The 500,000 inhabitants of Mayo Tsanaga River Basin are vulnerable to a "silent" fluorosis from groundwater consumption. For the first time, the groundwater is investigated for the purpose of identifying the provenance of fluoride and estimating an optimal dose of fluoride in the study area. Based on the fluoride content of groundwater, fluorine and major oxides abundances in rocks from the study area, mean annual atmospheric temperature, and on-site diagnosis of fluorosis in children, the following results and conclusions are obtained: Fluoride concentration in groundwater ranges from 0.19 to 15.2 mg/l. Samples with fluoride content of <1.5 mg/l show Ca-HCO(3) signatures, while those with fluoride >1.5 mg/l show a tendency towards Na-HCO(3) type. Fluor-apatite and micas in the granites were identified as the main provenance of fluoride in the groundwater through water-rock interactions in an alkaline medium. The optimal fluoride dose in drinking water of the study area should be 0.7 mg/l, and could be adjusted downward to a level of 0.6 mg/l due to the high consumption rate of groundwater, especially during drier periods.

Delivery of very preterm babies in maternity units with on-site neonatal intensive care (level III units) is associated with lower mortality and morbidity. This analysis explores risk factors for not delivering in a level III unit, using data from a population-based study of very preterm births in Paris and surrounding districts in 2003. The sample for analysis included resident women with a fetus alive at the onset of labour between 24 and 31 weeks of gestation (n = 641). Characteristics of women delivering in and those not in level III units were compared using logistic regression. Further analysis was carried out for the subgroup of women not already scheduled to deliver in a level III unit. Twenty-nine per cent of women did not deliver in level III units; in the subgroup scheduled to deliver in level I or II units, 43% were not transferred. Women were less likely to deliver in a level III unit if they had a singleton pregnancy, a gestation of <26 weeks or at 31 weeks, experienced antenatal haemorrhaging, lived in socially deprived neighbourhoods or at a greater distance from the nearest level III. Women scheduled to deliver in a maternity unit with a special care nursery were also less likely to deliver in a level III unit. In contrast, preterm rupture of membranes and fetal growth restriction increased the likelihood of a level III delivery. These results underline the importance of controlling for clinical characteristics when analysing perinatal outcome by place of delivery and show how socioeconomic factors, known to impact on the risk of having a preterm birth, can also affect access to appropriate care.