Are shared fryers really off-limits for people with celiac disease?

By Dr. Liji Thomas, MDReviewed by Lauren HardakerOct 22 2026

French fries and tortilla chips cooked alongside gluten-containing foods may not always pose the risk previously assumed, as new research finds that most samples contained low gluten levels, with fryer oil changes emerging as a potentially important factor.

Study: Shared Fryers May Not Always Be as Risky for Gluten Cross-Contact in Celiac Disease as Dietary Guidance Suggests. Image credit: siamionau pavel/Shutterstick.com

People with celiac disease are commonly advised to avoid foods prepared in fryers also used for gluten-containing foods. However, a recent study published in Nutrients found that most French fries and tortilla chips prepared in shared fryers contained gluten levels below the US regulatory threshold for gluten-free labeling, suggesting that contamination risk may be lower than commonly assumed in some food-service settings.

Why shared fryers raise concerns about gluten contamination

Celiac disease is an immune-mediated disorder in which consuming gluten, a protein found in wheat, barley, and rye, can damage the small intestine. A strict gluten-free diet remains the primary treatment, but while food labels help people identify products containing gluten, avoiding accidental exposure can be much harder, especially when eating outside the home.

Restaurants present a particular challenge because naturally gluten-free foods may be prepared using the same equipment as gluten-containing products. Deep fryers are a common concern, as batter and breading can contaminate cooking oil and potentially transfer gluten to other foods. Consequently, dietary guidelines often recommend avoiding French fries, tortilla chips, and other gluten-free foods prepared in shared fryers, even though evidence of clinically significant gluten transfer under routine restaurant conditions remains limited.

Much of this precautionary advice is based on clinical experience, expert consensus, and regulatory standards intended to minimize gluten exposure. In the United States, the Food and Drug Administration (FDA) requires foods labeled gluten-free to contain less than 20 parts per million (ppm) of gluten. However, this threshold reflects population-level estimates of gluten tolerance and cumulative exposure rather than a definitive level at which gluten becomes harmful. Dietary recommendations have nevertheless tended to classify food-preparation practices as either safe or unsafe, without accounting for differences in the amount of gluten transferred.

Although this cautious approach is intended to protect people with celiac disease, unnecessarily restrictive advice may increase dietary anxiety, limit food choices, and reduce quality of life. The clinical significance of gluten exposure depends not only on its concentration in food but also on the quantity consumed, how frequently exposure occurs, and the total amount ingested over time. Research suggests that sustained consumption of as little as 10 mg of gluten per day can cause intestinal villous injury in some individuals, with higher daily doses associated with persistent intestinal damage.

Against this background, the researchers sought to establish how much gluten actually transfers to naturally gluten-free foods cooked in shared fryers. By testing French fries and tortilla chips prepared under routine restaurant conditions, the study aimed to provide real-world evidence to inform dietary guidance rather than relying solely on laboratory simulations or precautionary assumptions.

Researchers test 98 shared-fryer foods across 14 states

The researchers conducted a cross-sectional observational study involving French fries or tortilla chips prepared in shared fryers. The samples included 55 French fries and 43 tortilla chip samples, collected from food-service establishments in 14 US states between September 2025 and January 2026.

The samples came from a range of food-service establishments, including fast-food outlets, casual dining restaurants, cafés, and fine dining restaurants. The investigators collected data on fryer oil replacement frequency: daily, weekly, or unknown.

Gluten concentrations were measured using an enzyme-linked immunosorbent assay (ELISA), with <20 ppm as the reference point. The assay’s limit of quantification (LOQ) was 5 ppm, with results below this limit reported as <5 ppm. Thus, the exposure groups were <5 ppm, 5 to <20 ppm, and ≥20 ppm.

The measured values were interpreted in terms of the potential dietary exposure, based on established dose-response data.

Nearly 86% of samples contain minimal measurable gluten

Most samples contained little or no quantifiable gluten. Approximately 84 samples contained less than 5 ppm gluten, eight contained 5-<20 ppm, and six contained ≥20 ppm.

Thus, six of the 98 tested samples exceeded the 20 ppm threshold, although the relatively small, non-random sample means this proportion cannot be assumed to represent contamination risk across all shared fryers.

The researchers emphasize that the 20 ppm cutoff is a regulatory labeling standard, not a guarantee of clinical safety. The potential health effects of gluten exposure also depend on portion size, frequency of consumption, cumulative intake, and individual susceptibility.

Food type

All tortilla chip samples were below the LOQ. In comparison, six of the 55 French fries samples contained at least 20 ppm. Gluten concentrations in these samples ranged from 29.6 ppm to >80 ppm, with three samples exceeding the assay's upper quantification limit of 80 ppm. The authors interpret this to indicate that the quantity of food, frequency of consumption, and cumulative exposure all contribute to the final clinical interpretation.

The difference between French fries and tortilla chips was statistically significant. However, the study could not determine whether this reflected differences between the foods themselves or variations in preparation practices and fryer conditions.

Oil change frequency

Samples were collected from 39 establishments reporting daily oil changes, 35 with weekly oil changes, and 24 with unknown oil-change frequency. No samples from outlets with daily oil changes reached or exceeded the 20 ppm threshold, compared with four from those with weekly oil changes and two from those with unknown oil-change frequency.

The difference between daily and weekly oil replacement was statistically significant, although the small number of contaminated samples and observational design limit conclusions about whether more frequent oil replacement directly reduces gluten contamination.

Type of food service outlet

Samples were collected from 44 casual dining restaurants, 19 franchise restaurants, 15 fast casual restaurants, 12 fast food outlets, two cafes, two fine dining restaurants, and four other types of outlets.

Most samples had gluten levels below 5 ppm regardless of outlet type, except for those from fine dining restaurants. Of the few samples at or above 20 ppm, outlet type showed no clear pattern. For instance, among fast-food outlets, 9 of 12 were <5 ppm, and only one reached or exceeded the 20 ppm threshold. Among fast-casual restaurants, 13 of 15 samples were below the 5 ppm threshold, and only one reached or exceeded 20 ppm.

These results suggest that shared fryers can cause measurable gluten contamination, but it is uncommon in the sampled establishments, with concentrations varying by food type and food-service conditions.

The findings support the importance of restaurant practices. Four of the six samples reaching or exceeding the 20 ppm threshold came from establishments reporting weekly oil changes, while the remaining two came from establishments with unknown oil-change frequency. None came from establishments reporting daily oil changes. However, this is an observed association and does not support a causal inference.

The authors note that even among people with self-reported high adherence rates to a gluten-free (GF) diet, measurable ingestion of gluten occurs frequently. This could result from cumulative exposures to GF-labeled foods that contain gluten near the 20 ppm threshold, cross-contamination in shared kitchens, and eating outside.

Compared to a binary safe/unsafe classification, dose-response data might help patients and healthcare professionals to better understand how measured gluten concentrations correspond to clinical outcomes. Although gluten exposure from multiple foods can accumulate over the day, the authors note that consuming customary portions of foods meeting the gluten-free labeling standard would generally be expected to remain within accepted exposure ranges. Repeated low-level exposures may nevertheless be clinically relevant for susceptible individuals.

To illustrate the importance of portion size, the researchers calculated that consuming 500 g of food containing 20 ppm gluten would correspond to approximately 10 mg of gluten. However, this theoretical example does not establish an individual clinical safety threshold. For example, gluten-free pasta cooked in water previously used for gluten-containing pasta has been shown in earlier research to contain gluten concentrations above 20 ppm.

Single restaurant visits leave fryer variability unexplored

The study was observational, and the researchers measured gluten concentrations in individual food samples rather than assessing intestinal injury, symptoms, or long-term outcomes in people with celiac disease. Moreover, restaurants were not selected randomly, limiting generalizability. Fryer management information was based on reported practices that were not independently verified. The study collected only one sample per restaurant on a single occasion, ignoring variation within the same facility and fryer.

The study did not measure other fryer-related practices, including filtration type, temperature, and cleaning practices. The study cannot establish a precise contamination threshold for every type of fryer, food, or cooking process.

Gluten exposure assessment could improve celiac dietary guidance

The findings indicate that gluten cross-contact is associated with a risk of contamination under some circumstances, although most tested samples contained less than 20 ppm gluten. The study does not establish that all shared-fryer foods are safe for people with celiac disease.

Rather, gluten concentrations varied according to food type and food-service conditions, with lower contamination levels observed in establishments reporting daily oil changes than in those reporting weekly changes. The authors suggest discussing cumulative exposures from routine daily eating compared with infrequent, higher-level exposures from eating out in certain circumstances.

Based on these findings, blanket advice to avoid every food prepared in a shared fryer may be more restrictive than the available evidence warrants.

Applying this evidence-informed perspective in CeD care may help clinicians and dietitians address patient concerns, reduce unnecessary dietary burden, and support long-term adherence to the GF diet.

Future studies integrating measurements of dietary exposure and markers of intestinal injury would better elucidate the relationship between quantified gluten exposure and long-term nutritional impacts.