Secondary Logo

Journal Logo

Contents: Gynecologic Oncology: Original Research

Second Uterine Curettage and the Number of Chemotherapy Courses in Postmolar Gestational Trophoblastic Neoplasia

A Randomized Controlled Trial

Hemida, Reda MD; Vos, Elvira L. MD, PhD; El-Deek, Basem MD; Arafa, Mohammad MD, PhD; Toson, Eman MD; Burger, Curt W. MD, PhD; van Doorn, Helena C. MD, PhD

Author Information
doi: 10.1097/AOG.0000000000003232
  • Free
  • Correction

Postmolar gestational trophoblastic neoplasia,1 is typically diagnosed in asymptomatic women undergoing serum human chorionic gonadotropin (hCG) monitoring after evacuation of a complete or partial molar pregnancy.2 After diagnosis and after anatomical staging3 patients at low risk for failure of single-agent chemotherapy (score less than 7) can be treated with methotrexate or actinomycin D,4 resulting in complete response rates of 70–80% and survival rates approaching 100%.4,5

The role of a second uterine curettage in the management of postmolar gestational trophoblastic neoplasia is unclear. A debulking effect has been shown in two retrospective analyses, which reported that fewer chemotherapy courses were needed to reach undetectable serum hCG levels after second curettage.6,7 Recently, two prospective observational studies have been published demonstrating that selected patients may avoid chemotherapy entirely. A small prospective pilot study reported cures in 10 out of 12 patients after second curettage alone in postmolar gestational trophoblastic neoplasia.9 A larger study, performed by the Gynecologic Oncology Group, reported cure rates of 40% for low-risk nonmetastatic gestational trophoblastic neoplasia using second uterine curettage as a single treatment.10 The disadvantages include complications such as uterine perforation, infection, and bleeding, as well as a delay in starting chemotherapy when postcurettage hCG levels fail to normalize.10

We hypothesized that a second curettage would debulk the uterus and thereby reduce the number of chemotherapy courses needed to normalize the hCG level by two courses, hence reduce treatment time by four weeks. This single-center, randomized phase III trial in low-risk postmolar gestational trophoblastic neoplasia patients was designed to test this hypothesis. We also aimed to identify variables associated with the number of courses required.


In this randomized, controlled phase III trial, carried out in a single university center in Mansoura, Egypt, consecutive patients with low-risk postmolar gestational trophoblastic neoplasia and serum hCG levels of 5,000 international units/L or less were considered eligible. The Mansoura University Hospital provides tertiary health care for most of the Delta region of Egypt, with a population of about 12 million.

Eligible participants were women aged 18 years or older with postmolar gestational trophoblastic neoplasia having a score less than 7, a serum hCG level of 5,000 international units/L or less, and fit for standard treatment with methotrexate. Exclusion criteria were previous uterine perforation and life-threatening bleeding. At the visit to the clinic, all patients were given oral and written information about the study. Patients gave written informed consent before inclusion in the study. The study was approved by the ethical committee of the Faculty of Medicine, Mansoura University (Number R/48). The study was registered in the Dutch Trial Registry (NTR3390). Most of the required treatments (second uterine curettage, methotrexate, and folinic acid) were provided free of charge to reduce the cost burden on patients and to minimize drop-out.

Patients were randomly assigned via Trial Online Process, a web-based application, managed by the research center at the Erasmus MC Cancer Center, University Medical Center Rotterdam, the Netherlands. Random assignments to second curettage or not (1:1) were balanced with use of a biased-coin minimization procedure, with the bias dependent on the average imbalance between numbers of patients already assigned to each treatment arm overall and within the stratification factors of the new patient. The patients were stratified according to the presence or absence of vaginal bleeding and the level of baseline serum hCG, that is, less than 1,500 international units/L or 1,500–5,000 international units/L, because previous studies suggested that these factors might influence the effect of second curettage. The assigned treatment arm was sent immediately by email to the treating gynecologist.

At baseline, all patients underwent a pelvic transvaginal ultrasound examination to measure the dimensions of intrauterine disease (mm2), calculated as the product of the longest compared with the perpendicular diameter of the intrauterine lesion. In addition, patients underwent a staging chest X-ray and hCG measurement.

Second uterine curettage was performed under short-acting general anesthesia or sedation, and antibiotic prophylaxis. After cervical dilatation, the uterine cavity was evacuated using manual vacuum aspiration under ultrasound guidance (Ipas MVA EasyGrip Cannulae).11,12 Tissue obtained was sent for histopathologic analyses.

All patients received standard methotrexate treatment with folinic acid support, both given intramuscularly in a twice-weekly regimen. Methotrexate was administered at a dose of 50 or 75 mg (the latter for women exceeding 75 kg) on days 1, 3, 5, and 7, alternating with folinic acid at a dose of 12.5 mg on days 2, 4, 6, and 8. Women that underwent a second curettage were given the first methotrexate injection within 24 hours of the curettage. The first methotrexate course was given during admission, and subsequent courses were given at the woman's home by a trained nurse. After normalization of serum hCG, that is, hCG level less than 7 international units/L, two consolidation courses of methotrexate were given.

Before each course of chemotherapy, a complete blood cell count, kidney and liver function tests, and serum hCG were evaluated. Adverse events were defined and graded using Common Terminology Criteria for Adverse Events 3.0.13 In patients who experienced severe toxicity or failure of methotrexate, treatment was abandoned and further management was given according to tumor board recommendations. Methotrexate failure was deemed to be a plateau or decline of less than 10% in hCG levels over three courses, or a rise in the hCG level of more than 10% above the prior obtained value over two consecutive courses. In case of methotrexate failure or relapsed gestational trophoblastic neoplasia, restaging was performed. The choice of second-line chemotherapy was individualized according to disease status, tumor board recommendations, and drug availability. During and after treatment, patients were strictly advised not to become pregnant. Follow-up was performed by monthly testing of serum hCG levels for a year after normalization of hCG.

The primary endpoint was the mean number of chemotherapy courses needed to achieve hCG normalization. Secondary endpoints were: length of chemotherapy cycles, complications of methotrexate, treatment outcomes, histology, and follow-up. Treatment outcomes includes the measure of successful treatment, which was defined as normalization of serum hCG level by methotrexate only, no emergency surgery, and without relapse within one year. Second-line treatment was defined as treatment after methotrexate failure, the need for emergency surgery and any treatment for relapse during follow-up.

Based on previously published data, we expected to find a mean reduction from 4.8 to 2.5 chemotherapy courses, with an expected SD of 2.4, before hCG normalization.6,8 With a two-sided 5% significance level, a power of 99%, and an expected dropout rate of 10%, a sample size of 44 patients per group was necessary to detect this difference. We anticipated that a 36-month inclusion period would be required to recruit this number of patients.

Analyses were performed as intention-to-treat. Sociodemographic characteristics between the two arms were compared using the Mann-Whitney U test for continuous variables and the Kruskal-Wallis test for categorical variables. Primary and secondary outcomes between the two randomization arms were also compared by Mann-Whitney U and Kruskal-Wallis tests. Subgroup analyses were performed for the primary outcome according to: bleeding, hCG pattern, hCG level, weight, length of single course, and treatment outcome. Linear regression was performed to study the effect of second uterine curettage and other variables on the number of chemotherapy courses needed. A negative odds ratio (OR) means the variable was associated with fewer chemotherapy courses. To study the effect of several variables on treatment success, which is a dichotomous outcome, logistic regression was performed. An OR less than 1 means the variable was associated with less frequent treatment success. Multivariable regression analysis was performed with all relevant variables, with a maximum degree of freedom of 1 per 10 patients. Statistical tests were two-sided and P-values were considered statistically significant when P<.05. Statistical analyses were carried out using the statistical package SPSS 22 for Windows.

Authors' Data Sharing Statement

  • Will individual participant data be available (including data dictionaries)? Yes, deidentified participant data on the patient level will be available.
  • What data in particular will be shared? Any data.
  • What other documents will be available? Study protocol.
  • When will data be available (start and end dates)? Not applicable.
  • By what access criteria will data be shared (including with whom, for what types of analyses, and by what mechanism)? All study data will be available on request for systematic reviews and meta analyses. Requests made by physicians and epidemiologists will be granted in a timely matter and shared, obeying our hospital policies.


Between October 2011 and February 2016, 123 patients with postmolar gestational trophoblastic neoplasia were seen at the Mansoura Trophoblastic Clinic, of whom 34 were excluded for the following reasons; 26 patients did not fulfil the eligibility criteria, five eligible patients refused randomization, two started treatment in another hospital, and one needed immediate emergency surgery for severe uterine bleeding. Consequently, 89 consecutive patients were enrolled in the study. After randomization, three patients were excluded; one chose hysterectomy in another hospital, the second refused subsequent treatment when methotrexate failed and was lost to follow-up, and the third ultimately did not fulfill the diagnostic criteria of postmolar gestational trophoblastic neoplasia. Therefore, 86 patients were finally included in the intention-to-treat analyses (Fig. 1). The first patient was enrolled October 17, 2011. Five patients were enrolled before the trial registration (March 12, 2012). Patient groups were comparable for age, parity, hCG level, size of intrauterine lesion, performance score, weight, and body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) (Table 1).

Fig. 1.
Fig. 1.:
CONSORT (Consolidated Standards of Reporting Trials) flow chart of the study.Hemida. Second Curettage in Gestational Trophoblastic Neoplasia. Obstet Gynecol 2019.
Table 1.
Table 1.:
Sociodemographic Data

The mean number of chemotherapy courses required for hCG normalization was comparable for the control group, 4.4±2.2 SD, and the intervention group, 3.8±2.3 SD (P=.14) (Table 2). Mean cycle duration was 17.9 days in both groups. Methotrexate failure was similar for both groups—in three patients in the control group and four in the intervention group. During the first year of follow-up, disease relapse occurred in two patients in the control arm and in one patient in the intervention group. In the intervention group, one patient was diagnosed with choriocarcinoma after a subsequent full-term pregnancy.

Table 2.
Table 2.:
Treatment Outcome With Methotrexate

According to protocol, all patients received two consolidation courses after normalization of hCG, except for three patients who received only one consolidation course owing to use of an earlier protocol. None of the latter experienced a relapse.

No surgical complications occurred in the intervention group. Side effects of methotrexate occurred in 28 patients (32.6%), comprising 15 patients in the control group and 13 in the intervention group (Table 2), with predominantly grade 1 and 2 adverse effects. Two patients in the control group developed grade 3 adverse effects, that is, severe vaginal bleeding necessitating surgical intervention; one patient underwent emergency uterine curettage and the other underwent hysterectomy on day 6 of the second methotrexate course. In the intervention group, one patient requested a hysterectomy when methotrexate failure occurred after the third methotrexate course.

Tables 3 and 4 depict the number of courses needed to reach hCG normalization in relation to patient characteristics. Linear regression analyses showed that only serum hCG level was statistically significantly related to the number of chemotherapy courses needed, both when hCG level was considered as a continuous value (OR 4.69×10−4 [95% CI 2.32×10−4; 7.07×10−4]) and when hCG level between 1,500 and 5,000 international units/L was compared with less than 1,500 international units/L (OR 1.85 [95% CI 0.88; 2.81]). In addition, after adjustment for factors including baseline hCG level, vaginal bleeding, weight, and cycle duration in multivariable regression analyses, second uterine curettage did not significantly reduce the number of courses needed to reach normalization of hCG (OR −1.04 95% CI −2.54; 0.46, P=.17). There was also no relation between baseline serum hCG level and uterine dimensions (Pearson correlation coefficient 0.164 [P=.132]).

Table 3.
Table 3.:
Subgroup Analyses for the Mean Number of Methotrexate Courses Needed to Achieve hCG Normalization
Table 4.
Table 4.:
Univariable Linear Regression Analysis Showing the Effect on Total Number of Methotrexate Courses Needed to Achieve hCG Normalization

Histopathology after second uterine curettage was available for 42 of 43 patients, revealing molar tissue in 35 patients (83%) and only endometrial tissue in seven patients (17%). The histopathologic finding of molar tissue was not predicted by the dimensions of the uterine lesion; the mean dimensions for women with molar tissue was 208±293 mm2 vs 96±222 mm2 for women with only endometrial tissue (P=.34).

Patients were advised to undergo monthly hCG monitoring for a year and to refrain from pregnancy during this period. Three patients were lost to follow-up, and 13 (15.1%) got pregnant within 12 months (Table 2). One of them underwent termination, two suffered first-trimester pregnancy loss, one developed a second molar pregnancy, and in one pregnancy the fetus had congenital intestinal obstruction. The eight other pregnancies resulted in healthy neonates.

Treatment success rate was similar for both groups (Table 5). After adjustment for hCG level, vaginal bleeding, weight, and mean cycle duration, success rates of methotrexate treatment were similar in both treatment groups (OR 0.86 95% CI 0.24; 3.06, P=.81).

Table 5.
Table 5.:
Univariable Logistic Regression Showing the Effect on Treatment Success


Second uterine curettage did not significantly reduce the number of chemotherapy courses among postmolar gestational trophoblastic neoplasia patients in this randomized controlled trial. Complication rates remain unclear as the diagnosis is relatively rare and the use of a second curettage inconsistent, but when performed it may result in a treatment delay and prolongation of the treatment duration should hCG levels fail to normalize. Our results are in line with one retrospective study that concluded that dilatation and curettage of gestational trophoblastic neoplasia does not affect future chemotherapy requirements,14 yet in contrast with two other retrospective analyses.6,8 In the Gynecologic Oncology Group study, 26 of 64 (40%) patients did not need chemotherapy after a second curettage.10 Because results of this study were not available at the development and execution of our study, all women were treated with methotrexate, and a possible definitive curative effect of second curettage was not awaited. It is therefore conceivable that the inclusion of women cured by second curettage (ie, not receiving any chemotherapy) in some retrospective studies led to a lower overall number of chemotherapy courses. However, it should be noted that it is still unclear whether awaiting the curative effect of second curettage affects the total duration of treatment in patients for whom second curettage fails.

The few studies available at the time of study-protocol writing suggested that two factors influence the likelihood of reducing the number of chemotherapy courses required to reach hCG normalization, namely the presence of vaginal bleeding and the precurettage hCG level. Severe vaginal bleeding can lead to prompt surgical treatment, illustrated in our study by two women in the control arm undergoing emergency surgery for severe bleeding. Because we performed an intention-to-treat analysis and stratified for vaginal bleeding at presentation, this did not affect the study results. In United Kingdom guidelines, the presence of an intrauterine mass in combination with hCG levels less than 5,000 units/L will prompt second curettage.15 Although all women in the current study had an intrauterine mass, ultrasound findings did not correlate with the hCG level, or to the presence of molar tissue at curettage. Although the numbers included in the present study prohibit conclusions on the success or failure of second curettage in the absence of molar tissue in the pathology specimen, it would be reasonable to conclude that second curettage is not beneficial without histologic evidence for molar tissue. Although all patients in our study had relatively low hCG levels precurettage, initial hCG levels were still related to the number of courses required for normalization. However, when we compared women with hCG levels less than 1,500 with women with hCG levels between 1,500 and 5,000, the second curettage did not have an effect on the number of courses needed for normalization of serum hCG levels.

Failure of methotrexate occurred in seven patients and relapse after therapy in four, including a choriocarcinoma after a subsequent full-term pregnancy. Including the two women who underwent emergency surgery, 15% of the patients needed some form of second-line treatment. This is in line with previous reports.16,17

This study was performed in a single center, with a set protocol, which could limit the broader applicability of our findings. Although a multicenter setting would have been preferable, this study and the Gynecologic Oncology Group study are the first prospective trials completed, and both experienced slow accrual owing to the low incidence of this disease. Methotrexate chemotherapy is generally given in a twice-weekly cycle. In the current study, the mean treatment cycle duration was 17.9 days in both groups, and thus unlikely to have affected the primary study outcome. This observation prompted a substandard care analysis, in which we observed several factors associated with treatment delays. Patient-related factors included reluctance to perform blood tests on time, reluctance to switch to a second-line treatment, difficulties in reaching the clinic, and financial burdens; external factors such as poor availability of methotrexate or other second-line drugs in Egypt occurred as a result of political factors.

Methotrexate is available in vials of 50 mg, and the dose in our study was administered in a fixed dose (50 or 75 mg), allowing subsequent courses to be administered by a nurse in a home setting. The international discussion on the best methotrexate dosing schedule is not yet settled, and some favor a dose related to body weight (1 mg/kg) rather than a fixed dose. Because the two treatment groups were comparable regarding body weight and BMI, the schedule and dosing in this study is unlikely to have affected the primary outcome.


1. Berkowitz RS, Goldstein DP. Current management of gestational trophoblastic diseases. Gynecol Oncol 2009;112:654–62.
2. Niemann I, Vejerslev LO, Frøding L, Blaakær J, Maroun LL, Hansen ES, et al. Gestational trophoblastic diseases - clinical guidelines for diagnosis, treatment, follow-up, and counselling. Dan Med J 2015;62:A5082.
3. Ngan HY, Bender H, Benedet JL, Jones H, Montruccoli GC, Pecorelli S, et al. Gestational trophoblastic neoplasia, FIGO 2000 staging and classification. Int J Gynecol Obstet 2003;83(suppl 1):175–7.
4. Maestá I, Nitecki R, Horowitz NS, Goldstein DP, de Freitas Segalla Moreira M, Elias KM, et al. Effectiveness and toxicity of first line methotrexate chemotherapy in low-risk postmolar gestational trophoblastic neoplasia: the New England Trophoblastic Disease Center experience. Gynecol Oncol 2018;148:161–7.
5. Khan F, Everard J, Ahmed S, Coleman RE, Aitken M, Hancock BW. Low-risk persistent gestational trophoblastic disease treated with low-dose methotrexate: efficacy, acute and long-term effects. Br J Cancer 2003;89:2197–201.
6. van Trommel NE, Massuger L, Verheijen R, Sweep FC, Thomas CM. The curative effect of a second curettage in persistent trophoblastic disease: a retrospective cohort survey. Gynecol Oncol 2005;99:6–13.
7. Pezeshki M, Hancock BW, Silcocks P, Everard JE, Coleman J, Gillespie AM, et al. The role of repeat uterine evacuation in the management of persistent gestational trophoblastic disease. Gynecol Oncol 2004;95:423–9.
8. Hemida RA, Toson E, van Doorn HC. Impact of uterine recurettage, pre-evacuation, and week-1 hCG level on number of chemotherapy courses in treatment of postmolar GTN. J Exp Ther Oncol 2011;9:217–20.
9. Yarandi F, Jafari F, Shojaei H, Izadi-Mood N. Clinical response to a second uterine curettage in patients with low-risk gestational trophoblastic disease: a pilot study. J Reprod Med 2014;59:566–70.
10. Osborne RJ, Filiaci VL, Schink JC, Mannel RS, Behbakht K, Hoffman JS, et al. Second curettage for low-risk nonmetastatic gestational trophoblastic neoplasia. Obstet Gynecol 2016;128:535–42.
11. Tasmin N, Mahmud G, Fatima S, Sultana M. Manual vacuum aspiration: a safe and cost-effective substitute of electric vacuum aspiration for the surgical management of early pregnancy loss. J Pak Med Assoc 2011;61:149–53.
12. Lara-Ricalde R, Rodríguez-Bosch M, de la Jara-Díaz J, Tovar-Calleros G, Ahued-Ahued JR. Manual intrauterine aspiration in the treatment of molar pregnancy. Ginecol Obstet Mex 1999;67:438–41.
13. Cancer therapy evaluation program, common terminology criteria for adverse events. Available at: Retrieved February 26, 2018.
14. Growdon WB, Wolfberg AJ, Goldstein DP, Feltmate CM, Chinchilla ME, Lieberman ES, et al. Evaluating methotrexate treatment in patients with low-risk postmolar gestational trophoblastic neoplasia. Gynecol Oncol 2009;112:353–7.
15. McGrath S, Short D, Harvey R, Schmid P, Savage PM, Seckl MJ. The management and outcome of women with post-hydatidiform mole ‘low-risk’ gestational trophoblastic neoplasia, but hCG levels in excess of 100 000 IU/l. Br J Cancer 2010;102:810–4.
16. Gueye M, Ndiaye-Gueye MD, Kane-Gueye SM, Gassama O, Diallo M, Moreau JC. Diagnosis, treatment and outcome of gestational trophoblastic neoplasia in a low resource income Country. Int J MCH AIDS 2016;5:112–8.
17. Lawrie TA, Alazzam M, Tidy J, Hancock BW, Osborne R. First-line chemotherapy in low-risk gestational trophoblastic neoplasia. The Cochrane Database of Systematic Reviews 2016, Issue 9. Art. No.: CD007102. DOI: 10.1002/14651858.CD007102.pub4.

Supplemental Digital Content

© 2019 by the American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.