Most randomized clinical trials (RCT) evaluating 2nd generation tyrosine kinase inhibitors (TKIs) for the treatment of chronic myeloid leukemia (CML) used as comparator the 'standard' dose of 400 mg of imatinib daily (IMA 400). When major molecular remission (MMR) after 12 months of therapy is used as primary endpoint to measure short term outcome this might however not be the correct comparator. Several studies showed that treatment with a dose of 800mg of imatinib per day (IMA 800) achieves higher rates of MMR at 12 months compared to IMA 400. The objective of our work is to systematically review randomized trials comparing IMA 400 and IMA 800, to calculate a pooled estimator and to compare the result to a pooled estimator of trials evaluating a 2nd generation TKI in comparison to IMA 400.
Patients and Methods
We searched Medline for RCTs of IMA 400, IMA 800 and 2nd generation TKIs. We extracted data on study designs, treatments, patient numbers and characteristics, statistical analysis methods and MMR at 12 months. Relative risks and respective 95% confidence limits were calculated and pooled using a random effects model. Forest plots were printed to visualize the results. The Jadad score was used to assess the quality of the trials.
We identified three RCTs [1,2,3] comparing patients treated with IMA 400 and IMA 800 and reporting on MMR at 12 months. Overall 1284 patients were included in the trial arms of interest - 554 were randomized into the IMA 400 arm and 730 into the IMA 800 arm. One trial used the Sokal and two used the Euro prognostic score as a stratum for randomization to allocate patients with different prognosis equally into the treatment arms. Percentage of Euro score high/intermediate risk patients in the IMA 800 arms was 14% / 49% in  and 49% / 30% in  and in the IMA 400 arms 12%/53% in  and 49%/31% in . The percentage of Sokal score high/intermediate risk patients in  was 27%/34% in the IMA 400 arm and 23% / 35% in the IMA 800 arm. Median age of the patients in the IMA 400 / IMA 800 arms in [1,2,3] was 54/52, 45/48 and 50/52 years, respectively. Relative chance of being in MMR after 12 months of treatment in [1,2,3] was 1.78 (1.47 to 2.15), 1.15 (0.91 to 1.45) and 1.47 (1.01 to 2.14), respectively. The pooled estimate over all three studies was 1.45 (1.10 to 1.92) and thus indicates a significantly higher rate of MMR after 12 months by treatment with IMA 800 (Fig. 1, p = 0.0088).
We identified four RCTs comparing IMA 400 to a TKI of the 2nd generation [4,5,6,7]. In  an IMA 400 arm was compared to treatment with two doses of 300mg nilotinib daily. Trials  and  each compared an IMA 400 arm with 100 mg/d of dasatinib, while  compared IMA 400 to 500mg of bosutinib daily. The percentage of Sokal high/intermediate risk patients in  and  was 28%/36% and 18%/ 47% in both the 2nd generation and the IMA 400 trial arm. Relative chance for MMR at 12 month for the respective 2nd generation TKI treatment arm compared to the IMA 400 arm was 2.00 (1.55 to 2.59) for , 1.64 (1.24 to 2.17) for , 1.56 (1.22 to 2.00) for  and 1.34 (1.05 to 1.75) for . The pooled estimate for the relative chance was calculated as 1.61 (1.37 to 1.91) (Fig. 2, p<0.0001).
All trials were rated with three out of five possible points on the Jadad scale; all were missing two points as none of the trials were blinded.
However all trials were properly randomized and drop outs were well documented.
The pooled estimate indicates a 45% higher probability of achieving MMR after 12 month with IMA 800 compared to IMA 400. This difference is statistically significant. Consequently, when choosing MMR at 12 months as primary endpoint, new therapies need to be compared with the IMA 800 regimen and not with the inferior IMA 400 therapy - as it is commonly done.
The pooled efficacy estimates of the IMA 400 / IMA 800 and the IMA 400 / 2nd generation TKI trials cannot be compared directly. But given that the prognostic profiles of the patients of the different trials are fairly similar we can conclude that the MMR rates achieved with IMA 800 might be comparable to the ones achieved with 2nd generation TKIs as the confidence limits show a considerable overlap.
 Hehlmann et al. 2011, JCO 29(12):1634-42.
 Baccarani et al. 2014, IJH 99:616-24.
 Deininger et al. 2011, BJH 164:223-32.
 Saglio et al. 2010, NEJM 362:2251-59.
 Kantarjian et al. 2010, NEJM 362:2260-70.
 Cortes et al. 2012, JCO 30(28):3486-92.
 Radich et al. 2012, Blood 120(19):3898-905.
Hoffmann:Novartis Oncology Europe: Research Funding. Hasford:Novartis: Research Funding. Hehlmann:Novartis Pharma: Research Funding; BMS: Consultancy.
Asterisk with author names denotes non-ASH members.