S1 low risk = a,b
S2 na= not applicable, low risk = a S3 low risk = a
S4 low risk = a
C1 na= not applicable, low risk = a O1 low risk = a
O2 low risk = a O3 low risk = a,b
Outcome in all analysis is (men_rpts) adjusted for n (men_rptstot)
Main analysis:
Analysis 1:
Metaprop på alle
metaprop men_rpts men_rptstot, ftt random cimethod(score) label(namevar=auth, yearvar=year)
Analysis 2:
Metaprop kategorisert per follow-up kategori (fut_cat)
Sensitivity analysis Metaprop
Meniscal detection method, other ACL surgeries than ACL reconstruction, High risk of bias (S4, O1), meniscal repair at baseline (+ Davis (largest study) post hoc decision)
Subgroup analysis:
Analysis 3:
melogit
Children compared with adults (ped_ad) adjusted for follow-up time (fut) Remove mixed age studies.
Analysis 4:
melogit
Non-operated patients compered to ACL-reconstructed (tx) adjusted for follow-up time (fut).
Studies with mixed treatments removed
Metaregression Analysis 5:
Melogit
Meniscal injuries adjusted for follow-up time
GRADE Judgement rational for certainty of evidence
Grading of evidence:
For evaluating a risk or prognosis, prospective longitudinal studies (observational studies) provide the highest level of evidence1. However, the majority of included studies (67%) in this systematic review were retrospective case series. Therefore, we found it reasonable to downgrade one level to moderate certainty. (The GRADE evidence profile is displayed in table 2)
In our risk of bias assessment, we found that 68-84% of studies had high risk of bias
regarding the following domains: selection of cohort, demonstrating that the outcome was not present at the start of the study and assessment of the outcome. These aspects may directly under-or over-estimate the new meniscal injury rate, and therefore, indicate downgrading the certainty one level to low (table 2).
There were not consistent findings in the included studies. The I2 statistic was very high >
93%, but we have to be aware of the limitations in using the I2 statistic when there are large sample sizes and narrow overall confidence interval2. However, the individual study
confidence intervals did not overlap and suggest high inconsistency and warrant downgrading the certainty one level to very low (table 2).
The confidence interval of the overall point estimate is narrow (0.05 to 0.10). It is not likely that the treatment would change for the patient if either the lower or the upper limit represented the true estimate. The GRADE rule is to downgrade the certainty if patient treatment is altered by choosing the lower or upper limit of the confidence interval1. As a result, we did not find reasons to downgrade the certainty based on imprecision.
There are few studies published on active rehabilitation and also relatively few studies with long follow-up time. Furthermore, the outcome of interest in this systematic review, is not a common primary outcome. Assessment of the outcome is not appropriate in most studies.
Thus, the rate of new meniscal injuries in the included studies is not likely to adequately reflect the total picture of new meniscal injury rates after ACL injury. This indirectness, indicates downgrading the certainty one level, but our certainty is already downgraded to very low and further downgrading is not possible.
In conclusion, the certainty of the body of evidence regarding risk of new meniscal tears after ACL injury treatment is very low due to the GRADE assessment discussed above.
1. Iorio A, Spencer FA, Falavigna M, et al. Use of GRADE for assessment of evidence about prognosis:
rating confidence in estimates of event rates in broad categories of patients. BMJ 2015;350:h870. doi: 10.1136/bmj.h870 [published Online First: 2015/03/18]
2. Ferretti A, Monaco E, Ponzo A, et al. Combined Intra-articular and Extra-articular Reconstruction in Anterior Cruciate Ligament-Deficient Knee: 25 Years Later. Arthroscopy - Journal of
Arthroscopic and Related Surgery 2016;32(10):2039-47. doi:
http://dx.doi.org/10.1016/j.arthro.2016.02.006
Details Population Exposure Outcome
JJarvala 2001 PinczewskiSTG 2007 Reid 1992
Risberg 1999 Salmon 2006 Seitz 1996
Shelbournecontrols 2015 Shirakura 1995
YYoo 22009 Wall 2017 W
Willimon 2015
RCS RCS RCS
31 21 21
A P P
29(18-47) 11(9-15) 12(10-14)
13%
15%
0
ACL-Rmmr ACL-Rall-epi ACL-RITB-Mic
patientsi patientss patientss
4.2(1-11) 3.6(2-6.6) 3(1-6.9)
Supplementary file z Study characteristics, setting and outcome assessment for individual studies included for qualitative analysis.
Abbreviations and expainations:
Design
qRCT quasi randomized controlled trial (ie if randomized based on birth date or timepoint for inclusion)
RCT randomized controlled trial RCS retrospective case series PCS prospective case series
CS? Case series, unclear if prospective or retrospective RCCS retrospective case control study
PCCS prospective case control study RMCS retrospective matched case study PMCo prospective matched cohort PCo prospective cohort
Population
PopI population category at inclusion
P pediatric (age under 16 years in all patients or described as skeletally mature at inclusion) A adult (age over 15 years in all patients or specified that all patients were skeletally mature at inclusion
M mixed (both patients age under and 16 years old and over at inclusion and not specified wether the patients were skeletally mature or not )
Age
Age (mean, min-max or ± SD) at inclusion ( same as surgery in some surgical cohorts) if not stated otherwise.
AgeFU indicate age at follow-up (some studies report that and not age at inclusion) Age? Age is reported, but at unclear timepoint.
- Age is not reported
med indicate that the age was reported as median instead of mean
CI indicate that age was reported in confidence intervals rather than min-max
icr indicate that the age was reported in interquartile range rather than min-max.
Treatment
ACL-S is ACL surgery other than ACL-reconstruction ie repair ACL-R is ACL reconstruction
ST/G indicate semitendinosus and gracilis graft, PT patella tendon graft
SB or DB indicate single bundle or double bundle technique
SI or DI indicate single or double incision technique
all-epi indicate the physealsparing technique all-epiphyseal ACL-reconstruction, transphy indicate transphyseal technique, hybrid indicate a hybrid (physealsparing + transphyseal) technique and
ITB_Mic indicate the physealsparing extra articular technique described by Micheli.
mmr indicate that all patients had a medial meniscal repair
PT+ rem indicate use of patellar tendon graft combined with leaving the ACL remnant
M-repairmixed indicate that all patients had meniscal repair and that the treatment of the ACL was mixed; non-operative or ACL-reconstruction
Mixed indicate that some patients had non-operative treatment and some had ACL reconstruction. In the majority the study started out with non-operative and a proportion had delayed ACL-R.
Non-opb surg indicate that all the patients in the study were evaluated prior to having surgery
Reporting
Reporting indicates how new meniscal injuries were reported in the study
Patientss indicates that the study reported the number of patients with new injuries that had surgery Patientsi indicates that the study reported the number of patients with new injuries (unregardless of having surgery or not)
Injuries s the number of injuries that had surgery, patients are not reported and we are not able to calculate the number of patients who had these injuries
Injuriesi indicate the number of injuries unregardless if they had surgery or not, as above patients with new injuries are not reported.
Follow-up time
Years follow up, mean, min-max or ±SD. In some patients only min or up to a certain timepoint is reported.
– means it is not reported.
Follow-upmedian indicate that median follow up time was reported
Neusel: Only 24 of 35 patients in the study was included for qualitative analysis because 11 patients had partial tears and were removed.
Selection/representativeness Comparability Outcome
Jarvala -07
Wall Willimon
❶
❶
na na
⓿
⓿
❶
❶
na na
❸
❸
⓿
⓿
❸
❸
Supplementary file 7 Risk of bias of individual studies included for qualitative analysis by Newcastle Ottawa Scale domains (dichotomized to low/high or unclear).
❶High risk of bias: red ball marked with number 1
⓿Low risk of bias: green ball marked with number 0
❸Unclear risk of bias (not described): yellow ball marked with number 3 Studies also included for qualitative analysis are marked in bold letters.
Grindem -14; only the non-operated arm of the study is included for qualitative analysis, because the diagnosis in the operated arm of the study was clinical.
Shelbourne; only the controls in the study were included for qualitative and quantitative analysis, because the number of new injuries in the cases were unclear.
Subgroup No. studies Effect estimate
<10 years FU 1 0.31
Subgroup numbers of studies, point estimate, CI, I2
0.05.1.15.2se(ES)
-.4 -.2 0 .2 .4 .6
ES