ARTHRITIS & RHEUMATISM

Vol. 60, No. 11, November 2009, pp 3388–3399

DOI 10.1002/art.24883
© 2009, American College of Rheumatology

Macrophage Activation Syndrome in

Juvenile Systemic Lupus Erythematosus

A Multinational Multicenter Study of Thirty-Eight Patients

Alessandro Parodi,1 Sergio Davı̀,1 Alejandra Beatriz Pringe,2 Angela Pistorio,1

Nicolino Ruperto,1 Silvia Magni-Manzoni,3 Paivi Miettunen,4 Brigitte Bader-Meunier,5
Graciela Espada,6 Gary Sterba,7 Seza Ozen,8 Dowain Wright,9 Claudia Saad Magalhães,10
Raju Khubchandani,11 Hartmut Michels,12 Patricia Woo,13 Antonio Iglesias,14
Dinara Guseinova,15 Claudia Bracaglia,16 Kristen Hayward,17 Carine Wouters,18 Alexei Grom,19
Marina Vivarelli,16 Alberto Fischer,20 Luciana Breda,21 Alberto Martini,22 and Angelo Ravelli,22
for the Lupus Working Group of the Paediatric Rheumatology European Society

Objective. To describe the clinical and laboratory features of macrophage activation syndrome as a com-
plication of juvenile systemic lupus erythematosus
Dr Pringe is recipient of an Alpha Scholarship from the (SLE).
European Union (contract no. AML/B7-311/970666/II-0246-FI). Methods. Cases of juvenile SLE–associated mac-
1
Alessandro Parodi, MD, Sergio Davı̀, MD, Angela Pistorio, rophage activation syndrome were provided by investi-
MD, Nicolino Ruperto, MD, MPH: Istituto di Ricovero e Cura a
Carattere Scientifico G. Gaslini, Genoa, Italy; 2Alejandra Beatriz gators belonging to 3 pediatric rheumatology networks
Pringe, MD: Hospital General de Ninos Pedro de Elizalde, Buenos or were found in the literature. Patients who had
Aires, Argentina; 3Silvia Magni-Manzoni, MD: IRCCS Fondazione evidence of macrophage hemophagocytosis on bone
Policlinico S. Matteo, Pavia, Italy; 4Paivi Miettunen, MD: Alberta
Children’s Hospital, Calgary, Canada; 5Brigitte Bader-Meunier, MD: marrow aspiration were considered to have definite
Assistance Publique-Hopitaux de Paris, Hopital Necker Enfants macrophage activation syndrome, and those who did not
Malades, Paris, France; 6Graciela Espada, MD: Hospital de Ninos have such evidence were considered to have probable
Ricardo Gutierrez, Buenos Aires, Argentina; 7Gary Sterba, MD:
Hospital de Clinicas Caracas, Caracas, Venezuela; 8Seza Ozen, MD: macrophage activation syndrome. Clinical and labora-
Hacettepe University Children’s Hospital, Ankara, Turkey; 9Dowain tory findings in patients with macrophage activation
Wright, MD: Children’s Hospital Central California, Madera, Califor- syndrome were contrasted with those of 2 control groups
nia; 10Claudia Saad Magalhães, MD: Hospital das Clinicas, Faculdade
de Medicina de Botucatu, UNESP, Botucatu, Brazil; 11Raju Khub- composed of patients with active juvenile SLE without
chandani, MD: Jaslok Hospital and Research Centre, Mumbai, India; macrophage activation syndrome. The ability of each
12
Hartmut Michels, MD: German Center of Pediatric Rheumatology, feature to discriminate macrophage activation syn-
Garmisch-Partenkirchen, Germany; 13Patricia Woo, MD: Great Or-
mond Street Hospital, London, UK; 14Antonio Iglesias, MD: Univer- drome from active disease was evaluated by calculating
sidad Nacional de Colombia, Bogota, Colombia; 15Dinara Guseinova, sensitivity, specificity, and area under the receiver op-
MD: Children’s Clinical University Hospital, Riga, Latvia; 16Claudia erating characteristic curve.
Bracaglia, MD, Marina Vivarelli, MD: Ospedale Pediatrico Bambino
Gesù, Rome, Italy; 17Kristen Hayward, MD: Seattle Children’s Hos-
Results. The study included 38 patients (20 with
pital, Seattle, Washington; 18Carine Wouters, MD: University Hospital definite macrophage activation syndrome and 18 with
Gasthuisberg, Leuven, Belgium; 19Alexei Grom, MD: Cincinnati Chil- probable macrophage activation syndrome). Patients
dren’s Hospital Medical Center, Cincinnati, Ohio; 20Alberto Fischer,
MD: U.O. Pediatria, Ospedale di Acireale (CT), Acireale, Italy;
with definite and probable macrophage activation syn-
21
Luciana Breda, MD: Università degli Studi di Chieti, Chieti, Italy; drome were comparable with regard to all clinical and
22
Alberto Martini, MD, Angelo Ravelli, MD: Istituto di Ricovero e laboratory features of the syndrome, except for a greater
Cura a Carattere Scientifico G. Gaslini and Università degli Studi di
Genoa, Genoa, Italy.
frequency of lymphadenopathy, leukopenia, and throm-
Address correspondence and reprint requests to Angelo
Ravelli, MD, Pediatria II, Istituto G. Gaslini, Largo G. Gaslini 5, 16147 Submitted for publication March 9, 2009; accepted in revised
Genoa, Italy. E-mail: angeloravelli@ospedale-gaslini.ge.it. form July 11, 2009.

3388
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3389

bocytopenia in patients with definite macrophage acti- condition that can follow a rapidly fatal course, prompt
vation syndrome. Overall, clinical features had better recognition of its clinical and laboratory features and
specificity than sensitivity, except for fever, which was immediate therapeutic intervention are imperative.
highly sensitive but had low specificity. Among labora- However, the diagnosis of macrophage activation syn-
tory features, the best sensitivity and specificity was drome in patients with SLE may be challenging because
achieved using hyperferritinemia, followed by increased it may mimic the clinical features of the underlying
levels of lactate dehydrogenase, hypertriglyceridemia, disease or be confused with an infectious complication.
and hypofibrinogenemia. Based on the results of statis- Differentiation of macrophage activation syndrome
tical analysis, preliminary diagnostic guidelines for from these conditions is critical to select the appropriate
macrophage activation syndrome in juvenile SLE were therapeutic approach. Recently, preliminary diagnostic
developed. guidelines for macrophage activation syndrome as a
Conclusion. Our findings indicate that the occur- complication of systemic JIA have been developed (13).
rence of unexplained fever and cytopenia, when associ- However, it is unclear whether these guidelines may be
ated with hyperferritinemia, in a patient with juvenile applied to patients with juvenile SLE. Other potentially
SLE should raise the suspicion of macrophage activa- suitable diagnostic guidelines are those developed for
tion syndrome. We propose preliminary guidelines for hemophagocytic lymphohistiocytosis (14).
this syndrome in juvenile SLE to facilitate timely diag- Because little information exists on macrophage
nosis and correct classification of patients. activation syndrome in juvenile SLE, we undertook a
multinational, multicenter collaborative study, with the
Macrophage activation syndrome is a severe, primary aim of describing the clinical and laboratory
potentially life-threatening complication of childhood features of this complication in patients with juvenile
systemic inflammatory disorders. The hallmark of this syn- SLE. Secondary objectives were to investigate whether
drome is excessive activation and proliferation of T this complication may be regarded as underdiagnosed
lymphocytes and macrophages with massive hypersecre- in juvenile SLE and whether the diagnostic guidelines
tion of proinflammatory cytokines, including interleukin- for systemic JIA–associated macrophage activation syn-
1␤ (IL-1␤), IL-6, interferon-␥, and tumor necrosis factor drome or hemophagocytic lymphohistiocytosis may be
␣. Macrophage activation syndrome may occur sponta- used to identify macrophage activation syndrome in
neously, as a complication of active underlying disease, patients with juvenile SLE. An additional purpose of the
or may be triggered by an infection or a change in study was to attempt development of diagnostic guide-
therapy. Clinically, patients with macrophage activation lines for macrophage activation syndrome as a compli-
syndrome present with nonremitting high fever, pan- cation of juvenile SLE.
cytopenia, hepatosplenomegaly, hepatic dysfunction, en-
cephalopathy, coagulation abnormalities, and sharply
increased levels of ferritin. The pathognomonic feature PATIENTS AND METHODS
of the syndrome is seen on bone marrow examination, Patient selection. Clinical information on patients with
which reveals numerous morphologically benign macro- juvenile SLE–associated macrophage activation syndrome
phages actively phagocytosing hematopoietic cells (1–4). was collected from several sources. All investigators belonging
Because macrophage activation syndrome bears a close to the Italian Pediatric Rheumatology Study Group (IPRSG),
the Pediatric Rheumatology International Trials Organization
resemblance to the group of hemophagocytic lympho- (PRINTO), and the Pediatric Rheumatology Collaborative
histiocytosis syndromes, it is currently classified among Study Group (PRCSG) were contacted by e-mail and asked
the secondary, or acquired, forms of hemophagocytic whether they had seen any cases of macrophage activation
lymphohistiocytosis (5,6). syndrome in juvenile SLE and, if so, whether they were willing
Among pediatric rheumatic disorders, macrophage to enroll their patients in the study. Those who responded
positively were asked to complete a structured case report
activation syndrome occurs much more frequently, for form with each patient’s anonymous data and to send the
unknown reasons, in systemic juvenile idiopathic arthri- completed form to the coordinating center (Istituto G.
tis (JIA) (1–4). However, in recent years this syndrome Gaslini). In addition, the clinical database at the coordinating
has been increasingly reported in patients with juvenile center was scrutinized to identify patients with juvenile SLE
systemic lupus erythematosus (SLE) (7–11). Further- who had macrophage activation syndrome. Finally, a system-
atic review of the literature was conducted to identify pub-
more, it has been suggested that macrophage activation lished cases of macrophage activation syndrome in patients
syndrome in juvenile SLE may be underrecognized (12). with juvenile SLE with sufficient information available. The
Because macrophage activation syndrome is a serious Medline database was searched using a strategy that included
3390 PARODI ET AL

the following medical subject headings: systemic lupus ery- Statistical analysis. Comparison of the frequencies of
thematosus, children, childhood, pediatric, macrophage activa- demographic and clinical features and of laboratory findings
tion syndrome, and hemophagocytic syndrome. This screening between patients with definite macrophage activation syndrome
was supplemented by a manual search of references in the and patients with probable macrophage activation syndrome
articles. and between patients with definite macrophage activation
To be included in the study, patients had to have been syndrome and patients with probable macrophage activation
diagnosed as having SLE according to the American College of syndrome combined and controls was made by chi-square test
Rheumatology (ACR) 1997 revised criteria (15), have been or Fisher’s exact test in cases of expected frequencies of ⬍5.
younger than 18 years at diagnosis, and have had an episode of Comparison of mean values of laboratory parameters was
macrophage activation syndrome diagnosed and treated as performed using the Mann-Whitney U test. The ability of each
such by the attending physician. The diagnosis of macrophage feature to discriminate instances of macrophage activation
activation syndrome had to be based on the typical clinical and syndrome from instances of active disease was evaluated by
laboratory picture of the syndrome, irrespective of evidence of calculating sensitivity, specificity, and area under the receiver
macrophage hemophagocytosis in the bone marrow aspirate. operating characteristic (ROC) curve (16). The discriminative
However, patients who had such evidence were considered to ability of laboratory tests was assessed using either the stan-
have definite macrophage activation syndrome, whereas those dard threshold, i.e., the threshold reported in the literature or
who lacked it because bone marrow aspiration was not per- judged to be clinically meaningful, or the best threshold, i.e.,
formed or did not show hemophagocytosis were considered to the threshold obtained through the ROC curve analysis that
have probable macrophage activation syndrome. Patients who produced the most appropriate tradeoff between sensitivity
had an infection at the time of macrophage activation syn- and specificity.
drome were not excluded because it is known that infections To devise the preliminary diagnostic guidelines for
are common triggers of this syndrome in patients with rheu- macrophage activation syndrome as a complication of juvenile
matic diseases (1). SLE, we sought the best classification/diagnostic rule. This
Control groups. To identify the clinical and laboratory requires making a specific tradeoff between sensitivity and
features with the greatest sensitivity and specificity for the specificity, through changing the definition of a positive. To
diagnosis of macrophage activation syndrome in juvenile SLE, reach this goal, we used the “number of criteria present”
we followed the classification criteria approach, as was done approach, which is done by varying l, the minimum number of
for the development of preliminary diagnostic guidelines for criteria required to be present for a patient to be classified as
macrophage activation syndrome as a complication of systemic positive (16). In other words, if any l or more of a list of criteria
JIA (13,16,17). The purpose of this approach is to separate are present in a patient, then the patient can be classified as
patients with a particular disease from patients without the positive. Notably, all criteria must be able to be judged as being
disease. Ideally, classification criteria have high sensitivity for either present or absent (that is, they must be dichotomous) to
the disease in question and high specificity against other allow using this method. The lower the decision threshold, the
diseases (that is, a high proportion of patients with the disease larger the number of patients that will be judged to be positive,
are found to be positive and a high proportion of patients resulting in high sensitivity and low specificity. Conversely, if
without the disease are found to be negative). These criteria the threshold chosen is high, then more patients will be judged
are generally created by comparing patient groups with the to be negative, resulting in low sensitivity and high specificity.
index disease with control patients who have a “confusable” By varying the threshold, a table can be produced that allows
disease. In our study, the index disease was represented by for the selection of the combination of variables that shows the
macrophage activation syndrome as a complication of juvenile “best” diagnostic accuracy. For each combination of variables
SLE and the confusable condition by active juvenile SLE tested, sensitivity, specificity, and diagnostic odds ratio (OR)
without macrophage activation syndrome. (20) were calculated.
Two control groups of patients with active SLE without
macrophage activation syndrome were selected. The first group
consisted of 29 consecutive patients seen at the Istituto G. RESULTS
Gaslini (IGG control sample) between 2002 and 2006 who had
33 instances of active disease. Active disease was defined as A total of 38 patients with juvenile SLE and
either the time of diagnosis, before the start of a disease- macrophage activation syndrome were included in the
specific treatment; or the first disease flare, defined as an study; 14 patients were enrolled by PRINTO investiga-
increase in the SLE Disease Activity Index 2000 score of ⱖ3 tors, 7 patients were enrolled by IPRSG investigators, 5
points compared with a previous assessment (18), requiring an patients were enrolled by PRCSG investigators, 6 pa-
increase in prednisone dosage of ⬎0.5 mg/kg/day or 20 mg/day
or the start of cyclophosphamide therapy. The second group tients were seen at the coordinating center, and infor-
was composed of 387 patients enrolled in a multinational study mation on 6 patients was found in the literature (9–11).
to investigate cumulative damage in juvenile SLE (multi- Twenty patients had definite macrophage activation
national control sample) (19). These patients had clinical syndrome, and 18 patients had probable macrophage
manifestations recorded within the first month after disease activation syndrome.
onset, a time period that was thought to imply the presence of
active disease. Laboratory findings were not available in this The main demographic features and the fre-
control group. The study protocol was approved by the Ethics quency of ACR criteria for SLE in patients with macro-
Committee of the Istituto G. Gaslini. phage activation syndrome and controls at the time of
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3391

Table 1. Demographic characteristics and frequency of ACR criteria at diagnosis in patients with definite and probable macrophage activation
syndrome and in patients with active juvenile SLE without macrophage activation syndrome*
Patients with Patients with
definite probable All patients with
macrophage macrophage macrophage
activation activation activation Multinational
syndrome syndrome syndrome IGG controls controls
(n ⫽ 20)† (n ⫽ 18) (n ⫽ 38)‡ (n ⫽ 29) P§ (n ⫽ 387)¶ P#
Female 18 (90.0) 16 (88.9) 34 (89.5) 23 (79.3) 0.31 330 (85.3) 0.48
White 9 (69.2) 14 (77.8) 23 (74.2) 29 (100.0) 0.005 140 (36.2) ⬍0.0001
Age at diagnosis of SLE, mean 12.6 ⫾ 3.7 12.7 ⫾ 3.4 12.6 ⫾ 3.5 12.2 ⫾ 2.7 0.46 12.1 ⫾ 3.4 0.39
⫾ SD years
Time from SLE diagnosis to 0.4 ⫾ 0.9 1.0 ⫾ 1.7 0.7 ⫾ 1.3 – – – –
onset of macrophage
activation syndrome, mean ⫾
SD years**
ACR criteria at diagnosis
Malar rash 13 (68.4) 12 (66.7) 25 (67.6) 24 (82.8) 0.16 245 (63.3) 0.61
Discoid rash 1 (5.3) 2 (11.1) 3 (8.1) 1 (3.4) 0.62 20 (5.2) 0.44
Photosensitivity 8 (42.1) 5 (27.8) 13 (35.1) 7 (24.1) 0.33 118 (30.5) 0.56
Oral or nasal ulcers 8 (42.1) 11 (61.1) 19 (51.4) 7 (24.1) 0.01 77 (19.9) ⬍0.0001
Arthritis 14 (73.7) 13 (72.2) 27 (73.0) 19 (65.5) 0.51 200 (51.7) 0.013
Nephritis 15 (79.0) 11 (61.1) 26 (70.3) 8 (27.6) 0.0006 195 (50.5) 0.022
CNS disease 5 (26.3) 4 (22.2) 9 (24.3) 0 (0.0) 0.0036 28 (7.2) 0.0004
Serositis 9 (47.4) 10 (55.6) 19 (51.4) 5 (17.2) 0.004 53 (13.7) ⬍0.0001
Hematologic involvement 18 (94.7) 16 (88.9) 34 (91.9) 21 (72.4) 0.05 271 (70.0) 0.005
Positive immunoserology 19 (95.0) 18 (100.0) 37 (97.4) 27 (93.1) 0.57 351 (90.7) 0.16
Antinuclear antibody positive 20 (100.0) 18 (100.0) 38 (100.0) 29 (100.0) – 382 (98.7) 1.00

* Except where indicated otherwise, values are the number (%). There were no significant differences between patients with definite macrophage
activation syndrome and patients with probable macrophage activation syndrome. ACR ⫽ American College of Rheumatology; – ⫽ not applicable.
† Data were available for 13 patients for race and for 19 patients for malar rash, discoid rash, photosensitivity, oral or nasal ulcers, arthritis, nephritis,
central nervous system (CNS) disease, serositis, and hematologic involvement.
‡ Data were available for 31 patients for race and for 37 patients for malar rash, discoid rash, photosensitivity, oral or nasal ulcers, arthritis, nephritis,
CNS disease, serositis, and hematologic involvement.
§ All patients with macrophage activation syndrome versus Istituto G. Gaslini (IGG) controls, by Fisher’s exact test for sex, race, frequency of discoid
rash, frequency of CNS disease, and immunoserology (positivity for anti–double-stranded DNA, anti-Sm, or antiphospholipid antibodies), by the
Mann-Whitney U test for age at diagnosis, and by chi-square test for all other parameters.
¶ Data were available for 382 patients for age at diagnosis and for 386 patients for nephritis.
# All patients with macrophage activation syndrome versus multinational controls, by Student’s t-test for age at diagnosis, by Fisher’s exact test for
frequency of discoid rash and antinuclear antibodies, and by chi-square test for all other parameters.
** One patient was excluded from the probable macrophage activation syndrome group because macrophage activation syndrome occurred before
diagnosis of juvenile systemic lupus erythematosus (SLE).

diagnosis are shown in Table 1. Patients with definite Table 2 shows the frequency of typical clinical
macrophage activation syndrome and patients with features of macrophage activation syndrome, fulfillment
probable macrophage activation syndrome were compa- of systemic JIA–associated macrophage activation syn-
rable with regard to all demographic features and the drome or hemophagocytic lymphohistiocytosis diagnos-
frequency of all ACR criteria at diagnosis. All patients tic criteria, occurrence of macrophage activation syn-
with macrophage activation syndrome (patients with drome within 1 or 6 months after diagnosis of juvenile
definite macrophage activation syndrome and patients SLE, triggering factors, admission to the intensive care
with probable macrophage activation syndrome com- unit (ICU), death, and therapeutic interventions in
bined) were comparable with both control groups with patients with macrophage activation syndrome and in
regard to the proportion of female patients and mean control groups. All clinical features were comparable
age at diagnosis of juvenile SLE. Compared with control between patients with definite and probable macro-
groups, patients with macrophage activation syndrome phage activation syndrome, except lymphadenopathy,
had greater frequencies of the ACR criteria oral/nasal which was more common in the former group. The same
ulcers, nephritis, central nervous system (CNS) disease, clinical features were much more common in all patients
arthritis, serositis, and hematologic involvement at diag- with macrophage activation syndrome combined than in
nosis. the control groups.
3392 PARODI ET AL

Table 2. Frequency of clinical features of macrophage activation syndrome in patients with definite and probable macrophage activation syndrome
and in patients with active juvenile SLE without macrophage activation syndrome*
Patients with Patients with
definite probable All patients with
macrophage macrophage macrophage
activation activation activation Multinational
syndrome syndrome syndrome IGG controls controls
(n ⫽ 20)† (n ⫽ 18)‡ (n ⫽ 38)§ (n ⫽ 33)¶ P# (n ⫽ 386) P**
Fever 19 (95.0) 15 (83.3) 34 (89.5) 7 (21.2) ⬍0.0001 248 (64.2) 0.002
Hepatomegaly 9 (47.4) 10 (55.6) 19 (51.4) 4 (12.1) 0.0005 40 (10.4) ⬍0.0001
Splenomegaly 8 (42.1) 6 (33.3) 14 (37.8) 5 (15.1) 0.033 32 (8.3) ⬍0.0001
Lymphadenopathy 14 (70.0) 6 (33.3) 20 (52.6) 11 (33.3) 0.10 69 (17.9) ⬍0.0001
Hemorrhagic manifestations 8 (40.0) 6 (33.3) 14 (36.8) 3 (9.1) 0.006 – –
CNS disease 8 (40.0) 6 (33.3) 14 (36.8) 1 (3.0) 0.0005 33 (8.5) ⬍0.0001
Patients with macrophage 13 (65.0) 11 (61.1) 24 (63.2) – – – –
activation syndrome
within 1 month after
diagnosis
Patients with macrophage 17 (85.0) 12 (66.7) 29 (76.3) – – – –
activation syndrome
within 6 months after
diagnosis
Triggers 11 (68.8) 15 (83.3) 26 (76.5) – – – –
Active disease/flare 9 (81.8) 12 (80.0) 21 (80.8) – – – –
Infection 4 (36.4) 3 (20.0) 7 (26.9) – – – –
Therapeutic change 1 (9.1) 0 1 (3.8) – – – –
ICU admission 8 (57.2) 6 (33.3) 14 (43.7) – – – –
Death 2 (10.5) 2 (12.5) 4 (11.4) – – – –
Treatment
Corticosteroids 20 (100.0) 18 (100.0) 38 (100.0) – – – –
Cyclosporine 7 (36.8) 7 (38.9) 14 (37.8) – – – –
IV immunoglobulin 9 (47.4) 3 (16.7) 12 (32.4) – – – –
Cyclophosphamide 4 (21.1) 4 (22.2) 8 (21.6) – – – –
Azathioprine 1 (5.3) 3 (16.7) 4 (10.8) – – – –
Plasma exchange 3 (15.8) 1 (5.6) 4 (10.8) – – – –
Rituximab 2 (10.5) 0 (0.0) 2 (5.4) – – – –
Mycophenolate mofetil 0 (0.0) 1 (5.6) 1 (2.7) – – – –
Etoposide 1 (5.3) 0 (0.0) 1 (2.7) – – – –
Patients meeting systemic 19 (100) 17 (100) 36 (100) 22 (71.0) 0.0005 – –
JIA–associated criteria
Patients meeting 13 (81.3) 9 (52.9) 22 (66.7) 0 (0.0) ⬍0.0001 – –
hemophagocytic
lymphohistiocytosis
criteria

* Values are the number (%). The only significant difference between patients with definite macrophage activation syndrome and patients with
probable macrophage activation syndrome was for lymphadenopathy (P ⫽ 0.024). SLE ⫽ systemic lupus erythematosus; – ⫽ not applicable.
† Data were available for 19 patients for hepatomegaly and splenomegaly, for 16 patients for triggers, for 11 patients for active disease/flare,
infection, and therapeutic changes, for 14 patients for intensive care unit (ICU) admission, for 19 patients for death, for 19 patients for cyclosporine,
intravenous (IV) immunoglobulin, cyclophosphamide, azathioprine, plasma exchange, rituximab, mycophenolate mofetil, and etoposide, for 19
patients for systemic juvenile idiopathic arthritis (JIA) criteria, and for 16 patients for hemophagocytic lymphohistiocytosis criteria.
‡ Data were available for 15 patients for the triggers active disease/flare and infection, for 16 patients for death, and for 17 patients for systemic JIA
criteria and hemophagocytic lymphohistiocytosis criteria.
§ Data were available for 37 patients for hepatomegaly and splenomegaly, for 34 patients for triggers, for 26 patients for active disease/flare,
infection, and therapeutic change, for 32 patients for ICU admission, for 35 patients for death, for 37 patients for cyclosporine, IV immunoglobulin,
cyclophosphamide, azathioprine, plasma exchange, rituximab, mycophenolate mofetil, and etoposide, for 36 patients for systemic JIA criteria, and
for 33 patients for hemophagocytic lymphohistiocytosis criteria.
¶ The n value represents the number of assessments. (Some patients were assessed at both time of diagnosis and first disease flare.) Data were
available for 31 assessments for systemic JIA criteria and for 32 assessments for hemophagocytic lymphohistiocytosis criteria.
# All patients with macrophage activation syndrome versus Istituto G. Gaslini (IGG) controls, by chi-square test.
** All patients with macrophage activation syndrome versus multinational controls, by Fisher’s exact test for central nervous system (CNS) disease
and by chi-square test for all other parameters.

Overall, clinical features had better specificity tional control group. Hepatomegaly, splenomegaly,
than sensitivity, except for fever, which was highly hemorrhages, and CNS dysfunction were effective in
sensitive but had poor specificity versus the multina- discriminating macrophage activation syndrome from
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3393

Table 3. Laboratory findings in patients with definite and probable macrophage activation syndrome and in patients with active juvenile SLE
without macrophage activation syndrome*
Patients with Patients with
definite probable All patients with
macrophage macrophage macrophage
activation activation activation
syndrome syndrome syndrome IGG controls
(n ⫽ 20) (n ⫽ 18) (n ⫽ 38) (n ⫽ 33)† P‡
White blood cell count, ⫻ 10 /liter
9
2.4 ⫾ 1.4 4.5 ⫾ 2.4 3.4 ⫾ 2.1 4.4 ⫾ 3.1 0.12
Hemoglobin, gm/liter 79 ⫾ 16 88 ⫾ 18 83 ⫾ 17 109 ⫾ 20 ⬍0.0001
Platelet count, ⫻ 109/liter 95.7 ⫾ 78.3 146.3 ⫾ 99.3 119.7 ⫾ 91.3 221.4 ⫾ 109.1 ⬍0.0001
Aspartate aminotransferase, units/liter 327.1 ⫾ 376.7 170.9 ⫾ 129.6 246.5 ⫾ 284.4 38.5 ⫾ 30.7 ⬍0.0001
Alanine aminotransferase, units/liter 131.3 ⫾ 143.6 192.4 ⫾ 296.1 162.9 ⫾ 233.3 52.1 ⫾ 45.0 0.0002
Bilirubin, mg/dl 0.6 ⫾ 0.4 1.1 ⫾ 1.7 0.9 ⫾ 1.4 0.4 ⫾ 0.2 0.09
Lactate dehydrogenase, units/liter 1,953.5 ⫾ 1,649.9 1,254.9 ⫾ 624.7 1,604.2 ⫾ 1,277.5 430.5 ⫾ 134.0 ⬍0.0001
Albumin, gm/dl 2.3 ⫾ 0.5 2.7 ⫾ 0.7 2.5 ⫾ 0.6 3.7 ⫾ 0.7 ⬍0.0001
Fibrinogen, gm/liter 2.03 ⫾ 0.98 2.25 ⫾ 1.13 2.13 ⫾ 1.03 3.61 ⫾ 1.13 ⬍0.0001
Triglycerides, mg/dl 397.2 ⫾ 412.3 428.9 ⫾ 228.3 413.5 ⫾ 325.6 132.6 ⫾ 99.0 ⬍0.0001
Serum sodium, mmoles/liter 134.1 ⫾ 8.6 132.7 ⫾ 5.4 133.3 ⫾ 6.8 136.8 ⫾ 3.2 0.0054
Ferritin, ␮g/liter 3,829.9 ⫾ 5,039.1 2,071.8 ⫾ 2,603.4 2,840.9 ⫾ 3,892.4 84.6 ⫾ 78.0 ⬍0.0001

* Values are the mean ⫾ SD. The only significant difference between patients with definite macrophage activation syndrome and patients with
probable macrophage activation syndrome was for white blood cell count (P ⫽ 0.03). SLE ⫽ systemic lupus erythematosus.
† The n value represents the number of assessments. (Some patients were assessed at both time of diagnosis and first disease flare.)
‡ All patients with macrophage activation syndrome versus Istituto G. Gaslini (IGG) controls, by Mann-Whitney U test.

active disease. However, their sensitivity was low to poside were used in 12, 4, 2, and 1 patients, respectively.
moderate (data not shown). Diagnostic guidelines for Overall, frequency of therapeutic choices was similar in
systemic JIA–associated macrophage activation syn- patients with definite macrophage activation syndrome
drome (13) were fulfilled by all patients with macro- and patients with probable macrophage activation syn-
phage activation syndrome and by 22 (71%) of 31 IGG drome.
control patients. Hemophagocytic lymphohistiocytosis The mean values of laboratory parameters of
diagnostic criteria (14) were met by 22 (66.7%) of 33 macrophage activation syndrome in patients with mac-
patients with macrophage activation syndrome and by rophage activation syndrome and in control samples are
none of the IGG control patients. Diagnostic guidelines presented in Table 3. Patients with definite and probable
could not be assessed in the multinational control sam- macrophage activation syndrome were comparable with
ple due to the lack of laboratory data. regard to all laboratory values, with the exception of
As many as 63.2% and 76.3% of patients with white blood cell count, which was lower in the former
macrophage activation syndrome developed this syn- group. All laboratory values were markedly worse in all
drome within 1 and 6 months, respectively, after diag- patients with macrophage activation syndrome com-
nosis of juvenile SLE. A triggering factor was suspected bined than in the IGG control group, except white blood
in 76.5% of instances, with macrophage activation syn- cell count and bilirubin, which were comparable be-
drome occurring much more commonly in the setting tween the 2 groups.
of disease activity or flare; an associated infection was Table 4 shows the frequency of laboratory fea-
reported in 26.9% of instances. Fourteen (43.7%) of 32 tures of macrophage activation syndrome in patients
patients had to be admitted to the ICU, and 4 (11.4%) of with macrophage activation syndrome and in control
35 patients died (2 of multiple organ failure, 1 of acute groups. Patients with definite macrophage activation
respiratory distress syndrome, and 1 of pneumococcal syndrome and patients with probable macrophage acti-
sepsis). The frequencies of these events were comparable vation syndrome were comparable with regard to all
between the definite and probable macrophage activa- laboratory features, with the exception of a greater
tion syndrome groups. All patients received systemic frequency of leukopenia and a borderline greater fre-
corticosteroid therapy, most frequently intravenously. Cy- quency of thrombocytopenia in patients with definite
closporine was the most commonly administered immuno- macrophage activation syndrome. As with the mean
suppressive medication, followed by cyclophosphamide, values for the laboratory parameters described above,
azathioprine, and mycophenolate mofetil. Intravenous the frequency of laboratory abnormalities was much
immunoglobulin, plasma exchange, rituximab, and eto- greater in all patients with macrophage activation syn-
3394 PARODI ET AL

Table 4. Frequency of laboratory features of macrophage activation syndrome in patients with definite and probable macrophage activation
syndrome and in patients with active juvenile SLE without macrophage activation syndrome*
Patients with Patients with
definite probable All patients with
macrophage macrophage macrophage
activation activation activation
syndrome syndrome syndrome IGG controls
(n ⫽ 20)† (n ⫽ 18)‡ P§ (n ⫽ 38)¶ (n ⫽ 33)# P**
White blood cell count ⱕ4.0 ⫻ 10 /liter
9
18 (90.0) 8 (44.4) 0.003 26 (68.4) 21 (63.6) 0.67
Hemoglobin ⱕ90 gm/liter 14 (70.0) 10 (55.6) 0.36 24 (63.2) 5 (15.2) ⬍0.0001
Platelet count ⱕ150 ⫻ 109/liter 18 (90.0) 11 (61.1) 0.04 29 (76.3) 6 (18.2) ⬍0.0001
Aspartate aminotransferase ⬎40 units/liter 12 (80.0) 15 (93.8) 0.27 27 (87.1) 10 (30.3) ⬍0.0001
Alanine aminotransferase ⬎40 units/liter 11 (78.6) 14 (87.5) 0.43 25 (80.6) 15 (45.5) 0.0037
Bilirubin ⬎1.0 mg/dl 1 (12.5) 3 (21.4) 0.53 4 (18.2) 0 (0.0) 0.12
Lactate dehydrogenase ⬎400 units/liter 15 (93.8) 15 (93.8) 0.76 30 (93.8) 12 (52.2) 0.0003
Albumin ⱕ3.0 gm/dl 10 (100.0) 9 (64.3) 0.05 19 (79.2) 5 (20.0) ⬍0.0001
Fibrinogen ⱕ1.5 gm/liter 6 (37.5) 6 (42.9) 0.76 12 (40.0) 0 (0.0) 0.0003
Triglycerides ⬎160 mg/dl 12 (75.0) 15 (88.2) 0.30 27 (81.8) 3 (20.0) ⬍0.0001
Serum sodium ⬍135 mmoles/liter 7 (63.6) 10 (66.7) 0.60 17 (65.4) 5 (19.2) 0.0008
Ferritin ⬎500 ␮g/liter 13 (92.9) 17 (94.4) 0.69 30 (93.8) 0 (0.0) ⬍0.0001

* Values are the number (%). SLE ⫽ systemic lupus erythematosus; IGG ⫽ Istituto G. Gaslini.
† Data were available for 15 patients for aspartate aminotransferase, 14 patients for alanine aminotransferase, 8 patients for bilirubin, 16 patients
for lactate dehydrogenase, 10 patients for albumin, 16 patients for fibrinogen and triglycerides, 11 patients for serum sodium, and 14 patients for
ferritin.
‡ Data were available for 16 patients for aspartate aminotransferase and alanine aminotransferase, 14 patients for bilirubin, 16 patients for lactate
dehydrogenase, 14 patients for albumin and fibrinogen, 17 patients for triglycerides, and 15 patients for serum sodium.
§ Patients with definite macrophage activation syndrome versus patients with probable macrophage activation syndrome, by chi-square test for white
blood cell count, hemoglobin, and fibrinogen and by Fisher’s exact test for all other parameters.
¶ Data were available for 31 patients for aspartate aminotransferase and alanine aminotransferase, for 22 patients for bilirubin, for 32 patients for
lactate dehydrogenase, for 24 patients for albumin, for 30 patients for fibrinogen, for 33 patients for triglycerides, for 26 patients for serum sodium,
and for 32 patients for ferritin.
# The n value represents the number of assessments. (Some patients were assessed at both time of diagnosis and first disease flare.) Data were
available for 17 assessments for bilirubin, for 23 assessments for lactate dehydrogenase, for 25 assessments for albumin, for 26 assessments for
fibrinogen, for 15 assessments for triglycerides, for 26 assessments for serum sodium, and for 24 assessments for ferritin.
** All patients with macrophage activation syndrome versus controls, by Fisher’s exact test for bilirubin and by chi-square test for all other
parameters.

drome combined than in the IGG control group, with superior to alanine aminotransferase. An examination of
the exception of a comparable frequency of leukopenia the abnormalities of the 3 blood cell lines (leukopenia,
and bilirubin increase. anemia, and thrombocytopenia) in various combinations
The sensitivity, specificity, and area under the revealed that the greatest sensitivity and specificity (both
ROC curve for each laboratory feature, assessed using ⬃80%) were obtained when any 2 of the 3 abnormalities
the standard threshold or the best threshold obtained were simultaneously present (data not shown).
through ROC curve analysis, are shown in Table 5. The frequencies of traditional laboratory indica-
Platelet count, liver transaminases, serum albumin, tri- tors of SLE activity, including erythrocyte sedimentation
glycerides, serum sodium, and ferritin yielded similar rate (ESR), C-reactive protein (CRP) level, C3, C4, and
levels of sensitivity and specificity using either the antinuclear and anti-DNA antibodies, were comparable
standard or the best threshold. White blood cell count, between patients with definite macrophage activation
hemoglobin, bilirubin, lactate dehydrogenase, and fi- syndrome and patients with probable macrophage acti-
brinogen yielded different levels of sensitivity and spec- vation syndrome. All patients with macrophage activa-
ificity depending on whether the standard or best thresh- tion syndrome combined had a borderline lower fre-
old was used. Overall, hyperferritinemia had the best quency of increase in ESR, but had a much greater
sensitivity and specificity, followed by increased lactate frequency of increase in CRP level compared with the
dehydrogenase level, hypertriglyceridemia, and hypo- IGG control group. The frequencies of hypocomple-
fibrinogenemia. Thrombocytopenia was a better indica- mentemia and antinuclear antibody positivity were com-
tor of macrophage activation syndrome than leukopenia parable between patients with macrophage activation
or anemia, and aspartate aminotransferase was slightly syndrome and patients with active SLE without macro-
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3395

Table 5. Sensitivity and specificity of laboratory parameters analyzed for ability to discriminate macrophage activation syndrome from active
juvenile SLE without macrophage activation syndrome*
Standard Best
threshold Sensitivity Specificity threshold Sensitivity Specificity AUC (95% CI)
White blood cell count, ⫻ 10 /liter
9
ⱕ4.0 68.4 36.4 ⱕ1.9 34.2 100 0.61 (0.48–0.72)
Hemoglobin, gm/liter ⱕ90 63.2 84.8 ⱕ112 100 57.6 0.84 (0.74–0.92)
Platelet count, ⫻ 109/liter ⱕ150 76.3 81.8 ⱕ149 76.3 81.8 0.77 (0.66–0.86)
Aspartate aminotransferase, units/liter ⬎40 87.1 69.7 ⬎33 93.5 66.7 0.87 (0.76–0.94)
Alanine aminotransferase, units/liter ⬎40 83.3 54.5 ⬎48 80.6 66.7 0.77 (0.65–0.86)
Bilirubin, mg/dl ⬎1.00 18.2 100 ⬎0.34 81.8 52.9 0.66 (0.49–0.81)
Lactate dehydrogenase, units/liter ⬎400 93.8 47.8 ⬎567 90.6 95.7 0.94 (0.84–0.99)
Albumin, gm/dl ⱕ3.0 79.2 80.0 ⱕ3.4 91.7 68.0 0.87 (0.74–0.95)
Fibrinogen, mg/dl ⱕ150 40.0 100 ⱕ290 80.0 76.9 0.84 (0.71–0.92)
Triglycerides, mg/dl ⬎160 81.8 80.0 ⬎178 78.8 93.3 0.87 (0.75–0.95)
Serum sodium, mmoles/liter ⬍135 65.4 80.8 ⬍135 65.4 80.8 0.72 (0.58–0.84)
Ferritin, ␮g/liter ⬎500 93.8 100 ⬎249 96.4 100 0.99 (0.92–0.99)

* Sensitivity and specificity were obtained for the standard threshold (the threshold reported in the literature or judged to be clinically meaningful)
and for the best threshold (the threshold obtained through the receiver operating characteristic curve analysis that produced the most appropriate
tradeoff between sensitivity and specificity). SLE ⫽ systemic lupus erythematosus; AUC ⫽ area under the curve; 95% CI ⫽ 95% confidence interval.

phage activation syndrome, whereas the latter group had of 90.9%, and a diagnostic OR of 116.7 (95% confi-
a borderline greater prevalence of positive anti-DNA dence interval 21.9–621.6). Based on these results,
antibodies (data not shown). we set up diagnostic guidelines for macrophage acti-
Using the “number of criteria present” approach, vation syndrome as a complication of juvenile SLE
we sought the combination of clinical and laboratory (Table 6).
variables that had the greatest diagnostic accuracy, that
is, the best ability to discriminate macrophage activation Table 6. Preliminary diagnostic guidelines for macrophage activa-
syndrome from active disease without macrophage acti- tion syndrome as a complication of juvenile SLE*
vation syndrome. Only variables that revealed strong
Clinical criteria
discriminating properties, were not duplicative, and were 1. Fever (⬎38°C)
available for a sufficient number of patients were exam- 2. Hepatomegaly (ⱖ3 cm below the costal arch)
3. Splenomegaly (ⱖ3 cm below the costal arch)
ined. The clinical variables included were fever, hepato- 4. Hemorrhagic manifestations (purpura, easy bruising, or
megaly, splenomegaly, hemorrhages, and CNS dysfunc- mucosal bleeding)
tion, and the laboratory variables included were cytopenia 5. Central nervous system dysfunction (irritability, disorientation,
lethargy, headache, seizures, or coma)
(affecting 2 or more cell lineages), aspartate amino- Laboratory criteria
transferase increase, lactate dehydrogenase increase, 1. Cytopenia affecting 2 or more cell lineages (white blood cell
hypofibrinogenemia, hypertriglyceridemia, and hyper- count ⱕ4.0 ⫻ 109/liter, hemoglobin ⱕ90 gm/liter, or platelet
count ⱕ150 ⫻ 109/liter)
ferritinemia. For each laboratory variable, the most 2. Increased aspartate aminotransferase (⬎40 units/liter)
discriminating threshold, either standard or best, was 3. Increased lactate dehydrogenase (⬎567 units/liter)
used. Evidence of hemophagocytosis in the bone mar- 4. Hypofibrinogenemia (fibrinogen ⱕ1.5 gm/liter)
5. Hypertriglyceridemia (triglycerides ⬎178 mg/dl)
row was not included because it was regarded as a 6. Hyperferritinemia (ferritin ⬎500 ␮g/liter)
confirmatory criterion rather than a first-line diag- Histopathologic criterion
nostic criterion. Evidence of macrophage hemophagocytosis in the bone marrow
aspirate
The study variables were combined in the follow-
ing 3 ways: clinical variables only, laboratory variables * The diagnosis of macrophage activation syndrome requires the
only, and clinical and laboratory variables. For each simultaneous presence of at least 1 clinical criterion and at least 2
laboratory criteria. Bone marrow aspiration for evidence of macro-
combination of variables, sensitivity, specificity, and phage hemophagocytosis may be required only in doubtful cases.
diagnostic OR were calculated. The best results were These criteria were developed using patients with active juvenile
obtained using the simultaneous presence of any 1 or systemic lupus erythematosus (SLE) without macrophage activation
syndrome as a control group. As such, they may not be powerful
more clinical criteria and any 2 or more laboratory enough to distinguish macrophage activation syndrome from particular
criteria, which yielded a sensitivity of 92.1%, a specificity infectious complications.
3396 PARODI ET AL

DISCUSSION temic JIA–associated macrophage activation syndrome

was strengthened by the finding that its diagnostic
Although no specific data exist for juvenile SLE,
guidelines were met by 100% of juvenile SLE patients
the reported prevalence of macrophage activation syn-
with macrophage activation syndrome. However, these
drome in SLE ranges from 0.9% to 4.6% (21). However,
guidelines did not demonstrate sufficient diagnostic
it has been suggested that macrophage activation syn- specificity, since they were also met by 71% of patients
drome in SLE may be more common than previously with active juvenile SLE without macrophage activation
recognized. Morales et al (22) evaluated bone marrow syndrome. The main shortcoming of systemic JIA–
specimens from 28 patients with SLE obtained during associated macrophage activation syndrome guidelines
30 episodes of cytopenia. They found that 22 specimens is that certain thresholds of laboratory parameters do
(73.3%) exhibited hemophagocytosis, which was not not apply to patients with juvenile SLE. Because of the
correlated with severity of SLE, serum complement prominent inflammatory nature of systemic JIA, the
levels, or anti-DNA antibody titers. Tsuji et al (23) occurrence of a relative decrease in leukocyte count,
reported that 7 (9.6%) of 73 patients with SLE and liver platelet count, or fibrinogen may be more relevant in
dysfunction had hemophagocytic syndrome. making an early diagnosis of macrophage activation
We compared the typical clinical and laboratory syndrome in that disease. However, since cytopenia is a
features of macrophage activation syndrome in patients frequent feature of active juvenile SLE, an absolute
who had evidence of macrophage hemophagocytosis on decrease in blood cell lineages would be required. Other
bone marrow aspirate (definite macrophage activation diagnostic guidelines that might be useful to identify
syndrome) with those in patients who lacked such evi- macrophage activation syndrome in juvenile SLE are
dence (probable macrophage activation syndrome). It those developed for hemophagocytic lymphohistiocyto-
was hypothesized that the latter group of patients might sis (14). These guidelines proved highly specific, since
not have true macrophage activation syndrome or might they were met by no patient in the control group. How-
have a more subtle form that could have been over- ever, sensitivity was not satisfactory; 33.3% of patients
looked in the absence of a specific suspicion. However, with macrophage activation syndrome did not fulfill the
the 2 patient groups were comparable with regard to criteria for hemophagocytic lymphohistiocytosis.
demographic characteristics, most clinical and labora- In the patients with juvenile SLE in the present
tory features, frequency of ACR criteria at diagnosis, study, fever discriminated macrophage activation syn-
fulfillment of diagnostic criteria for systemic JIA– drome from active disease well when the IGG control
associated macrophage activation syndrome and hemo- sample was used as comparator. However, diagnostic
phagocytic lymphohistiocytosis, admission to the ICU, strength was lower in the comparison with the multina-
death, and drug therapies. tional control sample. Nevertheless, the fact that fever
This finding suggests that macrophage activation was present in as many as 90% of patients with macro-
syndrome in juvenile SLE may be diagnosed in the ab- phage activation syndrome suggests that it is a key
sence of evidence of macrophage hemophagocytosis in component of the clinical picture of macrophage activa-
the bone marrow. As noted in patients with hemophago- tion syndrome in juvenile SLE. All of the other clinical
cytic lymphohistiocytosis (24) and macrophage activa- features provided high specificity rates, but were not as
tion syndrome as a complication of systemic JIA (13), good in terms of sensitivity.
the bone marrow aspirate does not always show hemo- Laboratory features showed excellent discrimi-
phagocytosis, and furthermore, hemophagocytosis is not nating properties, with the use of the standard threshold
always demonstrable at onset. The failure to demon- being more advantageous in some cases, and the use of
strate hemophagocytosis does not negate the diagnosis the best threshold more advantageous in other cases.
of hemophagocytic lymphohistiocytosis. However, a bone This is in keeping with clinical experience indicating that
marrow aspirate would be required to rule out a condi- early suspicion of macrophage activation syndrome is
tion that may induce a macrophage activation syndrome most commonly raised by the detection of subtle changes
by itself, such as Leishmania infection (25,26). in laboratory parameters, whereas clinical symptoms are
The clinical and laboratory spectrum of macro- often delayed or not specific. The most notable excep-
phage activation syndrome in our patients with juvenile tion was leukopenia, which was detected with similar
SLE is similar to that observed in macrophage activation frequency in macrophage activation syndrome and ac-
syndrome occurring in the course of adult-onset SLE tive disease without macrophage activation syndrome.
(9,27) and systemic JIA (1–4). The similarity with sys- However, the frequency of leukopenia was greater and
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3397

its mean value was lower in patients with macrophage a difference from systemic JIA, in which macrophage
activation syndrome who had evidence of macrophage activation syndrome has been more commonly described
hemophagocytosis on bone marrow aspiration than in in patients with advanced disease (1–4). Furthermore, it
those who lacked such evidence. This suggests that the suggests that immune derangement induced by systemic
diagnostic validity of leukopenia is poor in the more disease is the major determinant of macrophage activa-
subtle or initial stages of macrophage activation syn- tion syndrome in juvenile SLE. In the present study,
drome, whereas it becomes a central laboratory feature macrophage activation syndrome represented a serious
in the acute phase of the syndrome. complication, since 43.7% of the patients required ad-
Thrombocytopenia was a better indicator of mac- mission to the ICU, and 11.4% died.
rophage activation syndrome than leukopenia or ane- The therapeutic strategies for macrophage acti-
mia. Of the laboratory features of macrophage activa- vation syndrome in juvenile SLE are not well estab-
tion syndrome, the strongest ability to discriminate this lished. With regard to systemic JIA–associated macro-
complication from active SLE was shown by hyperfer- phage activation syndrome, treatment is primarily based
ritinemia, whose sensitivity and specificity were both on the parenteral administration of high doses of cortico-
equal to or nearly 100%. As many as 78.6% of patients steroids. However, fatal cases of macrophage activation
with macrophage activation syndrome and only 33.3% of syndrome in spite of the use of massive doses of cortico-
patients with active disease had elevated CRP levels. steroids have been reported (2,4,7). Early introduction
This suggests that increased CRP level may be useful to of cyclophosphamide has been advocated as soon as
distinguish macrophage activation syndrome from active corticosteroids seem to be insufficient, since this drug is
juvenile SLE, although it might not discriminate it from a recognized treatment of severe SLE (9,10). Cyclospor-
an infection (28). ine has been found to be dramatically effective in severe
On the basis of the results of statistical analysis or corticosteroid-resistant instances of macrophage ac-
and considering the clinical importance of the different tivation syndrome in children with systemic JIA (30,31)
features, we selected 5 clinical criteria and 6 laboratory and children with juvenile SLE (32). Cyclophosphamide
criteria to be included in the preliminary guidelines and cyclosporine were given to 21.6% and 37.8% of the
for macrophage activation syndrome in juvenile SLE patients in our study, respectively. There were no differ-
(Table 6). Using the “number of criteria present ap- ences in outcome between patients who did or did not
proach,” we found that the best separation between receive such treatments. Etoposide, which is the main-
patients and control subjects occurred when any 1 or stay of therapeutic protocols in hemophagocytic lympho-
more clinical criteria and any 2 or more laboratory histiocytosis, was used in 1 patient. Recently, a favorable
criteria were simultaneously present. The strong dis- outcome was reported in an adult patient with refractory
criminating ability shown by this definition led us to lupus-associated macrophage activation syndrome who
suggest that demonstration of macrophage hemophago- was treated with infliximab (33). The only biologic
cytosis in the bone marrow aspirate should be reserved medication used to treat patients included in the present
for diagnostic confirmation only in doubtful cases. It study was the anti–B cell agent rituximab, which was
should be recognized, however, that the statistical power administered to 2 patients, both of whom had severe
provided by the relatively small size of the patient pancytopenia.
samples was limited. Furthermore, the guidelines were Our study should be interpreted in the light of
developed using patients with active disease as a control several potential limitations. We should acknowledge
group. It remains to be established whether these guide- that it is difficult to truly determine the sensitivity and
lines are powerful enough to distinguish macrophage specificity of diagnostic criteria without a gold standard
activation syndrome from other confusable conditions, for diagnosis, and considering that many clinicians al-
namely infectious complications. Future modifications ready use versions of these criteria when determining a
of guidelines should consider inclusion of novel para- clinical diagnosis of macrophage activation syndrome.
meters, such as soluble IL-2 receptor ␣ (CD25) and Data collection was conducted retrospectively. A retro-
soluble CD163, which reflect activation and expansion of spective study is subject to missing and possibly erroneous
T cells and macrophages, respectively, and may help data. Some laboratory measurements were not available
identify subclinical cases (29). in some patients. Although most of the laboratory tests
As observed in adult patients with SLE (9,27), examined are widely standardized routine procedures,
macrophage activation syndrome was associated with their execution in different laboratories and at different
disease onset in the majority of patients. This represents times may have affected their reliability.
3398 PARODI ET AL

The low frequency of renal and CNS disease seen REFERENCES

in the control group with active disease enrolled at the 1. Ravelli A, Martini A. Macrophage activation syndrome. In: Leh-
coordinating center may be partially due to most of man TH, Cimaz R, editors. Pediatric rheumatology. Amsterdam:
these patients having been assessed at disease onset. It is Elsevier; 2008. p. 55–63.
2. Prieur AM, Stephan JL. Macrophage activation syndrome in
known that kidney and CNS disease may occur late in children with joint diseases. Rev Rhum Engl Ed 1994;61:385–8.
the disease course (34). Another possible explanation is 3. Grom AA, Passo M. Macrophage activation syndrome in systemic
that patients with juvenile SLE who are prone to develop juvenile rheumatoid arthritis. J Pediatr 1996;129:630–2.
4. Sawhney S, Woo P, Murray KJ. Macrophage activation syndrome:
macrophage activation syndrome have a greater multi- a potentially fatal complication of rheumatic disorders. Arch Dis
system involvement at disease onset. That the control Child 2001;85:421–6.
group with laboratory data available comes from a single 5. Athreya BH. Is macrophage activation syndrome a new entity?
Clin Exp Rheumatol 2002;20:121–3.
center may raise a potential referral bias. We should 6. Ramanan AV, Schneider R. Macrophage activation syndrome—
acknowledge that findings in these patients may not be what’s in a name! J Rheumatol 2003;30:2513–6.
generalizable to patients seen in other tertiary pediatric 7. Stephan JL, Kone-Paut I, Galanbrun C, Mouy R, Bader-Meunier
B, Prieur AM. Reactive haemophagocytic syndrome in children
rheumatology centers. The multinational sample, which with inflammatory disorders: a retrospective study of 24 patients.
was more representative, lacked laboratory data. The Rheumatology (Oxford) 2001;40:1285–92.
8. Ravelli A, Viola S, De Benedetti F, Magni-Manzoni S, Tzialla C,
main strengths of our study lie in the multicenter nature
Martini A. Dramatic efficacy of cyclosporine A in macrophage
of the data collection and in the sample size of patients activation syndrome [letter]. Clin Exp Rheumatol 2000;19:108.
with juvenile SLE and macrophage activation syndrome, 9. Lambotte O, Khellaf M, Harmouche H, Bader-Meunier B, Man-
ceron V, Goujard C, et al. Characteristic and long-term outcome
which is the largest collected to date. of 15 episodes of systemic lupus erythematosus-associated he-
In conclusion, macrophage activation syndrome mophagocytic syndrome. Medicine (Baltimore) 2006;85:169–82.
is a life-threatening and probably underdiagnosed com- 10. McCann LJ, Hasson N, Pilkington CA. Macrophage activation
syndrome as an early presentation of lupus [letter]. J Rheumatol
plication of juvenile SLE. When a patient with juvenile 2006;33:438–40.
SLE presents with unexplained fever and cytopenia, an 11. Avcin T, Tse S, Schneider R, Ngan B, Silverman E. Macrophage
evaluation for macrophage activation syndrome, includ- activation syndrome as the presenting manifestation of rheumatic
disease in childhood. J Pediatr 2006;148:683–6.
ing assessment for hyperferritinemia, should be carried 12. Pringe A, Trail L, Ruperto N, Buoncompagni A, Loy A, Breda L,
out. Diagnostic confirmation through the demonstration et al. Macrophage activation syndrome in juvenile systemic lupus
of macrophage hemophagocytosis in the bone marrow erythematosus: an under-recognized complication? Lupus 2007;
16:587–92.
may not be necessary in the presence of the typical 13. Ravelli A, Magni-Manzoni S, Pistorio A, Besana C, Foti T,
clinical and laboratory features of the syndrome. Al- Ruperto N, et al. Preliminary diagnostic guidelines for macro-
phage activation syndrome complicating systemic juvenile idio-
though first-line treatment is based on the parenteral pathic arthritis. J Pediatr 2005;146:598–604.
administration of high-dose corticosteroids, cyclo- 14. Henter JI, Horne A, Arico M, Egeler RM, Filipovich AH,
sporine, and perhaps etoposide, may play a role when Imashuku S, et al. HLH-2004: diagnostic and therapeutic guide-
lines for hemophagocytic lymphohistiocytosis. Pediatr Blood Can-
disease is refractory to corticosteroid treatment. Further cer 2007;48:124–31.
studies in larger numbers of patients and using different 15. Hochberg MC, for the Diagnostic and Therapeutic Criteria Com-
control groups are needed to investigate the validity of mittee of the American College of Rheumatology. Updating the
American College of Rheumatology revised criteria for the clas-
these preliminary diagnostic guidelines for macrophage sification of systemic lupus erythematosus [letter]. Arthritis Rheum
activation syndrome as a complication of juvenile SLE. 1997;40:1725.
16. Bloch DA, Moses LE, Michel BA. Statistical approaches to
classification: methods for developing classification and other
AUTHOR CONTRIBUTIONS criteria rules. Arthritis Rheum 1990;33:1137–44.
17. Fries JF, Hochberg MC, Medsger TA Jr, Hunder GG, Bombardier
All authors were involved in drafting the article or revising it C, and the American College of Rheumatology Diagnostic and
critically for important intellectual content, and all authors approved Therapeutic Criteria Committee. Criteria for rheumatic disease:
the final version to be published. Dr. Ravelli had full access to all of the different types and different functions. Arthritis Rheum 1994;37:
data in the study and takes responsibility for the integrity of the data 454–62.
and the accuracy of the data analysis. 18. Bandeira M, Buratti S, Bartoli M, Gasparini C, Breda L, Pistorio
Study conception and design. Parodi, Davı̀, Pringe, Ruperto, Martini, A, et al. Relationship between damage accrual, disease flares and
Ravelli. cumulative drug therapies in juvenile-onset systemic lupus ery-
Acquisition of data. Parodi, Davı̀, Pringe, Magni-Manzoni, Miettunen, thematosus. Lupus 2006;15:515–20.
Bader-Meunier, Espada, Sterba, Ozen, Wright, Saad Magalhães, 19. Ravelli A, Duarte-Salazar C, Buratti S, Reiff A, Bernstein B,
Khubchandani, Michels, Woo, Iglesias, Guseinova, Bracaglia, Hayward, Maldonado-Velazquez MR, et al. Assessment of damage in juve-
Wouters, Grom, Vivarelli, Fischer, Breda, Ravelli. nile-onset systemic lupus erythematosus: a multicenter cohort
Analysis and interpretation of data. Parodi, Davı̀, Pistorio, Ruperto, study. Arthritis Rheum 2003;49:501–7.
Martini, Ravelli. 20. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The
MACROPHAGE ACTIVATION SYNDROME IN JUVENILE SLE 3399

diagnostic odds ratio: a single indicator of test performance. J Clin 28. Gaitonde S, Samols D, Kushner I. C-reactive protein and systemic
Epidemiol 2003;56:1129–35. lupus erythematosus. Arthritis Rheum 2008;59:1814–20.
21. Fukaya S, Yasuda S, Hashimoto T, Oku K, Kataoka H, Horita T, 29. Bleesing J, Prada A, Siegel DM, Villanueva J, Olson J, Ilowite NT,
et al. Clinical features of haemophagocytic syndrome in patients et al. The diagnostic significance of soluble CD163 and soluble
with systemic autoimmune diseases: analysis of 30 cases. Rheuma- interleukin-2 receptor ␣-chain in macrophage activation syndrome
tology (Oxford) 2008;47:1686–91. and untreated new-onset systemic juvenile idiopathic arthritis.
22. Morales PM, Jimenez BF, Yanes P, Rios D, Godinez R. Bone Arthritis Rheum 2007;56:965–71.
marrow (BM) with reactive histiocytosis (RH), hemophagocytosis 30. Mouy R, Stephan JL, Pillet P, Haddad E, Hubert P, Prieur AM.
and storage histiocytes (SH) in systemic lupus erythematosus Efficacy of cyclosporine A in the treatment of macrophage acti-
[abstract]. Arthritis Rheum 1992;35 Suppl:S239.
vation syndrome in juvenile arthritis: report of five cases. J Pediatr
23. Tsuji T, Ohno S, Ishigatsubo Y. Liver manifestations in systemic
1996;129:750–4.
lupus erythematosus: high incidence of hemophagocytic syn-
drome. J Rheumatol 2002;29:1576–7. 31. Ravelli A, De Benedetti F, Viola S, Martini A. Macrophage
24. Janka GE. Familial hemophagocytic lymphohistiocytosis. Eur J activation syndrome in systemic juvenile rheumatoid arthritis
Pediatr 1983;140:221–30. successfully treated with cyclosporine. J Pediatr 1996;128:275–8.
25. Gagnaire MH, Galambrun C, Stephan JL. Hemophagocytic syn- 32. Ravelli A, Viola S, De Benedetti F, Magni-Manzoni S, Tzialla C,
drome: a misleading complication of visceral leishmaniasis in Martini A. Dramatic efficacy of cyclosporine A in macrophage
children—a series of 12 cases. Pediatrics 2000;106:e58. activation syndrome [letter]. Clin Exp Rheumatol 2000;19:108.
26. Tunc B, Ayata A. Hemophagocytic syndrome: a rare life- 33. Ideguchi H, Ohno S, Takase K, Hattori H, Kirino Y, Takeno M,
threatening complication of visceral leishmaniasis in a young boy. et al. Successful treatment of refractory lupus-associated haemo-
Pediatr Hematol Oncol 2001;18:531–6. phagocytic lymphohistiocytosis with infliximab. Rheumatology
27. Dhote R, Simon J, Papo T, Detournay B, Sailler L, Andre MH, (Oxford) 2007;46:1621–2.
et al. Reactive hemophagocytic syndrome in adult systemic dis- 34. Petty RE, Laxer RM. Systemic lupus erythematosus. In: Cassidy
ease: report of twenty-six cases and literature review. Arthritis JT, Petty RE, Laxer RM, Lindsley CB, editors. Textbook of
Rheum 2003;49:633–9. pediatric rheumatology. Philadelphia: Elsevier; 2005. p. 342–406.