohorts have used 8 weeks with numerically larger but not definitively superior benefits.18,50 Pragmatic dose–response trials are warranted to define the minimal mm-distraction needed to achieve a significant clinical benefit.
Because pin-site infection is the most common adverse event after KJD (typically superficial and managed with oral antibiotics), the development of more “comfortable” options is of interest. The use of extra-capsular ISA partially could potentially resolve the issue while maintaining a substantial effect. In fact, previous studies demonstrated how ISA could provide superior pain and functional improvement versus high tibial osteotomy (HTO) at 2 years and encouraging arthroplasty-free “survival” in selected phenotypes (including knees with subchondral insufficiency fracture), suggesting a role in working-age patients who are not ready for arthroplasty or alignment osteotomy.20,21,49 Mechanistically, however, ISA differs from true distraction, getting closer to a medial unloader and the evidence on the effect on structural cartilage outcomes is still limited.
Regarding the construct itself, articulated/hinged distraction is another strategy intended to maintain range of motion during the distraction phase and potentially mitigate post-treatment stiffness. Early feasibility series report acceptable safety and improvements in PROMs, but whether articulating constructs compromise the capacity to maintain separation under load compared with rigid frames remains uncertain. 26 Comparative trials using objective in vivo measurements of inter-condylar distance (e.g., EOS/low-dose CT) are needed to determine equivalence in “distraction power” and to quantify any trade-off between motion and unloading.
The choice of observed outcomes also deserves particular attention. From a patient perspective, durable PROM improvement is fundamental; nonetheless, the biological plausibility of structural response matters when discussing disease modification. Increases in JSW and MRI-derived cartilage thickness gains after KJD have been reported, and early structural “repair” appears to predict longer-term clinical success and “survival” without arthroplasty in some cohorts.18,44 Further analysis with multiple endpoints pairing PROMs with objective measured both radiological and biological (i.e., MRI analysis, biomarkers) are warranted for future research.
Notably, the imaging and biomarker findings are derived from relatively small subsets of patients and from overlapping cohorts, and robust correlations between structural changes and long-term clinical benefit remain limited. Future studies should prospectively co-primary structural and clinical outcomes in larger, more diverse populations to clarify whether early tissue-level changes translate into durable disease modification.
Finally, repeatability is a pragmatic question for joint-preservation procedures. In Dutch regular-care implementations, re-distraction has been introduced for recurrent symptoms, and early experience suggests feasibility; however, prospective data quantifying the effectiveness, durability, and complication profile of repeat KJD remain limited.25,43
The major strengths of our work is the synthesis, unique to our knowledge, of clinical studies focusing on each phenotype of tibiofemoral OA. By extracting both PROMs and radiological endpoints we could collect symptomatic benefit alongside biological plausibility.
Nonetheless, the present study suffers from several limitations. First, the evidence base is heterogeneous in patient selection, frame design, pin configuration, hinge/axis alignment, distraction entity and duration, concomitant interventions and rehabilitation protocols. Second, outcome reporting was diverse and often not comparable enough to conduct a meta-analysis. Third, many studies included a small number of patient, single-center, and sometimes represented serial publications of overlapping participants at different time points. Additionally, follow-up durations varied, and reporting of strategies to handle missing data was limited, all of which reduce precision and external validity. Finally, by design we evaluated both external frames and ISA within a single synthesis to position these options within joint-preserving procedures. While clinically pragmatic, this introduces modality-level heterogeneity because ISA unloads without true articular separation and pursues different mechanobiological targets. Hence, inter-technique comparisons still remain difficult. Taken together, these factors indicate that the overall certainty of evidence supporting both KJD and ISA remains low to moderate at best. Most non-randomized studies were judged at moderate or serious risk of bias and none of the randomized trials were at low risk in all domains. Consequently, effect estimates may be imprecise and susceptible to residual confounding and publication bias, and the findings should be interpreted as hypothesis-generating rather than definitive.
As the evidence base expands, future reviews may be able to apply multivariate or multilevel meta-analytic methods that jointly model multiple outcomes and account for clustering of overlapping cohorts; however, the current dataset was too sparse and heterogeneous to support such approaches.
Clarifying indications in PF-dominant disease, optimizing the duration and quantity of distraction, conducting studies with comparable clinical and radiological outcomes will represent the next steps to confirm whether early structural repair can be translated into durable disease modification for this new novel joint-preserving procedure.18,53
Conclusions
KJD appears to be a promising joint-preserving option for carefully selected, typically younger patients with advanced tibiofemoral osteoarthritis, providing meaningful short- to mid-term improvements in pain and function with parallel structural signals in a subset of cohorts, at the cost of a high rate of mostly minor pin-site complications. ISA offer a mechanistically distinct alternative for patients with medial compartment–dominant disease and preserved lateral and patellofemoral compartments, with consistent clinical improvements and encouraging short- to mid-term arthroplasty-free survival but limited structural data and a small risk of device-related reoperations. Overall, the certainty of evidence for both strategies remains low to moderate because of small, single-center studies, overlapping cohorts, and moderate-to-serious risk of bias. KJD and ISA should therefore be regarded as complementary, investigational joint-preserving interventions for well-informed, carefully selected patients and, at present, as adjunct options ideally delivered within prospective registries or trials rather than routine substitutes for arthroplasty or high tibial osteotomy.
Footnotes
Acknowledgements
Artificial-intelligence–assisted language editing (ChatGPT, OpenAI) was used to refine grammar and style under the direct supervision of the authors, who take full responsibility for the content.
Ethical Considerations
This systematic review was conducted using previously published studies and did not involve new data collection from human partecipants or animals. As a result, ethical approval was not required.
Informed Consent
Informed consent was not applicable, as this study analyzed data from previously published literature.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Data sharing does not apply to this article as no datasets were generated or analyzed during the current study.