Interferon α and zidovudine therapy in adult T-cell leukaemia lymphoma: response and outcome in 15 patients


Dr Estella Matutes, MD, PhD, FRCPath, Academic Haematology and Cytogenetics, Royal Marsden Hospital NHS Trust/Institute of Cancer Research, Fulham Road, London, SW3 6JJ, UK. E-mail:


Adult T-cell leukaemia lymphoma (ATLL) is an aggressive disease caused by the human T-lymphotropic virus 1 (HTLV-I) with a short survival. Responses to interferon alpha (IFN-α) and zidovudine (AZT) have been documented but not with long-term follow-up. We treated 15 ATLL patients with IFN and AZT. Eleven patients had acute ATLL, two had lymphoma and two smouldering ATLL, with progression. The main features were: organomegaly (14), skin lesions (10), high white blood cell (WBC) count (11) and hypercalcaemia (9). Eleven patients had previously received chemotherapy and one had received an autograft. At the time of the study, seven patients had progressive disease and eight were in partial or complete clinical remission. Responses (PR) lasting 2+ to 44+ months were seen in 67%; 26% did not respond (NR) and one patient was not evaluable. Hypercalcaemia predicted a poor outcome but differences were not significant. Eight of the 15 patients have died 3–41 months from diagnosis. Median survival for the 15 patients was 18 months. Survival of the NR ranged from 4 to 20 months; six PR patients are alive 8–82 months from diagnosis. The differences in survival between NR (median: 6 months) and PR (55% of patients alive at 4 years) were statistically significant (P = 0·002). In conclusion, IFN and AZT improves the outcome of ATLL patients and helps maintain responses.

Adult T-cell leukaemia lymphoma (ATLL) is an aggressive T-cell malignancy aetiologically linked to the human T-lymphotropic virus 1 (HTLV-I) with median survival ranging from 7 to 13 months in both endemic and non-endemic regions (Shimamoto et al, 1990a, b; Shimoyama et al, 1991; Plumelle et al, 1993; Pombo de Oliveira et al, 1995; Matutes & Catovsky, 1998). In the United Kingdom (UK), the disease affects mainly immigrants from the Caribbean basin or Africa and more rarely from other countries (Matutes & Catovsky, 1994). Although the incidence of ATLL among the various T-cell malignancies in the UK is unknown, it is estimated to be around 20 newly diagnosed cases/year (Taylor et al, 2000).

Despite new insights into the pathogenesis of ATLL, there has been minimal improvement in the treatment and survival of these patients. Combination chemotherapy including that used in acute lymphoblastic leukaemias results in a low response rate of around 30% and the durability of responses is short (Shimamoto et al, 1990a, b; Shimoyama et al, 1991; Matutes & Catovsky, 1998; Pawson et al, 1998).

Novel strategies such as the use of unlabelled or radiolabelled anti-CD25 monoclonal antibody (mAb), deoxycoformycin and interferons (IFN) alpha, beta and gamma have also been used, but only occasional patients have achieved a durable response (Tamura et al, 1987; Ichimaru et al, 1988; Waldmann et al, 1988; Mercieca et al, 1994; Waldmann, 1994). Isolated case reports of a response to etoposide monotherapy (Kojima et al, 1993), allogeneic bone marrow transplantation (Ljungman et al, 1994) and, more recently, a case responding to all-trans retinoic acid have been documented (Maeda et al, 2000). Extremely rare spontaneous remissions have been reported (Schnitzer et al, 1983; Mattock et al, 1986).

A preliminary report showed that a combination of IFN-α and zidovudine (AZT) was effective in a single ATLL patient co-infected with HTLV-I and the human immunodeficiency virus (HIV), producing a sustained response (Shibata et al, 1989). Two subsequent studies of a total of 24 ATLL patients further suggested that IFN-α combined with AZT produced a measurable response with 67% (7 out of 24) of the patients achieving a response, 29% of which were considered clinical complete responses (CR) (Gill et al, 1995; Hermine et al, 1995). In the largest study, which included 19 patients, the overall median survival was 3 months and that of the complete and partial responders 13 months versus 1·3 months for the non-responders (Gill et al, 1995). Therefore, it could not be concluded that this drug combination was superior to standard chemotherapy.

We report the response rate, long-term follow-up and survival of 15 ATLL patients treated over a 4 year period with the combination of IFN-α and AZT. This study aimed to (i) determine the efficacy and tolerability of IFN-α combined with AZT in ATLL, and (ii) estimate the response duration and compare survival with the historical series of ATLL treated in the UK during the last 15 years.

Patients and methods

Patients Fifteen ATLL patients treated over a 4 year period since August 1996 have been included in this study. The diagnosis was based on cell morphology, immunological markers, tissue histology and HTLV-I serology and/or DNA.

The series included six men and nine women (M/F:0·66) with a median age of 54 years (range: 28–79 years). The region of origin was: Caribbean islands (10), Africa (2), South-America (2) and Middle East (1). The main disease features were: lymphadenopathy (93%), skin lesions (67%) and hepatosplenomegaly (23%). Peripheral blood counts showed lymphocytosis (7–120 × 109/l) with abnormal circulating lymphocytes in 11 cases (80%), anaemia (Hb < 10 g/dl) in three (20%) and thrombocytopenia (< 100 × 109/l) in two cases (13%). Liver and renal biochemistry showed hypercalcaemia in two thirds of the patients (9 out of 15; 60%) and a raised lactic dehydrogenase (LDH) in the majority (median: 727 IU/ml; range: 172–2025 IU/ml). Bone marrow aspirates and trephine biopsies showed a variable degree of lymphoid infiltration in six (40%) patients. All patients had serum antibodies to HTLV-I detected using enzyme-linked immunosorbent assay (ELISA) and Western-blot analysis.

Diagnosis of the various forms of ATLL was based on the criteria of Shimoyama et al (1991). This classification considers the following forms: acute leukaemia ATLL (rapidly raising lymphocyte counts, high LDH, often hypercalcaemia and presence or not of organomegaly), chronic ATLL (lymphocytosis with or without organomegaly, LDH less than twice the upper normal limit, no hypercalcaemia), smouldering ATLL (normal WBC count with less than 5% atypical lymphocytes, skin lesions, often lung infiltrates, no hypercalcaemia and normal LDH) and lymphoma-type ATLL (organomegaly, high LDH, normal or raised calcium and no evidence of peripheral blood involvement). Accordingly, 13 patients (86%) had the leukaemia form, all but two having acute ATLL; the two latter had smouldering ATLL with evidence of progression; the remaining two cases (14%) had the lymphoma form of ATLL, one of whom manifested predominantly as cutaneous ATLL.

The disease status and therapy received prior to IFN-α and AZT was as follows: one patient received chlorambucil with no response and 11 patients responded to combination chemotherapy with anthracycline-containing regimens: CHOP (cyclophosphamide, vincristine, adriamycin and prednisolone), M-BACOD (methotrexate, bleomycin, adriamycin, cyclophosphamide, vincristine, dexamethasone), acute lymphoblastic leukaemia-type schedule and BEAM (carmustine, etoposide, cytarabine, melphalan), the latter followed by an autologous stem cell transplant. At the time of starting IFN-α and AZT, four of these patients had progressive disease, eight were in partial/complete remission and all three untreated patients had active disease; the two patients with smouldering ATLL had evidence of disease progression (Table I).

Table I.  Response rate and duration of response to IFN and AZT.
CaseATLL typePrior therapy and responseResponse to IFN and AZT (months)Follow-up (months)
  • *

    Smouldering and progressing to an acute type.

  • Patients switched to a combination of IFN and combivir (lamuvidine and zidovudine).

  • Still maintaining response.

Responders (10/15–67%)
B-4LeukaemiaNo914 – Dead
D−6LeukaemiaYes – PR913 – Alive
E.R-7.Leukaemia*Yes – PR510 – Alive
L-13LeukaemiaYes – CR1941 – Dead
M−15Leukaemia*No4040 – Alive
Mc-16LeukaemiaYes – PR610 – Dead
O-17LymphomaNo1114 – Alive
P−19LeukaemiaYes – PR4447 – Alive
W−22LeukaemiaYes – PR22 – Lost
M−23LeukaemiaYes – PR1982 – Alive
Non-responders (4/15–26%)
A-3LeukaemiaYes – CR 8 – Dead
H-9LymphomaYes – NR 20 – Dead
L-11LeukaemiaYes – CR 6 – Dead
L-12LeukaemiaYes – PR 4 – Dead
Non-evaluable (1/15, 7%)
A-2LeukaemiaYes – PR 3 – Dead

Cell morphology, histology and immunological markers Examination of peripheral blood films showed all 11 acute-type ATLL cases to have circulating atypical lymphocytes with a convoluted nucleus. Lymph node histology, available in five patients, showed diffuse infiltration by pleomorphic lymphocytes and immunohistochemistry was consistent with the diagnosis of pleomorphic T-cell lymphoma. Skin biopsy in three patients showed extensive dermal lymphoid infiltration.

A full immunophenotypic analysis was carried out in the 11 cases with the acute ATLL form. Cells from all these cases had a mature post-thymic phenotype (TdT–, CD2+, CD5+, CD4+, CD8–). Reactivity with the other monoclonal antibodies (mAbs) was as follows: CD3+ (90% cases); CD7+ (33%); CD25+ (89%); CD38+ (80%); HLA-Dr (29%). Immunohistochemistry in tissue sections from all lymphoma cases demonstrated a T-cell phenotype (CD45R0+, CD20–, CD3+/–).

Treatment and criteria for responses IFN was administered subcutaneously at a dose of 3 MU to 5 MU/d or every other day and AZT orally at a dose of 1000 mg/d. Dose escalation of IFN was considered if well tolerated and both drugs were stopped if severe neutropenia or thrombocytopenia developed. Patients received prophylaxis with cotrimoxazole 480 mg orally/d or 960 mg three times a week, paracetamol to alleviate the ‘flu’-like symptoms resulting from IFN, metoclopramide and broad-spectrum antibiotics, acyclovir and fluconazole as indicated.

Complete response (CR) was considered when all clinical and radiological features normalized, the LDH and calcium levels and peripheral blood counts became normal, and there was no evidence of abnormal circulating lymphocytes maintained at least for 2 months. Partial response (PR): reduction of 50% or more in all disease features and normal calcium levels maintained for at least 2 months. Minor response: as CR or PR but maintained for less than 2 months. Progressive disease: increase of 20% of disease features.

Statistical analysis The impact of organomegaly, age, performance status, ATLL forms (leukaemia versus lymphoma), LDH and hypercalcaemia in the response rate was evaluated using the Fisher exact test. Survival was estimated according to Kaplan and Meier and comparison between survival curves was evaluated.


Response rate and duration of responses (Table I)

Ten out of the 15 patients (67%) achieved or maintained a response with a median duration of 10 months (range: 2+ to 44+). In two patients (M-15; M-23), therapy was switched to IFN and combivir (lamuvidine 150 mg and AZT 300 mg) owing to the intolerance of the full dose of AZT, and another patient (P-19) is being maintained with IFN only, 3 MU three times a week. Five of these patients are still in PR on maintenance treatment, one has relapsed at 11 months, one was lost to follow-up at 2 months and three PR patients have died. Four patients (26%) did not respond and the remaining patient was not evaluable as she died from a brain haemorrhage 1 week after therapy was started (Table I).

Treatment was well tolerated with the main side-effects being: ‘flu’-like symptoms, sickness, low-grade fever, grade II neutropenia and, rarely, alopecia or infections. There were no differences in toxicity between untreated patients and those who received prior chemotherapy.

Relationship between response rate and disease features

There were no statistically significant differences in the response rate according to the presence of hepatomegaly, lymphadenopathy, skin lesions, performance status and LDH levels. The presence of hypercalcaemia predicted a poor outcome: 5 out of 10 (50%) PR had normal calcium levels versus 1 out of 4 NR (25%), but differences were not statistically significant.

Outcome and survival

At the time of this analysis, eight patients have died with progressive disease and these include the four NR and the one non-evaluable; six patients, all PR, are alive at 2–82 months from diagnosis. The median survival for all 15 patients was 18 months (Fig 1). Differences in survival were statistically significant comparing the PR (median survival not reached; 55% are alive at 4 years) versus the NR (median survival: 6 months) (Fig 2) (P = 0·002).

Figure 1.

Kaplan–Meier plot illustrating the median survival (18 months) of the 15 patients.

Figure 2.

Comparison between survival in NR (6 months) versus responders (median survival not reached at 4 years; 55% of patients alive) (P = 0·002).


Our findings have extended previous studies (Gill et al, 1995; Hermine et al, 1995) by showing that over two thirds of ATLL patients achieve and/or maintain a response to the combination of IFN and AZT. In our study, all responses qualified as PR but none were complete. However, in our study, responses were only considered as PR or CR when sustained for over 2 months compared with only 1 month in the earlier reports (Gill et al, 1995; Hermine et al, 1995).

There were no differences in the response rate according to whether the patients had or had not received previous chemotherapy. Indeed, 7 out of 10 (70%) responders were previously treated with standard chemotherapy. Similar results were seen in the study by Gill et al (1995) in which 4 out of 7 patients who had received prior chemotherapy benefited from the treatment, including two with CR. Thus, a reduction of the tumour burden with chemotherapy prior to starting IFN and AZT does not preclude achieving or maintaining a response with this combination. Other factors, e.g. degree of organomegaly and lymphocyte count, performance status and LDH levels, did not influence the response. There was a trend for the responders to have normal calcium levels but differences were not statistically significant. Because the majority of patients in the previous reports and in our series were cases with the leukaemia form of ATLL, it cannot be concluded that this schedule is also equally effective in the lymphomatous form of ATLL.

Treatment was well tolerated with no major toxicities. At the dose given here, flu-like symptoms developed during the first weeks of IFN treatment in most patients but subsided with symptomatic therapy. IFN was stopped in one patient owing to depression and alopecia. Among other side-effects, neutropenia, mainly induced by AZT, and subsequent infections were easily controlled with antibiotic therapy and rarely required hospitalization; only a few patients required intermittent use of growth factors, e.g. granulocyte colony-stimulating factor (G-CSF) to improve neutrophil counts. In two patients, the dose of AZT was reduced and both were switched to a combination of AZT and lamuvidine.

ATLL is a lymphoid malignancy with very poor survival ranging from 6 to 13 months in most of the series reported from Japan, the Caribbean and other endemic countries (Shimamoto et al, 1990a, b; Shimoyama et al, 1991; Plumelle et al, 1993; Pombo de Oliveira et al, 1995; Matutes & Catovsky, 1998). This was also apparent in the patients treated with IFN and AZT (Gill et al, 1995; Hermine et al, 1995). Thus, in one study the overall median survival was 3 months, for the responders 13 months and for those responders who had been previously treated with chemotherapy 17·5 months (Gill et al, 1995). The median survival of our 15 patients was 18 months, significantly longer than the 7 months median survival of a historical series of 52 ATLL patients treated in the UK between 1982 and 1995 (Matutes & Catovsky, 1994) and that of other series of patients. Furthermore, responders to IFN and AZT had a significantly better survival (median survival not reached) than the NR (6 months) and those previously treated with this combination (Gill et al, 1995). This might be related to the fact that our patients continued on maintenance treatment with IFN and AZT after achieving a PR, unlike those in the study of Gill et al (1995) in which treatment was stopped.

The mechanism(s) by which IFN and AZT induces and/or maintain responses is unknown. The possibility that AZT could directly act as an anti-retroviral agent has been entertained but this mechanism has not been yet demonstrated. It is known that healthy HTLV-I carriers as well as patients with HTLV-I-induced spastic myelopathy (TSP/HAM) have a strong cytotoxic T-cell (CTL) response against lymphocytes that express the tax protein of HTLV-I and are capable of eliminating such infected cells (Daenke et al, 1996). This is in contrast to findings in ATLL patients in whom the CTL response is poor. Therefore, there is a possibility that the mechanism of action of IFN and AZT is by enhancing the CTL response against the neoplastic and/or normal HTLV-I + lymphocytes.

Although the combination of IFN and AZT clearly improves the outcome of ATLL patients, a search for new therapies or modifications of previous schedules to cure this disease is needed. The fact that responses were mainly seen in patients with low tumour burden following chemotherapy, although with active disease, and that two cases are maintaining a PR with the combination of IFN plus combivir (AZT plus lamuvidine) has prompted us to start a new trial that first considers the use of combination chemotherapy (e.g. 2–6 courses of CHOP and/or until maximal response) to decrease the tumour load. Once the patient achieves PR, they will then receive consolidation and maintenance therapy with a combination of IFN 3–5 MU/d according to tolerance and combivir (lamuvidine 150 mg plus AZT 300 mg) twice daily. Studies investigating the CTL response as well as monitoring residual disease will be carried out in the new trial and hopefully this will shed some light on the mechanism of action of this combination.


We are grateful to Mr A'Hern for performing the statistical analysis.