Endocrinology and Metabolism |
1 Department of Pathology and Laboratory Medicine,
University of North Carolina at Chapel Hill, School of Medicine, Chapel Hill,
NC.
2 Department of Pathology and Immunology, Washington University
School of Medicine, Saint Louis, MO.
aAddress correspondence to this author at: Department of Pathology and Laboratory Medicine, CB #7525, University of North Carolina at Chapel Hill, School of Medicine, Chapel Hill, NC, 27599. Fax 919-966-9490; e-mail dgrenach@unch.unc.edu .
|
|
Abstract |
|---|
 Top Abstract IntroductionMaterials
and MethodsResultsDiscussionReferences |
|---|
Background: Human chorionic gondadotropin (hCG) screening tests
are performed on nearly all female patients of childbearing age
before any medical intervention. Although older women usually have
negative hCG test results, positive results do occur and may cause
clinical confusion. We examined changes with age in serum hCG
concentrations in nonpregnant women and investigated the use of serum
follicle-stimulating hormone (FSH) measurements as an aid to
interpreting higher than expected ("positive") hCG results.
Methods: We used 240 serum specimens for each of 4 female cohorts:
pregnant, 
18 years; nonpregnant,
18–40 years (premenopausal); nonpregnant, 41–55 years
(perimenopausal); and nonpregnant, >55 years (postmenopausal).
Patients were excluded if they had an ectopic pregnancy, a history of
trophoblastic disease or a germ-cell tumor, or if no chart was
available for review. Quantitative hCG and FSH tests were performed
on each specimen.
Results: Serum hCG concentrations in nonpregnant women increased
with the age of the women. hCG results were higher and significantly
different (P <0.0001) for nonpregnant women >55 years
(<2.0 to 13.1 IU/L) compared with nonpregnant women 18–40 years (<2.0
to 4.6 IU/L) and 41–55 years (<2.0 to 7.7 IU/L). Nineteen nonpregnant
women >40 years of age had hCG concentrations
5.0 IU/L, all with an FSH concentration
>32.4 IU/L. The highest FSH concentration in pregnancy was 7.3 IU/L.
Conclusions: Serum hCG increases with age in nonpregnant women. A cutoff of 14.0 IU/L should be used when interpreting hCG results in women >55 years of age. Pregnancy is unlikely in perimenopausal women 41–55 years of age with an hCG between 5.0 and 14.0 IU/L if serum FSH is >20.0 IU/L.
|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and
MethodsResultsDiscussionReferences |
|---|
Qualitative urine human chorionic gondadotropin (hCG) screening tests
are frequently performed on female patients of childbearing age
before any medical intervention that could potentially harm a fetus.
Positive screening tests are typically followed by a quantitative hCG
test performed in serum. hCG screening tests are being used
increasingly for female patients, including older women who are
likely to be postmenopausal. Although most of the hCG tests in these
older women are negative, positive results do occur, and
postmenopausal women have been reported to have increased
concentrations of hCG (above a nonpregnant cutoff) in the absence of
pregnancy or malignancy (1)(2)(3)(4)(5).
Although these "positive" results likely reflect a physiologic
increase in hCG in postmenopausal women unrelated to pregnancy (6),
limited knowledge of hCG concentrations in this population of women
often creates clinical confusion.
Even more difficult to interpret are increased hCG concentrations in perimenopausal women who may or may not be pregnant, because the menopausal status of these women is often uncertain for both the clinician and the patient. These situations make the interpretation of a positive hCG result in women >40 years of age difficult.
hCG is a glycoprotein hormone composed of 2 dissimilar subunits
designated 
and ß. Normally produced
by the syncytiotrophoblastic cells of the placenta, hCG is commonly
used as a marker for pregnancy detection and monitoring. Serum
concentrations of hCG also have clinical use as tumor markers for
trophoblastic diseases, germ cell tumors, and other malignancies (7)(8).
Because hCG measurements have two clinical applications,
false-positive hCG results not only cause clinical confusion, but in
several instances have led to harmful therapeutic interventions (6).
In addition to its synthesis during normal pregnancy, trophoblastic disease, or cancer, a small amount of hCG is normally produced by the pituitary gland in conjunction with the structurally similar glycoprotein hormones luteinizing hormone, follicle-stimulating hormone (FSH), and thyroid-stimulating hormone (3)(4).
An hCG cutoff concentration of <5.0 IU/L is generally used for the diagnosis of pregnancy; however, the upper limit of serum hCG in nonpregnant peri- and postmenopausal women has been reported in only a few studies (1)(2)(3)(4)(5). The primary objectives of this study were to determine this upper limit of serum hCG in these two populations of women and to evaluate the clinical use of a quantitative serum FSH measurement to assist in the interpretation of a positive hCG result.
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials
and MethodsResultsDiscussionReferences |
|---|
sample collection and exclusion of patients
This collaborative retrospective cohort study used leftover serum
samples sent to the laboratory between September 2003 and June 2004
for physician-ordered hCG, FSH, or thyroid-stimulating hormone tests.
Samples sent to the laboratory for hCG tests were used first. To
obtain sufficient numbers of specimens from female patients who were
41–55 and >55 years of age, samples were also randomly chosen from
those sent for thyroid-stimulating hormone or FSH testing. Specimens
were selected for the study if they met one of the following
inclusion criteria: pregnant, 18
years; nonpregnant, 18–40 years (premenopausal); nonpregnant, 41–55
years (perimenopausal); or >55 years (postmenopausal). Patients were
excluded if they had a history of trophoblastic tumor or ectopic
pregnancy or if no chart was available for review. Menopausal status
was defined by age alone, and no attempt was made to confirm
menopausal status on the basis of additional clinical information.
This study received approval from the Institutional Review Board of
each institution.
analysis of collected specimens
Quantitative analysis of hCG and FSH tests was performed on each
specimen. At the University of North Carolina, quantitative hCG was
measured with the Roche Elecsys® (Roche Diagnostics), and
quantitative FSH was measured with the Vitros® ECi (Ortho-Clinical
Diagnostics), both according to the manufacturer’s instructions. At
Washington University, quantitative hCG and FSH were measured with
the Advia Centaur® (Bayer Diagnostics) in accordance with the
manufacturer’s instructions. The functional sensitivities (defined as
the lowest concentration with a CV 
10%) for
hCG and FSH with either assay system were 2.0 and 1.0 IU/L,
respectively.
statistical analyses
Statistical analysis of the data was performed with GraphPad Prism,
Ver. 4.00 (GraphPad Software) and STATA release 8.2 (Stata
Corporation). For data analyses, results that were below the assay
functional sensitivity were used as reported by the instrument. A
Bland–Altman analysis was performed to compare assay methods (9).
Correlation was calculated with Spearman rank correlation, and
differences between the groups were assessed by the Mann–Whitney U-test.
Locally weighted regression (lowess smoothing) was used to fit smooth
curves to hCG and FSH as a function of age (10).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials
and MethodsResultsDiscussionReferences |
|---|
Data collected at the University of North Carolina and Washington
University were combined to give a total of 240 values per study
cohort (120 specimens from each institution).
Because distinct isoforms of hCG and FSH can be detected preferentially
by different immunoassays (11), we sought to determine
concordance between the two methods used in this study. Patient
specimens for hCG (n = 23) and FSH (n = 25) were analyzed by both
methods. Comparisons were accomplished by calculating the ratio of
the Elecsys to the Centaur method for hCG and the ratio of the ECi
to the Centaur method for FSH with a Bland–Altman plot to chart
the ratio vs the mean hCG or FSH concentration (Fig. 1
). Although a decreasing trend in the ratio with increasing
concentrations of hCG was observed, concordance between the two
methods for both hormones was clinically acceptable, and the data
from both sites were combined.
|
We found a significant correlation between hCG concentrations and
age for all 720 nonpregnant patients (r = 0.333; P <0.0001);
however, when examined per cohort, correlation was significant
only for hCG vs age in the nonpregnant perimenopausal (41–55 years)
cohort (r = 0.156; P = 0.02). The correlation between
hCG and age for the nonpregnant pre- and postmenopausal cohorts was
r = 0.034 (P = 0.60) and r = 0.039 (P = 0.55),
respectively (Fig. 2A
). The correlation between hCG and FSH was examined in each
cohort and was significant only in the nonpregnant perimenopausal (r
= 0.286) and postmenopausal (r = 0.385) cohorts (P <0.0001).
|
Three (1.3%) women in the perimenopausal cohort and 16 (6.7%)
women in the postmenopausal cohort had a serum hCG concentration that
was greater than the conventional cutoff of 5.0 IU/L (12).
None of these women had evidence of pregnancy or cancer. Table
1 shows the hCG
concentration results and the 97.5 percentile concentration for each
of the 3 nonpregnant cohorts. The postmenopausal (>55 years) cohort
had an hCG concentration greater than and significantly different
from the peri- and premenopausal cohorts (P <0.0001). The
highest hCG concentrations seen in pre-, peri-, and postmenopausal
women were 4.6, 7.7, and 13.1 IU/L, respectively (Table 1
). These findings are similar to previous reports that cite
increased hCG concentrations in postmenopausal women (Table 2
).
|
|
Because FSH concentrations are increased in postmenopausal women
and suppressed during pregnancy, it was speculated that this hormone
may help to evaluate perimenopausal patients with increased hCG
concentrations. As expected, our data illustrate that FSH
concentrations increase with age (Fig. 2B
). When broken into cohorts, the postmenopausal cohort had the
highest median concentration of FSH (46.9 IU/L), followed by the
nonpregnant perimenopausal cohort (median, 23.5 IU/L), the
nonpregnant premenopausal cohort (median, 4.7 IU/L), and finally the
pregnant cohort (median, 0.6 IU/L). We divided the peri- and
postmenopausal cohorts by hCG status, using the conventional 5.0 IU/L
cutoff for a positive test (Fig. 3
). The minimum FSH concentrations in peri- and postmenopausal
patients with positive hCG tests were 51.5 and 32.4 IU/L,
respectively. The highest FSH value determined in the pregnant cohort
was 7.3 IU/L; therefore, an FSH cutoff between 7.3 and 32.4 IU/L
could be selected that would adequately distinguish between positive
hCG caused by pregnancy and positive hCG resulting from advancing
age. A value of 20.0 IU/L, midway between these two points, was
chosen as a conservative cutoff.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials
and MethodsResultsDiscussionReferences |
|---|
Women are frequently given an hCG test to rule out pregnancy before
any procedure that could harm a fetus, but positive pregnancy tests
can lead to delays in the procedure. The results presented here,
along with reports published previously, indicate that increased hCG
concentrations >5.0 IU/L are not uncommon in peri- and postmenopausal
women and should not be considered abnormal (1)(2)(3)(4)(5).
Our data clearly indicate that hCG and FSH concentrations increase with age, although both hCG and FSH are fairly constant during the stages of premenopause and postmenopause. Perimenopause is the start of the endocrine, biological, and clinical features of menopause. During this period of transition between the two stages, a significant correlation between hCG and age is observed. Herein lies the difficulty in interpreting the hCG test results of a woman 41–55 years of age who, hormonally, could be pre- or postmenopausal, or somewhere in between.
The highest hCG concentration in the nonpregnant premenopausal cohort was 4.6 IU/L, consistent with the conventional hCG cutoff of 5.0 IU/L. The highest hCG concentration in the nonpregnant peri- and postmenopausal cohorts was 7.7 and 13.1 IU/L, respectively; therefore, by rounding these values to the nearest whole number that would include all patients in the cohort, we recommend that the upper limits of normal serum hCG for nonpregnant peri- and postmenopausal women be 8.0 and 14.0 IU/L, respectively. In women >55 years of age, in whom pregnancy is unlikely, hCG concentrations <14.0 IU/L should be considered normal.
The use of FSH quantification to help identify menopause-associated
low-positive hCG concentrations has been suggested previously (7).
Similarly, we investigated the use of serum FSH to help interpret
low-positive hCG concentrations in women 41–55 years of age in whom
pregnancy is possible. On the basis of these data, we designed an
algorithm for the interpretation of hCG results (Fig. 4
). Pregnancy should be considered unlikely in all women with hCG
concentrations <5.0 IU/L. hCG concentrations between 5.0 and 14.0 IU/L
should be considered an indicator of possible pregnancy in
premenopausal women. Pregnancy is unlikely in women >55 years of age
with hCG concentrations between 5.0 and 14.0 IU/L, although
additional testing is recommended if the concentration is >14.0 IU/L.
FSH testing should be performed in women between 41 and 55 years of
age with hCG concentrations between 5.0 and 14.0 IU/L. When FSH
concentrations are >20 IU/L, pregnancy is unlikely. Note that this
algorithm is intended exclusively for hCG tests that have been
ordered to rule out pregnancy, and test results need to be
interpreted together with clinical history.
|
Between August 1, 2004, and February 2, 2005, a total of 3751 physician-ordered hCG tests were performed at the University of North Carolina and Washington University hospitals. Of these, 547 (15%) were ordered for women 41–55 years, and 138 (4%) were ordered for women >55 years. In the 41–55 years group, 7 (0.2% of total) of the results were between 5.0 and 15.0 IU/L. Therefore, if reflex testing were established, it would not create a burden for the laboratory.
Although the exact source of hCG was not known for the peri- and postmenopausal women with increased serum hCG who were in our study, the evidence suggests a pituitary origin. Concentrations of pituitary FSH have been shown to increase during perimenopause (13). As expected, we noted a significant correlation between hCG and age and FSH and age beginning in the perimenopausal cohort. We also found a correlation between hCG and FSH beginning in the perimenopausal stage and persisting through menopause. This finding is indirect evidence to support a pituitary origin of hCG. More direct evidence was reported by Suginami and Kawaoi (14), who found enhanced immunohistochemical staining for hCG in the pituitary glands of postmenopausal women. Likewise, Hoermann et al. (15) found hCG immunoreactivity in a fractionated pituitary extract from a postmenopausal woman. Additionally, hCG of pituitary origin is typically found at low-positive concentrations, similar to those of the menopausal patients in our study (3)(4). To definitively test whether increased hCG originates from the pituitary, patients could be placed on estrogen–progesterone therapy, which should suppress pituitary hCG (4). Although differentiating between the exact isoforms of hCG may aid in determining the source of increased hCG, insufficient evidence is available to support the use of hCG isoforms for this purpose.
Some limitations of our study need to be discussed. First, the
molecular heterogeneity of hCG as well as the various antibodies used
in clinical hCG assays are causes of interassay discordance (16).
As such, the upper limits of hCG reported here may not be
transferable to hCG assays different from the two that we used.
Second, it is possible that the increased hCG in some of our
postmenopausal women was caused by the interference of human
anti-animal antibodies; however, we believe that this scenario is
unlikely because interference by human anti-animal antibodies
generally produces hCG concentrations higher than those observed in
our cohorts (17). The third limitation is that in
determining the upper hCG concentration in the different cohorts, we
used age as the sole criterion for defining menopausal status. Other
previous studies of hCG in postmenopausal women used age, FSH
concentrations, and clinical features to define menopause (Table 2
) (1)(2)(3)(4)(5).
It is possible that some of our perimenopausal women were actually
postmenopausal and that some postmenopausal women were actually
perimenopausal; however, despite differences in the definition of
menopause, the mean hCG concentrations for postmenopausal women (>55
years) in our study are similar to those of other published
studies (Table 2 ) (1)(2)(3)(4)(5).
In addition, age is often the only criterion available when hCG
results are interpreted. Fourth, many nonpregnant peri- and
postmenopausal patients in our study had FSH concentrations below the
highest FSH concentration in pregnant patients. Some of the very low
FSH concentrations can likely be explained by the use of
estrogen–progesterone replacement therapies that inhibit FSH (18).
We did not exclude patients who were taking these medications from
our study. Because estrogen–progesterone medications also suppress
hCG of pituitary origin (4), it is possible that
the inclusion of patients on hormone therapy affected our final
results. This may explain the absence of a significant correlation
between hCG and age in postmenopausal women (>55 years); however,
because patients on hormone replacement therapy may be a large subset
of the postmenopausal population, we believe that inclusion of these
patients has allowed us to define a more accurate hCG threshold
for the postmenopausal population. Finally, our algorithm is intended
for use in women 41–55 years of age; however, the FSH cutoff in the
pregnant cohort contained only 5 pregnant perimenopausal women, and
it is possible that the distribution of FSH was different in these
women. We believe that this is unlikely because the state of
pregnancy should inhibit FSH regardless of patient age. Nonetheless,
additional studies are currently in progress to confirm an FSH
reference interval in pregnant perimenopausal women.
In summary, concentrations of hCG in nonpregnant women increase
with age and may produce false-positive pregnancy tests in nonpregnant
peri- or postmenopausal women. In the future, widespread knowledge
of this phenomenon coupled with expanded hCG thresholds may
eliminate confusion regarding positive hCG tests. Finally, our
demonstration that FSH concentrations may help determine pregnancy
status in perimenopausal women with an hCG concentration
5.0 IU/L might reduce the
clinical confusion caused by low-positive hCG test results.
|
|
Acknowledgments |
|---|
We thank Roche Diagnostics for supplying hCG reagent and Bayer
Diagnostics for supplying hCG and FSH reagents used in this study.
|
|
Footnotes |
|---|
1 These authors contributed equally to this work.
|
|
References |
|---|
|
TopAbstractIntroductionMaterials
and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article: (Search Google Scholar for Other Citing Articles)
 |
|
 |
|
  L. A. Cole "Background" Human Chorionic Gonadotropin in Healthy, Nonpregnant Women Clin. Chem., October 1, 2005; 51(10): 1765 - 1766. [Full Text] [PDF] |
|
|
|
|||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | SEARCH RESULT |
