Osteoporosis, fractures and osteonecrosis

GCs have been shown to stimulate osteoclastic activity initially (first 6–12 months of therapy), followed by a decrease in bone formation by suppressing osteoblastic activity in the bone marrow, decreasing osteoblast function and life span, and promoting the apoptosis of osteoblasts and osteocytes [11–13]. A meta-analysis of over 80 studies in adults found that use of ≥5 mg/day of prednisolone (or equivalent) was associated with significant reductions in bone mineral density (BMD) and an increase in fracture risk within 3 to 6 months of treatment initiation; this increased fracture risk was independent of patient age, gender and the underlying disease [14]. Kanis and colleagues examined 42,500 subjects from seven prospectively studied cohorts followed for 176,000 patient-years and found that prior and current use of corticosteroids increased fracture risk in both adult men and women, regardless of BMD and prior fracture history [15].

Osteonecrosis develops in 9–40% of adult patients receiving long-term GC therapy; it can occur as a result of systemic therapy or via intra-articular injections as well as in the absence of GC-induced osteoporosis [16]. Osteonecrosis is also being increasingly reported in children and adolescents treated for acute lymphoblastic leukemia (ALL) and non-Hodgkin lymphoma [17, 18].

Although the risk of osteonecrosis appears to increase with higher doses and prolonged treatment, it may also occur with low doses or after short-term GC exposure. Excessive alcohol intake, hypercoaguable states, sickle cell disease, radiation exposure and human immunodeficiency virus (HIV) infection have also been associated with the development of osteonecrosis [19].

A study of 270 adult cases of GC-induced osteonecrosis of the femoral head indicated that this condition was often misdiagnosed as lumbar disorders [20]. In this study, only one patient did not report any pain associated with osteonecrosis. Of the 269 patients who did report symptoms, 79% experienced pain due to osteonecrosis within 3 years of GC initiation (median 18 months) [20].

Adrenal suppression

Adrenal suppression (AS) refers to decreased or inadequate cortisol production that results from exposure of the HPA axis to exogenous GCs [21]. Duration of GC therapy and doses of GC treatment are not reliable predictors of which patients will develop AS [22, 23]. AS has been demonstrated after exposure to even 5-days’ duration of high-dose GC therapy [23]. It is important to recognize that inhaled, topical and intraocular GCs may also be absorbed systemically to the degree that they can cause AS [24–26].

Longer-acting GC formulations tend to be associated with a higher risk of AS [27]. Timing of GC administration may also influence the development of AS, with morning administration being potentially less suppressive than evening doses [27, 28]. Alternate-day therapy is also theoretically less suppressive than daily GCs based on the physiology of the HPA axis; however, there is currently no solid clinical evidence to support this proposition.

AS often occurs following abrupt discontinuation of GC therapy [29]. However, there are currently no evidence-based guidelines for tapering of GCs. Gradual GC tapering is frequently part of treatment protocols to reduce the risk of relapse and, therefore, comparative studies looking at AS without tapering would be difficult to perform. A study of patients with rheumatic disease found that rapidity of steroid taper did not make a difference in HPA-axis recovery [30]. However, many of these patients had undergone a gradual taper to prevent disease relapse and were all on “close to” physiologic doses of GC at the time of testing. In a study of children with ALL, GC tapering before discontinuation did not lead to complete resolution of AS [31]. Despite the lack of supportive evidence, many centres follow empiric tapering regimes based on the knowledge that AS is often seen following abrupt GC withdrawal.

Cushingoid appearance and weight gain

Prolonged corticosteroid therapy commonly causes weight gain and redistribution of adipose tissue that result in Cushingoid features (truncal obesity, facial adipose tissue [i.e., moon face], and dorsocervical adipose tissue). A survey of 2,167 long-term GC users (mean prednisone equivalent dose = 16 ± 14 mg/day for ≥60 days) found weight gain to be the most common self-reported AE (70%) [32]. An analysis of four prospective trials of GC use in patients with rheumatoid arthritis found a 4 to 8% increase in mean body weight with the use of 5–10 mg/day of prednisone or equivalent for >2 years [10].

Cushingoid features may develop within the first two months of GC therapy, and the risk of these complications appears to be dependent on both the dose and duration of treatment. One study evaluating the prevalence of Cushingoid abnormalities in 88 patients initiating long-term systemic corticosteroid therapy (initial daily dose ≥20 mg of prednisone or equivalent) found the cumulative incidence rates of these abnormalities to be 61% at 3 months and almost 70% at 12 months [33]. The risk of these complications was higher in younger patients, those with a higher baseline body mass index (BMI) and those with a higher initial caloric intake (>30 kcal/kg/day). Another study found the rate of Cushingoid features to increase linearly with dose: 4.3%, 15.8%, and 24.6% in patients receiving <5 mg/day, 5–7.5 mg/day, and >7.5 mg/day of prednisone (or equivalent), respectively [34].

Hyperglycemia and diabetes

Exogenous corticosteroid use is associated with hyperglycemia, and high-dose therapy increases insulin resistance in patients with pre-existing and new-onset diabetes. The effects of GC administration on glucose levels are observed within hours of steroid exposure [35], and appear to be dose-dependent. A population-based study of over 11,000 patients found that the risk for hyperglycemia increased substantially with increasing daily steroid dose; odds ratios (ORs) for hyperglycemia were 1.77, 3.02, 5.82 and 10.34 for 1–39 mg/day, 40–79 mg/day, 80–119 mg/day and ≥120 mg/day of hydrocortisone-equivalent, respectively [36]. GCs also appear to have a greater impact on postprandial compared to fasting glucose levels [37].

Glycemic targets and management strategies for patients with GC-induced hyperglycemia/diabetes are generally the same as in those with pre-established diabetes or glucose intolerance in the absence of GC therapy [38, 39] (see Hyperglycemia/Diabetes sections in Practical Recommendations for the Monitoring, Prevention and Management of Systemic Corticosteroid-Induced AEs). In general, GC-induced hyperglycemia improves with dose reductions and usually reverses when steroid therapy is discontinued, although some patients may develop persistent diabetes.

Cataracts and glaucoma

The risk of both cataracts and glaucoma is increased in patients using GCs, and this risk appears to be dose-dependent. GC use is typically associated with the development of posterior subcapsular cataracts (PSCC) [40], as opposed to nuclear or cortical cataracts. PSCC tend to be more visually significant and, therefore, usually require earlier surgical intervention/removal than other types of cataracts.

There is inter-individual variation in susceptibility to PSCC and the incidence varies per individual. Time until onset is at least 1 year with doses ≥10 mg/day of oral prednisone (or equivalent). Although PSCC are frequently seen in patients treated systemically, or even occasionally in those receiving inhaled corticosteroids (ICSs) [41], they are more commonly caused secondary to local treatment (e.g., topical eye drops and periocular or intravitreal administration).

Glaucoma is the more serious ocular complication of GC therapy. Systemic corticosteroids can painlessly increase intraocular pressure, leading to visual field loss, optic disc cupping, and optic nerve atrophy. Once systemic therapy is discontinued, the elevation in intraocular pressure often resolves within a few weeks, but the resultant damage to the optic nerve is often permanent.

While all patients using systemic steroids are at risk for elevation in intraocular pressure and glaucoma, certain groups appear to be at higher risk. Ocular hypertension and glaucomatous visual field defects have been reported in patients using systemic steroids with a personal or family history of open angle glaucoma, diabetes, high myopia or connective tissue disease (particularly rheumatoid arthritis) [42]. To reduce the risk of steroid-induced glaucoma, it is important to screen patients for these risk factors. All patients who may require long-term systemic GC therapy with a positive history for glaucomatous risk factors should be referred to an ophthalmologist for a comprehensive ocular assessment (see Ophthalmologic Examination section).

Central serous chorioretinopathy (CSCR) is also associated with systemic GC use. This type of chorioretinopathy is associated with the formation of subretinal fluid in the macular region which leads to separation of the retina from its underlying photoreceptors. This manifests as central visual blur and reduced visual acuity. Therefore, GCs should be used cautiously in patients with a history of CSCR [43].

Cutaneous adverse events

Corticosteroids induce atrophic changes in the skin that can lead to skin thinning and fragility, purpura and red striae. Skin thinning and purpura are usually reversible upon discontinuation of therapy, but striae are permanent.

Purpura generally affect the sun exposed areas of the dorsum of the hands and forearms, as well as the sides of the neck, face, and lower legs, and are usually not accompanied by palpable swelling [44, 45]. Red striae generally appear on the thighs, buttocks, shoulders and abdomen.

Impairment of wound healing is another common, and potentially serious, side effect of systemic GC use. Corticosteroids interfere with the natural wound-healing process by inhibiting leukocyte and macrophage infiltration, decreasing collagen synthesis and wound maturation, and reducing keratinocyte growth factor expression after skin injury [44]. Some topical and systemic agents may help counter the effects of corticosteroids on wound healing, including epidermal growth factor, transforming growth factor beta, platelet-derived growth factor, and tetrachlorodecaoxygen [45].

Gastrointestinal events

GC therapy has been associated with an increased risk of several adverse GI events including gastritis, ulcer formation with perforation and hemorrhage, dyspepsia, abdominal distension and esophageal ulceration. Despite the commonly held perception that steroid use increases the risk of peptic ulcer disease, large meta-analyses of randomized, controlled trials have failed to show a significant association between GC use and peptic ulcers [46, 47]. Recent evidence suggests that the risk of peptic ulcer disease due to corticosteroids alone is low, but increases significantly when these agents are used in combination with non-steroidal anti-inflammatory drugs (NSAIDS) [48]. One meta-analysis found a nearly four-fold increased risk of GI events among GC users who were also taking NSAIDS vs. those not using NSAIDS [49]. Messer et al. also found a four-fold increased risk of GI events with concomitant NSAID and GC use vs. non-use of either drug [50].

Acute pancreatitis has also been reported to be an adverse effect of corticosteroid use. A Swedish population-based, case–control study demonstrated an increased risk of acute pancreatitis after exposure to GC therapy [51]. Overall, the OR for developing acute pancreatitis was 1.53 (95% confidence interval [CI], 1.27-1.84) in GC users vs. non-users, with the risk of developing pancreatitis appearing to be greatest 4–14 days after subjects received treatment [51]. However, other evidence suggests that the underlying disease processes for which GC therapy is prescribed (particularly systemic lupus erythematosus [SLE]) may be more likely causes of pancreatitis than GC use [52].

Cardiovascular disease and dyslipidemia

GC use is associated with AEs that are known to be associated with a higher CVD risk, including hypertension, hyperglycemia, and obesity. A population-based study comparing 68,781 GC users and 82,202 non-users found the rate of CV events to be significantly higher in patients prescribed high GC doses (≥7.5 mg/day of prednisone or equivalent) vs. those who had not received GCs (adjusted relative risk [RR], 2.56; 95% CI, 2.18-2.99); CV risk was not increased in patients using <7.5 mg of prednisone daily [53]. Another large, retrospective case–control study found current GC use to be associated with a significantly increased risk of heart failure (adjusted OR, 2.66; 95% CI, 2.46-2.87) and ischemic heart disease (OR, 1.20; 95% CI, 1.11-1.29), but not ischemic stroke or transient ischemic attack (TIA). CV risk was found to be greater with higher GC doses and with current vs. past use [54].

Population-based studies conducted in Northern Europe have also noted an increased risk of new-onset atrial fibrillation (AF) and flutter in GC users [55, 56]. In these studies, the risk of AF was significantly greater with current or recent use (i.e., within 1 month) of high GC doses or with long-term use of these agents.

Serious CV events, including arrhythmias and sudden death, have also been reported with pulse GC therapy. However, these events are rare and have occurred primarily in patients with underlying kidney or heart disease [57]. Although it is unclear whether these serious AEs are due to GC use or the underlying condition, some experts recommend continuous cardiac monitoring in patients with significant cardiac or kidney disease receiving pulse therapy. Longer infusion times (2–3 h) for pulse GC therapy should be considered in patients who are at high risk of CVD [58].

Findings from studies examining the relationship between GC use and dyslipidemia have been conflicting. While clinical trials involving patients with SLE have shown prednisone doses >10 mg/day to be associated with hyperlipidemia [59, 60], another trial conducted in patients with rheumatoid arthritis found no adverse effect of prednisone (20 mg/day tapered to 5 mg/day over 3 months) on serum lipids after adjustment for other risk factors [61]. In fact, findings from a study examining data from 15,004 participants in the Third National Health and Nutrition Examination Survey suggest that GC use may have a beneficial effect on lipids in adults ≥60 years of age [62]. Despite the conflicting evidence, regular monitoring of lipids (as well as other traditional risk factors for CVD) is recommended in patients using GCs at high doses or for prolonged periods (see CV Risk and Dyslipidemia section).

Myopathy

Corticosteroids have direct catabolic effects on skeletal muscles that can lead to reductions in muscle protein synthesis and protein catabolism and, ultimately, muscle weakness. Myopathy generally develops over several weeks to months of GC use. Patients typically present with proximal muscle weakness and atrophy in both the upper and lower extremities; myalgias and muscle tenderness, however, are not observed. [58, 63].

Although there is some variation in the dose and duration of GC treatment prior to the onset of myopathy, it is more common in patients treated with ≥ 10 mg/day of prednisone or equivalent [64]. Also, the higher the GC dose utilized, the more rapid the onset of muscle weakness.

There is no definitive diagnostic test for GC-induced myopathy and, therefore, the diagnosis is one of exclusion. Symptoms generally improve within 3 to 4 weeks of dose reductions, and usually resolve after discontinuation of GC therapy [64]. Evidence also suggests that both resistance and endurance exercise may help attenuate GC-induced muscle atrophy [65].

Critical illness myopathy may also develop in patients requiring large doses of IV GCs and neuromuscular blocking agents. It is characterized by severe, diffuse proximal and distal weakness that develops over several days. Although it is usually reversible, critical illness myopathy can lead to prolonged intensive care unit (ICU) admissions, increased length of hospital stays, severe necrotizing myopathy and increased mortality [58, 63, 66]. Treatment is directed toward discontinuation of GC therapy or dose reductions as soon as possible, as well as aggressive management of underlying medical comorbidities.

Psychiatric and cognitive disturbances

GC use can lead to a wide range of psychiatric and cognitive disturbances, including memory impairment, agitation, anxiety, fear, hypomania, insomnia, irritability, lethargy, mood lability, and even psychosis. These AEs can emerge as early as 1 week after initiating corticosteroid therapy, and appear to be dependent on dose and duration of therapy [67, 68]. A family history of depression or alcoholism has also been reported as a risk factor for the development of GC-related affective disorders [69]. Individuals who develop psychiatric manifestations on short courses of GCs most commonly report euphoria, while those on long-term therapy tend to develop depressive symptoms [68, 70, 71].

GC therapy may also be associated with sleep disturbances and unpleasant dreams [72]; the risk of these events can potentially be decreased by modifying the timing of GC administration (e.g., a single morning dose) and/or night-time administration of drugs with sedative effects.

A decline in declarative and working memory has also been reported with GC therapy; these effects appear to be dose-dependent and frequently occur during the first few weeks of therapy [73]. Partial loss of explicit memory has been reported in patients treated with prednisone doses of 5 to 40 mg/day for at least 1 year [74]. Older patients appear to be more susceptible to memory impairment with less protracted treatment.

GC-induced psychosis usually only occurs with the use of high doses (>20 mg of prednisone or equivalent) for prolonged periods [75]. In patients with SLE, low serum albumin levels may also be predictive of GC-induced psychosis [76]. For patients with persistent symptoms of psychosis, antipsychotic therapy may be required [63].

Most patients with psychiatric reactions to corticosteroids usually recover from these symptoms with dose reductions or upon cessation of therapy. Lithium has also been found to be effective for both the prophylaxis and management of GC-related affective disorders [77].

Immunosuppression

The mechanisms by which corticosteroids inhibit the immune system and decrease inflammation may predispose patients to infection. A meta-analysis of 71 clinical trials involving over 2000 patients randomly allocated to systemic GC therapy found the overall rate of infectious complications to be significantly higher in patients using systemic corticosteroids vs. control subjects (RR, 1.6; 95% CI, 1.3-1.9; P < 0.001). However, the rate was not increased in patients given a daily dose of < 10 mg or a cumulative dose of < 700 mg of prednisone [78].

In addition to GC dose, other factors influencing the risk of infection include: the underlying disorder, patient age, and concomitant use of immunosuppressive or biologic therapies [48, 79]. A study comparing the infection risk posed by biologic therapies vs. non-biological disease-modifying antirheumatic drugs (DMARDs) in patients with rheumatoid arthritis found baseline GC use to be the factor most strongly associated with serious infections [80].

Patients using GCs appear to be particularly susceptible to invasive fungal and viral infections; this is especially true in bone marrow transplant recipients [45]. Older patients and those with lower functional status are also at higher risk for infections with steroid use. It is important to note that early recognition of infections in patients taking GCs is often difficult [48]. GC users may not manifest signs and symptoms of infection as clearly as non-users, due to the inhibition of cytokine release and associated reduction in inflammatory and febrile responses.

Growth suppression

Oral GC therapy has been associated with a delay in growth and puberty in children with asthma and other childhood diseases such as nephrotic syndrome [81–85]. Some evidence suggests that final height may also be compromised in children with a history of GC use [81, 86]. Lai and colleagues reported growth data on 224 children with mild-to-moderate cystic fibrosis who participated in a trial of alternate-day prednisone (1 or 2 mg/kg body weight) vs. placebo [86]. Subjects started prednisone treatment at a mean age of 9.5 years (range, 6 to 14 years); treatment was discontinued at mean ages of 12.9 and 13.8 years for the high-dose and low-dose groups, respectively, and growth was followed for an additional 6 to 7 years after prednisone discontinuation. At the time of final follow-up, 152 patients (68%) were older than 18 years of age. Mean height after age 18 years was found to be significantly lower in boys previously treated with either high- or low-dose prednisone vs. placebo. When indices of pulmonary status were controlled for, the negative association between the use of prednisone and height remained strong in boys. No persistent growth impairment was noted in female subjects.

It is important to note that although growth can be an independent adverse effect of corticosteroid therapy, it can also be a sign of AS.

Adrenal suppression

AS is the most common cause of adrenal insufficiency in children. The rate of adrenal crisis or death related to AS is unknown, however, adrenal insufficiency is associated with higher mortality in the pediatric population, highlighting the importance of recognition [29]. As in adults, the symptoms of AS are non-specific; therefore, the condition may go unrecognized until exposure to a physiological stress (illness, surgery or injury), which may result in adrenal crisis. Children with adrenal crisis secondary to AS may present with hypotension, shock, decreased consciousness, lethargy, unexplained hypoglycemia, seizures or even death (see Table 4) [87–91].

Several cases of pediatric AS have been reported in the literature, including adrenal crises in children requiring hospitalization and prolonged ICU stays [92, 93]. To our knowledge, there have been no published population studies looking at the frequency of symptomatic AS associated with systemic GCs. However, interim results from a national survey examining AS associated with any form of GC in the Canadian pediatric population over a two-year period have reported 44 cases of symptomatic AS, 6 of which presented as adrenal crisis [90]. A recent meta-analysis of AS in children treated with acute lymphoblastic leukaemia (ALL) found biochemical evidence of AS immediately following GC discontinuation in nearly all 189 patients [93]. AS resolved within a few weeks in most patients, but persisted for up to 34 weeks in others.

Although some studies have suggested that higher doses and longer durations of GC treatment may be risk factors for AS, these findings have not been consistent across trials [30, 93–96]. Even relatively low pharmacologic GC doses are significantly higher than physiologic doses, making AS a potential risk. For example, the standard dose of prednisone for the treatment of nephrotic syndrome in children is 2 mg/kg/day. When converted into dose/m², this dose is approximately 20 times the physiologic dose of GCs, highlighting the potential for AS with similar therapies. Currently, the Pediatric Endocrine Society recommends that AS be considered in all children who have received supraphysiological doses of GCs (>8-12 mg/m²/day hydrocortisone or equivalent) for greater than 2 weeks [29].

Hyperglycemia and diabetes

Most cases of medication-induced diabetes in children are associated with GC use. Steroid-induced hyperglycemia and diabetes have been reported in post-transplant patients, children with ALL, and those undergoing treatment for nephrotic syndrome [97, 98]. There have also been reports of diabetic ketoacidosis at presentation in these children [97, 98].

There is currently limited data describing risk factors for hyperglycemia and diabetes secondary to GC use in the pediatric population. Although obesity has been described as a potential risk factor, a retrospective Canadian study of children < 18 years of age found that, compared to those with established type 2 diabetes, those with medication-induced diabetes were less likely to be obese, have a positive family history of type 2 diabetes, and have obesity-related comorbidities (e.g., dyslipidemia, hypertension or elevated alanine aminotransferase levels) [97]. Therefore, evaluating for the typical type 2 diabetes risk factors may not be sufficient for identifying children at-risk of developing steroid-induced hyperglycemia or diabetes.

Cushing’s syndrome

GC therapy is by far the most common cause of Cushing’s syndrome in children. The clinical presentation in the pediatric population is similar to that in adults, and includes truncal obesity, skin changes and hypertension. In children, however, growth deceleration is also observed [99]. Children who develop features of Cushing’s syndrome as a result of GC therapy are at higher risk of experiencing AS. Therefore, HPA-axis function should be evaluated prior to discontinuing steroid therapy in children with Cushingoid features [88, 89].

Osteoporosis

A number of studies have reported decreased bone density in children taking oral corticosteroids [100–107]. Van Staa and colleagues examined the medical records of general practitioners in the United Kingdom to estimate the fracture incidence rates in children aged 4–17 years taking oral steroids (n = 37,562) and those taking non-systemic corticosteroids (n = 345,748) [108]. The risk of fracture was increased in children who received four or more courses of oral corticosteroids (adjusted OR, 1.32; 95% CI, 1.03-1.69). Of the various fracture types, the risk of humerus fracture was doubled in these children (adjusted OR, 2.17; 95% CI, 1.01-4.67). Children who stopped taking oral corticosteroids had a comparable risk of fracture to those in the control group [108].

Vertebral fractures are an under-recognized manifestation of osteoporosis in children, in part due to the fact that such fractures are often asymptomatic (even when moderate or severe) [109–112]. Similar to adults, vertebral fractures in GC-treated children are most frequently noted in the mid-thoracic region and at the thoracolumbar junction [109–112]. Recently, it has been shown that in children with GC-treated rheumatic disorders, 7% had prevalent vertebral fractures around the time of GC initiation, and 6% manifested incident vertebral fractures at 1 year [110, 111]. Children with rheumatic conditions and incident vertebral fractures at 12 months received twice as much steroid, and had greater increases in BMI and declines in spine BMD Z-scores [111]. These studies have played an important role in furthering our understanding of the osteoporosis burden manifesting as vertebral fractures in steroid-treated children.

BMD and fracture risk in adults

A lateral spine x-ray is also recommended in adults ≥65 years to assess for vertebral fractures.

BMD and fracture risk in children

In adults, a single BMD assessment can help predict the likelihood of fracture due to age-related osteoporosis. In children with GC-induced osteoporosis, however, this relationship is not as evident. Therefore, experts have recommended serial BMD assessments in at-risk children as well as in those displaying evidence of growth failure [119]. Since BMD results need to be carefully interpreted in relation to the child’s gender, age, height, and weight, as well as the underlying disease requiring GC therapy, referral to a specialist for assessment of bone symptomatology and BMD changes is recommended. At the same time, the bone health assessment of a child on chronic GC therapy needs to be extended beyond BMD in order to identify risk factors as well as early manifestations of osteoporosis. As such, bone health monitoring in pediatric chronic GC users includes an evaluation of calcium and vitamin D intake, back pain, physical activity, and disease-related risk factors for attenuated bone mineral accrual and bone loss (such as chronic inflammation and disuse). A spine radiograph should be considered in at-risk children with a prior history of vertebral fractures, back pain, chronic GC exposure (> 3 months), poorly-controlled inflammatory disease, significantly impaired mobility, or reductions in spine BMD Z-scores on serial measurements (Table 5).

Osteonecrosis (adults and children)

Because early diagnosis and appropriate intervention can prevent or delay the progression of osteonecrosis and the need for joint replacement, patients using high-dose GC therapy or those treated with GCs for prolonged periods should be evaluated for joint pain and decreased range of motion at each visit [58]. Magnetic resonance imaging should be considered in adult or pediatric patients presenting with these signs or symptoms [16].

Adrenal suppression (AS)

If AS is suspected, biochemical testing of the HPA axis should be considered after GC treatment has been reduced to a physiologic dose. Given the ease and practicality of a first morning cortisol measurement, it should be considered for the initial screening of patients at risk for AS. The test should be performed at 8:00 am or earlier given that cortisol levels decline throughout the day with natural circadian rhythm, and both evening and morning GC doses should be held prior to testing (see Table 8) [91]. If the 8:00 am cortisol value is below the normal laboratory reference range, AS is likely present and further GC withdrawal should occur only once testing has normalized. It is important to note that the specificity of the first morning cortisol test approaches 100% if a very low cut-off value (<85-112 nmol/L) is used. However, the sensitivity of this test is poor (~60%) [121]. Therefore, a normal cortisol value does not rule out the presence of AS. If a patient has signs or symptoms of AS and requires further testing, then referral to an endocrinologist should be considered. Clinicians must be aware that exogenous estrogen therapy, which affects cortisol-binding globulin levels, increases serum cortisol; therefore, the same thresholds for diagnosing AS do not apply in the setting of estrogen use.

The insulin tolerance test (ITT) is the definitive test for evaluation of the HPA axis, but performing this test is complicated and risky for patients since insulin is administered to achieve hypoglycemia. The ITT is contraindicated in children secondary to the risks of hypoglycemia on the pediatric brain. Therefore, in the setting of a normal morning cortisol result and the presence of AS symptoms, the low-dose adrenocorticotropic hormone (ACTH) stimulation test should be performed to confirm the diagnosis since it is a sensitive and specific test for AS [122–124]. The low-dose ACTH stimulation test involves IV administration of 1 μg of cosyntropin with measurements of baseline and stimulated serum cortisol levels to assess the function of the HPA axis. Cortisol levels are expected to peak between 20–30 min after cosyntropin injection, hence, cortisol measurements are recommended at 15–20 min and 30 min [124]. Many protocols also recommend measuring cortisol at 60 min. A peak cortisol of <500 nmol/L is diagnostic of AS, with both a sensitivity and specificity of approximately 90% [122–124] (note that a lower peak cortisol cut-off value may be required in neonates).

Growth in children

For children receiving GC therapy, growth should be monitored every 6 months (ideally by using stadiometry measurements) and measurements should be plotted on an appropriate growth curve (Table 5). If, after 6 months, growth velocity appears to be inadequate, the physician should consider all possible etiologies, including AS, as well as referral to an endocrinologist [91]. It is also important to rule out malnutrition as a cause of poor growth [9, 119].

CV risk and dyslipidemia

Hyperglycemia/diabetes

All patients should be educated about the classic signs and symptoms of hyperglycemia (polyuria, polydipsia, unexplained weight loss) so that they are screened for steroid-induced diabetes if symptoms arise. In adults, monitoring of glycated hemoglobin (A1C), fasting plasma glucose (FPG), 2-hour plasma glucose (2-h PG) (using a 75-g oral glucose tolerance test [OGTT]), or casual PG (any time of the day without regard to the interval since the last meal) are recommended (Table 5), although FPG, casual PG, and A1C may be less sensitive for diagnosing diabetes. If blood glucose or A1C is abnormal at baseline, then home blood glucose monitoring is also recommended.

Glucose investigations should be repeated after starting GC therapy. In patients taking prednisolone, some experts have recommended that blood glucose be monitored within 8 hours of the first dose (i.e., in the afternoon if once-daily prednisolone is administered in the morning) [37]. The 2013 Canadian Diabetes Association (CDA) guidelines recommend that glycemic parameters be monitored for at least 48 hours after initiation of GC therapy, regardless of whether or not the patient has diabetes [38]. Guidelines for blood glucose monitoring post-transplant suggest weekly monitoring for four weeks after transplant, followed by blood glucose checks at 3 and 6 months post-transplant, then annually thereafter [126]. A diagnosis of diabetes is confirmed if A1C is ≥6.5% (in adults), FPG is ≥7.0 mmol/L, 2-hour PG is ≥11.1 mmol/L or if casual PG is ≥11.1 mmol/L and the patient has classic symptoms of diabetes [38].

In the absence of screening guidelines for GC-induced diabetes in children, the authors recommend that physicians be aware of the risk of hyperglycemia in children receiving long-term supraphysiological GC doses and, at a minimum, screen for classic symptoms [98]. An annual FPG should also be considered.

In children presenting with symptoms suggestive of diabetes, FPG should be performed. If FPG is not diagnostic of diabetes in those with symptoms, OGTT is recommended. The diagnostic criteria for diabetes in children are the same as for adults [38]. More frequent screening of glucose parameters should be considered in children who are at potentially higher risk of developing hyperglycemia or diabetes, such as transplant recipients, obese patients, or those with conditions such as ALL or nephrotic syndrome. Annual OGTT is recommended in children who are very obese and/or who have multiple risk factors for type 2 diabetes since this test may be associated with higher detection rates.

Ophthalmologic examination

Patients on low-to-moderate doses of systemic corticosteroids for more than 6–12 months should undergo annual examination by an ophthalmologist (Table 5). An earlier examination is required in patients with symptoms of cataracts (namely blurred vision), however, this is generally not considered an ocular emergency that requires urgent treatment.

Early referral for monitoring of intra-ocular pressure (glaucoma) is recommended in patients at higher risk of developing steroid-induced glaucoma, such as those with a personal or family history of open angle glaucoma, diabetes mellitus, high myopia, or connective tissue disease (especially rheumatoid arthritis).

General strategies

Finally, whenever possible, GC-sparing agents should be considered. In patients with severe asthma, for example, use of the anti-immunoglobulin E (IgE) monoclonal antibody, omalizumab, has been shown to reduce the occurrence of asthma exacerbations requiring systemic corticosteroid therapy and to improve symptoms and asthma-related quality of life [127]. At present, omalizumab is reserved for patients with difficult to control asthma who have documented allergies and whose asthma symptoms remain uncontrolled despite ICS therapy [128].

Osteoporosis (adults)

Most guidelines and evidence support the use of bisphosphonates and teriparatide as first-line therapy for GC-induced osteoporosis in adults. A number of studies have demonstrated that the bisphosphonates alendronate, risedronate and zoledronic acid are effective for the prevention and treatment of GC-induced bone loss [132–138], although their long-term efficacy on fractures is not well established [132]. Teriparatide has been shown to be effective in improving BMD and reducing vertebral fractures in patients with GC-induced osteoporosis [139–141]. The ACR recommendations for the use of teriparatide and bisphosphonates are shown in Table 11[115].

Although other therapies such as calcitonin, raloxifene and denosumab may also play a role in the management of GC-induced osteoporosis in adults, they are not currently recommended as first-line therapy. Calcitonin has been found to prevent lumbar spine bone loss in the setting of GC use, but the same protection has not been observed at the femoral neck or with respect to fracture risk [142]. The European Medicines Agency (EMA) recently completed a review of the benefits and risks of calcitonin-containing medicines and concluded that there is evidence of a small, increased risk of cancer (0.7-2.4%) with long-term use of these agents [143]. Therefore, given this risk as well as its lack of efficacy in reducing fracture risk, calcitonin is not recommended as first-line therapy for GC-induced osteoporosis. However, it may be considered when bisphosphonates are contraindicated or in those patients who are intolerant to oral or IV bisphosphonates. Due to its analgesic effect, calcitonin can also be considered in patients who have sustained an acute fracture.

A study of postmenopausal women on ≤10 mg/day of prednisone (or equivalent) for ≥6 months demonstrated that treatment with raloxifene for 1 year improved spine and total hip (but not femoral neck) BMD [144]. However, the generalizability of these findings are limited since the study cohort was predominantly Asian and did not include patients on high-dose GC therapy. Furthermore, as a selective estrogen receptor modulator, raloxifene use for osteoporosis prevention and treatment is limited to the postmenopausal female population.

In animal models, denosumab has been shown to prevent steroid-induced bone loss and improve bone strength [145]. A phase 2 trial also found that denosumab improved lumbar spine BMD in patients with rheumatoid arthritis treated with corticosteroids and bisphosphonates [146]. The phase 3 FREEDOM trial found denosumab to be associated with a slightly increased risk of cellulitis [147], although the 2-year extension trial found no increased risk with longer term treatment [148].

In addition to pharmacologic therapy, current guidelines for GC-induced osteoporosis in adults recommend preventive measures such as smoking cessation, reduced alcohol consumption, participation in weight-bearing and strength-building exercises, falls risk assessment, and calcium and vitamin D supplementation [113–117, 129–131]. Cochrane investigators reviewed the available data on calcium and vitamin D use in GC-treated patients and found that supplementation prevented bone loss at the lumbar spine and forearm, but had no effect on femoral neck BMD or fracture incidence [149]. Adults on high-dose GC therapy should be taking 1200 mg/day of elemental calcium in divided doses and 800–2000 IU of vitamin D daily [113, 117].

Osteoporosis (children)

There are currently no evidence-based guidelines for the prevention and treatment of GC-induced osteoporosis in children. General measures are similar to those described above and include: using the lowest effective GC dose possible for the shortest period of time; proper nutrition and maintenance of a healthy weight; promotion of weight-bearing exercise; vitamin D supplementation to achieve at least 50 nmol/L, and ideally 75 nmol/L [150, 151]; calcium supplementation (if diet is inadequate to achieve the current, recommended dietary allowance) [150]; and the use of GC-sparing agents when possible. Most of the studies examining bisphosphonate use in GC-treated children have been observational in nature and have utilized the intravenous (IV) preparation, pamidronate [152]. Some evidence suggests that bisphosphonate therapy increases BMD, promotes reshaping and relieves back pain from previously fractured vertebral bodies, and is safe and well-tolerated in children with secondary osteoporosis [152–155], although long-term safety and efficacy data is still required. Currently, experts recommend consideration of bisphosphonate therapy in children with evident bone fragility associated with reductions in BMD parameters, particularly if there is a persistence of risk factors and, thereby, less likelihood of spontaneous BMD restitution and growth-mediated reshaping of vertebral bodies [153].

Osteonecrosis

Initial treatment for osteonecrosis includes bed rest and non-steroidal or other analgesics to relieve pain. For patients with early stage or less advanced osteonecrosis, joint-preserving strategies, such as reducing weight-bearing activities and core decompression (with or without marrow transplantation), have been utilized with varying levels of success. For more advanced disease, femoral head or total hip replacement surgery is usually required [16]. Since these replacements generally have a 10-year lifespan, strategies that delay the need for surgery are desired. Some evidence suggests that treatment with alendronate may reduce the risk of bone collapse and delay the need for surgery [156, 157]. However, a recent randomized, controlled trial found no benefit of alendronate vs. placebo in patients with osteonecrosis [158]. In children with osteonecrosis in the leukemia setting, IV pamidronate has been associated with significant improvements in pain and mobility [159, 160].

Hyperglycemia and diabetes

If blood glucose levels are >15 mmol/L, then insulin is usually required to achieve glycemic control. In the absence of a contraindication, metformin is often recommended in combination with insulin (Table 12). A reasonable starting dose for insulin is 0.15-0.3 units/kg/day. With once-daily morning administration of prednisone, fasting glucose may be unaffected, but blood glucose will be higher later in the day. If this occurs, then an intermediate-acting insulin (such as N or NPH) or a premixed combination of intermediate- and fast-acting insulin can be initiated in the morning. If blood glucose is elevated in the morning as well, then an evening insulin dose may also be required. If shorter-acting GCs are administered more than once per day, or if dexamethasone is used, then both fasting and non-fasting glucose levels are likely to be affected. In this case, twice-daily intermediate-acting insulin or long-acting insulin, such as detemir or glargine, are recommended; fast-acting insulin may also be required at mealtimes. In order to prevent hypoglycemia, it is important to remember to adjust diabetes medications if GC doses are reduced.

The treatment of GC-induced diabetes in children is best accomplished through the combined efforts of a multidisciplinary pediatric diabetes healthcare team [98]. As with adults, lifestyle interventions should be initiated; if glycemic targets are not met with these modifications, insulin must be considered. Many of the other glucose-lowering agents used in adult patients with type 2 diabetes have not been licensed for use in the pediatric population and may be contraindicated in children with complex medical issues [98]. Until the safety and efficacy of these medications in children are established, they cannot be recommended for routine clinical use in this population.

Adrenal suppression (AS)

To minimize the risk of developing AS, it is important to consider the relative suppressive effects of the various GCs (based on potency and duration of action) prior to initiating therapy (see Table 3). The lowest effective dose should be utilized for treatment of the underlying condition and the dose should be re-evaluated regularly to determine if further reductions can be instituted. If possible, the GC should be administered once-daily in the morning.

Screening tests should be considered to assess adrenal function as GC therapy is being withdrawn. Screening should occur before tapering to less than a physiologic dose (Tables 13 and 14) [161, 162]. When possible, screening should occur at least 1 week after the dose has been tapered to a once-daily physiological dose (preferably hydrocortisone, which has a shorter half-life).

Consideration should be made to educate patients about the risk of AS if they have been treated with GC therapy within the last year, but have not had testing to rule out AS. In the event of severe illness or surgery, stress dose steroids should be considered to prevent adrenal crisis.

Growth

The potency of dexamethasone and betamethasone in suppressing growth has been shown to be nearly 18 times higher than that of prednisolone [163]. Therefore, to reduce the risk of growth suppression in children, lower potency agents, such as prednisolone, should be used whenever possible. Consideration should also be given to alternate-day dosing (if possible) since evidence suggests that the use of lower doses of prednisolone (10–15 mg/day or < 0.5 mg/kg/day single dose) on alternate days does not significantly slow growth velocity [9].

There is currently insufficient evidence to support the use of recombinant human growth hormone (rhGH) for the treatment/prevention of GC-induced growth suppression. There has been some evidence of short-term benefits on growth velocity with rhGH therapy [164], however further study, including evaluation of final adult height, is required.

Gastrointestinal (GI) events

Consideration can be given to the use of proton pump inhibitors (PPIs) for GI protection in GC users at high-risk of GI bleeding or peptic ulcers, such as those using NSAIDS, patients with a history of ulcers or GI bleeding, and those with serious comorbidities (i.e., advanced cancer) [1].

Cutaneous events: Red striae

Although treatment of red striae is often disappointing, some success has been noted with topical vitamin A 0.1% cream, flashlamp-pumped pulsed dye lasers, and a combination of pulsed dye laser and Thermage (a non-ablative radiofrequency device) [44]. To help reduce the risk of striae, patients initiating systemic corticosteroid therapy should be advised to follow a low-calorie diet.