BACKGROUND AND OBJECTIVES: Compare anesthesia techniques with sufentanil and intraoperative infusion of remifentanil or dexmedetomidine, and to then analyze their associations with the incidence of pain and other complications in the post-anesthesia care unit. METHODS: A retrospective analytical observational study conducted in a reference center for bariatric surgery in São Paulo, São Paulo State, Brazil. Patients (n=120) included in the study were randomly selected using this site's anesthesia records for bariatric surgery. Four 30-patient groups were established: G1-induction with 0.5 µg.kg-1 sufentanil associated with the continuous infusion of remifentanil; G2-induction with 0.7 µg.kg-1 sufentanil associated with the continuous infusion of remifentanil; G3-induction with 0.5 µg.kg-1 sufentanil associated with the continuous infusion of dexmedetomidine; and G4-induction with 0.7 µg.kg-1 sufentanil associated with the continuous infusion of dexmedetomidine. Significance was set at p<0.05. RESULTS: The incidence of pain in the post-anesthesia care unit was 52.5% (n=63). It was considered intense in 36.11% of these cases. The incidence of pain in the post-anesthesia care unit was lower in patients receiving a continuous infusion of dexmedetomidine relative to those who received remifentanil infusion. Sufentanil dosage in anesthesia induction did not influence the incidence of pain (G1=G2>G3=G4; p<0.05). The occurrence of respiratory complications did not differ significantly between the groups. CONCLUSION: The use of dexmedetomidine proved to be better for analgesia in the post-anesthesia care unit, and it did not increase the risk of respiratory complications when compared to the continuous infusion of remifentanil, regardless of the sufentanil dose used for induction.