To determine whether psyllium must be mixed with food to lower serum cholesterol, 18 modestly hypercholesterolemic subjects were studied for three 2-wk periods, in random order, separated by a 2-wk return to a National Cholesterol Education Program Step 2 diet. Compared with values for subjects consuming control wheat-bran cereal (63 g/d), after 2 wk of 54 g psyllium-enriched cereal/d containing 7.3 g psyllium, serum total, LDL, and HDL cholesterol, respectively, were reduced by 8% (6.15 +/- 0.15 vs 6.71 +/- 0.19 mmol/L, P < 0.01), 11% (4.24 +/- 0.15 vs 4.78 +/- 0.19 mmol/L, P < 0.02), and 7% (0.99 +/- 0.05 vs 1.07 +/- 0.05 mmol/L, P < 0.01). When 7.6 g of the same type of psyllium as in the test cereal was taken between meals, serum total (6.50 +/- 0.19 mmol/L), LDL (4.50 +/- 0.21 mmol/L), and HDL (1.06 +/- 0.06 mmol/L) cholesterol were no different from control values, and total cholesterol was greater than after psyllium cereal (P < 0.05). We conclude that psyllium must be mixed with foods to have the maximum effect on serum cholesterol.