Testosterone Testing in Saliva

Mark Newman,Vice President of ZRT Laboratory Operations

There have been a few negative publications on testing testosterone in saliva lately. The North American Menopause Society (NAMS) published a paper in which testosterone was tested in serum and saliva. The conclusion was that “salivary testosterone does not correlate strongly with…measurements from the serum samples.” Interestingly, it was also NAMS (North American Menopause Society) who concluded that for salivary estradiol, as compared to serum measurements, “the correlation … is nearly zero.”

Here’s what we know about testosterone and estradiol in saliva:

Related Articles of Interest

• In some papers saliva and serum correlate. In some, they don’t quite so well. Are we to believe that at one point in time there is an agreement between saliva and serum, and at other times there is not? Perhaps it has to do with the phases of the moon or which political party is in office during the study. Perhaps the same cosmic force that keeps the Cubs losing makes some researchers studies turn out differently.

• There are two reasons why saliva wouldn’t correlate with serum.

1. Topical Supplementation (this is a long story that we’ll get into some other day), but suffice to say that serum testing severely underestimates for creams.

2. The test isn’t accurate. This is very significant for testosterone and estradiol in saliva (and even in serum). Some assays are sensitive enough to measure high level samples, but not at low concentrations. If you use a test that isn’t good enough to measure low-level samples and then conclude that saliva and serum don’t match, have you really accomplished anything? Consider the following for estradiol:

Observations:
• Results are accurate
• There is a 2-3 fold difference between days 1-5 and days 19-21… there should be
• Serum Correlation was good!
• Conclusion: Saliva Testing Works

Accurate Work = Good Saliva Testing

Observations:
• Results at day 1 are 5 X higher than above (inaccurate)
• There is no difference between days 1-5 and days 19-21… there should be
• Serum Correlation was Poor
• Conclusion: Saliva Testing Does NOT Work

Inaccurate Work = Bad Saliva Testing

There are two sets of results in the literature. These both propose to show women’s estradiol (estrogen) levels throughout their cycle. I recreated the data here.

Wong used a method that is FAR more sensitive than Chatterton (I’ll spare you the details on the difference between a 3H extracted RIA and a ‘direct’ 125I RIA but Wong’s method is far more sensitive).

Chatterton got results that are 3-5 times higher than Wong. Wong says salivary estradiol correlates to serum (r>0.9) and Chatterton says it doesn’t (negative r-value). Are we to believe that in one group of women there was terrific agreement between saliva and serum levels and not in the other? No, reality is Chatterton’s test has so much background “noise” (that’s why his levels are so much higher). Remember, the method has to be accurate and reproducible before anything else matters and for testosterone and estradiol, not all methods are good enough…some are really poor.

Let’s look at this one more way. Chatterton’s numbers have a much higher background level that obscures the specific estradiol binding sites in the test and raises the “apparent” estradiol concentration throughout the menstrual cycle. This causes the background to increase about 5-fold. To really compare them, let’s just divide by 5 (20% of original values) to make them “the same” and see what that does.

Observations:

• Bloating the numbers by a factor of 5 obscures the critical differences between the three phases of the menstrual cycle.
• Only the accurate test clearly distinguishes the ‘low,’ (day1-8) ‘normal,’ (21-23) and ‘high’ (day 14, ovulation) phases of the cycle.

Wong’s results look exactly as what you would see in a text book or from serum results. It’s subtle, but the lack of distinction between days 1-5 and days 21-23 in the other test means you can’t tell the difference between “low” and “normal.” You can see why there is no serum correlation and no clinically useful information using the inaccurate test. You can also see why a group like NAMS might pick the inaccurate method to conclude that saliva testing “doesn’t work” if it is politically expedient for them to reach such a conclusion.

Whether you are talking about estradiol or testosterone, study conclusions need to always be taken with a grain of salt. It would be better if studies like Chatterton’s were never done. Instead we have an inaccurate method reaching negative conclusions about saliva testing.

Fortunately for ZRT Laboratory, we have our own data and we can correlate our actual testing to clinical parameters. My last blog on testosterone showed lots of great data which should increase confidence in using such a test. Here is a little bit of data for estradiol to show its clinical utility. Remember these figures are from actual patient data tested over the past 10 years. This is a lot of information, but just scroll down through it, read the bullet points, and see if it isn’t impressive clinical utility for a test that some would have you believe is not relevant.

• Here we see results from nearly 50,000 women, and we can see the expected relationship between estradiol levels and age.

• We can see a nice correlation between levels of estradiol and the severity (self-reported) of their hot flashes and also vaginal dryness.

• We know that a higher BMI means more adipose tissue. More adipose tissue means more aromatase, which will convert testosterone to estradiol. We can see this relationship strongly from levels measured in premenopausal women. The graph looks much the same (except lower levels of course) with postmenopausal women.

• One criticism of testing is that there is no relationship between the hormones taken and the clinical levels in the patients. This strongly refutes that position. First, with oral estradiol…

Even more impressive data with topical hormone. Not only do we see very nice dose-dependent results, but we also know how severe these women's hot flashes were (we collect that information on our requisition forms). You can see two hugely important things.
1. Dosages of 0.1 mg get most women up and into the premenopausal range, and
2. Low dosages do a very good job of relieving symptoms.

If you are using 1-4 mg of topical estradiol, this begs the question, why? The answer to that will have to wait for next time. In short, topically measured hormones do not increase serum levels. This has lead to many physicians who are monitoring topical hormones to overdosing.

The above data shows tremendous clinical utility for salivary estradiol. ZRT’s assay for estradiol is uniquely sensitive, which is critically important for this test. Don’t assume that this type of data would fall out of another lab’s estradiol test. It only will if it is accurate. Many of the tests available are not optimally accurate, so if you use a different lab, ask for this type of data to instill confidence in using the test. As Chatterton showed, just because samples are measured and numbers are given doesn’t necessarily mean that the numbers have a high degree of clinical relevance.