The reclassification could change Alzheimer's research, diagnosis and approaches to treatment, according to the researchers, whose study was published on Monday in the journal Nature Medicine.
"Through these data we are saying that perhaps this is a genetic form of this disease, not merely a risk factor indication," study co-author Sterling Johnson of the University of Wisconsin's Alzheimer's Disease Research Center told reporters in a briefing on the study.
Scientists have known for three decades that people with two copies of APOE4 gene variant have significantly higher risk of developing the disease than people with the most common version of the APOE gene, known as APOE3.
About two to three per cent of the general population, or about 15 per cent of people with Alzheimer's, have two copies of the APOE4 variant.
"This study adds compelling data to suggest that people with two copies of this gene are almost guaranteed to develop Alzheimer's if they live long enough, and that they will develop Alzheimer's earlier than people without this gene," said professor Tara Spires-Jones, an Alzheimer's researcher at the University of Edinburgh who was not involved in the study.
Doctor Juan Fortea of the University of Barcelona and colleagues studied more than 3000 donated brains from the US National Alzheimer's Coordinating Center, as well as biological and clinical data on more than 10,000 individuals from five datasets in three countries.
They found that by the age of 65, at least 95 per cent of people with two copies of APOE4 - known as homozygotes - had abnormal levels of an Alzheimer's-related protein called beta amyloid in their spinal fluid, and 75 per cent had positive brain scans for amyloid.
Nearly all APOE4 homozygotes in the study had higher levels of amyloid at age 65 than people who did not carry the risk variant.
The findings suggest APOE4 homozygotes meet the three main criteria for being a genetic disease: nearly everyone with these two gene variants have Alzheimer's biology; they develop symptoms at about the same rate; and clinical and biological changes occur in a predictable sequence, the researchers said.
Professor David Curtis of the UCL Genetics Institute, who was not involved in the research, remained unconvinced. "I do not see anything in this paper to justify the claim that carrying two copies of APOE4 represents some 'distinct genetic form' of Alzheimer's disease'," he said in a statement.
"No matter how many (copies) of APOE4 one carries the underlying disease processes seem similar across cases of Alzheimer's disease," he said.
The findings could have implications for the recently approved Alzheimer's treatment Leqembi from Eisai and Biogen, a drug that removes amyloid from the brain.
In clinical trials, patients with two copies of the APOE4 variant have much higher rates of brain bleeding and swelling associated with the treatment. Because of this, some centres do not treat these patients, Dr Reisa Sperling, an Alzheimer's researcher at Mass General Brigham who worked on the study, said in a briefing with reporters.
"This research really suggests that we should be treating them quite early at a younger age and at an early stage of pathology because we know they're very, very likely to progress to impairment quickly," she said.
Spires-Jones said the work highlights the need for more research on how genes affect Alzheimer's disease as people age. Other genetic forms of Alzheimer's include Autosomal-dominant Alzheimer's Disease, which is caused by mutations in two different genes, and Down Syndrome.
Most of the study population involved people of European ancestry, and the team noted that more work needs to be done in people of African descent, a population in which APOE4 appears to convey a lower risk of Alzheimer's disease.
