Our study provides new and comprehensive data on pregnancy outcomes in a specific and homogeneous population of female survivors treated for acute leukemia during childhood or adolescence. In our previous study from the LEA cohort, uterine volume measured by MRI was significantly reduced after HSCT regardless of the myeloablative conditioning regimen used [14]. This reduction in uterine volume after HSCT may be attributable to several mechanisms. Radiotherapy can induce myometrial fibrosis and impair uterine growth, leading to a smaller, fibrotic, and less compliant uterus. In addition, post-HSCT hypoestrogenism secondary to POI, particularly when occurring around puberty, may further compromise uterine development. Three other studies have also reported a uterine volume decrease following alkylating conditioning, urging the need for studying pregnancy and perinatal outcomes after exposure to high doses of alkylating agents [9, 14, 19]. Additionally, a secondary outcome of our previous study showed a reduction of the LBR in women treated with alkylating conditioning before HSCT [14]. This also highlighted the need for a detailed study focusing on the obstetric prognosis of women treated with alkylating agents.
Concerning TBI, our results with lower LBR, increased risk of pregnancy loss, prematurity, Cesarean sections, post-partum hemorrhage and low birth weight are in accordance with previous literature [7,8,9, 13, 20, 21]. A history of TBI is associated with poor pregnancy outcomes, with a significantly lower LBR due to an increase in fetal loss, including both first and second trimester pregnancy loss. In our study, the LBR after TBI was as low as 37%, compared to ~85% in the general population. Moreover, a history of TBI is associated with higher risks of obstetrical and perinatal complications. Notably, postpartum hemorrhage and other hemorrhagic events were much more frequent after TBI, occurring in approximately one-third of pregnancies, highlighting the particularly high-risk nature of these pregnancies. In cases of POI, even with well-managed hormone replacement therapy—such as that used for oocyte donation—uterine volume remains significantly smaller than in non-irradiated patients [9, 14, 19, 22, 23]. In the subgroup analysis among women exposed to TBI according to the mode of conception, the pregnancy outcome was similar after oocyte donation and pregnancies with autologous oocytes. This result enhances the hypothesis that the adverse pregnancy outcomes after TBI are mostly due to irradiation rather than the oocyte donation. In a study of 2654 survivors following adult HSCT, Sockel et al. demonstrated that pregnancy rates were reduced but remained possible, with 50 patients achieving pregnancy [24]. These chances appeared to be improved in the absence of radiotherapy or with low doses ( < 8 Gy), suggesting a dose-dependent effect on fertility. Regarding obstetric and neonatal complications, the findings of this study are consistent with those of our HSCT cohort. The authors report a live birth rate of 77% (57/74), 28% of cesarean section, 23% of prematurity and 14% of low birth weight. The authors also suggest the potential use of ovarian shielding to reduce radiotherapy-related complications.
We did not find statistically significant increased rates of preeclampsia and fetal growth restriction after a history of TBI when compared with women exposed only to chemotherapy. However, preeclampsia was more observed among pregnancies obtained after oocyte donation among HSCT cases when compared to autologous oocyte pregnancies. The preeclampsia risk is probably more linked to the oocyte donation rather than the TBI history. Indeed, the meta-analysis of Keukenset et al. reported that preeclampsia was more frequent after oocyte donation than natural conception or after assisted reproductive technologies with autologous oocytes [25].
Women with a previous TBI exposure experienced significantly higher rates of preterm birth. This is in line with a retrospective single-center cohort study of childhood cancer survivors treated for various malignancies, where 40 individuals who had a singleton live birth were compared to controls from the national perinatal registry [20]. In the subgroup treated with abdominal radiotherapy (n = 6, with a median total cumulative dose of radiation of 19.8 Gy (2.0–26.5 Gy), the number of preterm deliveries was significantly higher (mean term of delivery in survivors 34.9 versus 39.2 weeks in controls, p = 0.001).
We did not observe a significant difference in the risk of delivering a small-for-gestational-age newborn between the three groups. The neonatal weight centiles integrate gestational age and maternal BMI, and our results suggest that the low birth weight of newborns to mothers who received radiotherapy is primarily attributable to prematurity or low maternal BMI rather than to placental dysfunction. This is in accordance with previous reported literature [7, 9, 20].
After alkylating conditioning, LBR, obstetric and perinatal outcomes were comparable to those of female survivors exposed to conventional chemotherapy without HSCT. Our results suggest that high-dose alkylating conditioning may have no clinically significant impact on the gravid uterus, despite a diminished uterine volume observed pre-conceptionally in previous studies [9, 14, 19]. These reassuring results could be attributed to several factors. First, the reduction in uterine volume may not be substantial enough to impact LBR and obstetric or perinatal outcomes. In most HSCT cases with POI, hormone replacement therapy was initiated before conception, possibly improving uterine volume. Finally, the hyper-estrogenic hormonal environment of pregnancy itself might exert a protective effect, allowing the uterus to expand and reach a volume closer to the general population.
To our knowledge, our study is one of the few to specifically investigate obstetric and neonatal outcomes in a subgroup of patients who underwent HSCT following myeloablative alkylating treatment with multivariable regressions. Indeed, few studies have specifically focused on the impact of a history of chemotherapy on pregnancy outcomes, and some adverse outcomes are debated. Green et al. reported a decreased live birth rate (RR 0.52, 95% CI: 0.36–0.76) among various childhood cancer female survivors treated with chemotherapy [13]. Van der Loo et al. highlighted an increased risk of preterm birth with an RR of 4.4 (95% CI 1.02–17.32), though no significant differences were observed for the birth weight [9]. After childhood cancer, Signorello et al. did not find a significant difference for preterm birth, nor birth weight, even among patients in the highest tertile of alkylating-agent exposure [6]. Unfortunately, data on the corresponding cyclophosphamide equivalent dose for each tertile were unavailable. Haggar et al. reported more Cesarean sections in adolescent and young adult female cancer survivors and more infant resuscitations [26]. No difference was observed for preterm birth or birth weight. In 2021, a national perinatal survey collected extensive demographic, obstetric, and neonatal data on nearly 12,000 births from 453 maternity units in France [27]. In the global population of pregnant women, they reported 5% fetal growth restriction, 2% preeclampsia, 7% preterm birth, 5% low birth weight, 10% small-for-gestational-age, 21% Cesarean sections, and 12% postpartum hemorrhage. This national study suggests that obstetric and perinatal outcomes in female leukemia survivors treated without irradiation are similar to those of the general population.
Despite the reduced uterine volume previously observed after alkylator-based HSCT [14], our findings provide reassuring preliminary evidence that pregnancy outcomes after alkylators may resemble those observed after conventional chemotherapy. These reassuring findings after alkylators may be explained by several mechanisms. In Courbiere et al., uterine volume was significantly improved in patients with an adequate hormonal balance, either through HRT or preserved gonadotropic function, whereas no such improvement was observed after TBI [14]. Thus, after alkylating conditioning, uterine volume reduction may be less severe, partially reversible, or sufficiently compensated by adequate hormonal exposure. In addition, in most HSCT patients with POI, HRT was initiated before conception, which may have improved uterine development. Finally, the hyperestrogenic environment of pregnancy itself may promote uterine expansion, allowing the uterus to reach a functional capacity sufficient to support fetal development. Although outcomes after alkylating conditioning were not significantly different from those observed after conventional chemotherapy, these results should be interpreted with caution. The limited sample size in this subgroup may have reduced the statistical power to detect modest increases, and larger studies are needed to confirm these reassuring findings. Our HSCT group included both autologous and allogeneic transplant recipients, reflecting the diversity of transplant practices over the study period; however, the limited sample size precluded separate analyses according to transplant type.
Surprisingly, in our study, 17% of pregnancies after HSCT were unintended, with 6% induced abortion among TBI cases and 13% among alkylating conditioning cases. In a previous qualitative study involving young female leukemia survivors from the LEA cohort, Vergier et al. observed widespread misconceptions and a perceived lack of information regarding the residual ovarian function, the role of hormone replacement therapy, and their long-term reproductive function. Many survivors believed they were permanently sterile [28].
Several patients without contraception experienced unintended pregnancies following HSCT, even after total body irradiation (TBI), reflecting insufficient awareness among patients and even some healthcare providers regarding the persistence of residual reproductive potential after HSCT. Indeed, POI does not equate to sterility [29]. Spontaneous pregnancies may still occur in women with POI, including those requiring hormone replacement therapy for chemotherapy-induced amenorrhea. However, the reproductive window is markedly shortened, and the probability of spontaneous conception remains very low. To avoid unintended pregnancies and subsequent elective terminations of pregnancy, effective contraception is essential. The psychological impact of an induced abortion in a context where many of these women will later be unable to achieve their desired parenthood is likely to be significant. In a large cohort of 178 adult survivor women who had undergone HSCT for leukemia before puberty, Chabut et al. reported that 12% of them had at least one spontaneous pregnancy [30]. Approximately 25% of women with a POI may present resumption of ovarian function, and a cumulative incidence of spontaneous pregnancy of 3-10% have been reported [30, 31]. It is essential that every childhood cancer survivor receive personalized counseling during adolescence or early adulthood from a physician specialized in endocrine and reproductive health. For HSCT survivors, these consultations could facilitate hormone replacement therapy compliance, prescribing effective contraception instead of hormone replacement therapy and providing guidance on oocyte donation for future pregnancies while reassuring female leukemia survivors treated without TBI.
The large French LEA cohort enabled the study of a homogeneous population of acute leukemia survivors treated with standardized protocols. We also carefully accounted for potential confounders, including BMI, smoking status, pre-existing periconceptional diabetes, mode of conception, and maternal age at each pregnancy, which are independently associated with adverse pregnancy and obstetric outcomes and could otherwise introduce bias. Multivariable analyses were performed to reduce the impact of some of these biases in our analyses. While relapsed disease was more frequent in the HSCT group, relapse history did not influence the choice of conditioning regimen. Consequently, the better outcomes observed with alkylating-agent-based HSCT are unlikely to be explained by differences in relapse status.
A limitation of the study is that pregnancy and perinatal data were collected retrospectively, which may introduce recall bias or incomplete reporting. Although our findings support efforts to reduce TBI-related long-term reproductive toxicity, current evidence, particularly from the FORUM trial, still supports TBI-based conditioning for disease control in high-risk pediatric ALL [32]. Future studies should determine whether TBI can be safely omitted in selected patients after optimized pre-transplant therapy, including immunotherapy-based approaches.
In conclusion, our results suggest that alkylating conditioning may not adversely affect uterine function or pregnancy outcomes. These treatments appear less detrimental than previously assumed, despite earlier reports of reduced uterine volume following chemotherapy. Our findings are reassuring for female leukemia survivors wishing to conceive after high-dose alkylating agent exposure during childhood or adolescence. In contrast, pregnancies following TBI are at high risk for complications and should be managed in specialized, multidisciplinary settings within tertiary care centers. Particular attention should be given to fetal growth, increased risk of prematurity, and postpartum hemorrhage. Given the high rate of unintended pregnancies among women who often believe they are sterile, every cancer survivor should receive individualized counseling from a physician specialized in endocrine and reproductive medicine. Significant inter-individual variability remains in the occurrence and outcomes of pregnancies after HSCT. International collaborations will be needed to improve our understanding of the factors—including genetic ones—that contribute to long-term toxicity affecting fertility and pregnancy outcomes.
