- Top of page
- PATIENTS AND METHODS
Ankylosing spondylitis (AS) is a rheumatic disease characterized by inflammation and radiographic changes, mainly localized to the spinal column. Therapies are aimed at reducing inflammatory responses, spinal stiffness, and pain (1–4). Additionally, fatigue is reported to be a common complaint among patients with AS (5–7); according to Calin et al, this complaint should be emphasized more in clinical practice (5). The importance and relevance of fatigue as an outcome measure is also highlighted by different research groups working on outcome measure consensus, including the Patient Perspective Workshop at Outcome Measures in Arthritis Clinical Trials 6 (8) and the Assessment in Ankylosing Spondylitis Working Group (9).
Fatigue in healthy persons occurs as a normal and temporary phenomenon, whereas disease-related fatigue may persist despite adequate amounts of rest and sleep, and is considered long lasting or chronic (10–12). Although the acute form of fatigue is regarded to be purposeful and protective against overexertion, chronic disease-related fatigue is recognized as unrelenting and dysfunctional, and its symptoms may be attributed to an underlying somatic condition (10, 13). Fatigue is frequently reported in ill people as well as in the general population (12, 14–16). Group comparisons of fatigue severity levels by appropriate instruments, as well as examination of associations with other variables, may elucidate fatigue as a disease-related symptom or as a normally occurring phenomenon.
In a group of patients with AS, van Tubergen et al found that the degree of fatigue varied greatly, covering almost the entire range of a scale from 0 (none) to 100 (severe) on the disease-specific Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) fatigue item (17). Additionally, female patients reported higher fatigue scores than male patients. However, sex differences in fatigue levels are also reported in the general population (16, 18, 19), and the sex differences among patients may thus not necessarily reflect a disease-related problem.
Previous studies have shown that fatigue in AS patients is associated with limitations in daily life functioning (5, 6, 17), pain, and stiffness (6), as well as with global wellbeing and mental health (17). Possible associations between fatigue and biologic signs of inflammation in AS are, to our knowledge, scarcely reported. Fatigue has been measured by the fatigue item of the BASDAI, dichotomized with a cutoff at 50 mm (scores ≥50 mm indicate fatigue as a major symptom) (17, 20). To our knowledge, this cutoff is not based upon comparisons with fatigue levels in the general population. The use of reference data for comparisons is an accepted standard for assessment of health status (21), and the generic health status measure Short Form 36 (SF-36) is regarded as a useful instrument for estimating the relative burden of disease (22). Thus, using the SF-36 vitality scores of the general population as reference data (external indicator) could be a valid approach for establishing cutoff values on the disease-specific measure of fatigue.
The aims of this study were to investigate the levels of fatigue in patients with AS compared with the general population and to examine how demographic, self-reported, and clinical measures were associated with fatigue. Furthermore, we wanted to explore the sensitivity and specificity of a disease-specific measure of fatigue using a generic measure as an external criterion.
- Top of page
- PATIENTS AND METHODS
The AS patients in this study were significantly more affected by fatigue than the general population. The results of the study support the relevance of the initiatives that have been taken to put fatigue on the agenda for further research as an outcome measure in rheumatic diseases (9, 31, 32).
Self-reported health measures contributed in explaining about half of the variation in the generic and disease-specific fatigue measures, whereas clinical measures did not add significant explanation to the variation. Similar relationships between fatigue and measures of self-reported disease activity, limitations in functional abilities, and mental health have also been reported by others (5, 6, 17, 33). No differences were found in joint mobility or the inflammatory markers between the VTlow and the VTnormal groups. However, in accordance with other studies, moderate values of the inflammatory markers were found (20, 34), and the validity of ESR and CRP in AS have been discussed (34–36). A recent study by Dernis-Labous et al (20) showed that nonsteroidal antiinflammatory drug therapy strongly reduced pain and functional impairment in a group of AS patients, whereas the change in fatigue level was of lower magnitude. Thus, the relationship between fatigue and inflammatory markers is unclear, and more clinical trials are needed to explore whether interventions, such as disease-modifying drugs and physical activity, influence self-reported fatigue.
In the present study, women were more likely than men to be classified in the low vitality group, even if the cutoff score based on the normative data was made specific for each sex. Corresponding sex differences with respect to pain and fatigue are found in other studies (17, 37, 38). Although the effect of sex tended to be weaker in the multivariate analyses (Tables 2 and 3), women reported significantly worse scores on both the disease-specific and the generic measures, and the assumption was supported that, in clinical practice, special attention should be paid regarding women and fatigue (5).
In previous studies, the fatigued patients have been defined as those having BASDAI fatigue scores ≥50 mm (17, 20); according to this method, 53% and 63% (respectively) of their samples were classified as fatigued. When applying the same cutoff on the BASDAI fatigue item in the present study, 62% fulfilled this fatigue criterion. However, we made a cutoff between “affected” and “not affected” patients based on the normative data from the general population. The 10th percentile was chosen as cutoff, i.e., suggesting that low vitality normally occurs in 10% of the general population. Using this cutoff, about one-third of the patients in this study reported to suffer from low vitality. Because the proposed level of 50 mm on the BASDAI fatigue item seemed to capture nearly all patients with fatigue (sensitivity = 0.90), but also many without fatigue (specificity = 0.46) (Figure 2A and 2B), a more conservative cutoff should be considered. In an ROC curve analysis, we found an appropriate tradeoff between sensitivity and specificity (0.77 and 0.82, respectively) with a cutoff at 70 mm on the BASDAI fatigue item for classifying patients with low vitality. The use of this cutoff will avoid classifying too many false-positive individuals into a fatigue group.
Several factors have been suggested to cause or to associate with disease-related fatigue. Examples are bodily changes due to the disease processes itself, e.g., immunologic responses; drugs; or inability to get enough rest related to strong pain and disruptive sleep (29, 33). Fatigue can also be associated with deconditioning (11), and the presence of fatigue is one of the criteria of depression (39). Thus, fatigue is a multidimensional issue, and VASs provide information of only the patients' experienced level of disease-related fatigue. The vitality scale and the fatigue item had similar associations to other measures in multivariate analyses, and a substantial correlation to each other, suggesting that they measure similar constructs. Concomitant use of a domain-specific instrument would have provided additional information, but was not considered feasible due to the number of questionnaires to be completed by the patients in this study. Additional studies are needed to explore the different dimensions of fatigue. More specific information about fatigue might be important for the clinical management and for establishing valid outcome measures in clinical trials.
In the present sample, the demographic variables, the disease duration, and the measures of the self-reported disease severity are comparable with other studies (7, 17, 40, 41). In accordance with previous studies (6, 17), we applied a hospital register for identifying patients with AS. It may be questioned whether a sample based on a hospital register is representative for the patient group or if it includes those being most severely affected by disease. However, the clinical routine in Norway is that patients should have the diagnosis of AS confirmed by a rheumatologist. Thus, this patient cohort comprises both patients being seen by specialists and patients being seen by family physicians, indicating that the sample is representative. A possible shortcoming might be that all patients lived in a city area, but whether or how this may influence the results is unknown.
In this study, about one-third of the AS patients reported severe levels of fatigue. The phenomenon is associated with self-reported disease activity, physical functioning, and mental health, but not with objective clinical measures. The disease-specific BASDAI fatigue item reached an acceptable specificity and sensitivity with a cutoff at 70 mm. Further examinations of the different aspects of fatigue are needed to improve the accuracy of fatigue as an outcome measure in AS.