Review of Centeno MSC safety paper: without controls, conclusions muted

The number of people around the world being injected with stem cells every day has never been higher and the heterogeneous group of medical providers doing these procedures is also at or near its highest level ever.

The cells used are also highly variable as are the procedures themselves, but most of these cells fall under the umbrella of mesenchymal stromal/stem cells (MSC). The field needs to learn more about these MSC procedures in terms of safety and efficacy. Preliminary, relatively small and most often open label studies have been generally encouraging on safety, but not very definitive. For these reasons, whenever I hear about a new study published by a stem cell business, I am interested to read it and see what it teaches us.

A new published study from Dr. Christopher Centeno reports findings from a number of affiliated clinics that use MSCs for orthopedic procedures.

It is good to see stem cell clinics collecting and publishing their data so that is a plus. I commend them for that. The number of adverse events and in particular stem cell-related, severe adverse events was seemingly relatively very low amongst the population studied, but there’s a catch.

Centeno Table 3
Table 3, Centeno, et al., 2016

The overriding problem with this study is that there are no matched controls, leaving the reader unclear as to relative risk of the stem cell procedures. The data from this paper’s population of stem cell transplant patients who received various kinds of MSCs for a variety of orthopedic conditions is only compared to data from other previous studies reporting rates of adverse events in other populations of patients getting standard of care treatment or to general patient populations.

This lack of controls is unfortunate, but not surprising. For-profit stem cell clinics are in the business after all of making money and most patients are not going to be interested in paying money to the clinics only to have a chance to be a control not getting the stem cells.

Further, clinics in general have not shown interest over the years to themselves pay the additional costs associated with having control subjects. But it could be done. For instance, a clinic could do a study at cost (no profit margin built in) and offer participants the chance to be research subjects at a relatively low cost. Another idea, and this may sound outlandish given where the stem cell arena is today but is perfectly logical and reasonable, is that clinics could even pay patients a small sum to be clinical trial subjects in a controlled study.

Overall, the lack of controls in the current paper greatly limit what we can learn from it. Of the adverse events reported in this new MSC paper some sound concerning such as cancer (termed “neoplasms”; see Table 3 above) in the paper, but again without controls there is no way to be certain if these tumors were stem cell-related. The authors do not believe they were. I can say, “most likely they are right”, but scientifically there’s no way to be sure at this point. It is possible a matched control population would have gotten even more tumors or conversely no tumors at all.

While this study had longer follow up than some past ones, the follow up here still was only around one and a half years on average and it is reasonable to imagine that some adverse treatment-related events could arise much later.

Another concern with the paper is that reading the methods seems to suggest that aspects of the protocols used to administer stem cells to patients varied considerably even within the three individual treatment groups (same day marrow MSCs, marrow and lipid aspirate combo, and culture expanded MSCs). For instance, the number of cells and the amount of co-injected platelet lysate had sizable ranges. An interesting side note to this is that this variability reflects a challenge facing the stem cell clinic industry of every clinic even amongst affiliates doing things somewhat differently.

Update: I also meant to say that I think this paper was largely well written and where appropriate included caveats, limitations, etc.

Finally, I’m not clear on the strength of the methods for assessment of adverse events and assignment of potential adverse events to be related to treatment. Online recruitment of assessors may not be the most reliable approach.

Overall, I’d say this paper is just a bit encouraging on autologous MSC safety for orthopedic use in the short term. It is also a step in the right direction in terms of the more data the better from for-profit stem cell businesses, but without controls it is hard to say much more. For those who say that this kind of study is worthless, consider the hypothetical scenario that the same patients had been given these therapies exactly as reported, but none of this was published. How could that be a better situation than having a publication of the data even without controls? As long as we as a community soberly evaluate such publications, I see how they serve a role.

Finally, I propose a friendly challenge to those doing these experimental stem cell therapies to devote some funds to controlled studies and to do much longer follow up. This would also be a reassuring sign of good faith that the clinics have long term patient best interest as a top priority.

22 thoughts on “Review of Centeno MSC safety paper: without controls, conclusions muted”

  1. @Dr. Centeno
    So I guess BMC would be your first choice to treat AVN, not PRP?
    After his AVN was healed, my brother has left cyst in his deep talus role (far away from the cartilage), is there any chance to heal this cyst, too?

  2. Chris Centeno, M.D.

    AVN is a good example of the physician innovation pathway, as Hernigou has linked outcomes to CFU content in bone marrow concentrate. This is a disease where an RCT against placebo is tough, as the placebo group in just 3-6 months could rapidly progress into a hip replacement. Hence, we have a long history of many published studies that are large case series. Given that there is a 80% chance of predictable collapse in the bone, any lowering of that rate is a high confidence that the effects are not placebo. Before/after MRIs showing rapid resolution after an injection is also interesting data, as spontaneous resolution is uncommon.

  3. Chris Centeno, M.D.

    There is data showing efficacy of BMC in AVN via Herigou going back to the 90s-see . Not much data on PRP, but I have no problem believing it may work. We have treated about 100 cases with BMC over the last 8-9 years via trocar into the bone lesion guided by fluoro-works well in ARCO grades 1-2, less in 3/4. Here’s a nice case with MRI evidence:

  4. @Dr. Centeno part 2
    … On your homepage I can only find data concerning other diseases, but not for avascular necrosis.
    I think you have the data and could give us a valid answer. So I would like to ask you to give us some more information about the outcome rate/healing rate for avascular necrosis by PRP-therapy.
    Thank you in advance

  5. @Dr. Centeno part 1
    I have personal experiences with PRP-therapy. My brother was suffering from avascular necrosis at his joint for 7 years. No other conservative therapy helped him. Then he received a PRP-treatment (6 times once a week in combination with a shock wave therapy) and he was healed within 6 weeks. It was very amazing. Of course this result is just n= 1, and so I would like to know, how is your success rate for the treating of avascular necrosis?
    We discussed this case/theme here already several times in this forum, most comments were sceptic, but nobody was able to tell more about a efficacy results/outcome results which were backed by data. … part 2 next comment

  6. Chris Centeno, M.D.

    @admin, the FDA process is very long and expensive. Given that we have companies in this space that began their quest for approval in the late 90s (Osiris for example which was later bought by Mesoblast) and we’re sitting in 2016 without a single approved US FDA indication for an MSC cell drug, can you blame the companies for wanting an easier pathway? At least for autologous products, countries like Japan have decided that it’s time to spur innovation by allowing these procedures to be performed in an alternate framework. If you’re a fan of more regulations and more studies, then REGROW is likely perceived as bad. If you’re a patient, you want more options. I had thought REGROW might be a great thing for physicians, but as I said, it’s really not.

    Who is behind REGROW? There’s a long list of orgs that have thrown their weight behind it. Am I personal fan? I have mixed emotions. It won’t help physicians or clean up the crazy stem cell space and will increase prices for therapies. On the other hand it’s at least some recognition that maybe the drug pathway was never the right place for these therapies to live.

    BYW, there seems to be an issue with the comment box in Chrome, as when you leave a longer comment it gets rid of the post comment button! Works fine in IE, so I just switched browsers to leave the comment.

  7. Chris Centeno, M.D.

    All published efficacy studies are here: There are a number of multi-site, prospective registry based studies. We have additional similar studies now being submitted on other areas such as spine (you’ll find shoulder, knee, hip-also tendon/ligament) on the site. All patient registry data by site is here: As discussed, 3 RCTs in the works, one done recruiting. Does this stuff work? We’ve seen it work in many patients, but unlike other clinics in this space, we recognize that it’s up to us to prove that to the best of our ability.

  8. Chris Centeno, M.D.

    Brian, you’ve framed a big issue nicely. RCTs are great, we always need more of them. Having said that, they purposefully create an artificial world where strict inclusion and exclusion criteria filter out 80-90% of the real world patients a physician is confronted with day in and day out. Hence, data from an RCT usually doesn’t generalize well to the usual mix of real patients. This is why registry data is often used to craft better RCTs, providing important clues as to what’s working and not working.

  9. … @Dr. Centeno, part two
    … For example:
    How high are the efficacy rates for …
    … repair of a damaged disc?
    … the repair of a meniscal tear?
    … the repair of a small cartilage defect?
    … the repair of a big cartilage defect?
    … the repair of an avascular necrosis?
    Do you have data about this outcome, so please let us know.
    Because as Dan Kaufman already said:
    “The real question for those MSC therapies is do the benefits outweigh the risks?”
    To answer this question anybody must know the efficacy rates.

  10. @Dr. Centeno Dear Dr. Centeno,
    thank you very much answering the questions above. Safety is one big issue, but the second one is efficacy. Where can I find efficacy rates? Do you already have done studies about this? see next comment …

  11. Paul, your friendly challenge was way out of date before you wrote it, see, for example:
    OK, it’s not, perhaps, a pure control in the sense that a lab-rat scientist might prefer because they offer the patient the option of a cross-over afterwards. In my opinion, what they do is the ONLY ethical thing to do.

    The issue of treating diverse vs carefully selected patients is an interesting one. Especially so when you consider that in the real world a medical doctor isn’t always confronted with conveniently selected patients.

  12. Chris Centeno, M.D.

    The BMC+Adipose Graft group had the highest short-term complication rates, but that was mostly slef-limited pain/swelling. The cultured group had an overall higher rate of AEs, but the problem is that the longer duration of the follow-up and the age (most patients were middle aged on entry and this group will have more disease events occur over longer periods as they entering into prime disease prone years (i.e heart disease, stroke, cancer, etc…).

  13. Chris Centeno, M.D.

    On Regrow, I think if you’re an early phase cell company pursuing FDA approval (as is one of the companies we licensed an intervertebral disc technology to), you love Regrow. If you’re an existing cell therapy company past phase 1-2, you hate Regrow. As a provider, I’m pretty agnostic as the new pathway it creates is not one that most providers would be able to achieve. As an example, the company mentioned above has spent 600K in pre-IND alone and will likely spend 3-5 M to finish phase 1-2. Physicians don’t have that kind of money. I also think the act will raise prices for patients into the stratosphere, as companies spending 5 M to play plus the cost of cGMP manufacturing of an autologous product will need to raise same day therapy prices from about 4-8K to 10-20K. In addition, the safety data generated by phase 1-2 is minimal compared to the paper we just published. On the other hand, autologous BMC and short-term cultured MSC therapies are likely pretty safe based on the existing literature, so safety is likely the least of the concerns. In addition, as a physician, if a patient is terminal or has a severe chronic disease, I’m not sure what we’re protecting them from?

    On the physician side, on the one hand the act doesn’t change FDA’s penchant to categorize everything as a drug. On the other, I’m no fan of the wild west of stem cells. Regrettably, the act also does nothing to try and rein in and control an out of control physician stem cell space, where doctors can take a weekend course and begin offering IV SVF treatments for every known human ailment. Or clean up the clinics (like the one you highlighted) that are offering “amniotic stem cells” which consists of a vial of dead tissue.

    So on the one hand, it would help a technology that we licensed get to market, but do I think it will help physicians or patients, not so much…

    1. @Chris,
      Thanks for sharing your take on REGROW, which I find quite interesting. Why do you think those who crafted the Act and support it are doing this?

  14. Chris Centeno, M.D.

    All, we have 3 RCTs ongoing, and have published more registry based efficacy data in orthopedics than anyone at this point (none of which can rule out a placebo effect). RCTs are great, but they will never have the statistical power nor duration to pick up rare events (like the neoplasms discussed above). This is why drug approvals rely on post-marketing surveillance, which is less intensive than what we did on these 2,372 patients. In conclusion, very large controlled trials where side effects can be compared to placebo is needed in addition to this type of work. Having said that, many problems with approved drugs are picked up after the approval using a comparison of reported events to those that happen randomly in the general population.

    1. @Chris,
      Thanks for the comment. Do you have a sense of which of the three treatments that were compared that you think is the safest/best?

      I’m also curious if you’d be comfortable weighing in on the REGROW Act and the Bipartisan Council report that both advocate for significant FDA changes including in some cases not requiring Phase III, conditional approval, delayed requirement for BLA, etc.

      We’ve been having some debate on this blog about these with some in favor and some opposed to them. What do you think?

  15. It seems the main conclusion is, “well, at least they tried.” That seems very “adequate.”

    The real question for those MSC therapies is do the benefits outweigh the risks? There seems to be relatively little downside. I agree the number of neoplasms is probably not significantly different than the general population. However, as you state, no way to know.

    However, there is no way to determine clinical efficacy without appropriate controls. Sure, patients probably feel better after getting injected and paying some fee. How much of this is placebo affect and maybe non-specific inflammation, that could lead to angiogenesis and repair?

    Getting a good, double-blind trial with sham injections is the only way to determine efficacy. Private clinics like this won’t do it, and seems academic groups not too interested, unless you know of some “real” trials going on in this area.

  16. Dr. Centeno’s study seems refreshingly hopeful and rounded when it comes to a true meta-analysis of actual clinical practice.

    I’ve wondered, though, why it is that our amazing research interests in this country are not taking full advantage of the ample safety (with animal study and controls) and efficacy research already established internationally? And why no mention of the international 7-8 year longevity studies establishing efficacy of MSC’s. Especially for orthopedic uses.

    The clinics seem to feel confident in their approaches for legitimate reasons. The factor making them “most likely” okay to let continue their practice.

    We should embrace the seeming successes of this arm of treatment. Engage it further with our regulatory system, and allow our regulatory system to adapt to the unique situation the U.S. stem cell research community finds itself in.

  17. Nice review. To be fair, there are for-profit clinics that simply don’t bother reporting anything and here we have one where the reporting is pretty well balanced and constructively critical of it’s own limitations. Obviously barely any of the subjects would have been in the same cohort in a real clinical trial due to diverse backgrounds, pretreatments, and so on. But in addition to true controlled studies, there is value in such a wild mix of patients, as one may observe adverse events at n=1, that you don’t otherwise see, and these can then be followed up more rigorously.

Comments are closed.