Drug reaction with eosinophilia and systemic symptoms (DRESS) in children.

Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS) is a severe reaction to drugs. Incidence of DRESS in children is not well known and the mortality rate seems to be lower than 10%. Anticonvulsants are the main drugs involved both in adults and in children. The treatment of choice is intravenous immunoglobulins and corticosteroids used in synergy. Today there are not controlled clinical trials regarding DRESS treatment in children. Anyway, the prompt withdrawn of the offending drug is of paramount importance for a better prognosis. DRESS sequels may occur, consequently, follow-up visits are required at least until the first year after the reaction.

Epidemiology

Drug reaction with eosinophilia and systemic symptoms (DRESS) is a serious and potentially fatal adverse reaction to therapeutic medications. Over the last 80 years, the nomenclature of this disease has been changing from drug-induced pseudolymphoma, anticonvulsant hypersensitivity syndrome, drug induced hypersensitivity syndrome (DIHS), drug induced de-layed multiorgan hypersensitivity syndrome to DRESS. DRESS is classified among severe cutaneous adverse reactions (SCARs) and in 1966 Bocquet et al. (1) identified it as a distinct clinical syndrome. Moreover, the meaning of “R” in DRESS acronym has been changed from Rash to Reaction due to the heterogeneity of skin eruptions (2). Initially, DRESS was thought to affect only adults, later it was diagnosed also in children (3).

The incidence of DRESS due to antiepileptics is in the range of 1:1000 to 1:10.000 in general population (4) and of 0.4:1000 (5) in hospital settings. In younger children the incidence of DRESS seems to be lower than in adults, although the real incidence is not known (6,7). Anyway, DRESS is more frequent than other severe immediate drug-induced reactions such as anaphylaxis (8), or exercise-induced anaphylaxis (9) but less common than food-induced anaphylaxis (10,11). The overall mortality rate is of 10% with a lower percentage in children than in adults (12-13).

Pathogenesis

DRESS is the result of a complex interplay of genetic factors [ethnic predisposition in people with certain human leucocyte antigen (HLA) alleles], immunological response, abnormalities in metabolic pathways (such as a deficiency or abnormality in epoxide hydroxylase, an enzyme that detoxifies the metabolites of aromatic amine anticonvulsants) and associated reactivations of herpes virus family members (HHV-6 and HHV-7, EBV and CMV) (14). In this context, African Americans are most likely to develop DRESS syndrome after initiation of aromatic anticonvulsants drugs whereas the Han Chinese are most likely to develop DRESS after allopurinol intake (15).

In fact, it has been found that DRESS syndrome is associated with certain human leukocyte antigens (HLAs), such as, HLA A*31:01 (aromatic anticonvulsant-induced DRESS); HLA A* 24:02 (lamotrigine-induced DRESS); HLA B*51:01, HLA B*15:13 and CYP2C9*3 (phenytoin-induced DRESS); HLA-B*57:01 and DRB1*01:01 and HLAB*35:05 (abacavir-induced DRESS) and HLA-B*58:01 (allopurinol-induced DRESS); HLA C*04:01 (nevirapine-induced DRESS) (16-19).

Apart from HLA, cytochrome P4502C9 marker has been reported to be involved in phenytoin induced SCARs (20-21).

Moreover, being a slow acetylator of drugs is thought to be a risk factor for DRESS syndrome (22).

Drugs may act as foreign antigens, binding to HLA/peptide/TCR complex and inducing hypersensitivity reactions. DRESS is a delayed type reaction according to Gell and Coombs classification (23).

There are four hypotheses regarding drug presentation mechanisms that have been suggested to explain how small drug molecules might interplay with HLA and TCR in drug hypersensitivity: (1) the hapten theory, (2) the pharmacological interaction with immune receptors (p-i) concept (i.e. carbamazepine directly interacts with HAL B*15:02) (3) the altered peptide repertoire model (i.e. abacavir binds to the F-pocket of HLA B*57:01), and (4) the altered TCR repertoire model (i.e. sulfamethoxazole directly interacts with TCR).

In delayed type reactions such as DRESS syndrome, drug antigens may activate specific T lymphocytes or natural killer cells with production of various cytokines/chemokines (i.e. TNF-α, IFN-γ, IL-2, IL-4, IL-5, TARC/CCL17, IL-6, IL-15, and IL-13) (16).

Furthermore, viruses have also been proposed to be involved in HLA/drug/TCR interactions and play an important role in drug hypersensitivity reactions, representing a source of exogenous peptides for drug presentation (24).

So far, the role of viruses in the pathogenesis of DRESS is unclear: a) Viral reactivation may be provoked by a cytokine storm secondary to an immune response against the drug (25); b) DRESS is a consequence of a strong immune response against an early viral reactivation (26). CD4+ and CD8+ drug-specific T cells proliferate after encountering the drug, but also anti-viral specific T cell can be cross-activated by drugs. In conclusion, the most common hypothesis is that the immunologic response to drugs induces a boost viral reactivation, consequently T lymphocytes and monocytes/macrophages release viruses that represents as an early marker of stimulation of these cells, rather than the triggering event in the pathogenesis of DRESS (27). In particular, toxic drug metabolites accumulation provoke an immunosuppression of B cells with hypogammaglobulinemia and subsequent viral re-activation (28). For example, in Asia and Europe the detection of HHV-6 copies in DRESS cases has been commonly reported with a frequency of 70-80%, making this data as an available diagnostic test (29, 30).

Clinical manifestations

The time onset of DRESS symptoms ranges from 2-6 weeks after initiation of treatment (2), anyway latency periods up to 105 days have been described (31).

Retrospective studies have found that the average age of occurrence of DRESS syndrome is 9 years of age in children (7,13). The most common clinical feature is fever, which is usually high grade ranging from 38-40°C. The second most common feature is macular erythema. This kind of rash later evolves in more violaceous and papular lesions with or without pruritus (Figure 1), and over time, the eruption becomes potentially exfoliative. Consequently, although a maculopapular rash is the most common initial cutaneous manifestation other eruptions may be described, including targetoid, urticarial, pustular, blistering, lichenoid, exfoliative, and eczematous lesions. The skin eruption typically begins on the face associated to facial oedema and then involves the upper trunk progressively spreading at lower extremities. The skin can be involved from less than 50% of body surface to diffuse erythroderma, making consistent the cutaneous distribution of the eruption. Moreover, mucosal involvement has been frequently (>50%) described (i.e. conjunctivitis, oral mucositis and/or genital lesions) in DRESS (7).

Figure 1.

Acute Rash in DRESS syndrome

The eruption can persist for months after the offending drug has been discontinued. Lymphadenopathy is the third most common presentation, which is seen in 70-75% of patients (32).

Haematological abnormalities, such as leucocytosis, eosinophilia, atypical lymphocytosis, thrombocytopenia and agranulocytosis commonly occur in DRESS.

Eosinophilia is typically reported in DRESS studies from both Asia and Europe with percentages ranging from 48 to 95% of patients (2, 7, 33).

Among visceral organ, liver (i.e. hepatitis) is involved in 50-80% of patients, followed by kidney (i.e. nephritis with haematuria or acute renal failure) in 11-28% of patients. Unfortunately, in some patients, hepatic injury can progress to widespread hepatic necrosis and fulminant liver failure (29, 34, 35) and it represents the leading cause of mortality in these patients (36).

Lungs (i.e. pneumonitis) are involved in 2.6-5% patients, but also muscle (myositis), gastrointestinal (i.e. colitis) heart (i.e. myocarditis), pancreas (i.e. pancreatitis), brain (i.e. encephalitis), thyroid (i.e. thyroiditis) and conjunctiva (i.e. conjunctivitis) involvements have been described (31, 37). In table 1 are reported the clinical features of DRESS syndrome (29, 35, 36, 38-44). Clinical manifestations were similar between children and adults, with the exception of pulmonary involvement (excluding asthma), which was more frequent in adults, and gastrointestinal involvement, which was more frequent in children (42).

Table 1.

Most common clinical features of DRESS syndrome and percentages of organ involvement

Fever (>38°C)	86.5%	(38)
Acute Rash	85%	(38)
Facial Swelling with periorbital involvement	27%	(38)
Lymphadenopathy	70%	(38)
Eosinophilia	60-80%	(29, 30, 38)
Liver:Hepatomegaly and/or increase liver enzymes (AST/ALT) and/or hyperbilirubinemia; elevated Alkaline phosphatase (30)	51-84%	(35, 36, 40-41)
Kidney:Elevation in creatinineDecrease in glomerular filtration rate (GFR)ProteinuriaHaematuria*Allopurinol is most commonly implicated with renal involvement (36)	11-57%	(35, 40-41)
Lungs:Interstitial pneumonitisPneumoniaPleural effusionAcute respiratory distress Syndrome (ARDS)*Minocycline, Allopurinol, Abacavir are most commonly implicated with lung involvement (26, 37)	2.6-5%	(29, 36)
Non specific Gastrointestinal Symptoms:ColitisDiarrhoea with or without electrolyte abnormalities	8%	(35, 42)
Heart: Late onset Myocarditis (Troponin and CKMB elevated)	4-27%	(43,44)
Tachycardia, arrhythmias, chest pain, non specific ECG changes, gross ST segment, elevation or depression, decrease in LV ejection fraction * Ampicillin is most commonly implicated with heart involvement

Drugs Involved

More than 40 medications have been described as triggers of DRESS and among various drugs, aromatic antiepileptics are reported to be the most common cause followed by antibiotics. Moreover, aromatic anticonvulsants show cross-reactivity in 40-80% of cases while non aromatic anticonvulsants are well tolerated as alternative choice in case of reactions to aromatics. Anyway, data about DRESS in children are scarce and mostly come from case reports. In the study of Misirlioglu et al (45), antibiotics were the most common (50%) medication in the aetiology; 87.5% of the suspected antibiotics were beta-lactams, and 12.5% were macrolides. Antiepileptics were second (37.5%, n. 6) most common class of drugs in the aetiology. In Table 2 we reported the drugs most frequently involved in DRESS syndrome in children in the last ten years. Studies where children were included but not clearly specified in terms of age and culprit drugs, were excluded.

Table 2.

Most frequently reported drugs causing DRESS syndrome in children

Single case or less than 10 children (mean age 7,6 years) (46-114)	carbamazepine 14/103 (13.6%) phenytoin 12/103 (11.7%) phenobarbital 9/103 (8.8%) valproic acid 6/103 (5.9%) vancomycin 5/103 (5%) lamotrigine 4/103 (4%) cefotaxime 4/103 (4%) trimethoprim-sulfamethoxazole 4/103 (4%) ceftriaxone 3/103 (3%) levetiracetam 3/103 (3%) dapsone 3/103 (3%) clindamycin 2/103 (2%) piperacillin-tazobactam 2/103 (2%) azithromycin 2/103 (2%) oxacarbamazepine 2/103 (2%) minocycline 2/103 (2%) sulfadiazine 2/103 (2%) oxacilline 2/103 (2%) penicillin 2/103 (2%) cefixime 1/103 (0.9%) naproxen 1/103 (0.9%) canakinumab 1/103 (0.9%) amoxi-clav 1/103 (0.9%) anakinra 1/103 (0.9%) tobramycin 1/103 (0.9%) ibuprofen 1/103 (0.9%) acetylsalicylic acid 1/103 (0.9%) griseofulvine 1/103 (0.9%) sulthiame 1/103 (0.9%) infliximab 1/103 (0.9%) fluoxetina 1/103 (0.9%) cefepime 1/103 (0.9%) allopurinol 1/103 (0.9%) perampanel 1/103 (0.9%) cefditoren-pivoxil 1/103 (0.9%) paracetamol 1/103 (0.9%) Ethambutol+rifampin+pyranzinamide 1/103 (0.9%) pyrimethamine 1/103 (0.9%) rufinamide 1/103 (0.9%)
32 children (mean age 8,9 y) (13)	13 carbamazepine 12 phenytoin 5 phenobarbital 5 lamotrigine 1 primidone 1 oxcarbamazepine
33 children (mean age 5,8 y) (115)	18 phenobarbital 15 phenytoin
29 children (mean age 11 y) (116)	10 trimethoprim-sulfamethoxazole 3 phenytoin 3 amoxicillin 2 cefalosporins 2 lamotrigine 2 minocyclin 2 macrolids 2 oxcarbamazepine 1 carbamazepine 1 clindamycin 1 zonisamide
11 children (mean age 6,6 y) (117)	4 lamotrigine 1 cefotaxime 2 carbamazepine 1 phenytoin + phenobarbital 3 amoxi-clav
16 children (mean age 8,2 y) (45)	3 amoxi-clav 1 ampicillin-sulbactam 2 cefdinir 1 cefotaxime 1 clarythromycin 3 carbamazepine 1 lamotrigine 1 phenytoin 1 phenobarbital 1 sulfasalazine 1 oxymetazoline nasal spray

Diagnosis

Due to the variability of its presentation, DRESS is known as “the great mimicker” making difficult the diagnosis (118). In particular, DRESS symptoms resemble those of cutaneous and systemic infectious diseases and can appear up to 3 months after the initial culprit drug exposure. The allergy work-up should start with a detailed record of clinical history by focusing on the chronology of drug assumptions and physical examination. Laboratory testing is fundamental, it should include liver, and kidney functions, search for viral infections, complete blood count and coagulation testing.

There are no clear and specific histopathological patterns in skin biopsy that are characteristic of DRESS Syndrome. Maculopapular exanthema (MPE) may be the initial presentation of SCARs including DRESS (119-120). When comparing DRESS with MPE, skin biopsies showed differences in terms of inflammatory infiltrate, atypical lymphocytes, keratinocyte damage, dermal involvement and leukocytoclastic vasculitis, these characteristics being more frequently observed in DRESS cases than in MPE cases (86, 121). Few necrotic keratinocytes were associated with non-severe DRESS cases, otherwise high amount of necrotic keratinocytes with confluent necrotic areas were associated with severe DRESS, respectively. Anyway, the role of skin or lymph node biopsies remains controversial (119).

Eosinophilia is a diagnostic criterion for DRESS. In physiologic conditions, eosinophils are not present in skin, liver, lungs or other internal organs otherwise in DRESS, eosinophils are typically increased in blood, in skin and in involved organs. Eosinophils infiltrate organs in response to cytokines and chemokines including eotaxin-1, TARC, IL-5 and granule release representing key factors of tissue damage (122).

The discovery of biomarkers of drug hypersensitivity could be useful for the diagnosis of DRESS syndrome. In DRESS cases, serum TARC levels have been reported to be significantly higher than those in patients with Steven-Johnson Syndrome (SJS)/Toxic epidermal necrolysis (TEN) and MPE during the acute phase and to be correlated with skin eruptions (122). For this reason, TARC could be considered a potential biomarker for the early phase and disease activity of DRESS syndrome.

Re-challenging with the offending drug has been considered the gold standard to diagnose drug eruptions, but in suspected DRESS cases, it should not be used because of the life-threatening nature of this syndrome (2, 123). Patch tests can be useful to prove a drug-specific immune response in DRESS syndrome (124). On the contrary, patch tests to different allergens such as foods have a low diagnostic accuracy (125). In vivo patch tests represent a low-risk method for reproducing delayed hypersensitivity by re-exposing patients to low amount of suspected offending drugs. Anyway, the sensitivity and specificity of patch tests are different according to the drug tested.

The lymphocyte transformation/activation test (LTT/LAT) measures the proliferation of T cells to a drug (126, 127). Unfortunately, it is not standardized for many medications and it is difficult to perform. Furthermore, it usually yields a negative result early in the course of the syndrome, and lacks sensitivity. A positive LTT/LAT is useful to confirm the diagnosis due to very low false positive results (only 2%), however a negative test cannot exclude the diagnosis (128). All these factors prevent widespread use of this test.

For the diagnosis of DRESS syndrome different criteria can be used such as Bocquet’s criteria (1), The European Registry of Severe Cutaneous Adverse Reactions to Drugs and Collection of Biological Samples (RegiSCAR) criteria (7) and the Japanese group of Severe Cutaneous Adverse Reactions to Drugs (SCAR-J) criteria (Table 3). The RegiSCAR is most often used to diagnose DRESS (129), it is based on seven independent parameters and three of them are required for the diagnosis of DRESS. According to RegiSCAR, the diagnosis of DRESS can be definite (score >5), probable (score 4-5), possible (score 2-3) and no (score <2) DRESS syndrome.

Table 3.

Three proposed diagnostic criteria of DRESS syndrome

	Bocquet et al. (1)	RegiSCAR (7)	J-SCAR (129)
Requirement for diagnosis	≥3 criteria	≥3 criteria of the following asterisk marks	all 7 criteria = typicalwithout 2 asterisk marks = atypical
History		- hospitalization- reaction suspected to be drug related	- symptoms persisting at least 2 weeks after drug discontinuation
Fever		- fever ≥38°C*	- fever ≥38°C
Cutaneous finding	- acute drug eruption	- acute rash	- macular rash developing 3 weeks after starting offending drug
Hematologicabnormalities	- eosinophilia >1.5×109/L or atypical lymphocytosis	one of the following hematologic abnormalities- eosinophilia over laboratory limits- lymphocyte count over and under normal limits- thrombocytopenia under laboratory limits	one of the following hematologic abnormalities- leucocytosis (>11×109 /L)- atypical lymphocytes (>5%)- eosinophilia (>1.5×109 /L)
Other organ involvements	- lymphadenopathy ≥2 cm in diameter- hepatitis with liver transaminases ≥2 times of the normal values- interstitial nephritis- interstitial pneumonitis- carditis	- lymphadenopathy involving ≥2 sites*- at least 1 internal organ involvement*	- lymphadenopathy*- liver abnormalities (ALT >100 U/L)
Viral reactivation			- HHV-6 reactivation*

Differential Diagnosis

DRESS should be differentiated from viral exanthemas especially EBV infectious mononucleosis, staphylococcal and streptococcal shock syndrome, meningococcemia, non infectious drug eruptions (e.g. SJS/TEN), autoimmune diseases (e.g. hypereosinophilic syndrome, Kawasaki disease, Stills diseases), urticaria vasculitis (130), neoplastic diseases (e.g. leukemia cutis, pseudolymphoma, mycosis fungoides), serum sickness like reaction, and atopic eritrodermia. In the last, for example, nephritis and hepatitis are lacking, being caused by bacterial infections (131).

Depending on organs involved, the differential diagnosis also includes viral hepatitis (liver), parasitic infection (gastrointestinal tract) and bacterial, viral and fungal pathogens (lung) (36).

Treatment

So far, there have been no prospective clinical trials on treatment of DRESS syndrome. Current recommendations are based on case reports and expert opinion (132). The first and most important step in treatment of DRESS is withdrawal of the causative drug, because a better prognosis is associated with an earlier discontinuation of the drug.

In mild forms, treatment is mainly supportive and symptomatic, consisting of topical steroids, systemic anti-H1 antihistamines and emollients for rash and itching. In patients with exfoliative dermatitis a prompt and appropriate fluid, electrolyte and nutritional support is of primary importance. In moderate cases without visceral involvement, corticosteroids are usually adequate (133).

In case of organ involvement, such as liver (transaminases >5 times upper limit of normal), kidney, lungs or heart, the expert opinion of French Society of Dermatology recommended to administer corticosteroid (prednisone, 1 mg/kg/day per os). Several aspects (i.e. optimal dose, route of administration, duration of treatment, and rapidity of dose tapering) of steroid treatment have not been compared in controlled trials (119). Tapering should take three to six months of time because rapid taper can be associated with relapse of DRESS (119, 134, 135). Systemic steroid therapy is advised to treat cases of moder-ate to severe disease taking into account the dramatic improvement in symptoms and frequent relapses of DRESS associated with quick prednisone taper. For all these reasons, intravenous pulses of methylprednisolone (1 g/d) are recommended especially in patients worsening despite adequate doses of oral corticosteroids (52).

Proposed mechanism by which corticosteroids benefit the patient is inhibition of IL-5, which attracts eosinophils, which are responsible for visceral organ damage by accumulation in DRESS syndrome (35). For the same reason, some authors proposed the use of mepolizumab (anti IL-5) in the treatment of DRESS (136).

Today, cyclosporine may be considered a second-line therapy for patients with severe organ involvement who do not respond to systemic corticosteroids and for patients in whom corticosteroids are contraindicated (137). Intravenous immunoglobulins (IVIG) have been reported to be useful in a few patients with DRESS and detrimental in others (138). Periodical controls (both clinical and laboratory parameters) are necessary to check progression of the skin eruption and/or development of clinical fatal life-threatening signs, which include hemophagocytosis with bone marrow failure, encephalitis, severe hepatitis, renal failure, and respiratory failure requiring treatment with steroids generally administered at a dose of 2 g/kg over 5 days with IVIG. The largest series of paediatric patients have been described by Marcu N et al. (62) who reported 7 patients with severe DRESS in whom treatment with IVIG (1-2 gr/kg) in addition to systemic corticosteroids was successful. One possible explanation is that IVIG preparations contain antiviral neutralizing antibodies that help clear the viral infection/reactivation found to be fundamental in the pathophysiology of DRESS. Anyway, IVIG should not be administered in the absence of steroids.

Due to the fact that there is a major viral reactivation along with presence of life-threatening signs, it has been proposed to administer anti-viral medications (e.g. ganciclovir) in combination with steroids with or without IVIG but the efficacy is unclear (139).

In severe and corticosteroid-resistant cases, more potent immunosuppressant medications includ-ing cyclosporine, azathioprine, rituximab, infliximab and mycophenolate have been used, sometimes alongside adjunctive treatment with IVIG and plasmapheresis (42, 66, 140, 141). N-acetyl cysteine (NAC), which acts as detoxifying drug, can also be used in DRESS.

Finally, the treatment of DRESS syndrome should be started immediately after diagnosis, even if the result of viral markers are still ongoing. Further studies with appropriate designs (i.e. randomized controlled trials) are needed to establish a standard of care in DRESS. Such studies should also assess the potential application of anti-viral drugs or probiotics for treating infections (142, 143, 144).

Prognosis

After withdrawal of the causative drug, most of the patients need some weeks to completely recover. The prevalence of sequelae is unknown. Long-term sequelae may be renal failure, chronic anaemia, autoimmune diseases (autoimmune thyroid disease, diabetes melli-tus type I, systemic lupus erythematous (SLE), systemic scle-rosis, adrenal insufficiency and autoimmune haemolytic anaemia). For example, thyroiditis has been reported in the 12.5% of children with a previous DRESS (7).

Moreover, recurrence of DRESS with unrelated drugs can be observed in 25% of cases, whereas very little or no flares are reported in patients after SJS/TEN (145, 146).

Those manifestations can occur months to years following the initial episode and awareness of association with a drug administration is crucial to promptly recognise and treat a possible DRESS. Fol-low-up visits at 2, 3, 4, 5, 6, 12 months and then once a year are recommended (146, 148).