Ipsen states in its press release that primary efficacy was met in men and in patients treated for a duration of at least four years. Schwabe Pharmaceuticals took the opportunity of the French study to put a more positive spin on the same data. Without mentioning the negative primary outcome, they claim in a release of their own, which is based on the Ipsen press release, that ginkgo extract protects against AD. “Their statement is inaccurate,” said Bruno Vellas, of the University Hospital Center, Toulouse, France, who is the lead GuidAge researcher. Vellas believes that further analysis is needed. “We were surprised by the data because after the GEM trial we expected a completely negative study,” Vellas told ARF. “But we have some strange positive signals and we need to understand what they mean.” Schwabe, he emphasized, does not have access to the full data. Schwabe heavily advertises Tebonin on German national television and elsewhere.

The strange positive signals come from looking at data from years four to five of the study and from analysis by gender. Among the 2,854 subjects enrolled in GuidAge, men assigned to the ginkgo extract were less likely to develop AD than were men on placebo. Also, of the men or women who stuck to their regimen of 120 mg ginkgo extract twice per day for four years, fewer converted to AD than among those taking placebo for four years. The Ipsen press release states this year four to five analysis was not post-hoc but pre-specified. This may raise a statistician eyebrow or two since it is not clear how the study authors knew in advance that four years would be the tipping point. “It would be very unique to pre-specify a four-year analysis based on adherence,” suggested DeKosky. The GuidAge scientists do not specifically mention this proposed analysis in their 2006 methodology paper (see Vellas et al., 2006). There are other reasons to question this way of sifting through the data, according to DeKosky. Adherence-to-medication analysis of an earlier ginkgo trial also suggested the extract has positive effects (see Dodge et al., 2008), limiting cognitive decline. This is in contrast to intent-to-treat analysis, which includes all randomized patients including dropouts and is considered a more stringent way of analyzing clinical trial data. But as DeKosky, and Curt Furberg of Wake Forest University School of Medicine, Winston-Salem, North Carolina, wrote in an accompanying Neurology editorial (see DeKosky and Furberg, 2008), treating the data in this way “undermines the benefits of randomization by introducing the effects of in-study behavior on the outcome.” There are all kinds of reasons why patients who stick to medication longer do better, noted DeKosky. They could be healthier, wealthier, or have better caregiver support, any of which could modulate progression of underlying disease.

For now, Vellas believes that the GuidAge data means that the ginkgo hypothesis is not dead. At least not yet. “We have mixed results. We need more investigation, perhaps another clinical trial,” he told ARF. In the absence of definitive clinical data, sales of gingko extract are continuing apace.—Tom Fagan.