ORIGINAL ARTICLES

Gestational trophoblastic disease following complete hydatidi-

form mole in Mulago Hospital, Kampala, Uganda
* Dan K. Kaye
Department of Obstetrics and Gynaecology, Makerere University Medical School,
P. O. Box 7072, Kampala, Uganda.

ABSTRACT
Objectives: To determine epidemiological characteristics and clinical presentation of complete hydatidiform mole (CHM) and complica-
tions associated with prophylactic chemotherapy with oral methotrexate.
Setting: Mulago hospital, Kampala.
Design: Prospective study
Methods: Ninety-four patients with clinically and histologically confirmed complete hydatidiform mole admitted between 1/9/1995 and
30/1/1998 were followed for periods ranging from 12 months to 30 months. Seventy eight (83.0%) received a total of 187 courses of oral
methotrexate (0.4 mg/kg daily in 3 divided doses) as prophylactic chemotherapy. The main outcome measures were pre- and post-
evacuation serum hCG levels and complications associated with oral methotrexate use.
Results: The prevalence of CHM was 3.42 per 1,000 deliveries. The mean age of subjects was 29.6 + 8.5 years. Eighteen women (19.1%)
were nulliparous and mean gravidity was 8.3. Many women presented with high-risk disease. Risk factors for persistent trophoblastic
disease were prior molar pregnancy, age<19 or >35 years and features of high-risk molar pregnancy. Twenty-four of the seventy-eight
patients (30.7%) developed complications, mainly mucositis and haematological toxicity (leucopenia, anaemia and thrombocytopenia),
commonly after 3 or more courses.
Conclusion: CHM was common and many patients presented with high-risk disease. Oral methotrexate for prophylactic chemotherapy
was tolerable and safe for the first 2 courses, but serious complications occur as the duration of treatment increases. Prophylaxis did not
prevent development of (or death from) metastatic trophoblastic disease.
Recommendations: Patients with CHM should be monitored for the development of post-evacuation trophoblastic disease. Those
on prophylactic chemotherapy require close monitoring for the toxic effects of the drugs.
African Health Sciences 2002;2(2):47-51

INTRODUCTION 1.073 per 1000 deliveries (1969-1973). He recommended

Complete hydatidiform mole (CHM), in which there is a policy of routine chemoprophylaxis for all cases of
total degeneration of the trophoblast, is the commonest CHM, a policy which began to be implemented in 1980.
gestational trophoblastic disease (GTD). Histologically, it At that time, many patients who presented with molar
presents with diffuse trophoblastic hyperplasia with vari- pregnancy were apparently lost to follow-up, and pre-
able degrees of cellular atypia 1. Worldwide, there is sented later with malignant GTD. A prevalence of 3.82
marked temporal, regional and ethnic variation in the preva- per 1000 deliveries was reported in 1997 4. Since 1970, no
lence of CHM 1,2. High prevalence is reported in China, study has analyzed the clinical epidemiology of GTD, yet
South East Asia, Northern Europe, the Middle East, and there has been marked improvement in diagnosis of the
Eastern and Central Africa. disease by use of ultrasound scan and assay of Human
In Uganda, Smith 3 found a prevalence of CHM chronic gonadotrophin hormone (hCG). For a drug to
be used for prophylaxis, it must be safe, effective and
convenient. No study has analyzed the complications as-
* Correspondence Author:
sociated with use of Methotrexate for prophylactic che-
Dr. Dan Kaye motherapy against GTD .
Department of Obstetrics and Gynaecology If this prospective study was to describe the epi-
Makerere University Medical School demiological factors and clinical presentation of CHM,
P. O. Box 7072 Kampala, Uganda and complications associated with prophylactic chemo-
E-mail: aogu@africaonline.co.ug Fax. 041-533451 therapy using oral Methotrexate.

African Health Sciences Vol 2 No 2 August 2002 47

METHODS RESULTS
The study was conducted in Mulago hospital, the national During the study period, there were 94 cases of CHM
referral hospital for Uganda, in the Obstetrics and and 27,485 deliveries were seen in Mulago hospital, giving
Gynaecology Department, between 1st September 1995 an incidence rate of 3.42 per 1000 deliveries. The mean
and 30th January 1998. Permission to carry out the study age was 29.6 +8.5 years (range 14 -52). The antecedent
was granted by Mulago hospital Ethics and Research pregnancy was also molar for 10 women (10.6%), while
Committee and the Department of Obstetrics and 16 (17,0%) and 4 (4.3%) had 1 and 2 prior molar preg-
Gynaecology. nancies respectively. Tables 1shows the socio-demographic
The subjects were patients with clinically and his- characteristics and the reproductive history of the 94 cases.
tologically -confirmed CHM admitted to the hospital. CHM was common in adolescents and those aged 35 or
These were followed for periods ranging from 12-30 more years. Six of these patients aborted spontaneously,
months, and were reviewed every 3 months for the first 48 (51.0%) had suction curettage, 16 (17.0%) had manual
year and every 6 months thereafter. On admission, the vacuum aspiration while 24 had dilatation and curettage.
patients had history taken and examination and investiga- None had hysterectomy, or induction of labour with oxy-
tion done (notably ultrasound scan, liver function tests, tocics.
renal function tests and chest X-ray to determine the clini- Table 2 shows the clinical features of the patients
cal features and exclude malignant GTD. Evacuation was both pre- and post-evacuation. Eighty one percent had
done by suction curettage or manual vacuum aspiration, amenorrhea and presented in the second trimester (mean
and formal dilatation and curettage was done to get speci- gestation age 17 weeks) with threatened abortion or his-
mens for histology for those who were referred after tory of molar abortion. Thirty -one (32.7%) had theca
abortion or evacuation. Data was also collected on the lutein cysts, which were bilateral in 25 cases, palpable per
patients’ socio-demographic characteristics (age, age of abdomen in 20 cases and complicated by torsion in 3
spouse, education level, parity, age at first pregnancy), cases. Six women had acute respiratory complications due
menstrual history and contraceptive ever-use. to trophobast pulmonary embolism, however 14(14.9%)
For all patients, a risk score for persistent tropho- had radiological metastases but only 3 developed
blastic disease was made. They were given oral Methotr- haemoptysis. These 3 later developed clinical cerebral
exate 0.4 mg/kg/day (25mg/day) for five days in 3 di- metastases. In those with extra-pelvic lesions, 2 had vagi-
vided doses (10mg, 5mg and 10mg at 6.00 a.m., 2.00 nal nodules, 2 had cervical lesions while 1 had a vulval
p.m., and 10 p.m. respectively) as directly observed therapy ulcer. One patient had profuse haemorrhage from a vagi-
(DOT). At intervals between courses ranging between 9 nal lesion. Twenty-four patients (25.5%) developed per-
and 16 days, 78 patients received 1-8 courses of oral sistent trophoblastic disease, of whom 9 (9.6%) died of
Methotrexate as a single agent (total 187 courses). A nor- GTD. The cause of death (all of whom had received
mal full blood count was a prerequisite for starting therapy. prophylactic chemotherapy) was severe anaemia 2, blood
On taking each dose, an inquiry was made of any com- transfusion reaction 1, malignant GTD with both pulmo-
plaint the patient had, especially related to drug and the nary and cerebral metastases4and pulmonary metastases 4.
staff on duty signed on the treatment sheet. Supportive Seventy-eight women received 187 courses of oral
treatment with intravenous fluids, antibiotics and anti-fungal methotrexate. One patient died after the first course, of
agents was given to those with toxic symptoms of che- severe anaemia (which arose from therapy). Table 3 shows
motherapy (severe mucositis, dermatitis, leucopenia or the courses used and the number who developed com-
hepatitis). Folinic acid was not used. At each review, evalu- plications. The minimum number was 1, but one patient
ation was done by history, examination, serum hCG assay died after the first course and another did not complete
and chest X-ray as a minimum. Remission was diagnosed the second course. Three women had 7 courses and 1
when the hCG levels fell to normal ranges, without clini- had 8 courses. Table 4 shows the nature of complications
cal or imaging evidence of disease by end of 60 days noted in 26 patients (33.3%) on methotrexate. The com-
post-evacuation. Three patients were lost to follow-up monest complications were mucositis and hematological
after 6 months, while 16 did not receive any chemopro- toxicity. Leucopenia, thrombocytopenia and hepatotox-
phylaxis. Data was collected using a coded questionnaire icity were the indication for discontinuation of treatment,
and occurred mainly after the second course.

African Health Sciences Vol 2 No 2 August 2002 48

Table 1. Socio-demographic characteristics and Table 3. Doses of Chemotherapy with Meth
reproductive history of 94 cases. otrexate for CHM patients

Characteristic (n=94) n(%) Courses No. of patients * Complications

(n=78) (%) (%)
Age (years) Less than 20 29 (31.0)
20-34 39 (41.4) 1 2 (2.6) 1 (50.0)
35 or more 26 (27.0)
2 23 (29.5) 2 (8.8)
Education None 17 (18.1)
level Primary 40 (42.5) 3 17 (21.8) 10 (58.8)
Secondary 23 (24.6)
Post-secondary 14 (14.9) 4 12 (15.4) 9 (75.5)
Parity Nulliparous 18 (19.1)
1-4 12 (12.8) 5 7 (9.0) 6 (85.7)
5 or more 64 (68.1)
6 or more 6 (7.7) 3 (50.0)
Antecedent Molar 10 (10.6)
pregnancy Abortion 19 (20.2) Complications of chemotherapy occurred mainly after
Live birth 47 (61.8) taking more than two courses

Complete hydatidiform mole was common among Key: * Refers to rate of any complications per group
women at extremes of reproductive age and parity.

Table 2. Clinical features of patients with CHM

Table 4: Complications associated with Prophylac-
tic chemotherapy
Characteristic (n=94) No (%)
Amenorrhea 76 (80.9) * Complication ( n=178) No.(%)
Vaginal bleeding 21 (22.3) Mucositis
Hyperemesis gravidarum 14 (14.9) Conjuctivitis 30 (16.0)
Haemoptysis 3 (3.2) Stomatitis 50 (26.7)
Theca lutein cysts 31 (32.7) Pharyngitis 56 (29.9)
Pulmonary metastasis 14 (14.9) Esophagitis 52 (27.8)
Cerebral metastases 3 (3.2) Colitis 8 (4.3)
Cystitis 6 (3.2)
Hepatic metastases 3 (3.2)
Pruritus vulvae/ani 24 (12.8)
Genital extra-uterine metastases 5 (5.3) µ Haematological toxicity
Uterus greater than dates 62 (66.3) Anaemia 40 (22.5)
Hypertension 38 (40.4) Thrombocytopenia 16 (8.8)
Hyperthyroidism 4 (4.3) Leucopenia 22 (11.8)
Spontaneous remission 5 (5.3) # Hepatic toxicity 6 (3.20
Required other chemotherapy 4 (4.3)
Deaths due GTD 8 (8.6) Others : Nausea and vomiting 16 (8.6)
Alopecia 4 (2.1)
Dermatitis/ Photosensitivity 6 (3.2)
The majority of patients presented with vaginal bleeding.
Presence of extra-uterine metastases indicates a diagnosis
The commonest complications were mucositis, hepato-
of Malignant metastatic GTD or Choriocarcinoma. None
toxicity and haematological (Mainly anaemia and
of the patients continued to have vaginal bleeding after
leucopenia)
evacuation.

African Health Sciences Vol 2 No 2 August 2002 49

Key or torsion. In prior studies, Montz et al 7 found them in
a) * Total number of courses given is the denominator 50% of 273 cases, 3 of which needed emergency surgery,
used, while the numerator is the number of courses while Mungan et al 8 found them in 50% of 310 cases,
during or after which a complication was noted. 1.8% of which required emergency surgery. Rare features
b) µ Anaemia was defined as Haemoglobin less than include acute pulmonary complications due to massive
10.0g/dl trophoblast embolisation, pre-eclampsia and hyperthyroid-
Thrombocytopenia was defined as less than 150,000 cells ism.
/ mm 3 Features associated with high risk molar pregnancy
Leucopenia was defined as less than 3000 cell/mm3 are very high hCG levels, uterus larger than dates of ges-
c) # Hepatotoxicity was defined as elevation of tran- tation or exceeding 20 weeks, large theca lutein cysts, ma-
saminase or alkalinephosphatase levels by ternal age greater than 40 years, prior molar pregnancy,
twice the upper limit. hyperthyroidism and pre-eclampsia 9. Histological find-
ings include marked trophoblast hyperplasia, presence of
extensive necrosis and haemorrhage, marked trophoblast
DISCUSSION immaturity and cytotrophoblastic proliferation 9. These
Gestational trophoblastic diseases (GTD) are a continuum are related to high hCG levels. Histology alone is not reli-
from the benign CHM, through to the malignant non- able in diagnosis of malignant GTD 9.
The diagnosis of persistent gestational tropho-
metastatic GTD (previously called invasive mole,
NMGTD) to the metastatic (commonly called choriocar- blastic disease was based on persistently elevated serum
hCG levels, with or without chemotherapy, at least 2
cinoma, MGTD). Both malignant types can arise as a
months after abortion or uterine evacuation. In the study,
CHM, a normal pregnancy or a non-molar abortion 1. All
24(25.5%) had persistent GTD. This is consistent with the
have similar epidemiological risk factors, which include
literature that about 20% of CHM result in persistent GTD
conception at extremes of reproductive age, high parity 10
. Other evidence of persistent GTD include persistent
and prior molar pregnancy 1,2.
amenorrhoea, persistent vaginal bleeding or theca lutein
The variation in the incidence of CHM in the lit-
cysts, development of massive trophoblast immobiliza-
erature arises from 3 factors: the definition used for CHM, tion several weeks after abortion or evacuation, and evi-
the denominator used (total pregnancies, total deliveries,
dence of extra-pelvic disease, at least 2 months after abor-
total live birth or total population of women of child- tion or uterine evacuation
1,2,8,9,10
. Therefore, patients on
bearing age), and the population at risk 2. The incidence follow-up should have serial chest X-rays, clinical evalua-
obtained in the study, though higher than in 1969-1973, tion and assay of serum hCG levels
1,2,10
. Progressively-
may be an under-estimate, as not all women with sponta- rising levels at any time post-evacuation, levels exceeding
neous abortion had specimens taken for histology. Like- 20,000I.U/ml four weeks or more post-evacuation, per-
wise, there is now improved diagnosis of CHM by use sistent levels even in absence of extra-uterine disease, are
of ultrasound and serum hCG assay. indication for chemotherapy 1.
On epidemiological risk factors, prior molar preg- In this study, oral methotrexate was used for pro-
nancy, ages < 20 years and > 35 years, and high parity phylactic chemotherapy in 78 women (187 courses, range
were associated factors. These are recognised risk factors 1-8 courses, without folinic acid or citrovorum rescue fac-
1,2,3,4
. Prior history of CHM is a recognised risk factor 6, tor). Overall, the drug was well tolerated for the first 2
the risk being higher for those who had more molar preg- courses, and toxicity occurred mainly after completion of
nancies. In this study, 10(10.7%) women had the anteced- the second course. Mucositis (mainly stomatitis, pharyngi-
ent pregnancy as molar, while 16(17.0%) and 4(4.3%) had tis, esophagitis, colitis, conjunctivitis and diarrhea) were
at least 1 and 2 molar pregnancies respectively, and 19 the commonest complications. Others were hematologi-
(20.2%) had unspecified spontaneous abortions prior, cal and hepatic toxicity. The rate of complications of
some of which could have been molar. 33.7% , was comparable to that of Kashimura et al 12,
On clinical features, 80.9% presented in the second tri- who used parenteral methotrexate. Therefore, care must
mester with threatened abortion. This is the commonest be taken if 3 or more courses are given for chemopro-
presentation 1,2. Theca lutein cysts, which are due to stimu- phylaxis, with close monitoring for mucositis, hepatic and
lation of the theca by high hCG were present in hematological toxicity. The primary remission that occurred
31(32.7%) of cases, of which 25 were bilateral, 20 were in 54 of 78 patients treated (69.2%) is acceptable and com-
palpable per abdomen and 3 required emergency surgery parable to that of oral methotrexate. This is one of the

African Health Sciences Vol 2 No 2 August 2002 50

few studies where oral methotrexate was evaluated as a several courses of the drug are used for prophylactic che-
single agent for chemotherapy for CHM and persistent motherapy, mainly mucositis and hematological toxicity.
GTD. It is also the first study to evaluate the Mulago policy Methotrexate does not eliminate the occurrence of, or
of prophylactic chemotherapy with oral Methotrexate for prevent death from, metastatic gestational trophoblastic
CHM. disease.
In a prospective study to evaluate the role of
single-agent prophylactic chemotherapy with Parenteral Recommendations
methotrexate for CHM, Kashimura et al 12 followed 420 Patients with CHM should be monitored closely for the
patients for 5-15 years. Twenty-two (7.5%) of 293 pa- development of post-evacuation gestational trophoblas-
tients on chemotherapy developed secondary trophoblas- tic disease. Those on prophylactic chemotherapy require
tic disease compared to 23(18.1%) of 127 who did not close monitoring for the toxic effects of the drugs, mainly
receive it. In the above study, 30.7% of the patients on for mucositis, haematological and hepatic complications.
chemotherapy developed toxic symptoms, mainly mu-
cositis, thrombocytopenia and leucopenia, with no deaths REFERENCES
or discontinuation of therapy due to the drug, and all 1. World Health Organisation Scientific Group. Gestational
received folinic acid. Three times as many patients who Trophoblastic Disease. Technical Report Series 692. World Health
got chemotherapy developed secondary trophoblastic Organisation. Geneva. 1983.
disease, but were cured by combined treatment with a 2. Bracken MB. Hydatidiform mole: an epidemiological review.
Br J Obstet Gynecol 1987; 94: 1123-1135
single agent. They concluded that chemotherapy of CHM
3. Smith J. Choriocarcinoma in Uganda: a case for active
did not entirely eliminate occurrence of metastatic disease chemoprophylaxis.Aust. N. Z. J. Obst Gynecol 1970; 10: 51-54
as a single agent for treatment13
of non -metastatic GTD, 4. Kaye D. Risk factors for molar pregnancy in Mulago hospital.
Lurain & Elfstrand found parenteral Methotr- In : Proceedings of the first meeting of the East, Central and Southern
exate (0.4mg/kg/day intravenously for five days per treat- African Association Obstetrical and Gynaecological Societies Conference
ment course, repeated every 14 days) extremely well tol- (ECSAAOGS), 5TH -9TH April 1997, 289-296
erated and effective in 89.3% of cases, with significant 5. Leighton PC. Trophoblastic disease in Uganda. Am J Obstet
toxicity in 12 (4,7%). These studies show that Methotrex- Gynecol 1973; 117:341-34
ate is effective and safe for chemotherapy. The variability 6. Sand PK, Lurain JR, Brewer JI. Repeat gestational tropho-
in the rate of complications and remission may be due to blastic disease. Obstet Gynecol 1984; 63: 140-143
7. Montz FJ, Schlaerth JB, Morrow CP. Natural history of
individual variability in the rate of drug absorption. Low
theca lutein cysts. Obstet Gynecol 1988, 2 : 247-251
doses are well absorbed, but high doses are incompletely 8. Mungan T, Kuscu E, Dabakoglu s et al. Hydatidiform mole:
absorbed especially when used over a long time, and meals clinical analysis of 310 patients. Int J Gynecol Obstet 1996; 52: 233-
may affect absorption rate 14. Barter et al, who used a 236
single oral dose (given 1 hour before or 2 hours after 9. Murad TM, Longley JV, Lurain JR, Brewer JI. Hydatidi-
meals to enhance absorption), found oral methotrexate form mole: clinico-pathological associations with the develop-
to be equally safe, tolerable and cost-effective. Oral meth- ment of post-evacuation trophoblastic disease. Int J Gynecol Obstet
otrexate is cheaper, more convenient in dosage and more 1990; 32 : 359-367
acceptable to patients than parenteral regimens, as it may 10. Morrow CP. Post-molar trophoblastic disease: diagnosis,
be given on out-patient basis. Mucositis, however, may management and prognosis. Clin Obstet Gynecol 1984: 27: 211-
interfere with drug absorption (increasing or increasing 220
rates), thus affecting drug bioavailability and toxicity, and 11. Kohorn IE. Criteria toward the definition of non-metastatic
meals may interfere with absorption. Sulphonamide and gestational trophoblastic disease after hydatidiform mole. Am J
non-steroidal anti-inflammatory agents, by displacing Obstet Gynaecol 1982; 142: 416-419
methotrexate from binding sites on albumin, may increase 12. Kashimura Y, Kashimura M, Sugimori H et al.
free serum drug levels, thereby increasing toxicity. Prophylactic chemotherapy for hydatidiform mole: five to 15
years’ follow-up. Cancer 1986 ; 58: 6 24-62
CONCLUSION 13. Lurain JR, Elfstrand EP. Single-agent methotrexate
Complete hydatidiform mole occurs in Ugandan women chemotherapy for the treatment of non-metastatic gestational
at prevalence 3.4 per 1,000 deliveries. Many women present trtrophoblastic disease
14. Balis FM, Saviteh JL, Bleyer WA. Pharmacokinetics of
with high-risk disease, associated with high parity and con-
oral methotrexate in children. Cancer Res 1983; 43: 2342-2345
ception at extremes of ages. Although oral methotrexate 15. Barter JF, Soong JS, Hatch KD et al. Treatment of non-
is cheap and tolerable, complications are common when metastatic gestational trophoblastic disease with oral methotr-
exate. Am J Obstet Gynecol 1987; 157: 1166-1168
African Health Sciences Vol 2 No 2 August 2002 51