Complications of Corticosteroid Therapy: A Comprehensive Literature Review

Abstract

Relevance to Patient Care and Clinical Practice: Corticosteroids are among the most prescribed medications, particularly during the COVID-19 era. The literature has clearly highlighted the dangers of prolonged, high-dose corticosteroid use, which is important for clinicians to consider before treating patients in their clinical practices. Objective: The objective of this article is to review the literature on complications of corticosteroid use, review corticosteroid pharmacokinetics, and provide an updated reference on risks associated with corticosteroid therapy, especially at higher doses. Data Sources: A conventional literature search of PubMed was conducted without restrictions on publication date. Search terms included “corticosteroids,” “avascular necrosis,” “gastrointestinal bleeding,” and “complications.” Study Selection and Data Extraction: Pertinent systematic review/meta-analyses and randomized controlled trials were reviewed for study inclusion. Data Synthesis: Corticosteroids were associated with complications including avascular necrosis, gastrointestinal bleeding, myocardial infarction, heart failure, cerebrovascular events, diabetes mellitus, psychiatric syndromes, ophthalmic complications, tuberculosis reactivation, and bacterial sepsis. Increased daily and cumulative doses were associated with increased excess risk of complications. Cumulative doses greater than 430 mg prednisone equivalent were shown to increase the excess risk of avascular necrosis, with progressively higher rates with higher doses. Risk of gastrointestinal bleeding was significantly increased with corticosteroid usage in the in-patient but not out-patient setting. Conclusion: Since corticosteroids have been associated with the aforementioned severe complications and frequent medicolegal malpractice claims, counseling and informed consent should be performed when prescribing moderate-high dosages of corticosteroids. Further research is needed to characterize the long-term effects of corticosteroid usage in COVID-19 patients.

Introduction

Corticosteroids (CS) are commonly utilized across many clinical specialties in medicine despite having well-known complications including avascular necrosis (AVN), osteoporosis, upper gastrointestinal (GI) bleeding, myocardial infarction, reactivation tuberculosis, cataracts, diabetes, and psychosis. ¹ In the era of SARS-Cov-2 (COVID-19), high-dose dexamethasone therapy has shown to significantly reduce 28-day mortality for select COVID-19 patients.^{2 -4} Consequently, a literature review on the complications of steroid therapy is timely. Since many of the adverse effects of corticosteroids are observed years after administration, the clinical trajectory for COVID-19 patients currently receiving high doses of dexamethasone remains unknown. However, prior investigations on corticosteroid use and their associated adverse effects may help elucidate the subsequent trends that follow patients during this pandemic. Our review will outline the major short- and long-term adverse effects of CS, with the goal to provide an updated reference when contrasting the risks and benefits of prescribing CS, especially at higher doses.

Methods

A PubMed search was conducted using the search terms corticosteroids, glucocorticoids, steroids, osteonecrosis, avascular necrosis, AVN, gastrointestinal bleed, GI bleed, complications, adverse effects, adverse events, and SARS. These terms were combined in various permutations, with modifiers such as “AND” and “OR.” There were no restrictions placed on publication date. Pertinent systematic reviews, meta-analyses, randomized controlled trials, cohort studies, and case-control studies were reviewed for study inclusion. Non-English manuscripts were excluded. Several investigations were subsequently added through snowballing. A previously established abstraction form (Table 1) was used by author E.K. for data extraction.

Table 1.

Abstraction Form for Studies Featuring Corticosteroid Complications.

Title	Author	Year	Complication	Incidence/prevalence	Hazard ratio/odds ratio/relative risk
Association between topical corticosteroid use and type 2 diabetes in two European population-based adult cohorts	Andersen et al 5	2019	Diabetes	0.57-0.66	OR 1.41 (1.28 to 1.32)
Corticosteroid-induced cutaneous changes: a cross-sectional study	Kannan et al 6	2015	Dermatologic	56	—
Low-dose corticosteroids and avascular necrosis of the hip and knee	Bauer et al 7	2000	AVN		[Cumulative dose <430 mg] RR 0 (0 to 5) [Cumulative dose 440-1290 mg] RR 6 (1 to 43) [Cumulative dose >1290 mg] RR Undefined (26 to infinity)
Hypertension and asthma: A comorbid relationship	Christiansen et al 8	2016	Hypertension	54.5	OR 1.30 (1.27 to 1.34)
Corticosteroids and peptic ulcer: meta-analysis of adverse events during steroid therapy	Conn and Poynard 9	1994	Infection	6.5	OR 1.2 (1.0 to 1.6)
Corticosteroid-induced adverse events in adults: frequency, screening and prevention	Fardet et al 1	2007	Adipose distribution	61-69	—
			Insomnia	50	—
			Infection	—	RR 1.6 (1.3 to 1.9)
			Myopathy	—	OR 6.7 (4.8 to 4.93)
Psychiatric complications of treatment with corticosteroids: review with case report	Kenna et al 35	2011	Psych	5-10	[Psychosis] OR 1.3 (0.8 to 2.1)4
Use of inhaled corticosteroids and the risk of tuberculosis	Lee et al 10	2013	TB	19.1	OR 1.20 (1.08 to 1.34)
Avascular osteonecrosis after treatment of SARS: a 3-year longitudinal study	Lv et al 11	2009	AVN	57.7	[Cumulative dose 1 g] OR 1.42 (1.14 to 1.77) [Maximal daily dose 100 mg] OR 2.22 (1.44 to 3.41)
High-dose corticosteroid use and risk of hip osteonecrosis: meta-analysis and systematic literature review	Mont et al 12	2015	AVN	6.7	[Cumulative dose >10 g] OR 2.4 (0.8 to 6.4)
Corticosteroids and risk of gastrointestinal bleeding: a systematic review and meta-analysis	Narum et al 13	2014	GI bleed	—	OR 1.43 (1.22 to 1.66)
Cataract prevalence and prevention in Europe: a literature review	Prokofyeva et al 14	2013	Ophthalmological	—	OR 3.25 (1.39 to 7.58)
Corticosteroid therapy for patients hospitalized with community-acquired pneumonia: a systematic review and meta-analysis	Siemieniuk et al 15	2015	Hyperglycemia	13.8	RR 1.48 (1.01 to 2.19)
Long-term outcome of short-course high-dose glucocorticoids for SARS: a 17-year follow-up in SARS survivors	Sing et al 16	2020	AVN	—	[VHD compared with LTHD] 7.50 (2.34 to 24.0)
The epidemiology of corticosteroid-induced osteoporosis: a meta-analysis	van Staa et al 17	2002	Osteoporosis	—	RR 1.33 (1.29 to 1.38)
Taking glucocorticoids by prescription is associated with subsequent cardiovascular disease	Wei et al 18	2004	Heart failure	—	OR 3.72 (2.71 to 5.12) for >7.5 mg/day prednisolone equivalent
			Cerebrovascular	—	OR 1.73 (1.22 to 2.44) for >7.5 mg/day prednisolone equivalent
			MI	—	OR 3.26 (2.60 to 4.09) for >7.5 mg/day prednisolone equivalent
Steroid therapy and the risk of osteonecrosis in SARS patients: a dose-response meta-analysis	Zhao et al 4	2017	AVN		[Per 5.0 g methylprednisolone-equivalent cumulative increase] RR 1.57 (1.30 to 1.89)
Long-term systemic corticosteroid exposure: a systematic literature review	Rice et al 19	2017	Cataracts	1-3	—
			Fracture/osteoporosis	21-30	—
Association between oral corticosteroid bursts and severe adverse events: a nationwide population-based cohort study	Yao et al 20	2020	Heart failure	0.13	—

Abbreviations: AVN, avascular necrosis; GI, gastrointestinal; LTHD, low to high dose; MI, myocardial infarction; OR, odds ratio; RR, relative risk; SARS, severe acute respiratory syndrome; TB, tuberculosis; VHD, very high dose.

Results

Pharmacokinetics of commonly utilized CS are reviewed in Table 2. Dexamethasone has a longer biological half-life and greater antiinflammatory potency than hydrocortisone, methylprednisolone, prednisolone, and prednisone at drug-equivalent doses. Summary of complications included in Table 3. Unless otherwise stated, all dosages have been converted into prednisone equivalency in reviewed articles.

Table 2.

Pharmacokinetics of Oral Corticosteroids.

Drug [Equivalent dosage]	Plasma half-life (hours)	Biological half-life (hours)	Metabolism	Antiinflammatory & mineralocorticoid potency a
Hydrocortisone [20 mg] 21	1.7-2.1	8-12	Metabolized by liver to inactive metabolites	Antiinflammatory 1 Mineralocorticoid 1
Prednisone [5 mg] 22	1	18-36	Metabolized by liver to prednisolone (active compound)	Antiinflammatory 5 Mineralocorticoid 0.8
Prednisolone [5 mg] 23	2.1-3.5	18-36	Metabolized by liver to prednisone (reversibly) and other inactive metabolites	Antiinflammatory 5 Mineralocorticoid 0.8
Methylprednisolone [4 mg] 24	1.3-3.7	18-36	Metabolized by liver to inactive metabolites	Antiinflammatory 4 Mineralocorticoid 0.5
Dexamethasone [0.75 mg] 25	1.8-3.5	36-54	Metabolized by liver to inactive metabolites	Antiinflammatory 25 Mineralocorticoid 0

^aPotency relative to hydrocortisone.

Table 3.

Summary of Corticosteroid Complications Listed in Alphabetical Order.

Complication	Source	Incidence/prevalence %	Hazard ratio/odds ratio/relative risk (95% CI)
Acne	Kannan et al6,a	56	—
Adipose tissue distribution	Fardet et al 1	61-69	—
AVN	Bauer et al 7	—	[Cumulative dose <430 mg] RR 0 (0-5) [Cumulative dose 440-1290 mg] RR 6 (1-43) [Cumulative dose >1290 mg] RR Undefined (26-infinity)
AVN	Sing et al 16	—	[VHD compared with LTHD] 7.50 (2.34-24.0)
AVN	Lv et al 11	57.7	[Cumulative dose 1 g] OR 1.42 (1.14-1.77) [Maximal daily dose 100 mg] OR 2.22 (1.44-3.41)
AVN	Mont et al 12	6.7	[Cumulative dose >10 g] OR 2.4 (0.8-6.4)
AVN	Zhao et al 4	—	[Per 5.0 g methylprednisolone-equivalent cumulative increase] RR 1.57 (1.30-1.89)
Bacterial sepsis	Conn and Poynard 9	6.5	OR 1.2 (1.0-1.6)
Cataracts	Prokofyeva et al 14	1-319	OR 3.25 (1.39-7.58)
Fracture or osteoporosis	van Staa et al 17	21-3019	RR 1.33 (1.29-1.38)
GI bleeding or perforation	Narum et al 13	2.9	OR 1.43 (1.22-1.66)
Heart failure	Wei et al 18	0.1320	OR 3.72 (2.71-5.12) for ≥7.5 mg/day prednisolone equivalent
Hyperglycemia	Siemieniuk et al 15	13.8	RR 1.48 (1.01-2.19)
Hypertension	Christiansen et al 8	54.5	OR 1.30 (1.27-1.34)
Insomnia	Fardet et al 1	50	—
Infectious events	Fardet et al 1	—	RR 1.6 (1.3-1.9)
Cerebrovascular event	Wei et al 18	—	OR 1.73 (1.22-2.44) for ≥7.5 mg/day prednisolone equivalent
Major psychiatric syndromes	Kenna et al 35	5-10	[Psychosis] OR 1.3 (0.8-2.1) 4
Myocardial infarction	Wei et al 18	—	OR 3.26 (2.60-4.09) for ≥7.5 mg/day prednisolone equivalent
Myopathy	Fardet et al 1	—	OR 6.7 (4.8-4.93)
TB Infection	Lee et al 10	19.1	OR 1.20 (1.08-1.34)
Type 2 DM	Andersen et al5,a	0.57-0.66	OR 1.41 (1.28-1.32)

Unless otherwise stated, all dosages have been converted into prednisone equivalency.

Abbreviations: AVN, avascular necrosis; CI, confidence interval; DM, diabetes mellitus; GI, gastrointestinal; LTHD, low to high dose (≤100 mg a day); OR, odds ratio; RR, relative risk; TB, tuberculosis; VHD, very high dose (>100 mg a day).

^aSource includes topical steroids.

Avascular Necrosis

Pietrogrande and Mastromarino (1957) were the first to report an association between CS and AVN in a 43-year-old man who had received 100 mg cortisone acetate daily for 4 years. ²⁶ In 1971, Fisher and Bickel ²⁷ published the first clinical study investigating 77 patients with CS-related AVN, concluding that both higher CS dosages and longer duration of CS therapy were correlated with CS-related AVN. In 2000, Bauer quantified 3-year cumulative doses that were correlated with an increased excess risk of AVN. While the article found no excess risk of AVN for 3-year cumulative doses of 1 to 430 mg prednisone (relative risk [RR] 0; 95% confidence interval [CI] 0 to 5), there was an increase in excess risk of AVN for 3-year cumulative doses >430 mg prednisone (RR 6; 95% CI 1 to 43 for cumulative doses 440-1290 mg) (RR undefined; 95% CI 26 to infinity for cumulative doses 440-1290 mg). ⁷ In 2009, Lv et al ¹¹ reviewed 71 former SARS patients (mainly health care workers) who had been treated with CS and found that 41 (57.7%) of them had developed AVN. A majority ( n = 28 [68.3%]) of those who developed AVN were diagnosed 3 to 4 months after the start of CS treatment. For the 60 SARS patients who received a 3-year cumulative prednisone-equivalent dosage of 1000 to 9999 mg, 30 (50%) developed AVN; for the 11 SARS patients who received a 3-year cumulative prednisone-equivalent dosage of >10 000 mg, all 100% developed AVN.

Mont et al ¹² published a systematic review that reviewed 57 studies with 23 561 patients and a meta-analysis involving 7 of those studies (1515 patients). The systematic review found the incidence of AVN to be 6.7% for patients on cumulative >2000 mg prednisone equivalent but the incidence of AVN in CS-treated SARS to be 21.8%. Also, the meta-analysis found that for CS-treated patients, a 10 mg/day prednisone dose was associated with a 3.6% increase in the incidence of AVN. Patients treated with >20 mg/day had significantly higher odds than <20 mg/day (odds ratio [OR] 9.1; 95% CI 4.6 to 1.98). Patients treated with cumulative prednisone-equivalent doses >10 000 mg also had increased odds of developing necrosis (OR 2.4; 95% CI 0.8 to 6.4), versus <10 000 mg (OR 0.4; 95% CI 0.25 to 0.54).

Gastrointestinal Bleeding

In a systematic review and meta-analysis, Narum et al ¹³ reviewed 159 randomized, double blind, placebo-controlled trials with 33 253 participants. They concluded that compared with placebo users, CS users had a 40% increase of experiencing GI bleeding or perforation (OR 1.43; 95% CI 1.22 to 1.66). Subgroup analysis of these patients revealed that hospitalized patients had a statistically significant increased risk of GI bleeding or perforation (OR 1.42; 95% CI 1.22 to 1.66), with an absolute risk of 37.9 events/1000 patients for CS versus 26.4 events/1000 patients for placebo. However, while there was an increased risk for patients in ambulatory care, the results were not statistically significant (OR 1.63; 95% CI 0.42 to 6.34), with an absolute risk of 1.8 events/1000 patients for CS versus 0.7 events/1000 patients for placebo. Of 8651 ambulatory patients, there were 11 GI bleeds (8 CS/3 placebo). The authors discussed that higher rates of bleeding and perforation in hospitalized patients may be due to complications of stress ulcers seen in critically ill patients.

For Narum et al’s review, 33.3% of the studies excluded patients with a history of peptic ulcer disease; when this group was excluded from the analysis, results were still significant (OR 1.47; 95% CI 1.21 to 1.78). Concomitant gastroprotective drugs were described in 9% of studies; when this group was excluded from the analysis, results were still significant (OR 1.42; 95% CI 1.21 to 1.67). Looking at the results of the 14 studies with concomitant gastroprotective drugs documented, rates were lower but still elevated (OR 1.29; 95% CI 0.62 to 2.69), events 38 CS/25 placebo (50.6 CS/33.6 placebo per 1000 patients). Concomitant aspirin/non-steroidal antiinflammatory (ASA/NSAID) use was documented in 12% of studies. Since ASA/NSAIDs are associated with an increased risk of peptic ulcers, this group was excluded from the analysis.^28,29 However, the results were still significant (OR 1.44; 95% CI 1.20 to 1.71). ¹³

Cardiovascular Event

Souverein et al ³⁰ reported an observational, nested, case-control study over a 10-year period, in which all patients were older than 50 years and had received at least one dose of CS: either oral (study patient) or topical/inhaled (control). A patient on oral CS was included in the study group (50 656 patients) when first diagnosed with a cardiovascular (ischemic heart disease, heart failure) or cerebrovascular event. Controls (50 656 patients) were matched at or near the same date of inclusion, stratified with similar confounding risk factors, and had no history of prior cardiovascular events. Corticosteroid therapy was associated with a large increase in heart failure (OR 1.91; 95% CI 1.79 to 2.03), a small increase in ischemic heart disease (OR 1.09; CI 1.03 to 1.15), and a slight decrease in cerebrovascular events (OR 0.95; CI 0.89 to 1.01). The highest rates were seen with current CS use as opposed to recent or past use.

In addition, the temporal relationship of CS to these events reflected the associations stated above. ³⁰ Current use of CS had a strong association with heart failure (OR 2.66; 95% CI 2.46 to 2.87) versus recent (OR 1.40; CI 1.27 to 1.55) and past (OR 1.19; CI 1.08 to 1.32) use. However, current use of CS had a statistically significant association with ischemic heart disease (OR 1.20; CI 1.11 to 1.29) but not recent (OR 0.93; CI 0.85 to 1.02) nor past (OR 1.07; CI 0.98 to 1.17) use. Finally, current and recent use of CS had a statistically significant decrease in cerebrovascular events (OR 0.91; CI 0.84 to 0.99) (OR 0.89; CI 0.80 to 0.99) but not for past use (OR 1.06; CI 0.96 to 1.17).

Wei et al ¹⁸ reviewed 68 781 patients exposed to at least one dose of CS (including inhaled, topical, oral, parenteral, nasal, and rectal) and 82 202 matched controls without any CS exposure during a 3-year period. Corticosteroid dose was calculated as low (inhaled, topical, nasal), medium (<7.5 mg/day prednisone equivalent), or high (≥7.5 mg/day prednisone equivalent). Heart failure had increasing rates with increases in CS dosing: low (OR 1.18: 95% CI 1.05 to 1.33); medium (OR 1.5; 95% CI 1.29 to 1.75); high (OR 3.72; 95% CI 2.71 to 5.12). Myocardial infarction was elevated for high-dose therapy only (OR 3.26; 95% CI 2.6 to 4.09). Cerebrovascular events were elevated for high-dose therapy only (OR 1.73; 95% CI 1.22 to 2.44). All case mortality was increased for medium dose (OR 1.18; CI 95% 1.12 to 1.24) and high-dose therapy (OR 7.41; CI 95% 6.90 to 7.98).

Diabetes Mellitus

A large case-control study (11 855 cases and 11 855 controls) by Gurwitz et al ³¹ showed an increased risk of developing new-onset treatment-required hyperglycemia after CS use (RR 2.23; 95% CI 1.92 to 2.59). This risk increased with increasing average daily steroid dose (RR 1.77 for 1-10 mg/day prednisone equivalent; RR 3.02 for 10-20 mg/day; RR 5.82 for 20-30 mg/day; RR 10.34 for >30 mg/day). In regard to CS use in known diabetic patients, Feldman-Billard et al ³² showed CS-induced worsening of glycemic controls in known diabetic patients, with 51 of 80 (64%) patients with diabetes requiring rapid insulin after a pulse of methylprednisolone prescribed for ophthalmological diseases. Those who received rapid insulin had a significantly higher baseline blood glucose (9.9±3.7 vs 7.3±2.3 mmol/L, P <0.001) and HbA1c (8.4±1.9 vs 6.6±0.9%, P <0.001). Probabilities of requiring rapid insulin were 28% for HbA1c values of <6.5%, 58% for HbA1c values of 6.5% to 8%, and 100% for HbA1c values of >8%. In a more recent study from 2015, Siemieniuk et al ¹⁵ also showed increased risk of hyperglycemia requiring treatment with CS use (RR 1.4; 95% CI 1.01 to 2.19).

In a meta-analysis by Conn and Poynard, ⁹ diabetes was reported 4 times more frequently in CS-treated individuals (OR 1.7; 95% CI 1.1 to 2.6). Andersen et al ⁵ also supported an association between CS use and type 2 diabetes mellitus. Even when adjusted for, this association was seen with both topical (OR 1.25; 95% CI 1.23 to 1.28) and systemic CS use (OR 1.28; 95% CI 1.23 to 1.32). In addition, this study had 4 potency categories based on the World Health Organization’s classification of drugs: mild topical CS (e.g., hydrocortisone), moderate topical CS (e.g., hydrocortisone-17-butyrate), potent topical CS (e.g., mometasone furoate), and very potent topical CS (e.g., clobetasol propionate). Similar to hyperglycemia, CS-associated diabetes followed a dose-response pattern (OR 1.17 for mild topical CS; OR 1.22 for moderate topical CS; OR 1.26 for potent topical CS; OR 1.33 for very potent topical CS).

Psychiatric Syndromes

In 1972, the Boston Collaborative Drug Surveillance Program reported “psychiatric reactions” in 1.3% of 463 patients treated with ≤40 mg/day of prednisone, 4.6% of 175 patients treated with 41-80 mg/day, and 18.4% of 38 patients treated with >80 mg/day. ³³ Then in 1983, Lewis and Smith reported an incidence of 5.7% of severe psychiatric symptoms in patients treated with CS, looking at 13 studies and involving 2555 CS-treated patients. ³⁴ Finally, in 2011, Kenna et al ³⁵ concluded that hypomania and/or mania are the most common psychiatric adverse effects of CS treatment, in addition to increased risk of depression in prolonged CS exposure.

Ophthalmic Complications

A case-control study by Garbe et al ³⁶ quantified the association between CS use and the ocular hypertension-open angle glaucoma spectrum (OR 1.41; 95% CI 1.22 to 1.63). Similar to other steroid-associated complications, the odds of ocular hypertension or open-angle glaucoma also increased with steroid dosage (OR 1.26 for <10 mg/day prednisone equivalent; OR 1.37 for 10-20 mg/day; OR 1.88 for >20 mg/day).

In addition, a literature review by Prokofyeva et al ¹⁴ revealed that CS-induced cataracts are usually posterior and subscapular: a rare presentation in patients not taking CS. The review also revealed that various forms of CS (i.e., inhaled, topical, parental, and oral) were associated with an increased risk of cataracts. Oral, parental, and inhalational CS led to cataracts more frequently than CS applied topically. The group at highest risk of CS-induced cataracts was oral CS use for >5 years (OR 3.25; 95% CI 1.39 to 7.58).

Infection

A nested case-control study by Lee et al ¹⁰ with 4139 active tuberculosis (TB) patients and 20 583 control patients showed an increased rate of active TB diagnosis in patients who used inhaled CS (OR 1.20; 95% CI 1.08 to 1.34). Corticosteroids also dose-dependently increased the risk of active TB development in patients with asthma or chronic obstructive pulmonary disease. A large case-control study by Jick et al ³⁷ also illustrated a dose-dependent correlation with CS and TB reactivation. With 497 cases and 1966 controls, the OR was 4.9 (95% CI 2.9 to 8.3). For <15 mg/day prednisone equivalent, the OR was 2.8 (95% CI 1.0 to 7.9). For ≥15 mg/day prednisone equivalent, the OR was 7.7 (95% CI 2.8 to 21.4). Additional risk factors included elevated body mass index, current smoking, and history of diabetes, emphysema, bronchitis, and asthma.

In addition, corticosteroids can mask infections until they reach a severe and advanced stage. Conn and Poynard found the incidence of bacterial sepsis to be 6.5% of 2868 CS-treated patients compared with 4.8% of 2776 placebo patients (OR 1.2; 95% CI 1.0 to 1.6). ⁹

Discussion

The risk of developing AVN from CS therapy is clearly outlined, with increased risk associated with high-dose therapies used in treating acute illnesses including COVID-19. There appears to be a dose-dependent response curve with higher doses leading to higher rates of AVN. Although there has been no consensus on risk-related cutoffs, it has been previously reported that 3-year cumulative doses of >430 mg prednisone were statistically significant in AVN development. ⁷ However, it has been firmly established that large cumulative doses, as used in SARS therapy, result in higher AVN rates.^4,11 In patients with chronic illnesses like asthma, which may require recurrent moderate or high-dose CS therapy, the cumulative dose level may predispose the patients to AVN.

GI tract bleeding due to CS use is always an underlying concern, with the data supporting discontinuing the concomitant use of NSAIDs with CS.^28,29 While on CS, initiation of gastroprotective medications like proton pump inhibitors may be of benefit and should be considered. However, the benefit of gastroprotective drugs utilized with CS has not been adequately documented in the literature. Although most of the earlier studies used antacids, improved options for acid control are now available with proton pump inhibitors. ¹³ Regardless, gastroprotective medications have not been shown to decrease the risk of bleeding and perforation in patients using CS and with a history of bleeding or peptic ulcer disease. Alternatives to CS should be considered in these patients. In addition, the low bleed rate for ambulatory patients in Narum et al may provide a false sense of security and should not be relied upon, since the CS dosage levels were not provided.

Use of high-dose CS therapy in patients with a significant history of cardiovascular/cerebrovascular disease must be weighed with the benefits of systemic therapy. In the study from Souverein et al, ³⁰ one difficulty is that after controls were matched with study patients, later CS therapy was permitted for control group patients. This makes the data difficult to interpret, since both study patients and controls show high rates of oral CS use. This could potentially hide higher rates of cardiovascular events in the comparison between the two groups. The study by Wei et al ¹⁸ does not have this issue, as all matched controls have no use of CS. Heart failure worsened as doses increased, with similar results reported by Souverein et al. ³⁰ In addition, while all cause mortality was increased with both medium and high CS dosages, myocardial infarction and cerebrovascular events were only substantially increased with high CS dosages. ¹⁸ These findings demonstrate the cardiovascular risks of using high CS dose therapy but also indicate that Souverein et al’s results may be underestimating the risks secondary to study methodology.

Medicolegal Considerations

With the various and significant adverse events associated with CS use, proper medication stewardship is required to protect both the patient and provider. In a retrospective review of 700 legal cases, medication errors were found to be the third most common complaint. ³⁸ Of these, 55% were associated with CS use. In another review, long-term CS use has been implicated as the third most common medication associated with malpractice claims across all specialties. ³⁹ When considering CS-related lawsuits, judgments may be disproportionately high ($25 000-$8 100 000). ⁴⁰ Nash et al’s ⁴¹ retrospective medicolegal review specifically examined CS-related litigation in 83 cases. The complications leading to litigation were the following: AVN (39%), mood alterations and psychosis (16%), vision changes or infection (14%), non-AVN skeletal complications such as osteoporosis (12%), dermatologic side effects such as glucocorticoid-associated lipodystrophy and scarring (10%), and the remainder metabolic or other complications. The most commonly sued physicians were dermatologists (12%), primary care physicians (10%), and neurologists or neurosurgeons (6%).

With severe complications from CS and the frequency of CS-related medical malpractice claims, both counseling and informed consent should be conducted when prescribing moderate-high doses of CS.^41,42 Certainly, low dose short-term therapy may have a limited risk profile, but counseling should be conducted along with providing educational handouts. In addition, repeat CS therapy may be administered, so initial counseling and documentation would be beneficial. Obtaining informed consent may be time intensive, but it allows for better patient care/education and increased protection for the physician. In addition, some of this time could be recaptured with smart phrases used in electronic medical records and standardized consent forms, while providing improved documentation and risk management. ⁴²

Limitations

Like other summative review, this article is limited by the quality of the included papers. While all the articles were reviewed, the results are based on data drawn from peer-reviewed publications. Consequently, the results are inherently skewed due to publication and possible reporting biases and may not represent general population trends. In addition, pertinent papers may have been excluded from the initial review process due to inadequate breadth of search terms.