| Study | Reason for exclusion |
|---|---|
| Chan 2014 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Chen 2010 | Case series |
| Chien 2010 | No useable data relating to corticosteroid use (only P values) |
| Chotpitayasunondh 2005 | Case series |
| Diaz 2012 | Overlapping cohort, superceded by later analysis |
| Dominguez‐Cherit 2016 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Fujikura 2014 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Gao 2013 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Garnacho‐Montero 2013 | Overlapping populations* |
| Hien 2009 | Overlapping populations* |
| Hong 2013 | Overlapping populations* |
| Hu 2013 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Jung 2011 | Overlapping populations* |
| Leon 2017 | Overlapping populations* |
| Martin‐Loeches 2013 | Overlapping populations* |
| McKenna 2013 | Overlapping populations* |
| Ning 2014 | Fewer than 10 participants in study |
| Okur 2013 | Overlapping populations* |
| Perez‐Padilla 2009 | Case series |
| Rios 2011 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Segaloff 2017 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Singh 2017 | No outcome data clearly comparing corticosteroid treatment versus no corticosteroid treatment |
| Smud 2010 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment |
| Sun 2010 | Case series |
| Torres 2012 | No data on clinical outcomes |
| Vilar‐Compte 2018 | No outcome data comparing corticosteroid treatment versus no corticosteroid treatment; corticosteroid use within 30 days of infection only |
| Yale 2015 | No data on clinical outcomes |
| Yang 2015a | Overlapping populations* |
| Zhang 2011 | Outcomes unclear |
| Zhang 2013 | Overlapping populations* |
*Overlapping populations refers to studies where data from cohorts included in the review were duplicated, either due to multiple reporting of the same cohort, or inclusion of the cohort within the study population of a larger study.