Table 2. The effect of ICS use in COPD on pneumonia risk in important clinical trials.
| Study | Population | Intervention (N) | Effect on pneumonia risk |
|
TORCH Calverly, 2007[48] |
Moderate to severe obstruction | Placebo (851) LABA (960) ICS (947) ICS/LABA (1 011) |
Risk of pneumonia over 3 years 18.3% and 19.6% in ICS/LABA and ICS
groups v. 12.3% and 13.3% in placebo and LABA groups (p<0.001). No increase in risk of death from pneumonia in ICS groups. |
|
SHINE Tashkin, 2008[49] |
Severe obstruction with exacerbations | Placebo (300) ICS (275) LABA (284) ICS/LABA (845) |
No increase in risk compared with placebo. |
|
INSPIRE Wedzicha, 2008[50] Calverly, 2011 reanalysis[51] |
Severe obstruction with exacerbations | ICS/LABA (658) LAMA (665) |
Hazard ratio of 1.94 for having pneumonia in ICS group (p=0.008); unchanged when analysis restricted only to patients with CXR. Higher risk in those with baseline severe dyspnoea and baseline raised CRP. |
| Dransfield et al., 2013[52] | COPD with exacerbations | LABA (818) | Incidence of non-fatal pneumonia increased in ICS group. |
| ICS/LABA (3 dose variations 820/806/811) | High-dose ICS discontinued. | ||
|
ILLUMINATE, LANTERN Pooled analysis Vogelmeier, 2013[53] Zhong, 2015[54] Vogelmeier, 2016[55] |
Moderate to severe obstruction | LABA/LAMA (259 and 372) ICS/LABA (264 and 369) |
Incidence of pneumonia 0.5% in LABA/LAMA group and 2.2% in the ICS/
LABA group (p=0.0074). Higher rate in more severe COPD. |
|
FORWARD Wedzicha, 2014[56] |
Severe obstruction with exacerbations | ICS/LABA (595) LABA (591) |
Incidence of pneumonia 1.8% in LABA group and 3.8% in ICS/LABA group. |
|
INSTEAD Rossi, 2014[57] |
Moderate obstruction | ICS/LABA (250) LABA (246) |
Incidence of pneumonia 0% in LABA group and 0.7% in ICS group, not statistically significant. |
|
WISDOM Magnussen, 2014[11] |
Severe obstruction with exacerbations | ICS/LABA/LAMA (1 243) LABA/LAMA (1 242) |
Hazard ratio for time to first exacerbation = 1.06. Incidence of pneumonia increased from 5.5% to 5.8% in ICS group, not statistically significant. |
|
SALFORD Vestbo, 2016[58] |
COPD with excerbations | ICS/LABA (1 396) Usual care (1 403) |
No excess pneumonia risk. |
|
TRILOGY Singh, 2016[15] |
Severe obstruction with exacerbations | ICS/LABA/LABA (687) ICS/LABA (681) |
Incidence of pneumonia in both groups = 3% |
|
FLAME Wedzicha, 2016[59] |
Moderate to severe obstruction with exacerbations | LABA/LAMA (1 675) ICS/LABA (1 679) |
Incidence of pneumonia 3.2% in LABA/LAMA group and 4.8% in ICS/ LABA group (p=0.02). |
|
IMPACT Wang, 2016[60] |
COPD with ICS use | Pneumonia cases (19 838) Pneumonia controls (74 849) |
Odds ratio for pneumonia = 1.25. Odds increased with increasing ICS dose. |
|
UPLIFT Tashkin, 2008[61] Morjaria, 2017 reanalysis[62] |
Moderate to severe obstruction | No ICS (2 292) Fluticasone roprionate (1 981) Other ICS (1 719) |
Incidence of pneumonia 5.6% in no-ICS group and 6.8% with ICS use, and was higher with fluticasone proprionate than ‘other ICS’ (p=0.012). |
|
OUTPUL Di Martino, 2014[63] Cascini, 2017 reanalysis[46] |
Moderate to severe COPD | Pneumonia cases (3 141) Pneumonia controls (12 564) |
Relative risk of pneumonia = 2.29 with current ICS use. Relative risk of pneumonia = 1.23 with past ICS use. Risk rates increased with increasing dose of ICS and with increasing age. ICS users had a numerically lower risk of death. |
|
FULFIL Lipson, 2017[17] |
Moderate to severe obstruction with exacerbations | ICS/LABA/LAMA (911) ICS/LABA (899) |
At 52 weeks the groups had a risk of pneumonia of 1.9% and 1.8%. |
ICS = inhaled corticosteroids
COPD = chronic obstructive pulmonary disease
LABA = long-acting beta agonist
LAMA = long-acting muscarinic antagonist
CXR = chest radiograph
CRP = C-reactive protein