OBJECTIVE Hepatic-directed vesicle insulin (HDV) uses a hepatocyte-targeting moiety passively attaching free insulin, improving subcutaneous insulin’s hepatic biodistribution. We assessed HDV-insulin lispro (HDV-L) versus insulin lispro (LIS) in type 1 diabetes (T1D). RESEARCH DESIGN AND METHODS Insulin Liver Effect (ISLE-1) was a 26-week, phase 2b, multicenter, randomized, double-blind, noninferiority trial. RESULTS Among 176 randomized participants (HDV-L n = 118, LIS n = 58), the difference in change from baseline A1C was 0.09% (95% CI −0.18% to 0.35%), confirming noninferiority (prespecified margin ≤0.4%). Overall, there were no statistically significant differences between treatments for hypoglycemia or insulin dosing. However, baseline A1C modified the treatment group effect (interaction P < 0.001) on clinically apparent hypoglycemia designated by treatment-blinded investigators as severe. Thus, at higher baseline A1C, there was less hypoglycemia and lower insulin dosing with similar A1C outcomes during HDV-L versus LIS, whereas greater risk of hypoglycemia despite similar A1C outcomes and insulin doses was observed with lower baseline A1C. Among poorly controlled participants (A1C ≥8.5%), incidence rates of severe hypoglycemia in the HDV-L and LIS arms were 69 and 97 events/100 person-years, respectively (P = 0.03), whereas with A1C <8.5%, respective rates were 191 and 21 events/100 person-years (P = 0.001). Similar A1C-dependent trends in hypoglycemia were seen with continuous glucose monitoring. Among poorly controlled participants, bolus insulin doses at end point were ∼25% lower with HDV-L (P = 0.02), despite similar A1C outcomes; in better-controlled participants, insulin doses and A1Cs were stable over time in both subgroups. No safety signals were identified. CONCLUSIONS Hepatic biodistribution of HDV-L appears to potentiate insulin effect in T1D, with divergent clinical outcomes in hypoglycemia dependent on baseline A1C.