LAUSR

Background Previous studies have shown that serum albumin is associated with prostate cancer (PCa), but not with prostate-specific antigen (PSA) levels in populations without PCa history. Therefore, we analyzed secondary data provided by the National Health and Nutrition Examination Survey (NHANES) (2003–2010). Methods In total, 5,469 participants were selected from the NHANES database (2003–2010). Serum albumin and PSA levels were serially considered independent and dependent variables, serially. A number of covariates were included in this study, including demographic, dietary, physical examination, and comorbidity data. Using weighted linear regression model and smooth curve fitting, the linear and non-linear relationship between serum albumin and PSA was investigated. Results After modulating underlying interference factors, the weighted multivariate linear regression analysis revealed that serum albumin did not independently predict PSA levels (β = −0.009 95%CI: −0.020, 0.002). Nevertheless, a non-linear relationship was found between serum albumin and PSA, with a point of 41 g/L. Left of the inflection point, the effect size, 95%CI, and P-value were 0.019 (log2 transformation) (−0.006, 0.043) and 0.1335, respectively. We found a negative association between serum albumin and PSA on the right side of the inflection point, with effect size, 95%CI, and a P-value of −0.022 (log2 transformation) (−0.037, −0.007), 0.0036. Conclusion In summary, serum albumin and PSA levels are not linearly related. When serum albumin levels exceed 41 g, serum albumin levels are negatively associated with PSA levels.