Incidence and Risk Factors for Perioperative Adverse Respiratory Events in Children Who Are Obese

Approval from the University of Michigan Institutional Review Board (Ann Arbor, Michigan) and waiver of informed consent were obtained for this prospective observational study. Children aged 2–18 yr undergoing any elective surgery, with the exception of repair of a cardiac anomaly, were included in the study. A nonprobability consecutive sampling technique was used in which all obese children presenting for elective surgery were included, plus a large sample of normal and overweight children selected at random (using tables of random numbers) from the daily surgical schedule. Preoperative height and weight were used to calculate BMI using the formula BMI = weight/height². Children were then classified as obese using the method of Cole et al. , which uses combined international data to adjust BMI according to the age and sex of the child.6,7In addition, weight, height, and age were collected on all children undergoing an elective procedure during a single 2-week period to estimate the overall prevalence of obesity and overweight in our population.

The following data were collected prospectively by trained research assistants: demographics (including weight, height, and age), type and duration of anesthesia and surgery, and comorbidities. Diagnoses of comorbidities such as asthma and OSA were obtained from the history and physical and from the medical record. Data regarding the incidence and severity of respiratory adverse events were collected prospectively at five time points throughout the perioperative period (i.e. , induction, tracheal intubation or placement of a laryngeal mask airway, during maintenance, at emergence, and during postoperative recovery) by the anesthesia provider managing the case. Furthermore, the cumulative incidences of any single adverse respiratory event and any single major (critical) event (see table 1for definitions) were recorded. Although none of the providers had a vested interest in the study, they were generally unblinded to the patient’s BMI category. The research assistant followed up with the anesthesia care provider to ensure completeness of the documentation. Although the anesthetic management of each patient was at the discretion of the individual anesthesia provider, the providers were asked if they had modified their anesthetic plan because the child was obese.

Other adverse events, including aspiration, postoperative nausea and vomiting, and any unplanned hospital admission, were noted. Per routine practice, parents were telephoned after discharge, and data related to postoperative nausea and vomiting, surgical site appearance, and any other complications were recorded.

Data were analyzed using SPSS® software (SPSS Inc., Chicago, IL). Parametric data such as age, anesthesia, and surgical times were analyzed using analysis of variance followed by post hoc  pairwise comparisons using Tukey or Dunnett C tests depending on equality of variances and sample size. Data are presented as mean and SD. Nonparametric data were analyzed using chi-square with Fisher exact tests (where appropriate) and are presented as n (%). Independent variables that were found by univariate analysis to be significantly associated with the dependent outcome measures were force entered into a multiple logistic regression model. P  values less than 0.025 were accepted as statistically significant.

An a priori  sample size was determined based on results from a previous study from our department, in which the overall incidence of adverse respiratory events in children was 24.2%.8Accepting a 35% increase in adverse respiratory events in obese over nonobese children as clinically significant and given that the ratio of normal to obese children in our population is approximately 4:1, we required minimum sample sizes of 1,181 and 294, respectively (α= 0.05, β= 0.2, two-sided). Overweight children were sampled in approximately the same proportions as the obese children.

Completed data were collected from a total of 2,025 children who presented for surgery during a 20-month period (April 2004 through December 2005). Two hundred ninety-four children (14.5%) were classified as obese, 351 (17.3%) were overweight, and 1,380 (68.1%) were normal weight. The age-adjusted BMI of this sample was similar to the distribution of weight classification obtained from the separate sample of all children who underwent elective surgery over a defined 2-week period (i.e. , 9.5, 21.9, and 68.6%, respectively). Initial analysis of the three groups showed that the normal-weight and overweight children were demographically similar and behaved in exactly the same way with respect to the outcomes. Therefore, for clarity of analysis and presentation, we decided to restrict our analysis to two groups only, i.e. , obese and nonobese (normal and overweight combined).

The demographics of the study sample are presented in table 2. There were no differences in the prevalence of upper respiratory tract infections between groups. However, children who were obese were significantly older than the nonobese children and had a significantly higher prevalence of coexisting conditions, including asthma, hypertension, OSA, gastric reflux, and type II diabetes mellitus. Table 3presents a description of the preoperative airway assessment for each cohort and shows that obese children were less likely to be classified as Mallampati I compared with nonobese children.

The airway and anesthetic techniques used in the care of these children are presented in table 4. Tracheal intubation was successful on the first attempt in 88–90% of cases, with no differences between groups. Approximately 12% of anesthesiologists caring for obese children reported an a priori  change in their anesthetic plan because of the child’s obesity. There was, however, no difference in the incidence of critical respiratory adverse events between those who had specifically changed their plan and those who had not (51.6% vs.  38.6%, respectively). Induction of anesthesia via  the intravenous route, use of muscle relaxants, and controlled ventilation were more commonly used in obese children compared with the nonobese. When controlling for these variables, the incidence of critical respiratory adverse events was greater in the obese group compared with the nonobese group: controlled ventilation (odds ratio [95% confidence interval]: 2.25 [1.52–3.34]; P < 0.01), muscle relaxant (2.45 [1.54–3.89]; P < 0.01), and intravenous anesthesia (2.49 [1.59–3.9]; P < 0.01).

Table 5compares the incidence of all perioperative respiratory adverse events between obese and nonobese children. There was an increased incidence of critical respiratory adverse events among the obese children. This observation was maintained when controlling for surgery involving the airway (odds ratio [95% confidence interval]: 1.32 [1.03–1.69]; P < 0.01), history of asthma (1.61 [1.22–2.15]; P < 0.01), OSA (1.25 [0.94–1.67]; P < 0.01), and age younger than 10 yr (1.98 [1.37–2.89]; P < 0.01).

Exploratory univariate analysis identified several factors that were associated with overall critical respiratory adverse events (table 6). These included asthma, obesity, age, tracheal intubation, procedures involving the airway, and histories of asthma, snoring, or OSA. Correlations between these variables were examined using the Spearman rank correlation coefficient. Moderate correlation was observed between OSA and snoring (ρ= 0.45; P < 0.001); however, analysis revealed only low correlations (< 0.15) between the other variables. Factors from the univariate analyses were subsequently forced into a logistic regression model, which yielded five independent risk factors for critical respiratory adverse events in this surgical population of children (table 6).

Four children, all in the nonobese group, experienced apparent aspiration intraoperatively but did not require treatment beyond suctioning of the oropharynx. Seven children vomited during induction of anesthesia (six in the nonobese group and one in the obese group); however, none developed adverse sequelae. There were no differences in the incidence of postoperative vomiting between the nonobese and obese groups (8.3% and 7.9%, respectively). Three children (all in the nonobese group) were noted to have grade 1 pressure sores in the recovery area. Four hundred sixty-two children (22.3%) were telephoned by the recovery room staff after surgery. The proportions of obese and nonobese children who were contacted were the same (21.8% and 22.9%, respectively). There were no differences in parent report of surgical site problems between the nonobese and obese groups (2.3% and 3.1%, respectively) or return to activity/ambulation (69.0% and 75.4%). However, children who were obese had less nausea and vomiting at home compared with the nonobese group (4.8% vs.  16.8%; P = 0.013). One child in the nonobese group returned to the emergency department after discharge home due to prolonged vomiting.

Given the increasing national and international upward trends in the incidence of obesity1and the fact that obese individuals are more likely to experience health-related problems, anesthesiologists are now presented with an increasing proportion of adult and pediatric patients who are overweight or obese. This observation is further evidenced by the increasing number of adults and even adolescents who are opting for bariatric corrective surgery9and is reflected in the results of our study wherein 32% of children presenting for anesthesia and surgery were overweight or obese. These results are consistent with national trends and are also similar to results from a recent retrospective audit in children.1,5As such, they lend some external validity to our findings.

Although the definition of adult obesity based on BMI (30 kg/m²) is well established, there is no universally accepted criterion for use in children. This is partly because childhood BMI changes significantly as a function of age and sex. The Centers for Disease Control and Prevention criteria10define overweight in children as BMI of greater than 85th to less than 95th percentile and obese as 95th percentile or greater, but some would argue that these cutoffs are too arbitrary and may not reflect the age-related changes in BMI that occur during childhood. In response to this, Cole et al.  7constructed age- and sex-specific centile curves based on BMI data from 192,727 children from six countries. Age- and sex-specific cutoffs for obesity were then defined as those on or above the BMI curve, which passes through the 30-kg/m²point at age 18 yr. This approach, therefore, provides internationally relevant data that corrects BMI for age and sex and links it to the current accepted definition of adult obesity.6,11 

Childhood obesity is associated with a number of important medical comorbidities, including type II diabetes, asthma, OSA, hyperlipidemia, hypertension, and heart disease.12–18In our study, obese children were more likely to present with histories of hypertension, OSA, asthma, snoring, reflux, and type II diabetes mellitus. Chung et al.  3identified several preexisting medical conditions as predictors of adverse events in patients undergoing elective surgery, including obesity, hypertension, asthma, smoking history, and gastroesophageal reflux. Consistent with these results, our study identified obesity and a history of OSA as independent predictors of perioperative critical adverse respiratory events in children.

Most adult studies have shown that overweight and obesity are risk factors for perioperative adverse events, including difficult mask ventilation,4,19laryngoscopy,20,21aspiration, postoperative atelectasis,22and surgical site infection.23However, despite the adult data, there is a paucity of information regarding outcome in overweight or obese children who present for anesthesia and surgery. In one recent retrospective study, there was a small increase in minor respiratory complications in children who were obese.6Compared with normal-weight children, those who were obese had a significantly greater risk of intraoperative oxygen desaturation (2% vs.  0.19%, respectively) and unexpected hospitalization (2% vs.  0.19%). In another database audit study, Nafiu et al.  5showed that compared with normal-weight children, obese children had a greater frequency of difficult mask ventilation (7.4% vs.  2.2%, respectively), difficult laryngoscopy (1.3% vs.  0.4%), and postoperative airway obstruction (1.6% vs.  0.07%). Consistent with these studies, our study revealed that children who were obese had a greater incidence of difficult mask ventilation, airway obstruction, bronchospasm, major oxygen desaturation, and overall critical respiratory events. Our study, however, demonstrated no association between obesity and difficult laryngoscopy. Indeed, tracheal intubation was deemed equally successful on the first attempt regardless of BMI. Although the results of these other pediatric studies were similar to ours in terms of trends, one should note that the absolute frequencies of adverse events were significantly lower than ours. The reasons for this are unclear but may reflect differences in study design particularly with respect to the potential for underreporting that can occur when analyzing retrospective data.

A few points regarding the limitations of this study deserve mention. First, this is a nonrandomized, nonblinded study and, as such, there is some potential for selection and/or observer bias. However, despite this, we believe that any effect would be minimized by the large sample size and the fact that all data were collected by individuals with no vested interest in the study’s outcome. We also recognize that because the determination of sample size was based on expected differences in the incidence of critical perioperative adverse events between normal-weight and obese children, the sample may not have been sufficiently powered to detect differences in all individual univariate and multivariate comparisons.

To our knowledge, this is the first large-scale prospective study examining the effect of overweight and obesity on perioperative outcome in children undergoing elective noncardiac surgery. Results of this study confirm the observed upward trend in the prevalence of overweight and obesity among children presenting for anesthesia and surgery. Furthermore, these results support our hypothesis that the incidence of perioperative adverse respiratory events is increased in children who are obese. Although all adverse respiratory events were easily managed and there were no serious sequelae, identification of risk factors will be important as a means to anticipate, recognize, and treat complications that may disproportionately occur in obese children, and thus optimize their anesthetic care.

The authors thank Jordan Fennema, B.S., and Jennifer Hsieh, B.S. (Research Assistants, Department of Anesthesiology, University of Michigan, Ann Arbor, Michigan), for help with subject enrollment and data collection.