A five-year prospective study of fatigue in primary Sjögren's syndrome

Abstract

Introduction: Fatigue is prevalent in primary Sjögren's syndrome (pSS), and contributes to the considerably reduced health related quality of life in this disease. The symptom is included in proposed disease activity and outcome measures for pSS. Several studies indicate that there is an inflammatory component of fatigue in pSS and other chronic inflammatory rheumatic diseases. The purpose of this study was to investigate fatigue change in pSS in a longitudinal study, and explore whether any clinical or laboratory variables at baseline, including serum cytokines, were associated with a change in fatigue scores over time.

Methods: A clinical and laboratory investigation of 141 patients fulfilling the American-European consensus criteria of pSS was undertaken in the period May 2004 to April 2005. Median time since diagnosis was 5.5 years. Examinations included the fatigue questionnaires: fatigue severity scale (FSS), fatigue visual analogue scale (VAS), functional assessment of chronic illness therapy-fatigue (FACIT-F) and medical outcome study short form-36 (SF-36) vitality, which were repeated in a follow-up investigation in January and February 2010.

Results: A total of 122 patients (87%) responded at both time-points. Thirty-five percent of patients experienced a clinically significant FSS increase. On the group level, fatigue measures did not change except that there was a slight deterioration in SF-36 vitality score. High serum anti-Sjögren's syndrome A antigen (anti-SSA) showed weak associations with high baseline fatigue, and patients with increasing fatigue had lower baseline unstimulated whole salivary volume. Weak associations between increasing fatigue and serum immunoglobulin G (IgG), and the pro-inflammatory cytokine interleukin-17 (IL-17), were observed. Baseline sicca symptoms correlated with higher fatigue both at baseline and with increasing fatigue over time. Linear regression analysis did not identify any predictive ability of clinical or laboratory measures on fatigue change over time.

Conclusions: Fatigue remained mainly unchanged over time. Using multivariate models did not reveal any clinical or laboratory predictors of fatigue change over time.

Figure 1

Frequency of high fatigue. Percentage of patients with high fatigue level, N = 122. A high fatigue level defined as mean fatigue severity scale (FSS) score > 4 was experienced by 70.7% at baseline and 72.1% at follow-up (P = 0.86, McNemar's test). Using an FSS value of 5 as cut-off, the frequencies were 47.9% and 57.4%, respectively (P = 0.08, McNemar's test).

Figure 2

Change in fatigue over time. Errorbars showing 95% confidence intervals for the change in four different fatigue measures over time. N = 122 for FSS, otherwise 121. The vertical line (zero) represents fatigue measures at baseline. FSS and FACIT-F data were normalised to a 0 to 100 scale, corresponding to fatigue VAS and vitality data. Mean change values (and standard deviations) without normalising were: FSS, 0.15 (0.12); fatigue VAS, -1.62 (2.20); FACIT-F, 0.78 (0.76); and vitality, -3.00 (1.48). Vitality (low values representing high fatigue) showed a statistically but not clinically significant decrease over time. The tendencies in the other measures were increasing fatigue by FSS and decreasing fatigue by VAS and FACIT-F. See Table 1 legend for definitions.