What alzheimers treatment is best

The short version: the drugs offered for Alzheimer’s produce modest, temporary symptom effects, not cures, and the newest antibody treatments carry real brain-safety risks that many families are not warned about clearly.

The standard drugs—donepezil, rivastigmine, galantamine, and memantine—slow cognitive decline by about 2–3 points on the ADAS-Cog scale over 6–12 months. That is a real but small average effect [1, 3]. One large 2009 caregiver satisfaction study of 811 patients found that donepezil scored highest on overall satisfaction, ease of use, and perceived tolerability, rivastigmine the lowest on side effects and convenience [3]. The main benefit of medication appears to be delaying institutionalization: hazard ratio 0.80 for treated versus untreated patients in a 2017 study of nearly 9,000 newly diagnosed patients, along with a mortality hazard ratio of 0.69 [1]. But the absolute numbers from that study show annual costs of $25,828 per treated patient versus $30,110 untreated, meaning the drugs pay for themselves partly by keeping people at home longer, not by reversing the disease [1].

The new anti-amyloid antibodies lecanemab and donanemab have a very different risk profile. The 2026 review of both drugs reports that lecanemab slowed decline by only 0.45 points on the CDR-SB scale over 18 months, a difference many clinicians consider clinically meaningless for daily life [2]. Donanemab, in the TRAILBLAZER-ALZ 2 trial, showed a 3.25-point benefit on the iADRS scale over 76 weeks in people with low or medium tau [2]. The trade-off is brain swelling and bleeding—ARIA—especially in people carrying the ApoE4 gene. In lecanemab’s trial, 45% of ApoE4/E4 carriers had some form of ARIA and 13% had the swelling type (ARIA-E, with 9% symptomatic). In donanemab’s trial, 55% of E4/E4 carriers had ARIA and 24% had ARIA-E [2]. Intracranial hemorrhage over 1 cm occurred in 0.7% of lecanemab patients versus 0.1% of placebo [2]. The drugs require regular MRI monitoring and infusions, and the cost is about $26,500 per year without confirmed cost-effectiveness [2].

My call: for a person with mild to moderate Alzheimer’s, donepezil or the generic donepezil oral disintegrating tablet has the best evidence for modest cognitive slowing, better tolerability, and delaying institutionalization, based on the largest head-to-head satisfaction data [3]. For early Alzheimer’s with confirmed amyloid, lecanemab and donanemab show measurable but small cognitive effects at the cost of real brain safety risks (ARIA, hemorrhage) that require frequent MRI monitoring—these are not a fit for most patients. Confidence is moderate: the key weakness is that no trial has shown improvement in functional independence or quality of life on a scale that families typically care about.