Article Text

Download PDFPDF

Original research
Severe methotrexate toxicity in elderly patients under diuretics
  1. Cara Kumar1,
  2. Matthias Kuhn2,
  3. Kristine Herrmann1,
  4. Nicolai Leuchten1 and
  5. Martin Aringer1
  1. 1Department of Medicine III and interdisciplinary University Centre for Autoimmune and Rheumatic Entities (UCARE), University Medical Centre and Faculty of Medicine TU Dresden, Dresden, Germany
  2. 2Institute for Medical Informatics and Biometry (IMB), TU Dresden Faculty of Medicine Carl Gustav Carus, Dresden, Germany
  1. Correspondence to Professor Martin Aringer; martin.aringer{at}


Objectives To explore the toxicity of low-dose methotrexate (MTX), an uncommon, but life-threatening event.

Methods We analysed the presentation, course and risk factors of all patients admitted to the rheumatology ward with severe low-dose MTX toxicity. These patients were compared with patients without signs of relevant MTX toxicity.

Results The 12 patients admitted for MTX toxicity included 7 patients with rheumatoid arthritis, 2 with psoriatic arthritis or psoriasis, 2 patients with giant cell arteritis and 1 with myositis. 1 patient died from infections, while 11 survived under folinic acid administration. All patients suffering from severe MTX toxicity were older than 70 years and were therefore compared with 400 patients who were also older than 70 years, but without MTX toxicity. Of these 400 control patients, the group of patients not on MTX (n=232) had more renal impairment than the group of patients on MTX (n=168). Compared with the 168 MTX-treated patients without toxicity, the 12 patients with life-threatening toxic events had a lower median estimated glomerular filtration rate (eGFR) at the routine visit preceding the acute event (64 (range 32–77) vs 69 (range 8 to >90) mL/min x 1.73, p=0.0251). A multivariate analysis found that patients with toxicity were more frequently treated with diuretics (6/12 vs 24/168), proton pump inhibitors (PPIs; 10/12 vs 70/168) and levetiracetam (2/12 vs 1/168).

Conclusions Patients older than 70 years with lower eGFR and being on diuretics, but also on PPIs and levetiracetam, have a significantly higher risk for MTX toxicity.

  • Methotrexate
  • Arthritis, Rheumatoid
  • Arthritis, Psoriatic
  • Giant Cell Arteritis
  • Polymyositis

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information. Not applicable.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Methotrexate (MTX) can occasionally lead to severe bone marrow toxicity, particularly in elderly patients and in patients with renal insufficiency.


  • This study found severe MTX toxicity only in patients older than 70 years and identified having impaired renal function and taking non-hydrochlorothiazide diuretics, proton pump inhibitors and/or levetiracetam as risk factors for severe MTX toxicity.


  • The results of this study help identify patients at higher risk of life-threatening MTX toxicity who can be more safely treated by switching to other disease-modifying antirheumatic drugs.


Over four decades, low-dose methotrexate (MTX) has been the number one treatment for patients with rheumatoid arthritis (RA),1–3 also playing a major role in the therapy of patients with psoriatic arthritis,4 connective tissue diseases5 6 and vasculitides,7 8 though off-label for the latter two. In essence, the only common rheumatic disease clearly not improving under MTX is axial spondylarthritis,9 be it ankylosing spondylitis or non-radiographic axial spondylarthritis.

The success of MTX is not only based on its efficacy, but also on its long-established excellent long-term safety profile.10

The results of a large MTX Cardiovascular Inflammation Reduction Trial with 2391 MTX patients recently backed that serious adverse events under MTX are uncommon.11

Several serious issues originally discussed for MTX, such as lymphoma or lung fibrosis, have since been found to be caused by the underlying disease, not the drug itself.12–14 Liver cirrhosis, another potential MTX side effect, is avoided if liver enzymes are monitored.15 16 A slightly increased risk of developing non-melanoma skin cancer, presumably linked to MTX-induced ultraviolet (UV) light sensitivity, is the only remaining tumour risk.17

While some nausea and slight elevations of liver enzymes are fairly frequent,11 18 the acute life-threatening adverse events, namely MTX pneumonitis and agranulocytosis or pancytopenia, are rare events.17 19 MTX pneumonitis usually occurs early in therapy and is mostly reversible if diagnosed in time and managed correctly. As with essentially all other disease-modifying antirheumatic drugs, agranulocytosis is found rarely and rather early in the treatment course.

In contrast, severe bone marrow depression due to true MTX toxicity, usually accompanied by mucositis, can occur at any time under MTX therapy and is often critical.20 Such toxicity is usually either caused by accidental overdosing, often due to inadvertent daily administration of the weekly MTX dose,21 or by a decline in renal function, usually acute renal failure, leading to the accumulation of MTX, because it is mostly eliminated via the kidneys.22 23

There is limited information on the risk profile for MTX toxicity of these patients. We therefore analysed the available data of all 12 patients admitted to our rheumatology inpatient ward for severe MTX toxicity since 2011 and compared them with 400 rheumatology outpatients of similar age, but without severe MTX toxicity.

Patients and methods

All patients admitted since 2011 to the rheumatology inpatient ward for toxic bone marrow suppression with or without severe mucositis associated with low-dose MTX were identified from the inpatient records. In addition, one recent patient was manually added. Complete data on age, sex, diagnosis, MTX dose and concomitant medication and on relevant laboratory data were extracted from the patient records. Previous laboratory values on renal function were obtained from outpatient records or letters from treating physicians.

Given that all patients affected by severe MTX toxicity were older than 70 years, and had RA, psoriatic arthritis or psoriasis, giant cell arteritis (GCA), or myositis as the underlying diagnosis, patients with these diseases older than 70 years at the date of their last recorded outpatient visit or hospital admission were identified, and their data similarly extracted from the outpatient records.

The data on the one patient who succumbed to septicaemia following MTX toxicity were collected and pseudonymised by a treating physician. All other patients gave their informed consent to the pseudonymised analysis of their routine data.

All statistical analyses were performed with Graph Pad Prism V.9.5.1 and R V.4.3.1. Summary statistics were obtained, and normal distribution was tested by Kolmogorov-Smirnov tests. Given that none of the parameters followed Gaussian distributions in all populations, non-parametric Mann-Whitney U tests were used for continuous variables and Fisher’s exact test and χ2 test were employed for testing categorical variables in 2 by 2 and larger tables, respectively. Variables with a p value<0.05 were included into a multiple logistic regression analysis with Firth’s correction for bias reduction in small samples.24 Multiple imputation via predictive mean matching was used to recuperate two toxicity patients that lacked estimated glomerular filtration rate (eGFR) measurements before suffering from MTX toxicity. The parameter estimates were pooled via Rubin’s rule.


A total of 12 patients were admitted for MTX bone marrow toxicity, via the emergency department or via the intensive care unit. These patients’ age ranged between 72 and 89 years, and two-thirds of the patients were women (table 1). In addition to supportive therapy (and to terminating MTX), all patients received different high doses of folinic acid intravenously. In total, 11 patients recovered and survived to be discharged. One 88-year-old female patient treated with 15 mg of weekly MTX for psoriasis in an outside private practice unfortunately succumbed to septicaemia.

Table 1

Demographics, MTX doses and blood levels, mucositis and key laboratory parameters of the 12 patients admitted for low-dose MTX toxicity

All 12 patients had severe mucositis. Some degree of leucopenia also occurred in all patients, but minimal leucocyte counts ranged between 0.49 and 3.42 x109/L (or 490 to 3420 leucocytes/mm3)(table 1). In total, 10 of the 12 patients had thrombocytopenia with values ranging from 2 to 140 GPt/L (or 2000 to 140 000 platelets per mm3) in the thrombocytopenic patients. Other than from the mucositis, there were no relevant bleeding events. Five patients had septicaemia and four had pulmonary infiltrates, three of them were associated with leucocytes<1 GPt/L (table 2).

Table 2

Infections following methotrexate toxicity in the admitted patients

Prescribed weekly MTX doses ranged between 7.5 and 15 mg. Accidental MTX overdosing was suspected in one case (patient (Pt) 8). Plasma MTX levels were determined in 11 of the 12 patients shortly after admission, and 4 of 11 patients had low, but detectable levels (table 1). At admission, the eGFR ranged between 9 and 73 mL/min x 1.73 m² (table 1). Two patients had close to normal levels, that is, chronic kidney disease (CKD) grade 2, six patients had CKD grade 3, three patients had CKD grade 4 and one patient had CKD grade 5. In this acute situation, 9 of the 12 patients had an eGFR of 40 mL/min x 1.73 m² or lower, incompatible with further MTX treatment as per local guidelines, and 3 of these patients had values below 30 mL/min x 1.73 m² (CKD4). For 10 of the 12 patients, the last eGFR value before toxicity could be retrieved. Based on these values, six patients had CKD grade 2 and four patients had CKD grade 3.

Next, we investigated the concomitant medication of the patients. As shown in table 2, only 4 of the 12 patients took non-steroidal anti-inflammatory drugs (NSAIDs) or cyclo-oxygenase-2 inhibitors. Only one patient took metamizole on an as-needed basis. Despite the low rate of conventional NSAIDs, 10 patients took a proton pump inhibitor (PPI), most commonly pantoprazole. Seven patients were on ACE inhibitors and/or angiotensin receptor blockers, with one being on the combination of both. Three patients had low doses of hydrochlorothiazide (HCT), but six had non-HCT diuretics, mostly torasemide, in their medication. Two patients were on levetiracetam, an antiepileptic medication with a disputed impact on MTX levels.25 Levetiracetam was not accompanied by measurable MTX in the plasma of the one patient (Pt 2) whose MTX level was measured. Cotrimoxazole and penicillins were not taken by any of the 12 patients in the weeks before admission.

Since all of these medications are fairly common in patients older than 69, we formed a comparator cohort of 400 outpatients with either RA, psoriatic arthritis, GCA or autoimmune myositis more than 69 years old at the time of their most recent outpatient clinic visit or admission. Of these 400 patients, 168 were on low-dose MTX, 232 were not or not any more, on MTX (table 3). Patients with and without MTX were of similar age (median 78 (range 70–91) years and 79 (70–95) years, respectively (figure 1), but differed in their eGFR (median 69 (8 to >90) vs 60 (4 to >90) mL/min*1.73 m², p<0.0001).

Figure 1

Age and estimated glomerular filtration rate (eGFR) distribution in the three patient populations of ≥70 years. Patients without toxicity who had never been or were no longer treated with methotrexate (MTX; circles) or who at that time were treated with low-dose MTX (open triangle) did not differ from patients with MTX toxicity regarding age (A), but with regard to their renal function as determined by eGFR (B). *In all 4 patients with an eGFR<40 mL/min*1.73 m², administration of MTX was immediately terminated.

Table 3

Potentially relevant comedication in the 12 patients with severe methotrexate toxicity

In the MTX group, 5% (8/168) had a normal renal function (CKD1), 72% (121/168) had a mild decrease (CKD2) and 22% (37/168) had a moderate decrease (CKD3) in eGFR. Four patients had eGFR values below 40 mL/min x 1.73 m². MTX was terminated in all four patients and none of the four developed MTX toxicity, but two died within days after the measurement from urosepsis, not associated with MTX bone marrow toxicity or mucositis, and advanced carcinoma, respectively, with therapy limited to the best supportive care according to patient preference. In the non-MTX group, half of the patients had CKD grade 3 or higher (table 3).

Compared with the 12 patients admitted for MTX toxicity, the 168 MTX-treated control patients without toxicity were of similar age and had a similar disease distribution (table 4), but had a slightly higher median eGFR (69 vs 63.5 mL/min x 1.73 m², p=0.0251). MTX doses ranged between 5 and 25 mg per week in the control group, which was not significantly different from the dose range of the 12 patients with MTX toxicity (table 4).

Table 4

Distribution of diagnoses, decreased renal function and MTX doses in patients with severe toxicity (cases) and in control patients without MTX toxicity who were either on MTX or off MTX

When analysing the concomitant medication, MTX-treated patients with and without MTX toxicity had NSAIDs and cyclo-oxygenase-2 inhibitors, low-dose aspirin, ACE inhibitors/angiotensin receptor blockers and HCT in similar proportions (table 5). However, patients with toxicity more commonly were on non-HCT diuretics (50% vs 14%), on PPIs (83% vs 42%) and on levetiracetam (17% vs 1%).

Table 5

Comparison of concomitant medications in MTX-treated elderly patients with severe toxicity (cases, n=12) and without toxicity (controls on MTX, n=168), using Fisher’s exact test

All parameters with a p value<0.05, that is, eGFR, non-HCT diuretics, PPIs and levetiracetam, were introduced into a Firth’s biased-reduced multiple logistic regression (table 6). The two missing data points for eGFR of patients with MTX toxicity were imputed by predictive mean matching. The resulting model had an area under the curve of 0.8636 (95% CI 0.731 to 0.963). All four parameters remained significant in this multivariate analysis (table 6).

Table 6

Firth’s logistic regression results after imputation of the two missing pre-event eGFR values in patients with toxicity


It is well established that low-dose MTX for rheumatic diseases can lead to severe toxicity.20 26 27 In a period of 12 years, we admitted 12 patients for such life-threatening events. Of note, all of these patients were above the age of 70 years. Not a single case occurred in a younger patient.

On the other hand, 9 of 12 patients admitted for MTX toxicity were previously cared for in our outpatient clinics. Compared with the 168 patients without toxicity who were 70 years or older, this means at least 5% of elderly patients treated with low-dose MTX suffered from MTX toxicity. We cannot fully exclude additional cases of MTX toxicity, which would potentially increase the percentage, but the local healthcare infrastructure makes it unlikely that hospital admissions of patients or deaths would go undocumented. Although all but one patient survived, severe MTX toxicity is life-threatening and, because of the severe mucositis, often very painful.

All patients had both severe mucositis and mild-to-severe leucopenia. All but two patients also had mild-to-severe thrombocytopenia, usually corresponding to leucopenia in degree. The clinical picture of cytopenia and severe mucositis combined with the rapid improvement under folinic acid therapy28 in the surviving patients demonstrates MTX toxicity, in contrast to immune-mediated forms of bone marrow depression. In addition, these events in part happened after long periods of uncomplicated low-dose MTX therapy.

Some of the observed cases of MTX toxicity could be due to errors in administration, that is, daily MTX dosing.21 In this regard, the 10 mg MTX tablets most commonly used in Germany are more critical if, for example, MTX is mistaken for metoprolol. Accidental overdosing was clinically suspected in one patient, and a total of four patients had measurable MTX plasma levels. While the absence of measurable plasma MTX does not exclude MTX toxicity, measurable levels in low-dose MTX therapy are suspicious.

In most instances, however, toxicity was likely due to an acute or subacute decline in renal function. This was exemplified by an episode of acute gastroenteritis preceding admission for MTX toxicity in one patient. Five patients had values corresponding to CKD stage 4 or 5 at the time of admission, while renal function was better, and had presumably already recovered in other patients.

In contrast to renal function at admission, in all of the patients with previous values available, eGFR at the most recent visit before toxicity was compatible with CKD stage 2 or 3, and above the values that mandate stopping MTX, even though we would recommend stopping already, at 40 mL/min x 1.73 m², in line with the exclusion criteria of the Cardiovascular Inflammation Reduction Trial.11 In fact, the clearly higher eGFR of the still MTX-treated patients without toxicity in our cohort compared with those not on MTX suggests that eGFR is usually carefully monitored and acted on.

Still, the eGFR values preceding the event of toxicity were slightly, but significantly, lower than in patients without toxicity. This reiterates the above recommendation of stopping MTX when the eGFR falls to≤40 mL/min x 1.73 m². However, most patients admitted for MTX toxicity at their most recent visits had eGFR values that were fully acceptable. Therefore, we aimed to find better predictors of MTX toxicity.

Of interest, neither NSAIDs nor cyclo-oxygenase-2 inhibitors found to be relevant in other cohorts,29 nor the ACE inhibitors and angiotensin receptor blockers very commonly used in these elderly patients were associated with severe MTX toxicity. HCT also appeared safe. In contrast, other diuretics, most commonly torasemide, were linked to MTX toxicity. It is very well possible that torasemide and other diuretics may provoke acute renal failure by exsiccation, and in combination with ACE inhibitors/angiotensin blockers and NSAIDs, in particular. On the other hand, it is equally possible that it is the internal organ problems necessitating such therapy that increase the risk of renal failure.

Internal organ problems are also the most likely reason why PPIs are associated with an increased risk of renal failure, neither explained by NSAID medication nor by pneumonia as a potential preceding event. While there are discussions on interactions between MTX and PPIs,30 this suspicion was not confirmed in larger populations.21 Still, the use of PPIs should be questioned for a clear indication.

The antiepileptic levetiracetam is not a common drug in elderly patients with inflammatory rheumatic diseases, illustrated by the fact that only 1 in 168 MTX patients without toxicity was on levetiracetam. Although not supported by measured MTX levels, it is somewhat suspicious that 2 of 12 patients with toxicity were on levetiracetam, even though original reports on a potential inhibition of MTX elimination by this drug were not confirmed in a larger sample size.25

The levetiracetam question illustrates the inherent limitations of this study, which is based on only 12 admissions. While the number of only 12 patients is somewhat reassuring with regard to the low MTX toxicity occurrence, the low number of patients and their variability in presentation and risk factors reduces the robustness of the statistical associations and the consecutive conclusions. Still, all the factors discussed were statistically significant in the multivariate analysis, and we believe that we have been able to combine all, or at least essentially all, of our patients affected.

Taken together, life-threatening low-dose MTX toxicity may occur in about 5% of patients treated at the age of 70 years or older. MTX treatment in these patients needs to be critically evaluated accordingly, and in those with impaired renal function, in particular. Lower eGFR as well as the use of non-HCT diuretics, PPIs and levetiracetam were significant predictors of toxicity. In contrast, ACE inhibitors and angiotensin receptor blockers as well as low-dose aspirin, NSAIDs and cyclo-oxygenase-2 inhibitors were not. Accordingly, MTX should probably be stopped in elderly patients with an eGFR of ≤40 mL/min x 1.73 m² and in those on diuretics other than HCT. PPIs should only be administered if indicated, and some of their indications may also be an argument against further MTX treatment. Caution should also be used when patients are on levetiracetam.

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information. Not applicable.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by the ethics committee at the TU Dresden EK 49022011. Participants gave informed consent to participate in the study before taking part.


The authors are grateful to Iris Steinebrunner, PhD, for language editing the manuscript.



  • Contributors CK, KH and MA have planned the study and extracted and collected the data and provided supervision.

    MA accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.

    All authors were involved in data analysis, have critically revised the manuscript and have approved the last version.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.