Mean visual analog scale (VAS) score for the placebo and local anesthetic (LA) groups during 24 hours.
Vasan NR, Stevenson S, Ward M. Preincisional Bupivacaine in Posttonsillectomy Pain ReliefA Randomized Prospective Study. Arch Otolaryngol Head Neck Surg. 2002;128(2):145-149. doi:10.1001/archotol.128.2.145
To determine the effect of preincisional bupivacaine hydrochloride infiltration on postoperative pain after tonsillectomy.
Prospective, randomized, double-blind clinical trial.
A secondary/tertiary referral center in Christchurch, New Zealand.
A volunteer sample of 70 patients, aged 16 to 42 years, with recurrent tonsillitis. Seven patients were excluded.
After randomization, one group received 5 mL of 0.5% bupivacaine hydrochloride in the peritonsillar space, with the patient under general anesthesia. The other group received 5 mL of isotonic sodium chloride solution, with the patient under general anesthesia. Both groups underwent surgery with a standardized surgical and anesthetic technique.
Main Outcome Measures
Postoperative pain was assessed with a visual analog scale at 15 minutes and 1, 4, 12, 16, and 24 hours after the procedure. Postoperative analgesic requirement, length of admission, and antiemetic requirement were also assessed.
No statistical difference was found between the 2 groups for postoperative pain by means of the visual analog scale at any time interval, nor was any statistical difference found for the other variables measured. A trend toward less pain in the immediate postoperative period in the group receiving bupivacaine was noted.
No statistically significant benefit is found for use of preincisional bupivacaine in tonsillectomy.
THE ROLE of local anesthetic (LA) infiltration in tonsillectomy is controversial. Studies have been published that support and refute the use of LA during tonsillectomy. Proponents of LA infiltration claim a reduction in postoperative pain that in some studies has shown a benefit up to 10 days postoperatively.1,2
Local anesthestic is thought to act by impeding noxious stimulation of C-fiber afferent neurons, thereby diminishing the excitability of dorsal horn neurons.1 The excitability produced by nociceptive stimuli may contribute to postoperative pain, even when procedures are performed under general anesthesia.1 Confirming the benefit of preincisional LA analgesia statistically has been difficult.
Tonsillectomy is known to cause severe pain postoperatively. The pain affects the patient's nutrition, ability to return to work or school, discharge from the hospital, and satisfaction with the whole process. Our study was designed to determine whether LA has an effect on postoperative pain after tonsillectomy. Only older teenagers and adults were recruited to ensure that participants could understand and complete a visual analog scale (VAS). We excluded children to avoid observer bias when assessing pain.
Ethical approval was obtained from the Canterbury Ethic Committee (Christchurch, New Zealand); 70 patients, aged 16 to 42 years, were recruited for the study. All patients had attended an otolaryngology outpatient clinic with a history of recurrent tonsillitis. Patients with allergies, those with bleeding disorders, those using regular analgesic medication, and those with significant comorbidities were excluded. Patients were enrolled between May 1, 1998, and September 30, 1999.
All patients gave consent and were instructed on how to complete a VAS before surgery. A 100-mm horizontal-line VAS was used, where 0 mm represented no pain, and 100 mm, the worst pain imaginable. Patients were given a new VAS at each testing interval and were instructed to mark on the line the approximate level of their pain at that moment. The VAS has been found to be reliable and easily used in a number of studies.3,4
A standard anesthetic protocol was constructed for the study patients and administered by a number of anesthetists. The protocol consisted of the following:
Premedication: acetaminophen, 20 mg/kg (up to 1.5 g) orally 40 minutes to 1 hour preoperatively; intravenous induction: fentanyl citrate (1 µg/kg), propofol (2 to 3 mg/kg), and mivacurium chloride (0.1-0.2 mg/kg).
Maintenance: oxygen and nitrous oxide in a ratio of 1:2 (oxygen saturation >94%), isoflurane (0.25%-2.00% end-tidal), and morphine, 0.1 mg/kg (maximum, 10 mg).
Antiemetic: cyclizine hydrochloride, 1 mg/kg (maximum, 50 mg).
Recovery: analgesia: morphine (0.02 mg/kg as needed to every 5 minutes if pain score ≥60 mm and respiratory rate ≥8/min); antiemetic: ondansetron (0.15 mg/kg as needed 1 time for nausea or vomiting; maximum, 8 mg).
The tonsillectomy was performed by 1 of 3 surgeons (N.R.V., S.S., and M.W.) with a standardized blunt dissection technique. No concurrent procedures were performed. All patients had been randomized into the LA group or placebo group (isotonic sodium chloride solution) by means of a sealed envelope to determine which solution was required before they entered the operating room. Surgeon, patient, anesthetist, and recovery and ward staff were all blinded to the solution used.
After induction and positioning of the patient, the tonsil was medialized by means of Denis Browne forceps. The lateral surface of the tonsil was identified submucosally. With the use of an aspiration-injection technique, 3 mL of a 0.5% bupivacaine hydrochloride solution (or isotonic sodium chloride solution in the control group) was injected into the peritonsillar area in approximately the same position where a peritonsillar abscess would be drained. Two milliliters of solution was injected into the peritonsillar space at the upper pole (total, 5 mL on each side). After a wait of 5 minutes, a blunt dissection snare technique was used to remove the tonsils, with hemostasis achieved with silk ligatures only.
Patients' pain scores were assessed by means of a VAS at fixed intervals after the end of the procedure. These times were 15 minutes and 1, 4, 12, 16, and 24 hours after extubation. Patients were blinded to their previous VAS scores. Postoperative analgesia was divided into regular medication and rescue analgesia. Regular medication was started 4 hours after premedication and consisted of acetaminophen (1 g orally or rectally every 4-6 hours) or codeine phosphate (30 mg orally every 6 hours). The rescue analgesia was administered by nursing staff according to a 0-to-10 verbal pain score, in which less than 3 indicated no extra medications; 3 to less than 6, codeine phosphate given as needed, 30 mg orally every 6 hours; and 6 or more, intramuscular morphine every 4 hours. Metoclopramide hydrochloride was given as an antiemetic on an as-needed basis. Outcome measures included VAS, analgesic requirement (recovery and ward), duration of admission, and antiemetic requirement. This study did not assess blood loss during tonsillectomy. A note regarding intraoperative difficulty was also made, as cases where interval tonsillectomy was being performed (after drainage of a peritonsillar abscess) were included in the study.
Statistical analyses were performed with SAS for Windows (Version 6.12; SAS Institute Inc, Cary, NC). The VAS scores of the placebo group and the treatment group were compared with a repeated-measures analysis of variance model. Analysis between continuous variables was done with either an unpaired t test or a Mann-Whitney test. The χ2 tests (with Yates correction) were used for categorical data. Results are expressed as means and SDs where appropriate, or as medians and ranges. To detect a difference between the 2 groups of 15 mm and to have a power greater than 0.81 (SD of 15 mm), 30 patients in each group were required, assuming a 2-tailed significance test at α = .05.
Seventy patients were randomized into the LA and placebo groups. Seven patients were excluded because of breaches of anesthetic protocol (2 patients), postoperative analgesia protocol (2 patients), and changes in surgical technique (3 patients). A total of 31 patients (8 male) received LA. Thirty-two patients (9 male) received isotonic sodium chloride solution. The mean (SD) age was 23.4 (6.7) years and 23.1 (6.4) years for the LA and placebo groups, respectively. Of the 63 patients in the study, only 46 patients had a VAS completed at each of the recording times. Data were missed because the patient was asleep and the VAS recording was not obtained, or because of nursing error. The number of VAS scores obtained for each time and the mean values recorded are shown in Table 1.
Results between the 2 groups for each time were compared by means of a repeated-measures analysis of variance. No statistical difference between the 2 groups was found (F = 3.03, P = .09). At 15 minutes, the LA group had a lower mean score (45 mm) than the placebo group (60 mm), but this was not significant. The overall mean for the LA group was less (36 mm) than that for the placebo group (42 mm), but again this was not significant.
Figure 1 shows a trend toward less pain in the LA group that was not statistically significant. All other variables (Table 2), including postoperative analgesic consumption, time to first codeine tablet, length of admission, and antiemetic requirement, showed no significant difference between the 2 groups. No patient suffered any adverse effect from LA infiltration (ie, respiratory obstruction or local anesthetic toxic effects).
This double-blinded, randomized, prospective study did not demonstrate any statistical benefit with preincisional bupivacaine treatment in patients undergoing tonsillectomy. The hypothesis behind preemptive analgesia is to prevent or reduce any "memory" of the painful stimulus in the nervous system.5 Subsequently, this reduction in pain memory should lower any subsequent analgesia needs.5 In this study, we were unable to confirm this hypothesis.
In this study, 3 surgeons performed tonsillectomy by means of a fixed anesthetic protocol, injection method, and operative technique. The method was constructed to closely resemble techniques in common practice in New Zealand. The VAS and postoperative analgesic requirements failed to show any benefit with LA.
Many studies have addressed the question of the effect of LA in tonsillectomy, with a marked variance in results. Jebeles et al2 assessed adenotonsillectomy in 1993 and found a statistically significant improvement in postoperative pain in 22 children with the use of LA. A reduction in pain in some cases within the LA group persisted until day 10. The same authors in 1991 had shown similar results with LA in 14 children undergoing tonsillectomy.1 Unfortunately, these studies suffer from inclusion of children and small sample sizes and lack detail regarding injection technique. Pain is a subjective and complex expression, and its assessment depends on personal experience, social and ethnic factors, and anxiety level as well as the patient's ability to describe the type and degree of pain on the basis of some frame of reference. The inclusion of children into such studies makes the assessment of pain even less precise. Our study was constructed to minimize the number of variables; only an older teenaged and adult patient population was used, who completed their own VAS.
Dynamic assessments of pain, such as drinking water or opening the jaw, have been used in past studies, in an attempt to measure pain objectively.2,6 We elected to assess the patients' pain by using a VAS at repeated intervals, the overall score of which would represent the contribution of constant and dynamic pain.
The majority of previous studies used electrocautery as the dissection method; however, recent evidence has shown that an electrocautery dissection technique increases postoperative morbidity in terms of pain, otalgia, and poor diet when compared with blunt dissection technique.7,8 The electrocautery dissection method used in other studies may have altered their results.
The peritonsillar region is innervated by fibers from the glossopharyngeal nerve, the lesser palatine nerves, and the lingual nerve.3 The premise for LA injection is to obtain blockade of these fibers. Descriptions of injection technique and amount of solution injected have been variable. In Jebeles and coworkers' initial 1991 article, no mention of injection technique was included.1 In our study, 5 mL of 0.5% bupivacaine hydrochloride was injected into each peritonsillar space. Our assumption was that a high volume of solution within this area would anesthetize the multiple pain fibers supplying the tonsillar bed. In a recent study,9 the injection technique attempted to block the glossopharyngeal nerve and lesser palatine nerve contributions to the fossa. That group did not mention the amount of LA administered at each site, however. In 2 studies by Schoem et al,6,10 small volumes of LA were used. In their adult study, only 1.8 mL of 0.5% bupivacaine hydrochloride was used. There was a concern in their study regarding toxicity, but the dose they used fell well short of the 225-mg adult toxic dose. Violaris and Tuffin11 assessed LA in patients by using the contralateral side as a control. We believe assessing pain in this way after tonsillectomy is difficult.
Studies investigating the role of LA in tonsillectomy are listed in Table 3. From reviews of similar studies, no conclusions regarding LA can be made. Ågren et al16 published an article comparing tonsillectomy performed with the patient under general anesthesia and LA. The patients in the LA group tolerated the procedure well and demonstrated less postoperative pain than did the general anesthesia group. Obviously, LA must exert an effect if tonsillectomy can be performed when the patient is awake. During our study, all surgeons noted that preoperative injection of fluid in the peritonsillar space aided in dissection by the blunt dissection technique.
Currently, the benefit of LA in controlling pain after tonsillectomy performed with the patient under general anesthesia remains debatable. The difficulty lies in our limited ability to assess pain accurately with the measures we use now. Pain is a multifaceted symptom, influenced by the patient's previous experiences and expectations, as well as his or her peripheral neural inputs. The question of what is a significant reduction in pain has yet to be determined. There is no clear advice in the literature on what is an acceptable reduction in pain, measured with a VAS, in any situation. Most authors use a 100-mm VAS; some then seek a difference of 20 mm between the means of the 2 groups, whereas others accept a difference of 15 mm. Many previously published articles lack power.17 In our study, a difference of 15 mm between the means of the 2 groups gives a power of 0.81. Research into ways to improve tonsillectomy continues to be troubled by issues of this nature, as institutions seek to evaluate different operative techniques, anesthetic protocols, the benefit of perioperative corticosteroids, the benefit of perioperative antibiotics, and the use of nonsteroidal anti-inflammatory drugs. Randomized, prospective, double-blind studies in all of these areas are limited in number; research using the structures outlined in this study may give answers to these issues.
Accepted for publication September 4, 2001.
Preliminary results were presented at the Annual Scientific Meeting of the New Zealand Society of Otolaryngology and Head and Neck Surgery, Apia, Samoa, September 7, 1999.
We thank the otorhinolaryngology nursing staff in the operating room and the ward in Christchurch Hospital, Christchurch, for helping to obtain the VAS scores for these patients. Special thanks to Patrick Graham, Jamie Sleigh, MD, and Teena West for the statistical analysis.
Corresponding author: Nilesh Raman Vasan, FRACS, Department of Otolaryngology–Head and Neck Surgery, Glasgow Royal Infirmary, 84 Castle St, Glasgow G4 0SF, Scotland (e-mail: firstname.lastname@example.org).
Reprints: Scott Stevenson, FRACS, Department of Otolaryngology–Head and Neck Surgery, Christchurch Public Hospital, Private Bag, Christchurch, New Zealand (e-mail: email@example.com).