Prediction and management of drug reaction with eosinophilia and systemic symptoms (DRESS)

KEYWORDS:

• Drug eruption
• drug-induced hypersensitivity syndrome
• drug reaction with eosinophilia and systemic symptoms
• viral reactivation
• regulatory T cells

1. Introduction

Drug-induced hypersensitivity syndrome (DiHS)/drug reaction with eosinophilia and systemic symptoms (DRESS) is a life-threatening, drug-induced, multi-organ system reaction, which is characterized by sequential reactivations of herpesviruses [1,2]. Because this syndrome is initiated by fever of >38°C and maculopapular morbilliform exanthema, viral infections could be a potential consideration at onset. The delayed onset of clinical symptoms of DiHS/DRESS, usually 2–6 weeks after starting therapy, often causes the potential diagnostic delay (Figure 1). Paradoxical worsening of clinical symptoms often occurs 3–4 days after withdrawal of the causative drug, a quite unusual finding for drug eruption: owing to increasing concern of infections, unnecessary empirical antibiotic therapy may be started, thereby increasing the risk of developing additional sensitization to these drugs. Importantly, patients with DiHS/DRESS often show unexplained cross-reactivity to multiple drugs with different chemical structures, including nonsteroidal anti-inflammatory drugs (NSAIDs) [1,3]. A detailed drug history is very important to identify the causative drugs. This editorial presents a management-based approach to the diagnosis and treatment.

Figure 1. Clinical course of DiHS/DRESS [3]. This syndrome usually begins with a fever shortly followed by a maculopapular rash >3 weeks after starting therapy with a limited number of drugs, such as anticonvulsants. Patients usually develop two or three features of symptoms followed by a step-wise development of other symptoms. These symptoms continue to deteriorate or several flare-ups can be seen even for weeks or months after stopping the offending drug. Serum Ig levels continue to decrease for a weak after withdrawal of the drug. Despite such a wide variety of clinical symptoms, HHV-6 reactivation occurs 2–3 weeks after onset.

2. Confirmation of the diagnosis of DiHS/DRESS

Skin manifestations of DiHS/DRESS are not specific, but most of erythematous macules do not evolve into blisters and no mucous membrane involvement is usually seen, which help distinguish DiHS/DRESS from Stevens–Johnson syndrome/toxic epidermal necrolysis (SJS/TEN). In the histology, severe epidermal necrosis observed in SJS/TEN is absent even in the fully evolved lesions. Leukocytosis with atypical lymphocytes and eosinophilia of various degrees is unique features of this syndrome. According to a recent RegiSCAR study [2], transient eosinophilia is far more often present　(95%) than is usually reported, suggesting a need for frequent monitoring of eosinophilia. Thus, DiHS/DRESS should always be considered early in the evaluation of unexplained eosinophilia. Other reports, however, described that eosinophilia can be observed at most in 60–70% of patients with DiHS/DRESS [3]; eosinophilia cannot necessarily be observed depending on the timing of sampling. The diagnosis of DIHS/DRESS could easily be missed or delayed, as far as the diagnosis is only established on the presence of eosinophilia. Because the list of drugs that have been associated with DiHS/DRESS is very limited [1,3], a diagnosis of DiHS/DRESS could be considered when these symptoms as described earlier are present in patients with anticonvulsant therapy for >3 weeks. The elevation in liver enzymes occurs in up to 70% of patients with DiHS/DRESS [3]. A marked decrease in serum immunoglobulin (Ig) levels is typically observed at the acute stage, although their levels eventually return to normal upon recovery [4]. Human herpesvirus 6 (HHV-6) reactivation can be widely used as a specific and sensitive diagnostic clue in Japan and also in EU [1,3–7], while the validity of this test has not necessarily been confirmed in other countries largely due to the lower testing rate: unfortunately, these tests are not yet routinely available in these countries. HV-6 reactivation as evidenced by the significant increase in serum IgG titers to HHV-6 and HHV-6 DNA in leukocytes can be detected in the vast majority of DiHS/DRESS at a certain time point, 2–3 weeks after onset (1, 3–7). According to our sequential analyses of viral loads in patients with DiHS/DRESS, the cascade of reactivation events is initiated by HHV-6 or Epstein–Barr virus extends, with some delay, to HHV-7 as well, and eventually to cytomegalovirus (CMV) [5]. Frequent deterioration or several flare-ups of clinical symptoms occurring beyond the point where the causative drug would be expected to be eliminated from the body could be due to sequential reactivations of herpesviruses, which would occur in various organs in a sequential order but totally independent of that observed in leukocytes.

The lymphocyte transformation test (LTT) is a reliable in vitro method to confirm the causative drug, but positive LTT reactions can only be obtained at the recovery stage, > 2 months after onset, but not the acute stage, unlike those in other drug eruptions, in which positive LTT reactions are usually observed when the test was performed at the acute stage [8]. Thus, negative LTT reactions at the acute stage followed by positive LTT reactions at the recovery stage are a key laboratory finding that can aid in the identification of the causative drug. The acute stage of DiHS/DRESS is characterized by expansion of regulatory T cells (Tregs) [9]. This expansion of Tregs would occur to counteract activation of drug-specific effector T cedlls (Teffs), thereby delaying the activation of Teffs and allowing sequential reactivations of herpesviruses. Consistent with this explanation, in vitro proliferation of drug-specific Teffs as evidenced by a positive LTT reaction can only be detected at the resolution stage, at which time expansions of Tregs have been contracted and are associated with a gradual loss of Treg function.

3. Prediction of DiHS/DRESS

Genetic factors predisposing to certain types of drug eruptions induced by limited drugs have been identified in several HLA alleles. The most important association between HLA and drug eruptions includes that between HLA-B*15:02 allele and carbamazepine (CBZ)-induced SJS/TEN and that between HLA-B*58:01 allele and allopurinol-induced DiHS/DRESS and SJS/TEN. Indeed, predictive genetic testing for HLA-B*15:02 allele in the Taiwanese population has been shown to reduce the incidence of CBZ-induced SJS/TEN [10]. We also have recently demonstrated a strong association between HLA-A*31:01 allele and CBZ-induced drug eruptions including DiHS/DRESS in the Japanese population [11]: this association was confirmed in the Caucasian and Han Chinese populations. We have conducted a prospective study to demonstrate whether prospective screening of HLA-A*31:01 before prescribing CBZ could reduce the incidence of the CBZ-induced drug eruptions: this pre-prescription genotyping significantly decreased the incidence in the Japanese population (Mushiroda T. et al. Manuscript submitted). These findings suggest that these HLA alleles could be among the important factors involved in determining a patient’s susceptibility to certain drugs and act at least as surrogate markers of predisposition.

4. Severity of DiHS/DRESS

The severity of clinical symptoms at onset provides only a guide to prognosis and various clinical symptoms develop at various time points after onset, indicating difficulties to predict its prognosis. There is a clear need to identify patients at high risk for progression to a more aggressive stage of the disease. Thus, as soon as the diagnosis of DiHS/DRESS has been made, the severity and prognosis of the disease should be sequentially evaluated so as to define how to manage the patients. Unfortunately, however, no previous studies have established a scoring system for sequentially evaluating the severity and prognosis of the disease. According to our retrospective series of 55 patients, 4 patients died during the course of DiHS/DRESS (mortality, 7.2%) and an infectious cause was highly probable in the 3: they included acute respiratory distress syndrome related to CMV infection, septic shock, and Pneumocystis jirovecii pneumonia. The three were >75 years old, suggesting that the age could be the most important predictive factors for the mortality. Allopurinol was the most frequently involved drug resulting in a fatal outcome. DiHS/DRESS patients receiving allopurinol are often associated with concurrent renal dysfunction, indicating that renal dysfunction may be one of the most important predictive factors for the prognosis. CMV disease such as intestinal pneumonia and abrupt gastrointestinal bleeding, which frequently occurs 3–7 weeks after onset [12] and may worsen the prognosis, is also among the most important factors for the prognosis.

Although antibiotics and antipyretics such as NSAIDs have been also implicated in some cases of DiHS/DRESS, attributing these cases to these drugs is usually difficult because they are often administered for the early manifestations of DiHS/DRESS and may only be cross-reactive with the original causative drug; however, further studies are needed to elucidate the mechanism of the cross-reactivity. Thus, patients with DiHS/DRESS should be advised to avoid not only the causative drugs but also any antibiotics and antipyretics unless there is definite evidence for infection: the decision to administer antibiotics and antipyretics must be carefully deliberated.

5. Specific therapy

Systemic corticosteroids are the first-line therapy for DiHS/DRESS and most patients respond rapidly to moderate- or high-dose therapy (40–60 mg prednisone equivalent daily): fever and rashes rapidly resolve and abnormal liver enzymes have normalized within several days. Systemic corticosteroids administered during the acute stage had differential effects on the complications and prognosis depending on the stage examined [13]. Various infections were noted in the corticosteroid treatment group in the acute phase (<6 months), including herpes simplex, herpes zoster, P. jirovecii pneumonia, and CMV diseases. Of note, most infectious diseases appeared within 3 weeks, coincided with a tapering of their corticosteroid dose.

This clinical course of DiHS/DRESS patients is reminiscent of immune reconstitution inflammatory syndrome (IRIS) [14]. Although IRIS is originally defined as a paradoxical deterioration in clinical symptoms of infectious diseases attribute to the recovery of the immune response following institution of antiretroviral therapy in HIV patients, IRIS-like symptoms have been reported to develop on withdrawal or reduction of immunosuppressive agents, such as prednisone, in the setting of therapy with immunosuppressive agents. Clinical illness consistent with IRIS includes tuberculosis, herpes zoster, herpes simplex, CMV disease. P. jirovecii pneumonia and DiHS/DRESS. To prevent rapid restoration of immune responses, systemic corticosteroids need to be gradually tapered at least over >8 weeks. Because CMV disease occurring during corticosteroid taper in patients with DiHS/DRESS often results in fatal outcomes [12], the direct anti-CMV medications, such as ganciclovir, while maintaining corticosteroid doses, may help to avoid disease progression to full manifestations of IRIS.

In contrast, the development of autoimmune diseases such as lupus erythematosus and autoimmune thyroiditis, along with the generation of autoantibodies, were preferentially observed in the noncorticosterid treatment group in the late phase (>6 months) of DiHS/DRESS [15]: severe liver damage and noncorticosteroid therapy during the acute stage were associated with the subsequent generation of autoantibodies against plakin family proteins [15]. Thus, corticosteroids, especially if administered in the acute stage, may improve long-term outcome.

It is recommended that systemic corticosteroids be initiated in a sufficient dose of 40–60 mg (figure 2) prednisone equivalent daily for most cases of DiHS/DRESS and be followed by a gradual dose reduction of prednisone given over 10 weeks so as to prevent rapid reconstitution of valid immune responses against various pathogens, although the mild form would resolve spontaneously over a period of weeks. Although it has been suggested that a short course pulse of high-dose corticosteroids may be of benefit, our retrospective study demonstrated that patients treated with pulsed corticosteroids tended to run a more aggressive or protracted course, probably because the therapy is intrinsically associated with the process requiring rapid reduction or withdrawal of corticosteroids. Plasma exchange and high-dose intravenous immunoglobulin (IVIG) (figure 2) shown to be effective for the treatment of SJS/TEN may have adverse consequences in DiHS/DRESS. This results may reflect the differential effects of these therapies on Tregs: SJS/TEN is characterized by marked dysfunction of Tregs while functional Tregs are expanded in DiHS/DRESS [9]; and pulsed corticosteroids and IVIG are suggested to enhance or restore impaired Treg function. However, these considerations do not exclude the possibility that pulsed corticosteroids, IVIG or plasma exchange may be of benefit only when used in the late phase of DiHS/DRESS characterized by dysfunction of Tregs, given the absence of other validated specific therapeutic alternatives.

Figure 2. Algorithm for prediction, diagnosis and treatment of patients suspected of having MP rash, SJS/TEN or DiHS/DRESS. MP, maculopapular; IVIG, high-dose intravenous immunoglobulin pulsed corticosteroids, IVIG and plasma exchange are of benefit for the treatment of SJS/TEN, but not DiHS/DRESS.

6. Conclusions

There are no evidence-based guidelines for the management of DiHS/DRESS. Although systemic corticosteroids remain the mainstay of treatment for most patients, the current data do not allow a conclusion to be drawn as to the efficacy of other treatment modalities such as cyclosporine due to the small number of patients treated, the lack of a survey of long-term outcomes and the presence of the frequent confounding factors. Thus, it becomes increasingly important to establish scoring systems or biomarkers, by which the severity, prognosis and efficacy of the treatment can be evaluated.

7. Expert opinion

Because DiHS/DRESS would occur upon tipping the balance between drug- or pathogen-specific immune responses and self-protective Treg responses, treatment should be aimed at restoring the balance by administrating anti-inflammatory medications while maintaining appropriate treatment for pathogens. Therapeutic choices should be guided not only by the severity of the clinical manifestations in the acute stage and IRIS occurring during a steroid taper but also by autoimmune responses and diseases as long-term consequences of IRIS or DiHS/DRESS. For better management of DiHS/DRESS patients, there is a great need to establish scoring systems and biomarkers.

Declaration of interest

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This work was supported in part by the Health and Labour Sciences Research Grants (Research on Intractable Diseases) from the Ministry of Health, Labour and Welfare, Japan (to T Shiohara).