Spontaneous clinical regression in chronic lymphocytic leukaemia


Professor D. Catovsky, Academic Department of Haematology and Cytogenetics, The Royal Marsden NHS Trust, London SW3 6JJ, UK. E-mail: d.catovsky@icr.ac.uk


Summary. Chronic lymphocytic leukaemia (CLL) is a B-cell disorder, which has a median survival of over 10 years from diagnosis for stage A disease. The natural history of stage A disease is generally indolent or only slowly progressive. It is less well known that CLL may undergo spontaneous regression. We report a series of 10 such cases (eight stage A and two stage B) followed at our institutions.

Chronic lymphocytic leukaemia (CLL) accounts for 25% or more of all cases of leukaemia. It is conventionally staged into three groups according to the Binet system. Stage A disease is variably progressive over time, with about 30% and 50% of patients developing progressive disease at 5 and 10 years respectively. Cumulative data on 2370 patients entered in the Medical Research Council (MRC) CLL trials since 1978 show median survivals of 9·2, 4·8 and 3·5 years for stages A, B and C respectively (personal communication, J. Halsey, CTSU, Oxford, UK). Patients with stage A CLL do not require therapy unless they progress or the doubling time is short. Indeed evidence for the lack of benefit of early treatment of stage A CLL was shown in a large overview of randomization trials (CLL Trialists, 1999). There are only a few cases in which spontaneous regression of CLL has been reported. We present 10 patients with CLL in whom the illness has undergone spontaneous regression, and in whom no treatment for CLL was given.

Case reports

Relevant data of the 10 cases are summarized in Table I.

Table I.  Summary of the clinical data in the 10 patients.
CaseAge at
Date of
Stage at
at diagnosis
Lymphocyte × 109/l
at diagnosis
Highest lymphocyte
count × 109/l
Marrow %
of lymphoid


× 109/l
Duration of
  • *

    Stage by Binet System and Rai stage in brackets in patients with stage A.

  • Case report, patient 10.

  • N/A, not available; l+, lambda light chain restriction; k+, kappa light chain restriction.

160/FemaleMay 93A (I)CD5+
7·47·424%YesTrisomy 12
April 01
 8In Feb 01 only
6% B cells
absence of
trisomy 12
269/MaleMarch 96A (I)N/A 818180%YesN/AMarch 00
blood: 84% T
cells, only 20%
B cells were
370/FemaleNovember 91A (0)CD5+
5·75·7N/AN/AN/AMarch 01
451/FemaleOctober 95A (I)CD5+
 558780%Yes47XY+12February 01
 530% of bone
still show
trisomy 12 by
554/MaleAugust 94A(0)CD23+
9·114·4N/AN/AN/AJune 01
 6Breast lump
showed CLL,
50% of cells in
06/00 CD19
& CD5+ve
673/FemaleOctober 89A (0)CD19+
 13·614·5N/AN/AN/AJuly 00
769/FemaleJanuary 97A (II)CD19+
 10·410·9N/AN/AN/ANovember 00
 4At diagnosis
cells were only
weakly CD5+ve
857/MaleFebruary 88A (II)CD19+
N/AN/ANovember 00
947/MaleMarch 79BCD19+
N/AN/AAugust 00
1047/MaleOctober 82BCD19+
 22 96Yes
N/AN/AApril 01
18No palpable
disease present
when last seen

Case 1, a 60-year-old woman, presented in 1993 with a 9-year history of recurrent attacks of toxoplasmosis. She was noted to have small volume left cervical and axillary nodes since 1985 which remained stable in size over that period. In 1993, when she was found to have monoclonal B lymphocytosis, a lymph node biopsy was performed that was consistent with a low-grade B-cell lymphocytic lymphoma. Since the diagnosis of CLL, she continued to have recurrent toxoplasmosis for which she received intermittent courses of ciprofloxacin. During follow up, she developed larger bilateral cervical and axillary lymph nodes, the largest measuring 2 × 1 cm, and these subsequently disappeared on clinical examination, while the lymphocyte count gradually decreased reaching a nadir of 1·1 × 109/l in July 1998. Immunophenotyping at this time revealed most lymphocytes to be T cells (CD2 : 87%) and only 4% of the lymphocytes expressed CD5/CD19 positivity, suggesting minimal residual disease in the peripheral blood (PB). A search for trisomy 12, previously present using fluorescence in situ hybridization (FISH) analysis in 1995, was negative in peripheral blood.

Case 2 is a 69-year-old man who presented with stage A CLL in March 1996 with a total lymphocyte count of 81 × 109/l. He had a few 2–3 cm lymph nodes in the left cervical region and both axillae. Bone marrow revealed 80% lymphoid infiltration. Over time his peripheral counts fell spontaneously to a lowest count of 2·5 × 109/l at his last visit, associated with diminishing lymphadenopathy. A repeat bone marrow in July 1999 showed evidence of persisting CLL with 25% CD5/CD19-positive cells and his trephine continued to show interstitial infiltration, but the infiltrate in the aspirate and the trephine had both markedly decreased from presentation. His most recent peripheral count revealed 84% of the lymphocytes to be T cells and only 20% of the B cells co-expressed CD5 and CD19.

Case 3 is a 70-year-old woman who presented in 1991 with a lymphocytosis of 5·7 × 109/l in Stage A (0) disease which persisted through 1992–3. Since 1994 her blood counts have been normal and, on last review in March 2001, she had a lymphocyte count of 1·4 × 109/l.

Case 4 is a 51-year-old woman who was referred for a cervical lymph node biopsy in 1995. At that time she had a peripheral lymphocyte count of 55 × 109/l and chromosome studies revealed trisomy 12 in 16% of her peripheral lymphocytes. The lymph node biopsy was consistent with CLL, as was the marrow aspirate and trephine. Initially, her lymphocyte count increased to a maximum of 87 × 109/l and then, along with a regression in the size of her lymph nodes, the lymphocyte count fell over the subsequent 5 years to 3·7 × 109/l in February 2001 when last seen. Despite this, FISH analysis still showed trisomy 12 in 30% of the bone marrow lymphocytes in August 1999 and the degree of bone marrow lymphocyte infiltration of 80% was similar to the marrow aspirate at presentation. The peripheral lymphocyte count at that time (August 1999) was 3·7 × 109/l.

Case 5, a 54-year-old man, was noted to have a mild lymphocytosis since 1994 but was not investigated until 1998, when he was referred to the surgeons for evaluation of a breast lump. CLL was confirmed in the peripheral blood by immunophenotyping when the total lymphocyte count was 14·4 × 109/l. The completely excised breast lump showed infiltration with CLL. His disease remained stable until May 1999, when he was admitted with a flu-like illness and abnormal liver function tests. He then had a lymphocyte count of 2·8 × 109/l and in June 2001 the count remained 3·1 × 109/l. Markers revealed that 50% of the cells expressed CD19 and CD5. Clinical examination was entirely normal. There was no recurrence of the breast lump.

The patient in case 6 presented in October 1989 with a fracture and a lymphocyte count of 13·6 × 109/l. Immunophenotyping confirmed B-cell CLL. Her lymphocyte count peaked at 14·5 × 109/l the following year. Subsequently, her lymphocyte counts gradually fell and, when last seen in July 2000, were 2·9 × 109/l. Clinical examination was unremarkable throughout.

Case 7, a 69-year-old woman, was diagnosed with B-cell CLL in January 1997 when she presented to her general practitioner (GP) with flu-like symptoms. Her lymphocyte count was 10·4 × 109/l and immunophenotyping suggested CLL, with CD 23 and CD 5 weakly positive. By June 1997 she had developed splenomegaly which was confirmed on ultrasound to be 24 cm long. Subsequent visits demonstrated shrinking splenomegaly and falling lymphocyte counts. When last seen, in November 2000, her splenomegaly had decreased to 3 cm below her costal margin and she had a lymphocyte count of 1·4 × 109/l. Repeat immunophenotyping showed only 1% B cells with no light chain restriction. A bone marrow revealed 12% lymphocytes, only 3% of the cells being B cells with no surface immunoglobulin restriction.

Case 8 is a 57-year-old man whose routine blood count in 1988 revealed a lymphocyte count of 25 × 109/l. A bone marrow aspirate showed increased cellularity and a diffuse infiltration of lymphocytes. The next year he developed palpable splenomegaly that progressed over the next 5 years to being palpable 8 cm below the costal margin. The peripheral lymphocyte count remained stable at 17 × 109/l. Over the next 12 years his lymphocyte count and spleen size decreased and, when last seen, the spleen was not palpable and he had a lymphocyte count of 3·4 × 109/l. Immunophenotyping, however, showed the persistence of a small clone of CD19/CD5-positive cells.

Case 9 is a 47-year-old man who presented in 1979 when a routine count revealed a lymphocyte count of 124 × 109/l. He had 1 × 1 cm bilateral cervical and inguinal lymphadenopathy and a spleen that was just palpable. A bone marrow aspirate revealed a heavy infiltrate of lymphocytes. Over the next 20 years his lymphocyte counts gradually decreased without any treatment. When last seen in August 2000, he had a lymphocyte count of 13·3 × 109/l and no lymphadenopathy or splenomegaly.

Case 10, a 47-year-old man, was noted to have a lymphocytosis in 1982 as part of a routine blood count (lymphocytes 22 × 109/l). In 1983 his lymphocyte count peaked at 96 × 109/l; of these, 90% were CD5/CD19 positive and demonstrated lambda light chain restriction. He had cervical, axillary and inguinal lymphadenopathy, the largest being 2 × 2 cm, and a spleen that was just palpable. His disease regressed over the next 18 years and, when seen in February 2000, he had a lymphocyte count of 14 × 109/l and no lymphadenopathy and splenomegaly. However, over the last few months his lymphocyte count has risen to 27 × 109/l, although he remains otherwise well with no organomegaly or lymphadenopathy.


Twelve cases of spontaneous remission in CLL have been reported in the literature since 1970 (Schmidt et al, 1995; Bernard et al, 1999). Most of these cases showed a significant decrease in the blood lymphocyte count and, in a few cases, this was associated with blood or bone marrow remission. Two of them had normal bone marrow immunophenotyping at the time of remission and one had molecular studies that showed the presence of residual disease in the bone marrow in the presence of molecular blood remission. We describe 10 cases seen at our institutions in which regression of the disease has occurred in the absence of treatment. Except for a solitary patient developing flu-like symptoms prior to the regression of disease (case 5), no causative factors were obvious.

In a review of the literature undertaken by Schmidt et al (1995), 37 CLL patients were reported as experiencing spontaneous remission. However, only 11 cases presented in the series were true spontaneous remissions (in the absence of any treatment thought to modify disease). The remaining patients received an assortment of treatments that included corticosteroids, chlorambucil, triethylenemelamine, injections of liver extracts, splenectomy and variable amounts of radiation. Surprisingly, about 40% of all cases followed up for more than 2 years developed secondary neoplasia. This is far in excess of the 10% increased risk of secondary malignancies reported by Greene et al (1978). It was postulated that this might be secondary to persisting alterations of immune function, despite disappearances in the manifestations of CLL.

Clinico-haematological remission in CLL is defined as (i) absence of lymphadenopathy, splenomegaly or constitutional symptoms; (ii) peripheral lymphocytosis less than 4 × 109/l, no anaemia, no thrombocytopenia; (iii) bone marrow lymphocytosis in aspirates < 30% and/or normal bone marrow histology, as defined by the absence of lymphoid infiltration and < two lymphoid nodules (Cheson et al, 1988). Six out of our 10 patients fit the first two criteria. Only three of these patients had a bone marrow aspirate or trephine performed at remission and, in two cases, it continued to show lymphoid infiltrates suggestive of residual CLL. With the advent of flow cytometry and molecular biology techniques, more stringent evaluation of residual disease can be achieved. Advances in detection of very small numbers of malignant cells using dual-marker flow cytometry with CD5 and CD19 monoclonal antibodies and polymerase chain reaction (PCR) study of gene rearrangements have shown that, despite the achievement of complete clinical remission, most patients continue to carry malignant B cells of the CLL clone (Vuillier et al, 1992).

Spontaneous regression of CLL is rare, as documented by the few cases in the literature. In an attempt to review this further, we reviewed the data on patients with Stage A disease registered with the MRC CLL 3A trial. These patients were registered but no treatment was offered unless the disease progressed or the doubling time was < 12 months. In all, 714 patients were registered and the blood counts were available for 1 year after diagnosis (personal communication, Dr Sue Richards, CTSU, Oxford, UK). Eleven (1·5%) of these patients had a 50% fall in lymphocyte count from a preceding highest value, but seven of these 11 patients progressed subsequently with increasing counts after the fall. Only three patients had a reduction in their counts to < 20% of a previous highest lymphocyte count and only two have maintained the decline within the period of a year. As the last follow-up blood count in CLL3A was taken after 1 year, one can only speculate as to the number of patients who show continuing spontaneous regression.

Our centres are referred a total of about 120 patients with CLL per year. A number of patients with CLL are followed up in the community by their GPs and are not referred to hospital. Conceivably, some of these patients could have undergone spontaneous regression of the illness without it having been brought to medical attention. Therefore, only a limited amount of information can be obtained by the annual referral rate, but it seems certain that spontaneous regression is a rare event.

The mechanism of spontaneous long-lasting regression in CLL remains speculative and poorly understood. In some cases they are associated with viral infections, smallpox vaccinations or the development of secondary neoplasms (Schmidt et al, 1995). Also, the pathogenesis of B-cell CLL is poorly understood but may be related to apoptotic pathways. B-cell CLL cells express high levels of the anti-apoptotic members of the Bcl 2 family and no pro-apoptotic molecules such as Fas and C–myc. However, interactions with other components of the immune system and with the microenvironment of the bone marrow and lymphoid organs might affect the course of the disease (Oscier, 1999). Immune-mediated control of B-cell proliferation by T cells has been suggested (Mu et al, 1997), but studies of the T-cell subsets in B CLL, the cytokines produced by these cells or functional studies have failed to elicit consistent results.

Two of our patients had cytogenetic studies performed and both had trisomy 12 identified using FISH. Although our patients appear to have done well, this cytogenetic abnormality is usually associated with a poorer prognosis and/or disease progression (Hjalmar et al, 1998; Juliusson & Merup, 1998). Other prognostic features have been reported in B-cell CLL, for example the somatic mutations of Ig VH genes and the expression of CD 38 (Damle et al, 1999; Hamblin et al, 1999). These are likely to be important in the evaluation of CLL patients, particularly those undergoing spontaneous regression, such as those we have described.

This report highlights the possibility of a ‘natural’ regulatory mechanism controlling B-cell CLL, which in rare instances leads to a downregulation of cell proliferation and, eventually, to a durable clinical remission. The published evidence indicates that remissions in CLL have occurred in early disease. Two of our patients had stage B disease and in both there was substantial regression in the illness.

Further reports of spontaneous remission in B-cell CLL including immunological, cytogenetic and molecular studies will hopefully lead to a better understanding of the self-regulatory mechanisms in this lymphoproliferative disorder.


We are grateful to Nick Telford, Christie Hospital, Manchester, UK, for helping with the cytogenetic work carried out on patient number 4.