Memantine

Memantine is not a general memory booster. It is a symptomatic drug for a specific stage of dementia. The 2019 Cochrane review assembled nearly 10,000 participants across 44 trials. In roughly 3,700 people with moderate-to-severe Alzheimer's disease, high-certainty evidence showed small benefits over six to seven months in global rating, cognition, activities of daily living and behaviour. The effect was present whether or not people also took a cholinesterase inhibitor [2].

The scale of benefit matters. In the pooled analysis, the average advantage was 3.11 points on the 100-point Severe Impairment Battery and 1.09 points on the 54-point ADL19 scale [2]. In the original 28-week monotherapy trial, memantine reduced deterioration on daily function and severe-cognition measures, but one of the two primary outcomes reached significance only in the observed-cases analysis, not the last-observation-carried- forward analysis [3]. It is a brake, not a reversal.

Added to stable donepezil, memantine produced statistically better cognitive, daily-function, global and behavioural outcomes over 24 weeks [4]. A later 52-week factorial trial found average advantages of 1.2 points on the 30-point SMMSE and 1.5 points on the 60-point BADLS, below its stated minimum clinically important differences; combining memantine with donepezil was not significantly better than donepezil alone [5].

Severity is the boundary. In mild Alzheimer's disease, pooled trials found no advantage on cognition, daily living, global rating or behaviour [2][6]. There are no cited trials showing cognitive enhancement in healthy people.

NMDA receptors are both essential and dangerous. Brief, well-timed activation supports synaptic plasticity and memory; prolonged activation can drive excessive calcium entry and excitotoxic stress. Memantine sits inside the open ion channel rather than competing with glutamate at its binding site. Its block is uncompetitive, voltage-dependent, moderate in affinity and quick to disengage [1].

That kinetic profile is the point. A high-affinity, slow NMDA antagonist would also suppress normal synaptic signalling and cause unacceptable central effects. Memantine preferentially limits channels that remain pathologically active, while leaving during normal voltage changes and short transmitter pulses [1]. Clinical trials show symptomatic benefit, but they do not establish that this mechanism changes the underlying course of Alzheimer's pathology.

Across dementia trials, the Cochrane review found no meaningful difference in the chance of having at least one adverse event. Dizziness was more common with memantine than placebo (6.1% versus 3.9%) and headache may have been slightly more common (5.5% versus 4.3%); falls did not differ [2]. A pooled safety analysis of 2,311 trial participants similarly found overall event and discontinuation rates close to placebo [10].

Confusion can matter in the population most likely to receive it. In the donepezil add-on trial, confusion occurred in 7.9% on memantine and 2.0% on placebo, though most cases on memantine were mild and resolved within two weeks [4].

Clearance is a practical safety issue. Renal impairment roughly doubled exposure in the most impaired group and prolonged half-life to about 124 hours [9]. Alkaline urine sharply reduces renal clearance [8], so substantial changes in renal function or urinary pH can change exposure even without a dose change.