Chapter 71. Neuraxial Drug Delivery

The use of neuraxial analgesia for the goal of functional analgesia is the next step beyond the optimum use of oral and transcutaneous analgesia. The goal of pain relief with functional restitution should be the guiding light of all approaches of analgesia. Once a patient with acute, chronic, or cancer-related pain has been prescribed a trial of opioids and adjuvant drugs, side effects having been either treated or avoided with opioid sequential trials, then consideration should be given to alternative delivery sources. Neuraxial analgesia, with either single agent or a combination of drugs, may allow the patient to achieve relief of intractable pain when opioid analgesia alone has its limits.

The discovery of the analgesic effects of neuraxial opioids, alkaloids, and peptides has led to the expanded use of regional anesthesia in long-term analgesic delivery systems. Today the choices of analgesic agents include opioids, α2-agonists, SNX-111, local anesthetic agents, and there are sure to be additional agents added in the future. The clinician will always be faced with a decision of when, what, and how. Patient selection, device selection, and route of delivery are and will likely continue to be the key questions.

In this chapter we approach the patient in pain with options for neuraxial analgesia. This includes discussions of drug choices and pharmacology, patient selection, and optimization of conservative therapy. Issues of delivery routes and device selection are explored, and lastly, we give the reader insights into identification and treatment of side effects and complications.

Spinal opioids have been used in clinical practice routinely since the late 70s. Epidural analgesia for postsurgical, obstetric, and intractable cancer-related pain is considered standard medical care. Intrathecal opioid therapy for chronic noncancer pain has become an increasingly routine therapeutic approach over the last 10 years, although reimbursement issues continue to be problematic in some states with some payers.

Perhaps the most exciting new developments are in the area of alternative agents that are being, or may soon be, used in spinal drug delivery systems. The escalation of basic science research exploring new spinally active agents has been brisk in the last decade, and with newly committed funding for pain research at the National Institutes of Health, the next decade will surely bring an explosion of research and discovery in neuroaxial pharmacology.

There continue to be major advances in systemic opioid management, particularly the advent of long-acting oxycodone, transdermal and transmucosal fentanyl, and the soon to be released long-acting hydromorphone. These advances, and the controversy surrounding the potential link between M3G and neurotoxicity,1 have left expert clinicians debating if morphine is still the standard drug of choice or just one of many choices for systemic therapy. There is not much debate, however, over morphine’s superiority when administered spinally, particularly for chronic cancer and noncancer administration.

Morphine

It seems clear that the potency of spinally administered opioids is inversely related to its lipophilicity. Spinal concentrations of opioid after epidural administration are a result of the balance between vascular and meningeal permeability, which is influenced by lipid solubility. Although epidural morphine, with its hydrophilic profile, is up to 10 times more potent that intravenous morphine, many lipophilic agents are essentially equipotent when given either intravenously or epidurally.2,3 In the arena of postoperative pain management, the epidural administration of lipophilic opioid alone may offer no marked clinical advantages compared with the intravenous route.4

The clearest advantage with spinally administered morphine is reduction of side effects associated with systemic blood levels. There is predictable vascular uptake with epidural morphine, but very little systemic impact with intrathecal morphine (Fig. 71-1). In the patient with opioid-responsive pain for whom systemic morphine provides good to excellent relief but causes unmanageable systemic side effects, epidural or intrathecal morphine will likely provide a good outcome. The hydrophilic properties of morphine allow for a long duration of action that makes it amenable to bolus dosing. These authors have taught hundreds of cancer patients to self-administer epidural morphine boluses on a scheduled basis at home. The average opioid-tolerant cancer patient will require 8 mg of epidural morphine on an every 8-hour schedule. This is generally a cost-effective way to provide epidural analgesia; however, it is only likely to succeed in those patients with nociceptive pain (i.e., no neuropathic or mixed pain character).

Bolus epidural morphine dosing can be problematic. Postoperatively, in the opioid naïve patient, the side effect that has the most potential to produce adverse outcomes with epidural morphine is delayed respiratory depression. Studies indicate that morphine’s cephalad migration after a lumbar epidural bolus follows a predictable time course to the brainstem, where respiratory depression can occur.5,6 Additionally, several authors make a good case that continuous infusion of morphine may provide better analgesia for postoperative pain management.7–9 At least one study, however, indicated that there were significantly more dose escalations in patients treated with continuous infusion of epidural morphine versus bolus administration, leaving open the question of a differential effect for mode of administration on tolerance development.10 Today, in managing chronic cancer or noncancer pain, this issue is rarely relevant. Patients with nociceptive pain can almost always be managed with less invasive therapy. Most advanced cancer patients requiring spinal therapy also need either local anesthetics or clonidine coadministered, which requires infusion, and most noncancer patients will likely be treated with implanted pump systems that are generally continuous infusion (intermittent dosing is possible with the computerized versions of the implanted pump systems).11

The choice of epidural versus intrathecal drug delivery is framed primarily around diagnosis, prognosis, severity of side effects, and cost-versus-benefit markers. Morphine is used widely both epidurally and intrathecally with internalized and externalized approaches. In the United States postoperative and terminal cancer patients have generally had epidural approaches and patients with noncancer pain syndromes have generally had implanted morphine pump systems. Over the past 5 years the European anesthesia community has increasingly published papers on the efficacy and safety of percutaneous intrathecal morphine infusions for terminal cancer patients.12–14

Morphine is the only Food and Drug Administration (FDA)-approved opioid for intrathecal delivery, and it is used widely in the long-term treatment of noncancer pain syndromes. A retrospective, multicenter, survey-based study reported by Paice et al examined outcomes for intrathecal morphine therapy in 429 patients (roughly two thirds noncancer and one third cancer).15 For all patients the mean percent pain relief was 61%. Patients with somatic pain syndromes tend to have greater relief than those with neuropathic pain syndromes. Roughly 20% of the patients had bupivacaine added to the morphine. Anderson and Burchiel reported prospectively on 40 patients with severe noncancer pain that had been poorly managed on systemic medications.11 Ten of the 40 patients failed a trial for intrathecal morphine therapy. Thirty patients had reported pain relief and were implanted with intrathecal delivery devices. Two years later, one half of the patients were still reporting at least a 25% reduction in pain as assessed by the visual analog scale (VAS).

There has been some concern about the neurologic sequelae related to high-dose morphine in the spinal canal. Cabbell and colleagues reported three cases of intrathecal granuloma formation at the spinal catheter tip, which they thought were associated with high-dose morphine (25 mg/day and up).16 North et al previously described a case of a granulomatous mass at the tip of an intrathecal catheter causing spinal cord compression.17 A recent survey, however, indicated that the overall incidence of this phenomenon was less than 1%.18

Alternate Opioids

Hydromorphone is used frequently as a second-line drug intraspinally, although it is not FDA approved for spinal use. Hydromorphone’s lipid solubility is intermediate between morphine and fentanyl.19 The rostral spread of hydromorphone in the cerebrospinal fluid (CSF) is similar to morphine after a bolus dose in the lumbar epidural space; it has a comparatively faster onset of action and shorter duration of action than morphine.20 Halpern and associates reported no difference between morphine and hydromorphone in pain relief or in the incidence and severity of side effects.21 Chaplan et al found the quality of analgesia with epidural morphine versus hydromorphone to be similar, but with a four-fold increase in pruritis in the morphine group.22 Liu and coworkers studied intravenous versus epidural hydromorphone in radical prostatectomy and noted that patients in the intravenous group required twice as much drug, whereas patients in the epidural group had a greater incidence of pruritus.23 Epidural administration did not appear to be associated with improved analgesia, patient satisfaction, or enhanced clinical outcomes when compared with the intravenous route.

Fentanyl has been used as an alternate opioid. Multiple studies have seemed to indicate no benefit for postoperative pain management for epidural versus intravenous fentanyl.3,24,25 Epidural fentanyl has been associated with less nausea and pruritus than morphine.26,27 Sufentanil is a highly lipophilic opioid with a high level of intrinsic efficacy that may require a relatively low-receptor occupancy. De Leon-Casasola and Lema make a strong argument for sufentanil’s utility in the setting of morphine-tolerant cancer patients with severe postoperative pain.8,28

Local Anesthetics

Opioid-local anesthetic combinations are used widely in postoperative and cancer-related pain management, and to a lesser extent in chronic noncancer pain management. In animal studies local anesthetics and opioids have demonstrated a synergistic potentiation of antinociceptive effects.29 Bupivacaine is the most commonly used local anesthetic in pain management, owing primarily to its long duration of action. Literally thousands of patients have successfully been treated with epidural morphine-bupivacaine combination therapy for postoperative pain. An excellent large prospective study done by de Leon-Casasola and associates reported on over 4000 patients who had received epidural morphine-bupivacaine therapy.30 The average length of therapy was 6 days, with nausea and pruritus each occurring in 22% of patients. Respiratory depression was reported in only three patients (0.07%).

Fentanyl-bupivacaine infusions are also used extensively in acute pain management and generally may have some side effect advantages over morphine-bupivacaine.26,27,31 The often-expected goal of reducing opioid when the bupivacaine is added may not always be achievable or may be dose-dependent. One study examined 0.1% bupivacaine as an “add-on” to epidural fentanyl in 40 patients after major abdominal surgery. The epidural fentanyl infusion of 10 μg/mL (with or without 0.1% bupivacaine) was titrated to adequate pain relief during forced inspiration. Patients reported similar median pain scores and were equally satisfied with pain relief in both groups, and the fentanyl infusion rate and plasma concentration were comparable. Respiratory and cardiovascular functions were preserved, and the incidence of nausea, pruritus, and drowsiness were similar. The investigators concluded that in low concentrations bupivacaine did not reduce the titrated dose of epidural fentanyl required for adequate pain relief during forced inspiration.32 Sufentanil and bupivacaine are also used in combination and have been reported to provide effective analgesia during labor at reduced doses compared with either single drug alone.33

Epidural morphine-bupivacaine combinations are also the mainstay of cancer-related pain. Epidural bupivacaine infusions administered long term have been safely administered even with plasma bupivacaine concentrations as high as 10.8 μg/mL.34 A combination of low-dose epidural morphine (0.5 to 1.5 mg/hour) and low-concentration bupivacaine (0.1% to 0.125%) in a 5- to 10-mL/hour volume effectively targets the most commonly seen intractable cancer pain syndromes: breast or lung carcinoma that has metastasized to the spine and is causing a severe mixed pain character syndrome. The severity of motion-related pain in these patients often causes the window between pain relief with function and side effects to be exceedingly narrow.35 The disadvantages of using epidural bupivacaine results primarily from the sensorimotor blockade. At low to moderate concentrations of epidural bupivacaine, patients often experience cutaneous numbness in the expected dermatomal distribution. This numbness can be transient, and is often positional. Many patients that sleep on one side will awaken with “downside” numbness that will resolve with repositioning. In patients who require high concentrations of epidural bupivacaine (≥0.16%), significant sensory loss and motor dysfunction can occur. Orthostatic blood pressure can also be significant, particularly in the end-stage dehydrated cancer population.

Bupivacaine-containing solutions are also used successfully with morphine via the intrathecal route in patients with cancer and chronic noncancer pain syndromes.35–37 Intrathecally administered bupivacaine is also limited by sensorimotor blockade or hemodynamic instability.38,39 Clinically relevant side effects usually are not observed with doses <15 mg per day.40

Ropivacaine has emerged as a potential alternate to bupivacaine for pain management. Ropivacaine has sensory block advantages, providing a greater degree of dissociation between sensory and motor effects.41 This agent may prove to be helpful in combination with epidural opioids in cancer pain management where the balance between pain relief and functionality can be narrow, particularly in patients with spinal nerve impingement from vertebral metastasis. Ropivacaine has also been used intrathecally in cancer pain management.42

Pharmacy Admixture Issues

The mixing of agents together, particularly for long-term administration, raises issues of stability, compatibility, and infection control. A number of investigators have reported on laboratory experiments. Christen and coworkers looked at bupivacaine and hydromorphone.43 Bupivacaine hydrochloride 625 and 1250 μg/mL and hydromorphone hydrochloride 20 and 100 μg/mL were mixed in 0.9% sodium chloride injection and found to be stable and compatible for up to 72 hours under fluorescent light when stored in polyvinyl chloride containers at 24°C. The physical and chemical properties of multiple drugs in solutions have also been examined.44 Bupivacaine, morphine, and clonidine were stored in reservoir bags for up to 90 days. The bags contained (1) bupivacaine 0.75% alone, (2) morphine 2% alone, and (3) morphine 6.66 mg/mL plus bupivacaine 3 mg/mL plus clonidine 30 μg/mL. No macroscopic or microbiologic signs of precipitation, change in color, or contamination were observed, and pH remained stable. None of the three drugs declined in concentration during the observation period. A small increase in concentration of all three drugs did occur over time, most probably due to evaporation processes (Fig. 71-2). Cook et al examined six mixtures of diamorphine-bupivacaine, representing the range of clinically useful concentrations.45 Solutions were studied for antibacterial activity against common contaminants: Escherichia coli, Pseudomonas aeruginosa, Enterococcus faecalis, and a coagulase-negative Staphylococcus species. Challenge experiments used an inoculum or approximately 5.0 × 106 cfu/mL. Mixtures were incubated at 30°C for 5 to 7 days. Viable counts of all organisms decreased with time for all the formulations tested. Formulations containing 0.5% bupivacaine were rapidly bactericidal, and increasing diamorphine concentrations increased this effect. The authors argue against frequent infusion changes. Another group of investigators combined diamorphine and bupivacaine in 100-mL bags for use with a patient-controlled analgesia pump and found no significant change in concentration of either drug over a period of 8 days.46 Several other studies on long-term stability have been done on morphine-bupivacaine and fentanyl-bupivacaine mixtures.47–50

Other Spinally Active Agents

α-Adrenergic Drugs

Clonidine is an α-adrenergic agonist that has been used for spinal administration in Europe for many years, but only received FDA approval in the United States in 1997.51 Spinal clonidine has been shown to demonstrate a dose-dependent duration of analgesia in the acute setting for more than 12 hours with a dose of 450 μg.52 There is evidence that clonidine has synergistic properties with both opioids and local anesthetics. Brunschwiler and colleagues reported that clonidine 150 μg produced a “local anesthetic sparing effect,” prolonging continuous spinal anesthesia with bupivacaine significantly longer than morphine 150 μg.53 Patients receiving 30 μg of intrathecal clonidine with 2.5 or 5 μg of sufentanil had significantly longer lasting analgesia.54

Clonidine has been shown to effectively impact neuropathic pain in animal models.55,56 This is perhaps its greatest potential. Eisenach and associates did a multicenter study looking at cancer-related pain. Eight-five cancer patients were treated with patient-controlled epidural morphine with placebo or epidural clonidine (30 μg/hour). Neuropathic pain was significantly reduced in patients who received clonidine, although the frequency or quantity of morphine was not influenced. Hypotension and bradycardia were the most frequently observed side effect.57 Rauke and associates studied epidural clonidine in the treatment of complex regional pain syndrome (CRPS).58 Twenty-six patients received 300 μg, 700 μg, or placebo. Results showed significant analgesia within 20 minutes of injection compared with placebo. Nineteen patients went on to have clonidine administered by continuous epidural infusion; weekly VAS scores were significantly reduced to an average of 5/10 compared with 8/10 prior to clonidine therapy.

Hassenbusch and coworkers treated 32 patients with intrathecal clonidine (all but three with CRPS) in doses ranging from 480 to 900 μg/day. Fifty-six percent of patients had improvements in both their pain rating scale scores and CRPS symptoms. Systemic hypotension was observed at intermediate doses, but was eliminated at doses ≥40 μg/hour, and mostly observed with doses between 17 and 25 μg/hour.40

Tizanidine (often called the “son of clonidine”) is another α-adrenergic drug that produces antinociception in a similar manner to clonidine but without the pronounced hemodynamic changes when administered intrathecally in animal studies.59,60 These drugs are likely to remain an important part of neuraxial treatment of neuropathic pain.

Somatostatin Analogs

Numerous investigators in the 70s and 80s published data describing the localization of somatostatin as a substance intrinsic to spinal cord interneurons in the superficial layers of the dorsal horn, displaying inhibitory effects on nociceptive neurons.61–64 Penn and coworkers described their experience with six cancer patients using octreotide, an analog of somatostatin.65 Patients experienced reduced pain and decreased oral opioid usage without central or systemic side effects. Somatostatin has since been isolated in the periaqueductal gray, in substantia gelatinosa of the spinal cord, and in descending modulatory systems from the brain, seeming to confirm a role in inhibition; however, two of eight patients appeared to have demonstrable histopathologic changes noted at autopsy.66

SNX–111

N-type calcium channel antagonists have been shown to be consistently antinociceptive in animal models.67–69 Early results in humans have continued to show efficacy.70,71 Staats et al reported on 102 patients treated in a double-blind, randomized, placebo-controlled trial of the efficacy and safety of SNX-111.72 Results indicated a significant decrease in pain intensity when compared with placebo. Adverse events included dizziness, nausea, nystagmus, abnormal gait, constipation, confusion, and urinary retention.

Ketamine

Spinally administered NMDA-receptor antagonists have been studied in animal models.73,74 The relationships between opioids and NMDA antagonists continue to be explored. In a double-blind, controlled trial, one group of investigators administered doses of alfentanil and ketamine, or their combination, to normal volunteers and found no advantage of the combination over a larger dose of either drug alone in relieving pain caused by chemical stimulation.75 These authors concluded that the opioid and NMDA-antagonist interaction was one of simple additivity and not synergy. Chia and others added ketamine 0.4 mg/mL to a standard patient-controlled epidural analgesia mixture of 0.02 mg/mL morphine, epinephrine 4 μg/mL, and 0.8 mg/mL bupivacaine in a postoperative double-blind study (N = 91). Patients in the group with ketamine had significantly lower VAS scores both at rest and with coughing and movement, and patients in the control group used more of the morphine-epinephrine-bupivacaine mixture at both the 24-hour and 48-hour assessment points.76 Wong and associates demonstrated that ketamine coadministration with epidural morphine potentiates morphine’s analgesic effect via a single bolus injection.77 Ketamine and its oral relative dextromethorphan have failed to gain much clinical acceptance as adjuvants because of CNS side effects.

Future Development

Allan Basbaum looked to the future in a recent 25-year anniversary monograph from the International Association for the Study of Pain and noted the promise of molecular biology in “deleting” genes from the different opioid receptors.78 He noted the ability to target drugs to the spinal cord through selective gene manipulation. One startling new discovery gives us a preview of what the future of neuraxial drug delivery could hold: orphanin FQ/nociceptin.79 The defining of the δ-opioid receptor gene sequence in the early to mid 90s led to cloning of three receptors, μ, κ, and a previously undescribed receptor referred to as an “opioid-like orphan receptor.” This orphan receptor has an endogenous ligand known as orphanin FQ/nociceptin or OFQ/N (an amino acid peptide) that is widely distributed throughout the nervous system. Investigators have reported an apparent lack of “motivational” effects,80 indicating the lack of abuse potential and the presence of antihyperalgesic and antiallodynic effects81 that may have efficacy for neuropathic pain.

The interesting pharmacologic aspect of this agent is that OFQ/N has potent opioid-antagonist actions in the CNS; however, at the spinal cord level, it has opioid-like activity and no anti-opioid effect. This combination of clinical properties has the potential for providing pain relief in the periphery while avoiding concurrent CNS side effects, and more importantly, avoiding centrally mediated opioid-induced hyperalgesia in the persistent pain model.

The patient selection process for neuraxial drug delivery is arguably the most significant step toward achieving a successful outcome. No patient should receive an implant without having first received optimum conservative therapy. The primary goals of patient selection are to confirm the indications for implantation, evaluate the psychological stability of the patient, and estimate the patient’s future responsiveness to the treatment plan. A comprehensive analgesic history should be part of the primary assessment of the patient. The comprehensive history will give the practitioner a detailed history of side effects and adverse reactions from the patient’s previous exposure to oral opioids and adjuvant therapy. Oral opioids are the mainstay of treating cancer pain, and are increasingly accepted as a long-term alternative for chronic nonmalignant pain. Opioids should be aggressively titrated to pain relief or to side effects. The optimization of adjuvant analgesic drugs should precede spinal drug delivery. Nonsteroidal anti-inflammatory drugs (NSAIDs) are indicated for bone pain at any level of intensity. The tricyclic antidepressants (TCAs) and the anticonvulsants are indicated for neuropathic pain syndromes. Several of these agents can and should be given a trial in an effort to lower pain intensity. If pain intensity continues to be unacceptable despite optimum treatment with adjuvant and aggressive opioid dose escalation, the pain is considered “intractable” and the patient should be considered for alternative routes of drug delivery. These uncontrolled pain situations are generally associated with neuropathic or severe intermittent pain (i.e., pathologic fractures in the advanced cancer pain patient).

Truly unmanageable side effects from systemic opioid therapy are the most common phenomenon leading to neuraxial implant. It is typical for patients to experience some side effects of opioids, most only transiently. Persistent sedation can be managed with the use of stimulant drugs. For drug-induced nausea, the concurrent administration of an antiemetic is recommended, or sequential trials of different opioids will often result in a better efficacy to toxicity profile. The algorithm for pain management in this population is depicted in Figure 71-3.

Neuraxial opioid analgesia should not be selected solely on the basis of failed systemic opioid analgesia. Opioid intraspinal analgesia is simply another opioid delivery system, which has some advantages over oral opioids, but the quality of achievable analgesia is similar. One clear advantage is to achieve a consistent delivery modality for low-dose opioid analgesia in patients who have truly intractable dose-related opioid side effects. Morphine and nonlipophilic opioids have a distinct dose-response advantage when administered intrathecally, and are used almost exclusively for somatic pain.82 More commonly, patients with severe nerve injury or bony destruction and neuropathic pain require the addition of local anesthetics or clonidine to achieve “functional” analgesia. Placement of the catheter tip to a specific spinal level will allow accurate delivery of local anesthetics and clonidine to the cord level of the pain processing.

Indications and Contraindications of Intraspinal Analgesia

The indications and contraindications of intraspinal analgesia are outlined in Table 71-1. There are no individual studies that indicate specific risk-factor frequency for epidural hemorrhage, hematoma formation, and other complications. The American Society of Regional Anesthesia published a consensus report in 1998 giving guidelines for epidural analgesia in the face of coagulation risk factors.83 The individual patient’s clotting factors should be close to normal before epidural catheter placement or catheter withdrawal, but it is clear from this report that there are no definable specific laboratory values that will protect the patient from adverse events.

A clear requirement for the use of these implantable techniques is the presence of support from an interested family member or significant other if the patient is to be cared for in the home. A caregiver-partner is critical for any externalized system, and the absence of a caregiver should alert the practitioner toward a fully implantable system. Although some patients may be able to manage independently in all other aspects of their lives, a partner to monitor and maintain aseptic technique, and assume accountability for intraspinal drug administration, is a prerequisite to implantation. The potential strain on this person should be discussed and factored into the decision-making process. Occasionally, family members will be reluctant to view and/or use any tubes coming out of the body. Other family members may be anxious and willing to do whatever they can to help. The caregiver should be present during the discussions of implantation and analgesic therapy initiation. Barriers to use of opioid analgesia include the attitudes of the caregivers and family. It is important for the implanting physician to understand the family dynamics before implantation. Negotiation of acceptable analgesic care includes all aspects of the device selection, implantation, drug infusion, and complications.

Home Care Issues

The home care agency or hospital pharmacy department must have a depth of experience in the nursing and pharmacy to embark on the care of a patient with an implantable device. Nurses will be asked to assess and titrate externalized infusions, and perform care of any exit site. If a device is to be implanted, special training will be required for clinic nurses to be able to refill and program the device. The pharmacy will be asked to mix infusion solutions with a variety of drug combinations, investigate drug stability and interaction information, and have necessary stock on hand. Managed care organizations and medical insurance organizations are now selecting home care and pharmacy agencies who may or may not have the skills to deliver the care required by these patients. It is important for the clinician to be constantly aware of the technical skills of the staff in agencies with whom he or she is working. This can be accomplished by having direct contact with the nursing and pharmacy staff as they interact with the patients.

The population of candidates for acute pain analgesia is the largest population of patients considered for access to a neuraxial space. Most of these patients will be experiencing postoperative pain. Also, within this class of patients are the acute trauma patients and those chronic pain patients experiencing an acute exacerbation of pain. Among the most challenging acute pain presentations involve patients with a history of drug addiction and opioid tolerance, and those who are hypersensitive to opioid side effects to the point of preferring the pain over the side effects. These patients may benefit from epidural local anesthetic with or without clonidine. Preemptive analgesic planning by clinical anesthesiologists in these difficult to treat patients will abort the postoperative crises of patients, caregivers, and the system at large.

Selection for Implantable Pump Systems

Patient selection for implantable pump systems differs from the selection criteria for externalized devices on the issues relating to duration of therapy and planned drug infusion. Implantable pumps are usually reserved for chronic pain and cancer patients who are expected to live and require infusion for more than 3 to 6 months. Bohme and Tryba point out the following criteria for pump implantation: pain of somatic origin, exclusion of mental diseases and psychogenic causes of pain, causal therapy is exhausted, insufficient effects of peripheral analgesics and coanalgesics, oral or transdermal opioids are insufficient despite dosages resulting in side effects, pain is sensible to opioids, and regional applications of opioids have been tested effectively before implantation.82 The main point of their article is that implantation in a patient who does not get relief from a test on neuraxial opioid analgesia and has somatic pain will not benefit from a pump implantation. There are significant cost issues for long-term externalized infusions, where an exteriorized pump is required for infusion. Rental on pumps, cost of drug mixing, delivery, and support may be cost prohibitive for long-term infusion in the United States. The cost advantage of the subcutaneously buried intrathecal morphine pump is the lack of ongoing infusion company charges. The pumps, however, are limited in their volume capacity and therefore are difficult to use with patients who have neuropathic pain or activity-related pain and require the infusion of clonidine or local anesthetic agents. Both of these agents are not commercially available in high enough concentrations for the low-volume infusions of the subcutaneous pump systems. Pumps are used less often for the control of cancer-related pain, as the aggressive use of oral opioids and adjunctive drugs are effective in most cancer-related pain states.

Chronic nonmalignant pain patients are more often candidates for pump implantation since their use is expected to be of long duration, and their pain is often somatic in origin. Before implantation, patients in our clinic are tested for opioid responsiveness with epidural analgesia trials of 2 to 12 weeks. Psychosocial issues are evaluated with a psychological consultation. The planned location of pump placement must be addressed with the patient, particularly as it applies to activities of daily living that may cause disruption in dress, exercise, or sexuality.

Analgesic Testing before Implantation

The process of neuraxial testing before selecting a patient for pump implantation is a controversial issue. Some implanters will test the patient with a single or multiple neuraxial injections of opioid and evaluate the patient’s response to determine the potential efficacy of this approach. The simple process of injection and questioning will illicite a potential placebo response. To avoid the placebo response and to determine the functional analgesia of this technique, the practitioner must evaluate the patient’s response over time and in his or her own environment. We support the process of a long-term (3 to 6 weeks) epidural or intrathecal infusion of analgesic agent to determine the functional analgesia of this technique. During the long-term infusion, not only can the functional capacity be measured but the risk of continuous escalating doses can also be estimated. Patients who do achieve long-term neuraxial analgesia during this test phase will in our experience do much better after implantation.

Patient selection is a negotiation with the patient. From the patient’s view this is treatment selection by the patient. Once the clinician has determined in his mind the appropriateness of the treatment plan based on the information available, the patient makes the ultimate decision. It is important for the clinician to be aware of the patient’s goals, fears, and perceptions and address these as thoroughly as possible, being very clear on the amount of relief that can be anticipated and any limitations set of supplemental oral analgesics once the titration period ends. Clear communication between the patient and the physician during this phase is essential.

Selection of an Implantable Device

Selection of an implantable device is a process that is concurrent with patient selection. The decision to implant a device for drug delivery should follow an algorithmic decision process starting with the optimum use of oral opioids and adjuvant drugs and minimization of side effects. When oral and transcutaneous analgesic delivery has been maximized without functional analgesia, then patient selection and device selection of an implantable device starts.

Device selection requires an understanding of the current status of the pain generators from malignant and nonmalignant sources. In nonmalignant pain states, the specific pain generators may be identified through the use of diagnostic blocks, studies, and imaging. Surgical consultation and second opinions may be obtained to determine if additional surgery may be indicated. Once the cause of the pain is identified, the treatment aimed at the causative condition has been optimized or exhausted, conservative multimodal pain management has failed, a screening for intraspinal analgesia can begin.

Pain associated with malignancy is usually a progressive process requiring continuous reassessments and updates of the pain treatment plan. The pain-focused consultation and physical examination should include documentation of the location of original malignancy, metastatic locations, future expectations for disease spread, and future antitumor treatment options. The correlation of metastatic sights with pain sources on the pain-focused examination will not only tie the pain generator to the metastatic disease process but will also give the pain practitioner a better idea of future possible pain generators. The site of the pain generators, correlated with the pain character, pattern, and intensity, will lead to the practitioner’s choice of delivery system. This is important information that will be needed in determining the choices of drug for delivery, spinal level for catheter tip location, and the method of delivery.

The presence of neoplastic spinal disease has previously been considered to be a contraindication for neuraxial drug delivery.84 Metastatic lesions of the spine, however, account for some of the most severe pain problems, particularly when nerve root impingement occurs secondary to tumor invasion and vertebral collapse. Most of these lesions affect the vertebral bodies, leaving the posterior spinous elements and lamina relatively unaffected.85 In nonmalignant states, trauma and/or degenerative changes may result in vertebral collapse, with extension to the epidural space, and intrathecal compression may occur. Neuropathic and activity-related pain generated from nerve root compression and disk space sources would be similar to that seen in malignant patients. When epidural space involvement is suspected, an epidurogram, magnetic resonance imaging (MRI), or contrast computed tomography (CT) scan may clarify the extent of external pressure versus epidural space-occupying lesion. Careful radiographic evaluation of vertebral involvement with identification of any current or potential epidural space compromise should be identified. When there is extensive epidural space invasion, epidural infusions may have limited efficacy due the incomplete spread of drug in the epidural space. Epidural infusions under these conditions may have a segmental effect with incomplete coverage of affected area. Intrathecal infusions may by considered both through externalized devices or subcutaneous pumps when the epidural space is compromised or when in the opinion of the practitioner the intrathecal space may have clinical advantages.

Types of Devices

Access to the neuraxial space may be achieved through several devices. Table 71-2 describes the advantages and disadvantages of intraspinal devices. The commercially available devices for epidural and intrathecal access include ports and three main categories of catheters:

• 1. Short-term catheters made from polyamide, polyurethane, and nylon with have friction adaptors attached to their distal end.
• 2. Non-kinking, wire-supported epidural catheters with friction adapters, which under ideal clinical conditions may be used for prolonged epidural access.
• 3. Long-term silicone rubber catheter with fixed integrated Luer-Lock adapters.
• 4. Implantable epidural and intrathecal ports.

Each of the short-term catheter materials differs in stiffness and tissue reactivity; some soft catheters will require a stylet for epidural placement. Figure 71-4 illustrates some of the short-term catheters on the market. Catheter stiffness, on the other hand, aids the practitioner during placement to a specific spinal level or nerve root location, when precise drug delivery to that location is desirable. Desired catheter stiffness may be obtained with soft catheters during catheter placement by the use of special wire stylets. Stiffness may play a role in catheter migration from epidural to intrathecal spaces or in the reverse direction; stiffness may also result in pressure and nerve root paresthesias. If the practitioner is concerned more about catheter kinking and obstruction than the location of the catheter tip, then the use of a wire-reinforced epidural catheter may be the best choice. Device selection is an individual choice, and the practitioner should select a device with which he is most adept and comfortable. This selection usually results in the most predictable outcomes.

Reliability is the next level of device selection. When the duration of implantation is expected to be less than 3 weeks, then the use of a temporary catheter may be the best choice; although expected duration of use is a difficult determination to make. The average short-term catheter, in the author’s hands, will remain in place for an average of 2 weeks of therapy before needing replacement. The durability of the externalized catheter and catheter adapter are the weakest segment of temporary catheters. Failure of the temporary catheters usually occurs with separation of the catheter adapter and contamination of the catheter. Less commonly, the temporary catheter may break or kink. There is a gain in durability and dependability by moving from the temporary catheters to the silicone rubber catheters. The two-part catheter system with a fixed Luer-Lok adapter and the small intraspinal catheter connected to the strong externalized catheter gives the durability. This catheter system requires surgical implantation, which is a definite disadvantage over the temporary catheters that may be placed at the bedside. The silicone rubber epidural catheters are placed for expected durations of therapy in the weeks to months range. Figure 71-5 depicts the Du Pen silicone rubber epidural catheter. Although there are advantages of this system over the conventual temporary epidural catheters, there are no clinical applications that cannot be managed effectively with a temporary catheter. Patient convenience and safety may be better addressed with the permanent catheter when long-term implantation is contemplated. Catheter failure will result in loss of the infusion and added risk of infection if the continuity of the system is broken. Loss of infusion usually means an extra trip to the hospital and the resurgence of pain.

The implantable pump manufacturing companies have workshops set up around the country to teach practitioners skills needed to manage these devices, including pump implantation and fixation, dose calculation, and identification and treatment of complications associated with therapy. These are valuable instruction sessions usually taught by experienced implanters that give the practitioner a more complete understanding of the issues of patient and device selection. There are two basic types of intrathecal pumps and one intrathecal patient-activated reservoir system (still investigational) in use today. There are variable-rate, battery-powered computerized pumps and the fixed-rate, Freon gas-powered pump.

• 1. Medtronic SynchroMed pump (Medtronic Neurological, Minneapolis, Minn) 10 mL and 18 mL sizes with or without side ports.
• 2. Arrow 3000 constant flow implantable pumps (15, 30, and 50 mL sizes) with high- (1.46 to 2.0 mL/day), medium- (0.86 to 1.45 mL/day), and low-flow rates (0.5 to 0.85 mL/day).
• 3. The Algomed patient-activated intrathecal reservoir delivery system (Medtronic Neurological, Minneapolis, Minn, not yet approved by FDA).

Selection of an implantable pump is usually a practitioner preference issue. The two classes of pumps available on the market today were outlined earlier. The choices are between the variable-rate Medtronic pump and the Arrow fix-rate device. There are advantages to each pump type. The SynchroMed pump has a small reservoir (10 and 18 mL) but is larger in size due to its internal mechanisms (Figs. 71-6 and 71-7). Its advantages include the simple and complex infusion rates operated by the computer, adjustability by the external computer station, and its ease of reservoir and side-port accessibility. These advantages are valuable when analgesic admixtures are used that require frequent titration adjustments in infusion rates. Titration of the intrathecal analgesic does not require refilling the pump. The side-port access to the intrathecal space is protected by a screen, which can only be accessed by a 25-gauge needle.

The Arrow constant flow pump has a choice of reservoir sizes (15, 30, and 50 mL), which allows lots of flexibility in physical size and reservoir size, the rate of flow (high, medium, and low) of the pump is also selected, or a higher infusion flow model that allows for the use of local anesthetic agents. Figure 71-8 shows two of the pump sizes. The Arrow pump has a double chamber for access to the reservoir and the side port. Infusions do not require the use of a computer, allowing use in areas were the computer is not available. The low profile of the Arrow pump allows placement in thin patients. Its interesting to note that a heat (such as a hot tub) can increase the infusion rate, a sort of patient-controlled intrathecal analgesia effect.

The choice of implantable pump should be based on the patient’s individual needs. A choice is dependent on the drug or drug combination chosen for infusion, infusion rate, need for dose titration, proximity to a pump computer, and patient choice.

The Algomed system is a patient-controlled intrathecal dose-delivery system that has no automatic or computer driven functions. The device is operated by a series of valves, which allow the patient to activate the injection chamber and administer a specified dose volume into the intrathecal space. The delay in refilling of the pump chamber determines the time allowed between doses. The Algomed system has been used in clinical studies in Europe and the United States, but has not been approved for use in the United States by the FDA at the time of this printing. Figure 71-9 shows the pump.

Ports are another alternative to accessing both the epidural and intrathecal spaces. The main attraction of the port system is the lack of externalization. This allows the patient to function normally when the port is not accessed. The pharmacokinetics of today’s intraspinal analgesic therapy does not lend itself well to intermittent dosage. The practitioner has a choice of long-acting intraspinal opioids, short-acting local anesthetic agents, and α2-agonists. The duration of injectable intraspinal opioids is from 8 to 12 hours, where the local anesthetic and α2-agonists agents require infusion because of their short duration of action. Most patients requiring intraspinal analgesia have severe neuropathic or activity-related somatic pain. These pain states require constant infusions and continuous access of the port. The continuous access of the port effectively negates the advantages of the port system. The controversy between ports and externalized catheters for intraspinal infusion is similar to the controversy between central intravenous ports and Hickman catheters. This mode of therapy may take on a whole new utilization profile if and when we have new drugs developed that have long durations of effect.

Implantation techniques vary from practitioner to practitioner, as there are no specific implantation standards in pain management. Individual practitioners have reported on implantation techniques, but there are no comparative studies looking at the outcomes from different techniques. There are many variables such as needle placement, location of incision, antibiotic irrigation, suture techniques, and pocket dissection. More exploratory work in this area could bring valuable information to the literature on complications.

The choice of technique depends heavily on the training and experience of the practitioner. Internally placed pumps and ports require additional technical skills. During these implantation procedures, the practitioner must be aware of and select the site of implantation, depth, and fixation. These technical differences may influence the duration of the implantation procedure, the ease of post-procedure care, patient comfort, and convenience.

Temporary Catheters

Temporary catheters of all types may be placed at the bedside into either the epidural or intrathecal space, with or without a short tunnel to add to catheter stability and possibly influence catheter track infections. Zenz et al86 reported on the use of a single nylon catheter for more that 1[1/2] years, using a special skin fixation technique; so longer-term use is possible. These catheters may be placed under fluoroscopic guidance, when a specific location is desired, but direction of the catheter movement is difficult to control. Catheters, which have wandered along a nerve root exit, will be less effective and may result in increased neuropathic pain. Epidurograms will give the practitioner information about catheter position, patency of the epidural space, and dye flow in that space. These films may also be used as a standard for future epidurograms that the patient may require in determining the cause of failure of analgesia. Myelograms may be easily obtained to prove intrathecal catheter placement and position.

Permanent Catheters

The permanent catheters differ from temporary catheters in that their two-part construction allows for radiographic-controlled specific epidural placement with a long subcutaneous, tunneled durable catheter for stability and infection protection. There is a Dacron cuff for catheter stability and fixation and a silver-impregnated cuff as a barrier to bacterial growth along the catheter. The 33% barium content of the Du Pen catheter allows visualization with simple lumbar spine films, but does not alter the breaking strength of the catheter. Other than occasional technical problems, which may result in occlusion, the only major risk of implantation is infection. Care must be taken in all aspects of catheter management and wound care. Short-term perioperative antibiotics are commonly used for most patients receiving implanted devices. Standard therapy would include administration of a broad-spectrum antibiotic 1 hour before surgery and two doses 6 to 8 hours apart postoperatively (see section, Complications).

The procedure for epidural or intrathecal catheter implantation should be undertaken in a sterile environment and fluoroscopy should be used during the catheter placement. The patient should be placed in a lateral position, allowing access of C-arm fluoroscopy from the side away from the surgical field. The 14-gauge Hustead epidural needle is placed in the epidural space using a long paramedian approach, which avoids the sharp angle created by a midline approach and allows a smooth entry into the epidural space for accurate needle positioning.87 The needle angle will allow the practitioner to withdraw the catheter for repositioning without a risk of catheter damage. The use of 10 to 15 mm of saline or an local anesthetic agent will not only dilate the epidural space but will give analgesia for the catheter tunneling procedure. The catheter should extend four vertebral levels above entry to ensure stability; this should be planned before the entry level is determined. The catheter should be advanced two vertebral levels above the planned catheter tip position using fluoroscopy. This allows for one vertebral segment catheter withdrawal during the process of positioning the tunneled catheter. Postoperative orders are written for an epidurogram or myelogram, postoperative antibiotics, analgesic therapy, fluids, diet, and opioid conversion. Postoperative catheter care includes daily wound cleaning and a dry 2 × 2 dressing, until the securing sutures have been removed. Orders for discharge should include comprehensive instructions on dressing care and the use of the filter system. All sutures must be removed at 10 to 14 days to avoid the added risk of infection. Activity instructions include avoidance of use of hot tubs or other possible contaminated exposure of the catheter track.

Ports

Port implantation, epidural or intrathecal, is similar to that of the permanent catheter except instead of an exit location, the catheter is connected to a subcutaneous port. The port should be placed over a bony structure to withstand the pressure required for needle access of the port. Postoperative care includes normal wound care. Care should be taken to use aseptic conditions to access the port to protect for infection. Special care should be taken if the port will be accessed for constant infusion. Under these conditions, the needle and area around the port should be constantly protected from bacterial contamination. Rotation of the needle access position and attention to health of the tissue over the port should be paramount. Dressings and skin care over the port should be optimized as long as the device is accessed.

Intrathecal Pumps

Implantation of the intrathecal pump systems requires the use of an operating suite and a full surgical setup, general or deep sedation anesthesia, and access to fluoroscopy. The patient is placed in a lateral position with access to the spine for intrathecal catheter placement and access to the lateral pelvis for pump implantation. An accomplished practitioner will have no difficulty entering the intrathecal space at any accessible level. A small 1-in. transverse incision is usually made at the L4-5 level, allowing a flat paramedian needle entry into the intrathecal space at the L3-4 or L2-3 interspace. The angle of the needle will decrease the risk of catheter damage during catheter manipulation and will avoid the risk of retrograde flow of spinal fluid along the outside of the catheter. If difficulty is encountered while attempting to enter the intrathecal space, the head of the table may be elevated to distend the dura. The intrathecal catheter is threaded into the intrathecal space; fluoroscopy may be used to ensure that the catheter remains in a central position and that it remains in the dorsal space. There are a number of devices supplied by both companies to fixate the intrathecal catheter; however, the author prefers to use a single silk suture attached to the intraspinous ligament and secured to the connection between the two catheters. This “dog leash’’ suture allows a gentle curvature of the catheter, but does not allow the weight of the pump to pull on the catheter.

The pump pocket location is individually selected for the greatest comfort to the patient. The incision is usually about 1 in. below the umbilicus and extends laterally about 4 in. in length. The depth of the pocket should be about 1.5 cm, but this may vary from patient to patient. Ideally, the pump is fixed to the fascia, but in most cases the layer of subcutaneous fat is too deep to allow this placement without making the pump almost inaccessible for refilling. The tunneling device is used to bring the connecting catheter from the pump to the posterior incision. After the catheters are connected and the dog leash is secured, the catheter unit is withdrawn into the tunnel, avoiding catheter kinking. Wound closure must be made watertight, as there will be a collection of fluid around the pump, and any openings will continue to “leak” until the closure is watertight.

Each company has a specific protocol to follow for the original pump filling and follow-up refills of the pump. Both companies have addressed the fear of instilling concentrated drug directly into the intrathecal space. Medtronic has a port screen, which will not allow the refill 22-gauge needle to enter the side port, whereas Arrow has a double chamber with a special bolus and refill needle.

Wound care is similar for both pumps. During the first 5 to 7 days, while the sutures are in place, the patient may shower but not use of hot tubs. In addition, providers should collaborate so that the patient is not having radiation to the site or surrounding tissues during this first week. After the sutures are removed, the patient may bathe as he or she wishes. We use an abdominal binder to support the pump for the first 3 weeks, until the scar tissue is well established and the device is fixed in place.

Epidurograms are a helpful radiographic technique. The purpose of the procedure is not only to confirm epidural space location of the epidural catheter, but as a diagnostic tool to note the extent of epidural flow of dye. This information may be used to predict volume requirements or restrictions of fluid spread within the space. A baseline study may be referred to later if analgesic problem arise, which may be solved by a new epidurogram and a comparison to the baseline study. It is important to obtain the study in the radiology department to ensure reproducibility.

Pump myelograms will prove intrathecal catheter placement if fluid cannot be aspirated from the canal due to fibrosis. Pumps without side ports cannot be tested for catheter position, either by aspiration or myelography. Computed tomography and myelograms may be helpful in determining catheter position in pumps without side ports.

Treatment planning must be a flexible and continuous process. The process is based on the original assessment of the patient’s pain experience and a reassessment of responses to the executed treatment plan. Responses to opioid infusions will be relatively predictable, but when neuropathic pain persists, a more complex treatment plan will likely be required. The additions of clonidine or local anesthetic agents are commonly used in epidural and intrathecal analgesia. Dilute local anesthetic solutions administered neuraxially are a comfortable therapeutic technique for the vast majority of anesthesiologists in the inpatient setting. These infusions can be easily managed at home safely with a proficient team of outpatient and home care support staff. Once stabilized, the patient can be managed with a range of infusion rate parameters for the home care nurse to titrate within. Clonidine, a relatively new entry in the field, has been extensively used for neuropathic pain states in Europe and enjoys a remarkable efficacy in some of the most intractable situations (i.e., sacral plexopathy).51 As one might expect, clonidine requires close monitoring of blood pressure during initiation and titration. These combination infusions, opioid-local anesthetic, opioid-clonidine, and even all three together, will require close outpatient follow-up. A balance of side effects including drowsiness (opioid), skin anesthesia (local anesthetic), and clonidine (orthostatic hypotension) need to be balanced against pain relief. Most advanced cancer patients and elderly chronic pain patients are not able to come into the clinic, and hence the presence of well-trained home care nurses is essential.

The selection of an opioid for spinal administration has typically been done on a “physician’s choice” basis. Some physicians are only comfortable with morphine and others prefer hydromorphone or fentanyl, perhaps because of their own experience and comfort level. Opioid selection as a part of treatment planning incorporates duration of action, efficacy and toxicity response to the opioid at screening, vertebral level site of injection, and optimum volume and concentration of injectate.88 A patient’s past history of opioid exposure will often allow the practitioner to select an opioid that has given the best performance historically for the individual.

Morphine offers the best advantage over systemic administration because of its hydrophilicity. The duration of action of bolus-administered epidural morphine is generally 6 to 10 hours when administered in the opioid-tolerant cancer population. Plummer et al in a large retrospective study looked at 313 cancer and noncancer patients treated with epidural morphine for a mean treatment period of 96 days and found that dose escalation happened less when the bolus technique was used.89 The bolus method has implications for ease of use in the home care setting as well as cost containment. However, as the volume and sophistication of options for oral and transcutaneous opioid therapy have become widely available, the need for epidural opioid therapy alone has become almost obsolete. It is much more likely that the patient requiring neuraxial analgesia will require combinations of opioid and adjuvant drugs that make infusion therapy mandatory.

Morphine can and is still used as a sole agent by infusion both epidurally and intrathecally. In managing morphine infusions, titration to effective analgesia with minimal side effects is the method of choice in determining the ideal dose for an individual patient. Hydromorphone and fentanyl are most often found in the practitioner’s second-line drug category. Either agent may be a good choice in patients where morphine has been previously efficacious without side effects; however, side effects become problematic with escalating doses. If patients seem to be very sensitive to epidurally administered morphine, switching to a second-line drug is a useful tool for ongoing treatment.

Clonidine and local anesthetic agents are added to the epidural infusion to control neuropathic pain while lowering the dose of opioid and its associated side effects. Clonidine is usually titrated to effect, starting at 10 to 20 μg//mL and working up to 30 to 40 μg/mL (Table 71-3). Blood pressure and pulse rates are monitored and treated if required. Local anesthetic agents added to the infusion may result in motor and sensory loss and postural hypotension. In our experience the use of 0.115% bupivacaine, with concentration titration between 0.1% and 0.25%, will give functional analgesia. Our method of titration is in increments of 0.005%.

The potential complications associated with spinal drug delivery systems, along with cost, are the biggest barriers to the widespread use of these techniques. Before considering the use of invasive devices, the practitioner must be fully aware of all the potential complications of the procedure, the devices, and the associated drugs. Complications associated with long-term spinal drug delivery systems fall into four major groups: infection, drug toxicity, practitioner errors, and mechanical problems. Physicians, patients, caregivers, and those offering home care services must be proficient in avoiding the avoidable, troubleshooting problems, and correcting or treating the complications. Practitioners who are aware of these potential problems will identify them earlier and resolve problems as they occur. The procedures themselves do not create the problems; the problems usually arise out of a misunderstanding of the implantable devices, their care, and use.

Infection

Intraspinal infections are the most feared consequence of chronic epidural and intrathecal catheterization. Epidural infections are predictable in those pain management practices that use significant long-term epidural catheters, particularly in the immunosuppressed cancer patients. Our initial infection rate with the Du Pen epidural catheter was reported at 5.4%.90 From 1990 to 1996 the epidural infection rate within our practice has ranged between 5% and 15% in our long-term patients. Patient care provided by multiple home care agencies, multiple caregivers, and with variable filtering, tubing, and reservoir change procedures has made tracking of infection risk difficult. However, frequency of tubing and filter replacement was noted as a variable that changed over time. A protocol of changing filters weekly resulted in our previously reported 5.4% infection risk in our practice. The hospital epidemiology committee established a protocol of changing the filter and tubing every 3 days; this resulted in an increased infection rate (10% to 15%). An internal study showed no correlation between infection rate and any known factors except for duration of implantation. A colleague (William Baumgartl, MD, personal communication, July 14, 1997) reported his low incidence of epidural infections with a dual-filter system technique. He reported using two PALL epidural filters in series, the inner filter never being changed and the outer filter and pump tubing changed on a monthly basis (Fig. 71-10). The reports of intrathecal catheters used for chronic pain management from Sweden also reported a very low incidence of meningitis (0.5%). In these reports the intrathecal filters were changed once a month. There have been no studies to prove the benefits of the double-filter system or to indicate that the PALL filter is superior to other epidural filters.

Identification and Treatment of Epidural Infection

Early diagnosis and treatment of epidural space infections are key to a successful outcome. The most common infection occurs at the exit site. The exit site appears red, may have an exudate, and will often be painful. These infections are treated effectively with aggressive site care, often with topical antibiotic ointment, and more rarely with systemic antibiotic therapy. Of greater concern is the potential for infection inside the epidural space. The epidural infection may manifest itself from one of three sources: the drug infusate, the catheter track, or from a hematologic or local abscess source. In the 1997 calendar year, 75 permanent epidural catheters were placed at Swedish Medical Center, which is a regional pain management referral center for the northwest. Nine confirmed epidural space infections (12%) were treated in 1997. Only one of the patients presented with signs of epidural tunnel-track spread (1.3%), three patients had bowel tumors and episodic septicemia that potentially seeded hematologically (4%), and five patients had no known etiology, no exit site or track signs, and were presumed to have contaminant sources (6.6%). We adopted the double-filter protocol in 1997, and in three years we have had only four epidural infections in a total population of 145 epidural catheters, a 2.7% epidural infection rate. Clinicians should examine the potential for contamination within their institutional policies very carefully. The finding of three patients in the sample with advanced colorectal cancer suggests a higher index of suspicion in patients with bowel tumors; however, this phenomenon requires prospective clinical study.

The signs of epidural space infection include acute loss of analgesia, fever, and white blood count (WBC) elevations.91 In the immunosuppressed cancer population, fever and WBC elevation may be delayed. Epidural space infection and catheter encapsulation present with three clinical symptoms:

• 1. Pain on injection
• 2. Retrograde flow of infusate with pooling in the paravertebral tissues
• 3. Loss of analgesia.

The pooling of fluid and exudate in the paravertebral tissues is the result of material tracking back out of the epidural space, often within a fibrin-like sheath surrounding the catheter. The catheter entry site through the ligamentum flavum becomes a drain, and the exudate leaks out around the catheter, causing a significant inflammatory process to occur in the soft tissues.92

Diagnosis is made by obtaining an epidural aspirate, Gram stain, and culture. The culture must be obtained from the epidural space. In the past, epidural catheters were immediately removed when signs and symptoms of infection were present. Catheter tip culturing was commonly used to identify infection in the epidural space and is still reported in the literature, despite being clearly problematic.93 Bevacqua and others examined cultures from spinal catheters used for postoperative pain management (mean indwelling time, 66 hours). Cerebrospinal fluid cultures were done concurrently with removal of the catheters and culturing of the catheter tip. Semiquantitative culture methods demonstrated a relationship between the positive catheter tip cultures and contamination of the catheter that occurred on removal.94

Culturing of the catheter tip after removal will result in a composite culture of the epidural space, paraspinal soft tissues, catheter track, and superficial exit site. An aspirate culture from inside the epidural space prior to catheter removal with subsequent identification of the organism and its sensitivity is mandatory for best treatment outcomes in these patients. When the volume in the epidural space is sparse and fluid cannot be aspirated, 2 mL of normal saline may be injected for sampling. Table 71-4 summarizes the procedure for performing an epidural aspirate culture.

Epidural space infections can result in accumulation of up to 20 mL of exudate in the epidural space. A syringe attached to the catheter can decompress the space of any exudate while antibiotic therapy is being initiated. Withdrawal of exudate can continue on an every 8-hour basis until no fluid can be withdrawn. This technique may help to avoid the risk of cord compression from the accumulation of fluid in the epidural space.

Epidural space infections have been treated primarily by removal of the catheter and systemic administration of antibiotic therapy directed toward the specified organism identified with the culture. Our policy, established 10 years ago in collaboration with our infectious disease staff, was that removal of the catheter (after no exudate could be withdrawn and culture results had been reported by the laboratory) was mandatory for complete resolution of the infection. This policy has been applied to over 70 patients with positive epidural space cultures over the last 12 years, during which time no patients had to be surgically decompressed. Appraisal of response to treatment can be done with a “wait-and-watch” approach to resolution of symptoms, or in cases where neurologic signs are present, it is best accomplished with either a MRI or a contrast CT scan.

Many clinicians have considered sterilizing the epidural catheter with epidurally infused vancomycin (with sensitive organisms) rather than removing the catheter, treating, and replacing the catheter. Hahn et al first reported a case of sterilizing an infected epidural catheter with epidurally infused vancomycin.95 We have treated three cases of epidural infections attributed to Staphylococcus aureus with systemic vancomycin (1 g every 12 hours, given intravenously) for 10 days and concomitant epidural infusion of 150 μg per hour of vancomycin for 3 weeks. This treatment was done in collaboration with the infectious disease staff and with full disclosure to the patient of unknown risks. All three patients wanted to proceed with the therapy in lieu of losing their epidural catheters. In each case we followed the patients with repeat epidural cultures on a monthly basis for 3 to 6 months. In each case the epidural cultures were negative until one patient developed a second epidural infection at month 4 from a different organism. The longest time to postepidural infection was in a patient who had an epidural infusion that lasted for 1½ years without infection. There is currently no data on potential toxicity of vancomycin in the epidural space. Clearly, further examination of risk versus benefit in using this therapy is indicated. The standard of care continues to be device removal, aggressive systemic treatment, and replacement of the device after antibiotic therapy is completed.

Filters

Filtering should be a standard of care for chronic intraspinal infusions. The acute pain specialist argues that filters are a site of obstruction during therapy, a barrier to aspiration of the catheter when testing for the presence of spinal fluid, and have an unproved outcome with added cost. The lack of controlled studies adds fuel to this controversy. The use of filtered epidural catheters to prevent infection in long-term epidural catheters has been generally accepted and more commercially available epidural catheter sets are now being manufactured with in-line filters. We have noted that most epidural space infections occur in the absence of exit site inflammation and negative exit site cultures. It would seem logical that the contamination is occurring through the catheter rather than alongside the catheter. Contamination occurs at the time of bag and tubing changes, and any policy that increases the incidence of those changes will increase the risk of contamination. Figure 71-10 illustrates the Du Pen epidural catheter with the double PALL epidural filter configuration.

Intrathecal Device Infections

The intrathecal space may be accessed either by an externalized intrathecal catheter, port, or by placement of an intrathecal pump device. This technique has had limited use because of the fear of meningitis. Meningitis is exceedingly rare with the implanted pumps. The process of refilling the pump may result in contamination unless sterile conditions are maintained. We have had one pump-pocket infection requiring explant that was time correlated to an incident where multiple sticks were done at the time of a difficult refill. Reports of infections in implantable pumps are rare. Albright reported on infection during intrathecal baclofen therapy in children, a 5% infection rate “requiring pump removal” when utilizing the Medtronic’s system.96 Intrathecal vancomycin has been used to treat meningitis associated with pump infections,97 and in unrelated meningitis.98,99

Nitescu and his colleagues at the University of Gothenburg, Sweden, have reported on the use of intrathecal exteriorized catheters to treat both chronic pain and cancer-related pain. This approach has not gained support in the US literature. The fear of meningitis has always been such a great deterrent that its use has been limited. These reports have brought a new focus to the use of intrathecal catheters. The first report described their technique in a report detailing 142 cancer patients treated with externalized spinal catheters.100 The Portex nylon catheters were inserted at the lumbar level and tunneled paravertebrally up and over the shoulder. This first report included patients treated for 1 month using a syringe driver pump system. There were no reported cases of infection so the authors expanded the study in a second report to include another 89 patients. In this study the authors used a very low infusion rate of high concentration drug and reused CADD pump cassettes, but the cassettes were only changed or recharged on a monthly bases.101 Again, the authors reported no infections and there was no affect on the drug concentrations when the cassettes were left in place for a full month. Nitescu et al in their report did not make a point of the reuse of the cassettes or the monthly filter changes. Their conclusions were that this is a safe and effective means of intrathecal analgesic infusion when trained nurses exchanged the cassettes and filters on a monthly basis.102 In that same study, Nitescu, now reporting on 200 patients receiving the infusion over 7242 catheter days of use, had only a 0.5% rate of meningitis development. This low infection rate is far better than we were able to achieve during epidural infusions with filter and tubing changes every 48 to 72 hours. The only variable seems to be the frequency of “breaking into the system” or filter and tubing changes.

Another European group of investigators cited the use of externalized intrathecal systems because they were the “least costly and easiest to apply” for short periods.103 Fifteen patients were treated with externalized intrathecal systems; 13 of the patients had the system implanted at home. No infections were reported. Devulder and colleagues studied 33 patients with tunneled intrathecal catheters connected to subcutaneous ports. Three patients developed meningitis (9+%). The authors were able to correlate the 9% infection rate to accidental disconnections in the external tubing of the pump systems with presumed contaminant events; they further recommended in-line filters.104 Wagemans and others used externalized intrathecal catheters in 40 terminal cancer patients and reported meningitis in two patients, a 5% infection rate.105 We have used externalized intrathecal catheters in five cancer patients over the last 3 years. These patients had no epidural space available (tumor or surgical ablation of the epidural space), or in one case the catheter was intended for epidural placement and was found post-procedure to be in the intrathecal space. One of these patients was treated for up to 1 year with no reported infections. In each case the infusion was maintained at the lowest possible infusion rate with highly concentrated drugs, to keep the infusion rates at or below 2 mL/hour. The lower infusion rates allow a single infusion bag to last several weeks, with bag changes only done by the visiting nurses. Long-term studies are needed, but there may not be a cost advantage in the United States. In the United States, pumps and infusion bags are supplied by home care agencies who rent the pump and charge for the drug preparation; these charges may exceed the expectation of the managed care organizations and providers.

Drug-Related Complications

Infusion of analgesic agents into the epidural and intrathecal spaces may be associated with complications from unexpected reactions to drugs or drug combinations, drug errors, mixing errors, or infusion errors. Analgesic titration to effect or side effects is a standard of analgesic therapy. Those side effects are identified, classified for treatment priority, and treated to allow continued analgesic titration. The practitioner must be concerned about the complications of intraspinal analgesic drug therapy and how to identify, treat, and avoid drug-related complications.

Morphine appears to cause suppression of hormone production in some individual patients, resulting in endocrine alterations. Studies in animals, heroin addicts, and patients receiving long-term intraspinal morphine have demonstrated a decrease in testosterone associated with opioids.106,107 Paice and coworkers showed a decrease in testosterone in men and a decrease in FSH in women receiving intraspinal opioids.108 Patients exhibit decreased libido, fertility, and fatigue. Male patients may have erectile dysfunction and female patients may have amenorrhea. Testosterone and hormone replacement therapy has been successful in reversing these effects. These findings are not predictable and may be a rare occurrence. This is an area that the pain management community clearly needs to further study because sexuality can be a significant factor in our patient’s lives.

Preservative content of many opioid and other drugs used for intraspinal analgesia must be watched carefully. The addition of preservatives to multidose drug preparations is common in the United States, but less common in other countries. Practitioners must be aware of the total content of any drug preparations considered for compounding into neuraxial infusions. We published a case report of a patient who received a morphine preparation containing phenol and formaldehyde in a long-term epidural infusion that was associated with neuropathic pain and the development of epidural space adhesions.109 In this case the symptoms were reversed by removal of the preservative-containing morphine preparation. These errors may be corrected by education and information availability.

The infusion of concentrated analgesic combination into the neuraxial spaces have been studied by Yaksh, who has voiced concern over the use of untested drug combinations and concentrations and complications that may arise.110 Neurotoxicity with high-dose intraspinal drug therapy has been reported in humans. Interestingly, two reports in the literature indicate a higher degree of risk might exist in patients with preexisting neuropathology.111,112 Neuropathologic findings in 15 cancer patients status-post intrathecal morphine-bupivacaine therapy were reported by Sjoberg et al.113 The majority of neuropathologic changes were cancer-related. There were changes identified within the spinal canal; however, no changes could be correlated to the cumulative dose of morphine, bupivacaine, or the oxidants contained in the formulation in this small sample. The use of subcutaneous intrathecal infusion pumps in chronic pain patients requires the use of high concentrations of the infusion drugs. Without such concentrations, the pump would require frequent access and refilling. The interventionist should maintain a high level of attention to the current literature on the neurotoxicity of concentrated drugs on the spinal cord.

Local anesthetic toxicity is rare when analgesic dose concentrations are used during epidural infusions.34 Local anesthetic side effects include postural hypotension, skin anesthesia, and motor loss. These side effects are treated by vascular volume replacement and drug concentration adjustments. Once normovolemia is achieved, postural hypotension is rarely a problem unless the patient becomes dehydrated. High concentration local anesthetic agents have been used in intrathecal pumps; no neurotoxicity has been reported, but there have also been no controlled animal or clinical studies.

Clonidine is increasingly being used epidurally and intrathecally. Hypotension can be problematic, particularly early; however, volume replacement and slow upward titration is generally sufficient to normalize blood pressure. Intrathecal clonidine, although not approved for intrathecal use, is being used in intrathecal pumps using high concentration preparations. Yaksh has completed animal studies with concentrations as high as 6000 μg/mL (personal communication, 1998). These studies have not shown any signs of neurotoxicity. Controlled clinical trials are needed in this area to better define safety and efficacy. The more problematic issue is potential for rebound hypertension with abrupt discontinuation of spinal clonidine. Slow, downward titration of spinal clonidine, or prompt replacement with systemic clonidine, is indicated. Patients with intraspinal pumps may have mechanical failure of the pump; this unexpected failure could lead to a hypertensive crisis. Patients should be protected with an emergency kit of oral and transcutaneous clonidine.

Practical experience indicates that human error in drug administration needs to be considered. Pharmacy errors in mixing the drug for infusion have and will occur. Patients within our institution have received concentrations of drugs that were many times more or less potent than ordered, as well as the wrong drugs mixed in solution. These pharmacy errors are not predictable, so the clinician must consider these possibilities when the patient has an unexpected response to a recent refilling of the pump. When symptoms of drug withdrawal, overdose, or toxicity are noted soon after pump refills, the pump should be drained and refilled. The withdrawn solution should be saved for future analyses, if necessary. The multiple bag changes for patients receiving epidural infusions are a perfect setup for drug errors. Nursing errors in hanging the right solution for the right patient will occur, more likely in the hospital than in the home care setting. Patients within our institution have had hyperalimentation, heparin, and antibiotics inadvertently administered through permanent epidural catheters, both on the hospital floor and in the emergency room. It is important for the provider to be aware of these possibilities and to check all possible causes of any unexpected reaction and to correct the errors as they occur.

Mechanical Complications

Mechanical problems may be associated with the catheter, pump, filter, or the implanted device. One clinical dilemma associated with use of short-term epidural catheters is the failure of the adapter and connector. The connectors may become dislodged or the catheter material may become more brittle over time and break near the connector. The second most common problem is unexpected catheter withdrawal or migration outside of the epidural space. Tunneling the short-term catheter may add stability, but the durability of both catheter and connector are the major issues in maintaining the catheter for long-term use.114 Mechanical complications of both epidural and intrathecal drug administration have been alleviated to a great extent with the advent of the permanent silicone-rubber epidural catheters.

Epidural hematoma is exceedingly rare under normal conditions.115 Cancer patients with low platelet counts or those receiving anticoagulation therapy are, however, commonly encountered. Our policy, developed in close collaboration with the medical oncology staff, is to proceed if the platelet count is at least 20,000 without any signs of subcutaneous ecchymosis. Reversal of anticoagulation therapy is done on a case-by-case basis with consultation from the oncologist. This policy is consistent with the guidelines established by the consensus conference on anticoagulation and regional anesthesia organized by the American Society of Regional Anesthesia.

The most common cause of loss of analgesia encountered during epidural infusion is catheter or filter obstruction. Filter failure should be ruled out first when catheter obstruction is encountered. After catheter obstruction, failure of the externalized pump due to program input errors or an air bubble “lock” in the system is the next most common cause of analgesic failure. Utilizing aseptic technique and preservative-free saline, inject in the line above the filter and then below the filter to determine the source of the obstruction. Catheter obstruction may be further investigated by lumbar spine radiography or an epidurogram. Kinking of the catheter is easily seen on simple x-ray films. We have had one case of “crystallized” infusate in the catheter causing obstruction. This event occurred in a patient who had not been using the catheter for several months after obtaining pain relief with radiation therapy. The catheter was removed and subsequent dissection of the lumen revealed a solid mass of what was believed to be “old drug.” Subsequently, our policy became to flush all unused epidural catheters with 3 mL of preservative-free saline at least once a week. Once the catheter occlusion is diagnosed, then a surgical repair or repositioning of the catheter can be accomplished. Fibrosis around the catheter in the epidural space is discussed, but in our experience this occurs only in the face of a subacute epidural space infection. A survey study of 519 implanters reported a series of patients with granulomatous catheter tip masses in the intrathecal space.18 A second paper reports three cases of granuloma formation on an intrathecal catheter, with neurologic deficit. This a rare complication for which a cause has not been established, but infection and high concentration morphine have been suspected.16

Anderson and Burchiel reported that 20% of patients experience device-related complications with the implanted pump that require a repeat operation.11 General system failure of implantable infusion devices was reported by Paice15 and Penn as 21.6%, which is consistent with reports in the literature that vary from 10% to 40%. These system failures include catheter, pump, and connection problems. As with all multicenter, retrospective survey studies, the data are dependent on the rate of response from the implanters and the memory and recall of those clinicians.

Both external and internal pumps may be misprogrammed. It is important that two individuals check all pump programming of concentrated drugs that could endanger patient safety. A careful determination of the drug order and content of the infusion bag, pump infusion orders, and actual settings should follow any unexpected clinical outcome. In our practice, we felt the practice standards were in place, so we assumed this could not happen. We have had a major programming error event, which was corrected without sequelae; this event could have been avoided by simply having another nurse check the pump settings. The opioid-tolerant patient will tolerate most drug infusion errors, but may be placed in danger by errors of both drug concentration and infusion volume. Any abnormal patient response with signs of oversedation, neurologic deficits (not explained by tumor progress or metastatic lesions), or a combination of symptoms should alert the staff to turn down the pump significantly or stop until the cause of the symptom complex is determined. Consistent education and rational policies and procedures can minimize these human errors.

Interventional practitioners play an important role in facilitating analgesia where conventional techniques fail. Taking on a population of patients that requires neuraxial techniques can be a professionally and personally satisfying career decision. However, the provider must be prepared to tirelessly commit to a comprehensive approach to patient assessment, constant continuing education in spinal pharmacology, careful treatment planning, 24-hour availability, and continuous monitoring and follow-up. The basic research currently being done holds exciting promises for the future of neuraxial analgesia.