bs_bs_banner

Pain Medicine 2014; 15: 548–555 Wiley Periodicals, Inc.

SPINE SECTION Original Research Articles Comparative Effectiveness of Lumbar Transforaminal Epidural Steroid Injections with Particulate Versus Nonparticulate Corticosteroids for Lumbar Radicular Pain due to Intervertebral Disc Herniation: A Prospective, Randomized, Double-Blind Trial David J. Kennedy, MD,* Christopher Plastaras, MD,† Ellen Casey, MD,‡ Christopher J. Visco, MD,§ Joshua D. Rittenberg, MD,¶ Bryan Conrad, PhD,** James Sigler, MD,†† and Paul Dreyfuss, MD‡‡ *Department of Orthopaedic Surgery, Stanford University, Redwood City, California; †

Department of Physical Medicine and Rehabilitation, University of Pennsylvania, Philadelphia, Pennsylvania;

Disclosure: This research was supported by a research grant from the International Spine Intervention Society (ISIS). All authors contributed significantly to and approved the contents of the manuscript. The manuscript has not been previously published. Abstract Background. Lumbar transforaminal epidural injections are commonly utilized to treat radicular pain due to intervertebral disc herniation.



Rehabilitation Institute of Chicago, Northwestern University, Chicago, Illinois; § Department of Rehabilitation Medicine, Columbia University, New York, New York;



Department of Physical Medicine and Rehabilitation, Kaiser Permanente, Oakland, California; **Nike Sport Research Lab, Beaverton, Oregon;

††

Department of Physical Medicine and Rehabilitation, University of Kansas, Kansas City, Missouri;

‡‡

Department of Rehabilitation Medicine, University of Washington, Seattle, Washington, USA Reprint requests to: David J. Kennedy, MD, Department of Orthopaedic Surgery, Stanford University, 450 Broadway Street, Pavilion C, MC 6342 Redwood City, CA 94063, USA. Tel: (650) 725-5905; Fax: (650) 723-9805; E-mail: [email protected].

548

Objective. This study aims to determine if there was a major difference in effectiveness between particulate and nonparticulate corticosteroids for acute radicular pain due to lumbar disc herniation. Design. A multicenter, double blind, prospective, randomized trial on 78 consecutive subjects with acute uni-level disc herniation resulting in unilateral radicular pain. All subjects received a single level transforaminal epidural steroid injection with either dexamethasone or triamcinolone. Repeat injections were allowed as determined by the blinded physician and subjects. Primary outcomes included: number of injections received, surgical rates, and categorical pain scores at 2 weeks, 3 months, and 6 months. Secondary outcomes included mean Oswestry Disability Index. Results. Both triamcinolone and dexamethasone resulted in statically significant improvements in pain and function at 2 weeks, 3 months, and 6 months, without clear differences between groups. The surgical rates were comparable with 14.6% of the dexamethasone group and 18.9% of the

Particulate Versus Nonparticulate in Epidural Injections triamcinolone group receiving surgery. There was a statistically significant difference in the number of injections received, with 17.1% of the dexamethasone group receiving three injections vs only 2.7% of the triamcinolone group. Conclusions. Transforaminal epidural corticosteroid injections are an effective treatment for acute radicular pain due to disc herniation, and frequently only require 1 or 2 injections for symptomatic relief. Dexamethasone appears to possess reasonably similar effectiveness when compared with triamcinolone. However, the dexamethasone group received slightly more injections than the triamcinolone group to achieve the same outcomes. Key Words. Transforaminal Epidural Injection; Lumbar; Corticosteroid; Dexamethasone; Triamcinolone; Disc Herniation; Radiculopathy; Radicular Pain; Particulate

Park et al. published the first randomized controlled trial comparing dexamethasone to triamcinolone for lumbar radicular pain via a TFESI in 2010 in 106 patients [20]. Subjects that received triamcinolone had a statistically significant reduction in visual analog scale as compared with those that received dexamethasone. However, the two groups did not differ significantly on the McGill Pain Questionnaire or the Oswestry Disability Index (ODI) after treatments, and follow-up was for only 1 month. Kim et al. performed a study of 60 patients comparing the effectiveness of dexamethasone to methylprednisolone in lumbar interlaminar injections for the treatment of lumbar radiculopathy. They found that dexamethasone was statistically equivalent in decreasing pain when compared with methylprednisolone [21]. However, there was a nonsignificant trend toward shorter duration of pain relief with dexamethasone. A recent large retrospective study by El-Yahchouchi et al. also failed to show a difference in effectiveness between particulate and nonparticulate corticosteroids in over 3,000 injections [22]. However, this study was retrospective and did not stratify the results by type of compressive lesion.

Introduction Lumbar radicular pain due to a herniated intervertebral disc is a common and debilitating problem. A transforaminal epidural steroid injection (TFESI) is frequently used for treatment of radicular pain in patients that have failed conservative treatment [1]. While large studies of consecutive patients have failed to show any major complications [2], case reports have emerged documenting paralysis as a potential complication of this procedure [3–7]. The mechanism of injury is felt to be secondary to inadvertent intra-arterial injection of a particulate corticosteroid causing an embolic effect [8]. This mechanism of injury is collaborated by anatomic studies showing the arterial supply to the spinal cord transverses the intervertebral foramen [9–11], as well as light microscopy data demonstrating that particulates corticosteroids are either larger or aggregate to clusters that are larger than red blood cells [12,13]. Animal studies involving direct arterial injection have shown permanent central nervous system injury with particulate corticosteroids and no consequences with nonparticulate corticosteroids [14,15]. These data, combined with the lack of any documented cases of paralysis with dexamethasone, have led some to promote its routine use [16]. Others have questioned its efficacy due to the soluble nature of the nonparticulate corticosteroids [16]. Several studies have assessed the comparative effectiveness of different corticosteroid agents for cervical radicular pain, including one prospective study by Dreyfuss et al. [17] and two small retrospective studies by Lee et al. [18], and Shakir et al. [19]. Both retrospective studies showed similar effectiveness between dexamethasone and triamcinolone in cervical TFESIs, with no statistical difference in self-reported patient pain scores. The only prospective study by Dreyfuss involved 30 subjects. It showed a trend toward better effectiveness of the particulate steroids at 1-month phone follow-up, but this did not reach statistical significance due to the sample size.

There remains limitations in the available literature as to which corticosteroid preparation offers the best risk to benefit ratio for patients with lumbar radicular pain. There is currently no prospective data on the relative effectiveness of these corticosteroids through 6-month follow-up, as available studies only have 1-month outcomes. Also, no data exist on the effect of these preparations in preventing surgery, or the number of injections required to achieve an acceptable outcome. The aim of this study was to determine if there were any major differences in the effectiveness of particulate vs nonparticulate corticosteroids inclusive of 6-month follow-up. Additional goals included evaluating the ability of these medications to prevent surgery, and the number of injections received to achieve an acceptable outcome. Methods This study was conducted at two academic medical centers with the full approval of each Institutional Review Board (IRB) with all subjects having signed a written informed consent. During the recruitment period (December 2008–December 2012), 81 consecutive subjects were offered enrollment, and a total of 78 subjects were randomized. Twenty-five subjects from the Rehabilitation Institute of Chicago and 53 from the University of Florida were enrolled. Inclusion criteria included: Adult subjects were capable of providing informed consent and deemed appropriate for an epidural injection by their treating Physical Medicine and Rehabilitation physician or spine surgeon for single leg radicular pain with pain intensity of ≥4/10 and pain of less than 6 months’ duration. MRI had to demonstrate a single level herniated nucleus pulposus below L3 that corresponded with the patient’s clinical features. All subjects were deemed appropriate for surgery if the lumbar transforaminal epidural steroid injection failed to provide adequate pain relief. Exclusion criteria included: Back pain greater than leg pain, nonradicular 549

Kennedy et al. pain, unclear diagnosis, or more than one potential pain generator demonstrated on MRI, lumbar stenosis, prior spine surgery, prior spine injection, any condition that increased injection risk such as bleeding tendencies, workers compensation, pregnancy, and litigation. At enrollment, baseline demographics and outcome measures were obtained including: 11-point pain Numeric Rating Scale (NRS), ODI, and other medical interventions tried for their current pain including medication use, physical therapy, chiropractic care, and massage. Each subject underwent a single, unilateral lumbar TFESI consistent with imaging and the patient’s clinical presentation. To facilitate double blinding, the injecting physician was not involved in research follow-up, or future treatment decisions including repeat injections or surgical referral. Procedures were all performed per the International Spine Intervention Society guidelines [23]. All subjects were placed prone on the injection table. After sterile preparation with betadine or chlorhexidine, the area was draped and the skin was anesthetized with 1% lidocaine. Using fluoroscopic guidance, a spinal needle was advanced in an oblique view to the safe triangle. Both anterior-posterior and lateral views were obtained to confirm precise needle placement within the intervertebral foramen at the 6 o’clock position under the pedicle; ideally, the needle was placed adjacent to the back of the vertebral body immediately inferior to the pedicle. At the target level approximately 1 mL of contrast medium (Omnipaque 240; GE Healthcare, Princeton, NJ, USA) was injected while being visualized with real-time fluoroscopy to assure target medication flow and the absence of vascular or subdural or subarachnoid flow. Digital subtraction technology was also utilized per the injecting physicians preference. In the absence of aberrant flow, 2 mL of 1% lidocaine was injected as an anesthetic test dose. After waiting 2 minutes to assure no adverse events, the subjects were injected with the treatment corticosteroid. All subjects were randomly assigned to receive either 1.5 mL of dexamethasone phosphate 10 mg/mL (Decadron; Merck, Whitehouse Station, NJ, USA) or 1.5 mL of triamcinolone acetonide (Kenalog; Bristol-Meyers Squibb, Princeton, NJ, USA) 40 mg/mL, with assignment by computer-generated random order (http://www.randomizer.org/form.htm). These doses were chosen to have equal volumes of injectate, as well as equal potency. Follow-up was by blinded treating physicians at 14–21 days, 3 months, and 6 months. In addition to obtaining outcome measures, the treating physician could order repeat injections or refer the subject to surgery. Further treatment was based upon subject’s pain and functional limitations. Up to a total of three injections were allowed, and follow-up restarted after each injection. Subjects were reimbursed $50 for the 2-week visit, $100 for the 3-month visit, and $150 for the 6-month visit. In the rare cases when in person visits were not obtained, phone follow-ups were conducted by a blinded research assistant to obtain the validated outcome measures. These phone follow-ups did not result in a decision for further care. Even if multiple 550

injections were done the total reimbursement was limited to $300, to not entice the subject to receive more interventional care for monetary gain. The primary outcomes were: the need for surgery, the number of injections received, and the categorical proportion of patients that obtained ≥50% pain reduction at 2 weeks and 3 and 6 months follow-up. Surgical intervention and loss to follow-up were considered failure of treatment and calculated as such. Secondary outcomes included the ODI. The proportions of each group obtaining such were compared using 95% confidence intervals (CIs) and to be considered significantly different the 95% CIs could not overlap. The Mann–Whitney test was used for baseline comparisons between groups, and a paired t-test was used for comparisons within groups. Results Seventy-eight subjects were enrolled and randomized into two groups, with 41 subjects receiving dexamethasone and 37 subjects receiving triamcinolone. Baseline characteristics did not differ in respect to age, gender, duration of pain, disc morphology (protrusion vs extrusion), segment treated, mean 3-day 11-point NRS, or ODI (Table 1). Table 2 denotes the treatments tried prior to enrollment in the study. Tables 3 and 4 contain pain data and outcomes for subjects receiving dexamethasone and triamcinolone, respectively. Table 5 contains the ODI and categorical pain data for both groups. At the 7–14 day follow-up, both groups showed clinically and statistically significant improvement in pain and functional measures. A greater percentage of subjects receiving triamcinolone achieved ≥50% pain relief at 2 weeks than those receiving dexamethasone (43.2 vs 31.7%); however, this did not reach statistical significance and the 95% CIs were overlapping. This trend disappeared by 3 and 6-month followup, with greater than 70% of both groups achieving at least 50% pain reduction with no differences between groups. ODI data also improved in each group without reaching a statistically significance difference between groups. With respect to disability as self-rated by the ODI, average ODI measures for both groups moved from the “severe disability” range (score of 40–60) to the “minimal disability” range (score of 0–20) from baseline to 6 months follow-up. The surgical rates also did not reach a statistically significant difference with 14.6% of those receiving dexamethasone, and 18.9% of those receiving triamcinolone having had surgery. The mean duration of pain was not different in the surgical and nonsurgical subgroups. Disc morphology did differ between the surgical and nonsurgical subgroups. The 53% (95% CI = 43.2%–82%) of those undergoing surgical intervention had disc extrusions vs 30.9% (21.1%–42.7%) extrusions in the nonsurgical group. The average number of injections received for each group was 1.6 for dexamethasone and 1.4 for triamcinolone (Table 6). However, when the data were analyzed categorically, it was found that 17.1% of subjects that receiveddexamethasone had three injections vs 2.7%

Particulate Versus Nonparticulate in Epidural Injections

Table 1

Baseline data

Female (%) Mean age (years) Mean duration of pain (weeks) Mean pain (0–10) Mean ODI Level injected L4/L5 L5/S1 S1/S2 Disc morphology Extrusion (%)

Dexamethasone (N = 41)

Triamcinolone (N = 37)

34.1% (N = 14) 35.9 Range (19–51) 10.1 weeks Range (3–20) 6.3 Range (4–10) 45.7 Range (14–80)

35.% (N = 13) 35.6 Range (20–58) 8.6 Range (1–20) 6.5 Range (4–10) 42.4 Range (14–74)

N=6 14.6% N = 23 56.1% N = 12 29.2%

N=5 13.5% N = 20 54.1% N = 12 32.4%

0.093

N = 11 26.8%

N = 17 45.9%

0.058

of those that received triamcinolone (P < 0.05 and nonoverlapping 95% CIs). Discussion A contemporary review found TFESI are a very effective treatment for disc herniation [24], and this study supports

Table 2

Treatments tried prior to injection Dexamethasone Triamcinolone (N = 41) (N = 37) P

N = 33 80.4.% NSAIDs N = 32 78.0% Acetaminophen N = 19 46.3% Oral opioids N = 17 41.4% Oral steroids N = 14 34.1% Muscle relaxers N = 10 24.4% Neuropathic pain N = 10 medicines 24.4% Chiropractic N=8 manipulation 19.5% Massage therapy N = 6 14.6% Physical therapy

N = 34 91.9% N = 26 70.2% N = 18 48.6% N = 15 40.5% N = 11 29.7% N=9 24.3% N=5 13.5% N=5 13.5% N=9 24.3%

P 0.660 0.840 0.188 0.599 0.473

0.098 0.094

that conclusion. At baseline, every subject failed at least one conservative therapy, with most subjects failing a variety of conservative therapies. Collectively, both groups had >70% of subjects with ≥50% pain reduction at 6 months follow-up. This is inclusive of subjects that were either loss to follow-up or received surgery and included as failures. In fact, when reviewing the data in Tables 3 and 4, it is apparent that the vast majority of subjects had near complete pain relief by 6 months. Of note in this study over 50% of subjects received only one single level TFESI, further clarifying that there is no indication for a routine series of three injections or multilevel injections for single disc herniation.

0.12 0.34 0.86 0.97 0.50 0.91 0.18 0.37

In terms of comparative data, this study further corroborates the existing literature. Most studies show no statistically significant difference in outcomes between dexamethasone and particulate corticosteroids, although many have trends favoring particulate corticosteroids [17,18,21,24], but the study by Park et al. [20] is the only study that reached statistical significance. These studies were all limited a follow-up of 1 month or less. This mirrors the data from this study showing a nonsignificant trend favoring triamcinolone at 7–14 days. However, when the subjects were followed for 3–6 months, all differences in effectiveness disappear. Furthermore, there is no significant difference in surgical rates between groups, but there was a trend of less surgery in those receiving dexamethasone.

0.19

The only statistically significant finding in this study was that a greater percentage of subjects that received

551

Kennedy et al.

Table 3 Distribution of 3-day average of 11-point pain Numeric Rating Scale (NRS) for those receiving dexamethasone group at baseline and all follow-up points Dexamethasone (N = 41) Baseline

7–14 Days

3 Months

6 Months

Mean 3-day NRS

Frequency

%

Frequency

%

Frequency

%

Frequency

%

0 1 2 3 4 5 6 7 8 9 10 Surgery Loss to follow-up Mean

0 0 0 0 4 6 7 5 8 8 3 0 0 7.05

0 0 0 0 9.7 14.6 17.1 12.2 19.5 19.5 7.3 0 0

1 4 6 5 7 9 4 3 1 0 1 0 0 4.07

2.4 9.8 14.6 12.2 17.1 22 9.8 7.3 2.4 0 2.4 0 0

15 8 2 3 1 2 1 1 1 0 1 6 0 1.61*

36.6 19.5 4.9 7.3 2.4 4.9 2.4 2.4 2.4 0 2.4 14.6 0

16 7 3 2 0 2 0 0 1 0 0 6 4 1.37*

39 17.1 7.3 4.9 0 4.9 0 0 2.4 0 0 14.6 9.7

* Average pain score minus surgical cases and loss to follow-up.

dexamethasone underwent three injections. Although it affected only a small number of subjects, this could be due to a large number of reasons. Possibly, dexamethasone is slightly less effective or has a shorter duration of effect than triamcinolone. This would help explain the trend favoring triamcinolone at 7–14 day

follow-up as well. The greater number of injections received by the dexamethasone group may also explain the lower surgical trends in that group. The ultimate question when determining the ideal corticosteroid preparation is based on a risk to benefit calculation

Table 4 Distribution of 3-day average of 11-point pain Numeric Rating Scale (NRS) for those receiving triamcinolone group at baseline and all follow-up points Triamcinolone (N = 37) Baseline

7–14 Days

3 Months

6 Months

Mean 3-day NRS

Frequency

%

Frequency

%

Frequency

%

Frequency

%

0 1 2 3 4 5 6 7 8 9 10 Surgery Loss to follow-up Mean NRS

0 0 0 0 5 4 7 5 9 5 2 0 0 6.86

0 0 0 0 13.5 10.8 18.9 13.5 24.3 13.5 5.4 0 0

2 5 7 5 3 4 2 3 0 2 3 1 0 4.08*

5.4 13.5 18.9 13.5 8.1 10.8 5.4 8.1 0 5.4 8.1 2.7 0

14 4 4 1 4 0 1 1 0 1 0 6 1 1.77*

37.8 10.8 10.8 2.7 10.8 0 2.7 2.7 0 2.7 0 16.2 2.7

16 3 4 2 1 0 1 0 1 0 0 7 2 1.25*

43.2 8.1 10.8 5.4 2.7 0 2.7

* Average pain score minus surgical cases and loss to follow-up.

552

2.7 0 0 18.9 5.4

Particulate Versus Nonparticulate in Epidural Injections

Table 5 Mean scores for Oswestry Disability Index (ODI) and categorical 50% pain reduction. Those that were lost to follow-up or received surgery were calculated as nonresponders. All values had P < 0.05 when compared with baseline but not when comparing between groups

While it is clear that particulate corticosteroids are associated with the risk of paralysis, the risk of this complication still pales in comparison with the rate of complications with lumbar spine surgery [31], which is often the next step if this treatment fails. Therefore, until even more definitive research on different patient populations is available, particulate corticosteroids may continue to have a role with TFESIs.

7–14 Days

There are several limitations for this study. First, it was not powered to find small differences in effectiveness between corticosteroid preparations. Initial power calculations called for N = 48 in each group to utilize dichotomous outcomes to show differences of 60% and 30%. This study was terminated before this target was reached because the statement “Not for epidural use” was added to triamcinolone’s label. The institutional IRB therefore recommended the study be ceased. Second, this study represents a highly selective group of subjects, and the results may not be generalizable to all compressive lesions. Additionally, firm criteria were not established for the decision to repeat an injection or proceed to surgical intervention. Given the variety of concomitant noninterventional treatments and different treating physicians involved, it is possible that this could have affected the rates of repeat injections or even surgical interventions. This lack of control of confounders simultaneously represents an intrinsic strength to this study as it represents real life clinical scenarios. It is also possible that scheduled repeat injections could result in different outcomes due to a cumulative effect; however, the majority of subjects in this study did very well with a limited number of injections. This study was designed as a comparative effectiveness study between two different steroids; a different study design would be required to explore the timing, frequency, and number of epidural injections. Also only one particulate corticosteroid formulation was studied (triamcinolone). This decision was based on the two published retrospective studies that showed betamethasone to be inferior to triamcinolone, calling into question its utility [32,33]. Methylprednisone, the other commonly utilized particulate corticosteroid, also has other potential added risks if the preservatives within that preparation is injected into the intrathecal space [14]. Lastly, it is possible that different doses could have varying effects, and the studies to date have used a variety of dosing.

3 Months

6 Months

Dexamethasone (N = 41) 51% Reduction in ODI 26.8% 50% Pain reduction 31.7%

68.2% 73.2%

70.7% 73.2%

Triamcinolone (N = 37) ODI 50% Pain reduction

67.6% 73%

64.9% 75.7%

35.1% 43.2%

for a given patient and society as a whole. This study does help add valuable information in determining the possible benefits for a select patient population. While placing a needle near the spinal column will never be completely risk free, there are several techniques available that can help reduce the risk of paralysis in a given patient. It is clearly essential to utilize live real-time fluoroscopy for all TFESIs [25]; however, there are several additional published techniques that may further reduce the risk of unrecognized intra-arterial flow. These include the use of digital subtraction technology [26], an anesthetic test dose [27], and an infra-neural approach [28]. If a practitioner feels that a particulate corticosteroid is ideal for a particular patient, these techniques should be considered to further reduce the risks associated with this injection. It is also clear that proper training does affect a practitioners ability to recognize arterial flow [29]; however, the ideal training method or quantity has not been formalized [30].

Table 6

Number of injections received

Number of Injections

Dexamethasone (N = 41) Dexa

Triamcinolone (N = 37) Triam

1 injection

N = 22 53.6% 95% CI = 38.8–67.9%

N = 23 62.1% 95% CI = 46.1–75.9%

0.33

2 injections

N = 12 29.3% 95% CI = 17.6–44.5%

N = 12 32.4% 95% CI = 19.6–48.5%

0.47

3 injections

N=7 17.1% 95% CI = 11.1–23.2%

N=1 2.7% 95% CI = 0–10.1%

0.0052

Bold italics indicate statistical significance if P

Comparative effectiveness of lumbar transforaminal epidural steroid injections with particulate versus nonparticulate corticosteroids for lumbar radicular pain due to intervertebral disc herniation: a prospective, randomized, double-blind trial.

Lumbar transforaminal epidural injections are commonly utilized to treat radicular pain due to intervertebral disc herniation...
150KB Sizes 0 Downloads 0 Views