The Epidemiology of Chronic Pain of Predominantly Neuropathic Origin. Results From a General Population Survey

The Journal of Pain, Vol 7, No 4 (April), 2006: pp 281-289 Available online at www.sciencedirect.com

The Epidemiology of Chronic Pain of Predominantly Neuropathic Origin. Results From a General Population Survey Nicola Torrance,* Blair H. Smith,* Michael I. Bennett,† and Amanda J. Lee* *Department of General Practice and Primary Care, University of Aberdeen, Scotland. † Clinical Teaching and Research Unit, St. Gemma’s Hospice and University of Leeds, England.

Abstract: Progress in the understanding of chronic pain with neuropathic features has been hindered by a lack of epidemiologic research in the general population. The Leeds Assessment of Neuropathic Symptoms and Signs score (S-LANSS) was recently validated for use in postal surveys, making the identification of pain of predominantly neuropathic origin possible. Six family practices in 3 UK cities (Aberdeen, Leeds, and London) generated a total random sample of 6,000 adults. The mailed questionnaire included demographic items, chronic pain identification, and intensity questions, the S-LANSS, the Level of Expressed Needs questionnaire, and the Neuropathic Pain Scale. With a corrected response rate of 52%, the prevalence of any chronic pain was 48% and the prevalence of pain of predominantly neuropathic origin was 8%. Respondents with this chronic neuropathic pain were significantly more likely to be female, slightly older, no longer married, living in council rented accommodation, unable to work, have no educational qualifications, and be smokers than all other respondents. Multiple logistic regression modeling found that pain of predominantly neuropathic origin was independently associated with older age, gender, employment (being unable to work), and lower educational attainment. Respondents with this pain type also reported significantly greater pain intensity, higher scores on the NPS, higher levels of expressed need, and longer duration of pain. This is the first estimate of the prevalence and distribution of pain of predominantly neuropathic origin in the general population, using a previously validated and reliable data collection instrument. Perspective: Chronic pain with neuropathic features appears to be more common in the general population than previously suggested. This type of pain is more severe than other chronic pain but distributed similarly throughout sociodemographic groups. © 2006 by the American Pain Society Key words: Neuropathic pain, S-LANSS, chronic pain, questionnaire, epidemiology, prevalence.

B

owsher8 estimated that around 1% of the population experience chronic neuropathic pain, but this figure is based on estimates rather than direct evidence. Around 27% of patients attending a pain clinic have neuropathic pain, and among cancer patients approximately 36% have pain that involves neuropathic mechanisms.11,17 Although previous studies have examined the epidemiology of specific neuropathic pain conReceived July 5, 2005; Revised November 17, 2005; Accepted November 19, 2005. Supported by an unrestricted educational grant from Pfizer. Address reprint requests to Dr. Nicola Torrance, Department of General Practice and Primary Care, University of Aberdeen, Foresterhill Health Centre, Westburn Road, Aberdeen AB25 2AY, Scotland, UK. E-mail: [email protected] 1526-5900/$32.00 © 2006 by the American Pain Society doi:10.1016/j.jpain.2005.11.008

ditions (such as painful diabetic neuropathy and postherpetic neuralgia),25 there is no published study that has examined the overall prevalence of chronic pain with neuropathic characteristics in a general population. Determining the cause of a pain is essential to good management. In practice this usually requires the clinician to identify whether neuropathic pain-generating mechanisms exist in any given patient. Evidence suggests that neuropathic pain is not only more severe than nociceptive pain,9 but also less responsive to conventional analgesic techniques.6,10,20 Because of this, coanalgesics are used to treat neuropathic pain either alongside or instead of standard analgesia. A recent placebo-controlled crossover trial in 50 patients with postherpetic neuralgia found that opioids and tricyclic antidepressants produced similar levels of pain relief. However, pain reduction from opioids did not correlate with that 281

282

Epidemiology of Chronic Neuropathic Pain 23

from tricyclic antidepressants. This suggests that opioids and tricyclics act via independent mechanisms and that in neuropathic pain optimal analgesia might result from the use of both, supporting the need to identify neuropathic pain mechanisms in patients with chronic pain. Progress in the further understanding of pain with neuropathic mechanisms in the general population is hindered by a lack of epidemiologic research. This information is vital in order to understand the distribution, etiology, and natural history of this type of chronic pain, to target interventions, to evaluate their outcomes, and to design randomized controlled trials with appropriate sample size calculations. One reason for the lack of population-based epidemiologic data has been the lack of a good case identification instrument for use in surveys. Existing studies have often relied on diagnostic categories, such as diabetic neuropathy or postherpetic neuralgia,24,31 rather than using direct evidence of the patients’ symptoms and signs on which a diagnosis of pain with neuropathic features is made. Although both the Neuropathic Pain Scale (NPS)15 and the Neuropathic Pain Symptom Inventory7 can measure the intensity and impact of neuropathic pain, neither is able to distinguish pain of predominantly neuropathic origin from other pain types. The Leeds Assessment of Neuropathic Symptoms and Signs scale (LANSS) was developed as a clinicbased instrument for identifying patients whose pain is dominated by neuropathic mechanisms.2 We recently developed a self-complete version of this, the S-LANSS score, and found this to be valid and reliable at identifying chronic pain of predominantly neuropathic origin (POPNO) on the basis of the patient’s current symptoms and signs in comparison with expert clinical judgement.3 The S-LANSS score was tested in both a clinic setting and in a postal survey. This now presents the opportunity to describe the prevalence and distribution of POPNO in a general population survey. We report the results of such a survey in the United Kingdom which included an assessment of the associated sociodemographic characteristics of patients with this type of pain.

Material and Methods

ample in the event of recent death or terminal illness). Each excluded patient was replaced by another, sampled randomly from the practice register. The study sample therefore comprised 6,000 individuals.

Case Identification Individuals with chronic pain were identified by affirmative answers to 2 questions: 1) Are you currently troubled by pain or discomfort, either all the time or on and off? and 2) Have you had this pain or discomfort for more than 3 months?18 Identical case identification questions have been used in previous population-based research on chronic pain.13,14 The main focus of the questionnaire was the S-LANSS.3 This is a 7-item questionnaire, including 5 questions about pain characteristics and 2 self-examination items, with responses weighted to provide a maximum score of 24. A score of 12 or more has been found in a pain clinic population to have a positive predictive value for neuropathic pain of 76% (95% confidence interval (CI) 66.8% to 84.2%) when compared with classification by a pain specialist. Its validity and reliability in identifying POPNO with this cut-off score in postal research has also been confirmed.3 For the purposes of this study, individuals with an S-LANSS score of ⱖ12 were therefore identified as cases and described as having POPNO.

Site and Cause of Chronic Pain Respondents were asked to indicate the site(s) of any chronic pain and the most troublesome site by selecting from a list which included back; neck or shoulder; head, face, or teeth; stomach or abdomen; arms or hands; chest; hips; and legs or feet.30 Subsequent questions about pain, including the S-LANSS questions, related to the single most troublesome site of pain identified from this list. They were also asked to indicate whether any of the following common causes of pain had been diagnosed: a surgical operation more than 3 months previously, back problems (such as a slipped disc, back surgery, or sciatica), diabetes, an accident that damaged a nerve, arthritis, amputation of a limb, leg ulcers, shingles, or cancer.

Study Sample

Sociodemographic Status

This study was conducted in 3 locations chosen to represent a cross-section of the UK population: Grampian (population approximately 500,000, semirural, mainly caucasian), Leeds (population approximately 800,000, urban, multiethnic) and London (population approximately 8 million, metropolitan, multiethnic). Two general practices in each location were approached and agreed to participate in the study. This provided a wide range of practice population characteristics within each location. From each list of patients registered with the practices, a random sample of 1,000 patients over 18 years of age was identified. The sample list was screened by the general practitioners (GPs) in each practice to exclude patients in whom inquiry might be insensitive or inappropriate (for ex-

Questions were included on age, gender, smoking, marital and employment status, educational attainment, and home ownership (as a proxy for social class28). These were similar to questions used in a previous study of chronic pain in the community.13

Severity of Pain The severity of chronic pain was assessed on an 11point numerical rating scale, with 0 ⫽ “no pain” and 10 ⫽ “pain as severe as it could be.” The severity of neuropathic characteristics of pain was measured using the NPS,15 an 11-item numerical rating scale that addresses specific features of the experience of the pain reported. The NPS composite score has been validated within groups of patients with neuropathic pain and is sensitive

ORIGINAL REPORT/Torrance et al

283 Chronic Pain Case Identification (a) Are you currently troubled by pain or discomfort, either all the time or on and off? (b) Have you had this pain or discomfort for more than three months?

Help-Seeking Behavior questions 1. 2. 3. 4.

Have you sought treatment for your pain recently? Have you sought treatment for your pain often? Have you taken painkillers for your pain recently? Have you taken painkillers for your pain often?

Level 0

Answered no to all 4 help-seeking behavior questions

Level 1

Answered yes to any 1 help-seeking behavior question

Level 2

Answered yes to any 2 help-seeking behavior questions

Level 3

Answered yes to any 3 help-seeking behavior questions

Level 4

Answered yes to all 4 help-seeking behavior questions

Figure 1. Level of Expressed Need.

to treatment changes.16 The significance of chronic pain to individual respondents was assessed using the Level of Expressed Needs (LEN) questionnaire,26 which assigns a level from 0 to 4 depending on the number of positive responses to questions about the seeking of treatment and the use of painkillers, recently and frequently (Fig 1). Level 0 indicates that no treatment or painkillers were used, and level 4 indicates that treatment was sought and painkillers were taken recently and frequently. The LEN has been found to correlate well with disability and the use of health care.26

Analysis Data were entered into a Microsoft Access 2000 database and analyzed using SPSS for Windows (version 13.0). Basic descriptive statistics explored the occurrence of chronic POPNO (ie, S-LANSS score of 12 or more). Chi-square tests were used to test for associations between categorical sociodemographic variables, t tests, or analysis of variance were used to explore the mean difference in normally distributed continuous variables, and the Mann-Whitney U test was used for nonparametric continuous data. To identify the sociodemographic factors that were associated with chronic POPNO, odds ratios (ORs) (unadjusted and adjusted for practice, age, and gender) and 95% CIs were calculated. Forward stepwise logistic regression modeling was then used to obtain a subset of sociodemographic factors that were independently associated with chronic POPNO. All sociodemographic variables were entered into this model. Finally, unadjusted and adjusted ORs were calculated to explore the relationship between chronic POPNO, LEN, and reported pain severity. All reported P values were from

2-sided tests, and a P value of ⬍.05 was used to denote statistical significance. Return of a completed questionnaire was taken as informed consent to participate in the study. The study was approved by the National Health Service Research Ethics Committees for Grampian, Leeds (East), and St Thomas’ Hospital, London.

Results Study Sample Of the 6,000 questionnaires that were mailed, 272 were returned because the address was incorrect or because individuals had moved away or were unable to complete them owing to illness or learning disability. These were subtracted from the denominator prior to the calculation of response rates. A total of 3,002 completed questionnaires were returned, with an overall corrected response rate of 52.4% (Table 1). Corrected response rates varied between practices and ranged from 38.0% to 67.3%. The response rate also varied with age and gender, with nonresponders younger than responders (P ⬍ .001) and women more likely to respond than men across all practices (57.2% corrected response rate vs 47.4%; P ⬍ .001). These differences were similar in each practice. The mean age of respondents was 50.3 years (SD 17.1, range 18-96 years) and 55.6% were female. Forty-five respondents did not complete the 2 case identification questions for chronic pain and were excluded from prevalence calculations. Of the 2,957 other respondents, chronic pain was reported by 1,420 (48.0%, 95% CI 46.2% to 49.8%) with no significant difference found in the proportions in each of the 6 practices (P ⫽ .32).

284 Table 1.

Epidemiology of Chronic Neuropathic Pain

Characteristics of Respondents and Nonrespondents

Practice list size Questionnaires completed Response rate Responders Gender, n (%) Male Female Age yrs, mean (SD) Nonresponders Gender, n (%) Male Female Age, mean (SD) Presence of chronic pain, n (%)

GRAMPIAN A (n ⫽ 970)

GRAMPIAN B (n ⫽ 947)

LEEDS C (n ⫽ 967)

LEEDS D (n ⫽ 968)

LONDON E (n ⫽ 939)

LONDON F (n ⫽ 937)

ALL PRACTICES (n ⫽ 5,728)

P VALUE*

14,942 584 60.2%

7,850 637 67.3%

13,500 444 45.9%

12,101 541 55.9%

10,609 440 46.9%

8,334 356 38.0%

67,336 3,002 52.4%

⬍.001†

244 (41.8) 340 (58.2) 49.3 (16.9) 386

304 (47.7) 333 (52.3) 53.1 (15.7) 310

203 (45.7) 235 (43.4) 205 (46.6) 142 (39.9) 241 (54.3) 306 (56.6) 235 (53.4) 214 (60.1) 51.2 (17.4) 51.5 (18.0) 47.4 (16.7) 47.9 (17.3) 523 427 499 581

1,333 (44.4) 1,669 (55.6) 50.3 (17.1) 2,726

219 (56.7) 167 (43.3) 46.6 (17.8) 291 (50.5)

170 (54.8) 140 (45.2) 45.9 (16.1) 292 (46.3)

260 (49.7) 228 (53.4) 277 (55.5) 325 (55.9) 263 (50.3) 199 (46.6) 222 (44.5) 256 (44.1) 41.4 (17.4) 44.1 (19.5) 41.0 (15.9) 42.4 (15.9) 225 (51.7) 250 (46.9) 201 (46.5) 161 (45.9)

1,479 (54.3) 1,247 (45.7) 43.2 (17.2) 1,420 (48.0)

⬍.001† ⬍.001‡

⬍.001‡ .322†

*Comparing response rates and mean age across the six practices. †␹2 test. ‡Analysis of variance.

Respondents With Chronic POPNO In total, 241 individuals reported chronic pain and had a positive S-LANSS score (ⱖ12, indicating chronic POPNO), representing 8.2% of the study sample (95% CI 7.2% to 9.2%) and 17.0% of those with chronic pain (95% CI 15.0% to 19.0%). There were significant differences in all of the measured sociodemographic characteristics between respondents who reported chronic POPNO and those who did not (Table 2). Individuals with chronic POPNO were more likely to be women, slightly older, no longer married, and living in council rented accommodation than individuals without chronic POPNO. They were also more likely to be unable to work owing to illness or disability, to have no educational qualifications, and to be smokers. Unadjusted and adjusted ORs for chronic POPNO were calculated for each measured sociodemographic factor (Table 3). Adjustment for practice alone made very little difference to the ORs (data not shown), but slight confounding from age and gender was apparent after further adjustment for these factors. Forward stepwise logistic regression identified being female, being aged 40 or over, being unable to work, and having little or no education as factors independently predictive of chronic POPNO (Table 3).

Site of Chronic Pain When asked to identify any areas of the body where they had suffered pain in the previous 3 months, individuals with chronic POPNO were significantly more likely to indicate peripheral areas, ie, arms or hands (66.5% vs 47.6%; P ⬍ .001), legs or feet (86.4% vs 64.6%; P ⬍ .001), and all other sites except the back or the hips (Table 4). Most respondents reported chronic pain at more than one site and many at multiple sites. Individuals who reported pain at 5 or more sites were more likely to have POPNO (␹2 test; P ⫽ .002) (Fig 2). When asked to identify

the single site where pain had “bothered them the most” in the previous 3 months, there was a significant difference found between the sites indicated, with SLANSS-positive individuals much more likely to select the legs or feet (44.4% vs 22.3%) and much less likely to select the back (15.8% vs 27.5%) or the hips (1.8% vs 8.0%). A surgical operation more than 3 months previously was reported by 10.1% of respondents with chronic POPNO (21/208), compared with 3.5% of the other chronic pain sufferers (35/996; P ⬍ .001). Similarly, 10.1% of those with chronic POPNO (20/208) reported an accident that had damaged a nerve, compared with 4.3% of other chronic pain respondents (43/996; P ⬍ .001). No significant differences were noted in the proportions of those with each pain type reporting other possible causes of chronic pain.

Severity of Pain Individuals with chronic POPNO had higher median pain severity scores than other chronic pain sufferers (Fig 3), and were more likely to have a higher LEN score (indicating a greater use of treatment and painkillers), and to have experienced their pain for a longer time (Table 5). The unadjusted OR (95% CI) for reporting a high LEN (3 or 4) rather than a low LEN (0, 1, or 2) was 3.14 (2.36 to 4.18) for those with POPNO compared with other chronic pain sufferers. After adjustment for age, gender, and practice, the OR for a high LEN was 3.00 (2.24 to 4.02), and after further adjustment for pain severity (numerical rating scale), the OR was 2.07 (1.51 to 2.85). Therefore some of the relationship between LEN scores and chronic POPNO is accounted for by pain severity, but an independent relationship remains after adjustment. The composite NPS scores were higher in the chronic POPNO respondents (median score 50, interquartile range (IQR) 35.5 to 61) than others with chronic pain (median score 28, IQR 18 to 39; P ⬍ .001).

ORIGINAL REPORT/Torrance et al

285

Characteristics of Respondents With Chronic Pain of Predominantly Neuropathic Origin (POPNO)*

Table 2.

CHRONIC POPNO (n ⫽ 241)

Gender Men Women Age, mean (SD) Marital Status Never married Living as married No longer married Housing tenure Owned/mortgaged Council rent Private rent/other Employment Employed Retired Unable to work Others Education No qualifications Secondary school certificate Higher education Smoking Smoker Ex-smoker Never smoked

NO CHRONIC POPNO† (n ⫽ 2,716)

n

%

n

%

P VALUE

87 154 52.9 (15.5)

36.1 63.9

1203 1513 49.1 (17.0)

44.3 55.7

40 150 50

16.7 62.5 20.8

514 1807 378

19.0 67.0 14.0

.015‡

148 74 17

61.9 31.0 7.1

2098 396 200

77.9 14.7 7.4

⬍.001‡

109 64 43 24

45.4 26.7 17.9 10.0

1656 624 96 320

61.4 23.1 3.6 11.9

⬍.001‡

79 95 59

33.9 40.8 25.3

528 923 1187

20.0 35.0 45.0

⬍.001‡

68 66 106

28.3 27.5 44.2

580 771 1357

21.4 28.5 50.1

.040‡

.017‡ ⬍.001§

*S-LANSS score ⱖ12. †Includes all other respondents, ie, those with no chronic pain and those with chronic pain that was not POPNO. ‡from ␹2 test. §from t test.

Discussion We have presented the first estimate of the prevalence and distribution of chronic pain with distinct neuropathic features in a general population using direct evidence. This was based on a self-administered questionnaire that has been found to be reliable and valid with good sensitivity, specificity, and positive predictive value for neuropathic pain.3 We found a population prevalence of chronic POPNO of 8.2%, which is considerably higher than previous population estimates of neuropathic pain prevalence.8 This represents approximately 1 in 6 of those with chronic pain. With the data presented it is not possible to be confident that those respondents who were S-LANSS positive were confirmed clinical cases of neuropathic pain and this is why we have chosen to report using the term POPNO. It has been suggested that neuropathic pain is not necessarily a dichotomous phenomenon, but may instead be a spectrum where chronic pain is “more or less neuropathic” in origin.1 In this more flexible model of chronic pain, identification of pain of predominantly neuropathic origin (POPNO) may be more appropriate than a binary neuropathic/nonneuro-

pathic phenomenon. Adding to the complexity of this debate is the lack of a recognized “gold standard” for diagnosing neuropathic pain, and therefore any case definition for epidemiologic research must be arbitrary to a certain extent. Our findings at least indicate that 8% of a general population sample has a positive S-LANSS score, and further confirmatory research is required to identify the proportion of these that would be diagnosed with neuropathic pain by a physician. Dworkin et al12 have estimated that within the United States population of approximately 250 million people, around 3 million have diabetic neuropathy and 1 million have postherpetic neuralgia, representing 1.6% of the population. We found no association between positive S-LANSS score and these 2 diagnostic categories as reported by respondents in this study. However, we did find a significant association between positive S-LANSS score and pain in the area of the chest, abdomen, and limbs, which are common features of these diagnoses. We also found that the reporting of chronic POPNO in our sample was associated with a surgical operation more than 3 months

286

Epidemiology of Chronic Neuropathic Pain

Association of Demographic Factors With Reporting of Chronic Pain of Predominantly Neuropathic Origin (POPNO)*

Table 3.

UNADJUSTED ODDS RATIO (95% CI)

ADJUSTED† ODDS RATIO (95% CI)

ADJUSTED‡ ODDS RATIO (95% CI)

1.0 1.41 (1.07-1.85)

1.0 1.42 (1.08-1.87)

1.0 1.42 (1.06-1.87)

1.0 1.63 (1.16-2.28) 1.62 (1.13-2.33)

1.0 1.84 (1.30-2.59) 1.82 (1.26-2.63)

1.0 1.82 (1.28-2.59) 1.93 (1.33-2.80)

1.0 1.07 (0.74-1.53) 1.70 (1.10-2.63)

1.0 0.97 (0.66-1.43) 1.23 (0.76-1.98)

—

1.0 2.65 (1.97-3.57) 1.21 (0.72-2.03)

1.0 2.42 (1.73-3.37) 1.48 (0.86-2.55)

—

1.0 6.81 (4.52-10.24) 1.56 (1.13-2.15) 1.14 (0.72-1.8)

1.0 5.76 (3.77-8.80) 1.45 (0.86-2.44) 1.17 (0.73-1.86)

1.0 5.04 (3.2-7.88) 1.34 (0.79-2.29) 1.03 (0.63-1.68)

1.0 2.07 (1.48-2.90) 3.01 (2.11-4.28)

1.0 1.89 (1.34-2.68) 2.42 (1.62-3.61)

1.0 1.86 (1.23-2.81) 1.73 (1.21-2.47)

1.0 0.73 (0.51-1.04) 0.67 (0.48-0.92)

1.0 0.72 (0.50-1.05) 0.69 (0.50-0.96)

— —

Gender Male§ Female Age group 18-39§ 40-59 60⫹ Marital status Never married§ Living as married No longer married Housing Owner/mortgaged§ Council rent Private rent/other Employment Employed§ Unable to work Retired Other not employed Education Higher education§ Secondary school certificate No qualifications Smoking Smoker§ Ex-smoker Never smoked *S-LANSS score ⱖ12. †Adjusted for practice, age and gender.

‡Forward stepwise model, with all sociodemographic variables entered. §Reference category.

Table 4.

Site of Chronic Pain Experienced in the Past 3 Months, n (%) PAIN EXPERIENCED

PAIN

SITE

Back Neck or shoulder Headache, facial, or dental Stomach or abdomen Arms or hands Chest pain Hips Legs or feet

IN

ANY SITE*

SINGLE PAIN SITE THAT HAS BOTHERED

THE

MOST†

CHRONIC POPNO‡

OTHER CHRONIC PAIN

P VALUE§

CHRONIC POPNO‡

OTHER CHRONIC PAIN

P VALUE储

145 (70.7%) 131 (69.3%) 122 (68.5%) 89 (51.7%) 129 (66.5%) 51 (31.3%) 83 (45.4%) 190 (86.4%)

719 (69.9%) 603 (62.6%) 556 (59.6%) 370 (41.7%) 443 (47.6%) 181 (21.2%) 387 (42.2%) 656 (64.6%)

0.872 0.095 0.031 0.019 ⬍0.001 0.005 0.473 ⬍0.001

27 (15.8%) 24 (14.0%) 12 (7.0%) 10 (5.8%) 17 (9.9%) 2 (1.2%) 3 (1.8%) 76 (44.4%)

275 (27.5%) 132 (13.2%) 88 (8.8%) 87 (8.7%) 78 (7.8%) 36 (3.6%) 80 (8.0%) 223 (22.3%)

⬍0.001

*Respondents were able to indicate more than one site. †Respondents were asked to indicate which one pain site had bothered them the most in the past three months. ‡Pain of predominantly neuropathic origin (S-LANSS score ⱖ12). §From ␹2 test of individual sites (1 degree of freedom). 储From ␹2 test across all sites (7 degrees of freedom).

ORIGINAL REPORT/Torrance et al

pathic origin with other chronic pain respondents for number of pain sites reported.

before or an accident that damaged a nerve. This suggests that a large and under-recognized burden of chronic POPNO in the general population may be due to surgical or accidental trauma rather than well defined neuropathies or neuralgias. Chronic POPNO was associated with older age, being female, having lower educational status, and being unable to work because of illness or disability. This is a common pattern for all types of chronic pain.5,13,29 Chronic POPNO was reported to be more severe than other chronic pain, as measured on a numerical rating scale and by the seeking of professional treatment and advice (LEN). Davies et al11 reported that 27% of patients attending a pain clinic had neuropathic pain though we found that the proportion of those with chronic POPNO was only 17% of all those reporting chronic pain. This suggests that patients with neuropathic pain may be over-represented in pain clinic populations relative to the population prevalence because of their greater pain severity. Previous research has also found that a diagnosis of neuropathic pain was significantly associated with a judgement by clinicians of poor pain outcome21 regardless of age or gender22 and was rated as more intense by patients too.9 We found that pain severity was also associated with LEN, a finding that has been reported previously for all chronic pain.26 This raised the possibility that pain severity was confounding the relationship between chronic POPNO and LEN, and that the S-LANSS was simply detecting the severe end of the chronic pain spectrum. However, although a confounding effect was confirmed by further exploratory analysis adjusting for pain severity, there remained a significant independent relationship between chronic POPNO and LEN. This suggests that regardless of pain severity, the particular features of chronic POPNO result in greater levels of health care need. The distinct nature of pain with neuropathic features, and the further validity of the S-LANSS in detecting it, was also suggested by the much higher NPS scores among the chronic POPNO group. The NPS was specifically designed to measure characteristics of neuropathic pain,15 and we have previously shown

that patients with clinician-diagnosed neuropathic pain have significantly higher NPS scores than patients with nociceptive pain.3 The corrected response rate of 52% was lower than desired, though comparable with other recent postal survey research.4,19 The response rate ranged among the 6 practice populations from 38% to 67%. It is possible that ethnic or other demographic differences in the practice populations contributed to this variation, though we could not assess this specifically. Despite these rates and this variation, however, the proportion of respondents with chronic pain was strikingly similar in all practices, and almost identical to the prevalence reported in a recent survey in the Grampian area, which had a corrected response rate of 82%.13 This suggests that our sample is at least representative of those with chronic pain, if not evenly represented by ages and genders. Only minor differences to the magnitude of odds ratios were seen when adjustment for practice was made. This suggests that despite differences in response rates between practices, the sociodemographic associations that we found were consistent across the 3 locations. Other research has suggested that individuals with chronic pain may be more likely than those without to respond to a survey about chronic pain.27 It is possible therefore that the prevalence of chronic POPNO was overestimated by the relatively low response rate in this study. In summary, however, we have found that chronic pain with neuropathic features is probably more prevalent in the population than was previously thought and is generally more severe than other chronic pain, though it tends to affect the same population groups. In order to provide more information about the performance of the S-LANSS in accurately estimating the prevalence of chronic POPNO in a general population sample, future research may look to performing a standardized clinical examination, by trained clinicians, on S-LANSS-positive cases and a random sample

Chronic pain category

Figure 2. Comparison of chronic pain of predominantly neuro-

287

Chronic POPNO

Other chronic pain

0

2

4

6

8

10

Pain severity scale

Figure 3. Comparison of chronic pain of predominantly neuropathic origin with other chronic pain respondents for pain severity (boxes show IQR and median scores).

288

Epidemiology of Chronic Neuropathic Pain

Comparison of Chronic Pain of Predominantly Neuropathic Origin (POPNO)* With Other Chronic Pain: Pain Severity, Duration of Pain, and Level of Expressed Need (LEN)

Table 5.

Pain severity, median (IQR) LEN, n (%) Level 0 Level 1 Level 2 Level 3 Level 4 Duration of pain, n (%) Less than 6 months 6 months to 1 year 1 to 3 years 3 to 5 years 5 to 10 years More than 10 years

CHRONIC POPNO* (n ⫽ 237)

OTHER CHRONIC PAIN (n ⫽ 1,138)

7.0 (5.0-8.0)

5.0 (3.0-7.0)

⬍.001†

28 (12.0%) 26 (11.1%) 37 (15.8%) 44 (18.8%) 99 (42.3%)

207 (18.2%) 228 (20.0%) 266 (23.4%) 173 (15.2%) 264 (23.2%)

⬍.001‡

9 (3.6%) 26 (10.8%) 65 (27.0%) 43 (17.8%) 46 (19.1%) 52 (21.6%)

81 (6.9%) 162 (13.9%) 321 (27.5%) 193 (16.5%) 192 (16.5%) 218 (18.7%)

.021‡

P VALUE

*S-LANSS score ⱖ12. †From Mann-Whitney U test. ‡From ␹2 test for trend.

of negatives in the general population. In addition, further research is required to measure the relative impact of chronic pain with neuropathic features in terms of health and disability, and to compare the characteristics and experience of this pain type with other chronic pain. The results of this research, in combination with the findings reported here, will inform the targeting of resources and interventions to manage the problem and the evaluation of these.

Acknowledgments We thank the staff and patients of Cults Medical Group and Peterhead Health Centre, Aberdeen; The Street Lane Practice and the Meanwood Group Practice, Leeds; and Crown Dale Medical Group and Lambeth Walk Group, London. Our thanks also go to Hazel Riley and Paul Fearn for their assistance with questionnaire mailing and data entry.

References

8. Bowsher D: Neurogenic pain syndromes and their management. Br Med Bull 47: 644-666, 1991

1. Attal N, Bouhassira D.: Can pain be more or less neuropathic? Pain 110:510-511, 2004

9. Caraceni A, Portenoy RK: An international survey of cancer pain characteristics and syndromes. IASP Task Force on Cancer Pain. Pain 82:263-274, 1999

2. Bennett M: The LANSS Pain Scale: the Leeds assessment of neuropathic symptoms and signs. Pain 92:147-157, 2001 3. Bennett MI, Smith BH, Torrance N, Potter J: The S-LANSS score for identifying pain of predominantly neuropathic origin: validation for use in clinic and postal research. J Pain 6:149-158, 2005 4. Biener L, Garrett CA, Gilpin EA, Roman AM, Currivan DB: Consequences of declining survey respone rates for smoking prevalence estimates. Am J Prev Med 27:254-257, 2004 5. Blyth FM, March LM, Brnabic AJM, Jorm LR, Williamson M, Cousins MJ: Chronic pain in Australia: A prevalence study. Pain 89:127-134, 2001 6. Bonica JJ. Introduction. In: Nashold BS, Ovelman-Levitt J (eds): Deafferentation Pain Syndromes: Pathophysiology and Treatment. New York, Raven Press, 1991, pp 1-19 7. Bouhassira D, Attal N, Fermanian J, Alchaar H, Gautron M, Masquelier E, Rostaing S, Lanteri-Minet M, Collin E, Grisart J, Biureau F: Development and validation of the Neuropathic Pain Symptom Inventory. Pain 108:248-257, 2004

10. Cherny NI, Thaler HT, Friedlander-Klar H, Lapin J, Foley KM, Houde R, Portenoy RK: Opioid responsiveness of cancer pain syndromes caused by neuropathic or nociceptive mechanisms: a combined analysis of controlled, single dose studies. Neurology 44:857-861, 1994 11. Davies HT, Crombie IK, Macrae WA, Rogers KM. Pain clinic patients in northern Britain. Pain Clinic 5:129-135, 1992 12. Dworkin RH, Backonja M, Rowbotham MC, Allen RR, Argoff CR, Bennett GJ: Advances in neuropathic pain. Arch Neurol 60:1524-34, 2003 13. Elliott AM, Smith BH, Penny KI Chambers WA, Smith WC: The epidemiology of chronic pain in the community. Lancet 354:1248-1252, 1999 14. Elliott AM, Smith BH, Hannaford PC, Smith WC, Chambers WA: The course of chronic pain in the community: Results of a four-year follow-up study. Pain 99:299-307, 2002 15. Galer BS, Jensen MP: Development and preliminary validation of a pain measure specific to neuropathic pain. The neuropathic pain scale. Neurology 48:332-338, 1997

ORIGINAL REPORT/Torrance et al 16. Galer BS, Jensen MP, Ma T, Davies PS, Rowbotham MC: The lidocaine patch 5% effectively treats all neuropathic pain qualities: results of a randomized, double-blind, vehicle-controlled, 3 week efficacy study with use of the neuropathic pain scale. Clin J Pain 18:297-310, 2002 17. Grond S, Radbruch L, Meuser T, Sabatowski R, Georg Loick. Lehmann: Assessment and treatment of neuropathic cancer pain following WHO guidelines. Pain 79:15-20, 1999 18. International Association for the Study of Pain: Classification of chronic pain. Pain Suppl 3: S1-S225, 1986 19. Lacey RJ, Lewis M, Sim J: Presentation of pain drawings in questionnaire survey: influence on prevalence of neck and upper limb pain in the community. Pain 105:293-301, 2003. 20. McQuay HJ: Neuropathic pain: Evidence matters. Eur J Pain; 6(Suppl A):11-18, 2002. 21. Mercadante S, Fulfaro F, Casuccio A, Barresi L: Investigation of an opioid response categorization in advanced cancer patients. J Pain Symptom Manage 18:347-352, 1999 22. Mercadante S, Casuccio A, Puma S, Fulfaro F: Factors influencing the opioid response in advanced cancer patients with pain followed at home: The effects of age and gender. Support Care Cancer 8:123-130, 2000 23. Raja SN, Haythornwaite JA, Pappagallo M, Clark MR, Travison TG, Sabeen S, Royall RM, Max MB: Opioids versus antidepressants in postherpetic neuralgia: A randomized, placebo controlled trial. Neurology 59:1015-1021, 2002 24. Rowbotham MC, Harden N, Stacey B, Bernstein P, Mag-

289 nus-Miller L: Gabapentin for the treatment of post-herpetic neuralgia: A randomized controlled trial. J Am Med Assoc 280:1837-1842, 1998 25. Schmader KE. Epidemiology and impact on quality of life of postherpetic neuralgia and painful diabetic neuropathy. Clin J Pain 18:350-354, 2002 26. Smith BH, Penny KI, Elliott AM, Chambers WA, Smith WCS: The Level of Expressed Need—a measure of help-seeking behaviour for chronic pain in the community. Eur J Pain 5:257-266, 2001 27. Smith BH, Hannaford PC, Elliott AM, Smith WCS, Chambers WA: The “number needed to sample” in primary care research. Comparison of two primary care sampling frames for chronic back pain. Fam Pract 22:205-214, 2005 28. Stoate HG: Can health screening damage your health? J R Coll Gen Pract 39:183-195, 1989 29. Verhaak PFM, Kerssens JJ, Dekker J, Sorbi MJ, Bensing JM: Prevalence of chronic benign pain disorder among adults: A review of the literature. Pain 77231-239, 1998 30. Von Korff M: Epidemiologic and survey methods: Chronic pain assessment, in Turk DC, Melzack R (eds): Handbook of Pain Assessment, 2nd ed. New York, NY, Guilford Press, 2001, pp 603-618 31. Watson CP, Moulin D, Watt-Watson J. Gordon A, Eisenhoffer J: Controlled release oxycodone relieves neuropathic pain: A randomized controlled trial in painful diabetic neuropathy. Pain 105:71-78, 2003