Introduction

Adverse reactions in the skin to medications are very common and are an important cause of iatrogenic illness. Drug rashes are usually self-limiting and resolve completely upon withdrawal of the culprit medication, but a small number (2%) can cause serious morbidity and mortality. This may not only have medicolegal and economic implications but may undermine the patient's confidence in the prescriber and affect future adherence. Diagnosis of drug-induced skin disease may be difficult for a number of reasons:

• Almost any drug can cause any rash.
• Unrelated drugs can cause similar reactions.
• The same medication can cause a different rash in different individuals.
• Patients may not volunteer information about medicines that they have taken, which they deem not to be relevant (over-the-counter [OTC] preparations and complementary medicines).

History

It is imperative to take a thorough history from patients in whom a drug reaction is suspected. Eliciting the temporal association of the ingestion of the drug and the onset of the eruption is key. Apart from noting any medications taken for the first time in the three months prior to the appearance of the rash, patients should be specifically asked about any recent changes to brand, dosing, or preparation of long-term medications. Patients may not volunteer information about drugs they have taken that they assume are not relevant, such as paracetamol taken for a headache, or an antihistamine taken for hayfever. Direct questioning about OTC preparations should always form part of a thorough drug history. Medications recommended by alternative/complementary health practitioners may not be revealed spontaneously and should be asked about directly. Both generic and brand names of all drugs should be recorded, and the patient should be asked about any history of sensitivity to medications. Knowledge of whether the patient has been previously exposed to suspected culprit drugs is also relevant.

Examination

The patient should be exposed fully to allow complete examination of the skin. The morphology of the rash should be described - for example, lichenoid, urticated, vasculitic, maculopapular, or bullous. The distribution of the rash should be noted: Is it widespread? Limited? Acral (hands and feet)? Photo-distributed? These features will help to classify the eruption and may give a pointer as to the causative drug. Special attention should be paid to the mucosal sites - eyes, mouth, genitalia - as involvement at these sites can indicate one of the severe cutaneous adverse reaction (SCAR) syndromes. Early diagnosis of these syndromes is crucial as patients may become unwell and deteriorate quickly. Careful examination of the appendageal structures such as hair, nails, and teeth should also be carried out as these can be affected by certain medications.

Investigation of a suspected drug reaction

In most cases, careful history and examination will provide all the necessary clues to make a confident diagnosis of a drug rash. A skin biopsy can be helpful to confirm the diagnosis, but the result of this is likely to be delayed following an acute presentation, and so action based on clinical assessment will usually precede this. Exclusion of differential diagnoses such as infection may require other investigations such as blood tests (white cell counts and CRP levels).

There are no consistently reproducible diagnostic tests which confirm specific drug allergy in the convalescent period; however, certain investigations such as measurement of specific immunoglobulin E (IgE), patch testing, intradermal testing, and in vitro tests such as lymphocyte transformation tests and cytokine release assays may be helpful. However, these investigations should be carried out by experts as their interpretation is highly specialised. Testing should also be carried out with great caution following an episode of one of the SCARs such as toxic epidermal necrolysis (TEN).

Patch testing

In this method of skin testing, an amount of the medication is applied to the skin in a suitable vehicle such as petrolatum. This is best performed at least six weeks post-resolution of the eruption. The best positive predictive value has been observed when testing for allergy to abacavir, anticonvulsants, and beta-lactam antibiotics. However, the sensitivity of patch testing alone is not great enough to conclusively exclude reactions.

Intradermal testing

Injection of the suspected culprit agent in to the dermis in the convalescent period is sometimes performed as an adjunct to patch testing. Using both modalities may improve sensitivity in confirming a causative agent.

In vitro testing

In experimental settings, a blood sample may be taken from a patient who is experiencing a drug reaction - the patient's lymphocytes are incubated with a series of suspected culprit drugs. Markers of activation of those lymphocytes are then measured, and this has been used as a testing modality in the investigation of drug reactions (lymphocyte transformation test). Measurement of cytokines such as IFN-, IL4, tumour necrosis factor (TNF), and granulysin released by patient lymphocytes in response to in vitro exposure to suspected culprit drugs have also been used to confirm causality (cytokine release assays). Blood samples are best taken as early as possible in the course of the illness. No standardised, commercially available in vitro testing is currently available in the UK.

Classification of drug reactions in the skin

Cutaneous reactions to medications are extremely varied. They may be classified in a number of ways. Pathogenetically, drug reactions in the skin may be classified as immune-mediated or non-immune-mediated. Immune-mediated rashes are the most common and include hypersensitivity reactions from types I to IV. Type I reactions (immediate reactions, usually mediated by IgE or drug-specific receptors bound to mast cells and other immune cell membranes) tend to manifest in the skin as urticaria or angio-oedema. Type II reactions (cytotoxic reactions) result in cutaneous purpura. Type III (immune complex mediated) reactions lead to cutaneous vasculitis. Type IV delayed hypersensitivity reactions are by far the most common group of drug rashes resulting in generalised exanthems, phototoxic rashes, and severe drug reactions such as TEN.

Non-immune-mediated rashes include accumulation of medications in the skin (causing pigment changes), instability of mast cells (causing histamine release), slow acetylators (metabolism of drugs affected) and photosensitivity reactions (increased susceptibility to ultraviolet [UV] light).

However, drug reactions in the skin may also be classified clinically, and this is the approach adopted here. Drug reactions in the skin are discussed under the following headings:

• Drugs which alter normal skin function.
• Drugs which exacerbate an existing dermatosis.
• Common drug-induced rashes - maculopapular exanthem, urticaria, vasculitis, lichenoid drug reaction, and fixed drug eruption.
• Severe drug-induced rashes - Stevens-Johnson syndrome (SJS) and TEN, acute generalised exanthematous pustulosis (AGEP) and drug reaction with eosinophilia and systemic symptoms (DRESS).

Drugs which alter normal skin function

Photosensitivity

Drugs may cause excessive sensitivity to light in two ways: phototoxic reactions and photoallergic reactions. Phototoxic reactions (Figure 7-1a, b) are the more common, and resemble sunburn and may blister. The reaction is confined to light-exposed sites and may be characterised by a sharp demarcation between covered and uncovered skin. The onset will typically be fast (within 5-15 hours of taking the drug and exposure to light) and recovery is quick on withdrawal of the medication. Photoallergic reactions are usually eczematous, but may be lichenoid, urticarial, purpuric, or bullous. The onset may be delayed by weeks or months following introduction of the medication, and similarly, recovery may be slow on withdrawal. Patients taking medication known to cause light sensitivity (amiodarone, tetracycline antibiotics, and retinoids) should be advised to avoid excess sun exposure, and to wear a broad-spectrum sunscreen year-round. Drugs causing photosensitivity are detailed in Table 7-1.

Pigmentation

Hyperpigmentation, hypopigmentation, and discolouration are all associated with certain drugs (Table 7-2). The pigmentary change may require light exposure to manifest. Common examples would include the development of melasma in female patients taking the oral contraceptive pill, or the facial blue-black pigmentation which may be caused by amiodarone. Tetracycline antibiotics may also cause a slate-grey pigmentation. Mechanisms for drug-induced pigmentation are unclear, but may involve deposition of the drug or its metabolite in the dermis, or enhanced melanin production.

Hair and nails

Excessive hair: Hypertrichosis is the growth of hair at sites which are not normally hair-bearing; hirsutism is excessive growth of hair in the male pattern of hair growth, especially in women. Both hormonal and non-hormonal treatments may bring about this effect; the most commonly implicated would include ciclosporin and phenytoin.

Hair loss: Loss of hair may be dramatic or insidious in onset, and if the latter, may not be immediately noticed by the patient. The temporal relationship between the onset of the hair loss and the introduction of the medication depends on the part of the hair cycle which the drug is interfering with. Cytotoxic agents interrupt the anagen ('growth') phase of the hair cycle, and so loss is rapid and complete; delayed, insidious hair loss generally results from interference with the telogen ('shedding') phase of the hair cycle. Drugs such as acitretin, statins, and anti-thyroid drugs may have this effect. Androgenic drugs promote shrinkage of the hair follicles and shortening of anagen, and so can cause hair loss. An example would be exogenously administered testosterone used to treat hypogonadism in male patients.

Nails

Nails may become discoloured with use of mepacrine or hydroxyurea. Onycholysis, which is separation of the nail plate from the nail bed, may be caused by cytotoxic agents.

Drugs which exacerbate pre-existing dermatoses

Medications may exacerbate skin conditions which the patient already has. The following summarises the most common associations:

Psoriasis - this is a common condition affecting approximately 2% of the population; however, some medications are known to worsen psoriasis. These are classically described as beta blockers, lithium, and antimalarial medications, though newer drugs such as ACE (angiotensin-converting enzyme) inhibitors can also worsen psoriasis. Non-prescribed drugs such as alcohol have a detrimental effect on psoriasis.

Eczema - statins and diuretics such as hydrochlorothiazide may worsen eczema.

Acne - some forms of the oral contraceptive pill, particularly progesterone-only pills, may worse acne. Corticosteroids, ciclosporin, and anti-epileptics such as phenytoin may also have the same effect.

Urticaria - non-steroidal anti-inflammatory drugs (NSAIDs) and opiate analgesics may worsen urticaria in a susceptible individual, by lowering the threshold for mast cell degranulation. ACE inhibitors and angiotensin receptor blockers may exacerbate angio-oedema. This is non-allergic urticaria/angio-oedema; allergic urticaria/angio-oedema is described in the next section.

Common drug-induced rashes

Drug-induced exanthems

The most common cutaneous reaction to a drug is an exanthem, meaning a widespread rash. Such rashes may be morbilliform (resembling measles) or maculopapular (consisting of a mixture of raised and flat areas) (Figure 7-3). The patient may be symptomatic with burning, itch, or discomfort arising from the skin. Onset is typically within 7-10 days of starting the drug, representing a delayed-type hypersensitivity. However, subsequent reactions on inadvertent re-exposure to a culprit drug may provoke a reaction in the skin more quickly, because of the presence of memory T-cells in the lymph nodes. The proportion of the body surface area (BSA) involved may vary, and in cases where it exceeds 90%, the patient may be described as erythrodermic. Following withdrawal of the culprit drug, application of a potent topical corticosteroid and emollient will help alleviate discomfort and itch, and hasten resolution of the eruption. Any drug may cause a drug-induced exanthem but antibiotics of any class, anti-hypertensive agents, and cholesterol-lowering drugs are among the most common precipitants.

Urticaria/angio-oedema

The appearance of raised, red itchy wheals in the skin (Figure 7-4) may occur alone or in combination with angio-oedema, which is head and neck soft-tissue swelling. The latter may be serious, and when it involves the soft tissue of the airway, may cause respiratory embarrassment. It may be non-allergic (described above) or allergic; in the latter, a reaction occurs between a drug or its metabolite and a specific mast cell-bound IgE. A drug may cause anaphylaxis, occurring rapidly after drug ingestion (type I drug hypersensitivity) or may be delayed by a number of days following exposure to the drug (type IV hypersensitivity).

Drug-induced lupus

Medication may produce an eruption indistinguishable from cutaneous lupus - in particular, the rash of subacute cutaneous lupus (SCLE) (Figure 7-5). The patient does not have any pre-existing autoimmune disease, and the condition remits on withdrawal of the culprit drug. The most common drugs to cause drug-induced lupus are listed in Table 7-1, but a recent study named terbinafine as the most common culprit. Antihistone antibodies are present in >95% of cases, but dsDNA is usually negative and complement levels are normal.

Drug-induced vasculitis

Medication may cause a purpuric eruption that is indistinguishable from vasculitis (Figure 7-6). The distribution is usually predominantly in the lower limbs. As viral and bacterial infections may also cause vasculitis, it is often difficult to ascribe causality to a drug, as in cases where an antibiotic is suspected, the patient may also have had a recent infectious episode. In practice, in the absence of overt clinical signs of infection, causality is best determined by withdrawing the suspected culprit drug; if this brings about resolution of the vasculitis, then this adds weight to the diagnosis of a drug-induced phenomenon. Drugs associated with vasculitic eruptions include antibiotics, anticonvulsants, and NSAIDs.

Drug eruptions

Lichenoid drug eruptions resemble idiopathic lichen planus, but may not be confined to the classic sites of predilection of the latter. They consist of purplish papules which may have a lace-like white change on their surface (Figure 7-7a, b). The sites of predilection are the forearms, the neck and inner thighs, but the eruptions can appear anywhere. Onset may be delayed by a number of months following introduction of the culprit medication, leading to difficulties in diagnosing the eruption as drug-induced. Resolution following drug withdrawal can be slow and take up to two months, and the post-inflammatory hyperpigmentation left behind may be dramatic. Table 7-1 illustrates the drugs which most commonly cause lichenoid eruptions.

Erythema nodosum

This is a tender, nodular eruption which classically appears on the anterior aspect of the legs. It is characterised histologically by septal panniculitis (inflammation in the subcutaneous fat). Although infective and inflammatory triggers are recognised (such as tuberculosis, Yersinia infections, rheumatoid arthritis, lupus, and inflammatory bowel disease), EN may also be a drug-induced phenomenon. Drugs which commonly cause this include the oral contraceptive pill, penicillin and sulfonamide antibiotics, and salicylates.

Fixed drug eruption

This is peculiar phenomenon whereby one or more inflammatory patches appear at the same cutaneous or mucosal site on each occasion that the patient ingests a culprit drug (Figure 7-8). The time frame for developing the lesion at the characteristic site can vary from 2 to 24 hours. Common sites include the torso, hands, feet, face, and genitalia. The patches resolve sometimes leaving post-inflammatory hyperpigmentation in the skin. Any drug can potentially cause a fixed drug eruption, but those more commonly associated are listed in Table 7-1.

Severe drug reactions in the skin

Stevens-Johnson Syndrome (SJS) and toxic epidermal necrolysis (TEN)

SJS and TEN are rare, life-threatening drug-induced hypersensitivity reactions in the skin and mucous membranes. This mucocutaneous disorder is characterised by widespread, painful areas of epidermal detachment, and erosions of the mucous membranes, including eyes, mouth, genitalia, and respiratory tract (Figure 7-9a, b). The appearance of the eruption may be preceded by a prodrome of fever, malaise, and coryzal symptoms, and skin pain is often the first cutaneous manifestation, prior to the appearance of the rash. The terms SJS and TEN represent points along a spectrum of severity, with SJS classically denoting 10% BSA detachment, TEN indicating >30% BSA involvement, and the term 'SJS-TEN overlap' being used to describe cases with between 10% and 30% loss. Mortality from TEN may be as high as 90% and is estimated using the SCORTEN tool Table 7-3 SCORTEN parameters.

In the patient in whom SJS/TEN is suspected, the priority is to stop the offending drug. Common culprits include anticonvulsants, allopurinol, human immunodeficiency virus (HIV) medications, and antibiotics, particularly sulfonamide antibiotics. Treatment in the acute phase consists of supportive care. The patient will require high-dependency care in an intensive care environment, with organ support as dictated by clinical state. In addition to expertise from dermatologists and intensive care physicians, specialist input from ophthalmology, oral medicine, urology, and gynaecology may be required for site-specific involvement. Skin loss and fragility demand specialist dermatology nursing care, with anti-shear handling, non-adherent dressings, and careful attention to antisepsis to prevent systemic infection. Expectant management of mucosal involvement will help prevent serious sequelae of the illness, described below.

The use of a number of active agents in the treatment of SJS/TEN has been described, including intravenous immunoglobulin, ciclosporin, corticosteroids, thalidomide, infliximab, and etanercept, but there is insufficient evidence to conclusively support the use of any of these.

If the patient survives the acute phase of illness, a number of sequelae may be experienced. Corneal involvement in the acute phase may lead to blindness, involvement of the genital tract may lead to stenoses and the patient may experience dry mouth as a consequence of oral cavity involvement.

Drug reaction with eosinophilia and systemic symptoms (DRESS)

DRESS is a drug-induced phenomenon comprising a constellation of clinical features: a characteristic rash (usually a maculopapular exanthema, associated with head and neck oedema; Figure 7-10a, b), fever, lymphadenopathy, eosinophilia and involvement of one or more solid organs (usually the liver). Mortality is estimated at 5%, this being largely attributable to the small number of cases who develop fulminant liver failure in the context of DRESS. Other solid organs may also be involved, including the pancreas, the kidneys, the lungs, the heart and thyroid gland. The latency period following drug exposure is generally more protracted that in other drug-induced syndromes, being 15-60 days. For this reason, the diagnosis is often overlooked, and symptoms of rash, fever and lymphadenopathy attributed incorrectly to infection. Management consists of withdrawal of the offending drug plus administration of corticosteroids. The latter may be given topically in the mildest cases, but generally either oral corticosteroids in the form of prednisolone or intravenously in the form of methylprednisolone will be required. Drugs with high notoriety for causing this condition are listed in Table 7-1, with the anticonvulsants and allopurinol accounting for a large proportion of cases.

Acute generalised exanthematous pustulosis (AGEP)

This is a rare pustular drug reaction recognisable by the appearance of sheets of non-follicular pustules which have a predilection for the major flexures (axillae, groin, and neck) appearing three to seven days after ingestion of a culprit medication (Figure 7-11a, b). The pustules resolve over three to seven days, in a phase characterised by post-pustular desquamation. The rash may be accompanied by fever and oedema, and in a small number of cases by systemic upset with involvement of the lungs or the liver. Recovery may be accelerated by the use of topical or oral corticosteroids. Antibiotics are the most common culprit drugs.