Dermatology Online Journal

Toxic epidermal necrolysis as a complication of treatment with celecoxib
Ashley G Perna MD, Christy A Woodruff MD, Ramsey F Markus MD, and Sylvia Hsu MD
Dermatology Online Journal 9 (5): 25

Department of Dermatology, Baylor College of Medicine, Houston. shsu@bcm.emc.edu

---

Figure 1
Diffuse, confluent bullae on legs.

An 85-year-old woman presented to her internist with a 3-day history of nausea and vomiting and a 1-day history of blanching erythematous macules on her back. Her liver enzymes were elevated, and she appeared jaundiced. She was subsequently admitted to the hospital for evaluation of the jaundice. Over the next 2 days, the erythematous macules rapidly spread to involve the anterior trunk, arms, legs, and face; they developed dusky centers and began to evolve into flaccid vesicles and bullae involving approximately 40 percent of her body surface (fig. 1). She did not appear to have involvement of the ocular, oral, or genital mucosa. An endoscopic exam of her upper gastrointestinal tract, done as part of the jaundice evaluation, demonstrated extensive mucosal erosions throughout the gastrointestinal tract. A mucosal biopsy was not done because it was thought that the erosive mucosal process was an extension of the cutaneous process. Ultrasound and CT imaging failed to reveal any hepatic-duct dilatation, and her elevated liver enzymes and bilirubin were consistent with a cholestatic jaundice. The diagnosis of toxic epidermal necrolysis (TEN) with severe internal mucosal involvement was made without a skin biopsy.

Intravenous immunoglobulin (IVIg) at a dose of 0.75 gram/kg/day for 4 days was initiated; this slowed but did not stop the vesicle formation. Her liver enzymes continued to rise during the IVIg course, and she began to produce bloody sputum. On chest X-ray, her lung fields became progressively more opaque. The patient died on the eleventh day of hospitalization.

The patient's medications prior to admission included labetolol, atorvastatin, isosorbide mononitrate, amiodarone, warfarin, ranitidine, furosemide, and thyroxine; occasionally she took acetaminophen, darvocet, and vicodin. She had taken all of these medications for over 1 year. In addition, she was taking phenytoin at a stable dose for 5 years. She denied taking any over-the-counter medications or herbs. To treat her lower back pain, a 2-week course of celecoxib (100 mg orally each day) was begun 18 days prior to the development of the rash. The last celecoxib pill was taken 2 days before the onset of the rash. Because of the close temporal relation between the initiation of celecoxib and the appearance of the skin eruption, and because celecoxib was the only new medication the patient had taken, the etiology of TEN was considered to be most likely a reaction either to the celecoxib alone or to celecoxib in combination with one of her other medications.

The agents most commonly implicated in TEN are sulfonamide antibiotics, anticonvulsant agents, and allopurinol [1]. This patient's mucosal involvement involved the gastrointestinal tract but spared the mouth, an uncommon presentation of TEN. The most common mucous membranes involved are, in order of decreasing frequency, the oropharynx, eyes, genitalia, and anus [2]. The internal involvement most likely preceded her cutaneous eruption, accounting for her presenting symptoms of nausea, vomiting, and the jaundice noted at the time of admission. This time frame is classic for TEN; mucosal lesions generally precede skin lesions by 1-3 days [2].

Celecoxib is not known to be one of the medications commonly associated with TEN. In the drug manufacturer's post-marketing-surveillance program, which included a total of seven million prescriptions for celecoxib, there were two cases of SJS and TEN [3]. The most common medications implicated in SJS and TEN are sulfonamide antibiotics; this association is likely a response to the aromatic amine moiety [3]. Other medications such as celecoxib, furosemide, and thiazide diuretics also contain a sulfonamide component, but the amine portion is not an aromatic amine, theoretically rendering them less likely to cause SJS and TEN [3].

Because TEN and SJS are much more common with sulfonamide antibiotics than with celecoxib, the mechanism of this severe skin reaction is likely different for these medications [3]. There are two possible explanations for TEN plus liver and lung involvement seen in our patient. The first possible mechanism is an idiosyncratic reaction, usually associated with a metabolite, which can lead to SJS and TEN and involve multiple organs [3]. The other possible mechanism is a hypersensitivity syndrome, which consists of fever, rash, and internal organ involvement [3]. Our patient did not have fever, so the idiosyncratic reaction is the more likely explanation.

There are two cases of toxic epidermal necrolysis related to celecoxib reported in the literature. Many medications have rare, idiopathic adverse effects. In this case, we could not prove that TEN was a result of the celecoxib because the patient was taking other medications, including phenytoin. Phenytoin is much more likely than celecoxib to cause TEN. In this case, phenytoin was not as likely as celecoxib to be the etiology both because of the close temporal relation of the onset of TEN with the initiation of celecoxib and because she had been taking the same phenytoin dose for 5 years. In addition, we considered the possibility that it was the combination of celecoxib with the patient's other medications that led to the development of TEN. It seems more than coincidental that our patient developed TEN 2 weeks after starting celecoxib.

References

1. Roujeau JC, Kelly JP, Naldi L, et al. Medication use and the risk of Stevens-Johnson Syndrome or Toxic Epidermal Necrolysis. N Engl J Med 1995;333:1600-7. PubMed

2. Becker DS. Toxic Epidermal Necrolysis. Lancet 1998;351:1417-20. PubMed

3. Knowles S, Shapiro L, Shear NH. Should celecoxib be contraindicated in patients who are allergic to sulfonamides? Revisiting the meaning of "sulfa" allergy. Drug Saf 2001;24(4):239-47. PubMed

© 2003 Dermatology Online Journal