NEUROPATHIC PAIN SECTION

Transcription

1 bs_bs_banner Pain Medicine 2013; 14: Wiley Periodicals, Inc. NEUROPATHIC PAIN SECTION Original Research Articles Predicting to Pregabalin from Pretreatment Pain Quality: Clinical Applications of the Pain Quality Assessment Scalepme_ Arnold R. Gammaitoni, PharmD,* Steven S. Smugar, MD,* Mark P. Jensen, PhD, Bradley S. Galer, MD, James A. Bolognese, MS, Achilles Alon, PharmD,* and David J. Hewitt, MD* *Global Center for Scientific Affairs (A. R. Gammaitoni); Global Scientific and Medical Publications (S. S. Smugar); Early Clinical Development (A. Alon); Clinical Research (D. J. Hewitt), Merck Sharp & Dohme Corp., Whitehouse Station, New Jersey; Department of Rehabilitation Medicine, University of Washington School of Medicine, Seattle, Washington; Pain Group, Nuvo Research, Inc., West Chester, Pennsylvania; Clinical Trial Services, Cytel, Inc., Cambridge, Massachusetts, USA Reprint requests to: David J. Hewitt, MD, Clinical Research, Merck Sharp & Dohme Corp., Mailstop UG4C-06, 351 N. Sumneytown Pike, North Wales, PA 19454, USA. Tel: ; Fax: ; Disclosure/Conflict of interest information: Drs. Hewitt, Alon, and Smugar are employees of Merck and may potentially own stock and/or hold stock options in the company. Mr. Bolognese and Dr. Gammaitoni are former employees of Merck and may potentially own stock and/or hold options in the company. Dr. Jensen has received grant/research support from and has served as a consultant to Merck and has received consulting fees from RTI Health Solutions, Covidien, Bristol-Myers Squibb, Endo, Schwartz Biosciences, and Smith & Nephew within the past 12 months. Drs. Jensen, Gammaitoni, and Galer receive royalties from industry-sponsored use of the Pain Quality Assessment Scale but do not receive royalties from nonsponsored use. Support: The original trial and the present analysis were supported by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc. Summary: Analgesic response in neuropathic pain is variable. We evaluated the ability of the Pain Quality Assessment Scale to predict response and found that several baseline items significantly correlated with response to pregabalin in neuropathic pain. Measuring these variables could help identify patients most likely to respond. Abstract Objective. The aim of this study is to assess the Pain Quality Assessment Scale (PQAS) in predicting pregabalin in peripheral neuropathic pain (NP). Study Design. Post hoc analysis of a double-blind, placebo-controlled, enriched enrollment, randomized withdrawal trial evaluating pregabalin in 99 patients with NP who completed the PQAS, which comprises 20 questions regarding individual pain domains and qualities that are scored into three scales: paroxysmal, deep, and surface. Methods. Patients rated the average pain intensity and pain quality using the PQAS at baseline; average pain intensity was assessed again after 40 days of treatment with pregabalin. Associations between pretreatment PQAS scores and treatment response were estimated using Pearson s r. Logistic regression was used to identify pretitration PQAS scores contributing unique variance to predicting treatment response. Results. Fifty participants provided baseline PQAS scores and received pregabalin for the entire length of the study. Nine of 23 PQAS baseline scales and items were significantly associated with treatment response to pregabalin: the paroxysmal and deep scales, and the items assessing the following pain 526

2 domains and qualities: intensity, electric, tingling, cramping, radiating, throbbing, and deep (P values range, ; rs range, ). The PQAS items assessing sharp, hot, and unpleasant pain items demonstrated nonsignificant trends (P < 0.10) to be associated with treatment response. In the logistic regression analysis, pretitration PQAS scores had 77% sensitivity and 83% specificity to correctly identify pregabalin responders. Significantly correlated PQAS items had a sensitivity of 85% and specificity of 76%. Conclusion. Pretitration PQAS scores reliably predicted pregabalin responders in patients with NP. Key Words. Diabetic Neuropathy; Neuropathy; Pain Assessment; Pain Quality; Neuropathic Pain; Pain Quality Assessment Scale; Pregabalin; Introduction Pain response to analgesic medication is highly variable and may be related to disease, unique patient differences, underlying pain pathophysiology, and drug-related characteristics [1 5]. As such, patients often require multiple sequential trials with different treatments or treatment combinations before achieving adequate pain relief. During these trial periods, which based on guideline recommendations typically last for 4 6 weeks, patients may continue to experience pain. The ability to accurately predict patient response more quickly after initiation of therapy could greatly reduce the duration of the trial periods. Studies of various types of analgesic medications have shown that patients who experience relief early in the regimen (e.g., as early as within 2 weeks) are likely to continue to experience more ongoing relief, whereas most patients who do not experience early response are relatively less likely to achieve a subsequent response [6 8]. Although these findings could lead to faster decision regarding changing therapy in the absence of an early response, they still require an initial treatment period and provide no information on which patients are most likely to respond to a given pharmacology. Having the ability to predict response from patient history and/or baseline characteristics could allow clinicians to choose a treatment that the patient is likely to respond to without a preliminary trial-and-error period, saving the patient unnecessary time and possible adverse events associated with trying a potentially unsuccessful therapy. However, to date, studies investigating factors predictive of response have been highly inconsistent [9 11]. An important factor in considering pain and predicting therapeutic response is the multidimensional nature of pain. The experience of pain comprises a number of discrete pain qualities in addition to intensity; the pattern and ratings of which can vary by disease state and among individuals with the same pain condition. As a result, not all patients with the same diagnosis respond to treatment Predicting Pregabalin in Neuropathic Pain equally [12 20]. Recognizing and considering these patterns at baseline may potentially allow for a more precise initial treatment selection based on what is known regarding the efficacy of specific treatments in a given disease state or condition with a specific pain pathophysiology. To our knowledge, no one has directly examined the ability of pain quality measures to predict a priori response to treatment, or to distinguish treatment responders from nontreatment responders. To address this knowledge gap, we analyzed data from an enriched enrollment randomized withdrawal (EERW) study of pregabalin in peripheral neuropathic pain (NP) [21] to 1) examine the associations between baseline pain quality scores using the Pain Quality Assessment Scale (PQAS [14]) and improvement in pain intensity after (a) 40 days of treatment with pregabalin or (b) 21 days of pregabalin followed by 19 days of placebo treatment and 2) explore the sensitivity and specificity of the PQAS scales and items to distinguish pregabalin treatment responders from nonresponders. Methods Subjects and Study Design The subjects (N = 99) for this secondary analysis were participants in a published enriched enrollment randomized withdrawal study examining the effects of pregabalin on NP [21] and who provided PQAS data at each study assessment point. Detailed, descriptive information about the study design and the inclusion/exclusion criteria for the study participants were described in the primary outcome paper from this study. Briefly, and germane to the present analysis, patients were 18 years of age with moderate to severe chronic NP resulting from peripheral diabetic neuropathy, postherpetic neuralgia, small-fiberpredominant neuropathy, or idiopathic sensory neuropathy. Patients taking chronic pain medication for 1 month prior to the prestudy visit were permitted to continue medication at the same dose. Patients taking pregabalin or gabapentin who had adequate pain control were permitted to enter the study if they were willing to reduce their current dose to a lower dose in order to meet entry criteria for pain intensity. Patients who had previously failed to respond to pregabalin at doses 450 mg/day or gabapentin at doses 1,800 mg/day were excluded. Rescue medication with acetaminophen (4,000 mg/day maximum) or, if necessary acetaminophen/hydrocodone (10/1,000 mg/day maximum, for no longer than 3 days), was allowed only during the single-blind dosetitration period. Rescue medication of any kind was not permitted during the maintenance period or the doubleblind period. The primary purpose of the original study was to determine the effects of enrichment on assay sensitivity in the context of an EERW design. To achieve this, eligible patients entered a 12-day single-blind dose titration period to optimize pregabalin dosage, which started at 50 mg t.i.d. and increased by 50 mg/dose every 3 days up to 200 mg t.i.d., or until dose-limiting adverse 527

3 Gammaitoni et al. experiences were encountered. Patients who had been on pregabalin or gabapentin at screening who tapered their medication to meet entry criteria began titration with the same pregabalin dose (or gabapentin-equivalent dose) as the last dose taken during the screening period. Patients then entered a 9-day maintenance phase, after which they were randomized in a double-blind fashion to either placebo (i.e., withdrawal, tapered one-third/day) or to continue pregabalin treatment for 19 days. Because the primary purposes of the current set of secondary analyses were to 1) examine the ability of pain quality measures to predict response to treatment and 2) determine if pain quality assessed at pretreatment can distinguish treatment responders from nontreatment responders, we wanted to study primarily the 50 subjects who received pregabalin for all 40 days of participation. However, to help ensure that the predictors were associated with response to active treatment specifically and not necessarily response to placebo, we included the 49 participants who were switched to placebo for the last 19 days of the study. For the purposes of this report, comparator group names are based on the treatment received during the randomized withdrawal phase: i.e., pregabalin or placebo. Study Measures Outcome was assessed using the first item of the PQAS [14], which asks patients to rate their average pain intensity over the past week on a 0 10 Numerical Rating Scale, with 0 = No pain and 10 = The most intense pain sensation imaginable. Pain quality was assessed using the other 19 items of the PQAS [14]. Each PQAS item assesses a different pain domain or quality. Fifteen of these can be used to create three PQAS scale scores that assess paroxysmal, surface, and deep pain [18]. Support for the validity of these three scales comes from an initial exploratory factor analysis and subsequent confirmatory factor analysis that demonstrated that the items clustered into these three overarching domains in a consistent manner (across three pain conditions) [18]. The responses to the PQAS items can also be analyzed individually. The study participants completed the PQAS at the baseline/prestudy visit as well as at the end of the study or the discontinuation visit. The current analyses used all of the baseline PQAS responses and the study discontinuation average pain intensity PQAS item. Responder Definition Patients were classified as treatment responders if they reported a 2-point improvement in average daily pain intensity (on a 0 10 scale) from baseline to study endpoint [22]. Statistical Analysis We examined the associations between baseline PQAS items and change in average pain intensity from baseline to study discontinuation separately for 1) the participants who received pregabalin for the entire study (N = 50; pregabalin group) and 2) the participants who received pregabalin for 21 days and were then switched to placebo for 19 days (N = 49; placebo group). We then used logistic regression analyses to identify the ability of the PQAS items to distinguish treatment responders from the nonresponders for those who took pregabalin for the duration of the study. Given the large number of potential predictors (19 PQAS items) and the small number of subjects in the treatment group, we used a stepwise forward entry procedure to identify the PQAS scales and items associated with responder group. To determine if a composite PQAS predictor score might be useful in identifying treatment responders, we first computed this score as a mean of individual items that emerged as significant predictors in the regression analysis using the following formula: Item# 1rating + Item# 2 rating + Item# 3 rating, etc.. Total number of items To identify potential cutoffs for classifying patients as responders and nonresponders, we divided the PQAS predictor score into tertiles and determined the rates of treatment response/nonresponse for each PQAS predictor score tertile. Chi-square tests were used to determine if PQAS response score tertile was associated with treatment response in the two treatment groups. No adjustment for multiple statistical testing was performed as the analyses were post hoc and exploratory. Results Participants Baseline and study discontinuation PQAS data were available for 99 of the participants who completed all phases of the study. Of these, 29 (58%) were identified as treatment responders in the pregabalin group and 22 (45%) were identified as treatment responders in the placebo group. Of the 99 participants, 87 (87.9%) were white, 6 (6.1%) were African American, and the remaining six were Native American (2.0%), multiracial (2.0%), Hispanic (1.0%), or Asian (1.0%). Fifty-four (54.5%) of the sample were men. The average age of the participants was 59.9 years (range, years). The majority (69, or 69.7%) presented with painful diabetic neuropathy. Subsets of the sample also had postherpetic neuralgia (8, or 8.1%), small fiber predominant neuropathy (5, or 5.1%), or idiopathic sensory neuropathy (17, or 17.2%). Thirty-one (31.3%) of the participants were taking either pregabalin or gabapentin when they began the study, and 64 (64.6%) of the participants were taking a concomitant analgesic throughout the study trial. Baseline PQAS scales and individual item scores were similar between the pregabalin group and the placebo group. 528

4 Table 1 Pearson correlation coefficients between baseline Pain Quality Assessment Scale (PQAS) scales and items baseline to study discontinuation change in average pain intensity for participants who received pregabalin (treatment group) and pregabalin followed by placebo (placebo group) PQAS Scale or Item Treatment Group Pregabalin Placebo N = 50 N = 49 PQAS scale PQAS paroxysmal 0.42** 0.14 PQAS surface PQAS deep 0.29* 0.08 PQAS item Intense 0.43** Sharp Hot Dull Cold Sensitive Tender Itchy Shooting Numb Electrical 0.40** 0.01 Tingling 0.28* Cramping 0.28* 0.00 Radiating 0.36** 0.01 Throbbing 0.36* 0.17 Aching Heavy Unpleasant Deep 0.35* 0.15 Surface * P < 0.05; ** P < 0.01 Associations Between Baseline Pain Quality and Treatment Outcome Table 1 presents the zero-order correlation coefficients between the baseline PQAS scale scores and items and change in pain intensity over the course of the study for the participants who received pregabalin (N = 50) and placebo (N = 49). As can be seen, two PQAS scales (paroxysmal and deep) and six items (electrical, tingling, cramping, radiating, throbbing, and deep) demonstrated a significant (P < 0.05) association with change in average pain intensity in the treatment group. The sharp, hot, and unpleasant items demonstrated nonsignificant trends (P < 0.10) associated with treatment response. However, no PQAS item or scale evidenced a significant association with response in the placebo group. Predicting Pregabalin in Neuropathic Pain Ability of Pain Quality to Discriminate Between Treatment Responders and Nonresponders In the logistic regression analysis, using all 19 PQAS ratings as potential predictors of responder group, only the PQAS paroxysmal scale demonstrated a significant (P < 0.05) association with treatment response classification. The model with this predictor had both a high sensitivity (76.9%) and specificity (83.3%). The overall accuracy rate was 80.0% (Table 2). When limiting the regression analysis to the aforementioned six PQAS items significantly associated with being a treatment responder, the model had 84.6% sensitivity, 70.8% specificity, and an overall accuracy rate of 78.0% (Table 3). Pregabalin Prediction based on PQAS Predictor Score Tertile As shown in Table 4, response to pregabalin was strongly and significantly (P < 0.001) related to the PQAS predictor score tertile: 100% of patients in the highest tertile were Table 2 Sensitivity and specificity of the Pain Quality Assessment Scale (PQAS) to predict response all PQAS items Observed Predicted Nonresponse Nonresponse Overall percentage 80.0 Percentage Correct Table 3 Sensitivity and specificity of the Pain Quality Assessment Scale (PQAS) to predict response significant factors from logistic regression only* Observed Predicted Nonresponse Nonresponse Overall percentage 78.0 Percentage Correct * Significant factors from the logistic regression included the PQAS paroxysmal and deep scales, and the intense, electrical, tingling, cramping, radiating, throbbing, and deep items. 529

5 Gammaitoni et al. Table 4 Percentage of pregabalin responders based on baseline Pain Quality Assessment Scale (PQAS) responder score tertile Baseline PQAS Pregabalin Predictor Score Tertile Pregabalin* Placebo* Responders Nonresponders Responders Nonresponders Lowest tertile 22.2% 77.8% 37.5% 62.5% Middle tertile 61.1% 38.9% 40.0% 60.0% Highest tertile 100.0% 0% 55.6% 44.4% Chi-square (P < 0.001) 1.33 (P = 0.515) * Groups refer to the treatment received during the randomized withdrawal period. Therefore, the pregabalin group received pregabalin during the titration/maintenance period and continued receiving pregabalin during the withdrawal period. The placebo group received pregabalin during the titration/maintenance period and placebo during the withdrawal period. However, response in the placebo group was for response to pregabalin for the titration/withdrawal period followed by placebo during the withdrawal phase. That is, it did not measure response to placebo. responders, compared with 61.9% in the middle tertile, and 22.2% in the bottom tertile. PQAS predictor score tertiles were not related to treatment response in patients who received placebo during the randomized withdrawal phase (P = 0.515). Discussion To our knowledge, this is the first published study demonstrating that a pretreatment pain quality measure can predict a priori response to treatment and to distinguish treatment responders from nontreatment responders. We hypothesized that if baseline pain qualities were able to predict response to pregabalin specifically (and not to treatment in general), there would be 1) significant associations between some of the PQAS scales and items and change in pain in the participants who received pregabalin for the entire duration of the study and 2) at least some of those PQAS scales and items identified as being associated with response to pregabalin would not be associated with response to pregabalin followed by placebo. Consistent with the study hypothesis, we found that in the group receiving pregabalin for both the titration/maintenance and withdrawal periods, baseline values for the PQAS paroxysmal and deep scales, and the intense, electrical, tingling, cramping, radiating, throbbing, and deep items significantly correlated with change in average pain intensity. We also found that response in the pregabalin group was directly and significantly related to a baseline PQAS predictor score tertile. None of the PQAS items or scales, including the PQAS predictor score, was significantly associated with response to pregabalin followed by placebo. Overall, the PQAS items had reasonably high sensitivity and specificity to predict treatment response. Perhaps not surprisingly, when only significant variables were used to model prediction, sensitivity increased from 77% to 84%, while specificity decreased from 83% to 71%. Thus, the PQAS items appear to indicate with a fair degree of accuracy who will respond to pregabalin in NP but are slightly less accurate when predicting nonresponse. Our results are consistent with a previous secondary analysis using the same dataset as the present analysis, in which we found that pregabalin had a greater effect on the PQAS paroxysmal pain scale than on the deep or surface pain scales, or on the PQAS pain intensity item in patients with NP [23]. Our results also corroborate those found in previous studies of pain quality scales in various pain states. Namely, that the experience of pain comprises a number of descriptive factors that drive response and that not all descriptors respond to treatment equally. For instance, Dworkin et al. [17] found substantial differences in Neuropathic Pain Scale scores (the progenitor scale of PQAS) in patients with NP compared with patients with either low back pain or osteoarthritis, whereas there were few differences between osteoarthritis and low back pain. Similarly, Jensen et al. s initial work on the PQAS [14] showed variable effects of lidocaine patch or steroid injection on pain qualities in patients with carpal tunnel syndrome. Further studies of the PQAS also demonstrated differential PQAS symptom patterns and treatment effects among various pain states [15,18,19]. Importantly, these findings also further confirm the importance of evaluating pain quality response in addition to pain intensity response in both clinical trials and in clinical practice. Our finding of a direct relationship between baseline severity and treatment response may be related to spontaneous improvement, regression to the mean, drug effect or a combination of these and other unknown factors. Although a relationship between more severe baseline disease and response has been shown in other pain states [9], the overall relationship is inconsistent [10,11,24]. In addition, baseline severity of pain quality is clearly not the only or even primary factor driving the significant associations between the PQAS items and treatment response. If it were, we would anticipate that all or at least most of the PQAS items and scale baseline scores would show significant associations with response for both the pregabalin/pregabalin and pregabalin/ placebo groups. The fact that a specific subset of PQAS items and scales were significantly associated with treat- 530

6 ment outcome in the pregabalin/pregabalin group alone (and not the pregabalin/placebo group) provides strong support that the PQAS items identified predict response to pregabalin specifically, and not to pain treatment in general. Although there has been a call in recent past to apply mechanism-based treatment approaches, doing so in the clinical setting is difficult, given the impracticality of using Quantitative Sensory Testing (QST) and other types of invasive tests to determine the potential physiological mechanisms underlying the pain in a given patient. We have proposed that it might be possible to assess pain qualities using a patient reported outcome instrument (PQAS) as a more practical approach to mechanismbased therapy [12,14]. The current findings support this possibility. Although the specific mechanisms of pregabalin are not known, and we cannot yet be certain that that PQAS items clearly distinguish different mechanisms, the fact that surface pain was not predictive of outcome and that surface pain is thought to represent pathophysiological events in peripheral neuropathies, our findings suggest the possibility that pregabalin may operate more through central than peripheral mechanisms. However, such speculation will need to be confirmed by further evidence supporting the PQAS items as reflective of pain mechanisms as well as further evidence showing that surface pain qualities do not predict outcome to pregabalin. The primary limitation of our analysis is that it was a post hoc analysis from a study that was not designed or powered to assess baseline PQAS predictors of response. Our results should, therefore, be considered as preliminary and in need of prospective replication and cross validation. Another limitation is that we divided the baseline scores by tertile rather than by actual score values. This means that if we confine our responder selection to the top tertile, we would eliminate 67% of screened patients. If we confine to the top two tertiles, we would eliminate 33%. Finally, the low sample size (N = 99) available for analyses precluded our ability to reliably test for potential moderation effects related to pain type, sex, or age. It is possible that the strength of the associations between pain quality and outcome might vary as a function of one or more of these variables. Future researchers should examine this possibility in studies with larger sample sizes. In conclusion, we were able to demonstrate that certain pain qualities measured prior to treatment for NP were able to predict subsequent response. Specifically, a number of subscale and individual items from the PQAS significantly predicted response to pregabalin in NP. Future research is needed to determine the reliability and generalizability of these findings. To the extent that baseline patterns of pain quality in patients with NP or other pain conditions are consistently found to predict treatment outcome to specific pharmacologic agents, clinicians will be in a better position to more accurately tailor Predicting Pregabalin in Neuropathic Pain initial treatment selection, and therefore help their patients obtain greater pain relief sooner with less risk of adverse events associated with unwarranted therapeutic drug trials and thus improving the overall benefit risk for the interventions selected. In addition, the ability to predict responders based on baseline pain qualities could enhance treatment effects seen in clinical trials. Acknowledgments This study was supported by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, NJ. All authors are responsible for the work described in this article, and they contributed to the conception, data interpretation, and drafting of the manuscript and/or revising the manuscript for important intellectual content. All authors provided final approval of the version to be published. This manuscript was written collectively and in entirety by all named authors. No external writing assistance was provided. The authors would like to thank Sheila Erespe of Merck for her editorial assistance and facilitating the submission. References 1 Furst DE. Are there differences among nonsteroidal antiinflammatory drugs? Comparing acetylated salicylates, nonacetylated salicylates, and nonacetylated nonsteroidal antiinflammatory drugs. Arthritis Rheum 1994;37: Huskisson EC, Woolf DL, Balme HW, Scott J, Franklin S. Four new anti-inflammatory drugs: s and variations. Br Med J 1976;1: Preston SJ, Arnold MH, Beller EM, Brooks PM, Buchanan WW. Variability in response to nonsteroidal anti-inflammatory analgesics: Evidence from controlled clinical therapeutic trial of flurbiprofen in rheumatoid arthritis. Br J Clin Pharmacol 1988;26: Brooks PM, Day RO. Nonsteroidal antiinflammatory drugs Differences and similarities. N Engl J Med 1991;324: Levy RA, Smith DL. Clinical differences among nonsteroidal antiinflammatory drugs: Implications for therapeutic substitution in ambulatory patients. DICP 1989;23: Bingham CO III, Smugar SS, Wang H, Tershakovec AM. Early response to COX-2 inhibitors as a predictor of overall response in osteoarthritis: Pooled results from two identical trials comparing etoricoxib, celecoxib and placebo. Rheumatology (Oxford) 2009;48: Battisti WP, Katz NP, Weaver AL, et al. Pain management in osteoarthritis: A focus on onset of efficacy A 531

7 Gammaitoni et al. comparison of rofecoxib, celecoxib, acetaminophen, and nabumetone across four clinical trials. J Pain 2004;5: Dougados M, Gueguen A, Nakache JP, et al. Ankylosing spondylitis: What is the optimum duration of a clinical study? A one year vs a 6 weeks non-steroidal anti-inflammatory drug trial. Rheumatology (Oxford) 1999;38: Bingham CO III, Smugar SS, Wang H, Peloso PM, Gammaitoni A. Predictors of response to cyclooxygenase-2 inhibitors in osteoarthritis: Pooled results from two identical trials comparing etoricoxib, celecoxib, and placebo. Pain Med 2011;12: Bradley JD, Katz BP, Brandt KD. Severity of knee pain does not predict a better response to an antiinflammatory dose of ibuprofen than to analgesic therapy in patients with osteoarthritis. J Rheumatol 2001;28: Case JP, Baliunas AJ, Block JA. Lack of efficacy of acetaminophen in treating symptomatic knee osteoarthritis: A randomized, double-blind, placebocontrolled comparison trial with diclofenac sodium. Arch Intern Med 2003;163: Galer BS, Jensen MP. Development and preliminary validation of a pain measure specific to neuropathic pain: The Neuropathic Pain Scale. Neurology 1997;48: Jensen MP, Dworkin RH, Gammaitoni AR, et al. Assessment of pain quality in chronic neuropathic and nociceptive pain clinical trials with the Neuropathic Pain Scale. J Pain 2005;6: Jensen MP, Gammaitoni AR, Olaleye DO, et al. The pain quality assessment scale: Assessment of pain quality in carpal tunnel syndrome. J Pain 2006;7: Jensen MP, Dworkin RH, Gammaitoni AR, et al. Do pain qualities and spatial characteristics make independent contributions to interference with physical and emotional functioning? J Pain 2006;7: Jensen MP, Friedman M, Bonzo D, Richards P. The validity of the neuropathic pain scale for assessing diabetic neuropathic pain in a clinical trial. Clin J Pain 2006;22: Dworkin RH, Jensen MP, Gammaitoni AR, Olaleye DO, Galer BS. Symptom profiles differ in patients with neuropathic vs non-neuropathic pain. J Pain 2007;8: Victor TW, Jensen MP, Gammaitoni AR, et al. The dimensions of pain quality: Factor analysis of the Pain Quality Assessment Scale. Clin J Pain 2008;24: Gould EM, Jensen MP, Victor TW, et al. The pain quality response profile of oxymorphone extended release in the treatment of low back pain. Clin J Pain 2009;25: Jensen MP, Chiang YK, Wu J. Assessment of pain quality in a clinical trial of gabapentin extended release for postherpetic neuralgia. Clin J Pain 2009;25: Hewitt DJ, Ho TW, Galer B, et al. Impact of responder definition on the enriched enrollment randomized withdrawal trial design for establishing proof of concept in neuropathic pain. Pain 2010;152: 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 2001;94: Jensen MP, Gammaitoni A, Bolognese J, et al. The pain quality response profile of pregabalin in the treatment of neuropathic pain. Clin J Pain 2011 Dec 7 [Epub ahead of print]. doi: /AJP. 0b013e31823f9e Detora LM, Krupa D, Bolognese J, Sperling RS, Ehrich EW. Rofecoxib shows consistent efficacy in osteoarthritis clinical trials, regardless of specific patient demographic and disease factors. J Rheumatol 2001;28: