Your immediate logical answer is that if they eat before the test then that will confuse the blood glucose levels. This is correct, but you just need to adjust your answer so it makes sense and relate it to the question. So how do we do this? Well lets imagine you did eat something during those 8 hours, then your body would digest the food and put glucose into the blood. When you go and take this special drink which has a specific concentration of glucose in it, you might give a false reading because you ate something 4 hours ago or 2 hours ago. By stopping eating and drinking for 8 hours you allow your body to settle down and you also allow the test to only look at this special drink you are given. This is an important point when making a biological experiment. You want the results to be only due to the special glucose drink and not due to a cupcake or sandwich that was eaten earlier. You can also imagine the great amount of variance if many people took this test, some could eat 2 hours before the test, some 8 hours before the test. This all influences the amount of glucose in the body. By making it exactly 8 hours it give everyone the same chance to get to a stable level before the test