Bullous drug reactions (StevensJohnson syndrome

[SJS] and toxic epidermal necrolysis [TEN])

Medications may induce known autoimmune bullous diseases
such as pemphigus (penicillamine) or linear IgA disease (vancomycin).
Acute generalized exanthematous pustulosis may
be so extensive as to cause a positive Nikolsky sign, and have
a background of purpura and targetoid lesions, simulating
SJS/TEN.
Cytokines may produce widespread
bullous eruptions, perhaps through physiologic mechanisms.
These are fortunately uncommon reactions to medications,
with an incidence of 0.41.2 per million person-years for TEN
and 1.26.0 per million person-years for SJS. These druginduced
forms of erythema multiforme are usually more
extensive than herpes-associated erythema multiforme or
mycoplasma-associated SJS, but at times the distinction may
be difficult. The more severe the reaction, the more likely it is
to be drug-induced (50% of cases of SJS and 80% of cases of
TEN).
The following definitions
are useful to classify cases: SJS has less than 10% body surface
area (BSA) involved, cases with 1030% are SJSTEN overlap
cases, and more than 30% BSA erosion is called TEN. SJS and
TEN are considered by some as parts of a disease spectrum
based on the following:
They are most commonly induced by the same
medications.
Patients initially presenting with SJS may progress to
extensive skin loss resembling TEN.
The histologic findings are indistinguishable.
Both are increased by the same magnitude in HIV
infection.
However, recently, genetic evaluations of Caucasians with SJS
and TEN showed distinct genetic predispositions for these
conditions, allowing for consideration of them as distinct
disorders.
The cause of SJS/TEN is not established. In Taiwan carbamazepine
causes up to one-third of cases, but only 5% of
cases in Europe. In Han Chinese the HLA haplotype HLAB*
1502 is present in the vast majority of cases of carbamazepineinduced
SJS/TEN patients and is present in about 10% of the
Han Chinese population in general. This HLA association is
NOT found, however, in patients with carbamazepine-induced
SJS/TEN of other ethnicities, suggesting that this marker for
risk is specific for Asians. HLA typing should be performed in
all Asians before starting carbamazepine, since the prevalence
of HLA-B*1502 is 510% in Asians in the USA and Asia. HHV-6
reactivation may also be seen in SJS/TEN patients.
More than 100 medications have been reported to cause
SJS and TEN. In adults, common inciting medications are

trimethoprimsulfamethoxazole (13 in 100 000), Fansidar-R,

sulfadoxine plus pyrimethamine (10 in 100 000), nevirapine,
lamotrigine (1 in 1000 adults and 3 in 1000 children), and
carbamazepine (14 in 100 000). Antibiotics (especially longacting
sulfa drugs and penicillins), other anticonvulsants,
anti-inflammatories (NSAIDs), and allopurinol are also
frequent causes. Currently, in Europe, allopurinol is the
most common cause of SJS and TEN. In children SJS/TEN
is most commonly caused by sulfonamides and other antibiotics,
antiepileptics, and acetaminophen. SJS/TEN from
trimethoprimsulfamethoxazole is significantly more common
in the spring. If the inciting drug has a short half-life, and the
drug is promptly stopped, the mortality is reduced from 26%
to 5%. This suggests that the use of agents with short half-lives
and the prompt discontinuation of the medication when the
first signs of an adverse reaction appear may be very important
ways to reduce the mortality from TEN.
Fever and influenza-like symptoms often precede the eruption
by a few days.
Skin lesions appear on the face and trunk
and rapidly spread (usually within 4 days) to their maximum
extent.
Initial lesions are macular and may remain so, followed
by desquamation, or may form atypical targets with purpuric
centers that coalesce, form bullae, then slough. Patients with
purpuric atypical targets may evolve more slowly, and usually
the skin lesions are clinically inflammatory.
In SJS, virtually
always, two or more mucosal surfaces are also eroded, the oral
mucosa and conjunctiva being most frequently affected. There
may be photophobia, difficulty with swallowing, rectal erosions,
painful urination, and cough, indicative of ocular,
alimentary, urinary, and respiratory tract involvement,
respectively.
In other patients,
macular erythema is present in a local or widespread distribution
over the trunk.
Mucosal involvement may not be found.
The epidermis in the areas of macular erythema rapidly
becomes detached from the dermis, leading to extensive skin
loss, often much more rapidly than occurs in the patients with
atypical targets and extensive mucosal involvement.
Patients with SJS/
TEN may have internal involvement very similar to patients
with DRESS/DIHS induced by the same medication (see
above). These most commonly include eosinophilia, hepatitis,
and worsening renal function.
A skin biopsy is usually performed. Frozen-section analysis
may lead to a rapid diagnosis. This is to exclude other

diseases (such as staphylococcal scalded skin syndrome,

pemphigus, graft versus host disease (GVHD), etc.) and to
confirm the diagnosis.
Management of these patients is similar to those with an
extensive burn. They suffer fluid and electrolyte imbalances,
bacteremia from loss of the protective skin barrier, hypercatabolism,
and sometimes acute respiratory distress syndrome.
Their metabolic and fluid requirements are less than
in burn victims, however. Survival is improved if patients are
cared for in a specialized burn unit, or on a special dermatology
unit with skill in managing these patients.
In addition to extent of skin loss,
age, known malignancy, tachycardia, renal failure, hyperglycemia,
and low bicarbonate are all risk factors for having a
higher mortality with SJS/TEN. The SCORTEN gives one
point for each of these findings. The SCORTEN total predicts
mortality, with a 3.2% mortality for 01 points, and a 90%
mortality for 5 or more points. However, respiratory tract
involvement, not included in the SCORTEN, is also a bad
prognostic sign. About one-quarter of TEN patients have
bronchial involvement. In TEN, epithelial detachment of the
respiratory mucosae and associated acute respiratory distress
syndrome are associated with a mortality of 70%.
IVIG is now frequently used to manage the more severe adult
and pediatric patients with bullous drug eruptions (TEN). A
dose of 1 g/kg/day for the first 4 days following admission is
effective. It is best used when detachment has not become
extensive, as total BSA of skin loss is an important predictor
of mortality.
Keratinocyte death in SJS and TEN is proposed to occur
via two potential mechanisms, and the relative importance of
each of these mechanisms in SJS and TEN is not known.
Activated cytotoxic T cells and natural killer (NK) cells produce
granulysin, perforin, and granzyme B, all of which can induce
keratinocyte necrosis. In addition, keratinocyte necrosis can
be induced by the binding of soluble Fas ligand (sFasL) to
Fas (also known as the death receptor or CD95). Soluble
Fas ligand is elevated in the blood of patients with TEN, and
its level correlates with BSA involvement. In addition, the
peripheral blood mononuclear cells of patients with TEN
secrete Fas ligand upon exposure to the incriminated drug.
The sera of patients with TEN induce necrosis of cultured
keratinocytes, and a monoclonal antibody to Fas ligand in
a dose-dependent fashion inhibits keratinocyte necrosis
exposed to TEN patient sera. This strongly supports Fas
expression by keratinocytes, and Fas ligand production by
immune cells, as the mechanisms by which TEN is mediated.
The proposed mechanism of action of IVIG in TEN is by IVIG
blocking the binding of sFasL to Fas, stopping keratinocyte
apoptosis.
The presence of cytotoxic T lymphocytes and NK cells
within the dermis subjacent to the necrotic epidermis suggests

that immunosuppressive agents that block immune function

could also be effective in SJS or TEN. The role of immunosuppressive
therapy, however, is very controversial. The benefit
of immunosuppressives would be to stop the process very
quickly and thereby reduce the ultimate amount of skin lost.
Once most of the skin loss has occurred, immunosuppressives
only add to the morbidity and perhaps mortality of the
disorder. In children with SJS, this adverse effect has been
documented, probably since their mortality is low and
immunosuppressives only add risk for infection. Because SJS
and TEN evolve rapidly (average 4 days to maximum extent),
very early treatment would be required to observe benefit.
If immunosuppressive treatment is considered, it
should be used as soon as possible, given as a short trial to see
if the process may be arrested, and then tapered rapidly to
avoid the risk of immunosuppression in a patient with substantial
loss of skin.
High-dose corticosteroids given intravenously
with reported success have included 100 mg of
dexamethasone per day for 3 or 4 days and methylprednisolone,
30 mg/kg/day for 3 days. Cyclosporine in doses of
35 mg/kg/day for 824 days, with or without an initial burst
of systemic steroids, has also been reported to stop SJS and
TEN abruptly. Anecdotally, both etanercept, 25 mg twice, and
infliximab, 5 mg/kg intravenously once, have led to rapid
termination of skin sloughing. Systemic and topical steroid
therapy for ocular involvement also appears to improve outcomes.
In patients with SJS/TEN who also have systemic.
High-dose corticosteroids given intravenously
with reported success have included 100 mg of
dexamethasone per day for 3 or 4 days and methylprednisolone,
30 mg/kg/day for 3 days. Cyclosporine in doses of
35 mg/kg/day for 824 days, with or without an initial burst
of systemic steroids, has also been reported to stop SJS and
TEN abruptly. Anecdotally, both etanercept, 25 mg twice, and
infliximab, 5 mg/kg intravenously once, have led to rapid
termination of skin sloughing. Systemic and topical steroid
therapy for ocular involvement also appears to improve outcomes.
In patients with SJS/TEN who also have systemic
For patients who survive, the average time for epidermal
regrowth is 3 weeks. The most common sequelae are ocular
scarring and vision loss. The only predictor of eventual visual
complications is the severity of ocular involvement during the
acute phase. A sicca-like syndrome with dry eyes may also
result, even in patients who never had clinical ocular involvement
during the acute episode. Other complications include
cutaneous scarring, eruptive melanocytic lesions, and nail
abnormalities.