Clinically, the natural course of chronic myeloid leukemia is progressive; the patient usually experiences a relatively stable chronic phase lasting 3 years, followed by the accumulation of additional cytogenetic abnormalities, an accelerated phase, and, inevitably, blastic transformation and death. The median survival from diagnosis is 3 to 4 years [3]. Traditionally, the conventional chemotherapy drugs busulfan or hydroxyurea have been used to treat chronic myeloid leukemia. Although these agents control leukocyte counts and ameliorate symptoms at nontoxic doses, they clearly do not eliminate the Ph-positive clone and do not prolong survival. Intensive chemotherapy during the chronic phase has likewise been unsuccessful [4-6]. Although the proportion of Ph-positive cells can be reduced in some patients, few become completely Ph-negative, and the observed responses are brief. Furthermore, toxicity is increased, and no evidence has suggested improved survival. Over the past decade, allogeneic bone marrow transplantation has become an attractive therapeutic option for patients with chronic myeloid leukemia; it has a 50% long-term disease-free survival in patients who receive transplantation in the chronic phase [7, 8]. Unfortunately, fewer than 20% of patients are candidates for allogeneic transplantation because of age restrictions or the lack of a suitable donor. To date, autologous transplantation for patients with chronic myeloid leukemia has had minimal success [9, 10], although promising studies using new techniques to isolate early bcr-abl-negative progenitors and polymerase chain reaction for detection of residual leukemic cells are ongoing [11].

For now, however, recombinant human interferon- remains the treatment of choice for most patients. Since natural, partially purified interferon was first used in 1981 and since the more recent use of recombinant forms, interferon- has consistently been shown to induce complete hematologic remissions in 50% to 70% of patients in the chronic phase. In addition, 30% to 40% of these patients have karyotypic responses, and, in as many as 20%, all Ph-positive metaphases completely disappear. Despite significant toxicity, most investigators use high doses of interferon- (3 to 5 x 10⁶ U/m² body surface area daily) to treat chronic myeloid leukemia. Noting that much lower doses (as low as 2 x 10⁵ U/m² three times weekly) [12] consistently induced complete responses in hairy cell leukemia, we studied the efficacy of low-dose interferon- (2 x 10⁶ U/m² three times weekly) in chronic myeloid leukemia. We present our findings in 41 patients with chronic-phase Ph-positive chronic myeloid leukemia treated for a median of 39 months. Our results compare favorably with those of previously published studies in which higher doses were used but with a considerable reduction in both cost and toxicity.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Forty-one consecutive patients with chronic-phase Ph-positive chronic myeloid leukemia, both newly diagnosed and previously treated with hydroxyurea or busulfan, entered our study from September 1986 to August 1993 after giving written informed consent as approved by the Human Subjects Committee at the University of Colorado Health Sciences Center. Table 1 lists characteristics of these 41 patients and 122 historical controls who were followed in the same clinic from 1968 to 1987 and treated with conventional chemotherapy. Before patients were enrolled in the study, the diagnosis of Ph-positive chronic myeloid leukemia was confirmed by examination of bone marrow core biopsy specimen and aspirate for cellularity, morphology, and cytogenetics. We censored 6 patients from survival data at the time of the event: Four had received bone marrow transplants, and 2 died while in complete remission from causes unrelated to the leukemia. The median time from diagnosis to the initiation of interferon- therapy was 7 months; the median duration of therapy was 39 months. We included in our analysis all patients who entered this study and received interferon- for 6 months or longer. No patient who started interferon- therapy withdrew from the study at any time because of toxicity or other problems.

View this table: [in this window] [in a new window]   Table 1. Clinical Characteristics at Diagnosis of Patients with Chronic-Phase Chronic Myeloid Leukemia Who Received Low-Dose Interferon- and Historical Controls Who Received Conventional Chemotherapy*

Study Design

Recombinant human interferon-, provided by the Schering Corporation (Kenilworth, New Jersey) and Hoffman La Roche, Inc., was self-administered subcutaneously at a dose of 2 x 10⁶ U/m² daily for 28 days and three times weekly thereafter. All patients were evaluated monthly with a physical examination and a complete blood count for response and toxicity. For patients achieving a complete hematologic response (defined below), bone marrow core biopsy specimens and aspirates were obtained at least at 6 monthly intervals for evaluation of cytogenetic response. For patients who did not achieve a complete response, we added hydroxyurea to their drug regimen to control the leukocyte count or splenomegaly as necessary; however, all patients continued to receive interferon- therapy.

Interferon- Response Criteria

We objectively classified responses as previously described [13]. Briefly, a complete hematologic remission is defined as a sustained normal leukocyte count and differential without splenomegaly. Patients achieving a complete hematologic remission and a 5% to 100% suppression of the Ph chromosome on more than one occasion (a minimum of 20 metaphases were examined each time) were classified as karyotypic responders. We subclassified karyotypic responses as minor, major, or complete according to the criteria outlined by Talpaz and colleagues [14]. A minor response indicates a reduction in the proportion of Ph-positive metaphases to 35% to 95% of the total, a major response reflects a decrease in Ph-positive metaphases to 5% to 34% of the total, and a complete response indicates a total elimination of the Ph chromosome. A partial hematologic remission is defined as a 50% or greater decrease in the leukocyte count from pretreatment levels to 20 x 10⁹/L or less or as persistence of splenomegaly despite normalization of the leukocyte count. Patients who have less than a partial hematologic response are considered nonresponders. Although we eventually added hydroxyurea to control the leukocyte count or splenomegaly in all patients who achieved only a partial hematologic remission or no response while receiving only interferon-, the reported responses are those attained while the patient received interferon- only. A cytogenetic relapse is defined as a return to 100% Ph-chromosome positivity in a patient who had previously achieved a karyotypic response.

Anti-Interferon- Antibodies

We collected serum samples from patients who had not received interferon- for at least 48 hours and screened them for the presence of anti-interferon- antibodies using an enzyme immunoassay. We then analyzed serum samples from patients testing positive for antibodies by the enzyme immunoassay using a vesicular stomatitis virus bioassay to determine if the antibodies were neutralizing. Both assays were done as previously described [15, 16].

Statistical Analysis

We generated a Kaplan-Meier survival curve and survival statistics using the SAS PROC LIFETEST [17]. The difference in the median survival between patients treated with interferon- and the historical controls treated with hydroxyurea was analyzed using the log-rank procedure [18].

Results

Top Methods Results Discussion Author & Article Info References

Patient Characteristics

Table 1 shows the clinical characteristics at diagnosis of the 41 patients treated with interferon- from 1986 to the present and those of 122 historical controls treated with conventional chemotherapy in the same clinic from 1968 to 1987. The median age, sex, and hemoglobin levels at diagnosis were nearly identical in the two groups. Eighty percent of the historical controls and 70% of the patients treated with interferon- had splenomegaly at diagnosis, whereas the median leukocyte count in our patients was higher—167 x 10⁹/L compared with 101 x 10⁹/L for the historical controls. Likewise, the median platelet count at diagnosis was higher in the patients who received interferon- (420 x 10⁹/L) than in the historical controls who received hydroxyurea (318 x 10⁹/L). Overall, clinical and hematologic measurements were similar for the two patient groups at the time of diagnosis.

Survival

The Kaplan-Meier survival curves for the 41 patients treated with interferon- and for the 122 historical controls are shown in Figure 1. At a median follow-up of 52 months (for the patients receiving interferon-), the estimated median survival from diagnosis is 84 months. This is significantly longer (P < 0.001) than the actuarial median survival from diagnosis of 43 months for the patients receiving standard chemotherapy. Because 14 of the 41 patients treated with interferon- were diagnosed at least 1 year before they began receiving interferon-, the survival of all patients treated with interferon- is also shown from the start of interferon- therapy to account for any selection bias (Figure 1). The median survival from the start of interferon- therapy has not been reached; however, the estimated 5-year survival rate is 54%. This also is significantly longer (P = 0.007) than the estimated 5-year survival rate of 31% for the controls treated with hydroxyurea.

View larger version (47K): [in this window] [in a new window]   Figure 1. The Kaplan-Meier survival curves shown from the time of diagnosis (dotted line) and from the start of interferon therapy (broken line) for 41 patients with chronic-phase chronic myeloid leukemia treated with low-dose interferon- compared with the actuarial survival from the diagnosis of 122 control patients treated with hydroxyurea (solid line). IFN = interferon-.

Interferon- Response

Table 2 summarizes the diagnostic characteristics, duration of interferon- treatment and follow-up, and survival status of all patients categorized according to response to interferon-. Twenty-five of 41 patients (61%) have had a complete hematologic response. Eleven of these (27% of the total) have had a partial or complete cytogenetic response, whereas the remaining 14 patients who achieved a complete hematologic remission have had little or no suppression of the Ph-chromosome. Eight patients (19.5%) have had a partial hematologic response, and another 8 patients (19.5%) have had little or no objective response. We eventually added hydroxyurea to control the leukocyte count or splenomegaly in all patients in the latter two groups. In no patient, however, did the addition of hydroxyurea result in a karyotypic response. Further, although some patients achieved a complete hematologic response after hydroxyurea was added, they were not considered in this category for statistical purposes; reported responses were those attained while the patients received only interferon-.

View this table: [in this window] [in a new window]   Table 2. Clinical Characteristics, Survival, and Response to Low-dose Interferon- Therapy in Patients with Chronic Myeloid Leukemia*

We found no correlation between age, leukocyte count, hemoglobin level, platelet count, or splenomegaly at diagnosis and the subsequent clinical response to interferon- therapy. There was, however, a significant difference in the interval from diagnosis to the initiation of interferon- treatment between responsive patients (that is, those attaining a cytogenetic response or a complete or partial hematologic response) and nonresponsive patients; responsive patients began interferon- therapy a median of 4 months after diagnosis, and nonresponsive patients did not start therapy for a median of 37 months after diagnosis. Similarly, there is a considerable difference in the overall survival between responsive and nonresponsive patients. Twenty-six of 28 (93%) responsive patients are alive compared with only 2 of 7 (29%) nonresponsive patients. Although the actual survival rate of the nonresponders is significantly worse than that of the responders, this probably reflects, at least in part, their longer follow-up. Responsive patients were followed for a median of 51 months from diagnosis compared with a median of 76 months for the nonresponders.

Because of the highly significant correlation between the time from diagnosis until interferon- therapy was initiated and the subsequent clinical response, we directly compared patients receiving interferon- within 1 year of diagnosis (n = 27) with those treated later (n = 14). These data are shown in Table 3. The median time from diagnosis to the start of interferon- therapy was 3 months for the former group and 37 months for the latter. At the time of diagnosis, the clinical characteristics of patients in the two groups were similar, including age, sex, leukocyte count, hemoglobin level, platelet count, and presence of splenomegaly. There was, however, a significant difference in the response to interferon- between the two groups (Table 2). For patients treated within 1 year of diagnosis, 71% (19 of 27) achieved a complete hematologic remission with or without cytogenetic improvement compared with only 43% (6 of 14) of the group that did not start interferon- therapy until more than 1 year after diagnosis. Likewise, only 15% (4 of 27) of patients treated within 1 year failed to respond compared with 36% (5 of 14) of those treated later.

View this table: [in this window] [in a new window]   Table 3. Comparison of Clinical Characteristics and Response to Interferon- Therapy in Patients with CML Beginning Treatment within or after 1 Year of Initial Diagnosis*

Although the actual survival rate for the group starting interferon- therapy more than 1 year after diagnosis is 50% (6 of 12) compared with 96% (22 of 23) of the group treated within 1 year, the former group has been followed for much longer (83 months compared with 38 months). The Kaplan-Meier survival curves for each group are given in Figure 2 to show the influence of this difference on survival. To account for any selection bias that may exist in the group of patients who had been diagnosed for more than 1 year when they began receiving interferon-, survival rates are shown from the start of interferon- therapy rather than from the time of diagnosis. The median survival has not been reached for the group who started interferon- therapy within 1 year of diagnosis. However, the estimated 5-year survival rate for this group is 73% (95% CI, 51% to 95%) compared with only 37% (CI, 9% to 65%) for the group treated later.

View larger version (20K): [in this window] [in a new window]   Figure 2. The Kaplan-Meier survival curves for 27 patients with chronic myeloid leukemia treated with interferon- within 1 year of diagnosis (broken line) compared with that of 14 patients who started interferon therapy more than 1 year after diagnosis (dotted line). Survival is shown from the start of interferon therapy. IFN = interferon-.

Cytogenetic Responses

The response characteristics of the 11 patients achieving a karyotypic remission are given in Table 4. The median time to the first documented cytogenetic response was 6 months; 1 patient responded as rapidly as 3 months, but another required 15 months for an initial response. This patient eventually achieved and remains in a complete karyotypic remission. In 2 other patients, Ph-positive cells have also been completely eliminated. Three patients thus had complete cytogenetic remissions. In addition, 5 patients have had a major karyotypic response, and 3 have had a minor karyotypic response. Of these 11 cytogenetic responders, only 2 have relapsed, and both have tested positive for neutralizing anti-interferon- antibodies. The median duration of response is 21 months for all cytogenetic responders (n = 11) and 46 months for the nonrelapsing patients (n = 9).

Patient Follow-up

The follow-up for all 41 patients categorized by response is shown in Table 5. Twenty-seven of the 41 patients (66%) remain on study. Although we have added hydroxyurea to control the leukocyte count or splenomegaly in all partial responders and nonresponders, these patients have continued to receive interferon- and thus remain on protocol. Of the 14 patients no longer on study, 7 died of blast crisis, 2 died while in complete hematologic and karyotypic remission, respectively, of causes unrelated to chronic myeloid leukemia, 4 received bone marrow transplantation, and 1 nonresponsive patient was removed when an in vitro analysis showed interferon- to have a significant stimulatory effect on his bone marrow. No patient withdrew from the study because of interferon- toxicity, and no patient has requested to be removed because of intolerable side effects.

View this table: [in this window] [in a new window]   Table 5. Patient Follow-up Correlated with Response to Interferon- Treatment*

Low-Dose Compared with High-Dose Interferon-

In Table 6, the objective responses obtained in our study are listed with those that have been reported by other investigators who have treated patients with long-term therapy at higher interferon- doses (3 to 5 x 10⁶ U/m² daily). The studies of Talpaz and colleagues [19], Ozer and colleagues [20], Kloke and colleagues [21], and the Italian Cooperative Study Group on Chronic Myeloid Leukemia [22] have all enrolled only untreated or minimally pretreated patients. Noting that the interval from diagnosis to interferon- initiation is a key determinant of subsequent clinical response, we compared only the 27 patients in our series who began receiving interferon- within 1 year of diagnosis. This provides a more accurate assessment of the efficacy of low-dose interferon- compared with higher-dose interferon- therapy. Table 6 also shows patient characteristics at diagnosis for the various series. Overall, the patients are similar in age, sex ratios, and characteristics at diagnosis. In all series, the median leukocyte count ranged from 126 to 186 x 10⁹/L, the median platelet count ranged from 390 to 448 x 10⁹/L, and the percentage of patients with splenomegaly ranged from 56% to 80%. The interferon- dosing schedules varied from as low as 2 x 10⁶ U/m² three times weekly in our study to as high as 5 x 10⁶ U/m² daily. The investigators for all the studies followed patients who received long-term therapy with interferon-; the median duration of treatment for all studies was similar (range, 30 to 38 months). The estimated median survival rate from the start of interferon- therapy and the percentage of patients achieving a karyotypic response, a complete hematologic response, a partial hematologic response, and no response for each study are also shown in Table 6. The complete hematologic response values include those of patients achieving a karyotypic response and those of patients entering a complete hematologic remission without cytogenetic improvement.

View this table: [in this window] [in a new window]   Table 6. Clinical Measurements, Interferon- Dose, and Response to Treatment by Various Investigators Who Have Treated Chronic Myeloid Leukemia with Long-Term Interferon-*

The response rates reported by the different investigators are remarkably similar regardless of differences in interferon- dosage. We report an overall complete hematologic response rate (with or without cytogenetic improvement) of 70%, which is identical to that of Talpaz and colleagues [19]; Ozer and colleagues [20] report a similar response rate of 59%. Kloke and colleagues [21] and the Italian Cooperative Study Group [22] report only cytogenetic responses. There is variability in the overall cytogenetic response rates reported by the various investigators. Most variability, however, was in the number of minor cytogenetic responses; the major and complete response rates were similar, ranging from 16% to 29% (22% in our study).

For the 11 cytogenetic responders in our study, the median time to the first documented Ph suppression was 6 months. Seventy percent of karyotypic responders in the Italian Cooperative Study had achieved some degree of cytogenetic improvement by 8 months, and both Ozer and colleagues and Talpaz and colleagues reported a median time to cytogenetic response of 9 months. Thus, the time to cytogenetic improvement was similar in all studies. Of the 11 cytogenetic responders in our study, only 2 have relapsed. For the 9 nonrelapsing patients (82% of the total cytogenetic responders), the median duration of response is 46 months from the time of the first documented Ph suppression. In the Italian Cooperative Study Group series, the response of 41 of 63 (65%) karyotypic responders with adequate follow-up data was maintained or improved, but the response disappeared in 22 (35%). For Kloke and colleagues and Talpaz and colleagues, only 41% (19 of 46) and 35% (7 of 20) of their total karyotypic responders, respectively, remained in cytogenetic remission at the time of their reports. However, Kloke and colleagues included only patients who achieved major or complete cytogenetic responses, and 83% (15 of 18) remained in remission, which is similar to the results of our study, in which 88% (7 of 8) of major and complete cytogenetic responders have not relapsed. In Talpaz and colleagues' series, 63% of patients (5 of 8) achieving major or complete karyotypic responses did not relapse. Ozer and colleagues do not report the cytogenetic response duration or relapse rate. Only 15% to 20% of patients failed to respond in each study. The estimated median survival from the start of interferon- therapy for the 27 minimally pretreated patients in our study has not been reached. The estimated 3-year survival rate, however, is 79% (CI, 61% to 97%), which is similar to the estimated 3-year survival rate of 76% reported by Talpaz and colleagues. The estimated 5-year survival rate from the start of interferon- therapy in our series is 73% (CI, 51% to 95%), which compares favorably with the projected 5-year survival rate (from diagnosis) of 55% reported by Kloke and colleagues. Ozer and colleagues and the Italian Cooperative Study Group reported estimated median survivals from the start of interferon- therapy of 66 months and 72 months, respectively.

Although the response rates and survival rates are similar in the different series regardless of the interferon- dose used, the estimated yearly cost differs considerably. On the basis of the current cost of $9.54 per million units at our institutional pharmacy, the estimated yearly cost for a patient with a calculated body surface area of 2 m² for the dose used in our study is $5953. The cost of therapy at doses used by other investigators ranged from $10 446 to $34 821 (median, $24 375).

Toxicity

The toxicities (World Health Organization grades 1 to 4) encountered in our study are shown in Table 7. The interferon- dose administered in our study was well tolerated, and, with the exception of alopecia, no patient experienced grade 3 or 4 toxicity. No patient withdrew from the study because of toxicity, and no patient asked to be removed because of intolerable side effects. Most patients experienced mild fatigue (93%) and occasional mild myalgias or arthralgias (56%), but all of our patients can carry out their daily activities; in fact, most continue to work full-time. Fewer patients report mild loss of appetite (40%), intermittent fever with a body temperature of less than 40 °C (25%), headaches (36%), and anxiety or depression (18%). Twenty-four percent of patients have experienced mild to moderate alopecia, and two patients have had almost total hair loss, although this has tended to wax and wane in severity. No patient has experienced cardiac or pulmonary toxicity, and hematopoietic toxicity has been limited to grade I leukopenia, anemia, or thrombocytopenia in only a few patients (2% to 7%).

View this table: [in this window] [in a new window]   Table 7. Toxicities Experienced by 41 Patients with Chronic Myeloid Leukemia Who Were Treated with Low-Dose Interferon-*

In our study, every patient has been able to continue long-term therapy at the prescribed dose. By contrast, investigators using higher interferon- doses consistently report substantial toxicity, most commonly hematopoietic or neurologic, and treatment must be discontinued or the dose reduced in many patients. Ozer and colleagues [20] reported that 79% of their patients experienced either grade 3 or 4 toxicity and that a major dose reduction of 50% or greater was required at some point during therapy in 38% of patients. Similarly, Kloke and colleagues [21] reported that 16 of 62 (26%) patients in their study no longer received treatment because of toxicity or patient request, whereas 16% of patients in the Italian Cooperative Study Group series were withdrawn because of toxicity, an additional 7% asked to be removed, and one patient died of causes secondary to interferon- toxicity. Finally, Alimena and colleagues [23], who have treated patients with different dosing schedules, reported that the incidence and severity of side effects was directly related to the interferon- dose that patients received.

Discussion

Top Methods Results Discussion Author & Article Info References

It is now well established that interferon- induces complete remissions in a significant percentage of patients with chronic myeloid leukemia who are treated in the chronic phase and, further, that it is the first agent shown to induce durable cytogenetic responses in this disease [13, 14, 19-24]. Moreover, it has recently been established in a randomized controlled trial that interferon- indeed prolongs survival in chronic-phase chronic myeloid leukemia compared with conventional chemotherapy [22]. Nevertheless, most investigators have used doses of interferon- that result in considerable toxicity, perhaps because early reports by Alimena and colleagues [23] and Freund and colleagues [25] have suggested that lower doses are inferior in this disease. Careful review of these analyses, however, suggests that factors other than dose are probably responsible for the poorer response rates reported in these studies.

Alimena and colleagues [23] randomly assigned patients to receive interferon- at a dose of either 2 x 10⁶ U/m² three times weekly or 5 x 10⁶ U/m² three times weekly. They reported that only 24% (8 of 33) of patients receiving the lower dose achieved a complete hematologic remission compared with 47% (14 of 30) of patients treated with the higher dose. However, patients assigned to receive the lower dose were crossed over to the higher dose after only 4 weeks if the leukocyte count increased or after 8 weeks if the leukocyte count remained unchanged. It is now clear, however, that complete remissions often take considerably longer than 4 to 8 weeks to occur—as long as 15 months for complete hematologic responses and as long as 5 years for major cytogenetic responses [14, 20, 22]. In fact, Alimena and colleagues themselves report a median time to complete hematologic response of 14 to 18 weeks, which is significantly longer than the 4 to 8 weeks patients were given to respond to the lower dose before being crossed over to the higher dose. Thus, their results may have been biased in favor of the higher dose. Further, their overall median duration of treatment was only 13 months, and the investigators included in their analysis patients treated for as short a duration as 1 month. Investigators reporting higher response rates [19-21], have treated patients for as long as 7 years, allowing those who are slower to respond to be scored as responsive rather than nonresponsive.

Freund and colleagues [25] treated 27 patients with chronic myeloid leukemia with an interferon- dose of 5 x 10⁶ U/m² three times weekly. Only 10 patients (37%) achieved a complete hematologic remission, with 5 of these (19% of total) achieving minor cytogenetic responses. No patients had major or complete cytogenetic responses. They compared these results with those of other investigators who used daily dosing schedules and concluded that lower doses (administered intermittently) are inferior to higher doses (administered daily). Complicating this interpretation, however, is the fact that 17 of their 27 patients (63%) were pretreated, with a median interval from diagnosis to initiation of interferon- therapy of 26 months. Our data for patients who began treatment with interferon- within 1 year of diagnosis and for those starting later show substantially fewer complete responders in the latter group (70% compared with 43%). The complete hematologic response rate of 43% for our significantly pretreated patients is similar to Freund and colleagues' complete hematologic response rate of 37%. Likewise, the nonresponder rates in these two groups are nearly identical at 36% and 41%, respectively. Others have also reported that the interval from diagnosis to the initiation of interferon- therapy is a critical determinant of subsequent clinical response [24]. Moreover, the median duration of interferon- therapy in the study by Freund and colleagues was only 12.6 months, which probably further negatively biases their results.

On the basis of these considerations, the conclusion that lower interferon- doses are inferior to higher doses in the treatment of chronic myeloid leukemia is speculative. Given the significant toxicities that occur with higher doses and the considerable increase in expense, the issue of whether low-dose interferon- is efficacious in treating this leukemia is not a trivial one. Doses of interferon- as high as 12 x 10 (⁶) U/m² daily were initially used to treat hairy cell leukemia [12, 26]. Subsequently, however, doses as low as 2 x 10⁵ U/m² three times weekly have been shown to similarly improve the peripheral neutrophil and platelet counts as rapidly as do higher doses, but with considerably less toxicity [12]. Although the mechanism of action of interferon- in these two diseases is unknown, it is reasonable to consider that doses lower than the commonly prescribed 3 to 5 x 10⁶ U/m² daily may be effective in treating chronic myeloid leukemia. The Italian Cooperative Study Group on Chronic Myeloid Leukemia [22] has recently reported the results of the first prospective, controlled, comparative study showing that interferon- induces more karyotypic responses, delays progression, and improves survival compared with conventional chemotherapy. These investigators specifically note, however, that their study does not address the relation of interferon- dose to the response rate, the durability of response, or patient survival but do point out that these issues are probably important both to the cost of treatment and to the quality of life.

We have treated 41 consecutive patients with newly diagnosed or previously treated chronic-phase, Ph-positive chronic myeloid leukemia with low-dose interferon- (2 x 10⁶ U/m², three times weekly) for a median of 39 months. Consistent with reports of others, we found the interval from diagnosis to the initiation of interferon- therapy to be an important determinant of subsequent response. A direct comparison of patients treated within 1 year of diagnosis with those beginning interferon- later showed that although the clinical characteristics of the two groups were similar at the time of diagnosis, the patients receiving interferon- earlier had significantly better response rates. Thus, to adequately assess the efficacy of low-dose interferon- in chronic myeloid leukemia, we directly compared the response rates and survival of our 27 patients treated within 1 year of diagnosis with those previously described by other investigators who have treated similar patients with higher interferon- doses. As shown in Table 6, we noted no significant differences in response or survival. The major and complete cytogenetic response rates, the time to induction, and the durability of these responses are nearly identical. Further, the estimated median survival in our series compares favorably with the survivals reported by other investigators. Indeed, the only significant difference found between our series and those of others was in the cost and the extent and magnitude of interferon--related toxicity. It would appear that, on the basis of the present and previous data, future investigators of interferon- therapy in chronic myeloid leukemia must therefore find methods other than simple dose escalation to enhance the activity of this agent in clinically nonresponsive patients. Whether this will be, or should be, the addition of other therapeutic agents, as well as which agents should be used, remains to be determined. Most important in this regard, however, is the future understanding of the mechanism of action of interferon- in this leukemia. For the moment, this remains unclear.

This is the first report of the results of long-term, low-dose interferon- therapy in chronic myeloid leukemia. We conclude that low-dose interferon- given over long periods of time to minimally pretreated patients is as effective as higher-dose interferon- therapy, which has been administered by other investigators under similar conditions, in inducing remissions and prolonging survival in patients with this leukemia. Finally, these results have been achieved with a considerable reduction in both cost and patient toxicity.