Endgames Picture Quiz

Rash and fever after sulfasalazine use

BMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g5655 (Published 08 October 2014) Cite this as: BMJ 2014;349:g5655
  1. Nisha Raithatha, foundation year 1 doctor, rheumatology,
  2. Sarah Mehrtens, core medical trainee,
  3. Maria Mouyis, specialist registrar, rheumatology,
  4. Jessica Manson, consultant rheumatologist
  1. 1Department of Rheumatology, University College Hospital, London NW1 2BU, UK
  1. Correspondence to: N Raithatha nisharaithatha{at}gmail.com

A 28 year old Afro-Caribbean woman presented to our hospital with a three day history of an acute, diffuse, urticarial, papular rash on her trunk, limbs, and face. She also had facial swelling and fever. The mucosa and conjunctiva were not affected. Her medical history was unremarkable until a recent diagnosis of seronegative inflammatory arthritis during her second trimester of pregnancy, for which she was initially treated with corticosteroids and hydroxychloroquine. After a miscarriage at 30 weeks, this was switched to sulfasalazine for the three weeks before this admission.

On examination she had a fever (38.5°C), tachycardia, bilateral cervical and groin lymphadenopathy, generalised facial swelling (fig 1), and a diffuse erythrodermic rash covering her trunk, face, and limbs (fig 2).

A full blood count showed haemoglobin 120 g/L (reference range 115-165), white cell count 5.4×109/L (4.5-11), neutrophils 3.48×109/L (2-7.5), lymphocytes 3.26×109/L (1.3-4), eosinophils 0.16×109/L (0-0.45), and platelets 150×109/L (150-440). A blood film showed a few atypical lymphocytes and reactive lymphocytosis.

Liver function tests showed bilirubin 9 mmol/L (5-17), alanine transaminase 415 IU/L (10-35), alkaline phosphatase 145 IU/L (35-104). Her previous liver function test results had been within the reference range. C reactive protein was 32 mg/L (0-5); a viral screen was negative; and an immunology screen showed that antinuclear antibodies, immunoglobulins, and complement were all within the normal range. A microbiology screen showed that urine and blood cultures were negative

A chest radiograph was clear. Skin biopsy showed leucocytosis and early vasculitis.


  • 1. What are the important differential diagnoses in patients admitted to hospital with acute rash after drug exposure?

  • 2. What is this patient’s diagnosis and why?

  • 3. How would this patient’s condition be managed and what is the prognosis?


1. What are the important differential diagnoses in patients admitted to hospital with acute rash after drug exposure?

Short answer

Toxic epidermal necrolysis, Stevens-Johnson syndrome, drug rash with eosinophilia and systemic symptoms (DRESS), acute generalised exanthematous pustulosis, and drug induced erythroderma are important differential diagnoses in patients with acute rash after drug exposure. It is also important to rule out non-drug induced causes, such as viral infections, vasculitis, burns, severe eczema or psoriasis, Still’s disease, Kawasaki disease, and staphylococcus scalded skin syndrome.

Long answer

Adverse cutaneous drug reactions are common, affecting 2-3% of all patients admitted to hospital.1 A drug induced reaction should be considered in any patient who suddenly develops a symmetrical cutaneous eruption on the background of longstanding or new drug treatment. Drug reactions cannot be predicted and they can be fatal, so rapid diagnosis will facilitate appropriate treatment, which can be lifesaving.

The differential diagnosis of an acute cutaneous drug reaction includes toxic epidermal necrolysis, Stevens-Johnson syndrome, drug rash with eosinophilia and systemic symptoms (DRESS), and acute generalised exanthematous pustulosis. The table outlines these four conditions for ease of differentiation.1 The diagnosis of the correct cutaneous reaction depends on a detailed and reliable history (including duration of drug exposure and interval between the introduction of the drug and the onset of the eruption), clinical examination of the rash, investigations to assess organ involvement, and a skin biopsy. Skin testing is available to help identify a suspected drug if needed but is not routine practice.2

Differential diagnosis of acute cutaneous drug reactions

View this table:

As shown in the table, toxic epidermal necrolysis and DRESS typically have different durations of onset after exposure to the precipitating drug, and steroid treatment is needed for longer in DRESS. Facial oedema, lymph node involvement, multiorgan involvement, lymphocytosis, and eosinophilia are key features of DRESS, whereas lymphocytopenia and neutropenia are commonly seen in Stevens-Johnson syndrome and toxic epidermal necrolysis.1 Acute generalised exanthematous pustulosis is another differential diagnosis but is usually recognised by the presence of multiple non-follicular pustules. Drug induced erythroderma—a condition characterised by generalised erythroderma, which is often associated with pruritus and oedema but does not have the extensive systemic and haematological abnormalities found in DRESS—is also a possibility.

It is also vital to exclude non-drug related causes of a cutaneous reaction in any patient with a presumed drug reaction. Such causes include acute viral infections, especially human herpesvirus 6 and Epstein Barr virus (classically a maculopapular eruption), burns, vasculitis, and erythroderma secondary to a pre-existing skin condition (such as eczema or psoriasis). Other conditions to exclude are Still’s disease, which often has a classic salmon pink rash; Kawasaki disease, which causes conjunctivitis, strawberry tongue, palmar erythema, fissured lips, and cardiovascular abnormalities; and staphylococcus scalded skin syndrome, which is usually seen in infants. Failure to differentiate these from a drug related cutaneous reaction would have serious consequences because the treatment is very different.

2. What is this patient’s diagnosis and why?

Short answer

This patient has DRESS. Patients with DRESS have a drug induced rash (classically three weeks to three months after starting treatment with the drug responsible), systemic organ involvement, and haematological abnormalities (eosinophilia or atypical lymphocytes, or both).

Long answer

The diagnosis is DRESS, also known as drug induced hypersensitivity syndrome. DRESS typically develops three weeks to three months after starting treatment with the precipitating drug. It is characterised by a high fever (>38°C), a cutaneous eruption, lymphadenopathy, and single or multiorgan involvement.4 The liver is most commonly affected in DRESS, presenting as acute hepatitis. Liver failure is the leading cause of death.5 Retrospective studies show a mortality rate of 5-10%, so prompt diagnosis and management are needed.6

The following three criteria are needed to make the diagnosis of DRESS5:

  • Drug induced rash

  • Haematological abnormalities (eosinophilia (>0.45×109/L) or presence of atypical lymphocytes, or both)

  • Systemic involvement: lymphadenopathy, hepatitis (transaminase at least twice the normal value), interstitial nephritis, pneumonitis, or carditis.

Contrary to the name of the syndrome, many patients with DRESS do not have eosinophilia. Eosinophilia is usually seen one to two weeks after the onset of the syndrome or may even occur when liver enzyme values have returned to baseline.7 Facial oedema is also a hallmark of DRESS, occurring in 25% of patients (fig 2).5

The incidence of DRESS ranges from one in 1000 to one in 10 000 drug exposures,4 and people of Afro-Caribbean origin have an increased risk of developing the syndrome. The exact pathogenesis of DRESS remains unknown but there are likely to be multiple causes. Links have been found between reactivation of herpes family viruses (HHV-6, HHV-7), cytomegalovirus, and Epstein Barr virus. This may occur by the drug stimulating mass T cell production, which reactivates the herpes viral genome in the cells. Alternatively, crossreactivity between the virus and the drug may lead to an expansion of specific T cells.4 Studies have shown that HHV-6 reactivation is common in the second or third week after the onset of rash.8 Other studies have suggested a metabolic response, whereby deficiency of the epoxide hydroxylase enzyme, which is involved in the detoxification of anticonvulsants, leads to accumulation of arene oxides, which cause cellular toxicity or an immune response.4 There is also evidence of a genetic predisposition to DRESS—for example, HLA-B15:02 has been implicated in carbamazepine induced DRESS and HLA-B58:01 in allopurinol induced DRESS.9

The aromatic anticonvulsants (phenytoin, phenobarbital, carbamazepine) and sulphonamides (sulfasalazine, co-trimoxazole) are the most common drugs that cause DRESS, but many other drugs have been associated with the syndrome. These include antidepressants, allopurinol, non-steroidal anti-inflammatory drugs, angiotensin converting enzyme inhibitors (such as captopril), calcium channel blockers, antidepressants (such as fluoxetine), antibiotics (metronidazole, minocycline), antifungal agents (terbinafine), and antiretroviral drugs (such as abacavir).10

Full blood count, liver function tests, urine analysis, and a blood film are key tests in the diagnosis of DRESS. Skin biopsy is usually carried out but is not specific for DRESS—dermal eosinophils are detected only rarely—and usual features include lymphocytosis, extravasation of erythrocytes, and focal interface changes, which can be interpreted generically as a drug induced dermatitis.11 Skin biopsy is useful for excluding similar presentations that may require different management (table).

3. How would this patient’s condition be managed and what is the prognosis?

Short answer

Management involves stopping the drug, use of oral corticosteroids, and symptom control with antipyretics, topical corticosteroid, emollients, dressings, and antihistamines. The mortality rate is 5-10% and hypothyroidism is a common complication.

Long answer

DRESS is considered a dermatological emergency, and patients should be admitted promptly to hospital for management. The most important step in the management of DRESS is to stop the precipitating drug, and early withdrawal of the drug may decrease mortality.12 Steps should also be taken to ensure that the patient never uses that drug again. Other drugs in the same class should be used with care because the patient may have a genetic predisposition to hypersensitivity to that class of drugs. The patient may also be hypersensitive to other drug classes that commonly cause DRESS. After withdrawal of the drug, supportive therapy is often needed—for example, fluid replacement, antipyretics for fever, antihistamines, topical emollient creams, dressings, and topical steroids. The empirical use of antibiotics should be avoided because of unexplained crossreactivity between the drugs.13

Systemic corticosteroids are often used to treat DRESS. They have been shown to be effective in many case studies, and one such study showed recurrence of phenytoin hypersensitivity syndrome on withdrawal of corticosteroids and remission on reintroduction of corticosteroids.14 These drugs have also shown to be effective in patients with multiorgan involvement.15 Patients who do not respond to steroids should be treated with immunosuppressants or plasmapheresis (or both). The use of steroids or immunosuppressants remains controversial because of the risk of secondary skin infections, reactivation of viruses (such as cytomegalovirus), and long lasting corticodependent DRESS.10

In 2010 the French Society of Dermatology published the results of a consensus of experts on the therapeutic management of DRESS.16 The box summarises the recommendations.

Recommendations on the management of drug rash with eosinophilia and systemic symptoms (DRESS)

  • In the absence of signs of severity: Topical glucocorticoids, emollients, and H1 antihistamines

  • In the presence of signs of severity (transaminases more than five times normal, renal abnormalities, pneumonia, haemophagocytosis, or cardiac symptoms): Glucocorticoids equivalent to 1 mg per kg per day of prednisone and multidisciplinary evaluation

  • In the presence of life threatening signs (haemophagocytosis with bone marrow failure, encephalitis, severe hepatitis, renal failure, or respiratory failure): Steroids with intravenous immunoglobulin at a dose of 2 g per kg over five days. Intravenous immunoglobulin should not be used without associated steroids

  • In the presence of signs of severity with confirmation of major viral reactivation: Combination of steroids and antiviral agents (ganciclovir) or intravenous immunoglobulin, or both

Recent studies have looked at the role of N-acetylcysteine in the management of DRESS. N-acetylcysteine could potentially neutralise drug derived reactive metabolites and replenish glutathione stores that counteract oxidative stress.17

The prognosis of DRESS depends on the patient’s age and pre-existing health status. Mortality is 5-10% and the most common cause of death is liver failure. Hypothyroidism is a late complication,18 and thyroid function tests are recommended three months after the acute presentation. The duration of symptoms correlate positively with leucocyte, lymphocyte, and eosinophil counts.19 It has been suggested that serum creatinine and ferritin are prognostic indicators of DRESS because concentrations of these markers are higher in patients who die than in those who survive.19

Patient outcome

Our patient was diagnosed with DRESS secondary to sulfasalazine use. She was treated with three doses of intravenous methylprednisolone (250 mg, 500 mg, and 500 mg), followed by a weaning oral regimen, starting at 40 mg of oral prednisolone. Supportive management included intravenous fluids, oral chlorphenamine, cetrizine, hydroxyzine, paracetamol, topical emollients, and topical corticosteroids. Her symptoms improved and she was discharged after five days to be closely monitored in clinic. At outpatient follow-up two weeks later she was found to be healthy, with no evidence of inflammatory arthritis, very subtle post-inflammatory skin changes, and normal liver function tests. Further investigations could include a study of her risk for developing DRESS, with the use of genetic studies and family history.

This case highlights the need to always review the drug history in patients with cutaneous skin reactions. Many common drugs can cause diffuse cutaneous skin reactions such as DRESS, and it is difficult to predict or prevent these reactions. A high index of clinical suspicion is therefore needed and the precipitating drug should be stopped promptly.


Cite this as: BMJ 2014;349:g5655


  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare that we have no competing interests.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

  • Patient consent obtained.


View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial