Duration decides the taper, and the cut-off sits at three to four weeks
The number written on the box matters far less than the number of weeks it has been taken.
Five days of 40 mg for an asthma flare ends on the last tablet. Ten days of 30 mg for an inflamed joint ends the same way. Recommendation 2.1 of the 2024 European Society of Endocrinology and Endocrine Society guideline suggests not tapering glucocorticoids given for less than three to four weeks, irrespective of the dose, and adds that these patients can stop without testing because sustained suppression of the hypothalamic-pituitary-adrenal axis is unlikely over that window. Australian Prescriber compresses the same evidence into a single line: short courses, even at relatively high doses, rarely leave lasting suppression.
The two mistakes made around that line run in opposite directions. Patients taper week-long courses nobody meant them to taper, having absorbed the general warning that steroids are never stopped suddenly, and spend an extra fortnight on a drug whose work was finished. Others stop nine months of treatment on the grounds that the last tablet was only 5 mg. The second error is the dangerous one.
Three weeks is also not a switch that flips on day twenty-two. Several categories of short course carry enough risk that British guidance asks for a written withdrawal anyway.
Some courses under three weeks still need a written withdrawal
Duration is the first filter. These categories are the second, and they apply even to a ten-day prescription.
Reasons to write a withdrawal schedule regardless of course length
- More than 40 mg prednisolone daily for over a week
- 2 mg/kg daily for a week, or 1 mg/kg daily for a month
- Repeat doses taken in the evening
- More than three weeks of continuous treatment
- Recent repeated courses, especially any beyond three weeks
- A short course within a year of stopping long-term therapy
- Any other recognised cause of adrenal suppression
British and New Zealand formularies publish a list here rather than a rule, and the list catches people who assume the short-course exemption covers them. A gradual withdrawal is advised for anyone who has taken more than 40 mg of prednisolone daily for over a week, who has been given repeat evening doses, who has had more than three weeks of treatment, who has recently had repeated courses - particularly courses running beyond three weeks - or who takes a short course within a year of stopping long-term therapy. Any other cause of adrenal suppression counts as well.
Evening dosing deserves separate attention because patients rarely mention it. Adrenal activity peaks between 2 am and 8 am, so a dose taken at night suppresses ACTH release during the hours the axis would otherwise be busiest. Splitting 20 mg across the day is not neutral convenience.
Repeated bursts are the other under-reported pattern. Four seven-day courses across a winter is twenty-eight days of steroid, and the axis reads cumulative exposure more faithfully than the prescribing record does.
Withdrawal symptoms and adrenal insufficiency overlap at the bedside
Feeling awful at 4 mg has three possible causes, and only one of them is an emergency.
Three explanations compete when someone feels unwell during a reduction. Glucocorticoid withdrawal syndrome produces fatigue, aching muscles and joints, poor appetite, low mood and disturbed sleep in a patient whose axis is working adequately, reflecting the loss of a supraphysiological state the body had adapted to. Relapse reproduces the original disease. Adrenal insufficiency brings postural dizziness, nausea, vomiting, abdominal pain and weight loss, and it kills.
The overlap is genuine and the guideline acknowledges it. Where withdrawal symptoms are severe, the recommended manoeuvre is to go back to the most recent tolerated dose and lengthen the whole schedule rather than to abandon it. That instruction is worth writing on the reduction sheet, because patients who feel terrible at 3 mg often assume the plan has failed and either stop or double.
Timing gives a clue but not a diagnosis. Withdrawal tends to follow each reduction and settle over one to two weeks. Symptoms that arrive during a second illness, or that include vomiting and postural collapse, point at the axis and need same-day assessment. Anyone taught to tell the two apart at home should also be told which one justifies an emergency call.
A morning cortisol answers the one question the calendar cannot
Two acceptable paths exist below 5 mg, and one of them involves a blood test with strict conditions attached.
At the physiologic dose the guideline offers two routes and treats them as equivalent. Either continue reducing gradually while watching for symptoms of adrenal insufficiency, or measure a morning serum cortisol and let the number shorten the process. The second route suits patients whose exposure was relatively brief and those developing symptoms as the dose comes down, since a reassuring result allows the drug to stop rather than dwindle for another six months.
Technique determines whether the result means anything. The sample is drawn between 8 and 10 am, before the morning tablet, and never while the patient is on dexamethasone - switch to prednisolone or hydrocortisone first. Above 300 nmol/L, roughly 10 micrograms per decilitre, the glucocorticoid can be stopped. Between 150 and 300 nmol/L the physiologic dose continues and the test is repeated in a few weeks. Below 150 nmol/L the axis is suppressed and the repeat belongs months away, not weeks.
Treat the value as a continuum rather than a pass mark: 305 nmol/L does not mean the patient will cope with pneumonia next week. Dynamic testing is not recommended as a routine step during tapering. Where the axis has not recovered after a year at the physiologic dose, or where there has already been an adrenal crisis, the guideline asks for endocrinology input rather than another schedule.
Illness during the reduction, and for a year afterwards, needs extra cover
Sick-day dosing applies while the dose is coming down and for a good while after it reaches zero.
A suppressed axis cannot mount the surge of cortisol that fever, vomiting, trauma or surgery demands. The prednisone label states the principle without numbers: patients on corticosteroid therapy subjected to unusual stress need increased doses of a rapidly acting corticosteroid before, during and after the stressful event. NICE guidance on adrenal insufficiency supplies the numbers - during significant physiological stress, at least 40 mg of oral hydrocortisone daily in two to four divided doses, or at least 10 mg of oral prednisolone daily in one or two doses, until the illness resolves.
Vomiting is the failure point. A dose lost within 30 minutes is repeated at double the original amount once vomiting settles; if that returns too, the patient needs intramuscular hydrocortisone and an emergency department. The Society for Endocrinology reduces this to two rules: double the daily oral dose during illness with fever needing bed rest or antibiotics, and switch to injection during prolonged vomiting or diarrhoea, before colonoscopy preparation, and around trauma or surgery.
The cover does not end with the last tablet. Recovery of the axis continues for months after cessation, so anyone facing anaesthesia, a serious infection or major injury within roughly a year of finishing long-term steroids should have that history on record. Surgical teams ask about current medication and often miss the course that ended in March.
Below 5 mg the decrements shrink to 2.5 mg and then to 1 mg
This is the half of the schedule patients find frustrating, and the half that prevents crises.
| Daily dose | Typical decrement | Interval | What limits the pace |
|---|---|---|---|
| Above 40 mg | 5-10 mg, or 30-50 per cent | Weekly to monthly | Disease relapse |
| 20-40 mg | 5-10 mg | Weekly to fortnightly | Disease relapse |
| 10-20 mg | 2.5-5 mg | Every 1-2 weeks | Disease relapse, early withdrawal symptoms |
| 5-10 mg | 2.5 mg | Every 2 weeks | Withdrawal symptoms |
| Below 5 mg | 1 mg | Monthly | Adrenal recovery |
Once the daily dose approaches the physiologic equivalent the pace changes for endocrine reasons. Hypothalamic releasing factors have to resume, ACTH has to return to a pulsatile rhythm, and only then does the atrophied cortex start producing cortisol again. That sequence takes weeks to months, and there is no way to accelerate it with a prescription.
The practical consequence is the 2.5 mg tablet. Standard schedules move in 2.5 mg decrements every two weeks between 10 and 5 mg, and then slow further. For someone who has taken prednisolone for more than six months, Australian Prescriber offers a concrete regimen: reduce from 5 mg to cessation by 1 mg every month. On hydrocortisone the equivalent is a 4 mg monthly reduction below 20 mg a day. Neither figure comes from a randomised trial. Both come from physiological reasoning and accumulated practice, and the guideline authors say so.
Rate follows cumulative exposure rather than current dose alone. Two people sitting at 5 mg have different prospects if one has taken steroids for four months and the other for six years, and the heaviest users spend months at each milligram.
The first phase falls quickly and the disease sets the speed
Nothing about the top half of a reduction is delicate. The axis is fully suppressed and stays that way.
Above the physiologic band the limiting factor is relapse, so reductions can be large and frequent. The joint endocrine guideline states plainly that the taper can be faster and in larger decrements when the daily dose is high, above 30 mg of prednisone or so. Published schedules reduce by 5 to 10 mg weekly, or by 30 to 50 per cent every two to four weeks, until 20 mg a day is reached, then by 5 mg fortnightly or 2.5 mg weekly down to 10 mg.
Australian Prescriber frames this phase as one instruction: bring the dose down as rapidly as the underlying disease allows, aiming to avoid a flare. Where glucocorticoid complications have already appeared - hyperglycaemia that will not settle, mood disturbance, a new fracture, steroid psychosis, herpetic keratitis - the guideline permits an even faster descent towards physiologic dosing, because the harm from staying high outweighs the inconvenience of a flare.
Cortisol testing has no role here. Routine adrenal function testing is recommended against in patients on supraphysiologic doses, and the reasoning is arithmetical: at 20 mg a day suppression is near certain, so a low result changes nothing. Anyone on dexamethasone or betamethasone should move to a shorter-acting agent before the second phase, partly because dexamethasone makes a later cortisol assay uninterpretable.
Adrenal output of 10 to 20 mg of cortisol a day sets the floor
Everything in the second half of a reduction is arithmetic against the body's own daily cortisol output.
Numbers a reduction schedule is written against
An average 70 kg adult produces somewhere between 10 and 20 mg of cortisol every 24 hours, which corresponds to roughly 2.5 to 5 mg of prednisolone. Guideline tables put the physiologic daily equivalent at 15 to 25 mg hydrocortisone, 4 to 6 mg prednisone or prednisolone, 3 to 5 mg methylprednisolone, or 0.25 to 0.5 mg dexamethasone. Every milligram above that band is supraphysiological, and several weeks of supraphysiological dosing tells the pituitary to stop releasing ACTH. The adrenal cortex then thins from disuse.
That is the whole mechanism behind the taper. Nothing about the disease requires a gentle finish once inflammation is controlled; the slow part exists because an atrophied cortex needs weeks to months of pulsatile ACTH before it can cover a normal day, let alone a febrile one.
The numbers also explain why sources quote different endpoints. Australian Prescriber aims the fast phase at prednisolone 5 mg a day, the joint endocrine guideline says 4 to 6 mg, the British National Formulary describes reducing rapidly to 2 to 2.5 mg per square metre daily and then slowing, and the prednisolone SmPC uses 7.5 mg as its marker for slowing down. These are approximations of one physiological band, and a schedule written to any of them is defensible.
Where the published guidance stops and judgment takes over
Some of this field is settled. A useful amount of it is not, and the difference matters when a schedule is written.
Two questions in this field remain genuinely open. Whether switching prednisolone to hydrocortisone before the final steps helps is one: the argument is that a shorter duration of biological action gives the axis more overnight room, the supporting evidence is thin, and a recent study found most patients wean off prednisolone without the switch. The rate of reduction below physiologic dosing is the other, and it rests on reasoning and case series. Anybody offering one correct schedule is claiming more than the literature supports.
Two things are not open. Reduction begins only when the condition that justified the steroid is controlled and the drug is no longer needed for it; if the disease still needs treating, the conversation belongs to steroid-sparing agents. And the plan has to exist on paper. Current dose, total duration of exposure, the next reduction with its earliest date, the rule for holding or stepping back up, sick-day doses in milligrams, and an out-of-hours number: that is what an emergency department needs at 2 am and what patients cannot recall when unwell.
One thing the guidance is unanimous about: reductions are individualised by cumulative exposure and disease, and a schedule written for someone else on the same tablet strength may be wrong by months. What accumulates while treatment continues is covered in the long-term safety briefing; the drug itself is on the prednisone page.