Risk factor analysis for bone marrow histiocytic hyperplasia with hemophagocytosis: an autopsy study

HPS, which is characterized by the proliferation of activated macrophages with hemophagocytosis, is a rare syndrome [4, 6, 7, 20]. In contrast, HHH was detected in 41.4 % of this study cohort. Studies performed at two different university hospitals identified HHH in 35 of 107 deceased patients (32.7 %) [8] and 102 of 230 autopsies (44.3 %) [2]. In addition, the patients with mild HHH form the largest cohort in these previous studies. These data are consistent with our results, and together, they suggest that HHH is a common phenomenon in patients who die in major hospitals.

Two major categories of risk factors for HHH have been identified: (i) intrinsic factors, such as advanced malignancies, sepsis, and multi-organ failure; and (ii) extrinsic factors, such as blood transfusions and treatment intensity [2, 8, 21]. However, some controversial results have been obtained. In comparison with non-HHH patients, HHH patients had a significantly increased frequency of hematological diseases, hematological malignancies, and sepsis. Independent risk factors for HHH identified by logistic regression analysis were hematological diseases, ≥15 % bone marrow macrophages, sepsis, and high IL-6 levels. Hematological diseases were the most striking risk factor, and they have been implicated in severe immunodeficiency. Nevertheless, an underlying immunodeficient state at the onset of HHH and HPS seems underrepresented, unlike the documented infections for the most common setting of secondary HPS [22]. However, HHH developed preferentially in our hematological subjects with significantly decreased nadir WBC counts, particularly in hematological patients with hypo-HHH. Hematological patients with leukemia, malignant lymphoma, myelodysplastic syndrome, or aplastic anemia are potentially immunosuppressed, which would make them susceptible to severe infections such as pneumonia, febrile neutropenia, and bloodstream infections [15, 23]. Hematological disorder itself is not a condition of excessive inflammation; however, its associated risk of infections may lead to inflammation. The importance of severe infection for HHH may be validated by the other independent risk factors of sepsis and high IL-6 levels. In fact, sepsis is defined as the systemic response to infection and is involved in the systemic inflammatory response syndrome [11, 12]. IL-6 concentrations are proportional to the severity of infection [24]. The kinetics of inflammatory mediators after death remains unclear. Indeed, few studies have provided sequential analyses of inflammatory cytokine levels after death. As for the IL-6 and IL-1β concentrations, the mean levels tend to gradually increase at 24 h after death in both septic patients and normal cohorts, but several cases have shown decreased levels after death [25, 26]. Because our autopsies were performed within 24 h of death (3.0 ± 2.6 h; range, 1–15.5 h), postmortem modifications are assumed to be slight. The results of previous studies together with the present results suggest that underlying diseases characterized by excessive inflammatory conditions are the most striking initiators of HHH.

The median time from onset to diagnosis of hemophagocytic lymphohistiocytosis is 3.5 months [19], and the activated inflammatory cytokines often persist for 100 days [5]. In addition, the impairment of the Th1 and Th2 balance promotes secretion of Th1 cytokines such as IFN-γ, IL-4, IL-12, and IL-18, followed by the elevation of pro-inflammatory TNF-α, IL-1β, and IL-6 cytokines, IL-8 chemokine, and anti-inflammatory IL-10 cytokine [5, 10, 19]. However, the profiles of inflammatory cytokines have not been previously analyzed in HHH patients. We found that HHH patients had significantly higher levels of IL-1β, IL-6, and IL-8 as compared with patients without HHH. Several mediators including TNF-α, IL-1β, IL-8, and IL-10 showed strong positive correlations, indicating that the secreted cytokines were closely associated. The released cytokines were similar to the cytokine profiles described for septic shock, severe febrile neutropenia, and in vitro sepsis model experiments that used peripheral blood monocytes [27‐29]. Moreover, significant correlations are noted among inflammatory cytokines in severe inflammatory diseases [28, 29]. These findings suggest that a series of inflammatory cytokine releases was a feature of HHH as well as other severe inflammatory diseases and that monocytes and macrophages were responsible for the production of these cytokines. On the other hand, the levels of IFN-γ did not show a statistically significant difference between the HHH and non-HHH patients. The levels of IFN-γ decrease rapidly with improvements in the HPS syndrome [5, 19], suggesting that the IFN-γ concentration had already attenuated at the time of death.

HHH patients have been reported to have normocellular bone marrow, but we identified eight HHH patients with hypocellular bone marrow. Hypo-HHH was characterized by an increased rate of migrated macrophages in the bone marrow, severe leukocytopenia, and high serum IL-8 levels; however, there was no increase in the absolute number of macrophages, indicating that the functional activation of macrophages resulted in the engulfment of bone marrow hematopoietic cells. High IL-8 concentrations may be associated with the seriousness of HHH. Indeed, the chemokine concentration was significantly increased according to the severity of HHH. Even in the patients with mild HHH, the IL-8 concentration was greater than that in the normal cohort (714.2 ± 649.0 pg/ml vs. 505.8 ± 893.7 pg/ml, p < 0.05), although their macrophage counts were in the same range. Conversely, the minimum WBC count of mild HHH was significantly lower than that of non-HHH (4.2 ± 2.9 × 10³/μl vs. 6.5 ± 3.2 × 10³/μl, p < 0.05). IL-8 is a CXC chemokine produced by monocytes and macrophages that attracts neutrophils to inflamed sites for phagocytosis and killing of pathogens [27, 30]. The IL-8 concentrations are inversely correlated with leukocytopenia, and the peak value is seen during the neutropenic phase of hematopoietic stem cell transplantation [31]. These findings suggest that the high IL-8 level in hypo-HHH is the consequence of severe leukocytopenia due to excessive activation of bone marrow macrophages.

Strauss et al. described HPS as the most severe variant of HHH [8]. However, none of the HHH patients in their study or in our study fulfills the diagnostic criteria for HPS. Indeed, HPS and HHH have hypercytokinemia in common, but HPS is a syndrome with defined clinical and laboratory features, one of which is the occurrence of increased histiocytes with hemophagocytosis. By contrast to HPS, the development of HHH is most often a multifactorial process rather than one related to a single underlying condition [2]. Therefore, HHH should be considered as a different disease entity from HPS. Additionally, we found a new feature of HHH in that the absolute number of BM histiocytes was in a similar range in the presence or absence of HHH, indicating that the essential feature of HHH is the augmentation of hemophagocytic histiocytes rather than histiocytic hyperplasia. Hence, we propose a tentative definition of this disorder as “polyhemophagocytosis (PHP)” to avoid confusion with HPS (Table 7) [2, 8].

In conclusion, we found that conditions of excessive inflammation, such as hematological disorders and sepsis, appear to be the strongest initiators of HHH development. In addition, the development of HHH was also associated with increased IL-6 and IL-8 concentrations. Although further investigations of HHH are needed, insights into HHH can be gained from the analysis of data from the present study.