Abstract
Introduction
Defining obesity in terms of percentage of body fat is not practical and hence skin fold thickness and body mass index (BMI) have been widely used for this purpose. In adults, obesity is defined as a BMI > 30. This definition cannot be used in children because the median BMI changes with age. For example, the median BMI at birth is 13 kg/m2, 17 kg/m2 at infancy, 15.5 kg/m2 at the age of 5 and 21 kg/m2 at the age of 18. The median BMI for girls and boys are different at the same ages. Hence age- and sex-specific BMI cutoff points defining obesity and its different grades have been developed and used in children between 2 and 18 years of age.
Obese children have a higher prevalence of medical comorbidities. During anaesthesia and the postoperative period, a higher incidence of respiratory and airway adverse events, and longer postoperative lengths of stay, have also been demonstrated in obese children.1–11
Though it has been shown that the presence of certain factors such as endotracheal intubation, airway surgery, obstructive sleep apnoea (OSA), reactive airway disease, active/recent upper respiratory tract infection (URTI), and a younger age increase the risk of perioperative respiratory adverse events in obese children, the independent role of childhood obesity in increasing the risk of perioperative respiratory adverse events is still being evaluated. 9 This is because in many previous studies the prevalence of severe and morbid obesity has not been quantified and in a few others, airway surgery, which is a major contributing factor for perioperative respiratory adverse events, has not been excluded.
The prevalence of childhood obesity among the paediatric surgical population in Southeast Asia and the incidence of adverse events among obese children during anaesthesia has not yet been studied.
Hence the aims of our study were to evaluate the prevalence of mild, moderate and severe obesity in children having anaesthesia in Singapore and the independent effects of obesity on perioperative respiratory adverse events.
We chose to study children having minor non-airway ambulatory surgery because studying this population, which has not been previously studied, would enable us negate the effects of other factors as mentioned above that could mask the independent contribution of obesity towards adverse events.
What we already know
Obese children who are of a younger age are at a higher risk of perioperative adverse events in the presence of comorbid conditions such as OSA, reactive airway disease, active/recent URTI and if they have endotracheal intubation or airway surgery.
What our study demonstrates
Severely obese children are at a higher risk of perioperative respiratory adverse events during non-airway day-case surgery, despite the absence of any comorbid conditions or endotracheal intubation.
Methods
After obtaining institutional review board approval (Singhealth CIRB reference number 2013/780/D), we used the Paediatric Anaesthesia Quality Assurance Database to identify children who had day-case anaesthesia in our institution between January 2012 and December 2014. There were more than 10,000 patients. Instead of choosing 1000 patients in succession, every tenth patient in the database was selected for study inclusion to enable a random selection of patients and to avoid selecting patients who were anaesthetised in a time interval of just a few months, as this might have led to selection bias.
Children who had a day-surgery procedure under general anaesthesia and were discharged home the same evening were included. Children younger than 2 years were excluded because of a lack of BMI cutoff points that define overweight and obesity in this age group. Children who did not have ‘height’ data in our database were also excluded, as this precluded us from calculating the BMI.
Data gathered from the quality assurance database, the patient anaesthetic records, and day-surgery unplanned admission records were patient’s age, sex, race, type of surgery, weight, height, medical comorbidities, airway device, difficult airway, difficult intravenous (IV) access, surgical procedure, postoperative nausea and vomiting (PONV), perioperative adverse events, postoperative length of stay, and unplanned overnight admissions.
International age- and sex-specific BMI cutoff points defining overweight and obesity for children aged 2 to 18 years, passing through BMI of 25 and 30 kg/m2 at the age of 18, derived by Cole et al., using pooled international data for the BMI of children, including data from Singapore, were used to categorise our patients into overweight and obese groups, respectively. 12 Newly derived age- and sex- specific BMI cutoff points by Bervoets and Massa, passing though BMI of 35 and 40 kg/m2 at 18 years, respectively, were used to define severe and morbid obesity. 13
The primary outcome was prevalence of perioperative respiratory-airway adverse events, defined as adverse events during induction, maintenance and/or recovery from anaesthesia with involvement of the respiratory system and reported in our database as a ‘respiratory adverse event’. These were events that resulted in unexpected significant desaturation to less than 90% due to laryngospasm, bronchospasm, airway obstruction and other airway and respiratory problems.
Secondary outcomes were difficult bag mask ventilation, difficult intubation, aspiration, PONV requiring antiemetics in the postanaesthesia care unit (PACU), postoperative length of stay and unplanned admission.
Based on national anthropometric data in Singapore, a prevalence of 10% obesity in our day-case population was assumed. With a respiratory adverse event reporting rate of between 0.5% and 1% in our day-surgery centre, a sample size of 1102 children was calculated to detect a three-fold increase in the incidence of perioperative respiratory-airway adverse events, with a power of 80% and an alpha error of 5%. Chi-squared and Fisher exact tests were used to compare proportions and results were expressed in odds ratio (OR), with independent sample
Results
A total of 1158 patients were included for final data analysis. Based on the age- and sex-specific BMI cutoff points, our sample had 101 (8.7%) obese, 181 (15.6 %) overweight and 876 (75.64%) normal-weight children. Age- and sex-specific BMI cutoff points were used to further classify obese children as severely obese, and our sample had 30 (2.58%) severely obese children. The anthropometric data are shown in Table 1. The distribution of BMI in our study population for girls and boys is shown in Figures 1 and 2.
Anthropometric data of children having day case procedures.
Age-specific body mass index (BMI) cutoff reference lines for girls with scatter of BMI from our sample.
Age-specific body mass index (BMI) cutoff reference lines for boys with scatter of BMI from our sample.
A total of 320 (27.6%) children had medical comorbidities. Most of these patients had a respiratory condition (
Comparison of medical co-morbidities between groups.
CI: confidence interval; N: normal; Ob: obese; OR: odds ratio; OSA: obstructive sleep apnoea; Ow: overweight; URTI: upper respiratory tract infection.
There was a significantly higher prevalence of medical co-morbidities in obese group when compared to normal group.
respiratory conditions included various co-morbidities such as asthma, obstructive sleep apnoea and recent upper respiratory tract infections.
Overweight group had higher prevalence of OSA when compared to normal group.
obese group had a higher prevalence of OSA when compared to normal group.
The surgical procedures were minor non-airway procedures such as inguinal herniotomy, orchidopexy, circumcision, excision of small lumps, minor hand-deformity corrections, minor burns dressing changes, oesophagoduodenoscopy and/or colonoscopy, examination of ears with myringotomy, hearing tests, joint injections, removal of k wires and small orthopaedic implants, or dental restorations and extractions. Most children had a general surgical procedure; an ear, nose and throat; or an orthopaedic procedure (Table 3). Overweight and obese children had a significantly higher number of orthopaedic procedures than did normal-weight children. There was no difference in the use of laryngeal mask airway or endotracheal tubes for airway management among the three groups (Table 4).
Comparison of surgical procedures between groups.
CI: confidence interval; ENT: ear, nose and throat; N: normal; Ob: obese; OR: odds ratio.
Children in overweight and obese groups had significantly higher number of simple orthopaedic procedures when compared to the normal group.
Comparison of airway devices used between the groups and outcomes.
CI: confidence interval; ETT: endotracheal tube; LMA: laryngeal mask airway; N: normal; Ob: obese; OR: odds ratio; PONV: postoperative nausea and vomiting; PRAEs: perioperative respiratory adverse events.
and b are expressed as p values, because length of stay data were analysed using t-tests.
Outcomes
On comparing the normal-weight with the overweight group, there was no significant difference in the incidence of perioperative respiratory-airway adverse events, overall adverse events, antiemetic requirement, length of stay in recovery and unplanned overnight postoperative admission to the ward (Table 4).
On comparing the normal-weight with the obese group, there was no significant difference in the incidence of perioperative respiratory-airway adverse events, overall adverse events, antiemetic requirement, length of stay in recovery and unplanned overnight postoperative admission to the ward (Table 4). There were no reports of difficult bag mask ventilation or aspiration.
On comparing normal-weight and severely obese children, the incidence of perioperative respiratory-airway adverse events was significantly higher in the severely obese group (one, 3.44%) than in normal-weight children (three, 0.34%), with an OR of 10.04 (1.01–99.5).
There were no differences in the demographic and anthropometric profiles, comorbidity profile, airway devices used or surgical procedure types between these two groups. There were no differences in the overall adverse events, antiemetic requirement, length of stay in recovery or unplanned overnight postoperative admission to the ward between the normal-weight and severely obese groups. There were no reports of difficult bag mask ventilation or aspiration, and the only difficult laryngoscopy occurred in a normal-weight child.
Discussion
This study is the first to review prevalence of obesity among children having general anaesthesia in Singapore. The prevalence obesity in our study is reflective of childhood obesity in Singapore’s paediatric population (8.7% vs 11%).14,15 About 3500 children are anesthetised as day cases undergoing minor non-airway procedures in our hospital every year. Our sample included 30 (2.58%) children with severe obesity.
Similar to previous studies, our study demonstrated that obese children had a higher prevalence of overall medical comorbidities and of OSA when compared to normal-weight children. Obese children in our study did not have a higher prevalence of asthma. None of the children in our study had gastroesophageal reflux, IDDM, or hypertension. The lower prevalence of certain comorbidities in our study could be indicative of a lower threshold for inpatient categorisation in the presence of a medical condition in our centre.
The overall adverse event rate stands at 3% in our department, with a rate of 0.5% to 1% in children undergoing day-case anaesthesia. About 70% of these events are respiratory adverse events. 16 This rate is lower than the respiratory adverse event rate of 2.8% reported in the study by Subramanyam et al. in their day-surgery population probably because of exclusion of airway surgery in our study. 11
A previous study similar to ours with 1133 children younger than 12 years having day-case dental procedures under general anaesthesia showed a higher incidence of minor respiratory complications, such as perioperative desaturation and unexpected overnight hospital admissions, in obese when compared to normal-weight children; the prevalence of URTI or reactive airway disease was not higher in the obese children in that study. 8 Nevertheless, the prevalence of severe obesity and its effect on perioperative adverse events was not quantified.
Another retrospective study of more than 6000 children having general anaesthesia both as inpatients and day cases demonstrated a higher incidence of difficult airway, postoperative upper airway obstruction and prolonged length of stay in PACU. Obese children had a higher prevalence of asthma, diabetes and hypertension than did other children. No association between the airway device, medical condition, or the type of surgery and the incidence of adverse events in obese children was demonstrated. 10 This study also did not quantify severe obesity and its contribution to perioperative adverse events, and it included airway surgery.
A prospective study of 2025 children having general anaesthesia for elective non-cardiac surgery either as inpatients or day cases demonstrated an increase in the incidence of overall and critical respiratory adverse events, including difficult mask ventilation, major airway obstruction, major oxygen desaturation and bronchospasm, in obese children along with a higher incidence of asthma, IDDM, hypertension and OSA. This study demonstrated that children who had asthma, OSA, snoring, endotracheal intubation, airway surgery and/or obesity had higher risk of critical respiratory adverse events than did other children. 9 This study again did not quantify severe obesity and its role in perioperative adverse events and did not exclude airway surgery.
In a recent study by Subramanyam and colleagues, morbid obesity was an independent predictor of perioperative respiratory and airway events in children having day-surgical procedures. Nevertheless, airway surgery was not excluded. 11
Our study, having excluded airway surgery and quantified severe obesity, demonstrated a significantly higher incidence of respiratory and airway adverse events in the severely obese group than in the normal-weight group, despite a lack of difference between prevalence of respiratory comorbidities between the two groups. We did not demonstrate a significantly higher incidence of perioperative respiratory-airway adverse events in obese children than that in normal-weight children.
Severely obese children had more-severe adverse events than did normal-weight children.
One of these children needed unexpected positive-pressure ventilation intraoperatively; desaturated postoperatively, needing continuous positive airway pressure (CPAP) ventilation; and had an unplanned overnight admission to the high-dependency unit. The other obese child experienced a dental injury during a grade 2 laryngoscopy.
The airway and respiratory adverse events in children who were not obese were considered to be minor. One was dislodgement of the laryngeal mask on movement of the head and two other cases of mild, intraoperative laryngospasm. One of the children who experienced laryngospasm had also had a recent URTI.
Our study did not demonstrate association between perioperative respiratory-airway adverse events and OSA because of the small number of children with OSA. Our study did not demonstrate an increase in the unplanned overnight admission rate among severely obese, obese, or overweight groups compared to that of the normal-weight group, probably because PONV was the most common reason for overnight admission among the day cases. Our study did not demonstrate an increase in the incidence of difficult bag mask ventilation or difficult laryngoscopy among severely obese and obese children, possibly because of the lower incidence rate (10 per 8000 cases) of difficult airways in our centre, exclusion of potential or proven difficult airways from day-case anaesthesia, low numbers of severely and morbidly obese children in our sample, and the retrospective nature of the study.
One of the major limitations of our study is its retrospective nature. Details of difficult bag mask ventilation, minor desaturations at PACU needing airway support or oxygen supplementation, mild laryngospasms treated with CPAP only, oxygen administration during transfer of patients to PACU, the number of attempts at IV cannulation, or difficulty with positioning and non-invasive blood pressure monitoring were not captured.
Conclusions
We conclude that severely obese children, even in the absence of any associated comorbidities, have a higher risk of perioperative adverse events during minor non-airway day-case anaesthesia.
Obese children who are not severely obese can have simple non-airway surgery without a higher risk of perioperative respiratory adverse events. Additional comorbidities such as moderate to severe OSA, untreated OSA, moderate to severe asthma or other such respiratory illness, IDDM, and left ventricular hypertrophy with hypertension may increase the risk of respiratory adverse events in these children.
Recommendations
We recommend the use of age- and sex-specific BMI cutoffs or BMI percentile charts to identify children who are obese and especially severely obese to anticipate and prevent major respiratory adverse events. Prospective research is needed to better understand the role of obesity in increasing the risk of perioperative respiratory-airway adverse events, the need for prolonged observation or postoperative admission, the avoidance of ambulatory care in severely obese children, and the utility of obesity classification in children having general anaesthesia.