Immune dysregulation is implicated in the pathophysiology of both bipolar and major depressive disorder, while immune cell ratios (IRCs) have recently been proposed as clinically applicable immune biomarkers. We investigated IRCs differences in affective disorders and their association with current mood episodes and clinical features. This retrospective cohort study analyzed neutrophil–lymphocyte (NLR), monocyte–lymphocyte (MLR), and platelet–lymphocyte (PLR) ratios upon admission in 135 affective disorder in-patients with mania (MA, n = 36), bipolar depression (BiD, n = 38), and unipolar depression (MDD, n = 61). Demographic, clinical, and immune data were extracted from medical records. Monocyte count was significantly higher in BiD compared to MDD (p < 0.001). Multivariable regression models suggested higher NLR in MA compared to MDD (p = 0.039), higher MLR in both MA and BiD compared to MDD (p < 0.001 and p = 0.004 respectively), while we found neither group differences in PLR nor an effect of type and duration of hospitalization, current psychotic, or suicidal features and psychiatric history on IRCs. Here, we show that IRCs are elevated in bipolar disorder versus MDD and affected by mood episode, while MLR could be especially valuable in the differential diagnosis between bipolar and unipolar depression. IRCs represent inexpensive, routinely accessible and clinically applicable biomarkers with diagnostic validity in affective disorders that could be easily implemented as illness activity indicators, to better follow the course of illness and eventually predict relapse or treatment response and, thus, guide therapeutic targeting.