This work examined the adequacy of the only food composition database avail
able in Croatia to estimate fat and fatty acid intake. A 7-day food record
with 39 participants was carried out. In addition to Croatian food composit
ion tables (FCT1), the Swedish food tables (FCT2) were used to convert food
records into quantities of total fat and saturated, monounsaturated and po
lyunsaturated fatty acids, respectively. The subjects simultaneously collec
ted duplicates of all consumed food. The chemical analysis of duplicates to
determine total fat and fatty acids included extraction of fat, followed b
y methylation of fatty acids and GC analysis. According to the comparison o
f methods, FCT2 are a better predictor of fat and fatty acid intake than FC
T1 because of their closer agreement with analysed values. The correlation
coefficients imply that both food tables estimate most of these nutrients s
imilarly to a reference method. However, food records combined with either
of the food tables could not predict intake of polyunsaturated fatty acids
by men. In order to improve applicability of FCT1 in dietary surveys concer
ned with fat and fatty acids, the next issue should include changes in resp
ect to an identified systematic error, as well as a greater choice of food
items and their varieties. Accuracy of food records based on FCTI could fur
ther be increased with subjects giving details on recipe ingredients, analy
sis of missing foods, etc.