As the incidence of Alzheimer's disease rises and genetic risk factors become better understood, patients are increasingly requesting genetic testing for risk assessment. Such testing is generally not recommended, since the results usually cannot definitively determine whether or not someone will develop AD, and because there are no clinically proven interventions that can modify the risk. However, in some cases, it may be reasonable to pursue genetic testing, especially to help reduce anxiety and/or to assist in financial and long-term care planning.

AD can be loosely divided into two overlapping categories: early-onset disease, or EOAD, and late-onset disease, or LOAD. EOAD starts before age 60-65 years and makes up less than 5% of all AD cases. LOAD begins after age 60-65 years and constitutes the vast majority of cases.

Genetically, most - but not all - EOAD is due to the effect of mutation in one of several genes and is inherited in an autosomal dominant pattern, meaning that there is a 50% risk for each child of an affected person to inherit the same mutation. Three genes are currently known to cause this EOAD - amyloid-beta precursor protein (APP), presenilin 1 (PSEN1) and presenilin 2 (PSEN2).

Although these are rare causes of AD overall, the disease is likely to occur in all patients with a mutation in APP or PSEN1 and in 95% of those with a mutation in PSEN2. Testing can identify a mutation in only 40%-80% of EOAD cases and is most likely to be useful when three or more first-degree relatives have AD and at least one of them is early onset. Ideally, testing should be done first on an affected patient, in order to conclusively identify the genetic cause of AD in the family. In the absence of such knowledge, there is the risk of both false-negative and false-positive test results when screening at-risk family members to assess their future risk.

LOAD occurs sporadically - in the setting of a negative family history - about 75% of the time, but familial clustering (two or more first-, second-, or third-degree relatives) is found in up to 25% of cases. The population risk of AD by age 80 years is 10%-12%, but this goes up to 15%-39% when there is an affected first-degree relative. So far, the only genetic factor known to contribute to this risk is the apolipoprotein E (apo E) gene, which can be present in the E2, E3, or E4 form.

The E4 form is associated with increased risk, and most AD patients have one or two copies of this variant. From a risk-assessment perspective, although estimates vary widely, healthy individuals with only one copy of E4 (heterozygotes) are thought to have 2- to 3-fold increased risk of AD, while those with two copies of E4 (homozygotes) have 2- to 10-fold increased risk. Some data suggest that the E4 allele hastens development of AD, rather than increasing total lifetime risk. There is evidence that the E2 allele may decrease the risk of AD. However, this form increases the risk of hyperlipoproteinemia type III and may also increase the risk of macular degeneration.