The Incidence of Lumbar Discectomy after Epidural Steroid Injections or Selective Nerve Root Blocks

Discussion

There is a debate in the literature about which type of spinal injection ESI versus SNRB is more effective for relieving pain and possibly preventing surgery. Both the Buttermann and Riew studies found a 50% crossover rate in their studies to the surgical treatment arm.^{2, 20, 21} Our examination of a much larger, but far less controlled population, found a much smaller overall progression rate to surgery, but similar rates between both injection methods utilized. This finding contradicts the possibility that one method of injection for lumbar disc herniations is more effective than another in the avoidance of surgery.

The reader should be cautioned that this study is limited by its design. An obvious flaw of this database review is that without radiographic confirmation of the disc herniation and stringent review criteria for the reporter's surgical indications, it is difficult to conclude the real number of herniated discs that would have been treated surgically as opposed to being simply offered an injection of any type. Keeping this in mind, the data in the current study demonstrates the rate of lumbar discectomy after either a SNRB or an ESI in this study is smaller than reported in prior randomized control trials. This may suggest that either ESIs or SNRBs are more beneficial than previously suspected. Another reason may be that physicians are administering more injections during the course of non-operative management for a disc herniation. In both the Buttermann and Riew studies, ^{2, 20, 21} the patients that were included demonstrated discrete radiographic evidence of an acute and compressive disc herniation. Conservative measures, excluding injection, were attempted for at least six weeks before inclusion in these studies.^{2, 20, 21} Our study found an average time from diagnosis to injection of approximately eight weeks, 56.8 days, for patients receiving an ESI and 61.7 days for patiently receiving an SNRB. However, a large portion of both groups had the patient's injection listed as their initial entry into the database. If our analysis had included those patients whose initial entry into the database was their injection procedure, the average time from diagnosis to injection would have been significantly lowered to 36.9 days for patients receiving ESI and 45.7 days for those receiving SNRB.

This review simply examined patients who had a diagnosis of a disc herniation from a physician. ESI and SNRB are both treatments intended to reduce the inflammation. Our study did not consider the type of injection chosen and may have depended on the perceived need, the patient's HNP, or physician preference. Typically, a translaminar injection is considered when multiple nerves are affected by a disc herniation or if there is a need to treat a larger area of compression as opposed to a selective nerve root injection where a clearly identified area of pathology or the response to an injection is needed to diagnose the affected area. Based on the current study design, there is no way to determine the criteria used by the diagnosing physician, nor is there a means to determine the methods of conservative treatment used prior to their procedures. As noted in Buttermann's study, the patients who had a successful ESI were twice as likely to have an extruded or sequestered disc (57% compared to 26%, p = 0.036), and were more likely to have a hydrated herniated disc (a high signal on T2-weighted magnetic resonance image, p = 0.0075). As pointed out in a recent review of MRI images from the SPORT trial ²², there is excellent inter-reader reliability for disc morphology; however, only moderate reliability for thecal sac compression, as well as, nerve root impingement. These characteristics, in particular, the compressive nature of the disc herniation, were important inclusion criteria in both the Riew and Buttermann studies. These variables could not be examined in our database analysis; and are therefore, a limitation to our analysis.

Additionally, our study found a statistically significant difference in the discectomy rate for patients receiving an ESI by year. In 2004, 10.7% of patients received a discectomy after an ESI injection, while only 7.8% of patients received a discectomy after an ESI in 2006. There was also a trend toward a decrease in the crossover rate for SNRB that did not reach statistical significance. This finding may perhaps indicate improved effectiveness for the injection procedures. It may also indicate a trend toward more injections in a population that otherwise would not have been in danger of proceeding to surgery and would have recovered with continued conservative measures that did not include injections.

There was a small gender difference in crossover rate for both ESI and SNRB that was found to be statistically significant. However, looking at gender differences in discectomy rates for the entire study population, only 4.8% of women proceeded to a discectomy, while 6.7% of men proceeded to discectomy.

Age was a significant factor influencing crossover rates for both ESI and SNRB. It appeared that patients were less likely to proceed to surgery after either type of injection if they were older, particularly if they were more than 50-60 years of age. This trend of a decreased conversion to discectomy was seen in the entire study population with the diagnosis of a disc herniation, but not to the extent that is was seen in the injection populations. This may indicate that patients who were older were more inclined to consider an injection, but less inclined to consider surgery. It also suggests a potential for injections to be more efficacious in treating radiculopathy in older patients. However, this cannot be concluded with this study design.

Geographic variation was also noted within our study. Similar to the conclusions of Friedly and colleagues, with respect to ESIs, patients treated in the southern region of the country were more likely to undergo an ESI.²³ There was also a statistically higher discectomy rate in the southern region of the country. In both the study by Friedly et al. and our study, there was a lower rate of ESI and discectomy in the northeast region of the country.