Guidelines on the diagnosis and management of chronic lymphocytic leukaemia

Transcription

1 guideline Guidelines on the diagnosis and management of chronic lymphocytic leukaemia Methods The purpose of this guideline is to provide a rational approach to the diagnosis and management of patients with chronic lymphocytic leukaemia (CLL). This guideline has been compiled by the Guidelines Working Group of the UK CLL Forum on behalf of the British Committee for Standards in Haematology (BCSH). Recommendations are based on a review of the literature using Medline/Pubmed searches under the heading, CLL, up to October 2003 and data presented at the American Society of Hematology in 2003 and at the 10th International Workshop on CLL in The results of meta-analyses and phase 3 studies that have been published or presented in abstract form are included. Treatment recommendations were influenced by current and proposed clinical trials in the UK and by guidance from The National Institute for Clinical Excellence (NICE). A draft guideline was reviewed by members of the UK CLL Forum, patient representatives, members of the BCSH and a panel of approximately 60 UK haematologists. Their comments were incorporated where appropriate. To ensure widespread dissemination, the guideline is available on the BCSH website. Criteria for levels of evidence and grades of recommendation are shown in Table I. The guideline will be reviewed and updated in 2005, and a full guideline revision is planned for Epidemiology Chronic lymphocytic leukaemia is the most common type of leukaemia in the western world, accounting for 40% of all leukaemias in individuals over the age of 65 years. The median age of presentation is between 65 and 70 years. CLL is extremely rare below the age of 30 years but 20 30% of patients present under the age of 55 years. The overall incidence is approximately 3 per per year. Studies on the racial and geographic distribution show that CLL is times commoner in Europe, Australasia and North American white and black populations than in India, China Correspondence: BCSH Secretary, British Society for Haematology, 100 White Lion Street, London N1 9PF, UK. janice@b-s-h.org.uk and Japan. The male/female ratio in all populations is approximately 2:1 (Sgambati et al, 2001). There is no good evidence that exposure to chemicals or radiation, diet, cigarette smoking, viral infections or autoimmune disease are risk factors for the development of CLL. However, there is an increase in both lymphoid malignancies, including CLL, and a subclinical monoclonal B-cell expansion in first and second degree relatives of patients with CLL (Houlston et al, 2002; Rawstron et al, 2002). The phenomenon of anticipation in which the disease presents earlier and in a more severe form in successive generations is seen in many families with CLL (Yuille et al, 1998). The incidence of second malignancies is increased both in treated and untreated CLL. Diagnosis and prognostic factors Diagnostic investigations Patients may present with lymphadenopathy, systemic symptoms such as tiredness, night sweats and weight loss or the symptoms of anaemia or infection. However, 70 80% of patients are now diagnosed as an incidental finding on a routine full blood count. The initial clinical evaluation should seek to elicit a family history of lymphoid malignancy, susceptibility to infection, significant co-morbid conditions, and the presence of peripheral lymphadenopathy, hepatosplenomegaly and bulky intra-abdominal lymphadenopathy. A definitive diagnosis of CLL is based on the combination of a lymphocytosis and characteristic lymphocyte morphology and immunophenotype. Blood count. Current criteria for the diagnosis of CLL require a lymphocytosis of > /l. Patients whose routine blood count shows a lower level of lymphocytosis may subsequently develop clinically significant CLL. Options for adult patients with a lymphocytosis of between 3 and /l and lymphocyte morphology consistent with CLL include immunophenotyping or a follow-up blood count. However, there is no evidence that early diagnosis of asymptomatic patients with minimal lymphocytosis confers clinical benefit. doi: /j x ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

2 Table I. Criteria for (A) levels of evidence and (B) grades of recommendation. (A) Levels of evidence Table II. Scoring system for the diagnosis of chronic lymphocytic leukaemia (CLL). Score points Level Type of evidence Marker 1 0 Ia Ib IIa IIb III IV Evidence obtained from meta-analysis of randomized controlled trials Evidence obtained from at least one randomized controlled trial Evidence obtained from at least one well designed controlled study without randomization Evidence obtained from at least one other type of well designed quasi-experimental study Evidence obtained from well-designed non-experimental descriptive studies, such as comparative studies, correlation studies and case control studies Evidence obtained from expert committee reports or opinions and/or clinical experiences of respected authorities Smlg Weak Strong CD5 Positive Negative CD23 Positive Negative FMC7 Negative Positive CD22 or CD79b Weak Strong Scores in CLL are usually >3, in other B-cell malignancies the scores are usually <3. scoring system, 92% of CLL cases score 4 or 5, 6% score 3 and 2% score 1 or 2. Most other chronic B-cell lymphomas and leukaemias score 1 or 2, but a minority score 3. (B) Grades of recommendation Grade Evidence level Recommendation A Ia, Ib Required at least one randomized controlled trial as part of the body of literature of overall good quality and consistency addressing specific recommendation B IIa, IIb, III Required availability of well-conducted clinical studies but no randomized clinical trials on the topic of recommendation C IV Required evidence obtained from expert committee reports or opinions and/or clinical experiences of respects authorities Indicates absence of directly applicable clinical studies of good quality Lymphocyte morphology. Two subgroups of CLL can be distinguished using morphological criteria (Bennett et al, 1989). In typical CLL >90% of cells are small or medium sized lymphocytes with clumped chromatin, indistinct or absent nucleoli and scanty cytoplasm. In 15% of patients, the morphology is atypical; due to the presence of >10% prolymphocytes (CLL/PL) or >15% of cells showing lymphoplasmacytoid differentiation and/or cleaved nuclei. Immunophenotyping. Immunophenotyping should be performed in all cases requiring treatment and is of particular value in the following situations: (i) in cases with low lymphocyte counts to confirm the diagnosis of CLL and exclude reactive lymphocytosis; and (ii) in patients with atypical lymphocyte morphology to exclude other B- or T-cell lymphoproliferative disorders. Typically, CLL cells express weak monotypic surface immunoglobulin, CD5 CD19, CD23 and weak or absent CD79B, CD22 and FMC7. A recommended panel of monoclonal antibodies and scoring system for the diagnosis of CLL is shown in Table II (Moreau et al, 1997). Using this Additional investigations Additional investigations that may be helpful either at presentation and/or during the course of the disease include: direct antiglobulin test (DAT) (essential in all anaemic patients and before starting treatment); reticulocyte count; renal and liver biochemistry (including urate levels); serum immunoglobulins; chest X-ray; bone marrow aspirate and trephine biopsy. Although marrow examination is not usually essential for the diagnosis of CLL, the presence of proliferation centres and absence of paratrabecular foci and cyclin D1 nuclear staining support a diagnosis of CLL in cases with atypical morphology and a low immunophenotype score. Marrow examination is also valuable for determining the cause of cytopenias, providing prognostic information and assessing the response to therapy. Lymph node biopsy. Lymph node histology is not required for the diagnosis of typical CLL but may be indicated where the diagnosis is uncertain, in patients who develop bulky lymphadenopathy (particularly if localized to one lymph node area) and to exclude transformation to lymphoma or an unrelated cause of lymphadenopathy. Cytogenetic/fluorescence in situ hybridization (FISH) analysis. As with bone marrow and lymph node biopsy, genetic studies are not essential for the diagnosis of typical CLL. However, they may be helpful when there is diagnostic uncertainty. It is particularly important to exclude a t(11;14) translocation in CD5 positive cases with a low immunophenotype score. Computed tomography scan and/or ultrasound. These investigations may be helpful when the presence of splenomegaly is uncertain on physical examination, where ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

3 the finding of intrathoracic or bulky intra-abdominal disease would influence the need for, or choice of, therapy and to determine remission status following treatment in patients with bulky nodes prior to therapy. Prognostic factors The variable clinical course in CLL is a consequence of the frequency with which the disease is diagnosed in a preclinical phase, differences in the rate of disease progression between cases and the varying response to treatment. The median survival is approximately 10 years but this figure is of little value to an individual patient due to the extraordinary heterogeneity in the natural history of this disorder. Two studies have shown that the median survival of CLL is independent of whether patients present above or below years (Montserrat et al, 1991; Moreau et al, 1997; Mauro et al, 1999). However, younger patients are more likely to die of CLL-related causes while older patients more commonly die of unrelated causes including second primary malignancies. Data from the Medical Research Council (MRC) CLL trials have consistently shown that response rates to treatment and survival is better in women than in men (Catovsky et al, 1989). Established prognostic factors and more recent tests, which appear to provide additional prognostic information, are shown in Table III and include the following. Clinical stage. The clinical staging systems devised by Binet et al (1981) and Rai et al (1975) are the simplest and bestvalidated means of assessing prognosis (Table IV). It is important to exclude haemolysis and unrelated causes of anaemia or thrombocytopenia before assigning a patient to Binet stage C or the Rai high-risk group. Limitations of both systems include the choice of a single (but different) haemoglobin level to define marrow failure regardless of the sex of the patient, and the inability to predict the rate of disease progression in patients presenting with a low tumour burden. A subgroup of patients with Binet stage A disease Table III. Prognostic factors in chronic lymphocytic leukaemia. Factor Low risk High risk Gender Female Male Clinical stage Binet A Binet B or C Rai OI Rai II, III, IV Lymphocyte morphology Typical Atypical Pattern of marrow Non-diffuse Diffuse trephine infiltration Lymphocyte doubling time >12 months <12 months Serum markers* Normal Raised CD38 expression <20 30% >20 30% Genetic abnormalities None del 11q23 del 13q (sole) Loss/mutation of p53 IgVH gene status Mutated Unmutated *See text for details. who have smouldering CLL can be identified based on the following parameters (Monteserrat et al, 1988; French Co-operative Group on Chronic Lymphocytic Leukaemia, 1990a): haemoglobin >13 g/dl; lymphocyte count < /l; minimal, or no lymphadenopathy; non-diffuse pattern of bone marrow involvement; a lymphocyte doubling time of >12 months. In one study, only 15% of these patients showed disease progression after 5 years and 80% were alive at 10 years (Monteserrat et al, 1988). Patients entered into the MRC 3 trial with progressive stage A disease have a similar life expectancy to those presenting with stage B disease. Serum markers. These include b2 microglobulin, lactate dehydrogenase (LDH), serum thymidine kinase and soluble CD23 (Knauf et al, 1997; Hallek et al, 1999; Di Raimondo et al, 2001; Schwarzmeier et al, 2002). All have been shown to predict progression and survival in Binet stage A patients, but their value is currently limited either by the lack of a standard assay method, variable cut-off points between series or the lack of validation in a prospective study. CD38 expression. Many studies have shown that a high level and/or intensity of CD38 expression on leukaemic lymphocytes is a poor prognostic factor both in univariate analysis and in patients of known clinical stage and b2 microglobulin levels (Damle et al, 1999; Del Poeta et al, 2001; Ibrahim et al, 2001; Durig et al, 2002; Hamblin et al, 2002; Ghia et al, 2003). The optimum cut-off level with greatest prognostic significance is uncertain. Different studies have chosen values of 7%, 20% or Table IV. Staging systems in chronic lymphocytic leukaemia. Features Binet stage A <3 lymphoid areas* 60 B >3 lymphoid areas 30 C Haemoglobin < 10Æ0 g/dl or 10 platelets < /l Rai stage 0 Lymphocytosis only 30 I Lymphadenopathy 25 IIà Hepato or splenomegaly ± 25 lymphadenopathy III Haemoglobin < 11Æ0 g/dl 10 IV Platelet < /l 10 % of patients *The five lymphoid areas comprise unilateral or bilateral cervical, axillary and inguinal lymphadenopathy, hepatomegaly and splenomegaly. Risk group at low level. àrisk group at intermediate level. Risk group at high level. 296 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

4 30%. CD38 expression may vary during the course of the disease (Hamblin et al, 2002) and although there is a correlation between high CD38 expression and unmutated IgVH genes, CD38 is not a surrogate marker for VH gene status. IgVH gene status. There is a highly significant difference in the survival between patients with or without mutated IgVH genes (Damle et al, 1999). In one study, patients with mutated IgVH genes had a median survival of 25 years compared with 8 years for patients with unmutated IgVH genes (Hamblin et al, 1999). However, there is still controversy as to the percentage of mutations which best correlates with clinical outcome (between 98% and 95% homology to the germline gene) (Lin et al, 2002). Recent data suggest that the use of particular IgVH gene segments such as the VH 3.21 gene may confer a poor prognosis regardless of mutational status (Tobin et al, 2002). Preliminary data suggest that expression of ZAP70 mrna and ZAP70 protein, measured using flow cytometry, correlates closely with IgVH gene mutation status (Crespo et al, 2003; Wiestner et al, 2003; Orchard et al, 2004). Cytogenetic abnormalities. Cytogenetic analysis has shown correlations between chromosome abnormalities and clinical features in CLL. Patients with a normal karyotype or an isolated deletion of chromosome 13q have a better prognosis than those with trisomy 12 as the sole abnormality or with a complex karyotype (Juliusson et al, 1990). Subsequently, both chromosome 11q deletions and structural abnormalities of chromosome 17p were shown to be associated with short survival (Dohner et al, 1995, 1997). In a univariate analysis, using a panel of FISH probes, patients with an isolated deletion of 13q, trisomy 12, deletion of 11q, or of loss of one copy of the p53 gene on 17p had a median survival of 133, 114, 79 and 32 months respectively (Dohner et al, 2000). Drug sensitivity testing. Non-randomized studies have suggested that pretreatment drug sensitivity testing using the apoptosis by morphology using octospot [AMO; previously the differential staining cytotoxicity (DiSC)] assay can identify drug sensitivity and resistance in individual cases (Bosanquet et al, 1999). The value of the assay in guiding second line therapy is being evaluated in the MRC/Leukaemia Research Fund (LRF) CLL4 trial. Although the finding of good risk prognostic factors can be reassuring to asymptomatic stage A patients, evidence that the application of the prognostic factors discussed above improves clinical outcome is currently lacking. This is being addressed in ongoing trials such as the MRC CLL4 study and the German CLL trials. In the interim, the choice of prognostic markers should take account of the following considerations: 1 Ideally prognostic markers should be inexpensive, widely available, quality controlled, validated in phase 3 clinical trials and shown to provide additional information to markers in current use. 2 The value of prognostic factors differs for different treatments, e.g. p53 abnormalities predict a poor response to alkylating agents, purine analogues (Dohner et al, 1995) 3 and rituximab monotherapy (Byrd et al, 2003a) but not to high-dose steroids or alemtuzumab (Campath 1H). 3 Some prognostic markers, such as VH gene status, remain constant throughout the disease while others, such as cytogenetic abnormalities and high CD38 expression, may be acquired during the disease. Both the timing and frequency of testing for prognostic factors are therefore important. 4 Many prognostic markers are closely linked and their value can only be assessed in multivariate analyses. Two recent multivariate analyses, which included assessment of cytogenetic and genetic abnormalities, CD38 expression and VH gene status, both showed that only VH gene status and loss or mutation of the p53 gene had independent prognostic significance using a cut-off of 98% homology to the germline sequence to define IgVH gene mutational status. Deletion of chromosome 11q was also significant in one study when a cut-off of 97% VH gene homology was chosen (Krober et al, 2002; Oscier et al, 2002). Management of CLL Indications for referral The management of CLL requires a collaborative approach between primary care and a haematology department. Input from a palliative care team may sometimes be valuable in the management of terminal drug resistant patients. Indications for referral to a haematology department include: symptomatic disease, the presence of lymphadenopathy or hepatosplenomegaly and the investigation of a lymphocytosis, particularly if the lymphocyte count is high or there is anaemia or thrombocytopenia. The subsequent management of these patients should be discussed at a multidisciplinary team meeting. The follow-up of patients seen initially in hospital, who do not require treatment, may be organized in primary care, hospital outpatients or via a homecare service depending on local resources and patient wishes. Asymptomatic elderly patients, with a slight lymphocytosis only, may be managed by primary care teams, providing indications for referral to a haematology department are clearly documented. Communicating with patients Chronic lymphocytic leukaemia is characteristically a condition for which the natural history is measured in years and it is therefore particularly important that patients are able to establish a relationship and trust with the clinician managing their condition. Patients who are referred for a haematological opinion and subsequent management decision will expect and are entitled to be honestly informed about their disease. ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

5 A frequent dilemma is whether to convey the diagnosis of CLL to an elderly asymptomatic patient with low count stage A disease diagnosed on a routine blood count. Anxiety generated by the use of the word leukaemia can almost always be prevented by a clear and careful explanation of the benign nature of the condition. If a decision not to inform a patient is taken this must be very clearly documented to ensure that other health care workers do not subsequently impart the diagnosis without explaining its significance. The majority of patients benefit from, and should be offered, information about CLL in general and about their specific management. The latter might include details of the topography of the haematology unit, the staff involved in their care, who to contact if problems arise, details of local support groups and whether their treatment may be influenced by NICE guidelines or budgetary constraints. General information may be obtained from a wide variety of sources, as follows: Books and pamphlets. CANCERBACUP publishes a wide range of pamphlets, including one on understanding CLL and others on living with cancer. For further information go to The LRF also produces a range of booklets and pamphlets. These tend to be rather more technical than those published by CANCERBACUP. A list of their publications is found at dspace.dial.pipex.com/lrf-//publications/index.htm. There are three books from the USA that offer comprehensive guides for patients. Non-Hodgkin s Lymphoma by Lorraine Johnson, published by O Reilly. Adult Leukemia by Barbara Lackritz, published by O Reilly. Bone Marrow and Blood Stem Cells Transplants: A Guide for Patients by Susan Stewart, published by Bone Marrow information, Video and audio. CANCERBACUP has several UK-based videos and audio programmes. It also produces a CD-ROM that contains a comprehensive guide to all aspects of cancer. The remaining sources are from the USA. There are several organizations that produce video or audio programmes for patients. They are kept current and usually involve well-known doctors. They are a source of reliable information. Healthology at Healthtalk International CLL education network at Lymphomafocus at General internet sites. There are numerous sites. It is suggested that patients should be directed to a few that are known to be reliable. These sites do contain links to other sources of information for patients who wish to extend their knowledge. GrannyBarb s Leukemia Links, CLL at leukemia/cll.html. An early site developed by Barbara Lackritz. CLL FAQ s, Glossary of terms and an ABC of acronyms at Summary of CLL research information at acor.org/leukemia/medical_news.htm. American Cancer Society at docroot/lrn/lrn_0.asp. UK Cancer resources at clinks44.htm. National Cancer Institute (NCI) site at The patients guide to CLL from the NCI at &cdrid¼ On-line discussion groups. Some patients may wish to join a group of others with the same condition where they can share their experience. There is a large international group for CLL, for more information on this and other groups that may be of interest the contact is Patients may also join the UK CLL Forum, which publishes a newsletter. Indications for treatment The indications for treatment recommended by the NCI sponsored working group are shown in Table V (Cheson et al, 1996). These criteria will encompass the majority of patients with Binet stage B or C disease and a proportion of patients with Binet stage A disease, showing features of disease progression. Stage A patients who develop autoimmune haemolytic anaemia (AIHA) or immune thrombocytopenic purpura (ITP) should be treated for their autoimmune cytopenia but may not require anti-leukaemic therapy. The minority of patients who present with either Binet stage B disease, with generalized non-bulky lymphadenopathy and/or minimal hepatosplenomegaly, or with Table V. Indications for treatment. Progressive marrow failure: the development or worsening of anaemia and/or thrombocytopenia Massive (>10 cm) or progressive lymphadenopathy Massive (>6 cm) or progressive splenomegaly Progressive lymphocytosis >50% increase over 2 months Lymphocyte doubling time <6 months Systemic symptoms* Weight loss >10% in previous 6 months Fever >38 C for 2 weeks Extreme fatigue Night sweats Autoimmune cytopenias *It is important to exclude other causes for these symptoms, such as infection. 298 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

6 Table VI. Response criteria. Criteria Complete response Partial response Progressive disease Symptoms None Lymph nodes None >50% decrease >50% increase or new nodes Liver/spleen Not palpable >50% decrease >50% increase or new nodes Haemoglobin >11Æ0 g/dl >11Æ0 g/dl or >50% improvement from baseline (untransfused) Neutrophils >1Æ /l >1Æ /l or >50% improvement from baseline Lymphocytes <4Æ /l >50% decrease >50% increase Platelets > /l > /l or >50% improvement from baseline Marrow aspirate <30% lymphocytes Marrow trephine No interstitial or nodular infiltrate May be residual lymphoid nodules stage C disease and mild cytopenias and who are asymptomatic, may be observed without treatment until there is evidence of symptomatic or progressive disease. Hyperviscosity due to extreme lymphocytosis is very rare in CLL and a high lymphocyte count in the absence of a rapid lymphocyte doubling time is not an indication for treatment. Neither asymptomatic hypogammaglobulinaemia nor the presence of a paraprotein are reasons for treatment. Assessment of response The criteria for assessing complete response (CR) or partial response (PR) to treatment recommended by the NCI working group are shown in Table VI (Cheson et al, 1996). Patients who fulfil the criteria for a CR but whose bone marrow trephines contain lymphoid nodules have been designated as having a nodular PR. Since the publication of the NCI criteria, new methods have become available for assessing minimal residual disease (MRD). A variety of techniques are available to detect MRD. A flow cytometric assay that differentiates CLL cells from normal B cells based on expression of CD19/CD5/CD20 and CD79b is rapid, applicable to all patients and can detect one CLL cell in 10 5 normal cells when cells are analysed (Rawstron et al, 2001). Comparable sensitivity may be achieved using realtime quantifiable allele-specific oligonucleotide polymerase chain reaction (PCR) (Pfitzner et al, 2002). However, this technique is labour intensive and expensive. Newer therapeutic approaches can result in MRD negative remissions. The presence of detectable MRD in patients who achieve a complete remission by NCI criteria following treatment with purine analogues, monoclonal antibodies or autologous transplantation predicts for clinical relapse. However, patients who initially remain MRD positive after allogeneic transplantation may subsequently become MRD negative (Esteve et al, 2002). A treatment strategy for CLL Before initiating treatment, consideration must be given to (i) patient-related factors, such as age, performance status, co-morbid conditions and patient wishes; (ii) disease-related factors, such as the severity of symptoms and the presence of adverse prognostic factors; and (iii) treatment-related factors including the degree and duration of response to prior treatments and contra-indications to, and side-effects from, particular treatment modalities. Pharmacoeconomic considerations (Table VII) are also important. Chronic lymphocytic leukaemia presents significant management problems by virtue of the heterogeneity in both the age of presentation, in the natural history of the disease and also the frequency with which CLL is diagnosed in a preclinical phase. The median survival of patients with advanced CLL is usually superior to that seen in many other haematological malignancies and solid tumours. In the absence of a curative treatment, most patients receive a number of different treatment modalities during the course of the disease. The response to a particular treatment varies according to its place in the overall treatment strategy, such that treatments which produce high response rates when used as initial therapy may also make a substantial contribution to prolonging survival when given later in the disease, particularly if they are administered after less active agents. Table VII. Cost of chemotherapy in chronic lymphocytic leukaemia. Drug Route of administration Approximate cost ( ) Chlorambucil p.o. 730 Fludarabine p.o i.v Fludarabine and p.o cyclophosphamide i.v Rituximab Fludarabine p.o./i.v. Cyclophosphamide Alemtuzumab (12 weeks) i.v High-dose methylprednisolone i.v. 730 The costs are those for the chemotherapy alone, based on 6 months treatment (apart from alemtuzumab) using standard regimens for a patient with a surface area of 2 m 2. The figures do not represent the total cost of administering these treatments. ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

7 The latter phenomenon is seen in trials in which the design permits crossover between two treatment arms for patients who have not responded to their initial therapy. Although it is relatively easy to measure the rates of response to any given therapy, proving that this ultimately translates into a survival advantage is more difficult. Current strategies for the management of CLL, particularly in those patients with good performance status, mirror those adopted in other haematological malignancies and seek to achieve MRD negative responses. An important consideration on beginning treatment in CLL is whether to adopt a palliative approach and treat symptomatic disease with regimens causing minimal treatment-related toxicity, or to aim for prolonged disease-free survival in the hope that this will translate into superior overall survival. Initial treatment Treatment of early stage CLL. In 1998, the French Co-operative Group on CLL reported the outcome of two trials comparing initial or deferred treatment until disease progression in Binet stage A CLL (Dighiero et al, 1998). One trial included 609 patients, randomized to receive either no treatment or chlorambucil 0Æ1 mg/kg/d until drug resistance. The second trial randomized 926 patients to no treatment or to chlorambucil 0Æ3 mg/kg and prednisolone 40 mg/m 2 daily for 5 d per month for 3 years. Early treatment with chlorambucil slowed the rate of disease progression but there was no difference in overall survival in either trial between early and delayed treatment. A subsequent meta analysis of 2048 patients in six trials of immediate treatment with chlorambucil plus or minus prednisolone vs. deferred treatment showed no significant difference in 10 years survival (CLL Trialists Collaborative Group, 1999). In the untreated arm of the first French Co-operative Group Trial 51% of patients showed disease progression and 27% of stage A patients died of disease-related causes. There is current interest in conducting clinical trials to evaluate whether asymptomatic stage A patients with poor risk disease might benefit from newer and more effective treatments than chlorambucil. The choice of prognostic markers and treatment options is currently under discussion. Treatment of early stage disease with chlorambucil is not indicated (grade A recommendation, level Ia evidence). Treatment of advanced or progressive disease There are no randomized studies comparing treatment versus no treatment in patients with advanced disease stage. Evidence indicating the need for treatment comes indirectly from the obvious symptomatic and clinical improvement, as well as the survival advantage, seen in those patients who respond to therapy compared with non-responders (Robak & Kasznicki, 2002). Alkylating agents. Early studies used low dose chlorambucil with or without prednisolone/prednisone. The combined CR and PR rates ranged from 45 86%. In no study was there any advantage in terms of progression-free interval or overall survival with the addition of prednisolone/prednisone to chlorambucil (Han et al, 1973). Similarly, there was no difference between continuous and intermittent chlorambucil therapy (Sawitsky et al, 1979). Patients who are intolerant of chlorambucil may respond to low dose daily cyclophosphamide. Four randomized studies have compared chlorambucil with COP (cyclophosphamide, vincristine, prednisolone) in 630 patients (Montserrat et al, 1985; French Co-operative Group 5 on Chronic Lymphocytic Leukaemia, 1990b; Catovsky et al, 1991; Raphael et al, 1991). COP resulted in more rapid responses and a higher response rate, but there was no difference in progression-free interval and overall survival. In 1986, the French Co-operative Study Group reported a significant survival advantage for patients with previously untreated advanced disease stage using a modified CHOP regimen (cyclophosphamide, adriamycin, vincristine, prednisolone) utilizing a lower dose of adriamycin than the CHOP regimen introduced for the treatment of lymphoma, compared with COP (French Co-operative Group on Chronic Lymphocytic Leukaemia, 1990b). In this study, however, the COP arm fared worse than in previously reported studies. A metaanalysis of 2022 patients in 10 trials, comparing combination therapy with chlorambucil with or without prednisolone, showed an identical 5-year survival of 48% in both groups (CLL Trialists Collaborative Group, 1999). Six of the 10 studies included an anthracycline and a subgroup analysis of these trials also showed no survival advantage compared with chlorambucil. A number of studies have evaluated the role of higher dose chlorambucil in CLL. A total of 279 patients were randomized to either high-dose chlorambucil (15 mg daily) or intermittent chlorambucil (75 mg every 4 weeks) with prednisolone. Patients receiving continuous high-dose treatment achieved a higher remission rate (70% vs. 30%) and longer median survival (6 years vs. 3 years, P < 0Æ01) (Jaksic & Brugiatelli, 1988). A subsequent study randomized 228 previously untreated patients to either high-dose chlorambucil at a fixed dose of 15 mg/d until either a CR, grade 3 toxicity or to a maximum of 6 months (Jaksic et al, 1997). This induction phase was then followed by maintenance therapy with chlorambucil given at a dose of 5 15 mg twice weekly according to haematological tolerability. The other arm of the study comprised the CHOP regimen using the doses proposed by the French Co-operative Group on CLL. Six cycles of induction treatment were given followed by maintenance therapy with six additional courses given at three monthly intervals. High-dose chlorambucil resulted in a higher overall response rate (ORR) (89% vs. 75%) and prolonged median survival (68 months vs. 47 months) after a median follow-up of 37 months. Studies on high-dose 300 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

8 chlorambucil were excluded from the meta-analysis of randomized trials described above. They are difficult to evaluate in comparison with other phase 3 trials of chlorambucil, as they utilize non-standard response criteria: namely a total tumour mass score (Jaksic & Vitale, 1981) and a diagnosis of marrow failure based on a haemoglobin of <10Æ5 g/dl for men or <9Æ5 g/ dl for women and/or a platelet count of < /l. There is no requirement for bone marrow examination post-therapy. Purine analogues. Fludarabine. Fludarabine produces ORR of 70 80% with a CR rate of approximately 30% and a median survival of 73 months when given as a single agent at a dose of 25 mg/m 2 intravenously for 5 d each month. Factors predicting prolonged survival included age <70 years, intermediate rather than high-risk disease and response to therapy, particularly the attainment of a molecular complete remission (Keating et al, 1998). The addition of prednisolone has no effect on overall survival (O Brien et al, 1993). Oral fludarabine at a dose of 40 mg/m 2 /d for 5 d achieves a comparable remission rate with intravenous administration (Boogaerts et al, 2001). Three studies (Table VIII) have compared fludarabine with regimens including an alkylating agent with or without an anthracycline (Johnson et al, 1996; Rai et al, 2000; Leporrier et al, 2001). In each case fludarabine produced a higher ORR, CR rate and longer duration of response than CAP (cyclophosphamide, adriamycin, prednisolone) or chlorambucil. However, there was no statistically significant difference in the survival for patients receiving fludarabine. In the above studies of Rai et al (2000) and Leporrier et al (2001), this was related to the crossover nature of the studies and the high response rate to second line treatment with fludarabine in patients who had failed CHOP, CAP or chlorambucil. The risk of severe infection was lower with chlorambucil than with fludarabine, or in combination chemotherapy regimens including an anthracycline, while the latter were associated with a higher incidence of nausea, vomiting and alopecia. In the USA intergroup study the incidence of therapyrelated myeloid leukaemia was 3Æ5% in patients receiving fludarabine and chlorambucil, 0Æ5% in patients given fludarabine alone and 0% in patients treated with chlorambucil alone (Morrison et al, 2002). An European Organisation for Research and Treatment of Cancer study comparing fludarabine with continuous highdose chlorambucil has reported similar response rates and comparable response duration and survivals (Zittoun et al, 1999). An interim analysis of a phase 3 German CLL study group trial, comparing chlorambucil with fludarabine in previously untreated patients over the age of 65 years with advanced CLL, shows a higher overall and CR rate in the fludarabine arm, but no difference in progression-free survival with a median follow-up of 11 months. Myelotoxicity was greater in patients receiving fludarabine, but there was no difference in the incidence of severe infections (Eichhorst et al, 2003a). An interim analysis of the CLL4 study of the German CLL Study Group, which randomized 375 patients aged <65 years with progressive Binet stage A or stage B/C disease to fludarabine alone or in combination with cyclophosphamide, has shown that the combined treatment results in a higher ORR (94% vs. 85%), a higher CR rate (20% vs. 9%), prolonged progression-free survival (28 months vs. 23 months) and a lower incidence of severe anaemia, but more severe neutropenia and thrombocytopenia. There was no difference in the incidence of severe infections (Eichhorst et al, 2003b). Cladribine. The use of cladribine in previously untreated patients results in ORR of approximately 75% with 38% CR (Saven et al, 1995; Juliusson et al, 1996). The duration of response is about 2 years. A phase 3 randomized study comparing cladribine + prednisone with chlorambucil (12 mg/m 2 for 7 d) + prednisolone, resulted in a higher ORR, CR rate and progression-free survival for patients receiving cladribine but no advantage in overall survival Table VIII. Comparison of fludarabine with alkylating agents (alone or in combination therapy) in previously untreated chronic lymphocytic leukaemia. Study Regimen No. of patients CR (%) PR (%) Duration (months) Median survival (months) Johnson et al (1996) Fludarabine Not reached Not reached vs Æ9 52Æ6 CAP Rai et al (2000) Fludarabine vs Chlorambucil Leporrier et al (2001) Fludarabine Æ1 31Æ0 31Æ7 69 vs Æ2 43Æ0 27Æ7 70 CAP vs. CHOP Æ6 41Æ9 29Æ5 67 CHOP, cyclophosphamide, adriamycin, vincristine, prednisolone; CAP, cyclophosphamide, adriamycin, prednisolone; CR, complete remission; PR, partial remission. ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

9 6 (Robak et al, 2000a). Sixty-seven per cent of patients resistant to chlorambucil respond to cladribine, but only 27% of patients who failed cladribine benefited from second line treatment with chlorambucil. Although these results were broadly similar to those achieved with fludarabine, the extent of the evidence is considerably less and it is currently not possible to recommend cladribine as a routine alternative to fludarabine for the initial therapy of CLL. To prevent transfusion related graft versus host disease (GVHD), all patients treated with a purine analogue should receive irradiated blood products indefinitely thereafter (BCSH, 1996). Steroids. There is no evidence that prolonged treatment with low, intermediate or high-dose steroids is an effective initial treatment for CLL. However, it is recommended that patients with stage C disease should be given a short course of prednisolone before receiving chlorambucil (grade C recommendation, level IV evidence). Monoclonal antibodies. Alemtuzumab, given intravenously to nine previously untreated patients resulted in three CR and five PR (Osterborg et al, 1997). A subsequent phase 2 trial, in which alemtuzumab was given subcutaneously at a dose of 30 mg three times per week for up to 18 weeks to 41 patients, produced an ORR of 81% with 19% CR (Table IX). The median time to treatment failure has not been reached (>18 months) (Lundin et al, 2002). Alemtuzumab therapy results in high ORR in untreated CLL but follow-up is short and benefit over conventional therapy has not been demonstrated. Ten per cent of patients develop cytomegalovirus (CMV) reactivitation. Use of irradiated blood products is recommended following alemtuzumab. Single agent rituximab therapy induces short-term PRs in previously untreated patients. The use of subsequent maintenance treatment with rituximab in patients responding to initial treatment with the same agent is being evaluated (Hainsworth et al, 2003). A randomized phase 2 study, comparing fludarabine given with either concurrent or sequential rituximab, showed a higher overall and CR rate for the concurrent regime but the median response duration and survival have not been reached for either arm after a median follow-up of 23 months (Byrd et al, 2003b). Response rates to rituximab in combination with fludarabine and cyclophosphamide are better than historical controls using any previously reported regimen (Table X). In a study of 135 patients, complete remission was demonstrated in 67%, and 57% were in molecular remission using a PCR technique. Thirteen of 41 evaluable PCR negative patients became PCR positive, usually within 6 months of follow-up but longer follow-up is required to assess the clinical benefit of attaining a PCR negative response (Keating et al, 2002a). Summary of recommendations for initial treatment For the majority of patients who are ineligible for a transplant procedure and in whom there is no contraindication to fludarabine (severe renal impairment or an autoimmune cytopenia), entry into the MRC CLL4 study should be offered. This trial randomizes patients to either chlorambucil, fludarabine, or fludarabine and cyclophosphamide and assesses the value of prognostic factors and quality of life issues as well as outcome. Both fludarabine and chlorambucil are options for patients who do not wish to enter the study. Patients in whom fludarabine is contraindicated and for whom a palliative approach has been adopted should be treated with chlorambucil (grade A recommendation, level Ia evidence). There is no survival advantage for including an anthracycline with chlorambucil in the initial treatment of advanced CLL (grade A recommendation, level Ia evidence). Table IX. Use of alemtuzumab alone or in combination in untreated and previously treated chronic lymphocytic leukaemia. No. of patients Study Treatment Regimen Untreated Treated OR (%) CR (%) Osterborg et al (1996) Alemtuzumab 30 mg, 3 weekly s.c. or i.v. for up to 18 weeks Lundin et al (2002) Alemtuzumab 30 mg, 3 weekly s.c. for up to 18 weeks Osterborg et al (1997) Alemtuzumab 30 mg, 3 weekly i.v. for up to 12 weeks Kennedy et al (2001) Alemtuzumab 30 mg, 3 weekly i.v Rai et al (2001) Alemtuzumab 30 mg, 3 weekly i.v. for up to 12 weeks Keating et al (2002c) Alemtuzumab 30 mg, 3 weekly i.v. to maximum response Kennedy et al (2002) Alemtuzumab 30 mg, 3 weekly i.v Fludarabine 25 mg/m 2, 3 d monthly till maximum response Nabhan et al (2001) Alemtuzumab 3 30 mg, 3 weekly Rituximab 375 mg/m 2,4 weekly for 5 weeks Faderl et al (2001) Alemtuzumab 3 30 mg, 3 weekly Rituximab 375 mg/m 2 weekly OR, overall response; CR, complete remission. 302 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

10 Table X. Use of Rituximab combined with other agents in untreated and previously treated chronic lymphocytic leukaemia. No. of patients Study Treatment Regimen Untreated Treated OR (%) CR (%) Byrd et al (2003b) Rituximab (i) Concurrent Fludarabine (ii) Sequential Schulz et al (2002) Rituximab 375 mg/m 2 * Fludarabine 25 mg/m 2 * Keating et al (2002a) Rituximab 500 mg/m Fludarabine 25 mg/m 2 Garcia-Manero et al (2001) Rituximab 500 mg/m Fludarabine 25 mg/m 2 Cyclophosphamide 250 mg/m 2 Gupta et al (2001) Rituximab Four weekly until maximum response Cyclophosphamide Dexamethasone OR, overall response; CR, complete remission. *For four cycles. For six cycles. Further studies using standard response criteria are required before high-dose chlorambucil can be recommended as an initial treatment for CLL (grade C recommendation, level IV evidence). Where a patient is considered suitable for entry into the MRC CLL5 trial or for allogeneic transplantation, then an initial treatment, such as fludarabine or fludarabine and cyclophosphamide, which is likely to result in a complete or very good partial remission, should be chosen (grade C recommendation, level IV evidence). Alemtuzumab is not recommended for untreated CLL (grade B recommendation, level IIb evidence). Rituximab monotherapy is not recommended in untreated CLL (grade C recommendation, level III evidence). Rituximab combined with fludarabine (with or without cyclophosphamide) requires further evaluation in untreated CLL (grade B recommendation, level Ib evidence). Second line and subsequent treatment There has been only a single randomized study comparing treatments for patients with relapsed or refractory disease. Evidence of benefit is largely based on historical control data and the results of randomized trials of initial treatment in which patients who fail to respond to one arm of the study are crossed over to another arm. The inclusion of patients who may be either minimally or heavily pretreated, and who may have drug resistant or responsive disease, makes the results of many second line studies difficult to interpret. The indications for second line and subsequent treatments are symptomatic and/or progressive disease, as for initial therapy, although treatments with curative rather than palliative intent such as allogeneic transplantation, may be considered in early relapse, or in first remission. The response to second line treatments depends on a variety of factors including clinical stage, adverse biological prognostic factors, the number of prior therapies and critically, refractoriness to the last treatment. The second randomization in the CLL4 study is designed to evaluate the effectiveness of the AMO assay in predicting optimal therapy for patients relapsing after initial treatment. Therapeutic options are listed below and a therapeutic strategy is shown in Fig 1. Alkylating agents alone or in combination Patients who have responded to an alkylating agent such as chlorambucil can often be successfully retreated on one or more occasions. However, the quality and duration of response is usually inferior to that achieved with the initial treatment and multiple courses of treatment often result in the emergence of drug resistance (Galton et al, 1961; Ezdinli et al, 1969). The response rate to chlorambucil in patients relapsing after initial treatment with purine analogues is low. In the US intergroup study, only 7% of patients responded to chlorambucil after failing fludarabine therapy (Rai et al, 2000). In a randomized trial between cladrabine and prednisolone versus chlorambucil and prednisolone, 27% of patients who failed cladrabine subsequently responded to chlorambucil (Robak et al, 2000b). COP is not superior to chlorambucil and prednisolone in patients relapsing after initial therapy with chlorambucil (Montserrat et al, 1985). Anthracycline-containing combination therapy Although widely used, evidence for the value of anthracycline-containing regimens in previously treated patients is limited. A randomized phase 3 study comparing CAP with fludarabine, in patients who had received an alkylating agent ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

11 Fig 1. Treatment options in CLL. CLL4 trial and CLL5 autograft trial are current UK trials. Proposed UK CLL Forum trials are: (i) poor risk stage A CLL; (ii) Alemtuzumab ± fludarabine for patients resistant to fludarabine; and (iii) reduced intensity conditioning Allograft Study. Treatment strategy for patients who are refractory or become resistant to fludarabine or fludarabine + cyclophosphamide therapy is shown in shaded boxes. chlor ¼ chlorambucil; F ¼ fludarabine; FC ¼ fludarabine + cyclophosphamide; RIC ¼ reduced intensity conditioning; HDMP ¼ high-dose methyl prednisolone; CHOP ¼ cylophosphomide, adriamycin, vincristine, prednisolone; FCR ¼ fludarabine, cyclophosphamide and rituximab. for more than 6 months and less than 3 years, showed an ORR of 27% in the CAP arm compared with 48% in the fludarabine arm (P ¼ 0Æ036) (French Co-operative Group on CLL, 1996). A subsequent study conducted by the French collaborative group (Leporrier et al, 2001) randomized patients between fludarabine and either CAP or CHOP; 39% of patients failing fludarabine subsequently responded to CHOP. The above data suggest that anthracycline-containing regimens are less effective than purine analogues in patients previously treated with chlorambucil, but do have activity in patients relapsing after purine analogue therapy. Purine analogues Numerous phase 1 and phase 2 studies have evaluated the response of previously treated patients to fludarabine (Grever et al, 1988; Keating et al, 1988, 1989, 2000, 2002b; Sorensen et al, 1997). ORR range from 13% to 73% and CR rates from 0% to 37%. Response rates are highest in patients who have not been heavily pretreated, were not resistant to their last treatment, have a normal serum albumin and are <70 years old. Response rates to fludarabine in patients previously responsive to an alkylating agent range from 60% to 68%, whereas response rates in patients refractory to alkylating agents range from 20% to 49% (O Brien et al, 1993; Montserrat et al, 1996; Keating et al, 2000). Durable responses to fludarabine are seen in 40% of patients who failed to respond to CHOP as first line therapy (Leporrier et al, 2001). NICE guidance recommends fludarabine alone as second line therapy for patients who have failed, or are intolerant of first line therapy with an alkylating agent and who would otherwise receive CHOP, COP or CAP (NICE, 2001). Eighty-five per cent of patients previously responsive to fludarabine respond to retreatment with the same agent but 304 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

12 the response rate falls to 12% in patients who are refractory to previous fludarabine therapy. Cladrabine has been less extensively evaluated than fludarabine and there are differences in opinion as to its correct schedule of administration. However, response rates are broadly similar to that achieved with fludarabine (Piro et al, 1988; Saven et al, 1991; Juliusson & Liliemark, 1993, 1994, 1996). The results of a small case series which reported benefit from cladrabine therapy in three patients who were refractory to fludarabine (Juliusson et al, 1992) was not confirmed by the experience of other investigators. However, a recent study sponsored by the cancer and leukemia group B (Byrd et al, 2003c) achieved responses in nine of 28 patients (ORR 32%) for cladrabine in fludarabine refractory patients. Small phase 2 studies show that when purine analogues are administered in combination with other chemotherapeutic agents to previously treated patients, the response rates are higher than those achieved with purine analogues alone. Examples are shown in Table XI. High-dose methyl prednisolone (HDMP) Although widely used, there is little published data on the efficacy of HDMP in relapsed CLL. A study of HDMP given intravenously or orally at a dose of 1 gm/m 2 /d for 5 d each month was performed in 25 patients, 15 of whom were 7 refractory to fludarabine (Thornton et al, 2003). An ORR of 77% was achieved with a median duration of 12 months. Responses were seen in five of 10 patients with loss and/or mutation of the p53 gene. The event-free and overall survival was significantly better in responders than in non-responders. Patients who relapse after HDMP frequently respond to further courses of the same treatment. HDMP is contraindicated in patients with an active gastric or duodenal ulcer Table XI. Combination of a purine analogue with other chemotherapeutic agents in previously treated chronic lymphocytic leukaemia. Study No. of patients Regimen Rummel et al (1999) 25 Fludarabine + epirubicin O Brien et al (2001) 20 Fludarabine + cyclophosphamide Hallek et al (2001) 18 Fludarabine + cyclophosphamide Bosch et al (2002) 37 Fludarabine + cyclophosphamide + mitoxantrone Montillo et al (2003) 23 Pentostatin + cyclophosphamide Overall response rate (%) and should be used with caution in patients with diabetes or heart failure. Monoclonal antibodies Alemtuzumab. A total of 341 patients refractory to fludarabine have been treated with alemtuzumab in five non-randomized studies (Table IX). The ORR was 39% (9Æ4% CR and 30% PR) with a prolonged median survival compared with historical data on fludarabine-resistant patients. In the pivotal study of 92 patients with fludarabine-resistant disease, best responses were seen in patients with a low b2 microglobulin, platelet count > /l and <3 cm lymphadenopathy. No case with lymph nodes >5 cm achieved complete resolution of lymphadenopathy (Keating et al, 2002b). Alemtuzumab may be effective in some patients with p53 mutations refractory to purine analogues (Stilgenbauer & Dohner, 2002). Preliminary data suggest that the combination of fludarabine and alemtuzumab may be effective in patients resistant to both agents given alone (Kennedy et al, 2002). Alemtuzumab has been administered to patients with residual disease following treatment with fludarabine (O Brien et al, 2003), and prior to stem cell mobilization (Montillo et al, 2002). Although CRs and MRD-negative responses have been achieved, some patients have experienced prolonged myelosuppression following standard doses of alemtuzumab, and this sequential regimen should only be used in a clinical trial setting. Rituximab. Although rituximab has some efficacy in CLL, the response rates to rituximab monotherapy in previously treated patients are poor. Even at very high doses (up to six times the standard dose), all the responses are partial. Response rates to rituximab, in combination with fludarabine or with fludarabine and cyclophosphamide, given to patients with relapsed or refractory CLL are superior to those seen with standard second line therapies such as fludarabine or CHOP (Table X). The use of fludarabine, cyclophosphamide and rituximab has been reported in 167 patients with previously treated CLL, of whom 102 were evaluable with more than 6 months of follow-up. Fifteen per cent of patients had received alkylating agents only, 59% had been sensitive to fludarabine-containing regimens and 26% had been resistant to fludarabine. Complete remissions, using NCI criteria, were achieved in 20% of patients who had received alkylating agents only, 30% of fludarabine-sensitive patients and 7% of fludarabine-resistant cases (Garcia-Manero et al, 2001). A historical comparison of previously treated patients receiving either fludarabine, with or without prednisolone, fludarabine and cyclophosphamide, or fludarabine, cyclophosphamide and rituximab showed a CR rate and median survival of 13% and 19 months, 12% and 31 months and 28% and not reached respectively (Wierda et al, 2003). ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

13 Transplantation in CLL The majority of patients with CLL are elderly and in older patients the increased morbidity and mortality of high-dose chemo-radiotherapy with allogeneic or autologous stem cell rescue do not justify this approach. About 20% of patients are aged <65 years. In these younger patients, it may well be justifiable to consider a more aggressive approach when standard treatment with alkylating agents or purine analogues offers a poor outlook. To date, the only potentially curative therapy for CLL is allogeneic transplantation. Autologous transplantation. Autologous transplantation is based on the premise that for some chemotherapeutic regimens, there is a dose effect and that superior antitumour efficacy can be achieved by dose escalation. There is some logic to this in CLL, as conventional dose escalation results in increased response rates (Gale & Montserrat, 1993). There are no randomized studies comparing autologous transplantation with other forms of treatment for CLL. The first published studies of autografting came from Boston (Rabinowe et al, 1993) and were subsequently extended (Gribben et al, 1998). In this series, patients with CLL were transplanted after achievement of a good response to prior (unspecified) chemotherapy. The conditioning regimen was cyclophosphamide and total body irradiation (TBI), and the stem cell source was bone marrow purged in vivo with a cocktail of monoclonal antibodies (B5, anti-cd10 and anti- CD20). The procedure-related mortality was 10%, but the most striking observation was that the disease-free survival was 63% at 4 years and the overall survival was 85%. Subsequent reviews have shown that transplant-related mortality ranges from 4 10% to 65 94% of patients are 8 alive at 4 years from transplantation (Van Besien et al, 2001; Dreger & Montserrat, 2002) (Table XII). Evaluation of this data is complicated by heterogeneity in patient selection, choice of conditioning regimen and use of in vitro or in vivo purging. In addition, the majority of studies report data either from the time of stem cell mobilization or from the date of transplantation. This introduces a substantial bias, as only approximately 50% of patients for whom an autograft Table XII. Autografting in chronic lymphocytic leukaemia. Study No. of patients TRM (%) 2-year EFS (%) 4-year EFS (%) Gribben et al (1998) Dreger et al (2000a) Esteve et al (2001) Dreger et al (2000b) Dreger and Montserrat (2002) year OS (%) TRM, treatment-related mortality; EFS, event-free survival; OS, overall survival. is intended actually receive the transplant, either because of a poor response to pretransplant chemotherapy or inadequate stem cell mobilization (Forsyth et al, 2000). Difficulty with stem cell harvesting is related to the heavy marrow involvement with disease and chemotherapyinduced stem cell damage. European Bone Marrow Transplantation group data have shown that more than six cycles of fludarabine is associated with improved stem cell mobilization, as is a delay of more than 2 months between completing chemotherapy and attempting mobilization (Michallet et al, 2000). The optimal mobilization and conditioning regimens and the value of peripheral blood or bone marrow purging remain uncertain. Granulocyte colony-stimulating factor appears insufficient to mobilize peripheral blood stem cells in patients treated with fludarabine and cyclophosphamide (Tournilhac et al, 2004). Evidence for the benefit of autologous transplantation is based on the longer duration of remissions that are achievable compared with those obtained with previous second line treatments in the same patients. In addition, a matched pair analysis showed improved survival of poor risk patients with unmutated IgVH genes who received an autologous transplant rather than conventional chemotherapy (Dreger et al, 2002). The outcome after autologous transplantation appears to be better in patients with mutated IgVH genes (Ritgen et al, 2003), in those who are in CR at the time of autograft and in patients who received a TBI-based conditioning regimen (Dreger et al, 2000a). Patients with del 11q23 have a lower chance of achieving an MRD negative state, postautograft (Stilgenbauer et al, 2000). The lack of a plateau in survival curves and the high risk of clinical relapse in patients who either fail to achieve a molecular CR or who develop a molecular recurrence, suggests that autologous transplantation is not curative in CLL. The optimal timing of autologous transplantation is uncertain. This is being investigated in the MRC CLL5 trial, which compares immediate with the possibility of deferred autologous transplantation in patients who achieve a complete, very good or nodular PR to first or second line treatments. Allogeneic transplantation. There have been no controlled trials evaluating the role of allogeneic transplantation in CLL. The results of registry data and larger single centre studies are shown in Table XIII (Van Besien et al, 2001; Dreger & Montserrat, 2002). The majority of patients undergoing allogeneic transplantation have received several previous types of chemotherapy and many are drug resistant. In contrast to autologous transplantation, allografting is a potentially curative procedure in CLL, based on the longterm disease-free survival of some patients who have achieved a complete remission. The inferior 4-year overall survival in registry data of patients who have undergone allogeneic compared with autologous transplantation, reflects 306 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

14 Table XIII. Allografting in chronic lymphocytic leukaemia. Donor Study No. of patients Median age Related Unrelated Conditioning TRM (%) OS Gribben et al (1998) Standard 22 50% (4 years) Michallet et al (1999) Standard 40 54% (3 years) Horowitz et al (2000) Standard 25 GVDH 45% (3 years) 27 treatment related Dreger and Montserrat (2002) Standard 39 41% (3 years) Dreger et al (2001) Low Intensity 19 TRM, treatment-rated mortality; OS, overall survival; GVHD, graft versus host disease. both the higher treatment-related mortality associated with allografting and the selection of higher risk drug resistant patients. Although most patients have received stem cells from human leucocyte antigen (HLA)-matched siblings, younger patients with high-risk disease have obtained a durable CR following transplantation from an unrelated HLA-matched donor. The observations that patients with MRD after allogeneic transplantation may subsequently become MRD negative (Esteve et al, 2002), and the clinical benefit of donor lymphocyte infusions, strongly suggests that the long-term disease-free survival achievable following allogeneic transplantation is immunologically mediated (Ritgen et al, 2002). These data, together with the high treatment-related mortality associated with standard allogeneic transplantation, have provided the impetus for studies using low intensity conditioning regimens. Treatment-related mortality is reduced but the non-relapse morbidity and mortality remains high in older patients with advanced disease. Disabling GVHD can result in considerable reduction in quality of life. The optimal conditioning regimen and approach to GVHD control is currently uncertain. In vivo T-cell depletion using Campath IH significantly reduces GVHD at the expense of a high incidence of CMV reactivation. In a study of 129 patients with lymphoproliferative disorders receiving a sibling non-myelo-ablative transplant and conditioning with fludarabine and melphalan, there was no difference in event-free or overall survival between patients receiving either Campath 1H and ciclosporin A or methotrexate and ciclosporin A for GVHD prophylaxis (Perez-Simon et al, 2002). A recent overview of 77 low intensity transplants for CLL in Europe has shown a cumulative treatment-related mortality of 18% (95% CI 9 27) with event-free and overall survival at 24 months of 56% (CI 43 69) and 72% (CI 61 83) respectively (Dreger et al, 2003). Radiotherapy Lymphoid neoplasms are exquisitely sensitive to the effects of ionizing radiation, although the systemic nature of CLL has meant that cytotoxic chemotherapy is the principal treatment approach. Radiotherapy, however, continues to play an important although limited role in the palliation of patients with this group of diseases. Splenic irradiation. Splenic irradiation was first reported in the treatment of CLL in For many years, it remained the only available anti-neoplastic therapy for leukaemias. With the advent of systemic chemotherapy, it has become restricted to the treatment of symptomatic splenomegaly unresponsive to chemotherapy, where splenectomy is contraindicated. Splenic irradiation remains a useful, generally well-tolerated and effective palliative treatment with 50 90% of patients experiencing a reduction in splenic size and relief of abdominal pain and discomfort. A complete haematological remission has been reported in 38% of patients in one series (Catovsky et al, 1991). The early MRC 1 and 2 trials reported equivalent survival for patients treated with splenic irradiation and conventional chemotherapy (Catovsky et al, 1991). The mean duration of response is typically 7 18 months and benefit may be seen with a further short course of splenic irradiation. Adverse effects include fatigue, nausea and transient thrombocytopenia and neutropenia. Cytopenia is not, however, a contraindication to therapy as all haematological indices generally improve following a response to radiotherapy (Chisesi et al, 1991). Remarkably low doses of irradiation may be effective and doses of 0Æ5 1 Gray (Gy) one to three times per week has become the conventional practice, as no significant dose response is seen above 10 Gy (Van Mook et al, 2001). External beam radiotherapy for bulky nodal masses. A dose of Gy in 2 Gy fractions has traditionally been used for bulky lymph node masses in the palliation of CLL. However, the radiosensitivity of this disease means that doses of 20 Gy or less may be effective in achieving this goal. A sustained ORR of 81%, with a dose of 4 Gy in two fractions over 3 d to an involved field, can be achieved in this setting (Girinsky et al, 2001). Based on these results, it is clear that lower doses than are conventionally given may be effective in palliation of nodal masses with a consequent reduction in treatment-related morbidity. ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

15 Splenectomy Indications for splenectomy are as follows: Symptomatic massive splenomegaly. Refractory cytopenias. Autoimmune cytopenias. Hypersplenism. There have been no randomized studies comparing splenectomy with other treatments for CLL. A case-controlled study in which 55 patients undergoing splenectomy were matched (sex, age, albumin and Hb levels, Rai stage, number of prior therapies, time from diagnosis) with 55 patients receiving fludarabine, showed no survival advantage in favour of either treatment (Seymour et al, 1997). In a series of studies each comprising more than 40 patients, the operative mortality ranged from 1Æ5 9% with a morbidity (particularly infections) of 26 54% (Christensen et al, 1977; Pegourie et al, 1987; Delpero et al, 1990; Neal et al, 1992; Cusack et al, 1997). No consistent predictive factors for morbidity or mortality were identified. For patients (many of whom were drug resistant) undergoing splenectomy to relieve anaemia or thrombocytopenia, the response rates were 50 77% and 61 88% respectively. These responses are durable, frequently lasting beyond 3 years. No predictive factors for haematological response have been consistently identified. Summary of recommendations for second line and subsequent treatment Patients who relapse after an initial response to low dose chlorambucil may be treated with a further course of chlorambucil (grade B recommendation, level IIb evidence). Patients refractory to low dose chlorambucil should be treated with fludarabine. CHOP is an alternative treatment for patients unsuitable for fludarabine (grade B recommendation, level IIb evidence). Patients who develop progressive disease more than 1 year after receiving fludarabine and whose CLL responded to fludarabine initially, may be treated again with fludarabine alone (grade B recommendation, level IIb evidence). Patients who develop progressive disease within 1 year of previous fludarabine therapy may be treated with a combination of fludarabine and cyclophosphamide (grade B recommendation, level IIb evidence). Patients who are refractory or become resistant to fludarabine currently have a poor prognosis. Therapeutic options include the following: High-dose methyl prednisolone is recommended as a treatment option for patients who are resistant to fludarabine, particularly in cases with bulky lymphadenopathy and/or p53 abnormalities (grade B recommendation, level III evidence). Alemtuzumab is licensed for and recommended in patients without bulky lymphadenopathy, previously treated with alkylating agents and refractory to fludarabine (grade B recommendation, level IIb evidence). Rituximab monotherapy is not recommended for previously treated CLL (grade C recommendation, level IIb evidence). Rituximab combined with fludarabine (with or without cyclophosphamide) may be effective in refractory CLL and warrants further evaluation in this setting (grade B recommendation, level IIb evidence). Autologous transplantation should be considered for patients in complete or good partial remission who are able to withstand high-dose chemotherapy and TBI. Autologous transplantation should be performed in the context of a randomized trial, such as the MRC CLL5 trial. The possibility of an allogeneic transplant procedure should be considered for younger patients with good performance status who have been previously treated and have poor risk disease. Suitable patients should be discussed with a transplant centre at an early stage in their disease before the development of drug resistant disease for inclusion into a clinical research protocol (grade B recommendation, level III evidence). Splenectomy may be beneficial in relieving symptomatic splenomegaly and in improving peripheral cytopenias secondary to hypersplenism or autoantibodies (grade B recommendation, level IIa evidence). Management of complications Prophylaxis and treatment of infections Infective complications are a common clinical problem in CLL, with an incidence of 0Æ26 0Æ47 per patient year, accounting for up to 50% of all CLL-related deaths. The increased susceptibility to infection is both intrinsic to the disease and therapy-related, resulting from multiple factors including hypogammaglobulinaemia, neutropenia, impaired T and natural killer cell function and defective complement activity (Molica, 1994). Risk factors for infection include advanced age, number of previous treatments and ongoing treatment (Perkins et al, 2002; Hensel et al, 2003). Most infections are bacterial (pneumococcus, haemophilus influenzae, staphylococcus) with upper and lower respiratory, septicaemia, pyelonephritis, soft tissue and urogenital infections being the most common. Fungal, viral and opportunistic infections were historically rare but are becoming increasingly prevalent with the introduction of purine analogues, HDMP, alemtuzumab and transplantation (Morrison, 2001). Prophylaxis Antibiotic therapy. Although the use of cycling antibiotics as infective prophylaxis is widely used for patients with recurrent chest infections due to bronchiectasis, or 308 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

16 recurrent urinary tract infections, there are no large studies to assess the efficacy and cost-effectiveness of this approach in patients with CLL. Prophylaxis against Pneumocystis carinii with septrin or nebulized pentamidine should be administered routinely for all patients receiving purine analogues or alemtuzumab, and continued for a minimum of 6 months after stopping purine analogues or alemtuzumab (grade C recommendation, level IV evidence). Prophylaxis against herpes zoster/simplex and fungal infections should also be considered for patients receiving purine analogues or alemtuzumab, particularly if there is a previous history of herpetic or fungal infection. Patients treated with high-dose methylprednisolone should receive prophylaxis against candidiasis with fluconazole. Reactivation of CMV occurs in 10% of patients treated with alemtuzumab, with the peak incidence of reactivation occurring 2 6 weeks after starting treatment. Regular weekly monitoring with CMV PCR testing should be performed in patients receiving alemtuzumab. A positive quantifiable CMV PCR result is an indication for treatment with intravenous ganciclovir. Granulocyte colony-stimulating factor may be useful in reducing the incidence of infection in patients receiving myelotoxic regimens such as fludarabine and cyclophosphamide, and in the early weeks of alemtuzumab treatment during which neutropenia is common (O Brien et al, 1997). Intravenous immunoglobulin (IVIG). Hypogammaglobulinaemia occurs in 20 70% of unselected patients with CLL, being more common in patients with advanced disease stage and in those with a long disease duration (Montserrat & Rozman, 1993). The incidence of infection, particularly with encapsulated organisms, correlates with serum immunoglobulin levels (Chapel & Bunch, 1987). Early studies using prophylactic intramuscular immunoglobulin failed to show consistent benefit. Randomized studies using IVIG have differed in both the dose and duration of immunoglobulin replacement therapy. In a double blind trial, 84 patients with IgG levels of <50% of the lower limit of normal, or a history of one or more serious infections since diagnosis, received IVIG 400 mg/kg every 3 weeks for 1 year or placebo (Griffiths et al, 1989). Patients receiving IVIG had significantly fewer bacterial infections (23 vs. 42, P ¼ 0Æ01), and a longer period from the time of entering the study to that of first infection. The benefit of IVIG at this dose was confirmed in a crossover study, in which 12 patients received IVIG for 1 year before being transferred to placebo. A multicentre multinational study randomized 34 patients to receive IVIG at a dose of either 500 or 250 mg/kg every 4 weeks for 1 year (Chapel et al, 1994). There was no significant difference in the number or severity of infections between the two therapeutic groups. Similarly in a Danish prospective study, 15 patients received low dose IVIG at a dose of 10 g every 3 weeks for 1 year. Compared with the 12 months prior to IVIG therapy, patients required fewer hospital admissions and had fewer febrile episodes (31 vs. 63, P ¼ 0Æ004) (Jurlander et al, 1994). No study has demonstrated an effect on the incidence of viral or fungal infections or on overall survival (Weeks et al, 1991). Although there is little doubt about the efficacy of IVIG therapy in selected patients with CLL there has been debate about its cost-effectiveness. Patients with hypogammaglobulinaemia and recurrent bacterial infections, especially those in whom prophylactic antibiotics have proved ineffective, should be treated with prophylactic IVIG (grade A recommendation, level Ib evidence). Immunization. Patients with CLL have been shown to have a suboptimal response to vaccination against diphtheria, typhoid, mumps, influenza, pneumococcus and haemophilus (Gribabis et al, 1994). Most vaccines are unconjugated and hence thymus independent. Conjugation to a carrier protein renders vaccines thymus dependent and more immunogenic. A recent study showed 21% of CLL patients had baseline immunity to haemophilus influenzae, but following vaccination with a conjugated haemophilus influenzae type b vaccine, this rose to 54% (Sinisalo et al, 2001). In contrast, pneumococcal antibody baseline levels were similar between CLL patients and normal controls, but whereas the latter responded to vaccination (with an unconjugated vaccine) most CLL patients did not. A new protein conjugated form of pneumococcal vaccination has recently been introduced but no data as to its potential benefit in CLL are available. While it is standard practice to recommend annual influenza vaccination for patients with CLL, the efficacy of this measure is uncertain, and further studies on vaccination against infections in CLL are required (Sinisalo et al, 2003) Treatment of infections Patients and their carers need to be educated about the risks of infection and the necessity to seek immediate medical attention as soon as suggestive symptoms occur. Patients with minor infections can be treated as outpatients, but those with major infections will require hospitalization and access to full resuscitation including respiratory support. As many infections are potentially life threatening, broad spectrum antibiotics covering the common likely pathogens should be started as soon as all essential cultures have been taken. Autoimmune cytopenias The incidence of autoimmune cytopenias is significantly higher than in the general population. Warm AIHA, ITP and pure red cell aplasia (PRCA) occur with the incidence of 4 40%, 1 2% and <1% respectively (Hamblin, 2001). These complications may be diagnosed at presentation, or more ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

17 commonly during the course of the disease, particularly in patients with advanced disease stage. Several studies have reported an association between AIHA and treatment with purine analogues. The incidence of AIHA following fludarabine ranges from 2% in previously untreated patients to more than 20% in heavily pretreated patients. There have been no controlled trials comparing different treatments for autoimmune cytopenias in CLL. In a recent study of 52 cases of AIHA, the rare cases with an IgM warm autoantibody had the poorest prognosis (Mauro et al, 2000). It is recommended that patients with warm AIHA or ITP are treated according to the protocols used for patients with idiopathic AIHA or ITP (grade C recommendation, level IV evidence). Autoimmune cytopenias developing following purine analogue therapy are frequently severe and may be fatal (Myint et al, 1995; Weiss et al, 1998). The majority of patients who have developed AIHA post fludarabine have had recurrent haemolysis on re-exposure to fludarabine. There have been reports of small numbers of patients in whom fludarabine has been re-introduced successfully while patients are on ciclosporin A (Cortes et al, 2001). However, retreatment with a purine analogue cannot be recommended in a patient who has previously developed a purine analogue-related immune cytopenia (grade B recommendation, level IIa evidence). The risk of AIHA in patients with a positive DAT without features of haemolysis associated with purine analogue therapy is unknown. Haemolysis does not inevitably occur but purine analogues should be used with caution in this situation, with regular monitoring of the haemoglobin and DAT. There are anecdotal reports of patients with autoimmune PRCA responding to steroids, ciclosporin A, IVIG and alemtuzumab (Castelli et al, 2002; Ru & Liebman, 2003). Recent case reports suggest that rituximab can be effective in patients with AIHA, ITP and PRCA refractory to other treatments (Ghazal, 2002; Gupta et al, 2002; Hegde et al, 2002). Lymphomatous transformation The occurrence of lymphomas in 5 10% of patients with CLL was originally described by Richter (1928). A minority are diagnosed concurrently with CLL, but most are diagnosed during the course of the disease. In the largest reported series of 39 patients, 64% had progressive lymphadenopathy, 23% asymptomatic abdominal mass, 38% had extra nodal involvement, 54% had either fever or weight loss and 80% had a >2-fold elevation of LDH. Histologically most cases are diagnosed as a diffuse large cell lymphoma but Reed Sternberg-like cells are present in 10 15%. Lymphomas may arise as a transformation of the CLL clone or be clonally unrelated. The pathogenesis of lymphomatous transformation is poorly understood but the Epstein Barr virus has been detected in the Reed Sternberg-like cells in lymphomas developing following treatment with purine analogues. Studies of the treatment of lymphomas developing in the context of CLL are largely anecdotal or consist only of small case series. The experience at the MD Anderson Cancer Center has been reported in greater detail (Giles et al, 1998). It suggests that treatment of patients whose histology is diffuse large B-cell lymphoma achieve a response rate of about 40% with CHOP-like therapy, and that response duration and survival are short, at <6 months (Robertson et al, 1993). Increasing the intensity of treatment by using cisplatin/ fludarabine based chemotherapy is not associated with higher response rates (Giles et al, 1999), but the use of highly intensive treatment may generate improved overall survival (Dabaja et al, 2001). In the small number of cases in which treatment of transformations resembling Hodgkin s disease has been reported, the use of standard regimens such as ABVD (adriamycin, bleomycin, vincristine, dacarbazine) is associated with response rates of 40% or less and overall survival of months. However, these results are not significantly different from those reported in patients with de novo Hodgkin s disease who are over 60 years old with stage III or IV disease (Giles et al, 1998). No standard treatment can be recommended for Richter s transformation of CLL; existing clinical reports fail to identify consistently effective therapy (grade C recommendation, level IV evidence). Disclaimer Whilst the advice and information contained in this guideline are believed to be true and accurate at the time of going to press, neither the authors nor the publisher can accept any legal responsibility or liability for any errors or omissions that may be made. D. Oscier 1 C. Fegan 2 P. Hillmen 3 T. Illidge 4 S. Johnson 5 P. Maguire 6 E. Matutes 7 D. Milligan 2 1 Department of Haematology, Royal Bournemouth Hospital, Bournemouth, 2 Department of Haematology, Heartlands Hospital, Birmingham, 3 Department of Haematology, Pinderfields General Hospital, Wakefield, 4 Department of Oncology, Southampton General Hospital, Southampton, 5 Department of Haematology, Taunton and Somerset Hospital, Taunton, Main Street, Nottingham, and 7 Academic Department of Haematology and Cytogenetics, Royal Marsden Hospital, London, UK 310 ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

18 References BCSH Blood Transfusion Task Force (1996) Guidelines on gamma irradiation of blood components for the prevention of transfusionassociated graft-versus-host disease. Transfusion Medicine, 6, Bennett, J.M., Catovsky, D., Daniel, M.T., Fladrin, G., Galton, D.A.G., Gralnick, H.R. & Sultan, C. (1989) The French American British (FAB) proposal for the classification of chronic (mature) B and T lymphoid leukaemias. Journal of Clinical Pathology, 42, Binet, J.L., Auquier, A., Dighiero, G., Chastang, C., Piguet, H., Goasguen, J., Vaugier, G., Potron, G., Colona, P., Oberling, F., Thomas, M., Tchernia, G., Boivin, P., Lesty, C., Duault, M., Monconduit, M., Belabbes, S. & Gremy, F. (1981) A new prognostic classification of chronic lymphocytic leukaemia derived from a multivariate survival analysis. Cancer, 48, Boogaerts, M.A., Van Hoof, A., Catovsky, D., Kovacs, M., Montillo, M., Zinzani, P.L., Binet, J.L., Feremans, W., Marcus, R., Bosch, F., Verhoef, G. & Klein, M. (2001) Activity of oral fludarabine phosphate in previously treated chronic lymphocytic leukemia. Journal of Clinical Oncology, 19, Bosanquet, A.G., Johnson, S.A. & Richards, S.M. (1999) Prognosis for fludarabine therapy of chronic lymphocytic leukaemia based on ex vivo drug response by DiSC assay. British Journal of Haematology, 106, Bosch, F., Ferrer, A., Lopez-Guillermo, A., Gine, E., Bellosillo, B., Villamor, N., Colomor, D., Cobo, F., Esteve, J., Altes, A., Besalduch, J., Ribera, J. & Montserrat, E. (2002) Fludarabine, cyclophosphanide and mitoxauthrone in the treatment of resistent or relapsed chronic lymphocytic leukaemia. British Journal of Haematology, 119, Byrd, J.C., Smith, L., Hackbarth, M.L., Flinn, I.W., Young, D., Proffitt, J.H. & Heerema, N.A. (2003a) Interphase cytogenetic abnormalities in chronic lymphocytic leukemia may predict response to rituximab. Cancer Research, 63, Byrd, J.C., Peterson, B.L., Morrison, V.A., Park, K., Jacobson, R., Hoke, E., Vardiman, J.W., Rai, K., Schiffer, C.A. & Larson, R.A. (2003b) Randomized phase 2 study of fludarabine with concurrent versus sequential treatment with rituximab in symptomatic, untreated patients with B-cell chronic lymphocytic leukemia: results from Cancer and Leukemia Group B 9712 (CALGB 9712). Blood, 101, Byrd, J.C., Peterson, B., Piro, L., Saven, A., Vardiman, I.W., Larson, R.A. & Schiffer, C. (2003c) A phase II study of cladribine treatment for fludarabine refractory B-cell chronic lymphocytic leukemia: results from CALGB study Leukemia, 17, Castelli, R., Vismara, A., Pavia, G., Dagani, R. & Porro, T. (2002) Relapsing pure red cell aplasisa associated with B-cell chronic lymphocytic leukemia successfully treated by intravenous immunoglobulin concentrate. Annali Italiani de Medicini Interna, 17, Catovsky, D., Fooks, J. & Richards, S. (1989) Prognostic factors in chronic lymphocytic leukaemia: the importance of age, sex and response to treatment in survival. British Journal of Haematology, 72, Catovsky, D., Richards, S., Fooks, J. & Hamblin, T. (1991) CLL trials in the United Kingdom. The Medical Research Council CLL Trials 9 1,2,3. Leukemia and Lymphoma, (Suppl.), Chapel, H.M. & Bunch, C. (1987) Mechanisms of infection in chronic lymphocytic leukemia. Seminars in Hematology, 24, Chapel, H., Dicato, M., Gamm, H., Brennan, V., Ries, F., Bunch, C. & Lee, M. (1994) Immunoglobulin replacement in patients with chronic lymphocytic leukaemia: a comparison of two dose regimes. British Journal of Haematology, 88, Cheson, B.D., Bennett, J.M., Grever, M., Kay, N., Keating, M.J., O Brien, S. & Rai, K.R. (1996) National Cancer Institute-sponsored Working Group guidelines for chronic lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood, 87, Chisesi, T., Capnist, G. & Dal Fior, S. (1991) Splenic irradiation in chronic lymphocytic leukemia. European Journal of Haematology, 46, Christensen, B.E., Hansen, M.M. & Videbaek, A. (1977) Splenectomy in chronic lymphocytic leukaemia. Scandanavian Journal of Haematology, 18, CLL Trialists Collaborative Group (1999) Chemotheraputic Options in Chronic Lymphocytic Leukemia: a Meta-analysis of the Randomized Trials. CLL Trialists Collaborative Group. Journal of National Cancer Institute, 91, Cortes, J., O Brien, S., Loscertales, J., Kantarjian, H., Giles, F., Thomas, D., Koller, C. & Keating, M. (2001) Cyclosporin A for the treatment of cytopenia associated with chronic lymphocytic leukemia. Cancer, 92, Crespo, M., Bosch, F., Villamor, N., Bellosillo, B., Colomer, D., Rozman, M., Marce, S., Lopez-Guillermo, A., Campo, E. & Montserrat, E. (2003) ZAP-70 expression as a surrogate for immunoglobulinvariable-region mutations in chronic lymphocytic leukemia. New England Journal of Medicine, 348, Cusack, Jr, J.C., Seymour, J.F., Lerner, S., Keating, M.J. & Pollock, R.E. (1997) Role of splenectomy in chronic lymphocytic leukemia. Journal of American College of Surgery, 185, Dabaja, B.S., O Brien, S.M., Kantarjian, H.M., Cortes, J.E., Thomas, D.A., Albitar, M., Schlette, E.S., Faderl, S., Sarris, A., Keating, M.J. & Giles, F.J. (2001) Fractionated cyclophosphamide, vincristine, liposomal daunorubicin (daunoxome), and dexamethasone (hypercvxd) regimen in Richter s syndrome. Leukemia and Lymphoma, 42, Damle, R.N., Wasil, T., Fais, F., Ghiotto, F., Valetto, A., Allen, S.L., Buchbinder, A., Budman, D., Dittmar, K., Kolitz, J., Lichtman, S.M., Schulman, P., Vinciguerra, V.P., Rai, K.R., Ferrarini, M. & Chiorazzi, N. (1999) Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood, 94, Del Poeta, G., Maurillo, L., Venditti, A., Buccisano, F., Epiceno, A.M., Capelli, G., Tamburini, A., Suppo, G., Battaglia, A., Del Principe, M.I., Del Moro, B., Masi, M. & Amadori, S. (2001) Clinical significance of CD38 expression in chronic lymphocytic leukemia. Blood, 98, Delpero, J.R., Houvenaeghel, G., Gastaut, J.A., Orsoni, P., Blache, J.L., Guerinel, G. & Carcassonne, Y. (1990) Splenectomy for hypersplenism in chronic lymphocytic leukaemia and malignant non- Hodgkin s lymphoma. British Journal of Surgery, 77, Di Raimondo, F., Giustolisi, R., Lerner, S., Cacciola, E., O Brien, S., Kantarjian, H. & Keating, M.J. (2001) Retrospective study of the prognostic role of serum thymidine kinase level in CLL patients with active disease treated with fludarabine. Annals of Oncology, 12, ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

19 Dighiero, G., Maloum, K., Desablens, B., Cazin, B., Navarro, M., Leblay, R., Leporrier, M., Jaubert, J., Lepeu, G., Dreyfus, B., Binet, J.L. & Travade, P. (1998) Chlorambucil in indolent chronic lymphocytic leukemia. French Co-operative Group on Chronic Lymphocytic Leukemia. New England Journal of Medicine, 338, Dohner, H., Fischer, K., Bentz, M., Hansen, K., Benner, A., Cabot, G., Diehl, D., Schlenk, R., Coy, J., Stilgenbauer, S., Volkmann, M., Galle, P., Poustka, A., Hunstein, W. & Lichter, P. (1995) p53 gene deletion predicts for poor survival and non-response to therapy with purine analogs in chronic B-cell leukemias. Blood, 85, Dohner, H., Stilgenbauer, S., James, M.R., Benner, A., Weilgum, T. & Bentz, M. (1997) 11q deletions identify a new subset of B-cell chronic lymphocytic leukaemia characterised by extensive nodal involvement and inferior prognosis. Blood, 89, Dohner, H., Stilgenbauer, S., Benner, A., Leupolt, E., Krober, A., Bullinger, L., Dohner, K., Bentz, M. & Lichter, P. (2000) Genomic aberrations and survival in chronic lymphocytic leukemia. New England Journal of Medicine, 343, Dreger, P. & Montserrat, E. (2002) Autologous and allogeneic stem cell transplantation for chronic lymphocytic leukemia. Leukemia, 16, Dreger, P., van Biezen, A., Brand, R., Esteve, J., Gratwohl, A., Kimby, E., Michallet, M., Milligan, D.W. & Niederwieser, D. (2000a) Prognostic factors for survival after autologous stem cell transplantation for chronic lymphocytic leukemia (CLL): the EMBT experience. Blood, 96(Suppl. 1), 482a. Dreger, P., von Neuhoff, N., Sonnen, R., Kuse, R., Seyfarth, B., Glass, B. & Schmitz, N. (2000b) Feasibility and efficacy of autologous stem cell transplantation for poor risk CLL. Blood, 96(Suppl. 1), 483a. Dreger, P., van Biezen, A., Brand, R., Hansz, J., Corradini, P., Finke, J., Lambertenghi-Deliliers, G., Russell, N., Michallet, M. & Niederwieser, D. (2001). Allogenic stem cell transplantation (SCT) for chronic lymphocytic leukaemia using intensisty-reduced conditioning. An EMBT survey. Blood, 98, 743a. Dreger, P., Stilgenbauer, S., Benner, A., Ritgen, M., Schmitz, N. & Dohner, H. (2002) High-dose therapy with autologous stem cell transplantation (ASCT) prolongs survival of patients with chronic lymphocytic leukemia (CLL): a matched-pair analysis. Blood, 100(Suppl. 1), 217a. Dreger, P., Brand, R., Hansz, J., Milligan, D., Corradini, P., Finke, J., Deliliers, G., Martino, R., Russell, N., Van Biezen, A., Michallet, M. & Niederwieser, D. (2003) Treatment related mortality and graft versus leukemia effect after allogeneic stem cell transplantation for chronic lymphocytic leukemia using intensity reduced conditioning. Leukemia, 17, Durig, J., Nascher, M., Schmucker, U., Renzing-Kohler, K., Huttmann, A. & Duhrsen, U. (2002) CD38 expression is an important prognostic marker in chronic lymphocytic leukaemia. Leukemia, 16, Eichhorst, B., Busch, R., Stauch, M., Kneba, M., Ritgen, M., Soling, U., Burkhard, O., Bergmann, M., Wendtner, M., Hiddermann, W., Emmerich, B. & Hallek, M. (2003a) Fludarabine induces higher response rates in first line therapy of older patients with advanced chronic lymphocytic leukemia than chlorambucil: interim analysis of a phase 111 study of the German CLL study group. Blood, 102(Suppl. 1), 109a. Eichhorst, B., Busch, R., Hopfinger, G., Pasold, R., Hensel, M., Soling, U., Siehl, S., Steinbrecher, C., Jager, U., Bergmann, M., Wendtner, M., Hiddermann, W., Emmerich, B. & Hallek, M. (2003b) Fludarabine plus cyclophosphamide induces higher remission rates and longer progression free survival than fludarabine alone in first line therapy of advanced chronic lymphocytic leukemia: results of a phase 111 study of the German CLL study group. Blood, 102(Suppl. 1), 72a. Esteve, J., Montserrat, E., Dreger, P., Meloni, G., Pavletic, S., Catovsky, D., Dearden, D., Scime, C., Sutton, L., Michallet, M., Desablens, B., Kimby, E., Coiffier, B., Brunet, S., Sanz, M.A., Besalduch, J., Cabellero, D., Juliusson, G., Conde, E., Del Potro, E. & Schmitz, N. (2001) Stem cell transplantation (SCT) for chronic lymphocytic leukemia (CLL): outcome and prognostic factors after autologous and allogenic transplants. Blood, 98(Suppl. 1), 482a. Esteve, J., Vilamor, N., Colomer, D. & Montserrat, E. (2002) Different clinical value of minimal residual disease after autologous and allogenic stem cell transplantation for chronic lymphocytic leukemia. Blood, 99, Ezdinli, E., Stutzman, L., William August, C. & Firat, D. (1969) Corticosteriod therapy for lymphomas and chronic lymphocytic leukemia. Cancer, 23, Faderl, S., Thomas, D.A., O Brian, S., Giles, F., Koller, C.A., Beran, M., Williams, M., Garcia-Manero, G., Kantarjian, H.M. & Keating, M.J. (2001) An exploratory study of the combination of monoclonal antibodies CAMPATH-1H and rituximab in the treatment of CD52- and CD20-positive chronic lymphoid disorders. Blood, 98(Suppl. 1), 365a. Forsyth, P., Milligan, D. & Davies, F.E. (2000) High dose chemoradiotherapy with autologous stem cell rescue for patients with CLL is an effective and safe means of inducing molecular responses: an MRC pilot study. Blood, 96(Suppl. 1), 843a. French Co-operative Group on Chronic Lymphocytic Leukaemia (1990a) Natural history of stage A chronic lymphocytic leukaemia untreated patients. British Journal of Haematology, 76, French Co-operative Group on Chronic Lymphocytic Leukaemia (1990b) A randomized clinical trial of chlorambucil versus COP in stage A chronic lymphocytic leukaemia. French Co-operative Group on Chronic Lymphocytic Leukaemia. Blood, 75, French Co-operative Group on CLL, Johnson, S., Smith, A.G., Loffler, H., Osby, E., Juliusson, G., Emmerich, B., Wyld, P.J. & Hiddemann, W. (1996) Multicentre prospective randomised trial of fludarabine versus cyclophosphamide, doxorubicin and prednisone (CAP) for treatment of advanced stage chronic lymphocytic leukaemia. Lancet, 347, Gale, R.P. & Montserrat, E. (1993) Intensive therapy of chronic lymphocytic leukaemia. Bailliere s Clinical Haematology, 6, Galton, D., Wiltshire, E., Szur, J. & Dacie, J. (1961) The use of chlorambucil and steriods in the treatment of chronic lymphocytic leukaemia. British Journal of Haematology, 7, Garcia-Manero, G., O Brien, S., Cortes, J., Faderl, S., Giles, F., Albitar, M., Lerner, S., Kantarjian, H. & Keating, M. (2001) Update of results of the combination of fludarabine, cyclophosphamide and rituximab for previously treated patients with chronic lymphocytic leukaemia (CLL). Blood, 98(Suppl. 1), 633a. Ghazal, H. (2002) Successful treatment of pure red cell aplasia with rituximab in patients with chronic lymphocytic leukemia. Blood, 99, Ghia, P., Guida, G., Stella, S., Gottardi, D., Guena, M., Strola, G., Scielzo, C. & Caligaris-Cappio, F. (2003) The pattern of CD38 expression defines a distinct subset of chronic lymphocytic leukemia (CLL) patients at risk of disease progression. Blood, 101, ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,

20 Giles, F.J., O Brien, S.M. & Keating, M.J. (1998) Chronic lymphocytic leukemia in (Richter s) transformation. Seminars in Oncology, 25, Giles, F.J., O Brien, S.M., Santini, V., Gandhi, V., Plunkett, W., Seymour, J.F., Robertson, L.E., Kantarjian, H.M. & Keating, M.J. (1999) Sequential cis-platinum and fludarabine with or without arabinosyl cytosine in patients failing prior fludarabine therapy for chronic lymphocytic leukemia: a phase II study. Leukemia and Lymphoma, 36, Girinsky, T., Guillot-Vals, D., Koscielny, S., Cosset, J.M., Ganem, G., Carde, P., Monhonval, M., Pereira, R., Bosq, J., Ribrag, V., Vantelon, J.M. & Munck, J.N. (2001) A high and sustained response rate in refractory or relapsing low-grade lymphoma masses after low-dose radiation: analysis of predictive parameters of response to treatment. International Journal of Radiation Oncology, Biology, Physics, 51, Grever, M.R., Kopecky, K.J., Coltman, C.A., Files, J.C., Greenberg, B.R., Hutton, J.J., Talley, R., Von Hoff, D.D. & Balcerzak, S.P. (1988) Fludarabine monophosphate: a potentially useful agent in chronic lymphocytic leukaemia. Nouvelle Revue Francaise d Hematologie, 30, Gribabis, D.A., Panayiotidis, P., Boussiotis, V.A., Hannoun, C. & Pangalis, G.A. (1994) Influenza virus vaccine in B-cell chronic lymphocytic leukaemia patients. Acta Haematologica, 91, Gribben, J.C., Neuberg, D., Soiffer, R.J., Fischer, D.C., Schlossman, R., Alyea, E.P., Kuhlman, C., Ritz, J. & Nadler, L.M. (1998) Autologous versus allogenic bone marrow transplantation for patients with poor prognosis CLL. Blood, 92(Suppl.), 322a. Griffiths, H., Brennan, V., Lea, J., Bunch, C., Lee, M. & Chapel, H. (1989) Crossover study of immunoglobulin replacement therapy in patients with low-grade B-cell tumors. Blood, 73, Gupta, N.K., Patel, D., Kavuru, S., Janson, D., Hoffman, M., Ramadoss, U., Ahmed, S. & Rai, K. (2001) Rituximab, cyclophosphamide and decadron (RCD) is effective in previously treated advanced chronic lymphocytic leukaemia (CLL). Journal of Clinical Oncology, 20, page 284a (abstract no. 1133). Gupta, N.K., Kavuru, S., Patel, D., Janson, D., Driscoll, N. & Ahmed, S. (2002) Rituximab-based chemotherapy for steroid-refractory autoimmune hemolytic anemia of chronic lymphocytic leukemia. Leukemia, 10, Hainsworth, J., Litchy, S., Barton, J., Houston, G., Hermann, F., Bradof, J. & Greco, F. (2003) Single agent rituximab as first line and maintenance treatment for patients with chronic lymphocytic leukemia or small lymphocytic lymphoma: a phase II trial of the Minnie Pearl Cancer Research Network. Journal of Clinical Oncology, 21, Hallek, M., Langenmayer, I., Nerl, C., Knauf, W., Dietzfelbinger, H., Adorf, D., Ostwald, M., Busch, R., Kuhn-Hallek, I., Thiel, E. & Emmerich, B. (1999) Elevated serum thymidine kinase levels identify a subgroup at high risk of disease progression in early, nonsmoldering chronic lymphocytic leukemia. Blood, 93, Hallek, M., Schmitt, B., Wilhelm, M., Busch, R., Krober, A., Fostitsch, H.P., Sezer, O., Herold, M., Knauf, W., Wendtner, C.M., Kuse, R., Freund, M., Franke, A., Schriever, F., Nerl, C., Dohner, H., Thiel, E., Hiddemann, W., Brittinger, G., Emmerich, B. & the German Chronic Lymphocytic Leukaemia Study Group (2001) Fludarabine plus cyclophosphamide is an efficient treatment for advanced chronic lymphocytic leukaemia (CLL): results of a phase II study of the German CLL Study Group. British Journal of Haematology, 114, Hamblin, T. (2001) Autoimmune Disease and Its Management in Chronic Lymphocytic Leukemia. In: Chronic Lymphoid Leukemias (ed. by B.D. Cheson), pp Marcel Dekker, New York. Hamblin, T.J., Davis, Z., Gardiner, A., Oscier, D.G. & Stevenson, F.K. (1999) Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood, 94, Hamblin, T.J., Orchard, J.A., Ibbotson, R.E., Davis, Z., Thomas, P.W., Stevenson, F.K. & Oscier, D.G. (2002) CD38 expression and immunoglobulin variable region mutations are independent prognostic variables in chronic lymphocytic leukemia, but CD38 expression may vary during the course of the disease. Blood, 99, Han, T., Ezdinli, E.Z., Shimaoka, K. & Desai, D.V. (1973) Chlorambucil vs. combined chlorambucil-corticosteroid therapy in chronic lymphocytic leukemia. Cancer, 31, Hegde, U.P., Wyndham, H.W., White, T. & Cheson, B.D. (2002) Rituximab treatment of refractory fludarabine-associated immune thrombocytopenia in chronic lymphocytic leukemia. Blood, 100, Hensel, M., Kornacker, M., Yammeni, S., Egerer, G. & Ho, A. (2003) Disease activity and pretreatment rather than hypogammaglobulinaemia, are major risk factors for infectious complications in patients with chronic lymphocytic leukaemia. British Journal of Haematology, 122, Horowitz, M.M., Montserrat, E., Sobocinski, K., Giralt, S., Khouri, I.F. & Schmitz, N. (2000) Haemopoietic stem cell transplantation for chronic lymphocytic leukaemia. Blood, 96(Suppl. 1), 522a. Houlston, R., Catovsky, D. & Yuille, M. (2002) Genetic susceptibility to chronic lymphocytic leukemia. Leukemia, 16, Ibrahim, S., Keating, M., Do, K.A., O Brien, S., Huh, Y.O., Jilani, I., Lerner, S., Kantarjian, H.M. & Albitar, M. (2001) CD38 expression as an important prognostic factor in B-cell chronic lymphocytic leukemia. Blood, 98, Jaksic, B. & Brugiatelli, M. (1988) High dose continuous chlorambucil vs intermittent chlorambucil plus prednisone for treatment of B-CLL IGCI CLL-01 trial. Nouvelle Review Francais d Hematologie, 30, Jaksic, B. & Vitale, B. (1981) Total tumour mass score (TTM): a new parameter in chronic lymphocyte leukaemia. British Journal of Haematology, 49, Jaksic, B., Brugiatelli, M., Krc, I., Losonczi, H., Holowiecki, J., Planinc- Peraica, A., Kusec, R., Morabito, F., Iacopino, P. & Lutz, D. (1997) High dose chlorambucil versus Binet s modified cyclophosphamide, doxorubicin, vincristine, and prednisone regimen in the treatment of patients with advanced B-cell chronic lymphocytic leukemia. Results of an international multicenter randomized trial. International Society for Chemo-Immunotherapy, Vienna. Cancer, 79, Johnson, S., Smith, A.G., Loffler, H., Osby, E., Juliusson, G., Emmerich, B., Wyld, P. & Hiddemann, W. (1996) Multicentre prospective randomised trial of fludarabine versus cyclophosphamide, doxirubicin, and prednisilone (CAP) for the treatment of advanced-stage chronic lymphocytic leukaemia. The French Co-operative Group on CLL. Lancet, 347, Juliusson, G. & Liliemark, J. (1993) High complete remission rate from 2-chloro-2 -deoxyadenosine in previously treated patients with B-cell chronic lymphocytic leukaemia: response predicted by rapid ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 125,