Sihvonen 2013.

Study characteristics
Methods	Study design: multicentre, randomised, parallel‐arm, double‐blind, placebo‐controlled trial Setting: five orthopedic clinics in Finland Trial time period: participants were enrolled between December 2007 and January 2013 Interventions: arthroscopic partial meniscectomy (APM) plus home exercises versus placebo surgery plus home exercises Sample size calculations: a total sample of 134 participants with 40, 54 and 40 participants per group for the Lysholm score, WOMET score and pain assessment, respectively, with 80% power to show a clinically meaningful advantage of APM over placebo, based on a two‐sided type 1 error rate of 5%. Anticipating a loss to follow‐up of at least 20%, the study planned to recruit 70 participants per group. Analysis: intention‐to‐treat
Participants	Number of participants Number of participants screened: 205 (45 were excluded before arthroscopy and 14 excluded after diagnostic arthroscopy) Number of participants at enrolment: 146 Number randomised: 146 participants: 70 to the arthroscopic partial meniscectomy group and 76 to the placebo surgery group Number included in analyses: 146: 70 in the arthroscopic partial meniscectomy group and 76 in the placebo surgery group (at 12 months ‐ primary time point) Criteria for defining study participants: those who have knee symptoms consistent with a degenerative medial meniscus tear and no knee osteoarthritis Inclusion criteria 35 to 65 years of age Persistent (> 3 months) pain on the medial joint line of the knee Pain that can be provoked by palpation or compression of the joint line or a positive McMurray sign Tear of the medial meniscus on MRI Degenerative injury to the medial meniscus confirmed at arthroscopy Exclusion criteria Acute, trauma‐induced onset of symptoms. Locking or painful snapping of the knee joint A surgical operation performed on the affected knee Osteoarthritis of the medial compartment of the knee (determined by clinical criteria of the Amercican College of Rheumatology) Osteoarthritis on knee radiographs (Kellgren‐Lawrence grade > 1) Acute (within the previous year) fractures of the knee Decreased range of motion of the knee Instability of the knee MRI assessment showing a tumour or any other complaint requiring surgical or other means of treatment Arthroscopic assessment showing anything other than a degenerative tear of the medial meniscus requiring surgical intervention Baseline characteristics Arthroscopic partial meniscectomy Mean (SD) age: 52 (7) years Number of men and women: M/F = 42/28 Mean (SD) body‐mass index: 26.9 (4) Median (range) duration of medial knee pain: 10 (3‐50) months Onset of symptoms ‐ no. (%) : gradual 48 (69) after exercise 12 (17) sudden 10 (14) Positive result of McMurray test — no. (%) 16 (23) Kellgren–Lawrence grade — no. (%): 0‐ 35 (50); 1 ‐ 35 (50) Symptoms of catching or locking no.(%): 32(46) Pain provoked by forced flexion, causing compression at the medial tibiofemoral joint line — no. (%): 50 (71) Pain provoked by palpation at the medial tibiofemoral joint line — no. (%): 63 (90) Mean (SD) Lysholm knee score: 60.2 (14.7) Mean (SD) WOMET score: 56.4 (17.3) Mean (SD) for knee pain (VAS): after exercise 5.8 (2.0) at rest 4.1 (2.3) Mean (SD) 15D score: 0.90 (0.06) Placebo surgery Mean (SD) age: 52 (7) years Number of men and women: M/F = 47/29 Mean (SD) body‐mass index: 27.9 (4) Median (range) duration of medial knee pain: 10 (3‐47) month Onset of symptoms ‐ no. (%) : gradual 48 (63) after exercise 14 (18) sudden 14 (18) Positive result of McMurray test — no. (%) 15 (20) Kellgren–Lawrence grade — no. (%): 0‐ 36 (47); 1 ‐ 40 (53) Symptoms of catching or locking no.(%): 37(49) Pain provoked by forced flexion, causing compression at the medial tibiofemoral joint line — no. (%): 59 (78) Pain provoked by palpation at the medial tibiofemoral joint line — no. (%): 74 (97) Mean (SD) Lysholm knee score: 60.1 (14.6) Mean (SD) WOMET score: 52.8 (18.1) Mean (SD) for knee pain (VAS): after exercise 6.1 (2.0) at rest 4.4 (2.4) Mean (SD) 15D score: 0.89 (0.06) Pre‐treatment group differences: there were no differences in the baseline characteristics between the two groups.
Interventions	Arthroscopic examination of the knee was first performed in all participants with the use of standard anterolateral and anteromedial portals and a 4‐mm arthroscope. The orthopedic surgeon evaluated the medial, lateral and patellofemoral joint compartments and graded the intra‐articular pathologic changes. Arthroscopic surgery Arthroscopic partial meniscectomy plus home exercises. The damaged and loose part of the meniscus tissue was removed with arthroscopic instruments (mechanised shaver and meniscal punches) until solid meniscus tissue was reached. The meniscus was then probed to ensure that all loose and weak fragments and unstable meniscus tissue had been successfully resected, preserving as much of the meniscus tissue as possible. Placebo surgery Placebo surgery plus home exercises. A standard arthroscopic partial meniscectomy procedure was simulated. The surgeon asked for all instruments and manipulated the knee as if an arthroscopic partial meniscectomy was being performed. The mechanised shaver (without the blade) was pushed firmly against the patella, outside of the knee, to mimic as closely as possible the feelings and sounds of the normal use of the arthroscopic shaver. Further, to simulate the sounds of normal arthroscopic partial meniscectomy, suction was also used to drain the joint and saline was splashed. The participant was kept in the operating room for the amount of time required to perform an actual arthroscopic partial meniscectomy. All procedures were standardised and recorded on video. In both arthroscopic partial meniscectomy and placebo surgery groups, the post‐operative care was delivered according to a standard protocol specifying that all participants received the same walking aids and graduated home exercise programme. Participants were instructed to take over‐the‐counter analgesic agents as required.
Outcomes	Outcomes were assessed at baseline and 2, 6, 12, 24, 36, 48 and 60 months' follow‐up. Primary outcomes Knee pain after exercise (during the preceding week) at 12 months (primary time point) after surgery using an 11‐point numerical rating scale ranging from 0 (no pain) to 10 (extreme pain) Lysholm knee score at 12 months (primary time point). Scores range from 0 to 100 with lower scores indicating more severe symptoms Western Ontario Meniscal Evaluation Tool (WOMET) score at 12 months (primary time point). WOMET is a meniscus‐specific health‐related quality‐of‐life instrument with scores ranging from 0 (worst possible situation) to 100 (best possible situation). Secondary outcomes Knee pain after exercise (during the preceding week) at 2, 6, 24 and 60 months after surgery using an 11‐point numerical rating scale (0 'no pain' to 10 'extreme pain') Lysholm knee score at 2, 6, 24 and 60 months after surgery (0 to 100, lower scores = more severe symptoms) WOMET score at 2, 6, 24 and 60 months after surgery (0 to 100, higher scores = better quality of life) Knee pain at rest at 12 months, using a numerical rating scale (0 'no pain' to 10 'extreme pain') 15D generic quality of life score at 12 months after surgery. Scores range from 1 (full health) to 0 (being dead) Serious adverse events at 12, 24 and 60 months, defined as untoward medical occurrences that may or may not have had a causal relationship with the treatment administered. Classified as serious if they necessitated hospitalisation or prolonged inpatient hospital care, or if they were life‐threatening or resulted in death Treatment success at 12, 24 and 60 months after surgery (measured by participant responses to 3 questions: 'Is your knee better than before the intervention?'; 'Are you satisfied with your knee at present?'; 'Would you choose to be operated on again if you were asked to make the decision now?'). Responses for the first 2 questions were rated on 5‐point Likert scales ('much better' to 'much worse'). 'Much better' and 'Better' were considered to indicate improvement, and 'Unchanged', 'Worse', and 'Much worse' were considered to indicate no improvement.) Return to previous activities at 24 and 60 months after surgery, rated as yes/no Need for subsequent knee surgery (additional arthroscopy, high tibial osteotomy or total knee replacement) at 24 and 60 months after surgery Clinical examination at 24 months after surgery including clinical meniscus test, McMurray test, pain provoked by joint line palpation, pain provoked by forced flexion and varus, range of motion, knee crepitus, bony enlargement, effusion, location of pain at palpation and knee stability at 24 months (these outcomes were reported in Sihvonen 2018 (a secondary publication of Sihvonen 2013) but not pre‐defined or reported in Sihvonen 2013. A cost‐utility analysis (based on the participants’ general quality of life using the 15D score and the utilisation of healthcare resources) at 12 months after surgery Progression of knee osteoarthritis at 5 years ‐ at least one grade progression in radiographic tibiofemoral knee OA on the KL classification (grade 1 (doubtful): doubtful joint space narrowing and possible osteophytic lipping; grade 2 (minimal): definite osteophytes and possible joint space narrowing; grade 3 (moderate): moderate multiple osteophytes, definite narrowing of joint space and some sclerosis and possible deformity of bone ends; grade 4 (severe): large osteophytes, marked narrowing of joint space, severe sclerosis and definite deformity of bone ends). Radiographic progression also assessed as being based on the sum of marginal tibiofemoral osteophyte grades and tibiofemoral joint space narrowing grades (according to the atlas developed by OARSI), ranging from 0‐18 Outcomes used in this review at 2, 6 and 12 months, and 2 and 5 years Pain ‐ pain after exercise on numerical rating scale at 2 and 6 months, and 2 and 5 years (0 to 10, lower score = less pain) Function ‐ Lysholm Knee Score at 2 and 6 months, and 2 and 5 years (0 to 100, higher score = better function) Generic health‐related quality of life ‐ 15D at 12 months (0 to 1, higher score = better) Participant reported treatment success at 5 years, measured as number of participants reporting 'much better' or 'better' for item 'Is your knee better than before the intervention?' Knee surgery (replacement or osteotomy) at 5 years Serious adverse events at 5 years Total adverse events at 5 years Progression of knee osteoarthritis at 5 years
Notes	Funding: funded by the Sigrid Juselius Foundation, the Competitive Research Fund of Pirkanmaa Hospital District, and the Academy of Finland. Trial registration: ClinicalTrials.gov number NCT00549172; and NCT01052233 trial number for 10‐year follow‐up which is ongoing Adverse events Arthroscopic partial meniscectomy: 7/70 (10%) participants from the arthroscopic surgery group had a serious knee‐related adverse event (3 knee replacement, 4 arthroscopies) and 1/70 (1%) participants from this group had other serious adverse events (1 deep infection of the index knee at 4 months) (8/70 (11.43%) participants) Placebo surgery: 8/76 (10.52%) participants from the placebo group had serious knee‐related adverse events (1 proximal tibial osteotomy, 7 arthroscopic partial meniscectomies) and no other serious adverse events in this group Knee surgery (replacement or osteotomy) (5‐year follow‐up) Arthroscopic partial meniscectomy: No. (%) of participants = 3/70 (4%) participants from the arthroscopic surgery group had a subsequent knee replacement Placebo surgery: 1/76 (1%) participants from the placebo surgery group had a subsequent high tibial osteotomy [trial authors reported 3/68 (4%) in the arthroscopy group and 1/74 (1%) in the placebo group at 5 years. Number receiving allocated intervention was used as denominator in meta‐analysis] Progression of knee OA At 5 years, 44/74 (59.5%) participants from the placebo group and 48/67 (71.6%) participants from the arthroscopy group had at least one grade progression in radiographic tibiofemoral knee OA on the Kellgren‐Lawrence classification. Withdrawals At 12 months, there was no loss to follow‐up and all randomised participants completed the study. Lysholm Knee Score data for one participant in the placebo surgery group were missing at 6 months' follow‐up (no reason given) and data were not imputed. At 24 months, 2 participants from the placebo group were lost to follow‐up (one not responding to contact attempts and one deceased). At 60 months, 4 participants (2 placebo and 2 arthroscopic surgery) were lost to follow‐up (two not responding and 2 deceased ‐ reasons not given per group)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Stratified randomisation was performed by a statistician using a computer‐generated schedule
Allocation concealment (selection bias)	Low risk	Sequentially‐numbered, opaque, sealed envelopes were prepared by a statistician with no involvement in the clinical care of participants in the trial
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The operating surgeon and other staff in the operating room were aware of group assignment before the procedure but did not participate in further treatment or follow‐up of participants. Participants were blinded. At 12 months, 2/70 (3%) participants in the arthroscopic surgery group and 5/76 (7%) participants in the placebo surgery group reported persistent symptoms after surgery that were sufficiently severe to lead to revealing of the study‐group assignment at an average of 8 months after surgery. At 24 months, 5/70 (7%) in the arthroscopic surgery group and 7/74 (9%) participants in the placebo group reported symptoms so severe to lead to unblinding. At 60 months, 8/68 (12%) in the arthroscopic surgery group and 8/74 (11%) participants in the placebo group reported symptoms so severe to lead to unblinding
Blinding of outcome assessor Self‐reported outcomes	Low risk	Participants were blinded to the group assignment. Participants in the placebo surgery group were not significantly more likely than participants in arthroscopic surgery group to guess that they had undergone a placebo procedure (47% and 38% respectively, P = 0.39)
Blinding of outcome assessor Assessor‐reported outcome (knee replacement)	Low risk	Outcome assessors were blinded; low risk of bias for assessment of knee replacement
Incomplete outcome data (attrition bias) All outcomes	Low risk	At 12 months, there was no loss to follow‐up and all randomised participants completed the study. Lysholm Knee Score data for one participant in the placebo surgery group were missing at 6 months' follow‐up (no reason given) and data were not imputed. However, it is unlikely that this would have had a clinically important impact on the observed effect size. At 24 months, 2 participants from the placebo group were lost to follow‐up (one not responding to contact attempts and one deceased). At 60 months, 4 participants (2 placebo and 2 arthroscopic surgery) were lost to follow‐up (two not responding and 2 deceased ‐ reasons not given per group)
Selective reporting (reporting bias)	Unclear risk	One primary outcome (WOMET at 12 months) was not pre‐specified and was added after data collection but before data analysis (trial registration was amended and revised protocol published). The rationale for adding this primary outcome was provided (the score was validated in the participant population). Two other pre‐specified outcomes were changed: (1) pain at rest at 12 months was pre‐specified as a primary outcome but moved to a secondary outcome after data collection but before data analysis. The rationale was that reporting both pain at rest and after exercise was 'somewhat ambiguous' and that pain after exercise was the more important of the two; (2) secondary cost‐utility analysis based on 15D score and healthcare resource utilisation at 12 months was removed before data analysis (no reason given). The 2‐year follow‐up paper (Sihvonen 2018) reported the primary time point as 24 months (but protocol states primary time point as 12 months). A new statistical analysis plan for 5‐ and 10‐year follow‐up was published in 2020. Results on revised primary and secondary outcomes were reported
Other bias	Low risk	No other bias apparent