Rationale: The impact of immunocompromising conditions on outcomes in critically ill patients with COVID-19 remains poorly characterized. We hypothesized that disparities in COVID-19 outcomes in immunocompromised patients may differ by viral strain and time course of the pandemic. Methods: We retrospectively reviewed SARS-CoV-2 RT-PCR positive patients admitted to intensive care (ICU) at Stanford from March 2020 to July 2022 using electronic medical records. Immunocompromised status included solid organ transplant; hematologic malignancy/transplant; solid cancer with metastases and recent chemotherapy; HIV/AIDS; and connective tissue disease on immunosuppression. SARS-CoV-2 strain periods were classified as pre-Delta (3/2020–8/2021), Delta (9/2021-12/2021), and Omicron (1/2022–7/2022). We assessed 90-day survival differences between immunocompromised and non-immunocompromised patients by strain using Cox-proportional hazards, adjusted for age, sex, and APACHE II score. Results: A total of 791 patients were included, 450 from the pre-Delta period, 114 from the Delta period, and 227 from the Omicron period. We compared 633 non-immunocompromised patients with 158 immunocompromised patients (77 during the pre-Delta, 22 Delta, and 59 Omicron periods). Patients admitted during the Omicron period were more likely to be immunocompromised (26%) compared to pre-Delta (17%) or Delta periods (19%; p = 0.008). Immunocompromise was associated with worse outcomes in the entire cohort after adjustment (HR 1.68, 95% CI 1.2–2.3, p < 0.001). However, when stratified by strain period, immunocompromised status was only significantly associated with higher 90-day mortality in patients from the Omicron period (42% vs 17%, HR 2.9, CI% 1.7–5.0, p < 0.001). Differences in mortality between immunocompromised and immunocompetent patients were not statistically different in either the pre-Delta (34% vs. 23%, HR 1.4, 95% CI 0.9–2.1, p = 0.15) or Delta periods (32% vs. 28% HR 1.0, 95% CI 0.4–2.5, p = 0.92). Notably, survival was significantly improved during the Omicron period compared to earlier strain periods in immunocompetent patients (HR 0.6, 95% CI 0.4–0.9, p = 0.02) but not for immunocompromised patients (HR 1.4, 95% CI 0.80–2.44, p = 0.24). Conclusion: In this retrospective cohort analysis of patients admitted to the ICU due to SARS-CoV-2 infection, we found that lower survival among immunocompromised patients was due to higher mortality during the Omicron period, with similar mortality between immunocompetent and immunocompromised patients during other periods.