Drug-induced hypersensitivity syndrome/drug reaction with eosinophilia and systemic symptoms (DRESS) is a drug-induced, adverse T-cell–mediated hypersensitivity reaction that most often involves skin. The pathologic findings of DRESS-related lymphadenopathy have been described infrequently in the literature.
To present a case series of DRESS-related lymphadenopathy with an emphasis on the morphologic spectrum.
We describe detailed clinical and pathologic findings along with the literature review. We focus on the differential diagnosis between DRESS lymphadenopathy and angioimmunoblastic T-cell lymphoma (AITL).
There were 4 men and 1 woman with a mean age of 41 years (range, 23–59 years). One patient (20%) died. Three lymph node biopsy specimens showed a pattern reminiscent of AITL (AITL-like pattern) and 2 cases showed necrotizing lymphadenitis (Kikuchi-like pattern), associated with vasculitis in 1 case. The AITL-like morphology of DRESS-related lymphadenopathy may be difficult to distinguish from genuine AITL. The clinical information is important for differential diagnosis, including history of drug exposure, age, and the rarity or absence of AITL-associated manifestations such as hemolytic anemia and hypergammaglobulinemia. Molecular analysis of the T-cell receptor genes is helpful, typically revealing a polyclonal pattern in DRESS-related lymphadenopathy.
In the literature, 4 histologic patterns of DRESS lymphadenopathy have been described: reactive lymphoid hyperplasia, necrotizing lymphadenitis, Hodgkin lymphoma–like, and AITL-like. These patterns, particularly those that resemble lymphoma, highlight the importance of correct diagnosis to avoid unnecessary therapies.
Drug-induced hypersensitivity syndrome (DIHS)/drug reaction with eosinophilia and systemic symptoms (DRESS) syndrome is a drug-induced severe adverse reaction that most often involves the skin. Drug-induced hypersensitivity syndrome/DRESS (subsequently referred to as DRESS) can be life-threatening, with cutaneous and internal organ involvement and a mortality rate of about 10%.1 Because cases may be underrecognized because of the variable clinical and laboratory features, the incidence of DRESS remains unclear, with an estimated overall population risk of at least 10 cases per million per year.2
Historically, DRESS was observed initially in patients receiving anticonvulsants in the 1930s.3 Subsequently, a case of fever, exfoliative dermatitis, and hepatitis after taking phenytoin was reported in 1950 and was referred to as Dilantin hypersensitivity.4 In 1996, Bocquet et al5 proposed the term DRESS, along with a concise description of the syndrome. Reactivation of human herpesvirus (HHV)–6 was identified in patients with DRESS syndrome6 in 1997.
The reported incidence of lymphadenopathy in patients with DRESS syndrome ranges from 31% to 88%.7 However, papers describing the pathologic findings of lymph nodes in patients with DRESS are limited in the literature. Importantly, the morphologic findings in DRESS syndrome can be florid and resemble, in part, lymphomas.
Here, we describe a case series of 5 patients diagnosed with DRESS-related lymphadenopathy and present a literature review. We provide detailed clinicopathologic features of these patients and describe 3 patients who had morphologic features reminiscent of angioimmunoblastic T-cell lymphoma (AITL), which we refer to as AITL-like pattern. We also discuss the differential diagnosis between AITL-like DRESS lymphadenopathy and genuine AITL.
MATERIALS AND METHODS
We enrolled 5 cases of DRESS syndrome from the archives of National Cheng Kung University Hospital, Tainan, Taiwan (cases 1, 2, and 4) and consultation files of 1 author (cases 3 and 5). The diagnosis of DRESS was based on the diagnostic criteria for DRESS developed by the Registry of Severe Cutaneous Adverse Reaction group or the diagnostic criteria for DIHS/DRESS established by a Japanese consensus group.2 This study was approved by the institutional review board (A-ER-110-037 and CE21178B) and was in accord with the Helsinki Declaration of 1975, as revised in 2013.
Histologic and Immunohistochemical Stains
Hematoxylin-eosin–stained tissue sections were cut and prepared from formalin-fixed, paraffin-embedded tissue blocks. Immunohistochemical analysis was performed using 4-μm-thick tissue sections deparaffinized with xylene. The procedures were performed using an automated immunohistochemistry stainer (BenchMark XT, Ventana Medical Systems, Inc, Tucson, Arizona). The primary antibodies and working dilutions are listed in the Supplemental Table (see supplemental digital content at https://meridian.allenpress.com/aplm in the September 2022 table of contents). Appropriate positive and negative controls were used.
In Situ Hybridization for Epstein-Barr Virus–Encoded RNA
In situ hybridization was performed to detect Epstein-Barr virus–encoded small RNA (EBER) using a polymerase chain reaction (PCR)–derived digoxigenin-labeled DNA probe.8,9 The test for nucleotide integrity was performed by using an RNA positive control probe (Ventana Medical Systems, Inc). The intended target was the polyA tail in messenger RNA found in nuclei. The EBER in situ hybridization was deemed positive when there were more than 10 EBER+ nuclei per 0.5 cm2 (>0.2/mm2).10,11
Sample Preparation and DNA Extraction
Genomic DNA was extracted from 10-μm-thick sections prepared from formalin-fixed, paraffin-embedded tissue blocks and was purified using a kit (DNeasy Blood & Tissue Kit, Qiagen Inc, Valencia, California). The tissue sections were washed first with phosphate-buffered saline (pH 7.2) to remove fixative and then deparaffinized with xylene and washed twice with ethanol. After extraction, DNA was stored at −20°C until analysis.
T-Cell Clonality Analysis by PCR
Multiplex PCR protocols and primers were used for analysis of TRB and TRG following standardized BIOMED-2 protocols, as described previously.12 DNA quality of the specimens was assessed using the multiplex PCR reaction for internal control DNAs. A T-cell lymphoma was used as a positive control and water was used as a negative control.
Detection of HHV-6 by PCR
We assessed for HHV-6 viral DNA in lymph node tissues using a PCR-based method with a commercial kit for HHV-6 detection (TIB Molbiol, Berlin, Germany) following the manufacturer's instructions. A 272-bp fragment of the antigenic virion protein 101K (U11) gene of the HHV-6 genome was amplified with specific primers. The resulting PCR fragment was analyzed with hybridization probes and detected in channel 640.
The study group was composed of 5 patients, 4 men and 1 woman, with ages ranging from 23 to 59 years (mean, 41 years). All patients were hospitalized and had fever, enlarged lymph nodes, eosinophilia, and organ involvement mainly with liver function abnormalities. Four of the 5 patients (80%) had a skin rash and elevated serum levels of lactate dehydrogenase. Patient 5, although he did not have a skin rash, still fulfilled the criteria of atypical DIHS/DRESS established by a Japanese consensus group.13 Three patients (60%) had leukocytosis and atypical lymphocytosis in the peripheral blood. Serology tests for anti-nuclear antibody and anti-neutrophil cytoplasmic antibody were carried out in 4 and 2 cases, respectively, and all were negative. Patients 1 and 2 underwent skin biopsy, which showed pathologic features suggestive of DRESS: one was purpuric superficial perivascular dermatitis with scattered eosinophils, and the other was purpuric interface dermatitis, compatible with drug eruption.14 The clinical features, including duration of illness, are summarized in Table 1. One patient (case 2) died of septic shock. Blood cultures prior to death yielded Klebsiella pneumoniae and Elizabethkingia meningoseptica.
Three lymph node biopsy specimens (cases 1–3) showed a pattern reminiscent of AITL (AITL-like pattern) and 2 cases showed necrotizing lymphadenitis (Kikuchi-like pattern), associated with vasculitis in 1 case (case 5). The morphologic features, including results of special studies, are summarized in Table 2.
The AITL-like pattern (n = 3) revealed nearly total effacement of nodal architecture by a dense polymorphic infiltrate with pale appearance (Figure 1, A, case 1; B, case 2; and C, case 3). The infiltrate was composed of small and medium-sized lymphoid cells and eosinophils admixed with some immunoblasts in a background rich in high-endothelial venules, which showed an arborizing pattern (Figure 1, D). Some clear cells infiltrating around blood vessels were noted in 2 cases. The infiltrating lymphocytes including clear cells showed mild nuclear atypia (Figure 1, E). Immunohistochemical analysis showed increased T cells in interfollicular areas (Figure 2, A) with minimal CD20-positive B cells (Figure 2, B). There was T-zone expansion but no T-cell nodules. The atypical lymphocytes were positive for CD2, CD3, CD7, and T-cell receptor βF1 and were negative for CD10, BCL6, CD20, and PD-1. The CD4:CD8 ratio was 5:1 to 6:1 in the T-cell infiltrates (Figure 2, C and D). Expansion of follicular dendritic cell meshwork as demonstrated by CD21 or CD23 immunostaining was absent in all 3 cases (Figure 2, E). The CD30 antibody highlighted some immunoblasts in B-cell areas in 2 cases (Figure 2, F), and the antibody specific for CD68 highlighted histiocytes including some with hemophagocytosis in case 1. Bone marrow biopsy was performed in case 2 and revealed lymphocytosis characterized by patchy infiltration by small lymphocytes positive for CD3 and negative for CD20. CD4-positive cells outnumbered CD8-positive cells by 3:1 to 4:1 (Supplemental Figure). No lymphoid aggregates were found, and EBER in situ hybridization was negative. Surprisingly, granulomatous inflammation was noted and highlighted by the CD68 stain; acid-fast and Gomori methenamine silver stains were both negative.
The Kikuchi-like pattern (n = 2) showed interfollicular expansion by several foci of necrosis composed of karyorrhectic debris and fibrin deposits surrounded by histiocytes (case 4; Figure 3, A). There was also infiltration of small and large lymphocytes, histiocytes, and some apoptotic cells indicative of plasmacytoid dendritic cells (case 4; Figure 3, B). Plasma cells and eosinophils were inconspicuous. Residual lymphoid follicles were also present. Immunohistochemical analysis showed that the interfollicular infiltrate was expanded by CD3-positive T cells, and aggregates of plasmacytoid dendritic cells were highlighted by the antibody specific for CD123 (Figure 3, C). In case 5, vascular hyperplasia was additionally noted in the interfollicular region (Figure 3, D). These blood vessels around the necrotic patches showed fibrinoid material and nuclear dust suggestive of vasculitis (Figure 3, E). In addition, hemophagocytosis in sinusoids was discerned. By immunohistochemistry, the infiltrating cells were composed predominantly of CD3-positive cells and a small subset of CD20-positive cells. Among the T cells, CD4-positive cells outnumbered CD8-positive cells. Around 10% to 20% of CD4-positive T cells expressed CD25 (Figure 3, F). Interspersed histiocytes and activated larger cells were highlighted by CD68 and CD30 stains, respectively. Immunostaining for herpes simplex virus–1/2 and HHV-8 was negative.
In all 5 cases, in situ hybridization for EBER was performed. In 1 patient specimen (case 3), scattered EBER-positive cells were identified. Molecular studies to assess the T-cell receptor genes were negative in all 4 cases assessed (cases 1–3 and 5). All 5 cases were negative for HHV-6 DNA by PCR on lymph node specimens.
In summary, all 5 cases we report fulfilled the Registry of Severe Cutaneous Adverse Reaction/Japanese DIHS criteria for DRESS/DIHS.2,15 In the literature, 10 cases of DRESS-related lymphadenopathy with mention of lymph node morphology have been described (Table 3). Patient age ranged from 16 to 58 years (mean, 34 years), and the male to female ratio was 3:7. The culprit drugs and clinical data, including leukocyte counts, serum liver function enzymes, and lactate dehydrogenase levels, are provided (Table 3). Polymerase chain reaction analysis to detect HHV-6 in lymph node tissues was performed in 4 cases, and 3 were positive. The patient with negative PCR results had an elevated serum titer for HHV-6 immunoglobulin G antibody. Morphologic evaluation of these cases showed 6 with features reminiscent of T-cell lymphoma (n = 5) or atypical T-cell proliferation (n = 1). Other lymph node specimens showed Hodgkin lymphoma–like features (n = 2), necrotizing lymphadenitis (n = 1), reactive lymphoid hyperplasia (n = 1) and dermatopathic lymphadenopathy (n = 1). Clinical follow-up of these patients showed that 1 patient died and the overall mortality rate was 10%.
Although the incidence of DRESS syndrome is not uncommon,2 the pathologic features of DRESS-associated lymphadenopathy have been mentioned rarely in the literature. In Table 3, we summarize the morphologic patterns described in previous cases and the 5 cases reported here.16–25 Over half of all cases reported previously had T-cell lymphoma–like morphologic features, perhaps because of the selection bias of accepted reports. The most frequent morphologic features of DRESS-associated lymphadenopathy, mimicking T-cell lymphoma, include marked expansion of the paracortex with distortion of lymph node architecture and the presence of a mixed inflammatory background with eosinophils.20–25 Atypical immunoblasts with Reed-Sternberg–like cells were present in 2 cases.19,20 Decreased expression of pan–T-cell antigens, including CD3 and CD7, has been reported in some cases.25 These features may raise concerns or further suggest a diagnosis of lymphoma, especially AITL.
Two of the 5 cases we report showed morphologic features similar to necrotizing lymphadenitis (Kikuchi disease–like); 1 of these biopsy specimens also showed vasculitis. The lymph node architecture was relatively preserved with patches of necrosis and nuclear dust. Sinusoidal areas were dilated by histiocytes showing hemophagocytosis. No evidence of infection or granuloma formation was present. Although drug-induced or drug-associated vasculitis usually attacks the skin and sometimes lungs and kidneys,26,27 necrotizing lymphadenitis with vasculitis has not been reported previously in DRESS lymphadenopathy. Further studies with more case series are warranted to clarify the significance of necrotizing lymphadenitis with vasculitis in DRESS lymphadenopathy. The other 3 cases we report showed AITL-like morphologic features. These lymph node specimens showed diffuse lymphoid proliferations with effacement of the nodal architecture. There were dense and massive infiltrates of polymorphic and atypical lymphocytes around hyperplastic high-endothelial venules. Some atypical cells with clear cytoplasm were evident.
The 3 cases with AITL-like features we report highlight that the differential diagnosis of true AITL versus DRESS lymphadenopathy can be highly challenging. The neoplastic cells of AITL typically express markers of follicular helper T cells such as CD10, BCL6, PD-1, ICOS, CXCL13, and CXCR5. In situ hybridization for EBER commonly shows many Epstein-Barr virus–positive immunoblasts of B-cell lineage,28 and usually clonal T-cell receptor gene rearrangements and RHOA mutations are present.9,29 In contrast, DRESS-related lymphadenopathy is mediated by expansion of regulatory T cells.22,30,31 Regulatory T cells are a subpopulation of CD4-positive T cells characterized by overexpression of FOXP3 and CD25.30,32 Lymphadenopathy related to DRESS is frequently negative for Epstein-Barr virus–positive cells by in situ hybridization and there is no evidence of clonal rearrangement of the T-cell receptor genes. Clinical information is also important for this differential diagnosis, including history of drug exposure (Naranjo score), age, and the absence of other AITL-associated symptoms such as hemolytic anemia and hypergammaglobulinemia.33,34 Finally, molecular analysis of the T-cell receptor genes rearrangement will reveal a polyclonal pattern in DRESS-related lymphadenopathy.
The diagnostic criteria of DRESS syndrome are skin rash, eosinophilia, atypical lymphocytes, liver abnormalities, and lymphadenopathy. Hence, the clinical differential diagnosis may include connective tissue disorders, idiopathic hypereosinophilia, viral hepatitis, Churg-Strauss syndrome, and Kimura disease.35,36 A detailed clinical history, including anti-nuclear antibody and anti-neutrophil cytoplasmic antibody serology tests, would be very helpful to reach a correct diagnosis of DRESS syndrome and exclude other entities such as idiopathic hypereosinophilia, viral hepatitis, and Kimura disease. In addition, the pathologic features of lymph node biopsy in Churg-Strauss syndrome and Kimura disease are also characteristic. The reported pathologic features of nodal Churg-Strauss syndrome are broad, including lymphoid and plasmacytic hyperplasia, hyperplasia with eosinophilia, and hyperplasia with allergic granulomas.36 Although the latter is the pathognomonic feature of Churg-Strauss syndrome, lymphoid hyperplasia with eosinophilia may be found in drug-induced lymphadenopathy.36 In these cases, clinical history is of paramount importance for differentiation. Kimura disease typically presents with eosinophilia and multiple lymphadenopathies predominantly in the head and neck region, which may be one of the differential diagnoses of DRESS syndrome clinically. However, pathologically, Kimura disease is characterized by follicular and interfollicular hyperplasia accompanied with eosinophils and eosinophilic microabscesses within the germinal centers.35 These findings are distinct from those found in DRESS lymphadenopathy. On the other hand, the Kikuchi-like morphology in our 2 cases may raise the possibility of lupus lymphadenopathy. In addition to a negative anti-nuclear antibody test and other clinical manifestations, it has been recently found that lupus lymphadenopathy shows a higher frequency of plasma cell infiltration and positive C4d endothelial staining in the necrotic area.37 Accordingly, the lymph nodes in cases 4 and 5 showed inconspicuous plasmacytic infiltration and were negative for C4d endothelial staining.
In this study, all 5 patients presented initially with fever, lymphadenopathy, and laboratory abnormalities, fulfilling the diagnostic criteria of DRESS.2,15 However, HHV-6 was not detected in any of the 5 lymph nodes in this study, in contrast to 3 of 5 cases positive in previous reports (Table 3). The cause for this discrepancy is unclear. So far, there is no other report from Taiwan regarding the prevalence of HHV-6 in DRESS lymph nodes, but one study from Taiwan showed HHV-6 DNA can be detected in 1 of 19 serum samples (5.3%) from DRESS patients.38 A similar low HHV-6 detection rate (1 of 24; 4.2%) has been reported from the serum of patients with DRESS in the United States.39 Interestingly, a higher positive rate of HHV-6 reactivation (30%–50%) can be observed in whole blood by PCR methods or by assessment of serologic antibody titers—that is, at least 4-fold elevations of anti–HHV-6 antibodies.40 It appears that the discrepancy of HHV-6 detection rates among different reports may be attributable to differences in patient populations, timing for detecting viral reactivation (within versus beyond 4 weeks after the initial diagnosis), and, most importantly, detection methods with samples used (serology versus PCR and serum versus whole blood).39,40 On the contrary, given that expansion of regulatory T cells is a key mechanism in the acute phase of DRESS syndrome, which subsequently leads to reactivation of herpesviruses,13 other viruses than HHV-6 may alternatively play a role. Accordingly, we have detected Epstein-Barr virus genome in 1 case of nodal tissue by in situ hybridization.
Although DRESS syndrome is uncommon in its incidence, this syndrome is associated with significant morbidity and mortality. The most important management is immediate discontinuation of the culprit medication.41 A timely diagnosis and prompt treatment significantly affects the prognosis of patients with DRESS syndrome. Further treatment with systemic corticosteroids or other immunosuppressants may be required at the acute phase.2 At the later phase, preventing the subsequent development of infection such as sepsis, cytomegalovirus, and autoimmune diseases is mandatory.2 Although the prognosis of patients with DRESS syndrome is variable and unpredictable, most patients recover completely after drug withdrawal and appropriate therapy. The overall mortality rate of patients with DRESS syndrome is approximately 10%, primarily from visceral organ complications.41
In conclusion, DRESS-associated lymphadenopathy is an uncommon biopsy specimen encountered by pathologists that can be diagnostically challenging, as these lymph nodes can frequently show morphologic features closely resembling AITL. A high index of suspicion and knowledge of the clinical history are essential for diagnosis. Ancillary studies are also very helpful in this differential diagnosis. Unlike true AITL, lymph nodes involved by DRESS syndrome usually have expanded regulatory T cells and lack EBER-positive B cells or clonal T-cell receptor gene rearrangements.
Supplemental digital content is available for this article at https://meridian.allenpress.com/aplm in the September 2022 table of contents.
This study was supported by grants from the Ministry of Science and Technology, Taiwan (MOST 109-2320-B-006-045-MY3), to Chang.
The authors have no relevant financial interest in the products or companies described in this article.