Genetically, it helps to think of Alzheimer’s disease risk as falling into three distinct categories, each behaving differently.

• APP, PSEN1, and PSEN2 — the early-onset, single-gene forms. A pathogenic variant in any one of these three genes causes early-onset familial Alzheimer’s disease, typically beginning in a person’s forties or fifties.

o APP accounts for about ten-to-fifteen-in-one-hundred cases of early-onset familial disease, with onset usually in the forties and fifties, occasionally into the sixties.
o PSEN1 is the most common cause, accounting for roughly twenty-to-seventy-in-one-hundred cases of early-onset familial disease, with onset usually in the forties or early fifties. This form tends to progress relatively quickly, over six to seven years, and is often accompanied by seizures, involuntary muscle jerks called myoclonus, and language difficulties.
o PSEN2 is the least common of the three, accounting for about five-in-one-hundred cases, with a wider age range of onset, from forty to seventy-five. Some people with a PSEN2 change remain symptom-free even past age eighty, a pattern called reduced penetrance, meaning the gene change does not guarantee illness will appear.
o In twenty-to-forty-in-one-hundred families with a clear pattern of early-onset familial Alzheimer’s disease, no change in any of these three genes is found, meaning additional, not-yet-identified genes likely play a role.

• APOE — a risk-modifying gene, not a deterministic one. APOE comes in three common versions, called e2, e3, and e4. Carrying one or two copies of the e4 version raises the lifetime chance of developing Alzheimer’s disease but does not guarantee it happens, and does not follow the same predictable, single-gene inheritance as APP, PSEN1, or PSEN2. About forty-two-in-one-hundred people who develop Alzheimer’s disease do not carry an e4 variant at all, while the e2 version appears to offer some protection.

• Susceptibility genes. Roughly twenty other genes, including ABCA7, BIN1, CLU, and SORL1 among others, have each been linked to small increases in risk through large research studies. None of these causes Alzheimer’s disease on its own, and combining them into a “polygenic risk score” is not currently considered useful for individual clinical decision-making.

Because these categories behave so differently, understanding which one applies, or whether none does, is the foundation for any conversation about inheritance or testing.