Cochrane in CORR®: Percutaneous Vertebroplasty for Osteoporotic Vertebral Compression Fracture

Importance of the Topic Osteoporotic vertebral fractures can occur without trauma, and cause pain and disability related to decreased mobility [14]. As the North American population ages in the coming decade, osteoporotic vertebral fractures will become more burdensome on patients and healthcare systems, since they are eight times more common among patients aged 85 to 89 than 60 to 64 [12]. The first published evidence regarding vertebroplasty was for the successful treatment of spinal hemangiomas [7]. Surgeons then used this intervention to treat patients with osteoporotic vertebral fractures even prior to the emergence of a randomized controlled trial supporting its use [2]. But in the era of “Choosing Wisely” [3], conscientious use of hospital resources, and evidence-based clinical decision-making, it is important to understand the evidence underlying osteoporotic vertebral fractures treatment. In this updated Cochrane review and meta-analysis (see appendix), which includes 21 randomized controlled trials totaling 2862 patients, the authors present the results from five trials (541 patients) comparing vertebroplasty to placebo (sham surgery) as their “primary comparison.” This comparison most likely approximates the true effect of the procedure by blinding patients to the intervention received [2]. In these five trials, when comparing vertebroplasty to sham surgery 1 month after the interventions, the authors found small improvements in pain (a 6% reduction) and Roland Morris Disability Questionnaire scores (a 7% improvement) [2]. Both of these findings are supported by a high Grading of Recommendations Assessment, Development and Evaluation (GRADE) for quality of evidence [2]. Seven of the included trials compared vertebroplasty to kyphoplasty and found similar benefits from these two treatments. The impact of vertebroplasty on patient survival was not assessed in this review. Upon Closer Inspection While pain and disability scores improved in the vertebroplasty groups, neither improvement achieved the minimal clinical important difference (MCID) for the metrics in question, which is defined as the “smallest difference that a patient perceives as beneficial” [2, 8]. The MCID is important for contextualizing the effect sizes of treatments that have achieved statistical significance but may not be clinically relevant to patients [8, 13]. The MCID for pain reported on a VAS scale is approximately 15% [4, 9, 11]. That being so, the 6% improvement observed in the Cochrane review was unlikely to have been recognized or appreciated by the patients treated with vertebroplasty. When the treatment effect of a commonly used intervention such as vertebroplasty is not demonstrated in aggregate, it is reasonable to determine whether any subgroups may benefit from the treatment to better refine surgical indications. Osteoporotic vertebral fractures have a variable natural history with some patients experiencing pain and disability for extended periods [19]. The Cochrane review performed one subgroup analysis and found no differences in pain and disability scores when the duration of pain prior to surgery was more or less than 6 weeks. However, there is some evidence that patients who experience pain for 6 months or longer [19] and who have bone edema on MRI may be more likely to achieve pain reduction from vertebroplasty [1, 20], but no subgroup analysis of these populations was possible in this Cochrane review. Future studies examining vertebroplasty in specific subgroups or more-tightly defined indications might be worth considering; any such studies should follow established principles to ensure that subgroup claims are credible. Specifically, there should be a small number of subgroups assessed, they should be specified a priori, and there should be a strong biological rationale for selecting the subgroups [18]. Take-home Messages It is rare to have sham surgery-controlled trials in orthopaedics, and even less common to have five such trials contributing to a high GRADE of evidence supporting clinical decision making [10, 21]. Since the pooled results of vertebroplasty did not provide patients with a clinically important improvement in pain or function in this Cochrane review, it does not support the routine use of vertebroplasty for osteoporotic vertebral fractures. The main findings of this meta-analysis are consistent with the one previously conducted by the same authors from 3 years ago and replicated in other recent meta-analyses of randomized controlled trials on this topic [15-17]. However, it is possible that patients with osteoporotic vertebral compression fractures experiencing pain for longer than 6 months who have MRI evidence of bone edema may benefit from vertebroplasty [1, 20], although this was not assessed in this review. Future trials could assess whether this particular subgroup of patients may benefit from vertebroplasty and additional large observational studies might seek to replicate the suggestion that vertebroplasty may be associated with improved survival after vertebral fracture [6]. While this commonly performed procedure still has its adherents [5], the best evidence we have as summarized by this Cochrane review suggests that vertebroplasty is ineffective in the general osteoporotic vertebral fracture patient population. Supporters of vertebroplasty should design appropriately controlled randomized clinical trials focusing on the subgroups they believe may experience a benefit from this invasive intervention. Until such subgroups are identified, there seems to be little high-quality evidence in support of vertebroplasty for osteoporotic vertebral compression fractures. In the context of shared decision-making with patients, physicians should present this recent high-quality evidence that showed no clinical benefit from vertebroplasty.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.