BudWiederman, MD, MA, Evidence eMended Editor, Grand Rounds
I love to read a well-designed and reported study, and an even bigger treat is if
I learn something new about disease pathophysiology and some general trivia to boot.
Here's an article that fulfills this trifecta for me.
This was a randomized, placebo-controlled, double-blind noninferiority crossover study
of oats in children with typical celiac disease, suggesting that oat ingestion does
not cause an increase in intestinal symptoms in this patient population.
Look at the adjectives in the first part of that sentence. It has 5 important concepts
in study design. First, it was randomized, and in fact, the individuals generating
the randomization sequence had no other role in the study, helping to ensure blinding
of those directly involved with the study subjects. Second, it was placebo-controlled,
meaning 1 group received oats and the other a placebo. (Actually, both groups received
both oats and placebo, that's the fifth point below.)
The third important feature is the double-blinding; neither investigators nor patients
and family knew which treatment arm they had been assigned to. Beyond just concealing
the randomization scheme, the oats and placebo had to be indistinguishable from one
another, e.g. children should be unable to determine when they were eating oats just
by taste or other feature of the meal. This apparently was the case. Fourth, this
was a noninferiority trial, which I've discussed on a few occasions in the past. It's a method that allows for a smaller study size
and, while not achieving a large enough sample to detect a statistically significant
difference in treatments, if such exists, it does assure that enough patients are
enrolled to state that, if a difference exists, it is too small to be noteworthy.
Thus, a particular treatment is "noninferior" to another, at least within a predetermined
The fifth element, crossover design, also has been mentioned before. This is a very useful study design technique when studying treatments of chronic
diseases that do not progress (or regress) quickly over time. Here, 1 treatment group
received a gluten-free diet plus oats added in for 6 months, then underwent a "washout"
period of 3 months on a gluten-free diet followed by a crossover to a gluten-free
diet without oats (the placebo arm) for another 6 months. Everyone was blinded to
the type of diet being administered, except during the washout period. The other treatment
group did the reverse: first gluten-free diet without oats (placebo), then washout,
then gluten-free diet plus oats.
Bottom line: based on measurement of clinical symptoms, serologic data, and a noninvasive
test of mucosal integrity (it was considered unethical to perform repeated intestinal
biopsies of children for research purposes), the investigators found that ingestion
of oats combined with a gluten-free diet was "noninferior" to a gluten-free diet with
no oats. However, it could be problematic implementing the addition of oats to diets
of celiac disease patients, because it's hard to be sure that oat products aren't
contaminated with other grains during processing.
The trifecta I mentioned came with the Discussion section of the article, where I
learned a great deal about differences in gluten based on the taxonomic classification
of cereals and the structure of different grain structural proteins. It provided a
biologic basis for why oat ingestion might be safe for celiac disease patients to
ingest, as well as some new trivia tidbits to bring up at the next neighborhood cookout
Also, please pardon my Mairzy Doats reference in the title, I think I'm going overboard in my new grandfather role.