AndroGel (Testosterone Topical) Erythrocytosis Severity Grading Rubric

Which hematocrit cutoffs are actually used?

A single-center review of 178 hypogonadal men treated for 3 years called hematocrit above 50% erythrocytosis and found it in 12.8% of men on gels, 66.7% of men on injectable testosterone and 35.1% of men on pellets [1]. A review of erythrocytosis in aging men used a cutoff of 0.52 (52%) and reported that men on testosterone replacement had a 315% greater risk of crossing it than controls [2]. A 2024 retrospective cohort of US veterans with type 2 diabetes defined erythrocytosis as hematocrit above 54% within a year of starting therapy and found it in 756 of 53,971 patients, or 1.4% [4]. Separately, a 2025 review of testosterone adverse-effect management states that therapeutic phlebotomy is indicated when hematocrit exceeds 52% [3].

One study reported results as three bands rather than one cutoff: among 247 patients on testosterone at a single center, 57% reached a hematocrit above 0.46, 23% above 0.50 and 5% above 0.54 [6]. Only 35 of those 247 patients used a gel [6], so those figures describe testosterone therapy generally rather than AndroGel specifically.

The 2018 Endocrine Society guideline recommends against starting testosterone in a man who already has an elevated hematocrit, and recommends checking hematocrit alongside symptoms, serum testosterone and prostate cancer risk during the first year of treatment [7].

How much does testosterone gel raise hematocrit?

The largest comparison by formulation is a network meta-analysis of 29 placebo-controlled randomized trials covering 3,393 men [5]. Mean hematocrit rose against placebo with every formulation tested: gel +3.0% (95% CI 1.8 to 4.3), intramuscular enanthate or cypionate +4.0% (2.9 to 5.1), oral testosterone undecanoate +4.3% (0.7 to 8.0), intramuscular testosterone undecanoate +1.6% (0.3 to 3.0) and patch +1.4% (0.2 to 2.6) [5]. Intramuscular enanthate or cypionate raised hematocrit significantly more than the patch; no other pair of formulations differed significantly, including gel versus injection [5]. These are figures for gels as a group, not AndroGel specifically.

The retrospective 178-man review shows a wider gap: erythrocytosis above 50% in 12.8% of gel users versus 66.7% of injectable users [1]. That study was not randomized, and the injection group was younger (mean 42.5 versus 54.1 years) with a higher starting hematocrit [1].

The generic testosterone gel 1% label, which is not an AndroGel label, lists "lab test abnormal" in its 180-day controlled trial, a category that included elevated hemoglobin or hematocrit among other laboratory changes, in 6% of men taking 50 mg, 5% taking 75 mg and 3% taking 100 mg, and in 9.3% of 162 men in its 3-year extension [8].

In a 6-month randomized trial of daily testosterone gel in older men with mobility limitation, hemoglobin and hematocrit rose by 7% and 10%; erythropoietin rose and hepcidin and ferritin fell at 1 and 3 months, and by 6 months erythropoietin and hepcidin had moved back toward baseline while hemoglobin and hematocrit stayed above baseline [9]. The investigators proposed that testosterone stimulates erythropoietin and resets the relationship between erythropoietin and hemoglobin [9].

What the labels say

The AndroGel 1.62% label handles this under the heading Polycythemia. It says to check hematocrit before starting treatment, re-evaluate it 3 to 6 months after starting and then annually, and that if hematocrit becomes elevated, therapy should be stopped until it decreases to an acceptable concentration [15]. It adds that an increase in red blood cell mass may raise the risk of thromboembolic events [15]. The generic gel 1% label has the same wording [8]. Neither sets a numeric hematocrit at which to stop. AndroGel 1.62% is a separate gel product, and it was the gel used in TRAVERSE [10].

Does a high hematocrit on testosterone mean more clots?

Reviews are careful here. A 2018 review concluded that erythrocytosis risk with testosterone therapy is well established and that short-acting injectables carry the highest incidence, but that the link between erythrocytosis and thromboembolic events remains unclear [11]. A 2015 review reached the same place: all formulations raise hemoglobin and hematocrit, injectables most, and the erythrocytosis-VTE link is inconclusive [2].

A claims-database case-crossover study of 39,622 men found testosterone use associated with a higher risk of VTE in men with hypogonadism (odds ratio 2.32, 95% CI 1.97 to 2.74) and in men without it (odds ratio 2.02, 95% CI 1.47 to 2.77) [12]. The authors list elevated hematocrit, which raises blood viscosity, among several possible mechanisms, without showing it was the cause [12].

TRAVERSE was a randomized cardiovascular outcomes trial of daily 1.62% testosterone gel in 5,246 men; major adverse cardiovascular events were 7.0% versus 7.3% on placebo, hazard ratio 0.96 (95% CI 0.78 to 1.17) [13]. The generic gel 1% label reports TRAVERSE venous thromboembolism rates of 1.7% on testosterone gel versus 1.2% on placebo, with deep vein thrombosis in 0.6% versus 0.5% and pulmonary embolism in 0.9% versus 0.5% [8].

Does dose or age change the grade?

Hematocrit response to testosterone is dose-dependent in the data that exist. In a dose-response study, 60 men aged 60 to 75 and 61 men aged 19 to 35 received weekly injected testosterone enanthate at 25, 50, 125, 300 or 600 mg for 20 weeks; hemoglobin and hematocrit rose linearly with dose in both groups, and the rise was significantly greater in the older men [14]. That study tested injected enanthate, not AndroGel, so it does not establish how changing an AndroGel dose affects hematocrit.

Body weight and starting hematocrit also appear in observational data. In the 247-patient single-center series, body mass index was significantly associated with reaching a hematocrit of 0.50 or higher, and a higher starting hematocrit was associated with crossing 0.46 [6]. Of the patients who crossed 0.46, 46% had their highest value during the first year of treatment [6].