Secondary Logo

Journal Logo

Original Articles – General

Comparison of intravenous and peritonsillar infiltration of tramadol for postoperative pain relief in children following adenotonsillectomy

Akkaya, Taylana; Bedirli, Nurdana; Ceylan, Tijenb; Matkap, Erkana; Gulen, Guvena; Elverici, Ozanb; Gumus, Haluka; Akin, Istemihanb

Author Information
European Journal of Anaesthesiology: April 2009 - Volume 26 - Issue 4 - p 333-337
doi: 10.1097/EJA.0b013e32831c8988
  • Free



Management of the postoperative period following adenotonsillectomy requires prevention of pain, bleeding, laryngospasm, sedation, nausea and vomiting. Establishment of good analgesia is a major concern in this period.

Different methods such as drugs [nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, corticosteroids], adjustment of surgical technique and peroperative local anaesthetic infiltration have been used to reduce pain. It is known that pain after adenotonsillectomy or analgesic usage may lead to unwanted effects [1]. For example, NSAIDs may interfere with bleeding and opioids may cause respiratory depression, sedation, or nausea and vomiting [2]. So prevention of pain with the fewest side-effects is also very important. For this reason, the role of local anaesthetic infiltration in reduction of posttonsillectomy pain reduction became popular. But there are some reported complications seen especially after deep and large volume infiltrations of local anaesthetics into tonsillar and adenoid beds. These are bilateral vocal cord paralysis, life-threatening deep cervical abscess and brain stem stroke from intraoperative cardiac asystole [3,4].

In addition to maintaining effective analgesia, apparent lack of sedation and respiratory depression compared with opioids fostered a keen interest in tramadol. Tramadol – a synthetic opioid of the aminocyclohexanol group – is a racemic mixture consisting of two isomers with a different spectrum of activity [5]. It causes activation of both opioid and nonopioid systems which are mainly involved in the inhibition of pain [6]. Tramadol is a centrally acting drug, which is effective in the treatment of moderate to severe pain [7]. With regard to its systemic action, the local anaesthetic effect of tramadol on peripheral nerves has been shown in both clinical and laboratory studies [8,9]. Also, it was shown that intravenous (i.v.) tramadol has similar analgesic properties to morphine for paediatric tonsillectomy patients [10]; however, tramadol is reported to have a high incidence of nausea and vomiting [11].

The present study is designed to test the hypothesis that peritonsillar tramadol infiltration for pain relief after adenotonsillectomy would be associated with decreased incidence of postoperative nausea and vomiting and decreased sedation scores when compared with intravenous tramadol.


After Institutional Ethics Committee approval and parents' written consent, 66 children aged 3–9 years, ASA 1–2 status, scheduled for adenotonsillectomy were enrolled in this randomized, prospective, double-blind clinical study. Patients with a history of allergy, pulmonary or cardiac diseases, peritonsillar abscess, and analgesic usage within 24 h prior to surgery were excluded.

The children were premedicated with 0.5 mg kg−1 midazolam orally, maximum 15 mg, 30 min prior to the surgery. Anaesthesia was induced with sevoflurane 8% and nitrous oxide 70% in oxygen. After induction, an i.v. cannula was inserted and mivacurium hydrochloride 0.2 mg kg−1 was administered for muscle relaxation before tracheal intubation. Anaesthesia was maintained with sevoflurane 2% and nitrous oxide 30% in oxygen. After intubation children were randomized into two groups of 33 each: group I received 2 mg kg−1 tramadol intravenously and group II received 2 mg kg−1 tramadol in 2 ml of normal saline (1 ml per tonsil) via peritonsillar infiltration. Randomization was provided by shuffled, sealed, opaque, numbered envelopes. For the infiltration, all injections were performed through the tonsillar capsule and anterior plica. All the injections were performed with a 25-gauge needle by the same surgeon. After the infiltration, a minimum of 3 min was allowed for the onset of action of tramadol before adenotonsillectomy. Adenotonsillectomy was performed with cautery dissection by the same surgeon in all cases. All patients received metoclopromide 0.2 mg kg−1 i.v. before the end of surgery. Heart rate, arterial pressure and oxygen saturations (SpO2) were recorded at 10 min intervals during the operation. At the end of the surgery anaesthesia was discontinued and all patients were extubated when airway reflexes had returned and were transferred to the postanaesthesia care unit (PACU).

In the PACU, scores for pain, nausea and vomiting, bleeding and sedation were recorded by an anaesthetist who was not aware of the group to which the patient was assigned at 0, 15, 30 and 60 min. The intensity of pain was assessed by using a modified Hannallah pain score [12] – an observational pain score (OPS) developed and tested for validity and reliability in children (Table 1). Intravenous metamizol sodium 15 mg kg−1 was planned for rescue analgesia if OPS was greater than 4. The scales used to assess sedation, nausea and vomiting, and bleeding are given in Table 2. Additional antiemetic (50 μg kg−1 ondansetron i.v.) was given if the patient vomited more than twice in the PACU. The Aldrete score was used to determine PACU discharge criteria. Patients with an Aldrete score of 9 were transferred to the postoperative surgical ward; also the time at which the patient reached an Aldrete score of 9 was noted.

Table 1
Table 1:
Observational pain scores
Table 2
Table 2:
Nausea/vomiting scale, bleeding scale and sedation scale

Pain, nausea and vomiting, bleeding and sedation were assessed in the postoperative surgical ward at 6, 12 and 24 h follow-up by the nurses who were unaware of the technique used and who were trained for the use of scores involved. For pain assessment a 4-point scale was used: 0, no pain; 1, mild pain; 2, moderate pain; 3, severe pain. Patients with a pain score greater than 2 received 15 mg kg−1 paracetamol orally. Paracetamol consumption at every 6 h period was assessed.

All data were analysed with SPSS 11.5 software (SPSS Inc., Chicago, Illinois, USA) applying the t-test to demographic data and Fisher's exact test and chi-squared test to data related to categorical comparisons. A P value of less than 0.05 was considered statistically significant.


A total of 66 children (33 in each group) were included in this study. No patient was excluded from the study. There were no statistically significant differences between the groups for sex, age, weight, height, duration of surgery and duration of anaesthesia (Table 3). No statistically significant difference was found between the groups in heart rate, systolic and diastolic arterial pressure, and oxygen saturation during the surgery (P > 0.05). In the PACU, groups I and II had comparable pain scores that were not statistically significant (P > 0.05) (Table 4). Also, in the PACU, none of the patients needed rescue analgesia.

Table 3
Table 3:
Patients and clinical data
Table 4
Table 4:
Observational pain scores in postanaesthesia care unit

There were significant differences between the groups for nausea and vomiting: four patients (12.12%) in group I and one patient (3.05%) in group II complained of nausea and vomiting in the first hour in the PACU. Three patients (9.09%) in group I received i.v. ondansetron (P = 0.238). Group I had higher sedation scores at 15 min; the difference was statistically significant (P = 0.012) (Table 5). Bleeding was assessed by the surgeon; all patients had minimal or no bleeding and there were no statistical differences between groups.

Table 5
Table 5:
Sedation scores of the groups in the postanaesthesia care unit (number of patients with sedation scores 0/1/2/3)

Both groups had a statistically similar time to reach an Aldrete score of 9 (group I = 21.21 ± 5.16 min, group II = 21.96 ± 5.58 min) (P = 0.590).

During the first 24 h, although there were no significant differences between groups in the number of patients requiring supplemental analgesia during the 1–6 h period and during the 12–24 h period, during the 6–12 h period patients in group II had significantly decreased requirement for analgesia (P = 0.022). Moreover, at the end of 24 h the number of patients requiring supplemental analgesics in group I was significantly higher than in group II (P < 0.001) (Table 6).

Table 6
Table 6:
Requirement for analgesia during the first 24 h

Pain scores in the postoperative ward at 6, 12 and 24 h were significantly higher in group I than in group II (Table 7).

Table 7
Table 7:
Pain scores of the patients in the postoperative ward (0–3 scale)

When the total additional analgesic paracetamol requirement during the first 24 h was compared, the difference between group I (141.81 mg) and group II (83.63 mg) was significant (P = 0.002).

At 6, 12 and 24 h, there were no statistically significant differences between groups regarding nausea and vomiting, bleeding and sedation.


Adenotonsillectomy, often performed on a day-case basis, is usually associated with postoperative pain and it is important to maintain pain control that allows normal life and nourishment. Moreover, children undergoing adenotonsillectomy are at high risk of airway obstruction and respiratory depression in the postoperative period. Negus and Street [13] reported posttonsillectomy upper airway obstruction in patients in whom oral midazolam was used as premediacation with morphine for postoperative analgesia. Paracetamol is a well tolerated and effective analgesic but, if used alone, often provides insufficient analgesia [14,15]. NSAIDs, including ketoprofen or ibuprofen, are also reported to provide good analgesia but they increase the risk of haemorrhage after tonsillectomy [16]. Although there have been a number of recent studies that have shown there is no significant effect of NSAID on bleeding, the potential for adverse effects of NSAIDs such as bleeding and gastrointestinal intolerability still remains. These studies reported that NSAIDs should be used at the lowest doses for the shortest time with ranitidine [17].

According to our results, peritonsillar infiltration of tramadol maintains analgesia comparable to i.v. tramadol in the PACU within the first hour. Moreover, tramadol applied by peritonsillar infiltration provides a better control of postoperative pain than i.v. tramadol during the first 12 h. But no evidence is provided from the previous studies that tramadol provides 12 h of topical analgesia. To prove this, further studies with large series may be needed. Also, our data showed that additional analgesic requirement in the peritonsillar infiltration group was lower during the first 24 h. On the contrary, some of our results, for example the results of sedation scores, are significant only during the 15 min period, so this sheds some doubt over the clinical significance and draws a distinction between statistical and clinical significance.

The infiltration of peritonsillar beds for posttonsillectomy pain control was suggested in 1953 by Allen [18]. There has been a keen interest in providing analgesia by peritonsillar infiltration of drugs to decrease side-effects. And studies evaluating the effects of local anaesthetics such as bupivacaine and ropivacaine on posttonsillectomy pain have been carried out [19,20].

Clinical studies have shown that tramadol had peripheral local anaesthetic properties [21,22]. By direct tramadol application to sciatic nerve in rats, it was proven that it exerts a local anaesthetic effect [23]. Altunkaya et al. [24] reported that a local anaesthetic action similar to that of lidocaine because of its antinociceptive effect; this effect extended into the postoperative period. Atef and Fawaz [25] reported that peritonsillar tramadol infiltration provided effective analgesia for posttonsillectomy pain. But they had used a large infiltration volume comparing with our study (1.5 ml per tonsil) and had compared this only with a placebo group. Also, they had not discussed the side-effects of tramadol. Another difference between groups was the infiltration time of the tramadol; although Atef and Fawaz infiltrated tramadol at the end of the operation, we preferred infiltration before the operation. And, in this study, we managed to maintain a longer time of pain relief than did Atef and Fawaz. We think that preincisional peritonsillar tramadol infiltration (the preemptive effect) provides this prolonged analgesia for up to 12 h, as many studies have shown that preemptive application of drugs provides more effective analgesia [26,27]. Although the potential advantages of preemptive analgesia in children remain a rather controversial issue, some studies showed that subcutaneous tramadol is associated with burning pain and local irritation [27]. So we think that this is an additional reason for injecting the tramadol early, rather than at the end of the surgery.

Dal et al.[28] applied N-methyl-D-aspartate, which is an antagonist ketamine, in patients undergoing adenotonsillectomy via peritonsillar infiltration and provided efficient pain relief with decreased sedation scores. Similarly, in our study, patients in the peritonsillar infiltration group had decreased nausea and vomiting compared with those in the i.v. group. And these results suggest that peritonsillar infiltration reduces side-effects of the analgesic drugs.

The adverse effects – facial nerve paralysis and vocal cord paralysis – of local anaesthetic infiltration (the obvious alternative to tramadol infiltration) are rather off-putting for postoperative pain control. Fortunately, studies with tramadol infiltration, including our study, did not report such complications, so we think that tramadol may have a real advantage.

A limitation of our study is that we did not have an intravenous saline group and a peritonsillar saline infiltration group. But Costas-Gastiaburo et al.[29] showed that peritonsillar saline infiltration provides significant analgesia. If there had been an intravenous saline group, it would have been possible to blind the study more effectively. Moreover, we did not have an i.v. nonopioid analgesic drug for a rescue analgesic, which is why we planned to use i.v. metamizol in PACU. Metamizol has side-effects including agranulocytosis and the risks should be taken into account.

In conclusion, peritonsillar application of tramadol in adenotonsillectomy patients provided analgesia as efficient as i.v. tramadol in the first postoperative hour. Moreover, this analgesic effect was prolonged up to 12 h in the peritonsillar infiltration group. Besides, peritonsillar tramadol infiltration significantly lowers the incidence of nausea and vomiting at the early postoperative period. So, we think that peritonsillar tramadol infiltration is a better choice than classical methods for adenotonsillectomy in children.


1 Umuroglu T, Eti Z, Ciftci H, et al. Analgesia for adenotonsillectomy in children: a comparison of morphine, ketamine and tramadol. Pediatr Anesth 2004; 14:568–573.
2 Moiniche S, Romsing J, Dahl JB, et al. Nonsteroidal antiinflammatory drugs and risk of operative site bleeding after tonsillectomy: a quantitative systemic review. Anesth Analg 2003; 96:68–77.
3 Shlizerman L, Ashkenazi D. Peripheral facial nerve paralysis after peritonsillar inflitration of bupivacaine: a case report. Am J Otolaryngol 2005; 26:406–407.
4 Weksler N, Nash M, Rozentsveig V, et al. Vocal cord paralysis as a consequence of peritonsillar inflitration with bupivacaine. Acta Anaesthesiol Scand 2001; 45:1042–1044.
5 Vickers MD, O'Flaherty D, Szekely SM, et al. Tramadol: pain relief by an opioid without depression of respiration. Anaesthesia 1992; 47:291–296.
6 Sacerdote P, Bianchi M, Manfredi B, et al. Effects of tramadol on immune responses and nociceptive thresholds in mice. Pain 1997; 72:325–330.
7 Jou IM, Chu KS, Chen HH, et al. The effects of intrathecal tramadol on spinal somatosensory-evoked potentials and motor evoked responses in rats. Anesth Analg 2003; 96:783–788.
8 Altunkaya H, Ozer Y, Kargi E, et al. Comparison of local anaesthetic effects of tramadol with prilocaine for minor surgical procedures. Br J Anaesth 2003; 90:320–322.
9 Pang WW, Huang PY, Chang DP, et al. The peripheral analgesic effect of tramadol in reducing propofol injection pain: a comparison with lidocaine. Reg Anesth Pain Med 1999; 24:246–249.
10 Engelhardt T, Steel E, Johnston G, et al. Tramadol for pain relief in children undergoing tonsillectomy: a comparison with morphine. Pediatr Anaesth 2003; 13:249–252.
11 Van den Berg AA, Halliday E, Lule EK, Baloch MS. The effects of tramadol on postoperative nausea, vomiting and headache after ENT surgery. A placebo-controlled comparison with equipotent doses of nalbuphine and pethidine. Acta Anaesthesiol Scand 1999; 43:28–33.
12 Hannallah RS, Broadman LM, Belman AB, et al. Comparison of caudal and ilioinguinal/iliohypogastric nerve blocks for control of postorchiopexy pain in pediatric ambulatory surgery. Anesthesiology 1987; 66:832–834.
13 Negus BH, Street NE. Midazolam–opioid combination and postoperative upper airway obstruction in children. Anesth Intensive Care 1995; 2:232–233.
14 Romsing J, Ostergaard D, Drozdziewicz D, et al. Diclofenac or acetaminophen for analgesia in paediatric tonsillectomy out-patients. Acta Anaesthesiol Scand 2000; 44:291–295.
15 Mather SJ, Peutrell JM. Postoperative morphine requirements, nausea and vomiting following anaesthesia for tonsillectomy. Comparison of intravenous morphine and nonopioid analgesic techniques. Pediatr Anesth 1995; 5:185–188.
16 Salonen A, Kokki H, Nuutinen J. The effect of ketoprofen on recovery after tonsillectomy in children: a 3-week follow-up study. Int J Pediatr Otorhinolaryngol 2001; 62:143–150.
17 Madhusudhan TR, Raugan A, Gregg PJ. Gastric protection and gastrointestinal bleeding with aspirin thromboprophylaxis in hip and knee joint replacements. Ann R Coll Surg Engl 2008; 90:332–335.
18 Allen RT. New method for relieving postoperative pain following tonsillectomy. Arch Otolaryngol 1953; 57:86–89.
19 Unal Y, Pampal K, Korkmaz S, et al. Comparison of bupivacaine and ropivacaine on postoperative pain after tonsillectomy in paediatric patients. Int J Pediatr Otorhinolaryngol 2007; 71:83–87.
20 Giannoni C, White S, Enneking FK, et al. Ropivacaine with or without clonidine improves pediatric tonsillectomy pain. Arch Otolaryngol Head Neck Surg 2001; 127:1265–1270.
21 Acalovshi I, Cristea T, Margarit S, et al. Tramadol added to lidocaine for intravenous regional anesthesia. Anesth Analg 2001; 92:209–214.
22 Kapral S, Gollmann G, Waltl B, et al. Tramadol added to mepivacaine prolongs the duration of an axillary brachial plexus blockade. Anesth Analg 1999; 88:853–856.
23 Pang WW, Mok MS, Chang DP, et al. Local anesthetic effect of tramadol, metaclopramide and lidocaine following intradermal injection. Reg Anesth Pain Med 1998; 23:580–583.
24 Altunkaya H, Ozer Y, Kargi E, et al. The postoperative analgesic effect of tramadol when used as subcutaneous local anesthetic. Anesth Analg 2004; 99:1461–1464.
25 Atef A, Fawaz AA. Peritonsillar infiltration with tramadol improves pediatric tonsillectomy pain. Eur Arch Otorhinolaryngol 2008; 265:571–574.
26 Hamunen K, Kalso E. A systematic review of trial methodology, using the placebo groups of randomized controlled trials in paediatric postoperative pain. Pain 2005; 116:146–158.
27 Dahl JB, Moiniche S. Preemptive analgesia. Br Med Bull 2004; 71:13–27.
28 Dal D, Celebi N, Elvan GE, et al. The efficacy of intravenous or peritonsillar infiltration of ketamine for postoperative pain relief in children following adenotonsillectomy. Pediatr Anesth 2007; 17:263–269.
29 Costas-Gastiaburo LA, Rajah V, Rubin J. Tonsillectomy and the value of peritonsillar infiltrations. S Afr J Surg 1998; 36:142–145.

adenotonsillectomy; postoperative pain; tramadol

© 2009 European Society of Anaesthesiology