Obstetric and neonatal interventions have improved the survival of preterm infants, but there has not been an equivalent reduction in long-term neurological disability. Thus, some effort must be invested in finding ways of preventing preterm birth. Numerous programmes have been promoted to address the matter of how the frequency of preterm birth could be prevented. Most interventions intended to prevent preterm labour do not have the desired effect, except for antibiotic treatment in cases of asymptomatic bacteriuria or bacterial vaginosis and progesterone administered prophylactically in high-risk women. Tocolytic drugs appear to delay delivery long enough for successful administration of corticosteroids in women in preterm labour, but without decreasing the risk of preterm birth. Some authors promote public health approaches that address all risk factors and affect the entire population of pregnant women, given that prevention programmes directed only at high-risk women have had little effect in preventing preterm births. However, the lack of progress in reducing the frequency of preterm births is also due to our limited understanding of the aetiology of preterm delivery. Although there is growing evidence that infection and neuroendocrine processes are involved, progress has remained slow. Recently, the hypothesis of a genetic predisposition to preterm delivery has been set up. Additional research exploring the pathophysiology of preterm labour is obviously needed, which will hopefully lead to the development of new therapeutic approaches.