Galantamine

Alzheimer's disease: the trials behind approval

Janssen held the original US approval (Drugs@FDA NDA 021169), and the 2014 two-year trial lists Janssen authors (PMID: 24591834).

• Raskind 2000 (USA-1), 636 patients, 6 months, 24 or 32 mg a day. Both primary endpoints were met: cognition (ADAS-cog/11) improved by 3.9 and 3.8 points against placebo (p<0.001), and global ratings (CIBIC-plus) were better than placebo (p<0.05) (PMID: 10881250).
• Tariot 2000 (USA-10), 978 patients, 5 months, 8, 16 or 24 mg a day with slow dose escalation. ADAS-cog scores were 3.3 points (16 mg) and 3.6 points (24 mg) better than placebo (p<0.001), with better global, daily-living and behavior scores at those doses; stopping for adverse events was 6 to 10 percent with galantamine against 7 percent with placebo (PMID: 10881251).
• Wilcock 2000 (GAL-INT-1), 653 patients in Europe and Canada, 6 months, 24 or 32 mg a day. ADAS-cog scores were 2.9 and 3.1 points better than placebo (p<0.001), global ratings were better (p<0.05), and the higher dose improved a disability scale (PMID: 11110737).
• Cochrane review 2024, 21 trials, 10,990 participants. At 16 to 24 mg a day for six months, galantamine improved ADAS-cog by 2.86 points against placebo (high-certainty evidence), with smaller gains in daily function and behavior and a probable gain in global ratings; the authors also noted that unequal dropout combined with last-observation-carried-forward analysis may bias cognitive results in galantamine's favor (PMID: 39498781).

Alzheimer's disease: longer and later trials

• Hager 2014, 2-year trial, 2,045 patients. Safety, including mortality, was one primary end point, and mortality was lower with galantamine (hazard ratio 0.58, 95% CI 0.37 to 0.89, p=0.011). The primary efficacy end point, change in MMSE score at 24 months, declined less with galantamine (-1.41 against -2.14, p<0.001). The study was stopped early on the advice of its data safety monitoring board; 32 percent of patients completed it, 27 percent withdrew, and 41 percent were still enrolled when it stopped (PMID: 24591834).
• SERAD 2009, 407 nursing home residents with severe Alzheimer's disease. One co-primary endpoint was met (Severe Impairment Battery, difference 4.36 points, p=0.006) and the other was not (activities of daily living, p=0.383) (PMID: 19042161).

Vascular dementia

• Erkinjuntti 2002, 592 patients with probable vascular dementia or Alzheimer's disease with cerebrovascular disease, 6 months, 24 mg a day. Both primary endpoints were met: ADAS-cog improved by 2.7 points against placebo (p<0.0001), and 74 percent against 59 percent were stable or improved on global ratings (p=0.0001) (PMID: 11965273).
• Auchus 2007, 788 patients with probable vascular dementia, 26 weeks. Cognition improved more than with placebo (p<0.001), but the second co-primary endpoint, activities of daily living, did not differ (p=0.783), so the trial did not meet both co-primary endpoints; 13 percent of galantamine and 6 percent of placebo patients stopped because of adverse events (PMID: 17664404).

Mild cognitive impairment

• Winblad 2008, two 24-month trials, 2,048 people (990 and 1,058), 16 to 24 mg a day. The primary endpoint, conversion to dementia, did not differ from placebo in either trial (22.9 against 22.6 percent, and 25.4 against 31.2 percent). Recorded mortality was higher with galantamine in the original assessment, while a post hoc cohort analysis was consistent with no increased risk (PMID: 18322263). The US prescribing information for galantamine reports 13 deaths on galantamine and 1 on placebo in these trials, a difference it calls significant but highly discrepant with other galantamine trials.
• Cochrane review 2024. On low-certainty evidence, galantamine may not improve cognition (mean difference -0.21 on an expanded ADAS-cog) or daily activities in mild cognitive impairment, and the reviewers found no evidence to support its use there (PMID: 39498781).

Lucid dreaming

• LaBerge 2018, randomized double-blind crossover, 121 people with high dream recall and an interest in lucid dreaming. Participants woke about 4.5 hours after lights out, took 0, 4 or 8 mg galantamine on separate nights, stayed up at least 30 minutes and practiced a mental technique for inducing lucid dreams. Lucid dreams were reported by 14 percent with placebo, 27 percent with 4 mg (odds ratio 2.29) and 42 percent with 8 mg (odds ratio 4.46), and galantamine also raised dream recall, vividness and complexity (PMID: 30089135). Mild side effects were reported by 14 participants (12 percent) on galantamine nights against 4 (3 percent) on placebo nights, and one participant dropped out because of nausea (PMID: 30089135). The first two authors are affiliated with the Lucidity Institute, participants were recruited from the institute's own 8-day lucid dreaming workshops, the institute's own research ethics committee approved the study, and the capsules were an over-the-counter galantamine supplement (PMID: 30089135).
• Sparrow 2018, 35 participants, 8-night home study. Galantamine was combined with waking at night, meditation and reliving a distressing dream; the authors concluded the protocol merits further study (PMID: 29960246).
• Systematic review 2022. Galantamine was named as a possible lucid dream induction method that needs further replication (PMID: 36408823).

Prescription drug versus dietary supplement

• Cohen 2024, JAMA research letter. Eleven US generic immediate-release galantamine products contained 97.5 to 104.2 percent of their labeled dose with no microbial contamination. Ten supplements labeled with galantamine and bought online contained from under 2 percent to 110 percent of the labeled amount, only 1 was within 10 percent of its label, 1 had essentially none, and 3 carried Bacillus cereus with diarrheal toxin genes (PMID: 38393714).

Common effects in trials

The most common adverse reactions in double-blind galantamine trials (5 percent or more) were nausea, vomiting, diarrhea, dizziness, headache and decreased appetite, and the reactions that most often led to stopping were nausea (6.2 percent), vomiting (3.3 percent), decreased appetite (1.5 percent) and dizziness (1.3 percent) (US prescribing information for galantamine). In the 2024 Cochrane review of Alzheimer's disease trials at 16 to 24 mg a day, nausea affected 20.9 percent on galantamine against 8.4 percent on placebo, and 22.7 percent against 17.2 percent stopped early (PMID: 39498781). In the two mild cognitive impairment trials, nausea affected 29 percent on galantamine against 10 percent on placebo (PMID: 18322263).

Heart rhythm and fainting

The US prescribing information for galantamine warns that cholinesterase inhibitors can slow the heart and cause heart block through vagal effects, in patients with or without known conduction problems, and reports a dose-related rise in fainting (syncope) from 0.7 percent on placebo to 2.2 percent at 12 mg twice daily. A 2023 meta-analysis of 53 randomized trials of cholinesterase inhibitors, 14 of them with galantamine, found a higher risk of syncope (risk ratio 1.50) and a lower risk of falls (risk ratio 0.84) than with placebo (PMID: 37993407).

Deaths in mild cognitive impairment trials

In two 2-year trials in mild cognitive impairment, 13 people on galantamine (n=1,026) and 1 on placebo (n=1,022) died, about half of the galantamine deaths from vascular causes such as heart attack, stroke and sudden death (US prescribing information for galantamine). The label calls the difference significant but highly discrepant with other galantamine trials: the placebo death rate in these two trials was far lower than in the placebo groups of galantamine trials in Alzheimer's disease and other dementias (0.7 against 22 to 61 per 1,000 person-years), and in pooled Alzheimer's and other dementia trials (n=6,000) the placebo group had numerically higher mortality than the galantamine group. The trial authors reported that a post hoc analysis of the whole cohort was consistent with no increased risk (PMID: 18322263), and the 2024 Cochrane review rated the evidence on deaths in mild cognitive impairment as low certainty (PMID: 39498781).

Deaths in Alzheimer's disease trials

In Alzheimer's disease trials the pattern ran the other way. The 2024 Cochrane review found 1.3 percent of galantamine patients and 2.3 percent of placebo patients had died at six months (PMID: 39498781), a 2-year trial in 2,045 patients found lower mortality with galantamine (hazard ratio 0.58) (PMID: 24591834), and in a severe Alzheimer's disease trial 4 percent of galantamine patients and 11 percent of placebo patients died (PMID: 19042161).

Other labeled warnings

The US prescribing information for galantamine lists serious skin reactions (Stevens-Johnson syndrome and acute generalized exanthematous pustulosis), possible gastrointestinal bleeding in people at risk of ulcers, a possibility of seizures, bladder outflow obstruction, respiratory effects in severe asthma or obstructive lung disease, and exaggerated muscle relaxation with succinylcholine-type drugs during anesthesia.

Supplement quality

In a 2024 analysis, 3 of 10 galantamine supplements bought online contained Bacillus cereus bacteria carrying genes for diarrheal toxins, and measured galantamine ranged from under 2 percent to 110 percent of the labeled amount; none of 11 generic prescription products was contaminated (PMID: 38393714).

Pregnancy

The US prescribing information for galantamine states that there are no adequate human data in pregnancy and that animal studies showed developmental toxicity at doses similar to or above those used clinically.