Skip to main content

Genome-wide transcription profiling of human sepsis: a systematic review



Sepsis is thought to be an abnormal inflammatory response to infection. However, most clinical trials of drugs that modulate the inflammatory response of sepsis have been unsuccessful. Emerging genomic evidence shows that the host response in sepsis does not conform to a simple hyper-inflammatory/hypo-inflammatory model. We, therefore, synthesized current genomic studies that examined the host response of circulating leukocytes to human sepsis.


Electronic searches were performed in Medline and Embase (1987 to October 2010), supplemented by additional searches in multiple microarray data repositories. We included studies that (1) used microarray, (2) were performed in humans and (3) investigated the host response mediated by circulating leukocytes.


We identified 12 cohorts consisting of 784 individuals providing genome-wide expression data in early and late sepsis. Sepsis elicited an immediate activation of pathogen recognition receptors, accompanied by an increase in the activities of signal transduction cascades. These changes were consistent across most cohorts. However, changes in inflammation related genes were highly variable. Established inflammatory markers, such as tumour necrosis factor-α (TNF-α), interleukin (IL)-1 or interleukin-10, did not show any consistent pattern in their gene-expression across cohorts. The finding remains the same even after the cohorts were stratified by timing (early vs. late sepsis), patient groups (paediatric vs. adult patients) or settings (clinical sepsis vs. endotoxemia model). Neither a distinctive pro/anti-inflammatory phase nor a clear transition from a pro-inflammatory to anti-inflammatory phase could be observed during sepsis.


Sepsis related inflammatory changes are highly variable on a transcriptional level. We did not find strong genomic evidence that supports the classic two phase model of sepsis.


Sepsis is characterised by a bewildering array of abnormalities in both innate and adaptive immune systems. To help explain this complex pathophysiology, a two-phase model has been used by investigators. This model postulates that sepsis consists of an initial phase of systemic inflammatory response syndrome, followed by a later phase of compensatory anti-inflammatory response syndrome. This two-phase model has been the reigning paradigm under which scientists develop new therapeutic agents, with new drugs targeting either the pro-inflammatory or the anti-inflammatory arm of the host response. However, clinical trials have consistently failed to demonstrate any survival benefit of drugs that target the inflammation pathway. As a result, concerns have been raised regarding the validity of treating sepsis simply as a pro-inflammatory or anti-inflammatory phenomenon.

Complicating this uncertainty is the limited evidence to verify the two-phase model. Cytokine studies have been the mainstay evidence that provide support for the inflammation-based model. However, increasingly conflicting findings have emerged from recent cytokine studies [13]. Furthermore, it is often difficult to determine the exact nature of the host response (for example, pro-inflammatory versus anti-inflammatory) on the basis of cytokine measurement alone, which is highly variable depending on the choice of the cytokine used and the timing of the measurements.

Given the limitations of the protein level studies, we assessed the validity of the inflammation-based model using transcriptional level data. Genome-wide transcriptional studies have recently emerged as a powerful investigational tool to study complex disease [4]. These studies avoid the selection bias inherent in most cytokine studies, where only a small number of pre-selected genes can be examined. In this systematic review, we synthesized genomic data of recent microarray studies where the transcriptional changes of circulating leukocytes were examined in both experimental and clinical sepsis in humans.

Materials and methods

Search strategy and selection criteria

We searched in Medline and Embase, without language restriction, all publications on gene-expression studies between January 1987 and October 2010. In 1987 DNA array technology was first described, hence this year formed the starting point of our search [5]. We hand-searched the reference lists of every primary study for additional publications. Further searches were performed by reviewing journal editorials and review articles.

The search strategy used the following search terms: (1) "gene-expression profiling", (2) "microarray analysis", (3) "transcription profiling", (4) "cluster analysis", (5) "Affymetrix", (6) "GeneChip", (7) "sepsis", (8) "sepsis syndrome", (9) "septicaemia", (10) "bacteraemia", (11) "septic shock", (12) "infection", (13) "systemic inflammatory response syndrome", (14) "SIRS", (15) "systemic inflammation", (16) "endotoxin".

We also performed searches in public repositories of microarray datasets, including the National Centre for Biotechnology Information (Gene Expression Omnibus), the European Bioinformatics Institute (ArrayExpress), and the Centre for Information Biology Gene Expression Database (CIBEX). Datasets from microarray database were then cross-referenced with publications retrieved from Medline and Embase. Only datasets published as full reports were included in the final analysis.

We included a broad spectrum of gene-expression studies, including ones that are (1) cross-sectional or longitudinal design, (2) on different microarray platforms, (3) on whole blood or purified leukocytes, (4) in healthy volunteers or infected human hosts, and (5) paediatric or adult patients. As we only sought data on a genome-wide scale, we have excluded studies that assayed only a small number of genes, such as (1) Northern blot or PCR, (2) single gene or individual pathway studies, (3) proteomic studies, and (4) single-nucleotide polymorphism studies. We included custom designed microarrays only if such arrays are designed to study changes in inflammation pathways. Since we were interested in host response on a systematic level, as reflected by circulating leukocytes, we have excluded studies that (1) focused on resident immune cells such as alveolar macrophages or lymphoid tissue cells, and (2) used solid organ tissues such as spleen or liver.

Data extraction

We extracted study level data according to a pre-specified template, which included participant demographics, country of origin, clinical setting and inclusion criteria. A separate template was used to collect details of microarray experiments, including sample collection procedures, cell separation techniques, target cell types, methods used to extract ribonucleic acids, cDNA synthesis and hybirdization, microarray platforms used, number of probe set on arrays, microarray data processing and normalization methods. We extracted the signature gene list from each published report or from the accompanied data file in the journal websites. Where available, results of functional analyses were also extracted. These included results of cluster analyses, principle component analyses or pathway analyses.

Quality assessment

We performed a quality assessment of each study based on criteria modified from published guidelines on the statistical analysis and reporting of microarray data [6]. The assessment was performed using a 14-item checklist covering three quality domains including data acquisition (three items), statistical analysis (six items) and validation of microarray findings (five items).

Data synthesis

We performed a narrative synthesis on genomic data extracted from each study. First, individual genes from the gene list of primary studies were manually annotated by cross-referencing with publicly available gene nomenclatures databases (for example, Genebank, Locuslink, Affymetrix gene identifiers). Where a gene list was not available, findings on functional analyses reported by the original authors were used. These included cluster analysis or gene network analysis performed on the original microarray data. All results were then collated and presented in evidence tables. Due to the heterogeneous nature of the included studies, meta-analysis of the microarray data was not performed.


The literature search yielded 7,548 citations in electronic databases and 142 datasets in microarray data repositories. Of these, 12 patient cohorts met the inclusion criteria and were included in the final analysis (Figure 1).

Figure 1

Study selection.

Clinical characteristics of the included studies are summarized in Table 1. The cohorts were drawn from a broad spectrum of clinical settings including hospital wards, intensive care units and university research centres. The majority of the study participants were critically ill patients diagnosed with sepsis or infection. Among patients with sepsis, a full range of sepsis syndrome was represented (for example, sepsis, severe sepsis and septic shock).

Table 1 Summary of studies characteristics

Details of the microarray experiments are summarized in Tables 2. The target tissue was either whole blood or purified leukocytes isolated from whole blood (for example, neutrophils or mononuclear cells). Affymetrix was the most common microarray platform used. In total, gene-expression profiling of 784 individuals were performed across four different microarray platforms.

Table 2 Microarray experiments in included studies

Results on the assessment of the methodological quality of each microarray study are presented in Table 3. Just over half of the studies fulfilled the MIAMI criteria (Minimum Information About Microarray Experiment, published guidelines on the design, conducting, analysis and reporting of the microarray experiments) [7]. Only seven studies performed internal validation of microarray data and independently validated their reported gene lists in separate data sets. Raw microarray data are available in only 7 out of the 12 cohorts.

Table 3 Methodological quality of microarray experiments

A wide range of statistical approaches were used by the included studies. Table 3 provides detailed information on the reporting of the statistical methods by each study. Most studies provided details on the method used for normalization. Normalization is a data processing method that ensures only genes, which are truly differentially expressed between phenotypes of interest, are detected, instead of those caused by experimental artefacts or variation in the microarray hybirdization process. Different statistical approaches were used for detecting statistically significant genes, depending on the study design used in each cohort (Table 3). Multiple testing corrections were used by most studies to minimize a false positive rate in the significant genes (Table 3). However, variance estimation was poorly reported in most studies. A variety of variance estimation techniques were used by the included studies; but details were lacking in most studies (conventional t-statistics based variance estimation methods under-estimate the true variance of microarray data, so several variance estimation methods for microarray data have been developed). Overall, the reporting of statistical methods was variable among studies.

Pathogen recognition

Sepsis activates pathogen recognition pathways in host leukocytes. This is evident in most studies. Up-regulation of pathogen recognition receptors, such as toll-like receptors and CD14, was observed (Table 4). This was accompanied by the activation of signal transduction pathways, a process essential for subsequent transcription of immune response genes. The signal transduction pathways include nuclear factor kappa-B (NK-kβ), mitogen activated protein kinase (MAPK), Janus kinase (JAK) and transducer and activator of transcription protein (STAT) pathways (Table 4). The up-regulation of both pathogen recognition and signal transduction pathway genes was observed in most cohorts, including experimental and clinical sepsis, paediatric and adult patients, early and late sepsis.

Table 4 Gene-expression changes in pathogen recognition

Inflammatory response

In contrast to the above findings, changes in inflammatory pathways were much less consistent. A distinctive pro-inflammatory or anti-inflammatory phase, as depicted in the classic sepsis model, was not seen during any stage of sepsis. The early, transient rise in some pro-inflammatory mediators was evident only in a minority of studies (Table 5). In some studies, the expression of anti-inflammatory genes dominated over pro-inflammatory genes. In others, changes in inflammatory genes were noticeably absent. No studies demonstrated a clear transition from a pro-inflammatory phase to an anti-inflammatory phase during the course of sepsis. Overall, the transcriptional changes in inflammation-related genes are highly variable in most cohorts.

Table 5 Gene-expression in inflammation and immunity

We next identified, in each cohort, genes that are well known in the sepsis literature (for example, tumour related factor (TNF), interleukin (IL)-1, IL-8, IL-10 and TGF-beta). In particular, we were interested to see whether there was any systematic difference in their expression patterns between cohorts (for example, early sepsis vs. late sepsis). We restricted our analysis to cohorts of comparable microarray platforms (for example, Affymetrix) and target tissues (for example, whole blood). In this analysis, we found no consistent pattern of gene expression in any of the well-established markers of inflammation (pro-inflammatory or anti-inflammatory). Further analyses by stratifying cohorts based on patient groups (paediatric vs. adults) or presentation (pneumonia or non-specified sepsis) yielded similarly negative findings.

Experimental sepsis

A major limitation of the above studies is that the findings could be confounded by the variable time from onset of sepsis (since the precise time of infection is often unknown). We, therefore, performed a separate analysis on studies that used an in vivo endotoxin challenge model. In these studies, endotoxin was injected into healthy volunteers and blood sampling was performed at regular intervals (up to 24 hours). Consequently, the exact time of onset of infection is known and the effect of timing on gene-expression changes can be clearly defined. We found three endotoxin challenge studies in our data set [810]. All three studies used similar experimental protocols. The analysis showed that endotoxin challenge elicited an activation of pathogen recognition and signal transduction pathways, similar to findings in other non-endotoxemia studies. However, the findings on the inflammatory markers were again conflicting. In one study, a predominantly anti-inflammatory profile was observed [8]. In the other two studies, a mixed profile (anti-inflammatory and pro-inflammatory) was observed [9, 10]. Hence, even after allowing for the effect of timing, we still could not find any discernible pattern in inflammation-related genes as described in the classic sepsis model.


Historically, cytokine studies suggested that there was a linear transition from pro-inflammatory cytokines to anti-inflammatory cytokines during the course of sepsis. However, these patterns are infrequently seen in clinical settings. In fact, only a few infections follow the classic two-phase model (for example, meningococcal sepsis or contaminated blood transfusions). Recently, studies have shown that inflammatory cytokines in sepsis follow a variable time course [2, 3]. Our systematic review extends this growing body of evidence by adding genome-wide data from a variety of clinical settings. In our review, we found that neither a distinctive pro/anti-inflammatory phase nor a clear transition from a pro-inflammatory to anti-inflammatory phase could be seen during sepsis. We also did not observe any discernible pattern in the behaviour of well-established inflammatory markers (for example, TNF-related genes) across the cohorts. Overall, we did not find strong genomic evidence that supports the classic two phase model of sepsis.

The negative finding of our review on the inflammation-related genes is unexpected, considering that the other two well-studied biological phenomena in sepsis, namely the activation of pathogen recognition (for example, toll-like receptors) and signal transduction pathways, are confirmed in most cohorts. The negative finding on inflammation related genes remained even after the cohorts were stratified by timing, patient groups or clinical settings.

The lack of clinical evidence to support the classical two-phase model has been known to many clinicians. The temporal relationship of an early pro-inflammatory phase followed by an anti-inflammatory phase, as depicted in the classical model, is rarely seen in clinical settings. However, this model remains the reigning paradigm under which many anti-sepsis drugs are being developed. The data outlined above therefore provide molecular evidence to validate the increasing concern among clinicians that the current inflammation-based definition of sepsis is too simplistic to describe a complex syndrome [1113].

While we did not find evidence to support the inflammation-based model of sepsis, we are not able to rule out the existence of other evidence that may support such a model. This is because of the limitations of our study. For example, our review has excluded other gene-expression studies that did not use microarray platform. As a result, our review is based on data from one particular methodology. Studies using other experimental approaches may repudiate/strengthen our findings. Furthermore, the observed gene-expression changes are restricted to circulating leukocytes. The changes in resident leukocytes in local tissue are likely to be very different from circulating leukocytes. Additional data from resident cells will provide a more complete understanding of the host response to sepsis. Another limitation is that our review does not provide information on changes occurring on a proteomic level, as they are not within the scope of this review. Lastly, most studies did not provide information on the leukocyte differential in the blood sample. The variability in leukocyte differentials could have confounded our findings. Given these several limitations, our findings need to be interpreted with caution. A more thorough evaluation of the sepsis model should involve integrating data from other experimental approaches, including in vitro studies, animal models and proteomic data.

Our review also revealed several significant methodological limitations of the current microarray studies in sepsis. First, many of the studies included in our review did not make their raw data publicly available. This makes it difficult for other researchers to verify their findings or to undertake meta-analysis. In addition, each study uses different statistical analysis approaches. In particular, different variance estimation methods were used by studies. However, most studies have adequate sample size; hence the impact of variance estimation on our findings is likely to be minimal. Another notable problem is that authors of each paper present their findings differently, making comparison or generalization of their data difficult. For example, some studies reported only a subset of the discovered genes, while others report functional analyses findings without actually listing the discovered genes. To better utilize the findings derived from gene-expression studies of sepsis, a uniform standard of reporting published microarray findings, such as those required for cancer studies [6], should be considered by all study authors in the future.


Our systematic review shows that sepsis-related inflammatory changes are highly variable on a transcriptional level. The arbitrary distinction of separating sepsis into pro-inflammatory and anti-inflammatory phases is not supported by gene-expression data.

Key messages

  • Sepsis-related inflammatory changes are highly variable on a transcriptional level.

  • These changes are not consistent with the established model of sepsis, where a biphasic pro-inflammatory and anti-inflammatory process is thought to underpin the host response.



cAMP responsive element binding protein


Janus kinase


mitogen activated protein kinase


nuclear factor kappa-B


transducer and activator of transcription protein


toll-like receptor.


  1. 1.

    Osuchowski MF, Welch K, Siddiqui J, Remick DG: Circulating cytokine/Inhibitor profiles reshape the understanding of the SIRS/CARS continuum in sepsis and predict mortality. J Immunol 2006, 177: 1967-1974.

    Article  CAS  PubMed  Google Scholar 

  2. 2.

    Osuchowski MF, Welch K, Yang H, Siddiqui J, Remick DG: Chronic sepsis mortality characterized by an individualized inflammatory response. J Immunol 2007, 179: 623-630.

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  3. 3.

    Gogos C, Drosou E, Bassaris H, Skoutelis A: Pro-versus anti-inflammatory cytokine profile in patients with severe sepsis: a marker for prognosis and future therapeutic options. J Infect Dis 2000, 181: 176-180. 10.1086/315214

    Article  CAS  PubMed  Google Scholar 

  4. 4.

    Christie J: Microarrays. Crit Care Med 2005, 33: S449-452. 10.1097/01.CCM.0000186078.26361.96

    Article  PubMed  Google Scholar 

  5. 5.

    Kulesh DA, Clive DR, Zarlenga DS, Greene JJ: Identification of interferon-modulated proliferation-related cDNA sequences. PNAS 1987, 84: 8453-8457. 10.1073/pnas.84.23.8453

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  6. 6.

    Dupuy A, Simon RM: Critical review of published microarray studies for cancer outcome and guidelines on statistical analysis and reporting. J Natl Cancer Inst 2007, 99: 147-157. 10.1093/jnci/djk018

    Article  PubMed  Google Scholar 

  7. 7.

    Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C, Aach J: Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet 2001, 29: 365-371. 10.1038/ng1201-365

    Article  CAS  PubMed  Google Scholar 

  8. 8.

    Talwar S, Munson PJ, Barb J, Fiuza C, Cintron AP, Logun C, Tropea M, Khan S, Reda D, Shelhamer JH, Danner RL, Suffredini AF: Gene expression profiles of peripheral blood leukocytes after endotoxin challenge in humans. Physiol Genomics 2006, 25: 203-215. 10.1152/physiolgenomics.00192.2005

    Article  PubMed  Google Scholar 

  9. 9.

    Prabhakar U, Conway T, Murdock P, Mooney J, Clark S, Hedge P, Williams W: Correlation of protein and gene expression profiles of inflammatory proteins after endotoxin challenge in human subjects. DNA Cell Biol 2005, 24: 410-431. 10.1089/dna.2005.24.410

    Article  CAS  PubMed  Google Scholar 

  10. 10.

    Calvano SE, Xiao W, Richards DR, Felciano R, Baker H, Cho R, Chen R, Brownstein BH, Cobb JP, Tschoeke S, Miller-Graziano C, Moldawer LL, Mindrinos MN, Davis RW, Tompkins RG, Lowry SF, Inflammatory and Host Response to Injury Large Scale Collaborative Research Program: A network-based analysis of systemic inflammation in humans. Nature 2005, 437: 1032-1037. 10.1038/nature03985

    Article  CAS  PubMed  Google Scholar 

  11. 11.

    Marshall J, Vincent JL, Fink MP, Cook D, Rubenfeld G, Foster D, Faist E, Reinhart K: Measures, markers, and mediators: towards a staging system for clinical sepsis. Crit Care Med 2003, 31: 1560-1567. 10.1097/01.CCM.0000065186.67848.3A

    Article  PubMed  Google Scholar 

  12. 12.

    Marshall J: Such stuff as dreams are made on: mediator-directed therapy in sepsis. Nat Rev Drug Discov 2003, 2: 391-405. 10.1038/nrd1084

    Article  CAS  PubMed  Google Scholar 

  13. 13.

    Carlet J, Cohen J, Calandra T, Opal S, Masur H: Sepsis: time to reconsider the concept. Crit Care Med 2008, 36: 1-3. 10.1097/CCM.0B013E318165B886

    Article  Google Scholar 

  14. 14.

    Prucha M, Ruryk A, Boriss H, Moller E, Zazula R, Russwurm S: Expression profiling: toward an application in sepsis diagnostics. Shock 2004, 22: 29-33. 10.1097/01.shk.0000129199.30965.02

    Article  CAS  PubMed  Google Scholar 

  15. 15.

    Tang B, McLean A, Dawes I, Huang S, Lin R: The use of gene-expression profiling to identify candidate genes in human sepsis. Am J Respir Crit Care Med 2007, 176: 676-684. 10.1164/rccm.200612-1819OC

    Article  CAS  PubMed  Google Scholar 

  16. 16.

    Tang B, McLean A, Dawes I, Huang S, Cowley M, Lin R: The gene-expression profiling of gram-positive and gram-negative sepsis in critically ill patients. Crit Care Med 2008, 36: 1125-1128. 10.1097/CCM.0b013e3181692c0b

    Article  CAS  PubMed  Google Scholar 

  17. 17.

    Ramilo O, Allman W, Chung W, Mejias A, Ardura M, Glaser C, Wittkowski KM, Piqueras B, Banchereau J, Palucka AK, Chaussabel D: Gene expression patterns in blood leukocytes discriminate patients with acute infections. Blood 2007, 109: 2066-2077. 10.1182/blood-2006-02-002477

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  18. 18.

    Tang B, McLean A, Dawes I, Huang S, Lin R: Gene-expression profiling of peripheral blood mononuclear cells in sepsis. Crit Care Med 2009, 37: 882-888. 10.1097/CCM.0b013e31819b52fd

    Article  CAS  PubMed  Google Scholar 

  19. 19.

    Payen D, Lukaszewicz A, Belikova I, Faivre V, Gelin C, Russwurm S, Launay J, Sevenet N: Gene profiling in human blood leucocytes during recovery from septic shock. Intensive Care Med 2008, 34: 1371-1376. 10.1007/s00134-008-1048-1

    Article  CAS  PubMed  Google Scholar 

  20. 20.

    Cobb J, Moore E, Hayden D, Minei J, Cuschieri J, Yang J, Li Q, Maier R: Validation of the riboleukogram to detect ventilator-associated pneumonia after severe injury. Ann Surg 2009, 250: 531-539. 10.1097/SLA.0b013e3181b8fbd5

    PubMed Central  PubMed  Google Scholar 

  21. 21.

    McDunn J, Husain K, Polpitiya A, Burykin A, Ruan J, Li Q, Schierding W, Lin N, Cobb JP: Plasticity of the systemic inflammatory response to actue infection during critical illness: development of the riboleukogram. PLoS One 2008, 3: e1564. 10.1371/journal.pone.0001564

    PubMed Central  Article  PubMed  Google Scholar 

  22. 22.

    Pachot A, Lepape A, Vey s, Bienvenu J, Mougin B, Monneret G: Systemic transcriptional analysis in survivor and non-survivor septic shock patients: A preliminary study. Immunol Lett 2006, 106: 63-71. 10.1016/j.imlet.2006.04.010

    Article  CAS  PubMed  Google Scholar 

  23. 23.

    Shanley TP, Cvijanovich N, Lin R, Allen GL, Thomas NJ, Doctor A, Kalyanaraman M, Tofil NM, Penfil S, Monaco M, Odoms K, Barnes M, Sakthivel B, Aronow BJ, Wong HR: Genome-level longitudinal expression of signaling pathways and gene networks in pediatric septic shock. Mol Med 2007, 13: 495-508. 10.2119/2007-00065.Shanley

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  24. 24.

    Wong HR, Shanley TP, Sakthivel B, Cvijanovich N, Lin R, Allen GL, Thomas NJ, Doctor A, Kalyanaraman M, Tofil NM, Tofil NM, Penfil S, Monaco M, Tagavilla MA, Odoms K, Dunsmore K, Barnes M, Aronow BJ, Genomics of Pediatric SIRS/Septic Shock Investigators: Genome-level expression profiles in pediatric septic shock indicate a role for altered zinc homeostasis in poor outcome. Physiol Genomics 2007, 30: 146-155. 10.1152/physiolgenomics.00024.2007

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  25. 25.

    Cvijanovich N, Shanley TP, Lin R, Allen GL, Thomas NJ, Checchia P, Anas N, Freishtat RJ, Monaco M, Odoms K, Sakthivel B, Wong HR, Genomics of Pediatric SIRS/Septic Shock Investigators: Validating the genomic signature of pediatric septic shock. Physiol Genomics 2008, 34: 127-134. 10.1152/physiolgenomics.00025.2008

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  26. 26.

    Wong H, Cvijanovich N, Allen GL, Lin R, Anas N, Meyer K, Freishtat RJ, Shanley TP: Genomic expression profiling across the pediatric systemic inflammatory response syndrome, sepsis and septic shock spectrum. Crit Care Med 2009, 37: 1558-1566. 10.1097/CCM.0b013e31819fcc08

    PubMed Central  Article  CAS  PubMed  Google Scholar 

  27. 27.

    Johnson S, Lissauer M, Bochicchio G, Moore R, Cross A, Scalea T: Gene expression profiles differentiate between sterile SIRS and early sepsis. Ann Surg 2007, 245: 611-621. 10.1097/01.sla.0000251619.10648.32

    PubMed Central  Article  PubMed  Google Scholar 

  28. 28.

    Lissauer M, Johnson S, Bochicchio G, Feild C, Cross A, Hasday J, Whiteford CC, Nussbaumer WA, Towns M, Scalea T: Differential expression of toll-like receptor genes: sepsis compared with sterile inflammation 1 day before sepsis diagnosis. Shock 2009, 31: 238-244. 10.1097/SHK.0b013e3181834991

    Article  CAS  PubMed  Google Scholar 

  29. 29.

    Tusher VG, Tibshirani R, Chu G: Significance analysis of microarrays applied to the ionizing radiation response. PNAS 2001, 98: 5116-5121. 10.1073/pnas.091062498

    PubMed Central  Article  CAS  PubMed  Google Scholar 

Download references


This research was supported by grants from the Nepean Critical Care Research Fund. The sponsor plays no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

Author information



Corresponding author

Correspondence to Benjamin M Tang.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

BT conceived of the study, collected data, performed analyses and drafted the manuscript. BT, SH and AM interpreted the data. All authors read and approved the final manuscript.

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Tang, B.M., Huang, S.J. & McLean, A.S. Genome-wide transcription profiling of human sepsis: a systematic review. Crit Care 14, R237 (2010).

Download citation


  • Microarray Platform
  • Pathogen Recognition
  • Clinical Sepsis
  • Endotoxin Challenge
  • Pathogen Recognition Receptor