OBJECTIVE To investigate if pregnancy represents a period of increased risk of non-cavitated dental caries related to changes in saliva and oral health behaviours. METHODS A non-randomised longitudinal study was performed with 27 pregnant women and 25 non-pregnant women, who were evaluated twice with the same time gap (24 weeks on average). At the first visit sociodemographic and oral health-related behaviours were assessed through a structured questionnaire. At the second visit changes related to eating sweet snacks and oral hygiene habits were also assessed. In both visits the surface-related caries status was evaluated according to ICDAS II criteria. Calculation of D0 (Sound), D1-2 (visual changes) and D3-4 (precavitated caries lesions) Index was based on data collected from clinical examination. Saliva pH and saliva flow rate were also assessed. RESULTS Throughout pregnancy, a statistically significant increase of eating sweet snacks between main meals was reported, with no effective adaptation of oral hygiene habits. In comparison to the non-pregnant group, pregnant women presented a lower saliva pH at both the first and second visit, p < 0.0005. During the follow-up period, a decrease in the frequency of caries-free surfaces was observed in the pregnant women (p = 0.004) and an increase in precavitated caries lesions (p = 0.011). CONCLUSIONS The main results support the hypothesis that during pregnancy women are prone to enamel demineralisation, namely, to exhibiting additional lesions characterised by precavitated caries lesions.