Untitled

glucose, starch, NSP) there is also negligible loss of cIE to
urine, and so cUE can also be ignored. Under these
circumstances (which do not apply to many polyols), the
following equation is valid
DE ¼ ME
For carbohydrates that are fully absorbed and fully metabolized
to CO2 and H 2O (‘available carbohydrates’, for
example, glucose, starch), the following equation also
applies:
cIE ¼ DE ¼ ME
For other types of carbohydrates, ME is less than cIE, either
because there is some loss of cIE to faeces (cFE) and/or urine
(cUE), as in the case of some polyols. Polyols, which are often
used as bulk-sweetening agents, have fewer detrimental
effects on teeth, lower glycemic indices and lower ME values
than conventional sweet sugars (for example, glucose,
fructose, sucrose). The proportion absorbed by the small
intestine varies considerably with the type of polyol (Bernier
and Pascal, 1990; Livesey, 1992; Finley and Leveille, 1996),
and tends to decrease with increasing molecular weight. As
much as 35% of the energy of polyols can be accounted for
in faeces (digestibility of energy, 65–100%; although polyols
are fermented and not recovered in faeces, some of the
products of this fermentation are recovered in faeces). In
addition, a variable proportion of absorbed polyols is lost in
urine. A detailed discussion about the extent of absorption
and metabolism of polyols by bacteria in the human colon
and by human tissues is beyond the scope of this paper.
However, the following brief points illustrate the variability
in their physiological handling. The small intestine absorbs
only about 2–5% of lactitol, which is almost completely
recovered in urine (95%), with little oxidation by human
tissues (suggested by studies involving intravenous administration
of C-labelled lactitol (Bernier and Pascal, 1990)).
The small intestine absorbs B25% of mannitol and, like
lactitol, it is almost totally excreted in urine without
metabolism by human tissues (again suggested by intravenous
administration of labelled (Nasrallah and Iber, 1969)
and unlabelled (Nasrallah and Iber, 1969) mannitol). In
contrast, glycerol is completely absorbed and metabolized by
human tissues, without excretion in urine. Sorbitol is
absorbed to the extent of B50%, of which the majority
(B85%) is metabolized by human tissues (Finley and
Leveille, 1996).
Most studies assessing ME of diets have been carried out in
subjects close to N and energy balance. The DE and ME
systems appear to overestimate DE and ME when energy
intake is low, especially when accompanied by high NSP
intake because apparent digestibility tends to be lower. Food
energy values reflect the supply of energy and not whether it
is spent. Some energy is deposited during growth and
development of obesity. The combustible energy that is
deposited in such situations corresponds to DE, which may
Physiological aspects of energy metabolism
M Elia and JH Cummings
not be made available to oxidative metabolism, if for
example it is retained during the growing process. Some of
it may not be made available even if it were mobilized
and metabolized, because incompletely combusted products
of metabolism, such as urea, are excreted in urine
(for protein, DE4ME). However, when carbohydrate is
mobilized, it is usually totally oxidized (combustible
energy ¼ DE ¼ ME). The issue is discussed further below—
‘Calculating energy balance’.
DE and ME values may be influenced by genetic factors
and disease. For example, although lactose is assigned a
digestibility coefficient of 1.0, so that cIE ¼ DE (also ¼ ME), in
subjects with lactose malabsorption (Matthews et al., 2005;
Montalto et al., 2006) (due to genetic factors that can affect
entire populations, or to damage of intestinal lactase by
gastroenteritis or enteropathy), not all the lactose is hydrolyzed
and absorbed by the small intestine. Some of it reaches
the colon, where it is fermented to produce faecal matter and
gaseous end products (both of which reduce DE and ME) and
SCFAs, which are absorbed and metabolized by human
tissues. Another example concerns fructose, which when
ingested in small quantities is fully absorbed by the small
intestine, but when ingested in large quantities, can reach
the large bowel, where it is fermented (Truswell et al.,
1988). This fermentation results in some loss of energy
to faeces (cFE) (for example, as bacterial biomass), which
again reduces its DE and ME values. Yet another example
involves loss of glucose in the urine of diabetic patients
(metabolizability of o1.0; and MEoDE). Since DE and ME
values vary between individuals and are affected by disease,
the values in general use are based on average results
obtained in ‘healthy’ subjects using doses of substrates that
are likely to be ingested. Any errors in establishing DE values
will not only affect ME values but also NME values, which is
discussed next.
Net metabolizable energy
A criticism that has been raised against the ME system is that
it fails to consider differences in the efficiency with which
substrates supply biologically useful energy. The NME system
is based on the concept of biological efficiency at the level of
ATP and takes into account two specific processes, both of
which have the effect of reducing metabolic efficiency. First,
the heat of fermentation does not yield ATP to the host.
Since this lowers the overall bioenergetic efficiency of the
fermentable substrate, some workers have suggested using
the term ‘true metabolizable energy’, which they have
defined as ME minus the heat of fermentation (Bar, 1990;
Bernier and Pascal, 1990). This approach, which is used in
animal nutrition, differs from that in human nutrition,
where the heat of fermentation is included in ME. Second,
different fuels yield net ATP (ATP gain) with different
bioenergetic efficiencies (Elia and Livesey, 1988, 1992;
Blaxter, 1989). Those with lower efficiencies (for example,
protein) will result in more heat production for the
S45
European Journal of Clinical Nutrition