Primum non nocere, or “first do no harm,” is the well-known maxim from the Hippocratic corpus advocating that a physician’s first priority should be to avoid precipitating further injury, or inadvertent worsening of conditions, in patients who are under his or her care. With this founding precept of the health-care tradition in mind, I commend the efforts of Dr. Mandel and his colleagues in bringing to our attention the potential for epidural steroid injections to elevate the risk of vertebral body fracture in elderly individuals. Although this association has been alluded to in previous publications, to the best of my knowledge the study by Mandel et al. is the first scientifically rigorous effort to quantify the fracture risk associated with epidural steroid administration.
The study was conducted with use of a propensity-matched cohort with an average age of sixty-six years. Comparison between the patients who received epidural steroid and the controls in the primary analysis revealed that the risk of fracture increased by 21% per episode of steroid administration. Moreover, in the secondary analysis in which the possibility of multiple fractures in a single patient was considered, a significant elevation in the risk of subsequent vertebral fracture following steroid administration was again demonstrated. The identification of this significant association, as well as the quantification of risk inherent with each injection event, represent novel contributions to the literature.
It is necessary to appreciate these findings, however, within the context of other works. As previously mentioned, many authors have surmised that the use of epidural steroids may increase the risk of fragility fractures, although the evidence for this contention has been scant up to this point. For example, although the recent work of Kang et al. showed that bone mineral density decreased following epidural steroid injection, no specific association was drawn between this apparent reduction in bone mass and the onset of fracture1.
Of greater concern, the definable fracture risk as documented by Mandel et al. should be set against the best available evidence regarding the long-term efficacy of these interventions, which is admittedly less than robust2,3. Early studies maintained that epidural injections had poor clinical effectiveness. A similar result was reported in the prospective randomized trial by Cuckler et al., in which the authors concluded that “…a decision to use epidural steroids must be made with the realization that we failed to demonstrate its clinical efficacy in this study…”3 Proponents of steroid injection frequently cite the work of Saal and Saal4, but it should be emphasized that that study involved patients with disc herniations, and the results are not necessarily translatable to an elderly population that is more likely to suffer from symptoms of lower-extremity claudication related to spinal canal stenosis. Furthermore, more current research, including that of Cosgrove et al.2, failed to show lasting effects of epidural injections in the majority of patients, let alone in individuals of advanced age.
Although the available literature fails to support a substantial effect of steroid injection beyond the short-term period, the impact of vertebral fractures on quality of life as well as mortality is well understood5. Puisto et al. and other researchers found that vertebral fractures were associated with a demonstrable elevation in the risk of mortality for both men and women5. These facts, particularly when combined with the reported 21% increase in fracture risk per injection episode, raise the concern for adverse fracture-related events occurring as a result of epidural steroid administration, especially in the elderly. Moreover, the importance of this 21% elevation in the risk of fracture can be appreciated when it is contrasted with the recent, and highly publicized, article on the potential for ischemic heart disease in women following radiation therapy for breast cancer, which indicated a 7.4% increase in risk per gray of radiation6.
The investigation by Mandel et al. has limitations that the reader should appreciate and that potentially diminish the breadth of the conclusions that may be drawn from the study’s findings. As the authors recognize, their study was a retrospective work that depended on data entered into a registry and a great deal of heterogeneity consequently existed in the indications for steroid administration. The database itself was also not comprehensive in scope, and it appears to me that patients who were treated with epidural steroid in the Henry Ford health-care system but subsequently treated elsewhere for a vertebral fracture would have escaped recognition in the data set. This fact in and of itself could potentially confound some of the findings, and the rate of vertebral fracture following epidural steroid injections may be underestimated as a result.
Given its design and limitations, I would caution against the utilization of the study by Mandel et al. as a rationale for curtailing epidural injection procedures among all patients at present. Such interventions have been shown to be effective in patients with acute disc herniation4, and the risk of vertebral fracture highlighted by Mandel et al. may not be inherent in all individuals, especially those who are younger in age and/or possess normal bone mineral density. At a minimum, this laudable effort should be considered to represent a first step in the scientific process, essentially a call for further prospective research on the influence of epidural steroids on the risk of subsequent vertebral fracture. In addition, the results of this investigation can be used to inform the discussions and consent process prior to epidural steroid administration, particularly with respect to the risk-benefit profile in patients of advanced age and reduced bone mineral density. Clearly, much more work on this issue remains to be performed before definitive conclusions regarding the influence of epidural steroids on vertebral body fractures can be made. In the interim, however, practitioners would do well to consider the results presented by Mandel et al., as well as the Hippocratic dictum primum non nocere, when recommending epidural steroid injections in elderly patients.