To The Editor:

Grades of recommendation, introduced to readers of The Journal in September 2005, (1), are a method of summarizing the quality of the literature. Grades of Recommendation are used in practice guidelines and in review articles to summarize the evidence on clinical questions.

In response to concerns raised by some that the designation of a Grade-C recommendation was not sufficiently clear, we revised the definition to read: “poor-quality evidence (Level-IV or V studies with consistent findings) for or against recommending intervention.” The amended Grades of Recommendation are presented in Table I. We hope that this minor revision will make the Grades of Recommendation even easier for authors to use.

TABLE I Grades of Recommendation for Summaries or Reviews of Orthopaedic Surgical Studies

References:

1. Wright JG, Einhorn TA, Heckman, JD. Grades of Recommendation. J Bone Joint Surg Am. 2005;87:1909-10

Dr. Rama states that Grades of Recommendation are “simple and useful for the busy clinician”. He is also correct that Grades of Recommendation are built on Levels of Evidence and, thus, are complementary to the efforts of JBJS to promote evidence-based orthopaedics. We appreciate Dr. Rama’s support for Evidence-based Orthopedics and the changes to JBJS.

In response to Dr. Rama’s specific comments, he first points out that if randomized trials have not been performed, then treatment recommendations arising from those studies cannot be Grade A. We agree.

Second, he correctly emphasizes that randomized trials, like all research studies can be generalized only to the specific study population. As such, any treatment recommendations must consider the study populations of interest and extrapolation beyond those studies would constitute expert opinion or Level V Evidence.

Thirdly, he points out that intermediate outcomes such as bacterial counts cannot be extrapolated to other endpoints such as surgical site infection. This is not an issue of Levels of Evidence or Grades of Recommendation but inappropriately extending results beyond conclusions specific to the study.

Finally, he comments on the difficulty of balancing risks and benefits of surgical procedures when evaluating the outcomes of surgical therapy. Although differing valuations of outcomes of surgical trials may lead to contradictory grades, more likely the trials themselves will have contradictory conclusions and hence require consistent Grade C Recommendations.

We thank Dr. Rama for his comments and look forward to wider adoption of Grades of Recommendation wherever orthopaedic review articles are published.

To the Editor:

The recent introduction of “Grades of Recommendations” for review articles by The Journal is greatly appreciated (1). Such a simple classification will be very useful for a busy orthopaedic surgeon in obtaining guidance about treatment recommendations. Moreover, such a grading standardizes the conclusions of the authors of review articles.

However, the proposed grading system is merely an extrapolation of the ‘JBJS Levels of Evidence’ (2). Henceforth, a few words of caution are necessary for readers before assimilating the grades of recommendation into clinical practice.

First, although randomized controlled trials are widely accepted as the gold standard for obtaining the highest level of evidence, such trials may not be feasible in all instances (3). For example, we may never obtain "Level-I" evidence for the effectiveness of laminar air flow theatres in reducing the infections in major joint arthroplasties and so the grade of recommendation for such an intervention may never be ‘A’ in the proposed grading system.

The second issue is generalizability of the evidence. Results of the randomized controlled trials conducted in a narrow group of patients cannot always be generalized to the wider population. It is essential to clearly define the specific settings and the particular groups of the patients to which the grades of recommendation are applicable.

The third issue is the directness of the evidence. For example, a rise in the bacterial count in the operative field may not necessarily prove an increased risk of infection. It is advisable to provide the directness of the evidence also while stating the grades of recommendations.

Finally, irrespective of the level of evidence and its consistency, a judgment about the balance between the benefits and harms of an intervention must influence the type of recommendation in some instances, and so the recommendation may be more subjective or author dependent. This means that consistent findings of Level-I studies have the potential to lead to contradictory recommendations of high grade (grade A) by different authors.

Such limitations should always be considered by readers while interpreting these grades of recommendations. Nevertheless, this initiative of The Journal in extending its evidence based approach is highly acclaimed.

References:

1. Wright J, Einhorn T, Heckman J. Grades of recommendation. J Bone Joint Surg Am. 2005;87(9):1909-10.

2. http://www.ejbjs.org/misc/public/instrux.shtml#levels (accessed on September 8, 2005)

3. McLeod RS, Wright JG, Solomon MJ, Hu X, Walters BC, Lossing A. Randomized controlled trials in surgery: Issues and problems. Surgery. 1996;119(5):483-6.