Drug-Induced Hypersensitivity Syndrome (DIHS)/ Drug Reaction With Eosinophilia and Systemic Symptoms (DRESS): Clinical Features and Pathogenesis

Abstract

Drug reaction with eosinophilia and systemic symptoms/drug-induced hypersensitivity syndrome (DIHS/DRESS) is one example of a severe delayed T-cell mediated adverse drug reaction. DIHS/DRESS presents with fever, widespread rash and facial edema, organ involvement, and hematological abnormalities, including eosinophilia and atypical lymphocytosis. DIHS/DRESS is associated with relapse 2-4 weeks following acute symptoms, often coinciding with reactivation of prevalent chronic persistent human Herpesviruses (HHV) such as HHV-6, Epstein-Barr Virus, and cytomegalovirus (CMV). The mortality of DIHS/DRESS is up to 10% and often related to unrecognized myocarditis and CMV complications, with longer-term consequences that contribute to morbidity including autoimmune diseases such as thyroiditis. It is essential that all potential drug causes, including all new drugs introduced within the 8 weeks preceding onset of DIHS/DRESS symptoms, are identified. All potential drug culprits, as well as drugs that are closely related structurally to the culprit drug, should be avoided in the future. Systemic corticosteroids have remained the mainstay for the treatment of DIHS/DRESS with internal organ involvement. Steroid sparing agents, such as cyclosporine, mycophenolate mofetil, and monthly intravenous immune globulin, have been successfully used for treatment, and careful follow-up for CMV reactivation is recommended. Strong associations between HLA class I alleles and DIHS/DRESS predisposition include HLA-B*13:01 and dapsone, HLA-B*58:01 and allopurinol, and HLA-B*32:01 and vancomycin. These have opened a pathway for prevention, risk stratification, and earlier diagnosis. Single-cell sequencing and other studies of immunopathogenesis promise to identify targeted treatment approaches.

Introduction

Drug reaction with eosinophilia and systemic symptoms/drug-induced hypersensitivity syndrome (DIHS/DRESS) is a severe adverse drug reaction classically associated with multiorgan involvement and often complicated by herpesvirus reactivation. ¹ Mortality, dependent on patient age, underlying comorbidities, and the drug involved, occurs in up to 10% of patients and is typically related to underlying organ involvement such as myocarditis, and cytomegalovirus (CMV) complications, which can often be occult or unrecognized Several of these features distinguish it from the more commonly encountered morbilliform drug reaction.

In the past, "allopurinol hypersensitivity syndrome" and "anticonvulsant hypersensitivity syndrome" were used to refer to what we now call DIHS/DRESS. The first detailed description of phenytoin hypersensitivity syndrome in 1950 described an African American man with fever, rash, facial edema, cervical, axillary, and epitrochlear lymphadenopathy with atypical lymphocytosis, eosinophilia, and transaminitis 17 days following initiation of phenytoin. ² The causality between phenytoin and this syndrome was only apparent five weeks following the acute episode, when the patient was rechallenged with phenytoin and, following three doses, developed fever, malaise, and rash. ² The Japanese SCAR (J-SCAR) diagnostic criteria were proposed for DIHS in 2006, which consisted of seven components (Table 1)³, while the European Registry of Severe Cutaneous Adverse Reactions to Drugs and Collection of Biological Samples (RegiSCAR) consortium established its criteria for identifying possible DRESS cases in 2007 (Table 1). ⁴ Despite the nomenclature differences, both diseases probably belong to the same disease spectrum. ^{5, 6} Notably, typical DIHS may represent a severe form of DRESS, as Human Herpesvirus-6 (HHV-6) reactivation is included as a defining criteria⁷. In contrast to the J-SCAR consensus, HHV-6 reactivation was not included among the diagnostic criteria of Regi-SCAR; however, it remains possible that DIHS/DRESS is associated with human herpesvirus (HHV) reactivation, particularly HHV-6 and cytomegalovirus (CMV). ⁸ Patients with DIHS/DRESS are also at risk for systemic autoimmune sequelae, which can appear months and even years after resolution of the cutaneous eruption and acute systemic involvement. ⁹ These phenomena of viral reactivation and autoimmune sequelae have been considered an essential factor in understanding the definition, classification, and immunopathogenesis of DIHS/DRESS. Namely, DIHS/DRESS and its complications appear to occur due to complex interactions between immune responses and HHV, including antiviral- and drug-specific immune responses¹⁰. Although the correlation of herpesviruses reactivation and DIHS/DRESS has been reported, the specific cellular and molecular basis of how viral infections contribute to the pathogenesis of DIHS/DRESS and why subsequent autoimmune sequelae arise are yet to be elucidated. ^{6, 11}

Table 1.

Comparison of criteria for the diagnosis of DRESS/DIHS

DIHS (Japanese Consensus Group)3	DRESS (RegiSCAR)4
Maculopapular rash developing >3 weeks after starting with a limited number of drugs Prolonged clinical symptoms after discontinuation of the causative drug Fever (>38 °C)) Liver abnormalities (ALT>100 U/L) * Leukocyte abnormalities (at least one present) Leukocytosis (>11 x 109/L) Atypical lymphocytosis (>5%) Eosinophilia (>1.5 x 109/L) Lymphadenopathy HHV-6 reactivation	Acute rash Reaction suspected drug-related Hospitalization Fever (>38 °C) Laboratory abnormalities (at least 1 present) Atypical lymphocytosis Eosinophilia Involvement of >1 internal organ Enlarged lymph nodes >2 sites
The diagnosis is confirmed by the presence of the seven criteria above (typical DIHS) or of five of the seven (atypical DIHS).	The first 3 criteria are necessary for diagnosis, and the presence of 3 out of the other 4.

DIHS, drug-induced hypersensitivity reaction; DRESS, drug reaction with eosinophilia and systemic clinical manifestations; HHV, human herpes virus.

^*This can be replaced by other organ involvement, such as renal involvement.

General presentation

DIHS/DRESS presents typically two weeks to two months following initiation of a drug (Figure 1A). DIHS/DRESS characteristically presents with fever, generalized rash, lymphadenopathy, hematological abnormalities, and involvement of one or more internal organs. ^{1, 12} The early phase of disease occurs within the first couple of weeks of presentation, however, even after remittance of initial DIHS/DRESS symptoms and signs, viral reactivation of HHV can occur, and the symptom relapse can occur 2-4 weeks following the acute phase or upon weaning steroids. Later phase disease, such as DIHS/DRESS relapse or autoimmune complications, can also occur after all acute symptoms are quiescent, following weaning of steroids or other immunosuppressants, and may occur up to 5 years following the initial presentation. The time between the onset of DIHS/DRESS disease, removal of the suspected culprit drug, and resolution of symptoms is highly variable and dependent on poorly understood host, culprit drug (e.g. half-life), and treatment factors. Facial edema, although not part of the DIHS/DRESS scoring systems, is present in more than 75% of patients (Figure 1B). In addition, it has been correlated with the severity of disease ¹³, and can initially be hard to place if the examiner is not familiar with the patient's pre-existing appearance. Occasionally, erosion of the mouth and mucous membranes may be prominent and mistaken for Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN)-like oral involvement. Hematological abnormalities are very frequently observed, with atypical lymphocytosis potentially the most frequent and earliest in presentation but easily overlooked. Eosinophilia is present in more than 90% of reported cases as a later and persistent finding. ¹⁴ Lymphadenopathy is seen in more than 50% of patients, and hepatitis represents the most common internal organ involvement, with nephritis (often interstitial nephritis) being second. Myocarditis, pneumonitis, and neurological involvement can occur in up to a third of cases but can also be missed early unless carefully assessed. Myocarditis and pneumonitis may also occur as CMV diseases in the late phase (Figure 2A).

Figure 1. Timeline and clinical presentation of DRESS.

(A) Timeline of reaction onset compared to other drug reactions (FDE, Fixed drug eruption; MPE, Maculopapular exanthema; SSLR, serum sickness-like reaction; AGEP, acute generalized exanthematous pustulosis; SJS/TEN, Stevens-Johnson Syndrome/Toxic epidermal necrolysis; DILI, drug-induced liver injury). (B) Clinical presentation of DRESS across patients of different ethnicities and skin tones.

Figure 2. Immunopathogenesis of DRESS.

(A) Symptoms, sequalae, and multiorgan involvement observed with DIHS/DRESS. (B) Cellular immunopathogenesis of DRESS in the skin with the dermis as the primary site of the inflammatory response, enriched for CD4+ and CD8+ lymphocytes, plasma dendritic cells, and monocytes. CCL17/TARC is produced by dendritic cells that recruit CCR4+ Th2 T cells to the skin. IL-5 produced by Th2 cells and ILC2 induces differentiation, activation, and migration of eosinophils to the peripheral blood and tissues. TNFα and IFN-γ (Th1) and IL-4, -5, and -13 (Th2) are elevated in acute DIHS/DRESS. Viral reactivation of human herpesviruses is also observed and Tregs are present in lesional skin. (C) Models of drug-induced T-cell activation. In the hapten model, drug or drug-derived antigen forms covalent bonds to self-proteins/peptides to form a neoantigen hapten, processed by antigen-presenting cells for presentation by HLA risk allele(s) at the cell surface for recognition by corresponding T-cell receptors (TCR). The Pharmacological-Interaction (PI) model proposes that drug-antigen forms non-covalent, labile, and processing-independent interactions with either the TCR or HLA directly inducing T-cell activation. The altered self-peptide repertoire model proposes that drug-antigen binds to either the HLA (altered HLA) or TCR (altered TCR), changing the conformation of binding such that a different repertoire of endogenous self- peptides may bind that are recognized as immunogenic. (Treg, T regulatory cell; APC, antigen- presenting cell; DC, dendritic cell; IFN, interferon; TNF, tumor necrosis factor; ILC, innate lymphoid cell; TCR, T-cell receptor; HLA, Human leukocyte antigen). Figure created using Biorender.com.

The diagnosis of acute DIHS/DRESS can be challenging as it may, especially in the early stages, have many features in common with infections, lymphoproliferative and autoimmune diseases. DIHS/DRESS can be confused with other severe cutaneous adverse drug reactions (SCAR) given its long latency period. However, prominent organ involvement, hematological features, and rash with lack of mucosal involvement differentiate DIHS/DRESS from SJS/TEN and acute generalized exanthematous pustulosis (AGEP) (Figure 1A, Figure 2A).¹⁵

Cutaneous features

Skin manifestations are not detailed in the diagnostic criteria for DIHS/DRESS. Indeed, the nature of the rash is often polymorphic and can include maculopapular exanthema (MPE), lichenoid, exfoliative, urticarial, purpuric, eczematous, and pustular manifestations. ¹² Widespread erythema often develops into erythroderma, which persists even if the causative drug is withdrawn (Figure 1B). Target lesions with erythema multiforme-like skin manifestation may be present in severe DIHS/DRESS, and skin involvement is accompanied by significant dermal edema and inflammation. Associated hematological and internal organ features, as well as lack of involvement of mucosal surfaces, differentiate DIHS/DRESS from SJS/TEN. Blistering is not a feature of DIHS/DRESS, although occasional Nikolsky sign negative blisters can be seen on dependent surfaces due to significant underlying edema. ¹⁵ Edema and erythema on the face are considered to be highly characteristic, and there are reports of purpura on the lower extremities ¹⁶. There is also a report that disease severity is correlated with the area of purpura on the lower legs.¹⁶ Early treatment may modify the natural history of DIHS/DRESS. In particular, there are cases reported in which the diagnostic criteria are not met because fever is not observed due to the use of immunosuppressive reagents, such as corticosteroids and anti-IL-6 therapy, at the time of onset. ¹⁷ Of further note, DIHS/DRESS with lamotrigine may not show typical clinical findings in that liver dysfunction is mild and the number of atypical lymphocytes is low. ¹⁸

Immunopathogenesis

The dermis of DIHS/DRESS patients is enriched for CD4+ and CD8+ lymphocytes, plasma dendritic cells, and monocytes (Figure 2B). CD4+ and CD8+ T lymphocytes are also identified in biopsies of the internal organs in DIHS/DRESS. ¹ CCL17, also known as thymus and activation-regulated chemokine (TARC), is produced by dendritic cells in the dermis that recruit CCR4+ Th2 T-cells to the skin. ^19-21 IL-5 is produced by the Th2 cells and innate lymphoid cells 2 (ILC2), which induces differentiation, activation, and migration of eosinophils to the peripheral blood and tissues. ILC2 expressing the IL-33 receptor, serum soluble ST2, are present in the skin and blood during acute DIHS/DRESS, and sST2 appears to correlate with DIHS/DRESS severity. ²² Tumor necrosis factor-α and interferon-γ, which convey a Th1 pattern, are also known to be elevated in acute DIHS/DRESS (Figure 2B).¹

Although three non-mutually exclusive models of T-cell activation by drugs have been proposed (Figure 2C), it is not known if these are specifically relevant to the immunopathogenesis of DIHS/DRESS or vary for drugs that are associated with DIHS/DRESS based on their specific pharmacological properties. In the hapten/prohapten model, the drug covalently binds to a protein or peptide before processing. In the altered peptide and pharmacological-interaction models, the drug non-covalently binds to immune receptors. This interaction is dose-dependent, aligning with what is clinically observed for drugs such as allopurinol, which is more likely to cause DIHS/DRESS or, indeed any SCAR, during chronic renal failure driving the accumulation of its long-acting active oxypurinol metabolite. ²³ Importantly, mechanisms may not be mutually exclusive, with both hapten and direct pharmacological interaction models proposed during sulfamethoxazole-induced SCAR, including DRESS. Lastly, and although recently implicated during HLA-B*13:01-restricted dapsone-induced DRESS,^{24, 25} the altered peptide repertoire hypothesis is epitomized by the story of abacavir and onset of HLA-B*57:01-restricted drug hypersensitivity syndrome that is clinically distinct from DIHS/DRESS. ²⁶ By this model, abacavir occupies part of the antigen-binding cleft of HLA-B*57:01 and alters the repertoire of peptides presented to favor antigens to which the host has not previously been tolerized and triggers a hypersensitivity response. ²⁶ The long latency of DIHS/DRESS also differentiates it from abacavir hypersensitivity, where initial hypersensitivity symptoms have been described as early as 1.5 days following first exposure (Figure 1A). ^{27, 28} In DIHS/DRESS, patients are typically taking a drug for a minimum of two weeks (on first drug exposure) before hypersensitivity symptoms develop (Figure 1A). This suggests that there may be priming of an initial drug-specific response, unlike abacavir hypersensitivity, which may be mediated by cross-reactive memory CD8+ T-cells. ^{29, 30}

Single-cell and multi-omic approaches will bring a higher resolution to our understanding of the immunopathogenesis at a tissue level and will drive targeted approaches for prevention, earlier diagnosis, and treatment as recently demonstrated by a recalcitrant case of DIHS/DRESS where sc-RNA seq was used to identify cellular and molecular signatures at the tissue level to help guide treatment. ³¹

Immunogenetic Predictors of DRESS

Strong HLA class I associations with many drugs causing DIHS/DRESS support that these reactions may be triggered by CD8+ T-cells in the presence of CD4+ help. Intriguingly, many drugs associated with DIHS/DRESS are also associated with SJS/TEN, in some cases with a shared HLA class I association for allopurinol DIHS/DRESS/SJS/TEN and HLA-B*58:01, and dapsone DIHS/DRESS/SJS/TEN and HLA-B*13:01 (Figure 3A, Figure 3B, Table E1). ³² The factors driving the occurrence of tolerance as the dominant outcome versus DIHS/DRESS or SJS/TEN in an individual carrying an HLA risk allele are currently unknown. ³⁰ Although there have been many HLA discoveries associated with severe T-cell mediated adverse drug reactions (Figure 3A), few of these are used routinely as a pre-prescription screening test to prevent a specific reaction (Figure 3B). In a recent study where 1539 Chinese patients with leprosy were screened for HLA-B*13:01, there were no cases of dapsone hypersensitivity syndrome in those negative for HLA-B*13:01. ³³ The broad utility of HLA as a screening marker has been somewhat inhibited by both the very low predictive value of the HLA risk allele for the severe immune- mediated adverse drug reaction, such as HLA-B*58:01 and HLA-A*31:01 for allopurinol and carbamazepine DIHS/DRESS, respectively. This equates to a high number needed to test to prevent one case of DIHS/DRESS. In addition, in the case of allopurinol and the HLA-B*58:01 risk allele, it will only identify 60% of African Americans at risk of getting allopurinol-induced SCAR. Therefore, the discovery of additional HLA risk allele(s) relevant to this population will be necessary. Abacavir, a cause of a distinct HLA-class I-restricted CD8+ T-cell dependent drug hypersensitivity syndrome (Figure 1A) and not DIHS/DRESS, is an exception to this rule being one of the few drugs for which a pre-prescriptive HLA screening test is in routine clinical use. ²⁸ For abacavir hypersensitivity, the 100% negative predictive value of HLA-B*57:01 showed generalizability across race^{27, 34}, and a randomized double-blind placebo control study supported the utility of HLA-B*57:01 as a screening test to prevent abacavir hypersensitivity²⁸, with HLA- B*57:01 pre-prescription testing now routine in HIV practice across the developed world. ²⁸ For the allergist and immunologist, the utility of using HLA as a screening test may be further hampered by the fact that they would not be the typical prescribers of these drugs and are much more likely to be involved in risk stratification, diagnosis, and advice for future drug safety. ^{35, 36} As the number of associations across different races and ethnicities widens it may be that HLA class I testing is utilized as an adjunctive diagnostic test to help improve the probability of drug causality.^{32, 36} HLA-A*32:01 has been strongly associated with vancomycin-induced DIHS/DRESS in those of European ancestry.³⁶ An example of such pre-emptive and diagnostic usage would be using HLA-A*32:01 to risk stratify those requiring several weeks of vancomycin therapy to identify DIHS/DRESS risk and the need for treatment cessation (Figure 3C). To date, the strongest associations described have been in association with HLA class I alleles and small molecule-associated DIHS/DRESS (Figure 3A, 3B; Table E1). One exception to this rule is the association between HLA-DRB1*01:01 and nevirapine rash-associated hepatitis.³⁷ The IL-1 and IL-6 inhibitors, anakinra and tocilizumab, respectively, have also been associated with a DIHS- DRESS-like delayed hypersensitivity reaction in children with systemic-onset juvenile arthritis (sJIA) and adult-onset Still's a disease (AOSD) that is strongly associated with HLA-DRB1*15:01 and the related class II HLA alleles HLA-DRB1*15:03 and HLA-DRB1*15:06.³⁸

Figure 3. HLA risk alleles and use in prediction and diagnosis of DIHS/DRESS.

(A) HLA class I and II associations with DRESS (top) and other drug-induced reactions (bottom). (B) Examples of HLA class I risk allele associations for which there are current or pre-emptive screening tests available for SCAR.^{28, 36, 88-91} For many drugs the near 100% NPV might apply to a specific race (e.g. HLA-B*15:02 and HLA-B*58:01 in carbamazepine SJS/TEN and allopurinol DIHS/DRESS/SJS/TEN respectively. HLA-B*58:01 may explain only 60% of allopurinol DIHS/DRESS/SJS/TEN in European populations. (C) Implications for translation of HLA-A*32:01 screening in European populations for risk of vancomycin-induced DIHS/DRESS. HLA-A*32:01 is expressed in 6.8% of the European population, of whom 20% develop DIHS/DRESS after vancomycin exposure, providing a total number needed to test (NNT) to prevent one case of 75 individuals (Cen, Centromeric; Tel, Telomeric; DILI, Drug-induced liver injury; SJS/TEN, Stevens-Johnson Syndrome/Toxic epidermal necrolysis.

Diagnosis

Causative drug

The symptoms of DIHS/DRESS usually occur three or more weeks after starting treatment and with a limited number of drugs, mainly anticonvulsants.³ Indeed, carbamazepine, phenytoin, phenobarbital, zonisamide, lamotrigine, mexiletine, dapsone, sulfasalazine, minocycline, allopurinol, and vancomycin are implicated as the cause in most cases of DIHS/DRESS.^{8, 39, 40} (Table 2). The top five drugs associated with DIHS/DRESS have not changed significantly over the last 15 years. In the United States, these include vancomycin, allopurinol, lamotrigine, carbamazepine, and trimethoprim-sulfamethoxazole, which still account for almost 50% of cases.⁴⁰ Recently, there have been many reports of drug eruptions similar to DIHS/DRESS, such as immune checkpoint inhibitor-induced DIHS/DRESS-like skin disorders.⁴¹ Unique features of hypotension, angioedema, cytopenia and oropharyngeal involvement were common, whereas eosinophilia and lymphadenopathy were absent on presentation.⁴². DIHS/DRESS and other SCAR have also been associated with many immunomodulatory drugs with novel targets, including immune checkpoint inhibitors and the mechanism of these reactions is not known.^{43, 44, 45} Patients with DIHS/DRESS are at risk for relapse and sequelae for weeks and sometimes months (on weaning steroids) following their acute episode. This may be interpreted as multiple drug hypersensitivity, and there have been reports of unexplained cross-reactivity to multiple drugs with different structures, including those used after onset.⁸ Given the prolonged inflammation and complications of patients with DIHS/DRESS often being on prolonged courses of steroids, any adjunctive testing such as delayed intradermal or patch testing is not recommended until after least six months following the acute episode. ^{12, 46}

Table 2.

Common Small Molecule Causes of DIHS/DRESS⁴⁰

Allopurinol

Antimicrobials *

Dapsone

Minocycline*

Nevirapine

Trimethoprim- sulfamethoxazole * Vancomycin

Antiepileptics

Carbamazepine (and related drugs e.g. oxcarbazepine)*

Lamotrigine*

Phenytoin*

Phenobarbital*

Anti-inflammatory medications

Sulfasalazine

DIHS, drug-induced hypersensitivity reaction; DRESS, drug reaction with eosinophilia and systemic clinical manifestations.

^*most common causes in pediatric DRESS

Clinical course associated with organ complications and herpesvirus reactivation

In the acute phase of DIHS/DRESS, systemic involvement may include hepatitis, interstitial pneumonia, interstitial nephritis, and eosinophilic myocarditis. (Table 4) Hematological manifestations are frequently observed, such as lymphadenopathy, striking blood and tissue eosinophilia, peripheral lymphopenia, and atypical lymphocytosis.

Table 4:

DIHS/ DRESS severity score^{^}

Parameters	Grade/extent	Score
FIXED
Age, y	≤40	1
	41-74/	0
	≥75	2
Duration of drug exposure after onset, days
	0-6	0
	≥7	1
Allopurinol exposure	Yes	1
VARIABLE
Pulsed prednisone* Yes		2
Skin involvement (%BSA)
Erythema <70		0
Erythema≥70		1
Erythroderma		2
Erosion < 10		0
Erosion 10-29		1
Erosion ≥30		3
Fever ≥38.58C, days duration
	0 or 1 days	0
	2-6 days	1
	≥7 days	2
Appetite loss (≤70% of regular food intake), days
	0-4	0
	≥5	1
Renal dysfunction (creatinine), mg/dL
	<1.0	0
	1.0-2.0	1
	≥2.1 or HD	3
Liver dysfunction (ALT), IU/L
	<400	0
	400-1000	1
	>1000	2
C-reactive protein, mg/dL
	≤2	1
	2-9	0
	10-14	1
	≥15	2

^{^}See Figure 4 for reference of how to use scoring system

Each variable parameter was determined at early (days 0-3 after the initial presentation) and later times (2-4 weeks after the initial presentation), and on an as-needed basis.

ALT, Alanine aminotransferase; BSA, body surface area; HD, hemodialysis.

^*Intravenous methylprednisone use 500 mg/day for 3 days.

HHV-6 reactivation occurs in typical DIHS cases. In addition, other herpesviruses such as Epstein-Barr virus (EBV), HHV-7, CMV, and varicella-zoster virus (VZV), are also reactivated during the course of DIHS/DRESS in sequential order.⁴⁷ It is thought that this occurs even if the patient is not under immunosuppression such as steroid treatment, and some reports suggest that this could be immune-mediated inflammation against various antigens, including pathogenic microorganisms, drugs, and unknown autoantigens, during recovery from immunosuppressed conditions. Another report suggested that EBV reactivation was less common in DIHS/DRESS patients who received steroids. In addition, persistent reactivation of EBV and HHV-6 were significant risk factors for the future development of autoimmune disease.⁴⁸ Because inflammation associated with severe tissue damage and potentially reactivation of some viruses such as EBV is thought to be the underlying or preceding events for the subsequent development of autoimmune diseases, severe DIHS/DRESS associated with severe tissue damage would be accompanied by a heightened risk of developing some autoimmune diseases. (Table 4) According to a recent report, the satisfaction of all DIHS diagnostic criteria is suggested as an important predisposing factor for DIHS/DRESS-mediated autoimmune or inflammatory disease.⁴⁹

Pathology, Severity Score, Sequelae and Biomarkers

The pathology of DIHS/DRESS itself can be variable, and similar to the clinical picture of the rash, the features on histopathology can be polymorphic. ¹ The rash itself in the absence of other features does not define the clinical entity of DIHS/DRESS. The most common histopathological pattern of DIHS/DRESS is spongiotic dermatitis with eosinophils with a perivascular lymphocytic infiltrate in the dermis. ³⁶

Cases of DIHS/DRESS have a wide range of severity, from those that improve with palliative care alone to those that are difficult to treat. In addition, the prognosis of DIHS/DRESS is highly unpredictable. Severe complications, either related or unrelated to CMV reactivation, are a highly probable cause of death. ⁵⁰ Therefore, an adaptation to standard immunosuppressive therapy for the disease is urgently required. Mizukawa et al. performed a retrospective analysis on 55 patients with DIHS/DRESS and established a DIHS/DRESS severity score that predicts CMV disease from clinical data. ^{8, 51} (Table 4). Scores ≥4 were associated with the later development of CMV disease and complications, while no patients with scores <4 developed complications. Based on clinical courses of 55 patients, they proposed a flow diagram for diagnosis and prognostication of DIHS/DRESS (Figure 4).

Figure 4.

Analysis of 3 groups (mild, moderate, and severe cases) based on early scores to predict the risk of CMV reactivation or CMV disease and complications (Table 4). CMV reactivation including CMV disease and complications occurred most frequently in the severe group. CMV disease and complications developed exclusively in the severe group.

Patients who have DIHS/DRESS are at risk for long-term, autoimmune sequelae. These sequelae may appear after a long, symptom-free interval after complete resolution of the acute phase of DIHS/DRESS, or maybe a continuation of organ involvement that appeared during the acute phase (Table 3). ⁵² The lag time between the resolution of the acute phase and the development of autoimmune sequelae can be as long as four years. ⁹ Various sequelae are included, such as autoimmune thyroiditis, fulminant type 1 diabetes, autoimmune hemolytic anemia, and alopecia. This strongly suggests the need for long follow-up even after clinical resolution of the acute phase of DIHS/DRESS.

Table 3:

Sequelae of DIHS/DRESS

Short-term sequelae (within weeks following acute DIHS/DRESS onset)

Fulminant type 1 diabetes mellitus 71, 72

Fulminant hepatic failure 73, 74

Autoimmune hemolytic anemia 52

Renal failure 75

Disseminated intravascular coagulation 76

Myocarditis 68, 77

Pneumonitis 78 79 80

Hemophagocytic lymphohistiocytosis (HLH)81, 82

Autoimmune thyroiditis 9 83 78

Long-term sequelae (persistent and/or months to years following acute DIHS/DRESS onset)

Arthralgia (rheumatoid arthritis) 9, 84

Autoimmune thyroiditis 9 83 78

Vitiligo,9 83

Alopecia areata 78, 83

Myocarditis 85 86

Pneumonitis78

Systemic lupus erythematosus87

Very recently Mizukawa et al investigated a retrospective study of 55 patients with DIHS/DRESS followed for the possibility of later development of autoimmune disease – 18 years after resolution. The combination of increases in lymphocyte counts, severe liver damage, a rebound increase in globulin, persistent reactivation of EBV and HHV-6 as well as low IL-2 and IL-4 at the acute phases, were significant risk factors for the future development of autoimmune diseases. Based on these factors, a scoring system that can identify high-risk patients for development of autoimmune disease was established to categorize three groups (low/intermediate and high).⁴⁸

TARC is a potential biomarker for identifying acute and early DIHS/DRESS and appears to be more specific for DIHS/DRESS since levels were found to be significantly higher in DIHS/DRESS than SJS/TEN. ¹⁹ The TARC value is about three times higher in HHV-6 reactivated patients in the early onset of DIHS/DRESS than in the non-reactivated group, indicating that TARC is useful to predict virus reactivation. ¹⁹ It is also reported that the serum TARC levels in DIHS/DRESS were correlated with the severity of skin and mucosal lesions, fever, liver and kidney dysfunction, levels of HHV-6 and CMV DNA, and also cytokines such as IL-5, IL-10, and sIL-2R. This suggests that TARC might be not only a diagnostic marker but also a useful marker for assessing the overall clinical and immunological condition of DIHS/DRESS patients.⁵³

As purpura is one of the characteristic skin manifestations, we assessed the relationship between the presence of purpura and disease severity using the DIHS/DRESS severity score. ⁵¹ We found that disease severity was significantly correlated with the amount of skin surface area in the lower legs covered by purpura, suggesting an association between purpura and disease severity in patients with DIHS/DRESS. In a recent retrospective analysis of 49 patients with DIHS/DRESS, the presence of purpura on the lower extremities was positively correlated with acute disease severity and mortality. ⁵⁴

Pediatric cases

Recently, a summary of pediatric DRESS has been reported. ⁵⁵ The mean time from the drug intake until the onset of the first symptom was 18.9 days. Allopurinol was an infrequent cause in children. Antiepileptic drugs, including the aromatic anticonvulsants carbamazepine, phenytoin, and phenobarbital are the main cause in children, followed by anti-infectious agents such as beta-lactams, trimethoprim-sulfamethoxazole, and vancomycin. These data suggest adult and pediatric DRESS cases are not significantly different in clinical manifestations. Notably, relapsing DRESS was reported in 17 of 354 cases (4.8%), and fever, facial edema, lymph node enlargement, leukocytosis, pharyngeal and visceral involvement, and chronic medical conditions seem to correlate most strongly with DRESS reactivation in children. A recent series of 49 pediatric cases from France also highlighted the onset of a DRESS syndrome in the pediatric population (related to beta-lactam antibiotics in particular) might occur less than two weeks after treatment starts (median 13 days for all antibiotics). ⁵⁶ Another report highlighted long-term sequelae such as autoimmune disease (most commonly hypothyroidism, 3.8%) that occurred in approximately 11% of children.⁵⁷ It will be of interest to analyze risk factors in relapsing DRESS in adults and evaluate the usefulness of the DIHS/DRESS severity score in pediatric cases.

Specialized Testing

Clinical diagnosis of DIHS/DRESS is still considered the gold standard and is based on a typical latency period (2-8 weeks; Figure 1A) associated with the use of a drug recognized to cause DRESS with the help of a scoring system such as the RegiSCAR or J-SCAR. All patients who present with a potential diagnosis of DIHS/DRESS should have a complete basic work-up including complete blood count (CBC) and differential with peripheral smear, renal and liver function tests, creatinine kinase, troponin, electrocardiogram, and chest-X-ray at initial presentation. Thyroid function testing and antibodies should be done at baseline, because autoimmune thyroid disease, including both hyperthyroidism and hypothyroidism, can occur as a late complication. CBC and differential, renal and liver function should be checked frequently while acutely unwell and monitored weekly while weaning steroids. The pursuit of an alternative diagnosis that is included in the RegiSCAR scoring system should be done as clinically relevant. If the patient meets the criteria for DIHS/DRESS in the setting of having started a high-risk drug in the time frame, then ruling out other infectious and inflammatory causes may be less relevant. In clinical allergy practice, the common adjunctive tests used to help ascribe drug causality for DIHS/DRESS are patch and delayed intradermal testing.⁴⁶ In general, the utility of such testing is dependent on the drug and the formulation. One study suggested that delayed intradermal testing, performed with a sterile IV solution and read at 24 hours, is more sensitive than patch testing.⁵⁸ For both patch testing, with drugs applied in a soluble vehicle under occlusion on intact skin, and delayed intradermal testing, the highest non-irritating concentration should be used.⁵⁹ For certain drugs, such as allopurinol, patch testing with the parent drug or oxypurinol metabolite appears to have no utility.⁴⁶ Patch testing for antibiotic and anticonvulsants and patch testing or intradermal testing for antibiotics have proven useful. HLA has typically been thought of as a pre-prescription strategy to prevent DIHS/DRESS and other T-cell mediated adverse drug reactions. However, HLA testing may also be useful in the early diagnosis or diagnosis of DIHS/DRESS in specific cases (e.g., vancomycin and HLA-A*32:01 or HLA-B*58:01 and allopurinol)³⁶. HLA may be particularly helpful where combinations of multiple drugs at high risk to cause DIHS/DRESS were started concurrently. Diagnostic drug challenges are generally not utilized in the setting of DIHS/DRESS unless the benefit outweighs the risk.^{12, 60} Ex vivo and in vitro assays, such as interferon-γ ELISpot and lymphocyte transformation test (LTT), have been useful for some drugs, in particular antibiotics and anticonvulsants. However, they remain research tools and need to be validated and generalized across large populations.^{12, 61, 62} Sensitivity and specificity of the LTT is dependent on the drug but is significantly better in the recovery phase of DIHS/DRESS as compared to the acute phase.⁶³ In one study, combinations of in vivo and ex vivo (interferon-γ ELISpot) approaches successfully identified a culprit agent in 17/19 (89%) cases of severe maculopapular exanthem or DIHS/DRESS associated with penicillins, cephalosporins, or glycopeptides.⁶⁴

Treatment

The most important step in management is an immediate cessation of all potentially implicated drugs and aggressive supportive care. Systemic corticosteroids have been accepted as the standard treatment for improving clinical symptoms of DIHS/DRESS at the acute phase; however, the evidence for this is largely through case reports and case series. There are no clinical trials currently to support the treatment of DIHS/DRESS. ⁸ Rapid resolution of rashes, fever, and liver dysfunction occurs after starting systemic corticosteroids, with a minimum dose of 1 mg/kg/day recommended. However, systemic corticosteroids need to be tapered over typically a minimum of 6-8 weeks to prevent the relapse of various symptoms of this syndrome and thus are administered for 2-3 months: once systemic corticosteroids have started, dose should be reduced gradually even upon resolution of clinical manifestations and in some cases protracted courses out to a year or longer are needed. This is because patients with DIHS/DRESS are at greater risk of subsequently developing a wide spectrum of short and long-term immune sequelae ranging from viral reactivation and CMV disease to autoimmune disease.⁶⁵ Ongoing research is examining the role of potent topical steroids for milder DIHS/DRESS. The risk algorithm outlined in Figure 2 has also been used to help risk stratify and guide DIHS/DRESS treatment. In the absence of internal organ involvement and signs of severe rash, it is reasonable to manage the patient symptomatically with topical steroids, antihistamines, and careful monitoring.

For patients who are recalcitrant to steroid treatment, who relapse, or who have dose-limiting sequelae associated with corticosteroids, successful treatment has been noted in case reports with steroids sparing agents such as cyclosporine, mycophenolate, rituximab, plasmapheresis, and cyclophosphamide. ⁶⁶ Pulse steroid treatment typically consists of 5-20 mg/kg methylprednisolone for three days followed by a slow taper. More information is needed about the acute use of pulse steroids as the question has been raised whether this is associated with a higher risk of autoimmune disease.⁸

For cases with ongoing relapse or who are recalcitrant, some success has been reported with monthly intravenous immunoglobulin administered at doses of 0.5 g/kg/day-1 g/kg/day over two days.⁶⁷ Although there is no direct antiviral agent with activity against HHV-6, patients are at risk for reactivation of other Herpesviruses and some patients may benefit from treatment with valganciclovir when there is evidence of or suspected viral reactivation.³¹

The DIHS/DRESS severity score established by Mizukawa has been proposed as a guide for treatment choice (Figure 4), with systemic steroids recommended for moderate (score: 1-3) and severe (score: ≥4) disease. Corticosteroids are typically not necessary in mild cases (score: <1).⁵¹

There is currently little evidence for the routine use of the IL-5 antagonist mepolizumab or the IL-5 receptor blocker benralizumab in the treatment of DIHS/DRESS. Although IL-5 fuels maturation and differentiation of eosinophils and in part explains the eosinophilia seen in DRESS, organ failure associated with DIHS/DRESS is typically associated with organ infiltration with antigen-driven T-lymphocytes related to the implicated drug. However, in some cases, morbidity and tissue damage may be directly related to eosinophils such as in eosinophilic myocarditis or massive eosinophilia with hyperviscosity syndrome and risk of vascular events. Case reports have suggested some potential benefits in this context.^{68, 69} Two recent reports of DIHS/DRESS in the setting of COVID-19 with myocardial involvement that was recalcitrant to pulse steroids had improvement in blood eosinophils, skin, and liver function with benralizumab, although only one patient survived and the other died from cardiac arrest 17 days following DRESS diagnosis.⁷⁰

Recently, in the spirit of targeted approaches to therapy, 3'single-cell-RNA-seq (10X Genomics, droplet-based) was performed on skin and blood from a 44-year-old man with refractory DIHS/DRESS associated with trimethoprim-sulfamethoxazole, despite the use of multiple immunosuppressants. Single-cell transcriptomic profiles of skin-infiltrating lymphocytes showed upregulation of genes in the JAK-STAT signaling pathways.³¹ Through identifying the JAK-STAT signaling pathway as a potential target, the JAK-1/3 inhibitor tofacitinib was utilized, enabling disease control and tapering of other immunosuppressive agents.³¹ Thus, this "precision medicine" approach using scRNA-seq analyses guided successful therapeutic intervention in the patient with refractory DIHS/DRESS.

There are few guidelines for managing the treatment of DIHS/DRESS and supporting evidence has been limited to observational data and case series. Currently, the J-SCAR group is conducting a national epidemiological survey and formulating treatment guidelines.

Conclusions

Here, we have highlighted several emerging views about the clinical symptoms, pathogenesis, and management of DIHS/DRESS. Characteristic skin and other organ involvement are well recognized and might distinguish DIHS/DRESS and ordinary drug exanthema. Systemic corticosteroids are still considered to be the mainstay of treatment of DIHS/DRESS and using risk stratification can help predict corticosteroid dose to prevent lethal CMV reactivation. DIHS/DRESS often shows severe symptoms in the acute phase. In addition, there are serious sequelae in the chronic phase, and therefore careful follow-up is required. Since there is no definite view of how to treat and how long to follow, it is an important issue in the future along with the development of management guidelines.