Skip to main content
. 2022 Mar 3;2022(3):CD014328. doi: 10.1002/14651858.CD014328

Roos 2018.

Study characteristics
Methods Study design: multicentre, prospective, double‐blind, randomised, placebo‐controlled trial
Setting: outpatient departments of the orthopaedic clinics in Region Zealand; Slagelse Hospital; Næstved Hospital, Denmark
Trial time period: 21 February 2011 to March 2015
Interventions: arthroscopic surgery versus placebo surgery
Sample size calculations: 36 individuals per group would be required to obtain a power of at least 80% to detect a minimal important change (MIC) of 10 KOOS 5 score units, assuming a common standard deviation of 15. The study team decided to include 80 individuals in total (40 participants in each group), allowing for a 10% dropout rate.
Analysis: intention‐to‐treat analysis was planned and both a per‐protocol and ITT analysis (using best observation carried forward for missing data) were executed.
Participants Number of participants

  • Number screened: 586 (351 not eligible, 135 declined to participate)

  • Number of participants at enrolment: 100 (32 showed no meniscal lesion in MRI, 24 declined participation)

  • Number randomised: 44: 22 in arthroscopic surgery group and 22 in the placebo surgery group, 8 participants crossed over from the placebo surgery group to the arthroscopic surgery group

  • Number included in analyses: 42 (21 in each group at 3 months); 22 in the arthroscopic surgery group and 20 in the placebo surgery group at 24 months


Inclusion criteria

  • Participants with knee pain for more than 2 months without significant trauma (only the index knee included)

  • MRI‐confirmed medial meniscus lesion (participants with a medial meniscus lesion and a lateral lesion also included)

  • Age 35 to 55 years

  • Eligible for outpatient surgery


Exclusion criteria

  • Participants in need of acute surgery e.g. locking knees, high‐energy trauma

  • Symptoms associated with other musculoskeletal comorbidities overriding the symptoms of the knee

  • Grade 3 or 4 knee OA on the Kellgren‐Lawrence classification

  • Knee surgery within the last 2 years

  • Obese participants with BMI > 35

  • Participants with ischaemic heart disease

  • Participants with diabetic late‐complications

  • Participants with thrombophilia

  • Pregnancy

  • Participants unable to speak Danish

  • Participants with drug or alcohol abuse


Baseline characteristics
Arthroscopic partial meniscectomy (n = 22)
Mean (SD) age: 47.2 (5.9)
No. (%) male/female: 13/9 (59/41)
Mean (SD) BMI: 27.6(3.6)
No. (%) Kellgren‐Lawrence grading: 0 ‐ 9 (41), 1 ‐ 9 (41), 2 ‐ 4 (18), 3 ‐ 0, 4 ‐ 0
No. (%) joint line tenderness: 21 (100)
No. (%) positive McMurray test: 17 (81)
No. (%) swelling present: 11 (52)
No. (%) full knee extension: 18 (86)
No. (%) small extension deficit < 10o: 3 (14)
No. (%) full knee flexion: 17 (81)
No. (%) small flexion deficit < 10o: 3 (14)
No. (%) large flexion deficit: 1 (5)
Median (IQR) duration of pain (months): 5 (2‐6)
Mean (SD) KOOS5 (0–100 worst to best): 51.2 (15.6)
Mean (SD) KOOS Pain (0–100): 55.1 (15.4)
Mean (SD) KOOS Symptoms (0–100): 62.8 (17.7)
Mean (SD) KOOS Function in daily living (0–100): 64.9 (19.9)
Mean (SD) KOOS Function in sport and recreation (0–100): 35.0 (23.0)
Mean (SD) KOOS Knee‐related quality of life (0–100): 38.7 (15.4)
Mean (SD) EQ‐5D VAS score: 69 (14)
Mean (SD) EQ‐5D 3L index value (0–1): 0.749 (0.108)
Mean (SD) SF‐36 Physical Component Summary (0–100): 38 (10)
Mean (SD) SF‐36 Mental Component Summary (0–100): 59 (7)
Placebo surgery (n = 22)
Mean (SD) age: 46.4 (5.5)
No. (%) male/female: 10/12 (45/55)
Mean (SD) BMI: 26 (3.9)
No. (%) Kellgren‐Lawrence grading: 0 ‐ 10 (45), 1 ‐ 8 (36), 2 ‐ 4 (18), 3 ‐ 0, 4 ‐ 0
No. (%) joint line tenderness: 20 (91)
No. (%) positive McMurray test: 17 (77)
No. (%) swelling present: 8 (36)
No. (%) full knee extension: 21 (95)
No. (%) small extension deficit < 10o: 1 (5)
No. (%) full knee flexion: 14 (64)
No. (%) small flexion deficit < 10o: 6(27)
No. (%) large flexion deficit: 2 (9)
Median (IQR) duration of pain (months): 3.5 (2.0 – 6.0)
Mean (SD) KOOS5 (0–100): 44.8 (19.9)
Mean (SD) KOOS Pain (0–100): 45.9 (22.0)
Mean (SD) KOOS Symptoms (0–100): 59.9 (20.6)
Mean (SD) KOOS Function in daily living (0–100): 56.5 (22.3)
Mean (SD) KOOS Function in sport and recreation (0–100): 25.2 (26.3)
Mean (SD) KOOS Knee‐related quality of life (0–100): 36.6 (20.2)
Mean (SD) EQ‐5D VAS score: 63 (SD not reported)
Mean (SD) EQ‐5D 3L index value (0–1): 0.642 (SD not reported)
Mean (SD) SF‐36 Physical component summary (0–100): 35 (SD not reported)
Mean (SD) SF‐36 Mental component summary (0–100): 57 (SD not reported)
Pre‐treatment group differences: the KOOS scores (KOOS5 and subscales) were higher in the arthroscopic group compared to the placebo surgery group.
Interventions Arthroscopic surgery
The arthroscopic partial meniscectomy was performed on an outpatient basis under general anaesthesia combined with local anaesthesia. Arthroscopic surgery was performed by experienced surgeons in their final year of residency or attending orthopaedic surgeons. Two standard portals on the lateral and medial sides of the ligamentum patella were created but no outflow cannula inserted. An arthroscope was used with a pressure‐controlled irrigation system. Tourniquet use was at the discretion of the surgeon. The strategy for the meniscectomy was to preserve as much tissue as possible. A standard operation protocol was used to document possible findings in cartilage, ligaments, synovium and the medial and lateral menisci. The type, and extent of meniscus lesion was registered and changes in the articular cartilage was classified according to the ICRS (International Cartilage Repair Society) classification.
Placebo surgery
The placebo procedure (skin incision only) was performed under the same conditions as the arthroscopic surgery. Participants were fully sedated with general anaesthesia. Local anaesthetic was applied and two skin incisions made at the same locations and of the same size as in the arthroscopic surgery group. Then the knee was manipulated as if a real arthroscopy was performed, the spillage of water and all other equipment needed for an arthroscopy was used. No instruments entered the arthroscopy portals to avoid the possibility of deep infection, osteochondral lesions or unwanted interventions by the surgeon. A pre‐recorded video of a standard arthroscopic partial meniscectomy was planned to be played during the placebo procedure (but was not done).
All participants in both intervention groups were given a folder including an exercise program for post‐operative participants after knee arthroscopy or placebo procedure. The exercise program included 7 different non‐weight‐bearing exercises to improve lower extremity function and knee range of motion (for the first week after surgery) and a further 3 weight‐bearing exercises thereafter. The exercises were for the participants to carry out at home and were recommended to be performed 10 to 15 times three times daily. The participants were also advised to start unloaded cycling, swimming or walking after 1 week, and jogging or loaded cycling after 2 to 3 weeks.
Outcomes Outcomes were measured at baseline, 3 and 24 months.
Primary outcome

  • KOOS5 composite score at 24 months. KOOS5 is derived from the Knee injury and Osteoarthritis Outcome Score (KOOS) and is calculated as a mean of the 5 subscale scores: (KOOS pain + KOOS symptoms + KOOS ADL + KOOS sport and rec + KOOS QoL / 5). Scores for each subscale range from 0 (extreme symptoms) to 100 (no symptoms)


Secondary outcomes

  • KOOS individual subscales ‐ pain, symptoms, ADL, sports/rec, QoL. Scores for each subscale range from 0 (extreme symptoms) to 100 (no symptoms)

  • Global Perceived Effect score ‐ a rating of overall improvement in knee symptoms after the operation (7‐point scale ranging from 'much worse' to 'much better'). A clinically important change is considered an improvement or worsening of at least 2 steps on the scale

  • SF‐36 Physical Component Summary (PCS) and Mental Component Summary (MCS) (0 to 100, higher = better)

  • EQ‐5D generic quality of life using 2 scales: EQ‐5D VAS global assessment of disease status scale (0 to 100, higher = better) and EQ‐5D 3L descriptive system (0‐1, higher = better; contains 5 domains: mobility, self‐care, usual activities, pain/discomfort, anxiety/depression; each rated on 3‐point scale: no problems, some problems, severe problems)

  • Single leg hop test (best of 3 trials, measured in centimetres)

  • Knee‐bend test (maximum number achieved in 30 seconds)

  • Isometric knee extension strength (highest of 3 contractions) measured sitting using a dynamometer

  • Radiography of knees to assess possible onset or progress of knee OA (score assigned based on joint space width and presence of osteophytes using standard atlas), at baseline and 5 years

  • Self‐efficacy using the modified Danish Arthritis Self‐Efficacy Scale

  • Participant expectations (single item)

  • Adverse events (e.g. superficial infection, nerve or vessel injury, deep infection, compartment syndrome, DVT, MI, stroke and death)


Outcomes used in this review at 3 and 24 months

  • Pain ‐ KOOS pain subscale (0 to 100, higher score = less pain) (data supplied by trial authors) (we multiplied the mean values by –1 so that lower score = less pain, as per Cochrane Handbook guidance)

  • Function ‐ KOOS ADL subscale (0 to 100, higher score = better function)

  • Knee‐specific health‐related quality of life ‐ KOOS QoL subscale (0 to 100, higher score = better QoL)

  • Generic health‐related quality of life ‐ SF‐36 MCS (0 to 100, higher = better QoL)

  • Participant reported treatment success ‐ Global perceived effect (rating of 'better' or 'much better') at 24 months (data supplied by trial authors)

  • Total adverse events at 24 months


Serious adverse events could not be included as the details of the group to which some serious adverse events belonged were not reported.
Notes Funding: University of Southern Denmark, Odense; the Orthopedic Departments of Slagelse and Næstved Hospital; the Research Unit of Hospital South, Region Zealand; Edith and Henrik Henriksens Memorial Fund; Region Zealand Health Scientific Research Fund; Research Fund of Hospital South
Clinical trial registration: NCT01264991
Adverse events: 4/22 knee‐related adverse events in the surgery group, of which two were serious (two re‐arthroscopies comprising one partial meniscectomy and one anterior cruciate ligament reconstruction), two were not serious (one cutaneous nerve lesion and one mild knee swelling). No reported knee‐related adverse events reported in the placebo group.
Total other adverse events ‐ 7 adverse events in 5 participants (2 participants in the surgery group and 3 participants from placebo group). The nature of events included: chest pain, finger injury, nausea, dizziness and kidney stone, and included two regarded as serious (abdominal surgery and malignant melanoma). Details of the group to which some serious adverse events belonged were not reported.
Knee surgery (replacement or osteotomy): none
Progression of knee OA: not reported
Withdrawals: 1/22 in the arthroscopy group and 1/22 in the placebo group at 3 months, and 2/22 participants in the placebo group only at 24 months
Post‐protocol changes: a pre‐recorded video of a standard arthroscopic partial meniscectomy was planned to be played during the procedure for both the surgery and placebo group, but was not performed in the placebo group.
Statistical power: trial was underpowered due to inability to recruit 40 participants per group as planned
Data analysis: SMD for generic quality of life at > 6 months up to 2 years (Analysis 1.4) back‐translated to SF‐36 PCS by multiplying the SMD by the standard deviation at baseline in the surgery group (as standard deviation in the placebo group was not reported) (SD = 10)
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Used a computer‐generated table of random numbers, prepared by an external co‐investigator
Allocation concealment (selection bias) Low risk Used consecutively‐numbered, sealed envelopes stored in a briefcase outside the operating theatre
Blinding of participants and personnel (performance bias)
All outcomes Low risk Participants were blinded to group allocation at short‐term follow‐up, there was low risk of bias at primary time point. Blinding broken for 16 participants (36%; 6/22 from arthroscopy group and 10/22 from placebo group) prior to 2‐year follow‐up. Prior to 3 months, two additional participants from the arthroscopy group (due to adverse events) and one from the placebo group (persisting pain) were unblinded by the treating surgeon. Between 3 and 24 months, two from the arthroscopy group and eight from the placebo group were unblinded by the treating surgeon because of persisting pain. In total, 10/22 in the arthroscopy group and 19/22 in the placebo group were unblinded. Operating surgeons and theatre staff were not blinded.
Blinding of outcome assessor
Self‐reported outcomes Low risk Low risk at short‐term follow‐up, unclear if knowledge of treatment in 6/22 and 10/22 influenced assessment of pain and function at 2 years
Blinding of outcome assessor
Assessor‐reported outcome (knee replacement) Low risk Outcome assessors were blinded to group allocation.
Incomplete outcome data (attrition bias)
All outcomes Low risk Self‐reported outcome data were missing for 2/22 participants in the placebo group at baseline. Data missing for 1/22 in the arthroscopy group and 1/22 in the placebo group at 3 months. 2/22 participants in the placebo group were lost to follow‐up at 2 years (crossed over to arthroscopic surgery between 3 and 24 months)
Selective reporting (reporting bias) Low risk 3‐month follow‐up data were provided upon request from the authors
Other bias Low risk No other bias apparent