Patterns and Outcomes of Cutaneous Adverse Drug Reactions in a Hospital Based Study

Introduction : Drug reactions are one of the common reasons for admission in the Dermatological beds. Data on the culprit drugs and patterns of reaction are very scarce in Nepal. Objective: To determine the patterns of drug reactions admitted in Tribhuwan University (TU) Teaching hospital, the causes of drug reactions, duration of hospital stay, duration of steroid use and the outcome of the treatment. Material and Methods: This was a retrospective study done in TU Teaching Hospital. The admissions and discharge record of admitted patients from 15th April 2008 to 14th April 2012 were analyzed for the variables mentioned above. Results: There were 61 new patients of drug reactions; however only files of 42 patients could be studied.The mean age of the patients was 32.26 +/15.26 with male to female ratio of 1.8:1. Stevens Johnson’s syndrome (35.7%) was the commonest cause of admission followed by drug induced erythroderma (16.7%) and toxic epidermal necrolysis (11.9%). Carbamazepine, Phenytoin and allopurinol were the most common drugs for the reactions. The median time for the appearance of the lesions was 20.07 days. The median duration of hospital stay was 7 days. Steroids were used for initial treatment for around 4.9 +/-1.7 days. 83.3% of the patients improved while 11.9% succumbed despite optimal management. Conclusion: Severe drug reactions were the common reasons for admissions in our hospital. Anti-epileptics were the leading cause for the reactions. Steroids were used for initial period and rapidly tapered off in short duration. Overall, the outcomes of the treatment were good.


Introduction
Drugs can cure, suppress or prevent a disease and are usually benefi cial to humans.However, they can also produce undesirable/harmful eff ects, which are known as adverse drug reac ons (ADRs).Cutaneous adverse drug reac ons (CADRs) can be defi ned as noxious, unintended morphological skin changes with or without systemic involvement, developed a er local or systemic administra on of drugs in dosages commonly used for preven on, diagnosis or treatment of disease or modifi ca on of physiological func ons, in accordance with the World Health Organiza on's general defi ni on of adverse drug reac ons (ADRs). 1 Cutaneous ADRs can be caused by a wide variety of agents.The spectrum of Cutaneous ADRs ranges from a transient maculopapular rash to fatal reac ons like SJS, TEN and DRESS. 2,3They account for about 5 % of all hospital admissions in the UK. 4 The incidence of cutaneous ADRs among inpa ents in developed countries ranges from 1 -3% whereas in India it is 2-5% 5 , however, the data from our country is limited.
ADRs are reportedly responsible for up to 7% of hospital admissions, and cutaneous ADRs alone contribute to 2-3% of the overall hospital admissions. 6,7p to 30-45% of the ADRs are reportedly cutaneous in nature, 2% of which may be severe and few may even end in fatali es. 8,9Incidence of fatal cutaneous ADRs in the US is 0.32%. 10e most common CADR is morbilliform exanthema (ME), but it is also the least specifi c manifesta on.
Other poten ally severe forms are more specifi c, but they are less frequently encountered. 11e aim of this study is to report the various reac on pa erns of the severe cutaneous adverse drug reac ons and their puta ve drugs, as well as the outcome among the pa ents who were managed in a ter ary hospital of Nepal from 15 th April 2008 to 14 th April 2012.

Material and Methods
In this mono centric study, a retrospec ve analysis of all hospital records of severe cutaneous adverse drug reac ons admi ed in the dermatology ward of the TUTH from 15 th April 2008 to 14 th April 2012 was conducted.The age and sex of the pa ent, the off ending drug, the me interval between the drug intake and the erup on, dura on of hospital stay, outcome and sequel were recorded.Based on morphology, distribu on of the lesions, and clinical examina on of the CADR, toxic epidermal necrolysis (TEN) and Stevens Johnson syndrome (SJS); Erythroderma (ED); and drug hypersensi vity syndrome (DHS) were dis nguished.

Results
There were a total of 283 new admissions in the Dermatology ward, out of which 61 pa ents (21.6%) were having drug reac ons, in a four years period, but only 42 fi les could be retrieved.There were 27(64.3%)males and 15(35.7%)females.The mean age of the pa ents was 32.26+/-15.26(ranging from 11 to 75 years).The commonest age group to be aff ected by drug reac ons was 26-35 years.(Table 1).Twenty four (57.1%) pa ents had history of intake of single medica on and 18 (42.9%)had taken mul ple medica ons.Among the mul ple drugs, an bio cs with NSAIDs were common combina ons.
Maximum number of cases were diagnosed as SJS (15 cases) followed by 7 cases each of erythroderma and extensive exanthematous drug erup on.There were 5 cases of TEN, 3 cases each of drug hypersensi vity syndrome and SJS-TEN overlap; 1 case each of FDE and EM (Table 2).The median me for appearance of lesions a er medica on was 20.07 days.(rangingfrom 2 days to 90 days) (Table 3).

Discussion
Adverse drug reac ons may aff ect any organ, and the skin is the commonly involved organ.CADRs are usually benign.Severe forms are rare, with an es mated propor on of 2%. 12,13The course can be fatal in 0.2 to 29.3% of cases requiring hospitaliza on when a severe evolu on is predicted. 12The recogni on of the dermatological reac on pa erns and the exclusion of diff eren als are of primary importance for which history and clinical examina on form the cornerstone of diagnosis. 10These various reac on pa erns have diff erent temporal rela onships between the me of administra on of the medica on and the onset of dermatoses, although this latency may be shortened in the event of a re-exposure.Retrospec ve studies of drug erup ons have been reported by many countries and regions but retrospec ve study of inpa ents with CADRs in Nepal has not previously been reported, to our knowledge.
Comparing the incidence of CADRs by gender, we found our inference to be confl ic ng with some studies repor ng a female preponderance 14,15,16 and was in line with other studies with male preponderance. 3,17,18This study showed male preponderance with male female ra o of 1.8.Our study revealed 54.8 % cases of CADRs being in age group of 15 to 35 years and similar age groups were aff ected by CADRs in few other studies. 3,14,17,18Literatures suggested maculopapular erup ons as the most frequent reac on pa ern 3,15,19,20 while according to our study SJS (35.7%) was the most common CADR, while maculopapular erup ons and erythroderma were second highest reac ons with 16.7% each.The reason behind this may be the management of less severe exanthematous drug reac ons in outpa ent basis.In our study the latency was 2 -90 days with an average of 20.07days, this was similar to the latency of 3 to 60 days in Tunisia. 15However, one study revealed the latency for SJS/TEN to be few hours to one week 3 where as in another study 71.3% of CADR cases had a latency of less than a week. 17It was 2 -14 days in 80% cases in a North Indian study. 21Almost 36% cases had latency of less than a week in our study as well, this reduc on in latency may be due to repeated exposure to medica ons.Common drugs associated with CADR were an convulsants namely carbamazepine (n=8) and phenytoin (n=7) followed by diff erent an bio cs and unknown medica ons.This result was similar to a Tunisian study but in contrast to many other studies where an bio cs were the common culprits followed by an convulsants. 5,10,12,14,18,20However unknown medica on made second largest group in a study 17 and it was 14.4%(n=6) in our study.Carbamazepine accounted for 24% SJS/TEN in a study from Malaysia. 19ecently, many studies on allele associa ons with cutaneous ADR induced by aroma c amine an convulsants have been reported in Asian and European popula ons.Current studies indicate that HLAB*1502 is a marker for carbamazepine-induced SJS⁄TEN in Southeast Asian popula ons, where the prevalence of HLA-B*1502 is rela vely high. 22he role of cor costeroids in the treatment of SJS/TEN is controversial.Earlier studies suggested increased morbidity and mortality. 23,24A retrospec ve analysis of 289 pa ents from the Euro SCAR study found no benefi t from cor costeroids or IVIg compared to suppor ve care alone. 25There are, however, few studies sugges ng benefi ts in administering high-dose cor costeroids early in the course of disease. 26,27,28ossible explana on for this contradictory evidence may be due to the unfavourable outcomes from the use of cor costeroids resul ng from inadequate doses, a delay in the ini a on of treatment and the increased risk of mortality from sepsis from prolonged courses of therapy.However, in our study, the use of systemic cor costeroids with rapid tapering has given favorable results with median dura on of hospital stay being 7 days, (minimum of 3 days to a maximum of 51 days) which was similar to a Chinese study 16 but slightly higher than that of Singapore 10 where it was 2-24 days (average 9days).Despite op mal management, there was 11.9% mortality; cases who succumbed were 3 cases of TEN( 2 caused by unknown medica on and 1 caused by allopurinol) 1 case of erythroderma caused by pheny on, 1 case of DRESS caused by lefl unomide.
In general, the mortality rate in Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) is high; approximately 5% for SJS and 30% for TEN. 29n one of the studies, 4 out of 7 cases of TEN expired, which comprised 40% of all ADR related mortali es. 20n another study, 5 out of 12 cases of TEN and 2 out of 3 cases of erythroderma succumbed. 15Overall mortality of CADR in another study was 5.1%, which consisted of 2 mortali es of SJS/TEN and 3 of TEN. 10 Mortality rate was 15.6% among all cases; 9% in SJS and 26.7% in TEN in another study. 30

Conclusion
Most of the results obtained in our study are comparable to what are reported in the literatures.The role of cor costeroids is currently under revision.Some earlier studies have shown their lack of effi cacy or increased mortality in their use but the use of high doses early in the course of the disease may actually reduce morbidity and mortality, which has been shown in our study too.
A retrospec ve study has several limita ons.Our data may not include all the pa ents with CADRs because some of them may not have been admi ed to our hospital, in addi on all data could not be retrieved due to poor record keeping.

Table 1 :
Age group of pa ents admi ed with CADRs

Table 2 :
Types of CADR

Table : 3
Latency period referred to other speciality for management of comorbidi es and one case le against medical advice.Cases who succumbed were 3 cases of TEN (2 caused by unknown medica on and 1 caused by allopurinol), 1 case of erythroderma caused by pheny on and 1 case of DRESS caused by lefl unomide.

Table 4 :
Various puta ve drugs and their associated clinical reac on pa ern