In a paper published earlier this year (View Paper ), Shapiro, et al. reported on the results obtained from a clinical study in which a patient’s bone marrow was concentrated (creating BMC) and injected into one of the patient’s arthritic knees, with the other knee being injected with sterile saline. It turns out that the study was not only cleared by the Mayo Clinic’s Institutional Review Board (IRB), but that the organization opted to initiate an IND (Investigation New Drug) for reasons that weren’t explained. And don’t ask me why a study of 25 patients in a prospective, single-blinded clinical trial was elevated to an IND. I admire the willingness of someone to pay for this hugely expensive approach to clinical studies, but I at least would have done a preliminary clinical study without the huge price tag associated with an IND. What makes the IND status interesting is that as stated by the authors, “…a similar degree of pain relief was observed for both BMAC- and placebo-treated knees…”, which could lead to an indication for use if the sponsor continues with the IND process, making the world safe for treating bilateral knee OA by injecting one knee with saline and the other with BMC.
Let me try to clarify the issues by starting with a review of the clinical study design. The key feature of this study is that they only enrolled patients who suffered from bilateral knee osteoarthritis that was painful (mild to moderate) for an extended period of time and refractory to conventional therapy. Some of the patients had had prior surgery on one or more of their knees, which I consider to be a significant complication in the study. The two treatments to be evaluated were autologous bone marrow concentrate (BMC) and sterile saline (usually a 0.9% solution of salt—NaCl). The authors referred to the saline injection as a placebo, but I suspect the non-physiological nature of saline elevated its impact above that of a traditional control. In fact, saline is mentioned as one of the agents that is injected when performing prolotherapy, which is a treatment protocol intended to enhance inflammation at the injection site.
In terms of the actual injections, 15 mL of saline was injected, while 5 mL of BMC diluted in 10 mL of platelet poor plasma from the bone marrow aspirate was injected. The knees were assigned randomly to receive one or the other treatments, with the patient being screened and therefore blinded to what prep was going into which knee. Clinical outcomes focused on pain scores and activity levels, which were collected by clinic personnel blinded to the knee injections. The data in the publication includes the 6-month follow up time point. The authors performed an analysis of each patient’s BMA and BMC and reported some of the data in the paper, which is rare for this type of publication.
Not surprisingly, in view of the fact that the patients reported that both the BMC and the saline injected knees got better (i.e., reduced pain) to a similar extent out to 6-months, the authors spend a fair amount of the Discussion section on trying to explain the results. While I won’t go through the Discussion point-by-point, I will go through what I think are significant problems with the protocol, clinical study design and patient selection.
The authors admit that their clinical study design, in which patients with bilateral knee OA had both knees treated, might have contributed to the unexpected equivalency of treating knee OA with saline or BMC. They make a valiant effort to explain just why this might have happened. But I think it was a stupid study design, since both knees are on the same patient. Everything we know about how progenitor cells do their thing involves biomolecule signaling and sensing. And biomolecules float around, so they could leave the immediate vicinity of the BMC-injected knee and potentially could influence the pathology in the other knee. And why might that happen? Because the other knee was injected with an agent that is characterized as a “placebo”, but in fact might have stimulated additional inflammation in the joint. This extra amount of localized inflammation might have attracted the attention of the BMC-derived progenitor cells in the other joint and influenced them to migrate or kick their secretion of biomolecules into overdrive.
In contrast, consider a double-blind, placebo controlled study published by Black, et al. (2007; View Study Here ), in which canine patients diagnosed with OA of the equivalent of the human hip joint were treated with their own fat-derived cell preparation (called SVF, which contains progenitor cells like MSCs). The study was small, with just 21 or so dogs, but the composite pain score metric was statistically significantly reduced for the dogs receiving SVF compared to the controls who received phosphate buffered saline. There are two differences between Black, et al. and Shapiro, et al. that I will note; 1) A dog was treated with either SVF or the placebo, but not both at the same time; and 2) Phosphate buffered saline, a more physiologically benign fluid, was used as the placebo in the dog study, compared to saline in the human study. Some might argue that another difference is humans were treated in one and dogs in the other, but for purposes of the natural healing response to a pathology like osteoarthritis, I doubt there is a substantial inter-species difference when using an autologous therapy known to contain progenitor cells. So, the confusing data presented in Shapiro, et al. might be in part a result of having each patient be their own control.
I will review in the next post a number of other issues I have with the Shapiro, et al. study that undermines its conclusion about BMC and knee OA.