Black A E, Ravenscroft C, Paul A A
Hum Nutr Appl Nutr. 1985 Feb;39(1):9-18.
The differences in nutrient intakes as calculated by two British food composition tables (those compiled by the Department of Health and Social Security in 1963 [DHSS] and the 4th edition of McCance and Widdowson's 'The composition of Foods' [MW4] were examined. Intakes of 43 dietitians who kept weighed diet records every sixth day for 3 to 12 months with 11 to 75 (mean 40) days' records per subject were calculated from each food table and compared in a loge tranformation. Mean per cent differences between intakes (delta x 100) were vitamin A, +23.3; vitamin D, + 21.1; riboflavin, + 16.0;calcium, +6.3; thiamin +4.8; niacin, +3.2 (MW4 higher) and protein, -1.7; fat, -3.5; energy, -3.6; carbohydrate, -6.0; vitamin C, -6.1; iron, -11.3; vitamin B6, -14.4 (MW4 lower). Correlation coefficients were: nicotinic acid, 0.84; vitamin B6, 0.86; other nutrients 0.90 to 0.99; thus differences between the two food tables in classifying individuals as high or low consumers were small. Mean intakes compared by paired t-test were significantly different (P less than 0.001) for all nutrients except nicotinic acid, indicating consistent differences in the food tables when applied to foods eaten by different individuals (nicotinic acid excepted). Mean intakes compared by one-tailed t-test -- as if the data came from independent surveys -- were significantly different/only for riboflavin/and vitamin/B6 (P less than 0.01), vitamin A and iron (P less than 0.05). For these nutrients, significant differences in intakes could be found between two surveyed populations due solely to the use of different food tables and not to differences in foods eaten. For the other nutrients studied, differences between the food tables were of little practical importance. Differences were due primarily to new analytical values in MW4 for riboflavin in milk, retinol in liver, iron in meats, vitamin B6 in many foods, and the use of margarine fortified with vitamins A and D in recipes. Major differences in the (high) nicotinic acid content of beer, instant coffee and specific breakfast cereals resulted in unpredictable differences in intakes of individuals.
研究了根据两份英国食物成分表（1963年卫生与社会保障部编制的那份[DHSS]以及麦坎斯和威德森的《食物成分》第4版[MW4]）计算出的营养素摄入量差异。计算了43位营养师的摄入量，他们每隔六天记录一次饮食重量，记录时长为3至12个月，每位受试者有11至75天（平均40天）的记录，分别根据每份食物表进行计算，并在对数转换后进行比较。摄入量之间的平均百分比差异（δ×100）为：维生素A，+23.3；维生素D，+21.1；核黄素，+16.0；钙，+6.3；硫胺素，+4.8；烟酸，+3.2（MW4更高）；蛋白质，-1.7；脂肪，-3.5；能量，-3.6；碳水化合物，-6.0；维生素C，-6.1；铁，-11.3；维生素B6，-14.4（MW4更低）。相关系数为：烟酸，0.84；维生素B6，0.86；其他营养素，0.90至0.99；因此，在将个体分类为高消费者或低消费者方面，两份食物表之间的差异很小。通过配对t检验比较的平均摄入量，除烟酸外，所有营养素均有显著差异（P<0.001），这表明在应用于不同个体所吃食物时，食物表存在一致的差异（烟酸除外）。通过单尾t检验比较的平均摄入量（就好像数据来自独立调查），只有核黄素、维生素B6（P<0.01）、维生素A和铁（P<0.05）有显著差异。对于这些营养素，仅由于使用不同的食物表，而不是所吃食物存在差异，就可以在两个被调查人群之间发现摄入量的显著差异。对于所研究的其他营养素，食物表之间的差异在实际中不太重要。差异主要是由于MW4中牛奶中核黄素、肝脏中视黄醇、肉类中铁、许多食物中维生素B6的新分析值，以及食谱中使用了添加维生素A和D的人造黄油。啤酒、速溶咖啡和特定早餐谷物中（高）烟酸含量的重大差异导致个体摄入量出现不可预测的差异。
