Foods labeled “gluten-free” line the shelves of special sections in conventional grocery stores and natural food stores alike. But what exactly does this label mean?

Groups such as the Gluten Free Certification Organization, the Celiac Sprue Association, the Canadian Celiac Association and the National Foundation for Celiac Awareness each suggest different guidelines to establish a gluten threshold level.

The U.S. Food and Drug Administration (FDA) does not yet have a rule for gluten-free labeling. It proposed one in 2007 – that foods could be marketed as gluten-free if they contain less than 20 parts per million (ppm), or 20 milligrams per kilogram of gluten. That’s consistent with labeling rules from Codex Alimentarius, which is also used in Europe.

The FDA acknowledges that the term “free” does not necessarily mean that a food contains none of a particular substance. For example, foods labeled as cholesterol-free foods may contain up to 2 mg of cholesterol.

Four years have elapsed since FDA proposed its labeling rules for gluten and a decision on whether to accept them has yet to be made. The agency recently re-opened a 60-day comment period – which ended in October – to get new input on the guidelines. Now the food and restaurant industry and those in the celiac and gluten intolerant community anxiously await the agency’s decision.

Gluten is found in wheat, barley and rye and therefore in foods processed using these grains. The protein components of gluten (gliadins and glutelins) can trigger an autoimmune reaction in some people called celiac disease, the most common food-sensitive intestinal problem. According to the FDA, approximately 1 percent of the American population has the inherited digestive disorder.

For those who can’t tolerate gluten, it can be hard to know what to trust when it comes to processed foods, or even some medicines or vitamins. If a celiac ingests gluten in a high enough concentration, it can cause deterioration of the small intestine when the immune system attacks the villi lining the small intestine. It can also lead to other digestive problems and malabsorption of nutrients.

Registered Dietitian Tricia Thompson from Massachusetts has gotten a lot of questions from consumers about the safety of so-called gluten-free foods. Over time, she says, it has become harder to justify the 20 ppm threshold – which FDA has said is the minimum level at which gluten can be reliably and consistently detected – because many manufacturers are already producing foods with gluten amounts well under that level. It’s now possible to precisely test amounts as minute as 5 ppm, she notes.

Thompson believes that if the FDA lowers the allowable limit of gluten it shouldn’t affect food cost or availability adversely. While she believes that the less-than 20 ppm gluten level is probably safe for the majority of people with celiac disease, she hopes that the FDA will consider lowering the 20 ppm rule – if not to 5, perhaps to 10 or 15 ppm to help ease the minds of those who believe they react to very low levels of gluten.

Peter Olins, a PhD biochemist who runs the website, agrees. In his comment on the FDA recommendation, Olins wrote, “Currently, it appears that a level of 5 ppm gluten can be reliably measured, and we see no benefit in setting a higher threshold.”

Olins also thinks an alternative to the ppm measure would be to instead label the amount of gluten per serving measurement. Such a change would increase the number of foods that can be classified as gluten-free, because of small serving sizes.

“A label such as 20 ppm gluten is really a measure of concentration; however, it seems most likely that the toxicity of gluten is related to total daily quantity ingested, not concentration,” Olins said in his letter.

He notes that people are already used to measuring calories by individual food servings.

While Thompson can see the reasoning behind this, she fears it could prove to be unworkable for manufacturers. According to her research, different batches of the same food can contain different amounts of gluten. If labels gave specific information about the amount of gluten per serving, it could not be trusted to be uniform across batches. Therefore manufacturers would need to produce many different labels for their foods.

One issue confusing the gluten labeling debate is the FDA’s own Health Hazard Assessment,  published in May 2011, which listed Tolerable Daily Intake (TDI) values so low that critics say people with celiac disease will be afraid to eat. The safety assessment found that TDI for gluten to be 0.4 milligrams per day for “adverse effects on the intestinal mucosa” and 0.015 milligrams per day for adverse effects on clinical symptoms.

In a post on Thompson’s website, she calls many of the TDI estimates “ridiculously low” and notes that based on those numbers, people with celiac disease would have to limit themselves to less than a 1 ounce slice of bread containing 20 ppm of gluten each day to prevent intestinal damage, and far less to prevent clinical effects. The FDA also seems to doubt the numbers, and tentatively concluded that “based on the [levels of concern] identified in the safety assessment approach, the Agency should not use that approach in defining ‘gluten-free’ because the estimation of risk to individuals with celiac disease associated with very low exposure levels may be conservative and highly uncertain.”

In her comment to the FDA, Thompson said the regulators should consider “explaining in much more detail why they believe the findings of the Health Hazard Assessment to be conservative and highly uncertain. An explanation may help decrease anxiety among those who are very concerned about the published TDI values.”

“They need to explain to consumers why they don’t have a lot of confidence [in the TDI values],” Thompson says. “They owe it to consumers to decrease fear … so that consumers are confident in the proposed gluten threshold level of less than 20 ppm gluten.”

Olins also wants more information for consumers. On items that are labeled as gluten-free, he would like to see a text box warning that the food may in fact contain trace amounts of gluten.

“This approach would provide the best ‘truth-in-labeling,’ and would satisfy the needs of the consumer/physician/dietitian in selecting appropriate foods,” Olins says in his letter.

In a joint submission to the docket, the Celiac Sprue Association (CSA) commented on the proposed labeling rules, along with the Celiac Disease Foundation, Gluten Intolerance Group of North America, Canadian Celiac Association and the National Foundation for Celiac Awareness. Their recommendation, which is posted on the CSA website, supports a single definition of gluten-free, saying a dual definition of gluten-free and low-gluten would be confusing.

A study published in June by the Journal of the American Dietetic Association, titled “Gluten Contamination of Grains, Seeds, and Flours in the United States,” found that nine out of 22 inherently gluten-free products, such as corn or millet, contained mean levels of gluten varying from 8.5 ppm to 2,925 ppm.  The 13 other products had gluten levels below 5 ppm.

Given those findings, the ADA Journal said, “Gluten contamination of inherently gluten-free grains, seeds, and flours not labeled gluten-free is a legitimate concern. The FDA may want to modify their proposed rule for labeling of food as gluten-free, removing the requirement that gluten-free manufacturers of inherently gluten-free grains, seeds, and flours must state on product labels that all foods of that type are gluten-free.”

Stephen King, press liaison for the FDA’s Center for Food Safety and Applied Nutrition, said in an email that the agency is reviewing the comments and evaluating the information to determine if any changes need to be made to the rule.

At this point, predicting which way the agency is leaning, King says, “would be speculative.”

King says the process is rather involved, but the agency hopes to see the rule published within three months of the comment period close, though that depends on the time it takes to review comments, and whether the rule needs to be edited.