Printer Friendly

Perioperative gene expression analysis for prediction of postoperative sepsis.

Severe sepsis is one of the leading causes of death after major surgery (1). Despite intensive therapeutic efforts, mortality remains high. Soft-tissue trauma leads to an immunologic reaction consisting of hyperinflammation and immunodepression at the same time. Although increased plasma concentrations of proinflammatory cytokines can be found, also present is an impaired monocyte function, which is characterized by a reduced ability to produce tumor necrosis factor (TNF) [4] after lipopolysaccharide stimulation and a decreased HLA-DR production that leads to a reduced antigen-presenting capacity (2, 3). Additionally, an impaired natural killer (NK) cell activity and a shift toward an immune reaction dominated by type 2 helper T cells is typical (4). Although antiinflammatory and immunosuppressive reactions are regarded as physiological and protective responses during stress, the same reactions may be harmful if aggravated (5). Major stress, such as that occurring after trauma, hemorrhage, burns, and cardiac surgery (6), is well recognized to substantially increase the susceptibility to infectious complications (7-9). Although these pathophysiological principles are known and acknowledged, it is still not possible to apply specific therapies. One major reason is the limited ability to adequately assess an individual's host response. Moreover, immunologic dysregulation does not occur in every patient, and even if such risk factors as comorbidities and age are known, it is not yet possible to predict an individual's susceptibility. Consequently, different attempts have been made to distinguish high- and low-risk patients by their individual immunologic reactions. Thus far, most studies have focused on clinical criteria (10) and measuring known circulating mediators or cell-bound receptors on peripheral blood cells (6, 11, 12); however, no reliable marker or method for risk prediction has been identified.

Real-time reverse-transcription PCR (RT-PCR) analysis allows measurement of mRNA production and thereby the functional status of immune cells. In recent years, RT-PCR analysis has proved its effectiveness in different diagnostic fields and has become readily available for clinical applications. In the present pilot case-control study, we examined the early perioperative expression of 23 selected inflammation-related genes in samples of whole blood from patients undergoing major surgery. We used the concept of postoperative immunodepression as a permissive condition and the results obtained from an analysis of gene expression to establish an assay to predict the risk of sepsis early in the postoperative period.

Materials and Methods

PATIENTS

The protocol was approved by the local ethics committee, and written informed consent was obtained from all patients before surgery. Adult patients with planned major abdominal or thoracic surgery were consecutively enrolled at 3 Berlin centers and monitored for the development of postoperative sepsis and severe sepsis until postoperative day 14. Sepsis and severe sepsis were defined according to the criteria of the American College of Chest Physicians and the Society of Critical Care Medicine (13). We planned to enroll at least 20 patients with postoperative sepsis or severe sepsis.

BLOOD SAMPLING, DATA COLLECTION, AND MATCHING PROCEDURE

Venous blood samples were collected at 3 times (before surgery and on the first and second postoperative days) with PAXgene[TM] Blood RNA Tubes [PreAnalytiX (BD/ Qiagen)]. PAXgene tubes were frozen immediately at -20 [degrees]C and kept at -80 [degrees]C until batch processing. Patients' demographic data, American Society of Anesthesiologists classification, main diagnosis, type of operation, concomitant diseases, laboratory values, microbiological data, vital signs (heart rate, temperature, urine output, blood pressure, and so forth), and clinical criteria of sepsis/severe sepsis were obtained from medical charts and collected into a database. At the end of the study, each patient with postoperative sepsis was matched with a patient without infection according to the following criteria: sex, age, main diagnosis, type of intervention, American Society of Anesthesiologists score, and concomitant diseases. The Sequential Organ Failure Assessment (SOFA) score and the Simplified Acute Physiology Score II (SAPS II) were calculated on the day of sepsis/severe sepsis diagnosis from routinely derived data.

RNA PREPARATION, cDNA SYNTHESIS, AND REAL-TIME RT-PCR

RNA was prepared with the PAXgene Blood RNA Kit (Qiagen) according to the manufacturer's instructions. RNA preparation included a 30-min incubation with a bovine deoxyribonuclease to eliminate genomic DNA. RNA concentration and integrity were evaluated with the 2100 Bioanalyzer (Agilent Technologies). cDNA was then synthesized with the QuantiTect Reverse Transcription Kit (Qiagen) according to the manufacturer's instructions; a maximum of 1 [micro]g RNA per sample was used in a total reaction volume of 20 [micro]L. The Quantiscript[R] reverse transcriptase included in the kit was used for cDNA synthesis. The reaction was performed for 28 min at 42 [degrees]C in a DNA thermal cycler (PerkinElmer); the reaction was then stopped with a 3-min enzyme-deactivation step at 95 [degrees]C. A control with no reverse transcriptase was performed to detect contamination with genomic DNA. The control was analyzed in the PCR step with an intron-specific primer combination. We discarded cDNA samples that had a positive signal in the control with no reverse transcriptase. For gene expression analysis, we selected 23 inflammation-related candidate genes. The first group of genes consisted of cytokine genes and genes related to cytokine signaling: TNF [5] [tumor necrosis factor (TNF superfamily, member 2)], IL1B (interleukin 1, beta), IL10 (interleukin 10), IL18 [interleukin 18 (interferon-gamma-inducing factor)], SOCS3 (suppressor of cytokine signaling 3), TGFB1 (transforming growth factor, beta 1), and IL6 [interleukin 6 (interferon, beta 2)]. We also evaluated T cell- and NK cell-related genes, including CD3D [CD3d molecule, delta (CD3-TCR complex)], CD69 (CD69 molecule), PRF1 [perforin 1 (pore forming protein)], GNLY (granulysin), CCR3 [chemokine (C-C motif) receptor 3], KLRK1 (killer cell lectin-like receptor subfamily K, member 1), IDO1 (indoleamine 2,3-dioxygenase 1), and KLRD1 (CD94) (killer cell lectin-like receptor subfamily D, member 1). For the assessment of monocyte antigen-presenting capacity, we analyzed 2 genes associated with major histocompatibility complex class II (MHCII), HLA-DRA (major histocompatibility complex, class II, DR alpha) and CD74 (CD74 molecule, major histocompatibility complex, class II invariant chain). We also evaluated chemokine genes that have shown altered expression in septic patients: IL8 (interleukin 8), CXCL10 [chemokine (C-X-C motif) ligand 10], PF4 (platelet factor 4), CCL3 [chemokine (C-C motif) ligand 3], and CXCL1 [chemokine (C-X-C motif) ligand 1 (melanoma growth stimulating activity, alpha)]. In addition, we analyzed HMOX1 [heme oxygenase (decycling) 1], which encodes a molecule with cytoprotective facilities, and S100A8 (S100 calcium binding protein A8), which encodes a granulocyte-derived molecule associated with chronic and acute inflammation. cDNA was stored at -20 [degrees]C for a maximum of 4 weeks before RT-PCR analysis. Real-time PCR was performed with the QuantiTect Probe PCR Master Mix[R] (Qiagen) on the GeneAmp 5700 Sequence Detection System[R] (Applied Biosystems) with the instrument's analysis software (GeneAmp 5700 SDS 1.3; Applied Biosystems). All primers and probes were designed with Primer Express software (Applied Biosystems) and validated by BLAST search, with the exception of the primers and probes for S100A8, CXCL1, IL18, and PF4, which were purchased commercially. The amplification primers or fluorogenic probes were designed to span exon borders to exclude cross-reactivity with genomic DNA. The PCR reaction was performed in a 13-[micro]L final reaction volume containing 1 [micro]L cDNA, 6.25 [micro]L QuantiTect Master Mix, 0.5 [micro]L fluorogenic hybridization probe, 3 [micro]L primer mix, and 2.25 [micro]L distilled water. After an initial step of 2 min at 50 [degrees]C to degrade any contaminating RNA sequences, we performed a denaturation and hot-start step with HotStarTaq [TM] DNA polymerase (Qiagen) at 95 [degrees]C for 10 min. A 2-step PCR thermal profile was used (40 cycles of 15 s of denaturation at 95 [degrees]C and 1 min of annealing/extension at 60 [degrees]C). In contrast to the 22 genes analyzed with fluorescently labeled probes, HLA-DRA mRNA production was measured with SYBR Green[R] dye. The production of HPRT1 (hypoxanthine phosphoribosyltransferase 1) mRNA was used for data normalization (14) according to the expression [2.sup.-[DELTA]Cq], where Cq is the quantification cycle. The [2.sup.-[DELTA]Cq] values of the septic patients were normalized to the arithmetic mean of the mRNA production values for the reference group ([2.sup.-[DELTA]Cq] method) (15). The mean Cq values for the genes of interest and HPRT1 were calculated from duplicate assays. Samples were considered negative for gene expression when Cq values were >40 cycles. Samples were tested for contamination with genomic DNA and were excluded from the study if they tested positive. Untreated controls (i.e., Master Mix without added cDNA) were included on each plate, untreated control samples with a positive signal (Cq < 40 cycles) were discarded. All tested samples were positive for the respective cDNA (Cq < 35 cycles). Additional information on the RT-PCR is available in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol56/issue4.

STATISTICAL ANALYSIS

We used SAS 9.0 (SAS Institute) and the SAS macro LD_F2 for nonparametric longitudinal statistical analyses (16, 17). This macro assesses intergroup differences and changes over time. For the genes with statistically significant differences, we used the Wilcoxon paired test in post hoc analyses for postoperative group comparisons and for comparison of preoperative and postoperative values. The method of Bonferoni and Holmes was used for correcting P values, with a starting P level of <0.05/23 for the LD_F2 macro (for 23 genes assessed), a P level of <0.05/10 for intergroup comparison (5 differentially expressed genes according to the LD_F2 macro), and a P level of <0.05/80 (20 genes showing significant differences, LD_F2) for comparison of preoperative and postoperative measurements.

An ROC curve analysis was done for all genes differentially expressed in the 2 study groups. Three genes (IL1B, TNF, and CD3D) that showed the largest median differences between the 2 groups were selected for a logistic regression analysis (SPSS 12.0; SPSS). To produce cutoff values for the combined gene expression assay, we then performed an ROC curve analysis based on the probability derived from the logistic regression. With the objective of optimizing clinical applicability, we included expression values of a single measurement (first day after surgery, because this was the first day with significant differences between the 2 groups) in the logistic regression.

The positive and negative predictive values were estimated from the sensitivity and specificity values for each gene, which were derived from the ROC curve analysis and the prevalence of sepsis in the study population (20 sepsis cases among a total of 220 study patients).

For correlation analysis, the Spearman Rho rank correlation test was used (SPSS 12.0; SPSS).

Results

PATIENTS

We enrolled 220 eligible patients between September 2003 and May 2005. Four of these patients developed sepsis, and 16 developed severe sepsis. The median time to the onset of sepsis/severe sepsis was 6 days (range, 1-14 days). In 14 patients, sepsis was diagnosed after the second postoperative day. In the case cohort, the mean SOFA score was 6.3 (95% CI, 4.4-8.2) and the mean SAPS II score was 36.9 (95% CI, 28.7-45.1) on the day of sepsis/severe sepsis diagnosis. To comply with recommendations of the peer reviewers, we retrospectively gathered preoperative SOFA scores and scores for the first and second postoperative days, when they were available. Mean SOFA scores for the sepsis and control groups were not significantly different preoperatively (n = 28) or on the first postoperative day (n = 21) (1 vs 0.86 and 2.8 vs 1.27, respectively). On the second day after surgery, the SOFA scores for the 2 groups were significantly different (4.1 vs 1.3, P < 0.05; n = 28). Ten patients had to be excluded from the control cohort before matching (n = 210) because of loss to follow-up, withdrawn consent, or an incomplete set of blood samples. Table 1 lists the characteristics of the matched pairs. The 2 groups were not significantly different with respect to either C-reactive protein concentration or leukocyte count at any measurement time. The site of infection was pulmonary in 11 cases, abdominal in 6 cases, and mediastinal, urogenital, and bloodstream in 1 case each. Bacterial sepsis was diagnosed in 10 cases, fungal in 4 cases, and both bacterial and fungal in 2 cases. In 2 cases, sepsis was diagnosed clinically without proof of microbes.

ALTERED POSTOPERATIVE INTERLEUKIN-1[beta] (IL-1[beta]) AND TNF mRNA PRODUCTION IN PATIENTS DEVELOPING SEPSIS

Peripheral blood monocytes from septic patients show a reduced release of proinflammatory cytokines, such as IL-1[beta], IL-6, and TNF (18). To ascertain whether decreased cytokine release occurs in sepsis patients in the early postoperative period before sepsis develops, we evaluated IL1B and TNF expression. Starting at comparable levels in the 2 groups, IL1B expression increased strongly in the control group, whereas expression in the septic group did not change significantly from preoperative values. Significant differences between the case and control groups were apparent on day 1 after surgery (Fig. 1A). TNF expression was reduced in the sepsis group, in contrast to the control group. This down-regulation was significant on the first postoperative day (Fig. 1B).

DECLINE IN EXPRESSION OF NK CELL- AND T CELL-DERIVED GENES

NK cells and T cells participate in the eradication of invading bacteria, not only by interacting with monocytes but also via direct mechanisms (19). Hence, a reduced activity or a reduction in the number of circulating NK cells and T cells could lead to an impaired defense against microorganisms. mRNA production by T cell-associated genes, including CD3D, CCR3, and CD69, and by NK cell-associated genes (KLRD1, KLRK1, and PRF1) shows similar time courses, with highly significant decreases in mRNA production apparent between the preoperative and postoperative time points in both investigated groups (Table 2). This observation suggests a general effect of surgical intervention on the 2 groups, a finding consistent with published studies (4, 20). We found decreased mRNA production in the sepsis group for CD3D, PRF1, KLRD1, and KLRK1 but noted no such differences for the remaining genes. CD3D and PRF1 expression showed the most pronounced differences (Fig. 2, A and B), which were statistically significant. Interestingly, we noted a strong positive correlation between the postoperative expression of CD3D, which encodes a T cell surface antigen, and the postoperative expression of other T cell- or NK cell-related genes (r = 0.471-0.884), which may reflect the influence of changes in lymphocyte cell counts on the quantity of mRNA in whole blood. The correlation between CD3D cDNA and leukocyte counts was not significant, however, whereas PRF1 expression was the only gene noted to exhibit a minor negative correlation with white blood cell counts preoperatively and on the first day after surgery [r = -0.36 (P = 0.021, preoperatively), and r = -0.34 (P = 0.031, on the first postoperative day)].

[FIGURE 1 OMITTED]

EXPRESSION MARKERS WITH NO SIGNIFICANT DIFFERENCES BETWEEN GROUPS

For most of the tested genes, we found no significant differences between the patients who developed sepsis and those who did not. These genes encoded pro- and antiinflammatory cytokines (IL10, IL18, TGFB1, IL6) and the gene encoding a suppressor of cytokine signaling (SOCS3), which is associated with intracellular TNF/IL-6 signaling. Interestingly, we also found no differences for HLA-DRA and CD74, which are associated with the MHCII. The genes encoding chemokines (IL8, CXCL10, PF4, CCL3, and CXCL1) were expressed similarly in the 2 groups.

PREDICTIVE PROPERTIES OF CANDIDATE MARKERS

Sepsis therapy should begin as early as possible (21). Discrimination of patients with and without risk of sepsis early in the postoperative course could help in initiating preemptive therapy, even before the occurrence of clinical sepsis symptoms. To evaluate the diagnostic capacity of the gene expression analysis, we performed an ROC curve analysis for the most promising genes in the study (Table 3). We included only the genes with statistically significant differences between the 2 groups in the analysis, as described above. This group consisted of the TNF, IL1B, CD3D, and PRF1 genes. To try to obtain a diagnosis as early as possible, we analyzed the relative expression of these genes on the first day after surgery. Cutoff values for the identification of case patients were derived from the ROC curve analysis. From the expression data for TNF, IL1B, and CD3D, we performed a logistic regression analysis and assigned a calculated probability value for each patient (Nagelkerke's [R.sup.2], 0.660; Hosmer-Lemeshow significance, 0.116). A subsequent ROC curve analysis based on these values (Fig. 3) defined a cutoff that identified patients with contingent sepsis with a sensitivity of 85% and a specificity of 90%.

[FIGURE 2 OMITTED]

Discussion

Real-time RT-PCR analysis is a relatively new technique in medical diagnostics. It allows highly sensitive quantification of mRNA in samples. In recent years, it has been applied to routine diagnostics in different ways, such as in quantifying viral load in HIV or Epstein-Barr virus infections (22). In the present study, we assessed RT-PCR analysis as a diagnostic tool for monitoring the immune system in perioperative patients by quantifying the expression of 23 genes related to inflammation. Cytokine-encoding genes TNF and IL1B and the T cell- and NK cell-related genes CD3D and PRF1 were differentially expressed in a group of patients who developed postoperative sepsis at a median of 5 days before the clinical diagnosis of sepsis. From a logistic regression analysis of TNF, CD3D, and IL1B expression, we were able to predict sepsis with a specificity of 90% and a sensitivity of 85%. Control patients matched for relevant baseline parameters exhibited similar SOFA scores preoperatively and on the first day after surgery. This matching minimizes the influence of disease severity as a confounding factor on gene expression.

TNF and IL1B play a pivotal role in the host's response to invading microorganisms. Although earlier concepts of the pathophysiology of sepsis regarded these cytokines mainly as mediators of shock and cell damage, newer findings have demonstrated that a certain level of production of these cytokines is necessary for a patient to survive sepsis. In this context, Riese et al. found an association between complications and a postoperatively reduced capacity of macrophages to secrete cytokines (23).

In a group of patients with an uncomplicated postoperative course, Hensler et al. found an increased monocytic release of IL-1[beta] postoperatively upon lipopolysaccharide stimulation, whereas the TNF production remained unchanged in these cells. Although secretion of the latter cytokine by type 1 helper T cells was down-regulated, these results are generally consistent with the present gene expression data. The similar results obtained for septic patients (18) underline the pathophysiological relationship between infectious and noninfectious inflammatory responses.

NK cells and cytotoxic T cells synergistically interact with macrophages in the clearance of bacterial infection. NK cell-produced interferon [gamma] potentiates the antimicrobial activity of macrophages (24). An increased number of NK cells is associated with a longer survival of patients with severe sepsis (25), and a reduced number of NK cells after severe injury promotes the development of subsequent sepsis (26). Granulysin has a direct antimicrobial effect in vitro against fungi and gram-positive and gram-negative bacteria, and the lysis of mycobacteria by granulysin is intensified in the presence of perforin 1 (27). In this context, the decreased expression of such genes as CD3D and PRF1 in T cells and NK cells in patients who developed sepsis in our study suggests a pivotal role of cell-mediated immunity in the risk for postoperative sepsis. A strong correlation between most of the expression data for NK cell/T cell-related genes evaluated in this study suggests an important influence of changes in leukocyte subpopulations on the measurements we have presented. Because the study design did not include data for differential blood cell counts or the measurement of leukocyte subpopulations by fluorescence-activated cell sorting or similar methods, the pathophysiological interpretation of the results is somewhat limited. Differences in the amounts of mRNA measured can be caused by up- or down-regulation of transcription in the leukocytes or by differences in the leukocyte subpopulations between the case and control groups. Interestingly, we found no significant differences in the expression of MHCII-related genes. Decreased expression on monocytes of genes encoding HLA-DR antigens is a common finding in septic patients and constitutes a reliable marker of immunodepression (28, 29); however, the surface expression of HLA-DR antigens is subject to several posttranscriptional mechanisms (30), which may explain differences in the expression of these genes and the quantities of these antigens on cell surfaces. Nevertheless, Pachot et al. found decreased production of mRNAs for MHCII-associated genes in whole blood samples from patients in a state of septic shock, compared with the production in samples from healthy volunteers (31). Similarly, we found the postoperative production of HLA-DRA mRNA in both groups to be reduced compared with preoperative values, a finding that reflects the situation before and after an inflammatory stimulus. Given that the HLA-DRA gene is expressed in different kinds of circulating blood cells, the down-regulation seen on monocytic cell surfaces in immunocompromised patients maybe blurred in samples of whole blood.

[FIGURE 3 OMITTED]

In the past, 2 different approaches have been taken for predicting the perioperative risk of postoperative sepsis. A number of investigators focused on the measurement of plasma cytokine concentrations, whereas others approached the problem by quantifying the expression of HLA-DR on monocytic surfaces via fluorescence-activated cell-sorting analysis. Whereas assays based on cytokine measurements do not allow identification of patients at risk with sufficient specificity and sensitivity, fluorescence-activated cell-sorting analysis of monocytic HLA-DR has proven effectiveness in different clinical settings, including in a multicenter approach (32-34). Nevertheless, this technique requires the preparation of blood samples directly after obtaining them, which limits its applicability in day-today clinical diagnostics. Blood withdrawal into PAXgene tubes requires rapid transport and cool storage (35), which are available in most hospitals in Western industrialized countries. This technique therefore provides a useful supplement to the established methods.

Despite the promising results, the data we have presented are subject to certain limitations. Because of the retrospective study design, all of the data and the derived cutoff values require a prospective validation. The study setting limits the application of the gene expression assay to perioperative patients. Furthermore, the number of patients is relatively small for logistic regression, indicating a risk of overfitting the data to the study population and possibility limiting the applicability of the assay to other patient populations. Therefore, a prospective multicenter approach with a larger study population would be preferable for validating the presented data.

Patients undergoing major surgery are at high risk for postoperative sepsis. Provided that the data presented here can be confirmed prospectively, the described RT-PCR assay offers the possibility of identifying patients at high risk for septic complications. It is based on a simple, easy-to-perform blood test, without having to consider such clinical symptoms as the criteria for systemic inflammatory response syndrome. This assay could open the way to the early application of preemptive therapies and thereby help in reducing postoperative mortality due to infectious complications.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures of Potential Conflicts of Interest: Upon manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest:

Employment or Leadership: None declared.

Consultant or Advisory Role: None declared.

Stock Ownership: None declared.

Honoraria: None declared.

Research Funding: We appreciate some financial support from PreAnalytiX, Hombrechtikon, Switzerland.

Expert Testimony: None declared.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.

References

(1.) Angus DC, Linde-Zwirble WT, LidickerJ, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001;29:1303-10.

(2.) Heagy W, Hansen C, Nieman K, Cohen M, Richardson C, Rodriguez JL, West MA. Impaired ex vivo lipopolysaccharide-stimulated whole blood tumor necrosis factor production may identify "septic" intensive care unit patients. Shock 2000; 14:271-6; discussion 276-7.

(3.) Monneret G, Lepape A, Voirin N, Bohe J, Venet F, Debard AL, et al. Persisting low monocyte human leukocyte antigen-DR expression predicts mortality in septic shock. Intensive Care Med 2006;32: 1175-83.

(4.) Toft P, Dagnaes-Hansen F, Tonnesen E, Petersen MS. Influence of surgery and endotoxin-induced sepsis combined on natural killer cell activity, oxidative burst of granulocytes and antigen presentation capability of monocytes. Acta Anaesthesiol Scand 2002;46:405-10.

(5.) Munford RS, Pugin J. Normal responses to injury prevent systemic inflammation and can be immunosuppressive. Am J Respir Crit Care Med 2001; 163:316-21.

(6.) Strohmeyer JC, Blume C, Meisel C, Doecke WD, Hummel M, Hoeflich C, et al. Standardized immune monitoring for the prediction of infections after cardiopulmonary bypass surgery in risk patients. Cytometry B Clin Cytom 2003;53:54-62.

(7.) Angele MK, Chaudry IH. Surgical trauma and immunosuppression: pathophysiology and potential immunomodulatory approaches. Langenbecks Arch Surg 2005;390:333-41.

(8.) Brune IB, WilkeW, HenslerT, Holzmann B, Siewert JR. Downregulation of T helper type 1 immune response and altered pro-inflammatory and antiinflammatory T cell cytokine balance following conventional but not laparoscopic surgery. Am J Surg 1999;177:55-60.

(9.) Koller M, Clasbrummel B, Kollig E, Hahn MP, Muhr G. Major injury induces increased production of interleukin-10 in human granulocyte fractions. Langenbecks Arch Surg 1998;383:460-5.

(10.) Sprung CL, Sakr Y, Vincent JL, Le Gall JR, Reinhart K, Ranieri VM, et al. An evaluation of systemic inflammatory response syndrome signs in the Sepsis Occurrence In Acutely Ill Patients (SOAP) study. Intensive Care Med 2006;32:421-7.

(11.) Chachkhiani I, Gurlich R, Maruna P, Frasko R, Lindner J. The postoperative stress response and its reflection in cytokine network and leptin plasma levels. Physiol Res 2005;54:279-85.

(12.) Roumen RM, Hendriks T, van der Ven-Jongekrijg J, Nieuwenhuijzen GA, Sauerwein RW, van der Meer JW, Goris RJ. Cytokine patterns in patients after major vascular surgery, hemorrhagic shock, and severe blunt trauma. Relation with subsequent adult respiratory distress syndrome and multiple organ failure. Ann Surg 1993;218:769- 76.

(13.) Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101:1644-55.

(14.) Carrol ED, Salway F, Pepper SD, Saunders E, Mankhambo LA, Ollier WE, et al. Successful downstream application of the PAXgene Blood RNA system from small blood samples in paediatric patients for quantitative PCR analysis. BMC Immunol 2007;8:20.

(15.) Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] method. Methods 2001;25: 402-8.

(16.) Brunner E, Langer F. Nichtparametrische Analyse longitudinaler Daten. Munich: R. Oldenburg Verlag; 1999.

(17.) Brunner E, Puri M. Nonparametric methods in factorial designs. Stat Papers 2001;42:1-52.

(18.) Kremer JP, Jarrar D, Steckholzer U, Ertel W. Interleukin-1, -6 and tumor necrosis factor-alpha release is down-regulated in whole blood from septic patients. Acta Haematol 1996;95: 268-73.

(19.) Ernst WA, Thoma-Uszynski S, Teitelbaum R, Ko C, Hanson DA, Clayberger C, et al. Granulysin, a T cell product, kills bacteria by altering membrane permeability. J Immunol 2000;165:7102-8.

(20.) Miller GC, Pritchard DJ, Ritts RE, Ivins JC, Pierre RV. Effect of surgery on the quantity of lymphocyte subpopulations. J Surg Res 1976;21:155-8.

(21.) Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368-77.

(22.) Muti G, Mancini V, Ravelli E, Morra E. Significance of Epstein-Barr virus (EBV) load and interleukin-10 in post-transplant lymphoproliferative disorders. Leuk Lymphoma 2005;46:1397-407.

(23.) Riese J, Woerner K, Zimmermann P, Denzel C, Hohenberger W, Haupt W. Association of a TNF-beta gene polymorphism with complications after major abdominal operations. Shock 2003;19:1-4.

(24.) Romani L, Puccetti P, Bistoni F. Interleukin-12 in infectious diseases. Clin Microbiol Rev 1997;10: 611-36.

(25.) Giamarellos-Bourboulis EJ, Tsaganos T, Spyridaki E, Mouktaroudi M, Plachouras D, Vaki I, et al. Early changes of CD4-positive lymphocytes and NK cells in patients with severe Gram-negative sepsis. Crit Care 2006;10:R166.

(26.) Ditschkowski M, Kreuzfelder E, Majetschak M, Obertacke U, Schade UF, Grosse-Wilde H. Reduced B cell HLA-DR expression and natural killer cell counts in patients prone to sepsis after injury. Eur J Surg 1999;165:1129-33.

(27.) Stenger S, Hanson DA, Teitelbaum R, Dewan P, Niazi KR, Froelich CJ, et al. An antimicrobial activity of cytolytic T cells mediated by granulysin. Science 1998;282:121-5.

(28.) Wolk K, Docke WD, von Baehr V, Volk HD, Sabat R. Impaired antigen presentation by human monocytes during endotoxin tolerance. Blood 2000;96:218-23.

(29.) Manjuck J, Saha DC, Astiz M, Eales LJ, Rackow EC. Decreased response to recall antigens is associated with depressed costimulatory receptor expression in septic critically ill patients. J Lab Clin Med 2000;135:153-60.

(30.) Fumeaux T, Pugin J. Role of interleukin-10 in the intracellular sequestration of human leukocyte antigen-DR in monocytes during septic shock. Am J Respir Crit Care Med 2002;166:1475-82.

(31.) Pachot A, Monneret G, Voirin N, Leissner P, Venet F, Bohe J, et al. Longitudinal study of cytokine and immune transcription factor mRNA expression in septic shock. Clin Immunol 2005;114: 61-9.

(32.) Haveman JW, van den Berg AP, van den Berk JM, Mesander G, Slooff MJ, de Leij LH, The TH. Low HLA-DR expression on peripheral blood monocytes predicts bacterial sepsis after liver transplantation: relation with prednisolone intake. Transpl Infect Dis 1999;1:146-52.

(33.) Docke WD, Hoflich C, Davis KA, Rottgers K, Meisel C, Kiefer P, et al. Monitoring temporary immunodepression by flow cytometric measurement of monocytic HLA-DR expression: a multicenter standardized study. Clin Chem 2005;51: 2341-7.

(34.) Oczenski W, Krenn H, Jilch R, Watzka H, Waldenberger F, Koller U, et al. HLA-DR as a marker for increased risk for systemic inflammation and septic complications after cardiac surgery. Intensive Care Med 2003;29:1253-7.

(35.) Chai V, Vassilakos A, Lee Y, Wright JA, Young AH. Optimization of the PAXgene blood RNA extraction system for gene expression analysis of clinical samples. J Clin Lab Anal 2005;19:182-8.

Carl Hinrichs, [1] [[[dagger]]] Katja Kotsch, [1] [[dagger]] Sandra Buchwald, [2] Marit Habicher, [2] Nicole Saak, [2] Herwig Gerlach, [3] Hans-Dieter Volk, [1] * [[dagger]] and Didier Keh [2] [[dagger]]

[1] Department of Medical Immunology, Charite Universitatsmedizin Berlin, Campus Mitte, Berlin, Germany; [2] Department of Anesthesiology and Critical Care Medicine, Charite Universitatsmedizin Berlin, Campus Virchow, Berlin, Germany; [3] Department of Anesthesiology and Critical Care Medicine, Vivantes Klinikum Neukolln, Berlin, Germany.

[4] Nonstandard abbreviations: TNF, tumor necrosis factor; NK, natural killer; RT-PCR, reverse-transcription PCR; SOFA, Sequential Organ Failure Assessment; SAPS II, Simplified Acute Physiology Score II; MHCII, major histocompatibility complex class II; Cq, quantification cycle; IL-1/[beta], interleukin-1/[beta].

[5] Human genes: TNF, tumor necrosis factor (TNF superfamily, member 2); IL1BB, interleukin 1, beta; IL10, interleukin 10; IL18, interleukin 18 (interferon-gamma-inducing factor); SOCS3, suppressor of cytokine signaling 3; TGFB1, transforming growth factor, beta 1; IL6, interleukin 6 (interferon, beta 2); CD3D, CD3d molecule, delta (CD3-TCR complex); CD69, CD69 molecule; PRF1, perforin 1 (pore forming protein); GNLY, granulysin; CCR3, chemokine (C-C motif) receptor 3; KLRK1, killer cell lectin-like receptor subfamily K, member 1; IDO1, indoleamine 2,3-dioxygenase 1; KLRD1(CD94), killer cell lectin-like receptor subfamily D, member 1; HLA-DRA, major histocompatibility complex, class II, DR alpha; CD74, CD74 molecule, major histocompatibility complex, class II invariant chain; IL8, interleukin 8; CXCL10, chemokine (C-X-C motif) ligand 10; PF4, platelet factor 4; CCL3, chemokine (C-C motif) ligand 3; CXCL1, chemokine (C-X-C motif) ligand 1 (melanoma growth stimulating activity, alpha); HMOX1, heme oxygenase (decycling) 1; S100A8, S100 calcium binding protein A8; HPRT1, hypoxanthine phosphoribosyltransferase 1.

* Address correspondence to this author at: Institut fur Medizinische Immunologie, Charite Universitatsmedizin, Campus Mitte, Chariteplatz 1, 10117 Berlin, Germany. Fax +49-30-450-524-932; e-mail hans-dieter.volk@charite.de.

[[dagger]] These authors contributed equally to the work.

Received August 17, 2009; accepted January 14, 2010.

Previously published online at DOI: 10.1373/clinchem.2009.133876
Table 1. Patients' demographics. (a)

Patient Sex Age, ASA Diagnosis
no. (b) years score

1a M 73 3 Pancreatic cancer
1b M 77 3 Pancreatic cancer
2a M 68 3 Esophageal cancer
2b M 68 3 Esophageal cancer
3a F 75 3 Pancreatic cancer
3b F 68 3 Pancreatic cancer
4a M 86 3 Esophageal cancer
4b M 75 2 Esophageal cancer
5a M 71 3 Pancreatic cancer
5b M 75 3 Pancreatic cancer
6a M 46 3 Pancreatitis
6b M 45 3 Pancreatitis
7a F 46 3 Biliodigestive
 insufficiency
7b F 59 3 Chronic cholangitis
8a M 67 3 Esophageal cancer
8b M 54 3 Esophageal cancer
9a F 55 3 Gastric cancer
9b F 58 3 Gastric cancer
10a M 71 3 Esophageal cancer
10b M 79 3 Cardia cancer
11a M 68 3 Rectal cancer
11b M 62 3 Rectal cancer
12a M 63 4 Esophageal cancer
12b M 57 3 Cardia cancer
13a M 66 3 Esophageal cancer
13b M 57 2 Cardia cancer
14a M 52 3 Bronchial cancer
14b M 69 3 Bronchial cancer
15a F 88 3 Cardia cancer
15b F 80 3 Gastric cancer
16a M 61 3 Bronchial cancer
16b M 60 2 Pancreatic cancer
17a M 58 2 Rectal cancer
17b M 57 2 Rectal cancer
18a M 68 2 Gastric cancer
18b M 76 3 Pancreatic cancer
19a M 49 3 Emphysema
19b M 30 2 Hemoptysis
20a M 64 3 Esophageal cancer
20b M 56 2 Esophageal cancer

Patient Operation Concomitant diseases
no. (b)

1a PPPD HP, COPD
1b PPPD HP, DM
2a EsR HP, COPD
2b EsR HP, COPD, DM
3a Pancreas resection HP, DM, CAD
3b PPPD HP, DM
4a EsR HP, emphysema
4b EsR HP, CAD
5a PPPD COPD, DM, apoplexy
5b PPPD HP, CAD
6a PPPD Alcohol abuse,
 cirrhosis
6b PPPD Alcohol abuse
7a Biliodigestive HP, morbid obesity
 anastomosis
7b Biliodigestive Rectal cancer
 anastomosis
8a EsR HP, COPD, cirrhosis
8b EsR HP, PAOD
9a Gastrectomy COPD
9b Gastrectomy HP, aortic aneurysm,
 CA
10a EsR COPD, HP, PAOD
10b Gastrectomy/distal COPD, hyperthyreosis
 EsR
11a Rectum resection HP, PAOD
11b Rectum resection COPD, CAD
12a EsR HP, CA, hypothyreosis
12b Gastrectomy/distal Pleural carcinomatosis
 EsR
13a EsR CAD, DM
13b Gastrectomy/distal Alcohol abuse
 EsR
14a Pneumonectomy COPD, alcohol abuse
14b Lung lobe resection CAD, HP
15a Gastrectomy Aortic valve
 insufficiency
15b Gastrectomy DM, CAD, HP
16a Thoracotomy COPD, HP, DM
16b Laparotomy
17a Rectum resection Diverticulitis
17b Rectum resection
18a PPPD CA
18b Biliodigestive HP, DM
 anastomosis
19a Lung lobe resection COPD, CA
19b Lung lobe resection
20a EsR COPD, HP, PAOD
20b EsR Alcohol abuse

(a) Matching criteria were age, sex, American Society of
Anesthesiologists (ASA) score, diagnosis, and type of
intervention. The matched individuals in the 2 groups
were highly similar for the characteristics shown.

(b) a, case group; b, control group.

(c) PPPD, pylorus-preserving pancreatoduodenectomy; HP,
hypertension; COPD, chronic obstructive pulmonary disease;
DM, diabetes mellitus; EsR, esophagus resection; CAD,
coronary artery disease; PAOD, peripheral arterial
obstructive disease; CA, cardiac arrhythmia.

Table 2. Overview of the results for 4 genes. (a)

 Preoperative First postoperative
 day
IL1B

 Case 0.57 (0.11-2.57) 0.82 (0.15-1.52) (b)
 Control 0.64 (0.27-2.66) 1.44 (0.52-4.14) (b)

TNF

 Case 0.33 (0.16-0.85) 0.14 (0.03-0.52) (b)
 Control 0.53 (0.03-1.62) 0.39 (0.16-0.8) (b)

CD3D

 Case 4.52 (1.68-10.89) 0.94 (0.18-2.44) (b,c)
 Control 5.01 (1.17-7.89) 2.22 (1.05-4.42) (b,c)

PRF1

 Case 5.69 (1.34-11.04) 1.06 (0.26-5.01) (b,c)
 Control 5.89 (1.12-13.5) 2.12 (0.55-9.48) (b,c)

 Second postoperative
 day
IL1B

 Case 0.83 (0.11-4.44)
 Control 1.93 (1.11-4.5) (c)

TNF

 Case 0.21 (0.04-0.46)
 Control 0.4(0.18-0.92)

CD3D

 Case 0.84 (0.14-12.64) (c)
 Control 1.87 (0.59-4.5) (c)

PRF1

 Case 1.23 (0.33-11.47) (c)
 Control 2.36 (0.61-9.85) (c)

(a) Data are presented as median [2.sup.-[DELTA]Cq]
values, as normalized to HPRT1 gene expression.
Ranges are in parentheses.

b Significant differences between the 2 groups
(P < 0.005).

c Significant differences between the pre- and
postoperative time points (P< 0.000 625).

Table 3. High sensitivity and specificity of the combined
gene expression assay. (a)

Gene (cutoff) Specificity Sensitivity PPV

CD3D (1.04) 100% 60% 100%
PRF1 (1.26) 85% 70% 35.4%
IL1B (1.03) 80% 70% 29.2%
TNF (0.15) 100% 60% 100%
Combination 90% 85% 47.2%
 assay (0.60)

Gene (cutoff) NPV AUC (95% CI)

CD3D (1.04) 95.5% 0.844 (0.723-0.965)
PRF1 (1.26) 96.0% 0.814(0.680-0.948)
IL1B (1.03) 95.8% 0.805 (0.671-0.939)
TNF (0.15) 95.5% 0.869 (0.758-0.979)
Combination 98.3% 0.918(0.826-1.009)
 assay (0.60)

(a) The gene expression assay for each gene was able to
predict sepsis with an average sensitivity and specificity.
Results for the combination assay (for 3 genes: CD3D, IL1B,
and TNF were defined as positive if the mRNA quantity fell
below the cutoff for [greater than or equal to] 2 genes.
The combination assay was able to predict sepsis with high
specificity and sensitivity. Sensitivity and specificity
represent maximal accuracy according to ROC curve analysis.
The negative predictive value (NPV) and the positive
predictive value (PPV) are estimated from the sensitivity
and specificity, as indicated in the table and a prevalence
of 20 cases of sepsis in a study population of 220 patients.
AUC, area under the ROC curve.
COPYRIGHT 2010 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Molecular Diagnostics and Genetics
Author:Hinrichs, Carl; Kotsch, Katja; Buchwald, Sandra; Habicher, Marit; Saak, Nicole; Gerlach, Herwig; Vol
Publication:Clinical Chemistry
Date:Apr 1, 2010
Words:6339
Previous Article:MicroRNA-21 is induced early in pancreatic ductal adenocarcinoma precursor lesions.
Next Article:Quantifying EGFR alterations in the lung cancer genome with nanofluidic digital PCR arrays.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters