The recommendation to treat canine lymphoma with a discontinuous protocol is based on small case numbers and mostly historic controls. This study compares duration of first remission (DFR) and overall survival time (ST) with a discontinuous protocol to the same protocol with maintenance phase. 408 dogs were treated with a CHOP-LAsp (C=cyclophosphamide; H=hydroxydaunorubicin; O=Oncovin; P=prednisolone; LAsp=l-asparaginase)-based 28-week induction protocol. In 75 dogs (cohort 1), this was followed by a maintenance phase consisting of vincristine, chlorambucil and actinomycin-D with a total treatment duration of two years. In the subsequent 333 dogs, therapy was discontinued after induction (cohort 2). Median DFR and ST in cohort 1 were 216 and 375 days and 184 and 304 days in cohort 2. 6-Month, 1-year and 2-year survival rates in cohort 1 were 73 per cent, 50 per cent, 24 per cent and 67 per cent, 39 per cent, 21 per cent in cohort 2. There was no significant difference between the two protocols (P=0.291 for ST, P=0.071 for DFR). On multivariate analysis, corticosteroid pretreatment (P=0.005), thrombocytopenia at diagnosis (P=0.019), stage (P=0.009), substage b at relapse (P<0.001), age (P=0.002) and incomplete or unstable remission necessitating intensification of therapy (P=0.004) were negatively correlated with ST in both groups. This study supports the use of a discontinuous protocol for canine multicentric lymphoma.