A total of 34 patients (aged 29-59 years, 4 men) completed the study as per protocol. Overall, 56% had human leukocyte antigen (HLA)-DQ2 and/or HLA-DQ8. Adherence to diet and supplements was very high. Of 19 patients (68%) in the gluten group, 13 reported that symptoms were not adequately controlled compared with 6 of 15 (40%) on placebo (P=; generalized estimating equation). On a visual analog scale, patients were significantly worse with gluten within 1 week for overall symptoms (P=), pain (P=), bloating (P=), satisfaction with stool consistency (P=), and tiredness (P=). Anti-gliadin antibodies were not induced. There were no significant changes in fecal lactoferrin, levels of celiac antibodies, highly sensitive C-reactive protein, or intestinal permeability. There were no differences in any end point in individuals with or without DQ2/DQ8.
The journalist Peter Dizikes, writing in The Boston Globe in 2008, notes that popular culture likes the idea of the butterfly effect, but gets it wrong. Whereas Lorenz suggested correctly with his butterfly metaphor that predictability "is inherently limited", popular culture supposes that each event can be explained by finding the small reasons that caused it. Dizikes explains: "It speaks to our larger expectation that the world should be comprehensible – that everything happens for a reason, and that we can pinpoint all those reasons, however small they may be. But nature itself defies this expectation."