On occasion you may read a fitness article in the newspaper. It perhaps
discusses body fat, how to measure it, and normative data. It might say that a 22-year-old
female is typically 25% body fat and should ideally measure 17%. The article also advises
that her body fat is too low if allowed to fall below 13% lest her menstrual cycle shut
down. (For the sake of discussion, I am inventing these numbers — just like a real
exercise physiologist. You will appreciate this more as you continue reading.) Where did
this normative data come from? How was it collected? What tools were used?
There are several different types of body fat assessment. They include isotope dilution,
photon absorptiometry, potassium-40, hydrostatic weighing, radiography, ultrasound,
nuclear magnetic resonance, total body electrical conductivity, bioelectrical impedance
analysis, infrared, and skinfold thickness. Supposedly, the most reliable of the practical
methods is hydrostatic weighing (also underwater weighing). Hydrostatic weighing is
typically mentioned with reverence by exercise physiologists as though it is the test par
excellence.
With hydrostatic weighing, the Archimedean principle of density (mass per volume) is
applied. But hydrostatic weighing has its inconveniences. Needed for this is a tank just
large enough to completely submerge a subject. Optional is a seat connected to a scale, so
that the body can be weighed while submerged.
The simpler of two methods involves two readings: dry weight and volume displacement (as
determined by water-level rise on the sides of the tank.) However, such
displacement-reading accuracy requires a tight tub. A subject displaces the same fluid if
he submerges into a swimming pool, but the rise of the water level on the walls of the
relatively-large pool is difficult to accurately measure. Therefore, a small cylinder is
required that permits maximum rise in water level for a given volume displacement. (In
1981, Ken Cooper reported a near drowning of a morbidly-obese man who became stuck in the
underwater-weighing tank at The Aerobics Center in Dallas. Such a vessel was too-perfectly
tight for his girth.)
The second method establishes volume as a ratio between the dry weight and the submerged
weight of the subject.
Other nuisances are associated with hydrostatic weighing. The subject requires an enema
just prior to the test to completely evacuate the alimentary canal. A 24-hour fast is also
required.
Maximum expellation of the lungs is required just immediately before submersion. The
remaining residual lung volume then must be estimated along with residual intestinal gas.
(Imagine the fear some subjects possess when being told to exhale as much air as possible
before submersion — though they have been taught all their lives to inhale before
going under water.)
In addition, hydrostatic weighing is messy. It requires a room that is a designated wet
area and strong technicians on the ready to instruct the subject, read the data, and be on
the alert to provide CPR.
Hydrostatic weighing is a questionable estimate of density. It is not a measurement of
body fat. Body fat is then estimated from the density estimate. To reiterate, body fat is
not a measurement. It is an estimate.
A tremendous assumption occurs when estimating body fat from density. Although a greater
density generally should indicate a greater muscle-to-body-volume ratio and a smaller
fat-to-body-volume ratio, it may instead reflect a greater bone-to-volume ratio (or some
other tissue-to-volume ratio). Observant SCUBA instructors report that the weight-belt
required to offset buoyancy (density deficiency, so to speak) appears to vary between
subjects with less respect to body fat than with respect to bone size (or bone density).
Additionally, exercise physiologists often treat their density/body-fat determinations as
though they are measurements. Actually these determinations are statistical estimates
based on statistical bell curves (averages) representing a variable range of values
collected from a group of individuals. These readings and calculations are not direct
measurements of an individual's body fat. They are merely comparisons to a norm. And this
statement further assumes the validity of the density estimate. (Note how easy this is to
do. I must constantly correct myself from the error of stating such readings as a
measurement.)
This density/body-fat norm would not be required if there were only two kinds of tissue in
the human body: fat and muscle. But there are many tissues in the body and each possesses
a unique density. The most influential determinant of density is perhaps the bone.
Ultimately, direct or indirect density determinations are correlated with the
density/body-fat ratios of cadavers. Cadaver studies are the fundamental reference for all
body composition determination methods.
The discussion of body fat determination by skinfold illustrates the error incurred in
cadaverous dissection. The use of skinfold calipers is the most popular and widespread
method to determine body composition. Skinfold is usually deemed inferior to hydrostatic
weighing; however, I underscore that it avoids the indiscriminate inclusion of all body
tissues under a common [density] value. After all, the calipers are more, though not
exclusively, directly applied to the subcutaneous folds of fat. Cumulative skinfold sums
are then plotted on a nomogram to reveal the subject's percentage of body fat per his age,
race, and sex. Again the nomogram is fundamentally referenced with cadaver studies.
To illustrate how outrageously inaccurate cadaverous reference is, pretend that I seek to
construct a new nomogram:
Pretend that I am an eminent exercise physiologist who seeks a new nomogram using three skinfold sites. I assemble a team of my graduate students to collect data. I decide that we require a statistical bell curve representing 20 subjects for each 5-year-age range beginning with 10-year-olds. The ranges will run 10-14, 15-19, 20-24 and so on to age 74 (13 ranges). A bell curve is require for each sex/age-range of each major race. If I include both sexes in each of the major races — Caucasoid, Negroid, Mongoloid — I require:
3 races x 2 sexes x 13 age ranges x 20 subjects per bell curve = 1560 subjects.
My graduate students have used skinfold calipers before in their undergraduate courses, but their skill is not proficient. And although I require a tremendous number of subjects, I know that I may find few if any subjects for many of the 78 bell curves.
I ignore these problems [After all, I can account for the errors in my statistical analysis later — HA!] and go to the dean of the anatomy department in the university medical school. Only there can I obtain the cadavers I require. I beg his permission to allow my team of incompetent graduate students to collect data from the department's cadavers before the year's dissection begins on them. Of course, I am forced to use the body weight on each death certificate. And of course, my graduates will have to record skinfolds on dead human flesh which lacks the elasticity of live flesh.
Then I must ask the dean and his medical students one more favor. Would each student be so kind as to scrape off and collect all the fat separated as he performs shared (two or three medical students share a single cadaver) dissection assignments? I understand that much of the fat will not make it into the plastic bowl I provide, because some fat may be under the hot lamps until it melts to oil and drips off the dissection table onto the floor. And unavoidably some must be discarded with parts into the trash. And much of the intramuscular fat and fat inside the organs is not practically dissected by medical students performing their own assignments under duress. Nor is the intramuscular fat detected by the calipers as skin fold.
I will also have to live with the limitation that the school possesses facilities for only 70 cadavers each year and it will take me over 20 years to collect the data I have only one year to collect. I will just have to make do with what I have.
My students will visit the lab again at the end of the year to collect whatever fat the medical students manage to scrape into the plastic bowls. My grads will then correlate the skinfolds with the body fat collected per body weight. From these data they will plot a bell curve. And collectively the bell curves are normalized via regression equations to present us with a neat and clean-looking nomogram. I believe that most of my students can do this correctly. If not, I'll just have to adjust the statistical standard to make it appear clean.
Now let's pretend that all this gobbledygook is perfectly accurate.
Suppose that you take skinfold measurements of a man and thereby determine his body-fat
percentage by the nomogram to be 20%. Then through strength training he adds five pounds
of muscle (not truly measurable) to his frame. Correspondingly, he add five pounds of body
weight. By this we know that his body-fat percentage has decreased. So we then caliper him
again and get the same skinfold readings and thus the same body-fat percentage as
determined originally by the nomogram. Did we lose body fat? Answer: No. But did the
percentage truly decrease? Answer: Yes. However, the calipers are based on a static,
non-dynamic, dead cadaver that is incapable of growing muscle under his fat. Nor is there
any way for the calipers to detect muscular growth in the living subject.
These problems related to skinfold have plagued our fat loss programs for years. Typically
we subtracted subjects' after body fat from their before body fat to determine total fat
loss. The difference between their total fat loss and their weight loss was, therefore,
their change in muscle. Frequently, it is disconcerting to calculate a muscle loss when a
subject's appearance and performance indicated stronger and larger muscles. Skinfold can
not detect tissue changes other than skin and fat.
Although low-tech, Arthur Jones' body fat monitoring technique is just as valuable as any
of the techniques already mentioned. He merely records the difference between the
circumference of a fully-contracted brachial biceps and its relaxed circumference. If this
difference increases, the subject is becoming leaner. Decreases => fatter. This method
provides no quantification of body-fat percentage. Nor does skinfold or any of the other
methods.
I prefer skinfold calipers. They are much more reliable than the girth measurements Arthur
used. The skinfold measurements, if properly performed, are reasonably accurate. I cannot
obtain a percentage from them nor can anyone else. If the skinfold(s) go down the subject
is leaner. If up, fatter. What else is there to know?
And if you are a woman and you decrease your body fat and your menstrual cycle shuts down,
eat more. But before you do so, get a pregnancy check and evaluate other changes that may
affect you. The idea that some magic number exists beneath which you should not go pending
menstrual shutdown is a fantasy. Your best indicators are your own biological functions.
Magnetic resonance imaging (MRI) and densiometry (photon absorptiometry) offer the
possibility to selectively and volumetricaly quantify soft and hard tissues. These tools
may provide a means to perform direct and accurate research with which to greatly
enhance nomograms to be used with skinfold calipers and other tools more practical for
field application. Nevertheless, any improvements in nomograms are of limited value
because of the muscular volume dynamics not detectable with such indirect means.
Bioelectrical impedance (BIA) has been an exciting alternative for body composition
analysis. I have used some of these tools extensively only to find that they are
unreliable. Part of this is due to their dependence on the same cadaver based data used
for skinfold. Additionally, BIA accuracy is influenced by transcutaneous conductivity,
nearby appliances, electrode placement consistency, body position and shape, body
hydration, and time relationship to exercise. In 1994, the National Institutes of Health
condemned BIA as giving distorted or useless information.
These BIA machines are frequently used in the offices of doctors who are treating obesity,
and yet there is little data to prove such measures are clinically useful. According to
NIH, "There does not appear to be an established role for the technique."
Additionally, at least one company is selling a bathroom scale that provides a body
composition readout by merely standing in place to obtain electrical contacts for BIA.
Just imagine how poor this device might be under high humidity bathroom conditions and wet
feet. Supposedly, humans are gifted animals in that we can profit from our mistakes.
Nevertheless, our great communication skills and technology are not enough to prevent
apparently novel ideas from being manufactured atop previously disproven concepts.
Even more recent has been the use of infrared devices that are applied to one convenient
location on the body and then a total body composition is extrapolated. Human subjects
have greatly divergent and individualistic fat depot configurations. A single measurement
with any device is inappropriate.