Methotrexate

What it is

Methotrexate is a synthetic folic acid analogue (an antifolate) taken as a weekly tablet, oral solution or injection. The US label describes it as a dihydrofolate reductase inhibitor indicated for several cancers, rheumatoid arthritis, polyarticular juvenile idiopathic arthritis and severe psoriasis. The same molecule is used at very different doses for these different jobs. It is a cytotoxic drug and the label requires special handling.

What the research says

For rheumatoid arthritis the evidence is reasonably strong and reasonably well quantified: a Cochrane review of 7 placebo-controlled trials (732 people) found about 15 more people in 100 reached a major response (ACR50) on methotrexate than on placebo at 52 weeks, with a number needed to treat of 7 - and about 9 more people in 100 stopped the drug because of side effects. For psoriasis and juvenile arthritis the trial base is thinner. Outside its approved uses, a large placebo-controlled trial (CIRT, 4,786 people) found low-dose methotrexate did not reduce cardiovascular events, and a placebo-controlled trial in tubal ectopic pregnancy found no significant benefit over expectant management at low hCG levels. It carries a long boxed warning and documented risks to the liver, blood counts, lungs and a fetus.

Evidence grade: Well established.

How it works

Drug class: Antifolate antimetabolite; dihydrofolate reductase inhibitor. Used as a cytotoxic chemotherapy agent at higher doses and as a conventional synthetic disease-modifying antirheumatic drug (csDMARD) at low once-weekly doses.

Methotrexate blocks the enzyme dihydrofolate reductase, which cells need to turn folate into the active form used to build DNA and RNA. Fast-dividing tissues - cancer cells, bone marrow, gut lining, a fetus - are hit hardest. Why it helps rheumatoid arthritis and psoriasis at low weekly doses is not explained by this, and the label says so outright. (Source 1)

Boxed warning

Methotrexate tablets can cause embryo-fetal toxicity, including fetal death. For non-neoplastic diseases, methotrexate tablets are contraindicated in pregnancy. For neoplastic diseases, advise females and males of reproductive potential to use effective contraception [see Contraindications (4), Warnings and Precautions (5.1), Use in Specific Populations (8.1, 8.3)]. Methotrexate tablets are contraindicated in patients with a history of severe hypersensitivity reactions to methotrexate, including anaphylaxis [Contraindications (4), Warnings and Precautions (5.2)]. Serious adverse reactions, including death, have been reported with methotrexate. Closely monitor for adverse reactions of the bone marrow, gastrointestinal tract, liver, lungs, skin, and kidneys. Withhold or discontinue methotrexate tablets as appropriate [Warnings and Precautions (5.3, 5.4, 5.5, 5.6, 5.7, 5.8)]. Methotrexate tablets when inadvertently administered once daily have resulted in death [Warnings and Precautions (5.9)].

(Source 2)

What it is used for

• Seven placebo-controlled trials pooled in a 2014 Cochrane review show a clinically important benefit over 12 to 52 weeks: ACR50 risk ratio 3.0, number needed to treat 7, 15 more responders per 100. Radiographic progression was also lower (RR 0.31, NNT 13). No one in the single trial measuring remission reached remission criteria. Evidence: established. (Source 3)
• Approved on the label, but the comparative trial base is small. A 2026 meta-analysis of four randomised trials (247 people) comparing methotrexate with cyclosporine found no significant difference in PASI 75 or PASI 90, with high heterogeneity; methotrexate was far more likely to raise liver enzymes (RR 13.35). Evidence: limited. (Source 4)
• On the label for pediatric patients. We did not reach a systematic review of placebo-controlled trials in this population in this session; the label reports adverse reaction rates in children (elevated liver tests 14%, gastrointestinal reactions 11%) rather than pooled efficacy. Evidence: unknown. (Source 5)
• Approved as part of combination chemotherapy regimens rather than as a single agent for most of these. We relied on the label for the indication wording and did not retrieve the pivotal oncology trials in this session. Evidence: established. (Source 5)
• Widely used worldwide, but the only placebo-controlled randomised trial we found (80 women with beta-hCG below 1500 IU/L) reported success in 83% on methotrexate versus 76% on placebo, a difference that was not statistically significant (P = 0.47). The authors say their results do not support routine use at low hCG. Evidence: disputed. (Source 6)
• The CIRT trial randomised 4,786 people with prior myocardial infarction or multivessel coronary disease to methotrexate 15-20 mg weekly or placebo. It was stopped early; the primary endpoint rate was 4.13 versus 4.31 per 100 person-years (hazard ratio 0.96, 95% CI 0.79 to 1.16) and inflammatory markers did not fall. Evidence: not-supported. (Source 7)

Interactions

• Folic acid and folinic acid (folate supplements) (label): In cancer treatment folate supplements can blunt methotrexate's effect, because methotrexate and reduced folates compete for the same transporter into cells. In rheumatoid arthritis, juvenile arthritis and psoriasis the label takes the opposite line and tells prescribers to give folate deliberately, to reduce side effects. (Source 8)
• Folic acid and folinic acid (folate supplements), in rheumatoid arthritis (clinical trial): Low-dose folate alongside weekly methotrexate reduced gut side effects and abnormal liver enzymes and reduced the chance of stopping the drug, without measurably reducing its effect on joints. (Source 9)
• Alcohol (cohort study): Heavy drinking increases the risk of methotrexate liver injury. A UK primary-care cohort put numbers on it: above 21 units a week the rate of significant liver enzyme rises was about 85% higher; below 14 units a week no increase was detected. (Source 10)
• Alcohol (label position) (label): The label states plainly that heavy alcohol consumption increases the risk of methotrexate hepatotoxicity, which can be irreversible. (Source 11)
• Aspirin and other non-steroidal anti-inflammatory drugs (label): Aspirin and other NSAIDs appear by name in the label's section 7.1 list of products that may raise methotrexate blood levels and so raise the risk of severe methotrexate toxicity. The label says to monitor closely if the combination cannot be avoided. (Source 8)
• Proton pump inhibitors and probenecid (label): Proton pump inhibitors and probenecid both appear as items in the label's section 7.1 list of products that may raise methotrexate plasma concentrations. The quoted sentence is the lead statement that governs that whole list; the individual items are separate bullet lines in the same section. (Source 8)
• Highly protein-bound drugs (warfarin and other oral anticoagulants, phenytoin, salicylates, sulfonamides, sulfonylureas, tetracyclines) (label): Methotrexate is itself protein-bound; drugs that compete for the same binding sites can displace it and raise free methotrexate levels. The label names this group as a single bullet item in its list of products that increase methotrexate exposure. (Source 8)
• Other antifolate drugs (dapsone, pemetrexed, pyrimethamine, sulfonamides including trimethoprim-sulfamethoxazole) (label): These block the same folate pathway methotrexate blocks, so the effects stack - more marrow suppression and mucosal toxicity. The label names them as one bullet item among products that increase methotrexate exposure and toxicity. (Source 8)
• Nitrous oxide (anaesthetic gas) (label): Nitrous oxide interferes with the same folate-dependent pathways methotrexate blocks, which can amplify methotrexate toxicity. The label says to avoid it in people taking methotrexate. (Source 8)

Stopping it

• Methotrexate has no described withdrawal or dependence syndrome in the sources we read. What the literature describes instead is loss of disease control. In six non-inferiority randomised trials of people with rheumatoid arthritis who were already at treatment target on a biologic plus methotrexate, stopping or tapering methotrexate raised DAS28 by 0.20 points and reduced the proportion in low disease activity (RR 0.88), but formal remission rates did not fall. (Source 12)
• The same review's authors summarise the trade-off: withdrawal slightly increases disease activity, and matters least for people in deep remission. (Source 13)
• Stopping is also what the label requires when toxicity appears: it directs prescribers to withhold or discontinue methotrexate for bone marrow, liver, lung, skin, gut or kidney reactions, weighing the severity of the disease being treated against the severity of the reaction. (Source 11)
• After an overdose, or when methotrexate must be cleared urgently, leucovorin or levoleucovorin is the antidote and glucarpidase is used for toxic plasma levels with impaired kidney function. Dialysis has not been shown to help. (Source 14)

What goes wrong

Methotrexate roughly doubled withdrawals for adverse events and tripled the overall rate of any side effect compared with placebo in rheumatoid arthritis trials. (Source 15)

• Systematic review, Moderate certainty.
• Size: 732 participants across 7 trials.
• Who: adults with rheumatoid arthritis.
• How long: 12 to 52 weeks.
• Result: withdrawal for adverse events 16% vs 8% (RR 2.1, 95% CI 1.3 to 3.3; NNH 13); any adverse event at 12 weeks 45% vs 15% (RR 3.0, 95% CI 1.4 to 6.4); serious adverse events 3% vs 2%, not statistically different.
• Funding: not stated in the abstract; Cochrane review.

Limit of this finding: One absolute figure in this paragraph is printed as '(ATB 30, 95% CI 13% to 47%)', without the per-cent sign that every neighbouring figure carries; it means 30%. The harm figures themselves reconcile: 16% versus 8% withdrawal with a risk ratio of 2.1, and 45% versus 15% for any adverse event with a risk ratio of 3.0.

Patients in the MTX monotherapy group were twice as likely to discontinue from the study due to adverse events compared to patients in the placebo group, at 12 to 52 weeks (16% versus 8%; RR 2.1, 95% CI 1.3 to 3.3; NNT 13, 95% CI 6 to 44). Compared to placebo, nine more people out of 100 who took MTX withdrew from the studies because of side effects (ATB 9%, 95% CI 3% to 14%). Total adverse event rates at 12 weeks were higher in the MTX monotherapy group compared to the placebo group (45% versus 15%; RR 3.0, 95% CI 1.4 to 6.4; NNT 4, 95% CI 2 to 17). Thirty more people out of 100 who took MTX compared to those who took placebo experienced any type of side effect (common or rare) (ATB 30, 95% CI 13% to 47%). No statistically significant differences were observed in the total number of serious adverse events between the MTX group and the placebo group at 27 to 52 weeks. Three people out of 100 who took MTX alone experienced rare but serious side effects compared to 2 people out of 100 who took a placebo (3% versus 2%, respectively).

In the same cardiovascular trial, methotrexate raised liver enzymes, lowered white cells and haematocrit, and produced more non-basal-cell skin cancers than placebo. (Source 7)

• Randomized trial, High certainty.
• Size: 4,786 patients.
• Who: adults with stable atherosclerosis, all given 1 mg folate daily.
• How long: median 2.3 years.
• Result: elevations in liver-enzyme levels, reductions in leukocyte counts and haematocrit, higher incidence of non-basal-cell skin cancers (rates not given in the abstract)
• Funding: National Heart, Lung, and Blood Institute (independent of industry)

Methotrexate was associated with elevations in liver-enzyme levels, reductions in leukocyte counts and hematocrit levels, and a higher incidence of non-basal-cell skin cancers than placebo.

People with psoriasis taking methotrexate had about 2.8 times the odds of non-melanoma skin cancer compared with those not taking it, in pooled observational studies. (Source 16)

• Systematic review, Low certainty.
• Size: 11,875 patients with psoriasis across 9 cohort and case-control studies, 2,192 of them on methotrexate.
• Who: people with psoriasis.
• How long: varied, observational follow-up.
• Result: odds ratio 2.8 (95% CI 1.47 to 5.39; p = 0.002)
• Funding: not stated in the abstract.

Limit of this finding: This pools observational studies - comparative and case-control, searched only to June 2019 - and the paper's own closing sentence states the risk as though methotrexate caused it. Data of this kind cannot show that: people prescribed methotrexate differ from those who are not, including in disease severity and in past phototherapy and ciclosporin use, both of which raise skin cancer risk on their own. Read it as an association that needs explaining, not as a cause.

A meta-analysis demonstrated an odds ratio of 2.8 (95% confidence interval = 1.47-5.39; p = 0.002) for nonmelanoma skin cancer development in patients with psoriasis taking methotrexate compared with those not taking methotrexate.

Drinking more than 21 units of alcohol a week while on methotrexate was associated with nearly double the rate of significant liver enzyme rises; below 14 units a week no increase was detected. (Source 10)

• Cohort study, Low certainty.
• Size: 11,839 patients, 530 episodes of transaminitis in 47,090 person-years.
• Who: adults with rheumatoid arthritis starting methotrexate in UK primary care, 1987-2016.
• How long: up to 29 years of routinely collected data.
• Result: adjusted HR 1.01 (1.00 to 1.02) per unit per week; more than 21 units/week adjusted HR 1.85 (95% CI 1.17 to 2.93); 15-21 units/week a possible increase.
• Funding: not stated in the abstract.

Limit of this finding: One of the two alcohol figures is weaker than the paper's wording suggests. The per-unit hazard ratio of 1.01 has a lower confidence limit of exactly 1.00, which is no increase at all, yet the sentence calls it an increased risk. The figure that carries the finding is the one for more than 21 units a week, 1.85 (1.17 to 2.93). This is a cohort study, so it shows an association; it does not establish that the alcohol caused the liver enzyme rises.

Increased weekly alcohol consumption as a continuous variable was associated with increased risk of transaminitis, adjusted HR (95% CI) per unit consumed 1.01 (1.00 to 1.02); consuming between 15 and 21 units was associated with a possible increased risk of hepatotoxicity, while drinking >21 units per week significantly increased rates of transaminitis, adjusted HR (95% CI) 1.85 (1.17 to 2.93).

In the label's own paediatric trial data, elevated liver tests were reported in 14% of children treated with weekly methotrexate for juvenile idiopathic arthritis, and gastrointestinal reactions in 11%. (Source 17)

• Official position, Certainty not rated.
• Size: patients 2 to 18 years of age with polyarticular juvenile idiopathic arthritis (number not given in this section)
• Who: children and adolescents aged 2 to 18 with pJIA, most also taking NSAIDs and some corticosteroids.
• How long: not stated in this section.
• Result: elevated liver tests 14%; gastrointestinal reactions (nausea, vomiting, diarrhea) 11%; stomatitis 2%; leukopenia 2%; headache 1.2%; alopecia 0.5%; dizziness 0.2%; rash 0.2%. These are rates in the treated children only; the label gives no placebo comparison for them.
• Funding: regulatory label (manufacturer-submitted data), Teva Women's Health, SPL version 14, published 29 June 2026.

were as follows: elevated liver tests 14%; gastrointestinal reactions (e.g., nausea, vomiting, diarrhea) 11%; stomatitis 2%; leukopenia 2%; headache 1.2%; alopecia 0.5%; dizziness 0.2%; rash 0.2%.

The methotrexate label's own list of common adverse reactions is ulcerative mouth inflammation, low white cells, nausea and abdominal distress. (Source 18)

• Official position, Certainty not rated.
• Size: not stated in this part of the label.
• Who: patients taking methotrexate across its approved uses.
• How long: not stated in this part of the label.
• Result: no frequency is attached to this list; the label gives frequencies separately for each indication, and for rheumatoid arthritis it does so in a two-column incidence table.
• Funding: regulatory label (manufacturer-submitted data), Teva Women's Health, SPL version 14, published 29 June 2026.

Limit of this finding: The passage recorded here ends with the three bare incidence bands of the rheumatoid arthritis table that follows it, because that table is laid out in two columns - bands on the left, reactions on the right - and the two columns cannot be reproduced as a single run of continuous text. The label's own pairing is: 10% or more - elevated liver tests 15%, nausea or vomiting 10%; 3% to under 10% - stomatitis and low platelets; 1% to under 3% - rash, itching or dermatitis, diarrhoea, hair loss, low white cells, pancytopenia and dizziness. No frequency band should be read off the quoted text, and the list quoted above carries no frequencies at all.

Common adverse reactions were: ulcerative stomatitis, leukopenia, nausea, and abdominal distress. Other clinically relevant adverse reactions were infection, malaise, fatigue, chills, fever, and dizziness.

In the label's adult rheumatoid arthritis trials, elevated liver tests were reported in 15% of patients and nausea or vomiting in 10%. (Source 19)

• Official position, Certainty not rated.
• Size: 128 patients with rheumatoid arthritis, plus a second set of two trials in 680 patients.
• Who: adults with rheumatoid arthritis given methotrexate 7.5 mg to 15 mg orally once weekly, most also taking NSAIDs and some corticosteroids.
• How long: 12- to 18-week double-blind studies.
• Result: in the 10%-or-more band, elevated liver tests 15% and nausea/vomiting 10%; in the 3% to under 10% band, stomatitis and thrombocytopenia; in the 1% to under 3% band, rash/pruritus/dermatitis, diarrhea, alopecia, leukopenia, pancytopenia and dizziness; interstitial pneumonitis 1% in the second set of two trials (n=680)
• Funding: regulatory label (manufacturer-submitted data), Teva Women's Health, SPL version 14, published 29 June 2026.

Limit of this finding: The label prints these rheumatoid arthritis reactions in a two-column table, with the incidence bands in the left column and the reactions in the right, and the two columns cannot be reproduced as a single run of continuous text. The label's own pairing is: 10% or more - elevated liver tests 15%, nausea or vomiting 10%; 3% to under 10% - stomatitis and low platelets; 1% to under 3% - rash, itching or dermatitis, diarrhoea, hair loss, low white cells, pancytopenia and dizziness. The quoted text therefore carries only the reactions and any percentage printed next to the reaction itself, and no frequency band should be read off it. These are rates observed in the label's own short trials, not placebo-subtracted rates.

Elevated liver tests 15%, nausea/vomiting 10%

Methotrexate can cause pulmonary toxicity, including interstitial pneumonitis that is sometimes irreversible or fatal. (Source 20)

• Official position, Certainty not rated.
• Size: not quantified in this section; the label's rheumatoid arthritis trials (n=680) reported interstitial pneumonitis at an incidence of 1%.
• Who: people taking methotrexate for any indication.
• How long: can occur acutely or chronically.
• Result: no rate given in this section; cases described as acute or chronic interstitial pneumonitis, some irreversible or fatal.
• Funding: regulatory label (manufacturer-submitted data)

Pulmonary toxicity, including acute or chronic interstitial pneumonitis and irreversible or fatal cases, can occur with TREXALL

The FDA label carries a boxed warning covering embryo-fetal toxicity, severe hypersensitivity, fatal organ toxicity and fatal dosing errors. (Source 2)

• Official position, Certainty not rated.
• Size: not applicable.
• Who: all users.
• How long: not applicable.
• Result: no rates given; the warning states serious adverse reactions including death have been reported, and that once-daily dosing by mistake has caused death.
• Funding: FDA-approved label, Teva Women's Health, SPL version 14, published 29 June 2026.

Serious adverse reactions, including death, have been reported with methotrexate. Closely monitor for adverse reactions of the bone marrow, gastrointestinal tract, liver, lungs, skin, and kidneys. Withhold or discontinue methotrexate tablets as appropriate

What the evidence supports

In placebo-controlled trials in rheumatoid arthritis, methotrexate monotherapy tripled the chance of a major (ACR50) response, with 15 more responders per 100 people treated. (Source 3)

• Systematic review, Moderate certainty.
• Size: 732 participants across 7 trials.
• Who: adults with rheumatoid arthritis who had failed prior treatment, mean disease duration 1 to 14 years.
• How long: 12 to 52 weeks.
• Result: ACR50 at 52 weeks RR 3.0 (95% CI 1.5 to 6.0); NNT 7 (95% CI 4 to 22); absolute treatment benefit 15% (95% CI 8% to 23%)
• Funding: not stated in the abstract; Cochrane review.

Limit of this finding: Not every statistic in this section of the review holds together. A later sentence calls a physical-function result a statistically significant improvement while printing an odds ratio of 2.8 with a 95% confidence interval of 0.23 to 32.2, a range so wide that it includes no effect at all and so cannot be significant. The ACR 50 figures quoted here are internally consistent and are the ones to rely on; the physical-function odds ratio is not.

MTX monotherapy showed a clinically important and statistically significant improvement in the American College of Rheumatology (ACR) 50 response rate when compared with placebo at 52 weeks (RR 3.0, 95% confidence interval (CI) 1.5 to 6.0; number needed to treat (NNT) 7, 95% CI 4 to 22). Fifteen more patients out of 100 had a major improvement in the ACR 50 outcome compared to placebo (absolute treatment benefit (ATB) 15%, 95% CI 8% to 23%).

Radiographic progression of joint erosion was less frequent on methotrexate than placebo, with 8 fewer people per 100 showing progression. (Source 3)

• Systematic review, Moderate certainty.
• Size: 732 participants across 7 trials.
• Who: adults with rheumatoid arthritis.
• How long: 12 to 52 weeks.
• Result: RR 0.31 (95% CI 0.11 to 0.86); NNT 13 (95% CI 10 to 60); absolute treatment benefit -8% (95% CI -16% to -1%)
• Funding: not stated in the abstract; Cochrane review.

Limit of this finding: The review's own signs do not line up. It says eight more people out of 100 showed less joint damage, but prints the absolute treatment benefit as a negative number, -8% (95% CI -16% to -1%), while writing its other benefits as positive numbers. The direction comes from the risk ratio of 0.31, which means fewer people on methotrexate had worsening erosions; the minus sign is not a benefit for placebo.

radiographic progression rates (measured by an increase in erosion scores of more than 3 units on a scale ranging from 0 to 448) were statistically significantly lower for patients in the MTX group compared with placebo-treated patients (RR 0.31, 95% CI 0.11 to 0.86; NNT 13, 95% CI 10 to 60). Eight more patients out of 100 showed less damage to joints measured by an increase in erosion scores compared to placebo (ATB -8%, 95% CI -16% to -1%).

Taking folic or folinic acid alongside methotrexate for rheumatoid arthritis cut gastrointestinal side effects by 9 percentage points and abnormal liver enzymes by 16 percentage points. (Source 9)

• Systematic review, Moderate certainty.
• Size: 624 patients across 6 trials.
• Who: adults with rheumatoid arthritis on methotrexate 25 mg/week or less, with low-dose folate (7 mg weekly or less)
• How long: not stated in the abstract.
• Result: GI side effects RR 0.74 (95% CI 0.59 to 0.92), 26% relative / 9% absolute risk reduction; abnormal transaminases RR 0.23 (95% CI 0.15 to 0.34), 76.9% relative / 16% absolute; withdrawal for any reason RR 0.39 (95% CI 0.28 to 0.53), 15.2% absolute.
• Funding: not stated in the abstract; Cochrane review, GRADE moderate (low for haematologic outcomes)

For patients supplemented with any form of exogenous folate (either folic or folinic acid) whilst on MTX therapy for rheumatoid arthritis, a 26% relative (9% absolute) risk reduction was seen for the incidence of GI side effects such as nausea, vomiting or abdominal pain (RR 0.74, 95% CI 0.59 to 0.92; P = 0.008). Folic and folinic acid also appear to be protective against abnormal serum transaminase elevation caused by MTX, with a 76.9% relative (16% absolute) risk reduction (RR 0.23, 95% CI 0.15 to 0.34; P < 0.00001), as well as reducing patient withdrawal from MTX for any reason (60.8% relative (15.2% absolute) risk reduction, RR 0.39, 95% CI 0.28 to 0.53; P < 0.00001).

What the evidence does not support

Methotrexate did not improve the mental component of quality of life, and in the one trial that measured remission nobody on methotrexate or placebo reached remission. (Source 3)

• Systematic review, Moderate certainty.
• Size: 732 participants across 7 trials; remission measured in 1 trial.
• Who: adults with rheumatoid arthritis.
• How long: 12 to 52 weeks.
• Result: no clinically important or statistically significant difference on the SF-36 mental component; 0 participants in either arm met remission criteria.
• Funding: not stated in the abstract; Cochrane review.

Limit of this finding: The same paragraph says no statistically significant differences were found in the radiographic scores and then that radiographic progression rates were significantly lower on methotrexate. Both are published as they stand and they measure different things: the average change in score, and the proportion of people crossing a worsening threshold. Neither statement cancels the other, and neither should be quoted alone as the review's finding on joint damage.

No clinically important or statistically significant differences were observed in the SF-36 mental component.Although no statistically significant differences were observed in radiographic scores (that is, Total Sharp score, erosion score, joint space narrowing), radiographic progression rates (measured by an increase in erosion scores of more than 3 units on a scale ranging from 0 to 448) were statistically significantly lower for patients in the MTX group compared with placebo-treated patients (RR 0.31, 95% CI 0.11 to 0.86; NNT 13, 95% CI 10 to 60). Eight more patients out of 100 showed less damage to joints measured by an increase in erosion scores compared to placebo (ATB -8%, 95% CI -16% to -1%). In the one study measuring remission, no participants in either group met the remission criteria.

Low-dose methotrexate did not reduce cardiovascular events or inflammatory markers in people with stable atherosclerosis. (Source 7)

• Randomized trial, High certainty.
• Size: 4,786 patients.
• Who: adults with previous myocardial infarction or multivessel coronary disease plus type 2 diabetes or metabolic syndrome.
• How long: median follow-up 2.3 years (trial stopped early)
• Result: primary endpoint 4.13 vs 4.31 per 100 person-years, hazard ratio 0.96 (95% CI 0.79 to 1.16); original primary endpoint HR 1.01 (95% CI 0.82 to 1.25); no reduction in IL-1beta, IL-6 or CRP.
• Funding: National Heart, Lung, and Blood Institute (independent of industry)

The trial was stopped after a median follow-up of 2.3 years. Methotrexate did not result in lower interleukin-1β, interleukin-6, or C-reactive protein levels than placebo. The final primary end point occurred in 201 patients in the methotrexate group and in 207 in the placebo group (incidence rate, 4.13 vs. 4.31 per 100 person-years; hazard ratio, 0.96; 95% confidence interval [CI], 0.79 to 1.16). The original primary end point occurred in 170 patients in the methotrexate group and in 167 in the placebo group (incidence rate, 3.46 vs. 3.43 per 100 person-years; hazard ratio, 1.01; 95% CI, 0.82 to 1.25).

Folate supplementation did not significantly reduce mouth sores, and did not measurably reduce methotrexate's effectiveness in rheumatoid arthritis. (Source 9)

• Systematic review, Moderate certainty.
• Size: 624 patients across 6 trials.
• Who: adults with rheumatoid arthritis on methotrexate.
• How long: not stated in the abstract.
• Result: stomatitis / mouth sores RR 0.72 (95% CI 0.49 to 1.06), not statistically significant; no statistically significant effect on methotrexate efficacy measures.
• Funding: not stated in the abstract; Cochrane review.

We analysed the effect of folic or folinic acid on the incidence of stomatitis / mouth sores, and whilst showing a trend towards reduction in risk, the results were not statistically significant (RR 0.72, 95% CI 0.49 to 1.06)It was not possible to draw meaningful conclusions on the effect of folic or folinic acid on haematologic side effects of methotrexate due to small numbers of events and poor reporting of this outcome in included trials.It does not appear that supplementation with either folic or folinic acid has a statistically significant effect on the efficacy of MTX in treating RA (as measured by RA disease activity parameters such as tender and swollen joint counts, or physician's global assessment scores).

A single dose of methotrexate was not significantly better than placebo for tubal ectopic pregnancy when hCG was below 1500 IU/L. (Source 6)

• Randomized trial, Low certainty.
• Size: 80 women (42 methotrexate, 38 placebo)
• Who: clinically stable women with ultrasound-diagnosed tubal ectopic pregnancy and serum beta-hCG below 1500 IU/L, two UK units.
• How long: recruited 2005-2014; follow-up to hCG resolution.
• Result: success 83% vs 76%, chi-square 0.53, P = 0.47; median hCG resolution 17.5 vs 14 days, P = 0.73; in the 14 women with hCG 1000-1500 IU/L success was 33% expectant vs 62% methotrexate, not statistically significant.
• Funding: not stated in the abstract.

The rates of success were similar for the two study arms: 83% with methotrexate and 76% with placebo.

Where the evidence is mixed

Methotrexate and cyclosporine cleared moderate-to-severe plaque psoriasis to a similar degree in four small trials, and the review reported more elevated liver enzymes with methotrexate, though it published no confidence interval for that estimate. (Source 4)

• Meta-analysis, Low certainty.
• Size: 247 participants across 4 randomised trials (methotrexate 128, cyclosporine 119)
• Who: adults with moderate-to-severe chronic plaque psoriasis.
• How long: 4 to 12 weeks follow-up reported.
• Result: PASI 75 RR 0.81 (95% CI 0.56 to 1.18, P = .27, I2 = 80%); PASI 90 RR 1.03 (95% CI 0.50 to 2.11, P = .94, I2 = 65%); elevated liver enzymes RR 13.35.
• Funding: not stated in the abstract.

Limit of this finding: This abstract contains published errors that should not be taken at face value. Its PASI reduction result reads 'mean difference = -1.15 [95% CI: -0.23 to 2.53]', but a confidence interval has to contain its own estimate and -1.15 lies outside that range, so either the estimate or one limit is wrong in the paper. The same sentence says there was no difference between the two treatments and then that the difference was not statistically significant. One estimate is given for 4, 8 and 12 weeks together, so nothing can be read off it for any single timepoint. And the 13.35 risk ratio for elevated liver enzymes is printed with no confidence interval and no p-value, from four small trials, so its precision is unknown: treat it as a direction, not a size.

Pooled analysis showed no significant difference in PASI 75 (risk ratio [RR] = 0.81; 95% confidence interval [CI]: 0.56-1.18;

Stopping or tapering methotrexate in people already at treatment target on a biologic slightly worsened disease activity but did not reduce remission rates. (Source 12)

• Meta-analysis, Moderate certainty.
• Size: 1,430 patients across 6 non-inferiority randomised trials (734 withdrawal, 696 continuation)
• Who: adults with rheumatoid arthritis at treatment target on bDMARD/tsDMARD plus methotrexate.
• How long: not stated in the abstract.
• Result: DAS28 increased by 0.20 (95% CI 0.09 to 0.32, I2 = 0%); low disease activity RR 0.88 (0.80 to 0.97); remission by DAS28 RR 0.90 (0.81 to 1.01), SDAI RR 0.93 (0.77 to 1.11), CDAI RR 0.90 (0.74 to 1.11), ACR/EULAR Boolean RR 0.95 (0.70 to 1.29)
• Funding: not stated in the abstract; PROSPERO CRD42022303891.

Compared with continuing combination therapy, tapering off or discontinuing MTX increased DAS28 by 0.20 (95% CI 0.09, 0.32, I2 = 0%) and decreased the percentage of patients with LDA assessed by DAS28 to <3.2 [risk ratio (RR) 0.88 (0.80, 0.97), I2 = 0%]. However, MTX withdrawal did not decrease remission rates assessed by DAS28, SDAI, CDAI or ACR/EULAR Boolean remission [RR 0.90 (0.81, 1.01), 0.93 (0.77, 1.11), 0.90 (0.74, 1.11), 0.95 (0.70, 1.29), respectively].

Where the research disagrees

Whether methotrexate should be used routinely for a stable tubal ectopic pregnancy with low hCG

• Trial authors (Jurkovic et al., Ultrasound Obstet Gynecol 2017), placebo-controlled randomised trial, 80 women: The rates of success were similar for the two study arms: 83% with methotrexate and 76% with placebo. (Source 6)
• Worldwide clinical practice, as the same authors describe it, authors' description of prevailing practice, not a study: Methotrexate is used routinely worldwide for the medical treatment of clinically stable women with a tubal ectopic pregnancy. This is despite the lack of robust evidence to show its superior effectiveness over expectant management. (Source 21)

Whether folate supplements should be taken with methotrexate

• FDA label, neoplastic disease section, regulatory position, label dated 29 June 2026: Products containing folic acid or its derivatives may decrease the clinical effectiveness of methotrexate. (Source 22)
• Cochrane review of folate with methotrexate in rheumatoid arthritis, systematic review of 6 placebo-controlled trials, 624 patients, GRADE moderate: Folic and folinic acid also appear to be protective against abnormal serum transaminase elevation caused by MTX, with a 76.9% relative (16% absolute) risk reduction (Source 9)

How much

• Reference intake: Dosing is set by the prescriber and differs completely between uses. As a position, the Trexall label (Teva Women's Health, SPL version 14, published 29 June 2026) states a starting dose for rheumatoid arthritis of 7.5 mg orally once weekly, escalated to response. (Source 23)
• Upper limit: As a position, the same label states that weekly doses above 20 mg raise the risk of serious adverse reactions including myelosuppression. There is no general population upper intake level for methotrexate; it is a prescription cytotoxic drug, not a nutrient. (Source 23)
• Studied: The trials pooled in the 2014 Cochrane rheumatoid arthritis review used weekly doses between 5 mg and 25 mg. (Source 24)
• Studied: The CIRT cardiovascular trial used a target dose of 15 to 20 mg weekly with 1 mg of folate daily. (Source 25)
• Studied: The ectopic pregnancy trial used a single systemic injection of 50 mg/m2. (Source 6)

A common belief, and what the research shows

The belief: Methotrexate is just a chemotherapy drug, so the low weekly dose used for arthritis works the same way and carries the same risk.

What the research shows: The label's own mechanism section says the arthritis and psoriasis mechanism is not known: "The mechanism of action in rheumatoid arthritis and in psoriasis is unknown." And folate, which is deliberately avoided with cancer doses, is deliberately given with arthritis doses: "Administer folic acid or folinic acid for patients with rheumatoid arthritis, pJIA, and psoriasis". The risks are real but differently distributed - the Cochrane arthritis review found serious adverse events were not statistically different from placebo over 27 to 52 weeks (3% versus 2%), while withdrawals for side effects were twice as common.

Questions and answers

What is it?

Methotrexate is a synthetic copy of folic acid that jams the enzyme folate needs to work. It comes as a weekly tablet, oral solution, or injection. The FDA label classes it as a dihydrofolate reductase inhibitor and approves it for certain leukaemias and lymphomas, rheumatoid arthritis, polyarticular juvenile idiopathic arthritis and severe psoriasis. (Source 5)

What does it do in the body?

It blocks dihydrofolate reductase, so cells cannot make the active folate they need to build and repair DNA. Rapidly dividing tissues - tumour cells, bone marrow, gut and mouth lining, a developing fetus - are the most affected, which is both how it treats cancer and why its side effects cluster there. Why low weekly doses calm rheumatoid arthritis and psoriasis is a separate question the label says is unanswered. (Source 1)

Is it good or bad for you?

Both, depending on the setting and the dose. In rheumatoid arthritis the pooled trials show a real benefit - about 15 extra major responders per 100 people - but also 9 extra people per 100 stopping because of side effects. Outside its approved uses it has failed: in 4,786 people with heart disease it did not prevent cardiovascular events. It is dangerous in pregnancy and in overdose. (Source 26)

How do you get more of it?

Methotrexate is prescription-only; there is no food or supplement source and no way to raise your level other than a prescription. Dose is set by the prescriber: the label's starting dose for rheumatoid arthritis is 7.5 mg once weekly, escalated to response, with doses above 20 mg weekly carrying extra risk. Several commonly taken medicines - NSAIDs, aspirin, proton pump inhibitors, some antibiotics - raise methotrexate blood levels, which is a risk rather than a way to get more. (Source 23)

If it is harmful, what reduces it?

When methotrexate levels are too high, the antidote is leucovorin or levoleucovorin (reduced folate), given as soon as possible, with serum creatinine and methotrexate levels guiding it. Glucarpidase is used for toxic plasma concentrations above 1 micromole per litre when the kidneys are not clearing the drug. Hydration and urinary alkalinisation are given; ordinary dialysis has not been shown to help. (Source 14)

Why might someone be low in it or missing it?

The question of being low in methotrexate does not arise the way it does for a nutrient - nobody has a natural requirement for it. People end up off it because it was stopped for toxicity (liver, blood count, lung, skin, kidney reactions), because of pregnancy or planned pregnancy, or because it was tapered once their disease was controlled on another drug. The label directs prescribers to withhold or discontinue it when adverse reactions appear. (Source 27)

Which whole foods contain it or feed it?

No whole food contains methotrexate. The food-related fact that matters is the opposite direction: folate, which is abundant in leafy greens, legumes, liver and fortified flour, is the nutrient methotrexate blocks. In cancer treatment the label warns folate products can reduce methotrexate's effect; in rheumatoid arthritis, juvenile arthritis and psoriasis folic or folinic acid is prescribed on purpose to reduce side effects. (Source 22)

What happens if you do not have it?

Nothing happens from not having methotrexate in itself - it is a medicine, not a nutrient. For someone whose rheumatoid arthritis was controlled by it, the relevant evidence is what happens on stopping: in six randomised trials of people already at treatment target on a biologic, withdrawing methotrexate raised disease activity slightly and reduced the proportion in low disease activity, but did not reduce formal remission rates. (Source 13)

How can you test for it?

There is no test for whether you need methotrexate; the tests are safety tests done while you take it. The label directs blood counts at baseline, periodically during treatment and as clinically indicated, and liver tests on the same schedule, because liver fibrosis or cirrhosis can develop in psoriasis without symptoms or abnormal liver tests. Blood methotrexate levels and serum creatinine are measured in overdose or when clearance is delayed, not routinely. (Source 11)

References

1. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 12.1 Mechanism of Action. 2026. Read the source
2. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet, film coated - FDA label, section: WARNING: EMBRYO-FETAL TOXICITY, HYPERSENSITIVITY REACTIONS, SEVERE ADVERSE REACTIONS, AND RISK OF MEDICATION ERRORS (boxed warning, Full Prescribing Information body text). 2026. Read the source
3. The Cochrane Database of Systematic Reviews. Methotrexate for treating rheumatoid arthritis (Cochrane review) - abstract section: Benefits. 2014. PMID 24916606, DOI 10.1002/14651858.cd000957.pub2. Read the source
4. Journal of Cutaneous Medicine and Surgery. Comparative Efficacy of Methotrexate Versus Cyclosporine in the Treatment of Moderate-to-Severe Chronic Plaque Psoriasis: A Systematic Review and Meta-Analysis - full abstract (unstructured). 2026. PMID 41914632, DOI 10.1177/12034754261431774. Read the source
5. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 1 INDICATIONS AND USAGE (Highlights excerpt). 2026. Read the source
6. Ultrasound in Obstetrics & Gynecology. Single-dose systemic methotrexate vs expectant management for treatment of tubal ectopic pregnancy: a placebo-controlled randomized trial - abstract section: Results. 2017. PMID 27731538, DOI 10.1002/uog.17329. Read the source
7. The New England Journal of Medicine. Low-Dose Methotrexate for the Prevention of Atherosclerotic Events (CIRT trial) - abstract section: Results. 2019. PMID 30415610, DOI 10.1056/nejmoa1809798. Read the source
8. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 7.1 Effects of Other Drugs on Methotrexate. 2026. Read the source
9. The Cochrane Database of Systematic Reviews. Folic acid and folinic acid for reducing side effects in patients receiving methotrexate for rheumatoid arthritis (Cochrane review) - abstract section: Main results. 2013. PMID 23728635, DOI 10.1002/14651858.cd000951.pub2. Read the source
10. Annals of the Rheumatic Diseases. Quantifying the hepatotoxic risk of alcohol consumption in patients with rheumatoid arthritis taking methotrexate - abstract section: Results. 2017. PMID 28341765, DOI 10.1136/annrheumdis-2016-210629. Read the source
11. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 5.5 Hepatotoxicity. 2026. Read the source
12. Rheumatology (Oxford). Withdrawal of MTX in rheumatoid arthritis patients on bDMARD/tsDMARD plus methotrexate at target: a systematic review and meta-analysis - abstract section: Results. 2023. PMID 36125185, DOI 10.1093/rheumatology/keac515. Read the source
13. Rheumatology (Oxford). Withdrawal of MTX in rheumatoid arthritis patients on bDMARD/tsDMARD plus methotrexate at target: a systematic review and meta-analysis - abstract section: Conclusions. 2023. PMID 36125185, DOI 10.1093/rheumatology/keac515. Read the source
14. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 10 OVERDOSAGE. 2026. Read the source
15. The Cochrane Database of Systematic Reviews. Methotrexate for treating rheumatoid arthritis (Cochrane review) - abstract section: Harms. 2014. PMID 24916606, DOI 10.1002/14651858.cd000957.pub2. Read the source
16. Cureus. Psoriasis Patients Treated With Methotrexate Have an Increased Risk of Nonmelanoma Skin Cancer: A Systematic Review and Meta-Analysis - full abstract (unstructured). 2023. PMID 37153318, DOI 10.7759/cureus.37174. Read the source
17. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 6.1 Clinical Trials Experience: polyarticular juvenile idiopathic arthritis and psoriasis subsections. 2026. Read the source
18. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 6.1 Clinical Trials Experience: opening paragraphs and the rheumatoid arthritis preamble, ending with the incidence-band column of the adverse-reaction table. 2026. Read the source
19. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 6.1 Clinical Trials Experience: the adverse-reaction column of the rheumatoid arthritis table, the second controlled trial set (n=680) with its interstitial pneumonitis table, and the less common reactions. 2026. Read the source
20. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 5.6 Pulmonary Toxicity. 2026. Read the source
21. Ultrasound in Obstetrics & Gynecology. Single-dose systemic methotrexate vs expectant management for treatment of tubal ectopic pregnancy: a placebo-controlled randomized trial - abstract section: Objective. 2017. PMID 27731538, DOI 10.1002/uog.17329. Read the source
22. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 5.10 Folic Acid Supplementation. 2026. Read the source
23. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 2.3 Recommended Dosage for Rheumatoid Arthritis. 2026. Read the source
24. The Cochrane Database of Systematic Reviews. Methotrexate for treating rheumatoid arthritis (Cochrane review) - abstract section: Main results. 2014. PMID 24916606, DOI 10.1002/14651858.cd000957.pub2. Read the source
25. The New England Journal of Medicine. Low-Dose Methotrexate for the Prevention of Atherosclerotic Events (CIRT trial) - abstract section: Methods. 2019. PMID 30415610, DOI 10.1056/nejmoa1809798. Read the source
26. The Cochrane Database of Systematic Reviews. Methotrexate for treating rheumatoid arthritis (Cochrane review) - abstract section: Authors' conclusions. 2014. PMID 24916606, DOI 10.1002/14651858.cd000957.pub2. Read the source
27. DailyMed / Teva Women's Health LLC (SPL version 14, published 29 June 2026). TREXALL (methotrexate) tablet - FDA label, section 5.3 Myelosuppression. 2026. Read the source