RESULTS
We identified 1,563,899 pregnancy episodes in 1,321,312 females. Among females without a history of cancer, the cohort included 1,251,935 first singleton live births, 71,618 first multiple live births, and 1,212,710 linked newborns. Among 46,064 female cancer survivors, the cohort included 2,440 first singleton live births, 214 multiple live births, and 2,590 linked newborns (Figure S1). Among cancer survivors, the most common cancer types were thyroid (21.8%), melanoma (20.0%), and breast (16.7%) (Table 1). Compared to those without a history of cancer, cancer survivors were older at delivery, had more comorbidities before and during pregnancy and were more likely to undergo ART. Cancer survivors had higher incidence of both preterm birth and SMM (Table 1).
In singleton births, the incidence of preterm birth was 14.8% in cancer survivors compared to 12.4% in females without cancer; the incidence of SMM was 3.9% in cancer survivors compared to 2.4% in females without cancer. The most common SMM conditions were cardiopulmonary (1.9% in cancer survivors vs. 1.1% in females without cancer) and end organ injury (1.6% in cancer survivors vs 0.9% in females without cancer) (Figure S2). Unadjusted and adjusted relative risks of preterm birth and SMM are depicted in Table 2. Due to the collinearity of race and ethnicity, income, and education, the adjusted regression model with the best fit accounted used race/ethnicity as a covariate. Adjusting for maternal age at delivery, race/ethnicity, chemotherapy, radiation therapy, year of birth, and ART, AYA cancer was associated with 1.19-fold higher risk of preterm birth (95% CI 1.06-1.34) and 1.44-fold higher risk of SMM (95% CI 1.13-1.83) (Model 1, Table 2).
We conducted mediation analysis in order to evaluate maternal comorbidity as a potential mechanism to explain associations between AYA cancer and adverse outcomes. Nearly all comorbidities in the Maternal Comorbidity Index occurred more frequently in cancer survivors compared to females without a history of cancer in both singleton and multiple births (Figure 1). In models for preterm birth and SMM accounting for the Maternal Comorbidity Index as a putative mediator (Model 2, Table 2)., AYA cancer remained significantly associated with preterm birth (aRR 1.21, 95% CI 1.06-1.39) and SMM (aRR 1.40, 95% CI 1.09-1.79). Using this approach to estimate indirect and total effects, result of mediation analysis suggested that maternal comorbidities before and during pregnancy explain 26% of the association between AYA cancer and preterm birth and 30% of the association between AYA cancer and SMM.
In multiple births, the incidence of preterm birth was 38.8% in cancer survivors compared to 34.4% in females without cancer; the incidence of SMM was 8.9% in cancer survivors compared to 6.8% in females without cancer. Unadjusted, confounder adjusted (Model 1), and mediation model (Model 2) based relative risks of preterm birth and SMM are depicted in Table 3; no significant associations were observed in cancer survivors relative to the females without cancer.
In adjusted models in singleton births, ART was associated with 1.4-fold and 1.6-fold higher risks of preterm birth and SMM, respectively (Model 2, Table 2). In adjusted models of multiple births, ART was associated with a 2.6-fold and 2.0-fold higher risk of preterm birth and SMM respectively (Model 2, Table 3). The association between AYA cancer and preterm birth or AYA cancer and SMM was not modified by whether ART was undertaken (PTB p-interactionSingleton=0.12, p-interactionMultiple birth=0.13), (SMM p-interactionSingleton=0.11, p-interactionMultiple birth=0.87).
Two subgroup analyses were undertaken. Excluding gynecological cancers (ovarian, cervical, uterine, other reproductive cancers) because of known associations between these and preterm birth36, AYA cancer survivors with singletons still had increased risk of preterm birth (aRR 1.14; 95% CI, 1.01-1.29) compared to females without cancer. This association was not significant in multiple births (aRR 0.85, 95% CI, 0.63-1.14). In subgroup analysis restricted to births between 2010 and 2019, AYA cancer’s association with preterm birth and SMM and mediation by maternal comorbidity did not materially change (data not shown).