Elsevier

Cytokine

Volume 136, December 2020, 155263
Cytokine

Dynamics of cytokines, immune cell counts and disease severity in patients with community-acquired pneumonia – Unravelling potential causal relationships

https://doi.org/10.1016/j.cyto.2020.155263Get rights and content

Highlights

  • Decrease in IL-6 levels and increase in thrombocytes is slower in elderly patients.

  • IL-8, VCAM-1, and IL-6 levels associate with disease severity.

  • IL-6 and bilirubin levels and rates of change correlate.

  • Lagged relationships between pairs of cytokines can be inferred using time series.

Abstract

Background

Community acquired pneumonia (CAP) is a severe and often rapidly deteriorating disease. To better understand its dynamics and potential causal relationships, we analyzed time series data of cytokines, blood and clinical parameters in hospitalized CAP patients.

Methods

Time series data of 10 circulating cytokines, blood counts and clinical parameters were related to baseline characteristics of 403 CAP patients using univariate mixed models. Bivariate mixed models were applied to analyze correlations between the time series. To identify potential causal relationships, we inferred cross-lagged relationships between pairs of parameters using latent curve models with structured residuals.

Results

IL-6 levels decreased faster over time in younger patients (Padj = 0.06). IL-8, VCAM-1, and IL-6 correlated strongly with disease severity as assessed by the sequential organ failure assessment (SOFA) score (r = 0.49, 0.48, 0.46, respectively; all Padj < 0.001). IL-6 and bilirubin correlated with respect to their mean levels and slopes over time (r = 0.36 and r = 0.46, respectively; Padj < 0.001). A number of potential causal relationships were identified, e.g., a negative effect of ICAM-1 on MCP-1, or a positive effect of the level of creatinine on the subsequent VCAM-1 concentration (P < 0.001).

Conclusions

These results suggest that IL-6 trajectories of CAP patients are associated with age and run parallel to bilirubin levels. The time series analysis also unraveled directed, potentially causal relationships between cytokines, blood parameters and clinical outcomes. This will facilitate the development of mechanistic models of CAP, and with it, improvements in treatment or surveillance strategies for this disease.

Trial registration: clinicaltrials.gov NCT02782013, May 25, 2016, retrospectively registered.

Introduction

Community-acquired pneumonia (CAP) is a serious disease with high morbidity and mortality, often requiring hospitalization. Even in case of proper clinical treatment, the disease course can be highly dynamic with rapid serious deterioration requiring intensive care treatment. Cytokines released during the inflammatory response of the host to the infection have been shown to be predictors of treatment failure and mortality [1], [2], and factors influencing the patterns of cytokine concentrations in CAP have been investigated [3]. However, the complex interaction of cytokines, immune cells and disease severity is still only partly understood and analyses in humans are hampered by the lack of detailed time series data. While cross-sectional data only allow the analysis of associations, time series data can be used to analyze not only the mean levels as is usually done, but also the rates of change over time of cytokines and other parameters and their predictive value regarding disease severity or single organ failure. A further limitation of the current research on CAP is that correlations computed across patients do not necessarily reflect the dynamic relations that link two or more parameters within patients over time [4], [5], and do not provide information about the direction of the inferred relationships, i.e., the causality between players of the immune response.

In our PROGRESS study, we close this gap by collecting comprehensive characteristics of CAP patients at baseline and over 4–5 days of hospitalization, including daily measurements of several cytokines, immune cell counts and parameters of disease severity included in the Sequential organ failure assessment score (SOFA), which was identified as an excellent operationalization of CAP severity [6]. Making use of this rich resource, in the present analysis of these data, we aim at answering the following questions of increasing complexity: First, we are interested in which time-invariant factors (e.g., age, sex, BMI) correlate with cytokine dynamics, immune cells, and SOFA parameters. This will reveal potential confounders of associations of cytokines with disease progression or provide hints towards modifiable or physiological risk factors such as age. Second, we study correlations between the time courses of any two cytokines, immune cell counts or SOFA parameters to identify relationships between immune response and organ function over time. Third, we establish potentially causal lagged relationships between parameters, in the sense that the value of one parameter influences the value of another parameter on the next day. This insight is relevant, e.g., to identify valid therapy targets.

Section snippets

Study subjects

Study subjects were recruited in the framework of the PROGRESS study (clinicaltrials.gov identifier: NCT02782013), which is a multi-center clinical observational study of hospitalized patients with CAP. In brief, these patients were recruited, if they had a working diagnosis of pneumonia, were 18 or more years old, and they or their legal representatives signed a consent to participate. To exclude nosocomial infections (hospital-acquired pneumonia), patients staying in hospital during the

Clinical characteristics of study population

We analyzed data of 403 CAP patients for whom measurements of circulating cytokines were available. Baseline characteristics of these patients are shown in Table 1.

Factors determining SOFA components and cytokines

First, we analyzed each of the repeatedly measured parameters (10 cytokines, SOFA components and blood parameters, see Additional file 2, Table S1) separately regarding their dependence on a set of patient characteristics and risk factors for severe disease courses. These factors include age, sex, BMI, (number of years of) smoking,

Discussion

Understanding disease dynamics in pneumonia patients is of high importance to develop successful surveillance and intervention strategies but requires appropriate time series data of patients. Here, we present and analyze data of the PROGRESS cohort for which a sufficiently rich set of time series of cytokines, immune cells and disease states is available.

In the first step of the present study, we analyzed associations of several baseline characteristics of our CAP patients with their time

Ethics approval and consent to participate

Ethics approval was issued by the ethics committee of the University of Jena (2403-10/08) and by locally responsible ethics committees of each study center.

Availability of data and materials

The data that support the findings of this study are available from PROGRESS steering committee but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of PROGRESS steering committee.

Funding

The PROGRESS study is funded by the German Federal Ministry of Education and Research, grant numbers 01KI07110 (Giessen), 01KI07111 (Jena), 01KI07113 (Leipzig), 01KI07114 (Berlin), 01KI1010I (Leipzig), and 01KI1010D (Greifswald). The German Federal Ministry of Education and Research has no influence on the design of the study, on collection, analysis, and interpretation of data, or on writing the manuscript.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We thank the patients, their relatives and legal guardians very much for participation in the PROGRESS study and The PROGRESS Study Group for recruiting and data collection.

References (53)

  • P.J. Curran et al.

    The separation of between-person and within-person components of individual change over time: A latent curve model with structured residuals

    J. Consult. Clin. Psychol.

    (2014)
  • P. Ahnert et al.

    Sequential organ failure assessment score is an excellent operationalization of disease severity of adult patients with hospitalized community acquired pneumonia - results from the prospective observational PROGRESS study

    Crit. Care

    (2019)
  • P. Ahnert et al.

    PROGRESS - prospective observational study on hospitalized community acquired pneumonia

    BMC Pulm. Med.

    (2016)
  • J.L. Vincent et al.

    The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure

    On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine, Intensive Care Med.

    (1996)
  • G. Nieman et al.

    A Two-Compartment Mathematical Model of Endotoxin-induced Inflammatory and Physiologic Alterations in Swine

    Crit. Care Med.

    (2012)
  • H.H. Yiu et al.

    Dynamics of a Cytokine Storm

    PLoS ONE

    (2012)
  • G. An et al.

    Toward computational identification of multiscale “tipping points” in acute inflammation and multiple organ failure

    Ann. Biomed. Eng.

    (2012)
  • C.C. Chow et al.

    The acute inflammatory response in diverse shock states

    Shock

    (2005)
  • Wakeland W, Mackowsky L, An G., A hybrid simulation for studying acute inflammatory response., in: M.J. Ades (Ed.),...
  • K. Oikonomopoulou et al.

    Interactions between coagulation and complement–their role in inflammation

    Semin. Immunopathol.

    (2012)
  • M.A. Kovach et al.

    The function of neutrophils in sepsis

    Curr. Opin. Infect. Dis.

    (2012)
  • A. Shah et al.

    A Random-Effects Model for Multiple Characteristics With Possibly Missing Data

    J. Am. Stat. Assoc.

    (1997)
  • R.C. MacCallum et al.

    Studying Multivariate Change Using Multilevel Models and Latent Curve Models

    Multivariate Behav. Res.

    (1997)
  • H. Chakraborty et al.

    Estimating correlation by using a general linear mixed model: evaluation of the relationship between the concentration of HIV-1 RNA in blood and semen

    Stat. Med.

    (2003)
  • S. Fieuws et al.

    Random-effects models for multivariate repeated measures

    Stat. Methods Med. Res.

    (2007)
  • Y. Benjamini et al.

    Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing, Journal of the Royal Statistical Society

    Series B

    (1995)
  • Cited by (6)

    View full text