UNDERSTANDING “RECOMMENDED RANGE” vs “REFERENCE RANGE” ON A BLOOD TEST
A recommended range is decided upon by medical experts and is considered a relatively reassuring range. Examples include LDL cholesterol (“bad cholesterol”) and Prostate-specific Antigen (PSA). Experts have decided, based upon available evidence, that an LDL level less than 100 is associated with a lower risk of cardiovascular disease (CVD) than LDL greater than 100. Keep in mind that this is just a line in the sand, and the risk of CVD increases relatively linearly as LDL increases. It is not like 99 is good and 101 is bad. In terms of PSA, a level less than 4 confers a lower risk of prostate cancer than a level greater than 4. As with LDL, risk increases as PSA rises. It always makes me chuckle when a PSA of 3.99 was called “good” and 4.01 is a “you better see a urologist STAT” level. These numbers are ballpark suggestions – you notice that they are nice round, whole numbers. You don’t see recommended values such as 97.3, you see 100.
Also, recommended values can change as more data regarding risk becomes available. When I was a resident, the upper range of recommended LDL was actually 120. Since then, experts have decided that less than 100 is recommended. These numbers are “made up” so to speak, and change as evidence changes. Similarly, when I was a resident we considered blood pressure of 140/90 to be the upper limit of recommended range, and now most experts recommend less than 130/80. But again you don’t see 136/78, you see rounded to nearest 10. It shows you that they are just a bit arbitrary and should be looked at as just part of the whole clinical picture.
Along these lines, since these numbers are essentially cutoffs suggested on a population level, they must be considered in the context of individual risk factors. For instance, if you are a diabetic (another risk factor for cardiovascular disease), or have a strong family history of heart attacks, your doctor would likely suggest an LDL cholesterol of much lower than 100, maybe less than 60. Similarly, if you are a young black male with a strong family history of prostate cancer, a PSA of 2.5 would concern me.
A reference range is entirely different, and is not the same as a normal or even a recommended range. Reference ranges are usually quite broad. There can be a 5 or 6-fold difference between the low end and high end of a reference range. Why are they so broad? The answer is all around you. Look at people in your gender and age range. The vary broadly in state of health. And a reference range represents a population sample. Take for instance, the free testosterone reference range for a 40-year-old male is 4.3- 24.0 ng/dL. This is a huge range. To understand how that is ascertained, a knowledge of simple statistics is required. When a large sample of a population is tested, the results take the shape of a bell curve (see below). The highest proportion of results are at the mean (peak of the curve). Two standard deviations above and below the mean are considered the upper and lower limits of the reference range. By definition, 95% of people sampled lie between the upper and lower boundaries. Also by definition, 2.5% are blow the range and 2.5% are above. So if you (incorrectly) use a reference range as a proxy for a “normal” range, a 40-year-old male has to be below 4.0 ng/dL, and in the bottom 2.5% of his cohort, to be considered low. This is silly for a number of reasons. First, there is no way I would tell a man in the 4th percentile, who feels multiple symptoms of declining testosterone, that he is not a reasonable candidate for supplementation. That is treating a number and not an individual human. In fact, he is very likely to feel better, perform better, and be healthier somewhere higher in the bell curve.
But the other important caveat of a reference range is that it shifts as society becomes less healthy. Because men have become more overweight, more subject to poor diets and lifestyles, and probably under more stress with time–all of which have decreased state of health–average age-adjusted testosterone levels have decreased significantly in the last 50 years (In addition, average sperm counts have decreased and average estrogen levels have increased). In fact, the entire bell curve has shifted lower. So a 40-year-old with free testosterone at the middle of the 2023 reference range has a level that would have been in the lower quartile in 1980. This is another reason that it is silly to be so inflexible about using a reference range to tell a patient that he is “normal” when he presents with symptoms of low testosterone and levels within the reference range.
In general, we should always treat the patient rather than the numbers. Whether we are using recommended ranges or reference ranges, we need to look at the whole clinical picture, patient priorities, risks & benefits of treatments in conjunction with the actual numbers.