As part of a study to assess zoonotic milk-borne health risks, seasonal survey data and unpasteurized milk samples were collected between January 1999 and February 2000 from randomly selected informal milk market agents (220 and 236 samples in the dry and wet seasons, respectively) and from households purchasing raw milk (213 and 219 samples in the dry and wet seasons, respectively) in rural and urban locations in central Kenya and screened for antibodies to Brucella abortus (B. abortus) and presence of Escherichia coli (E. coli) O157:H7. The latter was assessed based on samples from consumer households only. Antibodies to B. abortus were screened using the indirect antibody Enzyme Linked Immunosorbent Assay (ELISA) and the Milk Ring Test (MRT). The presence of E. coli O157:H7 was assessed by culture, biochemical characterisation, serological testing for production of verocytotoxin one (VT1) and two (VT2) and polymerase chain reaction (PCR) analysis for the presence of genes encoding for the toxins. The prevalence of antibodies to B. abortus varied considerably ranging from none in milk sold in small units and originating from intensive production systems to over 10% in samples that were bulked or originating from extensive production systems. E. coli O157:H7 was isolated from two samples (0.8%), one of which produced VT1. All urban consumers (100%) and nearly all rural consumers (96%) of marketed milk boiled the milk before consumption, mainly in tea, thus greatly reducing chances of exposure to live pathogens and potential health risks.