Monday 23 April 2012

BMPs are not BHPs

BMP stands for Bone Morphogenetic Protein. The name suggests that it makes bone, and that is what it does. To be precise, it turns the tissues around it into bone. However, people think that because it makes bone, it heals fractures, but those are two separate things, and the evidence for it healing fractures is much less than the evidence for it making bone. That is why it is called Bone Morphogenetic Protein, and not Bone Healing Protein.

There are several types of BMP, but the two currently on the market are BMP2 (marketed as Infuse by Medtronic), and BMP7 (marketed as OP-1, by Stryker Biotech). The BMP family of proteins has been researched for 50 years, and over recent decades, recombinant DNA technology has made them marketable.

They are widely used, and the business is probably worth around a billion dollars annually (I am working off a figure of $750 million in 2007 for Infuse alone). In a review of tissue engineering from 2009 (, Infuse was labelled as one of the biggest success stories in getting from lab to market (I will be addressing the other supposed success story in the near future).

BMP is used in oral, spine and fracture surgery to get bones to heal. My opinion is that it produces bone well enough – I can’t argue with that. It’s just that it doesn’t heal fractures. Let me explain.

Infuse was tested in a multicentre randomised trial called the BESTT study, the results of which were published by Govender in the Journal of Bone and Joint Surgery in 2002 ( The drug (BMP2, Infuse) was placed in a collagen sponge, in two different doses, and then inserted into open tibia fractures at the time of skin closure (soon after the injury), to see if the BMP could get the bone to heal more reliably and quicker. They compared the two doses to a placebo and found that fractures in the high dose group were more likely to heal, healed faster, and required less secondary surgery.

The FDA bought it, and surgeons have been buying it ever since, at about $6,000 a pop. However other (European) approving bodies were a little more suspicious, and requested more information. The concern was over aspects of the study that had been raised in various letters to journals, regarding the discrepancies between the treatment groups. The idea of a randomised trial is that the treatment groups that you are comparing are all pretty much the same, except for the thing that you are actually testing. It turned out that the group with the high dose BMP were much more likely to have been treated with a reamed intramedullary nail (41%) than the placebo group (27%). This could have explained the difference in healing rates, and despite repeated analyses, the skeptics were not placated.

Govender, to his credit, did another similar study comparing a high dose group to placebo, and used reamed intramedullary nails in all patients ( There was no difference in the rates of healing at 20 weeks (68% vs 67%) or in the rates of secondary surgery (12% in both groups), but the infection rates was higher in the BMP group (19% vs 11%). Now, I will admit that the difference in infection rate was probably due to sticking collagen in an open fracture (rather than the BMP), but you will have to admit that the BMP didn’t help the healing much.

Those studies were in fresh fractures. Before that, BMP7 (OP-1) had already been approved for the treatment of non-union of the tibia (where the bone hasn’t healed after a long time). This is based on a randomised trial of 124 patients by Friedlaender ( that compared OP-1 to autologous bone graft (taking bone from the patient’s hip). The union rate after the surgery was not statistically different between the two groups. The conclusion was that OP-1 is as good as autologous bone graft, but I disagree. You always have to be a little wary when there is no placebo group, and when there are “cointerventions”. The patients in this study were mainly treated with reamed intramedullary nails (the same things that probably did the healing in the Infuse studies above), and the bone graft or BMP was given on top of that treatment. So this leads to the question: What would have happened if they hadn’t had the BMP or the bone graft, and only had the reamed intramedullary nail? Well, what we can tell from several case series of tibial non-unions, is: pretty well. Probably better than the BMP group (75% union), because that number is on the lower end of the scale.

I know that it is difficult comparing a case series from one place, to a trial from another, but you will have to admit that this study did not tell us that using OP-1 was better than NOT using OP-1.

We have used BMP only once in about 10 years in our institution, and I am pleased to see that it is losing favour in other places, but for the average surgeon in the community, it is still being used because it doesn’t cost them or the patient anything (not directly, anyway) and it might work.

At least it doesn’t do any harm, right? Well you might have to ask the spine surgeons about that. They have been using it in the spine, often “off label”, and have noticed a few problems. This study ( looks at a few trends, but the number of papers out there reporting problems is considerable, and I notice on the internet that the lawyers are now getting in on the act.

Addit 17 Dec 2013: Another trial reporting no advantage from BMP in acute tibia fractures.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.