BJD

C L I N I C A L A N D L A B O R A T O R Y I N V E S TI G A T I O N S British Journal of Dermatology

Prognosis of generalized bullous xed drug eruption:

comparison with StevensJohnson syndrome and toxic
epidermal necrolysis
S. Lipowicz,1 P. Sekula,2 S. Ingen-Housz-Oro,1 Y. Liss,3 B. Sassolas,4 A. Dunant,5 J.-C. Roujeau1 and
M. Mockenhaupt3
1
Department of Dermatology, Hopital Henri Mondor, Universite Paris-Est-Creteil, Creteil, France
2
Institute of Medical Biometry and Medical Informatics (IMBI); 3Dokumentationszentrum schwerer Hautreaktionen (dZh), Department of Dermatology;
University Medical Center, Freiburg, Germany
4
CHU Brest, France
5
Biostatistics and Epidemiology Unit, Institut Gustave-Roussy, Villejuif, France

Summary

Correspondence Background Generalized bullous fixed drug eruption (GBFDE) is a rare cutaneous
Jean-Claude Roujeau. adverse reaction to drugs, and may resemble StevensJohnson syndrome (SJS) or
E-mail: jean-claude.roujeau@wanadoo.fr toxic epidermal necrolysis (TEN), but is usually considered less severe.
Objectives To compare the severity and mortality rate in cases of GBFDE and con-
Accepted for publication
3 November 2012 trol cases of SJS or TEN of similar extent of skin detachment.
Methods This was a casecontrol analysis of 58 patients with GBFDE matched by
Funding sources age and extent of skin detachment to 170 control patients with a validated diag-
None. nosis of SJS or SJS TEN overlap. Data for cases and controls were extracted from
the EuroSCAR and RegiSCAR databases resulting from two population-based stud-
Conicts of interest
ies of severe cutaneous adverse reactions conducted in Europe. Preselected out-
M.M., J.-C.R. and B.S. have been experts for drug
causality of severe cutaneous adverse drug reactions come criteria were death (primary), and fever, duration of hospitalization and
for several pharmaceutical companies, as well as transfer to an intensive care or burn unit (secondary).
expert witnesses for trials related to drug causality Results GBFDE affected mainly older patients (median age 78 years, interquartile
of severe cutaneous adverse reactions. range 6884 years); 13 of 58 cases died (22%). The mortality rate was slightly
but not significantly lower for patients with GBFDE than controls [28%, multivar-
DOI 10.1111/bjd.12133
iate odds ratio 06 (95% confidence interval 03014)]. Patients with GBFDE and
controls did not differ in other preselected criteria for severity.
Conclusions Although our study featured limited statistical power, we were not able
to confirm that GBFDE had better prognosis than SJS or SJS TEN of similar dis-
ease extent in older patients. Severe cases of GBFDE deserve the attention and
active management given to patients with SJS or TEN.

Fixed drug eruption (FDE) is a specific phenotype among
adverse drug reactions affecting the skin. It is characterized by
a limited number of round erythematous patches (usually a
few centimetres in size) that resolve in a few days, often leav-
ing long-lasting pigmentation. One characteristic feature is
recurrence on the same sites a few hours after rechallenge
with the same medication.1 Initially reported in association
with antipyrine, FDE can be associated with various analgesics,
antibiotics, laxatives and other medications.1 The incidence of
FDE is not known, but it is suspected to vary greatly by geo-
graphical area.
Occasionally, blisters may develop on the lesions. Rarely,
such blisters may constitute large areas of the body surface.
Reports of such events generally cite the diagnosis of general-
ized bullous FDE (GBFDE).2,3 Confusion between GBFDE and
toxic epidermal necrolysis (TEN) has historically been a prob-
lem. Lyell acknowledged that two of the four cases included
in his original description of TEN4 should have been diag-
nosed as GBFDE.5
Most dermatologists now consider that the features of
GBFDE are original enough to distinguish it from TEN
(Fig. 1).68 In contrast to TEN, GBFDE is considered to consist
of an absence or paucity of constitutional symptoms, well-
demarcated blisters and erythematous patches, absence of
small spots or target lesions, absence or paucity of mucosal
erosions, absence or paucity of visceral complications, history
of a similar eruption, and onset within hours of exposure
to the associated drug. Most dermatopathologists also

726 British Journal of Dermatology (2013) 168, pp726732  2013 The Authors
BJD  2013 British Association of Dermatologists
 Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al. 727

SJS or SJS/TEN overlap cases GBFDE cases

Widespread small or confluent Large well demarcated blisters without
spots or atypical targets spots or atypical targets.

Case 3360363, 12% Case 9170765, 12%

Case 9150097, 7% Case 9170765, 12%

Case 9150097, 7% Case 9170572, 8%

Fig 1. Examples of generalized bullous fixed

drug eruption (GBFDE) and StevensJohnson
syndrome (SJS) or SJS toxic epidermal
necrolysis (TEN) overlap. Left column, cases
of SJS or SJS TEN overlap with widespread
small or confluent spots or atypical targets.
Right column, cases of GBFDE with large,
well-demarcated blisters without spots or
atypical targets. An insert in each picture
provides the individual case number and the
percentage of body surface area detachment.

consider that the necrosis of the epidermis observed in more pigment incontinence.8 However, in the acute
GBFDE differs from that in TEN by more abundant infiltra- stage, histological differentiation may be difficult or even
tion by lymphocytes (including many regulatory T cells) and impossible.

 2013 The Authors British Journal of Dermatology (2013) 168, pp726732

BJD  2013 British Association of Dermatologists
728 Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al.

GBFDE suspected by EuroSCAR or RegiSCAR experts n = 72

Photographs available n = 57 Photographs non available n = 15

Non Compatible (1 Suggestive (2 2 or more Less than

suggestive skin criteria*) skin criteria*) criteria* 2 criteria*
(0 skin plus at least n = 43 n=6 n=9
criterion*) 1 non
skin criterion*
n=5
n=9

Validated GBFDE n = 58
Community n = 38
Hospitalized n = 20
*Criteria:
Non skin: Skin:
Prior similar reaction, Absence of spots/targets,
Less than 2 mucous membranes involved, Well demarcated blisters/erosions Fig 2. Selection of cases of generalized
bullous fixed drug eruption (GBFDE).

The distinction between GBFDE and StevensJohnson syn-
drome (SJS) or TEN is important because of the supposed bet-
ter prognosis of GBFDE than SJS or TEN.3,6,8 We lack evidence
to support its generally considered benign nature.
We aimed to confirm the better prognosis with GBFDE than
SJS or TEN using a casecontrol design, matching the main
validated prognostic factors for SJS or TEN, namely age, and
extent of skin detachment as a percentage of body surface area
(BSA).9
Patients and methods
This casecontrol design involved validated cases of GBFDE and
validated cases of SJS or TEN as controls. To compare severity
adequately in both groups, controls (three per case) were
matched to cases for age and the extent of blisters or detach-
ment, known to be major prognostic factors in SJS or TEN.
Selection of cases and controls
Data for cases and controls were obtained from the databases
of the EuroSCAR and RegiSCAR multinational studies of SJS
and TEN.10,11 EuroSCAR was conducted from 1997 to 2001
in Austria, France, Germany, Israel, Italy and the Netherlands,
and RegiSCAR started in 2003 in the same countries. Patients
were selected up to 2005. Both studies had been approved by
ethics committees in each participating country. Both studies
were based on facilitated reporting of potential cases of SJS or
TEN by a network of hospitals, and information had been col-
lected by the interviewing of patients, their relatives and phys-
icians by trained interviewers (medical professionals) who
used a standardized questionnaire.
For the present study, we included data for 72 cases of po-
tential GBFDE reported to EuroSCAR or RegiSCAR as potential
cases of SJS or TEN for which the experts had excluded a diag-
nosis of SJS or TEN and suspected a diagnosis of GBFDE.
Three authors (S.L., S.I.-H.-O. and J.-C.R.), who were
blinded to any information on potential disease causes and
outcome, reviewed the charts of all 72 cases of potential
GBFDE and retained data for 58 cases of probable or definite
GBFDE (Fig. 2); data for 21 cases were retrieved from the
EuroSCAR database and 37 from the RegiSCAR database.
In the absence of a clear consensual definition of GBFDE,
cases were evaluated on the basis of expert judgement. Diag-
nosis of GBFDE was retained on the presence of at least two
(probable case) or three (definite case) of the following pre-
defined criteria: (i) similar reaction in the past; (ii) fewer than
two mucous membranes involved, absence of spots or target
lesions; (iii) large and well-demarcated blisters and erosions;
and (iv) lesions and erosions on at least two different sites of
the body (a site being defined as part of the body area as used
for the rule of nine for estimating BSA) to distinguish gener-
alized from localized forms of FDE.
We are not aware of any cut-off related to BSA involve-
ment in previous definitions of GBFDE and did not use
any.
Photographs were available for 57 cases and were used to
evaluate the two latter criteria, which were considered suggestive

British Journal of Dermatology (2013) 168, pp726732  2013 The Authors

BJD  2013 British Association of Dermatologists

(both present, 43 cases), compatible (one present, nine cases)
or nonsuggestive (both absent, five cases). Cases without pho-
tographs were considered only probable.
Among the 58 cases validated as GBFDE, the extent of skin
detachment was re-evaluated from clinical photographs if
available (52 of 58 validated cases) or as originally reported
in the case report form.
Cases of SJS or TEN had been evaluated with definitions
previously published12 and validated.13 Diagnosis had been
determined by EuroSCAR or RegiSCAR experts who reviewed
clinical information (widespread spots or atypical targets with
blisters and erosions on the skin and erosions of mucous
membranes), photographs and skin pathology results with
blinding to risk factors and outcome. Only cases with a diag-
nosis considered definite or probable were included.10,11 The age distribution is shown in Table 1. More women than
For cases and controls the following information was
extracted from the databases: chronology and clinical descrip-
tion of lesions affecting the skin and mucous membranes;
classification of cases as hospitalized when occurring during
hospitalization for another disease or community when
occurring outside the hospital and leading to hospital admis-
sion; demographics; and outcome measures.
Frequency matching on age (categorized by 10-year inter-
vals) and proportion of BSA detachment (categorized as 0
9%, 1019%, 2029%, 30%) was used to select three con- erosive lesions on two or more mucosal sites. Because there were
trols per case from the same databases (n = 170; 44 from
EuroSCAR and 126 from RegiSCAR). For four cases, the selec-
tion was incomplete (only two controls available per case).
Outcome measures
Severity of disease was evaluated by four preselected criteria;
in-hospital death was the main endpoint and transfer to spe-
cialized wards (intensive care unit, burn unit) and fever were
secondary endpoints. Moreover, duration of hospital stay for
survivors of community-acquired reactions was assessed as a
secondary endpoint.
Statistical analyses
To detect an odds ratio (OR) of death < 033 for cases vs.
controls based on an expected mortality rate of 03 in controls
(unilateral test, a = 5%, power = 90%), we needed a sample
size of 53 cases and three times as many controls (http://
www.u707.jussieu.fr/biostatgv/).
Binary variables were compared by v2-test and continuous
variables by nonparametric KruskalWallis test.
For binary outcomes (death, transfer to an intensive care or
burn unit, fever), we used multivariable logistic regression
with adjustment for age and maximum skin detachment (pre-
specified). Further potential confounders (study, country, and
community or in-hospital patient) were considered by using
backward elimination (significance level for exclusion, P =
005). In addition, pairwise interactions were assessed by use
of forward selection (significance level for inclusion,
P = 001).
 2013 The Authors
BJD  2013 British Association of Dermatologists
Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al. 729
For the continuous endpoint, length of hospital stay, multi-
variable linear regression with a similar modelling strategy
was used. Because of a skewed distribution of length of hospi-
tal stay, we transformed the variable (x03). However, for con-
venience and because of difficult interpretations of the linear
regression coefficients with the transformed outcome variable,
we used the final selected variables with the original (untrans-
formed) length of hospital stay variable.
Results
The re-review process validated 58 cases as GBFDE (Fig. 2).
With the exception of three paediatric cases, patients with
GBFDE were older (median age 78 years, range 393 years).
men had GBFDE (n = 36, 62% vs. n = 22, 38%).
Most patients with GBFDE (30 58) had skin detachment on
< 10% of BSA. Only nine had 20% detachment and none
had > 30% (Table 1). Therefore, matching by extent of
detachment included patients with SJS (< 10% of BSA
involved) and with SJS TEN overlap (1029% BSA involved)
and none with TEN (> 30% BSA involved).
Although the paucity of mucous-membrane involvement was
one of the diagnostic criteria for GBFDE, 18 patients (31%) had
few available photographs of mucous membranes, we could not
determine whether these erosions were diffuse small erosive
lesions as commonly observed in SJS or TEN or a limited num-
ber of well-demarcated bullous lesions as is more usual in FDE.
In total, 32 of the 50 patients with data available (64%)
had a history of adverse drug reactions, including 19 (38%)
who reported a blistering eruption.
A total of 20 patients (34%) with GBFDE and 66 (39%)
with SJS or SJS TEN overlap (P = 06) reported potentially
severe associated diseases. More patients with SJS or SJS TEN
than GBFDE had several associated diseases (16 vs. 4).
In all, 13 of 58 patients with GBFDE died [22%, 95% confi-
dence interval (CI) 1335%]. Death was more frequent among
patients with disease that began in hospital, compared with
others (8 20 vs. 5 38, P = 002), and also more frequent for
those aged > 60 years (13 46 vs. 0 12, P = 004) or with
10% BSA detachment (9 28 vs. 4 30, P = 009).
The 58 cases of GBFDE were matched by age and extent of
detachment to 170 cases of SJS or SJS TEN overlap. Cases and
controls did not differ by sex, percentage of community cases
or frequency of serious associated conditions (Table 1). Not
surprisingly, diagnostic criteria such as involvement of fewer
than two mucous membranes and previous episodes were
more frequent in patients with GBFDE than others.
GBFDE cases and controls did not differ in preselected
severity criteria (Table 2) by univariate or multivariate analy-
ses. The mortality rate did not differ between patients with
GBFDE and SJS or SJS TEN overlap [univariate OR 08 (95%
CI 0415); multivariate OR 06 (95% CI 0314)].
The delay between onset of reaction and death did not dif-
fer between patients with GBFDE and SJS or SJS TEN overlap
British Journal of Dermatology (2013) 168, pp726732
730 Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al.

Table 1 Characteristics of cases and controls

GBFDE (cases) SJS TEN (controls)
n = 58 n = 170 P-value
Sex, female male 36 22 114 56 05
Age, yearsa (median, IQR) 78, 6884 76, 6583 05
019 3 9
2039 2 6
4059 7 21
6079 22 66
8099 24 68
Extent of detachment,a % (mean SD) 108 68 105 68 07
09% 30 90
1019% 19 56
2029% 9 24
Erosions of < 2 mucosab 40 (69) 33 (19) < 001
Prior episodeb,c
Any prior drug reaction 32 50 (64) 28 153 (18) < 0001
Reported as similar 19 50 (38) 2 153 (13) < 0001
Disease onset during hospital stayb 20 (34) 40 (24) 01
Any serious associated conditionb,d 20 (34) 66 (39) 06
Malignancy 9 33
Chronic kidney disease 4 30
Chronic liver disease 5 10
Collagen vascular rheumatism 6 10
Country: Germany France other 31 19 8 109 37 24 02
a
Matching criteria. bn(%). cMissing information in eight cases of generalized bullous fixed
drug eruption (GBFDE) and 17 cases of StevensJohnson syndrome (SJS) toxic epidermal
necrolysis (TEN). dMore than one present in the same patient in four cases of GBFDE and
16 cases of SJS TEN. IQR, interquartile range.

Table 2 Outcomes for cases and controls

For binary outcomes (full cohort)

GBFDE n = 58, SJS TEN n = 170, Logistic regression, OR (95% CI)c

number (%) number (%) Univariate Multivariated

Fever 385 Ca 20 (39) 71 (46) 08 (0414) 06 (0313)

Death 13 (22) 47 (28) 08 (0415) 06 (0314)
Intensive care or burn unit 23 (40) 73 (43) 09 (0516) 08 (0417)
For continuous outcome (only survivors of community-acquired reactions)

Linear regression,
GBFDE n = 33, SJS TEN n = 94, Betae
median (IQR) median (IQR) Univariate Multivariate
Hospital stay, days b
17 (1223) 205 (1429) )11 ()66+45) )09 ()62+45)
a
Seven patients with generalized bullous fixed drug eruption (GBFDE) and 16 patients with StevensJohnson syndrome (SJS) toxic epidermal
necrolysis (TEN) had to be excluded due to incomplete information on fever. bOne patient with SJS TEN was excluded because the date of
discharge was not known. cAll presented odds ratios (ORs) with confidence intervals (CIs) indicate although not statistically significant a
lower risk for what is stated in the group of GBFDE in comparison with SJS TEN. dAdjustment: age and extent of detachment for all out-
comes; country for fever and hospital stay; and country, study and extent of detachment by study interaction for transfer to intensive care or
burn units. eVariable selection based on transformed outcome. Beta: the interpretation regarding hospital stay is that patients with GBFDE
were, on average and after adjustment, 1 less day hospitalized, again not significant. In general, the coefficient represents how much the
dependent variable (here, hospital stay with scale in days) is expected to increase (if the coefficient is positive) or decrease (if the coefficient
is negative) when that independent variable increases by one (here, from GBFDE to SJS TEN). IQR, interquartile range.

(23 10 vs. 22 24 days). No information was collected on vs. 13%, P = 002) than for SJS or SJS TEN overlap (death rate
the direct cause of death. 30% vs. 27%). Additional inclusion of the variable community
The impact on death of already being hospitalized for another vs. in-hospital cases in the multivariable analysis did not sub-
disease seemed greater for patients with GBFDE (death rate 40% stantially change the OR for death (06, 95% CI 0314).

British Journal of Dermatology (2013) 168, pp726732  2013 The Authors

BJD  2013 British Association of Dermatologists

Further exploration of cases of GBFDE revealed no signifi-
cant relation between the number of mucosal sites involved
and the risk of death (data not shown).
Discussion
We could not confirm the generally accepted concept that
GBFDE has better prognosis than SJS or TEN. Our cases of
GBFDE did not differ from patients with SJS or TEN matched
for age and extent of skin detachment in any of the four pres-
elected criteria for evaluating severity.
Of note, the mortality rate was 22% for patients with
GBFDE, despite the moderate mean area of detached epider-
mis. The older age of our cases of GBFDE (and therefore of
matched cases of SJS or TEN) likely contributed to the poor
prognosis we observed. In both groups, 70% of patients were
older than 70 years. Several associated conditions, such as
recent malignancy or impaired renal function, imply poor
prognosis in SJS or TEN,9,14 which may also be the case for
GBFDE. Patients with GBFDE and SJS or TEN did not differ in
the prevalence of associated chronic diseases, so comorbidities
did not likely contribute to the poorer than expected out-
comes with GBFDE.
The cases of GBFDE that we included were not representa-
tive of bullous FDE because all were initially suspected to be
SJS or TEN. Therefore, our GBFDE cases represented unusually
widespread forms, also with increased probability of having
mucous-membrane erosions, and our results cannot be extrap-
olated to FDE in general.
The predominance of women that we observed among our
cases of GBFDE has been reported with many other adverse
drug reactions, including SJS or TEN.10 The median age of
patients with GBFDE (78 years) was strikingly older than that
for patients with SJS or TEN in general; in the EuroSCAR
study, the median age of patients with SJS or TEN was
50 years.10
Also, only one-third of our patients with GBFDE had a his-
tory of blistering reaction to drugs, which is considered a key
criterion for a diagnosis of GBFDE. Recall bias may have con-
tributed to underestimating the real prevalence of prior epi-
sodes, but our findings indicate that most cases show no
history of prior reaction.
The major strengths of our study are the collection of a
large number of cases of GBFDE and comparison with a con-
trol group with SJS or SJS TEN overlap, matched on the two
main identified prognostic factors for SJS or TEN: age and
extent of skin detachment. Standardized collection of data also
afforded a similar quality of information that was available for
cases and controls.
Our study has several limitations. The criteria used for
defining the cases as probable or definite GBFDE need further
validation with another group of patients, as was previously
done for SJS TEN. We limited the study to evaluating disease
severity and did not address other questions such as histopa-
thology or drug causality. Finally, our sample size allowed for
detection only of a large difference in mortality rates.
 2013 The Authors
BJD  2013 British Association of Dermatologists
Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al. 731
Although we cannot exclude a moderate difference in mor-
tality rates between the two disease types, our results argue
against the general belief that GBFDE has a better prognosis
than SJS or TEN. Keeping the above limitations in mind, these
results are not surprising. For burns and autoimmune blister-
ing diseases14 the extent of skin involved is a well-established
severity marker. In SJS and TEN, age and extent of skin
detachment are the two most important prognostic factors.9,11
GBFDE with extensive skin detachment was previously con-
sidered less severe than SJS or TEN because, firstly, the paucity
of mucous-membrane involvement leads to less pain and less
severe clinical presentation. However, in SJS or TEN the sever-
ity of initial mucous-membrane involvement was never found
to contribute to mortality.9 Secondly, the process of GBFDE
was suggested to be more limited than that of SJS or TEN,
being under the control of large numbers of regulatory T
cells8 that could block the release of inflammatory cytokines,
with milder constitutional symptoms and more rapid repair of
the epidermis. Our data did not support less frequent occur-
rence of fever with severe GBFDE or more rapid re-epithelial-
ization, which should lead to a shorter hospital stay.
Our findings strongly suggest that GBFDE, especially in
older patients already hospitalized for another condition, is a
severe cutaneous adverse reaction with a high mortality rate.
Therefore, the potential severity of GBFDE should not be
underestimated. This disease likely deserves the same symp-
tomatic treatment and care as SJS or TEN, and further efforts
towards a more clear definition of clinical features, with estab-
lishment of diagnosis criteria.
Whats already known about this topic?
Generalized bullous fixed drug eruption (GBFDE) is a
rare variant of FDE.
It is often confused with toxic epidermal necrolysis
(TEN) and is considered less severe than TEN.
What does this study add?
Comparison of GBFDE cases with StevensJohnson syn-
drome (SJS) or SJS TEN cases matched for age and
extent of skin detachment was carried out.
The mortality rate was slightly but not significantly
lower for GBFDE than for cases of SJS or SJS TEN.
The reputation of the benign nature of GBFDE should
be reconsidered.
Acknowledgments
The authors thank all members of the EuroSCAR and Regi-
SCAR studies who contributed to the inclusion of cases:
Hele`ne Bocquet, Malika Bounoua, Alain Dupuy, Jean-Paul
Fagot, Antoine Flahault, Laurence Valeyrie-Allanore (France);
Konrad Bork, Martine Grosber, Uwe-Frithjof Haustein, Alexan-
British Journal of Dermatology (2013) 168, pp726732
732 Prognosis of GBFDE vs. SJS TEN, S. Lipowicz et al.
der Hellmer, Christiane Marks (Germany); Arnon D Cohen,
Batya Davidovici, Sima Halevy (Israel); Luigi Naldi (Italy);
Christianne Bearda Bakker-Wensveen, Jan-Nico Bouwes-
Bavinck, Esther Lai A Fat, Sylvia Kardaun (Netherlands).
References
1 Kauppinen K, Stubb S. Fixed eruptions: causative drugs and chal-
lenge tests. Br J Dermatol 1985; 112:5758.
2 Baird BJ, de Villez RL. Widespread bullous fixed drug eruption
mimicking toxic epidermal necrolysis. Int J Dermatol 1988; 27:170
4.
3 Lin T, Hsu MM, Lee JY. Clinical resemblance of widespread bullous
fixed drug eruption to StevensJohnson syndrome or toxic epider-
mal necrolysis: report of two cases. J Formos Med Assoc 2002;
101:5726.
4 Lyell A. Toxic epidermal necrolysis: an eruption resembling scald-
ing of the skin. Br J Dermatol 1956; 68:3556.
5 Lyell A. Requiem for toxic epidermal necrolysis. Br J Dermatol 1990;
122:8378.
6 Valeyrie-Allanore L, Roujeau JC. Epidermal necrolysis (Stevens
Johnson syndrome and toxic epidermal necrolysis). In: Fitzpatricks
Dermatology in General Medicine (Wolff K, Goldsmith LA, Katz SI, Gilch-
rest BA, Paller A, Leffell DJ, eds), 7th edn. New York: McGraw-
Hill, 2007; ch. 39.
British Journal of Dermatology (2013) 168, pp726732
7 Mockenhaupt M. Epidemiology and causes of severe cutaneous
adverse reactions to drugs. In: Drug Hypersensitivity (Pichler W, ed),
Basel: Karger Verlag, 2007; 1831.
8 Shiohara T. Fixed drug eruption: pathogenesis and diagnostic tests.
Curr Opin Allergy Clin Immunol 2009; 9:31621.
9 Bastuji-Garin S, Fouchard N, Bertocchi M et al. SCORTEN: a sever-
ity-of-illness score for toxic epidermal necrolysis. J Invest Dermatol
2000; 115:14953.
10 Mockenhaupt M, Viboud C, Dunant A et al. StevensJohnson syn-
drome and toxic epidermal necrolysis: assessment of medication
risks with emphasis on recently marketed drugs. The EuroSCAR
study. J Invest Dermatol 2008; 128:3544.
11 Sekula P, Liss Y, Davidovici B et al. Evaluation of SCORTEN on a
cohort of patients with StevensJohnson syndrome and toxic epi-
dermal necrolysis included in the RegiSCAR study. J Burn Care Res
2011; 32:23745.
12 Bastuji-Garin S, Rzany B, Stern RS et al. Clinical classification of
cases of toxic epidermal necrolysis, StevensJohnson syndrome,
and erythema multiforme. Arch Dermatol 1993; 129:926.
13 Auquier-Dunant A, Mockenhaupt M, Naldi L et al. Correlations
between clinical patterns and causes of erythema multiforme ma-
jus, StevensJohnson syndrome, and toxic epidermal necrolysis:
results of an international prospective study. Arch Dermatol 2002;
138:101924.
14 Pfutze M, Niedermeier A, Hertl M et al. Introducing a novel Auto-
immune Bullous Skin Disorder Intensity Score (ABSIS) in pemphi-
gus. Eur J Dermatol 2007; 17:411.
 2013 The Authors
BJD  2013 British Association of Dermatologists