Study design and sample

This study was approved by the Ethics Committee of the Academia Cearense de Odontologia and was performed in accordance with the Helsinki statements. The research protocol followed a prospective, single-centre, split-mouth, randomized, triple-blind, placebo-controlled study design, and it was conducted on patients recruited from the Division of Oral and Maxillofacial Surgery, Walter Cantídio University Hospital, Federal University of Ceará (Brazil), who required lower third molar extraction. Patient recruitment was conducted between April 2011 and September 2012 according to the CONSORT statement. The sample unit used in the present study was the surgical site.

Healthy subjects (American Society of Anesthesiologists (ASA) classification I) of both genders, aged 18–35 years, with a clear indication for removal of two lower third molars, were invited to participate in this study. In order to control for the level of traumatic injury inflicted on the patient, the following inclusion criteria were also applied: (1) full coverage of lower third molars by osseous tissue, requiring bone removal and/or tooth sectioning for their extraction, and (2) similar patterns of root formation, position, and degree of impaction between the right and left impacted third molars for all study subjects. Other inclusion criteria were the absence of periodontal disease, swelling, hyperthermia, and trismus prior to surgery. Enrolment in the study required the ability and willingness to cooperate with the research protocol and the provision of appropriate written informed consent.

Patients were excluded if they fulfilled any of the following criteria: were smokers, pregnant or breast-feeding, using medications that could potentially interact with the drugs used in the study, had orthodontic bands on the second molars, had a known allergy to NSAIDs, had a systemic chronic disease, had signs of any pre-existing acute inflammatory or infectious condition, or had used NSAIDs within the past 21 days. Individuals who did not express an interest in participating in this clinical trial during subject recruitment were also excluded. Patients were removed from the study if there was intolerance to the pharmacological regimen, if they were unable to follow the study protocol, if the surgical time exceeded a duration of 2 h, or if they presented a postoperative infection.

Data were recorded preoperatively according to a standardized clinical examination and included gender, age, systemic conditions, periodontal status, haemogram parameters, platelet count, international normalized ratio (INR), and plasma glucose. A panoramic radiograph was required to evaluate variables such as the tooth position according to the Pell and Gregory and Winter classifications, degree of tooth/root development, and level of impaction.

Patients were scheduled for surgery in two separate clinical sessions (one side at a time) at least 3 weeks apart. Subjects were allocated to one of two groups through a computer-generated randomization code (Microsoft Excel), according to the medication received 1 h before surgery: group 1, etoricoxib 120 mg; group 2, placebo. Antibiotic prophylaxis was not given to the patients. After surgery, ibuprofen 300 mg at 8-h intervals was allowed in the case that a rescue analgesic medication was needed.

Sample size calculation

Initially, a placebo-controlled study with six patients (12 mandibular third molars) was performed to calculate the sample size required to conduct this clinical trial and statistically reject the null hypothesis with 80% power and a 95% confidence interval. Based on the mean pain scores of the pilot study (etoricoxib 0.3 ± 0.8 and placebo 2.3 ± 2.9), a minimum sample size of 18 surgical sites in each group was estimated.

Blinding

Information on the type of medication provided to each study subject was withheld from the patient, surgeon, clinical investigator (responsible for patient follow-up examinations and outcome measurements), and statistician. Prior to surgery, a list containing a randomized distribution of all surgical sites and pain medications to be administered was held in a sealed envelope by an external study collaborator, who was unaware of the study protocol and had no further participation in this clinical trial other than to guarantee a triple-blind study design. The statistical analysis was initially carried out with groups coded with the letter ‘A’ representing etoricoxib and ‘B’ representing placebo. The envelope decoding this information was only accessed once both the clinical trial and statistical analysis had been concluded. At this time, each patient and surgical site assigned to receive etoricoxib or placebo was identified.

Surgical overview

All patients underwent a standardized surgical technique performed in an outpatient setting under local anaesthesia, followed by strict biosafety control. One surgeon with 10 years of experience in oral and maxillofacial surgery performed all of the surgical procedures. The same surgical procedure was adopted for both sides of the mouth, aiming to reduce differences in the level of intraoperative trauma. The extraction of lower third molars was performed under local anaesthesia with mepivacaine 2% and epinephrine 1:200,000 (Mepivalem AD; Dentsply, USA), using two or three 1.8-ml cartridges. A full-thickness flap was raised, followed by bone removal using a drill cooled with bi-distilled water. The surgical wound was closed using a 4–0 silk suture.

Postoperative instructions were read and explained carefully to the patient, e.g. following a liquid and cold diet for 24 h, performing rigorous oral hygiene, and avoiding mouthwashes to prevent the occurrence of post-surgical bleeding. Patients were informed that they must contact the surgeon by telephone in the case of persistent bleeding or any other complications such as fever. In addition, patients were also asked to report any physical symptoms experienced during the study period, e.g., nausea, vomiting, dizziness, headache, insomnia, and signs of infection.

Outcome measures

The primary outcome of the study was the occurrence of postoperative pain. Measurements of this outcome considered pain intensity and the need for rescue analgesia. Postoperative pain intensity was measured using a 10-cm visual analogue scale (VAS), which consisted of an interval scale ranging from 0 (absence of pain or discomfort) to 10 (maximum pain or discomfort). Before surgery, each patient received an explanation on how to measure pain intensity on the VAS. After surgery, patients received a standardized VAS form to record the values of postoperative pain, and it was required that this form be returned to the researcher on the day of suture removal. Study participants were asked to record the pain intensity score at 0, 2, 4, 6, 8, 10, 12, 24, 48, and 72 h and on days 5 and 7 following surgery. Additional analyses included the length of time elapsed until the intake of a rescue analgesic by the patient. The number of patients requiring a rescue analgesic was also recorded.

The occurrence of postoperative inflammatory events was the secondary outcome measure adopted in the present study. Measurements ( Fig. 1 ) were performed to assess postoperative facial swelling on the side of surgery, including the distance from the mandibular angle to (1) the tragus (distance MA–Tr), (2) the external corner of the eye (distance MA–ECE), (3) the nasal border (distance MA–NB), (4) the labial commissure (distance MA–LC), and (5) the soft pogonion (distance MA–SP). Differences between preoperative values (baseline) and those obtained at 24 and 72 h and at 5 and 7 days after surgery were compared. In order to estimate trismus, the maximum mouth opening ability was measured pre- and postoperatively (after 24 h, 72 h, 5 days, and 7 days), by assessing the distance between the upper and lower central incisors in millimetres using a calibrated ruler.

Facial measurements for the assessment of postoperative swelling: mandibular angle to tragus (distance MA–Tr), mandibular angle to external corner of the eye (distance MA–ECE), mandibular angle to nasal border (distance MA–NB), mandibular angle to labial commissure (distance MA–LC), and mandibular angle to soft pogonion (distance MA–SP).

Statistical analysis

Data were initially submitted to the Kolmogorov–Smirnov normality test. Parametric data were analyzed by one-way or two-way analysis of variance (ANOVA)/Bonferroni test. Non-parametric data were analyzed with the Mann–Whitney or Wilcoxon and Friedman/Dunn post hoc tests. Quantitative variables were expressed as the mean ± standard deviation of the mean (SD). The Kaplan–Meier method was used to evaluate the rescue medication through the log-rank Mantel–Cox test. Pearson linear correlation was applied to correlate the total number of rescue medications taken and the sum of pain scores. All analyses were performed with GraphPad Prism 5.0 software (GraphPad Software Inc., San Diego, CA, USA). The level of significance was set as P < 0.05 for all of the evaluations.

Study design and sample

This study was approved by the Ethics Committee of the Academia Cearense de Odontologia and was performed in accordance with the Helsinki statements. The research protocol followed a prospective, single-centre, split-mouth, randomized, triple-blind, placebo-controlled study design, and it was conducted on patients recruited from the Division of Oral and Maxillofacial Surgery, Walter Cantídio University Hospital, Federal University of Ceará (Brazil), who required lower third molar extraction. Patient recruitment was conducted between April 2011 and September 2012 according to the CONSORT statement. The sample unit used in the present study was the surgical site.

Healthy subjects (American Society of Anesthesiologists (ASA) classification I) of both genders, aged 18–35 years, with a clear indication for removal of two lower third molars, were invited to participate in this study. In order to control for the level of traumatic injury inflicted on the patient, the following inclusion criteria were also applied: (1) full coverage of lower third molars by osseous tissue, requiring bone removal and/or tooth sectioning for their extraction, and (2) similar patterns of root formation, position, and degree of impaction between the right and left impacted third molars for all study subjects. Other inclusion criteria were the absence of periodontal disease, swelling, hyperthermia, and trismus prior to surgery. Enrolment in the study required the ability and willingness to cooperate with the research protocol and the provision of appropriate written informed consent.

Patients were excluded if they fulfilled any of the following criteria: were smokers, pregnant or breast-feeding, using medications that could potentially interact with the drugs used in the study, had orthodontic bands on the second molars, had a known allergy to NSAIDs, had a systemic chronic disease, had signs of any pre-existing acute inflammatory or infectious condition, or had used NSAIDs within the past 21 days. Individuals who did not express an interest in participating in this clinical trial during subject recruitment were also excluded. Patients were removed from the study if there was intolerance to the pharmacological regimen, if they were unable to follow the study protocol, if the surgical time exceeded a duration of 2 h, or if they presented a postoperative infection.

Data were recorded preoperatively according to a standardized clinical examination and included gender, age, systemic conditions, periodontal status, haemogram parameters, platelet count, international normalized ratio (INR), and plasma glucose. A panoramic radiograph was required to evaluate variables such as the tooth position according to the Pell and Gregory and Winter classifications, degree of tooth/root development, and level of impaction.

Patients were scheduled for surgery in two separate clinical sessions (one side at a time) at least 3 weeks apart. Subjects were allocated to one of two groups through a computer-generated randomization code (Microsoft Excel), according to the medication received 1 h before surgery: group 1, etoricoxib 120 mg; group 2, placebo. Antibiotic prophylaxis was not given to the patients. After surgery, ibuprofen 300 mg at 8-h intervals was allowed in the case that a rescue analgesic medication was needed.

Sample size calculation

Initially, a placebo-controlled study with six patients (12 mandibular third molars) was performed to calculate the sample size required to conduct this clinical trial and statistically reject the null hypothesis with 80% power and a 95% confidence interval. Based on the mean pain scores of the pilot study (etoricoxib 0.3 ± 0.8 and placebo 2.3 ± 2.9), a minimum sample size of 18 surgical sites in each group was estimated.

Blinding

Information on the type of medication provided to each study subject was withheld from the patient, surgeon, clinical investigator (responsible for patient follow-up examinations and outcome measurements), and statistician. Prior to surgery, a list containing a randomized distribution of all surgical sites and pain medications to be administered was held in a sealed envelope by an external study collaborator, who was unaware of the study protocol and had no further participation in this clinical trial other than to guarantee a triple-blind study design. The statistical analysis was initially carried out with groups coded with the letter ‘A’ representing etoricoxib and ‘B’ representing placebo. The envelope decoding this information was only accessed once both the clinical trial and statistical analysis had been concluded. At this time, each patient and surgical site assigned to receive etoricoxib or placebo was identified.

Surgical overview

All patients underwent a standardized surgical technique performed in an outpatient setting under local anaesthesia, followed by strict biosafety control. One surgeon with 10 years of experience in oral and maxillofacial surgery performed all of the surgical procedures. The same surgical procedure was adopted for both sides of the mouth, aiming to reduce differences in the level of intraoperative trauma. The extraction of lower third molars was performed under local anaesthesia with mepivacaine 2% and epinephrine 1:200,000 (Mepivalem AD; Dentsply, USA), using two or three 1.8-ml cartridges. A full-thickness flap was raised, followed by bone removal using a drill cooled with bi-distilled water. The surgical wound was closed using a 4–0 silk suture.

Postoperative instructions were read and explained carefully to the patient, e.g. following a liquid and cold diet for 24 h, performing rigorous oral hygiene, and avoiding mouthwashes to prevent the occurrence of post-surgical bleeding. Patients were informed that they must contact the surgeon by telephone in the case of persistent bleeding or any other complications such as fever. In addition, patients were also asked to report any physical symptoms experienced during the study period, e.g., nausea, vomiting, dizziness, headache, insomnia, and signs of infection.

Outcome measures

The primary outcome of the study was the occurrence of postoperative pain. Measurements of this outcome considered pain intensity and the need for rescue analgesia. Postoperative pain intensity was measured using a 10-cm visual analogue scale (VAS), which consisted of an interval scale ranging from 0 (absence of pain or discomfort) to 10 (maximum pain or discomfort). Before surgery, each patient received an explanation on how to measure pain intensity on the VAS. After surgery, patients received a standardized VAS form to record the values of postoperative pain, and it was required that this form be returned to the researcher on the day of suture removal. Study participants were asked to record the pain intensity score at 0, 2, 4, 6, 8, 10, 12, 24, 48, and 72 h and on days 5 and 7 following surgery. Additional analyses included the length of time elapsed until the intake of a rescue analgesic by the patient. The number of patients requiring a rescue analgesic was also recorded.

The occurrence of postoperative inflammatory events was the secondary outcome measure adopted in the present study. Measurements ( Fig. 1 ) were performed to assess postoperative facial swelling on the side of surgery, including the distance from the mandibular angle to (1) the tragus (distance MA–Tr), (2) the external corner of the eye (distance MA–ECE), (3) the nasal border (distance MA–NB), (4) the labial commissure (distance MA–LC), and (5) the soft pogonion (distance MA–SP). Differences between preoperative values (baseline) and those obtained at 24 and 72 h and at 5 and 7 days after surgery were compared. In order to estimate trismus, the maximum mouth opening ability was measured pre- and postoperatively (after 24 h, 72 h, 5 days, and 7 days), by assessing the distance between the upper and lower central incisors in millimetres using a calibrated ruler.

Facial measurements for the assessment of postoperative swelling: mandibular angle to tragus (distance MA–Tr), mandibular angle to external corner of the eye (distance MA–ECE), mandibular angle to nasal border (distance MA–NB), mandibular angle to labial commissure (distance MA–LC), and mandibular angle to soft pogonion (distance MA–SP).

Statistical analysis

Data were initially submitted to the Kolmogorov–Smirnov normality test. Parametric data were analyzed by one-way or two-way analysis of variance (ANOVA)/Bonferroni test. Non-parametric data were analyzed with the Mann–Whitney or Wilcoxon and Friedman/Dunn post hoc tests. Quantitative variables were expressed as the mean ± standard deviation of the mean (SD). The Kaplan–Meier method was used to evaluate the rescue medication through the log-rank Mantel–Cox test. Pearson linear correlation was applied to correlate the total number of rescue medications taken and the sum of pain scores. All analyses were performed with GraphPad Prism 5.0 software (GraphPad Software Inc., San Diego, CA, USA). The level of significance was set as P < 0.05 for all of the evaluations.