In chronic obstructive pulmonary disease (COPD), treatment with inhaled corticosteroids (ICSs) in combination with long acting beta-2-agonists (LABA) or LABA/long-acting muscarinic antagonists (LAMA) is used in order to reduce exacerbations. Treatment with ICS is, however, associated with side effects such as oropharyngeal candidiasis, skin thinning or easy bruising and pneumonia. The aim of this review was to investigate when to use ICS in COPD and to compare the effectiveness and safety of different ICSs. Studies comparing the effect of ICS/LABA and LABA/LAMA on exacerbations have shown divergent results, whereas most studies comparing ICS/LABA/LAMA (triple therapy) with LABA/LAMA have reported fewer exacerbations with triple therapy. Several investigations have shown that the number of eosinophils in blood predicts whether a patient will benefit from treatment with ICS. There is also data indicating that ICS has a small but significant positive effect on lung function decline and decrease mortality. There are four observational studies showing a better effect on exacerbations with budesonide/formoterol than fluticasone propionate/salmeterol and three observational studies showing less risk of pneumonia with budesonide than fluticasone propionate. Studies comparing the effect and safety of other ICSs such as fluticasone furoate and beclomethasone are too few to draw firm conclusions from. In conclusion, ICS together with LABA or LABA/LAMA reduces the risk of exacerbations in COPD. The indication of using ICS in COPD is stronger if the patient has increased blood eosinophils levels. There are data indicating that the choice of ICS matters, with studies showing a better effect-safety profile with budesonide compared to fluticasone propionate whereas it is not possible to make benefit-risk comparisons between the other licensed ICSs.