Complications of Intrathecal Drug Delivery Systems

Saadat Kamran MD, Ballard D. Wright MD

Institutional / academic affiliation:

Department of Neurology, Anesthesiology University of Kentucky and The Pain Treatment Center Lexington, KY

Corresponding author, address:

Saadat Kamran,

MD280 Pasadena Drive,

Lexington, KY 40503,

USAOFFICE: 606-278-1316; FAX: 606-276-3846


<< Previous Page


Objectives. To report complications associated with implant of the intrathecal drug delivery systems (IDDS).

Patients and Methods. A retrospective review of the implant database was undertaken to determine various complications associated with the IDDS.  A total of 122 patients were reviewed, 97 included in the final analysis, 25 excluded due to incomplete data. The complications of excluded patients were reviewed separately. The complications were divided in to categories pharmacological, equipment, procedural, programming errors and psychological. The drugs used were morphine, dilaudid, fentanyl, sufentanil, baclofen and marcaine.

Results. There were 60 men and 37 women. A total of 43 patients reported various complications. Most common were transient pharmacological side effects (n=33), procedural (n=5), equipment (n=16), programming (n=2) and psychological (n=3). Most serious procedural complication was Staphylococcus Aureus meningitis (n=1). Catheter disconnection and leakage was noted in (n=6) patients. Two programming errors were related to increase in drug concentration with failure to reduce the dose. Seven pumps were explanted including three due to distorted body image.

Conclusion. IDDS are effective and safe devices for pain management. The complications associated with implants are mostly pharmacological and transient. Careful attention to the implant technique is required to minimize the complications.


Key words: Intrathecal drug delivery system, implant, chronic pain, complications


Although systemic opioids are effective in non-malignant pain management but they are associated with significant side effects (1). The majorities of adverse drug reactions are dose-related and occur in patients taking standard doses of medications. These facts suggest that for many patients, standard drug doses may be excessive. Moreover, oral medications may not be effective in controlling pain. The efficacy and cost effectiveness of intrathecal medications for the treatment of chronic pain is well documented (2-5). The major advantages of implantable, programmable intrathecal drug delivery systems include targeted local delivery of drugs at a constant or variable rate, less drug requirement, minimization of possible side effects, and enhanced efficacy of treatment. As with any invasive procedure there are complications and side effects associated with intrathecal drug delivery systems (6-8). The purpose of this study was to share our experience of complications in a large cohort of patients.

Patients and Methods.

We retrospectively reviewed our pump implant database for complications related to the implantation of intrathecal drug delivery system. All implants were performed by one physician and data was collected by a another physician and a nurse. A physician not related to implants reviewed the data base. Patients with incomplete data, loss to follow up were excluded from final review, but patients with major complications are reported separately. The complications were divided in to following categories

1.      Pharmacological

2.      Procedural (Patients received intravenous antibiotic prophylaxis prior to the procedure and completed a seven-day course of oral antibiotic post implant.)

3.      Equipment (All equipment related complications that involved spinal catheter were confirmed by a injecting non-ionic contrast through the pump port.)

4.      Programming errors

5.      Psychological


The pump implant database had one hundred and twenty-two (n=122) patients. Twenty-five (n=25) patients were excluded, loss to follow up/transfer of care to another area (n=5), pump implanted at other facilities/operative data not available (n=12) and pump implants prior to 1991 (data base incomplete) (n=8). Data on the excluded patients was reviewed separately. Ninety-seven (n=97) patients from 1992 to 1999 were included in the study. There were sixty (n=60) men and thirty-seven (n=37) women with age range of 28 to 85 years. Average follow up 5.8 years with range of 6 months to 9 years.  

Indications for implant in patients with pump complications included, failed surgical back (n=15), cervical or lumber spondylosis (n=10), complex regional pain syndrome (n=5), compression fractures (n=8), non-operative disc herniation with radiculopathies (n=4), peripheral neuropathy (n=6) and spasticity (n=2). Fourteen (n=14) patients had more than one diagnosis. The following intrathecal drugs were used in patients with complications, morphine (n=10), morphine and marcaine (n=19), fentanyl and marcaine (n=2), dilaudid (n=3), dilaudid and marcaine (n=5), sufentanil (n=2) and baclofen (n=2).

Forty-three (n=43; 44.3%) patients reported various complications. A detail is provided in table 1. The transient pharmacological complications were common at the beginning of the treatment and resolved rapidly. Chronic pharmacological complications were more common with morphine and dilaudid where as acute complications of pruritis were more common with fentanyl. Pharmacological complications were not seen with baclofen. All patients with sexual side effects were men. There was no correlation between drug dose and chronic complications.

In spite of prophylactic antibiotics five patients developed infection, four (n=4) of these patients had diabetes.  Catheter complications were the most common equipment related complications. All mechanical catheter complications were confirmed by injecting non-ionic contrast through the pump port (pump myelogram).

In patients with programming errors the mistake was detected within minutes of programming in one and did not receive any excessive medication. The second patient reported increasing nausea and sweating, programming error was detected one month later. This patient received twice the dose of medication. Both programming errors involved increasing pump medication concentration without decreasing the rate of flow.

In the psychological category three women with average age of 32 years, reported problems with distorted body image due to the pump implant. These patients underwent explant, with average implant duration of 1.34 years.Seven pumps were expalnted due to, abdominal discomfort (n=1; 1.03%), chronic wound infection (n=1; 1.03) and meningitis (n=1; 1.03%), pump rotation/torsion with port occlusion (n=1; 1.03%) and distorted body image (n=3; 3.1%).Incomplete pump data was available in twenty-five patients. In thirteen (n=13) patients implanted at our facility the following complications were noted; superficial wound infection (n=1), catheter retraction and paraspinal coiling (n=1). Twelve (n=12) patients were referred from other facilities with various complications and pump management. The major complications were pump torsion (n=1), pump torsion with catheter coiling in the pump pocket (n=1), catheter shearing with subarachnoid fragment (n=1), paraspinal catheter disconnection (n=1) and chronic pump pocket infection with meningitis (n=1). Two patients with major complications underwent pump explant.


A large number of small series regarding IDDS have been reported, there are only few large studies describing the complications of intrathecal drug delivery systems (4,9-11). We report one of the largest single implanter/facility series of complications associated with the intrathecal drug delivery system implants.  As more patients will be implanted, physicians have to become familiar with the potential complications associated with the procedure. The purpose of this study was to report a complete spectrum of complications encountered at our facility. Although serious complications are not common, they can be life threatening requiring explant of the equipment and hospitalization.The pharmacological complications were most common but generally they were self-limiting. The rate of pharmacological complications in current series was comparable to those reported in the literature (4,9,10). We did not observe any serious side effects reported in literature such as respiratory depression, seizures, myoclonus, withdrawal, cranial nerve palsies and subdural hematoma (12-18). The principle of variability among individuals, which requires tailoring the dose to the patient, should be followed to reduce the incidence of side effects.The incidence of infections related to implants varies widely in different reports from 0.5% to 9% (4,9,10,16,).  Most of the infections in our patients were superficial wound infections seen in diabetic patients inspite of prophylactic antibiotics. Only one patient developed chronic pump pocket infection that was unresponsive to antibiotics leading to explant. The most feared infectious complication is meningitis. 

We encountered only one case of meningitis, the offending organism was Staphylococcus epidermidis, treated with intravenous antibiotics and explant (6,10,16,19). Another patient transferred to us had developed Staphylococcus epidermidis meningitis from pump pocket infection that led to explant with uneventful recovery. The majority of equipment complications involve catheters. The catheter related complications were the most common cause of repeat surgery in our series. The incidence of operative catheter revision has been reported to be 7% to 34.6% (4,8,10, 21,22).

In our experience catheter shearing was more common with midline approach. With paramedian approach catheter shearing was not seen. Moreover, the use of non-absorbable sutures (purse string suture at the catheter entry site) was found to be a contributing factor since it cuts through the tissue and catheter. We have been able to avoid this complication by using absorbable suture. In patients with a free floating fragment for over two years, no neurologic complications have been seen to date. Interestingly spinal fluid in one patient with free fragment showed mild protein elevation with normal cell count and cultures.

The incidence of catheter dislodgment varies in the literature (10). In spite of catheter suturing to the underlying fascia and the use of anchoring device, retraction with paraspinal coiling and leakage at the connection site remains a problem. The leakage at pump and catheter connection site was encountered in one patient. We believe it was the result of over tightening of the suture with slippage of catheter site from pump site that was recognized later when a seroma developed. The catheter was later reconnected successfully with resolution of the seroma. The catheter related complications have decreased with improved catheters (21).Although there are reports of spinal cord compression and distal catheter occlusion by granuloma (23-25), only one patient with distal catheter occlusion was found on pump myelogram. The patient complained of loss of pain control with normal neurological examination. During pump myelogram initial catheter occlusion resolved when more dye was injected with later return of good pain control. We presume this was the result of a small granuloma at the catheter tip. The CSF leakage with formation of seroma and spinal headaches has been attributed to large needle bore, multiple punctures, inadequate sealing of tissue around catheter, micro lesions of the catheter from guide wire, leakage during the implant, puncture at the mobile spinal segments (L4-5) etc (6,9,26-28). With adequate tissue sealing around the catheter the incidence of spinal headaches has declined significantly (26). The formation of seroma is not a frequent complication and usually resolves with conservative management (6,10). In the current series only one case of abdominal seroma was noted. 

All spinal headaches required pharmacological management and rest (28).Complications associated with pump itself were rare in our series. One patient had mechanical pump failure, not battery related, after two years of implant. Errors in pump programming have been reported (6,29-31). Two pumps were incorrectly programmed and involved mistakes in decreasing the flow rate after increasing the medication concentration. Pump refills and programming should only be done by trained and experience person. The implanting physician should be knowledgeable in anticipating and treating such complications (33,34). Three pumps had to be explanted in young women who while reporting good pain relief could not tolerate the pump due to distorted body image. These patients had undergone thorough psychological evaluation and education program that included showing the original equipment, videotape of implant procedure, and physician/implantar's satisfaction with patient selection. These complications can be reduced by meticulous attention to technique and better understanding of the intrathecal drugs.




1.Levy M. (1994).Pharmacological management of cancer pain. Semin Oncol. 21:718-728.

2.Winkelmuller W, Burchiel K, Van Buyten J. (1999)  Intrathecal opioid therapy for pain:Efficacy and outcomes. Neuromodulation. 2(2):67-76.

3. Krames ES, lanning RM. (1993) Intrathecal infusional analgesia for nonmalignant pain:  analgesic efficacy of intrathecal opioids with or without bupivacaine. J Pain Symptom Manage .8:539-548.

4. Winkelmuller M, Winkelmuller W. (1996)long term effect of continuous opioid treatment      in chronic pain of non malignant etiology. J Neurosurg. 85(3):458-467.

5Mueller-Schwefe G, hassenbusch SJ, Reig E. (1999) Cost effectiveness of intrathecal therapy for pain. Neuromodulation. 2(2):76-84.

6.Krames ES. (1993) Intrathecal infusional therapies for intractable pain:patient management guidelines. J Pain Symptom Managm. 8(1):36-46.6.      Chaney MA. (1995) Side effects of intrathecal and epidural opioids. Can J Anes. 42:891-903.

7.      Naumann C Edrine S, Koulousakis A, Schuchard M. (1999) Drug adverse events and systemic complications of intrathecal opioid delivery for pain: origins, detection, manifestations and management. Neuromodulation. 2(2):92-107.

8.      Paice J, Penn RD, Shott S. (1991) Intraspinal morphine for chronic pain: a retrospective, multicenter study. J Pain Symptom Manage. 11:71-80.( Erdine S, Aldemir T. Long term results of peridural morphine in 225 pain patients. Pain. 45:155-159.)

9.       Erdine S, Oztakcin S, Yucel A. (1996) Intrathecal morphine delivered by implantable manual pump for cancer pain. Pain Digest. 6:161-165

10.  Tutak U, Doleys DM. (1996) Intrathecal infusion systems for treatment of chronic low back and leg pain of non cancer origin. So Med J. 89:295-300

11.  Velarde CA, Zuniga RE, Leon RF, Abram SE. (2000) Cranial Nerve palsies and subdural hematoma following implantation of intrathecal drug delivery device. Reg Anesth Pain Med 25(1):76-78.

12.  Devulder J, Bohyn P, Castille F, DeLaat M, Rolly G. (1996)A case of uncommon withdrawl symptoms after a short period of spinal morphine administration. Pain. 64(3):589-591

13.  Kronenberg MF, Laimer I, Saltauri L, Bramanti P. (1998) Epileptic seizures associated with intracerebroventricular and intrathecal morphine bolus. Pain. 75(2):383-387

14.  Deconno F, Caraceni A, Martini C, Spoldi E. (1991)Hyperalgesia and myoclonus with intrathecal infusion of high dose morphine. Pain. 47(3):337-339

15.  Follet KA, Hitchon PW, Piper J. (1992) Response of intractable pain to continuos intrathecal morphine: a retrospective study . Pain. 49(1):21-25

16.  Foley KM. (1982) Practical use of narcotic analgesics. Med Clin N Am. 66:1091-1104

17.  Coombs DW, Saunders RL, Gaylor MS. (1983) Relief of continuous chronic pain by intraspinal narcotic infusion via an implantable reservoir. JAMA. 250:2336-2339

18.  Byers K, Axelord P, michael S. (1995) Infections complicating tunneled intraspinal catheter systems used to treat chronic pain. Clin Infect Dis.21(2):403-408

19.  Muller H, Luben V, Zierski J. (1988) Long term spinal opiate treatment . Acta Anaesth Belg. 39(suppl 2): 83-86

20.  Sjoberg M, Karlsson PA, Nordborg C. (1992) Neuropathological findings after long term intrathecal infusion of morphine and bupivacaine for pain treatment in cancer patients. Anesthesiology. 76:173-186.

21.Penn RD.(1996) Catheter implant systems for intrathecal drug delivery. J Neurosurg. 84:713.22.Hirsh LF, Thanki A, Nowak T. (1985) Sudden loss of pain control with morphine pump due to catheter migration. Neurosurgery. 17(6):965-96723.Cabbell KL, Taren JA, Sagher O. (1998) Spinal cord compression by catheter granulomas in high dose intrathecal morphine therapy: case report. Neurosurgery. 42(5):1176-1180

24.Langsman A. (1999)A case of spinal cord compression syndrome by a fibrotic mass presenting in a patient with an intrathecal pain management pump system. Pain. 83(1):97-99

25.Bejjani GK, Karim NO, Tzortzidis F. (1997) Intrathecal granuloma after implantation of a morphine pump:case report and review of the literature. Surg Neurol. 48(3):288-29126.Koulousakis A, Imdahl M, Weber M. (1998) Continous intrathecal  aplication of morphine in cancer pain. Proceedings of the 8th world congress.

27. Gestin Y, Vaino A, Pegurier AM. (1997) Long term intrathecal infusion of morphine in the home care of patients with advanced cancer. Acta Anaesthesiol Scand. 41(1):12-17 28. Choi A, Laurito CE, Cunningham FE. (1996) Pharmacological management of post dural puncture headaches. Ann Pharmacotherapy. 30:831-839.

29.Wu CL, Patt RB. (1992)  Accidental overdose of systemic morphine during intended refill of intrathecal infusion device. Anesth Analg. 75(1):130-132

30.Day FJ. (1993) Accidental overdose of systemic morphine during intended refill of intrathecal infusion device. Anesth Analg. 76(1):203

31. Belmans L, Van Buyten JP, Vanduffel L. (1997) Accidental overdosing with intraspinal morphine caused by misprogramming of a SynchroMed pump: a report of two cases. Acta Anaesthesiol Belg. 48(2):93-97.

32.Sauter K, Kaufman HH, Bloomfeild SM. (1994) Treatment of high dose intrathecal morphine overdose. J Neurosurg. 81:143-146.

33. Kaiser KG, Bainton CR. (1987) Treatment of intrathecal morphine overdose by aspiration of cerebrospinal fluid. Anesth Analg. 66(5):475-477.

Table 1.Complications associated with intrathecal drug delivery system

 Pharmacological (n=60; 61.8%)

Acute (immediate post transplant to 4 weeks) (n=33; 34%)

  • nausea (n=12; 12.3%)
  • urinary hesitancy (n=9; 9.2%)
  • pruritis (n=12; 12.3%).

Chronic (persistent for over 4 weeks) (n=27; 27.8%)

  • decreased libido, erectile dysfunction (n=6;6.1%)        
  • constipation (n=16; 16.5%)
  • peripheral edema (n=5; 5.1%)

Procedural (n=5; 5.1%),

                Minor infections

                superficial wound infection (n=3; 3.1%)

                Serious infections

  • meningitis (n=1; 1.03%)
  • chronic pump pocket infection (n=1; 1.03)


            Catheter related (n=16; 16.5%)

  • distal catheter occlusion (n=1; 1.03%)
  • shearing at spinal entry site (n=2; 2.06%)
  • shearing with subarachnoid segment (n=2; 2.06%) retraction, paraspinal coiling/kinks (n=4; 4.12%)
  •  leakage at connection site between catheters (n=6; 6.2%)
  •  spinal headache (n=3;3.09%)
  • CSF leakage at the catheter and pump connection site with a small seroma (n=1; 1.03%).

Pump related

  • Pump failure (n=1; 1.03%)
  • Pump torsion with port occlusion (n=1;1.03%)

Programming errors

             Pump mis-programming (n=2; 2.06%)

 Psychological category

distorted body image (n=3; 3%)

First Published November 2000

Home • Journals • Search • Rules for Authors • Submit a Paper • Sponsor us
Rules for Authors
Submit a Paper
Sponsor Us

Google Search

Advanced Search




Default text | Increase text size