Selegiline

Early Parkinson's disease

• Tetrud and Langston 1989 was a double-blind, placebo-controlled trial in 54 patients. The average time until levodopa was needed was 548.9 days on selegiline against 312.1 days on placebo (PMID: 2502843).
• In the DATATOP interim report (1989), with 800 patients in a two-by-two factorial design with tocopherol, 97 patients on selegiline reached the primary endpoint, disability requiring levodopa, against 176 not on selegiline (hazard ratio 0.43) over an average 12 months (PMID: 2509910).
• In the 1993 DATATOP report, the benefit, largely in the first 12 months, remained significant (hazard ratio 0.50), with a median delay of about nine months; Parkinson's ratings improved during the first three months on selegiline and worsened after treatment was withdrawn, the authors wrote that the action behind the benefit remained unclear, and tocopherol had no effect (PMID: 8417384). A 1992 reanalysis of the hazard functions concluded that the effect was transient and likely symptomatic rather than neuroprotective (PMID: 1471871).
• In the 1996 DATATOP follow-up, earlier treatment with selegiline or tocopherol did not reduce the later rate of wearing off, dyskinesias or freezing on levodopa (PMID: 8572664).
• The 2002 DATATOP extension re-randomized 368 levodopa-treated patients. Staying on selegiline did not change the primary outcome, first wearing off, dyskinesia or on-off fluctuation (hazard ratio 0.87, p=0.38), but motor decline was slower, dyskinesias more frequent and freezing of gait less frequent (PMID: 12112107).
• A 2004 meta-analysis of 17 trials with 3,525 patients found less need for levodopa (odds ratio 0.57) and fewer motor fluctuations (0.75) with MAO-B inhibitors (PMID: 15310558). A 2005 Cochrane review of 10 trials, 9 of them with selegiline, found benefits at one year that were statistically but not clinically significant, concluded that MAO-B inhibitors do not appear to delay disease progression, and did not recommend them for routine use in early Parkinson's disease (PMID: 16034956).

Add-on to levodopa

• In the trials behind the US label, selegiline added to levodopa/carbidopa beat placebo on change in levodopa dose, time spent off and patients' own rating of treatment success, and the label states that there is no evidence of benefit without levodopa (US prescribing information for selegiline hydrochloride tablets).
• In an open randomized trial of 520 patients by the UK Parkinson's Disease Research Group (1995), levodopa plus selegiline gave no clinical benefit over levodopa alone and was associated with higher mortality (PMID: 8555803).

Depression (skin patch)

• Bodkin and Amsterdam 2002 was a 6-week trial in 177 outpatients on a tyramine-restricted diet. The patch beat placebo on all depression measures, and application site reactions were the only adverse effect more common than on placebo (PMID: 12411221).
• Amsterdam 2003 was an 8-week trial with no diet restriction that analyzed 289 outpatients. The patch beat placebo on the MADRS (p=.001) and the 28-item Hamilton scale (p=.039) but not significantly on the 17-item Hamilton scale (p=.069), and the authors described the effect as modest (PMID: 12633131).
• Feiger 2006 was an 8-week flexible-dose trial in 265 patients. The primary 28-item Hamilton score improved more than on placebo, with a modest effect (PMID: 17017821).
• Amsterdam and Bodkin 2006 followed 322 responders for 52 weeks. Relapse occurred in 16.8 percent on the patch against 30.7 percent on placebo (PMID: 17110814).
• DelBello 2014 was a 12-week trial in 308 adolescents. Depression scores fell equally on the patch and on placebo (PMID: 24955812).

Alzheimer's disease

• In a 2003 Cochrane review of 17 trials (Birks and Flicker), some memory and cognition scores improved at 4 to 17 weeks, but the reviewers found no clinically meaningful benefit and concluded that there was no justification for using selegiline in Alzheimer's disease or for further efficacy studies (PMID: 12535396).

Lifespan studies in animals

• Knoll 1988, at Semmelweis University, studied 132 male rats treated from the end of their second year. Rats given selegiline under the skin three times a week lived on average 197.98 weeks against 147.05 weeks on saline, and the longest-lived treated rat died in its 226th week (PMID: 3147347). A 1989 report of the same design gives 191.91 weeks for the selegiline group (PMID: 2505007).
• Some studies found longer life: selegiline increased mean and maximum survival in aged male Fischer rats in Toronto (PMID: 2118586), lengthened life by 8.1 percent in male and 6.7 percent in female Fischer rats in Japan, where a higher dose had earlier shortened life slightly but not significantly (PMID: 16463106), and increased lifespan in female but not male Syrian hamsters (PMID: 9258898).
• Other studies found little or no effect: in aged C57BL/6J mice at the US National Institute on Aging, selegiline cut striatal MAO-B by up to 60 percent but did not change survival (PMID: 8247225); in male Fischer rats treated from middle age, no clear effect on longevity was seen, although water maze learning improved (PMID: 9263197); and in mice at the Jackson Laboratory, lifespans rose 6 to 9 percent, significant only when all groups were pooled, while treated old males sired far fewer pups (PMID: 8914495).
• In dogs (82 beagles), survival did not differ overall, while in a subset of dogs aged 10 to 15, 12 of 15 on selegiline survived to the end against 7 of 18 on placebo (PMID: 9307048); a 2025 reanalysis that accounted for age at enrollment and sex found no significant effect (PMID: 40816452).
• A 2025 meta-analysis of 22 rodent experiments found that selegiline increased average lifespan with a moderate effect size (standardized mean difference 0.68, p=0.0002), with no significant sign of publication bias but substantial heterogeneity, and with larger effects at higher doses and older ages at the start of treatment; the author called for clinical studies in older adults (PMID: 40816452).
• In a 2016 rat study by Knoll and Miklya, BPAP extended lifespan more potently than selegiline at low enhancer doses (PMID: 27777099).

Survival in people

An open, uncontrolled 1985 study of people with Parkinson's disease, co-authored by Knoll, reported longer life expectancy with levodopa plus selegiline (564 patients) than with levodopa alone (377) (PMID: 3935752). Randomized trials and their meta-analyses did not find longer survival with selegiline (PMID: 9506548, PMID: 9748034, PMID: 15310558).

The Emsam skin patch in depression

In placebo-controlled depression trials (817 patients on Emsam, 668 on placebo), application site reactions affected 24 against 12 percent, headache 18 against 17 percent, insomnia 12 against 7 percent, diarrhea 9 against 7 percent and dry mouth 8 against 6 percent, and 7.1 percent against 3.6 percent stopped because of an adverse reaction (US prescribing information for Emsam). Low systolic blood pressure occurred in 3.0 against 1.5 percent, mania in 8 of 2,036 patch-treated patients (0.4 percent), and the label carries a boxed warning that antidepressants increased suicidal thoughts and behaviors in pediatric and young adult patients in short-term studies (US prescribing information for Emsam).

Blood pressure crises

The oral label reports rare hypertensive reactions after tyramine-containing foods at the labeled dose and one hypertensive crisis in a patient also taking ephedrine (US prescribing information for selegiline hydrochloride tablets). The Emsam label states that hypertensive crises after tyramine-rich food or drink can be fatal and that intracranial bleeding has been reported with the rise in blood pressure (US prescribing information for Emsam).

Serotonin syndrome and drug reactions

Stupor, muscle rigidity, severe agitation and high temperature, and in some cases death, have been reported with selegiline plus meperidine, and severe toxicity including deaths with selegiline plus tricyclic antidepressants or selective serotonin reuptake inhibitors (US prescribing information for selegiline hydrochloride tablets). The Emsam label describes potentially life-threatening serotonin syndrome when MAO inhibitors are combined with serotonergic drugs (US prescribing information for Emsam).

Overdose

Some people exposed to 600 mg of racemic (d,l) selegiline during its development had severe low blood pressure and agitation, and the oral label expects overdose to inhibit both MAO-A and MAO-B (US prescribing information for selegiline hydrochloride tablets).

Deaths in Parkinson's disease trials

In the open UK trial of 520 people with early Parkinson's disease, mortality was 1.57 times higher with levodopa plus selegiline than with levodopa alone after 5 to 6 years (PMID: 8555803). With longer follow-up the hazard ratio was 1.32 and not significant in the original arms, 1.33 (95 percent CI 1.02 to 1.74) when all randomized subjects were included, and the excess was in deaths attributed to Parkinson's disease; the selegiline arm was stopped in 1995 (PMID: 9583926). DATATOP found no effect of selegiline on deaths over 8.2 years (PMID: 9506548), a 1998 meta-analysis of five trials found a hazard ratio of 1.02 (PMID: 9748034), a 2004 meta-analysis of 17 trials an odds ratio of 1.13, not significant (PMID: 15310558), and a 2005 Cochrane review an odds ratio of 1.15 whose confidence interval did not exclude a small increase (PMID: 16034956).