The epidemiological evidence associating dietary lipids, refined carbohydrates and coronary heart disease (CHD) is evaluated. Population studies often show a high correlation between dietary lipid and heart disease mortality. It is argued that most of this association is secondary to the high correlation that refined carbohydrates have with dietary lipids, on the one hand, and with CHD on the other. This becomes apparent only when examples are found of a weak or negative correlation between dietary lipids and refined carbohydrates. Evidence from diet and drug intervention studies support the hypothesis that dietary lipid is only of secondary importance in CHD. It is concluded that refined carbohydrates are of primary importance.