Cerebrospinal fluid-cutaneous fistula and pseudomonas meningitis complicating thoracic epidural analgesia

K. T. Abaza and D. G. Bogod*

Department of Anaesthesia, Nottingham City Hospital NHS, Hucknall Road, Nottingham NG5 1PB, UK

*Corresponding author. E-mail: dbogod@ncht.trent.nhs.uk

Accepted for publication: July 16, 2003


    Abstract
 Top
 Abstract
 Introduction
 Case report
 Discussion
 References
 
We report a case of delayed cerebrospinal fluid-cutaneous fistula that developed in a patient following removal of a thoracic epidural catheter used for perioperative analgesia. It was further complicated by the development of bacterial meningitis. Predisposing factors and management of this rare iatrogenic complication are discussed and the literature reviewed for similar reports.

Br J Anaesth 2004; 92: 429–31

Keywords: anaesthetic techniques, regional, epidural; complications, cerebrospinal fluid cutaneous fistula


    Introduction
 Top
 Abstract
 Introduction
 Case report
 Discussion
 References
 
Central neuraxial blocks are commonly used to provide effective perioperative analgesia with a good safety record, and are also administered for obstetric analgesia and relief of chronic pain. However, rare and life-threatening complications resulting in permanent disability continue to be reported in the literature.1 Cerebrospinal fluid (CSF)-cutaneous fistula is a very rare complication of epidural analgesia2 3 and thus specific recommendations for management of such cases are lacking. We describe a case of CSF-cutaneous fistula, which was evident 5 days after removal of a thoracic epidural catheter used for postoperative analgesia, and which was subsequently complicated by bacterial meningitis caused by Pseudomonas aeruginosa.


    Case report
 Top
 Abstract
 Introduction
 Case report
 Discussion
 References
 
A 30-yr-old male was admitted with acute intestinal obstruction and listed for an emergency laparotomy. He had a 10-yr history of Crohn’s disease, and had been treated for the last 6 yr with prednisolone, which had been increased in the 6-week period before admission to 30 mg daily. He had also been receiving azathioprine for the last 10 months. Two years before the present admission, he had undergone a hemicolectomy. An epidural proved ineffective in providing postoperative analgesia.

The anaesthetic plan on this occasion after discussion with the patient was for a combined general anaesthesia and epidural technique. The epidural was performed with the patient awake in the sitting position, at the T9–10 interspace, by the midline approach and using loss of resistance to saline with a 16G Tuohy needle from a disposable epidural pack. The anaesthetist wore sterile gown, gloves and cap, and used an autoclave-sterilized dressing pack. The thoraco-lumbar skin area was prepared with chlorhexidine 0.5% in alcohol 70%. The skin was infiltrated with lidocaine 1% and the epidural space was located easily on the first attempt. The catheter was threaded, a 0.2-µm pore size bacterial filter attached, and transparent polyurethane dressing applied. After negative aspiration, and a negative epidural test dose of 3 ml bupivacaine 0.5%, general anaesthesia was induced and the patient underwent an ileo-rectal anastomosis. Prophylactic cefuroxime 1.5 g and metronidazole 500 mg were given.

Intraoperatively, the epidural was topped-up with 15 ml bupivacaine 0.25% in divided doses and 100 µg fentanyl to provide analgesia. In the recovery room the patient complained of severe pain and the epidural block was found to be inadequate, with a sensory block at T12 and no motor block. A further 18 ml bupivacaine 0.25% were administered incrementally over a 60-min period to achieve a sensory block to T6. A continuous epidural infusion of bupivacaine 0.125%/fentanyl 4 µg/ml was commenced to provide postoperative analgesia and the patient was visited daily by the Acute Pain Service. The Pain Nurse gave two additional top-ups of 5 ml bupivacaine 0.25% on the first day to supplement the infusion, and the infusion rate had to be increased to the highest permissible limit of 12 ml h–1. Subsequently, the epidural provided satisfactory analgesia.

On day 5 the patient’s temperature was noted to be 38.6°C associated with a dry cough and the surgical House Officer suspected a pulmonary infection. A full blood count showed a WBC of 7.7x109 litre–1 and a chest x-ray film was inconclusive. Cefuroxime i.v. and paracetamol orally were commenced. The following day the patient was apyrexial and respiratory symptoms were improving (repeat WBC 4.1x109 litre–1). Antibiotics were continued for 5 days. The epidural catheter was removed intact on day 8. On day 13 the anaesthetist was called to review the patient because of headache and a ‘wet patch’ over the patient’s back. The headache was severe, occipital, and only partially relieved by recumbency. The patient also complained of photophobia and hyperacusis and his temperature again had increased to 38°C. Examination of the back showed clear fluid weeping slowly and continuously from the epidural skin puncture site, which tested positive with a Dextrostix® reagent strip. This was consistent with a CSF leak. Motor function and sensation in the lower limbs were normal. The patient was managed conservatively with bed rest, hydration, and simple analgesics, with some reduction in headache intensity and amount of CSF leak when reviewed the next day. However, on day 15 there was sudden clinical deterioration with severe fronto-occipital non-postural headache, photophobia, somnolence, and pyrexia of 38°C. The leak from the CSF-cutaneous fistula was now diminishing and meningitis was considered likely. Meropenem and gentamicin (microbiologist advice) were started and the patient transferred to the High Dependency Unit. Lumbar puncture demonstrated turbid CSF with a glucose content of 0.1 mmol litre–1, protein 11.5 g litre–1, RBC 610 µl–1, PMN 5175 µl–1, lymphocytes 400 µl–1, but negative Gram stain. However, CSF and skin swab cultures grew P. aeruginosa, which was sensitive to the antibiotics prescribed. MRI scan demonstrated a thin posterior subacute subdural haematoma extending from T3 to T10, with no evidence of cord compression or abscess. Prophylactic low molecular weight heparin, which had been started on day 1, was discontinued. The patient responded to the antibiotics over the next 72 h, and repeat lumbar puncture 1 week later was essentially normal, as was the MRI repeated 2 weeks later. The patient made a complete recovery and was discharged home on day 35.


    Discussion
 Top
 Abstract
 Introduction
 Case report
 Discussion
 References
 
CSF-cutaneous fistula is a very rare but potentially serious and life-threatening complication of epidural anaesthesia if compounded by development of meningitis as demonstrated in this report. A few cases of CSF-cutaneous fistula have been described in anaesthesia practice: following accidental dural puncture during epidural insertion,24 secondary dural penetration by epidural catheter,5 administration of epidural,6 and following continuous spinal anaesthesia.7 None of these cases were complicated by meningitis. Additionally, in these reports, the fistulae presented within 24 h of epidural/spinal catheter removal, as opposed to 5 days later in our case. The reason for this delay is unclear.

CSF-cutaneous fistulas can complicate a variety of other clinical situations, for example spinal surgery,8 and after removal of CSF lumbar drains used in neurosurgery to prevent9 or treat10 cranial CSF leaks.

Apart from the distressing headache, the main risk from CSF leaks is meningitis. Predisposing factors for development of CSF-cutaneous fistulas have been suggested.3 4 The use of epidural steroids may hinder healing along the needle track. Multiple attempts to locate the epidural space using the same needle may result in fibrin deposition along the tract and post-laminectomy scar tissue may increase the risk.7 In addition, in one report,3 the patient was receiving systemic steroids (as was our patient). We speculate that systemic steroids may have delayed healing, thus contributing to fistula formation.

Several approaches have been described for treatment of this rare complication: (i) bed rest in prone position, prophylactic antibiotic, fluid restriction, figure-of-eight suture in the skin puncture site;4 6 (ii) bed rest with slight trendelenburg,11 or lateral position;2 (iii) epidural blood patch;9 (iv) lumbar CSF drainage or surgical closure.8

In our patient, we decided against an epidural blood patch, because of the possible risk of bacterial contamination, especially with him receiving azathioprine and steroids that are known to increase the frequency of infections.12

Reports of bacterial meningitis following uncomplicated epidural analgesia are rare,13 and in a large retrospective review there were no reported cases in 505 000 women receiving obstetric epidural anaesthesia.14 Meningitis is more frequent after spinal anaesthesia, with breach of the dura recognizable as an added risk in combined spinal-epidural analgesia.1 A recent review15 found that serious neuraxial infections were reported predominantly in immunocompromised patients.

In our patient, we could not establish how the dura was breached. One possibility could be an unrecognized dural puncture with the needle at the time of epidural insertion, which was somehow ‘plugged’ by the epidural catheter. Another explanation may be delayed subarachnoid catheter migration from a subdurally placed catheter, which in this case, was left in situ for 7 days. This may have eroded into the subarachnoid space, as described previously by Barnes.16

It is important to identify the causative organisms early in these cases to enable administration of appropriate antibiotics. In a review of cases of iatrogenic meningitis in the last 20 yr, 83% were caused by Gram positive bacteria.17 In our patient, an opportunist pathogen such as Pseudomonas, which is associated with a higher mortality, was suspected in the presence of the immunosuppressed state.

The Acute Pain Service guidelines for epidural catheters at our hospital state that catheters may be left in situ for several days. It may be noteworthy here that the patient in this case strongly requested that the epidural be continued for the extended period of 7 days, having suffered severe pain following his previous surgery when the epidural was ineffective. These guidelines are currently under review to establish whether a time limit should be applied, after which epidural catheters should be removed.

In conclusion, we suggest that CSF-cutaneous fistula in immunosuppressed patients should be managed promptly by complete bed rest, strict asepsis, figure-of-eight skin suture, and probably prophylactic antibiotics. Unusual features of PDPH, such as visual and auditory symptoms as in our case, should raise suspicions of associated serious neurological complications. Early lumbar puncture for CSF microscopy and culture are indicated in CSF fistulas when associated with non-specific features of systemic infection (e.g. pyrexia, raised WBC). We would exercise caution in performing an epidural blood patch in such cases. We also question the safety of leaving short-term un-tunnelled epidural catheters in situ for extended periods postoperatively in immunosuppressed patients, as this may pose an increased risk of infection. If the CSF fistula remains unresolved, neurosurgical referral is required to consider insertion of a lumbar CSF drain or surgical closure.


    References
 Top
 Abstract
 Introduction
 Case report
 Discussion
 References
 
1 De Tommaso O, Caporuscio A, Tagariello V. Neurological complications following central neuraxial blocks: are there predictive factors? Eur J Anaesthesiol 2002; 19: 705–16[ISI][Medline]

2 Jawalekar SR, Marx GF. Cutaneous cerebrospinal fluid leakage following attempted extradural block. Anesthesiology 1981; 54: 348–9[ISI][Medline]

3 Howes J, Lenz R. Cerebrospinal fluid cutaneous fistula. An unusual complication of epidural anaesthesia. Anaesthesia 1994; 49: 221–2[ISI][Medline]

4 Longmire S, Joyce TH 3rd. Treatment of a duro-cutaneous fistula secondary to attempted epidural anesthesia with an epidural autologous blood patch. Anesthesiology 1984; 60: 63–4[ISI][Medline]

5 Motsch J, Hutschenreuter K. Cutaneous cerebrospinal fluid fistula associated with secondary puncture of the dura caused by a peridural catheter. Reg Anaesth 1984; 7: 74–6[Medline]

6 Ball CG, D’Alessandro FT, Rosenthal J, Duff TA. Case history number 86: an unusual complication of lumbar puncture: a CSF cutaneous fistula. Anesth Analg 1975; 54: 691–4[ISI][Medline]

7 Hullander M, Leivers D. Spinal cutaneous fistula following continuous spinal anesthesia. Anesthesiology 1992; 76: 139–40[ISI][Medline]

8 Maycock NF, van Essen J, Pfitzner J. Post-laminectomy cerebrospinal fluid fistula treated with epidural blood patch. Spine 1994; 19: 2223–5[ISI][Medline]

9 Katz J. Treatment of a subarachnoid-cutaneous fistula with an epidural blood patch. Anesthesiology 1984; 60: 603–4[ISI][Medline]

10 Kumar V, Maves T, Barcellos W. Epidural blood patch for treatment of subarachnoid fistula in children. Anaesthesia 1991; 46: 117–18[ISI][Medline]

11 Lieberman LM, Tourtellotte WW, Newkirk TA. Prolonged post-lumbar puncture cerebrospinal fluid leakage from lumbar subarachnoid space demonstrated by radioisotope myelography. Neurology 1971; 21: 925–9[ISI][Medline]

12 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[Abstract]

13 Ready LB, Helfer D. Bacterial meningitis in parturients after epidural anesthesia. Anesthesiology 1989; 71: 988–90[ISI][Medline]

14 Scott DB, Hibbard BM. Serious non-fatal complications associated with extradural block in obstetric practice. Br J Anaesth 1990; 64: 537–41[Abstract]

15 Wheatley RG, Schug SA, Watson D. Safety and efficacy of postoperative epidural analgesia. Br J Anaesth 2001; 87: 47–61[Abstract/Free Full Text]

16 Barnes RK. Delayed subarachnoid migration of an epidural catheter. Anaesth Intens Care 1990; 18: 564–6[ISI][Medline]

17 Yaniv LG, Potasman I. Iatrogenic meningitis: an increasing role for resistant viridans streptococci? Case report and review of the last 20 years. Scand J Infect Dis 2000; 32: 693–6[CrossRef][ISI][Medline]