Results

The osmolality of human milk and HMF dissolved in water was similar before and after storage at 4°C. After fortification with HMF1, osmolality at T₁ was similar to the theoretical value (T₀). In contrast, the osmolality of human milk fortified with all other HMF (HMF2 to HMF5) (T₁) was significantly higher than the sum of the respective values of HMF dissolved in water and human milk (T₀) (p<0.0001) (table 2). The increase of osmolality after mixing was proportionally related to the rise of concentration of HMF2 and HMF3 (2.5 to 5%).

After 24 hours at 4 °C (T₂), the osmolality of fortified human milk was significantly higher than that observed just after reconstitution (T₁). This increase in osmolality was also present with HMF1 and varied, when all the groups were considered, from 6.9 (2.3) to 63.8 (10.7) mosmol/kg H₂O.

The difference in osmolality of fortified human milk between T₁-T₀ and T₂-T₁, expressed as mean and 95% confidence interval is shown in table 3. With HMF1, only a relatively small, late increase was observed: 6.9 (2.3) mosmol/kg H₂O. For HMF2 and HMF3, an early (T₁-T₀) and late (T₂-T₁) increase in osmolality was directly proportional to the HMF concentration, and the late increase was twice that of the earlier one. By contrast, supplementation of human milk with HMF4 and HMF5 induced an early increase in osmolality that was significantly higher than the later one.

Linear regression analysis showed that total dextrin content (r=0.84) rather than carbohydrate content (r=0.75) was the main determinant of the total osmolality increase (T₂-T₀). Mean amylase activity in human milk was similar before (+1737 (474) U/l) and 24 hours after the addition of HMF (1644 (460) U/l).

Discussion

We have shown that the addition of most HMF to human milk induces a rapid and clinically significant increase in osmolality. This effect is evident with all the HMF with the exception of HMF1 which contains only proteins. An additional effect dependent on the HMF used was also observed during storage at 4°C for 24 hours.

Recently, Jocson et al reported a mean osmolality value for human milk, after fortification with Enfamil HMF, of 408 mosmol/kg H₂O, and measured a 4% increase after 72 hours of storage. This value of osmolality is not dissimilar from the 401 and 408 mosmol/kg H₂O we found in human milk with the same fortifier 10 minutes and 24 hours after supplementation, respectively. The authors of that study found that the increase in osmolality could not be attributed to nutrient degradation following bacteriological growth.

In contrast, the amylase activity of human milk may explain the rapid and continuous increase in osmolality after fortification as well as the relation observed between osmolality and the dextrin content.

Carbohydrate distribution varied in the different HMF. For instance, carbohydrates in HMF4 are made from lactose (24%), mineral sources (glycerol from calcium glycerophosphate, gluconate and citrate) (20%), and corn syrup (56%); its dextrose equivalent (defined as reducing power of a hydrolysate compared with that of anydrous dextrose) is 24 and about 5% of dextrin linkages are similar to that of corn starch. In HMF2 carbohydrates are represented exclusively by dextrins.

The amylase in human milk is an alpha amylase that hydrolyses starch and dextrins.9 However, its activity has little specificity for the α 1–4 linkages in the vicinity of the α 1–6 branching point.10 The amylase in human milk is relatively resistant to heat, and to pasteurisation specifically.11Hamosh et al showed that about 95–100% of initial activity was present after 24 hours of storage at 15–25°C.11 Similar data were observed in the 18 samples of pasteurised human milk which we examined before and after 24 hours. Therefore, the fortification of human milk may lead to oligosaccharide production, depending on carbohydrate composition of HMF (containing molecules of carbohydrate of different lengths, such as small molecules of glucose, maltose, maltotriose, glucose tetramers as well as a quantity of oligosaccharides with more than 20 units of glucose) and on the α 1–4 :α1–6 linkage ratio. Thus the rise of osmolality observed in human milk supplemented with HMF can be explained by the fact that polysaccharides, present in the HMF, are broken down into their constituent mono- and oligosaccharides and the difference observed between the various HMF could be related to the different carbohydrate content and the use of dextrins of different origin and equivalent glucose content. This could also explain why some fortifiers, such as HMF4 and HMF5, induce a relatively early increase in osmolality, while HMF2 induces a late increase. The effect of storage (T₂- T₁) on osmolality was smaller with HMF₁ containing only proteins, and with HMF₄which had the lowest content of dextrins compared with the other fortifiers. The late increase was close to methodological error and was not clinically significant.

In recent years, there has been a dramatic increase in the practice of fortifying breast milk, to give to low birthweight infants the nutrients and other biologically important factors contained in human milk.12 The addition of a fortifier to breast milk has many potential advantages, including improved growth and bone mineralisation as well as a reduction in the incidence of hyponatraemia. Nevertheless, a lower weight gain was observed in preterm infants fed fortified human milk compared with those fed preterm formula.13 Very low birthweight infants, fed fortified human milk, may develop abdominal distension, reduced gastric emptying, frequent stools, and more rarely, diarrhoea. The supplementation, therefore, often has to be reduced or abandoned altogether, with the attendant reduction in nutrient intake and of growth. Thus the considerable increase in osmolality of fortified human milk, compared with preterm formulas (about 300 mosmol/kg/H₂O) could be partially responsible for the slower growth rate in very low birthweight infants fed HMF rather than preterm formula.

An association between hyperosmolar preparations (higher than 400 mosmol/kg H₂O) and necrotising enterocolitis has been reported.14 In a study of 275 preterm infants, Lucaset al found that, compared with control infants fed human milk, the fortified group showed an increased incidence of necrotising enterocolitis (5.8% compared with 2.2%; p=0.12).15 Although this difference was not significant, they suggested that a larger study would be required to detect a two-to threefold increase in incidence at the 5% significance level.

In conclusion, the use of human milk fortifiers increases osmolality more than would be expected from composition alone. To improve human milk feeding in very low birthweight infants, it is important to produce new fortifiers with a lower propensity to increase osmolality, but able to assure good growth, and to avoid the side effects responsible for the reduction in nutrients supplied by enteral feeding.