Management of pregnancy of unknown location. Diagnostic protocols, clinical outcomes and psychosocial aspects

Abstract

Background: Ectopic pregnancy (EP) is the leading cause of pregnancy-related first trimester mortality. Many EP remain undiagnosed beyond an initial transvaginal ultrasound assessment. Noninvasive diagnosis of EP in women with a pregnancy of unknown location (PUL) involves serial serum human chorionic gonadotropin (hCG) levels. Published hCG-based protocols, including risk-prediction models, can be of great interest to Swedish patients and physicians as they may improve PUL management. The performance of protocols is commonly assessed by their diagnostic and discriminatory abilities, but comparisons of clinical outcomes are limited. Aim: To determine a well-functioning strategy for PUL management in Swedish gynecological emergency units and to investigate the psychological outcome of having a PUL. Methods: A comparative study of four hCG-based protocols and a validation study of risk-prediction models (M4 and M6NP) were conducted at the gynecological emergency unit at the Sahlgrenska University Hospital. M4 and the NICE (National Institute for Health and Care Excellence, in the UK) algorithm were selected for a multicenter randomized controlled trial (RCT). Psychological distress was measured with the Hospital Anxiety and Depression Scale (HADS) in women with a PUL and in control participants in antenatal care. Health-related quality of life was assessed with the 36-item short form (SF-36). Results: In the first study M4 and the NICE algorithm had higher sensitivity for EP but lower specificity for non-EP than the two other protocols. In the validation study M6NP made more accurate predictions of EP, had higher discriminatory ability for EP vs non-EP, and higher sensitivity for EP, but lower specificity for non-EP than M4. In the RCT the NICE algorithm had higher discriminatory ability for EP vs non-EP and specificity for non-EP than M4 when tested on cross-sectional data. There was no significant difference in diagnostic performance for M4 vs NICE when analyzed according to randomized groups and no significant between-group differences in number of hCG or ultrasounds undertaken, time to diagnosis of EP, length of follow-up, first-line treatment success of EP, or adverse events. Women with a PUL had more frequently an anxiety score ≥8 (58.6% vs 29.2%, p=.0002) than women in antenatal care. No significant differences in anxiety or depression were observed for women with a viable vs a nonviable pregnancy within the PUL group after 1 or 4 weeks. For SF-36, the mental component summary was below the norm (50) for women with a viable (38.5) or a nonviable pregnancy (33.5), but did not differ between groups, p=.20. The physical component summary was lower in women with a viable pregnancy (47.7 vs 52.2, p=.020) and below the norm (50). Conclusion: The diagnostic and discriminatory abilities of the NICE algorithm were mostly superior to M4, but no significant differences of clinical outcomes were observed. The RCT was not powered to detect differences in these secondary outcomes, which should be considered when interpreting the results. The M6NP seems to be preferable for managing PUL, given its high sensitivity for EP and discriminatory ability, although it was not tested in clinical practice. Symptoms of anxiety and depression were substantial among women with a PUL, and health-related mental quality of life was low. This aspect of PUL is novel and requires more research.