Glucocorticoid withdrawal

INTRODUCTION — Chronic glucocorticoid therapy is used in the treatment of a variety of disorders because of its potent antiinflammatory effects and, occasionally, because it is thought to have immunosuppressive activity [1]. Among the rheumatic diseases in which glucocorticoids are often used are rheumatoid arthritis, large- and small-vessel vasculitis, systemic lupus erythematosus, polymyalgia rheumatica, and, in some cases, the arthritis associated with inflammatory bowel disease [1].

Despite its efficacy, steroid-induced side effects generally require tapering of the drug as soon as the disease being treated is under control. Tapering must be done carefully to avoid both recurrent activity of the underlying disease and possible cortisol deficiency resulting from hypothalamic-pituitary-adrenal axis (HPA) suppression during the period of steroid therapy. (See "Pharmacologic use of glucocorticoids", section on 'HPA axis suppression'.)

This topic discusses the major issues related to tapering, the regimen(s) we use in most patients, and other glucocorticoid tapering regimens that have been reported in the literature.

The clinical manifestations, diagnosis, and treatment of adrenal insufficiency are presented separately. (See "Clinical manifestations of adrenal insufficiency in adults" and "Treatment of adrenal insufficiency in adults".)

INDICATIONS FOR WITHDRAWING GLUCOCORTICOIDS — It is helpful to briefly review the indications for glucocorticoid withdrawal before discussing the different glucocorticoid withdrawal regimens. These indications include the following:

●When the maximum desired therapeutic benefit has been obtained

●When inadequate therapeutic benefit has been obtained after an adequate trial

●When side effects, such as osteoporosis or hypertension, become serious or uncontrollable with medication

In addition, there are two complications that require immediate cessation of glucocorticoids or a significant rapid reduction, rather than tapering:

●Steroid-induced acute psychosis that is unresponsive to antipsychotic medications

●Herpesvirus-induced corneal ulceration, which can rapidly lead to perforation of the cornea and possibly permanent blindness

If immediate cessation is not possible (eg, because of a critical clinical need), use of the lowest necessary dose, then discontinuing steroids as soon as possible, is strongly advised.

GLUCOCORTICOID PREPARATIONS — Prednisone will be the agent discussed, although other glucocorticoid preparations are available (table 1) (see "Pharmacologic use of glucocorticoids"). These preparations are available in various strengths (eg, 1, 2.5, 5, 10, 20 mg for prednisone) and in various formulations (tablet, intravenous preparations, intramuscular preparations, and rectal suppositories). Differences in the absorption or metabolism of these various strengths and formulations could affect the ability to taper steroids. Fortunately, most commercially available prednisone and prednisolone preparations appear to be bioequivalent. This can be illustrated by the following observations:

●In vivo studies using healthy males in a crossover study revealed no statistical difference in any pharmacokinetic parameter with five different oral prednisone preparations [2].

●Rectal and oral absorption of methylprednisolone are equivalent, with the relative bioavailability of oral to rectal administration being 90 percent [3].

●The systemic bioavailability of prednisone is nearly equivalent to that of prednisolone (0.77 to 0.80) [4]. Prednisone itself is biologically inactive, but it is rapidly converted to the active form prednisolone. However, patients with severe liver disease may have difficulty converting prednisone to prednisolone; in such patients, it is possible that one might not get the same effect from prednisone as from prednisolone. In addition, certain drug interactions can affect the metabolism and bioavailability of prednisone. As an example, barbiturates, phenytoin, or rifampin attenuate the biologic effects of glucocorticoids by enhancing their metabolism (table 2).

Steroid pharmacokinetics — Tapering regimens could potentially be influenced if drug distribution changed at varying prednisone doses. Although there is a trend toward dose-dependent kinetics, with larger doses being cleared more rapidly, the effect is relatively small and usually not of great clinical importance [5-7].

Other factors can influence pharmacokinetics. One interesting study examined 54 patients of varying ages who were given oral and intravenous methylprednisolone and prednisolone [8]. Eleven patients (20 percent) demonstrated unusual kinetics. Their drug clearance was approximately twice that of the rest of the population without an identifiable cause. Other findings included incomplete absorption of glucocorticoids in four patients and an inverse correlation (r = -0.88) between prednisolone clearance and age, which means that a given dose may have a greater effect in older persons. This relation to age has been confirmed in other studies [9]. In addition, prednisolone clearance is also slower in African Americans compared with White Americans [10].

With this degree of interpatient variability in kinetics, it is conceivable that some patients may show greater withdrawal symptoms than others. This may occur despite the known difference between plasma glucocorticoid kinetics and biologic activity.

HYPOTHALAMIC-PITUITARY-ADRENAL AXIS SUPPRESSION — Administration of exogenous glucocorticoids can suppress the hypothalamic-pituitary-adrenal axis (HPA). Abrupt cessation, or too rapid withdrawal, of glucocorticoids in such patients may cause symptoms of adrenal insufficiency. HPA suppression and the clinical manifestations of adrenal insufficiency are presented separately. (See "Pharmacologic use of glucocorticoids", section on 'HPA axis suppression' and "Clinical manifestations of adrenal insufficiency in adults".)

Identifying patients with HPA suppression — The potency, dose, and duration of glucocorticoid use are important but imperfect predictors of the presence of HPA suppression. A clinical decision about a particular patient's risk of having HPA suppression guides subsequent decision making about glucocorticoid tapering. Patients exposed to glucocorticoids can be classified and managed as follows:

●Patients suspected of having HPA suppression who are otherwise candidates for glucocorticoid withdrawal can be cautiously tapered. (See 'Tapering regimens' below.)

●Patients determined to have a low risk of clinically significant HPA suppression may have glucocorticoids weaned as dictated by control of disease activity rather than by concern for adrenal insufficiency.

●Patients for whom the likelihood of HPA suppression is uncertain, but for whom the consequences of developing acute adrenal insufficiency are serious (eg, the patient who is about to undergo major surgery), may benefit from testing of the HPA functional reserve to guide further therapy.

Characteristics of patients who are likely to have HPA suppression and those who are unlikely to, or for whom the likelihood of HPA suppression is uncertain, are discussed in more detail separately (see "Pharmacologic use of glucocorticoids", section on 'HPA axis suppression' and "The management of the surgical patient taking glucocorticoids", section on 'Evaluation of HPA axis suppression'). The following is a brief summary:

HPA suppression likely — The question of the amount and duration of exogenous glucocorticoid that may lead to adrenal suppression is not well answered in the literature because of the heterogeneity of studies and differing results. A systematic review of 73 studies found evidence for adrenal insufficiency across all average and cumulative doses of exogenous glucocorticoid [11]. HPA suppression may occur regardless of whether glucocorticoids are administered in the morning or evening [12].

While it may be reasonable to assume that any patient who has a Cushingoid appearance has HPA suppression, there is no literature to support this assumption. In addition, several studies have shown that neither the dose nor the duration of glucocorticoid treatment nor a random plasma cortisol measurement are reliable indicators of probable adrenal insufficiency [13,14].

Patients with suspected HPA suppression do not need testing to evaluate their HPA function, but should be treated like any patient with secondary adrenal insufficiency, including the wearing of a medical alert bracelet or necklace and carrying an emergency medical information card. We advise patients on glucocorticoids with presumed HPA suppression to seek immediate medical attention for diagnostic and therapeutic intervention if they develop symptoms of adrenal insufficiency. It is reasonable for patients without access to nearby medical care to carry a preloaded 1 mL syringe containing 4 mg of dexamethasone phosphate to inject in emergencies. (See "Pharmacologic use of glucocorticoids", section on 'Complications of chronic use'.)

If withdrawal from glucocorticoids is otherwise indicated, gradual reduction in dose is appropriate. Such patients do not need to be tested for HPA functional reserve unless abrupt discontinuation is being considered or the patient is facing an acute stress such as surgery. Management of the surgical patient taking glucocorticoids is discussed in detail elsewhere. (See "The management of the surgical patient taking glucocorticoids" and "Diagnosis of adrenal insufficiency in adults".)

HPA suppression unlikely — Patients who meet one of the following criteria regarding glucocorticoid use are less likely to have a suppressed HPA and therefore may have glucocorticoids weaned as is appropriate for the underlying disease:

●A patient who has received glucocorticoids for less than three weeks

●Patients treated with alternate-day prednisone at a dose of less than 10 mg (or its equivalent)

Estimation of HPA suppression — Identifying the degree of HPA suppression is not simple clinically. Thus, in practice it is unusual to perform any testing of HPA function prior to beginning the glucocorticoid withdrawal process. However, as noted above, in certain settings (eg, the patient for whom elective surgery is planned) such testing may be warranted. (See "The management of the surgical patient taking glucocorticoids", section on 'Evaluation of HPA axis suppression'.)

The response to administration of synthetic ACTH is the preferred method to assess adrenocortical function. Although the cosyntropin test does not provide information about hypothalamic function, it has the advantage that it can be performed in the office or clinic setting over the course of approximately one-hour. Test results should be available within hours to days thereafter. (See "Diagnosis of adrenal insufficiency in adults".)

Recommendation — Testing for HPA-axis function is appropriate when patients are using <5 mg/day of prednisone and there is difficulty reducing the dose further because of symptoms related to adrenal insufficiency. (See "Diagnosis of adrenal insufficiency in adults", section on 'Low-dose ACTH stimulation test not recommended'.)

OTHER FORMS OF GLUCOCORTICOID DEPENDENCE — Other forms of steroid dependence (beyond symptomatic and biochemical evidence of hypothalamic-pituitary-adrenal axis [HPA] suppression) have been identified which can hinder steroid tapering. These include [15,16]:

●Psychologic dependence on steroids

●Recrudescence of the disease for which the drug was prescribed

●Symptoms of apparent adrenal insufficiency despite normal HPA function and lack of disease recrudescence

TAPERING REGIMENS — There is a paucity of clinical evidence to support any particular regimen of glucocorticoid tapering. Published, controlled, trial-derived data do not specifically address the issue of weaning patients from long-term moderate or high-dose glucocorticoids in chronic rheumatic or other inflammatory disorders. The following illustrative reviews and selected studies have principally addressed the effect of glucocorticoid tapering regimens on the activity of the underlying disease:

●A 2013 systematic review of glucocorticoid tapering regimens for rheumatoid arthritis found six randomized trials and four long-term extension trials with limited data and great variability in the tapering regimens [17]. Glucocorticoid tapering was less successful in patients with longer disease duration. Success rates were 31 to 42 percent in patients with 6.3 and 9.3 years of disease, respectively, whereas success rates in patients with less than two years of disease duration ranged from 78 to 92 percent. Few trials described a specific tapering regimen for glucocorticoid withdrawal. Two randomized trials recommended a variant of tapering by skipping more days between doses but keeping the dose unchanged (eg, 7.5 mg every other day for two weeks, then every third day for two weeks, then discontinued). Another study decreased the prednisone dose by 1 mg every four weeks at doses below 5 mg daily.

●A 2002 systematic review found nine controlled trials with random patient assignment that compared different glucocorticoid tapering regimens [18]. The published reports reviewed dealt almost exclusively (seven of nine trials) with the treatment of asthma or chronic obstructive pulmonary disease (COPD) exacerbations and generally assessed relapse-rates and/or physiologic measures (such as peak expiratory flow rates, forced expiratory volumes, or oxygen saturation) as patients were weaned from systemic to inhaled glucocorticoids.

The authors of the review concluded that there was no significant difference between abrupt, rapid, or slow glucocorticoid tapering regimens for these outcomes in asthma or COPD. The other two trials reviewed were 7- versus 15 week-tapering regimens in Crohn disease and 10- versus 21-week tapering in bone marrow transplant patients with graft versus host disease [19,20]. Again, in these latter two studies there were no clinically significant differences in outcomes between the shorter and longer tapers.

●A trial evaluated 46 children with nephrotic syndrome who were treated with high doses of prednisolone (60 mg/m² per day) [21]. Tapering was performed over eight weeks or seven months. Twenty-nine children in the short-taper group received 60 mg/m2 per day for four weeks, followed by 40 mg/m² per day on three days a week for four weeks. Prednisolone was then discontinued. Seventeen children in the long taper group received 60 mg/m2 per day for four weeks, 60 mg/m² per day on alternate days for four weeks, followed by tapering by 10 mg/m² every other day every four weeks over seven months. The number of patients who relapsed with the nephrotic syndrome within six months after initial taper was significantly higher in the rapid taper group.

RECOMMENDED TAPERING REGIMEN — Short-term glucocorticoid therapy (two weeks or less) can simply be stopped and need not be tapered. Hypothalamic-pituitary-adrenal axis (HPA) suppression due to glucocorticoid use of this duration is unlikely, but can occur [13].

In patients who have taken a glucocorticoid for a longer time, we suggest a regimen which is largely based upon experience and rests upon the following assumptions:

●Factors of age, frailty, concomitant illnesses, dangerousness and likelihood of flare of underlying illness, psychological factors, and duration of previous use of glucocorticoids are taken into account.

●The disease is sufficiently stable so that tapering of the dose is appropriate.

●The patient has received long-term steroid therapy, not recurrent "pulses" as might be used in asthma.

●The observation that HPA suppression is uncommon at prednisone doses below 5 mg/day.

The regimen we recommend also assumes that repeated morning cortisol determinations are too expensive for routine use (see 'Other published tapering regimens' below) and that the appropriate endpoints are the patient's signs and symptoms.

The goal of tapering is to use a rate of change that will prevent both recurrent activity of the underlying disease and symptoms of cortisol deficiency due to persistent HPA suppression. We generally aim for a relatively stable decrement of 10 to 20 percent, while accommodating convenience and individual patient response. The dose is tapered by:

●5 to 10 mg/day every one to two weeks from an initial dose above 40 mg of prednisone or equivalent per day.

●5 mg/day every one to two weeks at prednisone doses between 40 and 20 mg/day.

●2.5 mg/day every two to three weeks at prednisone doses between 20 and 10 mg/day.

●1 mg/day every two to four weeks at prednisone doses between 10 and 5 mg/day.

●0.5 mg/day every two to four weeks at prednisone doses from 5 mg/day down. This can be achieved by alternating daily doses, eg, 5 mg on day one and 4 mg on day two.

This regimen will generally prevent symptoms of cortisol deficiency. At some point, however, many patients with rheumatic diseases complain of recurrent symptoms of the underlying disease. In this setting it may be difficult to distinguish between mild symptoms of glucocorticoid withdrawal (ie, arthralgia and myalgia or "pseudorheumatism") or recrudescence of the underlying rheumatic disease.

If the symptoms are not major, we try to wait 7 to 10 days, and use a nonsteroidal antiinflammatory drug (NSAID) or other analgesic. Resolution of symptoms during this period of time suggests pseudorheumatism. If the symptoms do not subside within this time frame, we increase the 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, the above tapering regimen can be resumed, using two to four weeks between decrements rather than one to two weeks.

Should this modest increase in dose not be sufficient to alleviate symptoms, we double the prednisone dose. The disease flare is allowed to subside and the taper is reinstituted at a slower rate (eg, once monthly) or at smaller decrements (eg, one-half of the original decrement).

It should also be appreciated that incremental change is inappropriate if life-threatening flares occur (as in acute recurrence of lupus nephritis, severe hemolysis, acute polymyositis, or vasculitis). In these settings, a return to the original, highest dose of steroids should be instituted. Tapering which is slowed in rate or decrement can be undertaken after the flare subsides, but specific guidelines become both convoluted and impractical in the latter situations.

Alternate-day regimen — We are not aware of any evidence-based data relating to steroid tapering on an alternate-day regimen. We do, however, use the following alternate-day approach (in which the entire dose is given on the alternate days) in some patients. After the daily regimen has reached 20 to 30 mg of prednisone per day, we decrease the alternate day dose by 5 mg every one to two weeks until the dose is 20 to 30 mg alternating with 10 mg. We then reduce the alternate-day dose by 2.5 mg every one to two weeks until the prednisone dose on the alternate day has fallen to zero. At that point we decrease the remaining drug in the same manner as was suggested for the daily dosing regimen.

Although this regimen is generally effective in most rheumatic diseases, patients with rheumatoid arthritis often do not tolerate alternate-day dosing.

OTHER PUBLISHED TAPERING REGIMENS — Other published glucocorticoid tapering regimens include the following:

●A report in 1976 used plasma cortisol measurements to gauge withdrawal [22]. Patients returned to the clinic at two to four week intervals for morning plasma cortisol measurement. Tapering was done at a rate of 2.5 mg of hydrocortisone/week down to a single morning dose of 10 mg of hydrocortisone (equivalent to 2 mg of prednisone). Steroid therapy could be discontinued when the morning plasma cortisol concentration rose to greater than 10 mcg/dL. Stress doses of steroids might be needed for infections. This approach, however, has not gained much popularity and is generally not used.

●Another report in patients with rheumatic disease suggested either switching to alternate-day therapy or gradually lowering the daily dose [23]. To switch to an alternate-day regimen, the dose was doubled on alternate days and then tapered as described above (see 'Alternate-day regimen' above). This regimen, however, might result in synovial and serosal flare and symptoms on the alternate days.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Side effects of anti-inflammatory and anti-rheumatic drugs".)

SUMMARY AND RECOMMENDATIONS

●Indications for glucocorticoid withdrawal – The indications for glucocorticoid withdrawal include achieving the maximum desired therapeutic benefit or inadequate benefit despite an adequate trial, and the development of serious or uncontrollable side effects from the medication. Immediate cessation or a significant rapid reduction (in those in whom complete cessation could lead to even more serious consequences) as soon as possible is required in patients with steroid-induced acute psychosis and those with herpesvirus-induced corneal ulceration.(See 'Indications for withdrawing glucocorticoids' above.)

●Hypothalamic-pituitary-adrenal axis (HPA) suppression – Administration of exogenous glucocorticoids can suppress the hypothalamic-pituitary-adrenal axis (HPA). Abrupt cessation or too-rapid withdrawal of glucocorticoids in such patients may cause symptoms of adrenal insufficiency. The potency, dose, and duration of glucocorticoid use are important but imperfect predictors of the presence of HPA suppression. (See 'Hypothalamic-pituitary-adrenal axis suppression' above and 'Identifying patients with HPA suppression' above and "Pharmacologic use of glucocorticoids", section on 'HPA axis suppression'.)

●Identifying patients with HPA suppression – The response to administration of synthetic adrenocorticotropic hormone (ACTH [cosyntropin]) is the preferred method to assess adrenocortical function. Testing for HPA function is appropriate when patients are using <5 mg/day of prednisone and there is difficulty reducing the dose further because of symptoms consistent with adrenal insufficiency. (See "Diagnosis of adrenal insufficiency in adults".)

●Other forms of glucocorticoid dependence – Forms of glucocorticoid dependence other than symptomatic and biochemical evidence of HPA suppression that can hinder glucocorticoid tapering include psychologic dependence on glucocorticoids, disease recrudescence, and symptoms of apparent adrenal insufficiency despite normal HPA function and lack of disease recrudescence. (See 'Other forms of glucocorticoid dependence' above.)

●Approach to glucocorticoid tapering – There is a paucity of evidence to support any particular regimen of glucocorticoid tapering. Short-term glucocorticoid therapy (up to three weeks) can usually be stopped without a taper. In patients who have taken a glucocorticoid for a longer time, have a Cushingoid appearance, or received evening dosing, we suggest a tapering regimen that is largely based upon experience and considers the patient's general health status, stability of the disease being treated, and the drug regimen that has been used. The usual tapering endpoints are the patient's signs and symptoms of the underlying glucocorticoid-requiring illness. (See 'Tapering regimens' above and 'Recommended tapering regimen' above and 'Other published tapering regimens' above.)

•Stable dose reduction – The goal of tapering is to use a rate of change that will prevent both recurrent activity of the underlying disease and symptoms of cortisol deficiency due to persistent HPA suppression. We generally aim for a relatively stable decrement of 5 to 10 percent every one to four weeks, while accommodating convenience and individual patient response. (See 'Recommended tapering regimen' above.)

•Alternate-day regimen – There is a paucity of data regarding alternate-day tapering regimens. These are generally effective in most rheumatic diseases, but patients with rheumatoid arthritis often do not tolerate alternate-day dosing. (See 'Alternate-day regimen' above.)