Response to Letter to the Editor Regarding “Antiresorptive Medications Prior to Stereotactic Body Radiotherapy for Spinal Metastasis are Associated With Reduced Incidence of Vertebral Body Compression Fracture”

We thank the authors of this letter for their interest in our work. ¹ They discuss a few questions about our work, which we are happy to answer and address.

First, the authors note that the spinal metastasis patients receiving antiresorptive therapy initiated after SBRT had a higher risk of vertebral compression fracture (VCF) in comparison to those taking no antiresorptive medication at all. However, the log rank test comparing these two groups found that this difference was not significant. Specifically, VCF incidence for those initiating antiresorptive drugs after SBRT compared to those taking no antiresorptive drugs at all were 15% vs 8% at 6 months (P = .077), 17% vs 12% at 12 months (P=.167), and 17% vs 23% at 24 months (P=.580). Nonetheless, we agree that this trend is interesting. Our speculation is that the correlation may be directly related. In particular, patients that develop fracture may be more likely to receive an antiresorptive agent in the first year following radiation therapy.

Second, the authors ask to clarify components of the statistical analysis, including the reported P-value of .008 on Figure 1. This P-value was generated using a log rank test between two groups: the group of those taking antiresorptive agents before SBRT and the group taking none at all. As mentioned in the text, the P-value of .008 is reflective of the significant difference between the VCF incidence of 4% vs 23% at 2-years post-SBRT in the group initiating antiresorptive agents before SBRT compared to those taking none at all, respectively. This reported P-value is the generated value, and if it was modified according to the Bonferroni correction, the adjusted P-value would still reflect statistical significance. However, due to the substantial conflicting nature of the Bonferroni correction, we report the directly generated P-values to avoid several analytical weaknesses of its use, including the increased likelihood of type II errors. ² All analyses were performed using Stata (Version 15.1, College Station, Texas 77845). ³

In conclusion, although our study suffers from the limitations inherent to all retrospective analyses, the data suggests that antiresorptive agent administration prior to spinal SBRT may help to reduce the risk of VCF. Larger multi-institutional datasets and prospective studies will be important to confirm and better define these findings.