Fatigue in adult coeliac disease
Prof. C. Ciacci, Gastroenterologia, via Pansini 5, 80131 Napoli, Italy.
Background : Fatigue is reported by many adults at the moment of diagnosis of coeliac disease and during follow-up.
Aim : To evaluate the prevalence, characteristics and associations of fatigue in adult coeliac disease patients.
Methods : The investigated sample comprised adults from Campania, Italy. A total of 130 coeliac disease patients were consecutively recruited in both treated (59 on gluten-free diet) and untreated conditions (71 on normal diet). The control group was made up of 80 healthy controls. Coeliac disease patients and healthy controls underwent laboratory tests, a set of questionnaires for studying fatigue: visual analogue scale for fatigue, chronic fatigue syndrome questionnaire, fatigue severity scale and a modified version of the Zung self-rating depression scale.
Results : Coeliac disease patients showed a significantly lower body mass index than controls (P = 0.0001), lower serum iron (P = 0.04). The entire cohort of coeliac disease patients reported greater modified version of the Zung self-rating depression scale score (P = 0.001), greater visual analogue scale for fatigue score (P = 0.0001) and greater chronic fatigue syndrome questionnaire score (P = 0.0001) compared with healthy controls. Coeliac disease patients on a gluten-free diet had a significantly higher modified version of the Zung self-rating depression scale score than coeliacs on a normal diet (P = 0.001). The prevalence of pathological modified version of the Zung self-rating depression scale score was 17% in all coeliac disease patients and 0% in healthy controls. A significant correlation was found between modified version of the Zung self-rating depression scale score and fatigue scale scores in coeliacs on a normal diet. Presence/absence of gastrointestinal symptoms did not show any significant correlation with modified version of the Zung self-rating depression scale score and fatigue scale scores. In coeliacs on a gluten-free diet, modified version of the Zung self-rating depression scale and fatigue scales scores did not significantly differ from coeliacs on a normal diet and were not related to dietetic compliance.
Conclusion : In coeliacs, fatigue is a common finding, which ameliorates with the gluten-free diet and is strictly correlated to depression although coeliacs on a gluten-free diet showed more frequent and more severe depression symptoms than coeliacs on a normal diet.
Fatigue is a subjective perception that can be definite as difficult in initiation or sustaining regular activities. In patients’ definition, it can overlap the feelings of tiredness, muscle weakness, depression, fatigability and/or irritability and affects the quality of life of the subject.
As a symptom, fatigue is the result of a complex control system of work output that encompasses hormones, neural transmission, cognitive processing and sensory input. It is frequently reported in primary care,1 and can be related to systemic or neuromuscular disease.2, 3
Fatigue is reported by many adults at the moment of diagnosis of coeliac disease (CD).4, 5 It is known that CD patients also report psychic symptoms6 and depression symptoms,7 which do not regress with gluten-free diet.8 The aim of the present study was to evaluate prevalence, characteristics and associations of fatigue and depression in adult CD.
With the proposed sample of 60 subjects for each group the study would have the power of 95.9% to yield a statistically significant result. The study was conducted in an out-patient clinic. The study was approved by the Ethic Committee of the Federico II University of Naples, Italy.
The investigated sample was made up of Caucasian adults from Campania, Italy (41 males, 169 females). The Ethical Committee of the University Federico II of Naples approved the study protocol. All those asked and eligible gave their informed consent and participated in the study (100%). Coeliac patients were consecutively recruited. Patients were divided in two groups: the first was made up of newly diagnosed with CD and still on gluten-containing ‘normal’ diet (CD-ND); the second was made up of those who underwent the annual visit necessary in Italy to obtain gluten-free food under the Health System coverage and on gluten-free diet (CD-GFD). The control group was made up of volunteers recruited from among medical and non-medical hospital staff (HC). Coeliac patients and controls had comparable number of years of education (CD 13.2 ± 4.1, HC 13.4 ± 3.2) and type of work (number of students CD = 24, HC = 15; housewife CD = 19, HC = 10; blue collar CD = 21, HC = 12; white collar CD = 19, HC = 13; professional CD = 22, HC = 14; unemployed CD = 25, HC = 16).
Coeliac disease diagnosis was made on the presence of specific antibodies and an intestinal biopsy compatible with gluten-related damage. HC were all negative for antitransglutaminasi antibodies, and had no gastrointestinal symptoms or anaemia.
Inclusion criteria were: age 18 years and greater. Exclusion criteria were: age lower than 18 years, lack of informed consent, presence of major psychiatric disease, presence of active thyroid gland disease.
A set made up of four different questionnaires was administered to all patients.
- (i) A modified version of the Zung self-rating depression scale (M-SDS) to investigate the presence of depressive symptoms.5 The M-SDS contains 17 of 20 items belonging to the original version of the Zung scale.9 The three items evaluating gastroenteric symptoms of depression (i.e. decreased appetite, weight loss and constipation) had been ruled out to avoid bias because of the illness. The items belonging to the scale have been classified as ‘biological’ and ‘psychological’, and two separate score concerning the two classes of symptoms have been considered.10 In the M-SDS scale, 44 is the cut-off score for pathological depression.
- (ii) The ‘chronic fatigue syndrome’ (CFS) questionnaire was validated to investigate the chronic fatigue. The questionnaire includes eight questions for physical fatigue (PhF) and five questions for mental fatigue (MF) including questions about memory and concentration. The total score is from 0 to 26.3
- (iii) The fatigue severity scale (FSS) evaluates and quantifies the fatigue because of chronic disease. In the original study, the FSS score was independent from the presence and severity of depression, the questionnaire includes nine questions and the patients score the answers from 1 to 7 according to how low or high he/she agrees with the sentence. The scores range from 9 to 63.11
- (iv) The last questionnaire was made up of a single question: ‘How did you feel in the last week?’ The patients were asked to mark a visual analogue scale (VAS) consisted of a 10-cm line at the left extreme the sentence ‘I never feel tired’ and at the opposite extreme the sentence ‘I always feel tired’. Thus, the possible score ranges from 0 to 10.
Data collection and statistical analysis were performed using SPSS 10 package for Windows.
Chi-square analysis, Dunn's post hoc analysis, Mann–Whitney U-test and Kruskall–Wallis test for non-parametric data were used when appropriate.
Table 1 shows the characteristics of the entire study cohort. The entire cohort of CD patients (CD-ND + CD-GFD) was significantly different from HC for main nutritional indices, such as body mass index, ferritin and haemoglobin.
Table 1. Gender, age and descriptive analysis of data from the study cohort: controls and CD patients were significantly different for BMI, haemoglobin and ferritin
|Age (years)||29.20 ± 5.7||31.25 ± 9.6||7.748||0.096|
|With diagnosis in childhood||–||15||–||–|
|BMI (kg/h2)||24.0 ± 2.5||21.5 ± 3.2||19.069||0.0001|
|Ferritin (ng/dL)||50.2 ± 13.8||18.5 ± 27.0||0.0001||0.04|
|Haemoglobin (g/dL)||13.4 ± 2.6||11.7 ± 1.9||0.0001||0.05|
Table 2 shows mean and s.d., median and interquartile range (IQR) of the questionnaires scores for CD patients and controls and the difference between the two groups. The scale scores were positively skewed and statistical analysis was carried out with tests for non-parametric data. Significant differences between the two groups were found for all scales, being fatigue and depression prevalence and score more relevant in CD patients. For depressive symptoms it must be noted that 17% of CD patients scored above 44 in the M-SDS scale, which is the cut-off score for pathological depression. None of the HC group scored more than 44. Three patients with diagnosis in childhood and 15 patients with diagnosis in adulthood showed M-SDS score above 44. Separate analysis in CD patients and HC group, showed no significant correlation of age with fatigue scales scores (rank correlation coefficient <0.145, P > 0.10), indicating that the difference between patients and controls is not related to age differences. Depression measured by M-SDS, conversely, was significantly correlated with age in both patients and controls (P = 0.04). With regards to the diet, 12 CD-GFD patients and six CD-ND patients scored more than 44.
Table 2. Scores of fatigue scales and analysis of variance of data in CD patients and controls. Data are expressed as mean ± s.d. and median (IQR). Statistical analysis was carried out by Mann–Whitney U-test for non-parametric data
| Mean ± s.d.||34.3 ± 9.5||28.53 ± 4.0||2736||0.001|
| Median (IQR)||33 (15)||30 (5.0)|
| Mean ± s.d.||5.4 ± 4.1||1.79 ± 1.5||1940||0.0001|
| Median (IQR)||5.0 (6.0)||1.0 (2.0)|
| Mean ± s.d.||2.8 ± 2.1||1.72 ± 1.5||2938||0.003|
| Median (IQR)||2.0 (3.0)||1.50 (2.0)|
|CFS (PhF + MF)|
| Mean ± s.d.||8.1 ± 5.7||3.51 ± 2.5||1113||0.0001|
| Median (IQR)||7.0 (7.0)||3.0 (3.0)|
| Mean ± s.d.||35.5 ± 15.2||16.65 ± 7.7||1955||0.0001|
| Median (IQR)||35 (24)||15 (8.0)|
| Mean ± s.d.||5.8 ± 2.5||4.38 ± 2.1||2712||0.0001|
| Median (IQR)||6.0 (4.0)||4.50 (2.0)|
Table 3 shows findings from fatigue scales and M-SDS in CD-ND (before diagnosis) vs. CD-GFD (on gluten-free diet). Values are also given as mean ± s.d. and median IQR. Significant differences were found for all fatigue scales and M-SDS scale among groups. Dunn's post hoc analysis indicated that the main difference accounted for the CD-ND and CD-GFD groups compared separately with the HC group. In fact, no differences were noted for fatigue scales mean scores in CD-ND vs. CD-GFD. In the coeliac patients group, the only difference was noted for M-SDS scores, which were significantly higher in CD-GFD than CD-ND (P = 0.002).
Table 3. Data from fatigue scales in coeliac patients before diet (CD-ND) and CD-GFD and HC. Data are expressed as mean ± s.d. and median ± IQR. Statistical analysis was carried out by Kruskall–Wallis test for non-parametric data (*)
| Mean ± s.d.||32.0 ± 8.8||37.2 ± 9.6||28.53 ± 4.1||24.83||0.0001|
| Median (IQR)||27 (13)||37 (12)||30 (5.0)|
| Mean ± s.d.||5.9 ± 4.0||4.8 ± 4.1||1.8 ± 1.5||35.794||0.0001|
| Median (IQR)||6.0 (5.0)||4.0 (6.0)||1.0 (2.0)|
| Mean ± s.d.||2.8 ± 2.0||2.4 ± 2.2||1.7 ± 1.45||11.467||0.0003|
| Median (IQR)||2.0 (3.0)||2.0 (3.0)||1.50 (2.0)|
|CFS (PhF + M F)|
| Mean ± s.d.||8.7 ± 5.5||7.2 ± 5.7||3.51 ± 2.4||65.737||0.0001|
| Median (IQR)||8.0 (7.0)||7.0 (8.0)||3.0 (3.0)|
| Mean ± s.d.||36.3 ± 13.3||34.4 ± 17.2||16.65 ± 7.0||35.569||0.0001|
| Median (IQR)||35 (23)||33.0 (27.0)||15 (8.0)|
| Mean ± s.d.||5.7 ± 2.3||5.9 ± 2.6||4.38 ± 1.9||12.986||0.002|
| Median (IQR)||5.0 (4.0)||7.0 (5.0)||4.50 (2.0)|
Table 4 shows findings from fatigue scales and M-SDS in CD patients with diagnosis in childhood vs. patients with CD diagnosis in adulthood. No significant differences were found for all fatigue scales in patients on gluten-free diet since childhood vs. patients with late diagnosis and more recent gluten-free diet.
Table 4. Analysis of variance of data (Mann–Whitney U-test) from fatigue scales in CD patients with diagnosis and gluten-free diet since childhood compared with patients with CD diagnosis adulthood
|CFS (PhF + MF)||1214||0.606|
Self-reported memory and ability to concentrate (items of CFS scale, section mental fatigue, 0 = as usual, 1 = worse than usual, 2 = worst than usual) were analysed in a separate analysis. No differences were noted among groups for both items (memory score – CD-ND 0.5 ± 0.6, CD-GFD 0.4 ± 0.6, HC 0.3 ± 0.5, F = 1.5, P = 0.218; concentration score – CD-ND 0.6 ± 0.6, CD-GFD 0.5 ± 0.7, HC 0.4 ± 0.5, F = 1.5 P = 0.229).
Table 5 shows linear regression analysis performed with M-SDS as a dependent variable. FSS scores and VAS scores appear to be directly related to M-SDS scores.
Table 5. Linear regression analysis: in CD patients only, M-SDS score analysed as dependent variable with relation to fatigue scales scores
|CFS (PhF + MF)||4.454||7.000||2.737||0.636||0.525|
Gender or symptoms reported at the moment of diagnosis were not related to fatigue or M-SDS scores. In fact, multiple regression analysis (using M-SDS score or fatigue scales scores as a dependent variable, and as independent variables gender, age, presence or absence of gastrointestinal symptoms and/or anaemia) showed that these variables did not influence depression and/or fatigue levels (data not shown).
Fatigue is a common compliant in patients with chronic disease. It is frequently reported in primary care,1 and can be related to systemic or neuromuscular disease.2
The findings of the present study indicate that fatigue is more frequent in coeliac patients than in controls and it is not related to gender, age at diagnosis, anthropometry or laboratory index examined or any of the symptoms of gluten sensitivity. The possibility exists that fatigue is a gluten-related symptom.5 Patients on gluten-free diet show not significantly lower scores of fatigue scales compared with patients on gluten-containing diet. We also explored the possible correlation of fatigue and depression in CD patients. In fact, psychic disturbances in CD patients are common and confirmed by a number of studies,5,6,12,13 and fatigue symptoms may overlap some psychic symptoms in patients’ perception. In the present study, depression was significantly more frequent and severe in coeliac patients than controls, and in untreated patients only, fatigue and depression were strongly related. In patients on gluten-free diet, in fact, fatigue tends to be reduced in intensity while depression seems to persist or even worsen, supporting recent observations that gluten-free diet alone fails to significantly reduce the percentage of CD patients affected by depression.7
It is conceivable that in CD patient's depression, initially present as a consequence of both the disease symptoms with a decreased sensation of general well-being,14 and malabsorption-malnutrition related to chronic gastrointestinal disease15,16 is sustained by the reduction in the quality of life11, 12 related to dietary restrictions that lead to difficulties in daily social relationships and is considered to be a feature of CD.
The findings of the present study allow the conclusion that fatigue is also, in CD, a feature of the disease, not related to symptoms and little improving with gluten-free diet. Its strong relationship with depression indicates it may have, at least in part, a cognitive/affective origin.
Fatigue may also be the only symptom of CD, and a test for gluten-induced autoantibodies should be requested in evaluating medically unexplained physical symptoms and related syndromes such as CFS.17
The strong relationship among fatigue and depression should not limit the possibility that other organic factors may contribute to sustaining both conditions when gluten-free diet is also initiated.
As gluten-free diet represents the most important treatment of coeliac patients and is the only treatment that prevents several associated non-malignant18 and malignant18,19 complications, the importance of recognizing and treating depression as a risk factor for poor outcomes among patients who might not be adhering to medical advice is of great importance.20 Treating depression from the very beginning and reassuring the patients reporting fatigue, will improve quality of life of coeliac patients. Addolorato et al.16 underlines the beneficial effect of psychological support for CD patients after the diagnosis and at the beginning of diet treatment, in order to correct psychological disorders and to improve the acceptance of the diet, limiting the lack of compliance to treatment and the related disease complications.
The authors are grateful to Ms Wendy Parsons for revising the manuscript. They are also grateful to SOFAR spa for partly supporting the Dr Siniscalchi project.