Special low protein foods (SLPFs) are essential in a low phenylalanine diet for treating phenylketonuria (PKU). With little known about their nutritional composition, all SLPFs on UK prescription were studied (n = 146) and compared to equivalent protein-containing foods (n = 190). SLPF nutritional analysis was obtained from suppliers/manufacturers. Comparable information about regular protein-containing foods was obtained from online UK supermarkets. Similar foods were grouped together, with mean nutritional values calculated for each subgroup (n = 40) and percentage differences determined between SLPFs and regular food subgroups. All SLPF subgroups contained 43–100% less protein than regular foods. Sixty-three percent (n = 25/40) of SLPF subgroups contained less total fat with palm oil (25%, n = 36/146) and hydrogenated vegetable oil (23%, n = 33/146) key fat sources. Sixty-eight percent (n = 27/40) of SLPF subgroups contained more carbohydrate, with 72% (n = 105/146) containing added sugar. Key SLPF starch sources were maize/corn (72%; n = 105/146). Seventy-seven percent (n = 113/146) of SLPFs versus 18% (n = 34/190) of regular foods contained added fibre, predominantly hydrocolloids. Nine percent of SLPFs contained phenylalanine > 25 mg/100 g and sources of phenylalanine/protein in their ingredient lists. Stricter nutritional composition regulations for SLPFs are required, identifying maximum upper limits for macronutrients and phenylalanine, and fat and carbohydrate sources that are associated with healthy outcomes.

A phenylalanine (protein)-restricted diet is the primary treatment for phenylketonuria (PKU). Patients are dependent on food protein labelling to successfully manage their condition. We evaluated the accuracy of protein labelling on packaged manufactured foods from supermarket websites for foods that may be eaten as part of a phenylalanine-restricted diet. Protein labelling information was evaluated for 462 food items (“free from”, n = 159, regular, n = 303), divided into 16 food groups using supermarket website data. Data collection included protein content per portion/100 g when food was “as sold”, “cooked” or “prepared”; cooking methods, and preparation instructions. Labelling errors affecting protein content were observed in every food group, with overall protein labelling unclear in 55% (n = 255/462) of foods. There was misleading, omitted, or erroneous (MOE) information in 43% (n = 68/159) of “free from” foods compared with 62% (n = 187/303) of regular foods, with fewer inaccuracies in “free from” food labelling (p = 0.007). Protein analysis was available for uncooked weight only but not cooked weight for 58% (n = 85/146) of foods; 4% (n = 17/462) had misleading protein content. There was a high rate of incomplete, misleading, or inaccurate data affecting the interpretation of the protein content of food items on supermarket websites. This could adversely affect metabolic control of patients with PKU and warrants serious consideration.