Sarasota Dermatology, Sarasota Dermatologist, O'Donoghue Dermatology, Dermatology Experts

Report of a case and review of the literature

J Morgan O’Donoghue, MD
Yvana P. Céspedes, MD
Paul F. Rockley, MD
Thomas P. Nigra, MD

CASE REPORT 

A 26-year-old white female with a past medical history remarkable for systemic lupus erythematosus and vasculitis developed grand mal seizures, which were controlled with oral phenytoin 300 milligrams per day. Three weeks later, she noticed the onset of generalized pruritus and skin tenderness. The following day, she developed fever, chills, mucous membrane swelling and a symmetrical, erythematous, morbilliform eruption on the extensor aspects of the distal extremities. During the next 24 hours, the eruption slowly evolved from discrete, reddish macules and papules to confluent, necrotic blisters involving the extremities, trunk, head, neck, and mucosal surfaces

At the time of examination the patient had a temperature of 40 C, widespread tender erythema, hemorrhagic vesiculobullous lesions and seropurulent, crusted erosions. Approximately 20 percent of the body surface area was covered with necrotic epidermis. A positive direct Nikolsky sign was noted on lesional skin only. Mucosal surfaces contained some inflammatory bullous and erosive lesions. Conjunctival involvement was characterized by bilateral cheimosis, redness and lacrimation. Further evaluation revealed that her white blood cell count was 0.6 x 103/mm3 with 34%

lymphocytes, 32% bands, 17% monocytes, 9% granulocytes, 8% metamyelocytes; erythrocyte count was 3.12 x 106/mm3; hemoglobin was 9.0 g/dl; hematocrit was 26.6%; and platelet count was 162,000/mm3. Serum chemistries were remarkable for blood urea nitrogen 28 mg/dl; creatinine 1.5 mg/dl; glucose 160 mg/dl; total protein 4.0 gm/dl; albumin 1.6 gm/dl; and LDH was 278 IU/L. Liver function tests, coagulation studies and chest x-rays were within normal limits. Urinalysis showed protein 20 mg/dl, red blood cell count 20-50/HPF, white blood cell count 3-6/HPF, granular casts 20-50/HPF and bacteria 2+/HPF. Blood cultures grew coagulase-negative staphylococcus and beta-hemolytic streptococcus.

A skin biopsy from intact, erythematous abdominal skin revealed slight spongiosis; scattered necrotic keratinocytes; vacuolar degeneration of the basal layer; focal cleft formation at the dermal-epidermal junction; and a moderate lymphohistiocytic infiltrate at the dermal-epidermal junction and around superficial dermal blood vessels (Figure 1A). These histologic findings are characteristic of TEN. The clinical diagnosis of TEN was suggested by the skin biopsy. Phenytoin, implicated as the cause due to the temporal relationship between starting the drug and the development of TEN, was discontinued.

The patient was treated with intravenous methylprednisolone 100 milligrams (2 mg/kg) times one dose followed by 50 mg (1 mg/kg) every eight hours for three days. She also received intravenous fluids, antibiotics, antipyretics, analgesics, and blood transfusions. Topical therapy consisted of viscous lidocaine, colloidal oatmeal soaks, dressing exfoliated areas with mupirocin and petrolatum covered with gauze, and ophthalmic application of lubricants, antibiotic ointment, and prednisolone eye drops.

At the end of the third hospital day, the patient was afebrile, the eruption was less erythematous, there were no new areas of necrolysis, and the patient's condition stabilized. A follow-up skin biopsy from abdominal skin after three days of therapy showed a reduction of the lymphohistiocytic infiltrate without evidence of progression of keratinocyte necrosis (Figure 1B). Methylprednisolone was switched to hydrocortisone at physiologic (stress) doses. Over the next few days, sloughed areas began to re-epithelialize. A third abdominal skin biopsy obtained after one week of therapy revealed further reduction of the lymphohistiocytic infiltrate (Figure 1C). Repeat laboratory studies revealed that neutropenia resolved, BUN and creatinine normalized, urinalysis returned to baseline, and blood cultures were negative for growth.

Steroid therapy was gradually tapered to her usual maintenance dose and she was discharged in good condition after four weeks of hospitalization. On follow-up cutaneous examination, post-inflammatory pigmentary
changes were noted but there were no significant sequelae due to mucosal scarring.

DISCUSSION

In 1956 Lyell1 and Lang and Walker2 independently described TEN. Lyell coined the term "toxic epidermal necrolysis", referring to epidermal damage due to a presumed circulating toxin. Patients with brain tumors, systemic lupus erythematosus, human immunodeficiency virus infection or bone marrow transplant recipients are at increased risk for the development of TEN.3 Currently,

TEN is viewed as a mucocutaneous reaction to a variety of antigenic stimuli. 4 These stimuli include drugs, infections, collagen vascular disease, malignancy, pregnancy, menstruation and vaccination. The most commonly implicated medications are allopurinol, antibiotics, barbiturates, carbamazepine, hydantoins, sulfonamides and nonsteroidal anti-inflammatory drugs.5 These drugs
are typically begun 1 to 3 weeks prior to the onset of TEN. In our patient, the delay between the introduction of phenytoin and the onset of TEN may be attributed to long term steroid therapy.6

The pathophysiologic events involved in TEN are not completely understood. Pathogenic mechanisms which have been proposed include type IV hypersensitivity reactions (delayed-type hypersensitivity and T cell-mediated cytotoxicity), type II cytotoxic reactions and altered metabolism of the offending drug. In some cases of TEN, positive patch tests and positive lymphocyte transformation tests have been used to support the delayed-type hypersensitivity hypothesis.7

Immunohistologic studies stress the importance of cell-mediated cytotoxicity in which a T helper 1-like cytokine response may direct the traffic of effector CD8 T lymphocytes8 and/or macrophages into the epidermis thus causing keratinocyte damage in a process referred to as satellite cell necrosis.9 These findings are similar to that of skin graft rejection or of graft-versushost disease (GVHD) in which CD8 T lymphocytes and macrophages are the effector cells of an acute cell-mediated reaction against allogeneic antigens. Animal models10 and human data 11 suggest that acute GVHD and TEN share several biologic, immunologic, clinical and histologic features. In support of a role for type II cytotoxic reactions in patients with drug-induced TEN, indirect immunofluorescence studies and complement-dependent cytotoxicity assays have demonstrated common antigenic determinants on keratinocytes and monocytes that may serve as targets for drug-induced antibodies.5

In addition, impaired metabolism of an offending drug can produce reactive metabolites which may act as haptens by binding to keratinocyte surface proteins, thus generating an immune response.12 More recent data identifies fragmented DNA staining in lesional TEN epidermis which has been accepted as a feature of apoptotic cell death. It is theorized that an apoptotic mechanism could account for the generalized cell death created in TEN.  13 Further studies are needed to determine which of these pathogenic mechanisms is the predominant cause of necrolysis.

The use of systemic corticosteroids in the treatment of TEN is controversial because of a lack of randomized, controlled prospective studies and because the effects of steroid therapy probably vary depending on the dosage and time of their administration during the course of TEN.

In patients who present with early stage TEN (erythrodermic, slowly evolving cases with less than 25 percent total body surface area necrolysis), proponents of steroids recommend the administration of high dose methylprednisolone (1-3 mg/kg/day) to limit epidermal sloughing and occasionally arrest progression of disease.14-18 A recent study indicates that TEN can occur in some patients undergoing long term corticosteroid therapy for underlying disease, however, most of these patients were on low dose regimens.6 Although the mechanism of action of high dose corticosteroids in TEN remains unknown, the rationale for their use is based upon the theories that some cases of TEN are due to delayed hypersensitivity reactions or cell-mediated/cytotoxic reactions.4 These immunopathogenic mechanisms are mediated by T lymphocytes.

Our patient’s clinical improvement correlated with the administration of high dose systemic corticosteroids and the progressive reduction of lymphocytic infiltration. In patients who present with late stage disease (rapidly evolving cases with greater than 25
percent total body surface area necrolysis), the risks of using steroids, regardless of the dosage, probably outweigh their benefits.19,20

Many of the reports ascribing negative outcomes to the administration of steroids involved patients with late stage disease.21-23 Since the majority of patients with TEN have late stage disease when they receive medical attention, most patients should not be treated with systemic corticosteroids. Patients with late stage disease should be managed in burn centers with vigorous fluid resuscitation, topical antimicrobial agents, biologic dressings, debridement, wound care, nutritional support and physical and psychological therapy protocols.24

Adjunctive therapies include hyperbaric oxygen 2,25 and plasmapheresis.26 This case demonstrates that serial skin biopsies in patients with early stage TEN may be useful to assess the response to steroids, guide therapy and minimize potential complications. These biopsies can be particularly helpful in managing patients at high risk of infection such as those with neutropenia. Since inflammatory infiltrates can occur in response to necrolysis, it is important to obtain all biopsies from intact, erythematous, non-necrotic skin sites with similar clinical appearance.

If a follow-up biopsy after 48 to 72 hours of high dose steroids shows a good response in terms of decreasing the lymphocytic infiltrate and halting epidermal necrosis, then steroids can be tapered cautiously over a period of 1 to 3 weeks. Relapses have been reported following rapid withdrawal of steroids. 17,27-29 Patients with biopsies revealing a partial response may benefit from an additional 48 hours of therapy, however, high dose steroids are not recommended for more than a total duration of 5 days.30 If a follow-up biopsy shows progression of either the infiltrate or epidermal necrosis, steroids should be abruptly discontinued unless they are unequivocally indicated for a coexisting condition. Partial responders and non-responders can be treated with other immunosuppressive agents, including azathioprine 2 to 3 mg/kg/day27, cyclophosphamide 100 to 300 milligrams intravenously daily,31 or cyclosporine 3 to 4 mg/kg/day orally. 32

Cyclosporine and cyclophosphamide inhibit proliferation of T lymphocytes in patients with TEN.31,32 These agents may be effective as monotherapy in steroid-resistant cases or have additive/synergistic effects in combination with steroids. Further studies assessing the clinical and histopathologic response of early stage TEN to immunosuppressive therapy are needed in order to
establish treatment guidelines.

We wish to acknowledge the editorial and photographic assistance provided by Thomas D. Horn, M.D.

REFERENCES

1. Lyell A. Toxic epidermal necrolysis: an eruption resembling scalding of the skin. Br J
Dermatol 1956; 68:355-361.
2. Lang R, Walker J. An unusual bullous eruption. S Afr Med J 1956; 30:97-98.
3. Bourgault-Villada I, Roujeau J-C. Toxic epidermal necrolysis. Curr Prob Dermatol 1996; 3:
99-104.
4. Parsons JM. Toxic epidermal necrolysis. Int J Dermatol 1992; 31: 749-768.
5. Wolkenstein PE, Roujeau JC, Revuz J. Drug-induced toxic epidermal necrolysis. Clin
Derm 1998;16:399-409.
6. Guibal F, Bastuji-Garin S, Chosidow O, et al. Characteristics of toxic epidermal necrolysis
in patients undergoing long-term corticosteroid therapy. Arch Dermatol 1995; 131:669-672.
7. Tagami H, Tatsuta K, Iwatsuki K, et al. Delayed hypersensitivity in ampicillin-induced
toxic epidermal necrolysis. Arch Dermatol 1983; 119: 910-913.
8. Correia O, Delgado L, Ramos JP, et al. Cutaneous T-cell recruitment in toxic epidermal
necrolysis. Arch Dermatol 1993; 129: 466-468.
9. Paquet P, Pierard GE. Erythema multiforme and toxic epidermal necrolysis: a comparative
study. Am J of Dermatopath 1997; 19:127-132.
10. Merot Y, Saurat JH. Clues to pathogenesis of toxic epidermal necrolysis. Int J Dermatol
1985;24:165-168.
11. Takeda H, Mitsuhashi Y, Kondo S, et al. Toxic epidermal necrolysis possibly linked to
hyperacute graft-versus-host disease after allogenic bone marrow transplantation. J
Dermatol 1997; 24:635-641.
12. Friedmann PS, Strickland I, Pirmohamed M, et al. Investigation of mechanisms in toxic
epidermal necrolysis induced by carbamazepine. Arch Dermatol 1994; 130:598-604.
13. Sugimoto Y, Mizutani H, Sato T, et al. Toxic epidermal necrolysis with severe
gastrointestinal mucosal cell death: A patient who excreted long tubes of dead intestinal
epithelium. J Dermatol 1998; 25:533-538.
14. Bjornberg A. Fifteen cases of toxic epidermal necrolysis (Lyell). Acta Derm Venereol
(Stockh) 1973; 53:149-152.
15. Parsons JM. Management of toxic epidermal necrolysis. Cutis 1985; 36:305-312.
16. Sherertz EF, Jegasothy BV, Lazarus GS. Phenytoin hypersensitivity reaction presenting
with toxic epidermal necrolysis and severe hepatitis. J Am Acad Dermatol 1985; 12:178-
181.
17. Tegelberg-Stassen MJ, van Vloten WA, de la Faille HB. Management of nonstaphylococcal
toxic epidermal necrolysis: follow-up study of 16 case histories.
Dermatologica 1990; 180:124-129.
18. Stables GI, Lever RS. Toxic epidermal necrolysis and systemic corticosteroids. Br J
Dermatol 1993; 128:357.
19. Revuz J, Roujeau J-C, Guillaume J-C, et al. Treatment of toxic epidermal necrolysis,
Créteil’s experience. Arch Dermatol 1987; 123:1156-1158.
20. Herndon DN. Toxic Epidermal Necrolysis: A systemic and dermatologic disorder best
treated with standard treatment protocols in burn intensive care units without the prolonged