Abstract
Objective. To determine the characteristics of hemophagocytic syndrome (HPS) in adult Korean patients with systemic lupus erythematosus (SLE).
Methods. We reviewed the medical records of 1033 adult patients with SLE for a recent 14-year period and identified 15 patients who had developed HPS. Forty-two age- and sex-matched patients with SLE admitted for other manifestations were included as disease controls. Features of HPS in these patients were analyzed.
Results. Reactive HPS occurred from some distinct causes during the course of SLE. HPS was associated with SLE in 11 patients (4 at onset of SLE and 7 at SLE flare), infection in 3 patients (2 bacterial infection; 1 viral infection), and drug use (azathioprine) in 1 patient. Common clinical features included fever (93.3%), hepatomegaly (60.0%), and splenomegaly (60.0%). Steroid pulse therapy (46.7%), immunosuppressants (46.7%), and intravenous immunoglobulin (46.7%) were frequently used for treatment of HPS. One patient (6.7%) died. Compared with SLE patients without HPS, those with HPS showed a higher SLEDAI score (p = 0.003) and lower levels of plasma leukocytes (p < 0.001), hemoglobin (p = 0.013), and platelets (p < 0.001) as well as a higher serum C-reactive protein level (p = 0.039) and a lower serum albumin level (p = 0.004).
Conclusion. HPS was observed in 1.5% of adult Korean patients with SLE. The occurrence of HPS was most frequently associated with the SLE disease activity. Profound pancytopenia, a high SLEDAI score, and notable changes in the level of acute-phase reactants can be the characteristics of SLE patients with HPS.
Hemophagocytic syndrome (HPS) is a rare but potentially life-threatening disorder. The term refers to the characteristic pathologic findings of activated histiocytes engulfing erythrocytes, leukocytes, platelets, and their precursors in bone marrow and other reticuloendothelial systems1. Common clinical features of HPS include fever, cytopenia, hepatosplenomegaly, abnormal liver function tests, coagulopathy with hypofibrinogenemia, hyperferritinemia, and high blood triglyceride levels2. HPS is divided into primary and secondary HPS3,4. The primary (hereditary) form, also called familial hemophagocytic lymphohistiocytosis, is observed mostly in infants and is often caused by genetic mutations such as perforin gene mutations5. Secondary (reactive) HPS occurs at all ages and is associated with various clinical conditions including infections, malignancies, autoimmune diseases, and administration of certain drugs1,4,6,7,8,9,10. In addition, other miscellaneous conditions have been reported as the cause of reactive HPS11,12,13,14.
Among the systemic autoimmune diseases, systemic lupus erythematosus (SLE) is most frequently described as the underlying disease for HPS15,16,17. Some studies have reported the prevalence of HPS in SLE to be 0.9%–2.4%15,18. In the setting of SLE, reactive HPS can be associated either with an active infection, which is often a complication of an immunosuppressive treatment, or with immune dysregulation induced by the disease itself19. Although understanding the factors that suggest the presence of HPS may allow an earlier diagnosis and more prompt management, there is a lack of overall information about HPS in patients with SLE.
We reviewed demographic data, laboratory and clinical features, treatments, and outcomes of SLE patients with HPS. We also analyzed the characteristics that might distinguish SLE patients with HPS from those without HPS.
MATERIALS AND METHODS
Cases and controls
We retrospectively reviewed the medical records of 1033 adult Korean patients with SLE who were treated at the rheumatology department of Seoul St. Mary’s Hospital between January 1997 and June 2010. All patients satisfied the 1982 revised criteria for the classification of SLE20. Among them, we found 15 patients with HPS. The diagnosis of HPS was made in cases of otherwise unexplained cytopenia that affected at least 2 cell lines and abnormally activated hemophagocytic histiocytes identified in bone marrow, liver, spleen, or lymph node biopsies7,10. Forty-two patients with SLE matched for age and sex were included as disease controls. They were randomly chosen among patients with SLE who were admitted for other manifestations at the same time that SLE patients with HPS were admitted.
Methods
All medical records of these patients were reviewed and the following conditions were evaluated: demographic data such as age, sex, disease duration, potential trigger factors for HPS onset, various clinical features and laboratory findings, treatment modalities for HPS management, and outcomes. Disseminated intravascular coagulation (DIC) was defined as a DIC score ≥ 5 according to the scoring system of the International Society on Thrombosis and Hemostasis21. Terms for steroid doses came from the recommendations for the standardized nomenclature for glucocorticoid dosages and glucocorticoid treatment regimens as published in 200222: (1) pulse therapy ≥ 250 mg prednisone equivalent a day for 1 or a few days; (2) high dose (> 30 mg), but ≤ 100 mg prednisone equivalent a day; and (3) medium dose > 7.5 mg, but ≤ 30 mg prednisone equivalent a day. SLE disease activity was evaluated using the SLE Disease Activity Index (SLEDAI) score23.
Statistical analysis
Comparisons of descriptive data between groups were performed by Student t test or Mann-Whitney U test. Categorical data were compared using the chi-square or Fisher’s exact test. A p value < 0.05 was considered statistically significant.
Systematic review
We completed our study with a systematic review of the published cases of HPS that occurred in the course of SLE. We searched the Medline database using the terms “hemophagocytic syndrome,” “hemophagocytic lymphohistiocytosis,” “hemophagocytosis,” or “macrophage activation syndrome,” combined with “lupus,” between 2006 and 2010. The search was limited to well described adult cases. A manual search of all the references in the articles supplemented this screening.
RESULTS
Demographic and clinical features of HPS
Fifteen (1.5%) out of 1033 patients with SLE recruited for our study were diagnosed with HPS. The demographic and clinical data of 15 patients with HPS in the course of SLE are presented in Table 1. Age at the time of HPS diagnosis ranged from 17 to 50 years (mean 28.9 ± 9.9 yrs). The patients with HPS included 13 women and 2 men. The mean duration of SLE was 69.9 ± 88.9 months (range 0–270 mo). All of the patients were hospitalized for > 9 days because of HPS (mean duration of hospitalization 36.7 ± 30.3 days). It took an average of 15.8 ± 15.2 days to get pathological confirmation from the first day of admission.
All patients except 1 presented fever at the time of admission. The afebrile patient was admitted to our hospital only with general weakness. Hepatomegaly was observed in 9 patients (60%) and splenomegaly in 9 patients (60%); the 2 conditions did not necessarily occur in the same patient. Six patients showed the enlargement of superficial lymph nodes. Neuropsychiatric symptoms such as depressive mood, psychosis, and seizure were noted in 3 patients (20%). Most patients were diagnosed as having HPS from bone marrow analysis. Only 1 patient revealed excessive hemophagocytosis in spleen and liver and not in bone marrow.
Table 1 also shows the SLE manifestations that each patient had ever had before the onset of HPS. Among our 15 patients, 7 had lupus nephritis and 4 had neuropsychiatric lupus. Six patients, on the other hand, had only mild symptoms such as cutaneous or articular manifestations of SLE.
Trigger factors for HPS onset
Possible trigger factors for HPS are described in Table 1. SLE-associated HPS occurred in 11 of our 15 patients (73.3%). Among 11 patients with SLE-associated HPS, HPS and SLE were diagnosed simultaneously in 4 patients. The other 7 patients experienced a flare of SLE just before or almost simultaneously with the development of HPS. Among those 7 patients, 1 (Patient 5) had stopped taking drugs for more than 3 months because of poor compliance. No medication change was associated with Patient 7, but the patient was known to be under stress due to job and family problems before the SLE flare. Patient 8 was in the 12th week of pregnancy without any changes to her medication for at least 6 months. We could not determine any medication change or environmental changes in the other patients before the SLE flare. No evidence of infection was found in these 11 patients with SLE-associated HPS. Active infection triggered HPS in 3 patients (20%). Bacterial infection was identified in 2 patients. Bacillus species was isolated in both urine and blood in 1 patient. Another patient, who was under antibiotic therapy in a local clinic for 1 month before admission to our hospital, had a urinary tract infection caused by extended-spectrum ß-lactamase-producing Escherichia coli. Viral infection was diagnosed in Patient 14, who had a positive result for IgM antibody to Epstein-Barr viral capsid antigen. A specific drug (azathioprine) was suspected as a possible trigger factor in Patient 11, who developed HPS on the fifth day of readministration after transient discontinuation of the drug owing to mild symptoms of upper respiratory tract infection.
Laboratory characteristics of HPS
Table 2 shows the laboratory findings of 15 patients with HPS. Leukopenia or anemia were found in 13 patients (86.7%). Yet all patients showed thrombocytopenia ranging from 16,000 to 121,000/mm3. An elevated erythrocyte sedimentation rate (ESR) was observed in 12 patients (80%) and the level of C-reactive protein (CRP) was elevated in all but 1 patient. The transaminase levels were high in 14 patients (93.3%). All patients presented with hypoalbuminemia, hyper-lactate dehydrogenase (LDH)-nemia and hyperferritinemia. DIC was diagnosed in 6 patients (54.5%) out of the 11 whose data were available. Anti-dsDNA antibodies were present in 10 patients (66.7%).
Treatment and outcome of HPS
The treatments and outcomes of 15 patients with SLE and HPS are illustrated in Table 3. Steroid therapy was given to all patients for the treatment of HPS although the dosage or the type of steroid differed. Doses of steroids equivalent to > 30 mg/day of prednisone were administered to 12 patients, including 7 who received methylprednisolone pulse therapy. The other 3 patients (Patients 3, 6, and 15) received medium-dose steroid therapy (15 mg/day prednisolone). In addition to steroid therapy, immunosuppressants were added for 7 patients (46.7%); cyclosporine A in 5 (Patients 2, 7, 8, 9, and 15), intravenous cyclophosphamide in 2 (Patients 4 and 12), and azathioprine in 1 (Patient 4). On the other hand, Patient 11, who was diagnosed as having drug-induced HPS, recovered after he stopped taking the causative immunosuppressive drug (azathioprine). Intravenous immunoglobulin (IVIG) was used in 7 patients (46.7%). Splenectomy was performed in Patient 8, who was resistant to medical therapy. Plasmapheresis was used in 1 patient (Patient 4) in combination with other medical treatments, including steroid and immunosuppressive drugs. We had confidence that Patient 6 had no evidence of infection at the time of admission and she was the only patient who did not receive antibiotic therapy.
Four patients needed the intensive care unit because of their respiratory or hemodynamic problems. Two patients (Patients 1 and 9) required mechanical ventilation. Fourteen patients recovered from HPS, but 1 died of diffuse pulmonary alveolar hemorrhage.
Comparison of the SLE patients with and without HPS
The results comparing SLE patients with HPS and those without HPS are described in Table 4. Both groups had a statistically similar distribution of age and sex. There was also no significant difference in the duration of SLE between the groups. The SLEDAI scores were higher in SLE patients with HPS (p = 0.003). There were also significant differences in the severity of leukopenia, anemia, and thrombocytopenia between the groups (p < 0.001, p = 0.013, and p < 0.001, respectively). The level of CRP and LDH was higher in the HPS group (p = 0.039 and p < 0.001), although the level of ESR was comparable between the groups. Hypoalbuminemia was more prominent in the HPS group (p = 0.004). The complement level showed no difference between the groups.
DISCUSSION
Reactive HPS has rarely been described in connection with SLE. A few previous studies reported the prevalence and characteristics of HPS in the course of several overall systemic diseases7,17. HPS occurring in patients with SLE has most often been reported in the form of case reports. As shown in Table 5, our Medline search identified 42 adult HPS cases with SLE from 2006 to 201011,17,18,24,25,26,27,28,29,30,31,32,33,34,35. Those 42 patients consisted of 13 whites and 29 Asians. The demographic and clinical features of the literature cases did not differ from our cases, except for the frequency of using steroid pulse therapy (Table 6). There were 33 female and 9 male patients. The average age of the patients was 34.2 ± 14.4 years. Thirty-six patients (85.7%) had developed HPS associated only with SLE. Both SLE and infection (cytomegalovirus) caused the HPS at the same time in 1 patient (2.4%). Active infection was suggested as a sole trigger factor in 4 patients (9.5%; influenza B virus, cytomegalovirus, parvovirus B19, and Enterococcus faecalis in 1 patient each). Pregnancy was associated with the development of HPS in 1 patient (2.4%). All 42 patients in the literature and our 15 cases received steroid treatments, although steroid pulse therapy was used more frequently in the literature cases than in ours (p = 0.045). The frequency of using immunosuppressive agents did not differ between the groups (p = 0.358). Cyclosporine A was administered most frequently to the patients in both groups (5 patients among our 15 cases, 9 patients among the 42 literature cases). The second most common drug treatment was intravenous cyclophosphamide (2 patients among our 15, 6 patients among the 42 literature cases). IVIG was used as often as immunosuppressive drugs for the treatment of HPS in both groups. The overall mortality did not differ between the 2 groups.
Diagnosis of reactive HPS remains difficult in adult patients, although Henter, et al published revised diagnostic guidelines for hemophagocytic lymphohistiocytosis in 200436. Some studies asserted that reactive HPS in adults is distinct from HPS in children and that we should be prudent in adopting guidelines for adult patients37. Tsuda described a different frequency of hepatosplenomegaly, rash, and neurologic involvement, which were all more common in children38. Hypertriglyceridemia has been described as rare in adults39. Accordingly, we adopted definitions that could be accepted universally in the study of reactive HPS with autoimmune diseases in adult patients17,40. But more intensive and extensive studies of adults with HPS will be needed in order to create differentiated HPS guidelines for adults.
One afebrile patient (Patient 10) may seem to be a distinctive case. We raised that patient’s dosage of prednisolone from 25 mg/day to 40 mg/day when we noted mild leukopenia, thrombocytopenia, and abnormal levels of liver enzymes, under the assumption that she was experiencing an SLE flare. During the hospitalization, her body temperature was 37.8°C. Her fever might have been partially masked by a prompt dose of steroids. But profound cytopenia, high ferritin and triglyceride levels, hepatomegaly, and remarkable hemophagocytosis in the bone marrow were clinically sufficient to diagnose HPS and treat the patient with high-dose steroids despite the absence of high fever. This case made us reconsider the development of HPS, although fever is not prominent in steroid-treated patients with other features of HPS.
A notable finding in our study was that the onset or flare of SLE frequently caused the HPS. Adult-onset Still’s disease is the only other known disease that could be caused by the underlying disease itself7,40. In patients with other systemic diseases, infection was almost always present with occurrence of HPS7. This finding suggests that treatment decisions for HPS patients with SLE should differ from decisions made for patients with other systemic diseases.
Our data showed that SLEDAI scores of patients with HPS were higher than those of patients without HPS. The constituents of the SLEDAI score include leukopenia and thrombocytopenia, both of which can be manifestations of HPS. So we subtracted the scores of leukopenia and thrombocytopenia from the SLEDAI scores of patients with HPS and then compared the scores with those of patients without HPS. The corrected SLEDAI scores of SLE patients with HPS were significantly higher than the scores of those without HPS (p = 0.031). These findings suggest that a high SLEDAI score may be important when making assumptions about the development of HPS in patients with SLE.
The hematologic involvement of SLE can cause cytopenia, a condition that may lead clinicians to hesitate before evaluating a patient for HPS. That is also why the time from admission to HPS diagnosis varied from 2 to 54 days in our series. But it has been reported that only 1 or 2 cell lines are usually affected, and the patients have only mild pancytopenia15. Our results support the previous reports; clinicians should have doubts about the development of HPS in an SLE patient with longstanding, profound pancytopenia. One of the interesting findings is that all our patients with HPS had thrombocytopenia to a greater or lesser degree. It seems that thrombocytopenia is a more prominent feature of HPS in patients with SLE than leukopenia or anemia. We need further investigations and reviews of more cases to confirm this.
Slightly exaggerated hemophagocytosis can be a physiologic process that might occur in conditions including hemolytic and aplastic anemia, graft versus host disease, and after transfusions and cytotoxic treatments37. Despite various limitations, morphological evidence of hemophagocytosis is still considered the “gold standard” in the diagnosis of HPS37. Pathologic evidence of hemophagocytosis is usually obtained with bone marrow specimens41. Most of our patients’ diagnoses were confirmed with bone marrow studies, with 1 exception. That patient (Patient 8) had no abnormal findings in bone marrow biopsy, and hemophagocytosis was confirmed later through splenectomy. A negative bone marrow biopsy cannot always rule out HPS. In addition, histological hemophagocytosis is not always a pathognomonic finding in the diagnosis of HPS unless patients have no symptoms or signs of HPS.
HPS results from uncontrolled T lymphocyte activation that leads to macrophage activation and an increment of some cytokines. In the liver, increased inflammatory cytokines such as tumor necrosis factor-α42, interleukin 1 (IL-1)42, IL-643, IL-1844, and interferon-γ45 enhance the synthesis of acute-phase proteins, but they reduce the synthesis of albumin in the patients with HPS46,47. In our series, hypoalbuminemia and hyper-CRP-nemia were more remarkable in SLE patients with HPS than in those without HPS. Both albumin and CRP belong to the acute-phase reactants, a class of proteins whose plasma concentrations decrease or increase in response to inflammation48. Our study indicated that on average, HPS causes a more potent inflammatory condition than does infection and SLE flare, because the control group (SLE patients without HPS) included SLE patients who had undergone various infections and/or a flare of SLE including cutaneous vasculitis, mesenteric vasculitis, or thrombotic thrombocytopenic purpura. But this may be because our SLE patients with HPS almost always had combined conditions such as HPS with infection or HPS with an SLE flare.
It is common to administer antibiotics to febrile patients with SLE. This is because fever caused by disease activity can be diagnosed only after exclusion of infection by a thorough evaluation, although both infection and SLE disease activity have been reported to be the most common causes of fever in patients with SLE49,50. Similarly, it is difficult and time-consuming to identify the causative factors for HPS that mainly consist of SLE and infection. That situation may lead to the prescription of antibiotics for most of these patients and a delay in increasing the dosages of steroids or in adding the proper immunosuppressive agents.
HPS was observed in 1.5% of our adult Korean patients with SLE. The development of HPS was most frequently associated with SLE disease activity. Profound pancytopenia and a high SLEDAI score as well as remarkable changes of the level of acute-phase reactants were the characteristics of HPS patients with SLE. For prompt and appropriate therapy, physicians should be aware of the symptoms of HPS and understand the distinctions in laboratory and clinical findings between a flare of SLE and the occurrence of HPS.
- Accepted for publication September 15, 2011.