South Africa has a long history of uranium (U) production from gold mining and milling that has resulted in locally high-levels of environmental contamination from uranium and its decay products (radium 226 and radon 222). Due to short distances from mining tailings to residential areas, populations living around gold mine tailings of the Witwatersrand goldfields may be exposed to uranium and its decay products from the tailings through multiple pathways, including ingestion of contaminated water and food grown in contaminated areas, direct consumption of soil (geophagia) and inhalation of dust, raising concern about potential health risks associated with environmental U-exposure, and in particular regarding haematological malignancies (HM). We designed a case-series study of HM from Chris Hani Baragwanath Academic Hospital (CHBAH), South Africa, to assess the feasibility of an analytical study on the association between U-exposure and HM risk in this setting.

First, patient and cancer characteristics and spatial distribution were retrospectively reviewed from patient files of the HM cases diagnosed between 2004 and 2013 at CHBAH. Spatial relations between residential addresses of patients and location of mine tailings and other U-sources were analysed using a specifically designed GIS-supported virtual geographical environment. Then, we assessed the feasibility of collecting additional information, from patients newly diagnosed with a HM at CHBAH (in 2014–2015), on potential environmental U-exposure pathways and referral patterns from interviewer-administered questionnaires.

Among the 1880 cases aged 18–94 years and retrospectively identified from CHBAH's patient files, 44% were diagnosed with Non-Hodgkin lymphoma (NHL), 26% with leukemia and 17% with myeloma. Referral from distant provinces was not uncommon but more than 90% of patients lived within 50km of the hospital. Human immunodeficiency virus (HIV) status was known for 94%, of which 44% were HIV-positive. No clear and consistent spatial relations between patient's residences and mine tailings deposits could be established. A total of 196 haematological cancer cases aged 18–90 were prospectively diagnosed between 2014 and 2015 at CHBAH, with NHL (35%), leukemia (34%) and myeloma (16%). HIV status was available for all patients, 41% being HIV-positive. Working in gold mines was reported by 12 cases (6%), all men. Soil consumption was a habit for 53 cases (27%), mostly women, and came from backyards and road vendors. Soil consumption occurred more frequently during adulthood and pregnancy. Drinking water came from piped water; untreated water from nearby rivers was not drunk. Animal products and vegetables came from stores and backyards. At the time of interview, 95% of cases were living in Gauteng Province, in Soweto (58%). Half of the cases were living at their current residence for 20 years or more. Cases were referred to CHBAH by governmental clinic doctors (44%), referral hospitals (25%) and private doctors (20%). Most cases lived less than one hour from CHBAH (77%).

This case-series analysis showed the capacity of CHBAH to recruit HM patients and allowed the description of the HM burden of the area. Caution is required when interpreting the spatial distributions of patients; our results may not reflect population-incidence rates due to referral patterns to the hospital and inaccuracies in residential addresses retrospectively collected from clinical records. This needs to be considered when analyzing spatial relations related to various exposure pathways. The questionnaire administered to the prospective cases allowed collecting information on U-related exposure, although results generally showed low exposure prevalence in our study population. As infections, such as HIV, may be a risk factor for some HM, collecting clinical data on infections and subtypes of HM needs to be considered in the investigation of HM risk.