Management of Postoperative Nausea and Vomiting in Children

Transcription

1 Pediatr Drugs 2007; 9 (1): REVIEW ARTICLE /07/ /$44.95/ Adis Data Information BV. All rights reserved. Management of Postoperative Nausea and Vomiting in Children Anthony L. Kovac Department of Anesthesiology, University of Kansas Medical Center, Kansas City, Kansas, USA Contents Abstract Incidence, Pathophysiology, and Etiology of Postoperative Nausea and Vomiting (PONV) and Postoperative Vomiting (POV) Incidence Pathophysiology Etiology Anxiety Inhalation Anesthetic Agents Other Factors Antiemetics for PONV and POV Ondansetron Dolasetron Granisetron Dexamethasone Droperidol Non-Pharmacologic Antiemetic Approaches Isopropyl Alcohol P6 Acupuncture and Acupressure Postoperative Pain, Antiemetic Use, and Patient-Controlled Analgesia Specific Emetogenic Surgical Procedures in Children Strabismus Surgery Incidence Muscles Repaired and the Oculocardiac Reflex Anesthetic Techniques Antiemetics Tonsillectomy Incidence Anesthetic Techniques Antiemetics Additional Pediatric Surgeries and Procedures Tympanoplasty Ear Surgery Radiofrequency Catheter Ablation Burn Surgery Craniofacial Operations Neurosurgery Magnetic Resonance Imaging Development of PONV and POV Management Guidelines and Algorithms Guidelines for POV Prophylaxis in Children Conclusions...65

2 48 Kovac Abstract Postoperative nausea and vomiting (PONV) continues to be a frequent and important cause of morbidity in children. Postoperative vomiting (POV) is more commonly studied in children than postoperative nausea because of a child s inability to effectively express distress after experiencing nausea. POV is problematic in children and is one of the leading postoperative complaints from parents and the leading cause of readmission to the hospital. POV occurs twice as frequently in children as in adults, increasing until puberty and then decreasing to adult incidence rates. Gender differences are not seen before puberty. POV remains a main cause of morbidity in children because severe vomiting can be associated with dehydration, postoperative bleeding, pulmonary aspiration, and wound dehiscence. While children have an increased potential for dehydration and the resulting physiologic impairments, other associated results such as a delay in hospital discharge or an overnight or longer hospital admission also must be considered. The two most common emetogenic surgical procedures evaluated in children are strabismus repair and adenotonsillectomy. The approach to the management of PONV and POV in children is similar to that in adults. However, as the rate of POV is more frequent in children than in adults, more children are candidates for antiemetic prophylaxis. The management approach is multifactorial and involves proper preoperative preparation, risk stratification, rational selection of antiemetic prophylaxis, choice of anesthesia technique, and a plan for postoperative antiemetic therapy. It is important to identify children at moderate-to-high risk for POV as prophylactic antiemetic therapy is useful in these children. Antiemetics of choice for POV in children include dexamethasone, dimenhydrinate, perphenazine, ondansetron, dolasetron, granisetron, and tropisetron. The serotonin (5-hydroxytryptamine; 5-HT3) antagonists are the antiemetic drugs of first choice for POV prophylaxis in children because as a group they have greater efficacy for preventing vomiting than nausea. The 5-HT3 antagonists can be effectively combined with dexamethasone with an increase in efficacy. If possible, regional anesthesia should be considered. For those undergoing general anesthesia, the baseline POV risk should be reduced. Children at moderate-to-high PONV risk should receive combination therapy with two or three prophylactic antiemetics from different antiemetic drug classes. Reference to and the use of PONV guidelines and management algorithms help improve cost-effective postoperative care. It is estimated that following anesthesia and surgery, children have more complications in the postoperative anesthesia care unit (PACU) recovery area than adults. [1] The majority of these events are age related, occurring mostly in neonates and infants, and most involve the respiratory rather than the cardiovascular system. As children become older, postoperative effects include nausea, retching, and vomiting. Nausea is an unpleasant, subjective sensation that may or may not be associated with vomiting. Retching is the synchronous, rhythmic contraction of the abdominal, diaphrag- matic, and intercostal muscles that occurs with a closed mouth and glottis. Vomiting is the forceful expulsion of gastric contents from the mouth. Postoperative nausea and vomiting (PONV) and postoperative vomiting (POV) continue to be important causes of morbidity in adults and children, respectively. [2,3] They are among the main postoperative complaints from parents, and a leading cause of delayed discharge and/or re-admission to the hospital for the pediatric patient. A review [4] of children undergoing day surgery found that PONV was the fourth most common reason for unplanned hospital admission following pain, surgical complications, and surgery late in the day. More PONV clinical trials have been conducted in adults than in children. The limitations involved in analyzing and comparing older antiemetic studies has resulted in difficulty in applying the results of randomized clinical trials to actual clinical situations. The systematic review is an important method that has been used to understand the efficacy of an intervention in actual clinical situations and the likelihood of harm or adverse events. It is especially useful when there is a large amount of data from numerous clinical trials but still unresolved questions. Concepts such as the number needed to treat (NNT; the inverse of the absolute risk reduction) and the number needed to harm (NNH; the inverse of the absolute risk increase) have been summarized by Tramér [5-7] for PONV and are useful concepts that can be used to compare antiemetic drug efficacy and adverse events, respectively. The improvement resulting from treatment compared with placebo has been used as a measurement of antiemetic efficacy. Tramér [5] used the example that a 20% improvement in treatment efficacy above the placebo response indicated that 20% of patients who received the antiemetic medication would benefit (absolute risk reduction) from the treatment. If a perfect response is defined

3 Management of Postoperative Nausea and Vomiting in Children 49 Various investigators [8-19] have evaluated the reasons for POV and PONV in the pediatric population. Kotiniemi et al. [8] evaluated PONV symptoms in children occurring at home following day- case surgery. PONV occurred in 13% of all children evaluated, and emetic symptoms were most common following tonsillectomy, occurring in 31% of patients. Specific predictors that PONV would occur at home were: (i) emetic symptoms in the hospital; (ii) age >5 years; (iii) pain at home; and (iv) the use of postoperative opioids. However, these authors noted that the intraoperative use of opioids did not affect the incidence of PONV. as a 100% improvement, then a 20% response yields an NNT of five (100% divided by 20%), i.e. five patients would need to receive the medication for it to have a positive effect in one patient. Specifically, for an antiemetic medication, an NNT of five indicates that five patients at risk for PONV would need to receive the medication in order for one patient not to vomit that would have vomited had he or she not received the medication. From Tramér s [7] evaluation, in children, intravenous droperidol 75 µg/ kg has an NNT of 5 and 4 5 for early (0 6 hours) and late (0 24 hours) vomiting, respectively. Intravenous ondansetron 100 µg/kg has an NNT of 4 5 and 2 3 for early (0 6 hours) and late (0 24 hours) vomiting, respectively. Intravenous ondansetron 150 µg/kg has an NNT of 2 3 for early (0 6 hours) vomiting. Regarding NNH in children, droperidol has an NNH of 91 for extrapyramidal symptoms. The oculocardiac reflex in children receiving propofol has an NNH of 4. This review summarizes the latest data and information regarding the management of PONV in the pediatric patient specifically with respect to strabismus surgery and adenotonsillectomy. A summary of anesthetic techniques and application of the recent PONV consensus guidelines for pediatric patients is presented. Study references cited in this article were obtained by an Internet search of Google, OVID, and Netscape databases using keywords such as postoperative nausea and vomiting, PONV, postoperative vomiting, POV, pediatrics, antiemetics, 5-HT 3 antagonists, ondansetron, granisetron, dolasetron, dexamethasone, droperidol, meta-analysis, systematic review, strabismus repair, tonsillectomy, and adenotonsillectomy. 1. Incidence, Pathophysiology, and Etiology of Postoperative Nausea and Vomiting (PONV) and Postoperative Vomiting (POV) 1.1 Incidence Despite the introduction of new antiemetic medications, the incidence of POV in children is estimated to be twice the incidence for both nausea and vomiting after surgery in adults. It is difficult to estimate the true incidence of nausea in children who may not be able to express their degree of discomfort associated with this subjective feeling. This is the major reason why antiemetic studies in children have evaluated POV rather than PONV. However, failure to report nausea in these studies does not mean that nausea is not experienced in children. If the true incidence of nausea could be measured accurately, the incidence of PONV (see table I) would be even higher in the pediatric age group. [1,8-10] Investigators [8-14,16,17] have reported the overall incidence of POV in children to be between 8.9% and 42% (table I). Surgery- specific POV in children ranges from 9% to 80%. Interestingly, Table I. Incidence of postoperative nausea and vomiting (PONV) and postoperative vomiting (POV) in children Study Year PONV or Incidence (%) POV Overall Rowley and Brown [10] 1982 POV 42 Patel and Hannallah [14] 1988 POV 8.9 D Errico et al. [9] 1989 PONV 19 Schofield and White [16] 1989 POV 14 Karlsson et al. [17] 1990 POV 25 Byers et al. [12] 1995 PONV 18.1 Kotiniemi et al. [8] 1997 PONV 13 Villeret et al. [13] 2002 PONV 9.4 Khalil et al. [11] 2005 PONV 28 Surgery/procedure specific Strabismus Abramowitz et al. [21] 1983 POV 80 Kuhn et al. [22] 1999 PONV 37 Tonsillectomy Ferrari and Donlon [19] 1992 POV 70 Stewart et al. [20] 2002 POV 15.6 Plastic surgery: ears Ridings et al. [18] 1994 POV 63 Radiofrequency catheter ablation Erb et al. [23] 2002 PONV 60 Burn reconstruction surgery McCall et al. [24] 1999 PONV 69 Stubbs et al. [25] 1999 PONV 45 (non-scalp); 100 (scalp) Craniotomy Furst et al. [26] 1996 POV 66 Magnetic resonance imaging Murray et al. [15] 1995 POV 9

4 50 Kovac is coordinated by the vomiting center. Stimulation can be initiated from peripheral areas such as the oropharynx, mediastinum, gas- trointestinal tract, renal pelvis, peritoneum, or genitalia, and from central areas such as the cerebral cortex, and labyrinthine, otic, or vestibular apparatus. [27-29] It is hypothesized that PONV after strabismus surgery may be due to an altered visual perception and afferent impulses causing the oculoemetic reflex, which is analogous to the oculocardiac reflex. An increase in the number of ocular muscles that are repaired is reported to increase the risk of POV. Afferent stimuli are relayed from peripheral to central vomiting centers and the area postrema via the glossopharyngeal and vagal nerves, which may help explain the cause of PONV following adenotonsillectomy and hernia repair. [27,28,30-33] the study by Stewart et al. [20] in pediatric patients undergoing tonsillectomy revealed a POV incidence of 15.6% despite the fact that these children received intraoperative antiemetics. A study by D Errico et al. [9] evaluating the incidence and reasons for prolonged PACU stay and unplanned hospital admission determined that the most common cause of prolonged length of PACU stay was due to PONV (19% of children), followed by respiratory complications (16%). Unplanned hospital admissions following outpatient surgery were primarily due to respiratory and surgical reasons (32% and 30%, respectively), and these outcomes had a significant impact on hospital staffing, institutional costs, family convenience, and patient satisfaction; PONV accounted for 8%. A lower incidence of POV of 22 40% was observed by Rowley and Brown [10] in children aged <3 years compared with 42 51% in children >3 years. This was supported by Khalil et al. [11] who reported a POV incidence of 27% and 28% in children aged 1 12 months and months, respectively. Byers et al. [12] found that the highest incidence occurred in children undergoing ear, nose, and throat (ENT) procedures and increased with age. Avoidance of intraoperative opioids and the use of local anesthesia and/or NSAIDs for pain control were found to reduce the incidence of PONV. Villeret et al. [13] evaluated the incidence of PONV during the first 24 hours following elective ambulatory pediatric surgery, specifically excluding head and neck procedures. PONV occurred most frequently in the hospital during the first 3 hours after anesthesia but rarely during the journey home and was associated with: (i) increasing age; (ii) previous history of PONV; (iii) tracheal intubation; (iv) use of the laryngeal mask airway; (v) controlled or manual ventilation; and (vi) opioids. The type of surgery, premedication, type of anesthesia induction, regional anesthesia, use of nitrous oxide, anesthesia duration, length of PACU stay, and duration of the journey home after discharge were not found to be significantly associated with PONV. Cerebellum Area postrema and chemoreceptor trigger zone Nucleus of the solitary tract Fourth ventricle Vomiting centre Fig. 1. Anatomic location of the brain postoperative nausea and vomiting receptor area: vomiting center, nucleus of the solitary tract, area postrema, and chemoreceptor trigger zone (reproduced from Kovac, [29] with permission). Patients who have a history of motion sickness have a higher incidence of PONV, as stimulation of the vestibular apparatus of the inner ear due to movement of endolymph in the semicircular 1.2 Pathophysiology canals stimulates otolith cells in the utricle. Transmission of impulses to the chemoreceptor trigger zone and vomiting center Mechanisms of PONV and POV in children are similar to those occurs, causing the sensation of motion sickness with the occurin adults. However, they appear to be more procedure specific in rence of nausea and vomiting. Motion sickness or vertigo as a children as a result of swallowing of blood in adenotonsillectomy result of vestibular stimulation also can be a consequence of patients, stimulation of extraocular muscles in strabismus surgery, middle ear surgery. [32,33] and labyrinthine, otic, and vestibular stimulation in ear surgery. Sensory stimuli causing PONV include tactile stimulation of The vomiting areas in the CNS include the emetic center, the posterior pharynx (from oral or nasal airway devices, and nucleus of the solitary tract, area postrema, and chemoreceptor nasogastric or endotracheal tubes), operations on extraocular mustrigger zone. The chemoreceptor trigger zone is located in the area cles (stimulating the oculocardiac reflex), as well as stretching and postrema near the emetic center at the bottom of the fourth inflammation or injury to the airway, upper abdomen, gastrointesventricle (figure 1). The process of nausea, retching, and vomiting tinal tract, renal pelvis, bladder, or testes. [32-36]

5 Management of Postoperative Nausea and Vomiting in Children 51 The close proximity of areas associated with balance, vasomotor activity, salivation, respiration, and bulbar control to the vomiting center corresponds to the physiologic reactions often seen with POV and PONV, such as salivation, increased swallowing, sweating, pallor, tachypnea, tachycardia, cardiac dysrhythmias, and motion sickness. [36,37] Metabolic, biochemical, and environmental factors that are mediated by the vomiting center and the chemoreceptor trigger zone include uremia, diabetes mellitus (hypo- or hyperglycemia), electrolyte disturbances (sodium, potassium), hormonal imbalances (estrogen, progesterone), chemotherapy, and radiation therapy. [33,36,37] 1.3 Etiology Similar to adults, surgical-, patient-, and anesthesia-related factors play an important part in the etiology of PONV and POV in children. Lerman [37] noted that while a greater incidence of PONV is reported in children compared with adults, this must be inter- preted with caution; postoperative follow-up data were not collect- ed prospectively for 24 hours in all patients. However, as previously determined in adults by Apfel et al., [43] Eberhart et al. [44] prospectively determined risk factors for POV in pediatric surgery. Table II. Mid-brain neurochemical emetogenic receptor locations (reproduced from Kovac, [29] with permission) Mid-brain location Area postrema Chemoreceptor trigger zone Nucleus of solitary tract a Receptors a Opioid, dopamine, serotonin (5-hydroxytryptamine), neurokinin-1 Enkephalin, opioid, dopamine Enkephalin, histamine, neurokinin-1, muscarinic, cholinergic The vomiting center is the coordinator for these receptors to initiate the vomiting reflex. Table III. Simplified risk score for postoperative vomiting (POV) in children [44] No. of risk factors a POV risk (%) a Risk factors include: strabismus surgery; age 3y; surgery >30 min; history of POV or postoperative nausea and vomiting in relatives (mother, father, siblings). The chemoreceptor trigger zone contains high concentrations of enkephalin, opioids, and dopamine (D2) receptors. The area postrema has high concentrations of opioids, D 2, serotonin (5-hydroxytryptamine; 5-HT), and neurokinin-1 (NK-1) receptors (table II). The nucleus of the solitary tract has a predominance of enkephalin, histamine, muscarinic, cholinergic, and NK-1 receptors. These emetic neuroreceptor areas serve as sensors and are stimulated by drugs, electrolytes, and metabolic chemicals, causing impulses to be relayed to the vomiting center, thereby initiating the vomiting reflex. The mechanism of action of the antiemetic medications commonly used for PONV involves blockade of these multiple neurochemical receptor sites; this helps explain why a combination or multimodal antiemetic approach may be necessary in some high-risk patients. [20,28,32,37-42] These include: (i) strabismus surgery; (ii) duration of anesthesia >30 minutes; (iii) history of POV or previous history of POV, PONV, or motion sickness in relatives; (iv) age 3 years; and (v) use of postoperative opioids. When 0, 1, 2, 3, or 4 of these risk factors were present, the POV risk was 10%, 10%, 30%, 50%, or 70%, respectively (table III). A variety of authors have also evaluated the effects of other patient- and anesthetic-related factors. [45-59] Anxiety The relationship between anxiety and PONV appears to be minimal. Wang and Kain [46] determined that preoperatively controlling for anxiety had no predictive value for the occurrence of PONV in children in the PACU or during the first postoperative day Inhalation Anesthetic Agents Numerous studies have evaluated the effect of anesthetic technique and inhalation agents on PONV. [45-56,59] In an attempt to quantify the relative importance of operative anesthetic and patient-specific factors for the development of PONV, Apfel et al. [43] conducted a randomized controlled trial of 1180 children and adult patients at high risk for PONV. They concluded that inhalation anesthetic agents caused early but not delayed PONV, and that this effect was more significant than the effects of other risk factors. Their conclusion was that in adult and pediatric patients at high risk for PONV, it makes better sense to avoid inhalation anesthesia rather than simply adding an antiemetic, which may still be needed to prevent or treat delayed PONV. A similar conclusion was reached by Elliott et al. [48] who compared inhalation versus intravenous anesthesia techniques and determined that there was a higher incidence of pre-discharge PONV and resulting higher costs if sevoflurane was used for anesthesia induction and maintenance compared with using propofol. Goa et al. [49] reviewed the use of sevoflurane in pediatric anesthesia. While sevoflurane provided a more rapid induction and anesthesia emergence than halothane, postoperative pain and

6 52 Kovac PONV were more frequent with sevoflurane. This review brings holding oral fluids was seen in patients who received opioids, in attention to the possible correlation between pain and PONV or the whom POV was reduced from 73% to 36%. [57] Similarly, another treatment of pain with opioids and PONV. study [58] determined that not requiring pediatric patients aged 1 Use of regional anesthesia has been suggested to decrease month to 18 years to consume clear fluids postoperatively de- PONV. Oddby et al. [50] evaluated the use of sevoflurane alone creased the incidence of POV and time to discharge from the versus spinal anesthesia combined with propofol for sedation in PACU. pediatric ambulatory surgery. While a reduced number of POV Murat et al. [61] determined an increased correlation with POV in episodes and better immediate postoperative analgesia were found the PACU in children aged 8 years who were intubated for ENT with spinal anesthesia than with propofol sedation, there was no surgery. difference between the two regimens regarding time to discharge or overall patient satisfaction. 2. Antiemetics for PONV and POV Two studies [10,44] determined that length of surgery >30 minutes has a positive increased correlation with POV in children. The effect of nitrous oxide on PONV in adults and children has 2.1 Ondansetron been controversial, and most clinical data have come from adult As ondansetron is relatively free of adverse events, numerous studies. The results of three systematic reviews [51-53] concluded researchers [62-65] have concluded that ondansetron is a safe firstthat omitting nitrous oxide from general anesthesia decreases line antiemetic for children. Ondansetron is the only 5-HT3 antag- PONV (NNT = 5). In other words, five high-risk patients would onist with US FDA approval for use in children as young as 1 need to undergo a nitrous oxide-free anesthetic for one to not month. [11] vomit who would have done so if they had received nitrous oxide. Ondansetron has been determined to have good antiemetic Nevertheless, the results in children are controversial as two other efficacy for the prevention of POV in children, particularly when studies reached different conclusions. Splinter and Komocar [54] combined with dexamethasone. In several large, dose-ranging, and studied the effects of nitrous oxide on POV in children who placebo-controlled trials, [63-67] intravenous ondansetron underwent outpatient dental restorations under halothane anesthe- mg/kg or oral ondansetron 0.1 mg/kg was significantly more sia. Even though the POV rate in the PACU was slightly less for effective than placebo for the prevention of emesis in children the no nitrous oxide versus the nitrous oxide group (15% vs 24%), undergoing highly emetogenic surgery, which included tonsillecthey concluded that nitrous oxide did not significantly affect POV. tomy or strabismus repair. Intravenous ondansetron 0.05 mg/kg Similarly, another study [55] concluded that nitrous oxide in combi- was determined to be the lowest effective dose. [66] Prophylactic nation with sevoflurane was not associated with an increase in ondansetron 0.1 mg/kg (up to a total dose of 4mg) reduced POV in POV; the incidence of POV for the nitrous oxide and no nitrous pediatric patients regardless of surgical or anesthesia factors. [67] oxide groups was similar (14.3% and 15.5%, respectively). The ondansetron-treated children reached the criteria for home Hannallah et al. [56] evaluated speed and quality of recovery, readiness 30 minutes earlier than the placebo-treated patients. comparing propofol with thiopentone or halothane for induction Rapid intravenous administration of ondansetron 0.15 mg/kg or and maintenance of anesthesia. Children who received propofol metoclopramide 0.25 mg/kg was not associated with changes in had a faster recovery, were discharged home earlier, and had a vital signs or oxygen saturation. [68] lower POV incidence. However, it should be stressed that the risk In children undergoing strabismus surgery, a POV prophylactic of propofol-related bradycardia is particularly high in children intravenous study [69] concluded that ondansetron 0.1 mg/kg and undergoing strabismus surgery, due to stimulation of the oculo- droperidol mg/kg had similar antiemetic efficacy and were cardiac reflex. The NNH of propofol causing the oculocardic significantly more effective compared with placebo or reflex in children has been estimated to be four despite the prophy- metoclopramide 0.25 mg/kg. A meta-analysis [70] of 54 studies lactic use of anticholinergics. [60] compared the efficacy and safety of ondansetron, droperidol, or metoclopramide for preventing PONV in adults and POV in Other Factors children. While ondansetron was determined to be more effective Withholding oral fluids postoperatively from children undergoeffective than droperidol, both antiemetics alone were found to be more ing day surgery has significantly reduced the incidence of POV. than metoclopramide in preventing POV in children. This difference was seen whether or not a patient was at high risk Antiemetic studies in children have been limited in study for POV (including operations for strabismus, adenoidectomy, design and have had low power, as these studies have been and/or tonsillectomy). Interestingly, the greatest effect of with- powered to detect differences between an antiemetic drug and

7 Management of Postoperative Nausea and Vomiting in Children 53 placebo and not between anesthetic regimens. Prophylactic PONV evaluated the effectiveness of granisetron for the prevention of antiemetic studies are easier to conduct than treatment studies. POV following pediatric surgery with most clinical studies con- Consequently, as in adults, there have been fewer POV treatment ducted in Japan by Fujii et al. [76-79] studies in children. A large treatment study [71] evaluated the use of An oral dose of granisetron 40 µg/kg was determined to be the intravenous ondansetron in established POV in 2720 pediatric lowest effective dose for the prevention of POV following inguioutpatients undergoing general anesthesia with nitrous oxide. A nal hernia and phimosis-circumcision surgery. [76] A dose-ranging single intravenous dose of ondansetron 0.1 mg/kg (up to a maxi- study [77] determined that intravenous granisetron 40 µg/kg was the mum dose of 4mg) was concluded to be effective and well tolerat- minimally effective dose for POV prevention. After inhalation ed (as rescue medication) for prevention of further episodes of anesthesia induction, intravenous granisetron 40 µg/kg was found POV and resulted in a shorter time to PACU discharge. to be effective for preventing POV in children with a history of Ummenhofer et al. [72] conducted a double-blind, prospective motion sickness. [78] Intravenous granisetron 40 µg/kg was also the prophylactic study on the effect of intravenous ondansetron lowest effective dose for preventing POV and retching following 0.1 mg/kg or placebo administered before surgical incision, and strabismus repair and tonsillectomy surgery. [79] This conclusion the effect of rescue antiemetics in patients in whom prophylaxis was also reached by Cieslak et al. [80] who studied pediatric outpafailed. For rescue medication, patients received either intravenous tients and determined that intravenous granisetron 40 µg/kg was ondansetron 0.1 mg/kg or droperidol 0.02 mg/kg. As ondansetron more effective than intravenous granisetron 10 µg/kg or placebo, was found to be effective for the prevention of PONV for the first but that this dose had a higher acquisition cost. 4 hours after general anesthesia with lower sedation scores com- Kranke et al. [81] evaluated the influence of a dominating center pared with droperidol, this was judged to be advantageous, espe- in a quantitative systematic review of granisetron for preventing cially in ambulatory surgery. Interestingly, the incidence of late- PONV. A total of 27 randomized clinical trials were assessed; onset PONV occurring >4 hours postoperatively was not found to 2938 patients including children and adolescents were included in be influenced by the preoperative prophylactic administration of a the analysis. In the dominating center, low-dose granisetron was one-time dose of ondansetron. determined to be ineffective, while high-dose granisetron was found to be effective. In contrast, the other centers showed both 2.2 Dolasetron low- and high-dose granisetron to be effective. These researchers [81] concluded that the overall results and dose-response charac- Dolasetron has been recommended for POV prophylaxis in teristics of meta-analyses may be significantly altered by one children aged 2 years and older. An intravenous equivalence dosedominating center. A cautious statistical analysis was previously ranging study [73] determined the lowest effective dose of dolaseconducted by Kranke et al. [82] on the distribution of side effects of tron (45, 75, 350, or 700 µg/kg) that was equivalent to the US comparative groups reported by the dominating center and sug- FDA-approved ondansetron intravenous dose of 100 µg/kg. Intragested that the reported data are idealized. venous dolasetron 350 µg/kg was determined to be the lowest With these data in mind, the safety and efficacy of granisetron effective dose providing acceptable equivalent efficacy and patient has not been established in children for the prevention and treatsatisfaction scores to those with the intravenous ondansetron ment of PONV and does not have US FDA approval for PONV in 100 µg/kg dose. children. The 2003 Consensus Guidelines [83] contained no granise- In a strabismus study, Wagner et al. [74] determined that the tron dosing recommendations for PONV in children. efficacy of intravenous dolasetron 350 µg/kg in preventing PONV in children was equivalent to that of an intravenous dolasetron 12.5mg fixed dose. In a prophylactic intravenous study [75] of 2.4 Dexamethasone pediatric patients undergoing tonsillectomy who also received intravenous dexamethasone 1 mg/kg (up to 25mg), dolasetron 500 A quantitative, systematic review [84] on the use of dexametha- µg/kg (up to 25mg) plus intravenous dexamethasone 1 mg/kg (up sone for the prevention of PONV evaluated data from 1946 adult to 25mg) was found to be equivalent to a prophylactic ondansetron and pediatric patients studied in 17 randomized controlled trials in dose of 150 µg/kg (up to 4mg). which 16 different dexamethasone dose regimens were used. An 8 or 10mg intravenous dose in adults and a 1 or 1.5 mg/kg intravenous 2.3 Granisetron dose in children were the most frequently used dexamethasone doses. Using these doses, the NNT to prevent early Granisetron has also been recommended for POV prophylaxis (0 6 hours) and late (0 24 hours) POV was seven and four, in children aged 2 years and older. Numerous studies [76-80] have respectively. Late efficacy with dexamethasone was a more pro-

8 54 Kovac nounced effect in children than in adults. A single prophylactic within 30 minutes of the bolus dose. No associated arrhythmias dose of dexamethasone was a more effective antiemetic compared were observed. These researchers followed the ECG for a miniwith placebo without any clinically relevant evidence of toxicity mum of 1 hour and were able to evaluate the length of time (15 30 or adverse effects in otherwise healthy patients. While the best minutes) in which there was a dose-dependent prolongation of the PONV prophylaxis was achieved with the combination of dex- QT interval. The authors noted that a large prospective study is amethasone and a 5-HT3 receptor antagonist, the authors noted needed to identify the true risk of arrhythmias in the pediatric that optimal doses of this combination requires further investiga- population. tion. However, other studies [63,64] determined that intravenous The US FDA black box recommendation indicated that all ondansetron 50 µg/kg when combined with intravenous dex- surgical patients should undergo 12-lead ECG monitoring prior to amethasone 150 µg/kg was significantly more effective at reduc- the administration of droperidol to determine if a prolonged QTc ing POV than either medication used alone. interval was present and to continue ECG monitoring for 3 hours after droperidol administration. [87] Because of these recommendations, 2.5 Droperidol this situation places the practising anesthesiologist in a dilemma, as there can be a significant difference between standard Henzi et al. [85] systematically reviewed the efficacy, dosedroperidol. clinical practice and the package insert recommendation for response, and adverse effects of droperidol for the prevention of PONV in 76 randomized controlled trials involving 5351 patients An editorial by Berry [89] noted that the Stuth et al. [88] study cast receiving 24 different antiemetic regimens. The average incidence doubt about the recommendations in the black box warning for of early and late PONV in the control groups was 34% and 51%, ECG monitoring. Reasonable practice suggests that proper evalua- respectively. Droperidol was determined to be more efficacious tion of patients for potential problems should allow the use of a than placebo in preventing PONV in children with an NNT of four drug such as droperidol in an appropriate manner while monitor- and five to prevent early and late vomiting, respectively. Two ing for expected potential complications. children were noted to have had extrapyramidal symptoms, and Further research must be completed to resolve the role of low- the NNH in children was determined to be 91. The effect of dose droperidol as an antiemetic. The US FDA is exploring droperidol on nausea was short-lived but was more pronounced options to obtain data that satisfy regulatory standards for the than its effect on vomiting, with sedation and drowsiness being demonstration of safety and efficacy at doses lower than 2mg. dose dependent, and a small risk for extrapyramidal symptoms Chang and Rappaport [90] urged practitioners to participate in the being present. postmarketing safety assessment process by reporting all potential In 1994 it was reported that droperidol caused a dose-dependent drug-related adverse events. The website prolongation of the QT interval. [86] While previously there was medwatch [91] contains information on reporting adverse events. warning of potential sudden cardiac death regarding the use of droperidol when administered at high doses (>25mg) to psychiatric 3. Non-Pharmacologic Antiemetic Approaches patients, in December 2001 a black box warning [87] of cardiac effects regarding the use of droperidol for PONV, issued by the US 3.1 Isopropyl Alcohol FDA, was included in the package insert. The revised warning Isopropyl alcohol is a novel alternative method to alleviate cautioned that even low droperidol doses such as 0.625mg for PONV in children who are scheduled to undergo elective outpa- PONV should be used only when other first-line antiemetic meditient surgery under general anesthesia. One study [92] randomized cations are not effective. Data regarding the NNH of cardiac children to inhale an isopropyl alcohol wipe versus saline, repeateffects with droperidol were not available because of the low ing this for up to three times. After three sequences, 65% in the number of adverse events previously reported. The majority of the isopropyl group versus 26% in the saline group had a significant reports of cardiovascular events with droperidol were in adults; reduction of either nausea or vomiting. However, this reduction however, cardiovascular effects have also been observed in chilwas transient in children with established PONV; recurrent nausea dren. or vomiting occurred within minutes. Stuth et al. [88] conducted a retrospective analysis of 20 children of whom 18 had undergone cardiopulmonary bypass. An intravenous droperidol 100 µg/kg bolus was given for perioperative 3.2 P6 Acupuncture and Acupressure sedation. Droperidol caused a significant but transient increase in A meta-analysis [93] was conducted of 19 randomized trials on the QTc interval; it was still present at 15 minutes but had resolved the efficacy of preventing PONV with acupuncture, acupressure,

9 Management of Postoperative Nausea and Vomiting in Children 55 acupoint stimulation, electro-acupuncture, and transcutaneous electrical nerve stimulation in children and adults. The primary outcomes for the incidence of nausea, vomiting, or both were evaluated at 0 6 hours (early efficacy) or 0 48 hours (late efficacy) after surgery. While the results of these techniques in adults were found to be statistically significant, no benefit was found in children. In contrast, Wang and Kain [94] evaluated P6 acupoint injections versus droperidol for control of early PONV in children and concluded that the P6 acupoint injections were as effective as droperidol in controlling early PONV. Similarly, another study [95] concluded that laser P6 stimulation, when administered 15 minutes before anesthesia induction for strabismus surgery and 15 minutes after arriving in the PACU, resulted in a significantly lower incidence of POV. In addition, another study [96] evaluated the effect of P6 electroacupuncture prophylaxis following pediatric tonsillectomy with or without adenoidectomy and concluded that perioperative P6 stimulation in awake children significantly reduced nausea, but there was no reduction in emetic episodes or the need for rescue antiemetics. 5. Specific Emetogenic Surgical Procedures in Children 5.1 Strabismus Surgery Incidence Over the last 20 years the incidence of POV in children having strabismus surgery has ranged from 37% to 80%. [21,22] The effects of various anesthetic techniques and antiemetics (see also section 5.1.3) on POV in children undergoing strabismus surgery are summarized in tables IV and V, respectively. [69, ] Muscles Repaired and the Oculocardiac Reflex In an evaluation of children who received no prophylactic antiemetic medication for strabismus surgery, the overall incidence of nausea and vomiting was determined as 37% and 32%, respectively. [115] Splinter et al. [115] determined that while the incidence of POV was not affected by the use of intravenous midazolam, droperidol 50 µg/kg, or duration of anesthesia, the number of repaired eye muscles was a significant predictor of POV, with an incidence of POV 2.5-fold higher with surgery performed on both eyes compared with one eye. While the efficacy of P6 acupuncture for PONV prevention is The relationship between the oculocardiac reflex and PONV believed to be similar to that of commonly used pharmacotheranous atropine 0.02 mg/kg, alfentanil, and no nitrous oxide. The was studied in children receiving a prophylactic dose of intrave- pies, its appropriate role in the prevention and treatment of PONV investigators concluded that while a thiopental-isoflurane techin children requires further study. nique with alfentanil resulted in a moderate risk for POV, adding intravenous ondansetron 4mg significantly decreased this risk. 4. Postoperative Pain, Antiemetic Use, and The NNT in the early postoperative period was six (six children Patient-Controlled Analgesia needed to be treated for one to benefit). Propofol and the combination of intravenous propofol and lidocaine (lignocaine) 2 mg/kg Similar to adults, nausea and vomiting in children related to demonstrated no benefit in decreasing POV but increased the risk opioids is difficult to treat, and the effectiveness of antiemetics for of the oculocardiac reflex despite a high dose of prophylactic opioid-induced nausea and vomiting is controversial. intravenous atropine 0.02 mg/kg. [117] Children who received intravenous tropisetron 0.1 mg/kg (up to Anesthetic Techniques a maximum of 5mg) had a lower incidence and severity of vomit- Diazepam ing during patient-controlled opioid analgesia, with only one child The combination of diazepam and atropine mg/kg has vomiting more than twice, compared with nine children in the been shown to decrease POV following strabismus surgery. [101] control group. [97] Prophylactic intravenous dixyrazine was found The overall incidence of POV and the need for rescue antiemetics to significantly reduce the incidence and severity of PONV in was significantly higher for the first 24 hours after using a children who used patient-controlled analgesia with morphine sevoflurane-nitrous oxide technique compared with a propofolnitrous after major surgery. [98] In contrast, another study [99] evaluated the oxide technique. However, there was a significantly higher effectiveness of adding antiemetics to the morphine solution in incidence of bradycardia from the oculocardiac reflex, using a patient-controlled analgesia syringes used by children after appendectomy, propofol-nitrous oxide technique. [102] and determined that addition of prophylactic antiemetics Opioids such as ondansetron or droperidol did not reduce the incidence of Several studies [100,103,104] have evaluated anesthetic techniques PONV. Thus, this continues to be a controversial topic. combined with opioids for strabismus surgery. Rectal diclofenac

10 56 Kovac Table IV. Effects of anesthetic techniques on postoperative vomiting (POV) in children Study Surgery No. of Age Anesthetic technique Conclusions (effect of technique on POV) type pts (y) Wennstrom and S Rectal diclofenac vs morphine Diclofenac < morphine Reinsfelt [100] Ozcan et al. [101] S Diazepam + atropine premedication vs Diazepam + atropine POV vs placebo placebo Gurkan et al. [102] S Propofol-nitrous oxide vs sevoflurane- Propofol-nitrous oxide < sevoflurane-nitrous oxide nitrous oxide Standl et al. [103] S Propofol-sufentanil vs propofol-isoflurane Propofol-sufentanil < propofol-isoflurane Eltzschig et al. [104] S Fentanyl vs remifentanil Fentanyl = remifentanil (no change in POV) Pandit et al. [105] T Nitrous oxide vs no nitrous oxide Nitrous oxide = no nitrous oxide Ved et al. [106] T Halothane + nitrous oxide vs propofol Halothane + nitrous oxide = 3-fold POV vs propofol Zestos et al. [107] T Subhypnotic propofol 0.2 mg/kg dose No effect on POV (subhypnotic dose) Chhibber et al. [108] T Atropine-neostigmine vs glycopyrrolate- Atropine-neostigmine < glycopyrrolate-neostigmine neostigmine for muscle relaxant reversal pts = patients; S = strabismus repair; T = tonsillectomy; < indicates significantly less effect; > indicates significantly greater effect; = indicates similar effect; indicates increase; indicates decrease. 1 mg/kg was associated with less POV than intravenous morphine ly reduced the incidence of POV following strabismus surgery 0.05 mg/kg in children aged 4 16 years. [100] A propofol-sufentanil compared with opioid use. anesthetic technique compared with propofol-isoflurane resulted in less POV requiring fewer antiemetic rescues during the early Dimenhydrinate postoperative phase in the PACU, irrespective of the use of nitrous Numerous studies have evaluated the effectiveness of oxide. [103] One study [104] determined that the number of children dimenhydrinate in the management of POV. [ ,121] Two stud- who experienced POV did not differ significantly between groups ies [109,110] comparing the efficacy of rectal dimenhydrinate 50mg irrespective of whether or not they received fentanyl. The effects administered 30 minutes before the start of strabismus surgery and of ketorolac and fentanyl on POV and analgesic requirements placebo reported the incidence of POV to be significantly lower were evaluated in children who received no antiemetic prophylaxer, with dimenhydrinate than placebo (15 30% vs 60 75%). Howev- is. [120] It was concluded by the study investigators that opioids such the dimenhydrinate-treated patients tended to be more sedated as fentanyl should be avoided as intravenous fentanyl 1 µg/kg had and required observation in the PACU for a longer period than the a greater incidence of POV compared with intravenous ketorolac placebo group. [109] Another study [111] determined that even though 0.9 mg/kg. children who received intravenous dimenhydrinate 0.5 mg/kg had This gives further proof of the emetogenic effects of opioids less POV compared with those who received placebo, the time to and how the use of NSAIDs such as ketorolac help decrease arousal and hospital discharge did not differ between groups. PONV. Kranke et al. [121] conducted a meta-analysis of dimenhydrinate and determined that it was an inexpensive, older antiemetic that Clonidine was effective clinically. However, these researchers also conclud- The effect of oral clonidine on POV following strabismus ed that the dose-response curve, estimation of adverse effects, surgery is controversial. While Handa and Fujii [112] concluded that optimal time of administration, and benefit of repetitive doses oral clonidine 4 mg/kg enhanced the antiemetic effect of propofol, remains unclear. a study by Gulhas et al. [113] reported that premedication with oral Ondansetron clonidine 4 mg/kg 1 hour prior to surgery did not reduce POV. Intravenous ondansetron 75 µg/kg has been determined to be Antiemetics the optimum, lowest effective dose and to be as effective as 150 µg/kg in preventing PONV and improving outcomes for Dixyrazine strabismus surgery. [119] Antiemetic efficacy was similar with ad- Karlsson et al. [114] concluded that avoidance of opioids and the ministration of intravenous ondansetron 100 µg/kg either before use of intravenous prophylactic dixyrazine 0.25 mg/kg significant- (at induction) or after surgical manipulation of extraocular mus-

11 Management of Postoperative Nausea and Vomiting in Children 57 cles (at the end of surgery). [122] Children given ondansetron had to be more effective than placebo in decreasing the incidence of less than half the risk of POV compared with those given placebo, post-discharge vomiting. [69] The efficacy and safety of intraoperawith no difference between groups in the incidence of side effects. tive intravenous droperidol followed by an oral dose of There was a significant decrease in POV with a corresponding dimenhydrinate at home did not differ from the use of intravenous increase in dose (0.04, 0.1, or 0.2 mg/kg) of ondansetron. [117] ondansetron administered in the operating room followed by oral While intravenous ondansetron 100 µg/kg and droperidol ondansetron at home. [123] 75 µg/kg have been determined to be more effective than intrave- Splinter et al. [124] evaluated the effect of intravenous ondansenous metoclopramide 250 µg/kg, as compared with placebo, in tron 0.15 mg/kg (up to a maximum of 8mg) versus intravenous decreasing the incidence of pre-hospital discharge vomiting in propofol mg/kg on POV after strabismus surgery in chilchildren undergoing strabismus surgery, no antiemetic was found dren. Inhalation halothane, nitrous oxide, oxygen, or propofol was Table V. Effects of antiemetics on postoperative vomiting (POV) in children undergoing strabismus surgery Study No. of Age (y) Antiemetic Conclusions pts Wennstrom and Diclofenac 1 mg/kg Diclofenac < morphine sulfate Reinsfelt [100] Morphine sulfate 0.05 mg/kg Welters et al. [109] Rectal dimenhydrinate 0.50mg Dimenhydrinate > placebo. sedation with dimenhydrinate Schlager et al. [110] Dimenhydrinate 50mg Dimenhydrinate < placebo Vener et al. [111] Dimenhydrinate 0.5 µg/kg Dimenhydrinate > placebo Handa and Fujii [112] Diazepam 0.45 mg/kg Clonidine > placebo Clonidine 4 µg/kg Gulhas et al. [113] Clonidine 4 µg/kg Clonidine = placebo Karlsson et al. [114] Dixyrazine Dixyrazine > placebo Splinter et al. [115] Midazolam 50 µg/kg Midazolam = droperidol Droperidol 50 µg/kg in the number of eye muscles POV Bowhay et al. [117] Ondansetron 0.4 mg/kg Ondansetron 0.4 > ondansetron 0.2 > ondansetron 0.1 > Ondansetron 0.1 mg/kg placebo Ondansetron 0.2 mg/kg Scuderi et al. [69] Predischarge vs postdischarge Predischarge: droperidol = ondansetron > metoclopramide = placebo Metoclopramide 250 µg/kg Postdischarge: droperidol = ondansetron = metoclopramide Ondansetron 100 µg/kg = placebo Droperidol 75 µg/kg Sadhasivam et al. [118] Ondansetron 25 µg/kg Ondansetron 75 = ondansetron 100 = ondansetron 150 > Ondansetron 50 µg/kg ondansetron 25 = ondansetron = placebo Ondansetron 75 µg/kg Ondansetron 100 µg/kg Ondansetron 150 µg/kg Shende et al. [119] Droperidol 15 µg/kg Droperidol + ondansetron > droperidol = ondansetron Ondansetron 0.1 mg/kg Droperidol + ondansetron pts = patients; < indicates significantly less efficacy; > indicates significantly greater efficacy; = indicates similar efficacy; indicates increase; indicates results in.