- Email: [email protected]
Topiramate in Chronic Lumbar Radicular Pain
Topiramate in Chronic Lumbar Radicular Pain Suzan Khoromi,* Athos Patsalides,† Suzan Parada,‡ Vesta Salehi,* Jennifer M. Meegan,* and Mitchell B. Max* Abstract: Chronic lumbar radicular pain is the most common neuropathic pain syndrome. This was a double-blind, randomized, 2-period crossover trial of topiramate (50 to 400 mg) and diphenhydramine (6.25 to 50 mg) as active placebo to assess the efficacy of topiramate. Each period consisted of a 4-week escalation, a 2-week maintenance at the highest tolerated dose, and a 2-week taper. Main outcome was the mean daily leg pain score on a 0 to 10 scale during the maintenance period. Global pain relief was assessed on a 6-level category scale. In the 29 of 42 patients who completed the study, topiramate reduced leg pain by a mean of 19% (P ⴝ .065). Global pain relief scores were significantly better on topiramate (P < .005). Mean doses were topiramate 200 mg and diphenhydramine 40 mg. We concluded that topiramate treatment might reduce chronic sciatica in some patients but causes frequent side effects and dropouts. We would not recommend topiramate unless studies of alternative regimens showed a better therapeutic ratio. Perspective: The anticonvulsant topiramate might reduce chronic lumbar nerve root pain through effects such as blockade of voltage-gated sodium channels and AMPA/kainite glutamate receptors, modulation of voltage-gated calcium channels, and gamma-aminobutyric acid agonist-like effects. © 2005 by the American Pain Society Key words: Anticonvulsants, chronic sciatica, randomized controlled trials, neuropathic pain.
A
lthough radiculopathy caused by lumbar disc disease is the most common chronic neuropathic pain syndrome, with a prevalence of 4.5% in individuals older than the age of 30,17 there are no published repeated dose-controlled analgesic trials in this disorder. Reports are limited to a few subset analyses in studies of a broader selection of patients with back pain. Dellemijn et al6 reported that mean pain reduction in an open-label study of transdermal fentanyl in 18 patients with lumbar radiculopathy was 32%. Atkinson et al1 suggested that the small subgroup of patients with lumbar radiculopathy (n ⫽ 11) responded more favorably to nortriptyline than placebo within a larger study including mixed back pain etiologies. Remmers et al28 reported no relief of either leg or back pain by pregabalin in a placebo-controlled trial. Medrik-Goldberg et al24 found that acute lidocaine infusion significantly reduced pain Received February 23, 2005; Revised May 25, 2005; Re-revised July 13, 2005; Accepted August 2, 2005. From the *Pain and Neurosensory Mechanisms Branch, National Institute of Dental and Craniofacial Research; and the Departments of †Radiology and ‡Nursing, Warren Magnuson Clinical Center, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland. Supported by NIDCR Intramural Research Grant Z01 DE00366 in addition to partial support of a data technician by an Ortho McNeil educational grant. Dr Max had previously served as a paid consultant for Ortho McNeil but resigned before planning the study. Ortho McNeil had no input into the study design, data interpretation, or manuscript preparation. A preliminary report of these results was presented at the Tenth World Congress on Pain, San Diego, California, August 19, 2002. Address reprint requests to Suzan Khoromi, MD, MHS, National Institutes of Health Building 10, 4-1741, Bethesda, MD 20892-1302. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2005.08.002
in 30 patients with sciatica when compared with placebo and amantadine. Despite the fact that chronic lumbar radiculopathy is at least 20 times more common than diabetic neuropathy and postherpetic neuralgia, most of the current evidence for neuropathic pain treatment is derived from the latter 2 pain syndromes.9,32,40 Generalizing these results to lumbar radiculopathy pain might be inappropriate because the site and pathology of these conditions differ. The lesion in most lumbar radiculopathies is a focal compressive37 and inflammatory injury of the nerve root14,23 and dorsal root ganglion.25 Therefore the pain might have a different pathophysiology and treatment response than either diabetic neuropathy, where the more diffuse lesions have metabolic34 and vascular causes,39 or postherpetic neuralgia, where the multifocal lesions are infectious and inflammatory.4 For this study we chose the anticonvulsant topiramate, which combines multiple mechanisms of action including voltage-gated sodium channel blockade and modulation,31 gamma-aminobutyric acid agonist-like effects,2 AMPA/kainate glutamate receptor blockade,30 and modulation of voltage-gated calcium ion channels.41 Topiramate was more effective than placebo in relieving pain associated with diabetic neuropathy in 2 of 3 randomized trials in patients with painful diabetic neuropathy.10,27,33 A study of topiramate (50 to 400 mg) in 27 patients10 and a dose-response study of topiramate 100, 200, and 400 mg in 1259 patients33 found no difference in pain relief between topiramate and placebo, whereas another study found significant pain reduction in 323 patients on topiramate (50 to 400 mg) compared with
The Journal of Pain, Vol 6, No 12 (December), 2005: pp 829-836
829
830
Topiramate in Chronic Sciatica 27
placebo. In addition, anecdotal reports indicate that topiramate might reduce pain in trigeminal neuralgia16,42 and intercostal neuralgia.3 The aim of this study was to evaluate the efficacy of topiramate in the treatment of chronic sciatica, adapting methods that have been previously validated in studies of diabetic neuropathy22 and postherpetic neuralgia.
Materials and Methods Study Population This study was conducted at the National Institutes of Health (NIH) Clinical Center. The protocol was approved by the National Institute of Dental and Craniofacial Research Institutional Review Board. Patients were recruited between March 2001 and March 2002 through local newspaper advertisements. Of 500 phone responders, 455 had back pain alone, pain location and quality that were typical of myofascial pain in the lower extremities (localized stiffness and dull aching pain), or were not interested in a drug treatment study. Ninety percent of the candidates were excluded because their features were more consistent with myofascial pain, which is consistent with the reported frequency of this etiology for low back pain seen in the primary care setting.7,8 Fortytwo of the 45 respondents with radiating leg pain who were willing to participate in a drug trial met the inclusion and exclusion criteria, and all subjects signed an informed consent form.
defined by American College of Rheumatology criteria38; (7) polyneuropathy and peripheral vascular disease; (8) history of nephrolithiasis; and (9) narrow angle glaucoma.
Imaging and Laboratory Evaluation Patients submitted an MRI scan of the lumbosacral spine taken within a year before study enrollment, or this was performed on study entry. One neuroradiologist blinded to the patients’ symptoms reviewed the films, commenting on degenerative disc or joint disease and other findings contributing to low back pain or radiculopathy according to definitions and classifications proposed by the American Society of Neuroradiology.12 To scrutinize the relationship between treatment response and imaging diagnosis, we assigned patients to 9 radiologic categories. Patients were classified as lateral recess syndrome, neural foramen stenosis, canal stenosis, or their combination if their clinical findings and MRI scan offered a consistent anatomic explanation of their root symptoms resulting from compression of roots due to degeneration of joint, facet, and/or disc. Patients with no visible root compression were classified as degenerative disc disease and/or degenerative joint disease. Laboratory evaluation included a complete blood count with differential, sedimentation rate, antinuclear antibody titer, and rheumatoid factor to exclude inflammatory arthritis, cancer, and spinal infection.
Screening for Somatoform Disorder Inclusion Criteria First criterion was evidence of lumbar radiculopathy, on the basis of the presence of pain in one or both buttocks or legs for 3 months or greater for at least 5 days a week and at least one of the following features on the side corresponding to leg pain: (1) sharp and shooting pain below the knee; (2) pain evoked by straight leg raising to 60 degrees or less; (3) decreased/absent ankle reflex; (4) weakness of muscles below the knee; (5) sensory loss in L5/S1 distribution; (6) electromyographic evidence for L4, L5, or S1 root denervation; (7) imaging (magnetic resonance imaging [MRI], computed tomography/myelogram) evidence of nerve root compression in the lower lumbar region. The second criterion was average leg pain of at least 4 of 10 for the past month. The third criterion was willingness to refrain from making changes in non-study medications taken for sciatica. The fourth criterion was age between 18 and 75 years at the start of the study. The principal investigator performed the clinical examinations. Sensory testing included responses to pinprick, light touch, vibration, and joint movement.
Exclusion Criteria The exclusion criteria consisted of (1) hepatic and renal dysfunction; (2) pregnancy or lactation; (3) seizure disorder; (4) pain of greater intensity in any other location than the low back or leg; (5) narcotic abuse and/or drug or alcohol abuse during the past year; (6) fibromyalgia as
Patients completed a 15-item questionnaire, the PHQ15, devised by Kroenke et al21 to assess for multisomatoform disorder.
Study Design We carried out a double-blind, 2-period, crossover study comparing topiramate with active placebo, diphenhydramine. We chose diphenhydramine as active placebo because of its side effects, mainly sedation and anticholinergic effects, some of which overlap with topiramate. Drug randomization was performed by an NIH pharmacist according to blocked randomization in blocks of 8 and 4 within a table of 32 random numbers before patient enrollment. Each period consisted of 4 weeks of drug titration and 2 weeks of maintenance, followed by a 2-week washout, during which doses were gradually tapered (Fig 1). Topiramate was started at 50 mg at bedtime. Doses were escalated to 50 mg in 2 divided doses during week 1, by 50-mg increments in each of AM and PM doses during weeks 3 and 4 for a maximum of 400 mg. Doses of the study drug were based on a prior multicenter trial of topiramate in painful diabetic neuropathy.27 Diphenhydramine was started at 6.25 mg at bedtime, increased to 6.25 mg twice a day during week 1, and then increased by 6.25-mg increments in each dose during week 2 and by 12.5-mg increments in each dose during week 3 to a maximum of 50 mg a day in 2 divided doses. If side effects associated with drug escalation interfered with any activities or if the patient found the
ORIGINAL REPORT/Khoromi et al
831
Figure 1. Study design.
side effects unpleasant, we decreased the patient’s dose to the previous level.
Outcome Assessment Each day at bedtime, patients rated their pain during the last 24 hours by using a 0 to 10 numeric scale.19 They were asked to assess pain in 6 separate categories: average back, leg, and overall pain (leg and back) and worst back, leg, and overall pain. For people who had pain in both legs, pain on the worse side was recorded. During the patients’ follow-up visit to the Clinical Center at the end of each period, the study nurses collected the pain diaries and questionnaires from the patients. They then entered these data into a database from which all study outcomes were derived. Side effects elicited during semiweekly phone interviews were rated for severity as mild, moderate, or severe, and they were then entered into a database. For the primary outcome, we chose the comparison of mean scores for average leg pain during the 2 weeks of maintenance on topiramate and placebo. In the original study protocol, we defined the primary outcome to be the comparison of mean scores for average leg pain during the 2 weeks at the final maintenance doses for topiramate versus placebo. We chose leg pain over back pain because we are testing the principle that topiramate relieves nerve pain, and leg pain is more likely than back pain to reflect nerve pathology. We have examined the scores in the final 2 weeks in this study and in most of our group’s previous clinical trials for several reasons. (1) Jensen and McFarland20 showed that an average of 7 to 14 pain scores proves a more accurate estimate of true mean pain than a single measurement, but additional scores do not improve the fit. (2) Comparative treatment effect is often highest during the final weeks of treatment if the patients have been slowly titrated along a rising dose-response curve. (3) Exclusion of the first sev-
eral weeks of the treatment period lessens the chance of pharmacologic carryover effect in a crossover design. Other outcome measures were (1) global pain relief (leg and back pain combined) by using a categorical pain scale rating overall pain outcome as worse, no relief, slight, moderate, a lot, and complete relief; (2) the Oswestry Low Back Pain Disability questionnaire11; (3) the Beck Depression Inventory36; and (4) the 36-item Short Form of Health Survey (SF-36), a general health status instrument that measures the social, mental, and emotional dimensions of health and illness.35
Blinding The NIH outpatient pharmacy was in charge of drug randomization, and patients and research staff were blinded to the randomization order.
Statistical Analysis Pain score outcomes from the SF-36, the Oswestry Low Back Pain Disability questionnaire, and the Beck Depression Index were compared by using paired, 2-tailed paired t tests. We performed all statistical tests with the analyst function in Statistical Analysis Software (SAS, Version 8; SAS Institute Inc, Cary, NC). We assessed for treatment-period interactions (carryover effect) by using 2 methods: a t test variant proposed by Pocock26 and inspection of the interaction term from a factorial analysis of variance (ANOVA) model in which treatment and period were independent variables and average leg pain the outcome score. We used global pain relief (GPR) scores of moderate or better for each drug as the response criterion to calculate the number needed to treat, NNT ⫽ 1/([proportion of patients in the treatment group with GPR score of moderate or better] ⫺ [the proportion of patients in the placebo group with same GPR scores]).13 NNT, the reciprocal of the absolute risk reduc-
832 Table 1.
Topiramate in Chronic Sciatica
Demographics STUDY COMPLETERS
Sex (F/M) Age (y) Median Range Work status (retired/nonretired) Pain duration (y) Median Range MRI diagnosis Neural foraminal stenosis Canal stenosis Lateral recess syndrome DJD/DDD No MRI Physical signs Sensory change Weakness Straight leg raising Reflex change Nonsteroidal antiinflammatory drugs Opioids Anticonvulsants Antidepressants
DROP
OUTS
13/16
7/7
53 28–74 8/11
60.5 32–74 5/9
in situations in which the experimental treatment harms more patients, and the control treatment were also calculated on the basis of the number of patients who dropped out of each arm because of side effects. Because we pre-specified a single primary outcome measure, we considered all other statistical tests to be exploratory and did not correct for multiple comparisons except for the multiple SF-36 subscales.
Sample Size 8 0.5–40
4.5 1–28 Not available
11 5 14 7 2 26 10 21 10 23
13 3 9 4 13
8 0 7
3 0 4
Abbreviations: MRI, magnetic resonance imaging; DJD/DDD, degenerative joint disease and/or degenerative disc disease.
tion, is an easily understood estimate of the effect of treatment and represents the therapeutic effort required to realize a specific clinical benefit.18 Number needed to harm (NNH), the reciprocal of the risk increase
We estimated sample size to be 28 patients for alpha ⫽ 0.05 and beta ⫽ 0.20 by using the sample size formula for a crossover study.5 We used 2.9 as the standard deviation for the difference between drug and placebo on the basis of a drug trial29 of patients with chronic low back pain. We set effect size to 1.6 on a scale of 0 to 10. This corresponds to approximately 30% pain reduction, a clinically meaningful difference.13
Results Patients Table 1 shows the patients’ demographic and clinical features. Twenty-nine patients had evaluable responses to both topiramate and placebo. In addition to 26 patients who had completed both treatments and returned pain diaries, we included data from 3 additional patients; patients 18 and 29 had discontinued topiramate during period 2 after at least 21 days in this period. Both of these patients had been on a stable dose of medication for at least 1 week. For patient 2 who had completed both periods but who had not returned his pain diaries, we used pain scores entered in the end-of-period data sheets from the physical examination, bringing N to 29 (Fig 2). The mean
Figure 2. Clinical trial profile.
ORIGINAL REPORT/Khoromi et al Table 2.
833
Percent Pain Reduction for All Pain Scores in 29 Study Completers
PAIN
BASELINE
TOPIRAMATE
PLACEBO
TOPIRAMATEPLACEBO
Average leg* Average back Average overall Worst leg Worst back Worst overall
4.04 4.69 4.63 5.01 5.64 5.63
3.06 3.33 4.3 3.91 4.1 4.4
3.8 4.2 5.12 4.77 4.8 5.2
⫺0.74 ⫺0.87 ⫺0.82 ⫺0.86 ⫺0.7 ⫺0.8
% REDUCED 19 21 16 18 15 15
95% CONFIDENCE INTERVAL 41% to ⫺3% 38% to 4% 29% to 3% 37% to 0.5% 34% to ⫺4% 30% to 1%
P VALUE .06 .017 .02 .046 .13 .03
*Average leg pain was the primary outcome measure.
dose for topiramate was 208 mg and for diphenhydramine as placebo, 40 mg.
Differential Dropout Eleven patients dropped out of the study because of adverse effects associated with either topiramate (10 patients) or diphenhydramine (1 patient). Patient 27 dropped out before randomization for personal reasons, and patient 28 was excluded after screening baseline electrocardiogram showed an incidental cardiac abnormality.
was 0.84 between global pain relief category scores and change in average leg pain scores on a 0 to 10 scale. The Beck Depression Inventory scores (Table 4), the Oswestry Disability Index average scores (Table 5), and the SF-36 categories (Table 6), when corrected for multiple comparisons, showed no significant differences between topiramate and placebo.
Subgroups Based on MRI Diagnoses Table 7 shows the average leg pain scores at baseline, topiramate, and placebo, according to MRI diagnoses.
Effects of Treatment Table 2 shows the pain scores for the study completers on topiramate and placebo at baseline. Topiramate reduced average leg pain, the primary outcome variable, by 19% (confidence interval, 41% to ⫺3%) compared with placebo (P ⫽ .065). An intent-to-treat analysis with 31 patients, adding data from patient 14 who had finished 3 days of period 2 on topiramate and patient 10 who had finished 4 days of period 2 on topiramate, showed a similar result (P ⫽ .06). Percent pain reduction scores were significantly better on topiramate than placebo for average back pain, average overall pain, and worst overall pain, and they showed a trend toward pain reduction for worst back pain. Patients reported significantly better global pain relief ratings on topiramate as compared with placebo (P ⫽ .005, Wilcoxon signed rank test). Nine patients had a lot or complete pain relief on topiramate versus 1 patient on placebo (Table 3). Spearman rank correlation coefficient
Table 3.
Global Pain Relief
Worst pain No pain relief Slight pain relief Moderate pain relief Lot of pain relief Complete pain relief
TOPIRAMATE
PLACEBO
0 6 8 6 8 1
1 10 11 6 1 0
NOTE. Global pain relief reported after each treatment period shows that 54% of patients rated relief with topiramate as moderate or better, compared with 23% with placebo (P ⫽ .005, Wilcoxon signed rank test, for comparison across the 6-category scale).
Table 4.
Beck Depression Inventory BASELINE
TOPIRAMATE
PLACEBO
22 5 1 0
21 6 0 1
24 4 0 0
Normal Mild/moderate Moderate/severe Extremely severe
Table 5.
Oswestry Low Back Pain Disability
Index BASELINE
TOPIRAMATE
PLACEBO
30% 14%
25% 16%
27% 15%
Mean score Standard deviation
Table 6.
SF-36
Physical functioning Social functioning Role limitations (physical)† Role limitations (emotional) Body pain Mental health* Vitality General health perception
BASELINE
TOPIRAMATE
PLACEBO
59 74 35 78 40 79 50 71
67 78 53 67 51 74 54 72
63 80 47 78 50 80 56 72
NOTE. Mental health scores (*P ⫽ .019) were significantly worse for topiramate as compared with placebo, and physical role limitations (†P ⫽ .020) were significantly worse for placebo as compared with topiramate. Neither remains significant if one corrects the P values for testing SF-36 subscales.
834
Topiramate in Chronic Sciatica
MRI Diagnosis and Reduction in Pain With Topiramate as Compared With Placebo
Table 7.
AVERAGE LEG PAIN
Adverse Effects and NNH PAIN
DIAGNOSIS
N
BASELINE
TOPIRAMATE
PLACEBO
DIFFERENCE (TOPIRAMATEPLACEBO)*
DDD/DJD NFS, LRS LRS NFS No MRI CS CS, LRS NFS, LRS, CS
7 6 5 4 2 2 2 1
4.7 4.2 3.0 4.5 3.5 3.3 3.8 4.1
4.2 2.3 2.2 3.2 4.6 2.9 2.1 3.1
5.6 2.2 3.6 3.3 4.3 4.4 3.4 2.1
⫺1.4 0.1 ⫺1.4 ⫺0.1 0.3 ⫺1.5 ⫺1.3 1.0
MRI
of a total of 42 patients. Therefore NNT was 5.3 in these patients.
NOTE. Negative scores correspond to less pain with topiramate. Abbreviations: LRS, lateral recess syndrome; NFS, neural foramen stenosis; CS, canal stenosis; DDD, degenerative disc disease; DJD, degenerative joint disease.
Post hoc analysis showed that the 11 patients who had neural foramen stenosis alone or in combination with other lesions had on average a 0.1-point increase in pain scores on topiramate as compared with placebo. In contrast, the 16 patients with all other MRI diagnoses had an average decrease of 1.4 when comparing topiramate with placebo.
Period and Carryover Effect The interaction term between period order and treatment (partial F⫽ 1.54; P ⫽ .22) and the Pocock test for carryover effect for average leg pain (P ⬎ .32) with an unpaired t test were not significant.
Estimate of Dropout Bias Assuming the null hypothesis that an experimental drug is devoid of efficacy, differential dropout can cause spurious attribution of analgesia in the following manner. Patients whose pain improved by chance during the treatment might attribute this to the drug and remain in the study, despite troubling side effects. Those without improvement would tend to drop out, and the exclusion of these scores might create a bias simulating efficacy. We were unable to include in the completer or intentto-treat analyses the first-period 7 dropouts who were never exposed to placebo, and we estimated the maximum dropout bias by comparing the reduction of average leg pain scores by topiramate in the dropouts and completers. First-period topiramate treatment reduced pain by 15% from baseline in the 14 completers and 10% in the 7 dropouts. Therefore, if all 21 patients had been considered, the estimated mean value would have been a 13.3% reduction, just slightly less than the pain reduction in the completers.
NNT Fifteen patients had categorical pain relief scores of moderate or better on topiramate and 7 on placebo out
The adverse effects of topiramate leading to patient withdrawal were acral paresthesias in 2 patients, nausea and anorexia in 2, sedation and amnesia in 3, depression and anxiety in 1, and rash in 1. In study completers 86% experienced side effects on topiramate as compared with 72% on placebo. The side effects associated with topiramate and placebo are reported in order of frequency in Table 8. NNH was 4.4 among the 40 patients who had taken topiramate.
Discussion This randomized trial found that topiramate, titrated to maximal tolerated dose, reduced average leg pain by 19% relative to placebo. Although this primary outcome just missed statistical significance, several of the other pain scores, such as the average back, average overall, worst leg, worst back pain scores, and the global pain relief scores, showed statistically significant improvement for topiramate. None of the secondary outcome scores, including the scores of the Beck Depression Inventory, the Oswestry Disability Index, and the SF-36, differed between treatments. In addition, an excess dropout during topiramate treatment might have added a small amount to our estimate of topiramate analgesia, on the basis of our finding that topiramate dropouts had less pain reduction than topiramate completers. Another limitation to be considered in the data interpretation is
Table 8.
Adverse Effects SYMPTOM
TOPIRAMATE (%)
PLACEBO (%)
Paresthesias Fatigue/weakness Sedation Diarrhea Headache Constipation Depression Joint pain Leg cramps Amnesia Anorexia Frequent urination Thirst Eyes twitching Edema Speech difficulty Blurred vision Photophobia Bleeding gums Tremor Thick templed Urine extremely yellow in color Decreased libido Any side effect
38 34 34 30 10 7 7 7 7 3
21 31 3 10 10
3 17 3
3 3 3 3 3 3 3 3 3 3 3 3 86
72
ORIGINAL REPORT/Khoromi et al
835
the possibility that the patients might have back-filled their pain diaries before their clinic visits. Such retrospective recording, or recall bias,15 would tend to add to the error variance and make it more difficult to demonstrate statistical significance for an effective drug. Inclusion of all types of MRI classifications associated with lumbar nerve root irritation might have also added to the uncertainty of our estimate of the true effect of topiramate on any anatomic subset of chronic lumbar radicular pain. Our post hoc analysis of pain relief by MRI subgroups (Table 5) raises the possibility that patients with neural foramen stenosis might be relatively resistant to topiramate treatment. To our knowledge this is the first clinical trial of chronic sciatica in which MRI subgroupings have been correlated with treatment response, and because of the exploratory nature of this analysis no firm inferences can be made. Further studies are required to address the issue of whether there are any correlations between MRI findings and treatment response. Our dropout rate of 26% as a result of topiramate side effects is comparable with the other 2 major trials with topiramate in painful diabetic neuropathy in which the average dropout rate in the topiramate arm was 24%.27,33 The most commonly reported adverse events were similar among the 3 studies, with changes in sensorium, fatigue, paresthesias, and gastrointestinal disturbances being most frequent.27,33 Both of those other studies specified a maximum topiramate dose of 400 mg as well. The starting dose of topiramate was 25 mg in both studies, however, and the escalation was over 6 to 8 weeks, respectively. In summary, we cannot completely rule out the possibility that the apparent pain reduction we observed was due to chance or dropout bias, but we consider it more likely that topiramate has a small but real analgesic effect. A modest effect size might also explain the conflicting results in the diabetic neuropathy studies.27,33 Some pain researchers, particularly industry-funded multicenter groups, are reluctant to test new analgesics in a pain syndrome until there is published evidence for the sensitivity of the study methods. The ability of our methods to detect a modest degree of pain reduction on several of the outcome measures in a 29-patient crossover study should encourage other researchers to study this common condition. Our lumbar radiculopathy pa-
tients had relatively few confounding psychiatric comorbidities compared with patients with idiopathic back pain or many other chronic pain conditions. No patient met the PHQ-15 criteria for multisomatoform disorder. Only 6 of 42 enrolled patients had evidence of depression based on the Beck Depression Inventory. Similar to patients seen in primary care and general neurology settings, none of them had had a trial of gabapentin, and few had tried other medications commonly used to treat neuropathic pain syndromes (Table 1). It is difficult to integrate our marginally positive finding for topiramate into an evidence-based algorithm in the absence of other published randomized trials of drug treatment of chronic lumbar radiculopathy pain. Pain experts recommend that the choice among medications for treating neuropathic pain should be based on the comparison of the drug side effects and the patient’s comorbidities and preferences. On the basis of our observations of modest pain improvement, troublesome side effects, and frequent dropouts, we concluded that topiramate is at best marginally effective in the treatment of chronic lumbar radiculopathy in patients who can tolerate its prominent side effects. It is at best a second-line treatment in patients with lumbar radiculopathy refractory to first-line drugs such as gabapentin, tricyclic antidepressants, and opioids9 proven to be effective in other neuropathic pain syndromes. Future studies of topiramate should include a larger group of patients with homogenous MRI diagnoses by using a slower titration with smaller initial doses to minimize potential problems of differential dropout. In addition, prospective dose-response studies are needed to explore whether lower doses might achieve similar pain relief but improve topiramate’s therapeutic ratio.
References
4. Cohrs RJ, Gilden DH, Mahalingam R: Varicella zoster virus latency, neurological disease and experimental models: an update. Front Biosci 9:751-762, 2004
1. Atkinson JH, Slater MA, Williams RA, Zisook S, Patterson TL, Grant I, Wahlgren DR, Abramson I, Garfin SR: A placebocontrolled randomized clinical trial of nortriptyline for chronic low back pain. Pain 76:287-296, 1998
Acknowledgments We thank Kenneth Schmader and Carl Pieper for reviewing the manuscript, Kelly Cahill, Department of Nursing, Judith Starling of the NIH Pharmaceutical Development Service, and Sandra Jones of the NIH Patient Recruitment Branch for technical assistance; and Dr Nicolas Patronas, Department of Radiology, for his suggestions.
5. Colton T: Statistics in Medicine. 1st ed Boston, MA, Little Brown, 1974, p142
2. Avoli M: GABA-mediated synchronous potentials and seizure generation. Epilepsia 37:1035-1042, 1996
6. Dellemijn PL, van Duijn H, Vanneste JA: Prolonged treatment with transdermal fentanyl in neuropathic pain. J Pain Symptom Manage 16:220-229, 1998
3. Bajwa ZH, Sami N, Warfield CA, Wootton J: Topiramate relieves refractory intercostal neuralgia. Neurology 52: 1917, 1999
7. Deyo RA, Rainville J, Kent DL : What can the history and physical examination tell us about low back pain? JAMA 268:760-765, 1992
836
Topiramate in Chronic Sciatica
8. Deyo RA, Weinstein JN: Low back pain. N Engl J Med 344:363-370, 2001
the treatment of sciatica: A double-blind, randomized, controlled study. Reg Anesth Pain Med 24:534-540, 1999
9. Dworkin RH, Backonja M, Rowbotham MC, Allen RR, Argoff CR, Bennett GJ, Bushnell MC, Farrar JT, Galer BS, Haythornthwaite JA, Hewitt DJ, Loeser JD, Max MB, Saltarelli M, Schmader KE, Stein C, Thompson D, Turk DC, Wallace MS, Watkins LR, Weinstein SM: Advances in neuropathic pain: Diagnosis, mechanisms, and treatment recommendations. Arch Neurol 60:1524-1534, 2003
25. Murphy RW: Nerve roots and spinal nerves in degenerative disk disease. Clin Orthop Relat Res 129:46-60, 1977
10. Edwards KR, Glanz MJ, Button J: Efficacy and safety of topiramate in the treatment of painful diabetic neuropathy: a double-blind, placebo-controlled study. Neurology 54:A81, 2000 11. Fairbanks JCT, Davies JB, Couper J, O’Brien JP: The Oswestry low-back pain disability questionnaire. Physiotherapy 66:271-273, 1980 12. Fardon DF, Milette PC: Nomenclature and classification of lumbar disc pathology: Recommendations of the combined task Force of the North American Spine Society, American Society of Spine Radiology, and American Society of Neuroradiology. Spine 26:E93-E113, 2001
26. Pocock SJ: Cross over trials, in Clinical Trials, A Practical Approach. 4th ed. Chichester, UK, Wiley and Sons, 1983 27. Raskin P, Donofrio PD, Rosenthal NR, Hewitt DJ, Jordan DM, Xiang J, Vinik A: Topiramate vs placebo in painful diabetic neuropathy: Analgesic and metabolic effects. Neurology 63:865-873, 2004 28. Remmers AESU, LaMoreaux L, Young JP, Moore J, Poole M: Pregabalin treatment of patients with chronic low back pain. Presented at the American Pain Society Annual Meeting, Atlanta, Georgia, November 2-5, 2000 29. Schnitzer TJ, Gray WL, Paster RZ, Kamin M: Efficacy of tramadol in treatment of chronic low back pain. J Rheumatol 27:772-778, 2000 30. Severt CD, Sombati S: Topiramate selectively blocks kainate currents in cultured hippocampal neurones. Epilepsia 36:38, 1995
13. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM: Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain 94:149-158, 2001
31. Shank RP, Gardocki JF, Streeter AJ, Maryanoff BE: An overview of the preclinical aspects of topiramate: pharmacology, pharmacokinetics, and mechanism of action. Epilepsia 41(suppl 1):S3-S9, 2000
14. Franson RC, Saal JS, Saal JA: Human disc phospholipase A2 is inflammatory. Spine 17(Suppl):S129-S132, 1992
32. Sindrup SH, Jensen TS: Efficacy of pharmacological treatments of neuropathic pain: An update and effect related to mechanism of drug action. Pain 83:389-400, 1999
15. Gendreau M, Hufford MR, Stone AA: Measuring clinical pain in chronic widespread pain: selected methodological issues. Best Pract Res Clin Rheumatol 17:575-592, 2003 16. Gilron I, Booher SL, Rowan JS, Max MB: Topiramate in trigeminal neuralgia: a randomized, placebo-controlled multiple crossover pilot study. Clin Neuropharmacol 24:109111, 2001 17. Heliovaara M, Impivaara O, Sievers K, Melkas T, Knekt P, Korpi J, Aromaa A: Lumbar disc syndrome in Finland. J Epidemiol Community Health 4:251-258, 1987 18. Irvine EJ: Measurement and expression of risk: Optimizing decision strategies. Am J Med 117(Suppl 5A):2S-7S, 2004 19. Jensen MP, Karoly P, Braver S: The measurement of clinical pain intensity: A comparison of six methods. Pain 27: 117-126, 1986
33. Thienel U, Neto W, Schwabe SK, Vijapurkar U: Topiramate in painful diabetic polyneuropathy: findings from three double-blind, placebo-controlled trials. Acta Neurol Scand 110: 221-231, 2004 34. Tomlinson DR: The pharmacology of diabetic neuropathy. Diabetes Metab Rev 8:67-84, 1992 35. Ware JE: SF-36 health survey update. Spine 25:31303139, 2000 36. de C Williams AC, Richardson PH: What does the BDI measure in chronic pain? Pain 55:259-266, 1993 37. Wipf JE, Deyo RA: Low back pain. Med Clin North Am 79:231-246, 1995
20. Jensen MP, McFarland CA: Increasing the reliability and validity of pain intensity measurement in chronic pain patients. Pain 55:195-203, 1993
38. Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, Tugwell P, Campbell SM, Abeles M, Clark P: Criteria for the classification of fibromyalgia: Report of the Multicenter Criteria Committee. The American College of Rheumatology 33:160-172, 1990
21. Kroenke K, Spitzer RL, deGruy FV 3rd, Swindle R: A symptom checklist to screen for somatoform disorders in primary care. Psychosomatics 39:263-272, 1998
39. Woolf CJ, Mannion RJ: Neuropathic pain: Aetiology, symptoms, mechanisms, and management. Lancet 353: 1959-1964, 1999
22. Max MB, Lynch SA, Muir J, Shoaf SE, Smoller B, Dubner R: Effects of desipramine, amitriptyline, and fluoxetine on pain in diabetic neuropathy. N Engl J Med 326:1250-1256, 1992
40. Zeigler D, Lynch SA, Muir J, Benjamin J, Max MB: Transdermal clonidine versus placebo in painful diabetic neuropathy. Pain 48:403-408, 1992
23. McCarron RF, Wimpee MW, Hudkins PG, Laros G: The inflammatory effect of nucleus pulposus. Spine 12:760-764, 1987 24. Medrik-Goldberg T, Lifschitz D, Pud D, Adler R, Eisenberg E: Intravenous lidocaine, amantadine, and placebo in
41. Zhang XL, Velumian AA, Jones OT, Carlen PL. Modulation of voltage-activated calcium channels in dentate granule cells by topiramate. Epilepsia 41(suppl):S10-S16, 2000 42. Zvartau-Hind M, Din MU, Gilani A, Lisak RP, Khan OA: Topiramate relieves refractory trigeminal neuralgia in MS patients. Neurology 55:1587-1588, 2000
A
lthough radiculopathy caused by lumbar disc disease is the most common chronic neuropathic pain syndrome, with a prevalence of 4.5% in individuals older than the age of 30,17 there are no published repeated dose-controlled analgesic trials in this disorder. Reports are limited to a few subset analyses in studies of a broader selection of patients with back pain. Dellemijn et al6 reported that mean pain reduction in an open-label study of transdermal fentanyl in 18 patients with lumbar radiculopathy was 32%. Atkinson et al1 suggested that the small subgroup of patients with lumbar radiculopathy (n ⫽ 11) responded more favorably to nortriptyline than placebo within a larger study including mixed back pain etiologies. Remmers et al28 reported no relief of either leg or back pain by pregabalin in a placebo-controlled trial. Medrik-Goldberg et al24 found that acute lidocaine infusion significantly reduced pain Received February 23, 2005; Revised May 25, 2005; Re-revised July 13, 2005; Accepted August 2, 2005. From the *Pain and Neurosensory Mechanisms Branch, National Institute of Dental and Craniofacial Research; and the Departments of †Radiology and ‡Nursing, Warren Magnuson Clinical Center, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland. Supported by NIDCR Intramural Research Grant Z01 DE00366 in addition to partial support of a data technician by an Ortho McNeil educational grant. Dr Max had previously served as a paid consultant for Ortho McNeil but resigned before planning the study. Ortho McNeil had no input into the study design, data interpretation, or manuscript preparation. A preliminary report of these results was presented at the Tenth World Congress on Pain, San Diego, California, August 19, 2002. Address reprint requests to Suzan Khoromi, MD, MHS, National Institutes of Health Building 10, 4-1741, Bethesda, MD 20892-1302. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2005.08.002
in 30 patients with sciatica when compared with placebo and amantadine. Despite the fact that chronic lumbar radiculopathy is at least 20 times more common than diabetic neuropathy and postherpetic neuralgia, most of the current evidence for neuropathic pain treatment is derived from the latter 2 pain syndromes.9,32,40 Generalizing these results to lumbar radiculopathy pain might be inappropriate because the site and pathology of these conditions differ. The lesion in most lumbar radiculopathies is a focal compressive37 and inflammatory injury of the nerve root14,23 and dorsal root ganglion.25 Therefore the pain might have a different pathophysiology and treatment response than either diabetic neuropathy, where the more diffuse lesions have metabolic34 and vascular causes,39 or postherpetic neuralgia, where the multifocal lesions are infectious and inflammatory.4 For this study we chose the anticonvulsant topiramate, which combines multiple mechanisms of action including voltage-gated sodium channel blockade and modulation,31 gamma-aminobutyric acid agonist-like effects,2 AMPA/kainate glutamate receptor blockade,30 and modulation of voltage-gated calcium ion channels.41 Topiramate was more effective than placebo in relieving pain associated with diabetic neuropathy in 2 of 3 randomized trials in patients with painful diabetic neuropathy.10,27,33 A study of topiramate (50 to 400 mg) in 27 patients10 and a dose-response study of topiramate 100, 200, and 400 mg in 1259 patients33 found no difference in pain relief between topiramate and placebo, whereas another study found significant pain reduction in 323 patients on topiramate (50 to 400 mg) compared with
The Journal of Pain, Vol 6, No 12 (December), 2005: pp 829-836
829
830
Topiramate in Chronic Sciatica 27
placebo. In addition, anecdotal reports indicate that topiramate might reduce pain in trigeminal neuralgia16,42 and intercostal neuralgia.3 The aim of this study was to evaluate the efficacy of topiramate in the treatment of chronic sciatica, adapting methods that have been previously validated in studies of diabetic neuropathy22 and postherpetic neuralgia.
Materials and Methods Study Population This study was conducted at the National Institutes of Health (NIH) Clinical Center. The protocol was approved by the National Institute of Dental and Craniofacial Research Institutional Review Board. Patients were recruited between March 2001 and March 2002 through local newspaper advertisements. Of 500 phone responders, 455 had back pain alone, pain location and quality that were typical of myofascial pain in the lower extremities (localized stiffness and dull aching pain), or were not interested in a drug treatment study. Ninety percent of the candidates were excluded because their features were more consistent with myofascial pain, which is consistent with the reported frequency of this etiology for low back pain seen in the primary care setting.7,8 Fortytwo of the 45 respondents with radiating leg pain who were willing to participate in a drug trial met the inclusion and exclusion criteria, and all subjects signed an informed consent form.
defined by American College of Rheumatology criteria38; (7) polyneuropathy and peripheral vascular disease; (8) history of nephrolithiasis; and (9) narrow angle glaucoma.
Imaging and Laboratory Evaluation Patients submitted an MRI scan of the lumbosacral spine taken within a year before study enrollment, or this was performed on study entry. One neuroradiologist blinded to the patients’ symptoms reviewed the films, commenting on degenerative disc or joint disease and other findings contributing to low back pain or radiculopathy according to definitions and classifications proposed by the American Society of Neuroradiology.12 To scrutinize the relationship between treatment response and imaging diagnosis, we assigned patients to 9 radiologic categories. Patients were classified as lateral recess syndrome, neural foramen stenosis, canal stenosis, or their combination if their clinical findings and MRI scan offered a consistent anatomic explanation of their root symptoms resulting from compression of roots due to degeneration of joint, facet, and/or disc. Patients with no visible root compression were classified as degenerative disc disease and/or degenerative joint disease. Laboratory evaluation included a complete blood count with differential, sedimentation rate, antinuclear antibody titer, and rheumatoid factor to exclude inflammatory arthritis, cancer, and spinal infection.
Screening for Somatoform Disorder Inclusion Criteria First criterion was evidence of lumbar radiculopathy, on the basis of the presence of pain in one or both buttocks or legs for 3 months or greater for at least 5 days a week and at least one of the following features on the side corresponding to leg pain: (1) sharp and shooting pain below the knee; (2) pain evoked by straight leg raising to 60 degrees or less; (3) decreased/absent ankle reflex; (4) weakness of muscles below the knee; (5) sensory loss in L5/S1 distribution; (6) electromyographic evidence for L4, L5, or S1 root denervation; (7) imaging (magnetic resonance imaging [MRI], computed tomography/myelogram) evidence of nerve root compression in the lower lumbar region. The second criterion was average leg pain of at least 4 of 10 for the past month. The third criterion was willingness to refrain from making changes in non-study medications taken for sciatica. The fourth criterion was age between 18 and 75 years at the start of the study. The principal investigator performed the clinical examinations. Sensory testing included responses to pinprick, light touch, vibration, and joint movement.
Exclusion Criteria The exclusion criteria consisted of (1) hepatic and renal dysfunction; (2) pregnancy or lactation; (3) seizure disorder; (4) pain of greater intensity in any other location than the low back or leg; (5) narcotic abuse and/or drug or alcohol abuse during the past year; (6) fibromyalgia as
Patients completed a 15-item questionnaire, the PHQ15, devised by Kroenke et al21 to assess for multisomatoform disorder.
Study Design We carried out a double-blind, 2-period, crossover study comparing topiramate with active placebo, diphenhydramine. We chose diphenhydramine as active placebo because of its side effects, mainly sedation and anticholinergic effects, some of which overlap with topiramate. Drug randomization was performed by an NIH pharmacist according to blocked randomization in blocks of 8 and 4 within a table of 32 random numbers before patient enrollment. Each period consisted of 4 weeks of drug titration and 2 weeks of maintenance, followed by a 2-week washout, during which doses were gradually tapered (Fig 1). Topiramate was started at 50 mg at bedtime. Doses were escalated to 50 mg in 2 divided doses during week 1, by 50-mg increments in each of AM and PM doses during weeks 3 and 4 for a maximum of 400 mg. Doses of the study drug were based on a prior multicenter trial of topiramate in painful diabetic neuropathy.27 Diphenhydramine was started at 6.25 mg at bedtime, increased to 6.25 mg twice a day during week 1, and then increased by 6.25-mg increments in each dose during week 2 and by 12.5-mg increments in each dose during week 3 to a maximum of 50 mg a day in 2 divided doses. If side effects associated with drug escalation interfered with any activities or if the patient found the
ORIGINAL REPORT/Khoromi et al
831
Figure 1. Study design.
side effects unpleasant, we decreased the patient’s dose to the previous level.
Outcome Assessment Each day at bedtime, patients rated their pain during the last 24 hours by using a 0 to 10 numeric scale.19 They were asked to assess pain in 6 separate categories: average back, leg, and overall pain (leg and back) and worst back, leg, and overall pain. For people who had pain in both legs, pain on the worse side was recorded. During the patients’ follow-up visit to the Clinical Center at the end of each period, the study nurses collected the pain diaries and questionnaires from the patients. They then entered these data into a database from which all study outcomes were derived. Side effects elicited during semiweekly phone interviews were rated for severity as mild, moderate, or severe, and they were then entered into a database. For the primary outcome, we chose the comparison of mean scores for average leg pain during the 2 weeks of maintenance on topiramate and placebo. In the original study protocol, we defined the primary outcome to be the comparison of mean scores for average leg pain during the 2 weeks at the final maintenance doses for topiramate versus placebo. We chose leg pain over back pain because we are testing the principle that topiramate relieves nerve pain, and leg pain is more likely than back pain to reflect nerve pathology. We have examined the scores in the final 2 weeks in this study and in most of our group’s previous clinical trials for several reasons. (1) Jensen and McFarland20 showed that an average of 7 to 14 pain scores proves a more accurate estimate of true mean pain than a single measurement, but additional scores do not improve the fit. (2) Comparative treatment effect is often highest during the final weeks of treatment if the patients have been slowly titrated along a rising dose-response curve. (3) Exclusion of the first sev-
eral weeks of the treatment period lessens the chance of pharmacologic carryover effect in a crossover design. Other outcome measures were (1) global pain relief (leg and back pain combined) by using a categorical pain scale rating overall pain outcome as worse, no relief, slight, moderate, a lot, and complete relief; (2) the Oswestry Low Back Pain Disability questionnaire11; (3) the Beck Depression Inventory36; and (4) the 36-item Short Form of Health Survey (SF-36), a general health status instrument that measures the social, mental, and emotional dimensions of health and illness.35
Blinding The NIH outpatient pharmacy was in charge of drug randomization, and patients and research staff were blinded to the randomization order.
Statistical Analysis Pain score outcomes from the SF-36, the Oswestry Low Back Pain Disability questionnaire, and the Beck Depression Index were compared by using paired, 2-tailed paired t tests. We performed all statistical tests with the analyst function in Statistical Analysis Software (SAS, Version 8; SAS Institute Inc, Cary, NC). We assessed for treatment-period interactions (carryover effect) by using 2 methods: a t test variant proposed by Pocock26 and inspection of the interaction term from a factorial analysis of variance (ANOVA) model in which treatment and period were independent variables and average leg pain the outcome score. We used global pain relief (GPR) scores of moderate or better for each drug as the response criterion to calculate the number needed to treat, NNT ⫽ 1/([proportion of patients in the treatment group with GPR score of moderate or better] ⫺ [the proportion of patients in the placebo group with same GPR scores]).13 NNT, the reciprocal of the absolute risk reduc-
832 Table 1.
Topiramate in Chronic Sciatica
Demographics STUDY COMPLETERS
Sex (F/M) Age (y) Median Range Work status (retired/nonretired) Pain duration (y) Median Range MRI diagnosis Neural foraminal stenosis Canal stenosis Lateral recess syndrome DJD/DDD No MRI Physical signs Sensory change Weakness Straight leg raising Reflex change Nonsteroidal antiinflammatory drugs Opioids Anticonvulsants Antidepressants
DROP
OUTS
13/16
7/7
53 28–74 8/11
60.5 32–74 5/9
in situations in which the experimental treatment harms more patients, and the control treatment were also calculated on the basis of the number of patients who dropped out of each arm because of side effects. Because we pre-specified a single primary outcome measure, we considered all other statistical tests to be exploratory and did not correct for multiple comparisons except for the multiple SF-36 subscales.
Sample Size 8 0.5–40
4.5 1–28 Not available
11 5 14 7 2 26 10 21 10 23
13 3 9 4 13
8 0 7
3 0 4
Abbreviations: MRI, magnetic resonance imaging; DJD/DDD, degenerative joint disease and/or degenerative disc disease.
tion, is an easily understood estimate of the effect of treatment and represents the therapeutic effort required to realize a specific clinical benefit.18 Number needed to harm (NNH), the reciprocal of the risk increase
We estimated sample size to be 28 patients for alpha ⫽ 0.05 and beta ⫽ 0.20 by using the sample size formula for a crossover study.5 We used 2.9 as the standard deviation for the difference between drug and placebo on the basis of a drug trial29 of patients with chronic low back pain. We set effect size to 1.6 on a scale of 0 to 10. This corresponds to approximately 30% pain reduction, a clinically meaningful difference.13
Results Patients Table 1 shows the patients’ demographic and clinical features. Twenty-nine patients had evaluable responses to both topiramate and placebo. In addition to 26 patients who had completed both treatments and returned pain diaries, we included data from 3 additional patients; patients 18 and 29 had discontinued topiramate during period 2 after at least 21 days in this period. Both of these patients had been on a stable dose of medication for at least 1 week. For patient 2 who had completed both periods but who had not returned his pain diaries, we used pain scores entered in the end-of-period data sheets from the physical examination, bringing N to 29 (Fig 2). The mean
Figure 2. Clinical trial profile.
ORIGINAL REPORT/Khoromi et al Table 2.
833
Percent Pain Reduction for All Pain Scores in 29 Study Completers
PAIN
BASELINE
TOPIRAMATE
PLACEBO
TOPIRAMATEPLACEBO
Average leg* Average back Average overall Worst leg Worst back Worst overall
4.04 4.69 4.63 5.01 5.64 5.63
3.06 3.33 4.3 3.91 4.1 4.4
3.8 4.2 5.12 4.77 4.8 5.2
⫺0.74 ⫺0.87 ⫺0.82 ⫺0.86 ⫺0.7 ⫺0.8
% REDUCED 19 21 16 18 15 15
95% CONFIDENCE INTERVAL 41% to ⫺3% 38% to 4% 29% to 3% 37% to 0.5% 34% to ⫺4% 30% to 1%
P VALUE .06 .017 .02 .046 .13 .03
*Average leg pain was the primary outcome measure.
dose for topiramate was 208 mg and for diphenhydramine as placebo, 40 mg.
Differential Dropout Eleven patients dropped out of the study because of adverse effects associated with either topiramate (10 patients) or diphenhydramine (1 patient). Patient 27 dropped out before randomization for personal reasons, and patient 28 was excluded after screening baseline electrocardiogram showed an incidental cardiac abnormality.
was 0.84 between global pain relief category scores and change in average leg pain scores on a 0 to 10 scale. The Beck Depression Inventory scores (Table 4), the Oswestry Disability Index average scores (Table 5), and the SF-36 categories (Table 6), when corrected for multiple comparisons, showed no significant differences between topiramate and placebo.
Subgroups Based on MRI Diagnoses Table 7 shows the average leg pain scores at baseline, topiramate, and placebo, according to MRI diagnoses.
Effects of Treatment Table 2 shows the pain scores for the study completers on topiramate and placebo at baseline. Topiramate reduced average leg pain, the primary outcome variable, by 19% (confidence interval, 41% to ⫺3%) compared with placebo (P ⫽ .065). An intent-to-treat analysis with 31 patients, adding data from patient 14 who had finished 3 days of period 2 on topiramate and patient 10 who had finished 4 days of period 2 on topiramate, showed a similar result (P ⫽ .06). Percent pain reduction scores were significantly better on topiramate than placebo for average back pain, average overall pain, and worst overall pain, and they showed a trend toward pain reduction for worst back pain. Patients reported significantly better global pain relief ratings on topiramate as compared with placebo (P ⫽ .005, Wilcoxon signed rank test). Nine patients had a lot or complete pain relief on topiramate versus 1 patient on placebo (Table 3). Spearman rank correlation coefficient
Table 3.
Global Pain Relief
Worst pain No pain relief Slight pain relief Moderate pain relief Lot of pain relief Complete pain relief
TOPIRAMATE
PLACEBO
0 6 8 6 8 1
1 10 11 6 1 0
NOTE. Global pain relief reported after each treatment period shows that 54% of patients rated relief with topiramate as moderate or better, compared with 23% with placebo (P ⫽ .005, Wilcoxon signed rank test, for comparison across the 6-category scale).
Table 4.
Beck Depression Inventory BASELINE
TOPIRAMATE
PLACEBO
22 5 1 0
21 6 0 1
24 4 0 0
Normal Mild/moderate Moderate/severe Extremely severe
Table 5.
Oswestry Low Back Pain Disability
Index BASELINE
TOPIRAMATE
PLACEBO
30% 14%
25% 16%
27% 15%
Mean score Standard deviation
Table 6.
SF-36
Physical functioning Social functioning Role limitations (physical)† Role limitations (emotional) Body pain Mental health* Vitality General health perception
BASELINE
TOPIRAMATE
PLACEBO
59 74 35 78 40 79 50 71
67 78 53 67 51 74 54 72
63 80 47 78 50 80 56 72
NOTE. Mental health scores (*P ⫽ .019) were significantly worse for topiramate as compared with placebo, and physical role limitations (†P ⫽ .020) were significantly worse for placebo as compared with topiramate. Neither remains significant if one corrects the P values for testing SF-36 subscales.
834
Topiramate in Chronic Sciatica
MRI Diagnosis and Reduction in Pain With Topiramate as Compared With Placebo
Table 7.
AVERAGE LEG PAIN
Adverse Effects and NNH PAIN
DIAGNOSIS
N
BASELINE
TOPIRAMATE
PLACEBO
DIFFERENCE (TOPIRAMATEPLACEBO)*
DDD/DJD NFS, LRS LRS NFS No MRI CS CS, LRS NFS, LRS, CS
7 6 5 4 2 2 2 1
4.7 4.2 3.0 4.5 3.5 3.3 3.8 4.1
4.2 2.3 2.2 3.2 4.6 2.9 2.1 3.1
5.6 2.2 3.6 3.3 4.3 4.4 3.4 2.1
⫺1.4 0.1 ⫺1.4 ⫺0.1 0.3 ⫺1.5 ⫺1.3 1.0
MRI
of a total of 42 patients. Therefore NNT was 5.3 in these patients.
NOTE. Negative scores correspond to less pain with topiramate. Abbreviations: LRS, lateral recess syndrome; NFS, neural foramen stenosis; CS, canal stenosis; DDD, degenerative disc disease; DJD, degenerative joint disease.
Post hoc analysis showed that the 11 patients who had neural foramen stenosis alone or in combination with other lesions had on average a 0.1-point increase in pain scores on topiramate as compared with placebo. In contrast, the 16 patients with all other MRI diagnoses had an average decrease of 1.4 when comparing topiramate with placebo.
Period and Carryover Effect The interaction term between period order and treatment (partial F⫽ 1.54; P ⫽ .22) and the Pocock test for carryover effect for average leg pain (P ⬎ .32) with an unpaired t test were not significant.
Estimate of Dropout Bias Assuming the null hypothesis that an experimental drug is devoid of efficacy, differential dropout can cause spurious attribution of analgesia in the following manner. Patients whose pain improved by chance during the treatment might attribute this to the drug and remain in the study, despite troubling side effects. Those without improvement would tend to drop out, and the exclusion of these scores might create a bias simulating efficacy. We were unable to include in the completer or intentto-treat analyses the first-period 7 dropouts who were never exposed to placebo, and we estimated the maximum dropout bias by comparing the reduction of average leg pain scores by topiramate in the dropouts and completers. First-period topiramate treatment reduced pain by 15% from baseline in the 14 completers and 10% in the 7 dropouts. Therefore, if all 21 patients had been considered, the estimated mean value would have been a 13.3% reduction, just slightly less than the pain reduction in the completers.
NNT Fifteen patients had categorical pain relief scores of moderate or better on topiramate and 7 on placebo out
The adverse effects of topiramate leading to patient withdrawal were acral paresthesias in 2 patients, nausea and anorexia in 2, sedation and amnesia in 3, depression and anxiety in 1, and rash in 1. In study completers 86% experienced side effects on topiramate as compared with 72% on placebo. The side effects associated with topiramate and placebo are reported in order of frequency in Table 8. NNH was 4.4 among the 40 patients who had taken topiramate.
Discussion This randomized trial found that topiramate, titrated to maximal tolerated dose, reduced average leg pain by 19% relative to placebo. Although this primary outcome just missed statistical significance, several of the other pain scores, such as the average back, average overall, worst leg, worst back pain scores, and the global pain relief scores, showed statistically significant improvement for topiramate. None of the secondary outcome scores, including the scores of the Beck Depression Inventory, the Oswestry Disability Index, and the SF-36, differed between treatments. In addition, an excess dropout during topiramate treatment might have added a small amount to our estimate of topiramate analgesia, on the basis of our finding that topiramate dropouts had less pain reduction than topiramate completers. Another limitation to be considered in the data interpretation is
Table 8.
Adverse Effects SYMPTOM
TOPIRAMATE (%)
PLACEBO (%)
Paresthesias Fatigue/weakness Sedation Diarrhea Headache Constipation Depression Joint pain Leg cramps Amnesia Anorexia Frequent urination Thirst Eyes twitching Edema Speech difficulty Blurred vision Photophobia Bleeding gums Tremor Thick templed Urine extremely yellow in color Decreased libido Any side effect
38 34 34 30 10 7 7 7 7 3
21 31 3 10 10
3 17 3
3 3 3 3 3 3 3 3 3 3 3 3 86
72
ORIGINAL REPORT/Khoromi et al
835
the possibility that the patients might have back-filled their pain diaries before their clinic visits. Such retrospective recording, or recall bias,15 would tend to add to the error variance and make it more difficult to demonstrate statistical significance for an effective drug. Inclusion of all types of MRI classifications associated with lumbar nerve root irritation might have also added to the uncertainty of our estimate of the true effect of topiramate on any anatomic subset of chronic lumbar radicular pain. Our post hoc analysis of pain relief by MRI subgroups (Table 5) raises the possibility that patients with neural foramen stenosis might be relatively resistant to topiramate treatment. To our knowledge this is the first clinical trial of chronic sciatica in which MRI subgroupings have been correlated with treatment response, and because of the exploratory nature of this analysis no firm inferences can be made. Further studies are required to address the issue of whether there are any correlations between MRI findings and treatment response. Our dropout rate of 26% as a result of topiramate side effects is comparable with the other 2 major trials with topiramate in painful diabetic neuropathy in which the average dropout rate in the topiramate arm was 24%.27,33 The most commonly reported adverse events were similar among the 3 studies, with changes in sensorium, fatigue, paresthesias, and gastrointestinal disturbances being most frequent.27,33 Both of those other studies specified a maximum topiramate dose of 400 mg as well. The starting dose of topiramate was 25 mg in both studies, however, and the escalation was over 6 to 8 weeks, respectively. In summary, we cannot completely rule out the possibility that the apparent pain reduction we observed was due to chance or dropout bias, but we consider it more likely that topiramate has a small but real analgesic effect. A modest effect size might also explain the conflicting results in the diabetic neuropathy studies.27,33 Some pain researchers, particularly industry-funded multicenter groups, are reluctant to test new analgesics in a pain syndrome until there is published evidence for the sensitivity of the study methods. The ability of our methods to detect a modest degree of pain reduction on several of the outcome measures in a 29-patient crossover study should encourage other researchers to study this common condition. Our lumbar radiculopathy pa-
tients had relatively few confounding psychiatric comorbidities compared with patients with idiopathic back pain or many other chronic pain conditions. No patient met the PHQ-15 criteria for multisomatoform disorder. Only 6 of 42 enrolled patients had evidence of depression based on the Beck Depression Inventory. Similar to patients seen in primary care and general neurology settings, none of them had had a trial of gabapentin, and few had tried other medications commonly used to treat neuropathic pain syndromes (Table 1). It is difficult to integrate our marginally positive finding for topiramate into an evidence-based algorithm in the absence of other published randomized trials of drug treatment of chronic lumbar radiculopathy pain. Pain experts recommend that the choice among medications for treating neuropathic pain should be based on the comparison of the drug side effects and the patient’s comorbidities and preferences. On the basis of our observations of modest pain improvement, troublesome side effects, and frequent dropouts, we concluded that topiramate is at best marginally effective in the treatment of chronic lumbar radiculopathy in patients who can tolerate its prominent side effects. It is at best a second-line treatment in patients with lumbar radiculopathy refractory to first-line drugs such as gabapentin, tricyclic antidepressants, and opioids9 proven to be effective in other neuropathic pain syndromes. Future studies of topiramate should include a larger group of patients with homogenous MRI diagnoses by using a slower titration with smaller initial doses to minimize potential problems of differential dropout. In addition, prospective dose-response studies are needed to explore whether lower doses might achieve similar pain relief but improve topiramate’s therapeutic ratio.
References
4. Cohrs RJ, Gilden DH, Mahalingam R: Varicella zoster virus latency, neurological disease and experimental models: an update. Front Biosci 9:751-762, 2004
1. Atkinson JH, Slater MA, Williams RA, Zisook S, Patterson TL, Grant I, Wahlgren DR, Abramson I, Garfin SR: A placebocontrolled randomized clinical trial of nortriptyline for chronic low back pain. Pain 76:287-296, 1998
Acknowledgments We thank Kenneth Schmader and Carl Pieper for reviewing the manuscript, Kelly Cahill, Department of Nursing, Judith Starling of the NIH Pharmaceutical Development Service, and Sandra Jones of the NIH Patient Recruitment Branch for technical assistance; and Dr Nicolas Patronas, Department of Radiology, for his suggestions.
5. Colton T: Statistics in Medicine. 1st ed Boston, MA, Little Brown, 1974, p142
2. Avoli M: GABA-mediated synchronous potentials and seizure generation. Epilepsia 37:1035-1042, 1996
6. Dellemijn PL, van Duijn H, Vanneste JA: Prolonged treatment with transdermal fentanyl in neuropathic pain. J Pain Symptom Manage 16:220-229, 1998
3. Bajwa ZH, Sami N, Warfield CA, Wootton J: Topiramate relieves refractory intercostal neuralgia. Neurology 52: 1917, 1999
7. Deyo RA, Rainville J, Kent DL : What can the history and physical examination tell us about low back pain? JAMA 268:760-765, 1992
836
Topiramate in Chronic Sciatica
8. Deyo RA, Weinstein JN: Low back pain. N Engl J Med 344:363-370, 2001
the treatment of sciatica: A double-blind, randomized, controlled study. Reg Anesth Pain Med 24:534-540, 1999
9. Dworkin RH, Backonja M, Rowbotham MC, Allen RR, Argoff CR, Bennett GJ, Bushnell MC, Farrar JT, Galer BS, Haythornthwaite JA, Hewitt DJ, Loeser JD, Max MB, Saltarelli M, Schmader KE, Stein C, Thompson D, Turk DC, Wallace MS, Watkins LR, Weinstein SM: Advances in neuropathic pain: Diagnosis, mechanisms, and treatment recommendations. Arch Neurol 60:1524-1534, 2003
25. Murphy RW: Nerve roots and spinal nerves in degenerative disk disease. Clin Orthop Relat Res 129:46-60, 1977
10. Edwards KR, Glanz MJ, Button J: Efficacy and safety of topiramate in the treatment of painful diabetic neuropathy: a double-blind, placebo-controlled study. Neurology 54:A81, 2000 11. Fairbanks JCT, Davies JB, Couper J, O’Brien JP: The Oswestry low-back pain disability questionnaire. Physiotherapy 66:271-273, 1980 12. Fardon DF, Milette PC: Nomenclature and classification of lumbar disc pathology: Recommendations of the combined task Force of the North American Spine Society, American Society of Spine Radiology, and American Society of Neuroradiology. Spine 26:E93-E113, 2001
26. Pocock SJ: Cross over trials, in Clinical Trials, A Practical Approach. 4th ed. Chichester, UK, Wiley and Sons, 1983 27. Raskin P, Donofrio PD, Rosenthal NR, Hewitt DJ, Jordan DM, Xiang J, Vinik A: Topiramate vs placebo in painful diabetic neuropathy: Analgesic and metabolic effects. Neurology 63:865-873, 2004 28. Remmers AESU, LaMoreaux L, Young JP, Moore J, Poole M: Pregabalin treatment of patients with chronic low back pain. Presented at the American Pain Society Annual Meeting, Atlanta, Georgia, November 2-5, 2000 29. Schnitzer TJ, Gray WL, Paster RZ, Kamin M: Efficacy of tramadol in treatment of chronic low back pain. J Rheumatol 27:772-778, 2000 30. Severt CD, Sombati S: Topiramate selectively blocks kainate currents in cultured hippocampal neurones. Epilepsia 36:38, 1995
13. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM: Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain 94:149-158, 2001
31. Shank RP, Gardocki JF, Streeter AJ, Maryanoff BE: An overview of the preclinical aspects of topiramate: pharmacology, pharmacokinetics, and mechanism of action. Epilepsia 41(suppl 1):S3-S9, 2000
14. Franson RC, Saal JS, Saal JA: Human disc phospholipase A2 is inflammatory. Spine 17(Suppl):S129-S132, 1992
32. Sindrup SH, Jensen TS: Efficacy of pharmacological treatments of neuropathic pain: An update and effect related to mechanism of drug action. Pain 83:389-400, 1999
15. Gendreau M, Hufford MR, Stone AA: Measuring clinical pain in chronic widespread pain: selected methodological issues. Best Pract Res Clin Rheumatol 17:575-592, 2003 16. Gilron I, Booher SL, Rowan JS, Max MB: Topiramate in trigeminal neuralgia: a randomized, placebo-controlled multiple crossover pilot study. Clin Neuropharmacol 24:109111, 2001 17. Heliovaara M, Impivaara O, Sievers K, Melkas T, Knekt P, Korpi J, Aromaa A: Lumbar disc syndrome in Finland. J Epidemiol Community Health 4:251-258, 1987 18. Irvine EJ: Measurement and expression of risk: Optimizing decision strategies. Am J Med 117(Suppl 5A):2S-7S, 2004 19. Jensen MP, Karoly P, Braver S: The measurement of clinical pain intensity: A comparison of six methods. Pain 27: 117-126, 1986
33. Thienel U, Neto W, Schwabe SK, Vijapurkar U: Topiramate in painful diabetic polyneuropathy: findings from three double-blind, placebo-controlled trials. Acta Neurol Scand 110: 221-231, 2004 34. Tomlinson DR: The pharmacology of diabetic neuropathy. Diabetes Metab Rev 8:67-84, 1992 35. Ware JE: SF-36 health survey update. Spine 25:31303139, 2000 36. de C Williams AC, Richardson PH: What does the BDI measure in chronic pain? Pain 55:259-266, 1993 37. Wipf JE, Deyo RA: Low back pain. Med Clin North Am 79:231-246, 1995
20. Jensen MP, McFarland CA: Increasing the reliability and validity of pain intensity measurement in chronic pain patients. Pain 55:195-203, 1993
38. Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, Tugwell P, Campbell SM, Abeles M, Clark P: Criteria for the classification of fibromyalgia: Report of the Multicenter Criteria Committee. The American College of Rheumatology 33:160-172, 1990
21. Kroenke K, Spitzer RL, deGruy FV 3rd, Swindle R: A symptom checklist to screen for somatoform disorders in primary care. Psychosomatics 39:263-272, 1998
39. Woolf CJ, Mannion RJ: Neuropathic pain: Aetiology, symptoms, mechanisms, and management. Lancet 353: 1959-1964, 1999
22. Max MB, Lynch SA, Muir J, Shoaf SE, Smoller B, Dubner R: Effects of desipramine, amitriptyline, and fluoxetine on pain in diabetic neuropathy. N Engl J Med 326:1250-1256, 1992
40. Zeigler D, Lynch SA, Muir J, Benjamin J, Max MB: Transdermal clonidine versus placebo in painful diabetic neuropathy. Pain 48:403-408, 1992
23. McCarron RF, Wimpee MW, Hudkins PG, Laros G: The inflammatory effect of nucleus pulposus. Spine 12:760-764, 1987 24. Medrik-Goldberg T, Lifschitz D, Pud D, Adler R, Eisenberg E: Intravenous lidocaine, amantadine, and placebo in
41. Zhang XL, Velumian AA, Jones OT, Carlen PL. Modulation of voltage-activated calcium channels in dentate granule cells by topiramate. Epilepsia 41(suppl):S10-S16, 2000 42. Zvartau-Hind M, Din MU, Gilani A, Lisak RP, Khan OA: Topiramate relieves refractory trigeminal neuralgia in MS patients. Neurology 55:1587-1588, 2000