Determinants of disability in everyday activities differ in primary and cervicogenic headaches and in low back pain

Authors


Dániel Bereczki, md, pHd, dsci, Department of Neurology, Health Science and Medical Center, University of Debrecen, Debrecen, Nagyerdei krt. 98, H−4012, Hungary. Email: bereczki@jaguar.unideb.hu

Abstract

Abstract  The aim of this study was to test whether the association between disability and depressive symptoms in patients with cervicogenic headache is similar to that found in primary headaches or to the pattern found in low back pain. During a 2-year period, 716 consecutive patients with the clinical diagnosis of cervicogenic headache (n = 182), low back pain (n = 116), migraine (n = 231), tension-type headache (n = 176), and cluster headache (n = 11) filled in the Beck Depression Inventory (BDI). Disability was scored by the migraine disability score questionnaire reflecting the number of days with lost or decreased work, household and social activities. Non-parametric tests and multiple general regression were used for statistical analysis. In multivariate testing, significant independent determinants of disability were pain frequency, pain intensity and the severity of depressive symptoms in migraine and tension-type headache; pain frequency and the BDI score in cervicogenic headache, and pain frequency alone in low back pain. Disability is related to pain frequency in all pain syndromes evaluated in the present study. The level of disability is associated with the severity of pain only in primary headaches, but not in pain syndromes of vertebral origin (cervicogenic headache and low back pain). Disability is associated with the severity of depressive symptoms in all headache types but not in low back pain. Both the location and the etiology of pain have importance in determining the interrelationship between pain characteristics, depression and disability.

INTRODUCTION

Pain is often associated with depression: in a neurological outpatient setting the odds of having depression were increased to over twofold in those with pain.1 Depression in patients with pain is associated with more pain complaints and greater impairment.2 The relationship among pain characteristics, depression and disability can be affected by several factors. For example, disability due to low back pain might be mainly a result of decreased ability to walk, whereas disability due to headache might be mainly due to the effect of pain on mental function. Disability due to low back pain might be socially more accepted than disability due to headache, and this difference in social acceptance, namely the refusal of headache as a real cause of disability by society, might result in a different psychological reaction – possibly a more severe depression – in those with headache than in those with low back pain.

In a previous study, the authors found that although a considerable depression and disability were associated both with primary headaches and with low back pain, disability was the most important predictor of the severity of depressive symptoms in patients with primary headache but not in those with low back pain.3 Therefore, the question arises whether the location of pain or the etiology of pain has more importance in determining the relationship between the severity of disability and depressive symptoms in pain syndromes. Headache attributed to disorders of the neck (cervicogenic headache) can be considered a disease bearing some characteristics of primary headaches (as far as the location of the pain is concerned) as well as of low back pain (as far as etiology is concerned, i.e. a syndrome due to the disorder of the spine and the paravertebral structures). Therefore, the authors used cervicogenic headache as a model to test if location or etiology is a more important factor in determining the association between the severity of disability and depressive symptoms in pain. The authors hypothesized that location of pain has more importance, therefore, the relationship in cervicogenic headache will follow the pattern seen in primary headaches.

PATIENTS AND METHODS

As part of a larger study on headaches,3,4 all consecutive patients with migraine, tension type headache, cluster headache, cervicogenic headache and low back pain, seen by two neurologists in a 2-year period between 1 February 2002 and 31 January 2004 were asked to fill in a questionnaire containing the age and gender of the patient, the frequency, and the severity of pain. Low back pain patients with coexisting headache and headache patients with coexisting low back pain were excluded. The diagnosis of cervicogenic headache was assigned using the clinical criteria for headache attributed to disorders of the neck by the Headache Classification Committee of the International Headache Society.5 The MIDAS (migraine disability score) questionnaire was used to evaluate the impact of pain on work, household and social activities.6 Pain frequency (number of days with pain in the last 3 months) and intensity (from 0 to 10, where 0 means no pain and 10 means the most intensive pain) were scored as recommended by the MIDAS questionnaire. Severity of depressive symptoms was evaluated by the Beck Depression Inventory (BDI7). The study was approved by the Ethics Committee of the University of Debrecen, and all participants signed an informed consent after reading and understanding the purpose of this study.

Data are presented as means ± SD or medians with 25 and 75 percentiles. Normality of continuous variables was checked by the Shapiro–Wilk test. Due to the distribution of the variables the Mann–Whitney U-test, the Kruskal–Wallis anova and the Spearman correlation tests were applied. General linear models (multivariate general regression models) were used to identify the independent determinants of the severity of disability. Factors proven to be significant predictors in univariate tests were included in the multivariate analysis. Statistical significance was assumed if P < 0.05. Analysis was performed using Statistica for Windows v.6.1 (StatSoft, Tulsa, USA).

RESULTS

Characteristics of the study sample

The characteristics of the study groups are presented in Table 1. Of the 716 subjects, there were 205 men and 511 women. Refusal rate to participate in the study was <5%. Mean age of the five groups significantly differed (P < 0.001); the low back pain group and the cervicogenic headache group was somewhat older and the primary headache groups were somewhat younger than the total mean of the 716 study subjects (Table 1). The gender distribution significantly differed among the five patient groups (P < 0.001); except for the small cluster headache group there were more women in the headache groups and more men in the low back pain group (Table 1). In the total study sample of 716 subjects, the severity of depressive symptoms (i.e. the BDI score) marginally significantly differed between men and women (mean ± SD: men, 11.2 ± 9.3; women, 12.2 ± 9.9; Mann–Whitney test P = 0.021) and there was a significant correlation between age and the BDI score (Spearman R = 0.205, P < 0.001). Neither gender (P = 0.66), nor age (Spearman R = 0.072, P = 0.053) were significantly associated with disability (i.e. the MIDAS score) in the total sample.

Table 1.  Demographic features of the patient groups
GroupAge (years, mean ± SD)nMen : women
Migraine37 ± 10231 36 : 195
Tension-type headache39 ± 15176 48 : 128
Cluster headache36 ± 8 11  8 : 3
Cervicogenic headache42 ± 12182 32 : 150
Low back pain48 ± 12116 81 : 35
Total40 ± 13716205 : 511

Univariate correlates of disability

There was statistically significant difference in the frequency and intensity of pain as well as in the severity of depressive symptoms and the level of disability among the five patient groups (Kruskal–Wallis anovaP < 0.001 in all comparisons; Table 2).

Table 2.  Characteristics of pain, depressive symptoms and disability
FeatureMigraineTension-type headacheCluster headacheCervicogenic headacheLow back pain
  1. Values are medians and 25 and 75 percentile values.

  2. BDI, Beck Depression Inventory; MIDAS, migraine disability score.

n23117611182116
Pain frequency 10 (5–15) 40 (20–70)20 (6–60) 50 (20–80) 40 (18–90)
Pain intensity  8 (7–10)  6 (5–8)10 (9–10)  7.5 (5–9)  7 (5–8)
BDI score  8 (4–14) 11.5 (6–21) 5 (4–8) 12 (7–19) 10 (6–16)
MIDAS score 16 (6–33) 27.5 (0–75)20 (2–55) 29 (0–70) 54 (6.5–97)

Univariate tests (Table 3) were used to select factors for inclusion in the final multivariate model. Factors that were found to have a significant effect on disability in any of the five patient groups in univariate tests were entered in the multivariate model. Gender and age had significant effect on disability only in migraine. Pain frequency appeared to be a significant factor in all five groups. Pain severity and the severity of depressive symptoms seemed to have important effects in univarate tests in three and four of the five patient groups, respectively (Table 3).

Table 3.  Univariate correlates of disability
FeatureMigraineTension-type headacheCluster headacheCervicogenic headacheLow back pain
  1. P-values are for Spearman correlation (R) or the Mann–Whitney U-test for gender.

Gender0.00080.7270.18P = 0.624P = 0.35
AgeR = 0.225R = 0.016R = 0.35R =−0.07R = 0.004
P = 0.0006P = 0.86P = 0.29P = 0.356P = 0.969
Pain frequencyR = 0.553R = 0.354R = 0.63R = 0.36R = 0.37
P < 0.0001P < 0.0001P = 0.039P < 0.0001P < 0.0001
Pain severityR = 0.340R = 0.479R = 0.08R = 0.2R = 0.11
P < 0.0001P < 0.0001P = 0.81P = 0.002P = 0.232
Beck scoreR = 0.451R = 0.446R =−0.02R = 0.40R = 0.29
P < 0.0001P < 0.0001P = 0.94P < 0.0001P = 0.0013

Multivariate correlates of disability

The authors used general regression models to test which are the factors that, after controlling for the effect of other factors, independently determine disability in patients with migraine, tension type headache, low back pain and cervicogenic headache. Due to the small number of patients, the cluster headache group (n = 11) was excluded from this analysis. According to the results of univariate tests, age, gender, pain frequency, pain intensity and the severity of depressive symptoms (i.e. the BDI score) were included in the final model as possible predictors of the level of disability.

Table 4 demonstrates the results of multivariate analysis in the patient groups. Age and gender are significant independent predictors of disability only in tension type headache. Pain frequency is an unequivocal determinant of disability in all patient groups. In low back pain only frequency of pain, whereas in cervicogenic headache pain frequency and the severity of depressive symptoms remained the significant independent predictors of disability. After controlling for the effect of other variables, in contrast to headaches the severity of depressive symptoms (i.e. the BDI score) was not an independent predictor of disability in low back pain (P = 0.77).

Table 4.  Multivariate correlates of disability
FeatureMigraine headacheTension-type headacheCervicogenic headacheLow back pain
  1. Regression coefficients (beta) and their 95% confidence intervals are given. P-values are for the regression coefficient (beta) in the general regression model.

Gender−0.015 (−0.115–0.084)−0.266 (−0.388– –0.145)−0.028 (−0.164–0.108) 0.090 (−0.083–0.262)
P = 0.76P < 0.0001P = 0.688P = 0.304
Age 0.047 (−0.050–0.145)−0.187 (−0.312– –0.062)−0.026 (−0.162–0.111)−0.127 (−0.31–0.059)
P = 0.34P = 0.004P = 0.711P = 0.178
Pain frequency 0.497 (0.394–0.601) 0.279 (0.153–0.405) 0.332 (0.190–0.474) 0.455 (0.266–0.643)
P < 0.0001P < 0.0001P < 0.0001P < 0.0001
Pain severity 0.159 (0.060–0.257) 0.316 (0.186–0.446) 0.114 (−0.027–0.255) 0.179 (−0.004–0.361)
P = 0.002P < 0.0001P = 0.112P = 0.055
Beck score 0.274 (0.171–0.377) 0.284 (0.146–0.422) 0.218 (0.075–0.362) 0.027 (−0.157–0.211)
P < 0.0001P < 0.0001P = 0.003P = 0.774

DISCUSSION

In the present study, the authors compared non-malignant conditions to study the relationships between pain, the severity of depressive symptoms and disability. Previously the authors reported that the determinants of the severity of depression differ between patients with primary headaches and those with low back pain.3 In the current analysis, the authors evaluated predictors of disability in primary and secondary headache and non-headache pain syndromes.

Pain is associated with disability and depression. In a Hungarian population based study, pain-related disability affected one-third of the adult population, and depressive symptomatology was present in 30% of those reporting pain-associated disability.8 It is of question, whether the level of disability in everyday activities depends mostly on pain characteristics (intensity and frequency), pain etiology (like pain related to malignant or non-malignant causes, primary or secondary pain syndromes), pain location (like headache or non-headache pain), or on depression associated with pain. Pain distribution, pain intensity and depressive mood were significant predictors of pain-related disability in a cohort with facial pain.9 In patients with advanced cancer, depression did not correlate with the severity of pain, and impaired activity of daily living in these patients was not related to depression.10

Significant interrelationships were found between pain-related subjective disability and pain intensity as well as pain location in patients with low back pain, and disability overlapped with both pain and work-related factors supporting the multidimensionality of low back disability.11 Pain duration and pain severity were significantly associated with function and disability in low back pain.12 In contrast to the previous studies, no strong relationship was found between pain intensity and the level of reported disability by Kovacs et al.13 in patients with low back pain. Longitudinal findings suggested that pain/disability and psychological distress might be causes and consequences of each other in low back pain.14 Although depression was found to be an independent predictor of chronic pain conditions, including low back pain and cervical pain,15 pain associated depression in most cases was found to be the consequence rather than the cause of pain.16

Perceived disability as well as psychological reactions like depression or pain tolerance might differ in patients with headache and non-headache pain. Such a difference was recently reported by Zwart et al.17 who found that chronic headache was more strongly associated with frequent intake of analgesics than other common pain conditions like chronic neck and chronic low back pain.

Not only differences but also similarities were reported between low back pain and headaches. In a study of primary headaches and low back pain, Jensen18 did not find any differences in stressful life events between the patient groups. Chronic low back pain patients were as fatigued as chronic neck pain patients, and female gender was associated with the complaint in a recent study.19

A different pattern of disability was reported in patients with primary headache than in those with cervicogenic headache.20 Location within the spine was reported to influence the strength of association between self-reported health factors and vertebral deformity.21 In the current analysis, the authors tested whether the relationship between the severity of depressive symptoms and disability in cervicogenic headache has its special pattern, or it resembles to that seen in primary headaches or to that in low back pain. Therefore, the authors tested if pain location (i.e. headache or non-headache pain) or the etiology of the pain (i.e. primary headache or a pain of vertebral origin irrespective of its location) has a more important role in determining disability in pain syndromes. In cervicogenic headache, the authors found that, similarly to primary headaches but in contrast to low back pain, the severity of depressive symptoms was a significant determinant of disability. However, in cervicogenic headache – as in all other pain conditions in this study – the severity of disability was also affected by the frequency of pain. Based on these results and on the authors’ previous analysis,3 the authors conclude that cervicogenic headache bears some characteristics of primary headaches as well as that of low back pain, therefore, both pain location (i.e. headache) and the etiology of pain (i.e. spinal disorder) represent its effect in determining the relationship between pain characteristics, depressive symptoms and disability.

The present study has several limitations. As a cross-sectional study, the data are appropriate to evaluate associations among factors, but direct conclusions about causes and effects can not be drawn. The number of factors analyzed in this study had to be limited due to technical considerations: due to time constraints several possible factors that could influence either depression or disability like socioeconomic status, workplace conditions, personality traits etc., could not be examined. The MIDAS questionnaire used to quantify the severity of disability was originally tested in migraine patients. This questionnaire evaluates general aspects of disability, defined as the number of days lost from work, household and social activities, and has been used in headache syndromes other than migraine.22,23 Therefore, the authors also felt it appropriate to score disability by this questionnaire in pain conditions other than migraine.

In the present study, patient selection was restricted only to those seeking neurological care. Although inclusion of consecutive patients limited selection bias, the study was not a population-based survey, therefore, the results might not be valid for all patients with low back pain or primary and cervicogenic headache. Individuals seeking medical care for neck or back pain were found to have worse health status than those who do not seek care,24 therefore, the results represent subjects with more severe signs or symptoms. However, as the patient sample represent those who meet the health care system, the results are relevant to everyday practice.

As a conclusion, based on the findings, the authors suggest that:

  • 1disability is affected by different factors in different pain syndromes;
  • 2pain frequency is more important than pain severity in determining the level of disability in non-malignant pain syndromes;
  • 3both pain location (i.e. headache) and the etiology of pain (i.e. spinal disorder) represent its effect in determining the relationship between pain characteristics, depressive symptoms and disability; and
  • 4the location of pain (i.e. headache) is more important than the underlying pathophysiology (i.e. pain due the disorders of the vertebral and paravertebral structures) in determining the relationship between disability and the severity of depressive symptoms.

ACKNOWLEDGMENTS

The work was supported by grant No. ETT 130/2003 from the Ministry of Health, Republic of Hungary.

Ancillary