Background: Clinical trials leading to the approval of immune checkpoint inhibition (ICI) have almost exclusively been performed in patients with good performance status (ECOG PS of 0-1). However, ICI remains an attractive option for patients with advanced tumors and poor performance status, considering their overall tolerability. While use of ICI in patients with poor PS (ECOG PS of 2 or greater) has been rapidly adopted, whether these patients derive the same benefits as expected in the studied populations is largely unknown. We therefore performed an institutional retrospective analysis of all patients treated with palliative single agent anti-PD1 or anti-PDL1 to determine response and survival for those with poor performance status. Methods: We retrospectively identified patients with advanced solid tumor malignancies who were treated with ICI monotherapy with palliative intent at our institution between 2015-2019. The primary objective was to compare overall survival (OS) for patients with good PS (ECOG PS 0-1) with those with poor PS (ECOG PS 2 or 3-4). The log-rank test compared the survival among patients with different ECOG PS. In addition, we used a proportional hazards model to assess association between ECOG PS and the OS with adjustment for age, gender, and smoking status. A secondary objective was to compare overall response rates (ORR) of the three ECOG PS groups which were evaluated with a binary rate model. Results: We identified 266 patients treated with ICI, 87 with NSCLC, 34 with melanoma, 33 with RCC, 24 with bladder cancer, 22 with head/neck cancer, and the rest with other histologies. 187 (70%) were ECOG PS 0-1, 62 (23%) were ECOG PS 2, and 17 (7%) were ECOG PS 3-4. 89 of these patients (33%) were still alive at time of last follow-up. Across all tumor types, patients with ECOG PS 0-1 had superior survival compared to ECOG PS 2 (median survival 12.4 months vs 4.6 months, HR 0.41, p < 0.001). Median survival for ECOG PS 3-4 was lower at 2.3 months. The ORR for ECOG PS 0-1 (23%) was significantly higher to that of ECOG PS 2 (6%, p = 0.02). ORR for ECOG PS 3-4 was 12%. Conclusions: Despite the appeal of ICI for patients with advanced malignancy and poor performance status, outcomes were poor. Survival and objective response rates for patients with ECOG PS 2 and higher were significantly worse than those with ECOG PS 0-1. ICI treatment comes with cost, including potentially forgoing early hospice referral or optimal support at the end of life. Prospective trials defining the activity and role of ICI in poor PS are urgently needed.