Influence of chemotherapy for gestational trophoblastic disease on subsequent pregnancy outcome
Correspondence: Mr R. P Woolas, Department of Gynaecological Oncology, St Mary's Hospital, Portsmouth PO3 6AD, UK.
The aim of this study was to determine the influence of cytotoxic chemotherapy on subsequent reproductive performance. Details of post-treatment reproductive intent and outcome were requested from 1211 survivors registered at The Charing Cross Hospital gestational trophoblastic disease centre; a response rate of 96% was achieved. Seven hundred and twenty-eight women had tried to become pregnant; 607 reported at least one live birth, 73 conceived but had not registered a live birth, and 48 did not conceive. No differences were apparent between the 392 women who received methotrexate as single agent chemotherapy and the 336 treated with multi-agent chemotherapy. Women who had registered a live birth were younger (P < 0.0001) and the duration of follow up was significantly less among those who did not achieve pregnancy at all (P < 0.0003). A higher than expected rate of caesarean section and stillbirth was recorded. The chemotherapy protocols used by this unit have minimal impact on the subsequent ability to reproduce.
Gestational trophoblastic disease (GTD) occurs predominantly in young women of reproductive age and is usually cured by modern cytotoxic chemotherapy regimens1. The subsequent ability of these women to conceive, or the health of their offspring, may be modified by exposure to these drugs. An earlier report suggested that in general, women treated for GTD can be reassured that they can anticipate a normal reproductive outcome following chemotherapy2. However, this favourable outcome may have been attenuated by the introduction of etoposide in 1979 into the treatment schedules for women with medium and high risk disease, and by cisplatin, which has been used in salvage regimens1. This paper provides information on the most recent experience of The Charing Cross Hospital referral centre for trophoblastic disease, the largest series of such patients in the world.
Between 1957 and October 1990, 1397 women were treated by this centre with cytotoxic chemotherapy for persistent trophoblastic disease, including invasive hydatidiform mole, choriocarcinoma and placental site trophoblastic tumours. During this period 137 women died of the tumour or the immediate complications of treatment. Eleven patients have subsequently died of a second malignancy. A further six were known to have died of a cause unrelated to their gestational trophoblastic disease. No forwarding address was available for 32 women who were resident abroad. The remaining 1211 patients (87%) were sent a questionnaire in October 1992 requesting details of their reproductive intentions and outcome in the years following their chemotherapy. The information returned was entered on the tumour registry database of this treatment centre.
Women presenting with GTD were stratified according to an established prognostic scoring system1 which divides patients into low, medium or high risk groups. Women with low risk disease were treated with single agent methotrexate chemotherapy with folinic acid rescue. Those with medium or high risk disease were treated with multi-agent combination chemotherapy regimens. The precise drug protocols have evolved during the duration of this study but have included various combinations of actinomycin D, hydroxyurea, methotrexate, 6 mercaptopurine, vincristine, cyclophosphamide, adriamycin and, after 1979, etoposide. In addition approximately one-quarter of the women with low risk disease switched from single agent to combination chemotherapy on account of drug resistance or toxicity. Patients who became resistant to combination chemotherapy received salvage therapy, including cisplatin after 1979.
Comparisons of parametric data were performed using Fisher's exact test and of non-parametric data in sample sets by the Mann-Whitney U test. The stillbirth rate in the general population has declined over the 30 years of this study. This has been accounted for by undertaking a Mantel Haenszel3 procedure comparing the GTD dataset and the annual national stillbirth statistics published by the Office of Population, Censuses and Surveys covering the years 1963–19924. During this period 21.7 million births were recorded, including 213,481 stillbirths; the stillbirth rate fell from 17.2 to 4.3 per 1000 births.
To date 1166 women (96.3%) have returned information concerning their subsequent reproductive history. Of these, 338 (28%) stated that they had not tried to become pregnant since completing their chemotherapy and these women are excluded from further analysis.
The median age at the time of diagnosis of GTD of the 728 women who wished to conceive was 25.5 years (range 13.7–43.7) and the median follow up was 12.3 years (range 2.7–32.0). Among these 728 women, up to 1 January 1994, 680 (93%) have been successful in achieving pregnancy, of whom 607 (83%) had registered at least one live birth; 73 (10%) had an unsuccessful pregnancy only, defined as a conception with any outcome except a live birth. Only 48 (7%) women treated with chemotherapy for GTD who stated that they wished to become pregnant, failed to conceive at all.
The age at which the disease was diagnosed in the women who have had subsequent live births (median 25.2; range 13.7–38.3 years) was significantly younger than those who had conceived but not achieved a live birth, (median 27.1, range 16.5–41.4 years: Mann-Whitney U test, P= 0.004) or failed to conceive (median 28.0, range 15.543–7 years: Mann-Whitney U test, P= 0.0003). The duration of follow up after completion of treatment was shorter in women who had not conceived (median 9.0, range 2.7–28.4 years) than in those who had live births (median 12.5, range 2.7–32.0 years: Mann-Whitney U test, P= 0.0003)or failed pregnancies (median 13.4, range 3.0–8.8 years: Mann-Whitney U test, P= 0.0007).
Three hundred and ninety two women received single agent methotrexate only, whilst 336 women were treated with multi-agent chemotherapy. There was no difference in the ages at which the disease was diagnosed (medians 25.6 and 25.5 years: Mann-Whitney U test, P= 0.8), or duration of follow up (medians 11.5 and 12.8 years: Mann-Whitney U test, P = 0.12) between these two groups. Furthermore, there was no significant difference in either the conception rate or pregnancy outcome between the women treated with single agent methotrexate, compared with those receiving multi-agent chemotherapy (x2= 0.21, P= 0.90) (Table 1).
Table 1. Age, follow up and pregnancy outcome according to treatment received. Values are given as median [range] or n (%).
|No. of patients||392||336|
|Age (years)||25.6 [13.7–43.7]||25.5 [15.8–39.8]|
|Follow up (years)||11.5 [2.7–27.2]||12.8 [2.7–32]|
|Live births||327 (83.4)||280 (83.3)|
|Unsuccessful pregnancies||38 (9.7)||35(10.4)|
|No pregnancy||27(6.9)||21 (6.2)|
A total of 1313 conceptions were reported, including repeat molar pregnancies on 18 occasions in 17 women. These gestations resulted in 1000 live births, 11 of which were multiple pregnancies and 19 stillbirths; 190 miscarriages or ectopic gestations were documented. On 104 occasions a termination of pregnancy was registered. The stillbirth rate in the women treated for GTD was significantly higher than that in the general population, with an odds ratio of 2.87 (95% CI 2.44, 3.03: Mantel Haenszel χ2= 22.6, P < 0.001). Only 18 major or minor congenital malformations were reported, affecting 1.7% of the total births.
In the pregnancies progressing beyond 24 weeks, 7.9% were delivered before 37 weeks gestation. The caesarean section rate for the whole group was 17%. This rate varied in relation to the date of delivery. No births were registered until 1963; prior to 1973 a caesarean section rate of 29% was recorded. For the subsequent two decades 1973–1983 and 1983–1993, the rates were 15% and 17%, respectively.
Of the women attending this national centre for gestational trophoblastic disease, 90% have been treated successfully. The proportion of patients cured has increased with time. Almost half (49%) of the mortality related to gestational trophoblastic disease occurred among the 25% of patients who were treated prior to 1973.
More than 70% of the survivors expressed a wish to conceive again. The pregnancy rates achieved by these women to not appear to differ from those that might be expected in the general population. Furthermore, the duration of follow up in the group of women who did not achieve pregnancy was significantly shorter than among those that did conceive. A reduction in the number of women in this group may be anticipated with time. Indeed, further follow up information received by the registry has now revealed that at least four of the women in the ‘no pregnancy’ group have been delivered of healthy infants over the last two years. The chemotherapy prescribed to these four patients was methotrexate only in three and a combination of drugs in one.
The results shown in Table 1 confirm that methotrexate prescribed as single agent with folinic acid rescue for low risk disease appears to have no influence on subsequent reproductive performance. Combination chemotherapy for malignant disease has been previously implicated in a reduction of fertility. However, the multi-drug regimens used for GTD in this study do not appear to have adversely affected conception rates.
Maternal age is a powerful determinant of reproductive success. It is therefore not surprising to find that women achieving a live birth were significantly younger than those who did not. Although age is a component of the Charing Cross scoring system and thus may have influenced the drug protocol prescribed, the median age of the patients in the two treatment groups was identical (Table 1). Among women treated with a high risk protocol using combination chemotherapy 25 progressed to a salvage regimen. Twenty-one of these women have been successfully treated, and all six who subsequently tried to become pregnant have been delivered of a healthy baby.
The rate of repeat hydatidiform mole recorded in these patients was > 1%. This is at least 10 times greater than the background incidence. Consequently, any woman with a history of gestational trophoblastic disease should undergo an early ultrasound scan if thought to be pregnant.
The stillbirth rate of 18.6 per 1000 births was greater than that which would be expected today. Five of these stillbirths were registered by two individual women. Twelve of these patients were treated with single agent chemotherapy. National statistics for England and Wales during the time period 1963 to 1992 record a fall in the rate of stillbirths from 17.2 to 4.3 per 1000 births4. A Mantel Haentzel procedure was performed to compare the stillbirth rate in the women treated for GTD with that of the general population. The stratification of these data in relation to the period in which the stillbirths occurred confirmed that a significant excess was present among the patients with GTD. It is of note that the stillbirth rate rises with maternal age above 25 years. Women with GTD are advised not to become pregnant for at least one year following completion of chemotherapy. The delay of over 12 months among this cohort, whose median age at diagnosis was 25.5 years, suggests that many women were reproducing at an age where the chance of a stillbirth is greater. Reassuringly, the congenital malformations reported were consistent in both frequency and type with those occurring in the general population. Premature delivery has been previously noted as an association with previous exposure to radiotherapy for childhood cancer5. We could discern no effect of previous cytotoxic drug treatment on premature delivery.
The caesarean section rate was higher than expected6, and this was especially apparent in the women delivering in the earliest years of this study, almost one-third of whom had a caesarean section. The rate fell to 15% between 1973 and 1983, but was still approximately twice the national rate, and it was still high between 1983 and 1993, 17% compared with 12% in the general population. Among the 607 women who were delivered of a live infant after treatment, 260 were delivered on 482 occasions before their diagnosis of GTD; their rate of caesarean sections was 6.2%. Three hundred and forty seven woman (57%) were primigravidae. The increased rate of obstetric intervention in this population, especially in the earlier years of this study, could in part be explained by increasing maternal age, the general upward trend for caesarean delivery, or, more likely, obstetric concern in relation to the unknown reproductive potential of such women at the time of their subsequent pregnancy. In those women who received salvage chemotherapy the rate of caesarean section was 63%.
The chemotherapy protocols used by this unit for the treatment of gestational trophoblastic disease appear to have little effect on subsequent reproductive performance. A past history of treatment with chemotherapy for this disease may increase the likelihood of a subsequent subsequent caesarean delivery or a stillborn child.
This study would not have been possible without the assistance of the patients who supplied their information and the gynaecologists who referred cases to this centre. The authors would like to thank Professor K. D. Bagshawe, Dr G. J. R. Rustin, Professor R. H. J. Begent and Ms J. Dent for their part in the management of many of these women and in establishing the gestational trophoblastic disease registry at Charing Cross Hospital.