Background: Illicit use of anabolic androgenic steroids (AAS) has emerged as a public health concern among men, but the long-term effect on gonadal function is still unresolved. Serum insulin-like factor 3 (INSL3) has emerged as a novel and potentially superior marker of Leydig cell function than serum testosterone per se. INSL3 synthesis and secretion exhibit far less daily variation than testosterone. Further, serum INSL3 levels are not related to body composition. The objective of this study was to investigate INSL3 as a marker of Leydig cell function in former AAS users. Methods: Community-based cross-sectional study including men aged 18 - 50 years, involved in recreational strength training and allocated to one of three groups: current (n = 46) or former AAS users (n = 42) or controls (n = 44). Mean age (SD) of all participants were 32 (7) years and the elapsed duration since AAS cessation, geometric mean (95% CI), was 32 (23; 45) months in former AAS users. All procedures were performed during one visit in the morning hours following overnight fasting. We drew blood through a cannula placed in an antecubital vein following 30 minutes of supine rest. Medical records, testicular size, questionaries, and detailed history of strength training and AAS use were obtained in a structured interview. Serum INSL3 and testosterone were measured using liquid chromatography mass spectrometry. Results: Serum INSL3 was markedly suppressed among current AAS users compared with former AAS users and controls, P < 0.001. Additionally, former AAS users also displayed lower serum INSL3 concentrations than controls, mean (SD), 0.43 (0.31) versus 0.60 (0.22) µg/L, P = 0.006 and the difference remained significant in a multivariate linear regression, (B) (95%CI), -0.17 (-0.28;-0.55) µg/L, P=0.004, adjusted for plasma LH, plasma sexual hormone-binding globulin, age, body fat %, smoking and use of other illicit drugs. Longer accumulated duration of AAS use (log2) was associated with reduced serum INSL3 levels in former AAS users, (B) (95%CI), -0.08 (-0.14;-0.01), P=0.022, suggesting a dose-response relation between AAS use and suppression of serum INSL3. We evaluated the association between INSL3 and total testosterone levels and they were not associated among former users and controls in a multivariate linear regression, P=0.821. We noted recovery of serum inhibin B levels among former AAS users reaching the mean plasma level of controls after elapsed duration since AAS cessation of ≈ 21 months; (B) (95%CI), 2.2 (0.7; 3.7) months, P = 0.006. In contrast, we did not note any recovery of serum INSL3, P = 0.541, or total testosterone, P = 0.861, among former AAS users. Conclusions: Serum INSL3 is decreased years following AAS cessation in former AAS users, independently of plasma testosterone, suggesting persistent impaired Leydig cell function, which should be investigated further.