For the first time, body composition has been studied in a very large sample of female anorexia nervosa patients (n 200) using two methods: anthropometry (skinfold thickness) and densitometry (underwater weighing). The concurrent validity of both methods appeared to be good (r 0·84). Although the mean percentage of body fat (13·5) in our sample corresponds with most of the previous reports, the range (4·3–24·8) is great. Our study shows that BMI is not a good measure of fatness. Age and duration of illness were not related to body fat. Percentage fat was found to be different (t 2·76, P <0·01) according to the diagnostic subtype of anorexia nervosa: restricters (12·9) v. bingers and/or purgers (14·7).

In view of the importance of long-chain polyunsaturated fatty acids (LCP) for growth and development of fetal and infant neural tissue, the influence of the dietary n-3 and n-6 LCP intake onthe LCP status of forty-three preterm infants (birth weight<1800 g) was studied. Thirty-one formula-fed infants were randomly assigned to receive a conventional formula lacking LCP (n 16), or an 22:6n-3-and 20:4n-6-enriched formula (n 15); twelve infants received their own mother's breast milk. Fatty acid compositions of plasma and erythrocyte (RBC) phospholipids (PL) were determined in umbilical venous blood, in weekly postnatal samples until day 35 of life and, for the formula-fed infants, at 3 months of corrected age. Both in plasma (P < 0·001) and RBC (P < 0.01) PLY, the changes with time until day 35 for 22: 6n-3 and 20:4n-6 in the two groups of formula-fed infants were significantly different, with higher values, comparable with those of human-milk-fed infants, in the LCP-enriched-formula group. At 3 months of corrected age, differences between the two formula-fed groups were even more pronounced. In conclusion, adding 22: 6n-3 and 20:4n-6 to artificial formulas in balanced ratios and in amounts similar to those found in preterm human milk raises both the 22:6n-3 and the 20:4n-6 status of formula-fed preterm infants to values found for human-milk-fed preterm infants. Additional studies are necessary to evaluate the potentially favourable effects of this combined addition on the neurodevelopmental outcome of preterm infants.

The aim of the present study was to evaluate the acute effect of ethanol on insulin sensitivity, and glucose, insulin, free fatty acid (FFA), and triacylglycerol responses in ten patients with non-insulin-dependent (type 2) diabetes. In the test study an oral dose of 0·66 g ethanol/kg followed by continuous intravenous infusion of 0·1 g ethanol/kg per h was given to maintain a constant ethanol level in the blood. In the control study identical volumes of oral water and intravenous saline (9 g NaCl/l) were given. After 90 min insulin sensitivity was determined by the hyperinsulinaemic, euglycaemic clamp technique. Ethanol caused no change in blood glucose or insulin concentrations. The FFA level was suppressed by ethanol while the triacylglycerol level was unaffected. The insulin sensitivity was not affected by ethanol. No major acute effect of ethanol on the glycaemic control in fasting type 2 diabetic patients was found in comparison with what is seen in healthy people. The present study, along with the sparse literature, indicates that the ability of ethanol to induce hypoglycaemia is attenuated or absent in diet-treated type 2 diabetes. Furthermore, we found no change in insulin sensitivity. Consequently, the risk of acute ethanol-induced aberrations in carbohydrate metabolism in diet-treated type 2 diabetes seems to be less than previously expected, when alcohol is not taken as a part of a meal.

The absorption of Zn or Fe from breakfast meals containing oat porridge prepared from malted and soaked oats and a control porridge made from untreated oats was measured in human subjects. The effect on Zn and Fe absorption of reducing the phytate content of oat-porridge meals was examined in each subject by extrinsic labelling of porridge with 65Zn and of bread rolls with 55Fe and 59Fe, and measuring whole-body retention and the erythrocyte uptake of isotopes. Each experiment comprised nine to ten subjects. The absorption of Zn from malted-oat porridge with a phytate (inositol hexaphosphate) content of 107 μmol was 18·3%, and significantly higher (P < 0·05) than from the control porridge containing 432 μmol phytate (11·8%). Fe absorption from the meal containing malted-oat porridge with 107 μmol phytate (Expt 2) was also significantly improved (P < 0·05) compared with that from the meal containing control porridge with 437 μmol phytate. The average increase in Fe absorption was 47%, or from 4·4 to 6·0%. In the breakfast meal containing malted porridge with 198 μmol phytate (Expt 3) the increase in Fe absorption was not significantly improved. Even though the phytate content was reduced to a greater extent in Expt 3 than Expt 2, the average increase in Fe absorption in Expt 3 was only 25% more than that from the meal containing control porridge (with 599 μmol phytate), depending on the higher absolute amount of phytate. In conclusion, an improvement in Zn and Fe absorption from oat products can be achieved by practising malting and soaking in the processing of oats. This may be of importance in the prevention of mineral deficiency in vulnerable groups.

The absorption of lysine synthesised by the gastrointestinal microflora was estimated by comparing the15N incorporated into body lysine in four germ-free (15N-GF) and four conventional (15N-CV) rats. They were fed for 10d on a protein-free diet containing fermentable carbohydrates and 15NHM4Cl; another four conventional rats (control), fed on the same diet but with unlabelled NH4Cl, were used to estimate the natural abundance of 15N. The eviscerated carcass of each rat was homogenized and a sample hydrolysed. Lysine was isolated by ion-exchange chromatography and its 15N enrichment was measured by isotope-ratio mass spectrometry. The 15N-CV rats significantly incorporated 15N into their body lysine. The 15N-GF rats had a statistically significant, although small, incorporation of 15body lysine, probably arising from a measurement artifact. It was concluded, therefore, that all [15N]lysine was of microbial origin. The total lysine content in the body and the 15N enrichment of lysine in the microbial fraction of the faeces of the 15N-CV rats were also determined. The amount of microbial lysine absorbed by the 15N-CV rats was estimated by dividing the total amount of [15N]lysine in the body by the enrichment of microbial lysine. It was estimated that the daily absorption of microbial lysine by the conventional rats was 21·3 (SE 2·04) mg/kg body weight0·75

Four rats were housed in cages with mesh floors; another four rats were housed in tubular anti-coprophagy cages, in which they could not turn round to reach their own faeces. Both groups were fed for 6 d on a low-protein diet containing fermentable carbohydrates and 15NH4Cl. At the end of the experiment the rats were killed and their carcasses were homogenized, lysine was isolated by ion-exchange chromatography and its 15N enrichment measured by isotope-ratio mass spectrometry. The 15N enrichment in the lysine of the microbial fraction of faeces and the total amount of lysine in the body were also determined in order to estimate the amount of microbial lysine absorbed. The 15N enrichment in body lysine of non-coprophagic rats was not different from that previously measured in rats given unlabelled NH4Cl, but in coprophagic rats it was significantly higher. The daily absorption of microbial lysine by the coprophagic rats accounted for 20·7 (SE 2·55) mg/kg body weight0·75 but was only 0·5 (SE 1·04) mg/kg body weight0·75 for the non-coprophagic rats. This value was not significantly different from zero. The utilization of microbial amino acids via coprophagy resulted in a higher weight gain (adjusted for intake) in the coprophagic group (15·5 g/6 d) than in the non-coprophagic rats (3·1 g/6 d). It was concluded that, in rats, the utilization of microbial lysine occurred exclusively via coprophagy.

Daily (24 h) rates of water and electrolyte turnover were measured in a conventional balance study in ten adult female pastel mink (Mustela vison) given free access to a standard mink feed for a 1-week conditioning period, followed by a 4 d experimental period and a 2 d fasting period. Drinking water was available throughout. In addition, the completeness of urine collection and the fraction of urine collected with the faeces were determined using a new experimental technique based on 24 h recoveries of specific urinary markers such as tritiated p-aminohippuric acid ([3H]PAH) or 14C-1abelled inulin ([14C]IN) continuously delivered by small Alzet® osmotic pumps implanted intraperitoneally. During feeding the mean individual percentage recovery in urine of [3H]PAH released from tbe osmotic pumps ranged from 68 to 88% (median 78%). Tbe mean percentage of urinary [3H]PAH recovered from faecal collections was 6% (range 3–12%). In response to fasting the mean individual percentage recovery of [3H]PAH in urine ranged from 62 to 78% (median 68%). For urinary [14C]IN the mean percentage recoveries in fed and fasted animals were 79 and 63% respectively. Furthermore, during fasting, withdrawal of the supplies of dietary water caused a slight but insignificant (P = 0·17) increase in the daily intake of drinking water and, hence, the animals maintained their normal water balance by a dramatic reduction in urine excretion (P < 0·001). At the same time urinary solute excretion declined significantly (P < 0·001), due in part to the cessation of dietary electrolyte intake and in part to reduced formation of urea, whereas urinary osmolality decreased only moderately. The mean 24 h balances of Na, K, Ca, Mg, Cl and P were close to zero and only minor differences between the feeding and fasting periods were observed. When corrected for the measured inaccuracies in urine collection the balance data obtained in the present study represent useful reference standards for normally fed and fasted non-growing mink and, to some extent, useful guidelines for future studies in experimental animals.

An in vitro dialysis method was employed to determine the effect on the Fe and Zn absorption of the type (beef, pork and soyabean) and the amount (10 and 30 g/kg) of protein present. In addition, the effects of low- and high-molecular-weight (LMW and HMW respectively) digestion products were investigated. After in vitro digestion and dialysis a lower percentage of N, Fe and Zn was found in the LMW fractions from beef, pork and soyabean proteins when the protein level was increased from 10 to 30 g/kg; the higher level of protein being associated with a lower percentage of hydrolysed protein. The highest percentage levels of intrinsic Fe were always found in the HMW fractions, independent of the type and the level of proteins studied, while in the case of Zn, both HMW and LMW fractions gave similar values. An interaction was found between inorganic Zn and non-haem-Fe. The addition of inorganic Zn (10 μg/ml) caused a significant decrease in the in vitro availability of Fe from soyabean protein, while it did not affect the dialysability of intrinsic Fe from beef and pork proteins. Our results showed that the type and the level of the protein had a positive effect on the dialysability of extrinsic Fe. We postulate that the effect of a protein on the absorption of extrinsic Fe could be accounted for by free amino acids and/or small peptides released during the digestion process and also by the undigested or partially-digested HMW fractions of hydrolysed proteins which could play a fundamental role in the availability of this essential element.

Two experiments were conducted to investigate whether the utilization of lysine in growing pigs is affected by the level of excess protein in the diet. Nine lysine-deficient diets containing 100,200 or 300g crude protein/kg and between 1·2 and 6·8 g ileal digestible lysine/kg were prepared. In the first experiment the apparent ileal digestibility of lysine in three of the nine diets was determined using pigs with simple T-cannulasand Cr2O3, as an indigestible marker. Ileal digestibility of lysine in the other diets was calculated by interpolation. In the second experiment N retention, as a measure of lysine utilization, was determined in all nine diets using growing pigs over the weight range 30–50 kg. The effect of excess protein on lysine utilization was assessed by comparing the regression of N retention v. lysine (ileal digestible) intake at the three levels of protein. Increasing ileal digestible lysine in the diets resulted in a linear increase in N retention with all three protein levels and there was no significant difference amongst the three regressions, indicating that lysine utilization was not affected by the level of protein.Therefore, all data were pooled together to calculate a single regression for all treatments.An increase of 1·0 g ileal digestible lysine led to an increase of 1·43 g N or 8·96 g protein (N x 6·25) retained. Assuming a lysine concentration in the retained body protein of 65–72 mg/g, lysine was utilized with an efficiency of 0·58–0·65

Boiled coffee contains the lipid compounds cafestol and kahweol, which raise cholesterol strongly in man. These lipids are retained by paper filters. In a search for an animal model for the effect of coffee lipids on serum cholesterol concentrations, we fed hamsters (Mesocvicetus aurutus) and rats on mash diets consisting of a purified base diet and either boiled water, unfiltered boiled coffee or filtered boiled coffee. After a feeding period of 8 weeks there was no statistically significant effect of unfiltered boiled coffee on serum total cholesterol and triacylglycerol concentrations in either the hamsters or the rats. The level of serum cholesterol did respond predictably to the addition of cholesterol and/or saturated fatty acids to the diet. The lack of effect of unfiltered boiled coffee in the hamsters and the rats, when compared with the previously reported activity in humans, could not be explained by dosage, duration of treatment, mode of administration or by insufficient statistical power. It is concluded that hamsters and rats are insensitive to unfiltered boiled coffee and thus are unsuitable models for investigating its hyper- cholesterolaemic effect.

Supersaturation of bile with cholesterol is a prerequisite of the development of gallstones. With the intention to study the integrated response of enzymes regulating hepatic cholesterol metabolism during gallstone formation we used an established model for the induction of cholesterol gallstone disease in mice. Ten mice were fed on a lithogenic diet containing 10 g cholesterol/kg and 5 g cholic acid/kg for 8 weeks and were compared with ten mice fed on a standard pellet diet. Cholesterol crystals or gallstones developed in 90% of gallbladders in treated mice. The lithogenic diet had an inhibitory effect on the rate-limiting enzyme of cholesterol biosynthesis, hepatic 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase (EC 1.1.1.88) activity, 39·6 (SEM 2·8) v. 171·0 (SEM 47·3) pmol/min per mg protein. Cholesterol 7α-hydroxylase (EC 1.14.13.17) activity, regulating bile acid synthesis, was decreased by 80%, and this was assumed to be due to cholic acid in the diet. The cholesterol-enriched diet also induced a tenfold increase in cholesterol esterification rate in the liver, i.e. acyl-CoA: cholesterol acyl transferase (ACAT; EC 2.3.1.26) activity. The total, as well as esterified, cholesterol contents of liver homogenates were significantly higher in cholesterol- and cholic acid-treated mice and correlated well with the ACAT activity (rs 0·72 (P < 0·005), and rs 0·68 (P < 0·01) respectively). A significantly higher ACAT activity was obtained in mice given cholesterol and cholic acid even when the enzyme was saturated with exogenous cholesterol, thus indicating an increased amount of the enzyme. The formation of gallstones is dependent on a delicate balance between lithogenic factors (increased absorption of cholesterol and reduced secretion of bile acids) and defence mechanisms (decreased synthesis and increased esterification of cholesterol). In the specific animal model studied here the two defence mechanisms cannot compensate for the increased absorption of cholesterol and the reduced synthesis of bile acids.

Effects of feeding soluble dietary fibre on apparent Ca absorption and the contribution of the caecum to Ca absorption were examined in five-sixths nephrectomized (NPX)and normal rats with or without caecectomy in four experiments. It is known that Ca absorption is lowered by renal failure. In the first experiment the amounts of femur Ca increased linearly with increasing dietary Ca up to 3·0 g Ca/kg diet in intact rats. Partial nephrectomy decreased apparent Ca absorption in rats fed on diets containing 30 and 4·5 g Ca/kg diet. In the NPX groups, Ca absorption in rats fed on the diet containing guar-gum hydrolysate (GGH; 50 g/kg diet; 3·0 g Ca/kg diet) was significantly higher than that in rats fed on a fibre-free diet, and the increase in Ca absorption with GGH feeding was completely abolished by caecectomy. Also, ingestion of GGH increased Ca absorption in normal rats, but not in normal, caecectomized rats. Mg absorption was also increased with GGH feeding and was decreased with caecaectomyin NPX and normal rats. In experiments which used caecectomized rats, coprophagy wasprevented with an anal cup to avoid re-ingestion of faecal Ca. We conclude that ingestion of the soluble dietary fibre, GGH, increased apparent Ca absorption in NPX and non-NPX rats, and the caecum was responsible for these increases in Ca absorption.