Analysis of oxygen saturation levels recorded during dental intravenous sedations - a retrospective quality assurance of 3500 cases

Abstract

ABSTRACT Background: The death of a patient under dental intravenous sedation in New South Wales, Australia, in 2002 while being treated by a dentist with appropriate training in intravenous conscious sedation (ICS), (Graduate Diploma in Clinical Dentistry; Conscious Sedation and Pain Control, University of Sydney) has once again brought into question the safety of ICS. The cause of death was irreversible cerebral hypoxia following a cardiac arrest, which was precipitated by numerous periods of ever-deepening hypoxaemia. Aim: This retrospective, quality-assurance audit investigated whether safe oxygen saturation levels could be maintained during single operator/sedationist dental sedations, when operating within the joint Royal Australasian College of Dental Surgeons (RACDS) and Australian and New Zealand College of Anaesthetists (ANZCA) PS21 guidelines for conscious sedation. Methodology: Safe oxygen saturation levels were defined as pulse oximeter readings of 94% and above. The recording of two or more readings of less than 94% during a sedation procedure was defined as the outcome of interest. The association of the variables of age (eighteen and over), gender, weight, the American Society of Anesthesiologists (ASA) Classification I or II and the use of propofol in addition to midazolam and fentanyl, to low saturations, was examined. Two sub-cohorts were randomly generated: 1,750 patients were sedated with iv midazolam and fentanyl and 1,750 patients received propofol, in sub-anaesthetic increments, in addition to midazolam and fentanyl. All patients received supplemental oxygen. Initial sedation was established using midazolam and fentanyl in both sub-cohorts. The second sub-cohort received sub-anaesthetic increments of propofol during times of noxious stimulation. Statistical analysis of the data used cross-tabulation of the variables by outcome, an associated chi- squared test and corresponding logistic regression analysis, together with odds ratio (OR) and a 95% confidence interval (95%CI). Results: Patient exposure to two or more oxygen de-saturations below 94% was uncommon. The null hypothesis could not be rejected because there was no significant difference between the saturation levels recorded for each sub-cohort (μ1 - μ2 = 0), where μ1 was the cohort that received propofol in addition to midazolam and fentanyl, and μ2 was the cohort that received midazolam and fentanyl only. Analysis of the two population groups found them to be quite different in make-up. Had the two population groups been similar in make-up, then μ1 - μ2 may not have equaled 0, and the outcome of the hypothesis test may have been to reject H0. The variables that were significantly associated with low saturations were age, gender and weight. The data showed that males were three times more likely than females to experience low saturations. Patients 45 years and older were nearly eight times more likely to experience low saturations than patients 25 years or younger. Patients classified as being in the gender- specific high weight group were twice as likely to experience low saturations than those in the low and medium weight groups. Neither the dose of midazolam, nor the additional use of propofol were significant risk factors, even after adjusting for the variables of age, gender and weight. ASA Classification (I or II) was not a determinant of risk. Conclusion: Within the limitations of this study, the data support that a single operator/sedationist, working within the RACDS/ANZCA document PS21 guidelines, and supported by a team of experienced dental nurses, can consistently maintain safe oxygen saturation levels when working on ASA I or II patients, regardless of age, gender, gender-specific weight, dose of midazolam, or the additional use of propofol.