OUP user menu

Patient-reported Outcomes in Subjects with Neuropathic Pain Receiving Pregabalin: Evidence from Medical Practice in Primary Care Settings

Ana Navarro MD, María T. Saldaña MD, Concepción Pérez MD, Sandra Torrades PhD, Javier Rejas MD
DOI: http://dx.doi.org/10.1111/j.1526-4637.2010.00824.x 719-731 First published online: 1 May 2010


Objective. To evaluate the effect of pregabalin on different patient-reported outcomes in subjects with neuropathic pain treated under usual medical practice conditions in primary care settings.

Patients and Methods. Secondary analysis of a 12-week, multicenter, naturalistic, and prospective study on 18 years of age or older patients of both genders with chronic pain (of at least 6 months) due to diabetic neuropathy, post-herpetic, or trigeminal neuralgia refractory to the previous analgesic treatment (at least one drug).

Subjects. Assessed at baseline and end of the study visits by the following questionnaires: Short Form McGill Pain Questionnaire, Sheehan Disability Inventory, Medical Outcomes Study Sleep Scale, Hospital Anxiety and Depression Scale, and EQ-5D.

Results. The analysis included 1,354 patients not previously exposed to pregabalin; 598 patients (44%) received monotherapy with pregabalin as a substitute of the previous treatment, in 589 patients (44%) pregabalin was added to the existing therapy, and the treatment schedule of the other 167 patients (12%) did not include pregabalin. After 12 weeks of treatment, significant improvements in all effectiveness assessments were observed in all of the three groups, these being significantly greater in the groups receiving pregabalin, with large effect sizes in most health outcome measures.

Conclusion. Under usual medical practice conditions, patients with chronic pain of peripheral neuropathic origin receiving pregabalin both in monotherapy and as add-on therapy showed substantial improvements in severity of pain and in the spectrum of associated symptoms, such as sleep disturbances, mood disorders, disability, and health-related quality of life. Further clinical trials are needed to confirm these findings.

  • Pregabalin
  • Patient-reported-Outcomes
  • Pain
  • Primary Care Settings
  • Medical Practice
  • Clinically Meaningful


Neuropathic pain (NeP) is a frequent type of chronic pain in clinical practice [1]. Due to its severity, chronicity, co-morbidities, and impact on the individual and society, NeP is especially challenging. Compared with patients with chronic non-NeP and regardless of pain intensity, patients with NeP show poorer health status and greater levels of disability [2]. In addition, anxiety, sleep, and mood disorders are frequently present among these patients [3,4]. Furthermore, a reciprocal relationship seems to exist between pain and the above-mentioned disorders; the more intense the pain is, the more severe the symptoms of anxiety and depression and sleep disturbances [3]. In turn, the presence of anxiety, depression, or sleep impairment may increase the severity of pain [5–7]. NeP significantly reduces the quality of life of affected patients [1] to a greater extent than background disease (e.g., diabetes) [8] and chronic non-NeP [2]. Work and domestic productivity of patients is also affected [9,10], and health resource consumption has been shown to increase [11,12].

Symptoms of NeP vary considerably from patient to patient and, in general, are quite resistant to treatment with commonly prescribed analgesic drugs [13]. Furthermore, patients with NeP often receive suboptimal treatment (i.e., inappropriate drug therapy and/or use of sub-therapeutic doses) [13,14], which increases the disease burden [15–17]. In fact, although treatments are available for NeP, including antidepressant drugs, tramadol, opiods, and several anticonvulsive drugs, the results of a recent systematic review suggest that, due to their well-balanced efficacy and tolerability, anticonvulsive drugs, such as gabapentin or pregabalin (PGB), should be considered first-line treatments for peripheral NeP [18–20].

PGB is an alpha2-delta ligand that displays analgesic, anxiolytic, and anticonvulsive properties [21]. In randomized, placebo-controlled clinical trials, PGB has demonstrated efficacy for pain relief of patients with diabetic neuropathy and peripheral postherpetic neuralgia, significantly improving affective symptoms, sleep, and the quality of life [22]. However, the use of drugs in clinical trials markedly differs from the use in clinical practice in a variety of aspects, limiting the generalization of clinical trial results to the more heterogenous population and treatment conditions in clinical practice [23]. In this sense, naturalistic studies may provide important information on the effectiveness of a particular treatment in real clinical practice settings [24]. So far, only two studies have evaluated the use of PGB for NeP under routine clinical practice [25,26]. Freynhagen et al. report that PGB decreased pain and improved sleep in a small sample of patients with heterogenous neuropathic conditions treated in a tertiary care setting [25]. In the second study, 4-week PGB treatment improved pain, sleep, and anxiety in patients with diabetic peripheral neuropathy or post-herpetic neuralgia [26].

This work presents evidence from clinical practice showing the effect of PGB on different patient-reported outcomes (PROs), including pain severity, and representing the multidimensional nature of this condition in a wide group of patients with NeP treated under usual clinical practice conditions in primary care settings (PCS).

Patients and Methods

Study Design

This work presents the results of a secondary analysis of a multicenter, observational (naturalistic), and prospective 12-week study, whose objective was to determine the cost of treatment of refractory NeP (due to painful diabetes neuropathy, post-herpetic neuralgia, or trigeminal neuralgia) under real-life conditions in PCS: the LIDO study [27]. The study was carried out between September 2005 and April 2006, and 391 primary care physicians representative of all Spanish territories participated. Each investigator was asked to recruit the first four consecutive patients who met the selection criteria. The study was of non-intervention nature, and the analgesic treatment prescribed was determined by the clinical judgment of the doctor responsible for patient management. Doctors could substitute the previous treatment by one or several other drugs or add a new drug to the existing therapy. In accordance with the Spanish recommendations, the study was approved by the Ethics Committee of Clinical Research of Hospital de la Princesa (Madrid), and it was conducted in agreement with the principles contained in the Declaration of Helsinki for studies in humans.

The objective of this secondary analysis was to compare the effect of two patterns of PGB treatment, add-on and monotherapy, when compared with a pattern of NeP treatment not including PGB.

Study Population

The study included patients 18 years of age or older from both genders, and a diagnosis of NeP secondary to diabetic neuropathy, post-herpetic neuralgia, or trigeminal neuralgia. The subjects were refractory to previous analgesia and suffered chronic pain with a duration of over 6 months. Refractory was defined as no pain reduction after treatment with at least one course of an analgesic in monotherapy. Additionally, study requirements included a DN4 NeP questionnaire (see below) score greater than or equal to 4, a cultural and educational level sufficient to complete health questionnaires written in Spanish, and patient informed consent to inclusion of their data in a database for posterior analysis. The secondary analysis only included those patients who had fulfilled the previously mentioned selection criteria and who had not received PGB treatment before study initiation.

The sample size of the LIDO study was defined in accordance with its primary end point, i.e., to determine health care resources use and costs after a 12-week follow-up period under usual medical practice conditions in PCS. Thus, no sample size was predetermined for the secondary analysis presented in this work.

Clinical Assessments and Instruments of Measurement

During the 12-week follow-up period, patients were assessed two times, at baseline and at the study end. During baseline visit, the Spanish version of the NeP diagnostic questionnaire DN4 was completed [28], selection criteria were verified, and socio-demographic and disease and treatment duration data were collected, in addition to data related to the use of health care and non-health care resources during the 12 weeks prior to the inclusion in the study, in accordance with patient medical records. These data were collected directly by means of chart review audits and patient interview. At baseline and at the end of the 12-week follow-up visits, patients had to complete the Short Form McGill Pain Questionnaire (SF-MPQ), the Sheehan Disability Inventory (SDI), the Medical Outcomes Study Sleep Scale (MOS-Sleep), the Hospital Anxiety and Depression Scales (HADS), and the EQ-5D questionnaire. Furthermore, patients had to complete a diary in which they recorded weekly pain intensity (SF-MPQ visual analog scale [VAS]) and patient's health status (health status VAS from the EQ-5D).

The NeP diagnostic questionnaire DN4 (previously called Inventory of Neuropathic Pain Symptoms) consists of 10 items describing different pain characteristics and allows the distinction between NeP and non-NeP. A score of at least 4 out of 10 possible points is considered NeP with an 83% sensitivity and a 90% specificity [28–30].

The primary component of the SF-MPQ consists of 15 descriptors (11 sensory and 4 affective) assessed in a scale of intensity ranging from 0 = no pain to 3 = severe pain [31]. By adding the different intensities, three pain scores are obtained: sensory, affective, and total. The second part of the SF-MPQ consists of a 100-mm VAS to assess the pain intensity of patients during the previous week. The third part measures pain intensity at the time of assessment by an ordinal six-point scale, ranging from 0 = no pain to 5 = unbearable. The SDS is a simple and common instrument assessing functional impairment of patients in three different domains: work, social life, and family life/home responsibilities [32]. Each domain is evaluated on an 11-point scale, from 0 = no impairment to 10 = extremely impaired. The instrument contains two additional items assessing perceived stress and perceived social support; an 11-point scale is used to assess perceived stress, while an 11-point scale representing percentages, from 0% = no support at all to 100% = ideal support, is used to assess perceived social support. This inventory provides three scores: a disability score consisting of the sum of the scores of the first three items ranging from 0 to 30, the score of the fourth item or perceived stress with a range of 0–10, and the score of the fifth item or social support, which, opposite to the other four, is inverted.

The MOS-Sleep is a self-assessed questionnaire evaluating key aspects of sleep [33]. It consists of 12 items composing six subscales or domains: sleep disturbances, snoring, shortness of breath or headache upon awakening, adequacy of sleep, day somnolence, and amount of sleep. In addition, the MOS-Sleep Scale provides a summary index of sleep disturbances that can be obtained from nine of its item scores; the higher the score, the worse the sleep, with the exception of the amount of sleep and adequacy of sleep dimensions, which are scored in the opposite direction. In patients with NeP, this scale has shown appropriated psychometric properties [34]. The HADS is a self-assessed instrument as well, consisting of 14 items, seven items exploring depression symptoms and the other seven exploring anxiety symptoms [35]. Each item score ranges from 0 to 3, where 0 represents the absence of that symptom and 3 the highest severity or frequency of the symptom. By adding the seven items of each subscale, two scores ranging from 0 to 21 are obtained, for depression and anxiety (HADS-D and HADS-A), respectively.

The EQ-5D is designed to assess patient's perceived health status [36]. This is a five-item generic measure of health status to assess the level of abnormality on five dimensions: movement, self-care, day life activities, pain/discomfort, and anxiety/depression. The scores of these five items may be used to calculate a utility index, ranging from −0.6 to 1.0, with the highest scores representing better health status. The instrument also includes a 20-cm VAS (EQ-5D) ranging from 0 = the worst imaginable health status to 100 = the best imaginable health status.

Statistical Analysis

For statistical analyses, patients were classified into three groups depending on the treatment initiated according to clinical judgment at baseline visit: patients to whom one or several drugs other than PGB were substituted or added to the previous treatment (other-treatment group), patients to whom monotherapy with PGB was prescribed as a substitute of the previous therapy (PGB monotherapy group), and patients to whom PGB was added to the previous therapeutic schedule (PGB add-on group).

Patients' baseline characteristics were described by means and standard deviations (SDs) for quantitative variables and by distributions of absolute and relative frequencies for qualitative variables. The Kolmogorov–Smirnov test was used to verify the normal distribution of quantitative variables, and variance analyses (anova), Kruskall–Wallis tests, and chi2 tests were used to ensure the homogeneity of baseline variables in the three assessment groups.

The proportion of patients with a reduction of at least 50% in pain intensity as rated by the SF-MPQ pain VAS was calculated and defined as responder. Calculations were made of the cumulative number of days with no or mild pain (<40 mm in the SF-MPQ VAS). Furthermore, changes from baseline in the previously mentioned scales and subscales scores, including pain intensity VAS, were used as outcomes variables. A between-group comparison of the change from baseline of quantitative variables was carried out using an analysis of covariance (ancova), adjusted to the baseline score of each scale and to the number of previous drugs. For ordinal qualitative variables, the change from baseline was assessed using a model of multivariate logistic regression, adjusted to the baseline score and to the number of previous drugs. All analyses were performed on the population of included patients completing the 12-week follow-up, for which the change in baseline variables could be calculated. The statistical significance of between-group comparisons was adjusted using the method of Tukey for multiple comparisons. All statistical tests were two sided and were considered significant when p < 0.05.

In order to evaluate the relevance or clinical significance of the change in the different measures, the obtained effect size was calculated through the difference in means, before and after treatment, of a particular measure divided by the combined SD of that measure before treatment [37]. For effect size interpretation, the established criterion of considering 0.20–<0.50 small size effect, ≥0.50 and <0.80 moderate, and ≥0.80 large was used [37].


Patient Disposition

A total of 1,845 patients were included in the LIDO study, 1,354 of which had not been previously exposed to PGB and fulfilled the secondary analysis inclusion-exclusion criteria. The analysis was run on 1,309 patients (96.7%) who completed the 12 weeks of the study. Lost to follow-up were due to adverse events in 11 (0.8%) cases, unknown in 8 (0.6%), patients' decision in 17 (1.3%), and to other causes in the remaining 9 (0.7%) cases (Figure 1). The most frequent cause for NeP was diabetic neuropathy (54.4%), followed by post-herpetic neuralgia (33.8%) and trigeminal neuralgia (11.8%). No differences were found on the distribution of diagnoses between the two treatment groups.

Figure 1

Patient disposition. PGB = pregabalin; I/EC = Inclusion /exclusion criteria.

Baseline Demographics and Clinical Characteristics

The recruited population was about 59 years of age, with a predominance of women, and approximately half of the patients were not working (Table 1). The mean duration of NeP was 2 years; most patients were treated with at least one drug, and the most frequent previous treatments were non-steroidal anti-inflammatory drugs (NSAIDs). Compared with the other groups, patients from the PGB-monotherapy group had received a lower mean number of previous treatments (Table 1). Furthermore, significant differences were found in the frequency of exposed patients according to the previous type of analgesic. Thus, NSAIDs were more frequently used in the PGB add-on group and anti-epileptic drugs in the other-treatment group, while opiates were less frequently used in the PGB monotherapy group (Table 1).

View this table:
Table 1

Demographics and clinical characteristics

CharacteristicOther treatments
(N = 167)
PGB monotherapy
(N = 598)
PGB add-on
(N = 589)
Gender (female), n (%)83 (55.3)309 (60.4)282 (57.8) 0.4874
Age, mean (SD)60.4 (12.4)58.6 (12.5)59.7 (13.0) 0.1486
Body mass index (kg/m2)26.9 (3.5)27.3 (3.8)27.3 (3.9) 0.4673
Civil status (married or with couple), n (%)107 (69.0%)378 (66.4%)386 (67.6%) 0.6041
Working status, n (%) 0.3835
  Active54 (32.5)204 (34.6)181 (31.0)
  Housewife31 (18.7)75 (12.7)87 (14.9)
  Off sick10 (6.0)45 (7.6)62 (10.6)
  Unemployed3 (1.8)15 (2.6)11 (1.9)
  Retired59 (35.5)208 (35.3)216 (37.0)
  Does not practice9 (5.4)42 (7.1)27 (4.6)
Time of progress (years), mean (SD)2.1 (3.1)1.9 (3.4)2.0 (3.4) 0.7130
Diagnosis (%) 0.2852
Diabetic neuropathy47.958.253.0
Post-herpetic neuralgia37.231.236.9
Trigeminal neuralgia14.910.710.1
Number of previous treatments
  Mean (SD)2.4 (1.3)2.0 (1.1)2.6 (1.4)<0.001
  1, n (%)48 (28.7)239 (40.0)127 (21.6)
  2, n (%)48 (28.7)190 (31.8)182 (30.9)
  3, n (%)40 (24.0)112 (18.7)156 (26.5)
  4, n (%)17 (10.2)42 (7.0)71 (12.1)
  ≥5, n (%)14 (8.4)15 (2.5)53 (9.0)
Previous treatments, n (%)
  NSAID109 (65.3)388 (64.9)479 (81.3)<0.001
  Paracetamol73 (43.7)232 (38.8)273 (46.3) 0.030
  Opiods74 (44.3)159 (26.6)250 (42.4)<0.001
  AED46 (27.5)133 (22.2)111 (18.8) 0.043
  Tricyclic drugs18 (10.8)61 (10.2)69 (11.7) 0.704
  Other13 (7.8)36 (6.0)81 (13.8)<0.001
  • Total number of analyzed patients; some patients did not report all data.

  • Patients might be receiving more than one previous treatment.

  • NSAID = non-steroidal anti-inflammatory drug; SD = standard deviation; AED = anti-epileptic drug; PGB = pregabalin.

Regarding the PRO assessments, subjects not receiving PGB, in general, showed better baseline clinical conditions, with lower levels of pain intensity and better social activity, working activity, and quality of life (Table 2). Likewise, when compared with patients who received PGB, those patients also showed significantly lower scoring in mood symptoms related with pain, but no significant differences were found in anxiety severity or sleep disturbances (Table 2).

View this table:
Table 2

Health patient-reported outcomes at baseline visit

Health outcomesOther treatments
(N = 167)
PGB monotherapy
(N = 598)
PGB add-on
(N = 589)
DN4 Questionnaire, mean (SD)6.4 (1.7)6.8 (1.8)6.8 (1.7) 0.033
SF-MPQ, mean (SD)
  Sensorial14.0 (5.5)16.0 (6.1)15.8 (5.9) 0.001
  Affective4.4 (3.4)5.0 (3.2)5.2 (3.3) 0.017
  Total18.5 (8.0)21.0 (8.4)21.1 (8.4) 0.002
  PPI (0–5)2.4 (0.9)2.7 (0.8)2.7 (0.9)<0.001
  VAS (0–100)66.8 (17.6)71.4 (15.2)72.6 (15.7)<0.001
Sleep-MOS, mean (SD)
  Summary index (0–100)47.1 (17.2)48.0 (19.0)50.1 (17.3) 0.063
  Sleep disturbance (0–100)49.6 (19)50.9 (21.4)53.8 (20) 0.018
  Snoring (0–100)43 (30.3)37.9 (29.3)38.1 (27.9) 0.140
  Shortness of breath (0–100)32.3 (29)32.3 (27.7)31.3 (25.6) 0.785
  Sleep quantity, hours5.8 (1.2)5.7 (1.5)5.6 (1.4) 0.191
  Adequacy of sleep (100–0)43.9 (22.5)41.4 (25.1)36.8 (22.5) 0.001
  Daytime somnolence (0–100)42 (20.2)41.1 (20.7)40.9 (18.7) 0.840
HADS, mean (SD)
  Depression (0–21)9.9 (4.9)10.3 (4.3)10.9 (4.5) 0.010
  Anxiety (0–21)10.2 (3.7)10.5 (4.1)10.9 (4.1) 0.087
SDI, mean (SD)
  Disability (0–30)16.5 (6.3)18.0 (6.3)19.4 (5.6)<0.001
  Perceived stress (0–10)5.5 (2.1)5.9 (2.1)6.3 (2.0)<0.001
  Perceived social support (0–100)56.5 (24.8)55.6 (23.2)58.7 (23.5)<0.001
  VAS, mean (SD)49.6 (18.1)42.2 (18.4)40.5 (18.0)<0.001
  Movement disturbances,§n (%)89 (58.2)218 (38.4)221 (39.3)<0.001
  Self-care problems,§n (%)96 (62.7)296 (52.3)264 (47) 0.002
  Problems of day life activities,§n (%)43 (28.3)103 (18.2)81 (14.4) 0.001
  Pain or discomfort of any type,§n (%)7 (4.6)14 (2.5)8 (1.4) 0.059
  Presence of anxiety/depression,§n (%)40 (26.1)143 (25.3)101 (18)0.006
  • Total number of analyzed patients; some patients did not report all data.

  • Sum of three disability item scores.

  • § Patients reporting “without problems.”

  • SD = standard deviation; VAS = visual analog scale; HADS = Hospital Anxiety Depression Scale; MOS = Medical Outcomes Study; PGB = pregabalin; PPI = present pain intensity; SDI = Sheehan Disability Inventory; SF-MPQ = Short Form McGill Pain Questionnaire.

Drug Treatment During the Study

Most patients from the other-treatment group (67%) received two or more drugs for NeP treatment [mean (SD): 2.2 (1.2)], the most frequent being paracetamol (44%, mean dose: 2,144 ± 1,010 mg/day), followed by gabapentin (33%, mean dose: 1,288 ± 543 mg/day), tramadol (29%, mean dose: 214 ± 130 mg/day), ibuprofen (19%, mean dose 1,438 ± 517 mg/day), and metamizol (17%, mean dose: 1,679 ± 606 mg/day). In the group receiving PGB as monotherapy, the mean dose was 208 ± 123 mg/day. Drugs used most frequently in the PGB add-on group (mean dose of PGB: 200 ± 113 mg/day) were paracetamol (40%, mean dose: 1,866 ± 999 mg/day), tramadol (20%, mean dose: 200 ± 111 mg/day), metamizol (19%, mean dose: 1,428 ± 641 mg/day), ibuprofen (15%, mean dose 1,148 ± 502 mg/day), diclofenac (9.5%, mean dose 110.6 ± 45.3 mg/day), amitriptyline (6.1%, mean dose 45.5 ± 29.1 mg/day), gabapentin (4.1%, mean dose 1,022.7 ± 630.1 mg/day), and ketorolac (3.2%, mean dose 25.0 ± 8.5 mg/day). In this latter group, the mean number of drugs used during the study was 2.7 (1.0).

Patient-reported Health Outcomes

Pain Reduction

Significant and clinically relevant reductions (effect size ≥0.8), both in pain intensity (previous and current) and in pain symptom scores (Table 3), were observed after a 12-week follow-up in the three analyzed groups. However, the reduction in pain intensity according to the SF-MPQ VAS was significantly greater in the PGB groups compared with that in the group not receiving PGB, with mean changes of 54% and 51% for the PGB monotherapy and add-on groups, respectively, compared with 34% in the group not receiving PGB (P < 0.0001, Table 3). Differences in pain intensity reduction between the two PGB-treated groups and the other-treatment group were significant from treatment initiation (at 5 weeks in the PGB monotherapy group and at 6 weeks in the PGB add-on group) and decreased even more throughout the study (Figure 2). At the end of the study, 57.9% and 52.1% of patients who received monotherapy with PGB and PGB add-on, respectively, showed a pain intensity reduction of 50%, compared with 30.2% in the other-treatment group (P < 0.0001, Table 3), resulting in a higher cumulative number of days with no or mild pain (<40 mm in SF-MPQ VAS) in the PGB groups compared with that in the group not receiving PGB (P < 0.0001, Table 3).

View this table:
Table 3

Mean change and magnitude (size effect) of different pain parameters measured by the SF-MPQ Questionnaire according to treatment groups

Health outcomeOther treatments
(N = 157)
PGB monotherapy
(N = 577)
PGB add-on
(N = 575)
  Sensorial−4.6 (5.4) [0.85]−8.8 (5.9) [1.44]−8.0 (5.8) [1.36]<0.001
  Affective−1.8 (2.8) [0.53]−3.3 (2.9) [1.03]−3.2 (2.8) [0.97]<0.001
  Total−6.4 (7.4) [0.80]−12.1 (8.1) [1.44]−11.2 (7.8) [1.33]<0.001
  PPI−0.9 (1.1) [1.00]−1.5 (1.0) [1.88]−1.4 (1.0) [1.56]<0.001
SF-MPQ intensity of pain (VAS)
  Mean change−22.8 (26.4) [1.27]−38.9 (22.9) [2.56]−36.8 (22.4) [2.34]<0.001
  Mean change in percentage (SD)34.2 (29.1)53.7 (27.7)50.6 (24.8)<0.001
Responders (%)§30.257.952.1<0.001
Cumulative number of days with no pain or mild pain (VAS <40mm), mean (SD)25.5 (29.4)35.0 (29.8)29.8 (28.9)<0.001
  • Total number of analyzed patients; some patients did not report all data.

  • Values expressed as means (standard deviation) [size effect].

  • § Responders: patients with at least 50% reduction in the pain intensity baseline value assessed by SF-MPQ VAS.

  • VAS = visual analog scale; PGB = pregabalin; PPI = present pain intensity; SF-MPQ = Short Form McGill Pain Questionnaire.

Figure 2

Mean weekly change in pain severity according to SF-MPQ VAS. VAS = visual analog scale; PGB = pregabalin; W1–W12 = weeks 1–12; SF-MPQ = Short Form McGill Pain Questionnaire. *P < 0.05 vs other treatments; P < 0.01 vs other treatments; P < 0.001 vs other treatments. Pain was measured with a VAS, which assesses the pain intensity of patients during the previous week, from 0 (no pain) to 100 (worst possible pain).

Other Health Patient-reported Outcomes

Regarding co-morbid symptoms of sleep, depression, and anxiety, the study results show significant changes from baseline scores in all groups. However, treatment effects in PGB-treated patients were significantly higher than those in patients who did not receive the drug (Table 4). In addition, while the PGB monotherapy and add-on groups showed large effect sizes of the mean change in the three symptoms, in the other-treatment group, the change was only of moderate magnitude for sleep and anxiety and small for depressive symptoms (Table 4). In the six dimensions of the MOS-Sleep, the effect on sleep followed a similar pattern, and clinically significant changes (effect size >0.8) in the “sleep disturbance” and “adequacy of sleep” dimensions were only observed in the groups exposed to PGB (Table 5).

View this table:
Table 4

Mean change and magnitude (size effect) of patient-reported outcomes at the end of study visit according to treatment groups

Health outcomeOther treatments
(N = 157)
PGB monotherapy
(N = 577)
PGB add-on
(N = 575)
  Depression (0–21)−1.7 (3.5) [0.43]−4.2 (4.1) [0.98]−4.0 (4.3) [0.89]<0.001
  Anxiety (0–21)−2.1 (2.9) [0.57]−4.3 (4.1) [1.05]−3.7 (3.7) [0.90]<0.001
  Disability (0–30)−5.0 (5.5) [0.79]−8.8 (6.4) [1.40]−8.8 (6.1) [1.57]<0.001
  Perceived stress (0–100)−1.7 (2.0) [0.81]−2.9 (2.2) [1.38]−2.9 (2.2) [1.32]<0.001
  Perceived social support, % (0–100)−1.1 (21.4) [0.04]0.5 (21.4) [0.02]−0.4 (21.3) [0.02]NS
  VAS (0–100)*13.1 (18.7) [0.72]27.9 (22.4) [1.52]27.8 (21.7) [1.54]<0.001
  Movement,§ N (%)100 (65.4)385 (67.8)369 (65.7)0.040
  Self-care,§ N (%)113 (73.9)446 (78.8)427 (76.0)0.017
  Day life activities,§ N (%)74 (48.7)341 (60.2)270 (48.1)<0.001
  No pain, N (%)25 (16.4)204 (36.0)148 (26.3)<0.001
  No anxiety/depression, N (%)74 (48.4)377 (66.6)312 (55.7)<0.001
  • * Values expressed as means (standard deviation) [size effect].

  • Analyzed patients; some patients did not report all data.

  • Sum of scores of the three disability items.

  • § Patients reporting “without problems.”

  • VAS = visual analog scale; HADS = Hospital Anxiety Depression Scale; MOS = Medical Outcomes Study; NS = non-significant; PGB = pregabalin; SDI = Sheehan Disability Inventory; QALY = quality-adjusted life year.

View this table:
Table 5

Mean change and magnitude (size effect) of various dimensions of sleep as assessed by the MOS Sleep Questionnaire according to treatment groups

MOS-Sleep dimensionOther treatments
(N = 157)
PGB monotherapy
(N = 577)
PGB add-on
(N = 575)
Summary index (0–100)−9.6 (15.2) [0.56]−19.1 (19.9) [1.00]−17.7 (18.7) [1.02]<0.001
Sleep disturbance (items 1, 3, 7, and 8) (rating 0–100)−10.1 (17.7) [0.53]−21.5 (21.4) [1.00]−20.3 (21.1) [1.01]<0.001
Snoring (item 10) (rating 0–100)−2.6 (18.2) [0.09]−7.9 (24.1) [0.27]−3.5 (22.8) [0.13]<0.001
Shortness of breath (item 5) (rating 0–100)−6.4 (25.2) [0.22]−14.4 (26.0) [0.52]−13.0 (25.1) [0.51]<0.001
Adequacy of sleep (items 4 and 12) (rating 100–0)10.9 (19.8) [0.48]20.4 (27.1) [0.81]20.3 (25.2) [0.90]<0.001
Amount of sleep (item 2), hours0,6 (1,1) [0.50]1,1 (1,4) [0.73]1,0 (1,3) [0.71]<0.001
Daytime somnolence (items 6, 9, and 11) (rating 0–100)−6.8 (19.0) [0.34]−13.7 (21.0) [0.66]−9.6 (19.1) [0.51]<0.001
  • Analyzed patients; some patients did not report all data.

  • Values expressed as means (standard deviation) [size effect].

  • MOS = Medical Outcomes Study; NS = non-significant; PGB = pregabalin.

In all the treatment groups, significant improvements were observed in the impact on NeP on social and working activities (disability) and perceived stress measured by the SDI (Table 4). These changes were significantly greater in the two groups of patients receiving PGB compared with those in the group of patients not receiving the drug, with large effect sizes. However, no significant differences between the groups were observed in perceived social support, with changes at end of study of limited effect sizes. The improvement in patients' general health status according to the EQ-5D questionnaire was significantly greater in the PGB-treated groups, and the clinical impact was greater as well: while the effect size was great in both PGB-treated groups (1.5 in each group), it was only moderate in the other-treatment group (0.7, Table 4). In the PGB monotherapy group, these differences were significant from study week 4 and progressively increased up to week 12 (Figure 3). In general, the benefits observed in the five dimensions of the EQ-5D were significantly greater in the PGB-treated groups, especially in the PGB monotherapy group (Table 4).

Figure 3

Mean weekly change in health status according to the EQ-5D VAS. VAS = visual analog scale; PGB = pregabalin; W1–W12 = weeks 1–12; *P < 0.05 vs other treatments; P < 0.01 vs other treatments; P < 0.001 vs other treatments. Health status was measured with a VAS (EQ-5D) ranging from 0 = the worst imaginable health status to 100 = the best imaginable health status.


Data presented in this work are likely the first evaluating the effectiveness of PGB for the treatment of NeP under usual clinical practice conditions in the PCS or “real world,” and thus different from existing clinical trial data. However, our study presents some limitations that should be kept in mind. Among them, its observational nature implies less control of confounding factors, in particular, due to the lack of randomization. Among these factors, there is the so-called confounding by indication bias, inherent to observational studies with drugs and explained by the fact that, in clinical practice, there is always a reason for a prescription, and that reason is often associated with the outcome of interest [38]. This would explain, for instance, the significant differences in baseline clinical characteristics of the three groups selected for this analysis, in intensity and descriptors of pain, and in associated co-morbid symptoms and level of disability as well. However, these differences would be expected to bias the results against PGB since, in general, patients from PGB monotherapy and PGB add-on groups showed significantly more severe symptoms. In the PGB monotherapy group, this could be due to the decreased use of drugs observed in the previous therapeutic schedule or to a decreased use of opiates before the study, compared with higher previous exposures to NSAIDs and paracetamol, and lower exposures to anti-epileptic drugs in the PGB add-on group. Interestingly, and according with recent guidelines and recommendations [39], it appears that a substantial proportion of subjects in the study were receiving inappropriate drug regimens for a neuropathic condition before starting the trial or even sub-therapeutic doses of some drugs, such as the average dosage of gabapentin (about 1,200 mg/day), which is clearly in the lower range of its therapeutic range. However, these findings are difficult to interpret, although one may speculate some possible reasons, among them failure in continuous medical education at the primary care level, scarcity of resources in these settings to devote enough time of care for patients, saturated waiting lists, and need for redundant physician visits to titrate drugs adequately. Whether these findings apply to our health context only or are extensible to the primary care practice in general is a limitation of this research that should be borne in mind for possible additional research on this matter.

Overall, in spite of these limitations, the results of this analysis as assessed by the examination of the effect size of changes in different PRO measurements suggest that PGB is effective in treating peripheral NeP due to diabetic neuropathy, post-herpetic neuralgia, or trigeminal neuralgia, this effectiveness also affecting their associated symptoms—sleep disturbances, anxiety, and depressive symptoms—in the usual medical practice. This improvement extends to less of an impact on the degree of disability for social and working activities and to a substantial improvement in the health-related quality of life as assessed by the EQ-5D questionnaire. PGB, both in monotherapy and as an add-on therapy, produced a marked reduction in pain (over 50%) with a large effect size. The observed percentages of responders (i.e., patients with a baseline pain severity reduction of at least 50%), 58%, and 52% for the PGB monotherapy and add-on groups, respectively, were very similar to those observed in clinical trials of PGB. Hence, patients with diabetic neuropathy showed values between 39% and 48% [40–42], with post-herpetic neuralgia from 28% to 50% [43–45]. In a sample of patients with such diagnoses, of which the 48% used a flexible-dosing regimen for PGB, the observed figures for responder rates were of similar magnitude [46]. These results are also consistent with those recently reported in patients with diabetic peripheral neuropathy or trigeminal neuralgia treated with PGB for 4 weeks under routine clinical practice [26].

Within the context of a non-randomized study, the apparent superiority of the PGB-treated groups vs the group not treated with this drug is difficult to interpret. However, differences were very marked, with effect sizes of almost double in the PGB-treated groups compared with those observed in the other-treatment group for pain and many other associated symptoms, including disability and quality of life. These differences were very consistent, particularly when taking into account the greater severity of symptoms in the PGB groups. In this sense, the effect of this drug on associated symptoms, including sleep disturbances, depression, and anxiety, should be noticed. Only a moderate or even small improvement in these symptoms was observed in the other-treatment group, while the effect size of the improvement in these three symptoms with PGB was much greater. Due to the frequent coexistence of these three symptoms in patients with NeP and the interrelations between them and pain [4], the clinically relevant effect of PGB on both pain and associated co-morbidities makes this drug particularly useful for the management of these patients under usual clinical practice conditions, in particular, when compared with other analgesic drugs. While opiates have shown their efficacy in NeP reduction [18], their benefits on mood, disability, and quality of life are not consistent [47,48]. Possibly, these benefits on co-morbidities may have also contributed to the differences we observed between PGB-treated groups and the group not receiving PGB in the reduction in the impact on the level of disability for social and work activities and in the improvement in the quality of life.

The use of combinations of drugs to treat NeP is very frequent in clinical practice [10,49], and it has been recommended by various authors for cases not responding to treatments in monotherapy [50]. The results of our analysis suggest that patients receiving PGB add-on treatment showed more severe symptoms, with greater disability levels, poorer health status, and previously higher exposures to NSAIDs/paracetamol and lower to anti-epileptic drugs, than patients who had not received PGB and, to some extent, than patients who received monotherapy with PGB. Data from clinical trials supporting the use of combined treatments for NeP are available; however, these are limited to the combination of gabapentin with morphine [51], venlafaxine [52], or nortriptyline [53]. In contrast, the combination of amitriptyline with fluphenazine [54] or of nortriptyline with morphine [55] did not show superiority over either treatment alone, although the recent trial from Gilron and co-workers showed that the combination of gabapentin and nortriptyline was more efficacious than either of those drugs alone [53]. Our study provides comparative data on the effectiveness of the addition of PGB to existing therapy under usual care conditions, although with comparable results observed for monotherapy with PBG, in spite of the observation of a moderate reduction in paracetamol use (from 46% to 40%) and a substantial reduction in NSAID prescription (from 81% to 28%) and tricyclic antidepressant drug utilization (from 12% to 6%).

In our study, mean doses of PGB in both groups, monotherapy and add-on, were in the lower limit of the recommended range for effective dosage (150–600 mg/day), with no significant differences in either group. In fact, the mean dose of PGB in our study (≈200 mg/day) is much lower than that reported in another naturalistic study run with PGB in patients with diabetic peripheral neuropathy or post-herpetic neuralgia (≈300 mg/day) [26]. A possible explanation for this discrepancy is that in our study, the dose schedule was left at the discretion of the treating physician, while it appears that in the previous study, the whole treatment plan (i.e., PGB dosing schedule, concomitant medications, etc.) was forced by protocol. The variability in the mean dose of PGB observed in both groups reinforces the general idea of the absence, in clinical practice, of a single drug or a single effective dose suitable for all patients with NeP [56]; this point is also reinforced by the observation of the same mean dose variability of the treatments used in the group not receiving PGB.

Another study limitation is the unavailability of a systematic assessment of treatment tolerability due to the lack of evaluation of the effect of a pharmacological intervention in the original study. However, the high percentages of patients who completed treatment in all the analyzed groups and the very low frequency of withdrawals due to adverse effects suggest that the three therapeutic strategies were well tolerated.

To conclude, our analysis suggests that under usual clinical practice conditions, patients with chronic refractory pain receiving PGB in monotherapy, and in combination as well, show substantial reduction in pain and in the co-morbid spectrum of associated symptoms, such as sleep disturbances, mood disorders, and associated disability, which translates into improvement in the health-related quality of life of patients with NeP of painful diabetic neuropathy, post-herpetic neuralgia, and trigeminal neuralgia origin. The possible superiority of this drug over other therapeutic options and the potential benefit of adding PGB to existing therapy should be further evaluated in randomized clinical trials.


This study was funded by an unrestricted grant from Pfizer Spain. The authors thank Fernando Rico-Villademoros, MD, for his contribution in the preparation of the manuscript.


View Abstract