arthroscopic surgery for knee OA??? not so helpful

there have been a couple of well-publicized studies finding that arthroscopic knee surgery for degenerative joint disease/osteoarthritis (OA) is not beneficial. however, there continues to be uncertainty, and "many specialists are convinced of the benefits of the procedure from their own experience".  in the US, 700K knee arthroscopies are done each year, and this number seems to be increasing. in this light, there is a welcome systematic review and meta-analysis just published (see knee arthritis arthroscop surgery bmj2015 in dropbox, or doi: http://dx.doi.org/10.1136/bmj.h2747 ).  details, for the 9 studies evaluated:

--1270 patients in total, receiving arthroscopic surgery with partial meniscectomy, debridement or both, vs control (sham surgery or exercise)

--mean age in the studies was 49.7-62.8. mean baseline pain (visual analog scale of 0-100mm) ranged from 36-63 mm.

--2 trials only had patients with radiologic OA (Kallgren and Lawrence grade 2 or more); 5 trials had some patients with radiologic OA; and 2 trials had patients without xray changes.

--time to primary endpoints in the studies varied from 3-24 months.

--results:

        --overall there was a small but statistically significant reduction in pain with arthroscopy (a difference of 2.4 mm on the visual analog scale), and only at 3 and 6 months, but not at subsequent evaluations. for reference, it is typically suggested that a change of 15-20 mm is likely to be clinically relevant.

        --no difference in physical function at any time-period between arthroscopy vs control

        --subgroup analysis:

                --no difference if there are xray changes of OA or not

                --no difference by type of surgery (partial meniscectomy with or without concomitant debridement) vs control

                --no difference by study design (sham surgery vs exercise for the control arm -- and, one might think that surgery would have more of a placebo effect than exercise as the control)

        ​--harms of surgery: DVT was most frequent, at 4.13 per 1000 procedures. followed by infection, pulmonary embolism, and death (death in 0.96/1000 procedures).

as we know, knee DJD is really common with aging, causes lots of disability, yet our diagnostic approaches are inadequate (the longitudinal studies, such as the Framingham Study, found a poor correlation between xrays and symptoms -- such that some reviews do not suggest getting routine xrays). in addition, the more revealing MRI exams show that cartilage and meniscus degeneration typical of OA can be found commonly in people without knee symptoms.  i am bringing up the current meta-analysis/systematic review because it reinforces the clinically insignificant benefit of surgery (though with attendant harms).  and, to me, one of our most important roles in primary care is to be a strong advocate for our patients -- so, for example, if we send a patient to an orthopedist for evaluation for knee replacement surgery and the orthopedist suggests arthroscopy based on OA with degenerative meniscal tears, this study supports our advising the patient to be evaluated by another orthopedist....