A testosterone of 200 and an ApoB of 120: what a chatbot cannot do with those numbers.

A total testosterone of 200 is not a diagnosis, it is a reason to look for one, and the thing worth ruling out is a pituitary tumor. An ApoB of 120 is not a rounding error. Read together they are often one problem, which is the part an AI asked about each number separately will never tell you.

Daniel Tagge, MDJuly 31, 20268 min read

Two numbers that show up together often enough to be a pattern: a total testosterone around 200 and an ApoB around 120. The man holding them is usually in his forties, feels flat, has been gaining weight slowly for a few years, and has done the reasonable modern thing. He bought a panel, pasted the results into a chatbot, and got back a careful, well-organized, confident explanation.

The explanation is mostly correct. That is what makes this worth writing about.

Here is what a testosterone of 200 and an ApoB of 120 actually mean, what the AI reliably gets right, and the specific things it cannot do that change what happens to you.

Start here: 200 is not borderline

Most labs put the bottom of the total testosterone range somewhere around 250 to 300 ng/dL. A 200 is below that floor. This is not the usual story on this site, where a number sits inside the normal range and is still wrong for you. This one is frankly abnormal by the lab's own standard, and it has been for a while.

Which means the important question is not "is this low." It is why.

And this is where the entire thing turns, because a low testosterone is not a diagnosis. It is a finding that has a differential, and one item on that differential is a tumor.

What a chatbot does well here

I want to be fair, because the fairness is the point.

Ask a current model about a total testosterone of 200 and it will tell you the value is below the reference range. It will explain what testosterone does. It will list the symptoms and they will match how you feel. It will usually mention that a single measurement is not enough and that testosterone should be drawn fasting in the morning, which is genuinely good advice that a lot of clinics skip. Push it and it will often name the right next tests.

That is a better answer than many people get in a fifteen-minute visit. I am not going to pretend otherwise.

What it cannot do

It cannot order any of it. And it does not know what it was not shown, which is the failure mode that matters, because the finding that explains your number is almost always on a test nobody ran.

For a total testosterone of 200, here is what I want before anyone writes a prescription:

A repeat, drawn fasting between 8 and 10 in the morning. Testosterone has a daily rhythm and single values are unreliable. Some 200s come back as 340s.
LH and FSH. This is the fork in the road. High LH with low testosterone points to the testicle. Low or inappropriately normal LH with low testosterone points upstream, to the pituitary or hypothalamus, and that is a completely different problem with a completely different workup.
Prolactin. If the picture is a secondary one, a prolactin-secreting pituitary adenoma is on the list. It is not common and it is not rare, it is treatable, it is frequently missed, and it will keep suppressing your testosterone for as long as nobody looks for it.
SHBG, with free or bioavailable testosterone. Total testosterone without SHBG is close to uninterpretable, and SHBG moves with insulin resistance, thyroid status, and alcohol.
Estradiol on a sensitive assay, TSH with a full thyroid read, and ferritin with iron studies, because iron overload is a real and reversible cause of hypogonadism that nobody thinks of.
A conversation about fertility, before anything is prescribed, because testosterone replacement suppresses sperm production and that conversation is much harder to have afterward.

Then the things that are not a lab at all. Sleep apnea. Opioids. Glucocorticoids. Alcohol. Cannabis. Visceral fat. Several of those are reversible, and reversing them can move the number without a prescription at all.

The scenario worth naming

The reason I labor this: the shortest path from a 200 to relief is a testosterone prescription, and that path is available in about four minutes online.

If you take it before the workup, three things happen. Your number goes up and you feel better, which is real. Your own production shuts down, so the diagnosis is now much harder to make. And if the cause was a pituitary adenoma, it is still there, still growing, and now masked by a gel that made you feel well enough to stop asking.

A chatbot cannot check your visual fields. It cannot order the prolactin. It cannot send you for the MRI. It is not able to notice that you sound different on the phone this month.

The other number

An ApoB of 120 is not a rounding error, and depending on the lab it may not even be flagged.

ApoB counts the atherogenic particles in your blood, every one of them, which is what makes it a better predictor than the LDL cholesterol number most people are handed. LDL-C estimates how much cholesterol those particles are carrying. ApoB counts the particles themselves. When the two disagree, and they disagree often in exactly this metabolic picture, the particle count is the one that tracks with what happens to you.

For most people in primary prevention, an ApoB under about 90 mg/dL is the goal, and lower for anyone with real risk. A 120 is meaningfully above that. It usually arrives alongside an LDL-C that looked acceptable, which is why it goes unaddressed for years.

Ask a model about it and you will get a competent explanation of all of that. What you will not get is the next part.

The part almost nobody puts together

Here is the thing a chatbot will not tell you, and it is not the model's fault. It is a consequence of how it was asked.

You asked about the testosterone. Then, in a separate message or a separate paragraph, you asked about the ApoB. You got two good answers to two questions.

But a testosterone of 200 and an ApoB of 120 in the same forty-something man are frequently not two problems. They are one problem showing up in two places.

Visceral fat and insulin resistance sit upstream of both. The visceral compartment raises aromatase activity, drops SHBG, and suppresses the signal from the pituitary to the testicle, which pulls testosterone down. That same insulin resistance drives the liver to overproduce the particles ApoB is counting, which pushes ApoB up. The low testosterone then worsens body composition, which worsens the insulin resistance, which is how a slow trajectory becomes a fast one.

Read separately, the answer is a prescription for each. Read together, there is one driver, and treating the driver moves both numbers at once.

Nothing about that requires me to be smarter than the model. It requires somebody to be looking at the whole person instead of answering the question that was typed.

Which questions a chatbot can actually close

The question a 200 raises	What answers it	Can a chatbot do it?
Is this real, or a bad draw?	A repeat, fasting, 8 to 10 a.m.	It can tell you to. It cannot order it.
Testicle or pituitary?	LH and FSH	No
Is there a prolactinoma?	Prolactin, then imaging if indicated	No
Is the total misleading?	SHBG with free testosterone	No
Is it iron overload?	Ferritin and iron studies	No
Is it sleep apnea, alcohol, or a medication?	History, exam, sometimes a sleep study	Partly, if you tell it everything
Do you want children?	A conversation, before prescribing	It can raise it. It cannot hold you to it.
Is the ApoB related to the testosterone?	Reading both against one metabolic picture	Only if you think to ask
Did any of it work?	Re-testing on a cadence, for years	No

The last row is the one I would underline. Every number on this page is a snapshot. The information is in the trend, and a trend requires someone who is still there in eight months.

What I would actually do

Repeat the testosterone properly. Add LH, FSH, prolactin, SHBG, sensitive estradiol, a full thyroid, and iron studies. Add fasting insulin and hs-CRP, because if the metabolic hypothesis is right they will say so. Add Lp(a) once, because it is genetic, it is measured a single time in your life, and it changes how aggressive I am about the ApoB.

Then treat the driver and the particle count at the same time, rather than choosing. And re-test, on a schedule, so we find out whether any of it worked instead of assuming.

If that turns out to be a statin, or replacement, or both, I will say so plainly. Lifestyle first, bioidenticals second, phytoceuticals third, pharmaceuticals as a last resort is an ordering, not an ideology. A pituitary adenoma does not care about my sequence.

Bring the chatbot output with you

I mean this without any edge on it. If you have run your labs through an AI, print it and bring it. Patients who have thought about their own data are easier to work with, not harder. Some of what it told you will be right and I will say so. Where I disagree, I will tell you why, and that is a better starting point than a blank page.

What I will add is the four things it structurally cannot: I can order what was never run, examine you, prescribe, and still be here next year to see whether the number moved.

If you are holding results and nobody is reading them, that is what the Precision Health Assessment is for. The deeper hormone work and the particle-count side are measured inside ongoing care, chosen by judgment rather than off a menu. Or bring all of it to me, and it is me reading it, not a coordinator.

Creating Health with Dr. Tagge

The science of health, connected with the way we live: sourced and translated, with no hype. A new issue by email every Friday.

Get Creating Health by email

Written by

Daniel Tagge, MD

Board-certified family physician and founder of Tagge Health. Certified in Health Optimization Medicine.

About Dr. Tagge

Lab reference

ApoB: the laboratory range, the evidence, and what moves itCardiovascular risk falls continuously as ApoB falls and no floor has been found. Guidelines set treatment goals below 100, 80 or 65 mg/dL depending on risk, population data put the lower-risk end of the healthy distribution below about 80, and the lowest-risk tertile in a US sample sat near 60.
Triglycerides: the laboratory range, the evidence, and what moves itLaboratories call triglycerides under 150 mg/dL normal, the Copenhagen cohorts found heart attack risk already higher above about 88 than below it, the American Heart Association has called a fasting value under 100 'optimal', and genetic studies show the risk travels with the number of ApoB particles, so a triglyceride number is best read as a signal about remnant particles and insulin resistance rather than a target of its own.
HDL cholesterol: the laboratory range, the evidence, and what moves itIn cohorts, mortality is lowest with HDL cholesterol in the middle, about 55 to 70 mg/dL in pooled data and 73 (men) to 93 (women) in Copenhagen, and rises again at both ends; but Mendelian randomization and four large drug trials show that raising HDL cholesterol does not lower heart risk, so the number is a marker of metabolic health, not a lever.
Lp(a): the laboratory range, the evidence, and what moves itLp(a) is set by genes, is usually measured once in a lifetime, and raises heart attack, stroke and aortic stenosis risk continuously with no floor found. Consensus statements read under 75 nmol/L (30 mg/dL) as lower risk, 75 to 125 nmol/L (30 to 50 mg/dL) as a grey zone, and 125 nmol/L (50 mg/dL) or more as high. Diet and exercise barely move it, so a high value is a reason to lower ApoB and blood pressure harder, earlier.
LDL cholesterol: the laboratory range, the evidence, and what moves itLDL cholesterol causes atherosclerosis, risk falls by about a fifth for every 39 mg/dL it is lowered with no floor found in trials, and the harm accumulates over years, so laboratories print under 100 mg/dL as optimal, guidelines set goals of under 116, 100, 70 or 55 by risk, and 190 or above is a treatment indication on its own; the number is a calculation that triglycerides can distort, and ApoB is the better count of the particles that do the damage.
Non-HDL cholesterol: the laboratory range, the evidence, and what moves itNon-HDL cholesterol is total cholesterol minus HDL cholesterol, the cholesterol carried by every particle that can enter the artery wall, and it predicted heart disease at least as well as LDL cholesterol in more than 300,000 people without fasting; laboratories print under 130 mg/dL as desirable, guidelines set secondary goals of under 85, 100 or 130 by risk, the lowest 30-year event rates sat below 100, and 220 or above parallels an LDL of 190.
Total testosterone (men): the laboratory range, the evidence, and what moves itIn pooled cohorts of 20,654 men, all-cause mortality was higher only below about 213 ng/dL and cardiovascular mortality only below about 153, with no difference between quintiles above that; sexual symptoms clustered below about 320 in the European Male Aging Study; and no trial has shown that raising a man without symptoms from the middle of the range to the top of it changes any outcome. The harmonized reference range for healthy nonobese men aged 19 to 39 is 264 to 916 ng/dL, and guidelines diagnose deficiency only with consistent symptoms and two low fasting morning values.
SHBG: the laboratory range, the evidence, and what moves itSHBG is the liver protein that decides how much testosterone and estradiol is free, and its level reads out the liver's metabolic state: in nested case-control studies of 359 women and 170 men, the highest SHBG quartile had about one tenth the odds of type 2 diabetes of the lowest, genetic variants that raise SHBG lower diabetes risk, and SHBG falls with liver fat and rises with age, thyroid excess and oral estrogen. In 149,436 UK Biobank men and nine pooled cohorts, higher SHBG went with higher, not lower, mortality, so no guideline names an optimal value; what SHBG changes is the reading of a total testosterone, which is why guidelines ask for a free testosterone whenever SHBG is abnormal or the total is borderline.
DHEA-S: the laboratory range, the evidence, and what moves itDHEA-S is the sulfated storage form of the adrenal steroid DHEA, a precursor the tissues convert to androgens and estrogens rather than a hormone in its own right; it peaks in the twenties and falls by about 70 percent by the fifties, so every value must be read against age and sex. In older men, the lowest quarter of values kept company with higher mortality (hazard ratio about 1.5 in 2,644 Swedish men aged 69 to 81), with no gradient above that and no consistent signal in women; two years of DHEA that restored youthful levels in 144 older adults changed no measure of body composition, fitness, insulin sensitivity or quality of life; and no study defines an optimal DHEA-S inside the age range.

Read alongside.

Measure it

Want these numbers on your own chart?

Testing is chosen by judgment rather than protocol, interpreted against your own history, and repeated only when it will change a decision. That reading is where care with me begins.

Learn for free

Explore working with Dr. Tagge