Although fish oil supplementation may prevent the onset of diet-induced insulin resistance in rats, it appears to worsen glycemic control in humans with existing insulin resistance. In the present study, the euglycemic, hyperinsulinemic (4x basal) clamp technique with [3-3H]glucose and 2-deoxy-[1-14C]glucose was used to directly compare the ability of fish oil to prevent and reverse sucrose-induced insulin resistance. In study 1 (prevention study), male Wistar rats were fed a purified high-starch diet (68% of total energy), high-sucrose diet (68% of total energy), or high-sucrose diet in which 6% of the fat content was replaced by menhaden oil for 5 wk. In study 2 (reversal study), animals were fed the high-starch or high-sucrose diets for 5 wk and then the sucrose animals were assigned to one of the following groups for an additional 5 wk: high starch, high sucrose, or high sucrose with 6% menhaden oil. Rats fed the high-starch diet for 10 wk served as controls. In study 3 (2nd reversal study), animals followed a similar diet protocol as in study 2; however, the reversal period was extended to 15 wk. In study 1, the presence of the fish oil in the high-sucrose diet prevented the development of insulin resistance. Glucose infusion rates (GIR, mg.kg-1.min-1) were 17.0 +/- 0.9 in starch, 10.6 +/- 1.7 in sucrose, and 15.1 +/- 1.5 in sucrose with fish oil animals. However, in study 2, this same diet was unable to reverse sucrose-induced insulin resistance (GIR, 16.7 +/- 1.4 in starch, 7.1 +/- 1.5 in sucrose, and 4.8 +/- 0.9 in sucrose with fish oil animals). Sucrose-induced insulin resistance was reversed in rats that were switched back to the starch diet (GIR, 18.6 +/- 3.0). Results from study 3 were similar to those observed in study 2. In summary, fish oil was effective in preventing diet-induced insulin resistance but not able to reverse it. A preexisting insulin-resistant environment interferes with the positive effects of menhaden oil on insulin action.