BACKGROUND: In the early nineties the increase of imported malaria in some European countries was temporarily halted, but it resumed in 1994. More Africans, more European travelers, and fewer long-term residents were counted amongst patients. A shift towards more subacute disease has been noted. This study intends to assess whether the same trends were observed in Belgium. METHODS: Clinical and epidemiological data of 128 patients treated for malaria in 1997 at the Institute of Tropical Medicine and the University Hospital of Antwerp were compared with 209 malaria patients treated in 1988/1989. Risk factors for clinical presentation and parasitemia were analysed. RESULTS: In Belgium the number of reported imported malaria cases remained almost stable between 1988 and 1997. In 1997, there were more African patients, less infections from Central Africa, and 50% less residents. Less patients reported prophylaxis use. The causative agent shifted from Plasmodium falciparum to other species. Subacute and atypical malaria became less frequent. In both years, there were no deaths, and severe malaria did not increase significantly. Mefloquine disappeared almost as a curative treatment, and was replaced by quinine, with or without a long acting agent, or by halofantrine. The ethnic origin, nor the use of chemoprophylaxis, influenced disease characteristics. In 1988, malaria attacks in the previous months predisposed to subacute disease; longer residence, and attacks in the previous months, protected against high parasitemia; longer symptom duration correlated with absence of fever, and with splenomegaly. None of these risk factors was correlated with severe malaria. CONCLUSION: The incidence of subacute malaria dropped significantly in the last decade. Although this presentation is almost limited to residents, the decline in malaria can not be explained by an overall shorter duration of stay, since the decline in this particular clinical presentation of malaria was also spectacular in residents. Apparently, insufficient treatment of malaria attacks in the previous months is the only independent risk factor.