Endogenous IgG hypogammaglobulinaemia in critically ill adults with sepsis: systematic review and meta-analysis

Manu Shankar-Hari, Nicholas Culshaw, Benjamin Post, Eduardo Tamayo, David Andaluz-Ojeda, Jesús F. Bermejo-Martín, Sebastian Dietz, Karl Werdan, Richard Beale, Jo Spencer, Mervyn Singer

Plasma immunoglobulin concentrations are acutely altered in critically ill patients with sepsis. However, the association between immunoglobulin levels on the day of sepsis diagnosis and subsequent mortality is inconsistent. Systematic review of studies that report immunoglobulin measurements and mortality among adults with sepsis managed in a critical care setting. Fixed and random effect meta-analyses were conducted using low IgG levels as primary exposure and acute mortality as the primary outcome. Both variables were used as defined in individual studies. The prevalence of a low immunoglobulin G (IgG) concentration on the day of sepsis diagnosis was variable [58.3 % (IQR 38.4-65.5 %)]. Three cut-off points (6.1, 6.5 and 8.7 g/L) were used to define the lower limit of IgG concentrations in the included studies. A subnormal IgG level on the day of sepsis diagnosis was not associated with an increased risk of death in adult patients with severe sepsis and/or septic shock by both fixed and random effect meta-analysis (OR [95 % CI] 1.32 [0.93-1.87] and 1.48 [0.78-2.81], respectively). This systematic review identifies studies of limited quality reporting heterogeneous sepsis cohorts with varying lower limits of normal for IgG. Although our data suggest that a subnormal IgG measurement on the day of sepsis diagnosis does not identify a subgroup of patients with a higher risk of death, further studies are needed to confirm or refute this finding, and whether optimal cut-offs and time windows can be defined for IgG measurement. This would determine whether patients receiving intravenous immunoglobulin therapy for sepsis could be stratified using IgG levels.