Infection of non-human primates (NHPs) such as rhesus and cynomolgus macaques with monkeypox virus (MPXV) or cowpox virus (CPXV) serve as models to study poxvirus pathogenesis and to evaluate vaccines and anti-orthopox therapeutics. Intravenous inoculation of macaques with high dose of MPXV (>1-2×10(7) PFU) or CPXV (>10(2) PFU) results in 80% to 100% mortality and 66 to 100% mortality respectively. Here we report that NHPs with positive detection of poxvirus antigens in immune cells by flow cytometric staining, especially in monocytes and granulocytes succumbed to virus infection and that early positive pox staining is a strong predictor for lethality. Samples from four independent studies were analyzed. Eighteen NHPs from three different experiments were inoculated with two different MPXV strains at lethal doses. Ten NHPs displayed positive pox-staining and all 10 NHPs reached moribund endpoint. In contrast, none of the three NHPs that survived anticipated lethal virus dose showed apparent virus staining in the monocytes and granulocytes. In addition, three NHPs that were challenged with a lethal dose of MPXV and received cidofovir treatment were pox-antigen negative and all three NHPs survived. Furthermore, data from a CPXV study also demonstrated that 6/9 NHPs were pox-antigen staining positive and all 6 NHPs reached euthanasia endpoint, while the three survivors were pox-antigen staining negative. Thus, we conclude that monitoring pox-antigen staining in immune cells can be used as a biomarker to predict the prognosis of virus infection. Future studies should focus on the mechanisms and implications of the pox-infection of immune cells and the correlation between pox-antigen detection in immune cells and disease progression in human poxviral infection.