www.elsevier.com/locate/jpedsurg

Journal of Pediatric Surgery (2013) 48, 2425–2430

Postoperative pain management in pediatric patientsundergoing minimally invasive repair of pectus excavatum:The role of intercostal block☆,☆☆

Laura Lukosiene a,⁎, Danguole Ceslava Rugyte a, Andrius Macas a, Lina Kalibatiene a,Dalius Malcius b, Vidmantas Barauskas b

aDepartment of Anesthesiology, Hospital of Lithuanian University of Health Sciences Kaunas Clinics, LithuaniabDepartment of Pediatric Surgery, Hospital of Lithuanian University of Health Sciences Kaunas Clinics, Lithuania

Received 17 August 2013; accepted 26 August 2013

P

S+

0h

Key words:Pectus excavatum;Intercostal block;Morphine;Children;Postoperative pain

AbstractPurpose: There are no published data regarding value of intercostal block following pectus excavatumrepair. Our aim was to evaluate the efficacy of intercostal block in children following minimallyinvasive repair of pectus excavatum (MIRPE).Methods: Forty-five patients given patient-controlled analgesia (PCA) with morphine postoperativelywere studied. Twenty-six patients were given bilateral intercostal blocks after induction of anesthesia(PCA-IB group), and nineteen patients were retrospective controls without regional blockade (PCAgroup). All patients were followed up 24 h postoperatively.Results: A loading dose of morphine (0,1 ± 0,49 mg/kg) before starting PCA was used in seventeenpatients in PCA group vs. no patient in PCA-IB group. Cumulative used morphine doses were lower upto 12 h after surgery in PCA-IB group (0,29 ± 0,08 μg/kg) than in the PCA group (0,46 ± 0,18 μg/kg),p b 0,01. There were no differences in pain scores, oxygen saturation values, sedation scores, and theincidence of pulmonary adverse events between the two groups. There was a tendency towards lessmorphine-related adverse effects in PCA-IB group compared to PCA group (p b 0,05). Nocomplications related to the intercostal blocks were observed.Conclusion: Bilateral intercostal blocks following MIRPE are safe and easy to perform and can diminishpostoperative opioid requirement. Double-blind randomized study is required to confirm the potential todiminish opioid related side effects.© 2013 Elsevier Inc. All rights reserved.

☆ Sponsoring Member: Walter J. Chwals, MD, FACS, FCCM, FAAP,rofessor of Surgery and Pediatrics, Tufts Medical Center, Boston, MA.

☆☆ Conflict of Interest. The authors declare that is no conflict of interest.⁎ Corresponding author. Hospital of Lithuanian University of Health

ciences Kaunas Clinics Anesthesiology Kaunas Clinics, Lithuania. Tel.:37069877511.E-mail address: lauralukos@gmail.com (L. Lukosiene).

022-3468/$ – see front matter © 2013 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.jpedsurg.2013.08.016

Pectus excavatum (PE) is a common chest wall deformityin children. The incidence is estimated to be from 1:300 to1:1000, depending on race and country [1–3]. Male gender isat higher risk than female [1–3]. PE in older children is oftenassociated not only with psychological stress, but also withreduced pulmonary and cardiovascular function, especiallywith exercise [4,5]. For over a decade a closed minimally

2426 L. Lukosiene et al.

invasive procedure introduced by Donald Nuss hasperformed instead of open Ravitch repair [6,7]. Nevertheless,pain following minimally invasive repair of pectus excava-tum (MIRPE) is significant. Although epidural analgesia iseffective and widely used after this kind of surgery, concernshave recently been raised over costs, invasiveness and safetyof the epidural approach [8,9]. Furthermore, studies indicatethat systemic analgesia with opioids can be as effective asepidural analgesia [10]. The advantages of systemic opioidsare low cost, technical simplicity and safety. However, evenwith high doses, sufficient analgesia may be difficult toachieve, especially in the early postoperative period andduring increased physical activity [9,11]. In turn, opioid-related side effects may effect patient well-being and reduceoverall quality of postoperative analgesia.

Our previous experience with a multimodal approachcombining systemic opioids and non-steroidal anti-inflam-matory drugs (NSAIDs) for pain management followingMIRPE indicated that acute postoperative period is extreme-ly challenging and requires additional treatment [11]. Anintercostal block is less invasive than epidural anesthesia, isrelatively simple to perform and has been shown to besuccessfully used in children after thoracic and upperabdominal procedures [12–14]. The advantages of intercos-tal block include good analgesia, an opioid-sparing effect,improved pulmonary mechanics, reduced central nervoussystem depression, and avoidance of urinary retention [15].

There are no data regarding the value of intercostal blockfollowing the pectus excavatum repair. Therefore, the aim ofthe present study was to evaluate the efficacy of intercostalblock in pediatric patients undergoing MIRPE. We prospec-tively studied 26 patients administered bilateral intercostalblocks and compared them with retrospective controls,treated using patient controlled analgesia (PCA) withoutregional anesthesia.

1. Methods

Participants of study were children undergoing MIRPE inHospital of Lithuanian University of Health Sciences KaunasClinics, Department of Pediatric Surgery from January 2008to March 2012.

After approval by local Regional Biomedical ResearchEthics Committee (Protocol Nr. BE-2-26), 26 patients wereprospectively examined. Consent was obtained from all thepatients or their parents. They were given bilateral intercostalblocks after induction of anesthesia (PCA-IB group). Thegroup of retrospective controls (PCA group) consisted of 19patients without regional blockade. Patients in both groupswere followed up 24 h postoperatively using the samefollow-up documentation chart.

In both groups patients underwent general anesthesia withtracheal intubation, following premedication with either oralor intravenous midazolam. Isoflurane (n = 8), sevoflurane

(n = 37), fentanyl (n = 45), atracurium (n = 24), or rocur-onium (n = 21) were used for maintenance of anesthesia.After induction of anesthesia, single shot intercostal blockswere performed in the PCA-IB group.

Intercostal blocks were performed using technique firstdescribed by Moore [16], later by Kopatcz et al, Jagannathanet al [15,17]. The block was performed at the mid-axillaryline with the patient lying in supine position. Using steriletechnique, the block sites were identified and the lower ribmargins were then located. The needle was inserted at an 80°angle with the skin and advanced until the rib was contacted.Afterwards the needle was gradually “walked” off the caudaedge of the rib. As soon as the needle lost contact with therib, the needle was slowly inserted 0.3–0.4 cm deeper thanthe lower border of the rib. After negative aspiration 4 to5 ml of bupivacaine containing 5 μg/ml of adrenaline wasinjected. The procedure was repeated bilaterally into eachintercostal space from Th4 to Th8. A short, beveled 22 gaugenerve stimulation needle was used in order to diminish therisk of pleural puncture. Ultrasound was not used.

No regional blocks had been given to retrospectivecontrol patients. After surgery, all patients were extubatedand PCA treatment was started. The PCA device settings formorphine were the same in both groups: bolus dose 20 μg/kg, lock-out time 5–6 min, max 4 h dose 400 μg/kg, and abackground infusion 5–6 μg/kg/h. The loading dose ofmorphine could be given before starting PCA at thediscretion of the attending anesthesiologist. This decisionwas made on the basis of patient's pain level (restlessness,agitation or crying on emergence).

All patients were given baseline analgesia of acetamin-ophen and ketoprofen by intravenous or oral/rectal routes.

All patients were continuously monitored for respiratoryfunction (oxygen saturation and respiratory rate), sedationand pain at rest until the discontinuation of PCA in theintensive care unit and surgical ward thereafter. The follow-up data were recorded at every 3 h, except for oxygensaturation, which was recorded hourly.

For evaluating and quantifying the degree of sedation a 4-point sedation scale was used: 0 – awake, 1 – drowsy, 2 –asleep, easy to arouse, 3 – asleep, difficult to arouse. Thechildren expressed their pain by a 10-point visual coloranalogue scale (VAS) from lower end of no pain to upper endof the worst possible pain. The backside of the scale wasgraded from 0 – no pain to 10 – worst possible pain [18].

The following adverse effects were scored as present orabsent: nausea/vomiting – retching or expulsion of gastriccontents, urinary retention – catheterization of the bladder.The need for oxygen therapy, any abnormal radiologicalfindings and other pulmonary adverse effects were noted.

1.1. Statistics

Continuous variables were checked for the normality ofdistribution. Normally distributed variables, such as patient's

Fig. 1 Median cumulative used morphine doses 3, 6, 12, and24 h after surgery. Error bars are 75 percentiles. *p b 0,0125(Mann-Whitney U-test with Bonferroni correction).

Table 1 Patient's demographics, intraoperative fentanyl,postoperative doses of non-opioid analgesics, prophylacticondansetron use and operating room time.

PCA-groupN = 19

PCA-IB groupN = 26

Age (years) 14,6 ± 3 14,3 ± 2,7Weight (kg) 53,8 ± 13,4 52,2 ± 13,5Gender: male/female 14/5 17/9Intraoperative fentanyl (μg/kg) 4,7 ± 1,3 ⁎2,4 ± 0,924-h paracetamol dose (mg/kg) 28 ± 14,4

(N = 10)(N = 24)⁎19,4 ± 6,7

24-h ketoprofen dose (mg/kg) 3,5 ± 1,4(N = 16)

3,0 ± 0,96(N = 26)

Prophylactic ondansetron 33% (N = 18) (N = 24)⁎79%

Operating room time ⁎⁎ (min) 73,3 ± 14,1 70 ± 8,4

Data are mean ± SD.⁎ p b 0,05 compared to PCA group.⁎⁎ Time between intubation and extubation.

2427The role of intercostal block

age, weight, non-opioid analgesic doses, and oxygensaturation values, were compared using the Student's t-test.Non-normally distributed variables, such as cumulativeopioid doses, were compared using the Mann–Whitney U-test. A Bonferroni correction was used for multiplecomparisons of cumulative morphine doses at differenttime intervals. Fisher's exact and z-test were used forproportions, such as gender and the incidence of adverseeffects. For the comparison of overall VAS scores during24 h, a general linear model for repeated measures was used.A p-value of less than 0,05 was considered statisticallysignificant. Results are presented either as means ± SD,median (25–75 percentiles), or as a number of patients (%).

2. Results

Demographic data, doses of analgesics used, prophylacticondansetron use and operating room time (between intuba-tion and extubation) are shown in Table 1. The two groupsdid not differ according to age, weight, gender, postoperativeketoprofen dose and operating room time, but lessintraoperative fentanyl was required in the PCA-IB group.Oral or rectal acetaminophen was used in half of the patientsin the PCA group, whereas intravenous acetaminophen wasused in almost all patients in PCA-IB group. Prophylacticondansetron usage was more frequent in the PCA-IB group.

Seventeen patients in the PCA-group required a loadingdose of morphine in the operating room before starting thePCA. Dosage ranged between 0.04 and 0.22 (mean ± SD0.1 ± 0.49) mg/kg. No patient needed loading dose in thePCA-IB group. Cumulative used morphine doses (excludingthe loading dose) 3, 6, 12 and 24 h after surgery are shown inFig. 1. Significantly lower median (25–75 percentiles) doses

up to 6 h after surgery were used in the PCA-IB group(0.15 (0.11–0.20) mg/kg) than in the PCA-group(0.26 (0.18–0.32) mg/kg), p b 0.0125. No differences wereobserved thereafter.

Pain scores every 3 h up to 24 h after surgery are shownin Fig. 2. There were no differences in pain scores betweenpatients in the PCA and PCA-IB groups.

Postoperatively observed oxygen saturation values,sedation scores, pulmonary and opioid-related adverseeffects are shown in Table 2. Oxygen saturation values,sedation scores and the incidence of pulmonary adverseevents did not differ between the two groups. Patients givenintercostal blocks suffered less morphine-related adverseeffects compared to those not given regional blocks: 9/26 vs.12/19, p b 0,05 (z-test), respectively. Postoperative oxygentherapy up to 8 h after surgery was required in 2 patients inthe PCA-IB group: one with atelectasis and one with residualpneumothorax. No other treatment, except for physiotherapywas applied in these patients. Two patients requiredadditional oxygen therapy up to 3 h postoperatively in thePCA group.

3. Discussion

Several decades ago, D. Moore described the technique ofintercostal block [19]. Studies in adults have shown thatintercostal blocks improve analgesia, reduce opioid intakeand have favorable effects on respiratory function, comparedto opioids alone [20,21]. To our knowledge, this is the firststudy describing analgesia when bilateral intercostalblocks were used to manage early postoperative pain afterthe Nuss procedure.

The main finding is that morphine use was significantlyreduced with intercostal blocks. A second important findingis that the total number of morphine-related side effects werealso decreased with the additional use of intercostal blocks.Reduced opioid doses were reported in children and adult

Fig. 2 Mean pain scores every 3 h after surgery in PCA and PCAIB group patients (error bars are standard deviations); p N 0,05(general linear model for repeated measures).

Table 2 Postoperative oxygen saturation, sedation scores,pulmonary and opioid-related adverse events in PCA andPCA-IB groups.

Variable PCA groupN = 19

PCA-IBN = 26

Spo2 96,6 ± 1,7 96,7 ± 1,0Sedation 0,7 ± 0,4 0,9 ± 0,4Pulmonary adverse eventsPneumonia (infiltration) 1 (5.3%) -Upper respiratory infection 1 (5,3%) -Respiratory insufficiencyowing to pain and bilateralpneumothorax

1 (5,3%) -

Segmental atelectasis - 1(3,8%)Haemothorax - 1 (3,8%)Postoperative oxygentherapy required

2 (10,5%) 2 (7,7%)

Total 5 (26%) 4 (15%)Opioid-related adverse effectsNausea/vomiting 8 (42,1%) 7 (26,9%)Urinary retention 3 (15,8%) 2 (7,7%)Respiratory rate b10/min - -Discontinuation of PCA therapyowing to persistentnausea/vomiting

1(5,3%) -

Total 12 (63,2%) 9 (34,6%) ⁎

Data are mean ± SD and number of cases (%).⁎ p b 0,05, compared to PCA group (z-test).

2428 L. Lukosiene et al.

studies where parasternal intercostal blocks were usedfollowing heart surgery [14,20,22].

We found that opioid doses were significantly decreasedup to 6 h postoperatively (Fig. 1). Similar findings have beenreported in patients using a single shot parasternal block withbupivacaine [20]. Rothstein et al. investigated relationship ofthe dose of bupivacaine for an intercostal nerve block to theobserved blood concentration. The elimination half-life inchildren from 3 months to 16 years old was reported to be147 ± 80 min [23]. Thus, it is unlikely that analgesia wouldlast longer than several hours after a single shot technique. Asystematic review of adult studies following thoracotomiesshowed that repeated blocks or continuous infusions weremore consistently associated with decreased opioid con-sumption than single shot techniques [24].

Reduced pain scores were reported to be decreased withparasternal blocks in children, and in a majority of adultstudies exploring single shot, repeat or continuous intercostalor parasternal blocs [14,20,22,24]. Pain scores, though, werenot significantly reduced in our study. PCA allows individualtitration of analgesic. In addition, all of our patients receivedintravenous NSAIDs, which have been reported to effec-tively reduce postoperative pain in children [25]. Manypatients in our study also received acetaminophen. Thus, it isnot surprising that pain scores over 24 h were not different inthe two study groups (Fig. 2).

Reduced opioid doses are important as opioids cause sideeffects in a dose-dependent manner. Since oxygen desatura-tion is uncommon with thoracoplastic surgery whenadequate analgesia is achieved [11], nausea/vomiting re-mains the most relevant side effect. It may prevent somepatients from achieving effective analgesia postoperatively.In the present study, we had to discontinue treatment withmorphine in two PCA group patients; one owing toprotracted nausea/vomiting and one on the second postop-erative day who complained of pain and shortness of breath.These findings further support the concept that the additionaluse of intercostal blocks can improve analgesia. Further-more, we observed that overall opioid-related side effects,

including nausea/vomiting, were reduced in patients addi-tionally given intercostal blocks, likely because of thedecreased opioid requirement. We cannot with certainty,however, attribute this decrease solely to the reducedmorphine requirement in the PCA-IB group as intravenous,rather than oral, acetaminophen and more frequent prophy-lactic administration of ondansetron in this group versus thePCA group might have also contributed to this finding.

Although we found intercostal blocks to be simple toperform, meticulous technique and safety are important. Theincidence of pneumothorax is reported from less than 1% inadults [19] to 3,4% in children [12]. Residual pneumothoraxfollowing MIRPE was reported to occur in up to 30% ofcases, however, most are subclinical and resume spontane-ously [26]. According to the recent meta-analysis by A. Nasret al., clinically significant pneumothorax occurs in less than5% of cases [27]. The incidence of pneumothorax was 15%(4 patients) in the PCA-IB group vs. 11% (2 patients) in thePCA group, with only one patient in each group presentingwith clinical symptoms. Thus we believe that the incidenceof pneumothorax was not affected by the intercostalblockade. There was also one relevant clinical case ofsegmental atelectasis with desaturation, which could not beattributed to the method of analgesia but rather insufficientlung distension at the end of surgery.

We had one case of clinically significant pleural effusionin the PCA-IB group. Pleural drainage confirmed the

2429The role of intercostal block

diagnosis of hemothorax. The information regarding theincidence of bleeding following intercostal block is scarce.The intercostal artery courses along with the intercostal nervewithin the costal groove and is covered by a rib laterally.Thus, a theoretical possibility of arterial damage exists,especially when the needle is advanced too cephalad beyondthe lower edge of the rib. In addition, pleural injury couldalso occur in association with the vascular damage.However, only one case of hemothorax was reported inliterature after intercostal catheterization [28]. We cannotknow if this hemothorax occurred as a result of the blockadeor the ensuing surgical procedure, but utilization ofultrasound while performing intercostal blocks would behelpful in preventing complications as pneumo- orhemothorax. While hemothorax may occur followingMIRPE, the incidence has been reported to be less than 1%[27]. Regarding the use of NSAIDs for adjunctive paincontrol, the literature in the adult population suggests anassociated increased the risk of bleeding [29] which has notbeen found to exist in children [25], so we continue to useNSAIDs with this kind of surgery.

Because of an excellent vascular supply, systemicabsorption after intercostal block is known to be faster thanafter other regional techniques. Furthermore, the rate ofabsorption in children has been reported to be highercompared to adults [23]. This may result in high plasmaconcentrations of local anesthetic and systemic toxicreactions. Bupivacaine can cause severe cardiac and/orneurologic side effects if toxic concentrations are reached.While neurologic manifestations may be masked duringgeneral anesthesia, cardiac complications should be kept inmind and appropriate measures to reduce absorption and therisk of toxicity applied. Johnson et al. found, that epinephrineadded to bupivacaine for intercostal blockade decreasedplasma concentrations up to 50% [30]. Rothstein et al. foundthat bupivacaine in doses of 2–4 mg/kg with epinephrine1:200 000 was safe for intercostal nerve blockade in childrenunder general anesthesia without demonstrable side effects[23]. We used 0,25% bupivacaine with 5 μg/ml (1:200 000)of epinephrine to reduce absorption and did not exceed 2 mg/kg of total bupivacaine dose. Accidental intravascularinjection can be another cause of systemic toxicity. We,therefore, always perform an aspiration test before andduring injection. Safer local anesthetics, such as levobupi-vacaine or ropivacaine, can diminish the risk of toxicity andare used in children and adults [14,20].

It took on average 10 min to perform the bilateralintercostal blocks, but operating room time was notsubstantially prolonged. We think that lower doses ofintraoperative opioid in the PCA-IB group allowed for fasterextubation after procedure, which is consistent with theprevious observations [14,20].

The results of our present study might have been biasedby several methodological limitations: undefined strategiesfor the use of perioperative fentanyl and the loading dose ofmorphine, as well as different treatment patterns of

prophylactic antiemetics and non-opioid analgesics betweenthe two groups. Finally, the unblinded design of retrospec-tive–prospective study introduces potential bias.

4. Conclusion

We found that multiple intercostal blocks followingMIRPE are safe and easy to perform and can be applied insurgical settings to reduce the dose of PCA-administeredopiates. We also found a decreased incidence of opioid-related adverse effects in patients additionally givenintercostal blocks. However, methodological limitations ofthe present study warrant double-blind, randomized trials tobe performed in order to confirm these findings.

References

[1] Shamberger RC, Welch KJ. Surgical repair of pectus excavatum.J Pediatr Surg 1988;23:615-22.

[2] Goretsky MJ, Kelly Jr RE, Croitoru D, et al. Chest wall anomalies:pectus excavatum and pectus carinatum. Adolesc Med 2004;15:455-71.

[3] Kelly Jr RE. Pectus excavatum: historical background, clinical picture,preoperative evaluation and criteria for operation. Semin Pediatr Surg2008;17:181-93.

[4] Coln E, Carrasco J, Coln D. Demonstrating relief of cardiaccompression with the Nuss minimally invasive repair for pectusexcavatum. J Pediatr Surg 2006;41:683-6.

[5] Lawson ML, Mellins RB, Tabangin M, et al. Impact of pectusexcavatum on pulmonary function before and after repair with theNuss procedure. J Pediatr Surg 2005;40:174-80.

[6] Nuss D, Kelly Jr RE, Croitoru DP, et al. A-10 year review of aminimally invasive technique for the correction of pectus excavatum.J Pediatr Surg 1998;33:545-52.

[7] Nuss D. Minimally invasive surgical repair of pectus excavatum.Semin Pediatr Surg 2008;17:209-17.

[8] St Peter SD, Weesner KA, Sharp RJ, et al. Is epidural anesthesia trulythe best pain management strategy after minimally invasive pectusexcavatum repair? J Pediatr Surg 2008;43:79-82.

[9] St Peter SD, Weesner KA, Weissend EE, et al. Epidural vs patient-controlled analgesia for postoperative pain after pectus excavatumrepair: a prospective, randomized trial. J Pediatr Surg 2012;47:148-53.

[10] Butkovic D, Kralik S, Matolic M, et al. Postoperative analgesia withintravenous fentanyl PCA vs epidural block after thoracoscopic pectusexcavatum repair in children. Br J Anaesth 2007;98:677-81.

[11] Rugyte DC, Kilda A, Karbonskiene A, et al. Systemic postoperativepain management following minimally invasive pectus excavatumrepair in children and adolescents: a retrospective comparison ofintravenous patient-controlled analgesia and continuous infusion withmorphine. Pediatr Surg Int 2010;26:665-9.

[12] Shelly MP, Park GR. Intercostal nerve blockade for children.Anaesthesia 1987;42:541-4.

[13] Matsota P, Livanios S, Marinopoulou E. Intercostal nerve block withBupivacaine for post-thoracotomy pain relief in children. Eur J PediatrSurg 2001;11:219-22.

[14] Chaudhary V, Chauhan S, Choudhury M, et al. Parasternal intercostalblock with ropivacaine for postoperative analgesia in pediatric patientsundergoing cardiac surgery: a double-blind, randomized, controlledstudy. J Cardiothorac Vasc Anesth 2012;26:439-42.

2430 L. Lukosiene et al.

[15] Kopacz DJ, Thompson GE. Intercostal blocks for thoracic andabdominal surgery. Tech Reg Anesth Pain Manag 1998;2:25-9.

[16] Moore DC. intercostal nerve block in 4333 patients. Anesth Analg1962;41:1-11.

[17] Jagannathan N, Suresh S. Truncal blocks in children. Tech Reg AnesthPain Manag 2007;11:260-4.

[18] McGrath PA, Seifert CE, Speechley KN, et al. A new analogue scale forassessing children's pain: an initial validation study. Pain 1996;64:435-43.

[19] Moore DC. Intercostal nerve block for postoperative somatic pain followingsurgery of thorax and upper abdomen. Br J Anaesth 1975;47:284-6.

[20] McDonald SB, Jacobsohn E, Kopacz DJ, et al. Parasternal block andlocal anesthetic infiltration with levobupivacaine after cardiac surgerywith desflurane: the effect on postoperative pain, pulmonary function,and tracheal extubation times. Anesth Analg 2005;100:25-32.

[21] Taylor R, Massey S, Stuart-Smith K. Postoperative analgesia in video-assisted thoracoscopy: the role of intercostal blockade. J CardiothoracVasc Anesth 2004;18:317-21.

[22] Barr AM, Tutungi E, Almeida AA. Parasternal intercostal block withropivacaine for pain management after cardiac surgery: a double-blind,randomized, controlled trial. J Cardiothorac Vasc Anesth 2007;21:547-53.

[23] Rothstein P, Richard Arthur G, Feldman HS, et al. Bupivacaine forintercostal nerve blocks in children: blood concentrations andpharmacokinetics. Anesth Analg 1986;65:625-32.

[24] Joshi GP, Bonnet F, Shah R, et al. A systematic review of randomizedtrials evaluating regional techniques for postthoracotomy analgesia.Anesth Analg 2008 Sep;107(3):1026-40.

[25] Michelet D, Andreu-Gallien J, Bensalah T, et al. A meta-analysis ofthe use of nonsteroidal antiinflammatory drugs for pediatric postop-erative pain. Anesth Analg 2012;114:393-406.

[26] Futagawa K, Suwa I, Okuda T, et al. Anesthetic management for theminimally invasive Nuss procedure in 21 patients with pectusexcavatum. J Anesth 2006;20:48-50.

[27] Nasr A, Fecteau A, Wales PW. Comparison of the Nuss and theRavitch procedure for pectus excavatum repair: a meta-analysis.J Pediatr Surg 2010;45:880-6.

[28] Dangoisse M, Collins S, Glynn CJ. Haemothorax after attemptedintercostal catheterisation. Anaesthesia 1994;49:961-3.

[29] Elia N, Lysakowski C, Tramèr MR. Does multimodal analgesiawith acetaminophen, nonsteroidal antiinflammatory drugs, orselective cyclooxygenase-2 inhibitors and patient-controlledanalgesia morphine offer advantages over morphine alone?Meta-analyses of randomized trials. Anesthesiology 2005;103:1296-304.

[30] Johnson MD, Mickler T, Arthur GR. Bupivacaine with and withoutepinephrine for intercostal nerve block. J Cardiothorac Anesth 1990;4:200-3.