SAM-e (S-adenosylmethionine)

• Delle Chiaie 2002, two multicenter trials in Italy, 576 patients with major depression. One trial gave oral SAM-e at 1,600 mg a day (143 patients) or imipramine at 150 mg a day (138 patients) for 6 weeks; the other gave intramuscular SAM-e at 400 mg a day (147 patients) or imipramine (148 patients) for 4 weeks. The main efficacy measures, the final HAM-D score and the Clinical Global Impression responder rate, did not differ between SAM-e and imipramine in either trial, and SAM-e caused significantly fewer adverse events. Neither trial had a placebo arm (PMID: 12418499).
• Papakostas 2010, Massachusetts General Hospital, 73 patients who had not responded to serotonin reuptake inhibitors, 6 weeks. SAM-e (target 800 mg twice a day) or placebo was added to the antidepressant. The paper names response on the 17-item HAM-D as the primary outcome: 36.1 percent responded with SAM-e against 17.6 percent with placebo, and remission rates were 25.8 and 11.7 percent; the abstract gives no p-values for these comparisons (PMID: 20595412). The trial's ClinicalTrials.gov record (NCT00093847) instead lists remission as the primary outcome and posts different figures, 46.1 against 17.6 percent remission and 35.8 against 11.7 percent response. The 2016 Cochrane review, drawing on this one trial, found low quality evidence that SAM-e added to SSRIs beat placebo on change in depressive symptoms (PMID: 27727432).
• Mischoulon 2014, Massachusetts General Hospital and Butler Hospital, 189 outpatients, 12 weeks. Patients received SAM-e (1,600 to 3,200 mg a day), escitalopram (10 to 20 mg a day) or placebo. The primary outcome, the 17-item HAM-D, improved by about 5 to 6 points in all three arms with no significant differences between them, and response rates were 36, 34 and 30 percent. The authors concluded that the results did not support an advantage of either SAM-e or escitalopram over placebo (PMID: 24500245).
• Sarris 2018, Australia, 107 patients not in remission on antidepressants, 8 weeks. Adding SAM-e at 800 mg a day did not beat placebo on the primary outcome, the MADRS (p=0.51); response rates were 54.3 percent with SAM-e and 50.0 percent with placebo (PMID: 30115553).
• Sarris and colleagues, Australia, 49 unmedicated patients with mild to moderate depression, 8 weeks. SAM-e alone at 800 mg a day reduced MADRS scores 3.76 points more than placebo, a difference that was not significant (p=0.13), and 53 percent responded on placebo (PMID: 31712971).
• Sarris and colleagues, Australia, 158 patients, 8 weeks, combination product. A combination of SAM-e, folinic acid, omega-3 fatty acids, 5-HTP, zinc and co-factors did no better than placebo, with response rates of 40 and 51 percent and a non-significant difference (PMID: 30699842). The trial cannot separate the share of SAM-e.
• Utrecht trial (published 2026), 31 adults with depression linked to childhood trauma. Participants had 12 weeks of trauma-focused psychotherapy plus SAM-e at 1,200 mg a day or placebo. Depression scores fell in both groups with no clinical benefit from SAM-e, plasma SAM-e levels did not change, and SAM-e was associated with DNA methylation differences in 66 genomic regions (PMID: 42027147).
• Pooled analyses. The 2016 Cochrane review of 8 trials in 934 adults found no strong evidence that SAM-e alone differed from placebo (2 trials, very low quality evidence), similar results for SAM-e and imipramine (4 trials, low quality evidence), and low quality evidence that SAM-e added to SSRIs beat placebo (1 trial) (PMID: 27727432). A 2024 meta-analysis of 14 trials with 1,522 participants found no significant difference between SAM-e and placebo, imipramine or escitalopram (PMID: 38199136). A second 2024 meta-analysis, of 23 trials with 2,183 participants, found that SAM-e alone beat placebo (standardized mean difference -0.58) while SAM-e added to antidepressants did not, and that SAM-e did not differ from antidepressants (PMID: 38423354). A 2025 review of trials in milder depression found no significant improvement with SAM-e over placebo (PMID: 40014460).
• Guideline. A 2022 taskforce of the World Federation of Societies of Biological Psychiatry (WFSBP) and the Canadian Network for Mood and Anxiety Treatments (CANMAT) graded SAM-e added to antidepressants as weakly recommended and SAM-e on its own as not currently recommended for unipolar depression (PMID: 35311615).

Osteoarthritis

• Cochrane 2009, 4 placebo-controlled trials, 656 patients with knee or hip osteoarthritis. Trial quality was poor. The pooled pain effect was small (standardized mean difference -0.17, 95 percent confidence interval -0.34 to 0.01), equal to 0.4 cm on a 10 cm pain scale, and function did not differ; no trial reported on serious adverse events, and the authors called the review inconclusive (PMID: 19821403).
• Soeken 2002, meta-analysis of 11 studies. Against placebo, SAM-e reduced functional limitation (effect size 0.31) but not pain, a result based on only 2 studies. SAM-e was comparable with nonsteroidal anti-inflammatory drugs, and fewer people on SAM-e reported adverse effects (PMID: 12019049).
• Najm 2004, University of California, Irvine, 61 adults with knee osteoarthritis (56 completed), 16-week double-blind crossover. SAM-e at 1,200 mg a day was compared with celecoxib at 200 mg a day. Celecoxib reduced pain more in the first month (p=0.024), there was no significant difference by the second month, and function improved in both groups (PMID: 15102339).

Liver disease

• Mato 1999, Spanish-led multicenter trial, 123 patients with alcoholic cirrhosis, 2 years. Patients took SAM-e at 1,200 mg a day by mouth or placebo. Death or liver transplant occurred in 16 percent on SAM-e and 30 percent on placebo, a difference that was not significant (p=0.077). After the 8 patients with the most advanced (Child C) disease were excluded, the rates were 12 and 29 percent (p=0.025) (PMID: 10406187).
• Medici 2011, University of California, Davis, 37 patients with alcoholic liver disease, 24 weeks. SAM-e at 1.2 g a day or placebo was given to patients who were required to stay abstinent from alcohol. The primary endpoints, serum aminotransferases and liver biopsy scores, did not differ between groups (PMID: 22044287).
• Morgan 2015, US phase 2 trial, 110 patients with hepatitis C cirrhosis and a mildly raised alpha-fetoprotein (AFP), 24 weeks. SAM-e at up to 2.4 g a day did not change AFP, the primary outcome, compared with placebo, and did not improve liver tests, oxidative stress markers or serum glutathione; 87 patients completed the trial (PMID: 26130251).
• Guo 2015, meta-analysis of 12 trials, 705 patients with chronic liver disease. Total bilirubin and AST differed with SAM-e but ALT did not, adverse events and long-term outcomes did not differ from placebo, and two other treatments were more effective than SAM-e in some liver diseases (PMID: 25774783).
• Intrahepatic cholestasis of pregnancy. The 2020 Cochrane review of drug treatments for this condition found insufficient evidence to show whether SAM-e works (PMID: 32716060). A 2016 meta-analysis of 5 trials with 311 patients found that ursodeoxycholic acid lowered itch scores, bile acids and ALT more than SAM-e and was associated with fewer preterm deliveries (PMID: 27799965).
• Intrahepatic cholestasis outside pregnancy. A 2020 systematic review of 3 randomized and 6 non-randomized studies reported that both placebo-controlled randomized studies found lower ALT with SAM-e within 2 weeks (PMID: 32184941). One of its three authors worked for Abbott, the sponsor of registered trials of ademetionine under the name Heptral (ClinicalTrials.gov NCT02200029).

Other conditions

• Fibromyalgia, Copenhagen, 44 patients, 6 weeks. SAM-e at 800 mg a day improved clinical disease activity, pain, fatigue, morning stiffness and mood on a face scale compared with placebo, while tender point scores, muscle strength and Beck Depression Inventory scores did not differ (PMID: 1925418).
• Alzheimer's disease, Melbourne-led phase 2 trial, 63 participants, 180 days. SAM-e at 400 mg a day did not change the primary outcome, plasma phosphorylated tau 217, compared with placebo (p=0.288), and safety and cognitive outcomes did not differ (PMID: 41980907).
• Smoking, Mayo Clinic, 120 adults, 8 weeks. SAM-e at 800 or 1,600 mg a day plus counseling did not raise abstinence rates at 8 or 24 weeks compared with placebo (PMID: 22901308).
• Schizophrenia, Israel, 18 patients with the low-activity COMT genotype, 8 weeks. Some reduction in aggressive behavior and better quality of life were reported after SAM-e at 800 mg a day (PMID: 18824331).

How SAM-e is absorbed in people

• Basel, 14 healthy adults, one 400 mg oral dose. Plasma SAM-e rose from 38 to 362 nmol per liter and fell back with a half-life of 1.7 hours (PMID: 9262350).
• China Pharmaceutical University, 20 healthy volunteers, 1,000 mg a day for 5 days. The terminal half-life was about 6 hours after enteric-coated tablets and about 4 hours after intravenous infusion, oral exposure was about 2 to 3 percent of intravenous exposure, and SAM-e did not accumulate (PMID: 19302903).
• Guangdong, 24 healthy volunteers, one enteric-coated tablet. Food delayed the peak from 4.5 to 7.5 hours and gave higher total exposure (PMID: 34981547).
• Company phase 1 study of an experimental tablet (MSI-195). The developer reported about 2.8 times the exposure of a commercial SAM-e supplement; food delayed the peak from 4.5 to 13 hours and cut exposure to 55 percent of the fasting value (PMID: 33317621).

Animal and cell studies

These results come from animals and cells, not people: liver cancer in Mat1a knockout mice (PMID: 12060674), less alcohol-induced liver injury in SAM-e-treated baboons (PMID: 12163146), more cisplatin kidney damage with SAM-e in mice (PMID: 19064128), and circadian disruption by excess SAM-e in cell lines and mice (PMID: 35383287).

In two Italian multicenter trials, SAM-e caused significantly fewer adverse events than the antidepressant imipramine (PMID: 12418499). In a 6-month Alzheimer's disease trial of 63 people, safety did not differ from placebo (PMID: 41980907). In a trial of SAM-e added to antidepressants in 107 patients, SAM-e caused no significant adverse effects beyond placebo, although one participant had serotonin syndrome-like symptoms (PMID: 30115553). In 20 healthy volunteers given oral or intravenous SAM-e for 5 days, no adverse events were reported (PMID: 19302903).

Mania and other psychiatric reactions

• Bipolar disorder. In a 1989 British series of 29 patients given SAM-e or placebo, 9 of 11 bipolar patients, all of them on SAM-e, switched into hypomania, mania or euphoria, while no unipolar patient did (PMID: 2673478).
• Trial counts. The 2016 Cochrane review counted two reports of mania or hypomania among 441 participants given SAM-e (PMID: 27727432).
• Case reports. Published cases describe a manic episode triggered by SAM-e (PMID: 15014726), mania with psychotic features in a patient taking SAM-e with an SSRI antidepressant (PMID: 29950497), and hypomania in a 42-year-old man about a week after he started SAM-e together with 5-HTP (PMID: 29955516). Another report describes a 61-year-old woman with no history of suicidal thoughts who attempted suicide 4 days after starting SAM-e on her own for depressive symptoms (PMID: 22210371).
• Schizophrenia trial. In an 18-patient trial, two patients on SAM-e had some worsening of irritability (PMID: 18824331).

Serotonin-related reactions

A 1993 letter reported a toxic interaction between SAM-e and the antidepressant clomipramine (PMID: 8434674). One participant in a 2018 trial of SAM-e added to antidepressants had serotonin syndrome-like symptoms (PMID: 30115553).

Blood tests

In a Mayo Clinic trial in 52 healthy volunteers, 4 weeks of SAM-e at 800 mg a day did not change plasma homocysteine compared with placebo; alanine aminotransferase, a liver enzyme, rose slightly at week 2 and total cholesterol fell (PMID: 19422296). A single 400 mg oral dose did not change plasma homocysteine over 24 hours in 14 healthy adults (PMID: 9262350).

Observational and animal signals

• Breast cancer chemotherapy (observational). In a retrospective study of 1,013 women given adjuvant chemotherapy for breast cancer at one hospital in Zhengzhou, China, SAM-e use during chemotherapy was associated with shorter overall and disease-free survival, and the association held for use of 14 days or longer after propensity matching (PMID: 42243831). A retrospective study of this kind shows an association, not cause and effect.
• Cisplatin (mice). In mice given the cancer drug cisplatin, SAM-e increased kidney dysfunction, measured as blood urea nitrogen and creatinine, without changing cisplatin's anticancer effect (PMID: 19064128).
• Circadian rhythms (cells and mice). Excess SAM-e disrupted circadian rhythms in cell lines and changed activity rhythms in mice (PMID: 35383287).

Product quality and gaps

Reviews have noted that SAM-e supplement products vary in quality (PMID: 33900393, PMID: 18246887). The longest placebo-controlled trial summarized on this page ran for 2 years (PMID: 10406187). SAM-e has been given to pregnant women in trials for intrahepatic cholestasis of pregnancy; the 2020 Cochrane review of drug treatments for that condition included 26 trials, mostly small and at unclear to high risk of bias (PMID: 32716060).