| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anesthesia, Lucile Packard Children's Hospital, Stanford University School of Medicine, Stanford, California; and Department of Anesthesia, San Diego School of Medicine, University of California, San Diego, California.
Address correspondence and reprint requests to: Alice Edler, MD, Department of Anesthesia, Stanford University School of Medicine, 300 Pasteur Drive H3580, Stanford, CA 94305. Address e-mail to edlera{at}aol.com.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: We studied 166 children, aged 1–18 yr, who underwent tonsillectomy and adenoidectomy or tonsillectomy and adenoidectomy, and bilateral myringotomy with tube insertion under general anesthesia. The primary outcome measure was the time spent in the PACU until predetermined discharge criteria were met.
RESULTS: The number of episodes of postoperative nausea and vomiting, patient age, and number of oxygen desaturations contributed significantly (P < 0.05) to prolonged LOS. Each episode of postoperative nausea and vomiting (P < 0.05) or oxygen desaturation to <95% (P < 0.05) increased the patient's LOS by 0.5 h. History of upper respiratory tract infection, emergence agitation, and parental anxiety did not significantly predict increased LOS.
CONCLUSION: This investigation is the first composite view of LOS in pediatric patients. The significance of identifying patients at risk of prolonged LOS prior to anesthesia is of use not only in allocating PACU resource and staffing needs, but also for improving quality of care and ensuring a minimally traumatic anesthetic experience for our pediatric patients and their families.
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
There are several unique aspects to the postanesthesia care of the pediatric patient, such as emergence agitation, postintubation croup, apnea of prematurity, and parental presence that can prolong LOS (3–5). Unexpected postanesthetic adverse events result in prolonged occupation of valuable PACU bed space and nursing staff which can translate into higher hospital costs (6). Along with the recognized expectations for quality anesthesia care, the growing demand for pediatric outpatient surgery and same-day discharge requires a reliable method to identify patients at high risk for prolonged LOS.
Previous studies of LOS in pediatric patients have been limited to catalogues of adverse events such as PONV, pain, bleeding, and postoperative respiratory events (7–9). Other studies have identified individual preoperative risk factors such as history of obstructive sleep apnea and urgency of surgery and their association with LOS (10,11). None of these studies has explored the interaction between these independent factors, nor identified their composite impact on prolonging LOS.
Tonsillectomy and adenoidectomy (T&A) is the second most common surgical procedure in the pediatric population and is associated with many common pre- and postoperative comorbid factors known to influence LOS. With an estimated quarter million tonsillectomies preformed annually in the United States alone, these findings impact a large portion of our total patient population. The goal of this study was to identify independent predictors of prolonged LOS after T&A. We describe a population-based approach to reducing risks and improving anesthetic care by avoiding unnecessary prolonged recovery in pediatric patients.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patient Selection
Eligible study subjects were 1–18 yr of age, ASA physical status (PS) 1 or 2, and scheduled for T&A under general anesthesia at Lucile Packard Children's Hospital (LPCH), Stanford, CA, between August 2003 and September 2004. Data were collected prospectively. Race, gender, English proficiency, or preexisting comorbidity (if ASA PS 2 or less) were not reasons for exclusion. Only patients with planned outpatient status were included in the study. Exclusion criteria included parental refusal, ASA PS more than 2, and emergency surgery.
We elected to standardize the surgical procedure and anesthetic management by selecting children undergoing only T&A or T&A plus bilateral myringotomy and tube insertion. Limiting the types of surgical procedures and standardizing the anesthetic management allowed the investigators to examine only pre- and postanesthetic factors that impact LOS.
Anesthetic Protocol
All patients were anesthetized by staff pediatric anesthesiologists and received premedication with midazolam (0.5 mg/kg PO with an upper limit of 20 mg) for preoperative anxiolysis. General anesthesia was induced by inhalation of inspired sevoflurane (0%–8%) with 70% N2O, rocuronium (0.6–1.0 mg/kg IV) was administered to facilitate tracheal intubation. Anesthesia was maintained with isoflurane (0.5%–2% inspired concentration) in 50% nitrous oxide and fentanyl (1–2 µg/kg IV). All patients received an acetaminophen suppository (30–35 mg/kg PR). A proportion of these patients were also enrolled in a concurrent study of PONV conducted by the Stanford Department of Pediatric Otolaryngology. All patients received either ondansetron or dexamethasone. No other antiemetic was used in this study. The antiemetic regimen was adjusted to "ondansetron-equivalent" dose for analysis based on the model by Apfel et al. (12) All patients received reversal of the nondepolarizing muscle relaxant using neostigmine (0.07 mg/kg IV) with either atropine or glycopyrrolate regardless of the duration of anesthesia.
In the PACU, all patients were monitored according to ASA standards. All patients received supplemental oxygen via blow-by. The PACU at LPCH is composed of one room with no physical separation between Phases I and II. One nurse is responsible for each patient throughout the entire duration of recovery. At LPCH, parents/caretakers are allowed to remain with patients throughout their PACU stay.
Data Collection
After informed consent, both pre- and postoperative data were collected by one of the investigators (KP) using direct observation and parental survey (Table 1). Preoperatively, demographic and medical history variables were collected from a survey given to the parents. Parents were asked, "Has your child had a cold within the past 2 weeks?" At that time, each anesthesiologist made his/her own assessment of the child's upper respiratory infection (URI) history. Parents were questioned at the time of the preanesthetic interview regarding runny nose, cough (productive or nonproductive), nasal congestion, fever, child's complaint of "feeling sick," and administration of any medication (prescription or over-the-counter) to relieve these symptoms. If any of these questions where answered affirmatively, the assigned anesthesiologist was consulted and a decision was made by him/her regarding the continuation of planned anesthesia. Since there are no institutional guidelines for case cancellation, each case was considered on an individual basis. Those children who were felt to be unfit for general anesthesia were postponed to a later date and not asked to participate in the study. If the assigned anesthesiologist did not feel that the recent respiratory symptoms mandated case cancellation despite a history of URI, the child was considered eligible for inclusion in the study. These children were classified as "URI within the past 2 weeks" on the data collection form.
|
Parental anxiety was self-reported on a scale of 1–10. Although parental anxiety has not yet been shown to be associated with prolonged LOS, these data were included for analysis. Patient anxiety was graded by the admission nurse using the 2000 Standards of Perianesthesia Nursing Practice published by the American Society of Peri-Anesthesia Nurses (ASPAN, http://www.aspan.org). Each patient's baseline vital signs were collected from the admission nursing notes.
The choice of morphine sulfate or fentanyl as the postoperative rescue analgesic was left to the discretion of the anesthesiologist. If fentanyl was chosen, the dose was converted to a morphine-equivalent dose for analysis. A ratio of 100:1 fentanyl to morphine relative potency conversion was used to create a dummy variable for additional narcotic administration. This conversion was used to minimize both the number of confounds and the interaction effects in the final regression model. Opiod analgesics other than morphine and fentanyl were not used.
The modified pediatric Aldrete score from ASPAN was used. PONV was defined as an episode of frank emesis, dry heaves, or subjective complaint of nausea (older, verbal children). Scores were recorded from the nurse's notes and were the summation of the number of times that a patient complained of nausea plus the number of episodes of retching or vomiting. The number of oxygen desaturations (Spo2 <95%) was recorded from both the nurse's notes and direct observation by one of the investigators (KP). This investigator followed all patients throughout their phase I and II PACU stay.
Continuous pulse oximetry monitors at LPCH are set to alarm when Spo2 <95%. Any alarm event, when not attributed to artifact due to movement or removal of the saturation monitor probe, was recorded as a desaturation. Both the primary "one-on-one" PACU nurse and the investigator assigned to the subject confirmed each desaturation event. The PACU record and the data sheet were reconciled at the end of the day to reflect an accurate accounting of desaturations in the data.
Agitation scores were graded on a five-point scale that defined the patient's behavior in the PACU. The scores 1–5 corresponded to "sleeping, awake and calm, irritable and crying, inconsolable, and restless and disoriented" (3). Vital signs, dosage, and time of administration of opiod pain medication and additional antiemetic medication were recorded from the nurse's notes.
Duration of recovery in the PACU was measured from time of arrival in the PACU until the nurse's notation of "home ready" in the patient's chart. Discharge criteria from the PACU are based on guidelines from ASPAN; a standard Aldrete score modified for pediatric patients was used. Pre- and postanesthesia nursing scores include the assessment of ventilation, pain, hemodynamic stability, level of consciousness, mobility, and presence of PONV. Sufficient discharge criteria according to LPCH protocol require the return to at least 80% of the preoperative nursing assessment. Both the time of nurse's notation of home ready and the actual time leaving the PACU, if different than 5 min, were recorded.
Outcome Measures
The primary outcome measure was the total duration of time (minutes) spent in the PACU. Additionally, if the child was deemed home ready by virtue of his/her discharge score but remained in the PACU, the time between home ready and actual departure was recorded along with an explanation for the delay. Only time to home ready was used in the analysis.
Statistical Analysis
The sample size was estimated using Power Calculator software (http://calculators.stat.ucla.edu/powercalc, Department of Biostatistics, UCLA, Los Angeles, CA). Assuming a power of 80% and 
of 0.05 for significance in the final regression equation, it was determined that a sample size of 177 patients would be needed to account for double the variance reported in previous literature (r = 0.12). SPSS Version 13.0 (SPSS Inc., Chicago, IL) was used for the analysis of all data.
Initially, all pre- and postoperative variables were entered into univariate correlation analysis with LOS using Pearson's or Spearman's correlation coefficient for parametric or nonparametric data. Only those variables showing significant (P < 0.05) or moderately significant correlations (P < 0.10) were entered into the preliminary regression model. Forward stepwise regression was then performed. The requirement for entry into the model was a probability of F change 
0.05. To account for the effect of varied antiemetic regimes on PONV and the effect of additional PACU opiod administration on oxygen desaturation, two first-order interaction terms were entered into the model at all trials and remained in the final model. An antiemetic interaction term was calculated from the number and relative strength (ondansetron-equivalent dose) of antiemetic regimes used intraoperatively based on the model presented by Apfel et al. (12). An interaction term used to consider the effect of supplemental postoperative dosage of opiod (expressed as mg/kg morphine-equivalent dose) and number of oxygen desaturations was also entered into the model.
Logistic regression was used to estimate the relative contributions of these variables to prolonged LOS. The regression model used the same model-building techniques previously mentioned. However, only final linear model elements (age, number of episodes of PONV or retching, and number of oxygen desaturations) were entered into the logistic equation. Both first level interaction effects were retained in the model. In all final models, P < 0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
The mean total LOS from PACU admission until home ready was 106 ± 52 min. Initial univariate analysis revealed that age, PONV, and oxygen desaturation were significantly correlated (P < 0.05) with increased LOS. Preoperative parental anxiety was moderately correlated (P < 0.1). Because the literature reported recent URI and postoperative agitation as factors that could prolong LOS, these variables were forced into the initial model. However, history of recent URI, emergence agitation, and parental anxiety failed to meet minimum entry level requirements (P > 0.05) when entered stepwise and were therefore withdrawn.
First-level interaction effect between PONV and antiemetic regimen and the interaction effect between oxygen desaturation and the supplemental administration of opiod were retained in the model. Regression diagnostics were performed. There was no significant collinearity or heteroscedasticity found (Fig. 1). The final model resulted in significant associations (P < 0.05) between increasing age, number of episodes of PONV or retching (accounting for antiemetic treatment), and number of oxygen desaturations when supplemental PACU opiod administration was considered.
|
The most significant factor was the presence of PONV (Table 4). Each episode of PONV or retching prolonged LOS by 
h (ß = 31 min, 95% CI 7–54 min, P < 0.05). Oxygen desaturation played a near-equivalent role in increasing PACU LOS. Each oxygen desaturation to <95% resulted in approximately h increase in LOS (ß = 28 min, 95% CI 10–47 min, P < 0.05). Age played a small, but statistically significant, role in the total PACU LOS. For each year of age over 1, the chances for prolonged LOS decreased by 2.2% (95% CI 0.01–0.05, P < 0.05).
|
The relative risk of having prolonged LOS with PONV or pulmonary complications compared to prolonged LOS without adverse events can be estimated using adjusted odds ratios (Table 5). Patients without PONV had more than 80% less risk of being in the highest quartile (>75th percentile) for PACU LOS (adjusted OR 0.19, 95% CI 0.078–0.537, P < 0.05). Patients who did not experience oxygen desaturation had more than 50% less risk of being in the highest quartile for PACU LOS (adjusted OR = 0.46, 95% CI 0.29–0.8, P < 0.05).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In younger patients, oxygen desaturation is a good predictor of prolonged LOS. Patients without episodes of desaturation in our study required 50% less recovery time in the PACU. Children who undergo otolaryngology procedures are 1.5 times more likely to develop adverse respiratory events compared to other surgical procedures, and this risk is inversely proportional to age (13). In addition to patient safety issues, oxygen desaturations in the PACU, leading to prolonged LOS, expend valuable PACU resources, increase parental anxiety, and negatively impact operating room flow.
Interestingly, parents' report of recent URI (less than 2 wk) was not found to be a factor contributing to prolonged LOS in our study. Tait et al. (14) found an association between recent URI and short-term respiratory events (postoperative breath-holding and desaturation), but none of the patients in that study demonstrated long-term sequelae. The authors concluded that children with recent URI should be eligible for outpatient procedures. In our study, patients with a history of URI who were evaluated and cleared by their anesthesiologist on the day of surgery did not develop oxygen desaturation or require prolonged LOS.
Patients in our study who did not experience PONV had up to 80% shorter recovery times in the PACU. PONV has been identified in past studies as a cause for prolonged postanesthesia recovery and unplanned hospital admission in children (15,16). Several investigations in adults have shown that the use of more than one antiemetic drug for prophylaxis may be superior to single drug therapy (12). In patients undergoing T&A, the routine administration of dexamethasone significantly reduces both the incidence of PONV and postoperative airway complications associated with airway edema (17–19). Aggressive prophylaxis against PONV, including combination antiemetic therapy and adequate replacement of fluid deficits, is indicated for pediatric patients undergoing T&A to prevent prolonged LOS and unplanned admission. Adequate prevention of PONV also increases patient and parent satisfaction after outpatient surgery (20).
Our model was able to account for 43% of the variability in LOS, which was significantly more than past investigations. Our focus on a specific anesthetic protocol and one type of surgical procedure most likely contributed to this relatively large proportion compared to previous studies. The inclusion of only pediatric patients undergoing T&A may have limited the generalizability of our results to the overall pediatric surgical population. However, T&A is a common surgical procedure in the pediatric population and is performed in hospital and day-surgery venues. Therefore, we believe that our conclusions regarding the prevention of oxygen desaturation and PONV can be useful to anesthesia providers and perianesthesia nurses in many settings.
Identifying the most likely causes of prolonged LOS and at-risk groups before surgery can impact PACU staffing and operating room efficiency. Factors prolonging LOS prevent timely PACU discharges, occupying valuable bed spaces and nursing staff, and potentially delaying scheduled surgical cases. Scheduling of pediatric outpatients requires careful patient selection. A cost-comparison of scheduled outpatient surgery and planned overnight admission has shown outpatient surgery to be less costeffective when the cost of prolonged PACU LOS is considered (21). Our findings suggest that aggressive prophylaxis of PONV and careful airway management can potentially reduce LOS and prevent these unanticipated admissions.
Although our statistical model accounts for 43% of the variability in LOS, 57% remains unaccounted for. This finding confirms the presence and importance of individual variability in postanesthetic recovery. For example, a factor such as emergence agitation may affect PACU LOS in one patient but not another for reasons that may be particular to each patient.
In conclusion, the safe and efficient scheduling of pediatric outpatient surgery requires the ability to identify appropriate candidates and anticipate postoperative needs. Our study demonstrates age, PONV, and oxygen desaturation as the major factors associated with prolonged LOS, and consideration of these factors should precede the decision to schedule high-risk pediatric patients on an outpatient basis. Focused efforts to decrease the risks of PONV and oxygen desaturation, such as multidrug antiemetic prophylaxis, meticulous airway management by skilled anesthesiologists and PACU nurses, provision of supplemental oxygen, and routine use of dexamethasone to decrease airway edema, can improve outcomes in pediatric surgical outpatients.
|
Footnotes |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. F. White, O. Sacan, N. Nuangchamnong, T. Sun, and M. R. Eng The Relationship Between Patient Risk Factors and Early Versus Late Postoperative Emetic Symptoms Anesth. Analg., August 1, 2008; 107(2): 459 - 463. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Edler and B. Golianu Factors Influencing Postanesthesia Recovery After Pediatric Ambulatory Tonsillectomy and Adenoidectomy Anesth. Analg., November 1, 2007; 105(5): 1513 - 1513. [Full Text] [PDF]
|
|
|
|
|
|
M. H. Lane, F. Flynn, L. H. Hepburn, and S. Jaggar Factors Influencing Postanesthesia Recovery After Pediatric Ambulatory Tonsillectomy and Adenoidectomy Anesth. Analg., November 1, 2007; 105(5): 1512 - 1513. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
