This review of the safety of the co-administration regimens to be used in programmes to eliminate lymphatic filariasis (albendazole + ivermectin or albendazole + diethylcarbamazine [DEC]) is based on 17 studies conducted in Sri Lanka, India, Haiti, Ghana, Tanzania, Kenya, Ecuador, the Philippines, Gabon, Papua New Guinea, and Bangladesh. The total data set comprises 90,635 subject exposures and includes individuals of all ages and both genders. Results are presented for hospital-based studies, laboratory studies, active surveillance of microfilaria-positive and microfilaria-negative individuals, and passive monitoring in both community-based studies and mass treatment programmes of individuals treated with albendazole (n = 1538), ivermectin (9822), DEC (576), albendazole + ivermectin (7470), albendazole + DEC (69,020), or placebo (1144). The most rigorous monitoring, which includes haematological and biochemical laboratory parameters pre- and post-treatment, provides no evidence that consistent changes are induced by any treatment; the majority of abnormalities appear to be sporadic, and the addition of albendazole to either ivermectin or DEC does not increase the frequency of abnormalities. Both DEC and ivermectin show, as expected, an adverse event profile compatible with the destruction of microfilariae. The addition of albendazole to either single-drug treatment regimen does not appear to increase the frequency or intensity of events seen with these microfilaricidal drugs when used alone. Direct observations indicated that the level of adverse events, both frequency and intensity, was correlated with the level of microfilaraemia. In non microfilaraemic individuals, who form 80-90% of the 'at risk' populations to be treated in most national public health programmes to eliminate lymphatic filariasis (LF), the event profile with the compounds alone or in combination does not differ significantly from that of placebo. Data on the use of ivermectin + albendazole in areas either of double infection (onchocerciasis and LF), or of loiais (with or without concurrent LF) are still inadequate and further studies are needed. Additional data are also recommended for populations infected with Brugia malayi, since most data thus far derive from populations infected with Wuchereria bancrofti.