Wolfberg, Adam J. MD, MPH4,5,7; Berkowitz, Ross S. MD1,2,4,6,7; Goldstein, Donald P. MD1,2,4,6,7; Feltmate, Colleen MD1,2,4,6,7; Lieberman, Ellice MD, DrPH3,4
Women diagnosed with a complete hydatidiform molar pregnancy are typically counseled that their risk of developing persistent gestational trophoblastic neoplasia (GTN) requiring further management with chemotherapy is 15–28%.1–5 To diagnose persistent GTN promptly, patients are followed up with serial human chorionic gonadotropin (hCG) levels after molar evacuation. A plateau or rise in hCG levels indicates persistent disease and the need for chemotherapeutic management. Complete remission is typically declared if the hCG level spontaneously declines to undetectable levels and remains there during a 6-month follow-up period.6
Investigators have tried to identify factors that would predict either persistent disease or remission to inform women and their physicians about their risk of developing persistent disease or reduce the interval between the diagnosis of molar pregnancy and the diagnosis of persistent disease. A means to predict a high risk of persistence might allow empiric chemotherapeutic management in the noncompliant patient at high risk for developing persistent disease; a method of predicting remission could reassure a woman at low risk for persistent disease.
Published reports describing factors that predict persistent disease have not met these criteria: Uterine size greater than dates and bilateral ovarian enlargement (> 8 cm) on presentation were found to be weakly predictive of developing persistent disease.4 Histopathologic characterization of the molar gestation was not consistently shown to be helpful in predicting persistent disease.7 Although the ratio of free β-hCG to hCG was higher in molar pregnancies than in normal pregnancies, the ratio was not helpful in distinguishing persistent disease from nonpersistent disease.8
Most investigators have focused on hCG as a tumor marker. Preevacuation hCG levels above 100,000 mIU/mL have been found to predict persistent disease, but never with sufficient precision to guide management.7 In an effort to find hCG levels predictive of outcome, several investigators have mathematically evaluated the curves developed by plotting hCG values against time from evacuation to determine whether patients who go on to have persistent disease have hCG regression patterns that are different from those who go into remission.9–12 However, none have been sufficiently accurate to guide management. In a previous article, we reported that women with complete molar pregnancies were very unlikely to develop persistent neoplasia after their hCG values spontaneously declined to undetectable levels.1 Using the same group of women with complete molar pregnancies, we attempted to identify hCG criteria that were predictive of persistent disease or of remission.
MATERIALS AND METHODS
We used the Donald P. Goldstein M.D. Trophoblastic Tumor Registry of the New England Trophoblastic Disease Center, a clinical unit founded in 1965 to care for patients with gestational trophoblastic neoplasia and to serve as a referral center for physicians in the community caring for their own patients with these diseases. We analyzed data from 1973, the year the hCG radioimmunoassay became commercially available in the United States, through 2001 and identified 2,120 women with a diagnosis of complete molar pregnancy on the basis of a pathology report.
Patients with partial molar pregnancies or other gestational trophoblastic neoplasms were not included in our analysis. Women with less than 6 months of follow-up (N = 877) were excluded from the analysis. Of the 877 subjects, 817 (93.2%) were lost to follow-up after their hCG assay fell to undetectable, and 60 (6.8%) were lost to follow-up without their hCG falling to undetectable. We have no record of any patient who was lost to follow-up subsequently being diagnosed with persistent neoplasm. We also excluded the 173 women with data missing for either their evacuation or their pathology report. Finally, we excluded 41 subjects from the analysis because they did not meet current hCG criteria for persistent neoplasm but received chemotherapy anyway. The 1,029 remaining women comprised our study population.
We then categorized patients with complete molar pregnancies into 2 groups on the basis of their subsequent hCG testing: persistent disease and spontaneous remission. Persistence was diagnosed when a patient's hCG level increased for 2 or more values, or if it failed to decline by more than 10% over 3 weekly hCG values (plateauing).13,14 Remission was diagnosed after 6-months of follow-up hCG testing after molar evacuation with undetectable hCG values. Comparison between these groups was made using χ2 or Fisher exact test as appropriate.
We evaluated all available hCG data for women who had more than 6 months of hCG data. Patients who subsequently became pregnant during the 6-month follow-up interval were diagnosed with remission if they had evidence of a normal pregnancy without evidence of choriocarcinoma or placental site tumor.
Next, we calculated the proportion of women developing persistent disease with an hCG level within a specific numeric range during 1-week intervals after molar evacuation, and calculated 95% confidence intervals for these risk levels. All analyses were performed using software from SAS Institute, Inc (Cary, NC) (Table 1). We focused on the first 8 weeks after evacuation, before most women had either been diagnosed with persistent disease or had reached undetectable hCG levels. hCG values measured after a woman had been diagnosed with persistent gestational trophoblastic disease were not included in the analysis. This study was approved by the human research committee at our institution.
Overall, persistent neoplasm was diagnosed in 153 of the 1,029 women (15%) on the basis of hCG criteria. Women developing persistent GTN were older than women who spontaneously entered remission (mean 27.7 years compared with 25.9 years, P < .01), and were somewhat more likely to have an hCG level on presentation above 100,000 mIU/mL (20.3% compared with 14.7%, P = .08), although the difference did not reach statistical significance. Women who developed persistent disease were not more likely to have uterine size greater than dates on presentation (62.7% compared with 56.3%, P = .4), to be multiparous (48.2% compared with 49.9%, P = .7), or to have had a prior mole (1.5% compared with 1.9%, P = .8) than women who spontaneously entered remission (Table 2).
Women whose hCG level declined below 50 mIU/mL during their follow-up were found to be at no more than 1.1% risk for developing persistent GTN, irrespective of when this level was reached. Women whose hCG level was below 200 mIU/mL in the fourth week after evacuation (59.8% of all women) or below 100 mIU/mL in the sixth week after evacuation (65.8% of all women), had a risk of persistence below 9% (Table 1).
Women with hCG levels above 2,000 mIU/mL in the third or fourth week after evacuation were at a 59.2% and 63.8% risk of developing persistent disease, respectively (Table 1). In the fourth week after evacuation, 13.3% of women had hCG above or equal to 2,000 mIU/mL (Figure 1).
More than one half of the women with hCG measured in the third week after molar evacuation had the estimate of their risk of developing persistent GTN modified: either increased above 50% or reduced to below 7% (Table 1).
Although an initial hCG more than 100,000 mIU/mL may be a risk factor for developing persistent GTN, women with hCG within a given range after evacuation were not more likely to develop persistent GTN if their initial hCG level was greater than 100,000 mIU/mL than if it was initially less than 100,000 mIU/mL. In the fourth week after molar evacuation, none of the women with hCG less than 50 mIU/mL developed persistent disease. Women with an initial hCG less than 100,000 mIU/mL and a hCG in the fourth week of 2,000 or greater had a 80% risk of developing persistent disease compared with a 60% risk of developing persistent disease for women with an initial hCG more than 100,000 mIU/mL and a hCG that week of 2,000 or greater (P = .6). We had initial hCG data (within 48 hours of evacuation) on 272 of the 518 women with hCG values measured in the fourth week after evacuation. Initial hCG values were available on a similar proportion of subjects with hCG measured in other weeks. Irrespective of the period evaluated or hCG range, the initial hCG level failed to affect the risk of persistent GTN in our analysis of hCG values in the third through eighth week after molar evacuation.
We previously reported our finding that women diagnosed with a complete molar pregnancy who have a spontaneous decline in their hCG to undetectable levels may not be at risk for developing persistent GTN.1
In our current analysis, we found that women who achieve hCG levels below 50 mIU/mL at any time during their follow-up may be reassured that their risk of developing persistent GTN is exceedingly low (≤ 1.1%). An additional group of women with hCG levels below 200 mIU/mL in the fourth week after evacuation or below 100 mIU/mL in the sixth week after evacuation may be reassured that their risk of developing persistent GTN is significantly diminished (≤ 9%) from the 15% risk they faced at the time of their initial evacuation.
Our analysis may allow clinicians to stratify by risk their patients with complete molar pregnancy during the initial weeks after molar evacuation and to provide more accurate counseling based on hCG levels obtained in the weeks after molar evacuation.
We corroborated prior reports describing an association between older maternal age and an increased risk of persistent GTN15 and found a modest trend toward an initial hCG level above 100,000 mIU/mL and an increased risk of persistent GTN (P = .08).7 We did not find the previously reported associations between uterine size greater than dates4 and a history of prior mole,16 and an increased risk of persistent GTN.
Our analysis also identified a small group of women—those with hCG levels above 2,000 mIU/mL in the third or fourth week after evacuation—whose risk of developing persistent GTN exceeded 50%. These criteria may be used by clinicians to identify a group of women at high risk for developing persistent GTN who may benefit from empiric chemotherapy if hCG follow-up is either unavailable or unreliable, circumstances that may be particularly common in geographic areas where limited resources or long distances make a weekly hCG testing regimen impossible. In these circumstances, the benefit of empiric chemotherapy might outweigh the risk of overtreatment.
In the third week after evacuation, more than one half of women with complete mole will have the estimate of their risk of developing persistent GTN modified—either halved or tripled—from their baseline risk at the time of diagnosis. This may significantly alter counseling that can be offered to these patients after they are diagnosed with this condition.
We believe that our data describing the decline in hCG levels after molar evacuation have an underlying biologic basis. Women with molar pregnancies who have trophoblastic tissue confined to the endometrial cavity are likely to have a rapid fall in their hCG level after the trophoblastic tissue is removed at surgery and a low risk of developing persistent disease. In contrast, women whose trophoblastic tissue has invaded the uterine wall, or metastasized beyond the uterus, will have a slower fall in their hCG level after surgery due to the presence of the residual hCG-producing tissue. These women are probably at a higher risk of developing persistent disease. A slow decrease in hCG levels after molar evacuation may, therefore, reflect the presence of invasive or metastatic trophoblastic tissue.
The number of women with hCG levels in each of the weeks after molar evacuation are far fewer than the 1,029 in the overall population, reflecting poor compliance with weekly hCG testing among our patients. In later weeks, the small numbers begin to reflect those women who switch from weekly to monthly hCG testing after reaching undetectable levels, those women who have been diagnosed with persistent disease, as well as poor compliance with testing recommendations.
By 6 weeks after evacuation, approximately one half of women who will ultimately develop persistence have been diagnosed with this condition. This may explain the weekly decline in the risk of developing persistent disease at high hCG levels, because the number of women who have yet to be diagnosed with persistent GTN (the numerator) decline, whereas the number of women who are being followed up with serial hCG and will ultimately be declared in remission (the denominator) remains relatively static.
Our data are specific to complete hydatiform molar pregnancies and do not address partial molar pregnancies that are one fifth as likely to develop persistent GTN after evacuation as complete molar pregnancies.13
We also acknowledge that our results could be confounded by referral bias, because it is conceivable that more of our subjects had the kind of complicated—and possibly prolonged—follow-up course necessitating referral to a tertiary care center that specializes in gestational trophoblastic disease. It should be noted that we have no record of a subject re-presenting to our center or any other center with persistent disease after being lost to follow-up.
After evacuation of a complete mole, hCG levels may predict the risk of persistent GTN. Levels of hCG after evacuation may be employed by clinicians to either reassure patients that they are low risk of persistence, or to identify patients at high risk of persistent GTN. If patients are at high risk of developing persistent GTN, and are also at high risk of incomplete follow-up, they may benefit from empiric chemotherapy.
1. Wolfberg AJ, Feltmate C, Goldstein DP, Berkowitz RS, Lieberman E. Low risk of relapse after achieving undetectable hCG levels in women with complete molar pregnancy. Obstet Gynecol 2004;104:551–4.
2. Lurain JR, Brewer JI, Torok EE, Halpern B. Natural history of hydatidiform mole after primary evacuation. Am J Obstet Gynecol 1983;145:591–5.
3. Kohorn EI. Hydatidiform mole and gestational trophoblastic disease in Southern Connecticut. Obstet Gynecol 1982;59:78–84.
4. Curry SL, Hammond CB, Tyrey L, Creasman WT, Parker RT. Hydatidiform mole: diagnosis, management, and long-term followup of 347 patients. Obstet Gynecol 1975;45:1–8.
5. Yuen BH, Cannon W. Molar pregnancy in British Columbia: estimated incidence and post-evacuation regression patterns of the beta subunit of human chorionic gonadotropin. Am J Obstet Gynecol 1981;139:316–9.
6. Goldstein DP, Berkowitz RS, Bernstein MR. Management of molar pregnancy. J Reprod Med 1981;26:208–12.
7. Ayhan A, Tuncer ZS, Halilzade H, Kucukali T, Predictors of persistent disease in women with complete hydatidiform mole. J Reprod Med 1996. 41:591–4.
8. Berkowitz R, Ozturk M, Goldstein D, Bernstein M, Hill L, Wands JR. Human chorionic gonadotropin and free subunits' serum levels in patients with partial and complete hydatidiform moles. Obstet Gynecol. 1989;74:212–5.
9. Smith EB, Szulman AE, Hinshaw W, Tyrey L, Surti U, Hammond CB. Human chorionic gonadotropin levels in complete and partial hydatidiform moles and in nonmolar abortuses. Am J Obstet Gynecol. 1984;149:129–32.
10. Morrow CP, Kletzky OA, Disaia PJ, Townsend DE, Mishell DR, Nakamura RM. Clinical and laboratory correlates of molar pregnancy and trophoblastic disease. Am J Obstet Gynecol 1977;128:424–30.
11. Schlaerth JB, Morrow CP, Kletzky OA, Nalick RH, D'Ablaing GA. Prognostic characteristics of serum human chorionic gonadotropin titer regression following molar pregnancy. Obstet Gynecol 1981;58:478–82.
12. Shigematsu T, Kamura T, Saito T, Kaku T, Nakano H, Kinugawa N. Identification of persistent trophoblastic diseases based on a human chorionic gonadotropin regression curve by means of a stepwise piecewise linear regression analysis after the evacuation of uneventful moles. Gynecol Oncol 1998;71:376–80.
13. Berkowitz RS, Goldstein DP. Chorionic tumors. N Engl J Med 1996;335:1740–8.
14. Kohorn EI. Negotiating a staging and risk factor scoring system for gestational trophoblastic neoplasia. A progress report. J Reprod Med 2002;47:445–50.
15. Xia ZF, Song HZ, Tang MY. Risk of malignancy and prognosis using a provisional scoring system in hydatidiform mole. Chin Med J 1980;93:605–12.
16. Berkowitz RS, Bernstein MR, Laborde O, Goldstein DP. Subsequent pregnancy experience in patients with gestational trophoblastic disease: New England Trophoblastic Disease Center, 1965–1992. J Reprod Med 1994;39:228–32.