Issue # 40: Phytoestrogens and
hot flushes
Daily soy supplementation is effective
in reducing the number of hot
flushes in post-menopausal women,
according to the results of a
double-blind placebo-controlled trial
from Italy just published.
Researchers studied 104 post-menopausal
women who had requested treatment
for severe hot flushes. All had a
history of moderate to severe hot flushes
occurring at least 7 times daily for
no less than 2 weeks prior to
commencing the study. None were taking
hormonal or other therapy for their
symptoms.
Fifty one of the women were placed
on 60 g/day of isolated soy protein
supplement for 12 weeks, whilst the
remaining 53 patients took 60g/day of
casein. Both supplements contained
the same amount of protein, but the soy
supplement had in addition 76 mg of
isoflavones. Hot flushes were monitored
by means of diaries, and the data
analysed using an intention-to-treat model.
Soy supplementation was associated
with significantly more reduction in
number of hot flushes compared with
placebo, with the difference beginning
after two weeks and increasing through
to the end of the study. By the 12th
week, the drop in flushes for the
soy group was 45% of baseline, compared
with 30% for placebo. Subtracting
placebo response left a statistically
significant drop in mean frequency
of 1.59 flushes/day attributable to
supplementation (from a baseline of
11/day, 95% CI: 1.2 - 1.95).
There were some less positive findings
as well. A noticeable lessening of
effectiveness of active treatment
was associated with a drop in compliance
after the 8th week. Whilst this may
have been partly related to the timing
of the trial (during the Italian summer
holidays), side-effects were
reported by more than two thirds of
the women. These were mostly GIT
related (constipation, bloating) and
were seen more in the placebo than the
active group. Drop-out rates were
21% for the active and 26% for the
placebo group.
Ref:
Obstetrics & Gynecology: 9;1998:6-11
COMMENT
The story of soy as a food source
of considerable interest continues to unfold. The fact that substituting
soy for other sources of protein favourably impacts on hyperlipidaemia
is now well established.
But more interest of late has been
focused on the phytoestrogen side of the
story. In particularly, can the phytoestrogens
contained in soy help
prevent or treat oestrogen-related
disorders, such as hormone-dependent
cancers and post-menopausal symptoms?
The evidence has been slowly forthcoming.
First the relatively crude
epidemiological observations linking
countries like Japan, which have very
high levels of soy consumption, to
lower rates of menopausal symptoms and
other oestrogen-related disorders
including breast cancer (e.g. J Nutr
1995;125: 709S-712S) .
Then rather more specific associations,
such as case-control studies,
correlations between cancer incidence
and consumption of specific soy foods
and between urine isoflavone excretion
and incidence of breast and prostate
cancer. A very recent paper from Australia,
for example, reported a link
between excretion of urine levels
of phytoestrogen excretion metabolites
(equol and enterolactone) and lower
risk of breast cancer. (Lancet
1997;350: 990-994). Laboratory and clinical work has elucidated
plausible
mechanisms for clinical action (e.g.
the effects of isoflavones on the
female hormonal axis in humans).
It is heartening, therefore, that
we are beginning to see controlled
clinical trials, which up to now have
been in short supply, despite the
rise of a burgeoning `health food'
industry in phytoestrogen supplements.
At the Second International Symposium
on the Role of Soy held in Belgium in
late 1996 a number of controlled trials
were reported to be underway in
high risk subjects for prostate, breast
and colon cancer, but as yet no
published results are to hand (see
Web Site of the Week below) .
At that same Symposium, some small
clinical trials on women with hot
flushes were reported, but with equivocal
or weakly positive outcomes.
This paper is therefore amongst the
first good peer-reviewed trials to show
that soy supplementation has significant
clinical benefits to women
suffering from symptoms of oestrogen
deficiency. Clearly there are many
more steps to be taken, the first
of which is to confirm these results in
subsequent trials. This is particularly
important when dealing with
symptoms such as hot flushes where
a high placebo effect is to be expected
(and which was seen in this study).
Should the benefits be confirmed,
one of the most interesting issues will
be to determine whether we are seeing
an effect purely related to
pharmacological actions of specific
isoflavones (e.g. genistein, daidzein,
glycitein), or whether soy as a food
is important in the equation. If
so,
which soy preparations? What doses?
At present, much of the phytoestrogen
supplement industry is based on
extracted and processed forms of the
humble soybean, and doses are aimed at
replicating the kind of isoflavone
intakes seen in Japan. But it is not
clear whether these doses are adequate
or necessary. Entirely different
plant sources of isoflavone are also
under consideration by the
pharmaceutical industry.
The high proportion of women experiencing
side-effects and the difficulties
in compliance taking these food-based
supplements highlights the question
of whether women will be best off
taking isoflavone supplements (and if so
what preparations have the least side-effects)
or simply eating more soy
food in their normal diet.
Return to index
Issue # 41: Hospital malnutrition
Malnourishment
is still remarkably common in hospital in-patients at
admission,
and adversely affects the incidence of illness complications,
according
to the results of a new study from Holland.
Researchers
studied 155 patients hospitalised for internal or
gastrointestinal
diseases. Three measures of nutritional status were
employed:
* the Subjective
Global Assessment - a physical
examination plus
questionnaire
* Nutritional
Risk Index - a formula involving serum albumin and weight
* Maastricht
Index - a similar formula which also includes prealbumin and
lymphocyte
count
The proportion
of patients assessed as being malnourished on admission was
a little
less than a half to two thirds, depending on the method used:
Prevalence of malnutrition
------------------------------------------------------------
Subjective
Global Assessment: 45%
Nutritional
Risk Index 57%
Maastricht
Index 62%
The risk
of suffering any complication during the course of hospitalisation
was two
to three times greater for malnourished compared with
well-nourished
patients. The risk was reduced but still significant when
adjusted
for the nature and severity of the admitting condition itself. The
table below
shows the exact ratios.
Odds ration for risk of complications
(95% CI)
---------------------------------------------------------------------------
Index
Crude Adjusted for disease
----------------------------------------------------------------------------
Subjective
Global 2.7
1.7
Assessment (1.4 - 5.3) (0.8 - 3.6)
Nutritional
Risk 2.8
1.6
Index (1.5 - 5.5)
(0.7 - 3.3)
Maastricht
Index 3.1
2.4
(1.5
- 6.4) (1.1 - 5.4)
Ref: Am J Clin Nutr 1997;66:1232-9.
COMMENT
What impresses
most about this new finding is not so much that it
represents
a new discovery, but that the problems of hospital malnutrition
which it
confirms are so persistent.
We have
known for many years that a significant proportion of patients
admitted
to hospital are malnourished, that the proportion rises the longer
they spend
within the hospital's hallowed walls, and that this
malnourishment
adversely affects clinical outcomes. Had the patients in
this study
been from a surgical rather than a medical service, the problem
would have
been even worse.
The real
issue for nutritionists is perhaps no longer just to identify the
problem,
but to ask why it is so resistant to change.
Admittedly,
the situation is improving in some countries, such as the USA,
where the
role of clinical nutritionists in hospital care is now much
better
established. Indeed as Dr. Roland Weinsier from the University of
Alabama
points out in his editorial accompanying this new research, there
are grounds
for thinking that these new results reflect some improvement
compared
with previous studies.
Yet in
other countries, such as Australia, where reports on malnutrition in
routine
hospital admissions go back 20 years (e.g. see Aust N Z J Surg
1980;5:516-9
and Aust N Z J Med, 1987:2:234-40), but where there is no
clear path
for training physician nutritionists, there is little to suggest
the message
has made much impact.
In thinking
about this dilemma, two obvious
questions spring to mind:
1. Have we produced convincing evidence that treating hospital
malnutrition produces
cost-effective benefits in patient outcome?
Although
there is more work to be done, the evidence in favour of this
notion
is now quite good. It has recently been well reviewed in the
consensus
statement published in the Journal of Enteral and Parenteral
Nutrition
(J Parent Ent Nutr 1997; 21:133-56)
2. Do generalist physicians have the knowledge and confidence
to identify and
treat these patients?
This may
well be the crucial issue. The present paper makes it clear that
there are
a variety of simple models which can help clinicians who are not
nutrition
specialists to identify malnutrition. Whilst the results suggest
that the
less sophisticated methods are not so sensitive, it is still clear
that a
couple of routine clinical measurements and a few clicks of the
calculator
can do a good job.
Do generalist
physicians understand the treatment options available to
them? There
is little evidence to encourage optimism about this. To many
clinicians,
the gamut of possibilities - ranging all the way from a
dietetic
consult through the various modes of nutritional support all the
way to
TPN - can seem confusing and intimidating.
The most
fruitful avenues for future research may well lie in demonstrating
that simple
protocols for diagnosis, combined with user-friendly options
for management,
can effectively change the habits of the average primary
care and
hospital-based physician so that they will do something about it.
Return to index
Issue # 42: Iron in children
Study one: Iron and low fat diet in
children
Low fat
diet does not need to place children's iron or zinc status at risk,
according
to the findings of a new Finnish study.
Researchers
in the Special Turku Coronary Risk Factor Intervention Project
for Babies
studied a sub-sample of 79 out of 540 children (over 7 months of
age) whose
families were given counselling to reduce fat intake,
as part
of a general
approach to minimising preventable coronary heart disease risk
factors.
Four day food records were made, along with blood testing for
iron/zinc
status. These data were compared with those from a sub-sample of
39 out
of 522 control children, whose families received no specific
counseling
concerning dietary fat.
The children
in the intervention group consumed less saturated fat than
those in
the control group and had continuously higher ratios of dietary
polyunsaturated
to saturated fatty acids. There were no significant
differences
in mean daily iron or zinc intakes, nor blood measures of iron
or zinc
status, between intervention and control groups (see table).
Intervention Control
------------------------------------
Iron intake 8.8 mg 8.6 mg
(± 4.2) (±
2.8)
Zinc intake 7.5 mg 7.4 mg
(± 1.2) (±
1.3)
S.Ferritin 21.8 µmg/L 19.2 µmg/L
(± 11.6) (±
12.4)
Transferrin 2.34 mg/L 2.29 mg/L
receptor ( ± 0.46)
(± 0.39)
Serum zinc
11.2 µmol/L 10.5 µmol/L
(± 1.9) (±
1.6)
There were
no significant differences between groups in any of these parameters,
nor
in relation
to blood screen (Hb, MCH, MCV) or transferrin levels.
Ref: Am J Clin Nutr 1997;66:569-74.
COMMENT
The blood
tests used here were certainly sensitive enough to pick up any
adverse
effect on iron status - including the relatively newly available
transferrin
receptor assay. On the other hand, the measures of zinc status
were not
quite so sensitive.
These results
will reassure dietitians and physicians that there is
no reason
to sacrifice iron (or zinc) status in the pursuit of a low fat
diet in
children.
This is
the other side of the now well established fact that high-iron
foods,
such as lean meat, chicken and fish are also relatively low fat and
can safely
be included in a cholesterol-reducing diet.
Australian
research in adults has shown, for example, that provided the
diet has
less than 10% energy from saturated fat and adequate
monounsaturated
and omega 3 fats, lean red meat (which can be quite rich in
monounsaturated
and omega-3 fats) is quite compatible with a cholesterol-
reducing
diet (J Am Diet Assoc 1993;93:644-648).
So in theory
there is no reason to trade off iron or zinc status for lipid
reduction.
This study confirms the theory in reality, showing that
carefully
counselled families are quite capable of reducing their
children's
saturated fat intake without adversely affecting iron or zinc
status.
Study two: Iron supplementation in
pregnancy
Iron supplementation
in pregnancy improves both maternal and infant iron
status,
according to recent results from an African trial.
Two groups
of pregnant women were studied from 28 weeks gestation at
antenatal
clinics in Niger. Half received 100 mg elemental iron/day for the
rest of
the pregnancy, whilst the remainder were given only placebo.
There was
a marked reduction in the prevalence of both third trimester
anaemia
and iron deficiency in the iron-supplemented women, and none in the
placebo
group.
At birth,
there were no differences between the iron status measures in
cord blood
iron from either group. However, the babies of iron-supplemented
mothers
had significantly greater average length and Apgar scores, compared
with placebo.
Three months
post-partum, the prevalence of anemia remained significantly
lower in
the mothers of the supplemented women than placebo, whilst serum
ferritin
concentrations were significantly higher in infants of the
supplemented
mothers.
Ref: Am J Clin Nutr 1997;66:1178-82
COMMENT
The finding
that iron supplementation during pregnancy is effective in
protecting
mothers' iron status is not new, The impact of iron deficiency
on the
`pregnancy' itself - i.e. on pregnancy outcomes - is also consistent
with previous
work which has shown adverse impacts of iron deficiency on
rate of
premature delivery, low birth weight etc.
What is
notable here is the finding that iron supplementation also improves
the baby's
iron status in the first three months post-partum. This seems to
run counter
to the conventional wisdom that babies are efficient `leeches'
of maternal
iron stores and (except in cases of prematurity or complete
maternal
iron depletion) are unlikely to suffer directly from iron
deficiency
within the first few months of life.
It remains
to be seen what the exact mechanism of this effect might be,
especially
in light of the lack of difference in cord iron status between
the two
groups. However, this study lends strong support to the growing
re-emphasis
that pregnant women should routinely be offered iron
supplementation.
Return to index
Issue # 43: Diabetes and magnesium
Study #1
Magnesium
levels are associated with the presence of
microalbuminuria
and clinical proteinuria in diabetics, according to the
results
of a recent Italian study. This suggests that decreased platelet
magnesium
concentration could be an additional risk factor in the
development
of the microvascular complications of diabetes.
Plasma,
erythrocyte and platelet magnesium levels were assayed in 10 IDDM
patients
and compared with 10 healthy controls.
Not only
were the readings lower in diabetics compared with controls, but
the platelet
magnesium values were lower still in those diabetics with
microalbuminuria and clinical proteinuria, compared to those
diabetics
without
these urine findings.
Moreover,
in the patients with these urine findings, there was a negative
correlation
between magnesium status (plasma and platelet) and diabetic
control,
as determined by HbA1c levels.
Table:
Clinical status vs platelet magnesium
in IDDM subjects
Finding Present?
Yes No
------------------------------------------------------------------------
Platelet magnesium (mumol/100 million cells)
Microalbuminuria 1.86 2.34 p < 0.05
(0.47) (0.46)
Clinical proteinuria 1.52 2.34 p < 0.01
(0.19) (0.46)
--------------------------------------------------------------------------
Ref: J
Trace Elem Med Biol 1997 Nov;11(3):154-7
Study #2
The known
diurnal fluctuation in extracellular magnesium levels
is quite
independent of blood glucose levels, according to the results of a
recent
study from Switzerland. Researchers assayed ionized magnesium levels
in healthy
subjects during an oral glucose loading test. There was no
impact
at any stage up to 210 minutes post-glucose load.
Ref: Acta
Diabetologica 1997;34:235-237
Study #3
Correction
of magnesium status in diabetics is difficult, despite
intensive
supplementation. This was the finding of researchers from Belgium
who submitted
eleven magnesium-depleted diabetic patients with stable IDDM
to a 10
week intensive oral and intravenous magnesium supplementation
regimen.
Although
there were significant increases in both ionized and erythrocyte
magnesium
levels, and a magnesium-retention test showed increased storage
levels,
it was not possible to correct magnesium status to consistently
normal
values and there was no change in diabetic control.
Ref: Magnesium
Res 1997;10:135-41
COMMENT
Few clinical
nutritionists would doubt that there is now good evidence that
magnesium
deficiency is common in diabetics (adult and children) and that
it may
well be clinically significant, particularly in pregnant women and
patients
with established cardiovascular complications.
Magnesium
deficiency could be relevant not only because it might cause or
aggravate
complications (e.g. hypertension and ventricular arrhythmias) and
but because
it can increase insulin resistance. On top of that magnesium
also has
a key role in energy metabolism. Thus is might turn out to be
implicated
in the very mechanisms of diabetes itself.
The first
of our three studies is typical of many that lead us to believe
there is
a clinical link worth taking seriously. It ties magnesium
deficiency
at the intracellular level with microvascular complications as
manifested
through the kidney. Whilst the evidence is only circumstantial
(it does
not show cause and effect) it is consistent with other studies
that show
an association between worse diabetic control and magnesium
deficiency.
The second
and third study illustrate why there are still problems with the
magnesium-in-diabetes
scenario.
The Swiss
study highlights the fact that we still do not have a decent,
cheap and
easily performed test which accurately reflects physiological
magnesium
status, and especially one that will sort out for us which
patients
have a clinical problem and which do not.
Part of
the difficulty is that the action of magnesium is to a large extent
intra-cellular.
It is interesting that, in a recently published and related
research
paper (involving some of the same researchers),
it was shown that
circulating
ionized magnesium levels could be reduced in adults with NIDDM,
despite
the fact that those subjects had normal total plasma magnesium
readings
(Miner Electrolyte Metab 1997;23:121-4).
Since insulin
enhances cellular magnesium uptake, the researchers in the
current
study thought that carbohydrate intake might directly affect serum
magnesium
measurements by driving the magnesium into cells. Their results
suggest
that this is not a factor, but highlights the need for clinicians
to take
diurnal variation into account when interpreting a patient's serum
magnesium
values.
The third
study reminds us that, despite a
couple of decades of interest
in the
role of magnesium in diabetes, we are still waiting for convincing,
randomised,
controlled clinical trials that demonstrate firstly that it is
possible
to make long term correction of magnesium deficiency with
supplementation,
and secondly that there are clinical benefits to doing so.
Despite
what they refer to as "intense" efforts at supplementation,
the
authors
here were not able to convince themselves that they had corrected
the magnesium
deficit in their diabetic patients.
Hence,
one wonders how much further we have come since 1992, when the
American
Diabetes Association in its Consensus Statement concluded that
"the
benefit of treatment is uncertain" and that "only diabetic
patients at
high risk
of hypomagnesemia should have total serum magnesium assessed, and such
levels should be repleted only if hypomagnesemia can be demonstrated."
Return to index
Issue # 44: Diabetes and fish oil
Study #1
Fish oil
supplementation offers potential for a profound
triglyceride-lowering
effect, particularly in NIDDM patients, without any
adverse
effects on diabetic control. This is the conclusion of a new
meta-analysis
study published this month.
The authors
combined the results of 26 intervention studies using fish oil
supplementation
in both IDDM and NIDDM subjects, using a weighted (by study
numbers)
linear regression analysis. Of these, 10 were blinded or
crossover,
the rest before-and-after design. In most studies the dose of
fish oil
used provided about 1.8g of EPA and 1.2g of DHA per day.
The table
below shows the combined change in various clinical parameters
due to
intervention. There was some evidence of a dose-response effect.
According
to the authors, this meta-analysis suggests that 3 g/day of
omega-3
fatty acids would be a safe and effective dosage for lowering
triglyceride
levels in NIDDM. The observed elevations in LDL cholesterol
levels
were so small that they may not have warranted treatment, but if
treatment
had been indicated, they commented that a statin drug could have
be given.
Table:
Meta-analysis of 26 trials on fish oil supplementation in diabetic
subjects
(Quantative results are p<0.05 and in mmol/L)
NIDDM IDDM Combined
-------------------------------
Fasting NS
-1.86 NS
glucose
HBA1c NS
NS NS
Triglyceride -0.81 -0.29 -0.60
Tot
cholest NS +0.19 NS
LDL chol 0.2 NS +0.18
HDL chol NS +.08 NS
Ref: Diabetes Care 1998;21(4):494-501
Study #2
Adding
extra fiber can potentially increase the beneficial impact
of fish
oil supplementation
on lipid levels. This includes lowering of LDL
cholesterol.
American
researchers added a 15gm/day pectin supplement midway through an eight
week trial of fish oil in 11 non-insulin dependent diabetics.
The effect,
compared with fish oil alone, was a further drop in both total
(p<
0.001) and LDL cholesterol (p< 0.05), as well as more than 40% additional
drop in triglyceride levels.
Ref: Am J Clin Nutr, 1997 Nov, 66:5, 1183-7
COMMENT
Based on
theory alone, the range of possible biological actions of fish oil
that could
potentially affect a diabetic patient is so great that it is
hard to
predict what the overall clinical outcome of supplementation might
be.
Those potential
effects include:
- changes
in lipid levels (both cholesterol and triglycerides)
- lowered
blood pressure
- decreased
platelet aggregation and therefore affects on development of
microvascular
disease
- alterations
in vitamin levels , including vitamin E
- effects
on ecosanoid and other prostaglandin metabolism.
The meta-analysis
paper is useful in clarifying what the actual outcome is
from giving
fish oil to diabetic patients in relation to two important
clinical
parameters - diabetic control and lipid levels.
The studies
in the meta-analysis are not without limitation. The total
number
of subjects in all of them is just over 400, and the authors rightly
point out
the limitations of methodology in many of the trials.
However,
even so the combined weight of the data does provide some
reassurance
that significant benefits might be expected in relation to
triglyceride
levels, and that this will occur with no compromise to
diabetic
control.
At the
same time it suggests that an eye should be kept on LDL levels.
Although
the patient sample in the second study was different
to those in
the meta-analysis,
the second study does add information about the benefits
of a two
pronged approach to supplementing diabetics with fish oil. Their
results
suggest that giving additional fiber may not only increase the
triglcyeride
reduction, but also counter any possible adverse effect on LDL
levels.
Return to index
