OBJECTIVE To propose guidelines for clinical practice for tocolysis in preterm labor without premature preterm rupture of the membranes (PPROM). METHODS Bibliographic searches were performed in the Medline and Cochrane databases and gynecologist and obstetricians' international society guidelines. It is important to note that most studies included women in preterm labour with and without PPROM. RESULTS Compared with placebo, tocolytics are not associated with a reduction in neonatal mortality or morbidity (LE2). Compared with betamimetics, nifedipine is associated with a reduction in necrotizing enterocolitis, intraventricular hemorrhage and respiratory distress syndrome (LE2). There is no difference between nifedipine and atosiban regarding neonatal prognosis, except a modest reduction in NICU transfer with nifedipine (LE2). Betamimetics, atosiban and nifedipine are equivalent to prolong pregnancy for more than 48hours (LE2). Compared with betamimetics, nifedipine reduces delivery before 34 WG and is associated with a longer pregnancy (LE2). Atosiban and nifedipine are equivalent to prolong the pregnancy over 7 days (LE2), but in women with spontaneous preterm labour without PPROM, nifedipine reduces deliveries before 37 WG and pregnancy prolongation is longer, without improving neonatal prognosis (LE2). Maternal severe adverse effects may occur with all tocolytics (LE4). Betamimetics cardiovascular adverse effects are frequents (LE2) and may be serious (maternal death) (LE4). Nifedipine and atosiban reduce maternal adverse effect compared with placebo (LE2). Cardiovascular adverse effects are moderately increased with nifedipine compared with atosiban (LE2), without increasing treatment discontinuation (LE2). Regarding their benefits on pregnancy prolongation and good maternal tolerance, atosiban and nifedipine can be used for tocolysis in spontaneous preterm labour without PPROM (Grade B), for singleton and multiple pregnancies (Professional Consensus). Advantageously, nifedipine is orally taken and is inexpensive (Professional Consensus). Nicardipine should not be used for tocolysis (Professional Consensus) and betamimetics should not be prescribed anymore for tocolysis (Grade C). All tocolytic treatment should be prescribed for up to 48hours (Grade B). In case of initial tocolysis failure, another treatment may be proposed with the other class of tocolytic (Professional Consensus). Different class of tocolytics should not be combined (Grade C). Scientific data are lacking to propose guidelines regarding a rescue tocolysis, after a first previous successful tocolysis with complete antenatal corticosteroid therapy (Professional Consensus). There is no scientific evidence to propose a tocolysis in women with advanced dilatation (GradeC), nor prescribe a tocolysis after 34 WG (Professional Consensus). There is no evidence to define a gestational age lower limit for tocolysis (Professional Consensus). CONCLUSIONS Nifedpine and atosiban can be used for tocolysis (Grade B), including for multiple pregnancies (Professional Consensus). Maintenance tocolysis is useless (Grade C) and potentially harmful (Grade C). Betamimetics should not be used for tocolysis (Professional Consensus).