- Top of page
- Materials and methods
- Conflict of interest, source of funding and authorship
Coeliac disease (CD) is unique amongst chronic disorders in that diet is the only recognised treatment. It is expected that, after the initiation of a gluten-free diet (GFD), the enteropathy will improve (Fasano & Catassi, 2001; Williamson & Marsh, 2002; Pietzak, 2005) and the restoration of absorptive surface area will enable the normal absorption of nutrients to occur. Strict compliance to a GFD is considered to be an essential part of patient management and is advocated in all patients with CD (Hill et al., 2005). The GFD taught in Australia is a ‘no detectable gluten’ standard (Shepherd & Gibson, 2006), comprising of foods that have a gluten content of <3 p.p.m. (i.e. gluten not detectable in currently available assays; The Coeliac Society of Australia, 2011).
Any restrictive diet is inherently at risk of nutritional inadequacy (Thompson et al., 2005). Although nutritional targets from the foods consumed by the Australian population have been set by governmental bodies [Commonwealth Department of Health & Ageing, 2006; National Health & Medical Research Council (NHMRC), 1991], no assessment has been made regarding the adequacy of the GFD, apart from a preliminary report of fibre intake in 40 patients (Tatnell et al., 1985). Previous studies in other countries have generally (Bode et al., 1991; McFarlane et al., 1995; Thompson, 1999, 2000; Ciclitira et al., 2001; Grehn et al., 2001; Hallert et al., 2002; Thompson, 2005; Zarkadas & Case, 2005; Hopman et al., 2006; Kinsey, 2008; Niewinski, 2008; Lee et al., 2009; Wild et al., 2010) but not always (Robins et al., 2008) found the GFD to be nutritionally inadequate. For health professionals educating patients with CD on the GFD, it is vital to know whether people can meet nutritional targets with a GFD, how that ability compares with the general population, and what implication any inadequacies may have on the health of people with CD.
The present study aimed to examine the nutritional adequacy of the ‘no detectable gluten’ diet in people with CD. Accordingly, three strategies were employed. First, the intake of nutrients by patients with at least 2 years experience of the GFD was examined. Second, the nutrient intake of a cohort of patients recently educated in the GFD by a dietitian with expertise in CD was assessed and compared with that of experienced patients. Finally, the effect of a diagnosis of CD and change to a GFD on nutritional adequacy of the diet was examined.
- Top of page
- Materials and methods
- Conflict of interest, source of funding and authorship
The present study identified that a complex relationship existed between the dietary habits of the individual, the characteristics of the GFD and the nutritional adequacy of what is being consumed. Many of the inadequacies of dietary intake, such as a deficient intake of fibre and folate, may originate in the food choice of the individual, whereas some deficiencies of intake, such as thiamin, appear to be coeliac-specific.
Best-practice methods were applied to obtain and analyse dietary data; however, they do have limitations. First, potential recording errors including inaccurate estimates of portions consumed and omission of foods (deliberate or unintentional) can result in an underestimation of nutritional intake. Behavioural changes altering intake have also been reported in patients who are keeping food records (Mela & Aaron, 1997). However, there is no assessment tool that is without limitation or specifically designed for the adult population with CD. Second, comparing the results from two different methods for the quantification of dietary intake introduces potential inaccuracies. The use of prospective food diary entries (after 12 months of GFD) and comprehensive dietary history (used to obtain prediagnosis data) was unavoidable because it would be unethical to request patients to continue on a gluten-containing diet and record intake in a prospective food diary after the diagnosis of CD. However, the accuracy of comprehensive dietary history has been validated (Van Staveren & Ocke, 1990). Third, it is possible that data entered for new products were associated with error because these relied upon information provided by food manufacturers. Finally, patient characteristics might introduce bias to the results. For the newly-diagnosed group, selection bias was minimised by recruiting patients who were consecutively referred; however, follow-up was more intense than usual as a result of their involvement in a prospective clinical study. Indeed, all patients were adherent to the diet. For the diet-experienced patients, the act of volunteering for the study and the prerequisite strict adherence to the GFD might be indicative of more motivated patients who may be more likely to adopt a more nutritious GFD than people not in the study. Data from both patient groups probably represent best-case scenarios. The observation that few differences in nutrient intake were observed between those experienced in the GFD, in whom gluten-free food habits have been formed over median 6 years of unsupervised free choice, and those who were taught the diet within a prospective study was reassuring with respect to these potentially confounding factors not having a significant impact upon the findings.
With such limitations in mind, the results gave clear indications of excesses and inadequacies. Nutritional adequacy data can be accurately and validly extrapolated from the data obtained in the present study because under- and over-reporting were not evident. Reported energy intake exceeded estimated energy expenditure for most patient groups. This is consistent with the weight gain observed in newly-diagnosed patients during the prospective follow-up as reported in the World Health Organization (2003); for example, the mean weight gain over 12 months was 2.7 kg in women and 3.6 kg in men, equating to an excess energy consumed per day of 250 kJ and 335 kJ, respectively. Saturated fat consumption was also well above the AMDR recommendation for patients in both cohorts. This is not a coeliac-specific phenomenon; it was similar to the intake of the Australian population and has been associated with chronic disease (Gross et al., 2004). Because the dietary intake data were not analysed according to the actual foods consumed, sources of saturated fat were not identified.
Summary data did suggest that people eating a GFD in Australia generally achieve population (EAR) nutritional targets (Table 1). The exceptions are fibre and folate for women, although this is not different from the general Australian population. Interestingly, the majority met nutritional targets for nutrient density. By contrast, the major proportion of patients did not meet their RDI or AI requirements (Table 3). Therefore, it is questionable that nutrient density equates to sufficient daily intakes of RDI or AI.
The findings in both cohorts were highly consistent. Inadequacies could be classified into three groups. First, the inadequate intake of magnesium was the only nutrient that appeared to reflect that of the population in general. Although whole grain wheat is a good source of magnesium, the amount is minimal in refined wheat products commonly consumed in the community (Gross et al., 2004). Other good sources of magnesium, such as leafy green vegetables, legumes, meat and seafood, are available to both gluten-free and gluten containing diets. This may explain why the trends are similar in both groups.
Second, inadequacies in fibre, folate, calcium, iron (women) and zinc (men) were frequently present before diagnosis by as much as 12 months on the GFD. The frequency of poor intake with untreated CD might be a manifestation of being unwell per se, although almost identical prevalence of such deficiencies were observed in patients who had been well on the GFD for at least 2 years. Such inadequacies may reflect the effect of deficiencies specific to gluten-free foods. There are reasons why the GFD might additionally compromise the adequate intake of some nutrients, such as folate, for example, because very few gluten-free breakfast cereals are fortified with folate compared to the common practice of fortification in wheat-based cereals [Food Standards Australia & New Zealand (FSANZ), 2008].
Third, inadequacies of the intake of thiamin and of vitamin A in women only were observed more commonly with the GFD than in the diet before diagnosis. The lack of availability of gluten-free foods fortified with thiamin may account for the reduced intake because, in Australia, only fortification of thiamin in wheat flour for bread making is mandatory [Food Standards Australia & New Zealand (FSANZ), 2008]. The intake of vitamin A may have been overestimated when reporting prediagnosis diet as a result of generalisations about the consumption of vitamin A-rich ‘orange-coloured vegetables’. However, this was less likely because it was relevant only to women and was not observed for any other nutrient. An inadequacy of vitamin A intake in association with the GFD has not been reported previously.
Because gluten-containing cereals are a major source of dietary fibre, patients with CD may be at higher risk of consuming an inadequate fibre intake, which is consistent with other studies (Thompson, 1999; Thompson et al., 2005; Wild et al., 2010). Compounding this is the potential that gluten-free foods contain less fibre than gluten-containing equivalent foods; many are prepared from refined maize flour and white rice, which are lower in fibre (0.5 g and 0.8 g per 100 g, respectively) than wheat (3.8 g per 100 g) and brown rice (3.2 g per 100 g). The most commonly consumed bread in the present study was the most readily available white loaf containing 3.3 g per 100 g of fibre. Wholegrain wheat-based breads are readily available and typically contain between 4.8–7.8 g per 100 g. Furthermore, it is likely that the volume of breads consumed will be less with the GFD as a result of their inferior texture. By contrast, no similar explanations can be mounted for deficient intake of calcium or of iron in women because the best dietary sources of both are gluten-free. Whether CD specifically reduces meat-seeking eating habits has not been investigated.
The ‘no detectable gluten’ GFD (<3 p.p.m.) adopted in Australia lies between the very strict zero gluten tolerance approach (as adapted by some groups in the USA) and the more liberal Codex Alimentarius Standard (Codex Alimentarus, 2008) adopted by many European countries and Canada. It is likely that the risk of nutritional adequacies will parallel the strictness of the dietary approach in line with variety of food available to individuals, although this notion has not been formally tested. For example, the Australian GFD permits fruit and nut bars that use wheat-derived glucose syrup as an ingredient, and these are good sources of fibre, vitamins and minerals. By contrast, the more liberal guidelines of the Codex Alimentarius Standard permits up to 20 p.p.m. of gluten in a food that ‘is prepared under good manufacturing conditions aimed at achieving the lowest possible levels of gluten resulting from cross contamination’ (Health Canada, 2012). Coeliac Australia does not recommend consumption of foods that indicate on the package that they are at risk of cross-contamination (e.g. cereals, nuts and dried beans), therefore limiting the availability of such nutrient dense foods. Additionally, the Australian GFD does not permit the consumption of any oats (even pure uncontaminated), which may further impact on ability to achieve nutritional adequacy because oats contain good amounts of fibre, folate and minerals.
The prevalence of observed dietary nutritional inadequacy in GFD-compliant patients has three important implications. First, it emphasises a key element of the delivery of dietetic education to patients with CD in that it should not only involve the teachings of how to choose gluten-free foods, but also emphasise the importance of nutritional adequacy over the long-term. The dietitian should encourage intake of nutrient-dense foods including wholegrain foods, enriched where possible, legumes, fruits, vegetables, lean meat, fish, chicken and eggs. The long-term use of micronutrient supplements should not be prioritised over achieving nutritional adequacy from dietary intake alone. Whether special dietary recommendations (nutritional targets) should be set, such as the calcium recommendations set for CD and osteoporosis in the UK (British Society of Gastroenterology, 2007) is not assessable on the basis of the present study because it was not designed to address the relationship between the effect of dietary nutritional adequacy without the impact of supplementation. Second, it highlights the need to promote the nutritional risks of the GFD identified in the present study. The CD-specific nutritional inadequacy of thiamin, particularly, and the high risk of inadequate folate intake should be addressed. Finally, the implications for the food manufacturing industry are clear with regard especially to the development of recommendations for the micronutrient fortification with, for example, thiamin and folate of gluten-free foods in consultation with key stakeholders. Two major studies undertaken in America found that many gluten-free foods contained significantly less thiamin, riboflavin, niacin (Zarkadas & Case, 2005), iron and folate (Thompson, 1999) compared to their fortified gluten-containing comparative products. The Codex Standard indicates that gluten-free foods that are dietary staples (e.g. flour, bread) should supply approximately the same amount of vitamins and minerals as the food they replace (Case, 2005), although there are no regulations governing this in many countries such as Australia.
In conclusion, patients consuming a GFD are at risk of nutritional inadequacies and excesses. Some may be a result of general community eating habits, some to preexisting individual eating habits, and some specifically to the GFD, although such attributions require evaluation in further specifically targeted investigations. Nevertheless, it is reasonable that all aspects, and not just how to avoid gluten, should be addressed by those educating patients on the GFD and further attention be paid to the fortification of gluten-free foods, so that they at least match the micronutrient content of the foods they replace.