Letter to the Editor: Editorial: The New AAOS Guidelines on Knee Arthroscopy for Degenerative Meniscus Tears Are a Step in the Wrong Direction

To the Editor,

The American Academy of Orthopaedic Surgeons (AAOS) has developed and funds a considerable quality program managed by its Evidence-based Quality and Value Committee. The AAOS is invested in this process and takes the effort very seriously. The output of the committee includes various evidence-based guidelines that are meant to aid member surgeons in choosing the best treatment for their patients.

The clinical practice guidelines (CPGs) produced by the committee represent the latest in the organization’s evidence-based treatment recommendations. It would be helpful to clinicians if there were one respected entity whose process and product documents were recognized as the last word in musculoskeletal treatment recommendations. The Evidence-based Quality and Value Committee process, however, has been criticized on its methodology, transparency, and real-world applicability [4, 11, 13]. It has been cited as the cause of limited patient access to safe and effective treatments for osteoarthritis of the knee [4, 12]. It has been and will remain controversial because most of the evidence supporting orthopaedic surgical treatments is lower on the Level of Evidence hierarchy and therefore is excluded from consideration from the CPG process when one or more high-level studies are available for review on a specific topic.

However, even high-level surgical studies have major methodologic flaws. Simply stated, it is difficult to transport evidence of the same rigor from pharmacologic to surgical research; issues of clinical equipoise, difficulty in blinding, and one-way crossover from nonsurgical to surgical treatments all complicate that effort [10, 13, 17]. Because of this, the science is not settled; CPGs should serve as an educational tool and guideline for surgeons and nothing else. Surgeons should still use their judgement and experience when determining which patients might undergo surgery. Surgeon experience is one of the three tenets of evidence-based medicine [14]. It is vital that the guidelines are used by clinicians and not by third parties to limit access to care. Such use defeats their value as a scientific and educational pursuit, and the AAOS recognizes this in a disclaimer that accompanies all clinical practice guidelines.

The latest iteration of the CPG for osteoarthritis of the knee, we think, represents an improvement in the process by recognizing the strengths and limitations of the current state of high-level orthopaedic research and the imperative to provide real-world guidance to practicing surgeons [1]. That guidance must consider that osteoarthritis is a debilitating and chronic disease with no cure and only limited and temporizing treatments.

Added to the list of criticisms of the CPG process is a recent editorial by Leopold [8]. He found great fault and error in the latest CPG on the treatment of knee osteoarthritis. Specifically, he lamented the recommendation on arthroscopic partial meniscectomy in patients with osteoarthritis. The recommendation states: “Arthroscopic partial meniscectomy can be used for the treatment of meniscal tears in patients with concomitant mild to moderate osteoarthritis who have failed physical therapy or other nonsurgical treatments” [1]. Dr. Leopold claimed that the latest iteration of the CPG disregards the latest high-level evidence and cited several well-designed studies the committee considered.

The editorial discussed the outcomes of various studies using sham surgical interventions. Placebo interventions are known to exhibit treatment effect sizes that can confound the results of Level I studies, and sham surgical procedures most likely have this characteristic [2, 3, 7, 16, 18]. Therefore, high-level studies that use such placebo and sham treatments in their design must be analyzed with potential bias in mind. The highly referenced study by Sihvonen et al. [15] took 5 years to identify 205 patients who met the inclusion criteria of the study and excluded patients with trauma-induced onset of symptoms or a recent history of a locked knee. Are these patients truly representative of all patients with knee pain and mild to moderate knee osteoarthritis, or does potential selection bias limit the applicability of these and other Level I studies to all patients with the disease? Additionally, a psychiatric study has demonstrated that the placebo effect can grow with increases in the patient’s “perceived invasiveness” [5]. Therefore, sham surgical interventions may place the surgical treatment being studied at a disadvantage compared with less-invasive placebo-controlled studies used for pharmaceutical research. This critical point can be highlighted by an underreported subset of patients involved in sham control studies. In a 2018 study [9], the authors reported outcomes for a subgroup of 22 patients who opted out of their sham surgical trial and elected to undergo arthroscopic treatment. The authors reported the change in Western Ontario Meniscal Evaluation Tool score was greater for those who declined participation (43.2 ± 22.4) than for those who were randomized (29.5 ± 21.1), with a between-group difference of -13.7 (95% confidence interval, -25.6 to -2.9). In other words, non-participants receiving standard treatment in a real-world clinical setting did better than patients in both study arms [9].

Another issue in high-level surgical studies is crossover, in which patients are removed from one treatment group and enrolled in another. Randomized controlled trials attempt to mitigate this problem by performing intention to treat analyses. Intent to treat dictates that patients are included in their original group designation even though they received the comparison treatment. Crossover can confound results if there is a greater percentage of crossover in one group than in another or if crossover is one way. One-way crossover presents a real problem in the design of randomized controlled trials for surgery because there is no going back when leaving a nonoperative group for an operative one. The crossover percentages in recent studies on surgical meniscectomy varied from 19% to 33%, a significant confounder and source of potential bias [10, 17]. In a recent analysis of the Level I evidence regarding knee osteoarthritis and arthroscopic partial meniscectomy, Forbes et al. [6] found that “the extensive crossovers from one group to another diminish confidence in the precision of the results” in Level I studies. Finally, only two of the randomized controlled trials had a defined and mandatory prerandomization period of nonoperative care. Because patients with knee osteoarthritis and meniscal tearing in the real world undergo such a trial of care, Lieb et al. [9] have contended that these studies lack external validity; they don’t address situations that occur in day-to-day clinical practice, again limiting applicability to clinical situations encountered by surgeons in practice.

Perhaps the most important point in considering high-level studies on arthroscopic partial meniscectomy is that all studies reported clinically meaningful improvement, with non-inferiority to the comparison groups. They are of equal efficacy to arthroscopic partial meniscectomy by outcome metrics. To contend that the science maintains that meniscectomy is never indicated in patients with mild to moderate knee osteoarthritis is not supported, because these patients experienced equal efficacy in the studies used in the CPG process. Other treatment factors should be included in the decision to perform surgery, such as cost and risk, but these should not prohibit patients from accessing surgical interventions, especially when less-invasive treatment options have not worked. Fundamental to the debate is how to define the subset of patients who will benefit from arthroscopic partial meniscectomy? In the outcomes of any clinical study, there are outliers who defy statistical means and respond exceptionally well or poorly to the treatments studied. Research focused on the outliers of treatment would contribute to a useful and pragmatic guideline.

We think that the treating orthopaedic surgeon is still in the best position to consider the evidence and patient characteristics and expectations to arrive at a reasonable course of therapy. The CPG on arthroscopic partial meniscectomy accurately reflects the current research and supports physician education and autonomous decision-making. Leopold [8] implied that the updated CPG is recommending meniscectomy as a routine treatment for patients with osteoarthritis of the knee. A more accurate reading of the guideline is that mild to moderate knee osteoarthritis is not a contraindication to arthroscopic partial meniscectomy for patients who have not responded to nonoperative modalities. This was a critical question overlooked by the 2013 guidelines. In 2021, the guidelines addressed this question, which we think is a step in the right direction. It is a guideline that clinicians can use to the benefit of their patients.

Had the guideline not recognized this important concept, we expect that third-party interest would have used similarly flawed reasoning to identify osteoarthritis as an absolute contraindication to arthroscopic surgical treatments. Perhaps Dr. Leopold would have preferred this conclusion. If the CPGs were purely used for education, we would not have to work so hard to debate the nuances of the science and attributes of clinical practice. The reality is that the guidelines exist where insurers will use any means possible to deny clinical care in the name of cost containment. The AAOS should support the autonomy of its members by identifying clinical decision-making as the surgeon’s scope of practice, and we believe our patients would prefer we work to maintain this scope of practice.