According to the World Health Organization (WHO), an adverse drug reaction is defined as any harmful and unintended drug use response that occurs at doses normally used in humans for the prophylaxis, diagnosis, or treatment of diseases.1 Adverse cutaneous drug reactions frequently occur, accounting for 10%-15% of all reported cases of adverse drug reactions.2 In the hospital setting, the reported incidence ranges from 2% to 3% and varies from mild, self-limiting cases to severe reactions, which can cause significant morbidity and mortality.3

In Brazil, few studies have been published on the subject, and only two of them evaluated the prevalence and the clinical forms of cutaneous drug reactions in the hospital environment. In the first study, published in 1984 by Weissbluth et al.,4 12.3% of the patients evaluated by the interconsultation service received a diagnosis of cutaneous drug reactions, while in the second study, published in 2015 by Botelho et al.,5 this percentage was 8.2%.

This study aims to analyze the profile of cutaneous drug reactions in patients hospitalized in a quaternary hospital, with the characterization of the demographic profile (gender, age and phototype), clinical forms, suspected medications, histopathological alterations, presence of systemic effects, implemented treatment, and course.

This is a clinical, retrospective, observational and descriptive study, which started after approval by the institution’s Research Ethics Committee (protocol number CAAE 59323616.0.0000.5505). All patients attended by the Dermatology Interconsultation service of a quaternary hospital with a diagnosis of cutaneous drug reaction from January 2013 to December 2016 were included, totaling 48 months in the analysis.

The collected sample consisted of 219 patients. Cases in which essential data were incomplete and whose recovery was not possible due to loss of medical records or other reasons were excluded. Data collection was carried out by reviewing the electronic database organized by the Hospital Dermatology Sector of the Institution and by searching the manual and electronic medical records of patients on the Hospital server. When necessary, complementary tests performed during hospitalization and histopathological review of skin biopsies were collected. Photographs were also consulted for better characterization of the clinical form of the cutaneous drug reaction. To define the causes of the cutaneous drug reaction, the drugs used by the patient within a period of up to 14 days before the onset of skin lesions were taken into account.

Analyzing the demographic data of this study, a slight female predominance (50.6%) was observed, differently from what is evidenced in the most recent literature, whose male to female ratio ranged from 1.14/1 to 1.55/1.3,6,7 Regarding the age of the patients, the mean age of 47 years was consistent with the literature data, between 45 and 55 years.2,3,5,8,9

The mean time between symptom onset and the request for interconsultation was six days, very similar to the result of 5.9 days found in a study published by the same group in 2016.5 This time interval may reflect the non-identification or non-appreciation of skin changes by the assistant team, which delays the request for dermatological evaluation. It is important to note that the delay in the diagnosis of drug reactions can be fatal, especially in severe forms such as SJS/TEN and DRESS, in which the prognosis is directly related to the early withdrawal of the suspected drug and the implementation of specific treatment when indicated.10,11

The class of drugs that most frequently caused cutaneous drug reactions was antibiotics, at 36.5%, followed by anticonvulsants, at 10%, data very similar to those in the literature.2–4,7,8,12 Among the antibiotics, broad-spectrum drugs (vancomycin, meropenem, polymyxin B) were involved, which are commonly prescribed in a high-complexity quaternary hospital environment and are rarely used in primary and secondary care services, outlining a distinct profile of eliciting drugs compared to non-hospital environments.

The same was observed among severe cutaneous drug reactions (SJS/TEN, DRESS and AGEP), whose main causes were antibiotics (32%) and anticonvulsants (21%), in agreement with a recent study that included 928 cases of the Chinese population, in which the main eliciting drugs were antibiotics (26%), followed by anticonvulsants (21.6%).13

However, there are conflicting data in the literature: in studies carried out in the Brazilian population, a statistical association was found between the use of anticonvulsants and severe cutaneous drug reactions.5 In these studies, anticonvulsants were still identified as the main agents (40.4%) of severe cutaneous drug reactions, followed by antibiotics (22.8%).14 In a study carried out in China in 2008,12 the class of anticonvulsants was also the most frequent cause of severe skin reactions, accounting for 38.5% of cases.

Regarding anticonvulsants, it is worth mentioning that the three medications most frequently involved in the present study (phenytoin, carbamazepine, phenobarbital) are aromatic anticonvulsants belonging to the same class. Therefore, in these cases, replacement therapy should be based on a non-aromatic anticonvulsant such as valproic acid or levetiracetam, since changing to a drug of the same class will perpetuate the immunological process triggered in the cutaneous drug reaction.15,16

In the present sample, however, it is worth noting that, among the 93 patients (42.3%) in whom it was not possible to identify the eliciting drug, 86 used two or more classes of medication. This probably reflects the complexity of cases admitted to a quaternary hospital.

Regarding the clinical manifestation of the adverse drug reaction, 65.7% of the patients had mild forms, with exanthema being the most common (77% of the mild forms and 50.6% of the total sample). This clinical form of cutaneous drug reaction was also found in other studies in the literature.4,5,7–9,11,17–19 It is noteworthy that there were no cases of fixed drug eruption in the evaluated sample. It is possible that very localized lesions were not observed by the assistant team, which possibly failed to request the Dermatology evaluation.

The severe forms represented 34.3% of the cases, a high frequency when compared to most of the other studies, in which the severe forms represented 3% to 25% of the cases,2,3,7–9,19 justified by the quaternary profile of high hospital complexity. Among the severe forms, DRESS was the most prevalent one (53.3% of the severe forms and 18.2% of the total), in agreement with a previous study carried out in the same country (45.6% of the severe forms)14 and with a study carried out in Malaysia in 2017 (63.6% of severe forms).3 These data, however, were different from others found in the literature, in which DRESS significantly showed lower prevalence rates (0.03% to 14.5%).2,5,6,9,17

Histopathological findings in retrospective studies on the frequency and clinical aspects of cutaneous drug reactions have been poorly described. In a series of 25 cases of DRESS, only 7 cases (28%) underwent skin biopsy and histopathological analysis.20 This study shows histopathological data from the vast majority of cases (82.6%), the main finding being interface dermatitis, both in mild and severe forms of cutaneous drug reactions. However, in 17% of the patients, this study was nonspecific, and most of them were cases of DRESS.

Among the histopathological findings of cutaneous drug reactions, those related to DRESS raise further discussion in the literature. The presence of atypical lymphocytes in the dermal infiltrate has been associated with the use of anticonvulsants and is most frequently seen in DRESS.5 Additionally, the intensity of keratinocyte necrosis, inflammatory infiltrate, and the presence of vasculitis was more related to DRESS compared to exanthema with or without systemic repercussions.21 It is postulated that histopathological findings may have prognostic value. The erythema multiforme-like histopathological pattern could indicate greater severity of liver involvement in cases of DRESS.22

As therapeutic approach, in addition to the withdrawal of the suspected drug(s) in all cases, systemic corticosteroid therapy was indicated in almost 50% of them. It is important to note that even among severe cutaneous drug reactions, the use of systemic corticosteroids is controversial.

Regarding DRESS, systemic corticosteroid therapy has been considered the main treatment, especially when there is severe involvement of one or more vital organs (liver, kidneys, lungs, heart). Among the 40 cases diagnosed with DRESS, 34 (85%) received systemic corticosteroids, 29 (72.5%) at an immunosuppressive dose (prednisone 1 mg/kg/day or equivalent). In other studies, systemic corticosteroid therapy was prescribed in 34% to 52% of the cases, whereas cases with mild cutaneous and visceral involvement were treated with high-potency topical corticosteroids alone.20,23

Among patients in the SJS/TEN spectrum, systemic corticosteroid therapy was given to all patients in the SJS group, and IVIg in association or not with corticosteroids in 58% of the TEN cases, similar to data from a Korean study in which 87% of SJS patients were treated with corticosteroids and 80% of TEN cases with IVIg.24

In the literature, the systemic treatment of SJS/TEN remains controversial. Two cohort studies and a review of the literature have not demonstrated any benefit with the use of systemic corticosteroids in the treatment of the SJS/TEN spectrum.25–27 There is ever increasing evidence on the use of cyclosporine in SJS/TEN, suggesting benefits such as reduced mortality, shorter time required for re-epithelialization, and shorter length of hospital stay.27–29 In a systematic review published in 2021, which included 67 studies and 2,079 patients, the combined use of corticosteroids and IVIg showed a possible reduction in mortality in patients with the SJS/TEN spectrum.30

The frequency of cutaneous drug reactions among the evaluated patients was 13.6%, with an average of 6 days between the onset of the condition and the request for consultation by the assistant team.

Antibiotics and anticonvulsants were the main classes of drugs involved, and the most frequent clinical form was exanthema. One-third of the cases were considered severe (53.3% DRESS), and there were ten deaths related to the skin condition, which reinforces the role of Dermatology in the monitoring of cutaneous drug reactions in the hospital environment.

The vast majority of cases (82.6%) underwent histopathological analysis of the lesions, confirming the diagnosis and helping to differentiate among the different clinical forms.

Given the important role of the dermatologist in the diagnosis of cutaneous drug reactions, identification of the suspected drug, the differentiation of potentially severe forms, investigation of possible systemic involvement, and implementation of adequate treatment, the ideal situation would have the dermatologist involved as soon as possible in patient evaluation. However, the interval verified in this sample was of almost one week between the onset of the first symptoms and specialized evaluation.

None declared.

Júlia Chain Martins: Design and planning of the study; collection, analysis and interpretation of data; statistical analysis; drafting of the manuscript; collection, analysis and interpretation of data; critical review of the literature.

Camila Arai Seque: Analysis and interpretation of data; statistical analysis; drafting of the manuscript and critical review of the content; collection, analysis and interpretation of data; participation in research orientation; critical review of the literature; approval of the final version of the manuscript.

Adriana Maria Porro: Design and planning of the study; analysis and interpretation of data; statistical analysis; drafting of the manuscript and critical review of the content; collection, analysis and interpretation of data; participation in research orientation; critical review of the literature; approval of the final version of the manuscript.

None declared.