Drug-induced hypersensitivity syndrome (DIHS), also referred to as drug reaction with eosinophilia and systemic symptoms (DRESS), is a severe cutaneous adverse reaction (SCAR) characterized by the presence of a rash, fever, and involvement of hematologic and visceral organs. The etiology of DRESS syndrome is currently not well-defined. Proposed mechanisms include aberrations in drug detoxification enzymes leading to the buildup of reactive drug metabolites, sequential reactivation of herpesviruses, such as cytomegalovirus, Epsteine-Barr virus (EBV), human herpesvirus (HHV) -6 and -7, and genetic predisposition linked to specific human leukocyte antigen alleles. However, a comprehensive understanding of the pathogenesis of DRESS syndrome is still lacking (1).

DIHS/DRESS is a highly debilitating and potentially life-threatening condition, historically associated with a mortality rate estimated at 10%, while recent studies have indicated mortality rates ranging from 1.2% to 6.1% (2). In the pediatric population, systematic reviews have reported mortality rates ranging from 3% to 5.4% (2).

To our best knowledge, so far there has been no article summarizing reports of death cases of DRESS. Therefore, we analyzed the signals of DRESS in FDA Adverse Event Reporting System (FAERS) and made a systematic review of the death case reports on DRESS, in order to provide reference for the security of drug usage in clinical practice.

DRESS is a life-threatening multi-organ system reaction. DRESS was often overlooked in clinical practice due to its unique clinical presentation, such as delayed onset and similarity to symptoms of other conditions. DRESS has a heterogenous clinical presentation making it challenging to diagnose, typically developed by assessing the timing of drug exposure in relation to the onset of skin rash and the resolution of symptoms following discontinuation of the offending drug. Recently, Brüggen et al. (30), reported a delphi-based international consensus on diagnostic workup, severity assessment, and management of DRESS, but it did not cover all aspects of DRESS management in detail and some topics remained controversial. Therefore, the development of more researches and future guidelines are necessary. The main advantages of our research is that we integrated the FAERS database analysis with case report assessments and a systematic review, effectively reducing the risk of bias that might come from using just one single database.

Results showed that allopurinol has the highest probability to induce DRESS and related death statistically. A comprehensive analysis (31) including 52 case reports revealed a mortality rate of 25.0% of allopurinol-induced DRESS, with septic shock, gastrointestinal bleeding, and multiple organ failure identified as the primary causes of death. Factors associated with increased risk of mortality included advanced age, underlying cardiovascular disease, chronic kidney disease, high allopurinol dosage, infection, and internal organ involvement such as the kidney, heart, lung, and gastrointestinal tract. In our systematic review, two patients (50%) had chronic renal failure, that may be related to their poor prognosis. The most common visceral manifestation were liver and kidney injuries. The human leukocyte antigen (HLA) class I alleles HLA-B*58:01 and HLA-C*03:02 have been implicated in allopurinol-induced SCAR, with HLA-C*03:02 being predominantly associated with DRESS (32). These findings underscore the importance of pre-treatment HLA genotyping in patients with compromised renal function to mitigate the risk of DRESS. Chronic renal failure was the cause of 2 out of 4 deaths caused by allopurinol in our systematic review. HLA associations with DRESS have provided insights into immunopathogenesis. In some SCAR cases, HLA oligoclonal CD8+ T-cell responses restricted by HLA are observed at the tissue level (33). Nevertheless, for the majority of drugs, the specific HLA risk alleles and antigens that drive this response remain unidentified. Additionally, HLA risk alleles exhibit incomplete positive and negative predictive values, rendering comprehensive screening a current challenge. Recently, single-cell RNA sequencing has been proved helpful in guiding therapeutic decisions on DRESS patients (34).

A systematic review (35) identified 1072 cases of psychotropic drug-induced DRESS indicated that carbamazepine, lamotrigine, phenytoin, valproate, and phenobarbital were the most implicated drugs. Studies have examined the genetic contributions to cutaneous adverse drug reactions caused by AEDs and discovered significant associations with HLA molecule genes, the cytochrome P450 enzyme and complement factor H genes (36). One case (1/26, 3.85%) in our systematic review were related to phenytoin, but the patient did not do any genetic test. Significant genetic factors have been identified in association with phenytoin-related SCAR. Patients with SCAR, particularly those carrying the CYP2C9*3 allele, exhibited delayed plasma phenytoin clearance (37). The alleles HLA-B*13:01, HLA-B*15:02, and HLA-B*51:01 demonstrated significant associations with phenytoin hypersensitivity, each exhibiting distinct phenotypic specificities (38). Therefore, integrating the evaluation of HLA and CYP2C9 risk alleles enhances the efficacy of predictive genetic testing for the prevention of phenytoin hypersensitivity in Asians. Although there were no deaths caused by lamotrigine among the deaths included in this article, lamotrigine ranked second in drugs calculated to induce DRESS and fifth statistically related to fatal outcome. In Asian populations, HLA-B*1502 increases the risk of lamotrigine- induced bullous lesions, while HLA-A*2402 is linked to susceptibility to SCARs (39). Conversely, HLA-A*3303 serves as a protective allele against lamotrigine- induced SCARs (39).

A review conducted in the USA (1980 to 2016) found that antibiotics, particularly vancomycin (39%), β-lactams (23%), fluoroquinolones (4%), tetracyclines (4%), and sulfonamides (3%), were implicated in total 74% of DRESS syndrome cases (40). In our study, vancomycin ranked third among the drugs related to DRESS and ranked second among the drugs statistically related to death. Vancomycin related DRESS syndrome happened within 21 days (41). A systematic case review (42) which analysis the immune-mediated reactions to vancomycin indicated that the incidence of DRESS syndrome was 16/71, 23%, men represented a majority (56%) with median latency of 21 days [IQR 17 days, 28 days] for DRESS syndrome. Our results showed that the time from the onset of the rash to the death with vancomycin were 40 to 56 days, which means once rash begins to appear, the next 20-30 days are very valuable for treatment. It is noteworthy that other glycopeptide antibiotics are also possible to cause DRESS syndrome in limited number of patients and may lead to allergic cross-reactivity with vancomycin (43). Vancomycin-associated DRESS may be due to genetic factors. A study identified that 20% of HLA-A*32:01 patients would develop vancomycin DRESS if exposed for more than two weeks (44). Another study reported the cross-reactivity of alternative glycopeptide antibiotics, namely teicoplanin and telavancin, in patients exhibiting HLA-A*32:01–restricted vancomycin-induced DRESS, who also share the expression of an HLA class II haplotype, specifically DQA1*01:01 and DQB1*05:03 (45). A PCR-based assay specific to HLA-A*32:01 has been developed to screen patients for the risk of vancomycin-induced DRESS (46).

Among the deaths included in this article, 3 patients were related to DRESS of minocycline, and 2 of them had cardiac involvement. Dyspnea and chest pain are the most prevalent symptoms in patients with DRESS and cardiac involvement, with minocycline being identified as one of the most frequently implicated medications (47). Due to the high mortality rate of DRESS with cardiac involvement (45%) (47), doctors and pharmacists should pay special attention to the chest pain or abnormal heartbeats of patients. In the systematic review, the cardiac damage lead to death in DRESS cases included sudden cardiac arrest, heart transplantation after giant cell myocarditis (GCM), cardiac insufficiencies and cardiogenic shock. When dealing with DRESS patients, ECG, echocardiogram, cardiac enzymes and troponin should be performed, If necessary, perform endomyocardial biopsy as well. Physicians should pay attention to the occurrence of chest pain, sinusoidal, ischemic, depressed injection fraction, giant cell myopathy, and fulminant eosinophilic necrotizing myopathy, which may be associated with a higher possibility of death.

The antibiotics mentioned in this study include minocycline (3/13, 23%), vancomycin (2/13, 15%), amoxicillin and clavulanate, azithromycin, clindamycin, colistimethate, oxacillin, pyrazinamide, sulfasalazine and sultamicillin (1/13, 0.69% each). Previous research (41) has shown that the severity of antibiotic-induced DRESS syndrome compared to other triggers remains a topic of debate, but our research suggests that antibiotic related DRESS has a relatively high mortality rate (50%). Therefore, clinicians should exercise caution when prescribing antibiotics to patients with a history of DRESS syndrome.

So far, rash remained the most frequent clinical findings. The skin disorders in fatal cases were mostly severe, systemic and diffuse, with features that included rash, erythema, maculopapulars, urticarial rash, pustules. The generalized rash could be with or without itching, edema, esquamation, blisters, scabs, erosions. The rash could appear on the trunk, extremities and face, and could started from one part of the body and spread to the entire body. Out of 26 cases, 10 have undergone skin biopsy and all were consistent with DRESS. Six skin biopsy (6/10, 60%) mentioned dermatitis or perivascular inflammatory infiltrate with elevated eosinophils. One skin lesion (1/10, 10%) showed liquefaction degeneration in the epidermis. One skin biopsy (110, 10%) for direct and indirect immunofluorescence revealed linear C3 deposits and IgG deposits, respectively. ELISA testing for BP 180 and 230 was positive, consistent with a diagnosis of bullous pemphigoid. A study retrospective study on 50 skin biopsies from 36 patients with DRESS syndrome, and demonstrated that 28% with typical lymphocytes (most with CD8 expressed) and 6% with T-cell (48). Different types of SCARs may differ in terms of the specific effector T cells involved and the chemokines/cytokines released, resulting in cell homing to skin and specific tissue targeting (33).

Virus reactivation were observed in 4 fatal cases in the systematic review, of which 2 cases referred to HHV reactivation (including HHV-6 and HHV-7), and 2 cases mentioned cytomegalovirus (CMV) reactivation. In addition, there were 5 cases reported as virus infection. A study prospectively assessed 40 patients exhibiting DRESS found that EBV, HHV-6 or HHV-7 reactivation in 76% of the patients (49). In our systematic review, this proportion was lower, that possibly due to two reasons. On the one hand, some cases did not undergo virus testing, and on the other hand, some cases that tested positive for the virus were not determined whether they belonged to reactivation. Virus reactivation is related to the onset of DRESS, because the viral reactivation in DRESS cases is thought to lead to the circulating CD8+ T lymphocytes activation (49). HHV reactivation is potentially linked to increased severity of reactions, with viral reactivation correlating to the level of inflammation (50). A reactivation can be asymptomatic, or it may cause prolonged symptoms in DRESS, long after stopping the causative drugs (50).

Among the drugs mentioned in this article, allopurinol, minocycline, vancomycin, amoxicillin and clavulanate, leflunomide, levetiracetam, nabumetone, sulfasalazine, zonisamide and azithromycin have warning signs for DRESS in their drug instructions, while colistimethate, dolutegravir, lansoprazole, omeprazole, oxacillin, phenoin, pyrazinamide do not have warning signs in their drug instructions. It is necessary to remind patients in the drug labels that serious skin reactions, including dress may develop, and advise patients to stop taking suspect drug immediately if they develop any type of rash or fever. The patients should discontinue the suspected drug immediately if symptoms occur and contact their healthcare provider as soon as possible.

The initial step in managing DRESS involves discontinuing the suspected drug responsible. Systemic corticosteroids continue to be the preferred treatment option. Systemic corticosteroids continue to be the preferred treatment option. However, several retrospective studies indicated that systemic corticosteroids associated with higher risk of DRESS relapse (51), viral reactivation (51, 52), various infections (53), and development of autoimmune diseases (53). Although the role of steroid-sparing therapies in DRESS is still not well-established yet, there is increasing evidence supporting the use of steroid-sparing therapies such as cyclosporine, intravenous immunoglobulin (IVIG), interleukin (IL)-5 axis inhibitors, and Janus kinase (JAK) inhibitors (54). Although biomarkers not measured in the cases reviewed in this article, in light of the ongoing diagnostic challenges associated with DRESS, there is a growing interest in investigating biomarkers. Ogawa’s study revealed a significant increase in serum thymus and activation-regulated chemokine (TARC/CCL17) levels among patients with DRESS compared to those with morbilliform drug eruption and Stevens-Johnson syndrome/Toxic epidermal necrolysis (SJS/TEN) (55). In fact, a prospective case control study demonstrated comparable efficacy between serum TARC levels and a RegiSCAR score (≥2) in distinguishing cases of DRESS from controls with maculopapular drug rash (56). Miyagawa reported that Th2-associated chemokines were markedly upregulated in DIHS/DRESS (57). IFN-γ ELISpot assay (58, 59) and IFN-γ combined with IL-4/IL-2 (56, 60) could offer potential for use as rapid diagnostic tests.

It must be admitted that the study encountered various limitations. Firstly, the FAERS data utilized in our analysis were subject to voluntary reporting by healthcare professionals and consumers, potentially resulting in variations in report quality due to the skills and diligence of the reporters. This could introduce bias into our findings through missing, inadequate, or incomplete reports. Additionally, FAERS lacks comprehensive information on patient morbidity related to drug use, as it does not provide data on the total number of patients using the drug. Furthermore, the small sample size of case reports limited our ability to do risk factor analysis. Lastly, methodological constraints prevented us from establishing a causal relationship between drug and the observed positive signals or reported cases.

To our knowledge, this article is the first study that focus on fatal DRESS cases. Our research indicates a relatively high mortality rate associated with antibiotic-related DRESS. Cardiac involvement is the most common comorbidities among DRESS deaths, and patients taking high-risk medication should be carefully monitoring.