Defining the Decline in Human Chorionic Gonadotropin (hCG) for Complete Resolution of Chemical Pregnancies

Abstract

To provide a detailed description of the expected rate of hCG decline in patients presenting with very low initial hCG levels and ultimately confirmed to have spontaneous abortions (sAB). Retrospective cohort. Using a computerized database of all women presenting to the emergency room in the first trimester of pregnancy with symptoms of pain and/or vaginal bleeding, a cohort of women who presented between January 1, 1990 and July 31, 1999 was evaluated. Inclusion required that patients had a non-diagnostic transvaginal ultrasound and follow-up with serial hCG measurements to determine the definitive diagnosis. We extracted those patients with decreasing levels and the ultimate diagnosis of sAB, defined as a spontaneous decline of hCG levels to less than 5 mIU/mL in the absence of surgical or medical intervention, or as the presence of products of conception on dilation and curettage. Using a model created by hCG values for the entire cohort of sAB, we performed a sub-analysis of those patients with initial hCG values < 500 mIU/mL. The model derived from the entire cohort of sAB patients was used rather than re-fitting the model with the sub-group of interest in order to take advantage of the precision achieved with the original model. The total cohort of 710 patients contributing 2150 hCG values was used to define the decline of hCG. The best model to explain these data was a parametric random effects model for the natural log of hCG, including terms for a fixed quadratic effect and random linear and intercept effects for number of days since presentation. An interaction term between initial hCG and the slope (quadratic and linear fixed effects) was needed to describe that faster rates of fall were associated with higher starting concentrations. For this analysis, there were 447 patients who presented with initial hCG values < 500 mIU/mL, contributing 1619 hCG values. The slowest rates of decline were associated with the lowest initial hCG values. Patients with an initial hCG of 500 mIU/mL experienced a mean reduction of 71% in two days, while those with an initial hCG of 50 mIU/mL experienced a smaller mean reduction of 68% in two days. When initial levels were between 50 and 500 mIU/mL, the slowest expected percent decline (defined from the 95% CI) ranged from 12% to 24% in two days and 34% to 68% in seven days. The hCG values predicted by the model will be presented as expected percentage decline over the first seven days. Because the expected decline of hCG can not be described adequately by a single curve but rather, differs depending on the initial hCG, a simple clinical rule is not applicable. Using a model fit from a large cohort of patients, we report precise estimates for a sub-group of patients presenting with very low initial hCG levels. These data can be used by clinicians for comparison when managing patients with chemical pregnancies. The information presented is likely to be particularly useful in the setting of early surveillance of serial hCG levels, a common practice when caring for infertility patients.