Friday, August 14, 2009
As I mentioned briefly last week, NEJM has published two blinded, sham-procedure controlled trials of vertebroplasty for painful osteoporotic vertebral compression fractures, both of which indicated VP was no better than the sham procedure (article 1 here, which I will refer to as Buchbinder, its primary author's last name; article 2 here, Kallmes). (Supporting editorial here.)
Buchbinder was an Australian multicenter study of ~70 patients with non-malignant, painful VCFs less than 12 month old. Everyone, including all the investigators were blinded to treatment group (only the radiologists who did the procedures were unblinded; they weren't involved in data collection, patient care afterwards, etc.). Patients were randomized, brought to the procedure suite, and underwent identical procedures with the exception that the sham patients just had their vertebral bodies lightly tapped with a blunt instrument and obviously no 'cement' inserted (for the sham patients the cement was opened and mixed in the room, the patients had local anesthetic, etc. etc.): they did a very careful job of blinding the patients to which procedure they had. The study was powered at 24 patients in each group to show a 2.5/10 point reduction in pain in the active group over sham.
A lot of outcome data were generated that's not worth belaboring: at no point (1 week to 6 months) did the active group patients have superior pain relief to the sham group: both groups reported rapid and durable pain relief (on the order of 2/10 in their overall pain ratings). Notably, the results were the same just looking at those with the freshest fractures (less than 6 weeks of symptoms) - although the study was not adequately powered to determine a difference just in this subgroup. Secondary outcomes about quality of life, etc. were similar between groups.
Kallmes was a multinational trial involving similar patients (N=131, all with fractures less than 12 months old, pain at least 3/10, not cancer related VCFs). A similar protocol was used, including anesthetizing the periosteum of the vertebral pedicles, face activity to suggest they were actually doing the procedure, etc.; blinding was similar as well. The study was powered at an N of 250 to find a 1 point difference in pain outcomes; when they had accrual problems they changed their targets to an N of 130 with '80%' power to detect a 1.5/10 difference in pain at one month.
The results for pain and disability were basically negative at all times, with some notable caveats. Baseline pain was ~7/10 in both groups; both active and sham groups had an immediate reduction in pain down to ~4/10 at 3 days, which continued out to 1 month (active was 3.9, sham was 4.6 - not statistically significant).
Kallmes allowed cross overs after one month (both ways - those who had real VP originally were crossed over to the sham procedure). By 3 mo 43% of sham patients had crossed over (asked for and received the alternative treatment - all of this was still blinded to the patients), compared to 12% of the VP. By one month after the repeat procedures there was no difference in pain outcomes between groups either. Overall, all of those who sought cross over did worse at all times than those who didn't. Interestingly, 63% of the control group patients correctly guessed at 14 days that they had received the control intervention; only 51% of active VP patients had.
Kallmes notes in the discussion:
These results suggest that factors aside from the instillation of PMMA may have accounted for the observed clinical improvement after vertebroplasty. Such factors may include the effect of local anesthesia, as well as nonspecific effects, such as expectations of pain relief (the so-called placebo effect), the natural history of the fracture, and regression toward the mean. The possible role of the placebo effect on outcomes in this trial remains unclear. Previous studies have documented pain reduction in placebo groups, on the order of 6 to 7 mm on a 100-mm scale. The treatment effect in our trial was substantially larger than those in previous studies, even though the previous studies included both pharmacologic and psychological interventions in addition to physical interventions.Compare this to the unblinded, randomized, conservative therapy controlled trial (of balloon kyphoplasty) I blogged about a few months ago which showed marked benefit of KP over conservative treatment.
What does all this mean?
When it comes down to it I think it does mean that the marked pain improvements noted after VP are in fact not due to installation of cement, just as Kallmes remarks. It's complicated though, to say the least, and unanswered questions remain - which don't however invalidate these new findings which are by far the most rigorous investigations of VP.
- Why was the pain relief in Buchbinder half of that in Kallmes (2/10 vs. 4/10 on average)?
- There is a sense from the Kallmes trial (with the differences between groups in rates of guessing correctly which treatment they initially had and rates of seeking cross over) that real VP had some benefit over sham which wasn't noticable at all on pain and disability questionairres. Or pain 'really was' improved in the VP group compared to placebo and the study didn't have the power to show it. Even if that's the case it is not good news for VP - if the benefits (over sham) are so minimal (or rare) that you need more than 130 patients to demonstrate it - this is not a ringing endorsement of the procedure. Needing more than 130 patients to prove a treatment saves a few lives a year is fine (and typical in cardiovascular research); for an analgesic modality it means it's pretty worthless.
- It is possible the sham procedure is actually an effective, active treatment for VCFs which they inadvertently demonstrated in these trials. It seems physiologically unlikely (outside of the initial hours after anesthetizing the pedicle) but one could invoke some 'placebo' response explanation? Such as - patients think this will help; anesthetizing the pedicle gives them rapid and effective pain relief; patients now 'know' if fact this will help and despite the fact that nothing else has been done which effectively relieves pain (including, apparently, the installation of cement into the vertebral body); this expectation/belief provides the patients durable (months long) pain relief.
- Balloon kyphoplasty, a related procedure, was not investigated and one can't conclude anything about KP from these trials (despite the compelling** non-blinded controlled research I cited above). However, these papers indicate the effectiveness of KP can no longer be assumed without properly blinded studies.
- It's possible that certain patients would receive benefit from VP over sham; patients with very acute fractures, specific anatomy, etc. (there has been for instance a clinical impression that more acute fractures respond better). Perhaps...but we can no longer assume that without it being demonstrated properly.
- Who knows about cancer patients? All I can say, and to join most of the world here, VP frequently seems to help. And in fact it does help, even if it's placebo - see below.
- These papers, if you accept their findings as they are (which I do - they are as good as you can get when it comes to this sort of symptom research), shouldn't actually lead us to conclude that VP is as good as medical/conservative management. VP is better than medical management, and has been shown to be so (as has KP). The results very specifically mean VP is no better than 'sham,' which is a different conclusion than no better than medical management. The difference is now we have reason to believe VP is better than medical management due to that thing called 'the placebo effect' as opposed to, eh, 'real analgesic effects' of instilling PMMA into a vertebral body. Real in quotes there because I don't mean to suggest that 'placebo' is not a real analgesic. It is, god knows, I just wish I had it in a bottle as opposed to a fancy fluoroscopy suite.
- I have thought VP and KP are pretty darn good things the last few years as I've seen many patients feel a lot better afterwards. It's frustrating and you wonder what would have happened if rigorous, blinded studies were done early on. E.g. the technique is developed and a few case series suggest a powerful treatment response to VP (this in fact happened). Next, a properly blinded trial was done which showed no benefit. Maybe this was repeated with larger trials - still no benefit - VP is never adopted and never became as wide-spread and accepted as it has now become. Instead we had large, unblinded trials; VP still looks great, everyone loves it, and then this. All of this makes a good argument (if we ever needed one, which we do, as we constantly forget) for remaining skeptical of unblinded treatment trials and also of doing blinded trials early, particularly with novel therapy.
- But coming back to #1 there, what do you do clinically given that it still remains that VP is 'superior' to medical management (at least in the short term) even if it's because of placebo. It's not ethical, supposedly, to prescribe placebos, right? But we do it all the time - at least I do with, for instance, lidocaine patches for pain syndromes highly unlikely to 'actually' respond to a lidocaine patch - I do it because it's almost ridiculously safe and a substantial minority of my patients report they feel a heck of a lot better. An invasive procedure is different though; while quite safe the tiny risk of a catastrophic adverse event remains. Frankly, for patients with disabling pain that I cannot effectively treat medically/conservatively without dose limiting side effects, I think I will continue to recommend this, or maybe kyphoplasty to make myself feel better.
- Please comment as to what your plan to do based on these studies. I'm also curious to the community's larger response and if this will in fact be the death of VP. As long as people get reimbursed for it though I imagine it will continue to be practiced.
- I Googled vertebroplasty to see if there were any open source images for this post and found this link (www.youhaverights.com/medical-devices/spinal-cement/) was at the top of my search (it was a paid link, and I am not embedding an actual link here because I don't want to improve this website's Google rank by doing so). Yes, people are already trying to cash in on this via litigation (thus this post's title). God bless America.
(I am traveling next week and won't be posting. Christian is back [missed you buddy] and we should soon be ending our lighter, 'summertime' posting schedule.)