Prognostic Factors and Life Expectancy in Myelodysplastic Syndromes Classified According to WHO Criteria: A Basis for Clinical Decision Maki
http://www.100md.com
《临床肿瘤学》
the Division of Hematology and the Department of Internal Medicine, University of Pavia Medical School
Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo, Pavia, Italy
ABSTRACT
PURPOSE: The aim of this study was to evaluate the prognostic value of the WHO proposal, to assess the role of the main prognostic factors in myelodysplastic syndromes (MDSs) classified into WHO subgroups, and to estimate mortality (standardized mortality ratio [SMR]) and life expectancy in these groups as a basis for clinical decision making.
PATIENTS AND METHODS: Four hundred sixty-seven patients who were diagnosed as having de novo MDS at the Division of Hematology, University of Pavia (Pavia, Italy), between 1992 and 2002, were evaluated retrospectively for clinical and hematologic features at diagnosis, overall survival (OS), and progression to leukemia (leukemia-free survival).
RESULTS: Significant differences in survival were noted between patients with refractory anemia (RA), refractory cytopenia with multilineage dysplasia, RA with excess blasts, type 1 (RAEB-1), and RAEB-2. The effect of demographic factors on OS was observed in MDS patients without excess blasts (age, P = .001; sex, P = .006), as in the general population. The mortality of RA patients 70 years or older did not differ significantly from that of the general population (SMR, 1.62; P = .06). Cytogenetics was the only International Prognostic Scoring System variable showing a prognostic value in MDS classified into WHO subgroups. Transfusion-dependent patients had a significantly shorter survival than patients who did not require transfusions (P < .001). Developing a secondary iron overload significantly affected the survival of transfusion-dependent patients (P = .003).
CONCLUSION: These data show that the WHO classification of MDSs has a relevant prognostic value. This classification, along with cytogenetics, might be useful in decisions regarding transplantation. MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, for whom a conservative approach is advisable.
INTRODUCTION
Myelodysplastic syndromes (MDSs) are a heterogeneous group of disorders clinically characterized by peripheral cytopenia, followed by a progressive impairment in the ability of myelodysplastic stem cells to differentiate and an increasing risk of evolution into acute myeloid leukemia (AML).[1] The natural history of MDS, ranging from indolent conditions spanning years to forms rapidly progressing to leukemia, complicates clinical decision making regarding therapeutic modalities and timing of intervention.
In 1982, the French-American-British (FAB) cooperative group proposed a classification of MDS based on morphologic criteria.[2] For the last two decades this classification has been the gold standard for patient management and clinical investigations. However, significant heterogeneity has been noted within FAB subgroups.[3-6] Various prognostic systems have been proposed with the aim of improving the ability to predict survival and progression in MDS patients, based on clinical variables, including age, peripheral cytopenias, bone marrow blast count, lactate dehydrogenase level, and cytogenetic pattern.[7-10] In 1997, an International MDS Risk Analysis Workshop defined the International Prognostic Scoring System (IPSS), based on bone marrow blast percentage scored into four ranges, number of peripheral cytopenias, and karyotype categorized in three groups.[11] Despite some discrepancies,[12] the IPSS has been extensively validated in independent patient populations,[13,14] and has become a benchmark for clinical trials and clinical decision making.
Recently, the WHO has formulated a new proposal for the classification of MDSs. This new classification is based on variables that had the demonstrated ability to stratify survival and leukemia progression in MDS patients, such as uni- or multilineage hematopoietic involvement, blast count split into narrower ranges, and peculiar cytogenetic abnormalities.[15] The clinical utility of this proposal is now under investigation,[16-18] and attempts are being made to refine the stratification of MDS classified according to the WHO criteria, based on clinical and morphologic criteria.[19]
The aim of this study was to evaluate the prognostic value of the WHO classification in a series of 467 patients with de novo MDS from our institution, to assess the role of the main prognostic factors in MDS patients classified in the WHO subgroups, and to estimate mortality rates and life expectancy of these groups as groundwork for evidence-based clinical decision making.
PATIENTS AND METHODS
Patient Characteristics and Clinical Procedures
All patients who were diagnosed as having de novo MDS at the Division of Hematology, University of Pavia Medical School, Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo (Pavia, Italy), between 1992 and 2002, were retrospectively evaluated for clinical and hematologic features at diagnosis, survival, and progression to leukemia. The procedures followed were in accordance with the ethical standards of the Institutional Committee on Human Experimentation and with the Helsinki Declaration of 1975, as revised in 1983. The cytologic diagnosis of MDS was initially made according to the FAB criteria2]; all patients were then reclassified according to the WHO classification.[15]
Four hundred sixty-seven patients entered the analysis. Two hundred eighty-eight were male and 179 were female. The median age at diagnosis was 66 years (range, 22 to 97 years). According to the FAB criteria,[2] 187 patients were classified as having refractory anemia (RA; 40%), 53 were classified as having RA with ringed sideroblasts (RARS; 11.3%), 131 were classified as having RA with excess of blasts (RAEB; 28.1%), 43 were classified as having RAEB in transformation (9.2%), and 53 were classified as having chronic myelomonocytic leukemia (11.3%).
To reclassify patients according to the WHO criteria,[15] two independent cytologists, who were blinded to the FAB classification, reanalyzed the marrow specimens morphologically. Five hundred bone marrow nucleated cells per sample were assessed. The evaluation of the erythroid lineage was based on the detection of megaloblastic changes, nuclear lobulation, multinuclearity, and cytoplasmic granules/inclusions. In the myeloid lineage, the following abnormalities were considered: bizarre nuclear shape, hypo- or agranularity, nuclear/cytoplasmic asynchrony, and pseudo-Pelger anomaly. Micromegakaryocytes, small binucleated megakaryocytes, megakaryocytes with small round separated nuclei, and megathrombocytes were considered signs of megakaryocytic dysplasia. A bone marrow lineage dysplasia was considered in the presence of 10% or more abnormal cells.
The concordance between the two reviewers of bone marrow morphology was evaluated using Cohen's test, which resulted in a 95% agreement (kappa, 0.932; P < .001). Discordant results were resolved by a joint review of the specimens. In this classification, 76 patients were given a diagnosis of RA (16%), 34 were given a diagnosis of RARS (7%), 80 patients were given a diagnosis of refractory cytopenia with multilineage dysplasia (RCMD; 17%), 13 patients were given a diagnosis of RCMD with ringed sideroblasts (RCMD-RS; 3%), 59 patients were given a diagnosis of RA with excess blasts, type 1 (RAEB-1; 13%), 72 patients were given a diagnosis of RAEB-2 (15%), 30 patients were given a diagnosis of MDS associated with isolated del(5q) (7%), and 10 patients were given a diagnosis of MDS unclassified (2%). Forty-six patients were classified as having chronic myelomonocytic leukemia (10%), whereas 47 patients were considered to have an AML from MDS (10%). The clinical and hematologic features of patients according to WHO subgroups are listed in [Table 1].
Cytogenetic analysis was performed at diagnosis, using standard G-banding with trypsin-Giemsa staining,[20] and karyotypes were classified using the International System for Cytogenetic Nomenclature Criteria.[21] Cytogenetic data were available for 386 of the 467 patients, and clonal cytogenetic abnormalities were detected in 164 patients.
The IPSS was calculated according to Greenberg et al.[11] The score could be assessed in 310 of 374 patients diagnosed has having MDS according to the WHO criteria: 106 were classified as low risk (34%), 120 were classified as intermediate 1 (39%), 53 were classified as intermediate 2 (17%), and 31 were classified as high risk (10%).
Statistical Analysis
Numerical variables were summarized by their median and quartiles or range. Categoric variables were described by counts and relative frequencies. Actuarial probability of survival and leukemia-free survival (LFS) were estimated using the Kaplan-Meier product limit method. Overall survival (OS) was defined as the time between diagnosis and death (as a result of all causes) or end of follow-up (censored observations). LFS was calculated from diagnosis to progression to acute leukemia or end of follow-up. Patients dying as a result of any cause before leukemic evolution were considered as censored at the time of death. Twenty patients who underwent allogeneic stem-cell transplantation, as well as 14 patients who received AML-like chemotherapy after progression to leukemia, were censored at the time of therapeutic procedure. Comparisons between Kaplan-Meier curves were carried out by the Gehan's Wilcoxon test. Uni- and multivariate analyses were performed by means of Cox proportional hazards regression to identify the most significant independent prognostic factors affecting survival.
We assessed the effect on survival of developing transfusion dependency and of the onset of secondary iron overload by applying Cox models with time-dependent covariates. Standardized mortality ratios (SMRs) were calculated to compare the patients' mortality with the mortality of the general population in Italy. The SMR is the ratio between the number of deaths observed and the number of deaths expected in the study group according to a set of reference mortality rates. The Italian population mortality rates by age, sex, and calendar year were provided by the Italian Institute of Statistics. An SMR > 1 indicates a higher mortality rate than that expected in the general population. SMR values were tested by means of the score test; a P value lower than .05 indicates that the SMR is statistically significantly different from 1. All analyses were performed using Statistica software version 6.0 (1995; Statsoft Inc, Tulsa, OK) and Microsoft Excel 2000.
RESULTS
Survival Analysis of WHO Subgroups
There was a significant difference in OS between patients with RA and those with RCMD; the median survival of patients with unilineage dysplasia was 108 months, whereas it was 49 months in patients with RCMD (P < .001; [Fig 1]). Patients with RA also showed a significantly longer LFS (P = .005; [Fig 2]).
No significant difference in either OS or LFS was noted between patients with RA and those with MDS with isolated del(5q) (P = .94 and P = .89, respectively). We tested the effects of RS on survival in RA and RCMD. No significant differences were noted when the analysis was limited between RA and RARS (P = .64 and P = .19, respectively), or between RCMD and RCMD-RS (P = .12 and P = .23, respectively). To maximize the number of patients entering the analysis, we also compared patients with refractory cytopenia without ringed sideroblasts (including together RA and RCMD) versus refractory cytopenia with ringed sideroblasts (RARS and RCMD-RS). No significant differences were found in OS and LFS between the two groups (P = .09 and P = .95, respectively).
A nonsignificant difference in OS was observed between patients with RCMD and those with RAEB-1 (P = .06), but the difference in LFS between these two groups was statistically significant (P = .005). Significant differences in both OS and LFS were observed between patients with RAEB-1 and RAEB-2 (P < .001). Finally, a significant difference in OS was noted between patients with RAEB-2 and AML from MDS (P = .004).
Demographic Prognostic Factors in MDSs Classified According to WHO Criteria
Cox proportional hazards regression was applied to assess the effects of age and sex on OS and LFS of our series of patients.
Age had a significant effect on OS of the MDS population analyzed as a whole (P < .001) and stratified by WHO subgroups (P < .001): the older the age, the worse the prognosis. No significant influence was observed on the risk of progression to leukemia (P = .31). If we focus the analysis on WHO subgroups, the effect of age was statistically relevant within RA and RARS patients (P = .001), and RCMD and RCMD-RS patients (P = .002), whereas it was not significant within RAEB-1 and RAEB-2 subgroups (P = .91 and P = .52, respectively).
A significant influence of sex on OS of MDS patients was observed; male patients had the worse prognosis (hazard ratio [HR], 0.62; P = .006). Focusing the analysis on WHO categories, the effect of sex was relevant in refractory cytopenias (RCs; HR = 0.58; P = .02), but did not affect the outcome of MDS patients with excess blasts (HR = 1.2; P = .52).
We calculated the SMR in MDS patients, comparing the observed mortality with the mortality of the general Italian population standardized by sex, age, and calendar year. Considering the entire MDS population, the SMR was 7.30, which is significantly higher than 1 (P < .001). Male patients had a significantly lower SMR than did females (6.48 v 9.48; P = .008).
If we focus the analysis on patients with RA, RARS, and MDS with del(5q) (n = 132), the SMR was 2.6; again this value was significantly higher than 1 (P < .001), and again the males had a significantly lower SMR than did the females (1.95 v 3.89; P = .04). If we stratify patients by age, patients younger than 65 years (n = 68) had an SMR of 6.55, indicating a higher mortality rate than among the general population (P < .001). Patients from 65 to 70 years old (n = 22) showed a global SMR of 3.02, significantly higher than 1 (P = .009). Finally, the life expectancy of patients 70 years or older (n = 42) was not significantly lower than that of the general population (SMR = 1.62; P = .06).
Considering patients affected with RCMD and RCMD-RS, the SMR of these groups (n = 93) was significantly higher than 1 (SMR = 6.12; P < .001), as was that of the patients divided into age subgroups (patients younger than 65 years, SMR = 24.89, P < .001; patients from 65 to 70 years, SMR = 5.99, P < .001; patients 70 years or older, SMR = 3.60, P < .001); there were no significant differences between sexes.
Disease-Related Prognostic Factors in MDSs Classified According to WHO Criteria
A significant association between WHO categories and IPSS subgroups was observed (2 test, P < .001; [Table 1]). Cox proportional hazard models applied to both OS and LFS showed a significant effect of IPSS on both OS and LFS in MDS patients grouped according to the WHO criteria (P < .001).
We then tested the prognostic value of IPSS variables via multivariate and univariate Cox regression analyses. Bone marrow blast count, cytopenia, and cytogenetic abnormalities were defined and scored according to IPSS criteria. Patients with isolated del(5q) were excluded from the cytogenetic good-risk group because they constitute an autonomous entity in the WHO classification. Data on the incidence of peripheral cytopenias and karyotypic abnormalities within WHO subgroups are summarized in [Table 1].
According to the multivariate analysis, blast count and cytogenetics showed a significant predictive value on both OS (P < .001 and P = .03, respectively) and LFS (P < .001 and P = .04, respectively) of MDS patients diagnosed according to WHO criteria, whereas peripheral cytopenia did not reach a statistical significance (P = .29 and P = .27 for OS and LFS, respectively).
To assess the effect of IPSS variables within WHO categories, we performed a stratified multivariate analysis. The highly significant likelihood-ratio test (P < .001) between the stratified and the pooled model indicated different effects of blast count and cytogenetics among WHO subgroups. Hence, we performed separate analyses in each WHO category. Given that blast count categorization is similar in WHO classification and IPSS, we used the percentage of bone marrow blasts as a continuous covariate in the stratified analyses.
Karyotype significantly affected OS of both MDSs without blasts excess and RAEB-1 (HR = 1.43, P = .046; HR = 2.79, P = .008, respectively), whereas in LFS the magnitude of this effect was maintained without achieving statistical significance. No effect of cytogenetics was noticed in RAEB-2. Qualitatively similar results were obtained after grouping patients with good (normal karyotype, del(20q), or –Y) and with intermediate cytogenetic risk according to IPSS ([Fig 3]). No blast count effect was found on OS and LFS in any WHO subgroup.
Finally, we tested the ability of the WHO classification to provide a substratification within IPSS subgroups. Patients with low IPSS risk were distributed within RA/RARS (58%), MDS del(5q) (21%), and RCMD/RCMD-RS (21%). The WHO classification significantly stratified OS of patients with a low IPSS risk (HR = 2.71; P = .01). A trend was noticed also on LFS (HR = 2.19), but was not statistically significant (P = .28).
Patients with intermediate 1 risk were distributed mainly within the RA/RARS (20%), RCMD/RCMD-RS (40%), and RAEB-1 (27%) groups. Patients in the WHO subgroups (RA/RARS, RCMD/RCMD-RS, and RAEB-1) showed a significantly different OS (HR = 2.06; P = .005). There was a trend toward a difference in LFS, although this was not statistically significant (HR = 1.71; P = .06): RA/RARS patients tended to have a longer LFS than did patients in the RCMD/RCMD-RS and RAEB-1 groups.
Patients with intermediate 2 risk were distributed within RCMD/RCMD-RS (13%), RAEB-1 (26%), and RAEB-2 (51%). No significant differences in OS or LFS were observed among patients with intermediate 2 risk grouped in RCMD/RCMD-RS, RAEB-1, and RAEB-2 (HR = 1.25, P = .35; HR = 1.83, P = .06, respectively).
Survival and Life Expectancy in MDS According to Transfusion Dependency
We applied Cox regression with time-dependent covariates to evaluate the prognostic value of developing RBC transfusion dependency among patients with MDS. The OS of transfusion-dependent patients was significantly shorter than that of patients who did not develop a transfusion requirement (HR = 2.16; P < .001). LFS of patients who developed transfusion dependency was also significantly worse than that of patients who did not need transfusions (HR = 2.02; P < .001).
We then tested the effect of developing RBC transfusion dependency in a multivariate analysis with cytogenetics scored according to IPSS criteria, which was shown to be the only significant disease-related prognostic factor in WHO subgroups. Both karyotype and transfusion dependency showed a predictive value on OS (HR = 1.44, P = .001; HR = 1.72, P = .007, respectively; [Fig 4]) and LFS (HR = 1.50, P = .003; HR = 2.64, P < .001, respectively).
We then tested the effect on the survival of the transfusion burden via a Cox regression with time-dependent covariates. First, we grouped patients receiving less than 20, 20 to 40, or more than 40 RBC units during their clinical course. The number of transfusions affected both OS (HR = 1.21; P = .02) and LFS (HR = 1.39; P < .001).
However, because the total number of RBC units depends on the duration of follow-up since the onset of transfusion dependency, we also considered the effect of transfusion burden calculated as the number of transfusions per month. A significant effect was found on both OS (HR = 1.35; P < .001) and LFS (HR = 1.75; P < .001). These effects were maintained after accounting for cytogenetics.
We then assessed the survival of transfusion-dependent and nontransfused patients via a Cox model with time-dependent covariates, considering nonleukemic death as an end point. These included mainly cardiac failure (51% of patients), infection (31% of patients), hemorrhage (8% of patients), and hepatic cirrhosis (8% of patients). Among MDS patients without excess blasts who died as a result of nonleukemic causes, cardiac failure was significantly more frequent in transfusion-dependent patients (P = .01). In the first 50 months of follow-up, transfusion-dependent patients had a significantly worse survival than those who did not require transfusions (HR = 1.98; P = .01), whereas no differences were noticed after this time period.
[Table 2] shows the clinical and hematologic characteristics at diagnosis of patients within WHO subgroups, according to the development of transfusion requirement during their clinical course. Significant differences between patients who developed a transfusion dependency and those who did not were noticed in hemoglobin levels at diagnosis (P values ranging from .04 to < .001), whereas no significant differences were noticed in cytogenetics.
The negative effect of developing a transfusion requirement on OS was observed both in patients with RA, RARS, or MDS with del(5q) (HR = 3.46; P = .005), and in those with RCMD or RCMD-RS (HR = 1.87; P = .04). In RAEB-1 patients, transfusion dependency had a borderline effect (HR = 2.34; P = .09); whereas it showed no effect on the survival of RAEB-2 patients (HR = 1.47; P = .22).
The total number of RBC units did not retain significance within WHO subgroups, whereas the transfusion burden calculated as number of RBC units per month significantly affected the survival of both patients with RA, RARS, or MDS with del(5q) (HR = 1.62; P = .007) and those with RCMD or RCMD-RS (HR = 1.51; P = .02).
In multivariate analysis with cytogenetics, the number of RBC transfusions per month retained a statistical significance on OS of both patients with RA, RARS, or MDS with del(5q) (HR = 1.54; P < .001) and those with RCMD or RCMD-RS (HR = 1.45; P = .03). A negative effect was also observed on LFS of both patients with RA, RARS, or MDS with del(5q) (HR = 1.44; P = .05), and those with RCMD or RCMD-RS (HR = 1.54; P = .02). In patients with excess blasts, transfusion dependency did not have a significant effect on LFS.
Finally, we applied the Cox model with time-dependent covariates to evaluate the prognostic value of patients with MDS developing iron overload during their follow-up. We chose a serum ferritin level of 1,000 ng/mL as a threshold that distinguishes between mild and clinically relevant iron burden.[22] Considering patients with transfusion needs, this threshold was reached after they received a median number of 21 RBC units (median time from development of transfusion requirement to onset of iron overload, 10.8 months). The development of secondary iron overload significantly affected the OS (P < .001), with an HR of 1.36 every 500 ng/mL of increase in serum ferritin above the threshold. After we added the transfusion needs (number of packed red cell units per month) into the model, the effect of iron overload was maintained (HR = 1.30; P = .003). If we focus the analysis on WHO subgroups, the effect of secondary iron overload was still present in patients with RA/RARS (HR = 1.51; P < .001), whereas it was not significant in those with RCMD/RCMD-RS (HR = 1.34; P = .20).
DISCUSSION
Our results show that the WHO classification of MDSs has prognostic relevance. Among patients with MDS without excess blasts, isolated involvement of the erythroid lineage rather than bi- or trilineage marrow dysplasia is associated with a significantly better prognosis in terms of both OS and LFS. The definition of two categories of refractory anemia with excess blasts identifies two groups of patients with significantly different OS and LFS.
As reported in other studies,[7,8,11,23] the OS of male MDS patients is shorter than that of female patients. However, the SMR is significantly higher in female patients than in male patients. The effect of sex is noticeable in patients with RA and refractory cytopenia, whereas it disappears in those with MDS with excess blasts, in whom the more aggressive disease probably overwhelms demographic predictors of life expectancy.
With regard to age, we found that older patients had a shorter survival, as observed in other MDS cohorts.[7,8,11] However, as previously reported by Morel et al,[23] the SMR is significantly higher in younger patients than in older patients.
The difference between sex- and age-specific SMRs can be due to the predominance of disease-related mortality in the patients' cohort. It is a demographic fact that males have a shorter life expectancy than females of the same age group. Given that both male and female MDS patients experience a much higher mortality than the reference population, the higher SMR in females is due mainly to the lower reference mortality rates. The same mechanism may apply to the higher SMR observed in young people.
MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, whose survival is significantly affected by demographic variables rather than by disease features. The life expectancy of such patients aged 70 years or older is not significantly shorter than that of the general population.
IPSS has a significant prognostic value in MDS patients stratified into WHO subgroups. However, both the systems were based on similar criteria, in particular with regard to the ranking of bone marrow blast. When we tested the significance of the IPSS variables in the WHO subgroups, as expected bone marrow blast count failed to show a prognostic value. Likewise, the number of peripheral cytopenias did not show a significant predictive value on the outcome of either low- or high-risk MDS. Although one may assume that peripheral cytopenia is an expression of more severe lineage involvement, in our analysis this failed to show an independent prognostic significance. The only IPSS variable that maintains a prognostic value in MDS patients classified into the WHO subgroups is cytogenetics. In our analysis, the effect of karyotype abnormalities is noticeable in RA and RC as well as in RAEB-1, whereas it seems to be negligible in patients with more than 10% of blasts.
The development of transfusion dependency significantly worsens the survival of patients with MDS. Although most clinical features at diagnosis do not differ significantly between transfusion-dependent and transfusion-independent patients within WHO subgroups ([Table 2]), the significant effect of cytogenetics in multivariate analysis suggests that this might partly reflect the severity of bone marrow failure. However, development of secondary iron overload (defined as serum ferritin > 1,000 ng/mL) significantly worsens survival: after adjusting for transfusion burden, we obtained a 30% increase in hazard for every 500 ng/mL of increase in serum ferritin above the threshold.
The effect of iron overload was noticeable mainly among patients with RA, who have a longer survival and are therefore more prone to develop long-term toxicity of iron overload. In contrast, secondary iron overload did not affect the survival of patients with refractory cytopenia, who have a median survival of about 50 months. On the basis of these results, it seems possible to recommend adequate iron chelation therapy for patients with RA whose transfusion burden exceeds 20 to 25 RBC units.
The critical clinical decision in MDS is whether to perform allogeneic stem-cell transplantation, which is the only potential cure for these syndromes.[24-26] There is wide agreement on performing allogeneic transplantation in adult MDS patients with high, intermediate 2, or intermediate 1 IPSS risk, whereas consensus is less definite for patients with low risk scores.[27] A recent decision analysis suggested that a delayed transplantation strategy for low and intermediate 1 IPSS groups would maximize survival, despite the worse outcome of the transplantation procedure. According to these results, in low-risk patients, transplantation should be performed in the presence of clinical events suggestive of disease progression.[28] Although such an approach reduces the relevance of predictive systems (given that the clinical decision is based on evidence of progression), individualized follow-up is essential to avoid progression to leukemia or nonleukemic events that could preclude transplantation. Our analysis shows that WHO classification is the most significant predictor of survival in patients with low-risk and intermediate 1 IPSS MDS, and can be used, in combination with cytogenetics, to guide the choice of whether to perform a transplantation procedure or, in the context of a delayed transplantation strategy, to define the follow-up schedule. In these patients, the onset of transfusion requirements is associated with an increased hazard of progression to leukemia. Moreover, among transfusion-dependent patients, the development of iron overload worsens survival. This might support the indication for transplantation in eligible patients with low-risk MDS and a transfusion requirement.
In summary, our data show that the WHO classification of MDSs has prognostic value. This classification, along with cytogenetics, seems to be relevant to the choice of whether to perform transplantation. In particular, MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, whose outcome seems independent of disease characteristics, and in whom a conservative approach seems to be advisable. The development of secondary iron overload significantly worsens the survival of these patients. Therapeutic approaches aimed at reducing transfusion needs and at preventing iron overload are warranted.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by grants to M.C. from Associazione Italiana per la Ricerca sul Cancro and Fondazione Cariplo, Milan, Italy.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Cazzola M, Malcovati L: Myelodysplastic syndromes: Coping with ineffective hematopoiesis. N Engl J Med 352:536-538, 2005
Bennett JM, Catovsky D, Daniel MT, et al: Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 51:189-199, 1982
Sanz GF, Sanz MA, Greenberg PL: Prognostic factors and scoring systems in myelodysplastic syndromes. Haematologica 83:358-368, 1998
Verhoef GE, Pittaluga S, De Wolf-Peeters C, et al: FAB classification of myelodysplastic syndromes: Merits and controversies. Ann Hematol 71:3-11, 1995
Rosati S, Mick R, Xu F, et al: Refractory cytopenia with multilineage dysplasia: Further characterization of an ‘unclassifiable’ myelodysplastic syndrome. Leukemia 10:20-26, 1996
Boultwood J, Lewis S, Wainscoat JS: The 5q-syndrome. Blood 84:3253-3260, 1994
Morel P, Hebbar M, Lai JL, et al: Cytogenetic analysis has strong independent prognostic value in de novo myelodysplastic syndromes and can be incorporated in a new scoring system: A report on 408 cases. Leukemia 7:1315-1323, 1993
Sanz GF, Sanz MA, Vallespi T, et al: Two regression models and a scoring system for predicting survival and planning treatment in myelodysplastic syndromes: A multivariate analysis of prognostic factors in 370 patients. Blood 74:395-408, 1989
Mufti GJ, Stevens JR, Oscier DG, et al: Myelodysplastic syndromes: A scoring system with prognostic significance. Br J Haematol 59:425-433, 1985
Aul C, Gattermann N, Heyll A, et al: Primary myelodysplastic syndromes: Analysis of prognostic factors in 235 patients and proposals for an improved scoring system. Leukemia 6:52-59, 1992
Greenberg P, Cox C, LeBeau MM, et al: International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 89:2079-2088, 1997
Estey E, Keating M, Pierce S, et al: Application of the International Scoring System for myelodysplasia to M.D. Anderson patients. Blood 90:2843-2846, 1997
Pfeilstocker M, Reisner R, Nosslinger T, et al: Cross-validation of prognostic scores in myelodysplastic syndromes on 386 patients from a single institution confirms importance of cytogenetics. Br J Haematol 106:455-463, 1999
Takahashi M, Takahashi H, Aizawa Y, et al: Usefulness of IPSS for the patients with refractory anemia. Am J Hematol 58:250-252, 1998
Vardiman JW, Harris NL, Brunning RD: The World Health Organization (WHO) classification of the myeloid neoplasms. Blood 100:2292-2302, 2002
Germing U, Gattermann N, Strupp C, et al: Validation of the WHO proposals for a new classification of primary myelodysplastic syndromes: A retrospective analysis of 1600 patients. Leuk Res 24:983-992, 2000
Nosslinger T, Reisner R, Koller E, et al: Myelodysplastic syndromes, from French-American-British to World Health Organization: Comparison of classifications on 431 unselected patients from a single institution. Blood 98:2935-2941, 2001
Howe RB, Porwit-MacDonald A, Wanat R, et al: The WHO classification of MDS does make a difference. Blood 103:3265-3270, 2004
Verburgh ER, Achten R, Brusselmans C, et al: A new disease definition of low grade MDS based on the expression of cytopenia and dysplasia in one or more lineages at diagnosis improves on the FAB and WHO classification. Hematol J 5:S90-S91, 2004 (suppl 2)
Bernasconi P, Astori C, Cavigliano PM, et al: Cytogenetic and FISH analyses in five patients with hypoplastic bone marrow. Leukemia 14:1322-1323, 2000
An International System for Human Cytogenetic Nomenclature: (1985) ISCN 1985—Report of the Standing Committee on Human Cytogenetic Nomenclature. Birth Defects Orig Artic Ser 21:1-117, 1985
Bowen D, Culligan D, Jowitt S, et al: Guidelines for the diagnosis and therapy of adult myelodysplastic syndromes. Br J Haematol 120:187-200, 2003
Morel P, Declercq C, Hebbar M, et al: Prognostic factors in myelodysplastic syndromes: Critical analysis of the impact of age and gender and failure to identify a very-low-risk group using standard mortality ratio techniques. Br J Haematol 94:116-119, 1996
Anderson JE, Appelbaum FR, Fisher LD, et al: Allogeneic bone marrow transplantation for 93 patients with myelodysplastic syndrome. Blood 82:677-681, 1993
de Witte T, Hermans J, Vossen J, et al: Haematopoietic stem cell transplantation for patients with myelo-dysplastic syndromes and secondary acute myeloid leukaemias: A report on behalf of the Chronic Leukaemia Working Party of the European Group for Blood and Marrow Transplantation (EBMT). Br J Haematol 110:620-630, 2000
Sierra J, Perez WS, Rozman C, et al: Bone marrow transplantation from HLA-identical siblings as treatment for myelodysplasia. Blood 100:1997-2004, 2002
Alessandrino EP, Amadori S, Barosi G, et al: Evidence- and consensus-based practice guidelines for the therapy of primary myelodysplastic syndromes: A statement from the Italian Society of Hematology. Haematologica 87:1286-1306, 2002
Cutler CS, Lee SJ, Greenberg P, et al: A decision analysis of allogeneic bone marrow transplantation for the myelodysplastic syndromes: Delayed transplantation for low risk myelodysplasia is associated with improved outcome. Blood 104:579-585, 2004(Luca Malcovati, Matteo Gi)
Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo, Pavia, Italy
ABSTRACT
PURPOSE: The aim of this study was to evaluate the prognostic value of the WHO proposal, to assess the role of the main prognostic factors in myelodysplastic syndromes (MDSs) classified into WHO subgroups, and to estimate mortality (standardized mortality ratio [SMR]) and life expectancy in these groups as a basis for clinical decision making.
PATIENTS AND METHODS: Four hundred sixty-seven patients who were diagnosed as having de novo MDS at the Division of Hematology, University of Pavia (Pavia, Italy), between 1992 and 2002, were evaluated retrospectively for clinical and hematologic features at diagnosis, overall survival (OS), and progression to leukemia (leukemia-free survival).
RESULTS: Significant differences in survival were noted between patients with refractory anemia (RA), refractory cytopenia with multilineage dysplasia, RA with excess blasts, type 1 (RAEB-1), and RAEB-2. The effect of demographic factors on OS was observed in MDS patients without excess blasts (age, P = .001; sex, P = .006), as in the general population. The mortality of RA patients 70 years or older did not differ significantly from that of the general population (SMR, 1.62; P = .06). Cytogenetics was the only International Prognostic Scoring System variable showing a prognostic value in MDS classified into WHO subgroups. Transfusion-dependent patients had a significantly shorter survival than patients who did not require transfusions (P < .001). Developing a secondary iron overload significantly affected the survival of transfusion-dependent patients (P = .003).
CONCLUSION: These data show that the WHO classification of MDSs has a relevant prognostic value. This classification, along with cytogenetics, might be useful in decisions regarding transplantation. MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, for whom a conservative approach is advisable.
INTRODUCTION
Myelodysplastic syndromes (MDSs) are a heterogeneous group of disorders clinically characterized by peripheral cytopenia, followed by a progressive impairment in the ability of myelodysplastic stem cells to differentiate and an increasing risk of evolution into acute myeloid leukemia (AML).[1] The natural history of MDS, ranging from indolent conditions spanning years to forms rapidly progressing to leukemia, complicates clinical decision making regarding therapeutic modalities and timing of intervention.
In 1982, the French-American-British (FAB) cooperative group proposed a classification of MDS based on morphologic criteria.[2] For the last two decades this classification has been the gold standard for patient management and clinical investigations. However, significant heterogeneity has been noted within FAB subgroups.[3-6] Various prognostic systems have been proposed with the aim of improving the ability to predict survival and progression in MDS patients, based on clinical variables, including age, peripheral cytopenias, bone marrow blast count, lactate dehydrogenase level, and cytogenetic pattern.[7-10] In 1997, an International MDS Risk Analysis Workshop defined the International Prognostic Scoring System (IPSS), based on bone marrow blast percentage scored into four ranges, number of peripheral cytopenias, and karyotype categorized in three groups.[11] Despite some discrepancies,[12] the IPSS has been extensively validated in independent patient populations,[13,14] and has become a benchmark for clinical trials and clinical decision making.
Recently, the WHO has formulated a new proposal for the classification of MDSs. This new classification is based on variables that had the demonstrated ability to stratify survival and leukemia progression in MDS patients, such as uni- or multilineage hematopoietic involvement, blast count split into narrower ranges, and peculiar cytogenetic abnormalities.[15] The clinical utility of this proposal is now under investigation,[16-18] and attempts are being made to refine the stratification of MDS classified according to the WHO criteria, based on clinical and morphologic criteria.[19]
The aim of this study was to evaluate the prognostic value of the WHO classification in a series of 467 patients with de novo MDS from our institution, to assess the role of the main prognostic factors in MDS patients classified in the WHO subgroups, and to estimate mortality rates and life expectancy of these groups as groundwork for evidence-based clinical decision making.
PATIENTS AND METHODS
Patient Characteristics and Clinical Procedures
All patients who were diagnosed as having de novo MDS at the Division of Hematology, University of Pavia Medical School, Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo (Pavia, Italy), between 1992 and 2002, were retrospectively evaluated for clinical and hematologic features at diagnosis, survival, and progression to leukemia. The procedures followed were in accordance with the ethical standards of the Institutional Committee on Human Experimentation and with the Helsinki Declaration of 1975, as revised in 1983. The cytologic diagnosis of MDS was initially made according to the FAB criteria2]; all patients were then reclassified according to the WHO classification.[15]
Four hundred sixty-seven patients entered the analysis. Two hundred eighty-eight were male and 179 were female. The median age at diagnosis was 66 years (range, 22 to 97 years). According to the FAB criteria,[2] 187 patients were classified as having refractory anemia (RA; 40%), 53 were classified as having RA with ringed sideroblasts (RARS; 11.3%), 131 were classified as having RA with excess of blasts (RAEB; 28.1%), 43 were classified as having RAEB in transformation (9.2%), and 53 were classified as having chronic myelomonocytic leukemia (11.3%).
To reclassify patients according to the WHO criteria,[15] two independent cytologists, who were blinded to the FAB classification, reanalyzed the marrow specimens morphologically. Five hundred bone marrow nucleated cells per sample were assessed. The evaluation of the erythroid lineage was based on the detection of megaloblastic changes, nuclear lobulation, multinuclearity, and cytoplasmic granules/inclusions. In the myeloid lineage, the following abnormalities were considered: bizarre nuclear shape, hypo- or agranularity, nuclear/cytoplasmic asynchrony, and pseudo-Pelger anomaly. Micromegakaryocytes, small binucleated megakaryocytes, megakaryocytes with small round separated nuclei, and megathrombocytes were considered signs of megakaryocytic dysplasia. A bone marrow lineage dysplasia was considered in the presence of 10% or more abnormal cells.
The concordance between the two reviewers of bone marrow morphology was evaluated using Cohen's test, which resulted in a 95% agreement (kappa, 0.932; P < .001). Discordant results were resolved by a joint review of the specimens. In this classification, 76 patients were given a diagnosis of RA (16%), 34 were given a diagnosis of RARS (7%), 80 patients were given a diagnosis of refractory cytopenia with multilineage dysplasia (RCMD; 17%), 13 patients were given a diagnosis of RCMD with ringed sideroblasts (RCMD-RS; 3%), 59 patients were given a diagnosis of RA with excess blasts, type 1 (RAEB-1; 13%), 72 patients were given a diagnosis of RAEB-2 (15%), 30 patients were given a diagnosis of MDS associated with isolated del(5q) (7%), and 10 patients were given a diagnosis of MDS unclassified (2%). Forty-six patients were classified as having chronic myelomonocytic leukemia (10%), whereas 47 patients were considered to have an AML from MDS (10%). The clinical and hematologic features of patients according to WHO subgroups are listed in [Table 1].
Cytogenetic analysis was performed at diagnosis, using standard G-banding with trypsin-Giemsa staining,[20] and karyotypes were classified using the International System for Cytogenetic Nomenclature Criteria.[21] Cytogenetic data were available for 386 of the 467 patients, and clonal cytogenetic abnormalities were detected in 164 patients.
The IPSS was calculated according to Greenberg et al.[11] The score could be assessed in 310 of 374 patients diagnosed has having MDS according to the WHO criteria: 106 were classified as low risk (34%), 120 were classified as intermediate 1 (39%), 53 were classified as intermediate 2 (17%), and 31 were classified as high risk (10%).
Statistical Analysis
Numerical variables were summarized by their median and quartiles or range. Categoric variables were described by counts and relative frequencies. Actuarial probability of survival and leukemia-free survival (LFS) were estimated using the Kaplan-Meier product limit method. Overall survival (OS) was defined as the time between diagnosis and death (as a result of all causes) or end of follow-up (censored observations). LFS was calculated from diagnosis to progression to acute leukemia or end of follow-up. Patients dying as a result of any cause before leukemic evolution were considered as censored at the time of death. Twenty patients who underwent allogeneic stem-cell transplantation, as well as 14 patients who received AML-like chemotherapy after progression to leukemia, were censored at the time of therapeutic procedure. Comparisons between Kaplan-Meier curves were carried out by the Gehan's Wilcoxon test. Uni- and multivariate analyses were performed by means of Cox proportional hazards regression to identify the most significant independent prognostic factors affecting survival.
We assessed the effect on survival of developing transfusion dependency and of the onset of secondary iron overload by applying Cox models with time-dependent covariates. Standardized mortality ratios (SMRs) were calculated to compare the patients' mortality with the mortality of the general population in Italy. The SMR is the ratio between the number of deaths observed and the number of deaths expected in the study group according to a set of reference mortality rates. The Italian population mortality rates by age, sex, and calendar year were provided by the Italian Institute of Statistics. An SMR > 1 indicates a higher mortality rate than that expected in the general population. SMR values were tested by means of the score test; a P value lower than .05 indicates that the SMR is statistically significantly different from 1. All analyses were performed using Statistica software version 6.0 (1995; Statsoft Inc, Tulsa, OK) and Microsoft Excel 2000.
RESULTS
Survival Analysis of WHO Subgroups
There was a significant difference in OS between patients with RA and those with RCMD; the median survival of patients with unilineage dysplasia was 108 months, whereas it was 49 months in patients with RCMD (P < .001; [Fig 1]). Patients with RA also showed a significantly longer LFS (P = .005; [Fig 2]).
No significant difference in either OS or LFS was noted between patients with RA and those with MDS with isolated del(5q) (P = .94 and P = .89, respectively). We tested the effects of RS on survival in RA and RCMD. No significant differences were noted when the analysis was limited between RA and RARS (P = .64 and P = .19, respectively), or between RCMD and RCMD-RS (P = .12 and P = .23, respectively). To maximize the number of patients entering the analysis, we also compared patients with refractory cytopenia without ringed sideroblasts (including together RA and RCMD) versus refractory cytopenia with ringed sideroblasts (RARS and RCMD-RS). No significant differences were found in OS and LFS between the two groups (P = .09 and P = .95, respectively).
A nonsignificant difference in OS was observed between patients with RCMD and those with RAEB-1 (P = .06), but the difference in LFS between these two groups was statistically significant (P = .005). Significant differences in both OS and LFS were observed between patients with RAEB-1 and RAEB-2 (P < .001). Finally, a significant difference in OS was noted between patients with RAEB-2 and AML from MDS (P = .004).
Demographic Prognostic Factors in MDSs Classified According to WHO Criteria
Cox proportional hazards regression was applied to assess the effects of age and sex on OS and LFS of our series of patients.
Age had a significant effect on OS of the MDS population analyzed as a whole (P < .001) and stratified by WHO subgroups (P < .001): the older the age, the worse the prognosis. No significant influence was observed on the risk of progression to leukemia (P = .31). If we focus the analysis on WHO subgroups, the effect of age was statistically relevant within RA and RARS patients (P = .001), and RCMD and RCMD-RS patients (P = .002), whereas it was not significant within RAEB-1 and RAEB-2 subgroups (P = .91 and P = .52, respectively).
A significant influence of sex on OS of MDS patients was observed; male patients had the worse prognosis (hazard ratio [HR], 0.62; P = .006). Focusing the analysis on WHO categories, the effect of sex was relevant in refractory cytopenias (RCs; HR = 0.58; P = .02), but did not affect the outcome of MDS patients with excess blasts (HR = 1.2; P = .52).
We calculated the SMR in MDS patients, comparing the observed mortality with the mortality of the general Italian population standardized by sex, age, and calendar year. Considering the entire MDS population, the SMR was 7.30, which is significantly higher than 1 (P < .001). Male patients had a significantly lower SMR than did females (6.48 v 9.48; P = .008).
If we focus the analysis on patients with RA, RARS, and MDS with del(5q) (n = 132), the SMR was 2.6; again this value was significantly higher than 1 (P < .001), and again the males had a significantly lower SMR than did the females (1.95 v 3.89; P = .04). If we stratify patients by age, patients younger than 65 years (n = 68) had an SMR of 6.55, indicating a higher mortality rate than among the general population (P < .001). Patients from 65 to 70 years old (n = 22) showed a global SMR of 3.02, significantly higher than 1 (P = .009). Finally, the life expectancy of patients 70 years or older (n = 42) was not significantly lower than that of the general population (SMR = 1.62; P = .06).
Considering patients affected with RCMD and RCMD-RS, the SMR of these groups (n = 93) was significantly higher than 1 (SMR = 6.12; P < .001), as was that of the patients divided into age subgroups (patients younger than 65 years, SMR = 24.89, P < .001; patients from 65 to 70 years, SMR = 5.99, P < .001; patients 70 years or older, SMR = 3.60, P < .001); there were no significant differences between sexes.
Disease-Related Prognostic Factors in MDSs Classified According to WHO Criteria
A significant association between WHO categories and IPSS subgroups was observed (2 test, P < .001; [Table 1]). Cox proportional hazard models applied to both OS and LFS showed a significant effect of IPSS on both OS and LFS in MDS patients grouped according to the WHO criteria (P < .001).
We then tested the prognostic value of IPSS variables via multivariate and univariate Cox regression analyses. Bone marrow blast count, cytopenia, and cytogenetic abnormalities were defined and scored according to IPSS criteria. Patients with isolated del(5q) were excluded from the cytogenetic good-risk group because they constitute an autonomous entity in the WHO classification. Data on the incidence of peripheral cytopenias and karyotypic abnormalities within WHO subgroups are summarized in [Table 1].
According to the multivariate analysis, blast count and cytogenetics showed a significant predictive value on both OS (P < .001 and P = .03, respectively) and LFS (P < .001 and P = .04, respectively) of MDS patients diagnosed according to WHO criteria, whereas peripheral cytopenia did not reach a statistical significance (P = .29 and P = .27 for OS and LFS, respectively).
To assess the effect of IPSS variables within WHO categories, we performed a stratified multivariate analysis. The highly significant likelihood-ratio test (P < .001) between the stratified and the pooled model indicated different effects of blast count and cytogenetics among WHO subgroups. Hence, we performed separate analyses in each WHO category. Given that blast count categorization is similar in WHO classification and IPSS, we used the percentage of bone marrow blasts as a continuous covariate in the stratified analyses.
Karyotype significantly affected OS of both MDSs without blasts excess and RAEB-1 (HR = 1.43, P = .046; HR = 2.79, P = .008, respectively), whereas in LFS the magnitude of this effect was maintained without achieving statistical significance. No effect of cytogenetics was noticed in RAEB-2. Qualitatively similar results were obtained after grouping patients with good (normal karyotype, del(20q), or –Y) and with intermediate cytogenetic risk according to IPSS ([Fig 3]). No blast count effect was found on OS and LFS in any WHO subgroup.
Finally, we tested the ability of the WHO classification to provide a substratification within IPSS subgroups. Patients with low IPSS risk were distributed within RA/RARS (58%), MDS del(5q) (21%), and RCMD/RCMD-RS (21%). The WHO classification significantly stratified OS of patients with a low IPSS risk (HR = 2.71; P = .01). A trend was noticed also on LFS (HR = 2.19), but was not statistically significant (P = .28).
Patients with intermediate 1 risk were distributed mainly within the RA/RARS (20%), RCMD/RCMD-RS (40%), and RAEB-1 (27%) groups. Patients in the WHO subgroups (RA/RARS, RCMD/RCMD-RS, and RAEB-1) showed a significantly different OS (HR = 2.06; P = .005). There was a trend toward a difference in LFS, although this was not statistically significant (HR = 1.71; P = .06): RA/RARS patients tended to have a longer LFS than did patients in the RCMD/RCMD-RS and RAEB-1 groups.
Patients with intermediate 2 risk were distributed within RCMD/RCMD-RS (13%), RAEB-1 (26%), and RAEB-2 (51%). No significant differences in OS or LFS were observed among patients with intermediate 2 risk grouped in RCMD/RCMD-RS, RAEB-1, and RAEB-2 (HR = 1.25, P = .35; HR = 1.83, P = .06, respectively).
Survival and Life Expectancy in MDS According to Transfusion Dependency
We applied Cox regression with time-dependent covariates to evaluate the prognostic value of developing RBC transfusion dependency among patients with MDS. The OS of transfusion-dependent patients was significantly shorter than that of patients who did not develop a transfusion requirement (HR = 2.16; P < .001). LFS of patients who developed transfusion dependency was also significantly worse than that of patients who did not need transfusions (HR = 2.02; P < .001).
We then tested the effect of developing RBC transfusion dependency in a multivariate analysis with cytogenetics scored according to IPSS criteria, which was shown to be the only significant disease-related prognostic factor in WHO subgroups. Both karyotype and transfusion dependency showed a predictive value on OS (HR = 1.44, P = .001; HR = 1.72, P = .007, respectively; [Fig 4]) and LFS (HR = 1.50, P = .003; HR = 2.64, P < .001, respectively).
We then tested the effect on the survival of the transfusion burden via a Cox regression with time-dependent covariates. First, we grouped patients receiving less than 20, 20 to 40, or more than 40 RBC units during their clinical course. The number of transfusions affected both OS (HR = 1.21; P = .02) and LFS (HR = 1.39; P < .001).
However, because the total number of RBC units depends on the duration of follow-up since the onset of transfusion dependency, we also considered the effect of transfusion burden calculated as the number of transfusions per month. A significant effect was found on both OS (HR = 1.35; P < .001) and LFS (HR = 1.75; P < .001). These effects were maintained after accounting for cytogenetics.
We then assessed the survival of transfusion-dependent and nontransfused patients via a Cox model with time-dependent covariates, considering nonleukemic death as an end point. These included mainly cardiac failure (51% of patients), infection (31% of patients), hemorrhage (8% of patients), and hepatic cirrhosis (8% of patients). Among MDS patients without excess blasts who died as a result of nonleukemic causes, cardiac failure was significantly more frequent in transfusion-dependent patients (P = .01). In the first 50 months of follow-up, transfusion-dependent patients had a significantly worse survival than those who did not require transfusions (HR = 1.98; P = .01), whereas no differences were noticed after this time period.
[Table 2] shows the clinical and hematologic characteristics at diagnosis of patients within WHO subgroups, according to the development of transfusion requirement during their clinical course. Significant differences between patients who developed a transfusion dependency and those who did not were noticed in hemoglobin levels at diagnosis (P values ranging from .04 to < .001), whereas no significant differences were noticed in cytogenetics.
The negative effect of developing a transfusion requirement on OS was observed both in patients with RA, RARS, or MDS with del(5q) (HR = 3.46; P = .005), and in those with RCMD or RCMD-RS (HR = 1.87; P = .04). In RAEB-1 patients, transfusion dependency had a borderline effect (HR = 2.34; P = .09); whereas it showed no effect on the survival of RAEB-2 patients (HR = 1.47; P = .22).
The total number of RBC units did not retain significance within WHO subgroups, whereas the transfusion burden calculated as number of RBC units per month significantly affected the survival of both patients with RA, RARS, or MDS with del(5q) (HR = 1.62; P = .007) and those with RCMD or RCMD-RS (HR = 1.51; P = .02).
In multivariate analysis with cytogenetics, the number of RBC transfusions per month retained a statistical significance on OS of both patients with RA, RARS, or MDS with del(5q) (HR = 1.54; P < .001) and those with RCMD or RCMD-RS (HR = 1.45; P = .03). A negative effect was also observed on LFS of both patients with RA, RARS, or MDS with del(5q) (HR = 1.44; P = .05), and those with RCMD or RCMD-RS (HR = 1.54; P = .02). In patients with excess blasts, transfusion dependency did not have a significant effect on LFS.
Finally, we applied the Cox model with time-dependent covariates to evaluate the prognostic value of patients with MDS developing iron overload during their follow-up. We chose a serum ferritin level of 1,000 ng/mL as a threshold that distinguishes between mild and clinically relevant iron burden.[22] Considering patients with transfusion needs, this threshold was reached after they received a median number of 21 RBC units (median time from development of transfusion requirement to onset of iron overload, 10.8 months). The development of secondary iron overload significantly affected the OS (P < .001), with an HR of 1.36 every 500 ng/mL of increase in serum ferritin above the threshold. After we added the transfusion needs (number of packed red cell units per month) into the model, the effect of iron overload was maintained (HR = 1.30; P = .003). If we focus the analysis on WHO subgroups, the effect of secondary iron overload was still present in patients with RA/RARS (HR = 1.51; P < .001), whereas it was not significant in those with RCMD/RCMD-RS (HR = 1.34; P = .20).
DISCUSSION
Our results show that the WHO classification of MDSs has prognostic relevance. Among patients with MDS without excess blasts, isolated involvement of the erythroid lineage rather than bi- or trilineage marrow dysplasia is associated with a significantly better prognosis in terms of both OS and LFS. The definition of two categories of refractory anemia with excess blasts identifies two groups of patients with significantly different OS and LFS.
As reported in other studies,[7,8,11,23] the OS of male MDS patients is shorter than that of female patients. However, the SMR is significantly higher in female patients than in male patients. The effect of sex is noticeable in patients with RA and refractory cytopenia, whereas it disappears in those with MDS with excess blasts, in whom the more aggressive disease probably overwhelms demographic predictors of life expectancy.
With regard to age, we found that older patients had a shorter survival, as observed in other MDS cohorts.[7,8,11] However, as previously reported by Morel et al,[23] the SMR is significantly higher in younger patients than in older patients.
The difference between sex- and age-specific SMRs can be due to the predominance of disease-related mortality in the patients' cohort. It is a demographic fact that males have a shorter life expectancy than females of the same age group. Given that both male and female MDS patients experience a much higher mortality than the reference population, the higher SMR in females is due mainly to the lower reference mortality rates. The same mechanism may apply to the higher SMR observed in young people.
MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, whose survival is significantly affected by demographic variables rather than by disease features. The life expectancy of such patients aged 70 years or older is not significantly shorter than that of the general population.
IPSS has a significant prognostic value in MDS patients stratified into WHO subgroups. However, both the systems were based on similar criteria, in particular with regard to the ranking of bone marrow blast. When we tested the significance of the IPSS variables in the WHO subgroups, as expected bone marrow blast count failed to show a prognostic value. Likewise, the number of peripheral cytopenias did not show a significant predictive value on the outcome of either low- or high-risk MDS. Although one may assume that peripheral cytopenia is an expression of more severe lineage involvement, in our analysis this failed to show an independent prognostic significance. The only IPSS variable that maintains a prognostic value in MDS patients classified into the WHO subgroups is cytogenetics. In our analysis, the effect of karyotype abnormalities is noticeable in RA and RC as well as in RAEB-1, whereas it seems to be negligible in patients with more than 10% of blasts.
The development of transfusion dependency significantly worsens the survival of patients with MDS. Although most clinical features at diagnosis do not differ significantly between transfusion-dependent and transfusion-independent patients within WHO subgroups ([Table 2]), the significant effect of cytogenetics in multivariate analysis suggests that this might partly reflect the severity of bone marrow failure. However, development of secondary iron overload (defined as serum ferritin > 1,000 ng/mL) significantly worsens survival: after adjusting for transfusion burden, we obtained a 30% increase in hazard for every 500 ng/mL of increase in serum ferritin above the threshold.
The effect of iron overload was noticeable mainly among patients with RA, who have a longer survival and are therefore more prone to develop long-term toxicity of iron overload. In contrast, secondary iron overload did not affect the survival of patients with refractory cytopenia, who have a median survival of about 50 months. On the basis of these results, it seems possible to recommend adequate iron chelation therapy for patients with RA whose transfusion burden exceeds 20 to 25 RBC units.
The critical clinical decision in MDS is whether to perform allogeneic stem-cell transplantation, which is the only potential cure for these syndromes.[24-26] There is wide agreement on performing allogeneic transplantation in adult MDS patients with high, intermediate 2, or intermediate 1 IPSS risk, whereas consensus is less definite for patients with low risk scores.[27] A recent decision analysis suggested that a delayed transplantation strategy for low and intermediate 1 IPSS groups would maximize survival, despite the worse outcome of the transplantation procedure. According to these results, in low-risk patients, transplantation should be performed in the presence of clinical events suggestive of disease progression.[28] Although such an approach reduces the relevance of predictive systems (given that the clinical decision is based on evidence of progression), individualized follow-up is essential to avoid progression to leukemia or nonleukemic events that could preclude transplantation. Our analysis shows that WHO classification is the most significant predictor of survival in patients with low-risk and intermediate 1 IPSS MDS, and can be used, in combination with cytogenetics, to guide the choice of whether to perform a transplantation procedure or, in the context of a delayed transplantation strategy, to define the follow-up schedule. In these patients, the onset of transfusion requirements is associated with an increased hazard of progression to leukemia. Moreover, among transfusion-dependent patients, the development of iron overload worsens survival. This might support the indication for transplantation in eligible patients with low-risk MDS and a transfusion requirement.
In summary, our data show that the WHO classification of MDSs has prognostic value. This classification, along with cytogenetics, seems to be relevant to the choice of whether to perform transplantation. In particular, MDS with isolated erythroid lineage dysplasia identifies a subset of truly low-risk patients, whose outcome seems independent of disease characteristics, and in whom a conservative approach seems to be advisable. The development of secondary iron overload significantly worsens the survival of these patients. Therapeutic approaches aimed at reducing transfusion needs and at preventing iron overload are warranted.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by grants to M.C. from Associazione Italiana per la Ricerca sul Cancro and Fondazione Cariplo, Milan, Italy.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Cazzola M, Malcovati L: Myelodysplastic syndromes: Coping with ineffective hematopoiesis. N Engl J Med 352:536-538, 2005
Bennett JM, Catovsky D, Daniel MT, et al: Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 51:189-199, 1982
Sanz GF, Sanz MA, Greenberg PL: Prognostic factors and scoring systems in myelodysplastic syndromes. Haematologica 83:358-368, 1998
Verhoef GE, Pittaluga S, De Wolf-Peeters C, et al: FAB classification of myelodysplastic syndromes: Merits and controversies. Ann Hematol 71:3-11, 1995
Rosati S, Mick R, Xu F, et al: Refractory cytopenia with multilineage dysplasia: Further characterization of an ‘unclassifiable’ myelodysplastic syndrome. Leukemia 10:20-26, 1996
Boultwood J, Lewis S, Wainscoat JS: The 5q-syndrome. Blood 84:3253-3260, 1994
Morel P, Hebbar M, Lai JL, et al: Cytogenetic analysis has strong independent prognostic value in de novo myelodysplastic syndromes and can be incorporated in a new scoring system: A report on 408 cases. Leukemia 7:1315-1323, 1993
Sanz GF, Sanz MA, Vallespi T, et al: Two regression models and a scoring system for predicting survival and planning treatment in myelodysplastic syndromes: A multivariate analysis of prognostic factors in 370 patients. Blood 74:395-408, 1989
Mufti GJ, Stevens JR, Oscier DG, et al: Myelodysplastic syndromes: A scoring system with prognostic significance. Br J Haematol 59:425-433, 1985
Aul C, Gattermann N, Heyll A, et al: Primary myelodysplastic syndromes: Analysis of prognostic factors in 235 patients and proposals for an improved scoring system. Leukemia 6:52-59, 1992
Greenberg P, Cox C, LeBeau MM, et al: International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 89:2079-2088, 1997
Estey E, Keating M, Pierce S, et al: Application of the International Scoring System for myelodysplasia to M.D. Anderson patients. Blood 90:2843-2846, 1997
Pfeilstocker M, Reisner R, Nosslinger T, et al: Cross-validation of prognostic scores in myelodysplastic syndromes on 386 patients from a single institution confirms importance of cytogenetics. Br J Haematol 106:455-463, 1999
Takahashi M, Takahashi H, Aizawa Y, et al: Usefulness of IPSS for the patients with refractory anemia. Am J Hematol 58:250-252, 1998
Vardiman JW, Harris NL, Brunning RD: The World Health Organization (WHO) classification of the myeloid neoplasms. Blood 100:2292-2302, 2002
Germing U, Gattermann N, Strupp C, et al: Validation of the WHO proposals for a new classification of primary myelodysplastic syndromes: A retrospective analysis of 1600 patients. Leuk Res 24:983-992, 2000
Nosslinger T, Reisner R, Koller E, et al: Myelodysplastic syndromes, from French-American-British to World Health Organization: Comparison of classifications on 431 unselected patients from a single institution. Blood 98:2935-2941, 2001
Howe RB, Porwit-MacDonald A, Wanat R, et al: The WHO classification of MDS does make a difference. Blood 103:3265-3270, 2004
Verburgh ER, Achten R, Brusselmans C, et al: A new disease definition of low grade MDS based on the expression of cytopenia and dysplasia in one or more lineages at diagnosis improves on the FAB and WHO classification. Hematol J 5:S90-S91, 2004 (suppl 2)
Bernasconi P, Astori C, Cavigliano PM, et al: Cytogenetic and FISH analyses in five patients with hypoplastic bone marrow. Leukemia 14:1322-1323, 2000
An International System for Human Cytogenetic Nomenclature: (1985) ISCN 1985—Report of the Standing Committee on Human Cytogenetic Nomenclature. Birth Defects Orig Artic Ser 21:1-117, 1985
Bowen D, Culligan D, Jowitt S, et al: Guidelines for the diagnosis and therapy of adult myelodysplastic syndromes. Br J Haematol 120:187-200, 2003
Morel P, Declercq C, Hebbar M, et al: Prognostic factors in myelodysplastic syndromes: Critical analysis of the impact of age and gender and failure to identify a very-low-risk group using standard mortality ratio techniques. Br J Haematol 94:116-119, 1996
Anderson JE, Appelbaum FR, Fisher LD, et al: Allogeneic bone marrow transplantation for 93 patients with myelodysplastic syndrome. Blood 82:677-681, 1993
de Witte T, Hermans J, Vossen J, et al: Haematopoietic stem cell transplantation for patients with myelo-dysplastic syndromes and secondary acute myeloid leukaemias: A report on behalf of the Chronic Leukaemia Working Party of the European Group for Blood and Marrow Transplantation (EBMT). Br J Haematol 110:620-630, 2000
Sierra J, Perez WS, Rozman C, et al: Bone marrow transplantation from HLA-identical siblings as treatment for myelodysplasia. Blood 100:1997-2004, 2002
Alessandrino EP, Amadori S, Barosi G, et al: Evidence- and consensus-based practice guidelines for the therapy of primary myelodysplastic syndromes: A statement from the Italian Society of Hematology. Haematologica 87:1286-1306, 2002
Cutler CS, Lee SJ, Greenberg P, et al: A decision analysis of allogeneic bone marrow transplantation for the myelodysplastic syndromes: Delayed transplantation for low risk myelodysplasia is associated with improved outcome. Blood 104:579-585, 2004(Luca Malcovati, Matteo Gi)