[General Spine] Medtronic (BMP-2) potentially linked to cancer?

[nolonger_superman]

Hey everyone. I may just be paranoid, especially since I'm not exactly sure if I am impacted by this, but I just read a story on Gizmodo (hey, I am a geek after all!) where BMP-2 is potentially linked to cancer. I'm not trying to scare anyone as our situations are scary enough, but just thought I'd share a link so everyone can make their own decisions as to what to believe. Some days I feel the world would be better off without the internet so we don't find so much random information to make us afraid of everything!

Surgical Device Company Failed to Report Its Spinal Implant Causes Cancer

[Justin]

Thanks for posting!

[MDE]

As an industry note: Medtronic tried to get Amplify through the FDA and the potential link to cancer is what caused the FDA to say no. Amplify was basically just InFuse with more highly concentrated rhBMP-2. The rest of it was pretty much the same.

While the Author at the Milwaukee Sentinel did a lot of research in putting this together, what was more damning was that there was entire edition of Spine journal (the most widely regarded peer reviewed spine related journal) regarding the potential of skewed study data based on surgeon compensation, and that there were a variety of potential symptoms including cancer, sterility in men etc that were being linked to these products now, that had been understated during the regulatory process. This may not seem like a big deal, but it is the first time a medical journal has ever spent an entire issue blasting a product ever.

This pushed Medtronic to sponsor an unbiased review of data (A doctor at Yale is working on it), including surgeon compensation into the "variables" effecting data which is supposed to come out next year. Their sales of Infuse also dropped 17% after the Spine journal issue was released.

[nolonger_superman]

That is great information MDE and I appreciate your time providing some excellent feedback. I requested a copy of my post-op medical write up just to see what exactly I've had used on me. So if I am reading everything correctly, the concern is with Amplify and that infuse has not been linked to a higher cancer risk, correct? I did have infuse used and honestly I can't complain about any side effects and my healing process has been pretty phenomenal imo (still have trouble bending/twisting too much, but I am only 9.5 months out -- but I'm able to run a 7:30 mile, do chores around the house, yard work, light weightlifting, etc. I am thrilled).

Also mostly off topic, I am amazed at the level of detail provided in post op reports. I was quite shocked my surgeon took time to document multiple

[MDE]

The studies that raised this issue were done with Amplify, however Amplify is essentially identical to Infuse, with the only difference being the concentration of rhBMP-2. Realistically, this isn't really a major difference, as the "quantity" of rhBMP-2 is dependent on the volume of product placed within the body. Using a lot of Infuse can provide the same end result as using a smaller amount of Amplify. Surgeons don't always use the entire amount supplied (as it's a fairly large amount) and will typically pack the cages with the patients own bone that's been harvested from the area as well.

Now, we also have to look at what exactly 'statistically significant' means.

I'm trying to keep this explanation fairly basic, but if you already understand statistics, it may be a bit oversimplified.

When doing any medical study, you need a control group. You've probably heard of this from your highschool or college science classes, but with medicine it gets a bit more complicated. When we look at things such as surgery, there are SOO many variables that can impact the "success" rate of a surgery (I won't go into what actually determines success vs failure at this point), that you want to eliminate as many of these potential differences as possible. This requires comparing patients that are undergoing almost identical procedures with and without the new component.

So now we have 2 patient populations, the control group, who was given the current "standard of care" for the pathology, and the test group, which is given the new device/drug/etc. Ideally the two patient populations are identical in all ways. Unfortunately, even with randomized studies, this is almost never the case. With large enough patient populations though, the differences can be minimized and the two populations can be determined to be "statistically equivalent".

To make that statement the study has to look at all of the potential co-variants (differences that may effect results) and determine if they are "close enough" to each other to be equivalent. There are different types of covariants, such as sex, age, smoker, BMI, operative level, etc etc. You can compare almost anything in the patient populations. You won't ever have exactly equal number of males and females, equal distribution of ages, all pathologies are identical etc. This is almost never the case, so we make a determination that within a certain margin, populations can be determined to be equivalent. This margin is derived from rather standard statistics practice of looking at a standard distribution, and comparing it to another distribution (remember that control population?)

Significance in statistics is determined by a "p-value" derived from evaluating the null hypothesis (what you expect to happen AKA control group distribution) to the test population. This Wikipedia article is very good at explaining this: Statistical hypothesis testing - Wikipedia, the free encyclopedia, but basically it's a percent likelihood that the control group would include as extreme a value of a given variable in question occurrences shown in the test group. Basically, how likely is it that a variable is within "normal" variance.

This same test is then preformed on results of tests. For the rhBMP-2 example, your control group ends up with 1.4% of patients getting cancer within a certain time frame, whereas your test group shows 2.4% of patients getting cancer within that same time frame. That data alone really doesn't mean anything because people have a certain risk of cancer no matter what, so who's to say that the 1.4% and 2.4% aren't variations within the standard risk of getting cancer.

What the P-value does is evaluate the percent chance that a variable is statistically the same between two patient populations. For example, the paper talks about the difference of the 8th vs 9th patient in the study. 8 patients getting cancer after receiving rhBMP-2 results in the P-value of being greater than 5%, whereas that 9th patient drops the P value to be below 5% (commonly shown as P<0.05) which is a the commonly accepted value (in medicine) of when something is deemed "Significant". Describing it a different way, there is a less than a 5% chance that 9 patients in the control group could get cancer within the same time frame if you repeated the experiment.

So end result is that statistics and significance of those statistics are really useful, but are never exact.

The other issue is whether rh-BMP-2 causes cancer to start, or if it just makes malignant tumors to grow at a more rapid rate.

[nolonger_superman]

Fantastic explanation MDE. While I did have to take a number of statistical courses in college, I think your explanation would make perfect sense to someone who has had none. Additionally, great info in general. I feel smarter for having just read your last post.