ELSEVIER

Journal of Back and Musculoskeletal Rehabilitation Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

Epidural steroid injections for the treatment of lumbosacral radiculopathy Stephen E. Abram* Department ofAnesthesiology, University of New Mexico School of Medicine, Albuguergue, NM 87131-5216, USA

Abstract

Objective: While there is an extensive body of literature concerning the use of epidural steroid injections in the treatment of sciatica, most of the literature is descriptive or anecdotal. There are few controlled studies regarding efficacy of this treatment modality. While there are few published reports of serious complications of this therapy, warnings about the hazards of epidural steroid injections occasionally appear in both medical and lay literature. It is the purpose of this review to assess the existing evidence for efficacy of epidural steroid injections for sciatica and to assess the risks of this procedure. Data sources: Peer reviewed medical literature from 1930 to the present was reviewed in order to survey reports regarding pathophysiology of radiculopathy, mechanism of action of epidural corticosteroids, controlled efficacy studies, reports on series of epidural steroid injections for sciatica, reports of adverse effects of epidural and intrathecal steroid injections, review articles of epidural and intrathecal steroid injections, and studies of the behavioral and histological effects of epidural steroids and their vehicle in animals. Study selection: Studies and review articles were selected from Medline search and from the author's files of older literature. Data synthesis: Results of this review are qualitative. It was felt that there was insufficient controlled data to analyze efficacy or safety studies in a quantitative fashion. Results: Radiculopathy following disc herniation appears to produce either mechanical or chemical nerve root inflammation. Epidurally injected corticosteroids most likely exert a beneficial effect through anti-inflammatory rather than direct analgesic mechanisms. Most descriptive studies report beneficial effects of epidural steroids in the majority of cases of radiculopathy, but not for other causes of low back pain. Most of the few controlled studies report epidural steroids to be more efficacious than placebo or epidural local anesthetic alone. Most patients who respond favorably continue to show improvement for many months. Several neurologic complications have been reported after intrathecal steroid injections, most following multiple intrathecal injections. Four cases of epidural abscess, one case of bacterial meningitis, and one case of aseptic meningitis have been reported following epidural steroid injections. Conclusions: The majority of the published literature supports the notion that epidural steroids provide relief of pain from lumbosacral radiculopathy. There is anecdotal evidence that multiple intrathecal steroid injections may be associated with neurological dysfunction, but there is very little evidence that epidural steroids are neurotoxic. © 1997 Elsevier Science Ireland Ltd. Keywords: Epidural; Corticosteroids; Radiculopathy; Efficacy; Complications

* Corresponding author. Tel.: + 1 505 2722610; fax: + 1 505 2721300. 1053-8127/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PH S1053-8127(96)00245-5

136

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

1. Introduction The use of epidural steroid injections for the treatment of lumbar radiculopathy has come into widespread use over the past two decades. Since the use of this therapy has continued to grow while other therapies such as chymopapain injection have been largely abandoned, it would seem that many are convinced of its safety and efficacy. Unfortunately, as with many established, widely used medical therapies, there have been relatively few studies that do more than describe the number of patients in a given population who appeared to get better when given the injections. While there is little evidence that epidural steroid injections, if used according to accepted practice, do any harm, there is also little rigorous evidence that they provide long-term clinical improvement. It is the purpose of this discussion to review the mechanisms of pain associated with radiculopathy and the rationale for the use of neuraxial steroids for sciatica, to present accepted treatment protocols, to present the evidence regarding the efficacy of epidural steroid injections, to evaluate the risk of side effects and complications, and to present some guidelines for the rational use of the technique. 2. Pathophysiology

The initial views regarding the pathophysiology of radiculopathy were based upon the concept that pain and neurologic dysfunction are the result of mechanical nerve root compression. Mixter and Barri were· the first to correlate the presence of disc herniation with symptoms of sciatica. A number of subsequent studies that examined the histological changes occurring in nerve roots adjacent to herniated discs prompted the conclusion that the ongoing mechanical compression of the affected root produced inflammatory and degenerative changes that may be responsible for the pain of sciatica [2,3]. Several authors have proposed that mechanical nerve root compression leads to inflammatory changes that produce increased intraneural pressure, osmotic and electrolyte changes and nerve root ischemia [4-6]. Parke

and Watanabe [7] emphasized the role of nerve root ischemia in the development of pathological changes. While theories that mechanical root compression produce ischemia and inflammation and can account for the histopathological changes seen in cadaver and tissue biopsy studies, they fail to account for symptoms of sciatica seen in patients who do not demonstrate any appreciable evidence for mechanical distortion of the nerve root or for the failure of development of symptoms among patients who are incidentally discovered to have substantial nerve root compression [8,9]. Murphy [4] and Marshall et al. [lO] proposed that the inflammation evoked in the nerve root by disc herniation may, in certain instances, be related to chemical irritation initiated by material from the extruded disc. Several investigators suggest that material from the nucleus pulposus sets up an immunologic response that triggers the subsequent pathological changes in the nerve root [10-14]. However, most recent studies have emphasized the role of the nuclear material in inducing regional inflammation. Studies involving epidural injection of material from the nucleus pulposus have demonstrated meningeal inflammation in both dogs [15] and primates [16]. McCarron's study [15] demonstrated that the inflammatory response involved the nerve roots and spinal cord as well. Similarly, Olmaker et al. [17] showed that autologous nucleus pulposus, implanted surgically in the epidural space in pigs, induced cauda equina and nerve root injury far in excess of that seen when retroperitoneal fat was implanted. The role of the nucleus pulposus in initiating an inflammatory response in the nerve root was further elucidated by Saal et al. [18]. They collected human disc samples removed at surgery from patients with symptomatic radiculopathy and analyzed the material for activity of phospholipase A2 (PLA 2), the enzyme responsible for initiating the release of arachidonic acid from cell membranes, thus activating the cascade of events leading to the production of prostaglandins and leucotrienes. They found extremely high PLA 2 levels in this material, 20-100-fold higher than activity from human inflammatory synovial effu-

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (J997) 135-149

sion, establishing further evidence that release of nucleus pulposus from the intervertebral disc may result in local inflammation. 3. Rationale for the use of epidural steroid injections At the time epidural and intrathecal injections of steroids were introduced into clinical practice, it was assumed that mechanical nerve root compression resulted in symptoms of sciatica because of the resultant inflammatory response. The test devised to determine whether the local application of steroids would reduce the inflammatory response was to simply inject steroids epidurally in patients with sciatica and determine whether the pain improves. Since the animal models of nerve root inflammation were not undertaken until epidural steroid injections had become accepted practice, there have been no studies to determine whether the inflammatory process is indeed reduced or whether patients experience improvement in symptoms for other reasons. There are reasons other than reduction of inflammation that corticosteroids might affect pain perception in the presence of nerve root pathology. Ectopic discharge originating from experimentally created neuromas is inhibited by corticosteroids whether administered at the time of injury or after abnormal neural activity is established [19]. It is possible that this membrane stabilizing effect of corticosteroids may be in part responsible for the improvement in symptoms associated with epidural steroid injections. In experimental models of neuropathic pain, loose ligation of the sciatic nerve or tight ligation of the L5 and L6 nerve roots in rats is associated with the subsequent appearance of thermal and mechanical hyperalgesia [20,21]. This hyperalgesia is thought to be at least partially related to impulses originating from the injured nerve segment, as evidenced by reduction in thermal hypersensitivity following low-dose systemic lidocaine [22]. In addition, there are N-methyl-Daspartate (NMDA) receptor-initiated changes in spinal cord function resulting in enhanced responses of dorsal' hom neurons following both noxious and non-noxious stimulation. This spinally

137

mediated hyperalgesia appears to be at least partially mediated by the production of prostaglandins. It is conceivable, therefore, that drugs which block production of prostaglandins in the spinal cord may have a beneficial effect on pain associated with nerve root irritation or injury. Small intrathecal doses of non-steroid antiinflammatory drugs (NSAIDs) have been shown to reduce the acute hyperalgesic behavior associated with spinal administration of NMDA or substance P in rats [23]. Spinally administered NSAIDs have also been shown to reduce the spinal sensitization induced by a noxious stimulus (subcutaneous formalin) [24]. Since corticosteroids are capable of reducing production of prostaglandins through the inhibition of PLA 2 [25], it is possible that epidural steroids may exert their analgesic effect through the blockade of noxious stimulus-induced spinal sensitization. To test that hypothesis, Abram et al. [26] examined the effect of either acute or chronic pretreatment with intrathecal triamcinolone diacet~te on the rat formalin test. There was minimal effect of this intervention on phase 2 flinching behavior, suggesting a lack of effect on pain-induced spinal sensitization. Similarly, Coderre- et al. [27] were unable to show an effect of a PLA 2 inhibitor on the second phase of the formalin test. 4. Epidural steroid treatment protocol While there is some variation among treatment protocols there is remarkable consistency among techniques reported in recent years, perhaps in part because of the widespread dissemination of information regarding the technique in the 1970s by a small number of individuals such as Carron [28] and Winnie [29]. Following is a suggested protocol, which represents an opinion derived from review of a large number of reports, several review articles [28,30-33] and personal experience.

4.1. Patient selection It is generally agreed that the principle indication for epidural steroid injections is the presence of pain related to nerve root pathology, i.e. pain in a radicular distribution. There is little evidence

138

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

for beneficial effects in the presence of mechanicalor non-radiating back pain. In addition, there are a number of non-radicular causes of pain that are likely to produce referred or radiating leg pain that are not related to nerve root pathology, such as myofascial pain syndrome and facet and sacroiliac arthropathy, which are unlikely to benefit from this procedure. Therefore, it is essential that there be a careful assessment prior to considering epidural steroid injections in order to determine whether the pain is of nerve root origin. For patients whose pain is secondary to lumbosacral radiculopathy, there is considerable variation in response. Pain related to disc herniation appears to be highly responsive to epidural steroids, although previous back surgery [34,35] and chronicity [36] are associated with lower success rates. Patients with spinal stenosis were reported to experience as much benefit from epidural steroid injections as those with a diagnosis such as disc herniation [37]. However, 'no discussion of the basis for the spinal stenosis was given in that study. When the diagnosis is based on symptoms of neural claudication, (multi-segmental pain, weakness and sensory loss which comes on over time- with walking) [38], epidural steroids appear to be of little help [33]. There is some controversy regarding the extent of patient workup required prior to performing an epidural steroid injection. Unfortunately, there are few studies that correlate treatment success with either physical examination findings or the results of imaging studies. Rowlingson and Kirschenbaum [39] found the highest correlation with treatment success with cervical epidurals in patients who exhibited dermatomalloss of sensation. Hickey [40] reported a high success rate (84%) from epidural steroids in a group of patients with a 51% incidence of dermatomal neurologic deficit and a 95% incidence of positive nerve stretch signs. A review by Rowlingson [32] suggests a rather poor correlation between imaging study results and response to epidural steroids. One might argue, therefore, that it would be most cost effective to perform epidural steroid injections on the basis of clinical signs and symptoms

of radiculopathy and reserve imaging studies for those patients for whom surgical intervention is contemplated. An additional issue, however, is the remote possibility that radicular symptoms are related to spinal tumor, infection or other condition that could be made worse by injections or corticosteroids. The issue of performing expensive diagnostic procedures for the sole purpose of ruling out rare conditions is one that the medical and legal community and the public will need to address jointly. Another medicolegal issue is the publication of contentions that epidural steroid injections can cause arachnoiditis [41-43]. While these claims are based on purely theoretical grounds, some physicians may feel that pre-treatment imaging studies are needed to rule out the presence of pre-existing arachnoiditis, which may occur as a result of disc herniation [44]. This precaution is taken not because arachnoiditis is a contraindication to epidural steroids, but to avoid blaming the procedure for producing that condition. There is evidence that epidural steroid injections are not beneficial for patients with nonradicular pain [36]. Unfortunately, they are frequently carried out in patients who lack convincing evidence of radiculopathy.

4.2. Injection technique The two preparations in common use for epidural steroid injections are both suspensions of insoluble acetates: methylprednisolone acetate (Depo Medrol, Upjohn, 'Kalamazoo, MI) and triamcinolone diacetate (Aristocort Intralesional, Fujisawa USA, Deerfield, IL). Depo Medrol is available in two formulations. The formulations for these products are shown in Table 1. It is postulated that these drugs provide high corticosteroid tissue levels for a period of about 2 weeks. Soluble preparations are generally avoided, as their effects locally are transient, and some soluble preparations are associated with the production of seizures when injected spinally in animals [45]. Local anesthetic is often injected with the steroid. The purpose is mainly to document correct drug placement. The addition of several milliliters of anesthetic will generally produce

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

139

Table 1 Commercial preparations of methylprednisolone acetate and triamcinolone diacetate Depo-Medrol multi-dose vials (Upjobn, Kalamazoo, MI) Methylprednisolone acetate 20 mg Polyethylene glycol 3350 29.5 mg Polysorbate SO 1.97 mg Monobasic sodium phosphate 6.9 mg Dibasic sodium phosphate USP 1.44 mg Benzyl alcohol 9.3 mg NaCI to adjust tonicity

40mg 29.1 mg 1.94 mg 6.Smg 1.42mg 9.16mg

Depo-Medrol single dose vials (Upjobn, Kalamazoo, MI) Methylprednisolone acetate 40 mg Polyethylene glycol 3350 29 mg Myristyl-gamma-picolinium chloride 0.195 mg NaCI to adjust tonicity

SOmg 2Smg 0.IS9 mg

Aristocort Intralesional (Fujisawa USA, Deerfield, IL) Triamcinolone diacetate Polyethylene glycol 3350 Polysorbate SO Benzyl alcohol NaCI

SOmg 2S.2mg 1.SSmg 6.59mg l.37mg S.SSmg

25mg 30mg 2mg 9mg S.5mg

prompt reduction of symptoms, hypoesthesia in one to three segments and improvement in nerve stretch signs if the affected nerve root is reached. In addition, if the local anesthetic is injected prior to the steroid, accidental subarachnoid placement of the steroid can be avoided, as the local anesthetic will produce rapid and extensive motor and sensory blockade if placed intrathecally. The author's usual technique is to inject 2 ml 1% lidocaine initially. If no appreciable motor or sensory block is present within 5 min, this is followed by 50 mg (2 ml) triamcinolone diacetate and an additional 2 ml 1% lidocaine. It is generally agreed that the drug should be injected as close to the affected segment as possible. Positioning the patient in the lateral decubitus position with the affected side down will theoretically allow the drug to remain in the vicinity of the affected root in high concentrations. Very often, the patient's radicular symptoms are transiently reproduced as the needle enters the epidural space and again as the drugs are injected. This finding would seem to indicate the presence of nerve root irritability, and is generally prognostic of a beneficial response. The typical response is

relief of symptoms for the duration of the local anesthetic effect, followed by 12-48 h of recurrence of symptoms, sometimes worse than preinjection, then gradual reduction in pain over the next 7-10 days. Patients are usually instructed to limit their activities during the first few postinjection days, but bed rest is not often prescribed.

4.3. Protocol for repeat injections Most protocols allow for 1-2 weeks before reconsideration of further treatment. If at the time of re-evaluation symptoms have subsided dramatically, with minimal residual pain, no further injections are indicated. A common practice is to perform a series of three injections even if there is marked improvement from the initial procedure. There is no evidence that such practice is beneficial, and there is clearly a negative impact on cost and risk to the patient. Such practice should be .discouraged unless well controlled outcome studies suggest otherwise. If there is no improvement at all at the end of that period, it is unlikely that repeat injections will be beneficial (though some authors argue

140

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

that a number of patients respond favorably to a second injection under those circumstances [46,47]. The exception is the situation in which the steroid did not reach the affected root, i.e. the drug was not injected epidurally or it was injected epidurally but failed to spread to the affected segment. In that case, it is reasonable to repeat the injection, possibly at an adjacent interspace. In the case of an S1 radiculopathy, it may be preferable to introduce the drug via the sacral

hiatus, adding normal saline to enhance spread to the upper sacral roots. If, at the time of re-evaluation, there is partial resolution of symptoms, it is reasonable to repeat the injection and re-evaluate after another 1-2 weeks. The same paradigm is usually carried out after the second injection, with cessation of therapy when the patient is greatly improved or when the preceding injection failed to produce any added benefit. Most protocols suggest cessation

History and Physical Examination: Radicular Signs and Symptoms?

yes/~NO~ Epidural s 4 0 n

No Epidural Injection

t

OtherRx

Return in 1-2 Weeks

t ~-------~ No Improvement ·Complete Relief

Partial Improvement

t

Repeat Epidural

t

Other Rx

t

Reevaluate in 2-4 weeks

t

Return in 1-2 Weeks

t~No Improvement" Complete Relief

Partial Improvement ~

,

t

Repeat Epidural

~

,

Other Rx

Recurrence of symptoms

4 after 1st or 2nd injection or

~

, r . > 6 weeks after 3rd injection

.

Reevaluate in 2-4 Weeks

.

Followup in 2-4 Weeks

t~-------~ No Improvemen~ Complete Relief

Partial Improvement

- - - - - - - -_ _...:::! .......

Table 2. Algorithm for treatment of sciatica with.epidural steroid injections

~/ Repeat Epidural x 1 ~

Further recurrence or dissatisfaction with result at any point in algorithm

t

Additional Workup or Alternative Rx

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

of therapy after the third procedure to minimize the risk of prolonged adrenal suppression and the development of cushingoid symptoms. In addition, while the risks of spinal neurotoxicity appear to be very low using such a protocol, there is little data available to show that repetition of these injections over long periods of time is safe. However, if a patient is improved over a several week period and symptoms begin to recur, it would seem reasonable to provide one more injection at that time. A suggested algorithm for repeat injections and follow-up visits is shown in Table 2. Patients with persistent or recurrent sciatica following lumbar disc surgery are less likely to respond to epidural steroid injections [35,36]. It is common to find a lack of immediate response to epidural local anesthetics in these patients, indicating the drug may have failed to reach the offending nerve root. Epidural scaring or adhesions are thought to account for at least some cases of persistent radicular symptoms in postoperative patients [33,48]. Techniques that involve the introduction of radioopaque epidural or caudal catheters under fluoroscopy into close proximity of the affected root have been described [49,50]. In addition to local anesthetic and corticosteroids, several other injectable substances have been proposed for use with this technique, including hypertonic saline [50] and hyaluronidase [33]. High fluid volumes are recommended to help break up adhesions. Epiduroscopy plus injection of hypertonic saline, hyaluronidase and steroids is also being used for such situations. Unfortunately, there are virtually no outcome studies of these techniques to determine their safety or long-term efficacy. It is essential that such studies be carried out before there can be general acceptance of these techniques. The price in terms of both financial costs and complications is potentially high. 5. Treatment outcomes There are over 50 studies in the English language literature, involving over 6800 patients, that report beneficial results from epidural steroid injections. None of these studies used placebo controls, and few involved more than short-term fol-

141

low-up. In contrast, there are only six controlled studies, involving about 200 patients, that compare efficacy of steroid epidurals to either local anesthetic injections or placebo. Three of these studies demonstrated that epidural steroids were more effective than epidural local anesthetic [52-54], two showed epidural steroids to be more effective than placebo [47,53], one showed no difference between epidural steroid and placebo [54], and one showed no difference between epidural steroid and epidural local anesthetics. This last study was criticized because patients were evaluated only 24 h after injection, before most patients exhibit any .improvement, and patients who had not responded were crossed over to the opposite treatment groups. The results of these controlled studies are discussed in greater detail in reviews by Benzon [30] and Rowlingson [32]. It should be emphasized that all of the controlled studies that showed improvement following epidural steroids involved therapy for radicular pain only. It has been suggested that, although many patients appear to show short-term improvement, there is little evidence that epidural steroid injections produce prolonged benefit. Abram and Hopwood [36] followed 212 patients treated with epidural steroid injections for lumbar radiculopathy over a 6-month period to determine whether patients who experienced short-term (at least 2 weeks) benefit continued to show improvement over a longer interval. The study reported a short-term success rate of 54%. At the end of 6 months patients who initially showed a favorable response to treatment were compared to patients who initially failed to benefit. After 6 months, 79% of patients in the initial success group reported a reduction in their pain level, while only 36% of the failure patients reported pain reduction over that time period. At 6 months, 66% of the success group were employed (up from 47% pretreatment), while 29% of the failure group were working (up from 25% pretreatment), and 52% of the success group reported marked reduction in pain medication use vs. 36% in the failure group. These data suggest that the patients who reported improvement soon after epidural steroid treatment continued to improve over the ensuing

142

S.E. Abram / ]oUnfal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

6 months and were still doing better than patients who failed to demonstrate initial benefit. A number of pretreatment variables are associated with variations in treatment outcome. The fact that social, psychological and vocational factors can influence treatment outcome attests to the multidimensional character of low back pain. A retrospective study by Abram and Anderson [35] reported that long duration of symptoms, previous back surgery, being unemployed as a result of pain, being injured at work, low educationallevels and several psychological factors such as helplessness and hopelessness were predictive of low treatment success. A more recent study by Hopwood and Abram [56] revealed a number of similar correlations. However, when these correlations were analyzed for the effects of multiple variables through a logistic regression analysis, only a few variables were found to be independently predictive of poor treatment outcomes. These factors were: unemployment as a result of the painful condition, long symptom duration, a diagnosis other than radiculopathy, and smoking. 6. Side effects and complications During the 30 years since the introduction of epidural steroid injections for the treatment of lumbar radiculopathy, there have been many published series reporting the clinical effects of this therapy. Aside from occasional cushingoid symptoms or side effects associated with the introduction of the epidural needle (backache, dural puncture headache) or caused by added local anesthetics, few of these reports have documented any complications. Only a handful of case reports of adverse effects have been published over the years. Nevertheless, a few publications have warned of potential serious complications from epidural steroid injections, citing mostly theoretical reasons for the perceived dangers. Despite the lack of epidemiological evidence for substantial risks from this treatment, the comments of these few authors, combined with sensationalistic television reporting, have led to all but abandonment of this treatment in Australia. One of the early reports of the potential haz-

ards of neuraxial steroid administration came in a publication by Nelson et al. [57], who cited two cases of adhesive arachnoiditis and one case of aseptic meningitis among 23 patients who had undergone 83 intrathecal (IT) injections of methylprednisolone acetate (MPA). Other reports of adverse reactions following IT administration of corticosteroid suspensions include a case of proliferative arachnoiditis [58] and a case of calcific arachnoiditis [59] following multiple IT injections of MPA, bacterial meningitis (2 cases) after IT steroid injections for sciatica [60], and conus medullaris syndrome following multiple (14) IT injections of MPA for sciatica [61]. A retrospective review by Roche [41] in 1984 reported the occurrence of myelographic changes consistent with arachnoiditis among 80 patients. Of these patients, 18· had received intrathecal steroid injections. In a subsequent review of these same patients [62], the author points out that only three of these patients had experienced clinical symptoms consistent with arachnoiditis. Roche concludes that the changes observed must have been related to the intrathecal steroid injections, pointing out that 13 of the patients had had no other drugs injected spinally. As indicated in a letter by Berrigan et al. [63], this study was seriously flawed. First, two of the three patients who experienced clinical symptoms compatible with arachnoiditis had undergone laminectomy, and their symptoms may have been associated with either disc disease, which itself has been associated with arachnoiditis in the absence of any intraspinal injections [44,64] or with changes occurring as a result of surgery [65]. Second, the patients in Roche's series had back pain severe enough to warrant myelography, and many of these patients may have had arachnoiditis prior to their steroid injections. Indeed, intrathecal steroid injections have been reported to have beneficial effects among patients with arachnoiditis [66,67]. Third, documentation of the dates of injection and the drugs given is vague for some patients in the series. Nelson has published several statements critical of the use of intrathecal steroids [43], and has extrapolated the presumed danger of intrathecal

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

steroids to include epidural steroid injections. In addition to citation of cases of arachnoiditis occurring in patients who had received intrathecal steroid injections, Nelson cited several lines of experimental evidence that the vehicle used in steroid suspensions, 3% .polyethylene glycol, is neurotoxic. His first reference, Margolis et al. [68], studied a preparation called Efocaine, a neurolytic agent containing butyl-p-aminobenzoate, procaine, 78% propylene glycol and 2% polyethylene glycol 300. In addition to studying the effect of the Efocaine, they also evaluated the effects of the solvent (propylene glycol 78% plus polyethylene glycol 2%) and propylene glycol 80%. Contrary to Netson's published statement [43], they did not evaluate polyethylene glycol alone. The second citation, by Hurst [69], evaluated the toxic effects of several detergents, not polyethylene glycol. The third citation, by Chino et al. [70], evaluated the neutoxic effects of 80% propylene glycol (the reference was cited incorrectly by Nelson in both the text and the reference section as polyethylene glycol). The fourth reference, by Mackinnon et al. [71], reported neuropathic changes from MPA only when it was injected intrafascicularly. The last citation, by Selby [42], is a two paragraph letter to the editor that stated a conclusion, it appears that both Depo-Medrol® and its sterile vehicle, polyethylene 4000, can immediately result in the dissolution of myelin ... ', with virtually no supporting data. Nelson's conclusion that intrathecal MPA is neurotoxic is based on purely anecdotal evidence. His conclusion that epidural MPA is neurotoxic is totally unfounded. He states: 'When this steroid compound is injected epidurally, it probably transudes the dura mater and the arachnoid membranes via the arachnoid villi.' No evidence for this statement exists. In fact, it has been shown that phenol, a drug that clearly has a neurolytic effect on the spinal cord when injected IT, and is a much smaller molecule than polyethylene gly·col, has no demonstrable histological effect on the adjacent spinal cord when injected epidurally [72]. The following section will present a review of the English language literature documenting side effects and complications of epidural steroid in-

143

jections and potential adverse effects of steroid suspensions and their vehicle in animal models.

6.1. Systemic corticosteroid effects Since steroid suspensions produce prolonged effects locally, it is not surprising that they can cause persistent systemic effects as well. Kay et al. [73] reported suppression of ACTH secretion, lasting as long as 7 weeks, following a series of three epidural steroid injections. Cushingoid symptoms, such as facial edema, buffalo hump, supraclavicular fat pads, easy bruising and scaly skin lesions, have been reported to last for periods of weeks to months following epidural steroid injections [74-77]. Diabetic patients frequently report elevated blood and urine glucose levels and increased insulin requirements for a few days to a few weeks after steroid epidurals.

6.2. Technical complications and minor side effects Large numbers of patients report a transient increase in radicular pain or paresthesias after epidural steroids, but these exacerbations do not appear to be associated with serious sequelae [78]. One case of an epidural hematoma requiring surgical decompression was reported followlng a cervical epidural steroid injection. No predisposing factors were present except for chronic indomethacin use. Accidental dural puncture with post-dural puncture headache is a commonly reported complication [78]. Introduction of subarachnoid or subdural air during attempted epidural injection can produce an immediate headache that usually lasts a period of hours [79,80]. There has been one documented case of a true allergy to triamcinolone diacetate following an epidural steroid reaction [81]. It is not known whether the steroid or a component of the vehicle acted as the triggering antigen. The reaction began a week after the injection and consisted of skin rash, itching, facial swelling, dyspnea abdominal pains and diarrhea. A repeat challenge of 0.1 ml of the same preparation caused recurrence of symptoms after a 12-h delay.

144

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (J 997) 135-149

6.3. Neurological complications There have been several reports of arachnoiditis occurring after intrathecal injections of methylprednisolone acetate. Almost all of these occurred after multiple injections over fairly long periods of time [57-59]. There have been no case reports of arachnoiditis following epidural steroid injections. Several cases of aseptic meningitis have been documented following intrathecal steroid injections [57,82,83]. There were no permanaent sequelae following any of these reported cases. Two prospective studies evaluated the potential for intrathecal steroids to cause aseptic meningitis. Goldstein et al. [84] found no increase in CSF protein, lymphocytes or polymorphonuclear cells in patients with multiple sclerosis treated with 40 mg intrathecal methylprednisolone acetate. There were no clinical signs of meningeal irritation. Seghal et al. [85] examined the effects of intrathecal methylprednisolone acetate in doses ranging from 40 to 200 mg. They reported a dose-related, predominantly polymorphonuclear pleocytosis beginning 24 h after injection and lasting about 6 days. Doses above 80 mg produced elevated CSF protein and clinical signs of meningeal irritation. There has been only one case report of aseptic meningitis following an epidural steroid injection. This occurred in a patient who received methylprednisolone acetate without local anesthetic, so it is impossible to know whether the drug was given intrathecally. Two cases of bacterial meningitis have been reported in association with epidural ster;oid injections [86]. In one case, a dural puncture was documented. There was no CSF return in the second case, but dural puncture could not be ruled out, as no local anesthetic was injected. There have been five reported cases of epidural abscess after steroid epidurals. In one case, which followed a series of four epidural methylprednisolone acetate injections, purulent material containing inflammatory cells and squamous carcinoma cells was drained at laminectomy, but cultures were negative [87]. It is not clear whether the abscess began as an infectious process that failed to produce positive cultures because of

antibiotic treatment or whether the material drained from the epidural space represented necrotic material from a metastatic tumor. A second case was reported in a diabetic patient treated with triamcinolone diacetate. Cultures grew Staphylococcus aureus [88]. Another epidural abscess occurred in a diabetic patient after three injections of methylprednisolone acetate [89]. Two weeks previously that patient had undergone a surgical sympathectomy which had been complicated by Staphylococcus aureus sepsis. Epidural cultures grew the same organism. A fourth case occurred in a patient treated with continuous epidural analgesia [90]. The patient received bupivacaine and 120 mg methylprednisolone acetate via an epidural catheter that was left in place for 24 h. Four days later a second epidural catheter was placed, and bupivacaine alone was infused over a 3-day period. The abscess became evident 3 weeks after the second catheter was removed. A fifth case occurred following three injections of methylprednislolone acetate plus bupivacaine in a patient with postherpetic neuralgia [91]. Cohen [61] reported a case of persistent urinary incontinence and sacral anesthesia in a patient who had received multiple epidural and intrathecal injections of methylprednisolone acetate. The patient was a 35-year-old woman who had previously undergone four myelograms and four lumbar spine operations. She was treated with 14 intrathecal and four epidural injections over an 18-month period. Her neurologic deficits were not evident until after the last IT injection.

6.4. Experimental neurotoxicity Several animal studies have evaluated the effect of depo-steroids or polyethylene glycol vehicle on the spinal cord, nerve roots and meninges. Cicala et al. [92] examined the effect of methylprednisolone acetate on the meninges and nerve roots in cats at 4 and 10 days after epidural injections. They found no evidence of inflammatory changes in either meninges or nerve roots. Delaney et al. [93] performed similar studies in cats using triamcinolone diacetate but carried out their histological studies at 30 and 120 days postinjection. They reported mild inflammatory

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (J997) 135,,-149

changes at 30 days, but no evidence of any pathological changes after 120 days. Abram et al. [26] treated rats with intrathecal triamcinolone diacetate or normal saline four times at 5-day intervals and histologically examined spinal cords 48 h after the last injection. They found occasional large, dark-staining neurons suggestive of neuronal injury in both steroid and saline-treated animals. The incidence was similar in the two groups and they were invariably seen in areas adjacent to the site of the spinal catheter where there was evidence of mechanical cord compression. None of the animals in either group showed any behavioral evidence of neurological dysfunction. Two studies examined the effect of steroid suspensions on peripheral nerve function and histology. McKinnon et al. [71] injected rat sciatic nerves with methylprednisolone acetate either perineurally or intrafascicularly. There was evidence of moderate neural degeneration with intraneural injection but no histopathology after perineural application. Wood et al. [94] found demyelinization, axonal degeneration and inflammatory changes following multiple sciatic nerve injections of methylprednisolone acetate in rats. Since the injections were done percutaneously, ,there is no way of knowing whether any of the injections were intraneural. None of the animals showed any behavioral evidence of neurological dysfunction [95]. One study has examined the effect of polyethylene glycol alone on nerve conduction. Benzon et al. [96] exposed both sheathed and de sheathed rabbit nerve fibers to different concentrations (3-40%) of polyethylene glycol. The 3% and 10% concentrations had no effect, 20% and 30% significantly slowed conduction, and 40% abolished conduction. Even with the highest concentration used, the compound action potentials returned to at least 80% of baseline following washout. 8. Conclusions It is likely that epidurally administered steroids provide clinical benefit in patients with symptoms of radiculopathy because of their anti-inflammatory properties as opposed to analgesic effects.

145

While most of the evidence for the salutory effects of epidural steroid injections for radiculopathy is anecdotal, a few controlled studies have been done, most of which indicated that epidural steroids are more effective than epidural local anesthetic or placebo. The few studies that have evaluated the long-term effect of epidural steroids suggest that the benefits are prolonged. There is no evidence that epidural steroid injections are effective for non-radicular low back pain. It is essential, therefore, that the treating physician perform a careful history and physical examination in order to determine the source and mechanism of the patient's pain. In complicated conditions in which mechanisms are not obvious, appropriate imaging studies should be utilized. Since most studies that attest to the efficacy of epidural steroids have utilized the procedure in conjunction with other conservative measures, such as appropriate back exercise, reconditioning and psychological interventions, epidural steroids should be used as a part of programs incorporating such measures, not as an alternative. This is particularly important for the more chronic cases. While there is little experimental evidence that intrathecal steroid injections produce inflammatory or neurotoxic changes, there are a few reports of aseptic meningitis and arachnoiditis following repeated intrathecal injections of methylprednisolone acetate. There is very little evidence that epidural steroid injections cause these conditions. Bacterial meningitis following intrathecal steroid injections and epidural abscess after epidural steroids are rare but real complications. 8. Recommendations for epidural steroid use The following recommendations, representing the author's opinion regarding the safe and effective use of this treatment, are based upon examination of the existing peer-reviewed literature and personal experience. They do not necessarily represent a consensus opinion, nor are they intended to serve as a guide to the use of therapy in any individual patient, whose treatment should be determined on the basis of an assessment of risks and benefits for that individual.

146

1.

2.

3.

4.

5.

6.

7.

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (J 997) 135-149

Epidural steroids should be used only in those patients who demonstrate clinical and/or laboratory evidence of radiculopathy. Patients should be carefully questioned about symptoms of infection, even if remote from the site of injection, coagulopathy or previous untoward reaction to coricosteroids. Potential benefit of epidural steroid injections should be weighed against the increased risk of infection and hyperglycemia in diabetic patients. The reduced potential for benefit from epidural steroid should be considered when making theapeutic decisions for those patients whose chances for improvement are low, e.g. patients with prolonged symptoms, particularly > 1 year, patients with previous back surgery, preoccupation with vocational or reimbursement issues, history of substance abuse or heavy smoking. Intrathecal steroid injection should be avoided in treating patients with radiculopathy, more because of medicolegal implications than because of potential harm to the patient. A test dose of local anesthetic will help rule out intrathecal needle position. If the dura is punctured it may be preferable to postpone steroid injection for a few days. Steroid doses should be kept at moderate levels, e.g. 80 mg methylprednisolone acetate or 50 mg triamcionolone diacetate, unless future studies show a greater benefit and no greater risk from higher doses. The number of injections given to an individual patient should be minimized and should be tailored to patient response. Patients doing well clinically after one or two injections should not be given repeat injections. In general, no more than three injections should be done during the initial series. A single repeat injection may be indicated if symptoms recur weeks to months after the initial treatment. The unknown risks of newer techniques, such as the addition of hypertonic saline or hyaluronidase, should be considered carefully in the therapeutic decision making process. As new techniques are described in the literature, it is important that animal models be

developed to help assess possible adverse reactions. It is also important that new techniques be introduced under hospital and medical school IRBs and be subjected to the scrutiny of controlled studies to determine safety and efficacy. While epidural steroid injections appear to have some benefit, particularly in patients with clearly defined radiculopathy who have failed to respond to initial conservative measures, there is probably considerable overuse and abuse of the technique. Use of lumbar epidural steroid injections for non-radiating or extremely chronic low back pain or the use of cervical epidural steroid injections for tension headache is probably widespread. Continued therapy despite resolution of symptoms, multiple repetition of procedures over months or years, and utilization for non-indicated procedures represent a waste of health care dollars. Under ideal circumstances, there is a 50-60% success rate from steroid epidurals. With inappropriate use, the overall success rate is probably much lower. As the technique is modified and extended to other indications, studies regarding safety and efficacy must accompany these modifications. References [1] Mixter WJ, Barr JS. Rupture of the inteIVertebral disc with involvement of the spinal cord. N Engl J Med 1934;211:210-215. [2] Lindahl 0, Rexed B. Histological changes in spinal neIVe roots of operated cases of sciatica. Acta Orthop Scand 1951;20:215-225. . [3] Kelly M. Pain due to pressure on neIVes? Spinal tumors and the inteIVertebrai disc. Neurology 1956;6:32-36. [4] Murphy RW. NeIVe roots and spinal neIVes in degenerative disc disease. elin Orthop 1977;129:46-60. [5] Rydevik B, Brown MD, Lundborg G. Pathoanatomy and pathophysiology of neIVe root compression. Spine 1984;9:7-15. [6] Kirkaldy-Willis WH. The relationship of structural pathology to the neIVe root. Spine 1984;9:43-52. [7] Parke WW, Watanabe R. The intrinsic vasculature of the lumbosacral spinal nerve roots. Spine 1985;10:508-515. [8] McRae DL. Asymptomatic inteIVertebrai disc protrusions. Acta Radiol (Stockh). 1956;46:9-13. [9] Hitzelberger WE, Witten RM. Abnormal myelograms in asymptomatic patients. J Neurosurg 1968;28:204.

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

[10). Marshall LL, Trethwie ER, Curtain Cc. Chemical radiculitis. Clin Orthop 1977;129:61-67. [11) Gertzbein SD. Degenerative disk disease of the lumbar spine: Immunological implications. Clin Orthop 1977;129:68-71. [12) Bobechko WP, Hirsch C. Autoimmune response to nucleus pulposus in the rabbit. J Bone Joint Surg 1965;47B:574-580. [13) Naylor A. Enzymatic and immunological activity in the intervertebral disc. Orthop Clin North Am 1975;6:51-58. [14) Pankovitch AM, Korngold.L. A comparison of the antigenic properties of nucleus pulposus and cartlage protein polysaccharide complexes. J Immunol 1967;99:431-437. [15) McCarron RF, Wimpee MW, Hudkins PG, Laros GS. The inflammatory effect of nucleus pulposus. A possible element in the pathogenesis of low back pain. Spine 1987;12:760-764. [16) Haughton VM, Nguyen CM, Ho K-C. The etiology of focal arachnoiditis. An experimental study. Spine 1993;18:1193-1198. [17) Olmarker K, Rydevik B, Nordborg C. Autologous nucleus pulposus induces neurophysiologic and ,histologic changes in porcine cauda equina and nerve roots. Spine 1993;18:1425-1432. [18) Saal JS, Franson RC, Dobrow R et al. High levels of phospholipase A2 activity in lumbar disc herniations. Spine 1990;15:674-678. [19) Devor M, Govrin-Lippman R, Raber P. Corticosteroids suppress ectopic neural discharge originating in experimental neuromas. Pain 1985;22:127-137. [20) Bennett GJ, Xie YK. A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 1988;33:87-107. [21) Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 1992;50:355-363. [22] Abram SE, Yaksh TL. Systemic lidocaine blocks nerveinjury induced hyperalgesia and nociceptor-driven spinal sensitization in the rat. Anesthesiology 1994;80:383-391. [23) Malmberg AB, Yaksh TL. Hyperalgesia mediated by spinal glutamate or SP receptor blocked by spinal cyclooxygenase inhibition. Science 1992;257:1276-1279. [24] Malmberg AB, Yaksh TL. Antinociceptive actions of spinal nonsteroid antiinflammatory agents on the foImalin test in the rat. J Pharmacol Exp Ther 1992;263:136-146. [25] DiRosa M, Caligano A, Carnuccio R et al. Multiple control of inflammation by glucocorticoids. Agents Actions 1985;17:284-289. [26) Abram SE, Marsala M, Yaksh TL. Analgesic and neurotoxic effects of intrathecal steroids in rats. Anesthesiology 1992;77:A810. [27) Coderre TJ. Contribution of protein kinase C to persistent nociception following tissue injury in rats. Neurosci Lett 1992;140:181-184.

147

[28) Carror EI. Epidural steroid therapy for low back pain. In Stanton-Hicks M, Boas R, eds. Chronic Low Back Pain. New York, Raven Press, 1982;193-198. [29) Winne AP, Eartman JT, Meyers HL et al. Pain Clinic II: Intradural and extradural corticosteroids for sciatica. Anaesth Analg 1972;51:990-999. [30) Benzoil HT.Epidural steroid injections for low back pain and lumbosacral radiculopathy.Pain 1986;24:277-295. [31] Kepes ER, Duncalf D. Treatment of backache with spinal injections of local anesthetics, spinal and systemic steriods. A review. Pain 1985;22:33-47. . [32) Rowlingson, C. Epidural steroids. Do they have a place in pain management? Am Pain Soc J 1994;3:20-27. [33) Stoiker RJ, Vervest ACM, Groen GJ. The management of chronic spinal pain by blockades: a review. Pain 1994:58:1-20. [34] Brown FW. Management of discogenic pain using epidural and intrathecal steroids. Clin Orthop 1977;L29:72-78. [35] Abram SE, Anderson RA. Using a pain questionnaire to predict response to steroid epidurals. Reg Anesth 1980;5:11-91. [36] Abram SE, Eopwood MB. What factors contribute to outcome with lumbar epidural steroids? In Bond ME, Charton JE, Woolf CJ, eds. Proceedings of the VIth World Congress on Pain. Amsterdam, Elsevier 1991;491-496. [37] Shadid H, Harms JJ, Chase J. Efficacy of epidural steroid injections for patients with herniated disk or spinal stenosis: a retrospective study. Carle Selected Papers 1986;38:34-38. [38] Veriest H. Lumbar spinal stenosis: morphology, classification and long term results. In Weinstein IN, Wiesel SW, eds. The Lumbar Spine. Philadelphia, WB Saunders, 1990;546-588. [39] Rowlingson C, Kirschenbaum LP. Epidural analgesic techniques in the management of cervical pain. Anesth Analg 1986;65:938-942. [40] Hickey RF. Outpatient epidural steroid injections for low back pain and lumbosacral radiculopathy. N Z Med J 1987;100:594-596. [41] Roche J. Steroid-induced arachnoiditis. Med J Aust 1984;140:281-284. [42] Selby R. Complications from Depo-Medrol®. Surg NeuroI1983;19:393-394. [43] Nelson DA. Dangers from methylprednisolone acetate therapy by intraspinal injection. Arch Neurol 1988;45:804-806. [44] Ransford AO, Harries BJ. Localized arachnoiditis complicating lumbar disc lesions. J Bone Joint Surg 1972;54B:656-665. [45] Oppelt WW, Rail CPo Production of convulsions in the dog with intrathecal corticosteroids. Neurology 1961;2:925-927. [46] Warr AC, Wilkinson JA, Bum JMB, Langdon L. Chronic

148

[47]

[48] [49]

[50]

[51]

[52]

[53]

[54]

[55]

[56] [57]

[58]

[59]

[60]

[61]

[62] [63] [64] [65]

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149 lumbosciatic syndrome treated by epidural injection and manipulation. Practitioner 1972;209:53-59. Dilke TFW, Burry HC, Grahame R. Extradural corticosteroid injection in the management of lumbar nelVe root compression. Br Med J 1973;2:635-637. Ross JS, Masaryk TJ, Modic MT et al. Imaging of arachnoiditis. Am J RadioI1987;139:1025-1032. Racz GB, Sabonghy M, Gintautas J et al. Intractable pain therapy using a new epidural catheter. JAmMed Assoc 1982;248:646-647. Racz GB, Holubec JT. Lysis of adhesions in the epidural space. In Racz GB, ed. Techniques of Neurolysis. Boston, Kluwer, 1989;133-144. Breivik H, Hesia PE, Molnar I et al. Treatment of chronic low back pain and sciatica: Comparison of caudal epidural steroid injections of bupivacaine and methylprednisolone with bupivacaine followed by saline. In Bonica JJ, Albe-Fessard D, eds. Advances in Pain Research and Therapy. New York, Raven Press, 1976;1:927. Beliveau P. A comparison between epidural anaesthesia with and without corticosteroid in the treatment of sciatica. Rheum Phys Med 1971;11:40-43. Yates DW. A comparison of the types of epidural injections commonly used in the treatment of low back pain and sciatica. Rheumatol Rehab 1978;17:181-186. Snoek W, Weber H, Jorgensen B. Double blind evaluation of extradural methylprednisolone for hemiated lumbar discs. Acta Orthop Scand 1997;48:635. Cuckler JM, Bernini PA, Wiesel SW et al The use of epidural steroids in the treatment of lumbar radicular pain. J Bone Joint Surg 1985;67A:63. Hopwood MB, Abram SE. Factors associated with failure of epidural steroids. Reg Anesth 1993;18:238-243. Nelson DA, Yates TS, Thomas RB. Complications from intrathecal steroid therapy in patients with multiple sclerosis. Acta Neurol Scand 1973;49:176-188. Bernat JL, Sadowsky CH, Vincent FM, Nordgren RE, Margolis G. Sclerosing spinal pachymeningitis. J Neurol Neurosurg Psychiat 1976;39:1124-1128. Carta F, Canu C, Datti R, Guiducci G, Pisani R, Silvestro C. Calcification and ossification of the spinal arachnoid after intrathecal injection of Depo-Medrol. Zent bl Neurochir 1987;48:256-261. Dougherty JH, Fraser RAR. Complications following intraspinal injections of steroids. J Neurosurg 1978;48:1023-1025. Cohen FL. Conus medullaris syndrome following multiple intrathecal corticosteroid injections. Arch Neurol 1979;36:228-230. Johnson A, Ryan MD, Roche J. Depo-Medrol and myelographic arachnoiditis. Med J Aust 1991;155:18-20. Berrigan TJ, Goucke CR, Graziotti P. Depo-Medrol and myelographic arachnoiditis. Med J Aust 1991;155:421. Bourne IHJ. Lumbosacral adhesive arachnoiditis: a review. J Roy Soc Med 1990;83:262-265. Smolik EA, Nash FP. Lumbar spinal arachnoiditis: a complication of the intelVertebral disc operation. Ann Surg 1951;133:490-498.

[66] Feldman S, Behar AJ. Effect of intrathecal hydrocortisone on advanced adhesive arachnoiditis and cerebrospinal fluid pleocytosis. Neurology 1961;11:251-256. [67] Seghal AD, Gardner WJ, Dohn DF. Pantopaque arachnoiditis. Cleve Clin Q 1962;29:177-188. [68] Margolis G, Hall HE, Nowill WK. An investigation of efocaine, a long-acting local anesthetic agent. Arch Surg 1953;61:715-730. [69] Hurst EW. Adhesive arachnoiditis and vascular blockage caused by detergents and other chemical irritants. J Pathol 1955;70: 167-178. [70] Chino N, Awad EA, Kottke FJ. Pathology of propylene glycol administered by perineural and intramuscular injection in rats. Arch Phys Med RehabilI974;55:33-38. [71] MacKinnon SE, Hudson AR, Gentilli G, Kline DG, Hunter D. Peripheral nelVe injection injury with steroid agents. Plastic Reconstr Surg 1982;69:482-489. [72] Racz GB, Heavner J, Haynsworth R. Repeat epidural phenol injections in chronic pain and spasticity. In Racz GB, ed. Techniques of Neurolysis. Boston, Kluwer Academic, 1989;193-212. [73] Kay J, Raff II, Findling JW. Epidural triamcinolone suppress~s the pituitary-adrenal axis in human subjects. Anesth Analg 1994;79:501-505. [74] Knight CL, Burnell JC. Systemic side effects of extradural steroids. Anaesthesia 1980;35:593-594. [75] Stambough JL, Booth RE, Rothman RH. Transient hypercorticism after epidural steroid injection. J Bone Joint Surg 1984;66A:115-116. [76] Wallace G, Solove GJ. Epidural steroid therapy for low back pain. Postgrad Med 1985;78:213-218. [77] Tuel SM, Meythaler JM, Cross LL. Cushing's syndrome from methylprednisolone. Pain 1990;40:81-84. [78] Abram SE, O'Connor TC. Complications associated with epidural steroid injections. Reg Anesth 1996;in press. [79] Abram SE, Cherwenka RW. Transient headache immediately following epidural steroid injection. Anesthesiology 1979;50:461-462. [80] Katz JA, Lukin R, Bridenbaugh PO, Gunzenhauser L. Subdural intracranial air: An unusal cause of headache after epidural steroid injection. Anesthesiology 1991;74:615-618. [81] Simon DL, Kunz RD, German JD, Zivkovich V. Allergic or pseudo allergic reaction following epidural steroid deposition and skin testing. Reg Anesth 1989;14:253-255. [82] Gutknecht DR. Chemical meningitis following epidural injections of corticosteroids Am I Med 1987;82:570-575. [83] Plumb VJ, Dismukes WE. Chemical meningitis related to intrathecal corticosteroid therapy. South Med J 1977;70:1241-1246. [84] Goldstein NP, McKenzie BF, McGuckin WF. Changes in cerebrospinal fluid of patients with multiple sclerosis after treatment with intrathecal methylprednisolone acetate: a preliminary report. Proc Mayo Clin 1962;37:657-668. [85] Seghal AD, Tweed DC, Gardner WH, Foote MK. Laboratory studies after intrathecal corticosteroids. Arch NeuroI1963;9:74-78.

S.E. Abram / Journal of Back and Musculoskeletal Rehabilitation 8 (1997) 135-149

[86] Dougherty JH, Fraser RAR. Complications following intraspinal injection of steroids. J Neurosurg 1978;48:1023-1025. [87] Shealy CN. Dangers of spinal injections without proper diagnosis. JAmMed Assoc 1966;197:156-158. [88] Chan S-T, Leung S. Spinal epidural abscess following steroid injection for sciatica. Spine 1989;14:106-108. [89] Goucke CR, Graziotti P. Epidural abscess following local anesthetic and steroid injection for chronic low back pain. Br J Anaesth 1990;65:427-429. [90] Strong WE. Epidural abscess associated with epidural catheterization: A rare event? Report of two cases with markedly delayed presentation. Anesthesiology 1991;74:943-946.. [91] Bromage PRo Spinal extradural abscess: pursuit of vigilance. Br J Anaesth 1993;70: 471-73. [92] Cicala RS, Turner R, Moran E, Henley R, Wong R,

[93]

[94]

[95]

[96]

149

Evans J. Methylprednisolone acetate does not cause inflammatory changes in the epidural space. Anesthesiology 1990;72:556-558. Delaney TJ, RowJingson JC, Carron H, Butler A. Epidural steroid effects on nerves and meninges. Anesth Analg 1980;59:610-614. Wood KM, Arguelles J, Norenberg MD. Degenerative lesions in rat sciatic nerves after local injections of methylprednisolone in sterile aqueous suspension. Reg Anesth 1980;5(1):13-15. Johansson A, Dahlin L, Kerns J. Long term corticosteroid application does not influence transmission or structure. Acta Anaesth Scand 1995;39:364-369. Benzon HT, Gissen AJ, Strichartz GR, Avram MJ, Covino BG. The effect of polyethylene glycol on mammalian nerve impulses. Anesth Analg 1987;66:553-559.