Rate of change of serial beta-human chorionic gonadotropin values as a predictor of ectopic pregnancy in patients with indeterminate transvaginal ultrasound findings
Rg. Dart et al., Rate of change of serial beta-human chorionic gonadotropin values as a predictor of ectopic pregnancy in patients with indeterminate transvaginal ultrasound findings, ANN EMERG M, 34(6), 1999, pp. 703-710
Study objective: To determine the predictive value of the rate of change of
serial beta-human chorionic gonadotropin (hCG) values in patients with sym
ptoms suggestive of ectopic pregnancy but who have indeterminate transvagin
al ultrasound findings, and to determine whether the predictive value was e
nhanced depending on whether the endometrial cavity was empty at ultrasound
examination.
Methods: A retrospective study was performed on consecutive emergency depar
tment patients from August 1, 1991, through August 1, 1998, presenting with
abdominal pain or vaginal bleeding, a positive beta-hCG test result, and i
ndeterminate transvaginal ultrasound findings. Patients were eligible for t
he study if they had a second beta-hCG assay performed within 7 days of the
initial visit and before either a diagnostic dilation and evacuation or la
paroscopy. Patients were excluded if they were lost to follow-up. Patients
were divided into 4 groups based on the rate of change of beta-hCG values o
ver a 48-hour interval (increase by >66%, increase by <66%, decrease by <50
%, decrease by >50%). In addition, the 4 main groups were further subdivide
d depending on whether the endometrial cavity was empty at ultrasound exami
nation. Intergroup differences in the frequency of ectopic pregnancy based
on the rate of change of the beta-hCG value were compared using logistic re
gression. Logistic regression also was used to determine whether addition o
f the ultrasound result improved predicative accuracy. A P value of less th
an .05 was considered significant. Odds ratios (ORs) were determined for ea
ch subgroup.
Results: Three hundred thirty-one eligible patients were identified; of the
se, 24 were excluded. Of the 307 enrolled patients, 33 (10.7%) had a final
diagnosis of ectopic pregnancy. Intergroup differences in the frequency of
ectopic pregnancy based on the beta-hCG rate of change were significant (P<
.0001). Addition of the ultrasound result to this model further improved pr
edicative accuracy (P<.0001). Overall, patients with increasing beta-hCG Va
lues were at increased risk compared with those with decreasing beta-hCG va
lues, and patients with empty uteri at ultrasound were at increased risk co
mpared with those with uteri that were not empty. Combining the beta-hCG ra
te of change with the ultrasound result identified 3 high-risk groups: pati
ents with beta-hCG Values that increased by less than 66% and an empty uter
us at ultrasound (OR 24.8); patients with beta-hCG values that decreased by
less than 50% and an empty uterus at ultrasound (OR 3.7); and patients wit
h beta-hCG values that increased by more than 66% and an empty uterus at ul
trasound (OR 2.6). Patients with beta-hCG Values that decreased by more tha
n 50% were found to be at low risk for ectopic pregnancy irrespective of th
e specific endometrial findings at ultrasound.
Conclusion: The rate of change of serial beta-hCG values, in patients with
an indeterminate pelvic ultrasound examination, is predictive of ectopic pr
egnancy. Addition of whether the endometrial cavity is empty at ultrasound
leads to a further improvement in predictive accuracy.