Approach to discontinuing systemic glucocorticoid therapy in adults and children

INTRODUCTION — 

Systemic glucocorticoids (eg, prednisone, dexamethasone), also commonly referred to as "steroids" or "corticosteroids," are used in the treatment of a variety of disorders because of their potent antiinflammatory effects and, in some cases, immunosuppressive activity [1].

Glucocorticoids are analogues of the adrenal steroid hormone cortisol and may cause a variety of adverse effects, including increased risk of cardiovascular disease, hypertension, type 2 diabetes [2], osteoporosis-related fractures, and overall mortality [3]. Glucocorticoids are therefore discontinued as soon as the disease being treated is under control. As an example, patients with many chronic inflammatory diseases (eg, rheumatoid arthritis [RA], inflammatory bowel disease [IBD]) often receive glucocorticoids as part of their initial therapy while transitioning to glucocorticoid-sparing immunosuppressive agents. When providers anticipate stopping systemic glucocorticoid therapy, the approach to discontinuation should be guided by the patient's risk of developing glucocorticoid-induced adrenal insufficiency (AI) as well as a flare of the underlying disease.

This topic discusses the indications for and risks of discontinuing treatment with systemic glucocorticoids, as well as a general approach to discontinuation based on the risk of AI. Additional guidance for using systemic glucocorticoids is provided in the appropriate disease-specific treatment topics. An overview of the use of glucocorticoids and related toxicities is also provided elsewhere. (See "Overview of the pharmacologic use of glucocorticoids" and "Major adverse effects of systemic glucocorticoids".)

The clinical manifestations, diagnosis, and treatment of AI in adults and children are presented separately:

●(See "Clinical manifestations of adrenal insufficiency in adults".)

●(See "Treatment of adrenal insufficiency in adults".)

●(See "Clinical manifestations and diagnosis of adrenal insufficiency in children".)

●(See "Treatment of adrenal insufficiency in children".)

TERMINOLOGY — 

For the purposes of this topic, we define the following terms as follows:

●Systemic glucocorticoids – This refers to enteral (eg, intake by mouth or feeding tube) and parenteral (eg, intravenous [IV], subcutaneous, and intramuscular) forms of glucocorticoids. While intraarticular and intramuscular formulations are also considered systemic, this topic describes discontinuing glucocorticoids that are administered daily or every other day for prolonged periods of time.

●Physiologic dose of glucocorticoids – This refers to the dose of glucocorticoid that is approximately equivalent to daily cortisol production by the adrenal glands. Physiologic replacement doses of various glucocorticoid formulations are provided in the table (table 1). For adults, physiologic replacement doses are 15 to 25 mg/day of hydrocortisone or 5 to 7.5 mg/day of prednisone [4]. For children, the physiologic replacement dose is approximately 8 to 10 mg/m²/day of hydrocortisone or 2 to 2.5 mg/m²/day of prednisone. However, in both adults and children, there is significant individual variability in the daily glucocorticoid dose required to prevent adrenal insufficiency (AI), and dosing must be adjusted to individual patient needs [5]. (See "Overview of the pharmacologic use of glucocorticoids", section on 'Natural patterns of endogenous cortisol'.)

●Glucocorticoid taper – This refers to a medication regimen in which the dose of glucocorticoids is gradually reduced (weaned) over time. Tapers are usually guided by the patient's therapeutic needs (eg, initiating therapy for a chronic rheumatic disease while transitioning to glucocorticoid-sparing agents) but may be adjusted to manage symptoms of glucocorticoid withdrawal or AI. (See 'Glucocorticoid withdrawal syndrome' below and 'Chronic adrenal insufficiency' below.)

These definitions are generally aligned with those in the joint guidelines on the diagnosis and treatment of glucocorticoid-induced AI issued by the European Society of Endocrinology and the Endocrine Society in 2024 [4].

INDICATIONS FOR DISCONTINUING GLUCOCORTICOIDS — 

Clinicians may stop systemic glucocorticoid therapy in the following scenarios:

●No clinical indication for ongoing therapy – Glucocorticoids are discontinued when they are no longer needed to treat the underlying condition (eg, the condition has resolved or is in remission).

●Intolerable side effects from therapy – Glucocorticoid therapy may need to be discontinued if patients experience serious adverse drug effects, such as hyperglycemia, osteoporosis, and/or Cushing syndrome. (See "Major adverse effects of systemic glucocorticoids", section on 'Organ-based toxicity of systemic glucocorticoids'.)

Several complications require immediate rapid dose reduction and, if possible, cessation of glucocorticoids:

•Glucocorticoid-induced acute psychosis that is unresponsive to antipsychotic medications (see "Psychosis in adults: Epidemiology, clinical manifestations, and diagnostic evaluation")

•Herpetic keratitis, which can rapidly lead to perforation of the cornea and possibly permanent blindness (see "Herpes simplex keratitis")

•Severe or persistent hypertension despite appropriate therapy (see "Overview of hypertension in adults")

•Severe or persistent hyperglycemia despite appropriate therapy or hyperglycemia-associated complications (eg, diabetic ketoacidosis or hyperosmolar hyperglycemic state) (see "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis")

●Medication interactions – Patients may rarely need to stop due to medication interactions (table 2), although dose adjustment is often sufficient.

If patients urgently need to stop glucocorticoids and immediate cessation is not possible (eg, because of a critical clinical need), we use the lowest necessary glucocorticoid dose and discontinue them as soon as possible.

When glucocorticoids need to be withdrawn and the underlying disease is still active, providers must carefully monitor patients for worsening disease and consider empiric escalation of alternative therapies (eg, other immunosuppressive agents). (See 'Monitoring' below.)

OVERVIEW OF GLUCOCORTICOID-INDUCED ADRENAL INSUFFICIENCY — 

Central adrenal insufficiency (AI) is a potential complication of prolonged glucocorticoid use. Patients who are at a higher risk of developing AI may require a taper prior to discontinuing glucocorticoids.

Physiology — Glucocorticoid-induced AI refers to isolated cortisol deficiency resulting from suppression of the hypothalamic-pituitary-adrenal (HPA) axis. It is a form of central (or tertiary) AI. HPA suppression leads to reduced secretion of corticotropin-releasing hormone (CRH) and adrenocorticotropic hormone (ACTH) due to feedback from long-term therapy with supraphysiologic doses of glucocorticoids. When CRH and ACTH are suppressed, patients are unable to endogenously increase cortisol production in times of stress (eg, fever, trauma) and can therefore develop symptoms of AI and even adrenal crisis. In addition, patients may also develop symptoms of AI if they significantly decrease their dose or stop exogenous glucocorticoids and are unable to produce a sufficient daily amount of cortisol.

More information on the physiology of the HPA axis and development of central AI in children and adults is provided separately:

●(See "Laboratory assessment of hypothalamic-pituitary-adrenal axis function", section on 'Basic physiology of the hypothalamic-pituitary-adrenal axis'.)

●(See "Causes of central adrenal insufficiency in children", section on 'Chronic high-dose glucocorticoid therapy'.)

●(See "Causes of secondary and tertiary adrenal insufficiency in adults", section on 'Chronic high-dose glucocorticoid therapy'.)

Clinical manifestations — Glucocorticoid-induced AI can be associated with a range of symptoms, ranging from mild to severe. Key differences between AI and glucocorticoid withdrawal syndrome are illustrated in the table (table 3).

Glucocorticoid withdrawal syndrome — Glucocorticoid withdrawal syndrome refers to symptoms (eg, fatigue, myalgias, arthralgias) that are not caused by the patient's underlying condition, occurring when glucocorticoid doses are lowered but remain supraphysiologic (table 4). This was historically called "pseudorheumatism" in patients with rheumatoid arthritis (RA) [6]. Abrupt reductions in glucocorticoid dose appear to be the primary factor associated with an increased risk of glucocorticoid withdrawal syndrome. This is supported by a high prevalence of glucocorticoid withdrawal syndrome in patients treated for Cushing syndrome with surgical resection of ACTH-secreting tumors [7]. Patients with glucocorticoid withdrawal syndrome may have co-occurring AI during times of physiologic stress (eg, serious infection, trauma) and should be managed accordingly. (See 'Management of complications' below.)

Chronic adrenal insufficiency — Chronic AI may develop in approximately half of patients treated with oral glucocorticoids [8]. The risk varies according to multiple factors (see 'Factors impacting risk' below). Symptoms of glucocorticoid-induced AI typically manifest when patients taper below physiologic doses of glucocorticoids. However, symptoms that occur during times of physiologic stress (eg serious infection, trauma) are indicative of AI even in patients on supraphysiologic glucocorticoid doses; in such cases, HPA axis suppression renders patients unable to appropriately increase cortisol requirements in response to increased glucocorticoid requirements.

Symptoms of AI are variable. In a case series of more than 200,000 adults treated with ≥3 months of oral glucocorticoids, the most common symptoms suggestive of AI after glucocorticoid discontinuation included hypotension, gastrointestinal symptoms (eg, nausea), hypoglycemia, and hyponatremia due to reduced free water clearance [9]. Other symptoms of chronic central AI may include fatigue, depression, and joint or muscle pain. Management of chronic AI is discussed elsewhere. (See 'Management of complications' below and "Clinical manifestations of adrenal insufficiency in adults", section on 'Central adrenal insufficiency' and "Clinical manifestations and diagnosis of adrenal insufficiency in children", section on 'Central adrenal insufficiency'.)

Adrenal crisis — Though rare, patients with glucocorticoid-induced AI can develop adrenal crisis [10,11]. Adrenal crisis is the most severe clinical presentation of AI, presenting with shock (hypotension, tachycardia) in addition to symptoms of AI [12] (see 'Chronic adrenal insufficiency' above). It is a life-threatening emergency requiring immediate management and prompt administration of stress-dose glucocorticoids and intravenous (IV) fluids as discussed elsewhere. (See "Treatment of adrenal insufficiency in adults", section on 'Adrenal crisis' and "Treatment of adrenal insufficiency in children", section on 'Adrenal crisis'.)

ASSESSING RISK OF ADRENAL INSUFFICIENCY BEFORE TAPERING GLUCOCORTICOIDS — 

Clinicians must determine whether patients are at risk of developing adrenal insufficiency (AI) prior to tapering glucocorticoids. Our general approach in adults is outlined in the algorithm and described in more detail below (algorithm 1). We use a similar approach in children.

Factors impacting risk — All patients taking systemic glucocorticoids have some degree of risk of developing AI, regardless of the formulation; the risk is influenced by many factors, as discussed in more detail below. The most important factor is cumulative glucocorticoid exposure, which is determined by both the dose and duration of therapy. However, glucocorticoid-induced AI has been reported in patients treated with relatively short courses of very low doses (ie, <5 mg/day of prednisone for four weeks) [13,14], and several studies have shown that neither the dose nor the duration of glucocorticoid treatment are reliable indicators of probable AI [15,16]. (See "Major adverse effects of systemic glucocorticoids", section on 'Glucocorticoid dose and duration'.)

●Cumulative glucocorticoid exposure – The risk of glucocorticoid-induced AI is primarily related to overall exposure to glucocorticoids, as determined by glucocorticoid dose and duration of therapy. However, data are lacking regarding a precise threshold beyond which the risk of glucocorticoid-induced AI will occur in most patients [13,17]. Reported risk factors for developing glucocorticoid-induced AI are described in detail below:

•Supraphysiologic glucocorticoid dose – Most treatment regimens utilize glucocorticoid doses that are "supraphysiologic," or in excess of adrenal cortisol production. Patients treated with supraphysiologic glucocorticoid regimens are at higher risk of developing AI (table 4). To determine if an individual patient's dose is supraphysiologic, the clinician must determine the potency of the glucocorticoid formulation (table 5). This is because a lower dose of a more potent formulation will be equivalent to a higher dose of a less potent one.

The importance of glucocorticoid dose was illustrated in a meta-analysis that examined the prevalence and risk factors for AI in adult patients being treated with glucocorticoids; AI occurred in 22 percent of those treated with higher doses of glucocorticoids (ie, >20 mg/day oral/enteral prednisolone/prednisone), compared with 2 percent of patients treated with lower doses (ie, <10 mg/day oral/enteral prednisolone/prednisone) [8].

•Longer duration of therapy – Those taking prolonged courses of glucocorticoids (ie, longer than three to four weeks) are also at higher risk of developing AI. As an example, in a cohort study of adult patients with asthma treated with glucocorticoids, AI occurred in 27 percent of those treated with prednisolone/prednisone for >28 days, compared with only 1.4 percent of patients treated for <28 days [8].

●Glucocorticoid formulation – Systemic forms of glucocorticoids (oral, intravenous [IV]) are much more likely to cause AI than topical or inhaled formulations, particularly when given intravenously. Additionally, duration of action varies among systemic formulations, as illustrated in the table (table 5); AI is more likely to develop in patients using longer-lasting formulations (eg, dexamethasone).

Topical or inhaled glucocorticoid preparations are less likely to cause hypothalamic-pituitary-adrenal (HPA) axis suppression since systemic absorption is limited. However, cases of AI have been reported in patients treated with glucocorticoid nasal sprays, skin creams, intraarticular injections, and inhaled glucocorticoids [8,18]. Therefore, using systemic glucocorticoids together with these other formulations may increase the overall glucocorticoid exposure and the subsequent risk of AI. The approach to discontinuation of topical and inhaled glucocorticoids is discussed elsewhere. (See "Major side effects of inhaled glucocorticoids", section on 'Systemic adverse effects' and "Topical corticosteroids: Use and adverse effects" and "Major adverse effects of systemic glucocorticoids".)

●Other factors – Factors that have been associated with a higher risk of iatrogenic AI include overweight and obesity, older age, and concurrent use of strong cytochrome P450 3A4 inhibitors (table 6). In addition, the development of clinical signs or symptoms suggestive of iatrogenic Cushing syndrome (table 7) (eg, progressive weight gain, facial rounding, striae) is often considered sufficient to make a presumptive diagnosis of AI; however, there is a lack of data describing the true incidence of AI in such patients. (See "Epidemiology and clinical manifestations of Cushing syndrome".)

Approach based on duration of glucocorticoid therapy — Since most patients are treated with glucocorticoids at supraphysiologic doses, we use the duration of therapy as the primary determinant of AI risk. This approach is generally consistent with the joint guidelines on the diagnosis and treatment of glucocorticoid-induced AI issued by the European Society of Endocrinology and the Endocrine Society in 2024 [4].

≤3 weeks of therapy (lower risk of adrenal insufficiency) — Patients are generally considered to have a lower risk of AI when they have used systemic glucocorticoids for <3 to 4 weeks at any dose. Testing for AI is generally not indicated in such cases, and patients can stop glucocorticoids without a taper (see 'No taper for most at lower risk of adrenal insufficiency' below). Although it is uncommon for patients treated with short-term glucocorticoids to develop AI, it has been reported. In a study of 481 patients with cancer treated with supraphysiologic dexamethasone for three to four days every two to four weeks to treat nausea associated with chemotherapy, 15 percent of patients had evidence of AI based on high-dose adrenocorticotropic hormone (ACTH) stimulation testing at three or six months after glucocorticoid treatment [19]. As a result, we still counsel patients about the symptoms of AI and glucocorticoid withdrawal syndrome and monitor clinically for these complications. (See 'Monitoring' below.)

>3 weeks of therapy (higher risk of adrenal insufficiency) — The risk of AI is high in patients who have received daily supraphysiologic glucocorticoid doses for >3 weeks. As a result, we do not routinely perform testing for AI prior to reducing the glucocorticoid dose for such patients, since iatrogenic AI is expected. In addition to tapering the dose prior to stopping therapy, we counsel patients about symptoms of glucocorticoid withdrawal syndrome, AI, and adrenal crisis. All such patients should receive instruction on the need for stress dose glucocorticoids during illness or trauma. (See 'Additional measures for patients at high risk of adrenal insufficiency' below.)

In some cases, patients may be treated with prolonged courses of glucocorticoids at subphysiologic doses. In such cases, we also taper glucocorticoids gradually over time prior to stopping. While the risk of AI is lower in this group, patients may have a nonphysiologic dependence on glucocorticoids. Because data on the risk of AI are limited in this population, some experts test for AI to guide management, while others only test if there is clinical suspicion for glucocorticoid-induced AI.

When is pre-taper testing for adrenal insufficiency needed? — Most patients taking systemic glucocorticoids do not require routine laboratory testing for glucocorticoid-induced AI prior to tapering the dose. This is because we assume that patients treated with supraphysiologic doses of glucocorticoid for >3 weeks have AI and treat accordingly (see 'Regimens in patients at risk of disease recurrence (most common)' below). However, testing for AI may be helpful in the following scenarios:

●Severe glucocorticoid-related adverse effects – Testing may be useful to assess for AI in patients with an intermediate to high risk of glucocorticoid-induced AI who are experiencing severe glucocorticoid-related adverse effects. If testing establishes normal HPA axis function, then glucocorticoids can be immediately discontinued without concern for the development of AI. (See 'Indications for discontinuing glucocorticoids' above.)

●Unknown duration of glucocorticoid treatment – Patients at indeterminate risk of AI (eg, when the duration of glucocorticoid treatment is unknown) may benefit from testing for AI to guide further management of glucocorticoid regimen. This is particularly true for individuals undergoing nonemergency surgery, who require stress-dose glucocorticoids if they have AI. (See "The management of the surgical patient taking glucocorticoids", section on 'Evaluation of HPA axis suppression'.)

However, pre-taper testing for AI may be performed in any patient in whom there is concern for HPA axis suppression. The details of how to perform this testing are discussed below. (See 'Monitoring' below.)

SELECTING A TAPER REGIMEN — 

We select a taper regimen based primarily on the risks of adrenal insufficiency (AI) and disease recurrence, as outlined in the algorithms for adults (algorithm 1) and children (algorithm 2)

No taper for most at lower risk of adrenal insufficiency — Most patients with a lower risk of AI (eg, those who have been treated with ≤3 weeks of glucocorticoid therapy) can discontinue glucocorticoids without a taper, provided that the underlying condition requiring glucocorticoids is unlikely to recur with rapid dose reduction (see 'Factors impacting risk' above). However, a taper may still be required for certain conditions (eg, gout flare, acute interstitial pneumonia) or other patient factors (eg, critical illness, history of glucocorticoid withdrawal symptoms). In such cases, the taper is similar to the approach used for patients with a higher risk of AI, including adapting the regimen based on patient needs. (See 'Taper for those at higher risk of adrenal insufficiency' below.)

While discontinuing therapy, we monitor all patients for glucocorticoid withdrawal symptoms and, if applicable, a recurrence of the underlying disease. (See 'Monitoring' below.)

Taper for those at higher risk of adrenal insufficiency — For most patients with a higher risk of AI (eg, those treated with >3 weeks of therapy), we taper the glucocorticoid dose before stopping and monitor clinically for symptoms of AI. Many factors discussed in detail below can impact the pace of taper, most importantly whether there is a reason to taper gradually over time (eg, patients who are taking glucocorticoids to treat a chronic condition) or whether a patient can immediately drop to a physiologic dose (eg, patients who no longer have indications to take glucocorticoids).

Choice of medication for taper — We generally continue using the original glucocorticoid formulation for the taper (ie, a patient treated with prednisone for nephrotic syndrome would be given a prednisone taper). However, long-acting formulations (eg, dexamethasone) may limit hypothalamic-pituitary-adrenal (HPA) recovery [17]. In such cases, we consider converting to a glucocorticoid with a shorter duration of action (eg, hydrocortisone), unless this is not possible based on the primary indication for glucocorticoid therapy (eg, when using dexamethasone for treatment of central nervous system [CNS] disease due to superior CNS penetration). For younger children taking dexamethasone, when possible, we convert to hydrocortisone while tapering to facilitate titration at very low doses and minimize growth suppression. This approach is similar to that used for patients who have established AI. (See "Treatment of adrenal insufficiency in adults", section on 'Hydrocortisone (preferred glucocorticoid)' and "Treatment of adrenal insufficiency in children", section on 'Choice of glucocorticoid'.)

Factors impacting the pace of taper — Several factors may impact the optimal tapering regimen, including the following:

●Reason for glucocorticoid treatment and discontinuation (most important) – The underlying condition and reason for stopping glucocorticoid treatment will both impact the pace of glucocorticoid dose reduction. Some patients may no longer have an indication for glucocorticoids and may be able to immediately decrease to a physiologic dose before tapering further. Similarly, patients experiencing serious glucocorticoid-related adverse effects may require a rapid taper. By contrast, patients with life-threatening rheumatic diseases may benefit from a gradual taper while watching for a recurrence of symptoms.

●Coadministration of other immunosuppressive therapies – Patients who are also taking glucocorticoid-sparing therapies to treat the underlying condition may be able to taper from higher doses more rapidly. As examples, patients with granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA) who are taking avacopan appear to tolerate a more rapid taper of daily glucocorticoids. If a patient is at higher risk of AI, precautions noted above are needed when approaching subphysiologic dosages during a taper from supraphysiologic ones. (See "Granulomatosis with polyangiitis and microscopic polyangiitis: Induction and maintenance therapy", section on 'Glucocorticoid dosing and taper'.)

In addition, several studies suggest that patients with certain types of lupus nephritis who receive pulse-dose glucocorticoids along with other immunosuppressive medications may be able to tolerate a faster taper of the dose of daily glucocorticoids. (See "Lupus nephritis: Initial and subsequent therapy for focal or diffuse lupus nephritis", section on 'Glucocorticoid dosing and taper'.)

●Glucocorticoid dose and formulation – This topic focuses on prednisone tablets, although there are other types of glucocorticoids (eg, methylprednisolone, hydrocortisone) and various preparations (eg, oral liquids and intravenous injections). The available tablet sizes and liquid concentrations can affect the optimal increments at which the dose may easily be tapered.

Differences in pharmacokinetics (eg, drug absorption, metabolism, and clearance) of various glucocorticoids could theoretically affect the optimal taper. Fortunately, most commercially available prednisone and prednisolone preparations appear to be bioequivalent. In addition, despite a trend toward dose-dependent kinetics where larger doses are cleared more rapidly, the effect is relatively small and usually not of great clinical importance [20-24]. The pharmacology of glucocorticoids is discussed in detail elsewhere. (See "Overview of the pharmacologic use of glucocorticoids".)

●Historical response to glucocorticoids – Some patients require multiple courses of systemic glucocorticoids over time. In such cases, the patient's prior tolerance of therapy and experience with tapers (eg, symptoms of glucocorticoid withdrawal at certain doses, disease flares with certain tapers) can help inform subsequent withdrawal regimens.

●General clinical status – Other factors including patient frailty and concomitant illness (especially the presence of comorbid conditions worsened by systemic glucocorticoids, such as infection, diabetes, and hypertension) may impact the pace of glucocorticoid withdrawal [17].

Regimens in patients at risk of disease recurrence (most common) — For most conditions (eg, rheumatic diseases), providers taper doses of glucocorticoids over time to ensure that the underlying condition remains controlled. Many factors influence the pace of the glucocorticoid taper, most importantly the type and severity of the underlying condition and the use of other immunosuppressive medications (see 'Factors impacting the pace of taper' above). However, we generally reduce the total daily dose by a certain percentage every one to four weeks. The percent reduction is often greater when patients are taking higher doses (eg, reducing by 25 percent when taking 80 mg/day of prednisone) and smaller when patients are taking lower doses (eg, reducing by 10 to 20 percent after reaching a dose of 40 mg/day of prednisone). More rapid tapers may be possible in some scenarios (eg, patients who are also receiving a glucocorticoid-sparing agent). Tapering regimens for children are based on weight or body surface area (BSA; eg, 2 to 1.5 mg/kg/day or 10 to 8 mg/m²/day). A calculator is available for determining BSA from weight and height (calculator 1).

Preferred and alternative strategies for tapering are discussed here:

●Gradual daily dose reduction (preferred) – When tapering to a physiologic dose of glucocorticoids, we prefer to gradually reduce the daily dose by 10 to 20 percent every few weeks while accommodating convenience and individual patient response. The table provides an example of this type of taper in adults (table 8).

The approach is similar in children. One important difference is the preferred use of hydrocortisone due to its short half-life, which minimizes glucocorticoid-induced growth suppression, and lower potency, which allows for easier titration at very low doses. This algorithm offers a general approach in children who have not completed linear growth (algorithm 2). Adolescents who have completed linear growth (usually those at least four years since onset of puberty) may often be treated using adult regimens. However, an approach based on BSA may be needed in adolescents in whom height or weight is significantly below average.

When patients take glucocorticoids in two divided doses per day, some providers prioritize tapering the second dose to better mimic natural diurnal cortisol patterns and eventually consolidate to once-daily dosing. As an example, a patient taking 10 mg of prednisone twice per day might continue taking 10 mg in the morning while tapering the afternoon dose by 2.5 mg every few weeks until stopping the afternoon dose.

●Alternate-day regimen — In some patients (eg, patients with rheumatoid arthritis [RA] on long-term therapy with <10 mg/day of prednisone or the equivalent), we use an alternate-day approach. Specifically, after reaching a low to moderate daily dose, the patient alternates every day between taking the same dose and a reduced dose. The table illustrates an example of this approach for an adult patient (table 9).

A report in patients with rheumatic disease suggested a different version of an alternate-day regimen, where the daily dose of glucocorticoids was doubled and decreased in frequency to every other day prior to tapering [25]. As with the regular approach to alternate-day tapering, this may result in suboptimal disease control on days without any glucocorticoid administration.

We are not aware of any evidence-based data relating to glucocorticoid tapering on an alternate-day regimen. Although this regimen is generally effective in most rheumatic diseases in the authors' experience, patients with RA may experience worsening symptoms on alternate days when they are receiving lower doses of glucocorticoids. (See "Use of glucocorticoids in the treatment of rheumatoid arthritis".)

●Tapering based on serial serum cortisol levels – Some experts use laboratory testing of cortisol to help guide the taper. Specifically, once patients reach a physiologic glucocorticoid dose, they can temporarily hold their dose for 24 hours and check an early-morning serum cortisol level. The level can be used to determine the likelihood of AI and therefore whether glucocorticoids can be stopped or else continued with repeat laboratory testing in the future (see 'Monitoring' below):

•Early-morning serum cortisol ≥10 mcg/dL – A cortisol level of 12 to 14 mcg/dL between 7 and 9 AM has been shown to be specific for ruling out AI [26,27]. Most clinicians feel that AI is also unlikely if the early-morning serum cortisol is ≥10 mcg/dL [17] and will proceed with tapering or stopping glucocorticoids [4]. However, results should be interpreted using age- and laboratory-specific reference ranges.

•Early-morning serum cortisol <10 mcg/dL – Low cortisol levels (<5 mcg/dL) on baseline assessment in patients with glucocorticoid-induced AI have been associated with a longer time until recovery from HPA axis suppression [28,29]. Levels <10 mcg/dL suggest a moderate likelihood of AI. If there is a moderate to high risk of AI based on the cortisol level, we continue glucocorticoids at physiologic doses. For patients with a moderate risk, we repeat testing after a few weeks; otherwise, we retest high-risk patients after two to three months [4].

Alternatively, there are also reports of using more frequent cortisol levels to guide the taper. As an example, a report from 1976 used serial plasma cortisol measurements every two to four weeks to gauge the pace of hydrocortisone withdrawal [30]; however, such frequent monitoring is likely unnecessary and largely impractical.

Our preferred approach of gradual daily dose reduction assumes that this method is easier for most patients to understand compared with alternate-day regimens, that repeated morning cortisol determinations are too impractical for routine use, and that patients should be monitored clinically for signs and symptoms of AI. In the authors' experience, this approach generally prevents symptoms of AI. In addition, there is ample evidence of how this type of taper can be used for a wide range of conditions, which is presented in the disease-specific treatment topics.

The approaches outlined above are generally consistent with those described in the European Society of Endocrinology and the Endocrine Society 2024 guidelines for glucocorticoid-induced AI [4]. There are very limited published, controlled, trial-derived data that specifically address the issue of discontinuing long-term moderate- or high-dose glucocorticoids in chronic rheumatic or other inflammatory disorders. Such studies are discussed in detail in the respective disease-specific treatment topics, but they generally address the effect of glucocorticoid tapering regimens on the activity of the underlying disease and underscore the importance of tapering based on disease activity and use of other immunosuppressive therapies. While there are few data to demonstrate that tapering prolonged courses of subphysiologically dosed glucocorticoids is advantageous from the perspective of preventing AI, this is commonly done to minimize symptoms that may be associated with tapering glucocorticoids even at very low doses.

In addition, data are very limited to guide the choice between tapering subphysiologic glucocorticoid doses based on monitoring symptoms alone versus symptoms plus laboratory testing. Some experts advocate for testing since symptoms of AI may not be present in all patients with AI. As an example, a meta-analysis examined the risk of AI in adults treated with glucocorticoids across a total of 521 patients; although only 10 reported symptoms, AI was identified based on biochemical testing in 98 patients [8]. However, there are limited data to guide the choice between tapering based on monitoring symptoms alone versus symptoms plus laboratory testing in specific patient groups. In addition, potential barriers to testing include inconvenience for the patient and health care costs. If clinicians monitor symptoms alone without laboratory testing, it is important to recognize that a lack of symptoms does not preclude the possibility of AI.

Regimens in patients at low risk of disease recurrence (uncommon) — Less commonly, glucocorticoids may be used to treat the acute phase of an illness that resolves or is of less clinical concern. As an example, patients treated with prolonged glucocorticoids in the setting of refractory septic shock will be at low risk of shock recurrence once sepsis resolves. If the underlying condition does not require a more gradual approach to tapering glucocorticoids, it is reasonable to reduce directly to a physiologic replacement dose as a first step (table 1). The subsequent taper of subphysiologic doses is the same as that outlined above for patients who are at risk of disease recurrence and includes close monitoring for symptoms of AI. (See 'Regimens in patients at risk of disease recurrence (most common)' above.)

ADDITIONAL MEASURES FOR PATIENTS AT HIGH RISK OF ADRENAL INSUFFICIENCY — 

Other aspects of management for patients with a high risk of developing glucocorticoid-induced adrenal insufficiency (AI) include the following:

●Reviewing the symptoms of AI and glucocorticoid withdrawal syndrome, as well as potential consequences (eg, adrenal crisis)

●Discussing ways to alert providers of their condition (eg, medical alert bracelets or necklaces, emergency medical information cards)

●Planning for the administration of stress-dose glucocorticoids (ie, oral and intramuscular rescue medications in the setting of acute illness)

We advise patients to seek immediate medical attention during periods of severe illness or if they develop symptoms suggestive of AI or adrenal crisis (eg, vomiting, altered mental status) at any time. This approach is consistent with international clinical practice guidelines for adults with AI [17].

The treatment of presumed AI in the setting of acute illness is discussed in detail elsewhere. (See "Treatment of adrenal insufficiency in children", section on 'Stress conditions' and "Treatment of adrenal insufficiency in adults", section on 'Acute illness'.)

MONITORING — 

All patients who are tapering chronic glucocorticoids are monitored for evidence of glucocorticoid withdrawal or adrenal insufficiency (AI). Patients who were previously using glucocorticoids to manage a chronic condition are also monitored for disease recurrence. Monitoring consists of regular assessments, which may include the following:

●Symptoms of cortisol deficiency – Symptoms of glucocorticoid withdrawal syndrome and AI are similar (eg, fatigue, arthralgia, myalgia), as outlined in the table (table 3) and described above. (See 'Clinical manifestations' above.)

●Laboratory testing for AI – Generally, we reserve laboratory testing for AI for patients experiencing extreme difficulty tapering glucocorticoids or symptoms suspicious for AI shortly after discontinuing therapy. Alternatively, some providers screen for AI with laboratory testing once patients reach a physiologic dose of glucocorticoids (table 4) (see 'Regimens in patients at risk of disease recurrence (most common)' above). Several forms of laboratory testing for AI are available, including first-morning cortisol and corticotropin (adrenocorticotropic hormone [ACTH]) stimulation (also known as a cosyntropin test). We typically start with a morning cortisol level and, if needed, proceed to an ACTH stimulation test:

•Morning cortisol – Once the glucocorticoid dose is at physiologic replacement, adrenal function may be assessed by measuring an early-morning (8 to 9 AM) cortisol level. This test should be performed 24 hours after the most recent glucocorticoid dose to reflect endogenous cortisol production rather than exogenous drug. Interpretation of testing is the same as when it is used to guide glucocorticoid tapering (see 'Regimens in patients at risk of disease recurrence (most common)' above). More information on cortisol testing in children and adults is provided separately. (See "Clinical manifestations and diagnosis of adrenal insufficiency in children", section on 'Basal (unstimulated) cortisol' and "Diagnosis of adrenal insufficiency in adults", section on 'Basal serum cortisol testing'.)

•ACTH stimulation testing – If an early-morning cortisol level suggests an indeterminate level of risk for AI (ie, 5 to <10 mcg/dL) and there is an urgent need to determine whether a patient has AI (eg, upcoming nonemergency surgery), patients may benefit from ACTH stimulation testing. Details of how to conduct this testing are discussed elsewhere. (See "Clinical manifestations and diagnosis of adrenal insufficiency in children", section on 'Adrenocorticotropic hormone stimulation test' and "Diagnosis of adrenal insufficiency in adults", section on 'Standard high-dose ACTH stimulation test'.)

●Disease recurrence – Patients must be monitored carefully while tapering systemic glucocorticoids for recurrence of the underlying disease. This can typically be distinguished from glucocorticoid withdrawal and/or AI based on a combination of symptoms, examination findings, and laboratory testing.

MANAGEMENT OF COMPLICATIONS — 

Our approach to managing complications of glucocorticoid withdrawal is outlined in the algorithm (algorithm 3) and discussed in more detail below.

●Glucocorticoid withdrawal syndrome – If patients have mild symptoms of glucocorticoid withdrawal (see 'Glucocorticoid withdrawal syndrome' above), we counsel patients to continue the same reduced dose of systemic glucocorticoids for 7 to 10 days and use a nonsteroidal antiinflammatory drug (NSAID) or acetaminophen for symptomatic relief. If symptoms persist, we increase the glucocorticoid (typically prednisone) dose by 10 to 15 percent (to the next convenient mg tablet regimen) and maintain that dose for two to four weeks. If the symptoms resolve, we resume tapering the dose at a more gradual pace (eg, using two to four weeks between decrements rather than one to two weeks) and consider an alternate-day approach as outlined above. (See 'Regimens in patients at risk of disease recurrence (most common)' above.)

Patients with glucocorticoid withdrawal syndrome may require increased ("stress-dose") glucocorticoids during times of physiologic stress (eg, severe infection, trauma, surgery). This is because such patients are at risk of hypothalamic-pituitary-adrenal (HPA) axis suppression due to prolonged exposure to exogenous glucocorticoids. (See "Treatment of adrenal insufficiency in adults", section on 'Circumstances requiring glucocorticoid dose adjustment' and "Treatment of adrenal insufficiency in children".)

●Adrenal insufficiency – If patients develop AI based on symptoms or laboratory testing, we pause the glucocorticoid taper and increase to the last tolerated dose (see 'Chronic adrenal insufficiency' above). Once symptoms have resolved, some providers send laboratory testing to confirm adrenal recovery prior to attempting a repeat taper. Alternatively, some providers resume tapering and monitor for a recurrence of symptoms.

Patients with glucocorticoid-induced AI require longer tapers (ie, decreasing the frequency and/or degree of taper). The time to recovery of adrenal function varies significantly between patients. Studies assessing recovery from HPA axis suppression in adults treated with glucocorticoid therapy have reported a median time to recovery of adrenal function ranging from 16 to 24 months [28,31].

These patients also require increased glucocorticoid doses ("stress-dose" glucocorticoids) during illness, injury, or surgery to prevent adrenal crisis. Stress-dose glucocorticoids and other aspects of the management of AI in adults and children are described elsewhere. (See "Treatment of adrenal insufficiency in adults", section on 'Chronic adrenal insufficiency' and "Treatment of adrenal insufficiency in children".)

●Adrenal crisis – If there is concern for adrenal crisis (eg, hypotension, tachycardia in the setting of a trigger such as infection) while tapering glucocorticoids, patients should receive empiric stress-dose glucocorticoids and be evaluated urgently (see 'Chronic adrenal insufficiency' above and 'Adrenal crisis' above). The clinical manifestations and treatment of adrenal crisis in children and adults are discussed in detail elsewhere:

•(See "Clinical manifestations of adrenal insufficiency in adults", section on 'Adrenal crisis'.)

•(See "Treatment of adrenal insufficiency in adults", section on 'Adrenal crisis'.)

•(See "Clinical manifestations and diagnosis of adrenal insufficiency in children", section on 'Adrenal crisis'.)

•(See "Treatment of adrenal insufficiency in children", section on 'Adrenal crisis'.)

●Disease recurrence – If patients develop a disease flare while tapering, we increase the dose of glucocorticoid as is appropriate for the specific condition and degree of disease flare and, when possible, adjust glucocorticoid-sparing therapies. Once the flare subsides, we often adopt a less aggressive taper (eg, slower rate, lower decrements).

WHEN TO REFER — 

Most patients who are discontinuing systemic glucocorticoids can be managed by the health care provider who is treating the underlying condition. However, a referral to or consultation with an endocrinologist is appropriate for patients who are unable to wean off of physiologic doses of glucocorticoids after a year and for those with a history of adrenal crisis [4].

SUMMARY AND RECOMMENDATIONS

●Indications for discontinuing glucocorticoids – To avoid adverse effects, glucocorticoids are generally discontinued as soon as possible after controlling the underlying disease, determining that an adequate trial was ineffective, or observing serious or uncontrollable adverse effects. (See 'Indications for discontinuing glucocorticoids' above.)

●Glucocorticoid-induced adrenal insufficiency (AI) – Glucocorticoid-induced AI refers to isolated cortisol deficiency resulting from the suppression of the hypothalamic-pituitary-adrenal (HPA) axis. It is associated with symptoms ranging in severity from mild (eg, glucocorticoid withdrawal syndrome) to severe (eg, adrenal crisis) (table 3). (See 'Overview of glucocorticoid-induced adrenal insufficiency' above.)

●Assessing the risk of AI before tapering glucocorticoids – AI risk increases in proportion to cumulative glucocorticoid exposure, which is determined by both the dose and duration of therapy. Since most patients are treated with glucocorticoids at supraphysiologic doses, we use the duration of therapy as the primary determinant of AI risk (algorithm 1). (See 'Assessing risk of adrenal insufficiency before tapering glucocorticoids' above.)

•≤3 weeks of therapy – These patients are generally considered to have a lower risk of AI; testing for AI is generally not indicated. Such patients can usually stop glucocorticoids without a taper unless a taper is needed based on disease-specific management considerations or clinical instability. We monitor clinically for symptoms of glucocorticoid-induced AI. (See '≤3 weeks of therapy (lower risk of adrenal insufficiency)' above.)

•>3 weeks of therapy – These patients have a high risk of HPA axis suppression. Therefore, we assume that AI is likely to occur, and we gradually reduce the glucocorticoid dose over time prior to stopping therapy. We counsel patients about symptoms of glucocorticoid-induced AI, the risk of adrenal crisis, and the indications for and administration of stress-dose glucocorticoids. (See '>3 weeks of therapy (higher risk of adrenal insufficiency)' above.)

While there are few data to demonstrate that tapering prolonged courses of subphysiologically dosed glucocorticoids is advantageous from the perspective of preventing AI, this is commonly done to minimize symptoms associated with tapering glucocorticoids.

●Selecting a taper regimen – When a taper is needed, we select a regimen based primarily on the risks of AI and disease recurrence.

•Choice of medication for taper – In adults, we generally use the original glucocorticoid formulation for the taper, unless patients are taking longer-acting formulations (eg, dexamethasone), in which case we convert to short-acting glucocorticoids (eg, hydrocortisone) when possible. In children, we attempt to convert to hydrocortisone to facilitate administration of lower doses (algorithm 2). (See 'Choice of medication for taper' above.)

•Regimens in patients at risk of disease recurrence (most common) – For many conditions (eg, rheumatic diseases), providers taper glucocorticoids over time to maintain disease control. While the taper should be individualized for the specific patient and condition, we generally reduce the dose by a certain percentage every one to four weeks. The percent reduction is typically greater at higher doses and smaller at lower doses. Tapering regimens for children may be based on weight or body surface area (BSA). (See 'Regimens in patients at risk of disease recurrence (most common)' above.)

While most providers gradually reduce the daily dose (table 8), an alternate-day regimen can also be used after reaching a moderate daily dose (table 9). Some providers incorporate regular laboratory testing for AI to guide the taper.

•Regimens in patients at low risk of disease recurrence (uncommon) – If the condition that was being treated with glucocorticoids resolves or is no longer of clinical concern (eg, resolved sepsis), we often reduce directly to a physiologic replacement dose before gradually tapering subphysiologic doses (table 1). (See 'Regimens in patients at low risk of disease recurrence (uncommon)' above.)

●Monitoring – We monitor all patients for symptoms of glucocorticoid withdrawal or AI and, when applicable, disease recurrence. We generally reserve laboratory testing for AI for patients experiencing extreme difficulty tapering glucocorticoids or symptoms suspicious for AI shortly after discontinuing therapy. Some experts obtain laboratory testing for AI once patients reach a physiologic dose of glucocorticoids or if there is an urgent need to determine whether stress-dose glucocorticoids are needed (eg, imminent surgery). (See 'Monitoring' above.)

●Management of complications – Mild glucocorticoid withdrawal symptoms can often be managed with nonsteroidal antiinflammatory drugs (NSAIDs) or acetaminophen. Severely symptomatic patients and those with AI are typically treated with a temporary increase in the glucocorticoid dose followed by a more gradual taper. Patients with suspected adrenal crisis (eg, unexplained hypotension, tachycardia in the setting of physiologic stress) should receive empiric stress-dose glucocorticoids and an urgent medical evaluation. (See 'Management of complications' above.)