Planned Pregnancy in Female Patients with Chronic Myeloid Leukemia Receiving Tyrosine Kinase Inhibitor Therapy

Women with chronic myeloid leukemia who want to have children are advised to discontinue tyrosine kinase inhibitor (TKI) therapy before conceiving; however, the duration of time off TKI cannot be predicted. A lengthy waiting period could result in disease recurrence or progression. This article addresses some unanswered questions about the best response that should be achieved by a patient before stopping TKI therapy, as well as possible adverse effects caused by short exposure to TKI on the developing fetus.

Abstract

Background.

The aim of this study was to explore outcomes of planned pregnancy in female patients with chronic myeloid leukemia (CML) on tyrosine kinase inhibitors (TKIs).

Materials and Methods.

Data of female patients proceeding with a planned pregnancy were retrospectively reviewed.

Results.

A total of 17 patients with CML who achieved at least a major molecular response (MMR) during imatinib (n = 13) or nilotinib (n = 4) therapy prior to a planned pregnancy were enrolled. At the time of TKI interruption, six were in MMR, two in molecular response 4 (MR⁴), and nine in molecular response 4.5 (MR^4.5). TKI therapy was discontinued 6 weeks (range, 2–15 weeks) before conception in 4 patients and at gestational age of 4 weeks (range, 2–5 weeks) after determination of pregnancy in 13 patients. Apart from 1 patient who suffered a spontaneous abortion, 16 patients delivered uneventfully. A total of 10 patients lost MMR after stopping TKIs; 8 lost molecular response 2, and 3 lost complete hematological response. Log‐rank analyses showed achieving MR⁴ (p = .030) or MR^4.5 (p = .031), complete cytogenetic response duration ≥3.5 years (p = .049), and MMR duration ≥3.5 years (p = .040) were significantly associated with longer MMR‐failure‐free survival during TKI interruption.

Conclusion.

Planned pregnancy might be pragmatic in female patients with CML on TKIs. Achieving deep molecular response and, importantly, MMR duration ≥3.5 years were significantly associated with maintaining MMR during pregnancy.

Implications for Practice.

Female patients with chronic myeloid leukemia on tyrosine kinase inhibitors (TKIs) wishing to conceive are currently advised to discontinue TKIs before conception. However, the ideal degree and duration of response before stopping TKI, in addition to whether there will be any adverse effect caused by a short exposure of TKI, is unknown. Data of 17 female patients, who achieved at least a major molecular response (MMR) before TKI interruption, was revised, and it was found that achieving deep molecular response and MMR duration ≥3.5 years was significantly associated with maintaining MMR during pregnancy. This provides direct evidence for a planned pregnancy strategy, and stopping TKI immediately after determination of pregnancy in female patients might be pragmatic.

Introduction

An almost normal life span has been achieved in the era of tyrosine kinase inhibitor (TKI) therapy in patients with chronic myeloid leukemia (CML), and improving quality of life has been one of the major treatment goals [1], [2]. The median age at diagnosis of CML is 60–65 years in Western countries; however, the age at presentation is 10–15 years younger in Asian counties [3], [4], [5], although it remains unclear whether this is because of the shorter life expectancy in Asia than that in Western countries. It is estimated that approximately 30%–50% of patients with CML are within childbearing age [6]; therefore, fertility and family planning of those patients has become an important issue.

According to National Comprehensive Cancer Network recommendations and experts’ opinions, female patients with CML who wish to conceive are currently advised to discontinue TKIs before natural conception as well as during pregnancy because of the recognized teratogenic effects of the medicine [1], [7], [8], [9]; however, the duration of time off TKIs cannot be predicted, and a lengthy waiting period might result in disease recurrence or progression. Several questions remain unanswered. First, the ideal degree and duration of response a female patient should achieve before stopping TKI therapy for pregnancy is not known. Second, it is unclear whether there will be any adverse effect caused by a short exposure of TKI on the developing fetus, and concrete infant outcomes for female patients who discontinued TKI therapy once pregnancy was confirmed have yet to be determined.

To address these questions, we retrospectively reviewed data of planned pregnancies in 17 female patients who achieved at least a major molecular response (MMR) and interrupted TKI therapy before or immediately after pregnancy was confirmed under close monitoring.

Materials and Methods

Patients and Monitoring

Clinical data were retrospectively reviewed for patients with CML diagnosed, treated, and serially monitored at Peking University People's Hospital from January 2008 to February 2018. The outcomes of female patients achieving at least MMR during TKI treatment who consulted on pregnancy were reviewed. Diagnosis, disease phase, and monitoring were based on European LeukemiaNet (ELN) recommendations [10]. Cytogenetic responses were assessed every 3–6 months until a complete cytogenetic response (CCyR) was achieved using bone marrow by the G‐banding technique. Molecular responses were assessed every 3 months until an MMR was achieved and then assessed every 3 or 6 months using peripheral blood. Real‐time quantitative polymerase chain reaction was used to measure BCR‐ABL transcript levels, and ABL was used as the control gene, as previous reported [11]. The raw BCR‐ABL transcript values were converted to international scales (BCR‐ABL^IS) with our laboratory‐specific conversion factor of 0.65, which was validated through sample exchange with the Institute of Medical and Veterinary Science international reference laboratory in Adelaide, Australia [11].

All female patients were fully instructed of the possible adverse effects of TKIs on pregnancy and infant outcomes, such as spontaneous abortion and teratogenicity. Patients were also informed of unknown risks for disease progression during the interruption of TKIs. Female patients who wished to get pregnant were advised to follow the recommendations: achieve at least an MMR, discontinue TKI therapy before conception or interrupt TKI therapy immediately after pregnancy was confirmed, and comply closely with disease and fertility monitoring with gynecology counseling. Apart from routine prenatal tests, a complete blood count, a peripheral blood smear, and a molecular response were monitored every 1–2 months after TKI interruption. All patients were advised to resume TKIs after delivery and thus avoid breastfeeding. This study was approved by the Ethics Committee of Peking University People's Hospital, and all patients gave written informed consent to participate in the study in accordance with the Declaration of Helsinki.

Definitions

Molecular responses of MMR and deep molecular responses (DMRs), including molecular response 4 (MR⁴) and molecular response 4.5 (MR^4.5), were defined according to ELN criteria [10]. Molecular response 2 (MR²) was used to refer to BCR‐ABL^IS transcript levels ≤1%. As mild leukocytosis and immature myelocytes in peripheral blood can be induced normally by physiologic stress during pregnancy [12], loss of complete hematological response (CHR) was defined as only when ELN criteria were fulfilled (loss any of the following criteria: white blood cells [WBC] <10 × 10⁹/L, basophils <5%, no myelocytes, promyelocytes, myeloblasts in the differential, platelet <450 × 10⁹/L, and spleen nonpalpable) and BCR‐ABL^IS transcript levels >10%. MMR‐, MR²‐ or CHR‐failure‐free survival was used to assess the maintenance of molecular or hematologic response, defined as the interval from TKI therapy interruption to loss of MMR, MR², or CHR, respectively.

Statistical Analysis

Descriptive statistics were used to summarize the characteristics at presentation of all patients. Median values were used as a cutoff to dichotomize continuous variables. The log‐rank test was used to assess statistical significance in the time‐to‐event analyses. Factors with p < .05 were interpreted as significant. SPSS 21.0 software (SPSS, Chicago, IL) was used for the analysis.

Results

Patients’ Characteristics

Data of 17 newly diagnosed female patients with CML including 16 (94.1%) in chronic phase and 1 in the accelerated phase were collected. Median age at diagnosis was 26 years (range, 19–30 years). All patients received imatinib (Gleevec, Novartis, Basel, Switzerland, n = 16) or nilotinib (Tasigna, Novartis, Basel, Switzerland, n = 1) as a first‐line treatment. Before achieving CCyR, three patients (18.8%) switched from imatinib to nilotinib because of imatinib failure (n = 1, Sokal high risk) at 12 months or intolerance (n = 2, one with Sokal high risk and one with low risk) at 1 and 6 months, respectively. The patient with imatinib failure had no ABL kinase domain mutation. Since the initation of TKI treatment, all patients achieved CCyR and MMR at a median of 6 months (range, 3–15 months) and 9 months (range, 3–40 months), respectively; 11 (64.7%), MR⁴ at 18 months (range, 6–60 months); and 9, MR^4.5 at 19 months (range, 6–48 months). Before TKI interruption, median TKI therapy and MMR durations were 49 months (range, 6–102 months) and 41 months (range, 1–78 months), respectively; median MR⁴ and MR^4.5 durations in those achieving DMR were 36 months (range, 2–65 months) and 31 months (range, 13–65 months), respectively. Patients’ characteristics are shown in Table 1.

Table 1. Patients’ characteristics and pregnancy and fetal outcomes.

^aTKI was stopped before conception.

Abbreviations: —, not available; AP, accelerated phase, CP, chronic phase; MMR, major molecular response; N, no; TKI, tyrosine kinase inhibitor; Y, yes.

Monitoring During TKI Interruption

A total of 4 patients conceived after a median of 6 weeks (range, 2–15 weeks) of TKI interruption, and 13 stopped TKI therapy immediately once pregnancy was confirmed at a median gestational age of 4 weeks (range, 2–5 weeks). Median age at pregnancy was 30 years (range, 25–35 years). Ten patients lost MMR at a median of 3 months (range, 1–9 months) after TKI interruption, and 8 of them subsequently lost MR² at 4 months (range, 2–7 months), among whom 3 lost CHR at 3 months (range, 3–6 months).

Four patients developed physiological leukocytosis with a median WBC count elevation of 14 (12–17) × 10⁹/L and maintained a median BCR‐ABL level of 1.4% (0.81%–5.4%) and normal platelet counts. In the other three patients who lost CHR during TKI interruption based on the definition in current study, WBC count reached 157 × 10⁹/L at 37 gestational weeks and leukapheresis was performed in one patient; the highest WBC counts were 13 × 10⁹/L and 24 × 10⁹/L, respectively, in the other two patients with a corresponding BCR‐ABL level of 34%–42%; platelet counts were all within normal ranges during the pregnancy period. No patient reinitiated TKIs after cessation or received interferon throughout the pregnancy, and none progressed to an advanced phase. For the six patients maintaining MMR throughout pregnancy, the optimal response before TKI interruption was MMR for one patient and MR^4.5 for five patients. Among the latter, two patients remained in MR^4.5 during TKI interruption: one had BCR‐ABL level rising to MR^4.0 at 7 months, and two had an MMR, both at 3 months.

Aside from 1 patient stopping nilotinib at gestational age of 4 weeks and reinitiating nilotinib therapy at 8 weeks because of spontaneous abortion, the median off‐TKI duration in the remaining 16 patients was 9 (8–20) months, among which the estimated rate of maintenance of MMR, MR², and CHR at 9 months after TKI interruption was 35.0%, 50.0%, and 81.3%, respectively (Fig. 1).

Figure 1.

Response maintenance after tyrosine kinase inhibitor interruption.

Abbreviations: CHR, complete hematological response; MMR, major molecular response; MR², molecular response 2.

Factors Associated with Response Maintenance During TKI Interruption

Variables, including age at diagnosis, Sokal score at diagnosis, age at pregnancy, interval from TKI therapy to achieving CCyR or MMR, duration of CCyR, MMR, TKI‐therapy, achieving DMR or not before TKI interruption, and maintaining an MMR or not at 2 months after TKI interruption, were analyzed to identify the factors associated with the maintenance of molecular responses and CHR during TKI interruption. Univariate analysis showed that achievement of MR⁴ (p = .030, p = .009, and p = .014) and MR^4.5 (p = .031, p = .005, and p = .033) were the common factors significantly associated with longer MMR‐, MR²‐, and CHR‐failure‐free survivals during TKI interruption. In addition, CCyR (p = .049) or MMR (p = .040) duration ≥3.5 years was significantly associated with MMR‐failure‐free survival; TKI therapy duration ≥4 years (p = .048) and maintaining an MMR at 2 months after TKI interruption (p < .001) were associated with longer MR²‐failure‐free survival (Table 2). However, the Sokal score (low risk vs. others; p = .014) was inversely associated with MMR‐failure‐free survival.

Table 2. Log‐rank test of factors associated with response maintenance after tyrosine kinase inhibitor interruption.

Abbreviations: CCyR, complete cytogenetic response; CHR, complete hematological response; MMR, major molecular response; MR², molecular response 2; MR⁴, molecular response 4; MMR^4.5, molecular response 4.5; TKI, tyrosine kinase inhibitor.

Maternal and Fetal Outcomes

Other than 1 patient who underwent a spontaneous abortion with an unknown fetal development, 16 patients delivered 16 babies (11 boys and 5 girls) at full term (≥37 gestational weeks) without any events, and 7 of them were by cesarean section. None of the patients took any concomitant medication during pregnancy. A total of three patients maintaining MMR before delivery breastfed for 4, 5, and 6 months, among whom one lost MMR at 4 months after delivery. One patient who lost MMR before delivery breastfed for 1.3 months, and one patient who lost CHR breastfed for 10 days.

Apgar scores for all infants at 1, 5, and 10 minutes after birth were ≥ 9. One infant whose mother stopped imatinib at a gestational age of 4 weeks was born with congenital intestinal atresia type I and fully recovered after corrective surgery. During a median follow‐up of 19 months (range, 7–78 months) after delivery, all 16 babies have continued normal growth and development.

Response After TKI Resumption

The patient with spontaneous abortion reinitiated TKI therapy immediately after this event and subsequently maintained MMR. For the other 16 patients with successful delivery, 15 reinitiated original TKI therapy at a median of 16 days (range, 3–133 days) after delivery. All of the 10 patients who lost MMR during TKI interruption regained MMR at a median of 4 months (range, 2–7 months) after resuming TKI therapy. One patient remained treatment free at 6 months after delivery while consistently maintaining MR^4.5.

Discussion

We retrospectively reviewed the outcomes of planned pregnancy in 17 female patients with CML achieving at least MMR, including 4 with TKI therapy that was interrupted before conception and 13 stopping TKI within 5 weeks of gestational age after pregnancy was confirmed. Apart from 1 patient who underwent a spontaneous abortion during the first trimester, 16 patients proceeded to a successful delivery while off TKIs, among whom 10 lost MMR, 8 lost MR², and 3 lost CHR. No patient reinitiated TKIs after cessation or received interferon throughout the pregnancy, and none progressed to advanced phase. Achievement of DMR before TKI interruption was significantly associated with maintenance of MMR, MR², and CHR during pregnancy. In addition, CCyR or MMR duration ≥3.5 years were significantly associated with MMR‐failure‐free survival; TKI therapy ≥4 years and maintenance of MMR at 2 months after TKI interruption were associated with MR²‐failure‐free survival. MMR was regained in all patients after TKI resumption. All of the 16 babies had normal growth and development during the follow‐up period.

Pregnancy in patients with CML remains a major concern worldwide. Since Pye et al. first reported the outcomes of 180 female patients exposed to imatinib during pregnancy [13], more evidence has accumulated to indicate the potential embryotoxic and teratogenic effect of TKIs including imatinib [14], nilotinib [15], [16], [17], and dasatinib [9] during pregnancy. Currently, for female patients with CML wishing to consider pregnancy, it is accepted that at least an MMR should be achieved prior to conception, and patients should remain off TKI during pregnancy, especially during the crucial first trimester [1], [6], [7], [14], [18], [19], [20], [21], [22], [23], [24]. As to the specific degree and duration of the response that should be achieved before planned pregnancy, there is no consensus. Apperley recommended that, ideally, 24 months in MMR should be achieved before discontinuing imatinib in 2009 [25], and later, Abruzzese et al. made a similar suggestion [14]. More recently, data of stopping TKI trials [1], [6], [26], [27], [28], [29], [30], [31], [32], [33], [34] indicated that achievement of DMR ≥2 years and a TKI therapy duration of 3–8 years before TKI interruption allowed for a successful treatment‐free remission (TFR) in just over half of the patients. In these TFR studies, the relapsing patients experienced molecular recurrence at 3–9 months of stopping therapy, meaning that pregnant women could safely remain off TKIs at least during the first trimester. However, it would be unrealistic for all female patients to meet these stringent criteria before conception. Several studies reported the adverse effect of imatinib on ovarian function [35], [36]. In our previous study, the incidence of hypomenorrhea or amenorrhea in women under 50 years of age was approximately 30% [37]. Therefore, long‐term TKI use might result in missing the optimum childbearing age and therefore failure of a pregnancy attempt. Our findings provided data for the safety of a planned pregnancy strategy in female patients achieving an MMR ≥3.5 years and TKI therapy ≥4 years.

Maintenance of MR², correlating with CCyR, and CHR could be an acceptable disease status in pregnant women, because TKIs would be resumed soon after delivery. It should be noted that the definition of CHR based on ELN criteria [38] could not be applied to pregnant women, as physiologic stress during pregnancy usually induces mild leukocytosis and myeloblasts in peripheral blood [39]. In our study, four patients developed a slight increase in WBC count with BCR‐ABL^IS level below 10%, which did not necessarily indicate loss of CHR.

Another question that remained to be addressed was the potential genotoxicity, embryotoxicity, and teratogenicity due to the off‐targets effects of all TKIs [40]. Preclinical animal studies showed that imatinib and nilotinib had no effect on the fertility of rats [41], [42], whereas for dasatinib, the fertility of female monkeys and rodents could be adversely affected [43]. For embryotoxicity, when referring to postimplantation loss, animal studies showed that all TKIs could result in embryo loss in a dose‐based manner, with imatinib and nilotinib at higher and dasatinib at lower dose [41], [42], [43]. Imatinib and dasatinib were shown to be teratogenic in animal studies, whereas nilotinib did not have similar effects [41], [42], [43]. Previous human studies reported that about 14% of the female patients who were exposed to imatinib during the first trimester or organogenesis period experienced spontaneous abortion [13]. Miscarriage or infant abnormalities after exposure to nilotinib or dasatinib within the first trimester were also reported [15]. Based on the significant concerns on embryotoxicity and teratogenicity induced by TKIs, it was recommended that TKI therapy should be interrupted before conceiving and avoided during the whole pregnancy period in female patients [1], [44]. However, it was still hard to decide whether a shortest TKI exposure period (within 5 weeks in our study) could be safe. Ault et al. [45] reported that brief exposure to imatinib until the gestational age of 4 weeks (range, 4–9 weeks) did not result in adverse effects on the fetus in 10 females. Abruzzese et al. also suggested that imatinib and nilotinib could be stopped at implant (10–15 days after ovulation, i.e., 5–7 gestational weeks). In our study, 16 out of 17 female patients exposed to imatinib or nilotinib with shortest exposure of TKI (within 4–5 gestational weeks) experienced uneventful pregnancy. We did observe 1 (5.9%) of the 17 patients who stopped nilotinib immediately after pregnancy was confirmed and experienced a spontaneous abortion. It should be noted that the frequency of early pregnancy loss for women aged 20–30 years is 9%–17% in the general population [46]. Thus, it was difficult to entirely implicate the nilotinib exposure for the loss. There was one (5.9%) baby born with mild atresia under 4 gestational weeks of exposure to imatinib, and the incidence was higher than that in general population (1.3–3.5 per 10,000 live births [47], [48], [49], [50]). Although fully corrected by surgery, the potential teratogenicity of TKIs should still be a cause for concern.

It was recommended that female patients avoid breastfeeding when undertaking TKIs because they could be distributed into breast milk [24], [51], but breastfeeding could be recommended for the first 2–5 days postpartum to allow the baby colostrum for the benefits of the immune system, allowing for the level of CML disease [14]. In our study, five patients breastfed for 4 months (range, 0.3–6 months) and responded well after resuming TKIs. Again, it is suggested that a short period of breastfeeding might be applicable under close molecular monitoring for those with an antenatal stable disease state.

Our study had some limitations in the relatively small number of patients, different time point when stopping TKI therapy, and different monitoring frequency during TKI interruption. One unusual observation in our study was that patients with a low‐risk Sokal score tended to lose MMR, perhaps because of statistical limitation of small cases.

Conclusion

Planned pregnancy might be safe and pragmatic in female patients with CML who achieve at least an MMR on TKI therapy. Close fertility monitoring under the guidance of a gynecologist and stopping TKIs as soon as possible after conception could be an approach to help female patients with CML on imatinib or nilotinib therapy with planned pregnancy. Achievement of DMR, CCyR or MMR duration ≥3.5 years, TKI therapy ≥4 years before conception, or maintenance of MMR at 2 months after TKI interruption were significantly associated with maintaining stable disease during pregnancy. The shortest possible exposure to TKIs by TKI interruption within 4–5 weeks of gestational age seems to be relatively safe for fetal development. These findings need to be confirmed in further studies with a larger number of patients. In addition, more cooperation with gynecologists is necessary to explore the feasibility of planned pregnancy.

Author Contributions

Conception/design: Xuelin Dou, Qian Jiang

Provision of study material or patients: Xuelin Dou, Yazhen Qin, Qian Jiang

Collection and/or assembly of data: Xuelin Dou, Yazhen Qin

Data analysis and interpretation: Xuelin Dou, Qian Jiang

Manuscript writing: Xuelin Dou, Yazhen Qin, Xiaojun Huang, Qian Jiang

Final approval of manuscript: Xuelin Dou, Yazhen Qin, Xiaojun Huang, Qian Jiang

Disclosures

The authors indicated no financial relationships.