Abstract Study question Do semen parameters predict long-term health outcomes in men? Summary answer Semen parameters do not have a good predictive value for mortality, diabetes, cardiovascular disease (CVD) or cancer. What is known already One in 3 men attending infertility clinics have suboptimal semen analysis. Assisted conception treatments are offered to them as treatment for infertility and their care usually ends there. There is now a growing body of evidence male infertility predicts future development of medical comorbidities. Multiple studies have looked into the association between clinical male infertility, involuntary childlessness or suboptimal sperm analysis and the risk of cancer, cardiovascular risks, diabetes, autoimmune diseases and mortality. However, there is still conflicting evidence regarding long-term health risk in men with suboptimal semen quality. Study design, size, duration A systematic review methodology was used using PRISMA guideline. The papers were selected from PubMed/MEDLINE, EMBASE and EBM from databases’ inception to May 2023. Reviews and case series were excluded. The selected studies were assessed for quality and risk of bias using QUADAS-2. Meta-analysis was done where appropriate. The scope of the analysis was to identify whether semen parameters could predict long-term health outcomes in men. Participants/materials, setting, methods Only papers reporting on health outcomes in men having at least one semen analysis result before diagnosis were included. A 2x2 contingency table was done for each study where data was available and true positives and negatives, false positives and negatives were calculated. The analysis was performed using MetaDTA, which generated estimates for sensitivity (Sn) and specificity (Sp), hierarchical summary receiver operating characteristic (HSROC) spaces and diagnostic odds ratio (DOR) with 95% confidence intervals (CI). Main results and the role of chance Twenty studies from three continents (USA, Europe and Asia) were included reporting on mortality (5), lifespan (1), cardiovascular disease (CVD) (2), diabetes (3), cancer (5), testicular carcinoma in situ (4), prostate-specific antigen (2), hospitalizations (2) and Charlson Comorbidity Index (1). Based on extracted data we could only analyse the predictive value of azoospermia and oligospermia for all-cause mortality, diabetes, CVD and cancer as there was insufficient data for other parameters. The results showed high specificity and low sensitivity of azoospermia and oligospermia for all outcomes (all-cause mortality: Sp 0.968, Sn 0.055 and Sp 0.839, Sn 0.162; diabetes: Sp 0.967, Sn 0.037 and Sp 0.902, Sn 0.046; cancer: Sp 0.964, Sn 0.135 and Sp 0.949, Sn 0.105; CVD: Sp 0.904, Sn 0.101 and Sp 0.932, Sn 0.074). HSROC model with 95% predictive region showed limited accuracy of sperm concentration as a test for any of the outcomes. DOR (95%CI) for azoospermia and oligospermia also showed that sperm concentration cannot be used as predictive test for any of the outcomes: all-cause mortality (1.72, 0.75-3.94 and 1.005, 0.59-1.693), diabetes (1.10, 0.77-1.56 and 0.44, 0.29-0.68), cancer (4.16, 1.45-11.97 and 2.18, 0.38-12.38) and CVD (1.05, 0.75-1.46 and 1.09, 0.83-1.44) when compared to normospermia. Limitations, reasons for caution Risk of Bias assessment showed high or unclear risk of bias in 14 studies mainly due to population selection. Not all laboratories adhered to WHO methodology, increasing the heterogeneity in reporting semen results and cut-offs. Concentration was the predominant parameter reported. Wider implications of the findings To answer the question whether poor semen quality is a predictor for adverse long-term health outcomes in men requires larger and well-designed studies. Until then, we should not be advising men that there are implications on long term health based on semen analysis results, with any confidence. Trial registration number not applicable