Following antiretroviral therapy (ART) initiation, HIV-associated immune reconstitution inflammatory syndrome (IRIS), indicated by an array of opportunistic infections, may occur, presenting as either paradoxical, a worsening of a previously treated infection, or unmasking, a flare-up of an underlying, previously undiagnosed infection. The impact of ART as the backbone of HIV treatment and prevention has prolonged the survival of people living with HIV. In pregnancy, benefits have been shown by slowing HIV progression and preventing perinatal transmission; however, there have been risks of adverse reactions with ART, including immune responses to both the fetus and mother. This study sought to estimate the incidence of HIV-IRIS cumulatively and by type either paradoxical or unmasking IRIS, determine the baseline maternal and HIV clinical markers as predictors of, and analyze the log-rank test for survival time to IRIS outcome assessed by relying on an increase in CD4 count and/or a rapid decrease in viral load. An active records study was conducted between June 2019 and March 2020 among ART-naïvepregnant women attending the antenatal care units (ANCu) at the Kenyatta National and MbagathiHospitals, Nairobi, Kenya. Participants were aged between 20 and 49 years and had a confirmed HIV-positive test. To ascertain a true case of IRIS, the diagnosis was adjudicated for accuracy and consistency by an independent review committee. Plasma HIV viral load, CD4 counts, and routine laboratory evaluations (hemoglobin, white blood cell count (WBC)) were performed by each hospital's clinical laboratory. The IRIS incidence was assessed using the International Network for Studies Against HIV-Associated IRIS (INSHI) during the first three months after ART initiation. Multivariate Cox regression with IRIS as the outcome, using the SPSSSurvival package, examined the relationship between baseline maternal characteristics and HIV clinical markers before ART initiation and IRIS, and finally, decision-tree analysis for predicting IRIS was performed. From a pool of 532 ART-naïvepregnant women, 133 (25%) developed IRIS, and 97 (72.9%) were in the unmasking category. The accumulated risk of experiencing IRIS symptoms increased from week two (hazard ratio (HR)=0.0287) to week 12 (HR=3.6158). Participants with a maternal BMI (MBMI) of 25-29.9 kg/m2 at baseline were at a higher risk of unmasking IRIS (HR=2.478; P=0.010). The WHO-HIV clinical stages 1 and 2 skewed towards paradoxical IRIS (regression coefficients =-0.111 and -0.276 (P<0.05)), while stage 4 skewed towards unmasking IRIS (regression coefficient=0.047, HR=1.048, P=0.941). A CD4 count > 500 cells/mm^3 skewed towards unmasking IRIS (regression coefficient=0.192, HR=1.211, P=0.416), while RNA-HIV viral loads >50 copies/ml towards paradoxical IRIS (regression coefficient=-0.199, HR=0.820, P=0.360. On decision tree analysis, 85% (P=0.001) of ART-naïve pregnant women with a baseline CD4 count below 500 copies/mm^3 presented with unmasking IRIS. For ART-naïvepregnant women, unmasking IRIS is the most common type, and an MBMI of 25-29.9 kg/m2, advanced HIV infection, a CD4 count <500 cells/mm^3, and a higher parity at baseline may be clinically useful predictors. The higher proportion of ART-naïvepregnant women experiencing unmasking as compared to paradoxical IRIS supports the need for earlier assessment based on potential predictors.