Vertebroplasty should not be offered as a treatment for spinal fractures due to osteoporosis

23 November 2018

Rachelle Buchbinder is an Australian NHMRC Senior Principal Research Fellow. She has been the Director of the Monash Department of Clinical Epidemiology since its inception in 2001 and Professor in the Monash University Department of Epidemiology & Preventive Medicine since 2007. She is a rheumatologist and clinical epidemiologist who combines clinical practice with research in a wide range of multidisciplinary projects relating to arthritis and musculoskeletal conditions. 

Vertebroplasty, a treatment that involves the injection of a type of acrylic cement, is a treatment first introduced in the late 1980s to treat acutely painful osteoporotic vertebral fractures. Early observations from case series published over the next decade and a half, indicated that it could quickly and dramatically improve the pain. On average, across 30 uncontrolled studies, the pain improved from a score of about eight on a scale from 0 to 10 (where 0 indicates no pain and 10 indicates very severe pain) to about a two and a half, and serious adverse events were reported in less than 1% of patients (1). The treatment was widely hailed as a miracle . For example in June 2003, the UK Telegraph reported it as an example of the ‘Lazarus effect’, enabling the bed bound to walk again, and it rapidly became standard of care.

The first two randomised placebo-controlled trials of vertebroplasty were published in 2009, more than twenty years after vertebroplasty had been introduced into practice (2, 3). In contrast to all previous studies, both participants and investigators (other than those performing the procedure) were blinded to treatment allocation, minimising selection and performance bias. Neither trial found a benefit of vertebroplasty compared with placebo, suggesting that previously observed benefits were explainable on the basis of contextual effects (e.g. favourable natural history, regression to the mean, placebo effects).

Three more placebo-controlled trials have now been completed (4-6). Addressing concerns about the two previous trials, all three included only people with severe pain that had been present for a short time (up to 6 to 9 weeks). The results of two of these trials were exactly the same as the previous trials (i.e. no demonstrable benefit from vertebroplasty)(4, 5), while the third trial was reported more favourably (6). When data from all the placebo-controlled trials were combined in pooled analyses within a Cochrane systematic review, there was high- to moderate-quality evidence that vertebroplasty provides no greater benefits than placebo (7). Subgroup analyses indicated that the results did not differ according to symptom duration (acute versus subacute).

Understanding the science leads to an understanding of how vertebroplasty could have initially appeared to confer such large benefits. When the outcome that is being assessed is based upon patient judgment, trials that do not blind participants and investigators on average overestimate the treatment benefit by about 25% (8). If this is taken into account, results of open trials comparing vertebroplasty with usual care are entirely consistent with the results from placebo-controlled trials (7).

Many hundreds of thousands of patients have received vertebroplasty since the first two high quality randomised trials were published, and many harms have been reported. These have included infection, rib fractures, further vertebral fractures, cement emboli lodging in the lungs, paralysis and even death. Doctors should now be convinced enough to stop using this treatment due to its unfavourable risk-benefit ratio. At the very least, they should fully inform their patients that evidence that does not support its use to treat acute vertebral fractures and its potential for harm.

References

  1. Hochmuth K, Proschek D, Schwarz W, Mack M, Kurth AA, Vogl TJ. Percutaneous vertebroplasty in the therapy of osteoporotic vertebral compression fractures: a critical review. Eur Rad 2006;16: 998-1004.

Comments

Add new comment

CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.