Evaluation of macrophage activation syndrome in hospitalised patients with Kikuchi-Fujimoto disease based on the 2016 EULAR/ACR/PRINTO classification criteria

Background To evaluate the impact of macrophage activation syndrome (MAS) on clinical features in patients with Kikuchi-Fujimoto disease (KFD) and to compare the features of MAS in KFD with those of adult-onset Still’s disease (AOSD) and systemic lupus erythematosus (SLE). Methods The medical records of febrile patients hospitalised with KFD between November 2005 and April 2017 were reviewed. Patients fulfilling the 2016 classification criteria for MAS were classified as having MAS. Clinical and laboratory features of patients with KFD with and without MAS were evaluated. Poor hospitalisation outcomes were defined as intensive care unit admission or in-hospital mortality. The treatment outcomes of MAS in KFD, AOSD, and SLE were also compared. Results Among 78 patients hospitalised with KFD, 24 (30.8%) patients had MAS during admission. Patients with KFD and MAS more frequently required glucocorticoid treatment (66.7% vs 40.7%, p = 0.036) and had longer hospital stays than patients with KFD without MAS (12.5 vs 8.5 days, p<0.001). In addition, patients with MAS had worse hospitalisation outcomes than patients without MAS (29.2% vs. 0.0%, p<0.001). Among patients with MAS in KFD, AOSD, and SLE, the number of patients requiring glucocorticoid treatment after 3 months was significantly lower among patients with MAS and KFD (KFD 33.3%, AOSD 88.9%, SLE 100%, p<0.001). Conclusions The presence of MAS in KFD was associated with adverse clinical outcomes including higher steroid usage and worse hospitalisation outcomes. However, compared to those with AOSD and SLE, patients with MAS and KFD were less likely to require long-term glucocorticoid treatment.


Introduction
Health System located in Shinchon and Gangnam district of Seoul, Korea. The inclusion criteria were as follows: (i) patients who were admitted with fever; (ii) patients with diagnosis of KFD that was confirmed pathologically; (iii) patients with ferritin level assessment; (iv) patients without concurrent autoimmune disease/infection. Finally, 78 patients were included in this study. The flowchart for patient selection is shown in S1 Fig. The clinical and laboratory data of patients with MAS in AOSD and SLE according to the 2016 classification criteria were obtained using a pre-existing dataset used in a previous study [25,26], and all data were fully anonymized for analysis. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Severance Hospital (IRB approval number 4-2018-0624). The requirement to obtain informed consent was waived because of the retrospective nature of the study.

Definition of patients with MAS according to the 2016 classification criteria
According to the 2016 EULAR/ACR/PRINTO classification criteria for MAS [23,24], we classified patients as having MAS when they presented with fever, a ferritin level of �684 ng/mL, and fulfilled more than 2 of the following 4 criteria: platelet count �181,000/mL, aspartate aminotransferase (AST) level >48 units/L, triglyceride level >156 mg/dL, and fibrinogen level �360 mg/dL. In addition, patients with MAS were divided into two groups: patients having MAS on admission and those who developed MAS after admission.

Collection of clinical and laboratory variables
The clinical data collected included age, sex, length of hospital stay, in-hospital mortality, intensive care unit admission, and treatment. Poor hospitalisation outcomes were defined as intensive care unit admission or in-hospital mortality. The clinical manifestations of KFD included the presence of hepatomegaly, splenomegaly, skin rash, and joint pain. Baseline laboratory variables included white blood cells, neutrophils, lymphocyte and platelet counts, erythrocyte sedimentation rate, and levels of haemoglobin, C-reactive protein (CRP), creatinine, AST, alanine aminotransferase (ALT), albumin, ferritin, lactate dehydrogenase (LDH), fibrinogen, and triglycerides. As fibrinogen and triglyceride levels were not measured regularly, the levels obtained within 3 days of the date of laboratory data selection were included in the analysis.

Calculation of the hemophagocytic syndrome score (HScore) and patients fulfilling the HLH-2004 criteria
The HScore was developed to evaluate individual risk for the presence of hemophagocytic syndrome [27] and is also useful to identify the risk of having reactive hemophagocytic syndrome in patients with rheumatic diseases [28]. The HScore and the probability of having hemophagocytic syndrome were calculated using an online calculator (saintantoine.aphp.fr/score). Furthermore, the proportion of patients fulfilling the HLH-2004 criteria was estimated as previously described [29].

Statistical analysis
Continuous variables were presented as medians with interquartile ranges (IQR), and categorical variables were presented as frequencies and percentages. Continuous variables were compared using the Kruskal-Wallis test or Mann-Whitney U test, and categorical variables were compared using the chi-square or Fisher's exact test, as appropriate. Kaplan-Meier analysis and the log-rank test were used to compare the survival probability among the groups. All statistical analyses were performed using MedCalc statistical software version 18.6 (MedCalc Software, Ostend, Belgium) or GraphPad Prism software version 5.0 (GraphPad Software, San Diego, California, USA). In all statistical analyses, a two-tailed p<0.05 was considered statistically significant.

Baseline characteristics of patients with and without MAS
Of the 78 patients included in the study, 24 (30.8%) were classified as having MAS during admission (Table 1). Among the 24 patients, 20 (83.3%) had MAS on admission, and 4 (16.7%) developed MAS after admission. Although the clinical manifestations were not different between patients with and without MAS, patients with MAS were older and had a longer hospital stays, and the proportion of patients with poor hospitalisation outcomes was significantly higher in patients with MAS (29.2% vs. 0.0%, p<0.001). In addition, the median HScore (97.0 [IQR 68.0-128.5] vs 49.0 [IQR 33.0-72.0]) and the probability of having hemophagocytic syndrome were higher in patients with MAS than in patients without MAS (all p<0.001) (Fig 1A and 1B). Furthermore, the proportion of patients requiring glucocorticoid treatment was significantly higher in patients with MAS (66.7% vs 40.7%, p = 0.036). Regarding laboratory findings, patients with MAS had higher CRP, ALT, and LDH levels and a lower lymphocyte count and albumin levels than those without MAS. Among the laboratory variables selected for MAS classification, higher AST and ferritin levels and lower platelet counts were observed in patients with MAS; however, the differences between triglyceride and fibrinogen levels were not significant ( Table 2). Of the 19/78 (24.4%) patients who had undergone a bone marrow study, only 2 patients in the MAS group were found to have hemophagocytosis.

Clinical and laboratory findings of patients with KFD and in-hospital mortality
As MAS in KFD is not a well-described complication, we further analysed the clinical and laboratory features of patients with KFD and in-hospital mortality ( Table 3). All patients in our study who died in the hospital had MAS, were older than 50 years of age and did not have any  specific clinical manifestations. Among MAS patients with KFD (n = 24), patients aged �50 years had a higher mortality rate than those aged <50 years (p = 0.003) (Fig 2). During the disease course, patients with MAS had increasing ferritin and AST levels, while platelet counts decreased. Most deaths occurred within 90 days, and among the five patients who died, the cause of death was multi-organ failure in 4 patients and sepsis in 1 patient (Table 3). Hepatosplenomegaly was only observed in one patient, and no patients had co-existing comorbidities with the exception of gout in one patient.

Clinical and laboratory features among patients with MAS in KFD, AOSD, and SLE
Next, we compared the clinical and laboratory features among patients with MAS in KFD, AOSD and SLE. Clinical and laboratory data of patients with AOSD or SLE and MAS used for comparison were obtained from the dataset of previous studies [25,26]. Patients with AOSD were older than those with SLE, and the proportion of female patients was the lowest in patients with KFD. The presence of lymphadenopathy was significantly higher in patients with KFD than in those with AOSD or SLE (Table 4). There were no differences in the median HScore ( Table 4).

Discussion
While KFD is pathologically characterised by a proliferation of activated cytotoxic T lymphocytes and histiocytes with central necrotic areas in lymph nodes, HLH is characterised by hemophagocytosis, engulfing erythrocytes, and leukocytes by histiocyte in the bone marrow or  Data expressed as median (interquartile range) or n (%). † Glucocorticoid treatment and the dosage of glucocorticoids in patients with mortality or who were lost to follow-up were counted on the last date of follow-up. ‡ Includes treatment with disease modifying anti-rheumatic drugs, biologics, cyclosporine, and etoposide/cyclophosphamide. liver. Therefore, both KFD and HLH are associated with abnormal histiocyte activation. In addition, inflammatory cytokines such as tumor necrosis factor (TNF)-α, interleukin (IL)-1, and IL-6 are also implicated in the pathogenesis of both MAS and severe cases of KFD [30,31]. Therefore, although it is possible that HLH can be accompanied by KFD, the association between KFD and HLH is not well understood. In this study, by using the 2016 EULAR/ACR/ PRINTO classification criteria, we found the incidence of MAS, a secondary form of HLH, was 30.8% in our study population of pathologically confirmed KFD. Because we included only febrile hospitalised patients with KFD, our study population might have been skewed towards those having severe laboratory abnormalities in tertiary referral centres; however, the association of MAS in KFD is an interesting finding that requires clinical attention.
KFD patients with MAS experienced longer hospital stays and worse hospitalisation outcomes, including higher rates of intensive care unit stays and in-hospital mortality than in patients without MAS. In addition, the proportion of patients with KFD requiring glucocorticoid treatment was significantly higher in patients with MAS than in patients without MAS. All 5 cases of mortality in our cohort were observed in patients with MAS who were over the age of 50. Compared to patients aged <50 years, those aged 50 years and older had significantly higher mortality in KFD with MAS. These findings suggest that older KFD patients with MAS might be especially associated with poor prognosis.
MAS presents as non-remitting fever, generalised lymphadenopathy, hepatosplenomegaly, pancytopenia, high ferritin, and low fibrinogen levels [32]. Since there are no disease-specific clinical manifestations or laboratory abnormalities of MAS, and many of the symptoms resemble those of KFD itself, the clinician's suspicion is critical for the diagnosis. In addition, even though hyperferritinemia is a prerequisite in the 2016 criteria for MAS, alternative diagnoses which can be associated with elevated ferritin levels such as sepsis and catastrophic antiphospholipid syndrome should be excluded [33]. In our study, the difference in clinical manifestations between patients with KFD with and without MAS was not significant. On the other hand, in terms of laboratory findings, patients with KFD and MAS had higher CRP, LDH, ferritin, AST, and ALT levels and lower platelet counts, lymphocyte count, and albumin levels than patients with KFD without MAS. Importantly, in our study, a higher proportion of patients with KFD and MAS were treated with steroids than patients without MAS. Monitoring for MAS in patients with KFD could be clinically important, as patients without MAS could be treated with supportive care, while those with MAS might require steroids or additional second-line treatments such as cyclosporine, anti-cytokine therapy, or chemotherapy [34].
Comparison of clinical and laboratory features between patients with MAS-associated KFD, AOSD, and SLE showed that they had different characteristics. Notably, a higher proportion of patients with KFD were able to stop steroids within 3 months compared to patients with SLE or AOSD. Additionally, fewer immunosuppressive agents were administered to patients with MAS and KFD than to patients with MAS and AOSD or SLE. Hence, the clinical course of MAS-associated KFD is milder than MAS with SLE or AOSD, suggesting that the treatment of MAS should be tailored to the underlying diseases.
In our data, 50% (12/24) of KFD patients with MAS received bone marrow evaluation and only 16.7% (2/12) patients were found to have hemophagocytosis. Although this is a small number of patients in a retrospective study, the incidence of hemophagocytosis in our KFD patients were lower than other studies with autoimmune diseases. In a recent review paper of MAS with autoimmune diseases, 42%-100% were positive for bone marrow hemophagocytosis in patients with SLE, Kawasaki disease, and sJIA [35]. In our recent MAS study with AOSD patients, 50% was positive for hemophagocytosis in bone marrow evaluation [26]. It is possible that our KFD patients might have received bone marrow biopsy in the early course which often makes the result negative [21]. It is also possible that macrophage activation in lymph nodes might be enough to drive MAS in KFD. Although in the autoimmune area, bone marrow analysis is not always required for diagnosis of MAS, it is of interest to validate our finding in future studies.
The strength of this study is that, to our knowledge, it is the first study to investigate the incidence and clinical outcomes of MAS in patients with KFD. We only included patients whose diagnoses were confirmed pathologically, further validating our results. Since KFD is prevalent in young Asian women, we have a large population of patients with KFD at our institution, enabling recruitment of a statistically significant number of patients for this study. However, our study also has several limitations. First, it is a retrospective study, and the data were collected by reviewing the hospital's medical records. Second, there were patients who did not have ferritin, fibrinogen, and triglyceride level results; this might have influenced patient selection and classification resulting in selection bias because clinically severe patients may undergo more blood tests. Therefore, the incidence of MAS in our study might be relevant in patients with KFD and severe clinical or laboratory features. Third, data regarding proinflammatory cytokines (TNF-α, interferon-γ, IL-1, IL-6, and soluble CD163) associated with MAS were not obtainable because the following tests were not available in our hospitals [36,37]. Because these cytokines are known to play an important role in the pathogenesis of MAS [30,31], assessment of such tests in future studies would provide additional significance. Fourth, because the 2016 classification criteria for MAS have only been validated in patients with sJIA, caution should be made when applying the criteria to non-sJIA diseases. HLH-2004 criteria may be more suitable for non-sJIA diseases. Because a majority of patients had not undergone tests of bone marrow study, natural killer cell activity, and soluble CD25, we could not validate our finding using HLH-2004 criteria.
In conclusion, our findings suggest that the application of the 2016 EULAR/ACR/PRINTO criteria for MAS in hospitalised patients with KFD can be useful in detecting MAS. In addition, although the clinical and laboratory findings were different based on the underlying autoimmune disease, the presence of MAS based on the 2016 EULAR/ACR/PRINTO criteria was associated with adverse patient prognosis. Especially older patients with KFD who presents with features of MAS may be associated with a high mortality rate and may need special attention. Therefore, monitoring patients with KFD according to the 2016 criteria for MAS might aid in predicting patient prognosis and the stratification of high-risk patients.