Gynecologic Oncology 145 (2017) 88–95 Contents lists available at ScienceDirect Gynecologic Oncology j ourna l homepage: www.e lsev ie r .com/ locate /ygyno Multiple pregnancies with complete mole and coexisting normal fetus in North and South America: A retrospective multicenter cohort and literature review Lawrence H. Lin a, Izildinha Maestá b, Antonio Braga c, Sue Y. Sun d, Koji Fushida a, Rossana P.V. Francisco a, Kevin M. Elias e,f,g,h, Neil Horowitz e,f,g,h, Donald P. Goldstein e,f,g,h, Ross S. Berkowitz e,f,g,h,⁎ a University of Sao Paulo Trophoblastic Disease Center, University of Sao Paulo Medical School, Sao Paulo, Sao Paulo, Brazil b Botucatu Trophoblastic Disease Center, Clinical Hospital of Botucatu Medical School, Sao Paulo State University, Botucatu, Sao Paulo, Brazil c Rio de Janeiro Trophoblastic Disease Center, (Maternity School of Rio de Janeiro Federal University, Antonio Pedro University Hospital of Fluminense Federal University, MaternityWard of Santa Casa da Misericórdia do Rio de Janeiro), Rio de Janeiro, Rio de Janeiro, Brazil d Sao Paulo Hospital Trophoblastic Disease Center, Paulista School of Medicine, Sao Paulo Federal University, Sao Paulo, Sao Paulo, Brazil e New England Trophoblastic Disease Center, Donald P. Goldstein MD, Trophoblastic Tumor Registry, Boston, MA, USA f Division of Gynecologic Oncology, Brigham and Women's Hospital, Boston, MA, USA g Department of Obstetrics, Gynecology and Reproductive Biology, Harvard Medical School, Boston, MA, USA h Dana Farber Cancer Institute, Boston, MA, USA H I G H L I G H T S • Pregnancies with mole and fetus in South America presented with more complications. • Pregnancies with mole and fetus were at high risk for persistent disease. • Elective termination did not influence risk of persistent disease. ⁎ Corresponding author at: Brigham andWomens' Hos E-mail address: rberkowitz@partners.org (R.S. Berkow http://dx.doi.org/10.1016/j.ygyno.2017.01.021 0090-8258/© 2017 Elsevier Inc. All rights reserved. a b s t r a c t a r t i c l e i n f o Article history: Received 28 December 2016 Received in revised form 12 January 2017 Accepted 19 January 2017 Available online 26 January 2017 Objective. To determine the clinical characteristics of multiple gestation with complete mole and coexisting fetus (CHMCF) in North and South America. Methods. Retrospective non-concurrent cohorts compromised of CHMCF from New England Trophoblastic Disease Center (NETDC) (1966–2015) and four Brazilian Trophoblastic Disease Centers (BTDC) (1990–2015). Results. Froma total of 12,455 cases of gestational trophoblastic disease seen, 72 CHMCFwere identified. Clin- ical characteristics were similar between BTDC (n = 46) and NETDC (n = 13) from 1990 to 2015, apart from a much higher frequency of potentially life-threatening conditions in Brazil (p= 0.046). There were no significant changes in the clinical presentation or outcomes over the past 5 decades in NETDC (13 cases in 1966–1989 vs 13 cases in 1990–2015). Ten pregnancies were electively terminated and 35 cases resulted in viable live births (60% of 60 continued pregnancies). The overall rate of gestational trophoblastic neoplasia (GTN) was 46%; the cases which progressed toGTN presentedwith higher chorionic gonadotropin levels (p=0.026) and higher frequency of termination of pregnancy due to medical complications (p= 0.006) when compared to those with spontane- ous remission. Conclusions. Themain regional difference in CHMCF presentation is related to a higher rate of potentially life- threatening conditions in South America. Sixty percent of the expectantlymanaged CHMCF delivered a viable in- fant, and the overall rate of GTN in this study was 46%. Elective termination of pregnancy did not influence the risk for GTN; however the need for termination due to complications and higher hCG levels were associated with development of GTN in CHMCF. © 2017 Elsevier Inc. All rights reserved. Keywords: Gestational trophoblastic disease Hydatidiform mole Pregnancy, Twin Multiple, Pregnancy Trophoblastic neoplasms Near miss, healthcare pital, 75 Francis Street, Boston, MA 02115, USA. itz). http://crossmark.crossref.org/dialog/?doi=10.1016/j.ygyno.2017.01.021&domain=pdf http://dx.doi.org/10.1016/j.ygyno.2017.01.021 mailto:rberkowitz@partners.org http://dx.doi.org/10.1016/j.ygyno.2017.01.021 http://www.sciencedirect.com/science/journal/00908258 www.elsevier.com/locate/ygyno 89L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 1. Introduction Multiple gestation with a complete mole and coexisting fetus (CHMCF) is a rare event, with an estimated incidence of 1 case for 20,000–100,000pregnancies [1–3]. These pregnancies are characterized by the coexistence of a complete mole and a potentially viable fetus with a normal placenta, in contrast to partial moles, which are com- posed of a single triploid conceptus with an abnormal placenta [4]. The differential diagnosis is important due to the fact that a partial mole's fetus is nonviable and the gestational trophoblastic neoplasia (GTN) risk is b5% [5]. There is still controversy on the rate of GTN pro- gression in CHMCF, with varying rates of GTN being reported [1,6–9]. CHMCF are classically associated with several pregnancy complica- tions, such as spontaneous abortions, intrauterine deaths, preeclampsia and hyperthyroidism [2,6,7]. In the past, due to the uncertain behavior of these pregnancies, they were commonly advised to be terminated, leaving little information on the natural history of this condition. One large study on CHMCF has shown that there is no association between elective pregnancy termination and the rate of GTN; similarly in single moles, gestational age of evacuation does not influence the progression to GTN [1,10]. A recent collaborative study described different characteristics of complete molar pregnancy in adolescents in North and South America, showing that regional differences may play a role in the clinical presen- tation of complete mole [11]. The reports around the globe describing clinical presentation and GTN risk of CHMCF are conflicting, either due to differences in hospital-based versus population-based data or actual regional differences on disease behavior. The aim of this study was to determine the clinical characteristics of CHMCF in North and South America and provide additional information on the natural history of this condition. 2. Methods 2.1. Study design and setting This is a retrospective cohort study consisting of all multiple preg- nancies with complete mole and coexisting fetus registered at the New England Trophoblast Disease Center (NETDC) from 1966 to 2015 and 4 Brazilian Trophoblastic Disease Reference Centers (Botucatu Tro- phoblastic Disease Center, Rio de Janeiro Trophoblastic Disease Center, Sao Paulo Hospital Trophoblastic Disease Center and University of Sao Paulo Trophoblastic Disease Center; all located in the southeast region of Brazil) from 1990 to 2015. This study was approved by the centers' respective institutional review boards and the manuscript was drafted in accordance with STROBE guidelines [12]. All electronic and paper charts for the patients were reviewed. The diagnosis of CHMCF was confirmed by histological evaluation by expe- rienced pathologists in all cases. In cases with uncertain histological di- agnosis, additional p57KIP2 immunostaining, ploidy or/and cytogenetic analysis was also performed. Patients with a diagnosis of partial mole or placental mesenchymal dysplasia were excluded from this study. The patients were compared regarding: regional location (NETDC vs Brazil) from 1990 to 2015; current cohort (1990–2015) vs historical co- hort (1966–1989) in NETDC; GTN progression vs spontaneous regres- sion (all cohorts' data). For temporally concurrent comparisons, a recently published NETDC cohort on single complete moles (1994– 2013) was compared to the NETDC CHMCF cohort (1990–2015) and all CHMCF data in recent years (1990–2015) [10]. 2.2. Variables and definitions The following variables were abstracted: pregnancy presentation, medical complications, obstetric outcomes, GTN progression and treat- ment, potentially life-threatening conditions and maternal near miss complications. Due to the nature of referral centers, some patients were outside consults without complete follow-up; patients with miss- ing information were included only in analyses of variables for which data were available. All patients followed in the referral centers were assisted during the course of pregnancy by maternal fetal medicine and trophoblastic disease team specialists. Medical complications evaluated in this study included vaginal hem- orrhage, preeclampsia, clinical hyperthyroidism and respiratory dis- tress. Vaginal hemorrhage was defined as excessive vaginal bleeding leading to hospitalization, hemodynamic instability, blood transfusion or termination of pregnancy. Preeclampsia was defined as hypertension (systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg) associated with proteinuria on dipstick or at least 300 mg in 24 h. Clinical hyperthyroidism was characterized as sup- pressed thyroid-stimulating hormone (TSH) with elevated serum free thyroxine (T4) levels in the presence of classical symptoms (such as tachycardia, tremor or elevated body temperature) or the need to be treated with beta-blocker. Respiratory distress was defined when the patient presented with acute tachypnea or dyspnea with a radiographic pulmonary infiltrate or pleural effusion. World Health Organization (WHO) defines a maternal near miss as “a woman who nearly died but survived a complication that occurred during pregnancy, childbirth, or within 42 days of termination of pregnancy”. Potentially life-threat- ening conditions and maternal near miss were established using the WHO criteria and classification [13]. According to theWHO recommen- dations on interventions to improve preterm birth outcome guideline, fetal viability depends on the local resources and before 24 weeks the chance of survival without considerable morbidity is low even in high- resource settings; therefore, viable pregnancies were considered the ones in which live infants were delivered at least at 24 weeks of gesta- tional age [14]. After the end of pregnancy the patients were followed with weekly serum human chorionic gonadotropin (hCG) measurements until nor- mal valueswere achieved and thenwithmonthly values up to 6months. All reference centers used hCG detection kits based on chemilumines- cent methods with a sensitivity of at least 5 mIU/mL. In cases where the hCG was above the detection range of the method, the highest re- corded value was identified. The patients were diagnosed with GTN if they presentedwith at least one of the following FIGO (2002) diagnostic criteria: (1) rise of at least 10% of hCG levels in 3weeklymeasurements, (2) plateau (b10% variation) for 4 weekly values of hCG, (3) metastatic disease in thepresence of positive hCG, (4) histological diagnosis of cho- riocarcinoma. Patients with GTN were staged according to the FIGO (2002) anatomic and prognostic system [15]. Remission was classified as the normalization of hCG levels for at least 3 consecutive weeks. Re- sistance was defined as a variation b10% in 3 consecutive weekly hCG levels or the rise of N10% in 2 weekly values of hCG during chemother- apy. Recurrence was characterized as hCG elevation after remission of disease after completion of chemotherapy in the absence of a new pregnancy. 2.3. Statistical methods The results were analyzed in means ± standard deviation for quan- titative variables and in proportions for qualitative variables. Themeans were compared using the unpaired Student's t-test for normally distrib- uted variables and a Mann-Whitney test for variables with non-normal distribution, while proportions were analyzed using Fisher's exact test. A p value b0.05 was considered statistically significant. The results were all shown as medians in the tables of this manuscript because of the non-normal distribution they displayed. 2.4. Literature review A literature reviewwas performed by searching inMedline, Pubmed and EMBASE databases for cohort studies or case series related to CHMCF in the literature, using the following MeSH (Medical Subject 90 L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 Headings) terms: [(“Gestational Trophoblastic Disease” OR “Hydatidiform Mole”) AND (“Pregnancy, Twin” OR “Multiple, Pregnan- cy”)]. Studies in the English language composed of N5 cases published between 1990 and 2016 and that displayed information on clinical pre- sentation, pregnancy outcomes and GTN progression were included in the review. Conference communications regardingupdated information on previously published series were also included in this review. 3. Results The complete study cohort consisted of 72 cases of CHMCF (46 cases fromBTDC, 13 cases from theNETDChistorical cohort and 13 cases from the NETDC current cohort). The overview of demographic and clinical information of the patients included is presented in Table 1. Seventy pregnancies consisted of multiple pregnancies with a complete mole and a coexisting normal fetus, one case in BTDC included a complete mole and two coexisting fetuses and another case in NETDC was com- posed by a completemole, a normal fetus and a blighted ovum. Thirteen percent of the cases were related to assisted reproductive technologies (ART). When comparing data from BTDC and NETDC from 1990 to 2015, most clinical characteristics (clinical presentation, diagnosis, obstetric outcome and GTN progression) were similar (Table 2). The rate of po- tentially life-threatening conditions in BTDC was significantly higher than in NETDC (63% in BTDC vs 30% in NETDC, p = 0.046). Although the difference in near-miss events was not significant, there was not a single case of maternal near-miss event in NETDC compared to a Table 1 Demographic, clinical characteristics and outcomes of the study population. Number of cases (n) 72 Location, n (%) Brazilian Trophoblastic Disease Centers 46 (64%) New England Trophoblastic Disease Center 26 (36%) Age, median (minimum–maximum) 28.5 (18–41)a Gravidity, median 2 (1–5)b Parity, median 0 (0–4)b Previous mole, n (%) 3 (4%)b Assisted reproduction treatment, n (%) 9 (13%)a Gestational age at diagnosis, weeks, median (minimum–maximum) 15 (9–30)c Gestational age at the end of pregnancy, weeks, median (minimum–maximum) 24 (11–40)d hCG levels at diagnosis, mIU/mL, median (minimum–maximum) 400,000 (1048–2,469,000)e hCG levels at the end of pregnancy, mIU/mL, median (minimum–maximum) 200,000 (1048–1,670,600)f Cytogenetic or ploidy analysis, n (%) 30 (54%)g Medical complications, n (%) 41 (63%)h Termination of pregnancy before 24 weeks, n (%)a Elective termination 10 (14%) Termination due to medical complications 7 (10%) No termination 53 (76%) Pregnancy outcomes, n (%)a Miscarriage 11 (16%) Intrauterine death 6 (8%) Preterm delivery 25 (36%) Term delivery 11 (16%) Termination of pregnancy 17 (24%) Fetal viability, n (%) 34 (49%)b GTN progression, n (%) 31 (46%)i Maternal death, n (%) 1 (1.4%) hCG: human chorionic gonadotropin hormone; GTN: gestational trophoblastic neoplasia. a Information not available in 2 patients. b Information not available in 3 patients. c Information not available in 7 patients. d Information not available in 5 patients. e Information not available in 12 patients. f Information not available in 15 patients. g Information not available in 6 patients. h Information not available in 9 patients. i Information not available in 4 patients. 21.7% rate in BTDC (p = 0.18). Even though the frequency of hysterot- omies was higher in Brazil (p = 0.015), the rate of GTN progression was the same between the two groups. Data examining the clinical course of CHMCF compared overfive de- cades at NETDC are presented in Table S1. There was no difference be- tween the two time periods in NETDC (1990–2015 vs 1966–1989) regarding clinical characteristics or outcome. Since there were no significant differences in the obstetric outcomes between the NETDC and BTDC cohorts, we combined these data to sum- marize the outcome of 70 cases of CHMCF (Fig. 1): in 2 cases the infor- mation on pregnancy outcome was not available, 10 cases were electively terminated in the first trimester, 36 viable infants were deliv- ered alive (60% of the 60 continued pregnancies), 7 patients had preg- nancy terminations due to severe pregnancy complications before fetal viability (12% of the 60 continued pregnancies), 17 cases were not viable pregnancies (28% of the 60 continued pregnancies were ei- ther spontaneous abortions, intrauterine deaths or extremepretermde- liveries before viability). There were no cases of fetal demise after 24 weeks of gestational age. Sixty-nine percent of the live births were preterm deliveries. In the combined cohort, the overall rate of GTN was 46%; the cases which developed GTN (n = 31) presented with higher hCG levels (250,000 mIU/mL vs 120,000 mIU/mL p = 0.026), lower gestational age (17 weeks vs 28.5 weeks, p b 0.001) at the end of pregnancy, lower fetal viability rates (27% vs 68%, p = 0.001) and higher need for termination of pregnancy due to pregnancy complications (20% vs 0%, p = 0.006) when compared to those with spontaneous remission (n= 37) (Table 3). No patients were given prophylactic chemotherapy nor were they subjected to a second uterine evacuation. To compare the outcomes between twin and singleton molar preg- nancies, data from a published study on single complete moles from NETDC (1994–2013) was compared to the CHMCF cases at the NETDC in a similar timeframe (1990–2015) (Table S2) [5,16]. Due to a similar presentation of CHMCF in NETDC and BTDC between 1990 and 2015, the data was combined to increase statistical power of the analysis for the comparison with single complete moles from NETDC (1994–2013) (Table 4). In both analyses, there was a statistically significant higher gestational age of diagnosis, gestational age at the end of pregnancy, preeclampsia and potentially life-threatening conditions rate in CHMCF than in single complete moles (Table 4 and Table S2). The com- bined data from NETDC and BTDC from 1990 and 2015 also showed higher levels of hCG at diagnosis (p b 0.001), higher rate of clinical hy- perthyroidism (p = 0.0003), higher frequency of maternal near miss events (p = 0.0001) and risk of GTN (p = 0.0001) when compared to single complete moles. Although not quite reaching statistical signifi- cance (p = 0.0625), the rate of progression to GTN in CHMCF in NETDC seemed to be higher (42%) than in the single moles (18%). 4. Discussion There is still controversy about disease presentation, obstetric out- comes and GTN progression concerning CHMCF, due to the rarity of this entity and different reports around theworld. In this currentwork, the re- gional differences in CHMCF were addressed by comparing cohorts from Referral Centers in Brazil and United States of America in the same time period (1990–2015). Most of the variables were not different, except for the higher frequency of potentially life-threatening conditions in Brazil. In single complete moles, more complications are expected when there is a later diagnosis of molar pregnancy, which would not explain the dif- ference seen in this study, since gestational age at diagnosis and at the end of pregnancy were similar in BTDC and NETDC [17]. This increased frequency of potentially life-threatening conditions in Brazil also could not be accounted only to healthcare access, since the rate of elective pregnancy terminations were similar and some of the conditions in the WHO classification may not be preventable in CHMCF (e.g. trophoblastic embolization, thrombocytopenia). A plausible explanation for this Table 2 Comparison of clinical characteristics and outcomes of multiple pregnancieswith completemole and coexisting normal fetus between Brazilian Trophoblastic Disease Centers (BTDC) and New England Trophoblastic Disease Center (NETDC) from 1990 to 2015. NETDC 1990–2015 BTDC 1990–2015 p value Number of cases 13 46 Age, median (minimum–maximum) 34 (18–41) 28 (18–41) 0.067 Gravidity, median (minimum–maximum) 2 (1–5) 1.5 (1–5) 0.028 Parity, median (minimum–maximum) 1 (0–3) 0 (0–4) 0.180 Previous mole, n (%) 1 (8%) 1 (2%) 0.395 Assisted reproduction treatment, n (%) 2 (15%) 5 (11%) 0.648 Gestational age at diagnosis, weeks, median (minimum–maximum) 15 (12−22)a 15 (9–23) Gestational age at the end of pregnancy, weeks, median (minimum–maximum) 21 (11–38)a 25 (13–40) 0.708 hCG levels at diagnosis, mIU/mL, median (minimum–maximum) 221,331 (34,366–1.365.080)b 500,000 (2545–2,052,500)b 0.142 hCG levels at end of pregnancy, mIU/mL, median (minimum–maximum) 114,633 (18,361–1,365,080)c 210,000 (28,764–1,670,600)c 0.183 Cytogenetic or ploidy analysis, n (%) 9 (82%)a 22 (48%) 0.051 Medical complications, n (%) 6 (60%)c 32 (70%) 0.711 Vaginal hemorrhage, n (%) 5 (50%)c 31 (67%) 0.468 Preeclampsia, n (%) 3 (30%)c 17 (35%) 0.511 Clinical hyperthyroidism, n (%) 0 (0%)c 9 (20%) 0.185 Respiratory distress, n (%) 0 (0%)c 7 (15%) 0.330 Termination of pregnancy before 24 weeks, n (%)d Elective termination 1 (8%) 6 (13%) 1 Termination due to medical complications 1 (8%) 5 (11%) 1 No termination 10 (83%) 35 (76) 0.716 Pregnancy outcomes, n (%)d Miscarriage 4 (33%) 7 (15%) 0.214 Intrauterine death 2 (17%) 4 (9%) 0.594 Preterm delivery 2 (17%) 18 (39%) 0.186 Term delivery 2 (17%) 6 (13%) 0.665 Termination of pregnancy 2 (17%) 11 (24%) 0.716 Fetal viability, n (%) 4 (33%)b 24 (52%) 0.336 Hysterotomy, n (%) 2 (18%)a 29 (63%) 0.015 Peripartum hysterectomy, n (%) 2 (18%)a 7 (15%) 1 Time to hCG regression, weeks, median (minimum–maximum) 16.14 (5.42–86)a 12 (5–33.14)e 0.186 GTN progression, n (%) 5 (42%)b 21 (47%)b 1 FIGO anatomic staging, median (minimum–maximum) 1 (1–3) 1 (1–3) 1 FIGO prognostic staging, median (minimum–maximum) 1 (0–5)b 3 (0–8) 1 hCG levels at GTN diagnosis, mIU/mL, median (minimum–maximum) 14,327 (9–76,644)b 15,000 (32–137,285) 0.979 Resistant GTN, n (%) 1 (25%) 2 (9%) 0.488 Recurrent GTN, n (%) 2 (50%) 1 (5%) 0.056 Potentially life-threatening conditions, n (%) 3 (30%)a 29 (63%) 0.045 Absolute number of potentially life-threatening conditions events 5 69 Maternal near miss events, n (%) 0 (0%)a 10 (22%) 0.180 Absolute number of maternal near miss events 0 22 Death, n (%) 0 (0%)b 1 (2%) 1 hCG: human chorionic gonadotropin hormone; GTN: gestational trophoblastic neoplasia. a Information not available in 2 patients. b Information not available in 1 patient. c Information not available in 3 patients. d Information not available in 1 patient in the NETDC series. e Information not available in 4 patients. 91L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 disparity could be the reflection of different disease behavior, possibly driven by genetic and environmental factors of regional location, which can seen in other models of disease [18]. It is also possible that after the diagnosis of a medical complication such as preeclampsia, the speed and intensity of medical intervention may vary in different healthcare systems. Furthermore, in Brazil theremay be delays of patients in seeking care until extreme situations. Analysing the cases in NETDC over the past 5 decades, there was no significant shift in disease presentation unlike that seen in singleton complete moles [10]. Even with the evident progress in perinatal med- icine over the years, an improvement of obstetric outcomes of CHMCF was not observed, supporting the notion that a determining factor in CHMCF could be related to trophoblastic biology and not simply the management of the disease. Some of the main questions addressed when caring for a patient with CHMCF are the possibility of reaching fetal viability, possible med- ical complications and the risk of GTN. The literature reports a chance of delivering a viable infant varying from 20 to 71%, when not taking elec- tively terminated pregnancies into account (Table 5) [4,6–9,19–22]. This study found a 60% chance of delivering a viable fetus if expectant management is adopted for CHMCF. Therewere no cases of intrauterine death after 24weeks, suggesting that there is no need to induce delivery after viability without an obvious maternal or fetal indication. Most studies in the literature report a similar rate of spontaneous abortion and intrauterine death between 20 and 30% [4,6–9,19–22]. The rate of medical complications (63%) and potentially life-threat- ening conditions (56%) were quite high in the overall data; however the need to medically terminate the pregnancy was 10%. The need for pregnancy termination due to medical complications in the literature varies between 0 and 71%, since the utilization of this method may de- pend on multiple factors, including the social and cultural acceptance of this practice in a particular setting and the resources of the hospital involved in the care of CHMCF (Table 5) [4,6–9,19–22]. Compared to single completemoles in a similar timeperiod, CHMCF in this study pre- sented with significantly higher rates of preeclampsia, clinical hyper- thyroidism, potentially life-threatening conditions and maternal near miss events. The only maternal death in this cohort was caused by a se- vere acute respiratory insufficiency during a medical termination of Fig. 1. Overview of the study population and obstetric outcomes. 92 L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 pregnancy due to severe preeclampsia; death was likely caused by a massive trophoblastic embolization associated with pulmonary edema. Across the studies in the literature, the frequency of clinical complications are variable, possibly due to different classification of the conditions and availability of the retrieved data, since some of the centers were not directly involved in the care of the pregnancy. The rate of vaginal hemorrhage in this study (59%) was higher than the other studies in the literature (14–34%) because we did not limit this definition to only the severe cases that led to termination of pregnancy [6,7,22]. The frequency of hyperthyroidism of this study was lower (14%) than reported by others in the literature (28–50%) due to the fact that only cases with clinical signs were included in the analysis and not all of those with laboratory finding of thyroid hormonal unbal- ance [7,20,22]. The preeclampsia rate in the literature ranges from 3% to 40%, and this study iswithin this range (32%) [4,6–9,19,20,22]. Since not all studies provided the definition used for preeclampsia the interpreta- tion of the variability of this finding was limited; some studies only had the information of this complication leading to pregnancy termination, whichwould represent amuch lower rate than the overall preeclampsia rate. In the study by Kihara et al., preeclampsia was found to be a factor associated with poor fetal outcome; however this finding was not seen in this study, since the presence of preeclampsia was not associated with fetal viability (p = 0.775) [19]. Another controversial issue lies in the risk of CHMCF progressing to GTN. The studies in the literature reported a GTN rate between 14 and 57% (Table 5) [1,4,6–9,19–22]. The largest study regarding this matter is from the Charing Cross and Weston Park Trophoblastic Disease Cen- ters, which included 90 cases of CHMCF with a risk of 27% for GTN, slightly higher than single complete moles [9]. A national survey con- ducted by Matsui et al. [6], also reported an increased GTN rate (31%) across 72 cases of CHMCF in Japan. The current study showed a higher risk (46%) when compared to single complete moles (18%) in a similar time period [5]. This disparity could be explained by at least two rea- sons: an actual different disease behavior or due to overrepresentation of GTNcases in a referral center versus population-based data. However, given the rarity and the doubts in the management of CHMCF, patients who are recognized to have this conditionwill tend to be treated in a re- ferral center. Similar to the findings in the Charing Cross experience, the present study did not find any difference related to elective terminations of pregnancy and GTN rate, confirming that the decision to terminate a CHMCF should not be based on the risk of developingGTN [1]. A striking finding in this study was the association between GTN and the need for medical termination due to severe complications and higher hCG levels at the end of pregnancy. The production of hCG reflects the behavior of trophoblastic cells and particularly high hCG levels are considered a marker of high-risk for developing GTN. In CHMCF, an increasing hCG trend during pregnancy was associated with poorer fetal outcome, as the fall of hCG levels may be correlated with the shrinkage of the molar mass [19,22,23]. The results suggest that trophoblastic cell Table 3 Comparison of clinical factors of multiple pregnancies with complete mole and coexisting normal fetus which progressed to gestational trophoblastic neoplasia (GTN) and those with spontaneous remission in all time periods and Referral Centers. GTNa Regression p value Number of cases 31 37 Age, median (minimum–maximum) 30 (18–41) 27 (18–41)b 0.054 Gravidity, median (minimum–maximum) 2 (1–5)b 2 (1–5) 0.522 Parity, median (minimum–maximum) 0 (0–4)b 0 (0–4) 0.960 Previous mole, n (%) 1 (3%)b 2 (5%) 1 Assisted reproduction treatment, n (%) 5 (16%) 4 (11%) 0.500 Gestational age at the end of pregnancy, weeks, median (minimum–maximum) 17 (11–39)c 28.5 (11–40)b 0.002 hCG levels at end of pregnancy, mIU/mL, median (minimum–maximum) 250,000 (33,648–1,670,600)d 120,000 (1048–1,365,080)e 0.026 Medical complications, n (%) 18 (64%)f 22 (65%)f 1 Termination of pregnancy before 24 weeks, n (%)g Elective termination 7 (23%) 3 (8%) 0.098 Termination due to medical complications 6 (20%) 0 (0%) 0.006 No termination 17 (57%) 34 (92%) 0.001 Pregnancy outcomes, n (%)g Miscarriage 8 (27%) 3 (8%) 0.095 Intrauterine death 1 (3%) 5 (13%) 0.213 Preterm delivery 7 (23%) 18 (49%) 0.043 Term delivery 1 (3%) 8 (22%) 0.035 Termination of pregnancy 13 (43%) 3 (8%) 0.001 Fetal viability, n (%) 8 (27%)b 25 (68%)b 0.004 Hysterotomy, n (%) 13 (43%) 19 (53%)b 0.470 Peripartum hysterectomy, n (%) 3 (10%) 6 (17%)b 0.494 Potentially life-threatening conditions, n (%) 17 (59%)c 18 (53%)h 0.800 Maternal near miss events, n (%) 5 (17%)c 4 (12%)h 0.721 hCG: human chorionic gonadotropin hormone; GTN: gestational trophoblastic neoplasia. a GTN progression was not available in 4 cases. b Information not available in 1 patient. c Information not available in 2 patients. d Information not available in 5 patients. e Information not available in 7 patients. f Information not available in 7 patients. g Information not available in 1 patient in the GTN group. h Information not available in 3patients. 93L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 biology may influence clinical presentation and behavior of disease and more aggressive molar pregnancies would have a higher likelihood to developing GTN [24]. The strengths of this study are the sample size, the availability of in- formation on pregnancy related events and the possibility of evaluating different regional settings by the same research team. One of the limita- tions of this study was its retrospective nature, however due to the rar- ity of CHMCF it is not feasible to perform a prospective study. Due to the extensive period of time evaluated at the NETDC there was substantial missing data, especially from the older cases which date back to 1966, Table 4 Comparisons of clinical characteristics of single completemoles from a recent published cohort f combined data from NETDC and Brazilian Trophoblastic Disease Centers (BTDC) on multiple p Sing Number of cases 194 Age, median (minimum–maximum) 30 ( Gravidity, median (minimum–maximum) 2 (1 Parity, median (minimum–maximum) 1 (0 Gestational age at diagnosis, weeks, median (minimum–maximum) 9 (5 Gestational age at the end of pregnancy, weeks, median (minimum–maximum) 9 (5 hCG levels at diagnosis, mIU/mL, median (minimum–maximum) 164 hCG levels at end of pregnancy, mIU/mL, median (minimum–maximum) 164 Preeclampsia, n (%) 2 (1 Clinical hyperthyroidism, n (%) 4 (2 GTN progression, n (%) 34 ( Potentially life-threatening conditions, n (%) 4 (2 Maternal near miss events, n (%) 0 (0 hCG: human chorionic gonadotropin hormone; GTN: gestational trophoblastic neoplasia. a Information not available in 17 patients. b Information not available in 2 patients. c Information not available in 13 patients. d Information not available in 6 patients. e Information not available in 3 patients. f Information not available in 6 patients and 2 patients were excluded for prophylactic chem limiting statistical comparisons and interpretation. One of the weak- nesses of the study designwas the fact that data was collected from dif- ferent centers, possibly with unequal resources and heterogeneous CHMCFmanagement; although all centers are University tertiary hospi- tals with a specialized multiprofessional team in Trophoblastic Disease and Maternal Fetal Medicine. In summary, the current study indicates that theremight be a differ- ence in disease presentation of CHMCF between geographical locations that could be explained by regional disparities. Importantly, 60% of the patients with expectant management of their CHMCF delivered a viable romNewEngland Trophoblastic Disease Center (NETDC) between 1994 and 2013with the regnancies with complete mole and coexisting normal fetus (CHMCF) (1990–2015). le moles NETDC 1994–2013 CHMCF in NETDC & BTDC 1990–2015 p value 59 22–35) 29 (18–41) 0.872 −10) 2 (1–5) 0.048 –8) 0 (0–4) 0.395 –22)a 15 (9–23)b b0.001 –22)a 23 (11–40) b0.001 ,579 (183–4,718,955)c 400,000 (2545–2,052,500)d b0.001 ,579 (183–4,718,955)c 200,000 (18,361–1,670,600)d 0.161 %) 18 (32%)e b0.001 %) 9 (16%)e b0.001 18%)f 26 (46%)b b0.001 ) 32 (56)b b0.001 ) 10 (17)b b0.001 otherapy. Table 5 Major studies reporting clinical presentation and outcomes of multiple pregnancies with complete mole and coexisting normal fetus. Matsui et al. [6] Kihara et al. [19] Lee et al. [22] Kutuk et al. [20] Niemann et al. [8] Massardier et al. [7] Niemann et al. [9] Wee & Jauniaux [4] Fishman et al. [21] Lin et al. (2017)d Number of cases 18 and 72a 15 6 7 8 14 90 8 7 72 Time frame NA 1991–2011 1998–2008 2007–2012 1986–2003 1999–2006 1998–2011 1996–2004 1966–1997 1966–2016 Setting Japan – national survey Japan – Chiba Korea - Seoul Turkey Denmark France-Lyon UK-Charing Cross & Weston Park Hospitals UK – UCLH & KCH USA - Chicago USA - Boston & Brazil Vaginal hemorrhage, n (%) 3 (17%)b NA 2 (34%)b NA 0 (0%) 2 (14%)b NA NA NA 37 (59%) Preeclampsia, n (%) 5 (28%)b 6 (40%) 1 (17%) 1 (14%) 0 (0%) 4 (29%) 3 (3%)b 1 (12%) NA 20 (32%) Hyperthyroidism, n (%) NA NA 3 (50%) 2 (28%) 0 (0%) 4 (29%) NA NA NA 9 (14%) Elective TOP, n (%) 5 (28%) 2 (28%) 1 (17%) 2 (28%) 5 (62%) 3 (21%) 39 (43%)e 1 (12%) 0 (0%) 10 (14%) TOP due to maternal complications, n (%) 8 (44%) 0 (0%) 2 (34%) 0 (0%) 0 (0%) 4 (29%) 5 (6%) 0 (0%) 5 (71%) 7 (10%) Miscarriage or intrauterine death, n (%) 2 (11%) 4 (57%) 2 (34%) 2 (25%) 2 (25%) 3 (21%) 17 (19%)f 2 (25%) NA 17 (24%) Preterm delivery, n (%) 1 (6%) 1 (14%) 0 1 (12%) 1 (12%) 3 (21%) NA 2 (25%) 2 (28%) 25 (36%) Term delivery, n (%) 2 (11%) 0 (0%) 1 (17%) 0 (0%) 0 (0%) 1 (7%) NA 3 of 8 (37%) 0 (0%) 11 (16%) Delivery of a viable infant in ongoing pregnancy, n (%) 3 of 13 (23%) 1 of 5 (20%) 1 of 5 (20%) 1 of 3 (34%) 1 of 3 (34%) 3 of 11 (27%) 29 of 51 (57%) 5 of 7 (71%) 2 of 7 (28%) 36 of 60 (60%) GTN progression, n (%) 9 of 18 (50%) 22 of 72 (31%) 1 (14%) 3 (50%) 1 (14%) 2 (25%) 7 (50%) 24 (27%) 3 (37%) 4 (57%) 31 (46%) NA: not available; TOP: termination of pregnancy;UCLH: University College LondonHospital; KCH: Kings CollegeHospital hCG: human chorionic gonadotropin hormone;GTN: gestational trophoblastic neoplasia. a The study by Matsui et al. [6] included analysis of 72 cases in the GTN progression rate and 18 cases for pregnancy presentation and outcomes. b Need for termination of pregnancy. d Missing data for clinical complications in 9 patients, for pregnancy outcomes in 2 patients and GTN progression in 4 patients. e The data on electively terminated pregnancies from the study by Niemann et al. [9] is mixed with cases of spontaneous miscarriages before 14 weeks. f Data on miscarriages and intrauterine deaths in the study by Niemann et al. [9] accounts only those that occurred after 14 weeks. 94 L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 infant. Elective termination of pregnancy did not influence the risk for GTN progression; however the need for termination of pregnancy due to medical complications and higher hCG levels were associated with development of GTN in CHMCF. These findings should contribute to a better understanding of the natural history andmanagement of CHMCF. Conflict of interest The authors declare no conflict of interest regarding the publication of this research. Financial support Wewish to acknowledge the support of the Donald P. Goldstein,MD Trophoblastic Tumor Registry Endowment; the Dyett Family Tropho- blastic Disease Research and Registry Endowment; University of Sao Paulo; and Carlos Chagas Filho Foundation for Research Support in the State of Rio de Janeiro/Brazil (FAPERJ) – an agency under the Brazilian Ministry of Science and Technology (AB). The funding agencies had no direct role in the generation of the data or the manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2017.01.021. References [1] N.J. Sebire, M. Foskett, F.J. Paradinas, R.A. Fisher, R.J. Francis, D. Short, E.S. Newlands, M.J. Seckl, Outcome of twin pregnancies with complete hydatidiform mole and healthy co-twin, Lancet 359 (2002) 2165–2166. [2] M.A. Steller, D.R. Genest, M.R. Bernstein, J.M. Lage, D.P. Goldstein, R.S. Berkowitz, Natural history of twin pregnancy with complete hydatidiformmole and coexisting fetus, Obstet. Gynecol. 83 (1994) 35–42, http://dx.doi.org/10.1016/0020- 7292(94)90494-4. [3] L.O. Vejerslev, Clinical management and diagnostic possibilities in hydatidiform mole with coexistent fetus, Obstet. Gynecol. Surv. 46 (1991) 577–588. [4] L. Wee, E. Jauniaux, Prenatal diagnosis and management of twin pregnancies com- plicated by a co-existing molar pregnancy, Prenat. Diagn. 25 (2005) 772–776, http://dx.doi.org/10.1002/pd.1272. [5] S.Y. Sun, A. Melamed, N.T. Joseph, A.A. Gockley, D.P. Goldstein, M.R. Bernstein, N.S. Horowitz, R.S. Berkowitz, Clinical presentation of complete hydatidiform mole and partial hydatidiform mole at a regional trophoblastic disease center in the United States over the past 2 decades, Int. J. Gynecol. Cancer 26 (2016) 367–370, http:// dx.doi.org/10.1097/IGC.0000000000000608. [6] H. Matsui, S. Sekiya, T. Hando, N. Wake, Y. Tomoda, Hydatidiform mole coexistent with a twin live fetus: a national collaborative study in Japan, Hum. Reprod. 15 (2000) 608–611, http://dx.doi.org/10.1016/S0020-7292(00)83458-X. [7] J. Massardier, F. Golfier, D. Journet, L. Frappart, M. Zalaquett, A.M. Schott, V.T. Lenoir, O. Dupuis, T. Hajri, D. Raudrant, Twin pregnancy with complete hydatidiform mole and coexistent fetus. Obstetrical and oncological outcomes in a series of 14 cases, Eur. J. Obstet. Gynecol. Reprod. Biol. 143 (2009) 84–87, http://dx.doi.org/10.1016/ j.ejogrb.2008.12.006. [8] I. Niemann, L. Sunde, L.K. Petersen, Evaluation of the risk of persistent trophoblastic disease after twin pregnancy with diploid hydatidiformmole and coexisting normal fetus, Am. J. Obstet. Gynecol. 197 (2007) 1–5, http://dx.doi.org/10.1016/j.ajog.2007. 02.038. [9] I. Niemann, R. Fisher, N. Sebire, M. Wells, D. Short, J. Tidy, B. Hancock, R. Coleman, P. Savage, M. Seckl, Update on UK outcomes for women with twin pregnancies comparimising a complete hydatidiform mole (CHM) and a normal co-twin, XVII World Congr. Gestation. Trophobl. Dis 2013, p. 57 (abstract). [10] S.Y. Sun, A. Melamed, D.P. Goldstein, M.R. Bernstein, N.S. Horowitz, A.F. Moron, I. Maesta, A. Braga, R.S. Berkowitz, Changing presentation of com- plete hydatidiform mole at the New England Trophoblastic Disease Center over the past three decades: does early diagnosis alter risk for gestational trophoblastic neoplasia? Gynecol. Oncol. 138 (2015) 46–49, http://dx.doi. org/10.1016/j.ygyno.2015.05.002. [11] R.R. Soares, I. Maestá, J. Colón, A. Braga, A. Salazar, R.C. Charry, S.Y. Sun, D.P. Goldstein, R.S. Berkowitz, Complete molar pregnancy in adolescents from North and South America: clinical presentation and risk of gestational trophoblastic neo- plasia, Gynecol. Oncol. 2016 (2016) 496–500, http://dx.doi.org/10.1016/j.ygyno. 2016.07.002. [12] E. von Elm, D. Altman, M. Egger, S. Pocock, P. Gøtzsche, J. Vandenbroucke, STROBE Initiative, The Strengthening the Reporting of Observational Studies in Epidemiolo- gy (STROBE) statement: guidelines for reporting observational studies, Lancet 370 (9596) (Oct 20, 2007) 1453–1457. [13] L. Say, J.P. Souza, R.C. Pattinson, Maternal near miss - towards a standard tool for monitoring quality of maternal health care, Best Pract. Res. Clin. Obstet. Gynecol. 23 (2009) 287–296, http://dx.doi.org/10.1016/j.bpobgyn.2009.01.007. http://dx.doi.org/10.1016/j.ygyno.2017.01.021 http://dx.doi.org/10.1016/j.ygyno.2017.01.021 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0005 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0005 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0005 http://dx.doi.org/10.1016/0020-7292(94)90494-4 http://dx.doi.org/10.1016/0020-7292(94)90494-4 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0015 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0015 http://dx.doi.org/10.1002/pd.1272 http://dx.doi.org/10.1097/IGC.0000000000000608 http://dx.doi.org/10.1016/S0020-7292(00)83458-X http://dx.doi.org/10.1016/j.ejogrb.2008.12.006 http://dx.doi.org/10.1016/j.ejogrb.2008.12.006 http://dx.doi.org/10.1016/j.ajog.2007.02.038 http://dx.doi.org/10.1016/j.ajog.2007.02.038 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0045 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0045 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0045 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0045 http://dx.doi.org/10.1016/j.ygyno.2015.05.002 http://dx.doi.org/10.1016/j.ygyno.2016.07.002 http://dx.doi.org/10.1016/j.ygyno.2016.07.002 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0060 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0060 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0060 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0060 http://dx.doi.org/10.1016/j.bpobgyn.2009.01.007 95L.H. Lin et al. / Gynecologic Oncology 145 (2017) 88–95 [14] World Health Organization, WHO Recommendations on Interventions to Improve Preterm Birth Outcomes, https://www.ncbi.nlm.nih.gov/pubmed/264472642015. [15] H.Y.S. Ngan, E.I. Kohorn, L.A. Cole, R.J. Kurman, S.J. Kim, J.R. Lurain, M.J. Seckl, S. Sasaki, J.T. Soper, Trophoblastic disease, Int. J. Gynecol. Obstet. 119 (2012) S130–S136, http://dx.doi.org/10.1016/S0020-7292(12)60026-5 (Suppl.). [16] S.Y. Sun, D.P. Goldstein, M.R. Bernstein, N.S. Horowitz, R. Mattar, I. Maesta, A. Braga, R.S. Berkowitz, Maternal near miss according toworld health organization classifica- tion among women with a hydatidiform mole: experience at the New England Tro- phoblastic Disease Center, 1994–2013, J. Reprod. Med. 61 (2016) 210–214. [17] M.S.F.S. Cagayan, Hydatidiform mole and its complications: review of patient pro- files and management at the University of the Philippines - Philippine General Hos- pital, J. Reprod. Med. (2014) 235–240 0024-7758/14/5905-06-235/$18.00/0. [18] P. Elliott, N. Best, Geographic patterns of disease, Wiley StatsRef Stat. Ref. Online 2014, pp. 1–7. [19] M. Kihara, H. Usui, H. Tanaka, H. Inoue, H. Matsui, M. Shozu, Complicating pre- eclampsia as a predictor of poor survival of the fetus in complete hydatidiform mole coexistent with twin fetus, J. Reprod. Med. 57 (2012) 325–328. [20] M.S. Kutuk, M.T. Ozgun, M. Dolanbay, C. Batukan, S. Uludag, M. Basbug, Sonographic findings and perinatal outcome of multiple pregnancies associating a complete hydatidiform mole and a live fetus: a case series, J. Clin. Ultrasound 42 (2014) 465–471, http://dx.doi.org/10.1002/jcu.22169. [21] D.A. Fishman, L.A. Padilla, P. Keh, L. Cohen, M. Frederiksen, J.R. Lurain, Management of twin pregnancies consisting of a complete hydatidiform mole and normal fetus, Obstet. Gynecol. 91 (1998) 546–550, http://dx.doi.org/10.1016/S0029- 7844(97)00720-5. [22] S.W. Lee, M.Y. Kim, J.H. Chung, J.H. Yang, Y.H. Lee, Y.K. Chun, Clinical findings of mul- tiple pregnancy with a complete hydatidiform mole and coexisting fetus, J. Ultra- sound Med. 29 (2010) 271–280. [23] M. Okumura, K. Fushida, R. Francisco, R. Schultz, M. Zugaib, Massive necrosis of a complete hydatidiform mole in a twin pregnancy with a surviving coexistent fetus, J. Ultrasound Med. 33 (2014) 177–179, http://dx.doi.org/10.7863/ultra.33.1. 177. [24] R.S. Berkowitz, D.P. Goldstein, Clinical practice. Molar pregnancy, N. Engl. J. Med. 360 (2009) 1639–1645, http://dx.doi.org/10.1097/01.SA.0000358600.18446.e2. https://www.ncbi.nlm.nih.gov/pubmed/26447264 http://dx.doi.org/10.1016/S0020-7292(12)60026-5 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0080 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0080 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0080 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0080 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0090 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0090 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0095 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0095 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0095 http://dx.doi.org/10.1002/jcu.22169 http://dx.doi.org/10.1016/S0029-7844(97)00720-5 http://dx.doi.org/10.1016/S0029-7844(97)00720-5 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0110 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0110 http://refhub.elsevier.com/S0090-8258(17)30065-3/rf0110 http://dx.doi.org/10.7863/ultra.33.1.177 http://dx.doi.org/10.7863/ultra.33.1.177 http://dx.doi.org/10.1097/01.SA.0000358600.18446.e2 Multiple pregnancies with complete mole and coexisting normal fetus in North and South America: A retrospective multicenter... 1. Introduction 2. Methods 2.1. Study design and setting 2.2. Variables and definitions 2.3. Statistical methods 2.4. Literature review 3. Results 4. Discussion Conflict of interest Financial support Appendix A. Supplementary data References