There has been recent interest in the possibility that undiagnosed coeliac disease (CD) might be the cause of diverse clinical symptoms, most particularly “tired all the time”.¹ A recent study reported a prevalence of three in 100 cases in a primary care environment in which samples were taken from patients with a range of symptoms and signs.² The second most frequent symptom reported by the endomysial antibody (EMA) positive patients was “being tired all the time”. We decided to examine the prevalence of EMA in patients attending our tertiary referral centre with the diagnosis of chronic fatigue syndrome (CFS).

We tested serum from 100 consecutive patients (47 men, 53 women; median age, 40 years; range, 18–57) referred to our specialist clinic and satisfying the standard CDC criteria for a diagnosis of CFS, and from 100 healthy control subjects (45 men, 55 women; median age, 40 years; range, 18–68) who were blood donors at the South East Thames Blood Transfusion Service. The CFS samples had been stored as part of other studies, and were analysed retrospectively. EMA of the IgA class were detected by indirect immunofluorescence (IF) using cryostat sections of distal primate oesophagus as substrate (Binding Site, Birmingham, UK). Positive samples were confirmed using an enzyme linked immunosorbant assay (ELISA) for the detection of antitissue transglutaminase antibodies³ (Menarini Diagnostics, Wokingham, UK), tissue transglutaminase being the autoantigen responsible for the IF pattern of EMA. To exclude selective IgA deficiency, serum IgA concentrations were measured by laser nephelometry using specific antisera according to the manufacturer's instructions (Behring Laser Nephelometer II; Dade Behring, Dortmund, Germany).

Two of the 100 CFS samples were positive for EMA using IF, and this was confirmed by ELISA, but none of the 100 control samples was positive. None of the subjects had selective IgA deficiency. Mean (SD) serum IgA concentrations among patients with CFS were 2.1 g/litre (0.98). Neither of the positive cases, both women aged 27 and 54, had reported symptoms typical of CD, although one had a history of constipation. Routine blood tests including serum proteins and full blood count were normal, and both had been seen by consultant physicians before referral. Both had histories of hypothyroidism, were taking long term thyroxine, and were currently euthyroid. Before the diagnosis of CD was made retrospectively, both had received cognitive behaviour therapy (CBT), a standard treatment for CFS. In both cases, CBT led to a substantial improvement in the quality of life and physical activity, but neither patient was symptom free at the end of treatment or at six months follow up. In both cases, CD was subsequently confirmed on jejunal biopsy after the retrospective identification.

In general, it remains true that although a wide range of physical illnesses can be misdiagnosed as CFS (see Wessely et al for review⁴), in practice this is uncommon. In particular, if basic physical examination, investigations, and history are unremarkable, misdiagnosis of CFS and other physical illnesses is very unusual. Until now there have only been two reports concerning three cases of CD being misdiagnosed as CFS.^{5, 6}

However, there is now evidence from primary care of a surprisingly high frequency of unsuspected positive EMA tests in people with non-specific symptoms and a suggestion that a higher index of suspicion is needed when assessing such patients.² We now extend that observation to our CFS clinic. Indeed, given our prevalence of 2%, and the fact that there is a treatment for CD, we now suggest that screening for CD should be added to the relatively short list of mandatory investigations in suspected cases of CFS.