The HAARLEM study suggests post-cycle therapy (PCT) doesn't work


During anabolic steroid use in high dosages, the endogenous production of testosterone will come to a screeching halt. Once halted, testosterone production will remain (severely) suppressed for a transient period of time after cessation of anabolic steroid use. During this period of time, the person will be hypogonadal, i.e. testosterone deficient. Anabolic steroid users commonly use certain drugs after their “cycle” in order to—presumably—speed up this process of recovering testosterone production. This practice is commonly referred to as post-cycle therapy (PCT).

Three types of drugs are frequently used for post-cycle therapy, namely:

  • Selective estrogen receptor modulators (SERMs, e.g. tamoxifen and clomiphene)
  • Aromatase inhibitors (AIs, e.g. letrozole, anastrozole, and exemestane)
  • Human chorionic gonadotropin (hCG, commonly also called Pregnyl)

The reasoning behind the usage of these drugs is quite simple. SERMs counteract the suppressive action estrogen has on testosterone production by blocking its receptors at the hypothalamus and pituitary. Similarly, aromatase inhibitors counteract it by decreasing the production of estrogen itself. Finally; hCG increases testosterone production directly by stimulating the testes to do so.

Of these drugs, SERMs are usually the mainstay in the ‘PCT world’. And, indeed, SERMs have proven to increase testosterone in various states of hypogonadism. However, no study to date had actually prospectively examined its efficacy in AAS-induced hypogonadism. Until now: let me introduce you to the HAARLEM study [1].

Want to learn more about anabolic steroids, PCT, and how they affect your endogenous testosterone production? Take a peek at my brand-new book 'Book on Steroids'. Chock-full of evidence-based information with well over 500 references to the scientific literature. A must-have on the bookshelf for anyone with a keen interest in anabolic steroids.

The HAARLEM study: a prospective and observational study in which 100 anabolic steroid users participated

The HAARLEM study is an initiative by the outpatient clinic for anabolic steroids users in Haarlem, the Netherlands. The outpatient clinic was established in 2010 and is run by the two endocrinologists dr. de Ronde and dr. Smit. (Dr. Smit has been so kind to proofread Book on Steroids and has provided me with invaluable feedback on the manuscript.)

The goal of the HAARLEM study was to obtain valuable information about the health risks involved in anabolic steroid use. Baseline characteristics of this cohort have been published previously [2].

In brief: a total of 100 subjects (all men) that were intending to start an anabolic steroid cycle within 2 weeks were included in the trial. Various health measurements, including blood tests, were performed on all participants before the cycle (T0), in the last week of the cycle (T1), 3 months after the end of the cycle (T2), and 1 year after the start of the cycle (T3). To be clear: subjects were using their own AAS, the endocrinologists didn’t prescribe it of course.

What’s of particular relevance for this article, is that they also measured testosterone levels and thus could see how it would recover after a cycle. Additionally, 80 of the subjects did PCT (and therefore, 20 didn’t). So this would be the first prospective trial in which the efficacy of PCT might become apparent. (In the end, data was available for 79 subjects doing PCT and 19 subjects not doing PCT).

PCT drugs also weren’t provided by the endocrinologists. People did their own thing. The authors noted that most PCT regimes consisted of tamoxifen (70% of the time) and/or clomiphene (55% of the time) for 4 weeks post-cycle. Which, indeed, would also be my guess as a stereotypical example of PCT.

Post-cycle therapy (PCT) did not aid in recovery of endogenous testosterone production after anabolic steroid use

I’m sure this rustles some jimmies, but here are the data:

Results from the HAARLEM study [1].
Testosterone (nmol/L) T0 T1 T2 T3
PCT (n=79) 16.6 81.6 14.4 15.8
No PCT (n=19) 16.5 68.8 16.3 15.6

Baseline (T0) testosterone values were pretty much identical, and, as to be expected, they were increased to supraphysiological levels during the last week of the cycle (T1). Then, 3 months after the end of the cycle, values were pretty much normalized in both groups again (albeit slightly, but not statistically significantly, lower in the PCT group).

Was this research perfect and thus the ultimate evidence to dispel PCT? No, it wasn’t. It wasn’t a double-blind placebo-controlled trial. But hey, it’s unlikely that such a trial will ever be performed. This is pretty much as good as it gets, they did a great job. So what kind of pitfalls are there? For one, there might have been a form of selection bias. That is, subjects who “know” they recover more easily, might’ve opted for not using a PCT. Although I consider it somewhat doubtful that it would lead to meaningful differences. Another reason could be that the average dosage was higher in the PCT group—which it was, 1.110 vs. 839 mg/week. However, both are well above the dosages required for maximal suppression of endogenous testosterone production (the minimum required dosage as a inclusion criteria was also 200 mg weekly). Moreover, the group that didn’t do PCT actually on average had a longer cycle duration (20 weeks vs. 18 weeks).

So, what else? In principle, perhaps the PCT group had recovered a bit earlier, which would've been visible if they had measured it 2 months after stopping anabolic steroid usage instead of 3. But yeah, I wouldn’t expect much of a difference either way still. If only because it would've taken most of these guys probably about a month before recovery would even start. After all, with high dosages and the long half-lives of the esters it will simply take this long before the AAS are cleared.

I’m also sure a lot of people will say that (most of) these subjects simply did their PCT wrong. But how would one know what would be a “good” PCT and what would constitute a “bad” PCT if there’s no evidence available that has examined the efficacies of various types of PCT? Of course, it might be that a certain type of PCT might have resulted in beneficial results. Nevertheless, this trial shows that when you follow a group of people who perform PCT as done in practice (SERMs for about a month) it simply doesn’t show to be effective. Or, in a sense, there's little to speed up if the non PCT group recovers in 3 months anyhow.

As a final note, some good news, the authors found that when gonadal function was normal at baseline, there was a 90% chance of having a normal total testosterone concentration after 3 months of recovery, and a 100% chance at the end of follow-up (on average about 8 months after stopping the cycle). Yay.

But why? Why isn’t a PCT working? SERMs show such good results in various types of hypogonadism! I’ll just paste a snippet here of a paragraph I had written about this in my book:

“Unfortunately, there's currently no good-quality evidence available in which SERMs are evaluated as a treatment for AAS-induced hypogonadism. The following sections below will therefore mainly focus on their use in hypogonadism due to other causes. But a word of caution is in place, as it's currently unknown how well these results translate to those suffering from AAS-induced hypogonadism. For the following sections, it's important to consider this difference with the populations of the studies I'm about to describe. The underlying cause of hypogonadism is vastly different. In principle, AAS-induced hypogonadism is a transient state post-cycle in which the hypothalamus and pituitary fail to adequately respond to decreased androgen and estrogen concentrations. After all, post-cycle, testosterone and estradiol concentrations are (very) low and thus the negative feedback they usually impose on the hypothalamus and pituitary is greatly diminished. So, while the stimulus (low estradiol and testosterone) to produce LH and FSH is there, the endocrine cells temporarily fail to respond adequately to this. It's unsure how making this stimulus larger by using a SERM will aid in HPGA recovery. Unlike AAS-induced hypogonadism, the hypogonadal study populations I'm going to cover are in a steady state of testosterone deficiency. Herein it would make sense that you can shift the steady state by making the stimulus larger with a SERM in order to increase testosterone. So again, be cautious when translating these studies into the post-cycle situation.”


The HAARLEM study was a prospective trial in which a large cohort of AAS users was followed over time. Several measurements, including testosterone levels, were performed before, during, and at two points after cessation of AAS usage. By comparing those subjects who performed a PCT to those who didn’t, some initial good evidence about PCT has finally surfaced. Unfortunately, the practice as commonly done, seems to be a bit useless.

Addendum 08-03-2021

I've heard this trial has gotten some criticism, I'll walk through a few of these:

  1. "Olivier de Hon is one of the authors. He's with the Dutch Doping authority." Yes, how does this exactly invalidate the results? Please point me what part got influenced by him. Moreover, I'm sure the Doping authority would've loved to see the steroid users NOT recover at all after 3 months already. But they did. I can assure you they would've loved different results from these. Jesus.
  2. "They didn't instruct the steroid users to do [x], [y], [z], which would've led to better results." Yes, it's an OBSERVATIONAL study, not an interventional study. If they would've instructed the AAS users to do certain things with the compounds they were using, it would be pretty hard to get it passed the medical ethical committee in the first place. The only way you get an intervention passed them is if the intervention encourages the users to take less steroids. The goal of this trial was to assess the health risks involved with androgen abuse in practice. An observational setup like this is EXACTLY what you would want to do in that case.
  3. "All androgens have been made equal on a milligram per milligram basis." Of course, there's no evidence to do otherwise. You could arbitrarily assign something like "2 mg testosterone = 1 mg trenbolone" or whatever, but that would be utter broscience. What would that number mean? Trenbolone is twice as potent in muscle-building? Twice as potent in suppressing the HPGA? Twice as potent in causing acne? Where the hell would you base these numbers on? Extremely flawed androgen assays? And how would this have affected the results anyway? ALL users had their endogenous testosterone production COMPLETELY suppressed during their cycles.
  4. "The subjects did their PCT wrong lulz." Well, first, back to point 2, they couldn't tell them to do otherwise in the first place. And second, please tell me which research can tell us what a "good PCT" entails. There is none. It's all broscience and guesswork based on extrapolating studies from study populations with different causes of hypogonadism. The guys in the trial simply did PCT like most people do: take SERM(s) for roughly a month.
  5. "They should've started PCT later." Ok, so what difference would you expect? The no-PCT group had the same testosterone levels that they had at baseline 3 months after their last injection anyways. How about just wait 3 months? That seems to work pretty well.
  6. "Lulz they didn't make a subanalysis based on who used compound [x] and who used compound [y]". Good luck with doing that, because only in 13% of the samples the vial exclusively contained the AAS that was on the label, and in 47% of the cases the vial didn't even contain the steroid claimed on the label at all but contained a different one (or ones):


  1. Smit, D. L., et al. "Disruption and recovery of testicular function during and after androgen abuse: the HAARLEM study." Human Reproduction (2021).
  2. Smit, Diederik L., et al. "Baseline characteristics of the HAARLEM study: 100 male amateur athletes using anabolic androgenic steroids." Scandinavian journal of medicine & science in sports 30.3 (2020): 531-539.