Differentiating Juvenile Myelomonocytic Leukemia From Infectious Disease

To The Editor:

Two recent articles in BLOOD^1,2 review the findings and outcome of juvenile myelomonocytic leukemia (JMML). Both omit an important aspect of JMLL: its differentiation from infectious disease. Several disseminated microbial infections of infancy can result in persistent fever, failure to thrive, hepatosplenomegaly, skin lesions, anemia, thrombocytopenia, and myelomonocytosis, including Epstein-Barr virus (EBV), cytomegalovirus (CMV), human herpes virus-6 (HHV-6), histoplasma, mycobacteria, and toxoplasma. Thorough investigation for infection is needed in infants with these findings to avoid erroneous diagnosis and mistaken interventions.

Herrod et al3 reported two infants with persistent EBV infection and findings consistent with JMML, including increased numbers of F and i cells and abnormal granulocyte-macrophage colony formation in vitro. Both recovered without treatment and remained well. This raises the possibility that some of the long-term survivors reported by Niemeyer et al and Arico et al had similar infections rather than leukemia. Neonatal CMV and HHV-6 infections can also mimic JMML.4,5 Might erroneous diagnosis account for the better prognosis reported for patients with JMML who are less than 6 months old?²

The excellent reviews of Niemeyer et al and Arico et al suggest that JMML represents a group of diseases rather than a single entity. Careful investigation for microbial associations, including more recently identified herpesviruses, might contribute to understanding pathogenesis as well as to diagnosis and management.

We are grateful to Dr Pinkel for his carefully considered remarks. We fully agree that some infectious disease can mimic JMML, thus jeopardizing the interpretation of some cases. In particular, Dr Pinkel raises the question of whether some of the long-term survivors we described, in fact, have JMML. Although such suspicion is obviously warranted, our experience indicates that the vast majority of cases that fit the diagnostic picture of JMML represent leukemia and not an infectious process. Even when the patients show rapid recovery with or without “minimal treatment,” reactivation of the disease may occur. In some cases these recurrences have an accelerated phase, mirroring that in patients with rapidly fatal JMML. We cannot rule out infectious diseases as the origin of some of the more unusual cases of JMML that have been reported in the medical literature, but we do not believe these exceptions account for more than 10% of the total number. Perhaps current advances in the diagnosis will help to resolve this intriguing issue.

We have recently published the results of a retrospective analysis of 110 children with chronic myelomonocytic leukemia (CMML).2-1There has since been an international consensus to rename the disease juvenile myelomonocytic leukemia (JMML). The new term JMML will include all leukemias of childhood previously classed CMML,2-1,2-2juvenile chronic myelogenous leukemia (jCML),2-3,2-4 or infantile monosomy 7 syndrome,2-3 because their clinical and biological similarities suggest that they are spectrums of the same disease. We believe that the broad agreement on nomenclature will facilitate cooperative treatment trials and hasten research on the pathogenesis of JMML.

As addressed by Dr Pinkel, the clinical and morphological picture of JMML can be mimicked by a variety of infectious organisms. In addition, granulocyte-macrophage colony-stimulating factor hypersensitivity of myeloid progenitor cells, thought to play a central role in the pathogenesis of JMML,2-5 has been noted in vitro in children with viral infections.2-6 A basic tenet for the definition of myeloid leukemias is the demonstration of the clonal origin from a malignant hematopoietic progenitor cell.2-7 The clonal nature is often inferred by evidence of a chromosomal abnormality or an activating mutation of a proto-oncogene. In this respect about half of the children with JMML have evidence of a clonal disorder; 35% are known to have a chromosomal abnormality2-1 and 15% to have a point mutation of the Nras or Kras oncogene in their hematopoietic cells.2-8 More recently, the study of X-chromosome inactivation patterns showed evidence for monoclonal origin of mononuclear cells in all female JMML patients analyzed.2-9 In the absence of a marker of clonality, the establishment of the diagnosis JMML and firm exclusion of an infectious origin can be difficult.

To address this issue in our retrospective study2-1 we collected data on the serology for cytomegalovirus (CMV; n = 56), herpes virus type I (HSV; n = 27), and Epstein-Barr virus (EBV; n = 51) from the time of diagnosis. Thirty-eight percent of children were positive for CMV, 44% for HSV, and 47% for EBV. The prevalences of antibodies to these viruses were similar to those observed in normal infant populations in Western Europe.2-10-2-13 There were no significant differences in age at diagnosis or length of survival between JMML patients with or without previous or recent CMV, HSV, or EBV infection. Dr Pinkel raises the concern that some of our long-term survivors might have had infections rather than leukemia. Of the seven patients with a survival of more than 5 years without bone marrow transplantation, four have succumbed to their disease (Table 7 in Niemeyer et al2-1). Of the three remaining patients, one girl currently alive with disease 6.5 years after diagnosis is known to have an Nras mutation (A. Biondi, personal communication). Another patient had monosomy 7 in his bone marrow cells documented twice within 6 months after diagnosis, whereas a normal karyotype was found 4 and 9 years later. He had no evidence of disease when seen last 9.6 years after diagnosis.2-14 The patient with the longest survival, currently 13 years after diagnosis, had no marker of clonality. His smears and clinical data were thoroughly reviewed, but it cannot be excluded that he suffered from an infection rather than from leukemia. Viral studies from the time of the diagnosis were not available.

We agree with Dr Pinkel that a careful investigation for an infectious cause is mandatory in all children suspected as suffering from JMML. Until the chromosomal and molecular abnormalities of the majority of JMML patients with so-called “normal karyotype” have been unraveled, the diagnosis of JMML will have to be based on a number of clinical and laboratory features (Table 1). The suggestive clinical features, the minimal laboratory criteria, and the criteria requested for definite diagnosis may prove to be a guideline in establishing the diagnosis of JMML.

For the European Working Group of MDS in Childhood (EWOG-MDS):