Selegiline

Selegiline is methamphetamine with a propargyl group on the nitrogen [10]. Joseph Knoll's group in Budapest created it in the 1960s, it became the first selective inhibitor of monoamine oxidase B, and it has been registered in more than 60 countries for Parkinson's disease, Alzheimer's disease and major depression, and used as an anti-aging drug [1].

Parkinson's disease. The DATATOP trial randomized 800 people with early, untreated Parkinson's disease to selegiline 10 mg a day, vitamin E, both or placebo. An interim analysis found that selegiline cut the risk of needing levodopa by 57% [11]; with longer follow-up the hazard ratio was 0.50 and the median delay about nine months (719 against 454 days). Vitamin E did nothing. The benefit was concentrated in the first year, with the hazard ratio rising from 0.35 in the first six months to 0.86 after 18 months, and motor scores improved on starting selegiline and worsened on stopping it, so the trial could not separate a protective effect from symptom relief [3]. Pooled across 17 trials, MAO-B inhibitors reduced the need for levodopa (odds ratio 0.57) and motor fluctuations (0.75) without changing mortality, and in the six selegiline trials with rating-scale data total scores were 2.7 points better than control at three months [4]. A 2023 meta-analysis of 27 randomized and 11 observational studies found the difference from placebo growing with time, from about 3.3 points at three months to 11 points at five years, at the cost of more adverse events, mostly neuropsychiatric [12]. In people already on levodopa with motor fluctuations, the dissolving tablet cut daily off time by about 2.2 hours at 12 weeks, against 0.6 hours on placebo [13].

The mortality scare. In 1995 the UK Parkinson's Disease Research Group reported that, in an open trial of 520 patients, those given selegiline with levodopa died at 1.57 times the rate of those on levodopa alone, with no clinical benefit [14]. Other evidence did not bear this out: DATATOP found no effect on survival over 8.2 years [15], a pooled analysis of five long-term trials gave a hazard ratio of 1.02 [16], and the 17-trial meta-analysis judged the UK result probably a chance finding, while noting that its confidence interval could not rule out a small increase [4].

Depression. At the 10 mg a day used for Parkinson's disease, oral selegiline did not lift depression in three weeks; at about 30 mg a day, a dose that is no longer selective, half of patients responded against 14% on placebo [17], and 60 mg a day helped treatment-resistant older patients [18]. Such doses bring back the tyramine reaction [19]. The skin patch was designed to inhibit MAO-A and MAO-B in the brain while largely sparing the gut and liver [20]. It beat placebo in three short adult trials, by a modest margin in two of them, and in a 52-week trial cut relapse from 30.7% to 16.8% [20][21][22][23]. In 308 adolescents it did no better than placebo [2]. A 2023 meta-analysis across psychiatric disorders found selegiline better than placebo for depression (standardized mean difference −0.96, ten trials) and atypical depression, with moderate confidence only for symptom reduction with the patch [5].

Other uses tried. For ADHD, two small Iranian trials found it no different from methylphenidate [24][25] and a crossover in 11 children reported gains in sustained attention and hyperactivity [26], but 20 or 60 mg a day did no better than placebo in 24 adults [27], a trial in children with Tourette's syndrome missed its primary outcome [28], and the pooled placebo-controlled trials showed no benefit [5]. In narcolepsy, 20 mg a day reduced sleepiness, sleep attacks and cataplexy over two weeks in 30 patients [29]. A multicenter trial reported fewer negative symptoms when selegiline was added to antipsychotics in schizophrenia [30], but pooled trials did not confirm it [5]. It failed for smoking cessation, by mouth and by patch, and for cocaine dependence [31][32][33].

Alzheimer's disease. In a two-year trial of 341 patients with moderately severe disease, neither selegiline nor vitamin E differed from placebo in the unadjusted analysis; after adjusting for higher baseline cognitive scores in the placebo group, selegiline delayed death, institutionalization, loss of basic daily activities or severe dementia by a median of about seven months, without any effect on cognitive test scores [34]. A Cochrane review of 17 trials found small short-term cognitive gains but no clinically meaningful benefit, and saw no justification for further trials [9].

Aging and the nootropic use. In 1988 Knoll reported that male rats given 0.25 mg/kg three times a week from two years of age lived on average 198 weeks against 147 for saline-treated rats, and that it restored sexual activity in 64 of 66 [7]. Later studies were mixed. Fischer rats lived about 7–15% longer in Kitani's studies [35][36], female but not male hamsters lived longer [37], mice gained 6–9% in pooled data [38] or nothing [8], Fischer rats given it in drinking water learned a water maze better but did not clearly live longer [39], male Wistar rats given 0.5 mg/kg died sooner [40], and old beagles lived longer only in a subgroup analysis [41]. Kitani concluded that the effect needs a dose within a certain range, and that higher doses may shorten life [36]. Knoll attributed the effect not to MAO-B but to a "catecholaminergic activity enhancer" action seen at much lower doses, and proposed 1 mg a day for life from sexual maturity [42][43]. No study has tested that in people.

MAO-B inhibition. Selegiline first binds monoamine oxidase B reversibly and then attaches covalently to its flavin cofactor, destroying the enzyme. PET shows it retained in regions rich in MAO-B, including the striatum, hippocampus, thalamus and substantia nigra [6]. Because the block is permanent, activity returns only as new enzyme is made: after people stopped taking it, brain MAO-B came back with a half-time of about 40 days [44]. MAO-B breaks down the trace amine phenylethylamine [43]; blocking it raises dopamine in the substantia nigra, and selegiline also inhibits dopamine reuptake [45].

Selectivity depends on dose and route. At 10 mg a day by mouth it leaves brain MAO-A, the enzyme that clears serotonin, norepinephrine and dietary tyramine, substantially intact [46]. At 30 mg a day tyramine sensitivity rose two- to fourfold and plasma levels of a norepinephrine metabolite fell, signs of MAO-A inhibition [19]; at 60 mg a day spinal-fluid metabolites of norepinephrine and serotonin fell by 51% and 17% [18]. Formulations that avoid first-pass metabolism give much higher selegiline levels: in healthy men, 10 mg a day of the dissolving tablet for 28 days inhibited brain MAO-A by 37% on PET, and the patch by 33% in three subjects, a difference that was not significant [46]. That loss of selectivity is thought to explain why the patch and high oral doses work in depression [46].

Amphetamine metabolites. Selegiline is metabolized to l-desmethylselegiline, l-methamphetamine and l-amphetamine, with no conversion to the d-forms [6][47]; the l-isomers have cardiovascular and psychological effects, weaker than those of the d-isomers [32]. After a 10 mg tablet the metabolite levels exceed those of selegiline itself [48]. Desmethylselegiline is itself an irreversible MAO-B inhibitor, but MAO-B inhibition tracked selegiline levels and not those of the metabolites [6][48]. The amphetamines have been discussed as a possible cause of selegiline's cardiovascular effects with levodopa [49].

The "enhancer" effect. Knoll's group argued that small doses of selegiline also make catecholamine neurons release more transmitter per nerve impulse, separately from MAO-B. Three weeks of 0.01–0.25 mg/kg a day raised dopamine and norepinephrine release from rat brain tissue and lowered serotonin release, and PPAP, an analog with no MAO-B activity, did the same [50][51][52]. In rat striatal slices, 10⁻¹⁰–10⁻⁹ M selegiline raised electrically evoked dopamine release without raising resting release; rasagiline, another MAO-B inhibitor, had no such effect and blocked selegiline's, as did the TAAR1 antagonist EPPTB, and the metabolite l-methamphetamine was also active at 10⁻⁹ M [53]. The group proposes that the enhancer effect works through TAAR1, a receptor for trace amines and amphetamines, and that it is what lengthened rat lifespans [42][53]. The evidence comes almost entirely from that laboratory's own assays, and how much the effect adds to MAO-B inhibition at clinical doses has not been tested in people.

Anti-apoptotic actions. In the cell and animal work reviewed by Tatton's group, selegiline and other propargylamines protected neurons independently of MAO-B by binding GAPDH and shifting protein synthesis away from cell death [54]. Selegiline also raised NGF and BDNF output from cultured astrocytes, but only at 0.4–0.5 mM [55]. These mechanisms were part of the rationale for neuroprotection trials, and no trial established protection in patients [3][9].

Three forms are in use, and they differ more than their doses suggest.

Tablets and capsules. A 10 mg oral dose is absorbed quickly, peaks within an hour and has a half-life of about 1.5 hours, but most of it is metabolized before reaching the circulation: absolute bioavailability is about 4–10%, and levels vary widely between people [45][56]. Taking it with food raised exposure at least threefold [45].

Orally disintegrating tablet (Zydis selegiline, Zelapar). Placed on the tongue, it dissolves, and almost a third of the dose is absorbed through the mouth within a minute. In seven studies in 156 healthy volunteers, 1.25 mg gave selegiline levels and MAO-B inhibition similar to a 10 mg tablet, with peak metabolite levels about a tenth as high: 0.93 against 12.9 ng/mL of l-methamphetamine and 0.34 against 3.6 ng/mL of l-amphetamine [48]. At 10 mg a day it no longer spares brain MAO-A [46].

Transdermal patch (Emsam). Through the skin, 73% of the dose reaches the circulation, against 4% for a tablet, and plasma levels stay high and steady through the day [2][56]. Bypassing the gut is meant to inhibit MAO-A and MAO-B in the brain while largely sparing intestinal and liver MAO-A, which clears dietary tyramine [20]. At 6 mg/24 h, tyramine sensitivity rose by a factor of 1.85, similar to 10 mg tablets and more than 20 times less than with tranylcypromine [57].

In Parkinson's trials selegiline was generally well tolerated, though not free of side effects. In DATATOP, nausea, injuries and mild cardiac arrhythmias were more frequent with selegiline, including bradycardia with atrioventricular block in five people on selegiline and none on placebo alone, and liver enzymes were raised more often; blood pressure and pulse did not change [3]. The 2023 meta-analysis found adverse events in 62.1% on selegiline against 54.7% on placebo, the excess mainly neuropsychiatric [12]. With levodopa, head-up tilt caused orthostatic hypotension in 9 of 16 patients on selegiline, two of whom lost consciousness, and in none of 9 on levodopa alone; stopping selegiline reversed it [49]. In the Alzheimer's trial falls and syncope were more frequent in the treated groups [34].

In psychiatric trials the patch and tablets caused more insomnia and dry mouth than placebo, and the patch more skin reactions [5]. No hypertensive crisis occurred in the adult patch trials run without a tyramine diet or in the adolescent trial [2][22][23]. In the adolescent trial, three patients on the patch withdrew because of suicidal ideation [2].

The mortality signal from the UK trial was not confirmed by DATATOP, a pooled analysis of five trials or the 17-trial meta-analysis [4][14][15][16].

Interactions follow from the pharmacology. Higher doses and the non-oral forms inhibit MAO-A [19][46]. Combined with the MAO-A inhibitor moclobemide, tyramine sensitivity rose 8.4-fold, more than with either drug alone [59]. The manufacturer advised against combining it with antidepressants because of serotonin syndrome; a survey of Parkinson's investigators found possible serotonin syndrome in 0.24% of 4,568 patients given both, serious in 0.04%, and no deaths [60]. Intravenous methamphetamine during selegiline treatment did not add to the rise in heart rate or blood pressure in a small study in methamphetamine users [61], and the patch blunted some effects of intravenous cocaine [62].

Its metabolites make urine screens positive for methamphetamine and amphetamine, and only chiral analysis shows that they are the l-isomers [63]. DATATOP and the Alzheimer's trial used the urine test to check that patients were taking the drug [3][34].