Share this article on:

Obstetric Outcomes in Cancer Survivors

Clark, H MSc1; Kurinczuk, J J. MD2; Lee, A J. PhD3; Bhattacharya, S MD1

doi: 10.1097/01.AOG.0000284458.53303.1c
Original Articles

OBJECTIVE: To assess obstetric and neonatal outcomes in women with a prior episode of cancer.

METHODS: Data were obtained from a linkage between the Scottish Cancer Registry and routinely collected data from Scottish maternity hospitals. Obstetric outcomes in a first pregnancy which ended between 1980 and 2005 were compared in 917 women with, and 5,496 women without, a previous history of cancer.

RESULTS: The mean age at delivery was 29 years (standard deviation 5.66) and 26 years (standard deviation 5.62) in the exposed and unexposed groups respectively (P<.001). Multiple logistic regression showed that cancer survivors had higher rates of postpartum hemorrhage (odds ratio [OR] 1.56, 95% confidence interval [CI] 1.09–2.23) and operative or assisted delivery (abdominal or vaginal) (OR 1.33, 95% CI 1.14– 1.54). Preterm delivery (at less than 37 weeks of gesation) was also found to be higher in this group compared with non-cancer women (OR 1.33, 95% CI 1.01–1.76).

CONCLUSION: While largely reassuring to women intending to become pregnant after surviving cancer, the results indicate areas of increased risk that require additional surveillance.


The risks of operative delivery, preterm delivery, and postpartum hemorrhage are increased in women with a history of cancer.

From the 1Department of Obstetrics and Gynaecology, University of Aberdeen; 2National Perinatal Epidemiology Unit, University of Oxford; and 3Department of General Practice and Primary Care, University of Aberdeen, United Kingdom.

This study was funded by the National Health Service (Grampian) Research Endowment Fund project number 06/42.

J.J.K. was partially funded by a National Public Health Career Scientist Award from the Department of Health and NHS R&D (PHCS02) and by a grant to the National Perinatal Epidemiology Unit from the Department of Health.

The authors thank the staff of ISD (Edinburgh) for their work on the record linkage.

Correspondence author: Heather Clark, Dugald Baird Centre, Aberdeen Maternity Hospital, Cornhill Road, Aberdeen, AB25 2ZL; e-mail:

Financial Disclosure The authors have no potential conflicts of interest to disclose.

The long-term survival rate for patients with cancer in childhood, adolescence, and young adulthood has improved substantially and now stands at over 75%. This is largely attributable to better methods of diagnosis and treatment.1 The major types of cancers occurring in women before middle age are of the skin, breast, and reproductive organs.2 Improved survival and greater access to fertility treatment now means that pregnancy and childbirth are real options for women with a history of cancer. This in turn demands greater awareness of the potential effects of malignancies and their treatments on reproductive outcomes. To date most research in this area has addressed the likelihood of fertility after cancer rather than obstetric and neonatal outcomes once pregnancy has been established.

With improved survival rates, it is likely that the number of cancer survivors facing pregnancy in the future will continue to increase. There is thus a growing need to understand any potential complications surrounding pregnancy and neonatal outcomes. By using national data from routinely collected maternity hospital admissions and cancer registrations, the aim of this analysis was to assess the risks of adverse obstetric and perinatal outcomes in women who have survived cancer.

Back to Top | Article Outline


This retrospective cohort analysis used data derived from a data linkage between the Scottish Cancer Registry, which holds data on all cancers diagnosed in Scotland regardless of place of care, and the Scottish Morbidity Record, which holds information on all women discharged from maternity hospitals in Scotland.3 Data from the Scottish Cancer Registry has been judged to be of high accuracy, whereas the maternity Scottish Morbidity Records are subject to regular quality assurance checks and have been more than 99% complete since the late 1970s.4,5 The exposed cohort included all women in Scotland with a diagnosis of cancer followed by a first pregnancy in the period 1980 to 2005. The comparison (unexposed) cohort comprised the remaining women whose first pregnancies ended between 1980 and 2005 and whose details were not present in the Scottish Cancer Registry. The linked data were extracted as an anonymized dataset by the National Health Service Scotland Information and Statistics Division. We anticipated that, with a minimum of 258 women exposed to a history of cancer and 1,032 (ratio of four to one) in the unexposed group, the analysis would have 90% power at the 5% level of statistical significance to show a doubling (relative risk 2.0) of preterm deliveries from 8% to 16% and cesarean deliveries from 13% to 26%.

Data for the following maternal variables were available: spontaneous fetal loss; legal termination of pregnancy; ectopic pregnancy; antepartum hemorrhage including placenta previa, placental abruption (separation of normally situated placenta), and antepartum hemorrhage of unknown origin; preeclampsia (blood pressure higher than 140/90 with proteinuria); and preterm premature rupture of membranes (rupture of membranes at less than 37 weeks of gestation). Labor and delivery characteristics included induction of labor, instrumental delivery by forceps or vacuum extraction, malpresentation or position at term, cesarean delivery (elective or emergency), and postpartum hemorrhage. Perinatal outcomes were preterm delivery before 37 and 34 completed weeks, birth weight of less than 2,500 g, Apgar score at 5 minutes, admission to a neonatal intensive care unit, stillbirth (antepartum or intrapartum death), and early neonatal death (death within the first 7 days after delivery).

The analysis was carried out in SPSS 14.0 (SPSS Inc., Chicago, IL). As the unexposed cohort numbered more than half a million, small differences between the exposed and unexposed women which are not clinically important would nevertheless be statistically significant. To avoid this effect of spurious statistical significance, we randomly sampled 1% of the unexposed cohort for inclusion in the main analysis. Frequencies of categorical variables were compared using the χ2 test, and crude odds ratios were calculated. Student t test was used to compare the means of the two groups for normally distributed continuous variables, and the Mann–Whitney test was used to compare medians of skewed data. Possible confounders identified by univariable analysis were adjusted for by means of multiple logistic regression and results presented as adjusted odds ratios with 95% confidence intervals (CIs). Due to the long observational period and improvements in treatment over time, the interactions of exposure and time period of delivery were examined for the statistically significant outcomes. No interaction term was found to be statistically significant.

Number needed to harm was calculated using the formula: ([PEER×{OR–1}]+1)/(PEER×[OR–1]×[1–PEER]) where PEER was the patient’s expected event rate and OR was the unadjusted odds ratio.6

This study was performed on an anonymized data set generated from the Information and Statistics Division and did not involve any direct contact with patients. The Information and Statistics Division is part of National Health Services Scotland. Linkage of the Cancer and Maternity registers was performed by Information and Statistics Division staff who generated an anonymized dataset. Permission to access the data were granted by the Information and Statistics Division Privacy Advisory Committee. As a consequence, the Grampian Research Ethics Committee did not feel that the project was one which required to be ethically reviewed under the Governance Arrangements for Research Ethics Committees in the United Kingdom.

Back to Top | Article Outline


A total of 756,752 women whose first pregnancies ended between 1980 and 2005 were identified. Of these, 1,327 were cancer survivors. Deliveries followed for 1,122 (85%) women with, and 578,780 (77%) without, a history of cancer.

After excluding women who had a multiple birth or delivered within 10 months of their cancer diagnosis (to exclude women whose cancer was diagnosed during pregnancy), 917 women with previous cancer and 568,802 women without were identified (Fig. 1). A 1% random sample of the unexposed cohort was drawn and the data for the resulting sample of 5,496 women were used in all subsequent analyses.



The descriptive characteristics of the two groups are given in Table 1. The mean age at which cancer was diagnosed in the exposed cohort was 24 years (standard deviation [SD] 7.33, range 0 to 43 years). Forty-six percent of women were diagnosed in the 1980s, 45% in the 1990s, and 9% between 2000 and 2005. Malignant neoplasms of the skin accounted for 33% of cancers diagnosed, lymphoid or hematopoietic neoplasms 19%, malignant neoplasms of the genital or urinary organs 16%, and breast cancer 9%. Table 2 details the breakdown of cancers by age (younger than 16 years compared with 16 years or older). The median interval between diagnosis and delivery was 4 years and 2 months (interquartile range [IQR] 2 years, 3 months to 7 years, 3 months).

Table 1

Table 1

Table 2

Table 2

Comparing pregnancy outcomes in the exposed and unexposed group, similar rates of miscarriage (22 [1.7%] compared with 16,878 [2.2%]) were recorded in both groups while ectopic pregnancies were more frequent in cancer survivors (38 [2.9%] compared with 6,534 [0.9%], P<.001). More legal termination of pregnancies were performed in the unexposed group (153,233 [20.3%]) compared with the exposed group (145 [10.9%], (P<.001)).

The mean age at delivery was 29 years (SD 5.66) and 26 years (SD 5.62) in the exposed and unexposed groups, respectively (P<.001). More women in the cancer group delivered between 1995 and 2004 (55%) compared with the comparison cohort—most of whom delivered between 1980 and 1989 (54%). The median birth weight and median gestational age of neonates born to the exposed group were 3,360 g (IQR 3,000–3,711 g) and 40 weeks (IQR 23–43 weeks) and in the unexposed group were 3,330 g (IQR 3,000–3,650 g) and 40 weeks (IQR 24–43 weeks) (P=.080 and P=.459, respectively).

Obstetric and perinatal complications in both groups of women are shown in Table 3. The unadjusted odds of preterm premature rupture of membranes, preterm delivery of less than 37 and 34 weeks of gestation, instrumental vaginal delivery (forceps and vacuum extraction), cesarean delivery, and operative delivery overall (abdominal or vaginal) were significantly higher in women with a history of cancer. Rates of induction of labor were not significantly different between the two groups. Of those women who delivered by cesarean, there was no significant difference in the proportion having an elective as compared with an emergency cesarean delivery between the two groups (OR 1.20, 95% CI 0.88–1.65). After adjustment for maternal age and year of delivery, the risk of postpartum hemorrhage, preterm delivery of less than 37 weeks of gestation, instrumental vaginal delivery, and operative delivery (abdominal or vaginal) were each significantly higher in women who had survived cancer. The association with delivery by cesarean was no longer statistically significant; the apparent excess risk being entirely explained by the effects of maternal age and year of delivery.

Table 3

Table 3

When simultaneous adjustment was made for all statistically significant confounders and outcomes (maternal age, year of delivery, preterm delivery of less than 37 weeks of gestation, postpartum hemorrhage, operative delivery), postpartum hemorrhage (OR 1.45, 95% CI 1.03–2.12, P=.042) and operative delivery (OR 1.26, 95% CI 1.08–1.47, P=.003) remained significantly higher in the cancer group. Preterm delivery of less than 37 weeks of gestation was no longer statistically significantly different between the two groups (OR 1.30, 95% CI 0.99–1.72, P=.063).

Numbers needed to harm were calculated for those outcomes where the (maternal age and year of delivery) adjusted odds ratios demonstrated statistically significant differences between the exposed and unexposed groups. This calculation estimates that, for every six women with cancer, one additional woman is delivered operatively; every 18 women with cancer results in an additional instrumental vaginal delivery; every 146 results in an extra case of postpartum hemorrhage; and every 46 with cancer leads to one extra preterm delivery.

Back to Top | Article Outline


This study suggests that women delivering their first pregnancies after surviving cancer are older and have a 56% higher risk of postpartum hemorrhage, a 33% higher risk of operative delivery (abdominal or vaginal), and a 33% higher risk of preterm delivery of less than 37 weeks of gestation than their peers.

This study is strengthened by its use of national, routinely collected data which allowed us to generate population-based figures. We are, however, limited by the paucity of data about the precise nature of cancer treatment which precludes subgroup analysis by mode of therapy (surgery, chemotherapy, and radiotherapy) and the small numbers in each group which preclude subgroup analysis by diagnosis. We were unable to adjust for confounders such as smoking, as the data were not available in either of the routine datasets. We were also unable to explore fertility in all survivors of cancer as details of fertility treatment including in vitro fertilization and gamete donation are not routinely available.

All malignant neoplasms have been included in this analysis resulting in a profile appropriate to this age distribution, which includes female survivors of cancer occurring not just in childhood, but up to the end of the reproductive years. While this has resulted in a relatively large proportion of skin cancers, we did not feel it was appropriate to eliminate any specific groups, as we did not have details of treatment on any group to allow us to be selective.

Practice both in oncology and obstetrics has changed substantially over the last 2 decades. Given that the two comparison groups we compared differed substantially in terms of their year of delivery and knowing that the frequency of both pregnancy complications and outcomes have changed over the 25 years covered by this study, we adjusted for year of delivery to alleviate the confounding associated with these temporal effects. In addition, for those outcomes which were shown to be statistically significant, we found no evidence of interaction between exposure and year of delivery.

There are a limited number of studies of pregnancy outcomes among cancer survivors, and many are based on small numbers of women. Female survivors of Wilm’s tumor treated with abdominal irradiation7–10 have been found to be at risk of a variety of reproductive problems, including miscarriage, low birth weight, preterm birth, and increased perinatal mortality rate. While we did not find a significant excess of miscarriage, these are known to be under reported in hospital-based data as many women are not admitted to hospital after such an event.

The largest study to date, the Childhood Cancer Survivor Study followed up 1,915 female survivors of childhood cancer with 4,029 pregnancies (63% resulting in live births) by questionnaire.11 The survivor group were less likely to have their pregnancy result in a live birth than the female sibling comparison group, and the risk of offspring of low birth weight was greater among women who had received pelvic radiation. Further analysis on the births of 2,201 children born to 1,264 female survivors showed these were more likely to be born preterm, but not smaller than the offspring of their siblings.12

A Norwegian study13 based on registry linkage data showed that infants born to women with a cancer history were on average 130 g lighter and delivered 6 days earlier compared with the general population and were more likely to be delivered by cesarean. In our study, the risk of offspring being born preterm showed a 33% increase among the cancer group. The odds ratio estimate remained relatively unchanged after adjusting for maternal age and year of delivery. However, since our analysis had only 50% power to detect an odds ratio of 1.3 as statistically significant, the possibility of a causal association cannot, on this basis, be excluded.

We had no specific prior hypotheses relating to postpartum hemorrhage, but it is of note that this finding cannot be explained by the excess of cesarean deliveries experienced by the cancer survivors group. Both factors were mutually adjusted for in the final model. Since we looked at multiple outcomes, this could be a chance finding. However, blood loss at delivery is often subjectively measured and generally underestimated14,15 and it may be that this finding reflects more accurate measurement in women perceived to be high risk.

Radiotherapy and chemotherapy have been identified as potentially harmful to the reproductive system. The level of damage is dependent on the age at time of treatment, the site and dose of irradiation, or drug regimen.16 Pelvic or abdominal irradiation has been shown to affect uterine volume,17 endometrial thickness, and blood flow18,19 while cranial irradiation may lead to damage to the hypothalamic–pituitary axis.20 While these biological changes after treatment could potentially result in poorer obstetric outcomes, reasons for complications such as postpartum hemorrhage are still largely speculative. A history of cancer also makes these women more likely to encounter increased surveillance and a lower threshold for intervention, as previously noted in other groups perceived as high risk.21,22 Further, the women themselves may see the pregnancy as particularly precious, possibly having considered that they would never achieve childbearing, and so may be less likely to be passive regarding the management of their pregnancy. This would help explain the higher rates of intervention shown in this population even after adjustment for maternal age.

This work provides information for health professionals involved in the effective obstetric management of the increasing numbers of women who become pregnant after treatment for cancer. It also provides data which will help address some of these women’s concerns before and during pregnancy. Understanding the effects of cancer on future childbearing may assist in the strategic targeting of resources to give these women the best care and access to treatment.

Back to Top | Article Outline


1. McVie JG. Cancer treatment: the last 25 years. Cancer Treat Rev 1999;25:323–31.
2. ISD Scotland cancer statistics. Available at: Retrieved June 27, 2007.
3. ISD Scotland. Available at: Retrieved April 25, 2007.
4. Brewster DH, Stockton D, Harvey J, Mackay M. Reliability of cancer registration data in Scotland, 1997. Eur J Cancer 2002;38:414–7.
5. Cole SK. Scottish maternity and neonatal records. In: Chalmers I, McIlwaine GM, editors. Perinatal audit and surveillance. London, (UK): Royal College of Obstetricians and Gynaecologists; 1980. p. 39–51.
6. Centre for Evidence-Based Medicine. NNT and NNH. Available at: Retrieved July 4, 2007.
7. Li FP, Gimbrere K, Gelber RD, Sallan SE, Flamant F, Green DM, et al. Outcome of pregnancy in survivors of Wilms’ tumor. JAMA 1987;257:216–9.
8. Byrne J, Mulvihill JJ, Connelly RR, Austin DA, Holmes GE, Holmes FF, et al. Reproductive problems and birth defects in survivors of Wilms’ tumor and their relatives. Med Pediatr Oncol 1988;16:233–40.
9. Hawkins MM, Smith RA. Pregnancy outcomes in childhood cancer survivors: probable effects of abdominal irradiation. Int J Cancer 1989;43:399–402.
10. Green DM, Peabody EM, Nan B, Peterson S, Kalapurakal JA, Breslow NE. Pregnancy outcome after treatment for Wilms tumor: a report from the National Wilms Tumor Study Group. J Clin Oncol 2002;20:2506–13.
11. Green DM, Whitton JA, Stovall M, Mertens AC, Donaldson SS, Ruymann FB, et al. Pregnancy outcome of female survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Am J Obstet Gynecol 2002;187:1070–80.
12. Signorello LB, Cohen SS, Bosetti C, Stovall M, Kasper CE, Weathers RE, et al. Female survivors of childhood cancer: preterm birth and low birth weight among their children. J Natl Cancer Inst 2006;98:1453–61.
13. Fossa SD, Magelssen H, Melve K, Jacobsen AB, Langmark F, Skjaerven R. Parenthood in survivors after adulthood cancer and perinatal health in their offspring: a preliminary report. J Natl Cancer Inst 2005;(34):77–82.
14. Razvi K, Chua S, Arulkumaran S, Ratnam SS. A comparison between visual estimation and laboratory determination of blood loss during the third stage of labour. Aust N Z J Obstet Gynaecol 1996;36:152–4.
15. Duthie SJ, Ven D, Yung GL, Guang DZ, Chan SY, Ma HK. Discrepancy between laboratory determination and visual estimation of blood loss during normal delivery. Eur J Obstet Gynecol Reprod Biol 1991;38:119–24.
16. Meirow D, Nugent D. The effects of radiotherapy and chemotherapy on female reproduction. Human Reprod Update 2001;7:535–43.
17. Larsen EC, Schmiegelow K, Rechnitzer C, Loft A, Muller J, Andersen AN. Radiotherapy at a young age reduces uterine volume of childhood cancer survivors. Acta Obstet Gynecol Scand 2004;83:96–102.
18. Critchley HO, Wallace WH, Shalet SM, Mamtora H, Higginson J, Anderson DC. Abdominal irradiation in childhood; the potential for pregnancy. Br J Obst Gynaecol 1992;99:392–4.
19. Holm K, Nysom K, Brocks V, Hertz H, Jacobsen N, Muller J. Ultrasound B-mode changes in the uterus and ovaries and Doppler changes in the uterus after total body irradiation and allogeneic bone marrow transplantation in childhood. Bone Marrow Transplant 1999;23:259–63.
20. Bath LE, Anderson RA, Critchley HO, Kelnar CJ, Wallace WH. Hypothalamic-pituitary-ovarian dysfunction after prepubertal chemotherapy and cranial irradiation for acute leukaemia. Human Reprod 2001;16:1838–44.
21. Thomson F, Shanbhag S, Templeton A, Bhattacharya S. Obstetric outcome in women with subfertility. BJOG 2005;112:632–7.
22. Bell JS, Campbell DM, Graham WJ, Penney GC, Ryan M, Hall MH. Can obstetric complications explain the high levels of obstetric interventions and maternity service use among older women? A retrospective analysis of routinely collected data. BJOG 2001;108:910–8.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.