drOPsite Login                                                                 Healthcare Professional Registration

You are here

What Saliva Testing Can (And Can’t) Tell Us When Hormone Creams Are Used

Written by Dr. George Gillson


For several decades, researchers, practitioners and patients have all been puzzling over the behavior of salivary hormone levels measured following application of hormone-containing skin creams and gels. When an individual is making her or his own hormones, the levels of these hormones measured in saliva correlate quite well to the levels of these same hormones measured concurrently in blood. Those numbers also correlate well to the clinical picture, in many cases. This happy arrangement does not apply for hormone levels measured after an individual has applied skin cream containing one or more hormones such as progesterone, estradiol, estriol, testosterone or DHEA. The numbers are much higher and may not correlate to the clinical picture. These levels do not have the same meaning as the endogenous hormone levels.
The Tale of Daphne and Her Progesterone
Consider Daphne, a woman having regular 28-day cycles, whom you instruct to take a saliva sample on the morning of Day 20 of her cycle. This is in the middle of the luteal phase, when women make a lot of progesterone. The level of progesterone measured in that sample comes in at 50 pg/mL.That’s low for a mid-luteal sample. A good number would be more like 150 pg/mL. Based on this result, and her estrogen-dominant symptoms, you reckon maybe she needs some progesterone so you ask her to start using 15 to 20 mg of progesterone in the form of a cream rubbed into her skin before bed each night, starting about Day 10 each cycle and stopping around Day 28, when her next menses starts.
Two months later, you take another saliva sample from Daphne again on the morning of Day 20. You are astonished and somewhat mortified to find that her progesterone level is now 3000 pg/ml, 60 times higher than what you saw when she wasn’t using hormones.
You have an impulse to call Daphne and tell her to stop using her progesterone cream immediately; you’re concerned that you have been subjecting her to a gigantic overdose of progesterone. But two things stop you.
First of all, you recall that 15 or 20 mg of progesterone is NOT a gigantic overdose; John Lee MD used 25 to 30 mg of progesterone in his thousands of patients for decades, without incident. You also recall that the normal daily production of progesterone is about 25 mg/day. You gave her less than that, and you also figured she needed it, based on her low saliva progesterone level last month. So you heave a sigh of relief as you tell yourself that you didn’t overdose her by giving her 15 or 20 mg/day. You were right.
The second thing that stops you from calling Daphne is that right after you saw the saliva hormone report, one of your staff came in and said that Daphne had called yesterday and left a message saying that she has been feeling “Great” since you asked her to start using that progesterone cream. This thought fleets through your mind: if she is overdosed, why would she say that she is feeling “great”?
Being the astute, yet cautious clinician that you are, you decide to hold off asking Daphne to stop the cream, but you arrange for her to have a progesterone level measured in her blood AND saliva the following month, again on Day 20. That saliva result comes back pretty much the same as the other one: 2800 pg/mL; however, the serum result is in the low end of the normal range for the mid-luteal part of the cycle.
You silently ask yourself this question: If the progesterone number has gone way up in saliva but the blood level is still low, how did that progesterone get into the saliva? You also wonder internally why the patient can feel “Great” when her serum progesterone level is low. Finally you clap yourself on the forehead and say out loud, to nobody in particular: “What the heck is going on here?”
The good news is that you’re not alone in your puzzlement; I too have been puzzling over this conundrum for almost twenty years-minus the forehead-clapping.

My Take On What Is Going On

What I think is going on is as follows, although it hasn’t been proven yet. Just remember that if it turns out that this is what’s going on, you read it here first.
Hormones applied to the skin are probably able to travel around the body OUTSIDE the bloodstream, by diffusing away from the application site through subcutaneous fat. This process is also assisted by shallow lymphatics, which can disperse particles and solutes laterally (over the body surface) away from the application site as opposed to carrying them straight back into the venous circulation, thence to the heart and back out to body tissues. Some of the hormone travelling in fat “under the radar” most likely fetches up in the vicinity of the saliva glands, many of which are not encapsulated. This means they are in close contact with, and can “read” surrounding tissues. This may be how saliva can be reflective of hormone stored in tissue following topical hormone usage, and also how hormones can get around the body without appearing in the blood.
I did experiments on myself years ago, applying the same dose of testosterone gel to a different body part once every few weeks, and then spitting in a tube for the next 24 hours after each application. I saw that the lowest salivary testosterone levels were seen when the hormone was appliedfarthest from the head; the highest level was seen when I applied the hormone near my collarbone. Levels from application sites in between those two extremes fell on a smooth curve.
This data supports the concept that hormones might disperse laterally away from the site of application, leaving less chance for the testosterone to get “lost along the way” when it is applied closer to the head. (That’s one interpretation; there might be others.)
Around that time, I also constructed plots of the severity of various hormone-related symptoms as a function of the high salivary hormone levels measured in women using topical testosterone and estradiol. These high levels were not correlated with symptoms of excess testosterone and estradiol; they were not clinically-relevant numbers. Note that each data point in the graphs below represents dozens to hundreds of patients.
So in my hypothesis, when you rub hormones on your skin (follow me closely here) saliva reflects hormone that is stored in fat as well as what is in blood. That’s why Daphne still feels great with a saliva progesterone number that is way outside the luteal range which applies when progesterone is homegrown. Only a fraction of that higher number (3000 pg/mL) is clinically active. The rest represents hormone stored in fat. It would stand to reason that the fat-stored hormone levels shouldn’t correlate to symptoms; they can’t because the hormones are in storage.
The main line of evidence supporting the notion that saliva testing can read fat-stored hormone after their topical application is that we still see elevated levels in people who were using topical hormones well after they stop using the hormone creams: this can be months afterward in the case of progesterone. This “hangover” level gradually diminishes with time. The simplest explanation for this (invoking Occam’s Razor) is that the test is reading stored hormone. This storage effect is most pronounced for progesterone, which also happens to be the most fat soluble of the commonly-supplemented hormones. We also know from research on pigs, that sex hormones shuttle into and out of fat stores when the animals make their own hormones. Body fat is regarded by many as an accessory human endocrine organ when it comes to steroid hormones.

Clinical Implication

The ranges we list for topically-supplemented hormones are not clinically-validated ranges.  Adjusting the hormone dose to get your patient’s level to the middle of one of these ranges does NOT ensure that your patient is on the “right” dose. You have to be guided by the patient’s symptoms.
You absolutely CANNOT use the ratio of topically-supplemented progesterone to topically-supplemented estradiol to say anything about whether the endometrium is being protected from excessive estrogen effects. That requires endometrial biopsy or transvaginal ultrasound.
You CAN use the numbers as a rough guide when things are not working.  In the example I gave, we saw a level of 3000 pg/mL for progesterone.  Remember that I think this number represents stored hormone, for the most part.
If the patient never saw any benefit and her progesterone result is 100 pg/mL when you’re expecting 3000 p/mL, then either there is something wrong with the cream (formulation too weak/formulation not stable/wrong carrier), or the patient is not using it as prescribed (patient has stopped using it, uses it infrequently, or was using too little). 
If the patient was doing well but has stopped responding, and your number is 50,000 pg/mL when you’re expecting 3000 pg/mL, then either there is something wrong with the cream (formulation too strong), the patient is not using it as prescribed (the patient uses too much or uses more frequently than prescribed) or the patient accumulates an excessive amount of hormone in their tissue on a modest dose/appropriate dosing schedule.  The high number may indicate that the patient has overloaded her system with progesterone and the receptors have stopped “listening.”
There’s simply a limited amount of information you can garner from saliva hormone levels measured when the patient is actively using hormones. You also have to be wary of testing too soon after discontinuation of topical hormones, if you seek to see what the patient’s endogenous (homegrown) production is. A typical scenario would be when you inherit a patient who was put on topical hormones by someone else.  That patient presents to you because the therapy isn’t working, and you want to start over again from scratch.
How long you need to wait depends on many factors including the dose, site of application, the type of cream the hormones were formulated in, and the hormone(s) in question.  It can be anywhere from a month to six months.
I have yet to see a month-long hormone profile from someone supplementing with topical hormones, which has yielded any useful information on estradiol and progesterone, beyond what you could get from doing a properly-chosen point sample, or spitting into the same tube on several mornings and submitting this pooled sample.  This is why I don’t recommend doing a Month-Long Hormone Profile in someone using topical hormones. I believe that the foregoing discussion has also made it clear why it is problematic to know when to test after a patient has stopped using topical hormones.
It’s interesting to note though, that in many cases, the luteal progesterone surge is still apparent in women using progesterone cream. It is simply amplified, as in the graph below, for a woman currently using topical progesterone. 
This suggests that the same signals which talk to the ovary are also talking to the sites where progesterone is stored (which could include the ovary). Progesterone cream use may turn fat into an “accessory ovary” or may help to “load” the ovaries with progesterone. In the meantime, here’s hoping other hormone researchers take this theory on and add more pieces to the puzzle.   
It seems that the forehead-clapping isn’t going to stop any time soon…