Knee osteoarthritis (OA) is the most common musculoskeletal disease estimated to affect
3.8% of the world's population.[1] Considered a disease of the whole joint, knee OA is characterized by loss of cartilage,
bone remodeling, and inflammation. Cumulative joint degeneration eventually leads
to substantial loss of function and quality of life, and represents a major cause
of global disability.[1]
[2] The burden of OA is set to increase with rising obesity levels and an aging population.[1]
[3] Gold standard treatment for OA of significant severity is joint arthroplasty after
initial conservative treatment. Beyond arthroplasty, no other treatment is proven
effective in halting or reversing disease progression. Globally, the prevalence of
knee OA peaks at 50 years.[1] However, both patients and surgeons are reluctant to replace joints where the patient
is expected to outlive the lifespan of the prosthesis as there is a greater risk of
revision surgery.[4]
[5]
[6]
Consequently, there is an increasing need for alternative treatments for this younger
OA population, not least because of the increased failure risk[5] but also because in some cases arthroplasty may result in poor clinical outcomes.[7] Following injury and osteoarthritis in the ankle, ankle joint distraction has provided
a useful means of reducing pain, improving function, and increasing radiological joint
space.[8] Likewise, basilar thumb arthritis has been effectively treated with joint distraction
and debridement in small prospective studies.[9] There are a certain risks of infection at pin sites and related bone infection often
observed in any surgical procedure using external frame and pins or wires; however,
such joint sparing alternatives are useful for patients who wish to preserve the native
joint.
A similar approach has been adopted to treat knee OA with knee joint distraction (KJD).
KJD uses an external fixator to unload the joint by distracting the tibia and femur.[10] It is reported that this temporary mechanical unloading allows natural intrinsic
repair processes to regenerate cartilaginous tissue evidenced by a sustained clinical
benefit and increase in joint width space.[11] With KJD being a joint sparing procedure aimed at postponing a first prosthesis,
successful clinical adoption could significantly improve patients' quality of life
and thus reduce the long-term health care costs associated with knee OA.
The aims of this systematic review are to identify and examine the current evidence
for the use of KJD focusing on clinical and radiological outcomes. This review will
also help to identify gaps in our understanding and so inform future clinical and
scientific studies.
Materials and Methods
Inclusion and Exclusion Criteria
Eligible studies included those involving patients aged 18 years or older with knee
arthritis that compared surgical KJD against other surgical procedures for knee arthritis.
There were no exclusion-based study designs or duration of distraction.
Information Sources and Search Strategy
Electronic databases (MEDLINE [Ovid], EMBASE [Ovid], Web of Science [ISI Web of Knowledge])
were searched from their inception until 25 February 2018 for studies meeting the
inclusion criteria. Searches were tailored to individual databases with the search
strategy for MEDLINE shown in [Appendix A]. In addition, reference lists of reviews and retrieved articles were assessed for
further studies as were registers of controlled clinical trials (metaRegister of controlled
trials [mRCT] [www.controlled-trials.com/mrct], clinicaltrials.gov [www.clinicaltrials.gov] and the World Health Organization [WHO] International Clinical Trials Registry Platform
[ICTRP] [http://apps.who.int/trialsearch/]). No restrictions were applied based on the publication status. Where necessary
authors were contacted for additional information.
Appendix A
MEDLINE (Ovid) Search Strategy
|
1. Knee joint/
|
|
2. distraction.mp. OR arthrodiatasis.mp
|
|
3. 1. AND 2.
|
Studies were assessed independently in duplicate for eligibility and data from eligible
studies extracted independently in duplicate into an electronic database (T.T., T.W.H.).
A risk of bias assessment was performed on included studies.
Outcome Measures Assessed
To assess the outcome of KJD, improvements from baseline to 1 year post intervention
were assessed. To compare KJD with other surgical interventions, outcomes at 1 year
post intervention were assessed.
The primary outcome assessed was functional outcome, assessed using a validated outcome
score, at 1 year following surgical intervention. Secondary outcomes included pain
scores, assessed using a validated pain score, structural assessment of the joint,
both radiographic and with magnetic resonance imaging (MRI), and assessment of adverse
events. All secondary outcomes were assessed at 1 year following surgical intervention.
Statistical Analysis
Heterogeneity of included studies was assessed using the I2 statistic and in the event of substantial heterogeneity (I2 > 85%), a meta-analysis was not be performed. As a degree of variability was expected
due to the subjectivity of the outcome measures, a random-effects model was used in
all cases. For continuous data, the mean difference (MD) was calculated along with
95% confidence intervals (95% CI), calculated using the inverse variance method. For
dichotomous data, the risk difference along with 95% CI was calculated using the Cochran-Mantel-Haenszel
method. Data analysis was performed using standard statistical techniques as described
in the Cochrane Handbook for Systematic Reviews of Interventions, using Review Manager-5.3
(The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).
Results
Three studies consisting of one cohort study and two randomized controlled trials
were identified as meeting inclusion criteria[11]
[12]
[13] ([Fig. 1]). The results of the cohort study were reported across three papers with relevant
data extracted where reported.[11]
[14]
[15] Included studies are outlined in [Table 1] with an assessment of risk of bias presented in [Fig. 2]. All studies were considered at high risk of performance and detection bias as it
was not possible to blind surgeons, participants, or outcome assessors as to the treatment
received. Attrition and reporting bias were assessed as low risk with no loss to follow-up
at 1 year reported. As all three studies originate from the same research group, it
was considered that this presented an unclear risk of bias.
Table 1
List of included studies
|
Author
|
Procedure
|
Age
|
Male gender
|
BMI
|
|
van der Woude et al 2017[11]
|
KJD vs. OA
|
48.5 vs. 51.2
|
55% vs. 45%
|
29.6 vs. 31.1
|
|
van der Woude et al 2017[12]
|
KJD vs. TKA
|
54.9 vs. 56.2
|
45% vs. 36%
|
27.4 vs. 29.4
|
|
van der Woude et al 2017[13]
|
KJD vs. HTO
|
51.2 vs. 49.4
|
73% vs. 60%
|
27.5 vs. 27.2
|
Abbreviations: BMI, body mass index; HTO, high tibial osteotomy; KJD, knee joint distraction;
OA, osteoarthritis; TKA, total knee arthroplasty.
Fig. 1 PRISMA flow diagram.
Fig. 2 Risk of bias summary.
Two studies were excluded as they reported the results of arthroscopic microfracture
in combination with KJD and it was the authors opinion that, as microfracture is already
an established treatment for cartilaginous loss, it would not be possible to delineate
any treatment effect seen.[16]
[17] The first of these studies by Deie et al reported the outcomes of six knees managed
with KJD and microfracture and found at a mean 3-year follow-up significant improvements
in Japanese Orthopaedic Association Score, VAS pain score, and radiographic joint
space width.[16] The second, by Aly et al, reported the outcomes of 61 knees, 19 managed with KJD,
joint debridement and microfracture and 42 managed with joint debridement and microfracture
and found that at a mean follow-up of 3 to 5 years the group managed with KJD, joint
debridement, and microfracture had significantly improved pain, walking capacity,
stair climbing, and radiographic joint space width compared with baseline, whereas
those treated with joint debridement and microfracture without KJD did not.[17]
Outcomes of KJD Improvement from Baseline to One Year Post Intervention
Primary Outcome
The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores
at baseline and 1 year post KJD were reported in all three studies, 62 patients, with
a significant improvement in WOMAC scores, MD 28.7 points (p < 0.001; 95% CI, 22.6–34.8), between baseline and 1 year post surgery observed ([Fig. 3]). Improvements were seen across all subdomains of WOMAC: pain (p ≤ 0.001; MD 29.3 points 95% CI, 21.9–36.5), stiffness (p ≤ 0.001; MD 19.5 points 95% CI, 8.4–30.6), and function (p ≤ 0.001; MD 29.5 points 95% CI, 23.6–35.4).
Fig. 3 Forest plot of improvement from baseline to 1-year Western Ontario and McMaster Universities
Osteoarthritis Index scores in knees managed with knee joint distraction. CI, confidence
interval; SD, standard deviation.
Knee injury and Osteoarthritis Outcome Score (KOOS), Intermittent and Constant Osteoarthritis
Pain (ICOAP) score, EuroQol 5 Dimensions (EQ-5D), and Short form (SF)-36 were reported
in two studies, 42 patients. Significant improvements between baseline and 1 year
scores were observed for KOOS (p < 0.001, MD 23.2 points 95% CI, 15.4–31.1), ICOAP (p < 0.001, MD 26.7 points 95% CI, 17.0–36.4), and EQ-5D (p < 0.001, MD 0.15 points 95% CI, 0.06–0.23) and all subdomains. Significant improvements
between baseline and 1 year SF-36 physical component score (p = 0.009, MD 7.8 points 95% CI, 1.9–13.7), but not mental component score (p = 0.41, MD −1.5 points 95% CI, −5.0–2.0) were observed.
Secondary Outcomes
Pain score, assessed using a pain visual analog score (VAS) 0 to 100 where 0 was equivalent
to no pain, was reported in all three studies, 62 patients. Patients managed with
KJD reported significant improvements in pain VAS of 33.3 points (p ≤ 0.001; 95% CI, 19.7–46.9) from baseline to 1 year post surgery ([Fig. 4]).
Fig. 4 Forest plot of improvement from baseline to 1-year visual analog scores for pain
(0–100) in knees managed with knee joint distraction. CI, confidence interval; SD,
standard deviation.
Structural assessment of the joint was performed radiographically in all three studies,
59 patients, and by MRI in one study, 20 patients. Between baseline and 1 year following
KJD, the radiographic minimum joint space width increased by 0.8 mm (p < 0.001; 95% CI, 0.5–1.0; [Fig. 5]) and mean joint space width increased by 0.8 mm (p = 0.003; 95% CI, 0.3–1.3). On MRI, the mean cartilage thickness over the total subchondral
bone area increased from 1.4 (standard deviation, SD 0.3) to 1.6 mm (SD 0.3; p = 0.03) on the tibia and from 1.0 (SD 0.4) to 1.4 mm (SD 0.3; p < 0.001) on the femur. The percentage of denuded subchondral bone decreased from 16.7
(SD 17.2) to 4.8% (SD 8.3; p = 0.006) on the tibia and from 27.3 (SD 25.6) to 4.2% (SD 10.2; p < 0.001) on the femur.
Fig. 5 Forest plot of improvement in radiographic minimum joint space width (mm) in the
affected tibiofemoral compartment from baseline to one year in knees managed with
knee joint distraction. CI, confidence interval; SD, standard deviation.
Adverse Events
Knee flexion was reported in two studies, 42 patients. No change in knee flexion between
baseline and 1 year following KJD was observed (p = 0.18; MD 2.4° 95% CI, −1.1–5.9) from baseline to 1 year post surgery. Across all
three studies, 62 patients, one patient was reported as requiring manipulation under
anesthetic (MUA) at 17 days following frame removal for stiffness.
Across all three studies, 62 patients, 42 patients developed single or multiple pin
site infection requiring antibiotics. Overall, the risk of developing pin site infection
was 69% (95% CI, 51–87) ([Fig. 6]). The risk of developing pin site infection requiring oral antibiotics was 57% (95%
CI, 33–82). The risk of developing pin site infection requiring intravenous antibiotics
was 10% (95% CI, 1–18). Overall two patients required surgical irrigation and debridement
with one developing osteomyelitis 3 weeks following frame removal.
Fig. 6 Forest plot of risk of pin site infection in knees managed with knee joint distraction.
CI, confidence interval.
Additional adverse events reported with the use of KJD included pulmonary emboli (2
of 20 patients [10%] in one study), postoperative foot drop managed with ankle foot
orthosis (1 patient), failure of the KJD device (1 patient), and breaking of a bone
pin during application (1 patient).
Outcomes of KJD Compared with Other Treatments
Primary Outcome
Two randomized controlled trials assessed the outcomes of KJD against other treatments
for arthritis, one against high tibial osteotomy (HTO) and one against total knee
arthroplasty (TKA). Both studies were conducted in patients aged 65 years and under.
At 1 year, no difference in total WOMAC score, or across subdomains, was seen between
knees managed with KJD and those managed with HTO (p = 0.25; MD −5.0 points, 95% CI, −13.5–3.5) or TKA (p = 0.53; MD −3.0 points, 95% CI, −12.5–6.5) ([Fig. 7]). At 1 year, no difference was seen in KOOS, ICOAP, EQ-5D or SF-36 between treatment
groups.
Fig. 7 Forest plot of 1-year Western Ontario and McMaster Universities Osteoarthritis Index
scores in knees managed with knee joint distraction (KJD) compared with knees managed
with high tibial osteotomy (HTO) and total knee arthroplasty (TKA). CI, confidence
interval; SD, standard deviation.
Pain score, assessed using a pain VAS 0 to 100, was reported in both studies. At 1 year,
no difference in pain VAS was seen between knees managed with KJD and those managed
with HTO (p = 0.17; MD 9.0 points, 95% CI, −3.8–21.8) or TKA (p = 0.13; MD 10.0 points, 95% CI, −3.0–23.0) ([Fig. 8]).
Fig. 8 Forest plot of 1-year visual analog scores for pain (0–100) in knees managed with
knee joint distraction (KJD) compared with knees managed with high tibial osteotomy
(HTO) and total knee arthroplasty (TKA). CI, confidence interval; SD, standard deviation.
Adverse Events
At 1 year, no difference in knee flexion was seen between knees managed with KJD and
those managed with HTO (p = 0.05; MD 4.0 degrees, 95% CI, −0.1–8.1) or TKA (p = 0.07; MD 5.0 degrees, 95% CI, −0.3–10.3). No difference in the rate of MUA was
seen between KJD and HTO (p = 0.40; risk difference (RD) 0.05 95% CI, −0.1–0.2). A higher rate of MUA was seen
with TKA compared with KJD (p = 0.04; RD 0.14 95% CI, 0–0.3).
The risk of developing infection requiring antibiotics was significantly higher following
KJD compared with both HTO (p < 0.01; RD 0.5 95% CI, 0.3–0.8) and TKA (p < 0.01; RD 0.6 95% CI, 0.4–0.8). This is likely to be secondary to associated risks
of using pins which provide a communication between the external environment and lower
limb bones into which they are placed.
Discussion
The main findings of this systematic review are that KJD is associated with significant
improvements in functional scores, pain scores, and radiographic measures of cartilage
thickness at 1 year postoperatively, and in patients aged 65 years or younger have
comparable functional outcomes to HTO and TKA. The main limitation of KJD is the occurrence
of pin-tract infection that was reported in 69% (95% CI, 51–87) of patients and was
significantly higher than that seen in HTO or TKA. At 1 year, no difference in knee
flexion, compared with baseline flexion and flexion 1 year following HTO and TKA,
was seen. While MUA following KJD has been reported (one case across three studies,
62 patients), the rate of MUA was found to be significantly lower than the rate observed
following TKA.
Compared with older patients, in young patients managed with arthroplasty, the risk
of implant failure and subsequent revision burden is high and any intervention that
can postpone or reduce the need for the index procedure in this group, and other groups
at risk of poor outcomes, is worth considering. This review has found that KJD appears
to be a potential alternative treatment option in managing knee OA, and in patients
aged 65 years or younger, the results appear to be as good as HTO and TKA at 1 year.
While these results are promising, the high rate of pin site infection following KJD
is a concern because both HTO and TKA can give lower rate of postoperative infection.
Despite in the majority of these cases, resolution of infection was achieved with
oral antibiotics. In a very few instances, osteomyelitis has been reported, and surgeons
may well have concerns about performing arthroplasty in these cases should KJD fail.
However, Wiegant et al[18] described the safety to perform TKA following KJD and concluded that it appears
safe to treat patients several years following KJD with a TKA.
The mechanism by which KJD works is unclear. In the clinical studies of KJD, increased
radiographic JSW and coverage of denuded bone assessed by MRI were reported. Biomarker
analysis has reported that following KJD a decrease in the collagen type II breakdown
marker is observed coupled with an increase in the collagen type II synthesis marker.[14]
[15] While these findings would suggest that KJD changes the intra-articular environment
to one that favors cartilage repair. It is likely that the conflicting results obtained
in animal experiments are due to a variety of reasons such as differences in experimental
setup, type of surrogate endpoints used to assess cartilage repair, and limited follow-up.
Some studies have shown promising results with evidence of bone and cartilage repair,
while others have failed to demonstrate any advantage with KJD, with some even reporting
adverse effect on the cartilage integrity. It is clear from these conflicting observations
that more work is needed to establish indeed when and how joint distraction works
and in which scenarios.[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
Alongside the mechanism of action of KJD, there are several other areas of uncertainty
around this treatment. In the present studies, static distraction was applied using
two 45 kg springs to permit some degree of joint loading. Whether this represent the
optimum distraction force, and whether a hinged distractor, which has been demonstrated
to be superior for ankle OA, still needs to be assessed.[27]
[28] Additionally, the patient population most likely to benefit from distraction and
optimum duration of distraction remains to be defined. Early reports suggest that
men with more severe arthritis are most likely to respond to treatment, and 6 weeks
distraction provides equivalent clinical outcomes to 8 weeks distraction; however,
these findings are based on limited data, and appropriately powered trials comparing
the outcomes of KJD to other treatments for knee OA are required.[29]
[30] Finally, further information on the long-term efficacy of KJD is required. Current
data suggest that at 5 years the functional outcomes and structural assessments of
joint remain improved compared with baseline, approximately 70% of the patients treated
still have their own knee instead of the initially planned joint prosthesis.[11] At 9 years post distraction, still 50% of the patients continue to manage with their
own knee and thereby the need for an artificial joint is avoided. Remarkably, mostly
women seem to drop out and opt for further intervention, although there is no clear
explanation for this gender difference.[31]
The strength of this systematic review is that it is a comprehensive assessment of
the efficacy of KJD for the treatment of knee arthritis. The weakness of this review
is that it is limited by the data available, with only three studies available for
inclusion, with all originating from the same research group.
This study has highlighted that KJD may be a valid alternative to HTO and TKA in the
treatment of knee arthritis in the young, resulting in improvements in functional
and pain as well as evidence of structural improvements within the joint lasting beyond
1 year. However, further work is required to optimize the technique of KJD, define
the optimum population for its use as well as develop methods to reduce the risk of
pin site infection, the major complication associated with this technique. Ultimately
KJD needs to be assessed pragmatically through appropriately powered multicenter studies
designed to assess its long-term effectiveness and comparative efficacy against other
established treatments for knee OA.
