We recommend against vertebroplasty for patients who present with an osteoporotic spinal compression fracture on imaging with correlating clinical signs and symptoms and who are neurologically intact.

There are two Level I studies that compare vertebroplasty to a sham procedure.1, 70 These studies report no statistically significant difference between the two procedures in pain using the VAS and function using the Roland Morris Disability scale (up to one month and six months respectively).

These studies have been criticized for a variety of reasons. It has been argued that one of the trials1 was underpowered. However, this study did have sufficient power to detect the minimally clinically important (MCII) difference in pain (see Supporting Evidence section for details). Although crossover of patients after one month may have influenced the results in one of these studies,70 there was no crossover in the other study1 which also found no statistically significant or clinically important differences. Furthermore, crossover does not affect the lack of benefit for pain and function that the authors measured at one month.

Another concern was the low participation rate of eligible patients. This is an issue of external validity (generalizability) and not internal validity. The work group discussed this flaw, but chose not to downgrade this study for applicability because the trial authors noted that the enrolled patients were comparable to patients seen in routine care.[ref]

Furthermore, it has been proposed that vertebroplasty works better with certain fracture types than others. There are no prospective studies that report significant differences in outcomes based on fracture type.

It has also been proposed that vertebroplasty works better in patients that have more pain than those that were included in these trials. The baseline pain in both these trials was approximately 7 on a scale from 0 to 10. Other comparative studies had a baseline pain of about 8 and also had a mainly negative outcome.55, 71, 72

We recognize that a sham procedure may still introduce bias in the results (e.g. surgeons who know they are performing a sham procedure can unintentionally convey expectations to their patients) but there are also three other Level II studies that do not use a sham procedure as a control and they report similar results. One of these studies found clinically important pain relief at 24 hours.72 At six weeks pain relief was still statistically significant but not clinically important. After six weeks the effect was not statistically or clinically important (observations to two years).  One study reported results for pain that were statistically significant and possibly clinically important at one day but inconclusive at two weeks.71 Another study found inconclusive results at three months.55

By making a strong recommendation against the use of vertebroplasty we are expressing our confidence that future evidence is unlikely to overturn the results of these trials.