Table of Content


Autumn 2015, Vol. 23 No. 3

Hong Kong J. Dermatol. Venereol. (2015) 23, 105-112

Original Article

Cutaneous adverse drug reactions in a tertiary hospital in South-east Asia


JYY Pang 彭亦瑜, SM Pang 彭紹明, HY Koh 許弘毅


The clinical spectrum of cutaneous adverse drug reactions range from benign, self-limiting reactions to life-threatening conditions. Patients with severe cutaneous adverse drug reactions require longer hospitalisation that cost both the hospital and health care system. Cutaneous adverse drug reactions are common complications of medical therapy. While most are benign and self-limiting after withdrawal of the offending drugs, serious cutaneous adverse drug reactions may cause significant morbidity and mortality. Physicians should be trained to recognise and manage cutaneous adverse drug reactions.


Keywords: Allergy, drug hypersensitivity syndrome, drug reaction, Steven-Johnson syndrome, toxic epidermal necrolysis

關鍵詞: 過敏、藥物過敏症候群、藥物反應、史蒂芬強生症候群、毒性表皮壞死溶解症


Severe cutaneous adverse drug reactions form the minority of all cutaneous adverse drug reactions (36.1%), but cause significantly more morbidity and mortality.1-3 In the United Kingdom (UK), the incidence of fatal adverse drug reaction is 6.5% with significant burden on hospital resources.1,4-6 An adverse drug reaction (ADR) is defined by World Health Organisation as a "noxious, unintended or undesired effect of a drug used for prophylactic, diagnostic or therapeutic use."7 Most ADR have cutaneous manifestations, which we will refer to as cutaneous adverse drug reactions (CADR). Common causative medications include antibiotics, anticonvulsants, allopurinol, complementary alternate medicine (CAM) and anti-retroviral treatments.

All physicians should have some knowledge of ADR, as it is inevitable that they would encounter ADR in some of the patients under their care. Although early recognition of CADR and interruption of causative agent can improve patient outcome and reduce the potential of development of SCAR, CADR remains under-diagnosed. CADR are not well reported in literature because few studies examine aspects of clinical presentations and outcomes of CADR in hospitalised patients. With the introduction of new medications and evolving prescription practices, patterns of CADR may vary through time.8,9

The current study aims to study the spectrum of CADR among patients who were managed in a tertiary hospital in Singapore between 1st January and 30th June 2010, and to compare our results to a similar study based on results obtained between 2005 and 2006.

Abbreviation used
ADR: Adverse drug reaction
AGEP: Acute generalised exanthematous pustulosis
CADR: Cutaneous adverse drug reaction
CAM: Complementary alternate medicine
DRESS: Drug rash with eosinophilia and systemic symptoms
FDE: Fixed drug eruption
GED: Generalised exfoliative dermatitis
SCAR: Severe cutaneous adverse drug reactions
SJS: Stevens-Johnson syndrome
TEN: Toxic epidermal necrolysis


Approval was obtained from the hospital's Institutional Review Board. Written informed consent was not required for the retrospective analysis. We reviewed records of hospitalised patients who were referred to the Dermatology Department at Singapore General Hospital from January 2010 to June 2010, and identified cases who were diagnosed with CADR.

Patient data was analysed for the following variables: age, race, gender, referring specialty, reason for admission, suspected drug causing CADR, outcome, length of stay and clinical reaction pattern.

The inclusion criteria include clinical diagnosis of CADR. Exanthems with uncertain or viral causes were excluded. In addition, to ensure reliability of data, a definite ADR was under final diagnosis in each patient's case file. The diagnosis of CADR was based on clinical features and reasonable temporal relationship between drug administration and the onset of symptoms. Drug causality was derived from case notes. The following were classified as SCAR: Stevens-Johnson syndrome (SJS), SJS/Toxic epidermal necrolysis (TEN) overlap, TEN, acute generalised exanthematous pustulosis (AGEP), generalised exfoliative dermatitis (GED) and drug rash with eosinophilia and systemic symptoms (DRESS).


A total of 692 inpatient dermatology referrals were reviewed during the study period. Seventy-six patients were diagnosed with CADR. Patients ranged from 14-92 years, with a mean age of 40 years. The racial distribution consisted of Chinese (81.5%), Malays (9.0%), Indians (7.9%) and other races (1.3%). Of the 76 patients, 89.5% were admitted into medical disciplines and 10.5% were admitted into surgical disciplines.

A detailed breakdown of the source of patient referrals is shown in Figures 1a & b. The highest number of referrals came from Internal Medicine, Haematology and Renal departments.

Figure 1 Referrals from (a) medical disciplines; (b) surgical disciplines.

Eleven different clinical reaction patterns were noted (Figure 2). These included: exanthems (68.3%), urticaria (14.5%), FDE (5.3%), AGEP (4.0%), fixed drug eruption, DRESS (4.0%), SJS (4.0%), TEN (1.3%), SJS/ TEN overlap (1.3%), GED (1.0%) and drug induced vasculitis (1.3%). The three most common presentations in this study were: drug exanthem, urticaria and FDE. Drug exanthem accounted for close to half of all CADR in the study while urticaria accounted for 14.5%. SCAR accounted for 11 patients making up 15.0% of all CADR.

Figure 2 Different patterns of drug reactions.

In this study, we found the three most common causes of CADR were antibiotics, NSAIDs and anticonvulsants (Table 1). Antibiotics accounted for close to half of the patients with CADR/SCAR. NSAIDs accounted for a little over 10.0% of patients in this study and anticonvulsants accounted for another 10.0% of total patients studied. ACE-inhibitors and allopurinol each accounted for 2.0%. Other less known causes such as antidepressants, CAM and proton pump inhibitors accounted for one fifth of the cases.

Table 1 Various putative drugs and their associated clinical reaction pattern
FDE: fixed drug eruption, SJS: Stevens-Johnson syndrome, DRESS: drug rash with eosinophilia and systemic symptoms, AGEP: acute generalised exanthematous pustulosis, TEN: toxic epidermal necrolysis, GED: generalised exfolative dermatitis, NSAIDs: nonsteroidal anti-inflammatory drugs, CAM: contemporary alternate medicine

Over the six-month period, ADR accounted for 11.0% of all inpatient dermatology consultations. The highest percentage of ADR diagnosed amongst patients referred to the dermatology department was in the months of April and June (17.0% and 11.0% respectively). Detailed breakdown of monthly inpatient dermatology consultations is shown in Figure 3.

Figure 3 Incidence of dermatology consultations over a six-month period.

Of the 76 patients diagnosed with ADR, 23 (30.0%) were admitted primarily for their cutaneous adverse drug reactions and required seven days of hospitalisation on average. All 23 patients were discharged after hospitalisation. The remaining 70.0% developed cutaneous ADR as a complication of their hospitalisation. A total of four deaths were reported in our study, giving a mortality rate of 5.3% amongst all cases of CADR. Only one death was directly related to ADR - a 61-year-old male developed SJS from phenytoin. Of the remaining three patients, one died from malaria, another from end stage real failure and the last patient died of pneumonia.

Patients with steroid-induced acneiform eruption and toxic erythema of chemotherapy (TEC) were not included in this study as they were considered side effects of drugs and not classified as ADR. There were seven cases of steroid-induced acneiform eruption and ten cases of TEC. In addition, patients for whom the exact cause of the rash is uncertain, such as "TEC versus drug exanthem" were excluded from the current study.


CADR is a common disease, with an incidence of 11.0% of all dermatological referrals in the current study. The incidence of CADR has remained unchanged since 1998, with no significant increase or decrease in the incidence of CADR patients.1,8 In a similar study done five years ago, the incidence of CADR was 13%.1 There is a decrease in the incidence of 2.0%. Several factors could explain this new finding. Firstly, toxic exanthems induced by chemotherapy (TEC) were no longer considered CADR as these are toxic reactions and not hypersensitivity reactions. Seven cases of exanthem related to chemotherapeutic agents were reported in a previous study done in the same institution but no such cases were reported in our study.1 Labelling chemotherapy exanthem patients as CADR would restrict patients from receiving chemotherapeutic treatment in the future. Five years ago, the term TEC was not introduced yet as Bolognia et al only coined this term in 2008.10 In our study we excluded six cases of TEC and four cases of TEC versus drug exanthem. Of the six cases of TEC, four cases exhibited exanthematous rash which clinically mimicked drug exanthem. The remaining two cases presented with palmar plantar syndrome, with rashes on the hands and feet. In the current study, it is possible that some reported cases of drug exanthems were actually TEC. More research is needed on drug exanthems and TEC in terms of histology, latency period from the intake of drug to the onset of symptoms, clinical manifestations and outcome to tell them apart.

The three most common causative drugs in our study were antibiotics, anticonvulsants and NSAIDs. In comparison, the most commonly implicated drugs in an earlier study based on 2005-2006 data were antibiotics, anticonvulsants and allopurinol.1 This may reflect a change in prescriptive behaviour of physicians over time.

Another finding is the lower incidence and fatality of ADR secondary to allopurinol in our study compared to its predecessor.1,11-13 Our study reported no fatality from allopurinol while Lee et al reported a mortality rate of 18.0% mainly due to allopurinol induced hypersensitivity syndrome.14 A possible explanation for this finding may be a heightened awareness among physicians of allopurinol's risk of inducing adverse reactions and a change in the management of patients with hyperuricaemia. As a result, prudence and restrictive use of allopurinol followed, accounting for a decline in the incidence and mortality of allopurinol hypersensitive reactions. The exact mechanism of how allopurinol causes hypersensitivity reactions remains unclear; more research on histology and immunology should be investigated.14,15

Valproate, macrolides, vancomycin, metronidazole, dapsone and anti-tuberculosis drugs were reported in Lee HY et al's study but the six drugs were not found in this study.

The influence of local culture on medical practice is reflected in the number of cases of ADR arising from CAM.16 In our study, CAM accounted for one case of urticaria and one case of vasculitis. Both were Chinese patients who required a mean hospitalisation of six days. Although both presentations were mild and non-life threatening, another study had highlighted that CAM might cause SCAR such as TEN.16 CAM should not be overlooked as a possible offending agent in cases of ADR and the history of any recent ingestion of CAM should be actively elicited.

In our study, 15.0% of cases were SCAR with a mortality rate of 6.7%. This appeared significantly lower compared with previous reports of an incidence of 36.0% and a mortality rate of 5.1%.1 There can be a few possible explanations. First of all, there is a decrease in the incidence of allopurinol as a cause of ADR. Allopurinol was previously reported to cause up to one-third of cases with DRESS.11 A reduction of allopurinol use resulted in a lower incidence of SCAR and hence reduced mortality.

Secondly, NSAIDs surpassed allopurinol as the third most common implicated drug in ADR in our study compared with the earlier study. NSAIDs are associated with milder forms of CADR such as drug exanthem and urticaria and this may have contributed to the reduced incidence of SCAR.

A detailed history and clinical examination form the backbone of diagnosis. Identifying the causal agent is important for both diagnosis and patient management. An understanding of temporal relationship between the intake of medication and the clinical reaction patterns would aid the physician in recognition, diagnosis and management of patients with ADR. The genetic predisposition is an area of research that is important in the pharmacogenetic of cutaneous drug reactions. One fatal recognisable SCAR is TEN (Figure 4), which is characterised by a spiking fever followed by a dusky appearance of the skin, that is painful to touch. TEN is the SCAR with the highest mortality rate. One predisposing factor for TEN is HLA-B 1502, which is more prevalent in Han Chinese compared with other races.17-19 Physicians should be aware of this genetic predisposition to TEN in their patient population. Prompt institution of appropriate treatments such as high dose intravenous immunoglobulin is important in improving survival of patients with TEN.20,21

Figure 4 Toxic epidermal necrolysis of patient.

In our current study, patients admitted primarily for CADR were hospitalised for seven days on average. This is less than a previous study with a mean hospitalisation of nine days.1 Evolving medical practices in recent years such as early discharge of patients could have led to shortened hospitalisation. Patients admitted with CADR had a median stay of four days while a median of eight days was reported in a similar study in UK.22 Improvement in the management of CADR patients may account for an evident drop in hospitalisation by reducing unnecessary ones. A reduction in the duration of hospitalisation decreases the burden on our healthcare services. In the UK, NHS spent £66 million annually on ADR patients.4-6,22

Singapore's racial composition in 2010, based on the information from the Singapore government's official Department of Statistics were Chinese 74.0%, Malay 13.0%, Indian 9.0% and other races 3.0%.23 In our study, we saw a similar distribution in our patients except for a slightly lower proportion of Indian, Malay and others. However, we do not have information on the racial composition of patients admitted to the Singapore General Hospital to be able to comment if there is any racial predisposition to ADR.

There were some limitations in our study. One was the short study period of six months, which may lead to reduced accuracy in our analysis. However, it should be noted that though our predecessors reviewed 12 months of data, the number of cases of ADR in the two studies are comparable. We had 76 patients in half a year compared to 96 patients over one year in 2005. Hence we believe our results are still representative of the types and distribution of ADR in our study population.

Our study is conducted in a single institution in one country in South-east Asia. Results from such a study may not be generalisable to the rest of the world. For example, Chinese formed the majority of the patients in our study, and genetic differences between Chinese and other races (e.g. prevalence of HLA-B 1502) may give rise to differences in the incidence and spectrum of ADR. Differences in management practices across institutions and countries may influence the length of hospitalisation for the same medical condition. In Singapore, CADR remains a predominantly a clinical diagnosis based on history and examination. In contrast, a Swiss study relied heavily on both clinical and histological examination.3


This paper hopes to raise the awareness of life-threatening cutaneous adverse reactions that physicians may encounter in their practice. Continuing research in this area is important because of high mortality associated with SCAR. Even with current medical advances, CADR and SCAR remain essentially a clinical diagnosis. Identification of the causative agent remains to be the most challenging and important aspect of ADR. Early recognition of reactions and interruption of causative agent could improve patient outcome (e.g. reduce length of stay and health-care costs). Anticonvulsants and antibiotics remain the main culprits of ADR while CAM and allopurinol must not be overlooked. Physicians should have the necessary training and experience in dermatology and develop an appreciation of common ADR and their causative drugs. This will help to reduce the incidence of ADR arising as a complication in hospitalised patients.


1. Singh G. Do no harm-but first we need to know more: the case of adverse drug reactions with antiepileptic drugs. Neurol India 2011;59:53-8.

2. Teoh PC. Hospital monitoring of adverse drug reactions in Singapore. Singapore Med J 1974;15:268-72.

3. Harr T, French LE. Toxic epidermal necrolysis and Stevens-Johnson syndrome. Orphanet J Rare Dis 2010;5:39.

4. Moore N, Lecointre D, Noblet C, Mabille M. Frequency and cost of serious adverse drug reactions in a department of general medicine. Br J Clin Pharmacol 1998;45:301-8.

5. Ghose K. Hospital bed occupancy due to drug-related problems. J R Soc Med 1980;73:853-6.

6. Col N, Fanale JE, Kronholm P. The role of medication noncompliance and adverse drug reactions in hospitalizations of the elderly. Arch Intern Med 1990;150:841-5.

7. Lee HY, Tay LK, Thirumoorthy T, Pang SM. Cutaneous adverse drug reactions in hospitalised patients. Singapore Med J 2010;51:767-74.

8. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998;279:1200-5.

9. Coopman SA, Johnson RA, Platt R, Stern RS. Cutaneous disease and drug reactions in HIV infection. N Engl J Med 1993;328:1670-4.

10. Bolognia JL, Cooper DL, Glusac EJ. Toxic erythema of chemotherapy: a useful clinical term. J Am Acad Dermatol 2008;59:524-9.

11. Fong PH, Chan HL. Current cutaneous drug reaction patterns in Singapore. Singapore Med J 1984;25:336-9.

12. Kidon MI, See Y. Adverse drug reactions in Singaporean children. Singapore Med J 2004;45:574-7.

13. Oh VM. Treatment of allergic adverse drug reactions. Singapore Med J 1989;30:290-3.

14. Lee HY, Ariyasinghe JT, Thirumoorthy T. Allopurinol hypersensitivity syndrome: a preventable severe cutaneous adverse reaction? Singapore Med J 2008;4:384-7.

15. Khoo BP, Leow YH. A review of inpatients with adverse drug reactions to allopurinol. Singapore Med J 2000;41:156-60.

16. Lim YL, Thirumoorthy T. Serious cutaneous adverse reactions to traditional Chinese medicines. Singapore Med J 2005;46:714-7.

17. Elzagallaai AA, Garcia-Bournissen F, Finkelstein Y, Bend JR, Rieder MJ, Koren G. Severe bullous hypersensitivity reactions after exposure to carbamazepine in a Han-Chinese child with a positive HLA-B*1502 and negative in vitro toxicity assays: evidence for different pathophysiological mechanisms. J Popul Ther Clin Pharmacol 2011;18:e1-9.

18. Ganesan S, Hussain N. Question 2 Should phenytoin and carbamazepine be avoided in Asian populations with the HLA-B*1502 positive genetic variant? Arch Dis Child 2011;96:104-6.

19. Goh TK, Pang SM, Thirumoorthy T, Goh SGN. Acute generalised exanthematous pustulosis and toxic epidermal necrolysis induced by carbamazepine. Singapore Med J 2008;49:507-10.

20. Teo L, Tay YK, Liu TT, Kwok C. Stevens-Johnson syndrome and toxic epidermal necrolysis: efficacy of intravenous immunoglobulin and a review of treatment options. Singapore Med J 2009;50:29-33.

21. Chan HL. Recognizing and preventing adverse drug reactions: with particular reference to drug eruptions. Singapore Med J 2000;41:519-20.

22. Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004;329:15-9.

23. Statistics Singapore - Publications [Internet]. [cited 2015 Jul 30]. Available from: