Purpose Methotrexate (MTX)/6-Mercaptopurine (6MP)-based maintenance therapy is crucial to cure childhood acute lymphoblastic leukemia (ALL). Cytotoxicity is mediated by incorporation of thioguanine nucleotides (TGN) into DNA (DNA-TG) with higher levels in leucocytes being associated with reduced relapse risk. To further understand the dynamics of DNA-TG formation, we measured DNA-TG levels in leucocyte subsets during maintenance therapy and in the months following its discontinuation. Methods DNA-TG levels were measured in leucocytes (DNA-TG(Total)), polymorph nucleated granulocytes (neutrophils, eosinophils, basophils [DNA-TG(PMN)]) and mononucleated cells (lymphocytes, monocytes [DNA-TG(MNC)]) in 1013 samples from 52 patients on ALL maintenance therapy (951 samples during therapy and 62 samples after therapy discontinuation, respectively). Results Median DNA-TG(Total), DNA-TG(PMN) and DNA-TG(MNC) during maintenance therapy were 539, 563 and 384 fmol/mu g DNA, respectively. DNA-TG(PMN) displayed more pronounced fluctuation than DNA-TG(MNC) (range 0-3084 [interquartile range IQR 271-881] versus 30-1411 [IQR 270-509] fmol/mu g DNA). DNA-TG(Total) was more strongly correlated with DNA-TG(PMN) (r(S) = 0.95, p <0.0001) than DNA-TG(MNC) (r(S) = 0.73, p <0.0001). DNA-TG(PMN) correlated less with DNA-TG(MNC) (r(S) = 0.64, p <0.0001) and to a much lesser extent with absolute neutrophil count (r(S) = 0.35, p <0.0001). Following discontinuation of therapy, DNA-TG(PMN) was rapidly eliminated, and not measurable beyond day 22 after discontinuation, whereas DNA-TG(MNC) was slowly eliminated, and five patients demonstrated a measurable DNA-TG(MNC) more than 365 days after therapy discontinuation. Conclusion Fluctuations in DNA-TG(Total) are predominantly caused by corresponding fluctuations in DNA-TG(PMN), thus DNA-TG(Total) measures recent TGN incorporation in these short-lived cells. Measurement of DNA-TG(Total) at 2-4 weeks intervals provides a reliable profile of DNA-TG levels.