Methotrexate-induced pancytopenia: serious and under-reported? Our experience of 25 cases in 5 years

To ascertain the extent of methotrexate (MTX)-related pancytopenia at the Norfolk and Norwich University Hospital (NNUH) between 1999 and 2004.

Patients were identified by a department database search, review of pharmacy records and personal communication. Pancytopenia was defined as white blood cell count (WBC) <3.5 x 10(9)/l, haemoglobin (Hb) <11 g/dl and platelet count <130 x 10(9)/l. Severe pancytopenia was defined as WBC <2.0 x 10(9)/l, Hb <10 g/dl and platelet count <50 x 10(9)/l.

Twenty-five patients had MTX-induced pancytopenia. Eleven patients were taking folic acid and one folinic acid. The median dose of MTX was 12.5 mg weekly (interquartile range 5.625 mg) and median duration of treatment 36 months (interquartile range 40.5 months). The severity of pancytopenia correlated with the dose (P = 0.04). The numbers of patients with potential risk factors were: renal insufficiency, 8; pre-existing folate deficiency, 7; age >75 yr, 15; hypoalbuminaemia, 18; pre-existing infection with hip prosthesis, 1; possible drug interactions, 18; dosing errors, 1; and polypharmacy, 15. Pancytopenia was detected by routine blood monitoring in nine patients. There were seven deaths (28% mortality), five from sepsis and two from acute myeloid leukaemia.

This is the largest reported individual case series of MTX-induced pancytopenia. With the increasing long-term use of MTX, it is important that patients be monitored for haematological side-effects as pancytopenia can be a late manifestation. Pharmacogenetics may hold the answer to predicting who is at risk of this potentially fatal complication of MTX.