A study that analyzed the interplay of Alzheimer's symptoms and genetic differences has found evidence for six subtypes of the deadly dementia.

The findings are important because they provide evidence that treatments that may work for some patients might not help others with a different genetic profile. This has implications for clinical trials -- the tested drug may look less effective if many study subjects don't have the right kind of dementia -- and could eventually lead to more personalized treatment. Genetic differences are a sign that the biological underpinnings of symptoms may be different and thus could respond to different medications.

"What we are saying is that there is a lot of heterogeneity in Alzheimer's," said Shubhabrata Mukherjee, a research assistant profession in general internal medicine at the University of Washington School of Medicine who led the study. "You cannot label everyone with Alzheimer's disease and say that's it."

It has been frustratingly difficult to find treatments that prevent or treat Alzheimer’s, the most common form of dementia. Until fairly recently, doctors could definitively diagnose the disease only by examining the brains of patients after they had died. New imaging techniques that make it easier to see key brain changes in living patients have improved diagnostic accuracy. Doctors have also learned that many patients have more than one type of dementia. For example, it is common to have Alzheimer’s disease as well as vascular dementia, a type of cognitive decline caused by poor blood circulation in the brain.

The new study was published this month in Molecular Psychiatry. Mukherjee’s team included researchers from several other institutions, including the University of Pittsburgh.

“This is sort of a first step in trying to understand the heterogeneity of Alzheimer’s disease,” said Oscar Lopez, a neurologist who heads the University of Pittsburgh Medical Center’s Alzheimer’s Disease Research Center.

“This is the beginning of a very long process.”

The researchers analyzed data on 4,050 people with late-onset dementia from five previous studies. They divided the patients into six groups based on their symptoms when diagnosed and then looked at whether the genetic profiles of the groups differed. They did.

The patients, who had an average age of 80 and mostly identified themselves as white, received scores on their functioning in four cognitive domains: memory, executive functioning, language and visuospatial. Most, Mukherjee said, were impaired to some degree on all fronts, but some showed more marked problems in a particular area. Patients in the biggest group, 39 percent of the total, were about equally impaired in all domains. Twenty-seven percent were considerably more impaired in memory. The percentages were smaller for the other groups: 13 percent for those whose language scores were exceptionally bad, 12 percent for especially poor visuospatial functioning and only 3 percent for particularly low scores in executive function or decision-making ability. Six percent of the patients had significantly poorer function in two domains.

The team then looked for genetic differences among the groups. One that stood out was that APOE4, a gene well known to greatly increase risk of late-onset Alzheimer’s disease, was considerably more common in the group where memory problems dominated. It was present in all the other groups as well, though. However, there were 33 other genetic markers known as single nucleotide polymorphisms or SNPs that were strongly related to each of the subgroups except the one for executive function. That subgroup was too small for meaningful genetic analysis.

Shubhabrata Mukherjee is a research assistant professor in general internal medicine at the University of Washington School of Medicine.
Courtesy of University of Washington School of Medicine
Shubhabrata Mukherjee is a research assistant professor in general internal medicine at the University of Washington School of Medicine.

Mukherjee said that the work is at a "very initial stage" but the study results provide "strong support for the biological coherence" of their Alzheimer's categories.

Mukherjee, who is a statistician and computer scientist, said he hopes that biological researchers will now study what role the newly identified genes play in the body.