Digital subtraction angiography versus real-time fluoroscopy for detection of intravascular penetration prior to epidural steroid injections: meta-analysis of prospective studies.

BACKGROUND: Neurological injury is a rare but devastating complication of epidural steroid injections (ESIs) generally thought to arise from neurovascular compromise. The use of real-time fluoroscopy (RTF) with contrast media is the most common preventative measure taken to avoid intravascular penetration. In 2002, it was proposed that digital subtraction angiography (DSA) might be more useful than RTF. Since then, several prospective studies have advocated for its use. OBJECTIVES: As DSA is not currently a "gold standard," a meta-analysis was performed to compare the efficacy of DSA versus RTF for detection of intravascular penetration during ESI. STUDY DESIGN: Meta-analysis of prospective observational studies. METHODS: A targeted Pubmed search was conducted, yielding 49 reports and 4 manuscripts, which were analyzed using Review Manager Software (Rev Man 5.2). Inclusion/exclusion criteria: peer-reviewed prospective reports comparing the sensitivity of DSA to RTF in the same individuals without change in needle position between comparative imaging. Pooled estimate of odds ratios with 95% confidence interval using a random effect model was applied. RESULTS: There were a total of 188 intravascular events from 1,290 ESIs performed. RTF was able to detect 148 events with DSA detecting an additional 40 events missed by RTF. No major neurological complications were reported. DSA had a statistically significant favorable odds ratio over RTF for detection of intravascular penetration during ESI (OR = 1.32 [1.05 - 1.67]; P = 0.02). LIMITATIONS: Although the major methodological aspects of each study assessed in this meta-analysis were quite similar, there were small differences in needle gauge and the selection of secondary outcome measures. Despite attempts to minimize it, concern for operator bias also exists. CONCLUSIONS: DSA had a 32% improvement (OR = 1.32) for detection of intravascular penetration with ESI when compared to RTF. Although this supports advocacy for use of DSA, it also suggests that there is a greater than 30% "missed-events" rate for detection of vascular penetration when using RTF for ESI, which does not correlate with the generally reported cumulative rates of complications (1%). This discrepancy suggests that factors other than vascular events also play a role in complications. Nonetheless, given the evidence, we advocate for the increased use of DSA over RTF for transformational ESIs.