Chronic fatigue syndrome (CFS) affects about 0.1–2.5% of the general adult population (Reeves et al., 2007). There are no characteristic clinical signs or laboratory abnormalities so CFS must be diagnosed by self-reported symptoms and ruling-out medical and psychiatric conditions with similar clinical presentations (Fukuda et al., 1994). Because the pathophysiology remains inchoate (Prins et al., 2006), current treatment modalities seek to alleviate symptoms. Critically, given the magnitude of the public health problem, no primary or secondary prevention strategies have yet been designed or tested. However, stress has been increasingly recognized as important in both the etiology and pathophysiology of CFS (Nater et al., 2010).

Various studies reported that persons with CFS experienced significantly more stressful life events in the year before illness onset than controls during the same calendar time-period (Reyes et al., 1996, Masuda et al., 2002, Faulkner and Smith, 2008). Very few studies focused on qualitative measures of experienced or perceived stress levels. One study found that the physical symptoms of CFS were exacerbated in people who experienced Hurricane Andrew (Lutgendorf et al., 1995). Another study found that about 15% of persons with CFS had posttraumatic stress disorder (PTSD), which reflects aberrant response to extreme stress (Nater et al., 2009).

Thus, both occurrence of recent stressful life events and chronic stress levels appear to be increased in individuals with CFS, but most studies have operationalized stress by quantitatively counting life events. Stress research has repeatedly shown that the qualitative evaluation of a life event as negative is crucial in its subsequent effect on health and no studies of which we are aware have measured stress perception, occurrence of stressful life events and stress levels within the same study. In addition, published studies have defined CFS by different criteria and with varying degrees of rigor. Finally, almost all published studies include persons with CFS enrolled from primary or tertiary care centres which may result in self-selection bias; only 50% of people with CFS have consulted a physician for their illness and fewer than 20% have been appropriately diagnosed (Solomon and Reeves, 2004).

The objectives of the current study were therefore to assess the occurrence of quantitative and qualitative measures of stress in a population-based sample of CFS, and examine the strength of the association of different aspects of stress with CFS.