Herein, we report on a 67-year-old Caucasian originating from Portugal admitted to the hospital for fever of unknown origin and asthenia. His medical history was relevant for hypertension, insulin-dependent type two diabetes, and vitiligo and was on aspirin cardio, lisinopril, torasemid, omeprazole, and insulin therapy. On clinical examination, we noticed anasarca with pleural effusion, ascites, and edema of the lower limbs. Laboratory tests revealed microcytic anemia (hemoglobin 100 g/l, MCV 75 fl) with thrombocytopenia (73 G/l), marked elevation of C-reactive protein (204 mg/l), renal insufficiency (creatinine 463 µmol/l), and cholestasis (alkaline phosphatase 300 U/l, gamma-GT 56 U/l), with normal transaminases. Urinalysis showed severe proteinuria (5.29 g/l) and glomerular microhematuria. Immunological evaluation revealed normal IgG, IgM, IgA, and complement (C3, C4) levels, anti-nuclear antibodies (1/640 speckled), anti-SSA, anti-actine, and anti-parietal cell antibodies were positive. The infectious work-up was negative (HIV, hepatitis B/C, CMV, TB spot, and blood culture) except for EBV that was slightly positive by PCR (3,560 copies/ml). A thoraco-abdominal CT scan detected multiple mediastinal, axillary and retroperitoneal adenopathies, pleural and pericardial effusions, hepatosplenomegaly, and ascites. There was no sign of peripheral hemolysis (absence of schsitocytes). Finally, a cytokine profile showed elevation of IL-6, VEGF, soluble IL-2 receptor, and TNF-α, whereas IL-8 was normal.

The etiology remained unclear and the patient underwent several supplementary investigations. An axillary lymph node was surgically removed and showed atreic secondary lymphoid follicles with hyalinized/vascular GCs. Interfollicular areas showed a marked vascular proliferation and contained a small number of CD138-positive plasma cells (Figures 1A–D). Immunohistochemistry for HHV-8 with latency-associated nuclear antigen-1 was negative. The histopathological findings were consistent with the newly defined diagnostic criteria for iMCD (11). The blood smear and the flow cytometry in the blood were normal without evidence for lymphoproliferative disease. A bone marrow trephine biopsy revealed megakaryocytic hyperplasia and mild reticulin myelofibrosis without abnormal lymphocytic infiltration (Figures 1E,F).

Renal biopsy showed (1) thickening of the peripheral basement membrane and the intern lamina rara, (2) mesangial proliferation without glomerular deposition of immunoglobulin, and (3) a band-like pattern of interstitial fibrosis associated with lymphohistiocytic inflammatory cell infiltrates. Malignancies were also actively searched for. Colon biopsies showed a collagenous colitis rich in eosinophils (Figure 2A). Gastric biopsies revealed chronic inflammation and glandular atrophia with intestinal metaplasia of the fundus suggestive of an atrophic gastritis (Figure 2B). Prostatic biopsies showed mild chronic inflammation without neoplasia (Figure 2C). Liver biopsy showed mild sinusoidal distension, unspecific portal and lobular inflammation, and rare epithelioid granulomas without necrosis (Figures 2D,E). Finally, the patient underwent a bronchial biopsy showing unspecific chronic inflammation without granuloma, which was not suspicious for sarcoidosis (Figure 2F).

Together, these clinicopathological findings are compatible with a diagnosis of iMCD (11) and TAFRO syndrome (9).

The best management of iMCD disease remains poorly defined with only one randomized clinical trial published in 2014 on siltuximab compared to best supportive care (12). Thus, this case required multidisciplinary discussion and consensus in 2011 when the diagnosed was made and judgment was based on available data with tocilizumab (13) and rituximab for the treatment of HHV-8-associated MCD (14). Thus, treatment was initiated with high dose methylprednisolone (1 g/day for 3 days) transitioned to prednisone 1 mg/kg for several weeks, one single infusion of tocilizumab (8 mg/kg), and 1 week later rituximab 375 mg/m² (total 4 weekly doses) (Figure 3A). The rational to use methylprednisolone and tocilizumab was to rapidly stop IL-6-mediated inflammation as induction regimen, whereas rituximab was used as maintenance treatment. We observed resolution of the inflammatory syndrome and slow clinical improvement over the course of 3 weeks, thrombocyte values normalized (>100 G/l) after 5 weeks and cholestasis parameters several months later. Corticosteroids were rapidly weaned off because of poor control of the diabetes, and the patient was hemodialyzed for end-stage renal disease. In the following 5 years, the patient relapsed twice, each time about 1 year after rituximab was discontinued (Figure 3A). The first relapse was successfully treated with rituximab alone and the second with methylprednisolone/tocilizumab (2 doses)/rituximab (4 weekly doses 375 mg/m²). Five years after the diagnosis was made (before the second relapse), the PET-CT showed reduction of the lymphadenopathies (Figures 3B,C). Auto-antibodies disappeared over the years and hypogammaglobulinemia developed, probably due to repeated rituximab treatments. Follow-up bone marrow biopsies were not performed in this case.

TAFRO syndrome is a rare clinical syndrome of unknown origin that has been described in patients with iMCD. Most of the reported cases were Japanese patients. To analyze the role of ethnicity, we have reviewed the literature for all known Caucasian cases and have identified eight including the present one. When analyzing major and minor criteria recently established by Masaki et al., we confirmed that Caucasian cases were HIV/HHV-8-negative and meet both major and minor criteria (9, 15–20). In comparison with the Japanese cohort, of 18 patients we find Caucasian TAFRO cases presented more frequently with kidney failure and auto-immune antibodies, although these results need to be interpreted with caution due to the low number of patients (Table 1). Recently, a patient with TAFRO was also reported in Latin America (21) and the members of the Castleman Disease Collaborative Network report treating iMCD patients with TAFRO from diverse ethnic backgrounds throughout the world for decades (11). As to IgG4-related disease primary reported in Japan (22), TAFRO syndrome is a rare entity affecting patients with different ethnic background without major differences in the clinicopathological presentation.

Recently, the international working group for the Castleman Disease Collaborative Network established the first-ever diagnostic criteria for HHV-8-negative-iMCD analyzing 244 patients with iMCD (128 patients from literature, 37 cases submitted by the working group members, and 79 from randomized control study) and more specifically the histopathological features of 85 lymph nodes with presumptive diagnosis of MCD. Importantly, this cohort included TAFRO and non-TAFRO-iMCD cases (11, 23). The authors argued that while there are clinical differences between iMCD patients with TAFRO syndrome and those who do not have TAFRO syndrome, there are a significant number of overlapping histopathological and clinical features. Furthermore, some patients with TAFRO clinical syndrome did not present with the more classical HV histopathology, and some patients with the more classical histopathology for TAFRO syndrome did not have TAFRO clinical features. There was consensus among the international working group that TAFRO syndrome is a clinical subtype of iMCD to be included in the diagnostic criteria, not a separate entity. Nevertheless, a consensus between both the international working group on iMCD and the Japanese working group on TAFRO is warranted to agree on criteria, prevalence, and incidence of TAFRO. We proposed a classification algorithm for MCD (Figure 4). An exhaustive list of major, minor, and exclusion criteria can been found elsewhere (11). More importantly, genomic, transcriptomic, and proteomic studies are needed to understand molecular differences that may exist. The clinical utility of subtyping the lymph node histological features into HV, plasmacytic, or mixed type also warrants further investigation.

The relationship between TAFRO/iMCD and auto-immunity is poorly understood. In our study, auto-antibodies such as anti-nuclear or anti-SS-A (SS-A) were found in all six patients with available data (Table 1). Albeit auto-antibodies alone are insufficient to make a definite diagnosis of connective tissue disease such as systemic lupus erythematosus (SLE), Sjögren syndrome, systemic sclerosis, or dermatomyositis/polymyositis, the classification criteria to diagnose undifferentiated connective tissue disease can be met (24). Furthermore, generalized lymphadenopathies are a common finding in about a quarter of patients with SLE (25, 26). When analyzing the histopathological findings in lymph node biopsies from 19 patients with active SLE, iMCD-compatible lesions were present in 5; 3 of mixed type and 2 of HV type (27). As to anti-SS-A52 auto-antibodies, they seem to be more prevalent in patients with TAFRO as illustrated by this study and reported by others (19, 20, 28). Interestingly, undiagnosed anti-SS-A 52 positive patients do not frequently develop definitive connective tissue disease (29). Finally, the chemokine CXCL10 (IP-10) involved in the pathogenesis of many auto-immune disease including Sjögren syndrome (30) seems to be increased during the flares-up in TAFRO (31). In conclusion, these data suggest that auto-immunity is an important aspect of TAFRO.

Kidney involvement is frequently observed in patients with TAFRO syndrome, the underlying mechanisms, however, are poorly understood. Depending on the authors, the R in TAFRO may refer to renal dysfunction or reticulin myelofibrosis (9, 20). Iwaki et al.’s recently proposed diagnostic criteria for TAFRO that do not include renal failure as major or minor criteria (20). Importantly, all Caucasian TAFRO patients who we have found had renal failure (Table 1). The clinical presentation of iMCD patients with renal failure is frequently associated with arterial hypertension, glomerular hematuria, and proteinuria. The most frequent lesions were endotheliosis of the small-vessel and glomerular double contours with glomerular/arteriolar thrombi, a pattern similar to what can be found in thrombotic microangioapthy (TMA) (32, 33). Of interest, 2/7 patients with TAFRO had renal insufficiency and lesions compatible with MAT. Furthermore, complement, an important trigger for MAT (34), can also be frequently decreased in TAFRO (Table 1). Thus, these results stress the importance of regularly monitoring kidney function in patients with TAFRO syndrome particularly in the presence of other risk factors such as diabetes as was the case for our patient. Further studies into the link between TAFRO, complement dysregulation, and MAT are needed.

TAFRO syndrome is a rare subtype of iMCD resulting in generalized organ inflammation of unknown origin. Diagnosis remains very challenging and the exclusion of infectious, auto-immune, and neoplastic disorders is necessary. Patients undergo a series of clinical investigations and biopsies, which are often non-contributive. In this article, we demonstrate a Caucasian case of TAFRO syndrome and lend further support to the notion that TAFRO syndrome can be seen in iMCD patients around the world. Genomic, transcriptomic, and proteomic investigations into the etiology, pathogenesis, relationship between auto-immunity, complement dysregulation, and kidney failure, and molecular differences between TAFRO-iMCD and non-TAFRO-iMCD are warranted.