OBJECTIVES:  To assess the tolerability and effectiveness of exenatide once-weekly (EQW) versus basal insulin (BI), among elderly T2D patients. METHODS:  From a large US EHR database, we selected injectable-naïve T2D patients initiating EQW or BI between 2012 and 2015, into propensity score matched cohort study. We analyzed a subgroup of 1,447 T2D patients aged 65+ years. Weight and HbA1c were identified at baseline and quarterly in the first year, and were compared using interval means, absolute change from baseline and 95% confidence intervals (CI). We identified nausea and vomiting using diagnostic codes, and hypoglycemia using diagnostic codes and clinical notes. Incidence per 1,000 person-years, relative risk (RR) and 95% CI of each were used to compare risk in EQW relative to BI. RESULTS:  Baseline characteristics were balanced in across comparators, [e.g., weight (kg), 65-74 years: 103.0 (EQW) and 101.6 (BI); 75+ years: 95.7 (EQW) and 92.8 (BI); HbA1c%: 65-74 years: 7.9 (EQW) and 8.0 (BI), 75+ years: 7.9 (EQW) and 7.6 (BI)]. In patients aged 65-74 years, EQW initiators lost 1.3 kg by Q2 increasing to 1.6 kg by year’s end, and weight did not change among BI initiators. Among patients 75+ years, EQW initiators lost more weight early, and by years end both groups lost weight (-1.98 kg (EQW), -1.45 kg (BI)). HbA1c% declined more among EQW initiators regardless of age (65-75 years: -0.5 (EQW), -0.34 (BI); 75+ years: - 0.5% (EQW), -0.14 (BI)). The hypoglycemia incidence was similar across comparators (65-74 years: RR=0.91, 75+ years: RR=0.97). Nausea and vomiting was more common among EQW versus BI in both age groups [65-74 years: RR= 1.77 (95% CI: 1.16, 2.69); 75+ years: RR=1.32 (95% CI: 0.57, 3.04)]. CONCLUSIONS: