Across 2,907 FAERS reports on methotrexate in pediatric leukemia, nervous system disorders gave the strongest signal, febrile neutropenia was the most reported PT, and 72.1% of evaluable onsets fell within 30 days

The study reports a two-level disproportionality signal profile for methotrexate-associated adverse events in pediatric leukemia: 7 of 26 SOCs met the criteria, with nervous system disorders having the largest report count (1,691 cases, ROR 2.87, 95% CI 2.69–3.05, PRR 2.53, χ² 1,126.79, IC025 0.92); among 1,280 PTs, 274 were valid signals spanning 15 SOCs, with febrile neutropenia (446 reports, ROR 1.44, 95% CI 1.29–1.6, χ² 44.27, IC025 0.25), neurotoxicity (238), and mucosal inflammation (170) most frequently reported. The authors describe it as the first systematic characterization of the complete spectrum of adverse drug events associated with methotrexate therapy in pediatric leukemia, placing SOC- and PT-level signals, sex stratification, age stratification, and time-to-onset within one FAERS analysis. Based on 2,907 FAERS reports from Q1 2018 to Q1 2025, with a signal requiring ≥3 reports, ROR 95% CI lower limit >1, PRR ≥2, χ² ≥4, IC025 >0, EBGM05 >2, and a >0; the authors state that signal strength reflects reporting disproportionality, not incidence or causality.

Sex stratification detected 149 sex-related signals, of which 12 PTs showed potential signals across sex subgroups: confusional state in 40 male versus 9 female reports (ROR 2.79, 95% CI 1.35–5.76), dehydration in 20 versus 1 (ROR 12.55, 95% CI 1.68–93.56), and epistaxis in 20 versus 2 (ROR 6.27, 95% CI 1.47–26.86). It moves sex differences from overall description to PT-level reporting disproportionality and names three specific events more concentrated in male reports. Report counts are small (20 male reports each for dehydration and epistaxis) with wide confidence intervals (e.g., dehydration 1.68–93.56); the authors stress that FAERS lacks sex-specific exposure denominators, so these differences describe reporting disproportionality within the database only and cannot be read as higher incidence, greater clinical risk, or increased biological susceptibility in males.

Age stratification showed different leading PTs by group: neurotoxicity in infants (<1 year) and adolescents (11–17 years), febrile neutropenia in toddlers (1–3 years) and school-age children (7–10 years), and mucosal inflammation in preschoolers (4–6 years); the five groups corresponded to 21, 77, 94, 101, and 139 PTs, and six PTs—neurotoxicity, leukoencephalopathy, mucosal inflammation, toxicity to various agents, drug clearance decreased, and drug level increased—were detected in all five strata. A five-stratum intersection analysis yields cross-age shared signals alongside age-specific leading signals, adding to prior reporting that treated the population as a whole. These are descriptive stratified results; the authors note the differences may reflect unequal methotrexate exposure, leukemia subtype, risk stratification, treatment phase or intensity, supportive care, concomitant therapy, or reporting behavior rather than true age-related differences in incidence or susceptibility.

Time-to-onset analysis included 944 evaluable reports (32.5% of 2,907), with 681 (72.1%) occurring within 0–30 days, a median onset of 11 days (4–50), and a Weibull shape parameter β of 0.55 (95% CI 0.53–0.56), entirely below 1, indicating an early reporting pattern. It quantifies the temporal dimension as median onset days and a Weibull shape parameter while explicitly framing the result as a reporting-time distribution rather than a causal risk window. Only 32.5% of reports had evaluable onset dates; the authors state that fewer reports after 30 days may reflect treatment phase transitions, follow-up duration, reporting behavior, and concomitant therapy and should not be interpreted as declining hazard over time.

The scope the study defines is FAERS spontaneous reports for pediatric leukemia (ages 0–17) in which methotrexate was the primary suspect drug, spanning Q1 2018 to Q1 2025. It can give clinicians and pharmacovigilance staff monitoring directions: watch hepatobiliary, gastrointestinal, nervous system, psychiatric, skin and subcutaneous tissue, renal and urinary, and eye SOC-related events during methotrexate use in children, with particular attention to nervous system events, and note febrile neutropenia, neurotoxicity, mucosal inflammation, and confusional state, dehydration, and epistaxis in male reports. The authors state that sex-, age-, and time-specific surveillance should follow established clinical protocols and be assessed prospectively rather than inferred directly from FAERS subgroup proportions.

Open questions the authors list include: FAERS does not provide the number of exposed patients, person-time, treatment duration, or a representative sampling frame, so reporting proportions and disproportionality metrics cannot estimate incidence, frequency, absolute risk, or comparative clinical risk; pediatric leukemia was analyzed as a single cohort even though methotrexate dose, route, schedule, cumulative exposure, treatment intensity, supportive care, and co-administered agents vary by subtype, risk-stratification group, and induction, consolidation, or maintenance phase, and these variables were unavailable or incomplete in FAERS and could not be controlled; concomitant medications were not systematically evaluated, so events such as febrile neutropenia, mucositis, neurotoxicity, and hepatotoxicity may reflect leukemia itself, multi-agent chemotherapy, supportive medications, infections, or interactions rather than methotrexate alone; limited follow-up restricts assessment of delayed or chronic toxicities; quantitative methotrexate concentrations and sampling times were unavailable, so relationships between serum levels and specific adverse events could not be evaluated; and dosing frequency, number of administrations, cycle number, and cumulative exposure could not be reliably evaluated, so the 0–30-day reporting pattern cannot be attributed to a specific dosing schedule or repeated exposure. In addition, the loaded text is an incomplete version: the contents of Figures 1 through 5 and Supplementary Tables S1, S2, S3, and S4 are not included in the body, so the full per-PT values in the figures and the supplementary table details cannot be checked here, which is an information gap to keep in mind while reading.