Postoperative Nausea and Vomiting in Children undergoing Radiofrequency Catheter Ablation : a 5-Year Review of 475 Patients

Postoperative Nausea and Vomiting (PONV) is an important cause of morbidity in children and is a leading cause of unplanned hospital admission following ambulatory surgery [1]. Patient satisfaction scores have demonstrated that reported quality of care increases significantly when the incidence of PONV decreases [2]. An overall incidence of 13-42% has been quoted for PONV in pediatric patients; however, certain procedures are considered high risk (30-80% probability) for PONV [3]. Two procedures commonly associated with a high risk for pediatric PONV are strabismus repair and tonsillectomy. In a prospective randomized blinded study of the effects of hydration on PONV in children undergoing strabismus repair, 38% of patients who did not receive intraoperative antiemetic medications experienced PONV in the first 24 hours [4]. A recent Cochrane review of steroids in pediatric tonsillectomy stated that patients who received steroid intraoperatively had a lower rate of postoperative emesis (21%) compared to patients who received placebo (48%) [5].

oxide and prophylactic ondansetron compared to an incidence of PONV of 63% in patients who received volatile-based anesthesia, nitrous oxide and prophylactic ondansetron [8].These numbers suggest that RFCA may represent an increased risk for PONV in the pediatric population.
The primary objective of the current study was to quantify the incidence of PONV in the pediatric RFCA population.The secondary objective was to identify those factors associated with increased risk and incidence of PONV and therefore define risk factors for PONV in the pediatric RFCA population.To this end, we designed and carried out a retrospective review that examined our RFCA population's experience of PONV over a five year timeframe.

Methods
Research approval was granted by our Institutional Ethics Committee to perform a retrospective review of the Electronic Patient Charts (EPC) of all patients scheduled to undergo RFCA during the 5 year period, July 2006-June 2011.Our institution is a university-affiliated quaternary referral pediatric centre with a Cardiac Diagnostic and Interventional Unit (CDIU) which performs in excess of 100 RFCAs each year.
Identifying and measuring nausea in children is both difficult and subjective due to the inability of younger children to adequately describe and verbalize the symptom.The only exclusion criterion in this study was children aged less than four years of age because troublesome PONV is rare in children under three years of age and they rarely require prophylactic antiemetic medication [9].
The EPC (Microdea Synergize™, Electronic Patient Chart, ON, Canada) in our institution is a complete record of nursing and medical notes from which we identified and recorded all documented incidences of nausea, vomiting or retching in the post operative period, both in the Post Anesthesia Care Unit (PACU) and up to 24 hours post procedure.Incidents of patient retching were included in the vomiting group as is consistent with previous studies examining PONV [8].We recorded 4 primary outcomes: nausea in PACU; vomiting in PACU; nausea in first 24 hours post-procedure; vomiting in first 24 hours postprocedure.In addition the following information was collected: gender, age, body weight, personal or family history of PONV, the type and duration of anesthetic technique administered and the administration of anti-emetics both intra-and post-operatively.Data sheets were specifically designed to incorporate the described data collection components.On completion of the first 20 data sheets by each member of our research group, we undertook an audit and review of the data to ensure consistency of data entry.Ongoing review was maintained to ensure data integrity.The data was encrypted and stored on a Microsoft ® Excel database.
The data was analyzed by a clinical statistician in order to identify factors associated with an increased risk of PONV.Descriptive statistics, means and standard deviation (for continuous variables), and percentages (for categorical variables) were used to describe the sample and to calculate the incidence of nausea and vomiting after RFCA.Chi-square tests were used to assess the association between two categorical variables, two sample t tests to assess the association between a categorical and a continuous variable, and Pearson/Spearman correlation to assess association between two continuous variables.Fisher's exact probability test was used to compare annual incidences of nausea, vomiting, nitrous oxide and antiemetic use between years studied.Multivariable logistic regression analysis was used to ascertain which of the variables were significantly and independently related to nausea and vomiting in the presence of other factors.Goodness of fit for logistic regression was assessed using the Hosmer-Lemeshow test.All tests were two-sided and p < 0.05 was considered statistically significant.The statistical software SAS 9.3 was used for data analysis.

Results
Five hundred and sixty seven patients scheduled for RFCA between July 1 st 2006 and June 30 th 2011 were identified from the CDIU database.Eighty-eight of these patients were excluded due to age less than four years.A further four patients were excluded because there was no anesthetic record within EPC (two patients) or because the procedure was abandoned due to an intraoperative complication (two patients), leaving a total of 475 patients for inclusion in this study.Demographic data are displayed in Table 1.
Details of the anesthetic and procedural characteristics are displayed in Table 2.The most common type of anesthetic provided for these cases comprised an intravenous induction followed by maintenance with an inhalational volatile agent that was supplemented by intraoperative opioids and neuromuscular blocking agents (NMB).Four hundred and thirty five patients underwent general anesthesia that was maintained throughout RFCA by inhalational volatile agents (91.5%), 25 patients received total intravenous anesthesia (5.3%), and 15 patients received a neuroleptic sedation technique (3.2%).Of the 475 patients included in this analysis, 123 (26%) received no intra-operative anti-emetic medication, 304 (64%) received one intra-operative anti-emetic (ondansetron in 300 of 304 cases), and the remaining 48 (10%) received two or more intra-operative anti-emetics.Two hundred and twenty seven patients (48%) received at least one rescue anti-emetic in PACU post RFCA.The mean duration of  anesthesia was 238 minutes (SD +/-73).Four hundred and seven patients (86.2%) received ablation of an abnormal pathway or focus of electrical activity.
The four primary outcomes are displayed in Table 3.One hundred and one patients (21%) complained of nausea in PACU post RFCA and a total of 140 patients (30%) complained of nausea at some point in the first 24 postoperative hours.One hundred and twenty four patients (26%) vomited in PACU and a total of 167 patients (35%) vomited at some point in the first 24 hours postoperative.Figure 1 demonstrates the annual incidence for each of these variables across the five years studied, in addition to presenting annual incidence for use of nitrous oxide and intraoperative antiemetics.There was a significant decrease in nitrous oxide use for maintenance of anesthesia between Years 1 and 5 (93% and 31%, respectively, p < 0.0001) There was no significant difference identified in the incidence of vomiting between Years 1 and 5 but there was a significant decrease in the incidence of nausea between years 1 and 5, both in PACU (40% and 15%, respectively, p < 0.0001) and, in the first 24 hours postoperative (48% and 17% respectively, p < 0.0001) Table 4 demonstrates the significant associations found with postoperative nausea and vomiting, on both bivariate and multivariable analysis.For bivariate analysis, the following variables were examined: age, gender, PONV history, family history of PONV, method of anesthesia induction, method of anesthesia maintenance, anti-emetics received/not-received intraoperativley, number of intraoperative antiemetic administrations, use of ablation, continued administration of nitrous oxide, duration of anesthesia, volume of intravenous fluid administered, use of midazolam, and, use of neuro-muscular blocking agents.Those predictors that were significant were entered into a multivariable regression model.No patient or procedural features were found to significantly affect the incidence of nausea or vomiting.For all four primary outcomes, prolonged use of intraoperative nitrous oxide and the omission of an intraoperative anti-emetic were associated with a significant increase in the incidence of nausea and vomiting post RFCA.In addition, duration of general anesthesia was associated with an increased incidence of vomiting, both in PACU and in the 24 hour period post RFCA (Table 4).

Discussion
Our results demonstrate that in pediatric patients undergoing RFCA under general anesthesia, 30% experienced nausea and 35% experienced vomiting in the first 24 hours post-procedure across a five year period of study.In 2009, the Association of Paediatric Anaesthetists of Great Britain and Ireland (APAGBI) published guidelines on the prevention of postoperative vomiting in children [9].These guidelines stated that strabismus surgery is perhaps the paediatric surgical procedure with the strongest evidence of PONV risk with a mean incidence for late vomiting of 59% reported in one systematic review [10].Similarly, children undergoing adenotonsillectomy are at increased risk of PONV; however, perioperative opioid administration may be a compounding risk factor as one study demonstrated as low as 11% incidence of PONV post pediatric tonsillectomy when perioperative opioids were avoided.The guidelines also stated that procedures such as groin and penile surgery represent a modest increased incidence of PONV but again, opioid administration may represent a compounding variable.
No guidelines or consensus statements on prevention of PONV in children refer to pediatric RFCA as a procedure associated with an increased risk of PONV, moderate, high or otherwise; however, this may be due to an under-appreciation of PONV in this population due to current lack of evidence.We demonstrate an incidence of 30% for nausea and 35% for vomiting in the first 24 hours post paediatric RFCA which represents at the very least a moderate increased risk of PONV.However, we also demonstrate that when nitrous oxide is delivered throughout the maintenance of general anesthesia for RFCA and no antiemetics are administered, the incidence of nausea in the first 24 hours post RFCA is as high as 48% and the incidence of vomiting is 39% (Figure 1, Year 1).These numbers are within the ranges of high risk of probability for PONV quoted in previous publications (30-80%) and are similar to quoted incidences for PONV in well documented high risk procedures such as strabismus repair and tonsillectomy [4,5,9].
A recent publication revisited guidelines for the prevention of PONV in both children and adults [11].This publication restated the four independent predictors of postoperative vomiting in children: duration of surgery > 30 minutes; age > 3 years; history of POV in patient, parent or siblings; and strabismus surgery.Strategies suggested to decrease baseline PONV risk were: avoidance of general anesthesia by the use of regional anesthesia; total intravenous anesthesia with propofol; avoidance of nitrous oxide; avoidance of volatile anesthetics; minimal opioid use; and adequate hydration.For the paediatric EPS population, regional anesthesia is largely unfeasible and exclusive use of propofol may decrease the ability to stimulate a tachycardia and re-entrant phenomena required during such electrophysiology studies.However, other factors listed could be modifiable.We found an association between several anesthetic features and the development of PONV after RFCA.The use of intraoperative nitrous oxide, maintenance of anesthesia with inhalational volatile agents when duration of anesthesia is prolonged, and the failure to administer an intraoperative anti-emetic were all associated with a significant increase in the incidence of PONV post RFCA.The APAGBI guidelines state that the use of nitrous oxide does not appear to be associated with a high risk of PONV in children, however, our analysis showed that in 65 patients who continued to receive nitrous oxide beyond induction of anesthesia (and who did not receive an intraoperative antiemetic), 26 of those 65 patients (40%) complained of nausea in PACU and 24 (37%) vomited.In comparison, amongst the 81 patients who did not receive nitrous oxide beyond induction of anesthesia (and who also did not receive an intraoperative antiemetic), 17 of those 81 patients (21%) complained of nausea in PACU and 18 (22%) vomited.In the current analysis, the continued administration of nitrous oxide represented a significant risk for increased incidence of PONV post RFCA.No patient factors, including age, gender or previous history of PONV demonstrated significant association with the increased risk of PONV post RFCA.
There is limited published evidence that RFCA is associated with PONV in the pediatric population.Tyndall and colleagues [12] in the late 1990s were the first to report PONV after RFCA in the adult population.In a retrospective chart review of 369 RFCA cases, all of whom received an intravenous prophylactic anti-emetic, they reported a 22% incidence of nausea and a 13% incidence of vomiting; however, the majority of patients in that study underwent RFCA with conscious sedation.The only previous study to investigate PONV after RFCA in the pediatric population prospectively randomised 56 children to undergo RFCA with either an intravenous propofol-based anesthetic or an inhalational isoflurane-based anesthetic [8].The incidence of nausea and of vomiting in the isoflurane group was 63% and 55%, respectively, despite the administration of a prophylactic anti-emetic (ondansetron) to all patients.However, all patients did receive nitrous oxide for the duration of their procedure.Of note, the respective incidences of nausea and of vomiting in the propofol group were significantly lower, 21% and 6%, respectively.Our findings also suggest an increased risk of PONV when prolonged inhalational maintenance of anesthesia is employed for RFCA, however, 91.6% of patients recruited to the current study received inhalational maintenance of anesthesia which makes direct comparison between choice of anesthetic techniques unreliable.Furthermore, because the majority of patients that received an intra-operative anti-emetic received ondansetron alone we were unable to examine whether patients who received an alternative anti-emetic experienced a different outcome for PONV.We did compare those patients who received no intra-operative anti-emetic to those who received one or more anti-emetic.The bivariate and multivariable regression analyses demonstrate that patients who received prophylactic anti-emetics while undergoing RFCA had lower incidence of PONV across the four outcomes studied.Therefore, while PONV may be an undesirable, and unanticipated, sequela of RFCA under general anesthesia, we demonstrate that it does respond to anti-emetic prophylaxis.Comments on the correct choice of anti-emetic, additive effects of antiemetic re-dosing and additive effects of a second or third anti-emetic are beyond the scope of the current study.However, it is important to note that RFCA is a procedure that represents an increased risk of PONV in the pediatric population and this increased risk can be attenuated by the use of antiemetics and the avoidance of prolonged administration of nitrous oxide.
We acknowledge that certain limitations exist within this study.Most notably, the retrospective design likely resulted in an under reporting of the true incidence of PONV as data collection was dependant on accurate recording of PONV in the patient record.Unlike pain or sedation, there is no standardized scale for nausea and vomiting employed in our institution's documentation, therefore we examined all medical and nursing notes throughout the electronic patient record for any documentation or mention of the phrases 'nausea', 'vomiting' or 'retching'.The anesthetic management was not standardized during the time period studied.More importantly, nursing documentation in the patient chart is also not standardized as our institution does not yet use electronic chart documentation.Therefore, we only counted a patient as experiencing nausea if it was explicitly documented by the nurse in the patient chart and there is only one location in the progress notes where a nurse might do so.All incidences of documented vomiting within the progress notes were used as evidence for incidents of POV, however, we also searched the patient fluid balance sheet because it is hospital policy for all incidences of vomiting to be clearly documented in this section.Vomiting is also likely to be documented more accurately than nausea because vomiting more often leads to a medication administration and therefore a justification in the patient chart for that administration.We did not 'overcount' our incidence of POV but it is likely that our incidence of POV is a true value whereas our incidence of nausea is likely underestimated.This explains why our incidence of vomiting is greater across all measured fields than that of nausea.Identifying and measuring nausea in children is both difficult and subjective due to the inability of younger children to adequately describe and verbalize the symptom of nausea.The risk of PONV is known to increase markedly above three years of age [3].Both of these factors lead us to exclude patients under four years of age from inclusion in this study in order to improve overall accuracy of reporting.Anesthesia technique was not standardized during the five year period we studied.Overall choice of anesthetic technique, use of antiemetic prophylaxis, and choice of agents for antiemetic prophylaxis was neither standardized nor mandated.This prohibits us from analyzing or commenting on the efficacy of individual agents (e.g.sevoflurane versus isoflurane, ondansetron versus diphenhydramine, one anti-emetic versus two, total intravenous anesthesia versus volatile inhalational anesthesia) in preventing PONV during RFCA.Furthermore, over the course of the five year time period studied we noted some changes in clinical practice with regards to the management of patients for RFCA.In the first year studied, 93% of patients received nitrous oxide for the duration of the procedure with an incidence of nausea in the first 24 hours post procedure of 48%.During the fifth year studied, only 31% of patients received nitrous oxide for the duration of the procedure with an incidence of nausea in the first 24 hours postoperative was 17%.In the first year studied, 98% of patients received inhalational maintenance of general anesthesia throughout RFCA whereas in the fifth year studied, 78% of patients received inhalational maintenance of general anesthesia throughout RFCA.These changes reflect a gradual shift in clinical practice over the five year timeframe studied, albeit heterogeneously so, nevertheless Figure 1 demonstrates a steady decrease in the incidence of the four primary outcomes for this study during the five years studied.The possibility of this study influencing the practices of our colleagues (the Hawthorne effect) while data collection was ongoing was excluded by the retrospective design of the study.
In conclusion, our results demonstrate that RFCA is a procedure that presents increased risk of PONV in the pediatric population.The administration of prophylactic antiemetic medications while avoiding the use of nitrous oxide should be considered for all pediatric patients who present for RFCA under general anesthesia.

Figure 1 :
Figure 1: Primary Outcomes, use of nitrous oxide and omission of intraoperative antiemetics for each year studied.

Table 2 :
Anesthetic and Procedural Characteristics.

Table 4 :
Significant Associations with Nausea, Vomiting in PACU and in first 24 hours post procedure.