Coeliac disease (CD) develops in response to unknown environmental factors in genetically susceptible individuals(Reference Lebwohl, Sanders and Green1). It is a T cell-mediated autoimmune chronic gastrointestinal disorder characterised by permanent intolerance to gluten, a protein composite found in wheat, barley and rye(Reference Farrell and Kelly2). It is histologically characterised by villous atrophy of the small bowel mucosa(Reference Cammarota, Cuoco and Cesaro3), leading to malabsorption of micronutrients(Reference Di Sabatino and Corazza4, Reference Reilly, Fasano and Green5) and when symptomatic leads to diarrhoea, weight loss and abdominal pain(Reference Green and Jabri6). Duodenal biopsy and serology offers objective classification of the severity of CD based on research-established criteria(Reference Ensari7, Reference Villanacci8).
CD is a multi-systemic disorder(Reference Kochhar, Singh and Gill9) and can include other organs such as the skin, liver, thyroid, pancreas, heart and brain(Reference Duggan and Duggan10–Reference Emilsson, Carlsson and Holmqvist14) and has potential for long-term complications, which might include osteoporosis(Reference Meyer, Stavropolous and Diamond15), anaemia(Reference Mahadev, Laszkowska and Sundstrom16) and more serious complications such as intestinal lymphoma(Reference Catassi, Bearzi and Holmes17). In the majority of cases, the condition responds to a gluten-free (GF) diet (GFD)(Reference Rubio-Tapia, Rahim and See18) only to relapse after reintroduction of gluten(Reference Silvester and Rashid19, Reference Jacobsson, Friedrichsen and Goransson20). An association between CD and increased mortality has been documented, whereby disease related mortality reduced after diagnosis and treatment with a GFD(Reference Ludvigsson, Montgomery and Ekbom21, Reference Logan, Rifkind and Turner22).
Worldwide prevalence of CD is estimated at 1·4 % and presently there is no known cure(Reference Singh, Arora and Strand23). Once considered a disorder of Europeans and people of European descent, it is now known to be a global condition with variations in presentation in people of different ethnicities, with some studies reporting a higher prevalence, for example approximately 3 % of patients from the Punjab region residing in the UK or the USA were reported as having CD(Reference Sher, Fraser and Wicks24–Reference Krigel, Turner and Makharia26). Rates in other countries that have reported the incidence of CD appear slightly lower; in Libya(Reference Alarida, Harown and Ahmaida27) and Iran. Additionally, there has been a reported prevalence of 0·8 % in Tunisia(Reference Ben Hariz, Kallel-Sellami and Kallel28) and Turkey(Reference Tatar, Elsurer and Simsek29). Furthermore, prevalence ranging between 0·5 and 0·6 % has been reported in Egypt(Reference Abu-Zekry, Kryszak and Diab30). Interestingly, CD has been described as being less evident in South East Asian countries(Reference Lionetti, Gatti and Pulvirenti31).
Dietary adherence to a GFD is paramount as this is the only treatment available for CD. Duodenal histological improvement, after removal of gluten from the diet(Reference Galli, Esposito and Lahner32), reverses the malabsorption state related to CD. Annibale et al.(Reference Annibale, Severi and Chistolini33) noted that recovery was dependant on various factors such as time between biopsies and starting GFD, severity of histopathologic changes at diagnosis, and age of the patients. Iron deficiency anaemia has been shown to improve following a GFD, suggesting increased iron absorption(Reference Annibale, Severi and Chistolini33, Reference Belei, Dobrescu and Heredea34). Histological recovery can take a long time and because CD can result in patchy villous atrophy of the duodenum of variable severity, hence, variable degrees of malabsorption is seen(Reference Lee, Lo and Memeo35). Adhering to a GFD can be very challenging(Reference Zarkadas, Dubois and MacIsaac36); it requires knowledge, skills and modified behaviours to undertake the substantial changes to dietary habits.
In CD there are two key aspects related to dietary adherence: patient focused aspects such as the challenges encountered when following the GFD(Reference Lee and Newman37, Reference Martinez-Martinez, Alegre-Martinez and Garcia-Ibanez38) and healthcare professional perspectives, whereby it is difficult to determine if patients are adhering to the diet. The WHO defines adherence as ‘the extent to which a person's behaviour – taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider’(Reference Sabaté39). Over the years research has shown that following a GFD not only reverses the duodenal pathology, but also improves quality of life, and reduces CD-related morbidity(Reference Bernardo and Pena40, Reference Hall, Rubin and Charnock41).
Methodological challenges of assessing dietary adherence
Histology from a duodenal biopsy is the gold standard when assessing GF dietary adherence, as villous atrophy due to ingested gluten is visible. However, due to the invasive, costly and time-consuming nature of the procedure a variety of other methods are used in clinical and research settings: serology, faecal or urine tests, dietitian's assessment, interviews, validated and non-validated questionnaires and patient reported adherence. On its own, symptomatic improvement may not present an accurate picture of dietary adherence as there is a subset of the coeliac population with significant villus atrophy and no symptoms(Reference Lähdeaho, Mäki and Laurila42, Reference Laurikka, Salmi and Collin43). Examination of haematological markers such as blood count, folate, vitamin B12, iron studies and liver biochemistry(Reference Rubio-Tapia, Hill and Kelly44) with improvement in parameters such as anaemia may indirectly result in improvement in absorption of micronutrients(Reference Kemppainen, Kosma and Janatuinen45).
Among the serology, antibodies in response to gluten consumption are measured; for example levels of anti-transglutaminase antibodies (anti-tTG) are used extensively in clinical practice for assessing GFD adherence. There is also evidence to suggest that persistently elevated anti-tTG denotes non-adherence with GFD(Reference Dipper, Maitra and Thomas46) and falling anti-tTG indicates adherence(Reference Leffler, Edwards George and Dennis47). There are studies which have examined the reliability of anti-tTG for this purpose, and reported a discrepancy between serological improvement and mucosal recovery(Reference Kaukinen, Sulkanen and Maki48–Reference Hopper, Hadjivassiliou and Hurlstone50). Other serological markers, such as endomysial antibodies and antibodies against deamidated gliadin peptides, also have reliability concerns in relation to measuring adherence to a GFD(Reference Monzani, Rapa and Fonio51–Reference Dickey, Hughes and McMillan53). Although these may not be reliable markers of histological recovery their base-line levels are important as they remain elevated with persistent dietary transgressions(Reference Dipper, Maitra and Thomas46). A promising advancement is the development of tests to measure gluten immunogenic peptides(Reference Moreno, Cebolla and Muñoz-Suano54). These peptides are involved in the immunogenic reaction of CD and anti-α-gliadin G12 antibody may be detected in body fluids, such as faeces and urine, of patients(Reference Comino, Real and Vivas55) and this has been used to monitor adherence to GFD in research settings(Reference Comino, Fernández-Bañares and Esteve56).
An assessment of GF dietary adherence by a dietitian is considered highly effective and is inclusive of assessing knowledge, behaviour while dining out, and intent to adhere(Reference Kurien, Trott and Sanders57, Reference Mehta, Pan and Riley58), indeed it is considered a gold standard by some authors(Reference Leffler, Dennis and Edwards George59, Reference Sainsbury, Mullan and Sharpe60). Leffler et al.(Reference Leffler, Dennis and Edwards George59) suggest that although serologic tests have very high sensitivities and specificities for the diagnosis of CD, they cannot replace dietitian evaluation in the assessment of GFD adherence. However, without a standard process it is difficult to replicate in clinical trials and as such an association between dietetic assessment and duodenal biopsy is yet to be published.
Differing questionnaire-based methodologies have been used to measure dietary adherence in patients with CD in several studies(Reference Leffler, Dennis and Edwards George59, Reference Butterworth, Banfield and Iqbal61–Reference Sainsbury, Mullan and Sharpe64). Leffler et al.(Reference Leffler, Dennis and Edwards George59) validated a seven item, CD-specific questionnaire to measure adherence with GF, the Coeliac Disease Adherence Test; with a score >13 deemed as not adhering to the GFD(Reference Villafuerte-Galvez, Vanga and Dennis65). This questionnaire has been subsequently used in several studies, allowing for comparison between population groups(Reference Muhammad, Reeves and Ishaq63–Reference Villafuerte-Galvez, Vanga and Dennis65). A Birmingham-based study evaluated adherence with the GFD with a 20-item questionnaire that holistically approached adherence employing clinical, social and economic terms(Reference Butterworth, Banfield and Iqbal61). The drawbacks of questionnaire-based studies include their inability to explore particular responses in depth, thereby giving it a static look in comparison to an interview. Some studies have used questionnaires to separate intentional and inadvertent gluten ingestion(Reference Hall, Rubin and Charnock41, Reference Casellas, Rodrigo and Lucendo66).
GFD adherence rates are variable depending on the methodology used and the population studied. Studies have defined adherence to a GFD using a range of terminology inclusive of: strict, partially or fairly strict and non-adherent. Hall et al.(Reference Hall, Rubin and Charnock41) reported adherence ranged from 42 to 91 % from thirty-eight studies published up to 2007, since then at least a further twelve studies have been published with adherence rate ranging from 53 to 76 %(Reference Holmes and Moor25, Reference Hall, Rubin and Charnock41, Reference Leffler, Dennis and Edwards George59, Reference Muhammad, Reeves and Ishaq63, Reference Villafuerte-Galvez, Vanga and Dennis65–Reference Sainsbury, Halmos and Knowles72). The reported variability in the adherence rate may be explained by the variability in the methodologies used by the studies. Three of them used dietitian assessment, the other three used validated questionaries and the remaining four used a non-validated self-reported measure of adherence.
Many studies do not report ethnicity of the participants and in particular there is a very limited body of literature about the dietary adherence of different ethnic groups with CD. Table 1 highlights six studies that report GFD adherence in different ethnic groups. In order to evaluate adherence rates and barriers to adherence in this population, a study in India recruited treatment-naive participants and those already following a GFD (n 146) and using a questionnaire; it was reported that the former group had an adherence rate of 65 % and the latter group 53 %. The main cause of non-adherence was reported to be poor availability of GF foods(Reference Rajpoot, Sharma and Harikrishnan70). However, the study may well have been affected by selection bias as patients were drawn from specialised CD clinics, which are attended by motivated patients and may not truly reflect the adherence of patients who do not regularly attend clinics. Garg & Gupta(Reference Garg and Gupta73) reported Indian children had slightly higher values for adherence (66 %) and this was related to age at presentation, nuclear families, mother's education, and parents having better knowledge of CD. Table 1 summarises studies concerning adherence in patients.
Studies reporting gluten-free (GF) dietary adherence in Caucasian and South Asian adults
Factors impacting upon dietary adherence
Research has indicated that causes of adherence and non-adherence to a GFD are numerous and multifactorial; these include socio-demographics, age of diagnosis, whether symptoms are present with gluten ingestion, practical difficulties associated with the GFD, and membership of advocacy groups(Reference Muhammad, Reeves and Ishaq63, Reference Leffler, Edwards-George and Dennis67, Reference Hall, Rubin and Charnock74).
Among the socio-demographic factors, age is significant and shows variability in the adherence rate as in childhood it tends to be higher. However, adolescents often have issues with adherence(Reference Arnone and Fitzsimons75) as they may have concerns about isolation and stigmatisation for following a GFD(Reference Olsson, Lyon and Hörnell76). In contrast, it has been reported that patients diagnosed later in life have relatively good adherence (77–90 %)(Reference Vilppula, Kaukinen and Luostarinen77, Reference Casella, Zanini and Lanzarotto78). Hall et al.(Reference Hall, Rubin and Charnock41) reported no difference between genders in relation to adherence with GFD from a systematic review of thirty-eight studies up to 2007, however, in 2018 a large study reported being male was associated with better adherence(Reference Halmos, Deng and Knowles79). This study of more than 5000 Australians with CD demonstrated ‘symptoms after gluten ingestion’ was an independent predictor of GF dietary adherence.
Strict adherence to a GFD has been associated with patients reporting ‘feelings of desperation’ or a need to gain or lose weight(Reference Dowd, Tamminen and Jung80), for these patients as well as patients who are struggling with adherence, dietetic counselling can be beneficial. Furthermore, dietary counselling and follow-up reviews for people with CD have been associated with better GF dietary adherence, resolution of disease specific symptoms, and improved quality of life(Reference Butterworth, Banfield and Iqbal61, Reference Rajpoot, Sharma and Harikrishnan70, Reference Hughey, Ray and Lee81–Reference Pietzak83). The studies, exploring the impact of follow-up reviews, do have methodological weaknesses and as such there remains a paucity of good quality studies.
Mental health conditions such as depression are common among patients with CD(Reference van Hees, Van der Does and Giltay84) and this may have a negative effect on adherence as suggested by a systematic review(Reference Sainsbury and Marques85); however, the quality of the systematic review is limited by the low number of studies included. Psychological traits associated with adherence include greater self-regulation, habit, self-efficacy, priority, facilitation and support, lower psychological distress, lower levels of conflict and fewer self-control lapses(Reference Holmes and Moor25, Reference Sainsbury, Halmos and Knowles72). One qualitative study highlighted that 54 % of people who reported their ethnicity as White (n 21) indicated motivation being a challenge compared with just 33 % of South Asian patients (n 7)(Reference Muhammad, Reeves and Ishaq86).
An important, but potentially modifiable, cause of low adherence is lack of knowledge about gluten-containing foods; studies have reported a positive association between food knowledge and dietary adherence(Reference Sainsbury, Mullan and Sharpe64, Reference Villafuerte-Galvez, Vanga and Dennis65, Reference Leffler, Edwards-George and Dennis67). Coeliac support groups offer practical advice and support to patients with CD, membership of such groups has consistently been associated with good adherence(Reference Butterworth, Banfield and Iqbal61, Reference Leffler, Edwards-George and Dennis67, Reference Silvester, Weiten and Graff87). In Canada, members had better knowledge of GF foods than non-members(Reference Silvester, Weiten and Graff71). However, members are a self-selected group of patients who may exhibit greater motivation to adhere to the GFD, which may in itself represent a confounding variable.
The ability to read and interpret food labels is a key skill patients need to master to enable them to choose appropriate GF foods. Patients who conveyed an understanding of food labels were more likely to adhere to the GFD(Reference Muhammad, Reeves and Ishaq63, Reference Pietzak83, Reference Butterworth, Iqbal and Cooper88). One of our own studies(Reference Muhammad, Reeves and Ishaq63) reported 76 % of South Asian patients agreed with the statement ‘I don't understand what foods I can eat’ and 53 % agreed with ‘I don't understand food labelling’ (n 38), a cause for concern and an area for clinicians to be aware of. There exists a paucity of food labelling in some countries including India with the exception of the main cities, and knowledge about a particular food item is often only a best guess as to whether it is GF or otherwise(Reference See and Murray89, Reference Saturni, Ferretti and Bacchetti90). In addition comprehension of food labels has been reported as being low in India(Reference Vemula, Gavaravarapu and Mendu91). Assessment of health literacy has not been studied in CD but could give valuable insight in to the ability of some patients to adhere to a GFD.
A GFD does also have associated financial costs(Reference Singh and Whelan92, Reference Hanci93) and the perceived affordability of GF foods is associated with dietary adherence(Reference Butterworth, Banfield and Iqbal61, Reference Villafuerte-Galvez, Vanga and Dennis65, Reference Leffler, Edwards-George and Dennis67, Reference Barratt, Leeds and Sanders69). In the UK, GF foods have been available through prescriptions since the 1960s to enable access to GF foods and reduce the financial burden to patients. Receiving GF foods on prescription has been shown to be associated with dietary adherence(Reference Hall, Rubin and Charnock41, Reference Butterworth, Banfield and Iqbal61, Reference Muhammad, Reeves and Ishaq63); however, these studies have not collected data on the amount of GF food received on prescription nor the reasons why patients were not receiving GF foods on prescription, thus there is scope for more detailed studies to be undertaken to explore this area. This research is time critical as the availability of prescribed GF foods is not uniform across the UK, with a general decline in availability over recent years due to the financial pressures within the National Health System, this is despite national guidance that GF foods should be available on prescription(94).
The ability to access GF food for the home, at work and whilst travelling have been reported as barriers to adhering to the GFD(Reference Villafuerte-Galvez, Vanga and Dennis65, Reference Barratt, Leeds and Sanders69, Reference Zarkadas, Cranney and Case95). Qualitative interviews have revealed both South Asian and Caucasian patients found eating out difficult (80 and 86 %, respectively), with the majority of each group indicating a lack of confidence in information from restaurant staff(Reference Muhammad, Reeves and Ishaq86). Surveys within the UK have indicated manufactured GF food staples (such as GF bread, starch or pasta) are rarely stocked in convenience stores, disproportionately increasing the burden of the GFD in socio-economically disadvantaged areas, for people who do not drive, the isolated elderly and those with physical disabilities(Reference Singh and Whelan92). Interviews with South Asians with CD who were not adhering to the GFD revealed 85 % were unable to find GF foods in their local Asian food stores(Reference Muhammad, Reeves and Ishaq86). Cross contamination of foods with gluten was also highlighted as a concern by South Asians(Reference Muhammad, Reeves and Ishaq86), since certain practices in grinding mills may encourage cross-contamination of GF products with gluten, and such starch could reach the UK and be sold in Asian shops(Reference Rajpoot and Makharia96). Studies exploring factors specific to South Asians in other health conditions, that can be extrapolated to adherence to the GFD, and these include dietary counselling not inclusive of specific details of the typical South Asian diet and social responsibilities to continue with a traditional diet(Reference Sohal, Sohal and King-Shier97) for example cultural pressure when visiting family members’ homes or attending celebratory events with ‘feelings of having to live up to cultural expectations of food and eating practises to avoid being alienated’(Reference Patel, Ferrer and Tyrer98).
Causes of low adherence are diverse and affected by many factors, and may even be different for particular ethnic groups. The evidence for different causes of low adherence, as suggested by the studies reported here is limited by the methodologies utilised and because the studies have used unreliable or non-validated instruments including subjective reports by patients about their own perceived adherence or non-adherence.
Interventions to improve adherence with GFD
It has been reported that there is a need to develop resources to help people with CD follow a GFD(Reference Dowd, Tamminen and Jung80). To date, only four well-designed interventions are reported in the literature, with two more underway, which target improving adherence to a GFD in adults with CD and Table 2 shows the details of the studies involving intervention to increase adherence to a GFD. Addolorato et al.(Reference Abenavoli, Proietti and Leggio99) conducted a study of sixty-six CD patients with state anxiety and depression and reported a greater proportion of participants adhered to a GFD after psychological counselling compared with a control group who received no counselling(Reference Addolorato, De Lorenzi and Abenavoli100). This study was conducted in Italy and no mention of the ethnicity of the participants was reported. Sainsbury and colleagues(Reference Sainsbury, Mullan and Sharpe64) devised a web-based intervention to improve dietary adherence to a GFD in adults with CD residing in Australia; the ethnicity of participants was not reported. All patients had biopsy-confirmed CD, and completed the intervention (n 46) or were on a wait list (n 64) and after 3 months completed a validated dietary adherence questionnaire(Reference Leffler, Dennis and Edwards George59). The intervention consisted of six, 30 min, online modules completed over a period of 6 weeks. The modules encompassed education, behaviour change and cognitive behaviour therapy to treat anxiety and depression and improve coping behaviour. This online course demonstrated a significant improvement in GF dietary knowledge and dietary adherence score among participants, and was sustained at 3 months from baseline. The improvements observed were greatest in those not previous adhering to the GFD (Coeliac Disease Adherence Test score <13 at baseline; n 18). The participants were recruited predominately through coeliac societies, of which membership is also associated with dietary adherence(Reference Muhammad, Reeves and Ishaq63), and thus could have introduced selection bias. The study employed web-based methodology and hence computer literacy might well be an issue in a groups of older adults that form a significant proportion of the adult UK population with CD(Reference West, Fleming and Tata101).
Intervention studies to improve gluten-free dietary adherence
CD, coeliac disease; anti-tTG, anti-transglutaminase antibodies; GFD, gluten-free diet.
Rajpoot and colleagues(Reference Rajpoot, Sharma and Harikrishnan70) aim to improve dietary adherence to a GFD in adults with CD through face-to-face nutrition counselling with a dietitian. The 45 min long counselling sessions were attended at 1, 3 and 6 month intervals for each patient and family member from baseline. The study recruited participants through hospital clinics in India, ethnicity of participants was not specified, 146 participants completed the intervention and 6-month follow-up period. Dietary adherence was assessed by participants completing a non-validated questionnaire. Dietary counselling increased the proportion of participants adhering to a GFD over time; adherence rates increased from 64 to 94 % at 6 months in the newly diagnosed group, and from 53 to 92 % in a group of patients with established CD. A limitation of the study was the lack of a control group, thus additional influences could have impacted upon the dietary adherence.
A recent study aimed to evaluate the role of text messages in relation to increasing adherence to the GFD in patients aged 12–24 years with CD living in the USA. The text message group received forty-five unique text messages over a 3-month study period, while the control group received standard care treatment(Reference Haas, Martin and Park102). Adherence was measured with serum anti-tTG and deamidated gliadin peptide levels and no significant difference was noted in either group. The study, however, could be criticised for the over-reliance on serology to detect dietary adherence.
Additional interventions have aimed to increase knowledge of patients about CD in general and GF foods(Reference Meyer, Fasshauer and Nebel103), although these studies did not assess dietary adherence. Associations between better knowledge and adherence have been reported(Reference Halmos, Deng and Knowles79), thus is could be inferred that increasing patients knowledge is likely to improve dietary adherence.
Conclusions
It is clear from the above discussion that a proportion of people with CD are not adhering to a GFD which is compromising their short- and long-term health. A range of methods are available to assess dietary adherence in the clinical and research setting, including serological tests, questionnaires and interviews, however it remains an area where validation studies are still needed. It has been shown that GF dietary adherence is affected by age at diagnosis, the symptoms experienced, dietetic counselling, mental health status, membership of a local support group or society, understanding and knowledge of food labels, the economic cost and availability of GF foods and currently whether GF foods are available on prescription.
Adhering to a GFD has its difficulties and there are a wide range of barriers for patients to overcome. Robust studies are needed to accurately assess the social and economic burden of undertaking a lifelong GFD in ethnically diverse populations of people with CD. Very few intervention studies have been conducted in adults with CD to improve GF dietary adherence. Such interventions have been traditionally based on both dietary and psychological counselling. Increasingly technological based solutions have been adopted with online module based training programmes, apps, text messages and telephone clinics presenting promising results. However further developments in this area would be welcomed.
Future interventions should include people of all ethnicities, with a focus on decreasing barriers to knowledge transfer, increase understanding and enabling behavioural change. These interventions should consider how they can be undertaken in clinical practice and consider cost-effectiveness in the healthcare environment.
Financial Support
None.
Conflict of Interest
The authors have previously received Dr Schär's coeliac disease award for the year 2016.
