Keywords chronic phase chronic myeloid leukemia - early molecular response - imatinib - randomized
controlled trial - RCT - structured dose hydroxyurea
Introduction
Chronic myeloid leukemia (CML) is a well-defined myeloproliferative disease where
the clonal cells originate from pluripotent hematopoietic progenitor cell (stem cell)
that exhibits growth advantage over normal hematopoietic progenitors.[1 ] The clinical hallmark of CML is uncontrolled production of mature and maturing granulocytes,
predominantly neutrophils, basophils, and eosinophils. Tyrosine kinase inhibitors
(TKIs) like imatinib have become the standard therapy for CML-chronic phase (CP).[2 ]
[3 ] Though adverse reactions are reported with it, no significant interactions have
been found from available literature.[4 ] Imatinib has a high interindividual variability in the pharmacokinetic parameters.
This interindividual variability is due to factors such as genetic polymorphisms in
transporters and metabolizing enzymes, body age, gender, white blood cell count, and
hemoglobin (Hb) value.[5 ]
[6 ] Several biosimilars of imatinib have shown promising results.[7 ]
Hydroxyurea (HU), a nonalkylating antineoplastic drug is often used for initial cytoreduction
in suspected CML patients.[4 ] The role of upfront structured total leukocyte count (TLC)-based dosing of HU with
imatinib in diagnosed CML has not been explored till date.[8 ] Hence, this randomized controlled trial was designed to compare the efficacy and
safety of HU and imatinib combination versus imatinib monotherapy (IM) for upfront
treatment of newly diagnosed CML-CP patients in terms of achievement of early molecular
response (EMR).
Materials and Methods
An open-label randomized controlled trial was conducted at a tertiary care center
in northern India. Study duration was 18 months (January 2019 to June 2020). Newly
diagnosed patients of CML-CP, age ≥ 18 years, were included. Exclusion criteria included
de novo accelerated phase or blast phase CML, who had already started either HU or
any TKI treatment prior to randomization, pregnancy, and active secondary malignancy.
Written informed consent was taken from all patients before enrolment.
Detailed history and physical examination were performed. The baseline investigations
and determination of disease staging were done. Sokal,[9 ] Hasford,[10 ] and European Treatment and Outcome Study (EUTOS)[11 ] risk scores were calculated for prognostication. Eligible patients were randomized
equally in 1:1 ratio to either of the two treatment groups:
Dose of imatinib was 400 mg orally once daily in both groups.
Initial dose of HU in the I-HU group was based on TLC at baseline: < 100 × 109 /L: 1000 mg, 100–199 × 109 /L: 2000 mg, 200–399 × 109 /L: 3000 mg, ≥ 400 × 109 /L: 4000 mg per day. HU daily dose was rounded off to nearest 500 mg multiples and
was administered as divided doses. The maximum daily dose of HU did not exceed 35 mg/kg
body weight. HU dose was subsequently adjusted on each follow-up visit according to
TLC, absolute neutrophil count, and platelet count. HU was discontinued when TLC was
less than 10 × 109 /L, or if patient developed grade ≥ 3 hematological or nonhematological toxicity as
per the National Cancer Institute Common Terminology Criteria for Adverse Events version
5.0.[12 ]
Patients were followed up every 2 weeks in the outpatient department for clinical
evaluation and monitoring complete blood count (CBC). After completion of 3 months,
5 mL of ethylenediaminetetraacetic acid-anticoagulated blood sample was collected
from each patient, and samples were sent for CBC and examination of peripheral blood
smear for hematological response. Quantitative real-time polymerase chain reaction
(PCR) for BCR-ABL1 (international scale; I.S.) was performed for assessment of EMR at 3 months of therapy
in both groups. EMR was considered as BCR ABL1 (I.S.) levels ≤ 10%.
Sample size was determined to be 90 (45 patients in each group). Statistical analysis
was performed using STATA version-12 (StataCorp, United States). Descriptive statistics
was used to summarize baseline characteristics of patients. Normality was assessed
using the Shapiro–Wilk test. Median and interquartile ranges (IQRs) were calculated.
Wilcoxon rank sum test was used to compare quantitative variables and chi-square test
and Fisher's exact test were used to compare categorical variables in the two treatment
groups. A p -value of < 0.05 was considered as statistically significant.
Ethical issues : Ethical clearance was obtained from the Institutional Ethics Committee prior to
the commencement of the study. Written informed consent was obtained from patients
and their confidentiality was maintained. The trial was prospectively registered in
Clinical Trial Registry India (CTRI) database (CTRI/2019/08/020879).
Results
Newly diagnosed CML-CP patients (n = 123) were consecutively assessed for eligibility. Among them, 33 patients were
excluded (13 were in pediatric age group, 9 had already started TKI treatment prior
to first visit, 7 had accelerated phase/blast phase disease, 2 were pregnant, and
2 did not follow-up after the first visit). Hence, 90 patients were randomized into
two treatment groups: I-HU and IM, with each group consisting of 45 patients ([Fig. 1 ]).
Fig. 1 Consolidated Standards of Reporting Trials (CONSORT) diagram of the trial.
Median patient age was 36.5 (IQR: 30–46) years (I-HU: 36 [IQR: 30–45] years and IM:
38 [IQR: 31–47] years, p = 0.58). Fifty-nine percent (n = 53) of the study population were males. There was no statistically significant
difference in the gender composition of the two groups (p = 0.83).
The patients presented with symptoms ranging from 1 to 3 months in duration, the median
duration being 1.5 months. About 37% patients (n = 33) presented with at least two symptoms. The most common symptom (60%) among the
study population was fatigue followed by fever, abdominal heaviness, and weight loss
(p = 0.34). Most common presenting sign was splenomegaly (89%) followed by pallor (34%)
and hepatomegaly (27%). The median spleen size at presentation was 10 cm (6–15 cm)
below left costal margin (I-HU: 11 cm [8–16 cm], IM: 10 cm [6–15 cm], p = 0.15).
Median TLC was 192.5 × 109 /L (IQR 116–278 × 109 /L) (I-HU: 224 [136–355] vs. IM: 173 [164.1–225], p = 0.01). Thirty percent patients had TLC in the range of 100 to 199 × 109 /L. Median Hb level was 9.7 g/dL (IQR 8.4–11.2 g/dL), and median platelet count was
338 × 109 /L (IQR 199–460 × 109 /L). The median reticulocyte count was 2.8% (1.99–3.5%). Median peripheral blood blast%,
basophil%, and eosinophil% were 3 (2–5), 5 (3–7), and 3 (2–5), respectively. Median
bone marrow blast% was 2 (2–4) (2 [2–3] % in I-HU, 3 [2–5] % in IM, p = 0.14). The World Health Organization grade 2 myelofibrosis on baseline bone marrow
biopsy was seen in 48% patients (23 [51%] in I-HU, 20 [44%] in IM). Risk score assessment
is reported in [Table 1 ].
Table 1
CML risk scores of patients in study
CML risk scores
Imatinib + HU
(n = 45)
Imatinib
(n = 45)
p -Value
Sokal score
Low-risk (< 0.8)
Intermediate-risk (0.8–1.2)
High-risk (> 1.2)
13 (29%)
18 (40%)
14 (31%)
15 (33%)
20 (45%)
10 (22%)
0.39
Hasford (EURO) score
Low-risk (< 780)
Intermediate-risk (780–1480)
High-risk (> 1480)
21 (46%)
16 (36%)
8 (18%)
26 (57%)
16 (36%)
3 (7%)
0.09
EUTOS score
Low (< 87)
High (> 87)
27 (60%)
18 (40%)
30 (67%)
15 (33%)
0.51
Abbreviations: CML, chronic myeloid leukemia; EUTOS, European Treatment and Outcome
Study; HU, hydroxyurea.
Baseline cytogenetics of 57 (63%) patients showed the presence of characteristic t (9; 22) (q34; q11) translocation, 5 (6%) had variant Philadelphia chromosome, and
1 (1%) had additional cytogenetic abnormalities. Cytogenetic reports were not available
in 27 (30%) who were either not willing for bone marrow study, had a dry tap, or had
undetectable metaphases. Treatment for these patients was started after confirmation
of CML diagnosis by BCR-ABL PCR.
At the end of 3 months, complete hematological response (CHR) and EMR were assessed.
CHR was seen in 74 (82.2) patients and EMR (BCR-ABL1 ≤ 10%, I.S.) was seen in 68 (76%). Data was not available in 5 patients who were
lost to follow-up (3 patients in the I-HU group and 2 patients in the imatinib group).
There was no significant statistical difference between the two treatment groups ([Table 2 ]).
Table 2
Complete hematological response (CHR) and early molecular response (EMR) in the study
population
Variable
Imatinib + HU
(n = 45)
Imatinib
(n = 45)
p -Value
Hematological response at 3 months
CHR
36 (80%)
38 (84%)
0.71
No CHR
6 (13%)
5 (11%)
BCR-ABL (I.S.) at 3 months
≤ 10%
34 (76%)
34 (76%)
0.53
> 10%
8 (18%)
9 (20%)
0.28
Median (IQR)
5.2%
(1.24–9.06)
6.5%
(1.66–9.54)
0.96
Abbreviations: HU, hydroxyurea; IQR, interquartile range; I.S., international scale.
The most common hematological adverse drug reaction was anemia in both groups. The
most common nonhematological toxicity was nausea and vomiting. Treatment discontinuation
due to adverse drug reactions was not observed in either of the groups. There was
no statistically significant difference in adverse events in the two groups ([Table 3 ]).
Table 3
Adverse effects of treatment in study population
Adverse effects
Imatinib + HU
n (%)
Grade ≥ 3
n (%)
Imatinib
n (%)
Grade ≥ 3
n (%)
p -Value
Hematological toxicity
Anemia
41 (91)
2 (4)
39 (87)
2 (4)
0.5
Leukopenia
24 (53)
1 (2)
22 (49)
1 (2)
0.67
Neutropenia
14 (31)
1 (2)
10 (22)
0
0.34
Thrombocytopenia
29 (64)
2 (4)
22 (49)
1 (2)
0.14
Febrile neutropenia
2 (4)
2 (4)
1 (2)
1 (2)
0.55
Nonhematological toxicity
Edema
6 (13)
1 (2)
4 (9)
1 (2)
0.50
Hyperpigmentation
5 (11)
0
4 (9)
0
0.72
Hypopigmentation
7 (16)
0
6 (13)
0
0.76
Stomatitis/Mouth ulcer
3 (7)
1 (2)
2 (4)
0
0.64
Diarrhea
4 (9)
1 (2)
2 (4)
0
0.39
Nausea/Vomiting
20 (44)
1 (2)
17 (38)
1 (2)
0.52
Muscle cramps
15 (33)
2 (4)
13 (29)
0
0.64
Grade ≥ 3 treatment-related AE (n )
0.08
Abbreviations: AE, adverse effect; HU, hydroxyurea.
Discussion
In this study, 90 patients were recruited. Age of CML presentation was similar to
a study by Bansal et al[13 ] and Mishra et al.[14 ] This younger age of Indian CML patients has been the most consistent fact confirming
that patients present at least a decade younger when compared to the Europeans (median
age: 55 years)[15 ] and Americans (median age: 66 years).[16 ] It might be due to nutritional deficiencies, exposure to herbicides/pesticides,
heavy metals mining, and carcinogens because of geographical variations and occupational
reasons. These may have implications related to fertility and childbearing in female
patients and keeping “treatment-free remission” as one of the goals of treatment in
the newly diagnosed patients.[17 ] Asians' studies have reported younger onset as compared to west but these have smaller
sample sizes and focused mainly on the treatment issues rather than risk factors associated
with the disease.[18 ]
[19 ] Male-to-female ratio was 1.43:1, similar to a study by Mishra et al (1.6:1)[14 ] and Goni et al (1.2:1).[20 ]
Most common presenting symptom was fatigue (62%), followed by fever (48%) and abdominal
heaviness (48%). Only 6% of them were asymptomatic and incidentally diagnosed to have
CML during workup for asymptomatic leukocytosis similar to Mishra et al.[14 ] Splenomegaly was detected in 89% patients which was similar to findings by Mishra
et al.[14 ] On the contrary, in the western population, Hoffman et al[15 ] reported that more than half of the patients did not have palpable splenomegaly.
Pallor was seen in 34%, similar to a meta-analysis by Singh et al.[21 ] In comparison to western data[22 ] where approximately 40% were asymptomatic and diagnosed on the basis of abnormal
blood counts, majority of Indian patients were symptomatic and mostly presented with
dull aching abdominal pain secondary to splenomegaly.
In our study, baseline TLC was higher as compared to Lange et al[8 ] (median: 59.5 × 109 /L), suggestive of a higher need for cytoreduction. Hb level at CML-CP diagnosis was
9.7 g/dL, similar to the study by Bansal et al[13 ] (9–11 g/dL) and varied from a study by Lange et al[8 ] (12.7 g/dL). This difference was due to early presentation in western patients.
Disease risk scoring showed majority had Sokal intermediate-risk scores, Hasford low-risk
scores, and EUTOS low-risk scores at baseline. This is similar to Bansal et al[13 ] and Malhotra et al[23 ] where majority had Sokal intermediate-risk score of 27 to 47% and 38.17%, respectively.
However, a study by Mishra et al,[14 ] 40% patients had Sokal high-risk scores at baseline. In the study by Lange et al,[8 ] 56% had Hasford intermediate-risk score and 91% had EUTOS high-risk scores. This
could be ascribed to the older age of presentation in the Western population. CHR
rates in the two treatment groups were comparable (80% in the I-HU group vs. 84% in
the imatinib group, p = 0.71). In the study by Lange et al,[8 ] 86% patients attained CHR at the end of 3 months (89.6% patients in the imatinib
group and 82.6% patients in the I-HU group).
A recent East German Study Group (OSHO) study by Lange et al explored the efficacy
of in vitro and in vivo study of potential additive effect of HU with imatinib. CHR
at 3 months was seen in 82% in I-HU and 89% in the imatinib arm with nonsignificant
difference. Our study used a TLC-based structured HU regimen from the initiation of
therapy as compared to OSHO study where HU was gradually increased over time based
on patient response. Moreover, our patients were young age group, hence genetic and
regional variation in the response to therapy can also be considered.[24 ] EMR was achieved in 76% patients. Lange et al[8 ] reported that there was no statistically significant difference between the two
groups in terms of major molecular response (MMR) rates at 18 months whereas the OSHO
study reported that major cytogenetic and molecular response at 6 months was higher
in the IM group. MMR at 12 months were 58.9 versus 41.9% in combination. These studies
did not mention 3 months EMR which is a preliminary predictor of EMR at 6 months and
MMR.[24 ] Side effect profile of drugs was similar to Bansal et al.[13 ] Overall higher treatment-related adverse effects were seen among the patients in
the I-HU group. Similar findings were reported by OSHO study where grade 1 to 4 adverse
events did not differ in each group.[24 ]
Conclusion
Addition of HU to imatinib in the present study was not found to significantly improve
the hematological response or EMR at 3 months in our cohort of newly diagnosed CML-CP
patients. However, HU is still a useful adjunctive therapy for initial, rapid cytoreduction
in patients who present with very high leukocyte counts in combination with first-line
imatinib therapy. Further studies can be considered to evaluate long-term molecular
response with structured dosing of HU.
Trial Registration
Clinical Trial Registry India, CTRI/ 2019/08/020879, Registered August
26, 2019, http://ctri.nic.in/Clinicaltrials/showallp.php?mid1=33760&EncHid=&userName=C
TRI/2019/08/020879
