In recent years, epidural analgesia has become increasingly popular in pediatrics. Usually of short duration, it plays an important role in perioperaive pain management [1-3]. In addition, epidural analgesia has found increased applicability for children with a variety of chronic pain management problems, such as cancer  and reflex sympathetic dystrophy [5,6].
Epidural abscess is a feared, but fortunately rare, complication of epidural catheter placement. Most epidural abscesses are not related to epidural catheters but are thought to be related to infections of skin, soft tissue, bones, or joints with hematogenous spread to the epidural space [7,8]. Baker et al.'s review  of spinal epidural abscesses between 1947 and 1974 totaled 39 patients, with only one infection thought to be secondary to a catheter placement. Ericsson et al.  reviewed 10 cases of epidural abscess occurring over a 10-yr period. Although five of these cases were associated with procedures on the spine, only one epidural catheter was implicated . Several other large retrospective series of epidural abscesses have not included any cases associated with epidural analgesia [7,10-12]. Conditions associated with epidural abscess include an immunocompromised state as from cancer or acquired immunodeficiency syndrome, concomitant steroid use, diabetes, debilitating disease, old age, and trauma, particularly in association with sepsis or a distal site of infection.
Reports of infection associated with epidural analgesia have generally been single cases, although Gastmeier and Gastmeier  and DuPen et al.  have published retrospective series of their experience with long-term epidural catheters used for chronic pain management in adults. Although epidural analgesia has found increased application in pediatric pain management, no studies have examined the risk of epidural abscess formation in pediatric patients after epidural catheter placement. We reviewed our experience with 1635 patients over a 6-yr period.
Protocols were approved by the Committee on Clinical Investigation at Children's Hospital. Data acquisition and chart review maintained patient confidentiality.
Epidural analgesia was initiated at Children's Hospital in February 1986. Complication records, maintained prospectively from the initiation of the acute pain service, and the overall denominator were taken from February 1986 through October 1992. Beginning in July 1991, detailed patient demographic data were entered prospectively in a SAS relational database, DBPain. Patient bills and the daily rounds information and sign-out list were created by DBPain, forcing high compliance with data entry. Demographic data are presented for the 580 patients treated since the initiation of DBPain. Statistical means, standard deviation, percentile rank, and 95% confidence intervals were calculated using DBPain.
Uniform management practices were maintained for epidural placement, infusion, and site care through the entire study period. The epidural catheters used were Portex Registered Trademark (Burron Medical Inc., London, England). Epidural catheters were placed in a sterile fashion and covered with a transparent sterile occlusive dressing permitting easy inspection of the skin entry site. No antibiotic ointment was routinely applied. A 0.2-mu filter was routinely placed between the catheter and the infusion tubing. Filters were not used with the Portex Registered Trademark 23-gauge catheters which were placed in infants and children less than approximately 2 yr of age. Attempts to use a bacterial filter with these smaller catheters resulted in malfunction of the IVAC Registered Trademark infusion pumps due to the high pressures required. After initial bolus dosing in the operating room, subsequent dosing was routinely maintained by continuous infusion. Solutions for epidural infusion were prepared by pharmacists under sterile, laminar flow conditions. Subsequently, infusion syringes were stored at 4 degrees C for not more than 7 days or at room temperature for not more than 48 h. Syringes were changed as needed by nurses on the postoperative units.
A brief neurologic examination emphasizing lower extremity strength and sensation was performed daily during the epidural infusions. If the patient developed a temperature greater than 39 degrees C by any route, the catheter was removed and the tip sent for culture. The insertion site was inspected for inflammation when the catheter was discontinued, if there was malfunction or leaking of the catheter, or if the patient complained of any pain or tenderness in the back. If the patient complained of back pain or tenderness subsequent to removing the catheter, the insertion site was reinspected for signs of inflammation or infection.
Epidural analgesia was used for 1620 patients aged less than or equal to 18 yr. Demographic data are shown in Table 1. Catheters were introduced at the lumbar level 93% of the time, at the thoracic level 3.3% of the time, and at the caudal level 3.7% of the time. Catheters were advanced cephalad to levels appropriate to the dermatomes involved in the surgery. Ninety percent of the patients received epidural analgesia for postoperative pain management. The remaining 10% of patients received epidural analgesia for a variety of indications including cancer, reflex sympathetic dystrophy, sickle cell painful crisis, rib fractures with cystic fibrosis, and trauma.
Since July 1991 the median duration of epidural catheter use has been 2.4 days. The 25th and 75th percentiles are 1 and 3 days, respectively. The overall range was from 1 to 9 days. The longest duration of single epidural catheter use in our hospital during the time of the study was 14 days. Longer term neuraxial catheterization in oncology patients has been primarily via subarachnoid catheters.
Among postoperative patients, none had evidence of either skin infection or epidural abscess. This provides a zero incidence of clinical infection with 95% confidence limits of 0% to 0.03% .
One patient, with osteosarcoma widely metastatic to chest wall, spine, and lungs, had candidal contamination of her epidural space associated with epidural catheter use. A thoracic epidural catheter was placed prior to chest tube placement and pleurodesis with adriamycin for metastatic pulmonary effusions. Bupivacaine 0.1% with fentanyl 2 micro gram/mL was administered for 10 days through the epidural catheter. Despite excellent pain control, the catheter was discontinued at this point because the patient was febrile to 39.5 degrees C and neutropenic. The exit site appeared clean and her back was not tender when the epidural catheter was withdrawn. The tip was sent for microbiologic culture and grew out a few colonies of Candida tropicalis. Over the next 4 days intravenous morphine was administered for pain control. This left her in inconsolable pain despite side effects of ileus, constipation, hypoventilation, hallucinations, and delirium. At this point, maximizing the quality of her remaining life by providing markedly improved pain control without impairment of her sensorium was considered the highest priority. Palliative radiation therapy was begun and a tunneled epidural catheter was placed at T9-10 to achieve this goal despite the risk of infection. A magnetic resonance imaging scan demonstrated a tumor in the epidural space deviating the spinal cord near T11-12. Over the next 2 wk she continued to be febrile with temperatures up to 39.5 degrees C. Her pain gradually worsened, consisting mostly of back and lower extremities pain. Two weeks after placement of the tunneled catheter, she developed an acute motor and sensory block at T2, Magnetic resonance imaging scan showed an epidural fluid collection from T1 to T10. She underwent emergent laminectomies which disclosed large amounts of necrotic epidural tumor and fluid under pressure containing neutrophils and C. tropicalis. After decompressive laminectomy, her pain markedly decreased and her sensory and motor loss resolved. She died 7 wk later, with adequate analgesia from systemic opioid infusions throughout her remaining life.
Three additional patients with chronic pain received evaluations for epidural infection, which were negative. These are summarized in Table 2 and included one other patient with cancer  and two patients with reflex sympathetic dystrophy receiving continuous sympathetic blockade through an epidural catheter . Radiographic and/or nuclear medicine scans ruled out epidural abscess in all three cases.
Epidural analgesia has gained acceptance for short-term use in the postoperative setting, not only because of improved quality of pain relief , but also because several adult [18-22] and pediatric [23-25] studies indicate that in certain high-risk patients there may be decreased morbidity, mortality, and overall cost. It is important to examine whether the risk of epidural catheters, including infection, are justifiably low. Studies in which epidural catheters were cultured upon removal demonstrate that epidural catheters placed in adults are at risk of contamination with skin flora over a few days [26-28], although Desparmet et al.  did not find catheter contamination after periods of only 12-36 h of epidural analgesia in children. None of these reports demonstrated any clinical evidence of infection.
Fifteen patients have previously been reported with epidural abscess associated with postoperative epidural analgesia, all as case reports [8,9,29-40]. Many of the patients had no obvious predisposing factor for epidural abscess. Epidural catheters inserted for pain control after trauma or primarily rib fractures show up in disproportionately large numbers, however (5 of 14 case reports). Unfortunately, case reports do not provide the total number of patients treated with epidural catheters, making it impossible to estimate the risk of postoperative epidural infection.
Our study has the advantage that the total number of patients using epidural catheters since the inception of the Pain Treatment Service is known. Our experience shows that with meticulous insertion technique and a low threshold for removing catheters, the risk of clinically significant infection in the postoperative setting is extremely low: there was a 0% incidence in this setting with 95% confidence limits of 0 to 3 per 10,000 cases. Overall, the number of epidural abscesses reported after short-term use of epidural catheters is comparable to the number of cases reported after single injections of local anesthetic or steroids epidurally, or after lumbar puncture and, in our study, is not statistically different from the spontaneous abscess rate of 0.2 to 1.2 cases per 10,000 hospital admissions . The low risk of infection and excellent quality of pain control associated with this technique supports the use of postoperative epidural analgesia. It is important to realize that epidural abscesses do occur, however infrequently, and that they may be associated with permanent nerve injury if not rapidly detected and treated. Meticulous sterile technique is recommended even for short-term epidural catheterization.
The literature on epidural abscess associated with chronic epidural catheterization contains two types of studies: case reports [41-46] and, of more interest, series assessing the complication rate in a large number of patients [13,14,26,28,47-50]. From the series it is possible to assess the risk associated with long-term catheter placement for chronic pain control. In the seven largest series [13,14,47-50], including the non-postoperative patients from the present study, the combined incidence of infection in the chronic pain patient reaches 2.4% with 95% confidence limits of 1.4%-3.4%. This risk is much lower than that suggested by Sollmann et al. , who quoted a 20% chance of infection with long-term catheter placement.
Series of spontaneous spinal epidural abscesses cite substantial mortality and morbidity including paraplegia often associated with delay in making the diagnosis [7,8,11,12]. Reports of epidural abscesses in the setting of chronic epidural catheterization have shown no mortality or long-term morbidity and the infections were successfully treated without laminectomy using only catheter removal and antibiotics. The authors obtained these results by maintaining a high index of suspicion for this complication [13,14,49]. We have also maintained a high index of suspicion for epidural abscess. Although our patient required laminectomy for decompression of tumor as well as infected fluid, this was performed rapidly enough to allow her complete recovery of neurologic function. For patients in the final stages of a terminal disease, the risk of infection with long-term epidural catheters appears acceptable. Careful monitoring of the catheters and a high index of suspicion for infection are required to prevent serious sequellae in the event of catheter contamination.
We would like to thank Liliana Goumnerova, MD, Zeev Kain, MD, Ariel Mizrahi, MD, Kari Pritchard, and Richard Smith for their help in translating articles from Russian, Hebrew, French, Norwegian and Danish, and Spanish, respectively. Special acknowledgment is also given to Minna Levine for creating DBPain.
1. Desparmet J, Meistelman C, Barre J, Saint-Maurice C. Continuous epidural infusion of bupivacaine for postoperative pain relief in children. Anesthesiology 1987;67:108-10.
2. Ecoffey C, Dubousset AM, Samii K. Lumbar and thoracic epidural anesthesia for urologic and upper abdominal surgery in infants and children. Anesthesiology 1986;65:87-90.
3. Murat I, Delleur MM, Levy J, et al. Continuous epidural anaesthesia for major abdominal surgery in children. Eur J Anaesthesiol 1987;4:327-35.
4. Berde CB, Sethna NF, Conrad LS, et al. Subarachnoid bupivacaine analgesia for seven months for a patient with a spinal cord tumor. Anesthesiology 1990;72:1094-6.
5. Wilder RT, Berde CB, Wolohan M, et al. Reflex sympathetic dystrophy in children. Clinical characteristics and follow-up of seventy patients. J Bone Joint Surg [Am] 1992;74A:910-9.
6. Berde CB. Regional analgesia in the management of chronic pain in childhood. J Pain Symptom Management 1989;4:232-7.
7. Danner RL, Hartman BJ. Update on spinal epidural abscess: 35 cases and review of the literature. Rev Infect Dis 1987;9:265-74.
8. Baker AS, Ojemann RG, Swartz MN, Richardson EJ. Spinal epidural abscess. N Engl J Med 1975;293:463-8.
9. Ericsson M, Algers G, Schliamser SE. Spinal epidural abscesses in adults: review and report of iatrogenic cases. Scand J Infect Dis 1990;22:249-57.
10. Hlavin ML, Kaminski HJ, Ross JS, Ganz E. Spinal epidural abscess: a ten-year perspective. Neurosurgery 1990;27:177-84.
11. Russell NA, Vaughan R, Morley TP. Spinal epidural infection. Can J Neurol Sci 1979;6:325-8.
12. Kaufman DM, Kaplan JG, Litman N. Infectious agents in spinal epidural abscesses. Neurology 1980;30:844-50.
13. Gastmeier K, Gastmeier P. The risk of infection in patients with long-term epidural catheters. Anaesthesiol Reanim 1991;16:267-70.
14. DuPen SL, Peterson DG, Williams A, Bogosian AJ. Infection during chronic epidural catheterization: diagnosis and treatment. Anesthesiology 1990;73:905-9.
15. Statistics for Social Sciences. New York: Holt, Rinehart and Wilton, 1973:379.
16. Berde CB, Sethna NF, Fisher DE, et al. Celiac plexus blockade for a 3-year-old boy with hepatoblastoma and refractory pain. Pediatrics 1990;86:779-81.
17. Youngstrom P, Cowan R, Sutheimer C, Eastwood DYJ. Pain relief and plasma concentrations from epidural and intramuscular morphine in post-cesarian patients. Anesthesiology 1982;57:404-9.
18. Bach S, Noreng JF, Tjellden NU. Phantom limb pain in amputees during the first 12 months following limb amputation after preoperative lumbar epidural blockade. Pain 1988;33:297-301.
19. Kehlet H. Epidural analgesia and the endocrine-metabolic response to surgery--updates and perspectives. Acta Anaesthesiol Scand 1984;28:125-7.
20. Mitchell D, Friedman RJ, Baker J, et al. Prevention of thromboembolic disease following total knee arthroplasty. Epidural versus general anesthesia. Clin Orthop 1991;102-12.
21. Rawal N, Sjostrand U, Christoffersson E, et al. Comparison of intramuscular and epidural morphine for postoperative analgesia in the grossly obese. Anesth Analg 1984;63:583-92.
22. Yeager MP, Glass DD, Neff RK, Brinck-Johnsen T. Epidural anesthesia and analgesia in high risk surgical patients. Anesthesiology 1987;66:729-76.
23. Bosenberg AT, Hadley GP, Murray WB. Postoperative ventilation requirements following esophageal atresia repair. J Pain Symptom Management 1991;6:209.
24. McNeely JM. Comparison of epidural opioids and intravenous opioids in the postoperative management of pediatric antireflux surgery. Anesthesiology 1991;75:A689.
25. Meignier M, Souron R, Lenee J. Postoperative dorsal epidural analgesia in the child with respiratory disabilities. Anesthesiology 1983;59(5):473-5.
26. Andersen BN, Damgaard B, Nielsen FB, Schroder SS. Bacteriological aspects of treatment with epidural morphine. Ugeskr Laeger 1984;146:2700-2.
27. Nickels JH, Poulos JG, Chaouki K. Risks of infection from short-term epidural catheter use. Reg Anesth 1989;14:88-9.
28. Shapiro JM, Bond EL, Garman JK. Use of a chlorhexidine dressing to reduce microbial colonization of epidural catheters. Anesthesiology 1990;73:625-31.
29. Aliaga Font L, Cuervo Argudin J. Epidural abscess: apropos of a case secondary to continuous epidural anesthesia. Rev Esp Anestesiol Reanim 1978;25:519-26.
30. Dawson P, Rosenfeld JV, Murphy MA, Hellyar AG. Epidural abscess associated with postoperative epidural analgesia. Anaesth Intensive Care 1991;19:569-72.
31. Ducrocq JL, Tchaoussoff J, Zarour J, et al. Infectious iatrogenic complications of continuous epidural analgesia. Agressologie 1983;24:179-80.
32. Herlevsen P, Pedersen MB, Engberg M, Sunde N. Epidural abscess caused by an epidural catheter. Ugeskr Laeger 1988;150:1295.
33. Konig HJ, Schleep J, Krahling KH. A case of transverse spinal cord syndrome following contamination of a peridural catheter. Reg Anaesth 1985;8:60-2.
34. McDonogh AJ, Cranney BS. Delayed presentation of an epidural abscess. Anaesth Intensive Care 1984;12:364-5.
35. Mikhailovich VA, Kotomina GL, Garmashev I, Zhdankina NE. Suppurative complication of prolonged peridural anesthesia. Vestn Khir 1974;112:109-10.
36. North JB, Brophy BP. Epidural abscess: a hazard of spinal epidural anaesthesia. Aust N Z J Surg 1979;49:484-5.
37. Saady A. Epidural abscess complicating thoracic epidural analgesia. Anesthesiology 1976;44:244-6.
38. Sonden B. Epidural abscess--a rare but severe complication after epidural anesthesia. Lakartidningen 1991;88:2401-2.
39. Sowter MC, Burgess NA, Woodsford PV, Lewis MH. Delayed presentation of an extradural abscess complicating thoracic extradural analgesia. Br J Anaesth 1992;68:103-5.
40. Wanscher M, Riishede L, Krogh B. Fistula formation following epidural catheter. A case report. Acta Anaesthesiol Scand 1985;29:552-3.
41. Sollmann WP, Gaab MR, Panning B. Lumbar epidural hematoma and spinal abscess following peridural anesthesia. Reg Anaesth 1987;10:121-4.
42. Fine PG, Hare BD, Zahnister JC. Epidural abscess following epidural catheterization in a chronic pain patient: a diagnostic dilemma. Anesthesiology 1988;69:422-4.
43. Knitza R. Subcutaneous abscess following long time continuous epidural anaesthesia. Anaesthesist 1981;30:198-9.
44. Nielsen TH, Husegaard HC, Joensen F. Tunnelled epidural catheter and infection. Ugeskr Laeger 1985;147:1548-9.
45. Pilon RN, Baker AR. Chronic pain control by means of an epidural catheter: report of a case with description of the method. Cancer 1976;37:903-5.
46. Reynolds PC, Hahn MB. Early diagnosis of a spinal epidural abscess. Reg Anesth 1991;16:57-8.
47. Batier C, Qui C, Blanchet P, et al. Iatrogenic infection of neurosurgical implanted equipment (epidural, intrathecal, intracerebroventricular catheters). Evaluation of 3 years of use. Agressologie 1983;24:173-4.
48. Downing JE, Busch EH, Stedman PM. Epidural morphine delivered by a percutaneous epidural catheter for outpatient treatment of cancer pain. Anesth Analg 1988;67:1159-61.
49. Neumann PB, Dahm PI. Epidural catheters and infection. Ugeskr Laeger 1982;144:2654.
50. Strong WE. Epidural abscess associated with epidural catheterization: a rare event? Report of two cases with markedly delayed presentation. Anesthesiology 1991;74:943-6.