Key takeaways
- Continuous exposure desensitises receptors; pulsatile release with recovery intervals keeps tissue responsive.
- Hourly GnRH restores gonadotropin secretion while the same total dose infused continuously shuts the axis down.
- Pulse frequency encodes information — fast pulses favour LH, slow pulses favour FSH, from one molecule.
- Steroid hormones show no frequency encoding, so the goal in replacement is a stable concentration, not a synthetic pulse.
- Testosterone and growth hormone pulses are tied to sleep stages, so fragmented sleep removes the release windows themselves.
Almost nothing in the endocrine system is delivered as a steady drip. GnRH arrives in bursts every one to two hours. Growth hormone comes in a handful of pulses concentrated in deep sleep. Cortisol is released in discrete episodes riding on a daily envelope. Even insulin oscillates. This is not engineering evolution failed to smooth out: the pattern carries information the average level does not, and delivering the same total amount continuously produces a different biological result — sometimes the opposite one.
Why the body bothers
The core reason is receptor behaviour. A receptor exposed continuously to its ligand desensitises: it uncouples from its signalling machinery, is internalised, and is expressed at lower density. Sustained stimulation ends in silence. Intermittent stimulation, with recovery intervals between pulses, lets the receptor reset and keeps the tissue responsive.
The experiment that established this remains one of the cleanest in endocrinology. Animals with the hypothalamic GnRH signal destroyed were given GnRH back in two ways — as an hourly pulse or as a continuous infusion at the same total dose. Pulsatile delivery restored normal gonadotropin secretion. Continuous delivery of the identical amount shut it down (Belchetz et al., Science 1978). Same molecule, same quantity, opposite outcome. The pattern was the active ingredient.
Which hormones are pulsatile, and on what timescale
- GnRH — pulses every 60-120 minutes from the hypothalamus, the master clock of the reproductive axis
- LH — follows GnRH almost one-for-one, which is why LH pulses are used to infer GnRH activity that cannot be measured directly in humans
- Testosterone — released in bursts, concentrated in the sleep period and tied to sleep architecture rather than to clock time
- Growth hormone — four to five pulses per night, the largest in early slow-wave sleep, with near-undetectable levels between them (Veldhuis et al., Growth Horm IGF Res 1998)
- Cortisol — frequent pulses through the day, superimposed on a strong diurnal envelope that peaks shortly after waking
- Insulin — rapid oscillations from pancreatic beta cells, on a scale of minutes
- TSH — modest pulsatility with an overnight rise
Note what this implies for testing. A single blood draw samples a waveform at one arbitrary point. For growth hormone that makes a random measurement close to meaningless, which is why IGF-1 — a stable integrator of GH exposure — is measured instead. For testosterone it is why a morning draw is specified, and why two low results are usually required before a diagnosis. Hormone diurnal rhythms covers the timing side.
How a pulse carries information
A pulsatile signal has more dimensions than a continuous one, and different tissues read different dimensions (Thompson et al., Mol Cell Endocrinol 2014):
- Frequency. The best-worked-out example is GnRH: rapid pulses preferentially drive LH beta-subunit transcription, slower pulses favour FSH beta. The pituitary is reading the interval between pulses, not the concentration, and it produces two different hormones in different proportions depending on the answer (Marshall et al., on GnRH pulse frequency and gonadotropin synthesis).
- Amplitude. Pulse height sets total signal magnitude and, for hormones with threshold effects, determines whether a given tissue responds at all.
- Pattern. Whether pulses are clustered, diurnally shaped or event-triggered carries meaning of its own — cortisol's morning surge is a different instruction from the same cortisol released at midnight.
So a single average value can describe two physiologically distinct states. This is how one signalling molecule runs the entire menstrual cycle: GnRH pulse frequency changes across the cycle, the LH-to-FSH ratio changes with it, and follicular development and ovulation follow. Persistently fast pulses favour LH over FSH — as in PCOS, where that contributes to anovulation. The HPG axis end to end traces the full circuit.
What continuous exposure does instead
The therapeutic use of this proves the principle in the other direction. GnRH agonists, given continuously rather than in pulses, initially stimulate the pituitary, then desensitise and downregulate it, shutting the axis off entirely. That shutdown is the point: it is how these drugs achieve androgen deprivation in prostate cancer, suppression in endometriosis, and control of central precocious puberty.
The same molecule that sustains the reproductive axis when pulsed abolishes it when delivered steadily. It is hard to construct a clearer demonstration that pattern is a variable in its own right.
What this does and does not mean for replacement therapy
The tempting conclusion is that all hormone replacement should reproduce natural pulses. That is right for some systems and wrong for others, and the distinction matters.
Where frequency is genuinely encoded — GnRH being the clear case — pulsatile delivery is required for the treatment to work at all, which is why pulsatile GnRH pump therapy exists for inducing fertility in hypothalamic hypogonadism. Where the receptor is a nuclear steroid receptor acting over hours on gene transcription, as with the androgen receptor, there is no evidence of frequency encoding, and reproducing hour-by-hour pulses is neither achievable nor obviously desirable.
For testosterone the relevant argument is therefore about stability, not pulsatility:
| Approach | Concentration profile | Practical consequence |
|---|---|---|
| Weekly intramuscular injection | A high peak in the first days, a low trough before the next dose | The widest swing, and the pattern most associated with cyclical mood, energy and libido complaints |
| Twice-weekly injection | Half the excursion, same weekly total | Smoother, and the most common first adjustment when someone reports a bad end-of-week |
| Daily subcutaneous micro-dosing | Near-flat, with small daily variation | The most stable; more injections, and not required by everyone |
| Cream or gel | Daily application with a diurnal-ish profile | Stable, with transference and absorption variability to manage |
Pharmacokinetic work comparing preparations supports the general point that formulation shapes the concentration curve substantially (Partsch et al., Eur J Endocrinol 1995). The claim that more frequent dosing produces better symptom control, though, rests mostly on mechanism and clinical experience rather than on randomised comparison, and it should be stated at that strength. Plenty of men do well on a weekly schedule. The men who benefit most from splitting the dose are the ones who can describe a predictable weekly cycle in how they feel — subcutaneous versus intramuscular and injection timing cover the practical side.
Elsewhere in practice
- Growth hormone replacement is given at bedtime to sit near the natural nocturnal pulse window, and secretagogues work by amplifying pulses that already exist rather than adding a flat background.
- Thyroid hormone is the exception that proves the rule: T4 has a half-life of about a week, so the system is already smoothed and daily consistency is about absorption, not rhythm.
- Sleep is the upstream variable for two of the most pulsatile hormones there are. Testosterone bursts and the largest growth hormone pulses both occur in specific sleep stages, so fragmented sleep removes the release windows themselves. No dosing schedule compensates for that — sleep and testosterone and sleep architecture and hormones cover the mechanism.
The clinical pearl: when someone feels different on Monday than on Friday at an unchanged weekly dose, the average concentration is not the problem — the shape of the curve is. That is a formulation and frequency conversation, not a dose-increase conversation, and increasing the dose usually makes the peak worse without fixing the trough.
Bottom line
Most hormones are released in pulses, and the pulse pattern carries information the average level cannot: frequency selects between outputs, amplitude sets magnitude, and continuous exposure desensitises receptors rather than stimulating them harder. GnRH is the definitive case, where continuous delivery shuts down the axis it would otherwise sustain. For replacement therapy the lesson is more limited than it is often made: steroid hormones do not appear to be frequency-encoded, so the practical goal is a stable concentration rather than a synthetic pulse, and smoother delivery is a reasonable response to cyclical symptoms even though the randomised evidence for it is thin. Protecting sleep does more for the genuinely pulsatile hormones than any schedule change. Decisions about formulation and frequency belong with a physician reading your labs alongside your symptoms.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
