We outline the epidemiology, clinical features, laboratory findings, outcome and the difficulties in diagnosis and management of these children, from a tertiary children’s teaching hospital in southern India.

This was a retrospective study of case-records of all the children who were diagnosed to have HLH during the period December 2006 to December 2008 at the Kanchi Kamakoti CHILDS Trust Hospital. Only those children who met the diagnostic criteria of the HLH-2004 protocol of the Histiocyte Society (Table I) were included. Genetic analysis, NK cell activity and soluble CD25 receptor assay could not be done in any child.

TABLE I



Diagnostic Guidelines for HLH-2004 

The case records of the children who fulfilled the criteria for HLH were analyzed and the following details collected: age at presentation, sex, relevant family history, clinical and laboratory data, course, treatment for HLH, and outcome (survival or death). For peripheral blood counts and biochemistry, nadir or peak values were recorded. In some children perforin expression was determined in cytotoxic cells using flow cytometry at the Institute of Immunohaematology, KEM Hospital, Mumbai, and compared with age matched controls.  

Forty three children were diagnosed with HLH over a 25-month period. Of these, only 33 met the inclusion criteria and were analyzed. The remaining 10 children presented with prolonged fever and had hemophagocytosis, either in the bone marrow or lymph nodes but were excluded since they did not meet the minimum diagnostic criteria.

Of the 33 children analyzed, 19 (58%) were male. The ages ranged from 50 days to 14 years (mean 46 months, median 33 months). Six children had a family history of consanguinity and three had a history of sibling death (anemia requiring blood transfusion and hospitalization elsewhere).

The most common reason for hospital admission was fever with or without additional symptoms or signs such as rash, hepatosplenomegaly, lymphadenopathy, respiratory symptoms [cough, tachypnea, dyspnea] or CNS involvement (irritability, altered behavior, meningeal signs or seizures). The interval between the onset of symptoms and diagnosis of HLH varied from 7 to 53 days (mean 16 days). The clinical and laboratory features are highlighted in Tables II and III. Eighteen (55%) children required intensive care for either respiratory distress or sepsis and 7 (22%) were ventilated.

TABLE II



Clinical and Laboratory Features 

Nineteen of the 21 children in whom ferritin was repeated showed a decreasing trend, and they survived. CSF analysis was carried out in all children with neurological involvement (n = 12) and was abnormal [cells >15 and/or elevated CSF protein (>50mg/dL)] in 4 of them. Bone marrow aspiration was performed in 31 children and showed hemophagocytosis (macrophages engulfing blood cells and their precursors) in 26 children (84%). There was no evidence of malignancy in any patient.

Infectious etiology was identified in 14 (42%) children, with viruses accounting for the majority (EBV-3, Dengue-5, CMV-1). Bacteria were isolated in 5 children – Mycobacterium tuberculosis (1), Klebsiella (1), Leptospira (1), Pneumococcus (1) and Sphingomonas (1). The child with tuberculosis has been reported earlier [4].

Two children required full 40 weeks treatment as per the HLH-2004 protocol and are doing well. In three children, chemotherapy was stopped after the initial 8-week course and they are doing well now. Perforin expression in cytotoxic cells was very low in 3 out of 10 children in whom the test was performed. Of these, two survived and are doing well.  

HLH is a rare disorder that is being increasingly recognized in children now. This is the largest case series of HLH from India. The median age at diagnosis of HLH in our series was 33 months, similar to other Asian studies [5,6]. The overall mortality (24%) in our series was lower than in series reported earlier [5-10]. Better survival rate in this series might be due to a high incidence of secondary HLH, early diagnosis and early institution of immunomodulatory treatment.

Mutations in the perforin gene have been reported to be present in 20-40% of FHL [1,11]. Perforin staining in cytotoxic cells by flow cytometry has been documented as a screening test to identify children with FHL, who can then be subjected to genetic analysis [12,13]. A very low perforin expression was detected in 30% children in whom the test was done but we could not perform genetic analysis in them. The most common presentation at admission was fever and hepatomegaly (100%). Hyperferritinemia, hypofibrinogenemia and high LDH levels were observed in the majority of our patients, suggesting that these tests could serve as important diagnostic clues to HLH. Similar findings were reported in an earlier study [14]. We had normal bone marrow histopathology in 16% of the patients in whom bone marrow was performed. Hemophagocytosis in bone marrow may not be present during the early phase of disease, and therefore the absence of hemophagocytosis does not exclude a diagnosis of HLH [15].

None of the clinical and laboratory parameters were significantly associated with mortality in our series. In survivors, recovery of laboratory parameters (increasing platelet count), decrease in ferritin levels, and clinical improvement (subsidence of fever and regression of organomegaly) was noted about 5-7 days after starting immunomodulatory treatment.

Dengue virus was the most common infectious agent identified in our series. Work up for HLH was done in these children as fever, thrombocytopenia and anemia, persisted for more than 2 weeks, without any proven secondary infection. Very few cases of dengue associated HLH have been reported in the literature [10]. EBV infection has been reported to be the commonest cause of infection associated HLH and has been observed to be a bad prognostic factor in earlier studies [16,17]. EBV accounted for only 20% of the cases in our series, which is much lower than in other reports [9,18]. All patients with EBV infection survived without cytotoxic drugs. 

Stem cell transplant is the standard treatment for FHL, once remission is achieved on immunomodulatory therapy [1].  However, genetic diagnosis is not available and stem cell therapy difficult in India. In our series, steroids with or without IVIg were used commonly in the early phase of the disease. A favorable response with IVIg has been observed in some studies [19,20], but a lack of efficacy was also reported [16]. Etoposide containing regimens have been reported to have better outcomes [16,17,21].

HLH should be considered in the differential diagnosis of children with sepsis or presumed sepsis that do not respond to the conventional treatment. Patients may not fulfill all the criteria during the early phase of the disease. Bone marrow examination may not show hemophagocytosis initially. Cytotoxic agents should be considered early in patients who do not respond to other therapies. Presence of an infection during illness does not exclude FHL. Prompt recognition and appropriate therapy may result in a good outcome.

Acknowledgment: We thank the individual treating Consultants at Kanchi Kamakoti CHILDS Trust Hospital, Chennai for their support and Dr Manisha Madkaikar, Research Scientist at the National Institute of Immunohaematology, Mumbai for performing the perforin assays.

Contributors: BR and KGR managed the cases, analyzed data and stand as guaranteers for the manuscript. SBS contributed to patient management and writing of the manuscript. AVR initiated the design of the study and helped in editing the manuscript. AN collected and analyzed the data, outcome assessment and prepared the manuscript. 

Funding: This study was supported by the CHILDS Trust Medical Research Foundation.

Competing interests: None stated.