How “Set” are Biomarker Reference Ranges?
Laura Boucai, MD
Laura Boucai, MD
During my clinical training as a fellow in the Division of Endocrinology, I noticed that we were seeing an unusual number of older white women diagnosed with high thyroid-stimulating hormone (TSH) levels and receiving thyroid hormone replacement therapy. My mentor Dr. Martin Surks and I began to wonder whether these women actually needed this hormone, or whether they were being overtreated. Was it possible that the reference range for these women was different from the “normal” reference range defined by the laboratory? As we further explored our hypothesis, we found exciting results: TSH reference ranges differ according to a person’s age, race, and sex.
One Range Does Not Fit All
Thyroid hormones usually get reported with minimum and maximum reference ranges, outside of which people tend to think they are abnormally high or low. Most primary care doctors will treat patients whose levels fall outside this reference range. Dr. Surks and I looked at huge numbers of people, both within Montefiore Medical Center and at a national level, through the National Health Nutrition and Examination Survey (NHANES database). We found that the reference range limits changed with age and race. For example, I might have a TSH reference range in the 0.3 to 4.5 range, but an 80-year-old white person might have a reference range completely different than mine. If we as physicians don’t pay attention to this, we might misclassify people as having abnormal thyroid hormones just because they don’t fall within the laboratory reference range, a range applied for everyone in the community.
TSH is a hormone made in the pituitary gland that stimulates the thyroid gland. TSH and thyroid hormones have an inverse relationship: the lower our thyroid hormone levels are, the higher our TSH levels are. Our TSH reference range increases as we age—we have proven this in people from the age of ten to the age of 90 or 100. Additionally, TSH reference ranges are higher in white people as compared to black people. This has been established not only at Einstein/Montefiore but also at a national level, and in other populations.
We often see patients in their 60s, 70s, and 80s with reports from their doctors stating that they have minimally elevated TSH levels. This means that their thyroid levels are low and they probably need medication for thyroid hormone replacement therapy. They come to us expecting to get the medication, but when we look at these patients more closely, sometimes we end up deciding that they are actually normal, without any thyroid condition, and in fact fall within their own age- and race-specific reference ranges, and that they may not need any thyroid medication or thyroid supplementation therapy.
Using a large, enhanced database with a number of variables (from thyroid hormones and family history of thyroid diseases to urinary iodine concentration, an excellent predictor of how much iodine we have in our diet), Dr. Surks and I created equations that predicted peoples’ reference range based on their age, sex, weight, race, and urinary iodine levels. We found that their urinary iodine and weight did not change things very much, but their age, sex, and race were significant predictors of changing this reference range. Our recent paper, published in Thyroid, supported the methodology that these predictors be implemented in the hospital’s patient record system so that we wouldn’t just get a regular value, like Quest or another diagnostic laboratory would provide, with a set reference range for the entire population. Instead, each patient’s sex, age, and ethnicity would be input into the system when they have blood drawn, which would then generate their own specific reference range, with a much greater accuracy than what we get right now.
To Treat or Not To Treat
Some studies show that when physicians prescribe thyroid hormone for the first time, we end up overtreating 40% of our patients. Overtreatment of thyroid hormone can manifest as palpitations which, over a long period of time, can cause congestive heart failure. Overtreatment also causes osteoporosis, which could damage our bone density. We don’t want to overtreat or undertreat; we try to find the balance. Interestingly, studies have found that 27-50% of people who fall outside the reference range but are not treated for it will return to the normal range again within five years of diagnosis.
The possibility of a link between minimally abnormal thyroid hormones and the risk of cardiovascular disease (CVD) is currently under debate. Some studies state that minimally abnormal thyroid hormone levels increase one’s risk of developing CVD, while others say that these same numbers might be protective against CVD. Despite numerous large, lengthy studies on this topic, the literature cannot come up with a final answer. As a result, we as physicians don’t know whether we should treat people with minimally abnormal levels of thyroid hormone, or whether if by treating them we are shortening their life expectancy. I wanted to find out whether there was a more sophisticated, scientifically rigorous way of looking at this issue. The work that Dr. Surks and I have accomplished so far provides a way to begin looking at things from a different angle.
Our methodology is applicable not only to thyroid diseases, but to many biomarkers that physicians use regularly. We seldom question the “normality” of the reference ranges provided by Quest or other diagnostic laboratories (for example, electrolytes, creatinin, or BMI), yet these biomarkers might be different for a population of young people than for a population of older people, or for whites compared to blacks, because these populations are genetically different.
The reason why TSH reference ranges change between different populations remains to be determined. Are there age-related changes in the negative feedback or the TSH glycosylation? Are these TSH changes a consequence of other medications being used? Is there indeed thyroid hormone deficiency as we age? And ultimately are these changes acquired or hereditary? In this regard, Dr. Nir Barzilai’s Longevity Genes Project has found that centenarians’ blood samples showed higher TSH levels compared to controls. We found this striking: that these higher levels might be protective and confer a longevity benefit. Two single nucleotide polymorphisms detected by Barzilai and Atzmon—genetic variations in the thyroid-stimulating hormone receptor—introduced a genetic pathway that might explain the laboratory abnormalities we often see. As our studies evolve, we would like to investigate the possibilities of a genetic explanation for these hormonal changes.
Laura Boucai is Assistant Professor of Medicine in the Division of Endocrinology.