Understanding the inflammatory cytokine response in pneumonia and sepsis: results of the Genetic and Inflammatory Markers of Sepsis (GenIMS) Study

Abstract

Background: Severe sepsis is common and frequently fatal, and community-acquired pneumonia (CAP) is the leading cause. Although severe sepsis is often attributed to uncontrolled and unbalanced inflammation, evidence from humans with infection syndromes across the breadth of disease is lacking. In this study we describe the systemic cytokine response to pneumonia and determine if specific patterns, including the balance of proinflammatory and anti-inflammatory markers, are associated with severe sepsis and death.

Methods: This is a cohort study of 1886 subjects hospitalized with CAP through the emergency departments in 28 US academic and community hospitals. We defined severe sepsis as CAP complicated by new-onset organ dysfunction, following international consensus conference criteria. We measured plasma tumor necrosis factor, IL-6 (interleukin 6), and IL-10 levels daily for the first week and weekly thereafter. Our main outcome measures were severe sepsis and 90-day mortality.

Results: A total of 583 patients developed severe sepsis (31%), of whom 149 died (26%). Systemic cytokine level elevation occurred in 82% of all subjects with CAP. Mean cytokine concentrations were highest at presentation, declined rapidly over the first few days, but remained elevated throughout the first week, beyond resolution of clinical signs of infection. Cytokine levels were highest in fatal severe sepsis and lowest in CAP with no severe sepsis. Unbalanced (high/low) cytokine patterns were unusual (4.6%) and not associated with decreased survival. Highest risk of death was with combined high levels of the proinflammatory IL-6 and anti-inflammatory IL-10 cytokine activity (hazard ratio, 20.5; 95% confidence interval, 10.8-39.0) (P<.001).

Conclusions: The circulating cytokine response to pneumonia is heterogeneous and continues for more than a week after presentation, with considerable overlap between those who do and do not develop severe sepsis. Unbalanced activation is uncommon, and mortality is highest when both proinflammatory and anti-inflammatory cytokine levels are high.

Figure 1

Subject disposition for the entire Genetic and Inflammatory Markers of Sepsis (GenIMS) cohort. CAP indicates community-acquired pneumonia.

Figure 2

Mortality rates for subjects with community-acquired pneumonia with and without sepsis. Severe sepsis was defined as sepsis with organ failure (Sepsis-related Organ Failure Assessment score ≥3 in any organ system).

Figure 3

Mean plasma cytokine concentrations (A) and proportion of subjects with community-acquired pneumonia (CAP) with elevated plasma concentrations (B). IL-6 indicates interleukin 6; ln, natural log; and TNF, tumor necrosis factor. A, Mean (ln) plasma cytokine concentrations for subjects with CAP by outcome and over time. B, Percentage of subjects with CAP with elevated concentrations compared with the normal range for each molecule by outcome. Elevated was defined for IL-6 as greater than 5.9 pg/mL; for IL-10, greater than 9.1 pg/mL; and for TNF, greater than 8.1 pg/mL.

Figure 4

Patterns of cytokine response to infection and resulting patterns of outcomes. A, Trajectory analysis of IL-6 (interleukin 6) and IL-10 concentration data suggests 3 distinct patterns of cytokine response during the first 7 days in the hospital. Combinations of 1, 2, 3, or 4 groups with quadratic, cubic, or linear trends of the cytokine levels, irrespective of outcome, were analyzed. For both IL-6 and IL-10, 3-group models (high, medium, and low cytokine concentration) with a quadratic trend for each group exhibited the best fit. ln Indicates natural log. B (top), The Kaplan-Meier analysis includes survival curves that represent different combinations of the IL-6/IL-10 cytokine responses. Although there are 9 such combinations, some groups have been collapsed together because of similar survival rates. Five distinct patterns of outcome were observed. The numbers of subjects in each group are as follows: low IL-6, n=667 (35.4%); medium IL-6, n=1003 (53.2%); high IL-6, n=216 (11.4%); low IL-10, n=1255 (65%); medium IL-10, n=509 (27%); and high IL-10, n=152 (8%).

Figure 5

Risk of death among patients with community-acquired pneumonia according to baseline characteristics and plasma cytokine concentrations. Results are from a Cox proportional hazards model for mortality. Point estimates are shown along with their 95% confidence intervals. While race and sex did not seem to contribute to the risk of death, age, comorbidity, and high concentrations of IL-6 (interleukin 6) or IL-10, even combined with low levels of the other cytokine, were associated with death. The combination of medium-level IL-6 and IL-10 also increased the risk of death severalfold, while high levels of both cytokines increased the risk of death by more than 20 times. Age indicates increase in age by 10 years; CCI, Charlson Comorbidity Index; in all low/medium/high concentration combinations, the first term designates IL-6 groups and the second term IL-10 groups, with low/low used as the referent for the hazard ratio of the rest.