Much evidence indicates that bacterial LPS (endotoxin) is removed from the bloodstream mainly by the liver, yet the hepatic uptake mechanisms remain uncertain and controversial. In plasma, LPS can be either ‘free’ (as aggregates, bacterial membrane fragments or loosely bound to albumin, CD14, or other proteins) or ‘bound’ (complexed with lipoproteins). Whereas most free LPS is taken up by Kupffer cells (KCs), lipoprotein-bound LPS has seemed to be cleared principally by hepatocytes. Here, we compared the liver’s ability to take up and deacylate free LPS aggregates and the LPS in preformed LPS-high density lipoprotein (HDL) complexes. In mice examined from 1 h to 7 d after a small amount of fluorescent (FITC-)LPS was injected into a lateral tail vein, we found FITC-LPS almost entirely within, or adjacent to, KCs. As expected, FITC-LPS complexed with HDL (FITC-LPS-HDL) disappeared more slowly from the circulation and a smaller fraction of the injected dose of FITC-LPS was found in the liver. Unexpectedly, the FITC-LPS injected as FITC-LPS-HDL complexes was also found within sinusoids, adjacent to, or within, KCs. In other experiments, we found that both free and HDL-bound radiolabeled LPS underwent enzymatic deacylation by acyloxyacyl hydrolase (AOAH), the LPS-inactivating enzyme that is principally produced within the liver by KCs. Our observations suggest that KCs and AOAH play important roles in clearing and catabolizing both free LPS and the LPS in circulating LPS-HDL complexes.