The most common psychiatric manifestations include agitation, anxiety, hypomania, insomnia, irritability, mood lability, and tearfulness, reported to occur in 27.6 % (range 13–62 %) of individuals. Severe reactions include mania, depression, or a mixed state which have been reported in 5.7 % (range 1.6–50 %) [5, 6].

Corticosteroid dosage has been found to be the most significant risk factor associated with psychiatric reactions. When symptoms were analyzed based on dose, there was a 1.3 % incidence in those patients receiving a daily prednisone dose ≤40 mg, a 4.6 % incidence in those receiving 41–80 mg of prednisone, and a 18.4 % incidence in those receiving >80 mg [7]. Reduction of the dose resulted in resolution of symptoms in all cases. Past reactions are not predictive of a future reaction, nor is past tolerance predictive of future tolerance [5]. No correlation between a history of psychiatric illness and a psychiatric reaction to corticosteroids has been established [8].

Cataract formation and increased intraocular pressure (glaucoma) are the most commonly encountered ophthalmologic side effects. It has been proposed that steroid molecules bond covalently with the lysine residues of the lens, leading to opacities, or that corticosteroids inhibit the sodium-potassium pump in the lens, leading to coagulation of lens proteins. Most studies report doses of 10 mg or more daily for at least one year before the onset of cataract formation [9].

The exact mechanism by which corticosteroids cause glaucoma is unknown; corticosteroids may exert a negative impact on the trabecular meshwork, leading to fluid retention and elevated pressures. Between 18 and 36 % of the population will develop at least a moderate (5 mmHg or greater) increase in pressure with prolonged steroid treatment [10]. Risk factors include a history of open-angle glaucoma, diabetes mellitus, high myopia, rheumatoid arthritis, hypertension, migraine headaches, and first-degree relatives with open-angle glaucoma [9, 10].

Corticosteroids increase blood sugars by increasing hepatic gluconeogenesis, decreasing glucose uptake in peripheral tissues, and decreasing the ability of adipocytes and hepatocytes to bind insulin. This effect can occur within hours of beginning therapy and tends to decrease with prolonged use [11]. Upon cessation of corticosteroids, the inhibition of glucose uptake and metabolism in peripheral tissues usually returns to normal [11].

The mechanism by which corticosteroids decrease inflammation may also lead to immunosuppressive effects. Although circulating neutrophils increase as a result of enhanced release from bone marrow and reduced migration from blood vessels, other leukocytes decrease due to migration from the vascular bed to the lymphoid tissue [12]. Corticosteroids impact neutrophil function by reducing their bactericidal activity as well as limit the function of macrophages and other antigen-presenting cells [12, 13].

A meta-analysis found that patients who received a daily dose of less than 10 mg per day or a cumulative dose of less than 700 mg of prednisone did not have an increased rate of infectious complications [14]. A second meta-analysis of over 8,700 patients reported bacterial sepsis occurred 1.5 times more frequently in patients using corticosteroids than in those using placebo (P <0.01). Mean daily dose was the equivalent of 35 mg of prednisone and the mean total dose was 2,200 mg of prednisone [15].

Despite the commonly held perception that steroid use increases the risk of peptic ulcer disease, several large meta-analyses of randomized, placebo-controlled trials have failed to show this association [16], although the studies did find that patients using corticosteroids complained of peptic ulcer-type symptoms more frequently than control patients [15]. In addition to gastric issues, pancreatitis has been reported with the use of corticosteroids, though the exact incidence and the mechanism are unknown [17].

Exogenous steroids increase the circulating corticosteroid levels, which can lead to a negative feedback on the hypothalamic-pituitary-adrenal axis at the levels of both the hypothalamus and the pituitary gland. This effect can lead to a decrease production of both corticotropin-releasing hormone from the hypothalamus and corticotropin or adrenocorticotropic hormone from the pituitary gland [18]. Decreased production of adrenocorticotropic hormone then leads to decreased cortisol secretion from the adrenal cortex.

The dose of exogenous corticosteroids that can lead to adrenal suppression is highly variable. The incidence of clinically evident adrenal insufficiency is unknown, yet it is believed to be much lower than the incidence based on objective measures [17].

The role of corticosteroids in bone loss is well described and may occur through several different mechanisms. They reduce intestinal calcium absorption and increase urinary calcium excretion. This stimulates parathyroid hormone which sacrifices bone mass, releasing calcium into the circulation [10, 19].

Corticosteroids also suppress the production of adrenal androgens and can cause apoptosis of osteoblasts and osteocytes. This effect can occur within 1 month of use; however it slows after 6–12 months and rapidly reverses with cessation of the corticosteroid [20, 21]. Several studies have demonstrated that supplemental calcium, vitamin D, and bisphosphonates can help reduce the steroid-induced bone loss [17].

Corticosteroid use has been associated with avascular necrosis or osteonecrosis, usually in the head of the femur, although all bones may be affected [19]. The etiology is not completely understood but is due to impaired perfusion of the bone, either as a result of embolic events, hyperviscous blood, or increased pressure in the femoral head, resulting in decreased blood flow [19, 22, 23].

One review identified 15 patients treated with a single course of corticosteroids who developed osteonecrosis of the femoral head [24]. Ages ranged from 20 to 41 years (mean 32.2 years), mean cumulative dose was 850 mg of prednisone (range 290–3,300 mg), and the mean duration of therapy was 20.5 days (range 6–39 days). The mean time from treatment to symptoms in the study was 16.6 months (range 6–33). A second series of 1,352 patients treated with corticosteroids identified 4 cases of avascular necrosis, a risk of 0.3 % [25]. The mean age was 26 years (range 21–31), the mean cumulative dose was equivalent to 673 mg of prednisone (range 389–990 mg of prednisone equivalents), and the mean duration was 20 days (15–27 days). The time to onset of symptoms in this group ranged from 4 to 27 months, with a mean of 14.5 months.

Despite the widespread use of corticosteroids in CRS, the data is remarkably lacking. Only four studies have evaluated the benefit of oral corticosteroids in patients with CRS without nasal polyps. Unfortunately, most include oral corticosteroids used in combination with other interventions such as antibiotics, topical steroids, and saline irrigations. Three of the four included CRS patients with and without nasal polyps.

The only study that evaluated the effect of oral corticosteroids alone on CRS symptoms was done by Ikeda et al. [26]. Twelve patients with CRS without nasal polyps, who had failed topical nasal steroids, underwent olfactory testing before and after a 10–14-day taper of prednisone. They found significant improvements in both detection and recognition thresholds following the prednisone course. Improved olfactory function occurred in ten patients with eight patients having persistent improvement for many months and only two reporting no improvement in olfactory function.