Portal de Periódicos da CAPES

Shwachman–Diamond Syndrome

2002; Wiley; Volume: 118; Issue: 3 Linguagem: Inglês

10.1046/j.1365-2141.2002.03585.x

1365-2141

Yigal Dror, Melvin H. Freedman,

Pneumocystis jirovecii pneumonia detection and treatment

In 1964, five patients showing evidence of exocrine pancreatic insufficiency and leucopenia were described (Shwachman et al, 1964). The syndrome, later termed Shwachman–Diamond syndrome (SDS), was described earlier by Nezelof & Watchi (1961) and a few months later by Bodian et al (1964). Burke et al (1967) and Pringle et al (1968) observed associated skeletal changes of the metaphyseal dysostosis type, which became the third fundamental feature of the syndrome. Aggett et al (1980) described 21 cases of SDS in detail and defined the scope of the syndrome as currently understood. SDS has also been called 'Shwachman syndrome', 'Shwachman–Bodian syndrome' and 'congenital lipomatosis of the pancreas. Although SDS is a rare disorder, it is the most common cause of pancreatic insufficiency in children next to cystic fibrosis and probably the third most common inherited bone marrow failure syndrome after Fanconi's anaemia and Diamond–Blackfan anaemia. It mainly involves the pancreas, bone marrow and skeleton, but the liver, kidneys, teeth and immune system may also be affected (Aggett et al, 1980; Mack et al, 1996; Ginzberg et al, 1999; Dror et al, 2001). Haematologically, SDS is characterized by varying degrees of cytopenia and a high risk of development of myelodysplastic syndrome (MDS) and leukaemia, making it an important model for genetic determinants of bone marrow failure and leukaemia. Studies of family pedigrees support an autosomal recessive mode of inheritance (Ginzberg et al, 1999). The SDS locus has recently been mapped to the centromeric region of chromosome 7 (7p10–7q11) (Goobie et al, 2001). The gene responsible for this complex, pleiotropic phenotype should be identified soon, and the molecular basis for the multisystem features, bone marrow failure and leukaemic transformation may be clarified. SDS manifests itself in many and various combinations, shown in Table I. Results of two retrospective studies (Mack et al, 1996; Smith et al, 1996) and a cross-sectional multicentre study (Ginzberg et al, 1999) have recently been published, adding significantly to our understanding of the clinical phenotype and probably representing the spectrum of clinical findings fairly accurately. Neutropenia. Neutropenia is the most common haematological abnormality, occurring in 88–100% of patients and has been identified as early as the neonatal period. In most cases the neutropenia is intermittent and fluctuates within the same patient from severely low to normal levels. Cyclic patterns without detailed information on the timing of cycling have also been reported (Dokal et al, 1997). However, cyclical haematopoiesis was not observed in the series described by Aggett et al (1980), Mack et al (1996) or Smith et al (1996) or among 16 SDS patients followed at our institution (unpublished observations). Impaired neutrophil chemotaxis. SDS neutrophils may have defects in mobility, migration and chemotaxis in most (Cipolli et al, 1999; Maserati et al, 2000) or all patients (Smith et al, 1996; Dror et al, 2001). Alterations in neutrophil cytoskeletal/microtubular integrity and function may play a prominent role in causing the defective chemotaxis. Rothbaum et al, (1982) demonstrated abnormal distribution of concanavalin-A receptors on polymorphonuclear leucocytes. Lithium appears to restore chemotaxis when used in vitro (Azzaràet al, 1988) and in vivo (Azzaràet al, 1991). Anaemia. Anaemia with low reticulocytes was recorded in up to 80% of patients in the reports reviewed. It was usually mild and normochromic-normocytic. However, macrocytosis has been described (Woods et al, 1981a) and three of 16 SDS patients we studied had macrocytic red blood cells (unpublished observations). Fetal haemoglobin was elevated in 80% of patients (Dror & Freedman, 1999). The elevation of heterogeneously distributed fetal haemoglobin reflects 'stress' haematopoiesis and/or ineffective erythropoiesis related to apoptosis and is also seen in all types of MDS. Thrombocytopenia. Thrombocytopenia, defined by platelets < 150 × 109/l, was seen in 24% to 88% of patients. Easy bruising may occur. Fatal bleeding while the patients had moderate to severe thrombocytopenia has also been reported (Nezelof & Watchi, 1961; Caselitz et al, 1979; Woods et al, 1981a; Gretillat et al, 1985; Okcu et al, 1998; Maserati et al, 2000). Pancytopenia. Trilineage cytopenia occurs in 10–65% of patients. Both anaemia and thrombocytopenia are usually mild to moderate in these cases but neutropenia tends to be more severe. Pancytopenia in SDS with hypoplasia of three bone marrow lineages might carry a poor prognosis, with a higher chance of developing symptomatic severe aplasia, advanced MDS or acute myeloid leukaemia (AML) (Mack et al, 1996). Severe aplasia requiring transfusion has been reported (Woods et al, 1981a; Tsai et al, 1990; Barrios et al, 1991). In the series reported by Aggett et al (1980), four of the 21 patients were transfusion dependent. Bone marrow findings. The severity of cytopenia does not always correlate with bone marrow cellularity. Varying degrees of marrow hypoplasia and fat infiltration are the usual findings (Aggett et al, 1980; Dror & Freedman, 1999), but marrows showing normal or even increased cellularity have also been observed (Smith et al, 1996; Ginzberg et al, 1999). Single-lineage hypoplasia is usually myeloid and occurs in 15–50% (Mack et al, 1996; Ginzberg et al, 1999). Left-shifted granulopoiesis or maturation arrest was also described (Aggett et al, 1980; Ginzberg et al, 1999). Scattered mild dysplastic changes in the erythroid, myeloid and megakaryocytic precursors are commonly seen in bone marrows of SDS patients and form part of the syndrome. These changes represent disordered haematopoiesis and may fluctuate. Prominent dysplasia is less common and, if it occurs, signifies malignant myeloid transformation. Myelodysplastic syndromes. Several findings suggest that SDS is a myelodysplastic disorder from its inception. MDS is defined as a preleukaemic, clonal, stem-cell disease with peripheral blood cytopenia, ineffective haematopoiesis and varying degrees of bone marrow cellularity and dysplasia (Harris et al, 1999). SDS easily meets at least five of these seven criteria: it is a stem-cell disorder with peripheral cytopenia, ineffective haematopoiesis, varying degrees of bone marrow cellularity and carries a significant risk of leukaemia. The fact that all marrow cells harbour two copies of the mutated SDS gene and that clonal marrow cytogenetic abnormalities are common probably satisfies the clonality criterion. In addition, scattered mild dysplastic changes in the erythroid, myeloid and megakaryocytic precursors are part of the syndrome. Finally, a close relationship between SDS and MDS is reflected in similar defects in marrow-stromal support of normal haematopoiesis (Dror & Freedman, 1999), increased apoptosis mediated through the Fas pathway (Dror & Freedman, 2001), similar prevalence of P53 protein overexpression to refractory anaemia patients (Elghetany & Alter, 2001) and high cluster-to-colony ratios (Dror et al, 1998a). Therefore, according to a newly devised classification system of childhood MDS (Mandel et al, 1999, 2002; Table III), we consider SDS as refractory cytopenia and when we refer to malignant myeloid transformation in SDS we mean more advanced MDS stages. Until recently, no comprehensive, all-inclusive classification scheme for paediatric MDS was available. We therefore developed the CCC system, which incorporates the 'Category' of MDS (denovo, syndromic or therapy-related), the 'Cytology' (refractory cytopenia, RC; refractory cytopenia with dysplasia, RCD; refractory cytopenia with ring sideroblasts, RCRS; and refractory cytopenia with excess blasts, RCEB) and the 'Cytogenetics'. Therefore, when we refer to malignant myeloid transformation in SDS, we mean stages beyond RC/CG– (RC without cytogenetic abnormalities): RC/CG+ (RC with cytogenetic abnormalities), RCD, RCRS, RCEB or leukaemia. Forty such cases have been reported in SDS (Nezelof & Watchi, 1961; Huijgens et al, 1977; Strevens et al, 1978; Caselitz et al, 1979; Aggett et al, 1980; Woods et al, 1981a,b; Gretillat et al, 1985; MacMaster & Cummings, 1993; Seymour & Escudier, 1993; Kalra et al, 1995; Passmore et al, 1995; Smith et al, 1995, 1996; Arseniev et al, 1996; Mack et al, 1996; Davies et al, 1997; Dokal et al, 1997; Dror et al, 1998b; Okcu et al, 1998; Faber et al, 1999; Sokolic et al, 1999; Maserati et al, 2000; Spirito et al, 2000; Cesaro et al, 2001) (Table IV). Twenty-five patients with SDS with cytogenetic abnormalities and/or morphological dysplasia and/or an increase in bone marrow blasts (5–30%) have been reported at a mean age of 11·8 years (median, 8 years; range, 2–42). Seventeen of the 25 were males, seven were females, and one was not documented. An additional 15 patients (all males) with SDS were diagnosed with malignant myeloid transformation at the stage of overt leukaemia at a mean age of 16 years (median, 14 years; range, 1·5–37). Cytogenetic abnormalities. Cytogenetic abnormalities have been reported in 25 SDS patients in various stages of malignant myeloid transformation: RC/CG+, RCD, RCEB and AML (Table V). Six of them (24%) had isochromosome 7q[i(7q)], an extremely uncommon cytogenetic abnormality rarely described in MDS, AML or acute lymphoblastic leukaemia patients without SDS (Mertens et al, 1994). Its high occurrence in SDS patients suggests that it is a fairly specific clonal marker in this syndrome and is probably related to the presence of the SDS gene on 7q(10–11) (Goobie et al, 2001). The relationship between the SDS gene and the i(7q) formation, which also involves breakpoints at 7q(10–11), may be deciphered when the gene is pinpointed, cloned and sequenced. Other chromosome 7 abnormalities were described in an additional 10 patients (40%), including monosomy 7, combined i(7q) and monosomy 7, and deletion or translocations involving part of the long arm of chromosome 7. These cases suggest that a substantial proportion of patients with SDS who develop MDS/AML will also acquire chromosome 7 abnormalities. However, complete monosomy 7 or partial deletions of its long arm are also characteristic of patients with other inherited bone marrow failure syndromes who develop MDS or AML, including Fanconi's anaemia, Kostmann's neutropenia and congenital amegakaryocytic thrombocytopenia (Luna-Fineman et al, 1995). Other cytogenetic abnormalities have also been described in SDS at various disease stages (Table V). The prognostic significance of the cytogenetic abnormalities in SDS will be clarified with prospective monitoring of patients. It is noteworthy that among the six patients with i(7q) no progression to RCEB or AML has been reported. In contrast, four patients with the other chromosome 7 abnormalities, one with RCEB and three with AML, either initially presented with these disorders or progressed to them from earlier stages of MDS. i(7q) has been associated with morphological dysplasia in one SDS patient, transfusion dependency in another and severe neutropenia requiring granulocyte colony-stimulating factor (G-CSF) therapy in a third. These findings may signify a different prognostic value of i(7q) in this disorder. Interestingly, seven SDS cases have been reported with cytogenetic abnormalities without prominent marrow dysplasia or an increase in blasts (RC/CG+). Of these patients, two had stable disease, two developed morphological dysplasia or followed a more severe clinical course, one developed AML and two could not be evaluated. Therefore, the finding of a clonal cytogenetic abnormality in morphologically 'benign' bone marrow aspirates may be of clinical significance and requires frequent monitoring. Whether allogeneic haematopoietic stem cell transplantation (HSCT) is indicated for patients with clonal marrow disease, but without symptomatic cytopenia or increased blast counts, is debatable. Cytology. Various morphological types of MDS have been described in SDS patients. RCD without previous RC/CG+ was reported in 12 patients, eight of whom could be evaluated for disease progression: three had non-specified clinical deterioration of their MDS, four developed AML and one had stable disease. Only one case of RCRS (without evidence of Pearson syndrome), in a patient with SDS who eventually progressed to AML, was reported. In the five reported cases with RCEB without previous RC/CG+ or RCD, three patients progressed to AML and two could not be evaluated. An additional patient developed RCEB from previous RC/CG+ and went on to develop AML. Leukaemia. Leukaemia was reported in 24 SDS patients (22 males, one female, one not documented) at a mean age of 16·8 years (median, 14 years; range, 1·5–43) (duplicate reports were again excluded). Of the 24 reported cases of AML, nine were preceded by an MDS phase (Huijgens et al, 1977; Strevens et al, 1978; Woods et al, 1981b; Gretillat et al, 1985; Seymour & Escudier, 1993; Arseniev et al, 1996; Smith et al, 1996; Dokal et al, 1997), while in the other 15 the malignant myeloid transformation was detected only at the AML stage (Nezelof & Watchi, 1961; Caselitz et al, 1979; Aggett et al, 1980; Woods et al, 1981b; MacMaster & Cummings, 1993; Mack et al, 1996; Dokal et al, 1997; Spirito et al, 2000). The crude risk is probably between 15% and 25%. Various types of leukaemia have been described in SDS patients: AML-M2 in three patients, AML-M4 in three, AML-M5 in three, AML-M6 in six, AML-non-specific in five, acute lymphoblastic leukaemia in three and juvenile myelomonocytic leukaemia in one. It is noteworthy that AML-M6 was particularly common in SDS, occurring in about 30% of cases. Earlier reports of acute lymphoblastic leukaemia are interesting (Strevens et al, 1978; Woods et al, 1981b; Gretillat et al, 1985). With newer diagnostic methods these children might have been diagnosed as having AML instead (possible AML-M0). Alternatively, the flawed early stem cell in SDS may be predisposed to malignant lymphoblastic transformation, as is sometimes seen in adults with MDS (Lima et al, 1997). The juvenile myelomonocytic leukaemia diagnosed in an 8-year-old SDS patient reported by Caselitz et al (1979) might now have been diagnosed as AML as well, as the case does not meet the current criteria for the diagnosis of this disease. SDS-related leukaemia carries a poor prognosis. In the 20 AML cases reported (as initial presentation or after a period with MDS), 17 patients died either because of refractory disease or treatment-related toxicity. The other three survived 24 and 27 months after chemotherapy and 97 d after HSCT respectively. Two of the three patients with acute lymphoblastic leukaemia died of refractory disease or treatment-related toxicity and one survived 1 year after diagnosis. The patient with juvenile myelomonocytic leukaemia died of treatment-related toxicity. The occurrence of MDS/AML in SDS patients receiving G-CSF therapy for severe neutropenia has raised concerns that the therapy might have played a role in malignant transformation. Malignant myeloid transformation in SDS patients while on G-CSF therapy has been reported (Davies et al, 1997; Freedman et al, 2000). To date, however, there is no strong evidence to incriminate the cytokine directly in the leukaemic conversion. G-CSF does not cause DNA damage, and no direct relationship has been found between the development of MDS/AML and the G-CSF dose or the duration of therapy (Freedman et al, 2000). G-CSF may simply be an 'innocent bystander' that corrects the neutropenia, prolongs patient survival and allows time for the malignant predisposition to declare itself. Alternatively, G-CSF may accelerate the propensity to malignant myeloid transformation in the genetically altered SDS stem cells, or rescue malignant clones that would otherwise be destined for apoptosis. SDS is only one among the many inherited bone marrow failure syndromes with a tendency to evolve into MDS and leukaemia. Syndromes such as Fanconi's anaemia, congenital amegakaryocytic thrombocytopenia, Kostmann's neutropenia and Diamond–Blackfan anaemia share this risk, although its prevalence varies. Patients with SDS are particularly susceptible to recurrent viral, bacterial and fungal infections, including otitis media, sinusitis, mouth sores, bronchopneumonia, osteomyelitis, septicaemia and skin infections. Overwhelming sepsis is a well-recognized fatal complication of this disorder, particularly early in life. The quantitative and qualitative defects in SDS neutrophils contribute to these infections. Numerous case reports describing B- and T-cell defects in patients with the syndrome have been published (Hudson & Aldor, 1970; Mäki et al, 1978; Aggett et al, 1979; Sacchi et al, 1982; Kornfeld et al, 1995). In a recent prospective study of immune function in 11 patients with SDS (Dror et al, 2001), we identified varying degrees of impairment. Seven patients had B-cell defects comprising one or more of the following abnormalities: low IGG or IGG subclasses, low percentage of circulating B lymphocytes, decreased in vitro lymphocyte proliferation in response to staphylococcal-antigen Cowan strain I and Staphylococcal protein, and a lack of specific antibody or isohaemaglutinin production. Six of nine patients studied had at least one T-cell abnormality, comprising a low percentage of total circulating T lymphocytes or CD3+/CD4+ cell subpopulations, an inverse CD4:CD8 ratio, or an abnormally low lymphocyte proliferative response to concavin A and pokeweed mitogen. Five of six patients studied had decreased percentages of circulating natural killer cells. Lymphopenia was observed in one patient. Immunodeficiency therefore seems to be a component of the syndrome. Varying severity of pancreatic dysfunction due to acinar maldevelopment is a hallmark of SDS. Its severity is not consistent with either haematological or skeletal abnormalities (Ginzberg et al, 1999). Pathological analysis shows extensive fatty replacement of pancreatic acinar tissue with relatively normal ductal architecture. Pancreatic stimulation tests using intravenous secretin and cholecystokinin confirm the presence of impaired enzyme secretion (including lipase, amylase and tryspsinogen) into the duodenal fluid, which is sampled through a nasogastric tube. These abnormalities are also reflected by the low serum trypsinogen levels found in 91% of patients (Ginzberg et al, 1999), low serum amylase in 73% (Aggett et al, 1980), low serum isoamylase (Ip et al, 2001) and fat quantification in a 72 h stool collection (Aggett et al, 1980; Ginzberg et al, 1999). Sonograms show that the pancreas is larger, with increased echogenicity. The features may be caused by either fibrosis or lipomatosis of the pancreas, which can easily be distinguished by computerized tomography (Genieser et al, 1982; Robberecht et al, 1985; MacMaster & Cummings, 1993). The symptoms of pancreatic insufficiency, malabsorption and steatorrhea are present in 86% of patients. These conditions can lead to malnutrition and fat-soluble vitamin deficiencies (A, D, E and K). The most severe manifestations of pancreatic insufficiency are seen in infancy (Ginzberg et al, 1999). With increasing age, up to 50% of SDS patients show improvement in enzyme production and can become pancreatic-sufficient with normal fat absorption. In these cases, pancreatic dysfunction can still be diagnosed by an abnormal pancreatic stimulation test in all patients and by low serum trypsinogen in most. The children who improve are able to decrease or even discontinue their replacement enzymes. Approximately half of children with SDS have metaphyseal dysostosis, which most often involves the femoral head and is usually asymptomatic (Taybi et al, 1969; Aggett et al, 1980). These changes can be seen on plain radiographs but may not be detectable until after 12 months of age. One third to one half of children with SDS have rib-cage abnormalities, including shortened ribs with flared ends, chostochondral thickening and a narrow rib cage. In a few cases, these abnormalities may lead to thoracic dystrophy and respiratory failure in the newborn period (Danks et al, 1976; Labrune et al, 1984). Other skeletal problems include clinodactyly, syndactyly, pes cavus, kyphosis, scoliosis (McLennan & Steinbach, 1974; Aggett et al, 1980), osteopenia, vertebral collapse, slipped femoral epiphysis (Ginzberg et al, 1999) and supernumerary thumb (Dror et al, 1998c) or toe (Sokolic et al, 1999). Failure to thrive is common in SDS patients and is caused by various factors, including metaphyseal dysostosis, pancreatic insufficiency, feeding difficulties and recurrent infections. Mean birth weight is at the 25th percentile, but by age 1 year and later over half of patients are still below the 3rd percentile for height. When treated with pancreatic enzyme replacement, most patients continue to show a normal growth velocity but remain consistently below the 3rd percentile for height and weight (Mack et al, 1996). Puberty is delayed in most children. Structural and functional abnormalities of the liver, including hepatomegaly or elevated serum liver enzymes, are seen in 50–75% of patients, most often in young children, and tend to resolve with age. The liver disease is usually mild and of little consequence. Its cause is not well understood. Delayed dentition of permanent teeth, dental dysplasia, increased risk of dental caries and periodontal disease may also occur. Kidneys, eyes, skin, testes, endocrine pancreas, heart, nervous system and craniofacial structures were reported to be infrequently involved (Aggett et al, 1980; Savilahti & Rapola, 1984; Cipolli et al, 1999; Ginzberg et al, 1999). The natural history of the disease is not yet defined. Morbidity and mortality in infancy are related to infections and thoracic dystrophy. At the same time, patients manifest symptoms of pancreatic dysfunction, causing great morbidity until diagnosed. Later in life, the main causes of morbidity and mortality are haematological. Cytopenias tend to fluctuate in severity but are never fully resolved. No remissions have been reported after symptomatic aplasia ensues. The outcome of patients with MDS and AML has been described above. On the basis of a literature review, Alter & Young (1997) calculated the projected median survival of SDS patients as more than 35 years. There is no specific biochemical or genetic test available for the diagnosis of SDS. Therefore, we currently diagnose the disorder by using strict criteria of clinical and laboratory features, which might be too narrow or too broad. Up to December 2000, more than 300 cases of SDS had been reported in the medical literature. Goobie et al (2001) estimated the incidence as 1/76 563, using cystic fibrosis data for the estimation. The ratio of males to females diagnosed with SDS is 1·7:1 (Ginzberg et al, 1999). The diagnosis, although occasionally delayed, is generally made in the first few years of life. Based on current knowledge of the SDS phenotype, the diagnosis requires evidence of exocrine pancreatic dysfunction and characteristic haematological abnormalities (Ginzberg et al, 1999). Short stature, skeletal abnormalities, hepatomegaly or biochemical abnormalities of the liver are supportive findings of the diagnosis. Table VI lists the clinical criteria for the diagnosis of SDS as used at our institution. Attention should be given to ruling out certain diseases, particularly cystic fibrosis (commonest cause of pancreatic insufficiency, with an abnormal sweat test), Pearson disease (pancreatic insufficiency and cytopenia, marrow ring sideroblasts and vacuolated erythroid and myeloid precursors) and cartilage hair hypoplasia (diarrhoea and cytopenia, and metaphyseal chondrodysplasia, common in the Amish population). Families with two or three affected siblings have been described and, except in rare cases, parents of patients do not manifest symptoms related to SDS. Nor are there consistent differences in the levels of serum trypsinogen between parents and healthy controls (Ginzberg et al, 2000). These data support an autosomal recessive mode of inheritance. Dale et al (2000) found no mutations in the neutrophil elastase gene (ELA2; 130130) in three patients with SDS. Inthe only reported case with constitutional karyotypic abnormalities in SDS, Masuno et al (1995) observed a denovo and apparently balanced reciprocal translocation, t(6;12)(q16.2;q21.2), in an 18-month-old girl. However, both 6q and 12q were excluded as potential sites for the candidate SDS gene by linkage studies of members of 13 SDS families with two or three affected children (Goobie et al, 1999). In 1998, we noticed an increased frequency of i(7q) in SDS (Dror et al, 1998b). We speculated that the 7(q10–11) locus was important in the pathogenesis of malignant myeloid transformation in SDS and that i(7q) was a specific marker for it. Subsequently, in a genome-wide scan of families with SDS, Goobie et al (2001) identified chromosome 7 markers that showed linkage with the disorder. Finer mapping revealed significant linkage across a broad interval that included the centromere. The maximum 2-point lod score was 8·7, with D7S473 at 7q(10–11), at a recombination fraction of 0·0. Evidence from all 15 families analysed by Goobie et al, (2001) provided support for the linkage, consistent with a single locus for SDS. However, the presence of several different mutations is likely. The data available on possible chromosomal instability in SDS are inconclusive. Tada et al (1987) and Hershkovits et al (1999) found increased frequencies of spontaneous chromosome aberrations in a patient's phytohaemagglutinin (PHA)-stimulated circulating lymphocytes, although the lymphocytes did not show increased sensitivity to mitomycin C. However, in subsequent studies, Fraccaro et al (1988) and Koiffmann et al (1991) were unable to confirm these observations. Initial studies in the late 1970s and early 1980s showed reduced bone marrow granulocyte–monocyte colony-forming units (CFU-GM) and erythrocyte burst-forming units (BFU-E) in most patients with SDS, findings compatible with a defective stem cell origin for the marrow failure (Saunders et al, 1979; Woods et al, 1981a; Suda et al, 1982). In addition, Saunders et al (1979) showed that marrow proliferative activity was normal, as assessed by determination of mitotic indices and tritiated thymidine uptake into granulocytic cells. Production of 'colony-stimulating activity' from patients' peripheral blood leucocytes appeared normal when tested on control marrow. No serum inhibitors against CFU-GM or 'colony-stimulating activity' were shown using both control and autologous marrow, and co-culture of patients' peripheral blood lymphocytes with control marrow did not inhibit CFU-GM growth. The authors concluded that committed granulocytic progenitors were proliferative in SDS and their frequency in vitro varied widely, as did the clinical neutropenia. The proliferative activity of mitotic granulocytic cells was normal and neither a deficiency of humoral stimulators nor the presence of serum or cellular inhibitors of granulopoiesis could be demonstrated. More recently, we determined that SDS is characterized by serious, generalized marrow dysfunction with an abnormal bone marrow stroma in terms of its ability to support and maintain haematopoiesis. To assess marrow stromal function, we established long-term marrow stromal cell cultures by plating normal marrow CD34+ cells over either SDS stroma (normal/SDS) or normal stroma (normal/normal) (Dror & Freedman, 1999). Non-adherent cells harvested weekly from the normal/SDS cultures were strikingly lower than normal/normal cultures. In addition to the stromal abnormality, bone marrow from patients with SDS is characterized by decreased frequency of CD34+ cells, which have a reduced ability to generate haematopoietic colonies of all lineages in vitro (Dror & Freedman, 1999). The stem cell defect comprises both the myeloid and lymphoid compartments (see above). The reduced numbers of marrow CD34+ cells and their decreased ability to produce colonies in vitro are probably related to their increased tendency to undergo apoptosis. Bone marrow mononuclear cells plated in methylcellulose cultures showed an increased percentage of apoptotic cells (mean ± SEM, 37% ± 3 vs 13% ± 2 in controls, P < 0·05) and a lower percentage of viable cells (mean ± SEM; 53% ± 7·2 vs 79% ± 3·1, P < 0·01) (Dror & Freedman, 2001). The increased tendency toward apoptosis was linked to hypersensitivity of cultured marrow cells to an agonist anti-Fas antibody, with markedly higher apoptosis and lower colony production than normal controls. Fas expression on marrow cells from patients was significantly higher than in normal controls. The difference between patients and controls in Fas expression was also significant for the subpopulations CD34−/CD38−, CD34−/CD38+ and CD34+ cell fractions, suggesting that the abnormal apoptosis process starts early during haematopoiesis (Dror & Freedman, 2001). We recently measured telomere length in SDS patients to see whether the increased apoptosis is also reflected in increased replicative stress. Compared with normal controls, mean telomere length of patients' marrow mononuclear cells was significantly shorter (Thornley et al, 2002). The mechanism for malignant myeloid transformation in MDS and the molecular steps involved in this process are not well understood. We recently compared data from four patients with malignant myeloid transformation (three RC/CG+ and one RCD/CG+) with those of the 10 other SDS patients followed at our institution (Dror et al, 1998a). The following tests did not distinguish patients with malignant myeloid transformation from the other SDS patients: severity of peripheral cytop