Ectopic pregnancy is the most common cause of pregnancy-related death in the first trimester and a common diagnostic and management challenge. Because of its relatively high incidence, ectopic pregnancy represents a small but real risk of mortality from ectopic rupture. When diagnosed early, ectopic pregnancy may be treated medically rather than surgically. Therefore, ectopic pregnancy should be considered and quickly ruled out in all women of reproductive age who present with abdominal pain or vaginal bleeding. Ultrasonography remains the diagnostic test of choice, but it is less specific in patients with low beta-subunit human chorionic gonadotropin (β-hCG) levels and may not be available 24 hours a day in some hospitals and clinics. Buckley and associates reviewed and refined a previously derived clinical model for estimating the risk of ectopic pregnancy in symptomatic first-trimester patients.

All hemodynamically stable patients of reproductive age who presented to the emergency department with abdominal pain or vaginal bleeding during the first trimester of pregnancy were enrolled in the prospective data registry. Clinical data were collected and a standard set of laboratory tests was obtained, including serum β-hCG and progesterone measurements. All patients were monitored until they met the criteria for a diagnosis of intrauterine pregnancy (IUP), nonviable IUP or ectopic pregnancy. Patients were categorized as high, intermediate or low risk according to their clinical presentation. For a description of the characteristics of each risk category, see the accompanying table. The clinical prediction model has two stages: the first stage distinguishes the high-risk group from the intermediate- and low-risk groups; the second categorizes those not in the high-risk group as intermediate or low risk.

A total of 429 patients was available for analysis. Of this total, 70 had ectopic pregnancies at an average gestation period of eight weeks. The presence of fetal heart tones on a handheld Doppler image or the presence of tissue extruding from the cervical os were the only single clinical variables that excluded ectopic pregnancy in all instances. Four percent of the patients with an ectopic pregnancy reported no abdominal pain at presentation, and 20 percent reported no vaginal bleeding. Definite cervical motion tenderness or peritoneal signs were the most specific findings for ectopic pregnancy. The clinical prediction model demonstrated a first-stage specificity of 95 percent and a second-stage sensitivity between 96 and 100 percent. Patients who met the low-risk criteria had a risk of ectopic pregnancy of less than 1 percent.

Management strategies for these patients can be determined by their risk classification. Those defined as low risk can be managed with nonurgent outpatient diagnostic evaluation, including pelvic ultrasonography, serial β-hCG and serum progesterone testing, and instructions to return if the symptoms become worse. Those classified as intermediate risk require a more intermediate evaluation, while those with a high-risk profile require emergency gynecologic consultation even if the immediate work-up is nondiagnostic.

The authors conclude that a clinical prediction model can provide useful information in the management of symptomatic first-trimester patients, especially when ancillary testing is unavailable or nondiagnostic. However, they caution that this model should not be extrapolated to asymptomatic patients in whom risk factors play a much greater predictive role.