AndroGel (Testosterone Topical) Erythrocytosis: When It Doesn't Go Away

AndroGel (testosterone 1% or 1.62% topical gel) is an FDA-approved testosterone replacement product for men with confirmed hypogonadism. It belongs to the broader class of testosterone replacement therapies (TRT), which also includes injectable esters, nasal gel, pellets, and other transdermal systems. This article is about one specific side effect: elevated hematocrit (erythrocytosis), and specifically what it means when that elevation does not resolve after the usual first steps of dose reduction or stopping the gel.

The useful question here is not whether AndroGel can raise red blood cell counts. It can. The more clinically important question is whether hematocrit that stays high after a dose cut or discontinuation is still "an AndroGel problem," or whether it has revealed something else, such as untreated sleep apnea or a primary bone marrow process, that needs its own diagnosis and treatment.

Why AndroGel Raises Red Blood Cell Counts

Testosterone stimulates red blood cell production (erythropoiesis) through mechanisms that are well established in the endocrinology literature, and transdermal formulations like AndroGel are not exempt from either pathway. The magnitude is generally smaller than with intramuscular testosterone, but a smaller average effect does not mean the effect is absent in any individual patient, and men who start with a borderline-high hematocrit can still cross a clinically significant threshold.

AndroGel raises hematocrit through two linked pathways: increased renal erythropoietin (EPO) secretion and suppression of hepcidin, the hormone that regulates iron availability for red cell production. In large placebo-controlled trials of testosterone gel in older men, hematocrit rose above 54% several times more often in the testosterone group than in the placebo group. Most men who stop or meaningfully cut their dose see hematocrit begin to fall within roughly six to ten weeks, because circulating red cells live about 120 days and existing cells have to complete that lifespan before turnover slows down. When hematocrit stays high well past that window, an independent driver, untreated sleep apnea, ongoing smoking, or an unmasked JAK2-positive marrow process, is often part of the explanation, and it needs its own workup rather than another dose adjustment.

The EPO pathway

Testosterone acts on the kidney to increase erythropoietin secretion, which in turn drives red blood cell progenitors in the bone marrow to mature and multiply faster. This effect is dose-dependent: higher free testosterone levels generally produce a stronger EPO signal. Because AndroGel is absorbed more gradually through the skin than an intramuscular injection is absorbed from muscle, peak free testosterone levels tend to be lower with gel, which is part of the reason transdermal testosterone is associated with a smaller average rise in hematocrit per milligram delivered compared with injections.

The hepcidin pathway

Testosterone also suppresses hepcidin, the hormone that controls how much iron is released into circulation. Lower hepcidin means more iron becomes available to feed the expanded pool of red cell precursors that EPO has already stimulated. This second mechanism is one reason hematocrit can keep rising for several weeks after a dose change: the iron supply stays elevated even as the EPO signal starts to taper.

How transdermal gel compares with injectable testosterone

Comparative data across testosterone formulations generally show a smaller average hematocrit rise with transdermal gels than with intramuscular injections, and current Endocrine Society guidance treats hematocrit elevation as formulation-sensitive, recommending dose reduction or a change in delivery route when hematocrit exceeds 54%. Even so, a subset of men on gel alone cross that threshold. The rest of this article focuses on what to do when the usual first response, cutting the dose, does not bring hematocrit back down.

A note on specific numbers in this section: exact percentage increases reported in individual trials vary by study population and definition of "elevated," and the underlying journal articles should be checked directly before quoting a precise figure to a patient.

How common is this, and who is at highest risk?

Exact prevalence of AndroGel-associated erythrocytosis depends on baseline hematocrit, prescribed dose, and duration of therapy. The general pattern across randomized trials in this area is consistent: testosterone gel increases the proportion of men who cross a hematocrit threshold like 52% or 54%, compared with placebo, over roughly six to twelve months of use. The Testosterone Trials program in older men and the earlier Testosterone in Older Men with Mobility Limitations (TOM) trial are the two most commonly cited randomized sources for this effect; both found a meaningfully higher rate of elevated hematocrit in the testosterone arm than in placebo, and the TOM trial was stopped early in part because of cardiovascular safety concerns in which elevated hematocrit was considered a contributing factor. Readers or clinicians who want exact percentages for citation should pull the original trial publications rather than relying on secondhand summaries, since figures are sometimes transcribed inconsistently across secondary sources.

Postmarketing adverse event reporting (through FDA's surveillance systems) has also captured erythrocytosis and polycythemia cases coded against testosterone gel products. These reports are subject to underreporting and reporting bias and cannot establish a rate, but the pattern is broadly consistent with what trials show: onset typically within the first several months of therapy, and a subset of cases where hematocrit did not normalize after stopping the drug, prompting further evaluation.

Men who start AndroGel with a baseline hematocrit already near or above 48%, men with undiagnosed obstructive sleep apnea, heavy smokers, and men living at high altitude carry higher baseline risk. Sleep apnea alone raises hematocrit through nocturnal hypoxia-driven EPO release, so starting testosterone on top of undiagnosed apnea adds a second erythropoietic stimulus on top of the first.

When does erythrocytosis resolve after stopping or reducing AndroGel?

This is the question most patients actually want answered. In general, hematocrit begins declining within about six to ten weeks of stopping or meaningfully reducing AndroGel, with fuller normalization over roughly three to six months for many men. A minority do not normalize within that window, and that minority is the focus of the rest of this article.

Typical resolution pattern

Because red blood cells survive roughly 120 days in circulation, hematocrit does not drop immediately even after AndroGel is stopped completely. New red cell production slows once the EPO stimulus falls, but the existing cells have to be cleared through their normal lifespan first. A rough trajectory after full cessation looks like this:

• Weeks 2 to 4: reticulocyte count (a marker of new red cell production) starts to fall, signaling reduced production pressure.
• Weeks 6 to 10: hematocrit itself begins a measurable decline in many patients.
• Months 3 to 6: hematocrit usually approaches pre-treatment baseline in men without a secondary driver.

A dose reduction rather than full cessation typically produces a shallower, slower decline, because the EPO signal is reduced but not eliminated. A patient who cuts dose in half may see hematocrit plateau at an intermediate level rather than trend clearly downward, and that plateau is not automatically abnormal, it may simply reflect a lower but still-present erythropoietic stimulus.

When it doesn't resolve

Persistent elevation beyond roughly three to six months after full cessation, or beyond about twelve weeks after a genuine dose reduction, is the clinical scenario this article is built around. A few explanations are worth ruling in or out:

An unmasked primary polycythemia. Polycythemia vera, a clonal bone marrow disorder, is associated with JAK2 gene mutations in the large majority of cases. Testosterone therapy can act as a trigger that reveals a previously subclinical JAK2-positive process. In this scenario, stopping AndroGel removes one erythropoietic stimulus, but the underlying marrow driver continues on its own.

A persistent secondary cause. Untreated sleep apnea, ongoing smoking, COPD with low oxygen levels, or living at high altitude will all maintain EPO drive independent of testosterone. Stopping AndroGel without addressing the comorbidity will not resolve the erythrocytosis.

Residual drug effect. Topical testosterone does not build the kind of tissue depot that intramuscular esters do, so lingering drug effect is an unlikely explanation for persistence beyond four to six weeks post-cessation, though it can slow the earliest phase of recovery.

A higher personal baseline. Some men have a naturally higher hematocrit set point. A value of 51 to 52% may be that person's normal range rather than a drug effect. Distinguishing this from pathology depends on having a genuine pre-treatment baseline value to compare against.

A 12-week persistence decision framework

Most general references stop at "reduce the dose and recheck labs." That advice does not tell a patient or a busy clinic what to do when the recheck at 12 weeks still shows an elevated number. The framework below is built specifically around that decision point.

Step 0, before anything else, confirm the number is real. Hemoconcentration from dehydration can inflate a single hematocrit reading. If a value is borderline, repeat the draw one to two weeks later, ideally in a well-hydrated, fasting morning state, before treating it as a true trend.

Step 1, classify the trajectory at 12 weeks after a genuine dose reduction or full stop:

Step 2, decide whether testosterone therapy should continue at all. If the man clearly benefits symptomatically from testosterone and hematocrit responds predictably to dose changes, a lower dose or a formulation switch may be reasonable. If hematocrit keeps rising independent of dose, or a secondary cause is found, the erythrocytosis is not primarily an "AndroGel problem" anymore, and management shifts to treating that underlying cause (for example, CPAP for sleep apnea, or hematology-directed care for a myeloproliferative disorder).

Step 3, exceptions that override the table above. Symptoms of hyperviscosity (severe headache, visual changes, chest pain, focal neurologic symptoms) at any hematocrit level warrant urgent evaluation, not a scheduled recheck. A hematocrit trajectory that looks reassuring on paper does not override new symptoms.

Structured workup when hematocrit won't come down

When hematocrit stays above roughly 54% more than twelve weeks after a real dose reduction, the following approach reflects the general shape of current professional guidance from bodies such as the Endocrine Society and the American Urological Association (AUA), which recommend evaluating for hypoxia and secondary causes alongside stopping or reducing testosterone, rather than treating dose adjustment as a first step to try before looking for other causes.

Confirm the elevation is real

Repeat the CBC after adequate hydration if the first result is borderline. If polycythemia vera is genuinely being considered, red cell mass testing may be part of a hematology referral.

Screen for secondary causes

A reasonable set of studies to order or discuss with the prescribing clinician before attributing the finding entirely to AndroGel includes:

• JAK2 V617F mutation testing (screens for polycythemia vera)
• Serum erythropoietin level (a low EPO level points toward a primary marrow process; an elevated EPO level points toward a secondary, oxygen-driven cause)
• Overnight pulse oximetry or a formal sleep study to rule out obstructive sleep apnea
• Pulmonary function testing in smokers or patients with respiratory symptoms
• Serum ferritin and iron studies

Phlebotomy

Therapeutic phlebotomy is the fastest mechanical way to lower hematocrit, typically producing a measurable drop within a day or two of a single unit removed. In polycythemia vera management generally, a hematocrit target below 45% is commonly used in men with cardiovascular risk factors; for testosterone-associated erythrocytosis specifically, many clinicians use a more practical target in the 50 to 52% range. Phlebotomy is usually repeated only as needed, guided by follow-up CBCs, and it is not a substitute for identifying why hematocrit will not come down on its own.

Consider a formulation change

If a patient clearly benefits from testosterone therapy and wants to continue it, switching delivery routes is sometimes reasonable:

• Testosterone nasal gel has been associated with less erythrocytosis than intramuscular testosterone in comparative data, likely reflecting its different absorption profile, though direct transdermal-gel-to-nasal-gel comparisons are limited.
• A lower AndroGel dose than the one that caused the problem may achieve adequate symptom control without crossing the hematocrit threshold, though this needs to be verified with repeat labs, not assumed.
• Clomiphene or enclomiphene, which stimulate the body's own LH/FSH and endogenous testosterone production rather than supplying exogenous testosterone, appear in limited data to carry a lower erythrocytosis burden, though head-to-head comparisons against topical testosterone are sparse.

Aspirin in men with ongoing elevation and cardiovascular risk

Low-dose aspirin is standard practice in true polycythemia vera to reduce thrombotic risk, and some clinicians extend the same reasoning to testosterone-associated erythrocytosis in men with additional cardiovascular risk factors. This is an extrapolation from polycythemia vera management rather than a practice tested in a dedicated randomized trial in TRT patients specifically, and it should be a discussion with the prescribing clinician, not a self-directed decision.

Reconciling the cardiovascular risk evidence

Patients often ask whether a high hematocrit from AndroGel is genuinely dangerous or a number their clinician worries about more than it deserves. The honest answer is that the evidence is not uniform, and the two most commonly cited studies point in different directions for different reasons.

An older, retrospective observational study of men with pre-existing coronary artery disease found an association between testosterone therapy and a higher rate of major adverse cardiovascular events, with erythrocytosis-driven blood thickening proposed as one possible mechanism. This was an observational analysis, not a randomized trial, and it cannot prove causation.

A large, prospective, randomized cardiovascular safety trial in men with hypogonadism and elevated cardiovascular risk, designed specifically to test this question, later found that testosterone therapy did not increase the overall rate of major adverse cardiac events compared with placebo over roughly three years of follow-up. Notably, that same trial found that erythrocytosis requiring dose reduction or phlebotomy was still substantially more common in the testosterone group than in the placebo group, confirming that the side effect itself is real even where the trial's overall cardiac outcome signal was reassuring.

The reconciling detail is monitoring. The reassuring trial excluded men with an already-elevated baseline hematocrit and checked blood counts on a fixed schedule, intervening whenever a threshold was crossed. That protocolized surveillance, not passive reassurance, is what the neutral cardiac outcome was built on. A hematocrit that is quietly climbing without anyone checking it is a different situation than one being actively monitored and treated.

Readers who want the exact hazard ratios and confidence intervals from either study should pull them directly from the published trial reports rather than relying on a rounded secondary summary like this one.

Monitoring: what a reasonable lab schedule looks like

Current professional guidance generally supports a CBC schedule along these lines for men starting testosterone therapy: a baseline check before starting, a follow-up around three months in, roughly every six months through the first year if hematocrit is stable, and annually after two consecutive normal results. Men who start with a baseline hematocrit near or above 48%, who have sleep apnea, a heavy smoking history, or a prior episode of erythrocytosis on any testosterone formulation, are reasonable candidates for closer monitoring, for example every six to eight weeks, during the first six months of therapy.

If hematocrit exceeds 54% at any point, the appropriate response is the structured workup above, not waiting for the next scheduled visit "in a few months."

Common misunderstandings worth correcting

Assuming gel carries no risk because it's "gentler." Topical testosterone produces smaller peak concentration swings than weekly injections, but a hematocrit of 56% carries the same physiologic risk regardless of how the testosterone was delivered. "Milder on average" is not the same as "safe to skip lab work."

Stopping AndroGel abruptly without a plan. Abrupt cessation removes the erythropoietic stimulus but does not lower hematocrit immediately, for the lifespan reasons described above. Meanwhile, sudden testosterone withdrawal can cause fatigue, mood changes, and sexual dysfunction that some men experience as worse than their original hypogonadism symptoms. A planned taper coordinated with serial CBC monitoring is generally preferable to an unplanned stop, unless hematocrit has reached an urgently dangerous range (roughly above 58 to 60%), in which case stopping immediately and seeking urgent evaluation takes priority.

Attributing all elevated hematocrit to AndroGel by default. As the secondary-cause workup above illustrates, a JAK2 mutation or untreated sleep apnea can independently drive erythrocytosis. AndroGel may have been the trigger that pushed hematocrit over a threshold while the underlying driver is something else entirely. Missing that distinction leads to a cycle of dose cuts and phlebotomy that never addresses the actual cause.

When cytoreductive therapy becomes relevant

In confirmed polycythemia vera, or when phlebotomy alone cannot control hematocrit, hematologists sometimes add marrow-suppressing therapy such as hydroxyurea or a JAK inhibitor. These drugs are not standard treatment for erythrocytosis caused purely by testosterone therapy, but they become relevant if the workup uncovers an underlying myeloproliferative disorder. General hematology guidance in this area treats frequent phlebotomy needs (commonly cited as more than roughly two per year) or high thrombotic risk as reasons to involve a hematologist rather than continuing to manage the problem through testosterone dose changes alone. Anyone meeting that description should get a hematology referral before restarting or continuing AndroGel.

Evidence boundary: what is established, what is plausible, what is not established

Established: Testosterone therapy, including transdermal gel, raises hematocrit on average, through EPO stimulation and hepcidin suppression, and this effect is dose-related. Randomized trials have shown a higher rate of clinically significant hematocrit elevation with testosterone gel than with placebo. A hematocrit above roughly 54% is a widely used clinical threshold for stopping or reducing testosterone therapy.

Plausible but not firmly proven for this specific population: That low-dose aspirin meaningfully reduces thrombotic risk specifically in TRT-associated erythrocytosis (this is extrapolated from polycythemia vera management, not tested in a dedicated trial in this population). That switching to nasal testosterone gel reliably prevents recurrent erythrocytosis in men who have already had the problem on transdermal gel (comparative data exist, but head-to-head trials in men with a documented prior episode are limited).

Not established: That AndroGel itself causes polycythemia vera. It does not; polycythemia vera is a clonal bone marrow disorder. What testosterone therapy can do is unmask a previously subclinical case. The exact individual-level thrombotic risk contributed by a given hematocrit value, independent of a person's other cardiovascular risk factors, is also not established with precision, current risk estimates come from mixed observational and trial data with real methodological limits, discussed above.

Summary thresholds

These thresholds reflect general professional guidance and common clinical practice patterns rather than a fixed rule for every patient; a clinician managing an individual's full risk profile may reasonably act earlier or later than these numbers suggest.

A note on sourcing: this draft intentionally avoids attaching specific journal citations or PMIDs to individual claims, because the identifiers associated with the prior version of this page could not be verified against the underlying papers before this rewrite. Before publication, an editor with primary-literature access should confirm and re-attach citations for the trial-level claims above (the Testosterone Trials, the TOM trial, the Vigen et al. observational study, and the TRAVERSE cardiovascular safety trial), the Endocrine Society and AUA guideline language, and the nasal testosterone gel comparison data, using the original publications rather than secondhand summaries. General regulatory information can be verified directly at fda.gov, and trial design details for large testosterone safety trials can be checked at clinicaltrials.gov.