Vertebroplasty Not as Effective as Advertised

Vertebroplasty, the percutaneous injection of a cement into a vertebral fracture, has become a widely used procedure for the treatment of spinal compression fractures. These fractures are a common problem: A quarter of persons in the US over 50 will have such a fracture during his or her lifetime. Treatment has been overall disappointing. Uptake of vertebroplasty over the last ten years has been brisk, with a doubling of the number of procedures in the last six years. Of interest, geographical variation in use is very high, suggesting some effect other than efficacy at least partially determines the prevalence of use. And, there is an unanswered question about increased fracture rates in persons undergoing this procedure.

Until now, randomized data were lacking. However, two randomized controlled trials were recently published in NEJM which call into question the value of vertebroplasty. The first trial was done in Australia. This was a parallel group trial, with blinded assessors. 78 patients were enrolled, of which 71 (91%) completed the 6 month follow up. The primary outcome was an overall pain score, and a variety of secondary outcomes were also measured. Outcomes were assessed at multiple times during the study. The results were fairly definitive: No outcome was statistically significantly different between groups, except for one secondary outcome measure which favored placebo. There was also no difference between groups in the rate of new spinal fractures after the procedure. Of note: The placebo group received a sham injection, and both group s showed modest improvement. This may suggest an effect of the injection, rather than the cement per se. The authors posit that this may also be due to "[r]aised expectations of an invasive intervention..."

The second trial, led by an investigator at the Mayo Clinic, was of similar, but not identical, structure. This was also a parallel group trial, which enrolled a total of 131 patients. Inclusion criteria, specified recruitment and power calculations were changed early in the trial due to slow recruitment, but this probably did not alter the fundamentals of the results. There were two primary outcome measures. The first was the change of score at one month on a disability scale, with a predefined change considered of clinical significance (30%). The second was a pain scale. As with the first trial, various secondary outcomes were also evaluated. Also as with the first trial, none of the primary or secondary outcome measures reached a difference of statistical significance. However, there was a trend towards clinically meaningful improvement in pain ratings at one month in the vertebroplasty group (64% v. 48% of patients). Both groups showed improvement at 3 days post-procedure, which was maintained at the one month outcome assessment. The investigators cited a variety of possible explanations for the improvement in both groups.

The accompanying editorial, cited above, suggests more patient involvement in treatment decisions would be of benefit.

So, these two trials, the first randomized controlled trials of vertebroplasty, did not show a benefit compared to sham treatment. Given this, it will be difficult in future to recommend this procedure with any enthusiasm. Fortunately, a large subset of patients with spinal compression fractures do improve to some degree spontaneously.