Epidural steroid injections have become one of the most frequently used conservative options in the management of acute and chronic back pain . As the indications for epidural steroid injections increase so do the reports of adverse responses associated with this procedure. We describe a series of 12 patients each of whom developed facial flushing and/or generalized erythema after receiving an epidural steroid injection for lumbar or cervical radiculopathy.
Between August 1, 1990, and August 13, 1993, we performed 1399 epidural steroid injections for patients with chronic radicular pain. Of those injections, 1021 were performed at the lumbar level whereas the remaining 378 were cervical epidural injections. During that time we encountered 12 patients who developed facial flushing and/or generalized erythema after epidural steroid administration. The reactions were typically described as a "sunburned" or a generalized reddish appearance of the skin which in some patients was present only on the face. Other patients experienced erythema primarily limited to the trunk and extremities although three patients developed cutaneous reactions involving the face, trunk, and extremities. No systemic manifestations occurred during any of these reactions to indicate an anaphylactic event. Of these 12 patients, only one had a known sensitivity to aspirin, while three patients had documented allergies to sulfa drugs. The eight remaining patients had no known drug allergies. Ten patients underwent lumbar epidural steroid injection and two patients had a series of cervical epidural steroid injections. In every patient, either 80 mg or 120 mg of Aristocort Registered Trademark (triamcinolone diacetate; Fujisawa Pharmaceutical Co., Deerfield, IL) was given for the initial epidural injection. All patients except one developed facial flushing or erythema after the initial injection. The patient who had no adverse reaction to the first injection developed a red face and a warm feeling after a subsequent injection of 80 mg triamcinolone approximately 1 mo later. Ten patients received a second injection. Of these 10 patients, 9 received either 80 mg or 120 mg of triamcinolone. In one patient, the drug was changed from 80 mg triamcinolone to 120 mg Depomedrol Registered Trademark (methylprednisolone; Lederle Parenterals Inc., Carolina, Puerto Rico).
Six of the nine patients receiving triamcinolone at the second injection experienced reactions similar to those that occurred after the first injection. Two patients who had developed a reaction after the first injection with triamcinolone did not exhibit a reaction after the second injection with the drug. One patient who had not experienced a reaction after the first injection with triamcinolone subsequently developed facial flushing after the second injection with the same dose of triamcinolone. The patient in whom the drug was changed to methylprednisolone also experienced a generalized erythematous reaction similar to that which developed after the first injection. However, the onset of the reaction (24 h) was more rapid than after the initial injection (48 h).
Seven patients underwent a third epidural steroid injection. Of the seven, four patients were given triamcinolone and the remaining three patients received methylprednisolone. Three of the four patients given triamcinolone had no adverse reaction after the third injection. One of these patients had been pretreated with diphenhydramine and ranitidine and expressed a "warm feeling" after the first and second injections. Prior to the second injection she had been given diphenhydramine alone, which had failed to prevent the erythema. The fourth patient who received triamcinolone for her third injection experienced facial flushing similar to that which developed after both previous injections. All three patients who received methylprednisolone for the third injection were treated previously with, and reacted to, triamcinolone. Two of these patients had no reaction after injection with methylprednisolone, while the third patient developed a sunburned-looking red face and neck.
In this group of 12 patients the mean time to onset of reaction was 27 h after epidural steroid injection. Once a reaction had occurred, an average of 72 h was required for complete resolution. The patients who had taken diphenhydramine after erythema or facial flushing had developed seemed to experience a slightly shorter duration of reaction (64 h) compared to the patients who did not take it once a reaction had occurred (82 h). However, due to the small sample size in the group treated with diphenhydramine (n = 4) and the wide variation in the overall duration of reactions, statistical analysis could not be applied to determine significance Table 1.
The occurrence of facial flushing after intra-articular distension with steroids for the management of capsulitis of the shoulder has been reported by Jacobs et al. . In this study, 50 intra-articular injections were performed on 47 patients over a 2-yr period, resulting in two patients developing temporary (<24 hr) facial flushing after steroid injection. The authors of this study did not propose a mechanism for the facial flushing that occurred after the injections. Facial flushing after cervical epidural steroid injection has been reported by Cicala et al.  in 9.3% of 142 patients. However, facial flushing and erythema after lumbar epidural steroid administration heretofore have not been reported.
We report a series of 12 patients who developed facial flushing and/or generalized erythema after an epidural steroid injection. In our series, facial flushing was far less frequent (1.4%) than the 9.3% occurrence reported by Cicala et al. . Furthermore, we observed a mean duration of erythema (3 days) greater than six times as long as that reported by Cicala et al. , who postulated that stiff neck and facial flushing were secondary to the large volume of the injectate that he used to ensure spread throughout the cervical spine (10-15 mL of 0.5% lidocaine and 1 mg/kg of methylprednisone acetate). Unlike the Cicala study, we used a smaller volume of injectate when performing cervical epidural injections (5-7 mL of normal saline with the steroid preparation). Furthermore, 10 of 12 of our patients received lumbar epidural steroid injections which consisted of 10 mL of either normal saline or 1% lidocaine with the steroid preparation. Although the relatively large volume that Cicala et al.  administered in the cervical epidural space could certainly account for the stiff neck that occurred in his study, it does not explain the facial flushing in our patients.
Our review poses more questions than it helps to answer. What is the true incidence of facial flushing or generalized erythema after an epidural steroid injection? Which patients are predisposed to developing an adverse reaction? Does changing steroid preparations after an initial reaction effect the chances of experiencing a subsequent reaction? What role does pretreatment with a histamine-1 and or a histamine-2 receptor blocker play in preventing an adverse reaction to epidurally administered steroids? As the indications for epidural steroid injections continue to increase, so will the number of injections being performed and undoubtedly the number of adverse reactions reported. The data presented here is by no means a scientific study, nor is it an attempt to provide concrete answers to the questions mentioned above. That would require a much larger prospective investigation. However, our review does provide some new information. Facial flushing and generalized erythema can occur after lumbar epidural steroid injection. The mechanism by which these reactions occur, although not fully understood, may be, in part, a histamine-mediated response and not a volume-related phenomenon as previously thought by Cicala et al. . Also, the duration of the reaction appears to be abbreviated by the administration of diphenhydramine once it has occurred; however, further investigation is needed to determine whether this is of statistical significance. By including this information in the discussion of the possible risks and side effects of an epidural steroid injection, we think that patients will be informed more completely as to what to expect should facial flushing or erythema develop.
1. Sandrock NJ, Warfield CA. Epidural steroids and facet injections. In: Warfield CA, ed. Principles and practices of pain management, vol 1. New York: McGraw-Hill; 1993:401-10.
2. Jacobs LG, Barton MA, Wallace, AW et al. Intra-articular distention and steroids in the management of capsulitis of the shoulder. BMJ 1991;302:1498-1501.
3. Cicala RS, Westbrook L, Angel JJ. Side effects and complications of cervical epidural steroid injections. J Pain Symptom Manage 1989;4(2):64-6.