LAUSR

Dear Editor, Following the recent publication by Clarençon et al [1] reporting our results of cervical vertebroplasty (cVP) procedures, we would like to temper the conclusions of the study in light of recent findings. The publication reported 140 cVP procedures on 176 vertebrae performed in 130 consecutive patients between 2001 and 2014. At the time of publication, no secondary fracture was depicted during the follow-up period in the study population. It was concluded that cVP is safe, with an acceptable major complication rate and effective in terms of pain relief at mid-term follow-up. Since the publication of our case series, four patients treated by cVP at our Institution suffered from secondary fractures in a short time-lapse (Fig. 1). Of note, all four of these cVPs were performed after the end of the inclusion period of the abovementioned publication. Interestingly, all these secondary fractures occurred after cVP of C2 vertebrae. We recall that C2 vertebra was one of the most frequently treated level in our previous cVP retrospective series (23% of the cases) [1]. The post-cVP C2 fractures recently observed at our Institution were all delayed (from 1 to 15 months; average 7.5 months). Only one of them resulted in neurological symptoms; the remaining 3 patients presented only pain recurrence. These fractures involved either the base of the C2’s dens or the dens itself. They were located above or below the cement cast in 3 cases, and through the cement in the remaining case. Maybe the specific biomechanical stress exerted on C2, especially on the dens through its articulation with the anterior arch of the C1 vertebra, could explain such fractures despite cementation with satisfactory lesion filing. Indeed, flexion-extension and rotation movements may exert specific constraints on the C2’s dens, which are not observed on other cervical levels, on which stress mostly consists in compressive loading [2]. It is well established that PMMA (polymethyl metacrylate) bone cement is mechanically resistant to compression stress but has a poor resistance to tension and flexion constraints [3]. This could thus explain the delayed fractures we report. This observation warrants further inves-