Is normal b-hCG regression curve helpful in the

diagnosis of persistent trophoblastic disease?

N. BEHTASH, F. GHAEMMAGHAMI, H. HONAR, K. RIAZI, A. NORI, M. MODARES &
A. MOUSAVI
Department of Gynecologic Oncology, Vali-e-Asr Hospital, Tehran University of Medical Sciences, Tehran, Iran
Abstract. Behtash N, Ghaemmaghami F, Honar H, Riazi K, Nori A,
Modares M, Mousavi A. Is normal b-hCG regression curve helpful in the
diagnosis of persistent trophoblastic disease? Int J Gynecol Cancer
2004;14:980983.
Objective: The aim of this study was to evaluate the probable useful-
ness of normal b-human chorionic gonadotropin (b-hCG) regression
curve in the diagnosis of persistent trophoblastic disease (PTD).
Methods: A log-value regression curve was developed from the means
and 95% confidence limits of serial weekly serum b-hCG titers of 43
patients with uneventful complete hydatidiform moles and 14 patients,
who were previously confirmed as PTD.
Results: All 14 PTD patients (100%) had abnormal values, beyond normal
range, within 4 weeks. b-hCG was in its upper values, compared to normal
regression curve at 2.290.19 weeks. This was earlier than plateau or rise
detection at 4.210.33 weeks (P<0.001). Within 3 weeks of evacuation, 13
of 14 (92.86%) PTD patients b-hCG values exceeded the normal range,
whereas only six of 14 (42%) showed a rise or plateau.
Conclusion: Our finding indicates that the normal b-hCG regression curve
may be useful for quicker detection of PTD than the plateau or rise of level.
KEYWORDS: human chorionic gonadotropin, persistent trophoblastic
disease, regression curve.
The spectrum of trophoblastic diseases extends from
benign molar pregnancy to a life-threatening meta-
static choriocarcinoma. Approximately, 92% of hydati-
diform moles resolve spontaneously after evacuation
and the remainder can be successfully treated with a
minimum of toxicity
(1)
. Virtually, almost all patients
are now potentially curable, if they are correctly diag-
nosed and the appropriate therapy is administered
early enough in the course of the disease.
After diagnosis of a molar pregnancy, follow-up is
essential to detect those who require chemotherapy.
Follow-up mainly relies on serial measurements of
b-human chorionic gonadotropin (b-hCG) serum
levels for at least 1 year after evacuation
(2)
. Serum
b-hCG level reflects the clinical course of the disease
and its serial measurement allows for the identifica-
tion of a rise or persistent plateau in the serum level of
b-hCG, which is usually indicative of the need for
treatment
(2)
.
This study was performed to establish a normal
b-hCG regression curve in a population of Iranian
patients diagnosed as having uneventful moles and
to evaluate the probable usefulness of this regression
curve in identifying patients with persistent tropho-
blastic disease (PTD).
Patients and methods
During the period from April 1999 through September
2001, a total of 72 patients with hydatidiform moles
were diagnosed and treated at Vali-e-Asr Hospital,
Address correspondence and reprint requests to: N. Behtash, MD,
Department of Gynecologic Oncology, Vali-e-Asr Hospital,
Keshavarz Blvd, Tehran (14194), Iran. Email: nadbehtash2@yahoo.com
Int J Gynecol Cancer 2004, 14, 980983
#2004 IGCS
Tehran, Iran. After evacuation, serum b-hCG was
measured weekly as a rule, until it reached a level
below 5 mIU/ml (normal). The serum b-hCG was
measured at the same laboratory with the same kit
(Radim, Rome, Italy).
Patients with the second curettage, prophylactic
chemotherapy, partial mole, and insufficient data
were excluded. In 43 patients, the b-hCG values
reached normal level within 13 weeks after evacua-
tion and remained at a normal level over a year. Dur-
ing follow-up period, all patients received low dose of
oral contraceptive pills.
Referring to the serial b-hCG in 43 molar patients,
we estimated the declining trend of the log value per
week with the help of stepwise linear regression ana-
lysis
(3)
. A log-value regression curve was developed
from the means and 95% confidence limits of serial
weekly serum b-hCG titers. During the construction
of the regression curve, patients with titers reaching
undetectable values were progressively excluded.
PTD after hydatidiform mole evacuation is diag-
nosed with the help of:
Four or more values of plateau of b-hCG over at
least 3 weeks on days 1, 7, 14, and 21.
A rise of b-hCG, 10% or greater, for three values or
longer over at least 2 weeks on days 1, 7, and 14
(4)
.
Few authors have reported the normal regression
curves of uneventful moles as a discriminator for the
early detection of PTD
(5,6)
.
Fourteen patients had a confirmed diagnosis of
PTD with the help of four values or more of plateau
of b-hCG titer. These changes happened over at least 3
weeks on days 1, 7, 14, and 21. Rising titer of 10% or
greater for three values was during at least 2 weeks on
days 1, 7, and 14.
In these patients, the time, when the b-hCG titer
exceeded the upper limit of the curve for the first
time, was compared to the time of diagnosis based
on a plateau or rise. We defined plateau or rise as a
less than 10% change or a more than 10% increase in
b-hCG level, respectively (based on three consecutive
measurements)
(6)
.
Using Students t-test, statistical analysis was per-
formed. P-values of less than 0.05 were considered
statistically significant.
Results
The mean age was 26.7 years for patients with
uneventful moles and 28.3 years for PTD patients
(not significant). In patients with uneventful mole,
40.9% were primigravida, 24.5% were gravida. Gesta-
tional age was 1115 weeks in 49.46%, <10 weeks in
19.35%, and 1620 weeks in 25.8%. The average pre-
evacuation serum b-hCG concentration in 43 patients
of uneventful patients was 23091 mIU/ml ranged
between 1600 and 178000 mIU/ml, whereas the aver-
age of 14 serum b-hCG values in the patients with
PTD, right before values of the evacuation, was
14257 mIU/ml and ranged between 545 and
95600 mIU/ml (not significant). The mean weekly
values of the serum b-hCG obtained from the 43
patients with uneventful moles have been shown in
Table 1. The time for undetectable value in these
patients has been shown in Table 2.
The regression curve of the mean weekly b-hCG
values and the upper 95% confidence limit has been
illustrated in Figure 1. The mean serum b-hCG value
of uneventful patients reached a level of less than
5 mIU/ml at 10 weeks after evacuation and the
upper 95% limit reached the same level in the 12th
week after evacuation. In this study, the weekly
values of the upper 95% confidence limit were
defined as the normal range and the regression
curve of the 95% confidence limit was defined as the
normal curve.
The b-hCG value beyond the normal range at a
given week after evacuation was defined as an abnor-
mal value. Based on this definition, the values of
b-hCG exceeding the upper limit of normal curve at
a given week after evacuation were considered abnor-
mal and thus all 14 PTD patients (100%) had abnormal
values beyond normal range within 4 weeks, which
never regressed to the normal range.
Table 3 shows the number of the weeks after eva-
cuation, when the serum b-hCG values exceeded the
upper limit (with a mean of 2.29 0.19), compared to
the weeks, when b-hCG values exhibited a plateau or
Table 1. Mean of b-human chorionic gonadotropin (b-hCG) in
weeks, standard deviation (SD), and confidence interval (CI) of
43 patients with complete mole
b-hCG Mean (mIU/ml) SD (mIU/ml) CI 95% (mIU/ml)
b1 4958.2 8580.8 2341.17575.3
b2 520.8 405.5 397.1644.5
b3 231.4 267.4 149.8312.9
b4 107.2 165.1 56.8157.6
b5 49.9 76.2 26.773.1
b6 20.8 24.4 13.428.2
b7 11.1 12.6 7.314.9
b8 6.7 5.8 4.98.5
b9 5.5 4 4.36.7
b10 4 1.7 3.54.5
b11 3.8 1.4 3.44.2
b12 3.6 1 3.33.9
b13 3.5 1.1 3.23.8
Role of b-hCG regression curve in the diagnosis of PTD 981
#2004 IGCS, International Journal of Gynecological Cancer 14, 980983
a rise (with a mean of 4.21 0.33) as defined earlier
(P<0.001).
b-hCG values exceeded the normal range in 13 of 14
(92.86%) PTD patients at 3 weeks of evacuation,
whereas only six of 14 (42.86%) PTD patients showed
a plateau or a rise in b-hCG levels.
Three weeks after evacuation, 13 of 14 PTD patients
had b-hCG values in upper region out of normal
regression curve. However, in only six of 14 PTD
patients, we could find a standard FIGO definition
of PTD at 3 weeks after evacuation.
Discussion
In the present study, we tried to construct a regression
curve of b-hCG titer in 43 uneventful molar pregnan-
cies. In the second part of this study, we depicted
b-hCG regression curve in 14 PTD patients and finally
we compared these two curves.
Serial measurement of serum b-hCG is the most
sensitive test in the assessment of clinical course of
trophoblastic disease. Titers are usually obtained on a
weekly basis, until these levels are normal for 3 con-
secutive weeks, followed by monthly determinations
up to 612 months
(7,8)
. In the cases of uneventful
moles, the average time to achieve the first normal
b-hCG level after evacuation is about 9 weeks
(9)
.
Diagnosis of PTD and indication of chemotherapy
are a rise or plateau of b-hCG values in at least 3
consecutive weekly measurements
(4,6)
.
Feltmate et al. have shown that with one undetect-
able b-hCG level after evacuation no cases of gesta-
tional trophoblastic tumor (GTT) will develop
(10)
.
A few studies have shown, by using the normal
b-hCG regression curve, that PTD patient can be iden-
tified at a given time after evacuation with the help of
a b-hCG titer
(5)
.
Yedema et al. reported 15% of uneventful moles to
express a temporary plateau or a rise in the b-hCG
changes
(11)
. Khanlian et al. define persistent low-level
b-hCG titers as a subset with preinvasive gestational
trophoblastic disease
(12)
.
Sometimes, it is not possible to check weekly serial
b-hCG titers regularly, because the patients are not
committed to follow-up. This problem arises mainly
from the costly follow-up sessions and not easily
available diagnostic facilities for many patients spe-
cially those from rural areas of the country. A normal
regression curve may help the physician to decide on
a single random value and to change the intervals of
follow-up visits in selected patients.
It may also facilitate the earlier recognition of PTD,
which, in turn, allows us to have prompt treatment.
In this study, we established a log-value normal
regression curve of postevacuation serum b-hCG in
our patients diagnosed as having uneventful moles.
Furthermore, by using the established normal curve,
we evaluated whether a PTD patient can be identified
precisely according to the level of b-hCG at a given
time and comparing it with the time plateau or rise
titer can be identified. We observed that in all (100%)
of our PTD patients, the b-hCG level had exceeded
the upper limit of the normal curve (the weekly
values of the upper 95% confidence limit) at a given
week within 4 weeks, before the diagnosis was made
Table 2. Frequency and rate of 43 patients with complete
mole, who got negative b-human chorionic gonadotropin
(b-hCG)
At the end
of the week
Number of patients,
who got negative b-hCG %
3 1 2.32
4 2 4.65
5 4 9.30
6 12 27.90
7 18 41.86
8 26 60.46
9 31 72.09
10 36 83.72
11 37 86.04
12 43 100
13 43 100
All 43 100
100,000
10,000
1000
100
10
5
1
0 2 4 6 8
Weeks after evacuation
S
e
r
u
m

-
h
C
G

(
m
I
U
/
m
l
)
10 12 14
Fig. 1. First values of b-human chorionic gonadotropin (b-hCG)
exceeding the normal curve in the 14 patients with persistent
trophoblastic disease (PTD) (circles). The open circles show the first
value to exceed the normal curve. The normal curve is shown as
mean (solid line) and upper 95% confidence limit (dotted line).
982 N. Behtash et al.
#2004 IGCS, International Journal of Gynecological Cancer 14, 980983
based on a rise or plateau levels of b-hCG. At 3 weeks
after evacuation, 93% of PTD patients had b-hCG
levels upper than the normal curve; at the same
time, only 43% showed plateau or rise level. This
finding may provide some evidence for the potential
ability of normal regression curve in the early diag-
nosis of PTD. It seems that it can help us to detect
PTD more quickly than a plateau or raise basis.
Shigematsu et al.
(4)
similarly reported an earlier
detection of PTD by using a normal regression curve
in comparison with identification based on a plateau
or a rise. They suggested that early diagnosis of PTD
by using a normal curve might reduce the need for
multi-agent chemotherapy due to the appearance of
resistance to the initial single-agent chemotherapy.
It has been reported that spontaneous regression in
serum b-hCG is more rapid in patients with partial
moles than those with complete and invasive
moles
(13)
. Thus, it seems reasonable to construct such
a curve for each type separately.
In conclusion, our observation supports the concept
that a PTD can be suspected more quickly by means
of a normal b-hCG regression curve than by means of
a plateau or rise method. This may encourage the
physicians to utilize similarly established curves as a
screening measure to distinguish high-risk patients
from low-risk patients.
More detailed information may be available from
larger prospective studies. Future randomized clinical
trials reveal whether using a single level of b-hCG titer
outside the normal regression curve may lead to a
higher proportion of patients receiving chemotherapy.
References
1 Smith DB, OReilly SM, Newlands ES. Current
approaches to diagnosis and treatment of gestational
trophoblastic disease. Curr Opin Obstet Gynecol 1993;5:
8491.
2 Cunningham FG, Gant NF, Leveno KJ, Gilstrap LC III,
Hauth JC, Wenstrom KD. Williams Obstetrics, 21st edn.
New York: McGraw-Hill, 2001, 842.
3 Akazawa K, Nakamura T, Palesch Y. Power of log-rank
test and regression model in clinical trial. Stat Med
1997;16:58397.
4 Kohorn EI. The New FIGO 2000 staging and risk factor
scoring system for GTD. Int J Gynecol Cancer
2001;11:737.
5 Shigematsu T, Kamura T, Saito T, Kaku T, Nakano H,
Kinogawa N. Identification of persistent trophoblastic
disease based on a human chorionic gonadotropin
curve by means of a stepwise piecewise linear regression
analysis after the evacuation of uneventful moles. Gyne-
col Oncol 1998;71:37680.
6 Kohorn EI. Hydatidiform mole and gestational tropho-
blastic disease in southern Connecticut. Obstet Gynecol
1982;59:7884.
7 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;128:42430.
8 Berek J. Novak Gynecology, 13th edn. Philadelphia, PA:
Lippincott, Williams & Wilkins, 2002, 136.
9 Genest DR, Borde O, Berkowitz RS et al. A clinicopatho-
logic study of 153 cases of complete H. mole. Obstet
Gynecol 1991;78:4029.
10 Feltmate CM, Batorfi J, Fulop V et al. Human chorionic
gonadotropin follows up in patients with molar preg-
nancy: a time for re-evaluation. Obstet Gynecol 2003;101:
7326.
11 Yedema KA, Verheijen RH, Kenemans P et al. Identifica-
tion of patients with persistent trophoblastic disease by
means of a normal hCG curve. Am J Obstet Gynecol
1993;168:78792.
12 Khanlian AS, Smith HO, Cole LA. Persistent low level of
hCG: a premalignant GTD. Am J Obstet Gynecol 2003;188:
12549.
13 Rob L, Robova H, Pluta M, Kulovany E et al. Regression
of b-hCG in various types of molar pregnancies: clinical
course and prognosis (article in Czech). Ceska Gynekol
2001;66:2305 (Abstract).
Accepted for publication March 26, 2004
Table 3. The number of weeks exceeding the normal range,
compared to the number of weeks exhibiting a plateau or rise in
the b-human chorionic gonadotropin (b-hCG) change in
patients with persistent trophoblastic disease (PTD)
Patient number
Weeks after
evacuation
Weeks showing a
plateau or rise
1 1 5
2 2 3
3 2 6
4 2 3
5 2 3
6 2 4
7 2 4
8 2 3
9 2 4
10 2 3
11 3 5
12 3 4
13 3 3
14 4 4
MeanSEM 2.29 0.19 4.21 0.33
Role of b-hCG regression curve in the diagnosis of PTD 983
#2004 IGCS, International Journal of Gynecological Cancer 14, 980983