Pharmacology and therapeutics

Severe adverse cutaneous drug eruptions: epidemiological

and clinical features
Inès Zaraa, MD, Meriem Jones, MD, Sondes Trojjet, MD, Rym Cheikh Rouhou, MD,
Dalenda El Euch, MD, Mourad Mokni, MD, Amel Ben Osman, MD

Dermatology Department, La Rabta Abstract

Hospital, Jabbari, Bab Saadoun 1007, Introduction Adverse cutaneous drug reactions (ACDR) are common, and some can be
Tunis, Tunisia
lethal. The aim of our study was to discuss the epidemiological and clinical features of
severe ACDR.
Correspondence
Inès Zaraa, MD Materials and methods We retrospectively analyzed 100 cases of ACDR from 1981 to
Dermatology Department 2007, collected in the Department of Dermatology of the La Rabta Hospital in Tunis, which
La Rabta Hospital is located in the north of Tunisia. Severity was defined on three criteria: hospitalization;
Jabbari
visceral involvement; and severity markers.
Bab Saadoun
Results Characteristics of the 54 included cases were: women (70%); mean age:
Tunis
1007 44.8 years; responsible drugs: anticonvulsants (28%), antibiotics (28%), and nonsteroidal
Tunisia anti-inflammatory drugs (15%). The most common dermatoses were maculopapular rash
E-mail: [email protected] (50%). We observed fever (76%), lymphadenopathy (31.5%), eosinophilia (35%), and
visceral involvement (50%). Twelve patients died directly related to the ACDR.
Conclusion This study underlines the polymorphous clinical presentation of ACDR
and the importance of researching some severity markers, which have important practical
implications.

presence of severity markers, such as fever, adenomegaly,

Introduction
involvement of more than 60% of body surface, or eosinophilia.1
Adverse cutaneous drug reactions (ACDR) are a frequent We recorded the age and sex of the patient, the offending
reason for consultations in dermatology.1,2 These reac- drug, the time interval between the drug intake and the
tions are usually benign and self-limited as long as the eruption, the duration of the eruption, the mucocutaneous signs,
causative drug is stopped. In rare cases, ACDRs can be and the results of the paraclinical investigations and
severe and even life-threatening.2,3 pharmacovigilance investigation.
The most common ACDR is morbilliform exanthema Based on morphology, distribution of the lesions, and
(ME), but it is also the least specific manifestation. Other biological exam of the ACDR, we distinguished: ME; acute
potentially severe forms are more specific, but they are generalized exanthematous pustulosis (AGEP); toxic epidermal
less frequently encountered.3 necrolysis (TEN) and Stevens–Johnson syndrome (SJS);
The aim of our study is to analyze the epidemiological erythroderma (ED); and drug hypersensitivity syndrome (DHS).
and clinical characteristics of severe ACDR.
Results
Materials and methods
During the 27 years of the study, 100 patients were admitted
In this monocentric study, a retrospective analysis of all hospital due to ACDR. Fifty-four of them had a severe form. The
records of ACDRs of the Department of Dermatology of La mean age was 44.8 years, ranging from 10 to 81 years. The
Rabta Hospital, located in the north of Tunisia (Tunis), during male to female ratio was 0.4. Drug exposure duration before
the years 1981–2007 was conducted. Only the severe forms of adverse reaction ranged from 3 to 60 days, with an average
ACDR were included. of 25 days. Winter preponderance was noted (40%).
The severity criteria used were based on the presence of at The most common presentation was ME (50%). The
least two criteria in addition to the necessity of hospitalization: a various presentations encountered are summarized in
visceral involvement (kidney, liver, lung, or heart); and the Table 1.
877

ª 2011 The International Society of Dermatology International Journal of Dermatology 2011, 50, 877–880
878 Pharmacology and therapeutics Severe adverse cutaneous drug eruptions Zaraa et al.

Table 1 Clinical subtype of ACDR

Severity Visceral
Hospitalized markers involvement
Clinical subtype cases (cases) (cases)

Morbilliform exanthema (ME) 45 27 7

Acute generalized 12 9 5
exanthematous pustulosis
(AGEP)
Drug hypersensitivity syndrome 7 7 7
(DHS) 9 3 3
Toxic epidermal necrolysis 12 3 2
(TEN)
Erythroderma (ED) 3 1 1
Erythème pigmenté fixe (EPF) 2 1 1
Stevens–Johnson syndrome 2 1 1
(SJS)
Othersa 8 2 0 Figure 2 Face edema in a drug hypersensitivity syndrome
Total 100 54 27 (DHS)

a after an ME. Phenobarbital was the most commonly

Photosensitivity, quincke edema, acute urticaria, vasculitis.
involved drug (six cases). In all other instances, the evolu-
tion was benign.
The anticonvulsants (15 cases: 28%), antibiotics (15
cases: 28%), and nonsteroidal anti-inflammatory drugs
(NSAID; eight cases: 15%) were the main causative drugs.
Other drugs were implicated in the remaining 16 patients,
including: allopurinol (three cases); paracetamol (three
cases); terbinafine (two cases); salazopyrine (two cases);
and chloroquine, captopril, glibenclamide, diltiazem,
iodine contrast product, and acebutolol in one case,
respectively. After pharmacovigilance investigation, the
offending drug was recognized in only 25% of the cases.

Discussion
Adverse drug reactions may affect any organ, and the
skin is a commonly involved organ. ACDR are usually
Figure 1 Numerous pustules of the trunk in an acute gener-
benign. Severe forms are rare, with an estimated propor-
alized exanthematous pustulosis (AGEP)
tion of 2%.2,4 The course can be fatal in 0.2–29.3% of
cases requiring hospitalization when a severe evolution is
Our study included four children (7.4%): three boys predicted.2
and one girl aged, respectively, 10, 10, 16, and 12 years. ACDR prevalence is different from one study to
Two of them had toxic TEN and SJS. another. It is estimated at 0.33–3% in admitted
Severity signs, such as fever, hypereosinophilia, and patients,2–6 0.14% in non-hospitalized patients, and
lymphadenopathy, were present, respectively, in 41, 19, 0.25% in general.7 The prevalence of ACDR in a French
and 17 patients. Visceral involvement was associated in hospital setting is estimated at 0.75/1000 (10 out of
27 cases. These patients had a mean age of 48 years, with 13,294 hospitalizations), and the incidence of ACDR in
a male to female ratio of 0.5. The liver was concerned in France is estimated at 0.4–1.2 per 1.2–6 million habitants
23 cases, the kidneys and lungs in three cases each, and per year.3 In our study, ACDR represented 1.5% of con-
the heart in two cases. Patients died of systemic complica- sultations in dermatology; 3.53% of them required hospi-
tions in 12 cases. The fatal course was mostly observed in talization, and 1.9% were considered as severe.
more specific ACDR, such as TEN in five cases, AGEP in The results obtained in our study are comparable to
three cases (Fig. 1) and DHS in one case (Fig. 2). In two what is reported in the literature: female preponder-
cases, death followed an ED and, in another, it occurred ance,2,7 lesional polymorphism,1 predominance of ME,

International Journal of Dermatology 2011, 50, 877–880 ª 2011 The International Society of Dermatology
Zaraa et al. Severe adverse cutaneous drug eruptions Pharmacology and therapeutics 879

and the possibility of association of these non-specific cal studies, the frequency of ACDR in children is not
cutaneous eruptions with fever, mucous involvement, known, but it seems inferior to that of adults.9 In our
and/or hypereosinophilia.1 study, four children had a severe cutaneous drug reaction.
As previously reported, symptoms appear usually The pattern of drugs causing ACDR is different accord-
within three weeks following the drug intake. ing to country because of differences in prescription hab-
In our study, the increase in ACDR observed in winter its.5,19 Drugs most frequently involved are antibiotic
is probably caused by the higher frequency of infections agents, anticonvulsants, and NSAIDs.1–19 New forms of
and therefore of drug intake. Besides, infections are con- drug eruptions will probably develop because new drugs
sidered as cofactors in the occurrence of ACDR. Recently, are continuously becoming available.9,18,19
authors suggested a possible causality link between ACDR can be life threatening,10 requiring hospitaliza-
HHV6A and HHV7 infection and DHS.8 tion,11 sometimes even intensive care units.2,10 Evolution
The most frequent ACDR encountered in the literature was fatal in 12 of our cases. The anticonvulsants were
are ME (30–51%,1–19 often induced by antibiotics, repre- the most common offending drugs (phenobarbital).
senting more than 3% of the cases.16,17 Despite their high Only a few studies have analyzed severity markers of
frequency, these non-specific ACDR remain unclear. Path- ACDR: Djien’s study1 suggests a possible association
ological findings are equally non-specific, revealing super- between the severity of the eruption and the occurrence
ficial perivascular and dermo-epidermal junction of fever (60%), adenomegaly (30%), eosiniphilia (44%),
lymphocytic infiltrate. The only helpful finding is isolated or visceral involvement (22%).
necrotic keratinocytes in the epidermis. This type of This study underlines the polymorphous clinical presen-
ACDR is usually mild, but certain ME may evolve into tation of skin reactions to drugs and proposes that for
more severe presentations such as SJS and TEN. Factors polymorphous drug reactions, research into severity clini-
predicting this evolution are fever, mucous membrane cal markers that have important practical implications is
involvement, and hypereosinophilia. This suggests that needed. ACDR represent a real public health problem,
ACDR are a spectrum of the same affection ranging from which all dermatologists may be confronted with. ACDR
non-specific and mild ME to the often fatal TEN, with management includes the description of skin manifesta-
factors inducing shifting from one to another.9,19 tions, evaluation of severity signs, and appropriate treat-
Some predisposing factors were identified in previous ment. A precise counsel has to be provided to the patient
studies. They are mainly non-hereditary, such as age of concerning his future drug intake. ACDR can be avoided
patient (over 60 years), female gender, obesity, drug dos- in 15% of cases.6 A careful and adequate prescription of
age, associated drug intake, and associated morbidity, medication with a reduction in the number of drugs may
including immune dysregulation (systemic lupus erythe- help reduce the frequency of drug reactions, especially in
matosus, dysthyroidism, and antiphospholipid antibody advanced-aged patients.3
syndrome), and concomitant viral infections.2,8,16 However, it is essential that any rash be carefully
More often, ACDR are described as morbilliform or monitored for possible, but rare, serious systemic events
maculopapular. Some rashes can be the prelude to much ensuing.
more severe outcomes, and serious systemic events have
been reported. Specific and severe presentations, such as
Conclusions
TEN, SJS, AGEP, and DHS are associated with high mor-
bidity and mortality rates (up to 10%).2,5,10 TEN and SJS Our study showed the importance of a thorough clinical
are the most specific and least frequent forms of ACDR, and biological examination in the management of patients
with the highest mortality rates reaching 30%.10,15 AGEP with specific forms of drug eruption, in search of severity
is a severe cutaneous adverse reaction, but after the makers enabling the detection of visceral involvement that
suspected drug is withdrawn, the skin heals rapidly, and require specific treatment. Besides, any prescription has to
mortality is low.15 be justified because it could be responsible for a severe
Although clinical presentation of ACDR does not seem life-threatening ACDR.
related to age,14,16,17 a higher frequency has been reported
at extreme ages. This is probably caused by a dysfunction
References
in the immune system or an insufficiency of drug metabo-
lism. In our study, the mean age was 44.8 years, with a 1 Djien V, Bocquet H, Dupuy A, et al. Symptomatology
maximum in the fourth and fifth decade. None of our and markers of the severity of erythematous drug
patients had a history of drug reaction, a finding not sup- eruptions. Ann Dermatol Venereol 1999; 126: 247–250.
ported by what has been reported in the literature (31% 2 Hernandez-Salazar A, Ponce-de-Leon Rosales S, Rangel-
vs. 9.3%). Because of the limited number of epidemiologi- Frausto S, et al. Epidemiology of adverse cutaneous drug

ª 2011 The International Society of Dermatology International Journal of Dermatology 2011, 50, 877–880
880 Pharmacology and therapeutics Severe adverse cutaneous drug eruptions Zaraa et al.

reactions: a prospective study in hospitalized patients. Indian Hospital. Pharmacoepidemiol Drug Saf 2005; 14:
Arch Med Res 2006; 37: 899–902. 567–570.
3 Roujeau JC, Stern RS. Severe adverse cutaneous reactions 12 Stern RS. Utilization of hospital and outpatient care for
to drugs. N Engl J Med 1994; 331: 1272–1284. adverse cutaneous reactions to medications.
4 Wolf R, Orion E, Marcos B, et al. Life-threatening acute Pharmacoepidemiol Drug Saf 2005; 14: 677–684.
adverse cutaneous drug reactions. Clin Dermatol 2005; 13 Zanocchi M, Tibaldi V, Amati D, et al. Adverse drug
23: 171–181. reactions as cause of visit to the emergency department:
5 Fiszenson-Albala F, Auzerie V, Mahe E, et al. A 6-month incidence, features and outcomes. Recenti Prog Med
prospective survey of cutaneous drug reactions in a 2006; 97: 381–388.
hospital setting. Br J Dermatol 2003; 149: 1018–1022. 14 Laroche ML, Charmes JP, Nouaille Y, et al. Is
6 Davies EC, Green CF, Mottram DR, et al. Adverse drug inappropriate medication use a major cause of adverse
reactions in hospital in-patients: a pilot study. J Clin drug reactions in the elderly? Br J Clin Pharmacol 2007;
Pharm Ther 2006; 31: 335–341. 63: 177–186.
7 Borch JE, Andersen KE, Bindslev-Jensen C. The 15 Davies EC, Green CF, Mottram DR, et al. Adverse drug
prevalence of acute cutaneous drug reactions in a reactions in hospital in-patients: a pilot study. J Clin
scandinavian university hospital. Acta Derm Venereol Pharm Ther 2006; 31: 335–341.
2006; 86: 518–522. 16 Sharma VK, Sethuraman G, Kumar B. Cutaneous adverse
8 Descamps V, Valance A, Edlinger C, et al. Association of drug reactions: clinical pattern and causative agents – a 6
human herpesvirus 6 infection with drug reaction with year series from Chandigarh, India. J Postgrad Med
eosinophilia and systemic symptoms. Arch Dermatol 2001; 47: 95–99.
2001; 137: 301–304. 17 Svensson CK, Cowen EW, Gaspari AA. Cutaneous drug
9 Roujeau JC. Toxic dermal reactions in the child. Arch reactions. Pharmacol Rev 2001; 53: 357–379.
Pediatr 2000; 7: 215s–217s. 18 Bigby M. Rates of cutaneous reactions to drugs. Arch
10 Campos-Fernandez Mdel M, Ponce-De-Leon-Rosales S, Dermatol 2001; 137: 765–770.
Archer-Dubon C, et al. Incidence and risk factors for 19 Wolkenstein P, Revuz J. Allergic emergencies encountered
cutaneous adverse drug reactions in an intensive care by the dermatologist. Severe cutaneous adverse drug
unit. Rev Invest Clin 2005; 57: 770–774. reactions. Clin Rev Allergy Immunol 1999; 17: 497–511.
11 Sushma M, Noel MV, Ritika MC, et al. Cutaneous
adverse drug reactions: a 9-year study from a South

International Journal of Dermatology 2011, 50, 877–880 ª 2011 The International Society of Dermatology