Gluten-Free, Casein-Free Diet and Autism Ranking:

Status Research

Studies

There are a number of limitations to all of the research studies published to date. For example

• The majority of studies (12) used single-case designs (such as case studies, ABAB or pre-test post-test designs) and six of these reported on a single participant.
• Three of the randomised controlled trials (Navarro et al, 2015; Hyman et al, 2016, Elder, 2006) had fewer than 20 participants. One of these studies (Navarro et al, 2005) had only six participants.
• One of the studies (Pusponegoro et al, 2007) had very high levels of participant drop out (almost 30%), meaning the results could be more open to bias.
• Some of the studies (such as such as Kinivsberg et al, 2002) did not provide enough details about the participants (such as whether they had a formal diagnosis of autism), meaning it is not clear if the diet would work for other people diagnosed with autism. 
• Some of the studies (such as Knivsberg et al, 2003) restricted the participants to those with unusual biochemical profiles (such as those with abnormal urinary peptides) meaning that the results cannot be generalised to autistic people who do not have those unusual profiles. 
• Some studies (such as Elder et al, 2006) did not restrict the participants to those with unusual biochemical profiles meaning (according to some supporters of the diet) that the diet could not have worked, since it is only designed to work on individuals with those profiles. 
• One of the controlled studies (Nazni et al 2008) was non-randomised and non-blinded, meaning the results could be more open to bias.
• One of the studies (Pusponegoro et al 2007) lasted no more than a week, meaning that there may not have been time for any effects of the diet to appear. 
• Some of the studies (such as Cade et al 1999) lasted for more than a year, mearing that any effects could have been caused by a range of other factors, including the normal development of the participants. 
• Some of the studies which used a crossover design (such as Elder et al, 2006) did not include a wash-out period. This means that those participants who received the placebo after the diet could have still been showing the effects of the diet when they received the placebo.
• Some of the studies (such as Reichelt, 1990) provided no information on how strictly the diet was implemented. Some studies (such as Johnson et al, 2011) reported that some of the participants struggled to adhere to the diet. In both cases, this could mean that the outcomes reported may not be valid.
• Some of the studies (such as Lucarelli et al, 1995) eliminated other, unspecified foodstuffs from the diet and some of the studies (such as Reichelt, 1990) included medications in the treatment, meaning it is difficult to know what caused any of the reported effects.
• Some of the studies (such as Cade et al, 2000) incorporated non-standard assessment tools used by lay raters (such as parents) meaning the reported outcomes could be less valid.
• Some of the studies (such as Whitely et al, 1999) had incomplete or partial outcome data (such as ratings by teachers) meaning it is difficult to interpret those data accurately.
• Three groups of researchers each published two different papers reporting on the study that they had carried out (Elder 2007 and Seung et al 2007; Knivsberg et al 2002 and Knivsberg et al 2003; Whitley et al, 2010 and Pedersen et al, 2014), meaning that care has to be taken not to count the results twice.

For a comprehensive list of potential flaws in research studies, please see ‘Why some autism research studies are flawed’

Scientific Reviews

There have been a number of scientific reviews of the gluten-free, casein-free diet as an intervention for autistic individuals. The majority of these have concluded that there is insufficient evidence to determine if the diet is effective for autistic people. For example, Mari-Bauset et al (2014) noted that 

“Methodologic limitations identified were associated with a range of factors: the lack of a control group and/or clear definitions of inclusion criteria, very small sample sizes, and analysis being based on single individuals or anecdotal information, groups being heterogeneous in terms of age, failure to control for phenotypic variability between individuals, interventions being of variable duration and generally short, as well as lack of preintervention postintervention comparisons. There was also a risk of bias in data on the behavioral variables attributable to memories of parents and other caregivers being distorted over time and that their perception of changes in the behavior of participants can be subjectively influenced by the fact of being included in nonblinded trials. Similarly, a placebo effect could have had an impact on the results. Lastly, alternative explanations were not always considered, such as the risk of confounding bias, in particular, it being possible that behavioral improvements were due to ongoing development and behavioral therapy given, rather than to gluten-free, casein-free diets per se”.

Updated

16 Jun 2022

Last Review

01 Aug 2017

Next Review

01 Nov 2023