How Is Bone Density Measured?
Bone mineral density is actually measured, based on how much of the x-ray beam is slowed while going through the bone, in units of grams per cubic centimeter of bone. This would be pretty meaningless to most of us, so it is converted into both a "T-score" and a "Z-score" (which are meaningless to all of us! ).
The "T-score" basically compares your density to that of an "average pre-menopausal woman". The "Z-score" compares you to women in your age, weight, and ethnic group. The "T-score" relates to your absolute bone density and how much you probably have lost over time, and therefore is related to your risk of having a fracture. The "Z-score", relating you to other women of similar risk, helps us to determine if you are exceptionally low in bone density - in other words, should we be looking for an underlying disease or risk factor. Typically we look for thyroid disease or elevated levels of parathyroid hormone. Certain medications, smoking, etc, can all increase your bone loss. In this page I will explian what these scores mean and relate them to percentiles, which generally we understand a little better. I guess we have to blame the statisticians for reporting these scores in terms of statistics instead of percentiles....
The Standard Deviation
This has nothing to do with what Bill and Monica did, but is a statistical term. Events and measured values which occur naturally tend to cluster around an "average value", or "mean". As an example, if we measured blood sugar in 150 people and charted the numbers of people within each range of values we obtained (a frequency distribution), we might have a chart which looked like:
As you can see, the values cluster around a central mean value. You can imagine that if we next measured 10,000 people and broke the data into smaller divisions, the chart of this "frequency distribution" would start looking more and more like what is called a "bell-shaped curve":
Statistics allows us to describe this type of frequency distribution for any group of such data by using only two measurements - the mean and a term which indicates how widely spread the data are - this term is called the "standard deviation". A mean of 100 with a standard deviation of five would be presented as a very high, narrow curve, showing most people having values near 100, but one with a standard deviation of 25 would be broad and flat, with many people having values far away from 100. Two curves with different spreads, or standard deviations are represented above.
In this type of curve there is a straightforward relationship between where a value is placed, in terms of standard deviation, and its relationship to the other values. As shown below, 68% of values lie between the mean +/- 1SD (that is plus-or-minus one standard deviation), as shown below:
Your DEXA scan reports your "T-Score", which is the number of standard deviations from the mean premenopausal value where your bone density lies. For example, as shown above, if your "T-Score" was "0", there would be half of the premenopausal population with greater bone density than you, and half with less. If your score was "-1", there would be 84% with greater bone density (the bottom half of the 68% between +/- 1 SD from the mean, plus the 50% above the mean). Fortunately, there are tables which allow an easy conversion of standard deviation to percentile. I have reproduced part of one below:
This table allows you to compare your "T-Score" or "Z-Score" to
an equivalent percentile. A "Z-Score" of -1.8, for example, means that
you are at the "4th percentile" - 96% of women in your category have greater
bone density than you do. A "T-Score" of -2.0 would mean you are at the
2nd percentile - that 98% of pre-menopausal women have greater bone
density than you do. The World Health Organization considers a "T-Score" of
-2.5 or below to represent osteoporosis. Many U.S. organizations consider
-2.0 or below to represent osteoporosis. |
|||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||