Gastrointestinal Symptoms Following Olestra Consumption

To the Editor.—Procter & Gamble's movie theater study¹ does a good job of answering the wrong questions. It fails to demonstrate a significantly increased incidence of gastrointestinal (GI) effects from a one-time, variable olestra exposure. But it sheds little light on what will happen when large numbers of Americans consume olestra over prolonged periods. Also, JAMA's "Editor's Note" was incorrect in stating that "controlled studies [on olestra's GI effects] have not been reported."²

In fact, several well-designed clinical trials reported to the Food and Drug Administration as part of the olestra approval process and published in the Federal Register³ show significant increases in GI symptoms during 5 to 56 days of daily olestra consumption. In 2 trials sponsored by Procter & Gamble, groups of 17 to 24 subjects fed 20 or 32 g of olestra (equivalent to about 56 g [2 oz] or 98 g [3.5 oz] of potato chips) daily for 8 weeks showed significant dose-related increases of diarrhea, loose stools, number of subjects experiencing 1 or more severe GI symptoms, and other GI effects.³ A clinical rechallenge study with 52 subjects confirmed those results, showing significant increases in GI symptoms and severity following 7 days of olestra consumption at 20 g/d.⁴ Similarly, an "oil loss study," with 1228 subjects consuming 34 g of olestra daily for 5 days, found significant increases in adverse GI effects.⁵

The statistical power of the movie theater study apparently was inadequate to detect GI effects following a single average dose of 17.5 g of olestra. With an incidence of "any GI event" of about 15%, 550 subjects in each group would have provided only about a 50% probability of detecting a 5% actual increase in the treatment group.

The 2 end points of greatest concern to the Food and Drug Administration based on the clinical trials—diarrhea and loose stools—were increased less than 1% in the olestra group over baseline levels of 2.6% and 1.1%, respectively. Maintaining 80% power to detect a 1% increase over a 2% baseline requires about 4000 subjects per group. If the true incidence of diarrhea and loose stools from 1 olestra exposure was 1% (Frito-Lay acknowledges that "roughly 2%" of people eating olestra snacks experience GI effects),⁶ then national marketing of olestra snacks would lead to hundreds of thousands of extra cases of those effects annually. Failure to detect the true incidence of GI effects following a single olestra exposure has more to do with lack of power than the absence of real effects and provides little assurance of safety.

Other methodological problems with this study include potential exposure misclassification (some "olestra eaters" may have eaten few or none of their chips) and delay of up to 10 days to assess symptoms while background rates increased.

The authors assert that their study does not support the notice required by the Food and Drug Administration on olestra products stating: "Olestra may cause loose stools and abdominal cramping." In fact, their study does not negate the earlier, better studies demonstrating that olestra causes such symptoms. Considering that national marketing of olestra products will likely cause widespread adverse effects—GI symptoms as well as carotenoid losses—in large numbers of consumers, a candid label notice, if not rescission of the additive's approval, is the minimum prudent response.