A medical food thought to promote synapse formation was found to improve memory test scores in patients with mild Alzheimer’s disease (AD) in a clinical trial presented at the 4th International Conference on Clinical Trials on Alzheimer's Disease (CTAD) in San Diego, California, on Friday, 4 November 2011. The 24-week-long Souvenir II trial extends the findings of an earlier proof-of-concept trial, Souvenir I, which lasted 12 weeks (see ARF related news story and Scheltens et al., 2010). Both trials are headed by Philip Scheltens, of the VU University Medical Center in Amsterdam, The Netherlands, who presented the data at CTAD.

The medical food, called Souvenaid, is a product of Nutricia, the medical foods division of Danone Research (which Americans will know as Dannon). It contains a mix of nutrients, including the omega-3 fatty acid docosahexaenoic acid (DHA), uridine monophosphate (UMP), and choline. All are dietary precursors in biochemical pathways responsible for the synthesis of phospholipids, the main components of cell membranes. “We don’t think Alzheimer’s patients are deficient in these nutrients, but we are adding an extra amount to give a boost to these pathways,” Scheltens told ARF. The formulation increased synapse formation in mouse brains and reduced cognitive deficits in transgenic mouse models of AD.

Souvenir II recruited 259 AD patients with a mean age of 74 at 27 European study centers. To be included in the trial, patients had to have a Mini-Mental State Examination (MMSE) score greater than 20, which is considered mild AD. They also could not be taking any medications or nutritional supplements, except for vitamins. The researchers then randomized participants to one of two groups: In the treatment group, 130 people took the 125 ml Souvenaid drink once a day, and in the placebo group, 129 people took another drink with as many calories but without the synapse-promoting ingredients; 118 people in the treatment group and 120 in the placebo arm completed the study.

Patients were examined at baseline, 12 weeks and 24 weeks for various neuropsychological measures. The primary outcome researchers calculated was a composite score of a handful of tests contained in the neuropsychological test battery (NTB). They included the Rey auditory verbal learning test (RAVLT, which measures immediate recall, delayed recall, and recognition performance) and the Wechsler memory scale (WMS) verbal paired associates test (which measures immediate and delayed recall). This primary outcome, which stresses episodic memory, is not one traditionally used in AD studies; Scheltens said it was chosen because more routine test batteries such as the ADAS-cog are insufficiently sensitive to pick up differences in this mild stage of AD.

There was no significant difference in scores between treatment and placebo at 12 weeks. Both improved somewhat, probably due to a learning effect. But by 24 weeks the Souvenaid group had a more favorable trajectory on the memory score than controls, leading to a Cohen's d effect size of 0.22 points that Scheltens said was significant. There were, however, no differences between drug and placebo on the trial’s secondary outcomes resulting from the total NTB, such as executive function domain, total composite score, and individual item scores. Souvenaid had no significant side effects, Scheltens reported.

Despite these favorable findings, it is too early to know what they mean for AD patients. “There is not enough evidence at this point to make a recommendation to patients,” said Paul Aisen at the University of California, San Diego. Aisen heads the Alzheimer's Disease Cooperative Study, which hosted the conference.

Although Souvenaid showed some benefit in mild AD, it failed to do so in more advanced disease. A parallel clinical trial to Souvenir II recruited 527 people with mild to moderate AD (MMSE scores 14-24) at 48 U.S. clinical centers. Unlike patients in the Souvenir II trial, who were not receiving any medications, these U.S. patients were taking cholinesterase inhibitors and/or memantine. At 24 months, there was no difference between the treatment and placebo group in 11-ADAS-cog scores, which was the primary outcome measure in this trial. “We think this supplement works by building membranes and synapses, but it only works if there are enough neurons there. Patients with more advanced AD may not have a sufficient number of neurons, and that is why we saw no effect,” Scheltens speculated.

Scheltens said the results warrant further studies, and two are ongoing. The group is conducting another trial in moderate AD patients and one in prodromal AD as defined by the “Dubois criteria” (see ARF related news story on Dubois et al., 2007). The prodromal AD study, funded by the European Union, will enroll 300 patients and last 24 months, with some centers extending to 36 months.

Medical foods are prescribed to patients by their doctors, but they are not typically tested as drugs since they do not have to go through the same regulatory process (see ARF news series on medical foods in AD). They have to show safety but not efficacy. Even so, the sponsor decided to mount a clinical trials program in an effort to build support among physicians. “If doctors are going to prescribe Souvenaid, I want them to have all the scientific information,” said Scheltens.

Other clinicians at CTAD were encouraged that there was a program of studies conducted on Souvenaid. They noted that the present trial in drug-naïve mild AD patients seemed as rigorously designed as trials of pharmaceuticals and cleanly met its prespecified primary endpoint. They also cautioned that the primary endpoint indicates an improvement at six months only in episodic memory, not in the much broader swath of symptoms that define AD. Others recalled that previous trials of individual components of this nutrient mixture have been negative (see, e.g., ARF related news story on DHA trial).

Lon Schneider of the University of Southern California, Los Angeles, noted that, because the bulk of preclinical research on the drink’s ingredients was done on cognitive aging rather than AD, a trial in age-related cognitive memory impairment might be informative. Schneider added that a robustness analysis looking at whether the participating centers contributed about equally to the effect on the NTB would be of interest. Other trialists said their confidence in the effect would grow once other evidence supporting Souvenaid’s effect on human synapses became available, such as FDG-PET, functional MRI, or encephalography (EEG).

In general, synaptic activity is getting more attention in AD clinical studies. At CTAD, a number of studies examined the use of event-related potentials as measured by EEG. Such measures of synaptic function could serve as markers of early AD as well as to measure disease progression and even treatment response. “More studies are looking at functional connectivity using functional MRI and EEG,” said Scheltens. “There is an awareness that early in disease, the network gets disrupted, so improving connections is a way for treating the disease.”

Scheltens and colleagues have collected EEG and magnetoencephalogram (MEG) data from a subset of Souvenir II participants; their analysis is underway. Scheltens did not present these results at CTAD, but once available they may provide further understanding on how Souvenaid acts.—Laura Bonetta and Gabrielle Strobel.


  1. Souvenaid represents an innovative approach to providing benefits for people with Alzheimer’s disease. People with Alzheimer’s exhibit a significant loss of brain synapses, and this loss correlates with the loss of cognitive function. Souvenaid is a combination of nutrients designed to improve the formation and functioning of synapses in the brain.

    The Alzheimer’s Association does have some concerns about “medical foods” in general. In contrast to FDA-approved drugs, no pre-market review process exists for medical foods. Instead, they are regulated after they have become available to consumers. However, new products do not always come to the attention of the FDA. Because medical foods do not undergo pre-market FDA review, the Association cannot be assured that data supporting their effectiveness (if such data exist) have undergone scientific scrutiny. In light of the lack of scientific data to substantiate the use of medical foods, the effectiveness of medical foods is difficult for the Association to assess. The Association is pleased that, in this case, the company is pursuing a rigorous research program for this product. Our statement on medical foods in general can be found at the Alzheimer's Association website.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Chicago: Phase 2 News—Therapeutic Breakfast Food?
  2. AD Diagnosis: Time for Biomarkers to Weigh In?
  3. Medical Foods—Food for Thought, But Think Twice
  4. Paper Alert: Negative DHA Trial Fuels Soul-Searching in AD Field

Paper Citations

  1. . Efficacy of a medical food in mild Alzheimer's disease: A randomized, controlled trial. Alzheimers Dement. 2010 Jan;6(1):1-10.e1. PubMed.
  2. . Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria. Lancet Neurol. 2007 Aug;6(8):734-46. PubMed.

Further Reading