Background: Active surveillance is a common management strategy for men with low-risk prostate cancer. However, there is some concern that delaying definitive therapy could lead to progressive disease and a missed opportunity to cure the cancer. We sought to study the impact of Gleason grade progression on the risk of metastases and prostate cancer-specific mortality (PCSM). We then sought to compare the risks of PCSM with the risk of competing mortality in this real-world cohort of men on active surveillance. Methods: This was a retrospective population-based cohort study using patient records from the Veterans Heath Administration Informatics and Computing Infrastructure of patients with a diagnosis of low-risk prostate cancer from January 1, 2001, to December 31, 2015, who were managed with active surveillance. Follow-up extended through March 31, 2020. Low-risk prostate cancer was defined as International Society of Urologic Pathology grade group (GG) 1, clinical tumor stage 2A or lower, PSA level of 10 ng/dL or lower. Active surveillance was defined as no definitive treatment within the first year after diagnosis with at least 1 additional re-staging biopsy after diagnostic biopsy. We identified men with pathologic upgrading to ≥GG2 and ≥GG3. We calculated the cumulative incidence of metastases, PCSM, and non-cancer mortality for men with and without pathologic upgrading. Results: The cohort included 8562 men with low-risk PC. Pathologic upgrading to ≥GG2 and ≥GG3 occurred in 3227 and 1200 men, respectively. For all patients, the 8-year cumulative incidence of metastases and PCSM was 0.8% (95% CI: 0.6-1.0%) and 0.4% (CI:0.3-0.6%), respectively. Pathologic upgrading to ≥GG2 was associated with an increased risk of metastases at 8 years (1.2% vs 0.5%, p < 0.001). Pathologic upgrading to ≥GG3 was also associated with an increased risk of metastases at 8 years (2.7% vs 0.5%, p < 0.001). In contrast, pathologic upgrading to ≥GG2 was not associated with an increased risk of PCSM at 8 years (0.5% vs 0.4%, p = 0.793). Similarly, pathologic upgrading to ≥GG3 was not associated with an increased risk of PCSM at 8 years (0.7% vs 0.4%, p = 0.82). The 8-year cumulative incidence of non-cancer mortality for all patients was 6.5% (95% CI:5.9-7.1%). Overall, the risk of non-cancer mortality was 16.25 (6.5/0.4) times higher than the risk of PCSM, regardless of upgrading status. Conclusions: For men with low-risk PC on active surveillance, pathologic upgrading was associated with a small increased risk of metastases but no increased risk of death from prostate cancer. Regardless of upgrading status, the risk of non-cancer mortality was substantially higher than the risk of PCSM. Future work is required to better tailor the intensity of follow up for surveillance protocols for men at risk of competing mortality.