Osteoarthritis (OA) of the knee is a noninflammatory degenerative disease of the knee joint consisting of 3 large categories: conditions that block normal synchronous movement, conditions that produce abnormal pathways of motion, and conditions that cause stress concentration resulting in changes to articular cartilage.1 Total knee arthroplasty is an established terminal treatment for late-stage OA of the knee. Yet, not all patients are appropriate because of age, comorbidity, or other factors. Intra-articular corticosteroid injection provides short-term pain relief,2 may require repeated treatment, and may cause cartilage damage over an extended exposure time.3 Anatomical studies have identified the superior lateral, superior medial, and inferior medial genicular nerves as possible targets for cooled radiofrequency ablation (CRFA) as a proposed therapeutic strategy to provide analgesia.4 Favorable outcomes have been reported using CRFA for sacroiliac joint pain5 and discogenic lumbar pain.6,7 Similarly, initial reports on knee denervation to provide extended analgesic effects in patients with OA were largely positive, but included a small cohort of patients or short follow-up.8–11
The current randomized controlled study was designed to test the hypothesis that CRFA was noninferior or superior to intra-articular steroid (IAS) injection at 6 months to treat OA-related knee pain. The primary efficacy end point was the proportion of subjects whose knee pain was reduced by 50% or greater from baseline at 6 months after treatment. Secondary end points included change in knee function, subjects' perception of treatment effect, and analgesic drug use 6 months following study interventions.
Study approval was obtained by the Western Institutional Review Board (Puyallup, Washington) and Rush University Medical Center Institutional Review Board (Chicago, Illinois). All patients properly consented prior to initiating screening activities. The study was registered in ClinicalTrials.gov (NCT02343003) on January 15, 2015.
Consecutive subjects presenting to study investigators with signs and symptoms of knee OA were considered for the trial. More specifically, radiographic confirmation of a patient having had OA within 12 months before study screening was required, with no other etiology demonstrated as the source of knee pain. Patients reported that they had suffered from OA pain for approximately 10 years (Table 1), and most entered the study with prior diagnoses of OA of the knee as evidenced by prior treatments received, medications taken, and independent radiographic assessment. The diagnosis was reconfirmed by study investigators before entry into the trial using this previous information, as well as patient presentation and physical examination. Subjects with bilateral knee OA were not excluded; only 1 knee was screened and enrolled as the “index knee” for treatment. Inclusion criteria included (1) knee pain for 6 months or more that was unresponsive to conservative treatments (physical therapy, oral analgesics, intra-articular injections with steroids, and/or viscosupplementation); (2) Numeric Rating Scale (NRS) pain score of 6 or greater for the index knee; (3) radiological confirmation of OA grades 2 to 4 noted within 12 months of enrollment; (4) Oxford Knee Score (OKS) of 35 or less; (5) positive diagnostic genicular nerve block (defined as a decrease of ≥50% in NRS score); and (6) if the patient was taking an opioid or other morphine-equivalent medication, the dose must have been clinically stable (<10% change in dosage for ≥2 months prior to the screening visit). Exclusion criteria included (1) body mass index of greater than 40 kg/m2; and (2) history of systemic inflammatory conditions such as rheumatoid arthritis or uncontrolled diabetes, cancer, previous total knee arthroplasty, previous knee RF block or ablation, or coagulopathy.
Described here is a prospective, randomized, open-label, multicenter (eleven sites) clinical study with a parallel-group design to compare CRFA utilizing the Coolief System (Halyard Health Inc, Alpharetta, Georgia) with IAS injection. Enrolled subjects underwent CRFA or corticosteroid injection in a 1:1 randomization scheme, and study follow-up visits occurred at 1, 3, and 6 months after study interventions. At the 6-month follow-up visit, patients randomized to the IAS cohort were allowed to “cross over” and receive CRFA treatment.
Subjects were permitted to use analgesics as needed during the study, but the total daily dose had to be the equivalent of or less than 60 mg of morphine at intake. Dosing for membrane stabilizers and antidepressants for pain remained constant throughout the study, unless approved otherwise by the investigator. Additional treatments to the index knee were prohibited during the study.
Subjects underwent fluoroscopically guided blockade of the superomedial and inferomedial branches of the saphenous nerve and the superolateral branch of the femoral nerve6 using 1 mL of local anesthetic (preferably Marcaine [bupivacaine] 0.5% or similar) at each site (with an ideal volume of 0.60–0.75 mL/site). A “positive” responder experienced a decrease in score on the NRS scale of 50% or greater8 at least 15 minutes after the injection. All positive responders were eligible for the study and randomized independently of the block results. Needle placement for diagnostic blockade was identical to Figure 1.4
Cooled Radiofrequency Ablation
Subjects randomized to the CRFA study arm underwent genicular ablation with the Coolief System. Cooled radiofrequency ablation of the index knee was performed using fluoroscopic visualization of anatomical landmarks for proper cooled radiofrequency (CRF) probe placement (Fig. 1). Subjects were placed supine on a radiolucent table, with the treatment knee slightly flexed on a bolster. After anesthetizing the skin and soft tissue, a 75- or 100-mm, 17-gauge CRF introducer was placed at the appropriate locations. If blood or fluid was observed after removing the stylet, the CRF probe was repositioned and reaspirated. True lateral fluoroscopic visualization, making sure to account for valgus or varus deformities, confirmed accurate probe positioning at 50% depth of the femur and tibia (Fig. 1B). A 4-mm, 18-gauge, internally cooled active tip electrode was placed into the introducer needle, and positioning was reaffirmed in the anteroposterior (Fig. 1A) and lateral (Fig. 1B) fluoroscopic views. Motor stimulation at 2.0 V established no muscular contractions, and sensory stimulation at less than 0.5 V in all 3 locations reproduced concordant knee pain. Motor stimulation was required in all cases solely as an additional safety check to ensure that there were no motor nerves present in the area being lesioned. The optional sensory stimulation provided supplemental confirmation that the target nerve locations identified during blocks were accurate and ensured proximity of the probe to each of these nerves prior to lesioning. Most of the initially treated subjects required local anesthesia, whereas 19 (28%) of the 67 required conscious sedation.
Each neural element was anesthetized with 1% lidocaine prior to CRFA at 60°C for 150 seconds. The heat generated from the RF energy produces thermal energy with average maximum tissue temperatures greater than 80°C.12 The needles were then removed, and study subjects were allowed to properly recover prior to discharge. All subjects were discharged to home with instructions for self-care.
Intra-Articular Steroid Injection
Study subjects randomized to IAS underwent 1 corticosteroid injection in the index knee. Subjects were placed supine, with the index knee prepared in a sterile fashion. A topical anesthesia was accomplished using ethyl chloride spray or local anesthetic. An appropriately sized needle per the investigator's routine practice was placed into the suprapatellar pouch, and a steroid dose equivalent to 40 mg Depo-Medrol (methylprednisolone acetate) was injected into the joint space. Across the study, Depo-Medrol, Kenalog (triamcinolone topical), and betamethasone were used in 70%, 18%, and 12% of treatments, respectively. Subjects were discharged to home with instructions for self-care.
The primary efficacy end point was the proportion of subjects whose knee pain was reduced by 50% or greater from baseline at 6 months after treatment. The 11-point NRS captured the amount of index knee pain at all study time points.13,14 Secondary end points included (1) change in knee function detected by OKS,15 (2) subjects' perception of treatment effect as reflected by the Global Perceived Effect (GPE) score,16 and (3) opioid and nonopioid (nonsteroidal anti-inflammatory drugs) analgesic use, as measured by subject self-reported average daily dosage used. Assessments of these study end points were made at baseline and at 1, 3, and 6 months following treatments. Outcome data were captured according to subjects' impressions made during the week preceding data collection at each study visit (baseline to 6 months).
All subjects were evaluated for adverse events (AEs) and serious AEs (SAEs) at each visit. Adverse events were any unfavorable and/or unintended sign (including abnormal laboratory finding), symptom, or disease temporally associated with the study procedure, regardless of relationship to the device. Serious AEs included events deemed life threatening or resulting in death, requiring inpatient hospitalization or prolonging existing hospitalization, resulting in persistent significant disability or incapacity, or warranting intervention to avoid the aforementioned SAEs.
A noninferiority evaluation was used to estimate the study sample size. The sample size was based on the estimated success rates of 59%8 (success = ≥50% NRS score reduction) and 47%17 (success = ≥30% NRS score reduction) in the CRFA and standard groups, respectively, and a noninferiority margin of 15%. Assuming an attrition rate of 20% and a 2-sided significance level of 5%, 144 subjects enrolled into the study would yield 114 subjects at the primary end point.
Continuous data were reported using descriptive statistics, and categorical data were summarized as counts and percentages with 95% confidence intervals (CIs). The statistical tests used for comparisons are described in the referenced tables and figures. Results of study end points that appear in text only are accompanied by descriptions of the statistical tests used for those comparisons. All statistical tests were 2-sided and performed at the 5% level of significance (α) except for the following prespecified end points: (1) change from baseline in OKS at 6 months: adjusted α = 0.0083, and (2) change from baseline in opioid analgesics usage at 6 months: adjusted α = 0.025.
Subjects in each study group may not have provided all observations at each follow-up time point. The number of subjects from which data were collected to make each assessment determination is indicated in the presentation of study results.
Two-hundred thirty-three subjects were screened, and 151 were randomized to study cohorts (CRFA, n = 76; IAS, N = 75) (CONSORT diagram; Fig. 2). Of those subjects, 138 proceeded to treatment (CRFA, n = 67; IAS, n = 71), and 13 others did not because of withdrawal (n = 9), being lost to follow-up (n = 2), and protocol deviations (n = 2). At 1 and 3 months after treatment, 136 and 133 subjects remained in the study, whereas at 6 months, 126 subjects (CRFA, n = 58; IAS, n = 68) were evaluated for study outcomes. Overall, a less than 20% dropout rate for the primary end point analysis occurred, with 87% (58/67) of the CRFA group and 96% (68/71) of the IAS group returning at 6 months to contribute data. Results presented in this article are according to those in the full analysis set. A last observation carried forward (LOCF) analysis was also conducted to accommodate missing data points, and selected data are shown to reinforce the LOCF achievement of the primary end point.
No differences at baseline existed between groups regarding age, sex, race, mean body mass index, mean duration of knee pain, and analgesic medication utilization. No baseline differences existed between groups in knee OA severity, mean index knee pain levels (NRS scores) before diagnostic block, and the extent of index knee pain reduction postdiagnostic block (Table 1). The mean reduction in pain by blocks was 83.2% in the CRFA group and 80.8% in the IAS group (P = 0.42). Logistic regression analysis did not identify a correlation between response to blocks and response to CRFA treatment. Regarding prior knee procedures, 48.7% and 54.7% of the CRFA and IAS groups, respectively, reported having a previous IAS injection in the index knee (P = 0.46), and 22.4% and 29.3% of these study groups reported prior viscosupplementation (P = 0.33).
Numeric Rating Scale
Mean baseline pain scores in the CRFA (n = 76; 7.3 ± 1.2) and IAS (n = 75; 7.2 ± 1) cohorts were not different (P = 0.55, Wilcoxon rank sum test) (Fig. 3). Within both study groups at 1, 3, and 6 months, mean pain score was reduced (P < 0.0001 at each data point, paired Student t test) relative to baseline.
At each follow-up interval, the mean knee pain score was less in the CRFA group than in the IAS group (Fig. 3) (1 month: CRFA [n = 67], 3 ± 2.3; IAS [n = 69], 3.9 ± 2.2, P = 0.025; 3 months: CRFA [n = 65], 2.8 ± 2.2; IAS [n = 68], 5.2 ± 2, P < 0.0001; 6 months: CRFA [n = 58], 2.5 ± 2.3; IAS [n = 68], 5.9 ± 2.2, P < 0.0001) (Wilcoxon rank sum test). Mean reductions in the average NRS scores from baseline in the CRFA group were greater than those in the IAS group at all follow-ups (1 month: CRFA [n = 67], −4.2 ± 2.5, IAS [n = 69], −3.3 ± 2.3, P = 0.02; 3 months: CRFA [n = 65], −4.4 ± 2.3, IAS [n = 68], −1.9 ± 2.1, P < 0.0001; 6 months: CRFA [n = 58], −4.9 ± 2.4, IAS [n = 68], −1.3 ± 2.2, P < 0.0001) (paired Student t test).
Seventy-four percent (43/58) (95% CI, 62.9–85.4) of the CRFA group and 16% (11/68) (95% CI, 7.4–24.9) of the IAS group met successful outcome criteria (≥50% reduction in NRS score) at 6 months (P < 0.0001, χ2 test), whereas 25.9% (15/58) and 83.8% (57/68) of these study cohorts, respectively, were nonresponders. None of the subjects in the CRFA group reported worse pain at 6 months, whereas 15% (10/68) of the IAS cohort experienced exacerbation of usual knee pain during the follow-up period (P < 0.0024, χ2 test). Twenty-two percent (13/58) (95% CI, 11.4–33.5) of the CRFA group and 4% (3/68) (95% CI, 0–9.4) of the IAS group reported “no pain” (100% reduction) 6 months after treatment (P < 0.0026, χ2 test). In the CRFA group, 6.9% (4/58) of the population had no change in usual pain following the 6-month time point compared with 23.5% (16/68) of the IAS group (Fig. 4).
Oxford Knee Score
The mean OKS in each study cohort was equivalent at baseline (Table 2). Mean OKSs improved at all end points within both study groups (P < 0.0001 at each data point, paired Student t test) relative to baseline scores. The mean OKSs were greater in the CRFA group than in the IAS group at 1, 3, and 6 months (Table 2). Increasingly higher OKSs in the CRFA group were noted over the course of the 6-month time frame, whereas scores in the IAS group exhibited an opposite trend from 1 to 6 months. Index knee function improved in the CRFA group, whereas it declined in the IAS group from 1 to 6 months after treatment.
Interestingly, while the number of subjects in each study group considered to have “severe (knee) arthritis” symptoms were not different at baseline, beginning at 1 month, more subjects in the CRFA group than in the IAS group had “satisfactory joint function,” and at 3 and 6 months, more subjects in the CRFA had “mild to moderate arthritis” or “satisfactory joint function” than in the IAS group. These latter OKS classifications were not applicable to any study subjects at baseline (Table 3).
Global Perceived Effect
At 1 month, there was no difference (P = 0.1, χ2 test) in the proportions of subjects who reported perceived improvement as a consequence of treatment for chronic knee pain between the CRFA (79% [53/67]; 95% CI, 69.1–89.1) and IAS (67% [46/69]; 95% CI, 55.3–78.1) groups. In contrast, at the 3-month time point, a higher proportion of the CRFA group reported improvement (80% [52/65]; 95% CI, 70.0–90.0) compared with the IAS group (31% [21/68]; 95% CI, 19.6–42.1) (P < 0.0001). At 6 months, 91% (53/58) (95% CI, 83.9–98.8) of the CRFA cohort versus 24% (16/67) (95% CI, 13.4–34.4) of the IAS group (P < 0.0001) reported improved global perceived effect.
Pain Medication Use
At baseline, 19 subjects in the CRFA group (25% [19/76]; 95% CI, 15.0–35.0) and 26 subjects in the IAS group (35% [26/75]; 95% CI, 23.6–45.7) required opioid analgesic medication. The average total daily dose (morphine equivalence) was not different between study groups (CRFA, 28 ± 28.9 mg; IAS, 27.2 ± 22.1 mg; P = 0.75, Wilcoxon rank sum test). Mean opioid drug use at each time point was not different (P ≥ 0.17 at each data point, Student paired t test) than baseline within each group, and mean changes in doses used were not different between cohorts at each follow-up.
Thirty-three patients in the CRFA group (43% [33/76]; 95% CI, 32.0–54.8) and 34 patients in the IAS group (45% [34/75]; 95% CI, 33.8–56.9) required nonopioid medication at baseline. The daily average dose of nonopioid medications at each time point was not different (P ≥ 0.06 at each data point, Student paired t test) from baseline within each study group. However, there was a difference in mean nonopioid drug dose use between study groups at baseline (CRFA [n = 33], 899.5 ± 625.1 mg; IAS [n = 34], 500.3 ± 443.5 mg; P = 0.002, Wilcoxon rank sum test). Although changes in such drug use were not different between groups at 1 month (CRFA [n = 33], 0 ± 0 mg; IAS [n = 33], 94.8 ± 354.5 mg; P = 0.08) or 3 months (CRFA [n = 31], −16.1 ± 89.8 mg; IAS [n = 32], 64.7 ± 201.4 mg; P = 0.03), the average changes were different between cohorts at 6 months (CRFA [n = 29], −34.5 ± 128.9 mg; IAS [n = 32], 135.5 ± 391 mg; P = 0.02) (Wilcoxon rank sum test).
Subjects were screened for all AEs regardless of relatedness to disease state or procedure. There were 61 and 65 AEs reported among 34 and 30 subjects in the CRFA and IAS cohorts, respectively, with the majority of them in each group having an “unrelated” or “unlikely” relationship to study intervention (CRFA, 77% [47 events/61 total events]; IAS, 97% [63/65]). Three subjects in the CRFA group experienced 4 SAEs, whereas 7 subjects in the IAS group experienced 8 SAEs. Three (75%) of the 4 SAEs in the CRFA cohort involved the respiratory system: (1) exacerbation of asthma, (2) severe acute asthma, and (3) acute respiratory failure), and 1 (25%) involved urogenital function (pyelonephritis). The majority (50% [4/8]) of the 8 SAEs in the IAS group involved gastrointestinal function: (1) nausea and vomiting, (2) worsening of hiatal hernia, (3) gastric volvulus, and (4) abdominal pain secondary to small bowel obstruction, whereas 2 (25%) pertained to the cardiovascular system (heart attacks, 2 subjects), and 2 (25%) were categorized as “other”: (1) opioid overdose and (2) death. None of the SAEs were related to the study treatments.
Cooled radiofrequency ablation reduced index knee pain by at least 50% at 6 months in 74.1% of treated subjects compared with 16.2% in the IAS-treated subjects. Consistent with this, the LOCF data revealed that the number of subjects who experienced at least a 50% drop in pain at 6 months was higher in the CRFA group (67.2%; 95% CI, 55.9–78.4) than in the IAS group (15.7%; 95% CI, 7.2–24.2). The CRFA group consistently experienced greater pain relief throughout the study, with a mean NRS reduction of 4.9 compared with 1.3 in the IAS group at 6 months. A greater proportion of CRFA study subjects (60%) reported low-level residual knee pain (NRS score ≤2 at 6 months) compared with the IAS study group (7%).
At 6 months, a greater proportion of the CRFA group (40%) reported satisfactory knee function on the OKS compared with the IAS group (3%). Patient self-perceived overall health status (GPE) improved in 91% of the CRFA compared with 24% in the IAS group. A greater proportion of the CRFA cohort (36%) compared with the IAS cohort (3%) experienced both low-level (NRS score ≤2) index knee pain and concurrent satisfactory knee joint function (OKS ≥40) at 6 months. Despite statistically significant reduction in knee pain and disability, concomitant opioid analgesic use was not different between the 2 groups and remained similar to baseline use. Possible reasons for this include the fact that (1) the study protocol did not include a plan to wean opioids; (2) 43% and 60% of patients in the CRFA and IAS groups who were taking opioids as of the study's baseline assessment were using such medication for medical indications beyond OA-related knee pain; and (3) patients who were using opioids at baseline indicated that they had been taking that medication for a significant period of time (mean, 2.3 [SD, 3.3] and 2.4 [3.4] years, CRFA and IAS, respectively) before the study's inception. Thus, the study's ability to identify a decrease in opioid usage was limited given these findings. Further research in a patient population using opioids strictly for OA-related knee pain would eliminate the confounding nature of opioid use in this current study and may reveal opioid use changes that are commensurate with the respective effects of CRFA and IAS on pain in this study. In contrast, a reduction in nonopioid analgesic medication did occur coupled with improvements in pain and function.
The results of IAS injection seen in this research are consistent with those that previously cited the limited effectiveness of corticosteroid injections to elicit analgesia for those with knee OA.18 These results suggest that, compared with a single IAS, CRFA provides a clinically meaningful reduction in knee pain associated with improved knee function.
Cooled radiofrequency ablation is a safe, target-specific treatment that can be performed on an outpatient basis with minimal sedation required and in a short period (typically <45 minutes). As ablated peripheral nerves regenerate,19 knee pain and disability may reemerge. Longer-term (12-month) data are being gathered to assess the durability of the treatment effect beyond 6 months. As has been reported with medial branch neurotomy,20 repeat neuroablative procedures reinstate pain relief. Therefore, if the index knee pain eventually did return, repeating the CRFA procedure would be reasonable and sensible, especially if in the interim the patient enjoys reduction in pain, disability, and the need for less specific oral analgesics.
The limitations of this study include the following: the comparison group (IAS subjects) underwent a singular injection rather than multiple IAS injections, and the 6-month time point at which the primary outcome was assessed is not consistent with the expected duration of effectiveness of a steroid injection.18 The IAS injections are not truly a “control” intervention, given that corticosteroids are analgesics.18 This was an open-label trial, and so not all study site observers were blinded to procedures. To help mitigate this bias, questionnaires (ie, NRS, OKS, GPE) were self-administered by study subjects. Medication diaries were not used to record medication usage in this study, which introduced potential for error and/or inability to identify acute changes in medication dosage during the study. The effect of each treatment on opioid use for OA-related knee pain could not be specifically measured, because patients in both study groups used opioids for medical indications other than OA-related knee pain. Although there were no specific mobility instructions for the affected joint or a medication utilization protocol, the study outcomes were not influenced by these standardization absences.
Nonetheless, the findings of this study indicate that CRFA (Coolief) for genicular nerve ablation is superior to a single corticosteroid injection in osteoarthritic subjects for managing knee pain. Cooled radiofrequency ablation is a safe and effective nonopioid option for managing pain and improving physical function and quality of life (ie, based on OKS and GPE results) for patients with OA-related knee pain compared with IAS injection.
1. Zhang W, Doherty M, Peat G, et al. EULAR evidence-based recommendations for the diagnosis of knee osteoarthritis. Ann Rheum Dis
2. Arroll B, Goodyear-Smith F. Corticosteroid injections for osteoarthritis of the knee: meta-analysis. BMJ
3. Raunauld J, Buckland-Wright C, Ward R, et al. Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial. Arth Rheum
4. Franco CD, Buvanendran A, Petersohn JD, Menzies RD, Menzies LP. Innervation of the anterior capsule of the human knee: implications for radiofrequency ablation. Reg Anesth Pain Med
5. Patel N, Gross A, Brown L, Gekht G. A randomized, placebo-controlled study to assess the efficacy of lateral branch neurotomy for chronic sacroiliac joint pain. Pain Med
6. Kapural L, Vrooman B, Sarwar S, et al. A randomized, placebo-controlled trial of transdiscal radiofrequency biacuplasty for treatment of discogenic lower back pain. Pain Med
7. Desai MJ, Kapural L, Petersohn JD, et al. Twelve-month follow-up of a randomized clinical trial comparing intradiscal biacuplasty to conventional medical management for discogenic lumbar back pain. Pain Med
8. Choi WJ, Hwang SJ, Song JG, et al. Radiofrequency treatment relieves chronic knee osteoarthritis pain: a double-blind randomized controlled trial. Pain
9. Ikeuchi M, Ushida T, Izumi M, Tani T. Percutaneous radiofrequency treatment for refractory anteromedial pain of osteoarthritic knees. Pain Med
10. Sari S, Aydin ON, Turan Y, Özlülerden P, Efe U, Kurt Ömürlü İ. Which one is more effective for the clinical treatment of chronic pain in knee osteoarthritis: radiofrequency neurotomy of the genicular nerves or intra-articular injection? [published online ahead of print August 12, 2016]. In: Int J Rheum Dis
11. Santana Pineda MM, Vanlinthout LE, Moreno Martín A, van Zundert J, Rodriguez Huertas F, Novalbos Ruiz JP. Analgesic effect and functional improvement caused by radiofrequency treatment of genicular nerves in patients with advanced osteoarthritis of the knee until 1 year following treatment. Reg Anesth Pain Med
12. Ball RD. The science of conventional and water-cooled monopolar lumbar radiofrequency rhizotomy: an electrical engineering point of view. Pain Physician
13. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain
14. Dworkin RH, Turk DC, Wyrwich KW, et al. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain
15. Dawson J, Fitzpatrick R, Murray D, Carr A. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br
16. Kamper SJ, Ostelo RW, Knol DL, Maher CG, de Vet HC, Hancock MJ. Global Perceived Effect scales provided reliable assessments of health transition in people with musculoskeletal disorders, but ratings are strongly influenced by current status. J Clin Epidemiol
17. Smith MD, Wetherall M, Darby T, et al. A randomized placebo-controlled trial of arthroscopic lavage versus lavage plus intra-articular corticosteroids in the management of symptomatic osteoarthritis of the knee. Rheumatology (Oxford)
18. Hepper CT, Halvorson JJ, Duncan ST, Gregory AJ, Dunn WR, Spindler KP. The efficacy and duration of intra-articular corticosteroid injection for knee osteoarthritis: a systematic review of level I studies. J Am Acad Orthop Surg
19. Choi EJ, Choi YM, Jang EJ, Kim JY, Kim TK, Kim KH. Neural ablation and regeneration in pain practice. Korean J Pain
20. MacVicar J, Borowczyk JM, MacVicar AM, Loughnan BM, Bogduk N. Lumbar medial branch radiofrequency neurotomy in New Zealand. Pain Med