Electroacupuncture Prophylaxis of Postoperative Nausea and Vomiting following Pediatric Tonsillectomy with or without Adenoidectomy

The study protocol was approved by Children's Hospital of Wisconsin Research and Publications Committee/Human Rights Review Board (Milwaukee, WI). Patients with American Society of Anesthesiologists physical status I and II, older than 4 and younger than 18 yr of age, undergoing tonsillectomy with or without adenoidectomy were eligible. Exclusion criteria included presence of skin lesions near acupuncture sites, previous and severe PONV, or a chronic history of nausea and vomiting. Patients were enrolled after written informed consent was obtained from their parents. A randomized block design procedure was used to assign enrollees to one of three treatment groups: acupuncture at P6 with electrical stimulation of needles (P6 Acu), sham acupuncture with stimulation (sham group), and a nontreatment control group.

Anesthetic and surgical techniques were standardized. Solid food was held for at least 8 h before surgery, and clear liquids were permitted up until 2 h before surgery for all groups. All patients received oral midazolam (0.5 mg/kg to a maximum of 10 mg) 20 min before induction of anesthesia with halothane or sevoflurane in oxygen (30%) and nitrous oxide (70%) via  mask. After intravenous cannulation, 0.2 mg/kg mivacurium, 0.1 mg/kg morphine sulfate, and at least 20 ml/kg lactated Ringer's solution were administered. Anesthesia was maintained with oxygen, 70% nitrous oxide, and isoflurane via  an endotracheal tube. Ventilation was controlled at a peak inspiratory pressure of 25–30 cm H²O to maintain an end-tidal carbon dioxide tension of 35–40 mmHg, until spontaneous return of neuromuscular function. Awake tracheal extubation was performed after orogastric suction. Postoperative pain was treated with 0.05 mg/kg morphine, repeated as needed. Lactated Ringer's solution was infused at a maintenance rate until oral clear liquids were accepted without vomiting. Thereafter, oral analgesics were administered as needed every 3 h (acetaminophen with codeine, 1 mg/kg). Surgical techniques were standardized, and dexamethasone was not administered. Details of treatment in three groups follow. All active, sham, and control treatments were performed by the same acupuncturist (LR).

After removal of the second tonsil, standard, disposable acupuncture needles (Serin No. 5, 0.25 × 30 mm, Shizuoka, Japan) were inserted to a distance of approximately 10 mm at a right angle to the skin, to a depth of 1 cm at the P6 point (AKA Master of the Heart 6-MH6). P6 is on the antebrachial region, 2 cuns (2 Chinese inches, equivalent to the width of the interphalangeal joint of the thumb of the patient) proximal to the anterior crease of the wrist, on the ulnar side of the tendon flexi carpi radialis. The needle was inserted between this tendon and the tendon of the palmaris longus, if it was present. A second needle was inserted at a neutral point with no known antiemetic properties midway up the forearm, on the same meridian but on a different dermatome (T1) (fig. 1). This second needle acted as a skin surface electrode to allow electrical current through P6, the point most commonly used for treatment of PONV 13,14and chemotherapy-induced nausea and vomiting. 16,17Needles were bent to lay flat against the skin, and the top end of the needle was taped to the skin. The negative lead (black) of an ITO electroacupuncture unit (LTD IC-1107; ITO-Co., Braintree, MA) was attached to P6, and the positive (red) lead was attached to the neutral point. The arm was then covered with full-length, soft restraints so the needle positions could not be seen. After placement of needles, the patient was allowed to emerge from anesthesia. Stimulation of the needles at low frequency, 4 Hz, was continued for 20 min as soon as the patient was awake.

Two needles were inserted high on the arm, on the ventral surface in the same acupuncture meridian at acupuncture point P2 (which is located in the anterior brachial region, 2 cuns distal to the end of the anterior axillary fold between the two heads of the biceps brachii muscle) and a neutral point. Both points are outside the dermatome of P6, or C5. These points, on the same meridian but outside the dermatome, do not exert the same antinausea effect that P6 exibits. 18The arms were covered, and stimulation occurred in the same manner.

No needles were inserted. Insulated wires were attached to the insides of the arm covers, which were applied to the arms as in the other groups. The stimulator box was activated so its indicator light was on to maintain blinding.

The primary outcome measures were the incidence of nausea, vomiting, and need for rescue therapy in the first 24 postoperative hours. PONV was defined as the presence of nausea, vomiting, or need for rescue therapy. Outcomes were recorded on a standard form beginning at entry to the recovery room. Experienced recovery room nurses, who were blinded to the treatment group, assessed nausea and vomiting. Vomiting was defined as forceful expulsion of gastric contents from the mouth. Retching was defined as an active attempt to vomit without expulsion of gastric contents. Repeated episodes occurring within 2 min were considered the same episode. 2Nausea that persisted for 15 min was treated. Rescue treatment for PONV, using ondansetron (0.15 mg/kg up to 4 mg × 1), was given if two or more episodes of vomiting occurred or if persistent nausea was reported. If PONV persisted, half this dose was repeated in 2 h, after which droperidol (10 μg/kg) was used. Assessment questionnaires with stamped, addressed return envelopes were given to parents when the patients were discharged to record all episodes of vomiting and all complaints of nausea over 24 h. The information on these forms was gathered by phone contact on postoperative day 1 and when the questionnaire was returned. Patients were not sent home with antiemetics.

Sample size estimation was based on a previous probability of nausea and vomiting of 65% and a reduction to 30% with acupuncture, based on literature in pediatric tonsillectomy patients and on reduction of PONV in adult patients after acupuncture. Using an α of 5% and a power of 80%, the estimated sample size for binomial proportions was 37 patients per group. To maximize power, equal size groups of 40 patients per group were generated using a randomized block design selection procedure.

Summaries of demographic and outcome data were expressed as mean ± SD for continuous variables, or as number and percent. Differences between groups were assessed by one-way analysis of variance and by chi-square tests as appropriate for data type, with Tukey correction for multiple comparisons. Mantel-Haenszel odds ratios for outcomes were computed relative to the control group, with significance assessed using the likelihood-ratio test with and without adjustment for demographic characteristics. Numbers needed to treat (NNT) for each outcome were computed from the inverse of the absolute risk reduction for P6 Acu and sham compared with controls. Differences in time to first episode of vomiting were analyzed using Kaplan-Meier survival analysis and the log-rank test for equality of the survivor functions. The cutoff for statistical significance was set at P < 0.05 for all tests.

One hundred twenty-two patients were asked to participate, 121 were enrolled after 1 refused, and 1 was disqualified after enrollment when propofol was administered during the anesthetic, generating 120 patients eligible for analysis, with 40 patients in each group. The groups were similar for age, sex, weight, analgesics administered, and surgical time (table 1), with no differences found.

The incidence of nausea was reduced by P6 acupuncture (24 of 40 or 60%;P = 0.0007) but not by sham acupuncture (34 of 40 or 85%;P = not significant) compared with controls (37 of 40 or 93%). The incidence of vomiting was not significantly different with P6 acupuncture (25 of 40 or 63%) or sham acupuncture (35 of 40 or 88%) compared with controls (31 of 40 or 78%). The incidence of rescue therapy for nausea and vomiting was not reduced by P6 acupuncture (23 of 40 or 58%) compared with control (24 of 40 or 60%), but was significantly higher with sham acupuncture (33 of 40 or 83%;P = 0.0153). The composite PONV incidence was significantly lower with P6 acupuncture (25 of 40 or 63%; NNT, 3.3;P = 0.0007) but not with sham acupuncture (35 of 40 or 88%;P = not significant) compared with controls (37 of 40 or 93%;fig. 2and table 2).

Antiemetic efficacy was also assessed by analysis of time to vomiting. The survival function for vomit-free interval was significantly unequal between groups (P = 0.0136), with sham acupuncture patients vomiting significantly earlier (fig. 3).

This study provides strong evidence that P6 electroacupuncture in awake patients reduces the feeling of nausea but that the effect may not be powerful enough to reduce the incidence of vomiting after tonsillectomy, a powerfully emetogenic procedure. A previous study in adult postoperative patients using acupressure at P6 also reduced nausea but not vomiting. 19Transcutaneous acupoint stimulation at P6 has been shown to reduce nausea but not vomiting after laparoscopic choleysystectomy in adults. 20These are similar to our findings in children. Although we suspect that vomiting may occur without nausea as a result of pharyngeal or gastric stimulation, this study does not address this hypothesis.

The overall incidence of nausea and vomiting in these patients was high and was similar to that reported in the surgical and anesthetic literature. 1,2Dexamethasone has been shown to have antiemetic effects with an NNT of 3 or 4 patients for PONV. 21–24Because of confounding antiemetic effects, dexamethasone was not used as a surgical adjunct. The incidences of nausea, vomiting, and rescue antiemetic treatment are often grouped in studies aimed at evaluation PONV 25,26or at establishing a risk score for the occurrence of PONV, 27and for this reason, we evaluated each variable independently and as a composite. The composite outcome was thought to be important because a higher incidence in vomiting, which did not in itself reach statistical significance, might have offset the significantly reduced incidence of nausea, or vice versa . However, we did not see this effect. The reduction in nausea with P6 acupuncture was robust but was not accompanied by a decrease in incidence of vomiting or need for rescue therapy compared with the control group. Therefore, active P6 acupuncture was effective in reducing PONV.

Postoperative nausea or vomiting remains difficult to treat in the pediatric patient undergoing tonsillectomy, adenoidectomy, or both. Although active acupuncture reduced the incidence of nausea and was effective in preventing PONV in 37% of recipients for an NNT of 3.3, we would not support its use in lieu of prophylactic intravenous antiemetics. White and Watcha 25recommended routine antiemetic prophylaxis for these high-risk surgical patients with single or combination drug therapy having a treatment effect yielding an NNT of 2.5–4.5. Furst and Rodarte 1showed that prophylactic ondansetron in this patient population reduced the incidence of vomiting from 62% to 27% for an NNT of 2.9. Welters et al.  28showed that prophylactic antiemetic was effective (NNT of 3.3) in reducing PONV in strabismus surgery. Rose et al.  29showed that 0.15 mg/kg ondansetron, administered before surgical manipulation of the eye muscles, was effective in reducing PONV in strabismus patients and speculated that the duration of hospital stay and costs due to PONV can be decreased with prophylactic use of this drug. Similar estimates of treatment efficacy have been found for ondansetron (NNT of 3) 30and droperidol (NNT of 4) 31after strabismus surgery. Therefore, the efficacy of P6 acupuncture was comparable with the most effective pharmacotherapies but was more labor intensive.

Pharmacologic therapies to manage nausea have been directed at the neurotransmitter receptors in the brain regions receptive to emetic stimuli. Ondansetron is a pure 5-hydroxytryptamine type 3 receptor antagonist, and it is this mechanism that is thought to be responsible for its antiemetic effects. Electroacupuncture has been thought to modulate serotonin, dopamine, substance P, and endogenous endorphins in the central nervous system. 17However, we do not know the neurochemical basis for the effect of electroacupuncture on nausea, but intact peripheral and central nervous system processing seems to be important for the efficacy of acupuncture because other studies in anesthetized patients 14have failed to show the positive treatment effect we found with stimulation in awake patients.

Interestingly, the odds of PONV was 4.2 times higher with sham puncture than P6 acupuncture, and sham acupuncture patients vomited significantly earlier than controls or P6 acupuncture patients. The worsened outcome with sham acupuncture raises the possibility that improperly administered acupuncture may be harmful, perhaps by stimulation of other active acupuncture sites or simply by increasing nociceptive input. The potentially harmful effect of improper acupuncture technique and apparent requirements for patient awareness are barriers to the widespread use of acupuncture for primary prophylaxis of PONV, especially in the needle-phobic pediatric patient.

In summary, this study presents evidence that active awake electroacupuncture reduces the incidence of nausea but not vomiting after tonsillectomy in the pediatric patient. Reduction in nausea was significant and robust enough that the overall incidence of PONV was also significantly reduced, with a treatment effect as powerful as that of pharmacologic therapies. Whether this efficacy is additive to effective pharmacotherapy or more useful in the treatment-resistant patient warrants further study.

The authors thank Joseph Kerschner, M.D., Valerie Flanary, M.D., Thomas Kidder, M.D., Steven Conley, M.D., (all from the Department of Otolaryngology, Medical College of Wisconsin, Milwaukee, WI), Charles Harkins, M.D., David Beste, M.D. (both from Pediatric Ear, Nose and Throat Specialists, Milwaukee, WI), and Jill Potter, R.N. (Nursing Department, Children's Hospital of Wisconsin, Milwaukee, WI), for facilitating patient enrollment.