Clinical features of haemophagocytic syndrome in patients with systemic autoimmune diseases: analysis of 30 cases

Abstract

Objectives. Haemophagocytic syndrome (HPS) is known as a relatively rare complication in autoimmune diseases. Here we analysed the clinical features of HPS in patients with systemic autoimmune diseases.

Methods. One thousand and fourteen patients with systemic autoimmune diseases admitted to Hokkaido University Hospital from 1997 to 2007 were recruited [350 SLE, 136 RA, 98 polymyositis/dermatomyositis (PM/DM), 88 SSc, 91 vasculitis syndrome, 37 primary SS, 26 adult onset Still's disease (AOSD) and 188 other diseases]. Clinical features and treatment outcomes were retrospectively analysed.

Results. Thirty cases (3.0%) fulfilled HPS criteria (progressive cytopenia in two or more lineages and haemophagocytosis in reticuloendothelial systems). Underlying diseases were SLE (18), RA (2), PM/DM (2), SSc (2), vasculitis (1), SS (2) and AOSD (3). Nineteen patients were diagnosed as having autoimmune-associated HPS, eight infection-associated, one drug-induced and one developed HPS after haematopoietic stem cell transplantation. For the treatment of HPS, high-dose corticosteroid monotherapy was given in 26 cases, being effective in 12 (46%). Ten out of 15 patients with corticosteroid-resistant autoimmune-associated HPS were treated with CsA, cyclophosphamide or tacrolimus, leading to the remission in 80%. The overall mortality rate was 20%. Multivariate analysis showed that the presence of infections and CRP level >50 mg/l on HPS related with poor prognosis.

Conclusions. The prevalence of HPS among in-hospital patients with systemic autoimmunity is not ignorable. Administration of immunosuppressants was effective in cases with autoimmune-associated HPS, whereas prognosis was poor in infection-associated HPS.

Haemophagocytosis, Haemophagocytic syndrome, Autoimmunity, Systemic lupus erythematosus, Mortality factors

Introduction

Haemophagocytic syndrome (HPS) is a clinicopathological entity characterized by the activation of macrophages and/or histiocytes with prominent haemophagocytosis in bone marrow and other reticuloendothelial systems [1]. Clinical features of HPS include fever, cytopenia, liver enzyme elevation, hepatosplenomegaly, lymphadenopathy and coagulopathy. HPS comprises primary (hereditary) and reactive forms. Primary HPS is observed mostly in infants, represented by cases with familial haemophagocytic lymphohistiocytosis (FHL), Chediac–Higashi syndrome, Griscelli syndrome and X-linked lymphoproliferative syndrome. Reactive HPS occurs in situations such as infections, malignant lymphomas, autoimmune diseases and administration of certain drugs [2–4]. In rare cases, HPS occurs after haematopoietic stem cell or organ transplantations [5–12].

Among the systemic autoimmune diseases, SLE is the most common underlying disease for HPS. In 1991, Wong et al. [13] proposed a disease entity of acute lupus haemophagocytic syndrome, in which haemophagocytosis occurs in association with the SLE flare and the corticosteroid therapy was effective in those cases. After Kumakura et al. [2] reported cases of reactive HPS associated with other autoimmune diseases such as adult-onset Still's disease (AOSD), MCTD and Evans syndrome, the concept has expanded to autoimmune-associated haemophagocytic syndrome (AAHS). Later on, it has been recognized that HPS and AOSD share the pathophysiology representing excessive activation of macrophages with cytokine overproduction [4, 8]. The prevalence of HPS in SLE and AOSD was reported as 0.9–2.4% and 12%, respectively, by some previous studies [4, 13, 14]. Recently, Dhote et al. [3] retrospectively investigated 26 patients with reactive HPS in adult autoimmune diseases in a French multicentre study and reported that the overall mortality was 38.5%. However, epidemiology of HPS in systemic autoimmunity has not been established, and clinical profiles of such patients are still obscure. To clarify the prevalence and the prognosis of HPS complicated with systemic autoimmune diseases, we retrospectively analysed the clinical features of HPS in the in-hospital patients with systemic autoimmune diseases in our hospital.

Patients and methods

Patients and data

One thousand and fourteen consecutive patients with systemic autoimmune diseases [350 SLE, 136 RA, 98 polymyositis/dermatomyositis (PM/DM), 88 SSc, 91 systemic vasculitis syndrome, 37 primary SS, 26 AOSD and 188 other autoimmune diseases] admitted to Hokkaido University Hospital during the period from January 1997 to December 2007 were recruited. All patients received in-hospital treatment for more than 7 days. RA was diagnosed using the ARA 1987 revised criteria [15]. Diagnosis of SLE was made according to the revised classification criteria of ACR [16]. PM/DM and SSc were diagnosed according to the diagnostic criteria proposed by Bohan and Peter [17] and to the preliminary criteria for the classification of SSc proposed by ARA, respectively [18]. AOSD was diagnosed using preliminary criteria for classification of AOSD proposed by Yamaguchi et al. [19]. Vasculitis syndrome was diagnosed according to the ACR 1990 criteria [20, 21], and SS using the revised European criteria [22]. Patients having malignant diseases were excluded from this study. Diagnoses of these autoimmune diseases were made or had already been made when our patients were in hospital.

Diagnosis of HPS was made in cases of otherwise unexplained progressive cytopenia affecting at least two cell lineages, and bone marrow showing mature histiocytes with prominent haemophagocytosis and/or haemophagocytosis in liver, spleen or lymph nodes [1, 3]. Diagnosis of AAHS was made according to the criteria proposed by Kumakura et al. [23], which require (i) cytopenia affecting at least two lineages, (ii) histiocytic haemophagocytosis in bone marrow or other reticuloendothelial systems, (iii) active phase of underlying disease at the occurrence of haemophagocytosis and (iv) exclusion of other reactive HPS such as virus or malignancy-associated HPS. Diagnosis of HPS/AAHS was made when these patients were under treatment.

Medical records of these patients were retrospectively reviewed and the following clinical conditions were evaluated: underlying diseases, triggers for the HPS onset, clinical features including fever, splenomegaly, lymphadenopathy and neuropsychiatric symptoms. Laboratory data such as complete blood cell (CBC) counts, serum liver transaminase levels, serum ferritin levels, serum lactate dehydrogenase (LDH) levels, serum CRP levels, serum triglyceride levels and plasma D-dimer levels were collected as well at the time of HPS diagnosis. Presence of active bacterial infections was diagnosed upon positive blood, urine or sputum cultures. CMV infection was defined by positive C7-HRP antigenaemia with at least one of the following clinical manifestations: fever, pneumonia, cytopenia or liver damage [24]. Pneumocystis jirovecii pneumonia (PCP) was diagnosed by CT scan with positive staining of bronchoalveolar fluid and/or positive PCR product from a respiratory specimen [25]. Disseminated intravascular coagulation (DIC) was diagnosed according to the algorithm proposed by International Society on Thrombosis and Haematostasis [26]. Lupus disease activity was evaluated using BILAG index [27]. Remission was defined as returning to >80% of the patients’ baseline CBC. Sex, age, duration of underlying diseases, clinical manifestations, presence of infections and laboratory data were explored for relationships with mortality.

This retrospective observation study has been approved by the ethical committee of Hokkaido University Hospital.

Statistical analysis

All factors were categorized, and then the relations with outcomes were examined using Fisher's exact test. P-values were further corrected by multiplying by the number of factors (n = 16) investigated (corrected P). The relation between global BILAG score and response to corticosteroids was examined using Mann–Whitney's U-test.

Results

Prevalence of HPS

Of 1014 in-hospital patients recruited in this study, 30 cases (3.0%) were diagnosed as HPS. The HPS patients included 24 females, and the mean age of HPS onset was 43 yrs (range 19–71 yrs) at the diagnosis of HPS. The mean duration of underlying autoimmune diseases was 5.7 yrs (from 0 months to 29 yrs) at the diagnosis of HPS. HPS was the most prevalent in patients with AOSD, followed by SS and SLE groups (Table 1). Nineteen patients (63% of HPS, 1.9% of autoimmune diseases) were diagnosed as having AAHS. AAHS was the most frequent in patients with AOSD (7.7%), followed by those with SLE (4.6%). Eight cases (27% of HPS, 0.8% of autoimmune diseases) were diagnosed as infection-associated HPS (IAHS). The remaining three patients had neither infection nor exacerbation of autoimmune diseases.

Table 1.

Prevalence of HPS and AAHS

Underlying diseases	SLE (n = 350)	RA (n = 136)	PM/DM (n = 98)	Vasculitis (n = 91)	SSc (n = 88)	SS (n = 37)	AOSD (n = 26)	Others (n = 188)	Total (n = 1014)
Age (mean ± s.d.)	37 ± 14	62 ± 12	54 ± 14	55 ± 19	44 ± 15	54 ± 12	34 ± 16	44 ± 17	46 ± 18
Sex, female/male	293/57	104/32	72/26	62/29	78/10	36/1	20/6	142/46	807/207
HPS (%)	18 (5.1)	2 (1.5)	2 (2.0)	1 (1.1)	2 (2.3)	2 (5.4)	3 (11.5)	0 (0)	30 (3.0)
AAHS (%)	16 (4.6)	0 (0)	1 (1.0)	0 (0)	0 (0)	0 (0)	2 (7.7)	0 (0)	19 (1.9)

Clinical features of HPS

The clinical manifestations of 30 patients with HPS are illustrated in Table 2. At the time when the diagnosis of HPS was made, fever at ≥38° was observed in 87% of the patients. Neuropsychiatric manifestations such as coma, seizure, meningeal irritation signs, mood disorder, cognitive dysfunction, delirium and psychosis were noted in nine cases (30%). Leukopenia (<4.0 × 10⁹/l), anaemia (haemoglobin <115 g/l) or thrombocytopenia (<150 × 10⁹/l) was found in 87, 83 or 87%, respectively. CRP was elevated in 83% of the patients. Liver enzyme elevation (aspartate aminotransferase or LDH), more than two times the upper normal limit, was observed in 60%. Median ferritin level was 980 ng/ml (range 17–27 842 ng/ml). Elevated plasma D-dimer levels were recorded in 96%, and overt DIC (DIC score ≥5) was noted in two patients (Patients 24 and 30).

Table 2.

Clinical manifestations and laboratory data at the onset of HPS

Patient	Age /sex	Underlying diseases	Spleno -megalya	Lymph -adenopathy	Neuropsychiatric symptoms	Fever (°C)	DIC	WBC (×109/l)	Hb (g/l)	Plt (x109/l)	AST (U/l)	LDH (U/l)	Ferritin (ng/ml)	D-dimer (ng/ml)	CRP (mg/l)
1	35/F	SLE	−	−	-	40.0	−	3.5	128	94	44	930	747	ND	218
2	26/F	SLE	−	−	Meningeal irritation signs	38.8	−	2.5	89	133	92	1012	1387	38 900	4.8
3	26/M	SLE	+	−	-	39.2	−	1.7	118	107	144	907	2949	4500	2.4>
4	21/F	SLE	+	+	Emotional incontinence	38.6	−	0.8	62	15	109	607	460	2200	2.4>
5	21/F	SLE	−	+	-	38.9	−	3.6	84	95	76	1278	88	3500	9.6
6	20/F	SLE	−	+	-	39.0	−	4.1	115	227	235	1172	1326	2900	8.8
7	22/F	SLE	−	−	-	41.0	−	1.6	78	161	243	3077	4606	ND	15.9
8	34/F	SLE	−	−	Coma, seizure	39.4	−	2.1	109	55	1724	4850	27 842	109 400	137.0
9	52/F	SLE	−	−	-	37.8	−	0.5	69	20	143	888	902	23 500	99.0
10	19/F	SLE	+	+	-	40.6	−	3.1	63	29	62	508	273	100 500	35.4
11	29/M	SLE	−	−	Mood disorder, cognitive dysfunction	38.8	−	2.1	107	62	20	529	653	8100	2.4>
12	41/F	SLE	+	−	-	40.2	−	4.1	88	63	109	563	3804	9500	104.2
13	29/F	SLE	−	−	-	38.9	−	0.5	65	79	300	1191	13 918	23 000	42.1
14	41/M	SLE	+	−	-	38.0	−	0.1	119	23	9	266	579	4100	167.2
15	60/F	SLE	+	−	-	37.9	−	0.5	72	64	99	249	92	23 400	6.6
16	40/F	SLE	−	−	Mood disorder, cognitive dysfunction	40.0	−	2.2	126	73	14	435	109	1100	3.2
17	51/F	SLE	+	−	Coma	36.9	−	1.9	75	9	54	314	17	400	2.4>
18	50/F	SLE	+	−	Psychosis	37.7	−	1.7	88	16	26	679	685	12 800	6.2
19	62/M	AOSD	+	−	-	39.5	−	0.2	59	14	23	220	1057	4800	84.1
20	71/M	AOSD	−	−	Delirium	38.2	−	2.7	76	89	16	335	1494	6500	172.2
21	31/F	AOSD	+	−	-	39.4	−	5.4	110	168	291	1213	4500	23 000	29.6
22	47/F	SS	+	+	-	39.2	−	1.3	108	201	45	708	321	2000	172.2
23	67/F	SS	−	−	-	39.2	−	0.7	56	17	41	393	2874	28 700	51.7
24	68/M	PM	−	−	-	38.0	+	6.3	65	29	44	608	4071	10 600	136.9
25	48/F	DM	−	−	-	38.5	−	3.5	103	51	49	969	123	ND	2.4>
26	52/F	SSc	−	−	-	38.9	−	3.6	79	106	129	2065	2013	48 600	18.4
27	19/F	SSc	−	+	-	38.0	−	0.1	74	5	26	317	598	6600	46.2
28	71/F	MPA	−	−	-	38.0	−	0.5	79	11	29	372	320	8100	217.8
29	64/F	RA	+	−	-	37.4	−	3.8	94	63	267	1604	10 183	197 000	141.3
30	68/F	RA	−	−	Coma	38.3	+	0.3	66	5	156	318	11 000	31 200	411.0

^aSplenomegaly was defined as major axis being ≥10 cm or ≥1 cm larger than that of previous examination. AST: aspartate aminotransferase; LDH: lactate dehydrogenase; Hb: haemoglobin; ND: no data; DIC: disseminated intravascular coagulation.

Triggers for HPS onset

Active infection triggered HPS in eight patients. Patients 1, 20 and 30 had single microbe infection with CMV, PCP and herpes simplex virus, respectively. Bacterial infection was identified in three patients (cholecystitis in Patient 9, pneumonia in Patient 24 and urinary tract infection in Patient 26). In two cases, patients had overlap infection (CMV and PCP in Patient 23, aspergillus pneumonia and PCP in Patient 28). Two cases were diagnosed as having drug-induced HPS in the series, zonisamide for Patients 22 and cyclophosphamide (large amount for autologous peripheral blood stem cell transplantation) for Patient 27. No trigger could be identified in Patient 29. The remaining 19 cases had active phase of underlying systemic autoimmune diseases without signs of infection or other triggers, being diagnosed as AAHS.

Treatments and outcomes

Treatments and outcomes are illustrated in Table 3 and Fig. 1. Two cases (Patients 22 and 26) recovered without additional immunosuppressive therapy. In Patient 22, discontinuation of zonisamide resulted in remission. Patient 26 had urinary tract infection and cefotiam led to remission of HPS as well as the manifestation of infection. In one AOSD case with PCP (Patient 20), who was on 40 mg/day of prednisolone (PSL) and 150 mg/day of CsA at the onset of HPS, sulbactam/trimethoprim treatment could not control the mortal infection. In another recurrent AOSD case (Patient 21), early corticosteroid pulse therapy with oral CsA was introduced, resulting in prompt remission.

Fig. 1.

Treatments and outcomes of our cohort of 30 HPS patients are sequentially illustrated. mPSL, methylprednisolone; *Transferred to another hospital.

Table 3.

Treatment and outcome

Patient	Age/sex	Underlying diseases	Durationa, months	Triggers	Therapies, effectiveness	Outcome
1	35/F	SLE	108	Infection	mPSL pulse + PSL + anti-viral drugs, effective	Alive
2	26/F	SLE	120	Autoimmune	mPSL pulse + PSL, effective	Alive
3	26/M	SLE	18	Autoimmune	PSL, effective	Alive
4	21/F	SLE	103	Autoimmune	mPSL pulse + PSL, effective	Alive
5	21/F	SLE	54	Autoimmune	PSL, effective	Alive
6	20/F	SLE	24	Autoimmune	mPSL pulse + PSL, effective	Alive
7	22/F	SLE	50	Autoimmune	mPSL pulse + PSL, effective	Alive
8	34/F	SLE	102	Autoimmune	mPSL pulse + PSL, effective	Alive
9	52/F	SLE	348	Infection	mPSL pulse + PSL + antibiotics, ineffective	Dead
10	19/F	SLE	60	Autoimmune	mPSL pulse + PSL, ineffective; CsA, ineffective; vincristine, ineffectived	Alive
11	29/M	SLE	46	Autoimmune	mPSL pulse + PSL, ineffective; IVCY, effective	Alive
12	41/F	SLE	25	Autoimmune	mPSL pulse + PSL, ineffective; CsA, effective	Alive
13	29/F	SLE	6	Autoimmune	mPSL pulse + PSL; ineffective; CsA, effective	Alive
14	41/M	SLE	48	Autoimmune	mPSL pulse + PSL, ineffective; IVCY, ineffective	Dead
15	60/F	SLE	240	Autoimmune	mPSL pulse + PSL, ineffective; CsA, effective	Alive
16	40/F	SLE	4	Autoimmune	mPSL pulse + PSL, ineffective; CsA, effective	Alive
17	51/F	SLE	187	Autoimmune	mPSL pulse + PSL, ineffective; CsA, ineffective;	Alive
18	50/F	SLE	0	Autoimmune	mPSL pulse + PSL, ineffective; CsA, ineffective; tacrolimus, effective; PE, effective	Alive
19	62/M	AOSD	2	Autoimmune	mPSL pulse + PSL, effective	Alive
20	71/M	AOSD	15	Infection	Antibiotics, ineffective	Dead
21	31/F	AOSD	31	Autoimmune	mPSL pulse + PSL + CsA, effective	Alive
22	47/F	SS	5	Drugb	Drug withdrawal, effective	Alive
23	67/F	SS	72	Infection	mPSL pulse + PSL + antibiotics + anti-viral drugs, effective	Alive
24	68/M	PM	7	Infection	mPSL pulse + antibiotics, ineffective	Dead
25	48/F	DM	3	Autoimmune	mPSL pulse + PSL, ineffective; IVCY, effective	Alive
26	52/F	SSc	96	Infection	atibiotics, effective	Alive
27	19/F	SSc	72	Drugc	mPSL pulse + PSL, effective	Alive
28	71/F	MPA	10	Infection	mPSL pulse + PSL + antibiotics + anti-fungal drugs, ineffective	Dead
29	64/F	RA	36	Unknown	mPSL pulse + PSL, effective	Alive
30	68/F	RA	168	Infection	mPSL pulse + PSL + anti-viral drugs, ineffective	Dead

^aDuration from the onset of underlying diseases to the onset of HPS. ^b Zonisamide; an anti-epileptic drug. ^c Cyclophosphamide. ^d Patient was transferred to another hospital on 25th day of admission. mPSL pulse: methylprednisolone pulse therapy; PE: plasma exchange; ND: no data.

Monotherapy with high-dose corticosteroid was given as an initial treatment in all of the remaining 26 cases (including methylprednisolone pulse therapy in 24 patients), being effective in 12 (46%). Four patients were lost due to infectious diseases. In 10 corticosteroid-resistant AAHS patients, immunosuppressants were introduced [oral/intravenous CsA in seven patients and intravenous cyclophosphamide pulse therapy (IVCY) in three cases. Two patients resistant to CsA received tacrolimus, resulting in remission], leading to remission in eight cases (80%). In a lupus patient (Patient 14) with HPS, glomerulonephritis and pericarditis with uncontrollable massive pericardial effusion, disease activity was not reduced by the treatment with high-dose corticosteroid including corticosteroid pulse therapy and IVCY. This patient died of multiple organ failure that is attributable to active SLE. In another lupus patient (Patient 10), HPS was resistant to corticosteroid pulse therapy, CsA and vincristine. Afterwards, rituximab was introduced in another hospital, resulting in remission. Four patients (Patients 9, 24, 28 and 30) did not receive additional immunosuppressants because they had active infection. In these cases, infectious diseases did not arrive at remission, resulting in death. In Patient 9, HPS was triggered by Enterococcus faecalis-induced septicaemia. Patient 24 suffered from bacterial pneumonia. In Patient 28, complication of CMV-induced HPS and CMV/aspergillus pneumonia was lethal. Patient 30, in whom herpes simplex virus was the suspicious cause of HPS, died of DIC.

For supportive therapies, eight patients received G-CSF when neutrophil counts were <0.5 × 10⁹/l. γ-Globulin was administered in five patients with severe infections. Plasma exchange was performed in Patient 18 for concomitant thrombotic microangiopathy that developed after corticosteroid pulse therapy for HPS, with preferable outcome.

In our study, 54% of all HPS cases, including 56% of AAHS cases, were resistant to corticosteroids. The overall mortality rate was 20% (6/30) in our series. Mortality rate of AAHS was 5.3% (1/19), whereas that of IAHS was 63% (5/8).

Lupus activity and HPS

Clinical data of the 18 SLE patients are illustrated in Table 4. SLE patient group comprised 15 females, with mean age of 34 (range 19–60) yrs. Mean disease duration at HPS diagnosis was 7.1 yrs, ranging from 0 months to 20 yrs. In all AAHS patients, the median global BILAG score was 24 (range 12–46), Grade B or higher disease-scoring items in more than one component being noted. Among 16 AAHS patients with SLE, there was no significant difference between global BILAG scores in corticosteroid responders and those in non-responders (P = 0.383).

Table 4.

HPS in SLE patients and BILAG index

^aDuration from the onset of SLE to the onset of HPS.

Risk factors for poor prognosis

Univariate analysis showed that age over 50 yrs, the presence of infection (IAHS), leucocyte count <0.5 × 10⁹/l, platelet count <50 × 10⁹/l and CRP level <50 mg/l at the onset of HPS were related with mortality, as shown in Table 5. Multivariate analysis confirmed the correlation of the presence of infection and high CRP level to the poor prognosis.

Table 5.

Factors associated with mortality in HPS in systemic autoimmune diseases

Factors	n	Mortality, %	P-values (univariate)	OR (95% CI)	Corrected P-valuesa (multivariate)
Gender
Male	6	50	0.068		NS
Female	24	13
Age, yrs
≥50	12	42	0.024	12.14
<50	18	6		(1.19, 123.62)	NS
Duration of underlying disease, months
≤12	22	18
>12	8	25	0.345		NS
Infection
IAHS	6	63		35.00
Non-IAHS	24	5	0.0024	(2.98, 411.48)	0.0384
Fever, °C
≥38.0	25	20
<38.0	5	20	0.447		NS
Splenomegaly
(+)	12	8
(−)	18	28	0.173		NS
Lymphadenopathy
(+)	6	0
(−)	24	25	0.227		NS
Neuropsychiatric disorder
(+)	9	22
(−)	21	19	0.363		NS
Leucocytes, ×109/l
≤0.5	8	50		10.0
>0.5	22	9	0.027	(1.34, 74.51)	NS
Hb, g/l
≤80	15	33
>80	15	7	0.075		NS
Platelets, ×109/l
≤50	12	42		12.14
>50	18	6	0.024	(1.19, 123.62)	NS
AST, U/I
≥80	14	14
<80	16	25	0.279		NS
LDH, U/I
≥260	28	21
<260	2	0	0.634		NS
Ferritin, ng/ml
≥10 000	4	25
<10 000	26	19	0.443		NS
D-dimer, ng/ml
≥10 000	13	15
<10 000	14	14	0.404		NS
CRP, mg/l
≥50	13	42
<50	17	6	0.0029	NA	0.046

^aP-values were corrected by multiplying by the number of factors investigated (corrected P). Hb: haemoglobin; AST: aspartate aminotransferase; NS: not significant; OR: odds ratio; NA: not available.

Discussion

Our series showed the prevalence of HPS in patients with systemic autoimmune diseases and demonstrated that a variety of diseases could underlie HPS. The frequency of AAHS was 4.6% in SLE patients and 7.7% in AOSD patients, being compatible with the previous reports [13]. Most of the HPS in other autoimmune diseases were IAHS. Considering the report by Dhote et al. [3] that almost all of HPS in systemic autoimmune diseases other than SLE and AOSD was complicated by infectious events, our data confirm that AAHS is rarely found in other systemic autoimmune diseases.

We found that 9/30 (30%) patients had neurological diseases. Although neurological abnormalities are documented mainly in FHL in childhood [28], few have been described in adult cases, especially in AAHS [29]. Previous studies have documented that ∼30–70% of HPS patients (mainly FHL) show neurological abnormalities such as seizures, conscious disorder, hemiparesis, nuchal rigidity and ataxia. Cytokines and other neurotoxic factors secreted by leucocytes and the microglial cells may be responsible for the cause of these neurological manifestations [30]. Although neurological involvement in our series of adult patients was not as frequent as that reported in FHL cases, we propose that neurological abnormalities are still one of the major symptoms in adult HPS patients.

Treatment strategy has not been fully established in HPS complicated with systemic autoimmune diseases. Corticosteroid monotherapy has been preferably administered for AAHS, based on the report by Wong et al. [13] in which corticosteroid was uniformly effective in HPS complicated with SLE. According to the multicentre report of 26 autoimmune cases, 19 were treated with corticosteroid alone and additional IVCY was administrated in 3 cases [3]. In the present report, 12 AAHS patients were treated with high-dose corticosteroid without immunosuprressants and 10 AAHS patients received immunosuppressive agents. Additional immunosuppressants in corticosteroid-resistant AAHS were effective in 80%. Nine out of 16 (56%) AAHS patients with SLE were resistant to corticosteroid monotherapy in our study, more frequent than previously reported (0–14%) [13, 14]. All of the corticosteroid-resistant acute lupus HPS patients in our series received CsA, IVCY or tacrolimus, being effective in seven out of nine (78%). Thus, we propose to introduce early and appropriate immunosuppressive treatment in corticosteroid-resistant AAHS cases. Although IVCY and CsA were equally effective in our cases, we prefer CsA to IVCY in two reasons. First, anti-cytokine effect in CsA is acceptable for the treatment of HPS, in which hypercytokinaemia is one of the basic aetiologies in this syndrome. Second, bone marrow suppression can be induced by IVCY, which may evoke serious infection in immunocompromised patients.

The treatment of IAHS presents difficulties and resulted in poor outcome compared with non-IAHS. For EBV-associated HPS, combination chemotherapy and immunotherapy are recommended as an initial treatment, whereas aciclovir does not appear to be useful [1, 31]. For IAHS associated with other viruses and microbes, treatment protocols have not been established. In general, treatments of the underlying infection with supportive therapies, including G-CSF and γ-globulin, are recommended. However, the efficacy of such treatments is still obscure.

We suggest that the presence of infections and high CRP level are related with mortality. There have been several reports concerning the risk factors for poor prognosis in patients with HPS. Kaito et al. [32] analysed 34 patients with HPS (13 with haematological malignancies or viral infections) and concluded that risk factors associated with death were age over 30 yrs, presence of DIC, increased ferritin and β₂-microglobulin or anaemia accompanied by thrombocytopenia and jaundice. Takahashi et al. [33] retrospectively analysed 52 adult HPS patients and reported that the prognosis depended on their underlying diseases, malignant lymphoma and opportunistic infections.

In the autoimmune setting, (i) absence of lymphadenopathy, (ii) thrombocytopenia (<50 × 10⁹/l), (iii) corticosteroid therapy at the time of HPS diagnosis, were listed as poor prognostic factors [3]. Taking these studies and our study together, infections seem to be the common risk factor in adult HPS patients. In this study, high CRP levels were associated with mortality and with infection (P = 0.02; Mann–Whitney U-test), thus high CRP level may relate to infection rather than to HPS itself.

Although this study provides novel information/suggestions for the treatment of HPS in autoimmune diseases, it has to be noted that retrospective observation studies have some shortcomings in the evaluation of the patients’ clinical features. Further studies, especially randomized prospective trials will be necessary to establish better management for the affected patients.

In conclusion, our study has clarified the prevalence of HPS in in-hospital patients with systemic autoimmune diseases. Among these patients, the prognosis was better in patients with AAHS than those with IAHS, although the presence of infections and high CRP level increase the risk for poor prognosis.

Acknowledgements

This study was supported by the Japanese Ministry of Health, Labor and Welfare; Japanese Ministry of Education, Culture, Sports, Science and Technology; Japanese Society for the Promotion of Science. We appreciate Dr Y. Fujieda, Dr K. Otomo, Dr N. Yoshida and Dr M. Kato for their clinical contributions. We thank Dr K. Ogasawara and Dr T. Terashita from the Department of Health Sciences, Hokkaido University for helpful comments on statistical analysis.

Disclosure statement: The authors have declared no conflicts of interest.