Eur J Clin Nutr. 1992 Oct; 46(10): 717-28.

Validity of repeated dietary measurements in a dietary intervention study.

Johansson G, Callmer E, Gustafsson JA.

Department of Medical Nutrition, Karolinska Institute, Huddinge University Hospital, Sweden.

The aim of the study was to evaluate the compliance in a dietary intervention study. When drawing conclusions about the relationship between dietary intake and disease occurrence/disease-related variables it is important to obtain valid dietary data. 20 healthy, non-smoking normal-weight omnivores changed from a mixed to a lactovegetarian diet. Dietary surveys (four 24 h recalls per person and time-period), urinary and faecal sample collections were performed before and 3, 6 and 12 months after the dietary shift. The validation of energy, protein, sodium and potassium yielded approximately the same ratio of dietary intake to biological marker at 0 and 3 months. This ratio decreased towards 6 months and continued to decrease towards 12 months. The fibre intake was compared to the total faecal weight directly and indirectly by calculating the fibre intake from the stool weight, the water content in faeces and the excretion of short-chain fatty acids (SCFAs). These four methods of fibre validation showed that the ratio of dietary intake to biological marker was always highest at 12 months, indicating an overestimation of the fibre intake at the end of the study. This is the first time these methods of validating fibre intake have been used in an epidemiological study. The ratio of dietary calcium intake to urinary and faecal calcium excretion did not show any statistical difference between the period before and 3 months after the dietary shift. To conclude, almost all investigated dietary data show approximately the same validity before and 3 months after the dietary shift, and show the least validity 12 months after the dietary shift. Thus, this study demonstrates that it is difficult to obtain valid dietary data 1 year after a drastic dietary change, indicating a decreased compliance to the new dietary regimen at the end of the 1 year study period. This represents important information when attempting to relate biological effects to dietary intake, and illustrates the importance of using biological markers for food intake in dietary surveys.