P-065 A systematic review of the accuracy of laboratory semen analysis as a test of fertility

Abstract

Abstract Study question Which semen parameters are predictive of treatment-dependent and treatment-independent reproductive outcomes? Summary answer Progressive motility predicted treatment-independent pregnancy, total motile sperm count (TMSC) predicted failed fertilisation with IVF and pregnancy following IUI, and morphology predicted pregnancy after IUI. What is known already Semen analysis, an essential component of any fertility work-up, is useful in identifying azoospermia and oligozoospermia, but its discriminatory capacity to identify fertile and infertile men has been questioned, including by the World Health Organisation (WHO). It has also been suggested that a combination of semen parameters has better predictive value than a single parameter. We conducted a systematic review of the literature to determine the association between different semen parameters and spontaneous (treatment-independent) pregnancy outcomes, as well as pregnancy outcomes following various fertility treatments (treatment-dependent outcomes). Study design, size, duration A protocol was registered prospectively with PROSPERO (CRD42021251988). A systematic search of MEDLINE, EMBASE, Web of Science, and the Cochrane Central Register of Controlled Trials was undertaken to include papers published from 1st January 2000 to 25th May 2021. Studies were not limited by language. Two reviewers independently screened the search results and undertook data extraction. Risk of bias assessment was undertaken with the QUADAS-2 or QUIPS tool, depending on study design. Participants/materials, setting, methods Studies were eligible if they included men undergoing semen analysis for any indication and reported the accuracy of one or more semen analysis parameters in predicting treatment-dependent or treatment-independent reproductive outcomes – these included fertilisation rate, clinical pregnancy, live birth, and time to pregnancy. Studies were excluded if semen analysis was not the primary test being assessed or no reproductive outcomes were reported. Main results and the role of chance Of 5236 publications and 139 full-text articles, 62 studies were included. 52 studies reported treatment-dependent outcomes and 10 reported treatment-independent outcomes. Most treatment-dependent studies lacked key information and were rated as unclear risk of bias in at least one domain (79.3% overall). Half of the treatment-independent studies were high risk of bias in at least one domain. Of all semen parameters, progressive motility and sperm concentration predicted treatment-independent pregnancy outcomes. Three studies concluded progressive motility predicted pregnancy rates at a range of thresholds (threshold 24%: area under the curve (AUC) 0.909; <32% vs ≥ 32%: adjusted odds ratio (aOR) 4.2 (95% CI 1.1-15) and <50% vs ≥ 50%: aOR 2.8 (95% CI 1.3-6.1); per 10% reduction in progressive motility, pregnancy chances reduced by 11%: hazard ratio (HR) 0.89 (95% CI 0.85-0.93)). Sperm concentration predicted pregnancy within 6 months (≥20 vs < 20 million/ml, adjusted rate ratio 0.68 (95% CI 0.52-0.91)) and 12 months (adjusted HR 1.35 (95% CI 1.09-1.66)). In those undergoing IUI, pre-/post-wash TMSC and morphology predicted pregnancy, but with no clear threshold values. TMSC predicted “total fertilisation failure” in an IVF population (AUC 0.631-0.75). In an ICSI population, no semen parameter predicted fertilisation rate, pregnancy, or live birth rate. Limitations, reasons for caution Meta-analysis was not possible due to a wide range of cut-off values used for each semen parameter. Due to the low quality of the evidence included, poorly detailed female populations, and WHO manual methods often not being used for reporting, the findings of this review should be viewed with caution. Wider implications of the findings Whilst this review has identified certain semen parameters in isolation are predictive of pregnancy in treatment-dependent and treatment-independent scenarios, clear threshold values have not been identified. This questions the reporting of morphology as it is labour intensive to assess and is only predictive of pregnancy in a population having IUI. Trial registration number Not applicable