Format

Send to

Choose Destination
Sci Transl Med. 2015 Apr 29;7(285):285ra61. doi: 10.1126/scitranslmed.aaa3636.

Trauma in silico: Individual-specific mathematical models and virtual clinical populations.

Author information

1
Immunetrics Inc., Pittsburgh, PA 15219, USA.
2
Department of Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA.
3
Critical Care Medicine, University of Pittsburgh, Pittsburgh, PA 15261, USA.
4
Department of Mathematics, University of Pittsburgh, Pittsburgh, PA 15260, USA.
5
Department of Surgery, Upstate Medical University, Syracuse, NY 13210, USA.
6
Department of Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA. Center for Inflammation and Regenerative Modeling, McGowan Institute for Regenerative Medicine, Pittsburgh, PA 15219, USA.
7
Department of Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA. Center for Inflammation and Regenerative Modeling, McGowan Institute for Regenerative Medicine, Pittsburgh, PA 15219, USA. vodovotzy@upmc.edu.

Abstract

Trauma-induced critical illness is driven by acute inflammation, and elevated systemic interleukin-6 (IL-6) after trauma is a biomarker of adverse outcomes. We constructed a multicompartment, ordinary differential equation model that represents a virtual trauma patient. Individual-specific variants of this model reproduced both systemic inflammation and outcomes of 33 blunt trauma survivors, from which a cohort of 10,000 virtual trauma patients was generated. Model-predicted length of stay in the intensive care unit, degree of multiple organ dysfunction, and IL-6 area under the curve as a function of injury severity were in concordance with the results from a validation cohort of 147 blunt trauma patients. In a subcohort of 98 trauma patients, those with high-IL-6 single-nucleotide polymorphisms (SNPs) exhibited higher plasma IL-6 levels than those with low IL-6 SNPs, matching model predictions. Although IL-6 could drive mortality in individual virtual patients, simulated outcomes in the overall cohort were independent of the propensity to produce IL-6, a prediction verified in the 98-patient subcohort. In silico randomized clinical trials suggested a small survival benefit of IL-6 inhibition, little benefit of IL-1β inhibition, and worse survival after tumor necrosis factor-α inhibition. This study demonstrates the limitations of extrapolating from reductionist mechanisms to outcomes in individuals and populations and demonstrates the use of mechanistic simulation in complex diseases.

PMID:
25925680
DOI:
10.1126/scitranslmed.aaa3636
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for HighWire
Loading ...
Support Center