There is still extensive debate on the best method of controlling the behavior of preschool children during dental treatment. Protective stabilization, moderate sedation and general anesthesia are advanced behavior control techniques indicated for the dental treatment of early childhood caries and offer advantages and disadvantages during the procedure or immediately after. Many children with early childhood caries require invasive dental treatment. According to the final report of a large epidemiological survey on the oral condition of Brazilians, five-year-old children had an average of 2.43 primary teeth with caries and fewer than 20% of these had been treated in 2010. This disease also remains a public health problem in most developed countries; 19.5% of 2-5-year-old American children have untreated cavities.

There is, however, a lack of the ideal sedative. Such drugs must, on the one hand, control the behavior of integral form, provide amnesia, minimizing physical discomfort, distress and pain, and, on the other, safeguard security, with minimal effect on the cardio-respiratory function, minimizing the occurrence of adverse events, as well as allowing the return of the patient to a State that allows high safely. The investigators thus performed this prospective study with the aim to assess the occurrence of adverse events during dental treatment and in the first 24 hours after sedation with midazolam, ketamine and sevoflurane in children aged four to six years. Our hypothesis was that no differences in adverse events among different association of drugs could be found.

Twenty one children aged 46-81 months (mean 62.8 months) completed a randomized clinical trial and were followed-up for adverse events from a total of 37 children (27 were excluded because of being outside the age range of the survey, did not present negative behavior or had have prior dental treatment). These children were classified as ASA PS 1 or 2, and were referred to the Dental Sedation Study Center (NESO) of the Federal University of Goiás School of Dentistry by public health services.

In this clinical trial, randomized controlled type triple-blind, two groups of children received specific interventions: a) MC group: composed of children undergoing sedation for association of midazolam and ketamine per oral, and placebo (oxygen) by inhalation; b) MCS: group composed of children undergoing sedation for association of midazolam and ketamine per oral, and inhaled sevoflurane. Inclusion criteria were: 1) age of four to six years (until the end of the dental procedures); 2) children ASA I or II; 3) need for restorative procedure at least a deciduous tooth; 4) patent airway and nasal effective breathing and 5) inadequate behavior in previous dental treatment. Exclusion criteria were 1) previous experience of dental treatment under sedation or 2) having completed seven years during data collection. Sample size was calculated based on primary study with 18 children, which were included in the final sample. In this, through two-tailed hypothesis test, it was found that the occurrence of adverse events in the first 24 hours after sedation for dental treatment presented proportion of 87.5% for the MC Group and proportion of 30% for the MCS Group, which if defined that, to achieve the probability of 80% of the study to detect a difference in a level of significance of 5%, we would need a sample of 10 participants for each group (n = 20). On the day of treatment, upon arriving at the clinic, after checking healthiness of children and NPO protocols, the child accompanied by their parent/guardian were redirected to a drug delivery room, where the anesthesiologist or the pediatrician prepared medications in a disposable syringe and administered the drugs: midazolam-Dormire ® (Cristália, São Paulo, Brazil), at a dose of 0.5 mg/kg (maximum dose 20 mg) and ketamine-Ketamin ® (Cristália, São Paulo, Brazil), at a dose of 3 mg/kg (maximum 50 mg). After 15 minutes, they were redirected to the dental office and vital signs were monitored (Infinity ® Vista XL (Drägerwerk AG & co., Lubeck, Germany). According to the randomization, the child received only oxygen or a mixture of oxygen and sevoflurane-Sevocris ® (Cristália, São Paulo, Brazil), provided through an anesthesia workstation (Fabius ® Plus - Drägerwerk AG & co., Lubeck, Germany). Such gases were provided through a mask placed over the nose of the child (Dynomite ® Nasal Hood - Matrx-Parker Instrument, Hatfield, USA) and analysed using an anesthetic gas analyser (Let's ® -Drägerwerk AG & co., Lubeck, Germany). Initially, the child received the 100% oxygen at a flow rate of 5 l/min, for 5 minutes. After this period, if the child had been randomized to the MCS Group, sevoflurane was added in an initial concentration of 0.1%, with 0.1% increment every 30 seconds, until a final expired concentration between 0.3 and 0.4 %. In cases where the child had been randomized to the MC Group (without sevoflurane), the anesthesiologist simulating the supply of sevoflurane, but the child received only 100% oxygen. After 15 minutes of placing the nasal mask and supply of gases, the pediatric dentist dental treatment began as planned.

At the end of each procedure, post-operative recommendations (written protocol) were given to parents/guardians. On the day following treatment, telephone contact was made to parents/guardians to gather information about possible adverse events in 24 hours following procedure. In both groups, during the sessions, children's behavior was evaluated through the Houpt scale, with an observer previously calibrated and masked. The scoring system suggested by Houpt et al. for this purpose included four scales: sleep scale (awake - score 1; asleep - score 2), crying scale (hysterical crying that demands attention - score 1; continuous, persistant crying that makes treatment difficult - score 2; intermittent, mild crying that does not interfere with the treatment - score 3; no crying - score 4), movement scale (violent movement, interrupting treatment - score 1; continuous movement make treatment difficult - score 2; controllable movement that does not interferes with the treatment - score 3; no movement - score 4) and overall behavior scale (aborted, no treatment rendered - score 1; poor, treatment interrupted, only partial treatment completed - score 2; fair, treatment interrupted, but eventually all completed - score 3; good, difficult but all treatment performed - score 4; very good, some limited crying or movement - score 5; excellent, no crying or movement - score 6). The evaluations were performed during all the treatment sessions. The scores were established in accordance with the Houpt Scale every minute of the session, and in three specific moments of the dental treatment (at the time of anesthesia, at the time of use of high rotation and at the end of treatment), in addition to the overall evaluation. For calibration, the examiner watched five minutes videos of three patients during the clinical examination. The same videos were watched by a researcher as a gold standard, in order to assess the correlation between the examiners. There was concurrence intra and inter examiner, obtaining values of kappa 0.9 and 0.8, respectively.

After the child was considered eligible to participate in the survey and your legal guardian have agreed with their participation, the same was assigned at random to a group. The randomization plan of children in groups was prepared through the website Randomization.com (http://www.randomization.com).

The data were analyzed using the Statistical Package for Social Sciences (SPSS for Windows, version 19.0 SPSS Inc. Chicago, IL, USA), having the following variables: a) dependents (occurrence of adverse events in 24 hours following procedure; occurrence of intraoperative adverse events; child's Behavior) and b) independents (gender; age; weight; dose of oral sedatives; need for physical restraint; duration of the session; recovery time; heart rate; oxygen saturation).

According to the distribution of the data obtained for the analysis Mann-Whitney test, Pearson's Chi-square and Fisher test were used. The choice of these non-parametric tests were due to the characteristics of the data of the study, which did not present normal distribution. Adopted as significance level for the statistical tests applied a value of 5% (p < 0.05).

Inclusion Criteria:

• Children ASA I or II;
• Need for restorative procedure on at least one deciduous tooth below;
• Airway patent and effective nasal breathing;
• Behavior not collaborator in previous dental treatment.

Exclusion Criteria:

• Previous experience of dental treatment under sedation;
• Have completed seven years during the data collection.

The research data can be shared with other researchers at a reasonable request.