Journal of Critical Care (2010) 25, 542.e1–542.e8

Polymorphisms of genes encoding tumor necrosisfactor-alpha, interleukin-10, cluster of differentiation-14and interleukin-1ra in critically ill patientsMaja Surbatovic a,⁎, Krasimirka Grujic b, Bojana Cikota c, Miodrag Jevtic d,Nikola Filipovic a, Predrag Romic a, Natasa Strelic c, Zvonko Magic c

aClinic of Anesthesiology and Intensive Therapy, Military Medical Academy, 11000 Belgrade, SerbiabAnesthesia and Critical Care Department, Clinical Center, 38220 Kosovska Mitrovica, SerbiacInstitute for Medical Research, Military Medical Academy, 11000 Belgrade, SerbiadClinic of General and Vascular Surgery, Military Medical Academy, 11000 Belgrade, Serbia

0d

Keywords:Sepsis;Trauma;Gene polymorphisms

AbstractPurpose: The aim of the study was to determine whether distributions of tumor necrosis factor (TNF)-α308, interleukin (IL)-101082, CD14159, and IL-1ra gene intron 2 genotypes in critically ill patients areassociated with outcome, underlying cause of sepsis, and type of microorganism.Materials and Methods: Blood samples from 106 critically ill white patients were genotyped by methodbased on polymerase chain reaction for TNF-α308, IL-101082, CD14159, and IL-1ra gene intron 2.Results: All patients with TNF-α308AA genotype survived; relative risk (RR) of death in patients with AGwas 3.250 and with GG, 1.923 (P b .01). In patients with Gram-positive sepsis, IL-101082AA and then AGgenotypes were the most frequent ones (odds ratio [OR], 18.67 and 7.20, respectively; P b .01). Whencomparing IL-101082AA with AG, RR of pancreatitis was 1.80 and OR was 3.40. When AA and GG werecompared, RR was 7.33 and OR was 20.00. In patients with GG, RR of peritonitis was 4.07 and OR was5.88 (P b .01). In patients with Gram-positive sepsis, CD14159CT was the most frequent one with OR of5.25. Distribution of 6 IL-1ra gene intron 2 genotypes showed no significant association.Conclusions: Distribution of TNF-α308 genotypes is associated with outcome, IL-101082 with type ofmicroorganism and underlying cause of sepsis, and CD14159 with type of microorganism.© 2010 Published by Elsevier Inc.

1. Introduction tremendous variability seen in the clinical profile and

The inflammatory response contributes significantly tothe morbidity and mortality of critically ill patients anddisplays high level of interindividual variation. There is

⁎ Corresponding author. Tel.: +381 11 2665 125; fax: +381 11 2665 125.E-mail address: maya@eunet.rs (M. Surbatovic).

883-9441/$ – see front matter © 2010 Published by Elsevier Inc.oi:10.1016/j.jcrc.2009.12.003

outcome in patients who encounter similar insults (traumaand/or infection). Despite significant advances in under-standing of the biology of inflammation, improvements inclinical outcomes have been more sporadic and, with fewnotable exceptions, are related to improvements in support-ive care rather than to specific therapies. As a result,morbidity, mortality, and cost remain high [1]. Therefore,focus of research has more recently shifted to identification

Table 1 Demographic data

Characteristic Finding

Total no. of patients 106Age (median, range) 50 (from 19 to 83 y)Sex, n (%)Male 75 (70.7%)Female 31 (29.3%)Simplified Acute PhysiologyScore II score, mean ± SD

58.38 ± 10.15

Acute Physiology and Chronic HealthEvaluation II score, mean ± SD

22.09 ± 4.33

Sequential Organ FailureAssessment score, mean ± SD

7.86 ± 2.47

Simplified Acute PhysiologyScore III score, mean ± SD

69.82 ± 11.84

Mechanical ventilation 57 patients (54%)Reason for ICU admission, nSevere trauma (ISS, 29.7 ± 10.2) 50 (47.16%)Secondary sepsis after trauma 20 (18.86%)Severe sepsis due to 56 (52.83%)Peritonitis 32 (30.18%)Pancreatitis 15 (14.15%)Other causes 9 (8.49%)

Mortality, n 53 (50.00%)

542.e2 M. Surbatovic et al.

of genetic determinants of the inflammatory response,especially to the group of genes that regulates immuneresponse to trauma and/or infection.

More than 20 years ago, Sorensen et al [2] conductedcrucial study of almost 1000 adoptees that indicated asubstantial inherited increase in the risk of premature deathfrom severe infection. The death of biologic parent beforethe age of 50 resulted in relative risk of death in the adopteesof 5.81 for infections, which is the highest compared to allother causes. A genetic polymorphism is an allelic variantthat exists stably in a population in a frequency that cannotbe accounted for by new mutations (generally N1%). Themost frequent type of polymorphism is the single-nucleotidepolymorphism (SNP), which can be a substitution, adeletion or an insertion of a single nucleotide. Approxi-mately 1 in every 300 to 500 bases of human DNA may bean SNP.

The proinflammatory cytokine tumor necrosis factor(TNF)-α is an essential component of the host immuneresponse to insult. Guanine to adenine transition at the 308site has been associated with variability in TNF-α secretionafter various stimuli. This polymorphism results in 2 alleleforms, one by which a guanine defines the common alleleTNF1 and one by which an adenine defines the uncommonallele TNF2 [3].

Interleukin-10 is potent antiinflammatory cytokine.Excess of IL-10 can induce immunosuppression and increasemortality. One of biallelic polymorphisms in the promoterregion of the IL-10 gene is at position 1082. Interleukin-101082 polymorphism results from the substitution of guaninewith adenine. Investigators focused on its involvement in thesusceptibility to and outcome of severe sepsis.

One of the most important link in understanding themolecular mechanisms of severe sepsis and septic shock hasbeen the discovery that cluster of differentiation (CD)-14 is areceptor for a very wide range of microbial productsincluding lipopolysaccharide (released from Gram-negativebacteria), peptidoglycans, and lipoteichoic acid (constituentsof Gram-positive bacteria). A genetic polymorphism hasbeen identified inside the CD14 promoter sequence. Itconsists of a cytosine to thymine transition at base pair 159.

Interleukin-1 receptor antagonist (IL-1ra) is a member ofthe IL-1 family that binds to the IL-1 receptor, thereby,inhibiting the proinflammatory actions of IL-1α and IL-1β.Polymorphisms within the IL-1ra gene have been reported inintron 2 and contains a variable numbers of tandem repeats of86 base pairs. Five alleles have been identified and includeA1 (4 repeats), A2 (2 repeats), A3 (5 repeats), A4 (3 repeats),and A5 (6 repeats) [4].

The aim of this study was to determine whetherdistributions of TNF-α308, IL-101082, CD14159, and IL-1ragene intron 2 genotypes in critically ill patients with traumaand/or sepsis are associated with outcome, underlying causeof sepsis (peritonitis, pancreatitis), or the type of infectingmicroorganism (Gram positive, Gram negative, and mixedGram positive and Gram negative).

2. Methods

2.1. Patients

The study was approved by local ethics committee, andinformed consent was obtained from a patient or a first-degree relative. Blood samples were obtained from 106critically ill white patients from Serbia and Montenegro withsevere trauma and/or sepsis on admission to surgicalintensive care unit (ICU) of Military Medical Academy(Belgrade, Serbia) from October 2006 until November 2008.The Simplified Acute Physiology Score III score [5] on ICUadmission, Simplified Acute Physiology Score II [6], AcutePhysiology and Chronic Health Evaluation II score [7], andSequential Organ Failure Assessment score [8] at 24 hoursafter ICU admission were calculated and recorded. Thepatients' demographics are shown in Table 1.

Determination of trauma severity was performed using theInjury Severity Score (ISS). Injury Severity Score wasdetermined using Abbreviated Injury Scale. Great majority oftrauma patients were casualties from motor vehicle accidentswith blunt and/or penetrating trauma. Sepsis patients enteredthe study if they had met the following criteria (according to2001 SCCM/ESICM/ACCP/ATS/SIS International SepsisDefinitions Conference): documented or suspected infectionplus presence of systemic inflammatory response syndromeand sepsis-associated organ dysfunction, hypoperfusion, orhypotension. In the absence of documented infection (sepsiscriteria do not require positive blood culture), the presence ofone or more of the following was used as entry criteria: (1)

Fig. 1 (A) TNF-α308 genotyping after 10% PAG electrophoresisand silver nitrate staining. TNF-α308 GG genotype was found inlanes 2, 3, 4, 5, 6, 8, and 9, whereas TNF-α308 AG genotype wasdetected in lanes 1, 7, and 10. (B) IL-101082 genotyping after 2%agarose gel electrophoresis and ethidium bromide staining. IL-101082 AG genotype was found in patient 1, AA genotype was foundin patient 2, and GG genotype was found in patients 3 and 4. In everylane, the band of 429 base pairs was present (internal control [IC]).(C) CD14159 genotyping after 2% agarose gel electrophoresis andethidium bromide staining. CD14159 genotype CCwas found in lane1, and genotype TT was found in lanes 2, 5, 6, and 7. Genotype CTwas found in lanes 3, 4, 8, 9, and 10. (D) IL-1ra genotyping after10% PAG electrophoresis and silver nitrate staining. IL-1ragenotype A3/A3 was found in lanes 1 and 4. A1/A1 genotype wasfound in lane 2 and A1/A3 genotype was found in lane 3.

542.e3Polymorphisms of genes

significant edema or positive fluid balance (20 mL/kg for 20hours), (2) hyperglycemia (plasma glucose, N120 mg/dL) inthe absence of diabetes, (3) inflammatory variables: plasmaC-reactive protein higher than 50 mg/L or plasma procalci-tonin higher than 1.5 ng/mL), (4) mixed venous oxygensaturation more than 70%, and (5) cardiac index more than3.5 L/min per square meter. Organ dysfunction variablesinclude arterial hypoxemia (PAO2/fraction of inspired oxygenratio of b300), acute oliguria (urine output b 0.5 mL/kg perhour), creatinine level more than 2.0 mg/dL, coagulationabnormalities (international normalized ratio N 1.5 oractivated partial thromboplastin time N 60 seconds), throm-bocytopenia (platelet count b 100 000), hyperbilirubinemia(plasma total bilirubin N 2 mg/dL), tissue perfusion variable:hyperlactatemia (N2 mmol/L), and hemodynamic variables:arterial hypotension (systolic blood pressure b 90 mm Hg,mean arterial pressure b 70 mm Hg, or a systolic bloodpressure decrease N 40 mm Hg).

Blood was collected in EDTA tubes and stored at −40°C.Each patient's DNA was extracted from whole blood with aDr GenTLE kit (TaKaRa, Otsu, Japan) according to themanufacturer's instructions.

2.2. TNF-308 genotyping

The oligonucleotide polymerase chain reaction (PCR)primers 5′-GGCAATAGGTTTTGAGGGCCAT-3′ and 5′-GAGCGTCTGCTGGCTGGGTG-3′ were used to amplify345 base pair fragment (94°C for 1 minute, 60°C for 1minute, 72°C for 1 minute; 30 repeats) [9]. The PCRproducts were digested with NcoI according to manufac-turer's recommendations (Fermentas, Vilnius, Lithuania).Digested DNA was analyzed on 10% polyacrylamide gels(PAG) after electrophoresis and silver nitrate staining. G atposition 308 was represented with the presence of 2 bands(325 and 20 base pairs), whereas A at 308 was visualized as345-base pair band on the gel (Fig. 1A).

2.3. IL-101082 genotyping

For each sample, 2 parallel PCR reactions were performed.The primer pair generic (5′-CAGTGCCAACTGAGAATTTGG-3′)/IL-10 A (5′-ACTACTAAGGCTTCTTTGGGAA-3′) was used to amplify IL-101082 A allele, whereas theprimer pair generic/IL-10 G (5′CTACTAAGGCTTCTTTGGGAG-3′) was used to amplify IL-101082 G allele. As aninternal control, 2 primers amplifying a human growthhormone sequences were added to each reaction (5′-GCCTTCCCAACCATTCCCTTA-3′ and 5′-TCACG-GATTTCTGTTGTGTTTC-3′). The PCR amplification (10cycles with 95°C for 1 minute, 65°C for 50 seconds, 72°C for50 seconds, followed by 20 cycles with 95°C for 1 minute,59°C for 50 seconds, and 72°C for 50 seconds) generatedproducts of 258 base pairs for IL-10 and 429 base pairs forinternal control. The presence/absence of PCR products were

visualized on 2% agarose gels after electrophoresis andethidium bromide staining (Fig. 1B).

2.4. CD14159 genotyping

CD14 gene promoter sequence was amplified with thefollowing PCR primers: 5′-TGCCAGGAGACACA-GAACCC-3′ and 5′-TGTCATTCAGTTCCCTCCTC-3′(96°C for 40 seconds, 56°C for 40 seconds, 72°C for50 seconds; 38 repeats) [10]. The generated product of 166base pairs was digested with Hae III according to manufac-turer's instructions (Fermentas, Vilnius, Lithuania). Restric-tion endonuclease cut PCR products with C at position 159,whereas products with the T polymorphic base at 159 werenot cut. Resulting fragments were electrophoresed on2% agarose gels and visualized with ethidium bromidestaining (Fig. 1C).

2.5. IL-1ra genotyping

The polymorphic region within IL-1ra gene intron 2contains a variable numbers of tandem repeats of 86 basepairs. It was amplified using PCR (95°C for 30 seconds,62°C for 30 seconds, 72°C for 50 seconds; 35 repeats) withthe following primers: 5′-CTCAGCAACACTCCTAT-3′

542.e4 M. Surbatovic et al.

and 5′-TCCTGGTCTGCAGGTAA-3′ [11]. The size ofamplified fragments was analyzed on a 10% PAG afterelectrophoresis and staining with silver nitrate. The allele A1,A2, A3, A4, A5, respectively, represent 4 repeats, 2 repeats,5 repeats, 3 repeats, and 6 repeats of 86-base pair unit, andtheir sizes are 410 base pairs, 240 base pairs, 500 base pairs,325 base pairs, and 595 base pairs, respectively (Fig. 1D).

2.6. Statistical analysis

Statistical analysis of the genotype distribution of theTNF-α, IL-10, CD14, and IL-1ra genes was performed byPearson χ2 test and Fisher exact test. The strength ofassociation between genotypes and outcome, underlyingcause of sepsis and infecting microorganism was expressedby the odds ratio (OR) an the relative risk. All statisticalanalysis was performed with SPSS 12.0 for Windows (SPSSInc, Chicago, Ill). A P b .05 was regarded as statisticallysignificant, and P b .01 as statistically highly significant.

3. Results

In this study, total of 106 critically ill surgical patientswere included consisting of 75 (70.7%) men and 31 (29.3%)women; age ranged from 19 to 83 with mean value of 50.20years. Fifty of them sustained severe trauma. Determinationof trauma severity was performed using the ISS. InjurySeverity Score (mean value, 29.7 ± 10.2) was determinedusing Abbreviated Injury Scale. Of 50 trauma patients, 20developed secondary sepsis. The other 56 patients hadsevere sepsis. Most patients with sepsis had peritonitis (n =32). The other major disease associated with sepsis wasnecrotizing pancreatitis (n = 15), and other causes includedabscesses at other locations, pneumonia, and mediastinitis(n = 9). The 6-month mortality rate in our study populationwas 50% (53/106 patients did not survive) (Table 1). OnTable 2, distribution of TNF-α308, IL-101082, CD14159, andIL-1ra genotypes in the group of Serbian critically illpatients is shown.

Table 2 Distribution of TNF-α308, IL-101082, CD14159, andIL-1ra genotypes in the group of Serbian critically ill patients

Gene Genotype

TNF-α308 AA AG GG7 (7.4%) 13 (13.7%) 75 (78.9%)

IL-101082 AA AG GG22 (21.8%) 55 (54.4%) 24 (23.8%)

CD14159 CC CT TT33 (34.7%) 41 (43.2%) 21 (22.1%)

IL-1ra A1/A1 A1/A2 A1/A355 (59.8%) 25 (27.2%) 3 (3.3%)A1/A4 A2/A2 A3/A33 (3.3%) 4 (4.3%) 2 (2.1%)

3.1. Polymorphism of TNF-308 gene

Distribution of 3 TNF-α308 genotypes (AA, AG, GG) isassociated with outcome with statistically high significance(Pearson χ2, 8.804; P b .01). All patients with AAgenotype survived (7/7). Relative risk of death in patientswith AG genotype was 3.250 and in those with GGgenotype was 1.923.

No significant association of any genotype with underly-ing cause of sepsis (peritonitis, pancreatitis) or with the typeof infecting microorganism (Gram positive, Gram negative,and mixed Gram positive and Gram negative) was found.

3.2. Polymorphism of IL-101082 gene

Distribution of 3 IL-101082 genotypes (AA, AG, GG) isassociated with the type of infecting microorganism withstatistically high significance (Pearson χ2, 15.863; P b .01).In patients with Gram-positive sepsis, AA and then AGgenotypes were the most frequent ones. In patients withGram-negative sepsis, GG genotype was the most frequentone. Relative risks of Gram-positive or Gram-negative sepsisand OR with 95% confidence interval (CI) for differentpolymorphisms are shown on Table 3.

Distribution of 3 IL-101082 genotypes is also associatedwith underlying cause of sepsis (peritonitis, pancreatitis)with statistically high significance (Pearson χ2, 15.862; P b.01). In patients with sepsis associated with pancreatitis, AAgenotype was the most frequent one. In patients with sepsisassociated with peritonitis, GG genotype was the mostfrequent one. When comparing AA with AG genotype, wefound that relative risk of pancreatitis as underlying cause ofsepsis was 1.80; OR, 3.40; and 95% CI, 0.524 to 22.027.When AA and GG were compared relative risk ofpancreatitis as well as OR have risen sharply: relative riskwas 7.33; OR, 20.00; and 95% CI, 1.390 to 287.614.Comparison of GG and AG genotypes showed that inpatients with GG genotype relative risk of peritonitis beingunderlying cause of sepsis was 4.07; OR, 5.88; and 95% CI,0.652 to 53.039.

When comparing AA genotype with combination of AGand GG, we found that there is statistically significantassociation with outcome (Fisher exact test, P b .05).Critically ill patients with G allele in either of genotypeshave relative risk of death, 1.46; OR, 2.43; and 95%CI, 0.894to 6.612.

3.3. Polymorphism of CD14159 gene

Distribution of 3 CD14159 genotypes (CC, CT, TT)showed that there is statistically significant association withthe type of infecting microorganism (Pearson χ2, 13.611; Pb .05). In patients with Gram-positive sepsis, CT genotypewas the most frequent one. In sepsis patients with CC and TTgenotypes, frequency of Gram-positive and Gram-negative

Table 3 Association of IL-10 gene genotypes with infecting microorganisms: relative risk and OR for causative bacteria

IL-10 polymorphisms

AA/AG AA/GG AG/GGGram positive/mixed Gram positive/mixed Gram positive/mixedRelative risk, 2.221 Relative risk, 2.96 Relative risk, 1.33OR, 12.00; 1.33-108.68 OR, 18.67; 1.56-222.94 OR, 1.56; 0.334-7.235Gram positive/Gram negative Gram positive/Gram negative Gram positive/Gram negativeRelative risk, 0.940 Relative risk, 2.67 Relative risk, 2.82OR, 0.830 OR, 6.00; 1.017-35.375 OR, 7.20; 1.352-38.327Gram negative/mixed Gram negative/mixed Gram negative/mixedRelative risk, 3.68 Relative risk, 1.42 Relative risk, 0.38OR, 14.40; 1.299-159.52 OR, 3.11; 0.281-34.420 OR, 0.22

542.e5Polymorphisms of genes

infecting microorganisms was similar. Relative risks ofGram-positive or Gram-negative sepsis and ORs with 95%CI for different polymorphisms are shown on Table 4.

When comparing CC genotype with combination of CTand TT, we found that there is statistically significantassociation with the type of infecting microorganism.Critically ill patients without T allele in either of genotypes,that is, with CC genotype have relative risk of Gram-negativesepsis, 1.25; OR, 1.45; and 95% CI, 0.381 to 5.572. In thiscomparison, in sepsis patients with CC genotype, Gram-negative bacteria were most frequent.

3.4. Polymorphism of IL-1ra gene

Distribution of 6 IL-1ra gene intron 2 genotypes (A1/A1,A1/A2, A1/A3, A1/A4, A2/A2, A3/A3) were analyzed. Nosignificant association of any genotype with underlying causeof sepsis (peritonitis, pancreatitis), with the type of infectingmicroorganism (Gram positive, Gram negative, and mixedGrampositive andGramnegative) orwith outcomewas found.

4. Discussion

The goal of our study was to determine whether genotypedistributions of TNF-α308, IL-101082, CD14159, and IL-1ragenes in critically ill patients with trauma and/or sepsis are

Table 4 Association of CD14 genotypes with infecting microorganis

CD14 polymorphisms

CC/CT CC/TTGram positive/mixed Gram positive/miRelative risk, 0.73 Relative risk, 1.85OR, 0.49 OR, 2.57; 0.468-1Gram positive/Gram negative Gram positive/GrRelative risk, 0.71 Relative risk, 1.17OR, 0.43 OR, 1.33; 0.204-8Gram negative/mixed Gram negative/miRelative risk, 1.08 Relative risk, 1.50OR, 1.14; 0.250-5.220 OR, 1.93; 0.387-9

associated with outcome, underlying cause of sepsis, or thetype of infecting microorganism.

We have found that TNF-α308 genotype is associated withoutcome with statistically high significance (P = .008). Allpatients with AA genotype survived. Relative risk of death inpatients with AG genotype was 3.250 and in those with GGgenotype was 1.923. Gordon with coworkers [12] performedprospective multicenter study on 213 patients recruited from8 ICUs. When TNF-α308 genotype and allele frequencies inICU survivors and nonsurvivors were analyzed, it turned outthat GG genotype was more frequent in nonsurvivors (75%vs 60%), AG genotype was more frequent in survivors(35.6% vs 23.1%) as well as AA genotype (4.4% vs 1.9%).So, in their study, G allele was more frequent innonsurvivors (86.5% vs 77.8%) and A allele was morefrequent in survivors (22.2% vs 13.5%). Although thoseresults did not reach statistical significance, trend was thesame as in our study.

Tumor necrosis factor-α is considered to be proximal andcentral cytokine in host immunoinflammatory cascaderegardless of initiating insult. One of the first studies todemonstrate association of the rare TNF2 (adenine) allele ofthe TNF-α promoter with outcome of severe sepsis andseptic shock was performed by Mira et al [13] on 89 patients.They concluded that TNF2 allele is strongly associated withsusceptibility to and death due to septic shock; patients withthe TNF2 allele had a 3.7-fold risk of death. Similarly, a USstudy [14] that looked at the occurrence of severe sepsis after

ms: relative risk and OR for causative bacteria

CT/TTxed Gram positive/mixed

Relative risk, 2.544.102 OR, 5.25; 1.093-25.211am negative Gram positive/Gram negative

Relative risk, 1.63.707 OR, 3.11; 0.526-18.382xed Gram negative/mixed

Relative risk, 1.39.601 OR, 1.69; 0.346-8.222

542.e6 M. Surbatovic et al.

trauma in 37 patients found that the A allele was associatedwith both occurrence of sepsis (OR, 4.6; 95% CI, 1.9-10.9)and death (OR, 2.7; 95% CI, 0.6-7.3). Watanabe withcoauthors [15] analyzed 150 consecutive critically ill patientsrecruited at admission to the ICU, regardless of diagnosis,and found association between TNF-α308 A allele carriersand increased mortality rate.

Also, there are studies in which no association of TNF2allele and increased mortality rate was found. In one of those,total of 112 postoperative critically ill infected patients wereprospectively enrolled. Authors found that patients carryingthe TNF2 allele were not more likely to develop septic shockor did they have a higher mortality rate. Only when septicshock had developed, in that subgroup of patients, themortality rate was higher in those carrying the TNF2 allele[9]. Barber with coauthors [16] showed that TNF-α308polymorphisms were significantly associated with anincreased risk for severe sepsis after burn trauma, but noneof SNPs were significantly associated with mortality. TheTNF-α genotype was not related to mortality in the study inwhich 88 critically ill patients with multiple organdysfunction syndrome were enrolled [17].

No significant association of any TNF-α308 genotypewith pancreatitis as underlying cause of sepsis was found inour study. This is in accordance with findings in otherstudies [18,19].

In our study, association between distribution of 3 IL-101082 genotypes and the type of infecting microorganism isstatistically highly significant (P = .009). In patients withGram-positive sepsis AA and then AG genotypes were themost frequent ones. In patients with Gram-negative sepsis,GG genotype was the most frequent one. There are very fewstudies in literature available to us that are similar to ours inthis aspect. In a review regarding genetic variability in thesystemic inflammatory response, only one study of IL-101082polymorphism and type of infecting microorganism in severesepsis is mentioned: GG genotype was associated withgreater risk for meningococcal disease [20]. Taking inconsideration that Neisseria meningitidis is Gram-negativebacterium, this is in concordance with our finding.

As far as underlying cause of sepsis (peritonitis,pancreatitis) is concerned, we found that distribution of 3IL-101082 genotypes is also associated with statistically highsignificance (P = .003). In patients with sepsis associatedwith pancreatitis, AA genotype was the most frequent one. Inpatients with sepsis associated with peritonitis, GG genotypewas the most frequent one. Although any genetic factor thatalters the expression of regulatory cytokines-chemokinescould potentially alter the inflammatory response topancreatic injury [21], there are few studies that addressthis issue. One of them showed, contrary to our results, thatgenetic factors are not important in determination of diseaseseverity or susceptibility to severe acute pancreatitis [22].

Comparison of IL-101082 AA genotype with combinationof AG and GG showed that there is statistically significantassociation with outcome in our study. Critically ill patients

with G allele in either of genotypes have relative risk of death1.46; OR, 2.43; and 95% CI, 0.894 to 6.612. Most studiesalso demonstrated that G allele is associated with highermortality rate [23-25]. Interleukin-10 is one of the mostimportant antiinflammatory cytokines. It down-regulates theproduction of proinflammatory cytokines and chemokinesby activated monocytes, polymorphonuclear leucocytes, andeosinophils; prevents antigen-specific T-cell activation;inhibits T-cell expression; and potentiates the release of theinflammatory modulator IL-1ra. This IL-10 polymorphism,associated with high IL-10 inducibility, might influence theoutcome via induced immunosuppression and impairedbacterial clearance. One study showed that allele frequenciesand genotype distribution of the IL-101082 polymorphism didnot differ between surviving and dead patients [26] and theother, contrary to our results but in slightly different patientpopulation (all patients developed acute respiratory distresssyndrome, whereas in our study, most, but not all, had acuterespiratory distress syndrome), showed that G allele wasassociated with lower mortality rate [27].

Distribution of 3 CD14159 genotypes in our study showedthat there is statistically significant association with the typeof infecting microorganism. In patients with Gram-positivesepsis, CT genotype was the most frequent one. In sepsispatients with CC and TT genotypes, frequency of Gram-positive and Gram-negative infecting microorganisms wassimilar. Furthermore, we found that critically ill patientswithout T allele in either of genotypes, that is, with CCgenotype have relative risk of Gram-negative sepsis, 1.25;OR, 1.45; and 95% CI, 0.381 to 5.572. In this comparison, insepsis patients with CC genotype, Gram-negative bacteriawere most frequent. Contrary to our results, Sutherland withcoworkers [28] found that CD14159 TT was associated withincreased prevalence of Gram-negative bacteria in a cohortof 252 critically ill whites and Yuan and coauthors [29]reported that prevalence of CD14159 CC genotype wassignificantly higher in 85 children with pneumococcal (ie,Gram-positive) sepsis, whereas, Calvano and Lowry [30]and found no association between CD14 polymorphisms andincreased risk of Gram-negative infections.

CD14 is an innate immunity receptor for lipopolysaccha-ride, peptidoglycan, and lipoteichoic acid. CD14 is found onthe surface of monocytes, macrophages, neutrophils, andhepatocytes and as a soluble form in serum. Transition of C toT at position 159 in the promoter of the CD14 gene has beenassociated with increased density of membrane-bound CD14on monocytes and increased serum levels of soluble CD14.This may alter host recognition and clearance of pathogens.

We found no significant association of any CD14polymorphism with outcome, same as Hubacek et al [31]and Nakada et al [32]. However, some authors reportedsignificant association of TT genotype with higher mortalityrate in patients with severe sepsis and septic shock [10] andsignificant association of C allele with decreased risk ofmultiple organ dysfunction and sepsis in major traumapatients [33].

542.e7Polymorphisms of genes

We analyzed distribution of 6 IL-1ra gene intron 2genotypes, and no significant association of any genotypewith underlying cause of sepsis (peritonitis, pancreatitis),with the type of infecting microorganism or with outcomewas found. Similar findings were reported by Powell et al[19] who found no association of IL-1ra polymorphism withpancreatitis and Fang et al [34] who found no associationwith outcome in 93 patients with severe sepsis.

IL-1ra is a naturally occurring competitive inhibitor ofIL-1–induced proinflammatory activity. The IL-1ra gene ispolymorphic, resulting in quantitative differences in IL-1raproduction. The relative levels of IL-1ra and IL-1 at aninflammatory site will determine whether a proinflamma-tory response will be initiated and persist or will beterminated. Persons homozygous for allele 2 of the IL-1ragene have a more prolonged and more severe proinflam-matory immune response than persons with other IL-1ragenotypes because they produce significantly lower levelsof IL-1ra [11]. There are studies in which significantassociation between A2/A2 polymorphism and highermortality rate was demonstrated in patients with severesepsis [35,36].

Evidently, there are inconsistent findings in currentstudies of genetic associations in human trauma and/orsepsis. Strict critics attribute that to methodological andanalytical problems (underpowered studies, inadequate sizeof the study, and others), but they also state that, for example,it has been calculated for a general ICU population withsepsis or septic shock, a sample size of 2000 patients wouldbe required to detect a mortality relative risk of 1.5 from anypolymorphism to confidently rule out false-negative associa-tions. To our knowledge, no genetic association studyrecruited this number of patients. Until then, relatively smallpopulation studies should be taken in to account.

Although most studies are of small more or lesshomogeneous populations, the identification of strongassociations between certain genetic polymorphisms andincreased mortality rate, underlying cause of sepsis, or thetype of infecting microorganism is intriguing and supportsfurther research using this approach. The establishment ofthese associations does not equal causation, and furtherresearch is required in both genetic and molecular aspect ofhost immunoinflammatory response to various insults incritically ill patients.

References

[1] Jean-Baptiste E. Cellular mechanism in sepsis. J Intensive Care Med2007;22:63-72.

[2] Sorensen TI, Nielsen GG, Andersen PK, et al. Genetic andenvironmental influences on premature death in adult adoptees. NEngl J Med 1988;318:727-32.

[3] Read RC, Teare DM, Pridmore AC, et al. The tumor necrosis factorpolymorphism TNF (-308) is associated with susceptibility tomeningococcal sepsis, but not with lethality. Crit Care Med 2009;37:1237-43.

[4] De Maio A, Torres MB, Reeves RH. Genetic determinants influencingthe response to injury, inflammation and sepsis. Shock 2005;23:11-7.

[5] Ledoux D, Canivet JL, Preiser JC, et al. SAPS 3 admission score: anexternal validation in general intensive care population. Intensive CareMed 2008;34:1873-7.

[6] Le Gall JR, Lemeshow S, Saulnier F. A new Simplified AcutePhysiology Score (SAPS II) based on a European/North Americanmulticenter study. JAMA 1993;270:2957-63.

[7] Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity ofdisease classification system. Crit Care Med 1985;13:818-29.

[8] Moreno RP, Vincent JL, Matos R, et al. The use of maximum SOFAscore to quantify organ dysfunction/failure in intensive care. Results ofa prospective multicentre study. Intensive Care Med 1999;25:686-96.

[9] Tang GJ, Huang SL, Yien HW, et al. Tumor necrosis factor genepolymorphism and septic shock in surgical infection. Crit Care Med2000;28:2733-6.

[10] Gibot S, Cariou A, Drouet L, et al. Association between a genomicpolymorphism within the CD14 locus and septic shock susceptibilityand mortality rate. Crit Care Med 2002;30:969-73.

[11] Witkin SS, Gerber S, Ledger J. Influence of interleukin-1 receptorantagonist gene polymorphism on disease. CID 2002;34:204-9.

[12] Gordon AC, Lagan AL, Aganna E, et al. TNF and TNFRpolymorphisms in severe sepsis and septic shock: a prospectivemulticenter study. Genes Immun 2004;5:631-40.

[13] Mira JP, Cariou A, Grall F, et al. Association of TNF2, TNF-αpromoter polymorphism, with septic shock susceptibility andmortality. A multicenter study. JAMA 1999;282:561-8.

[14] O'Keefe GE, Hybki DL, Munford RS. The G → A single nucleotidepolymorphism at the -308 position in the tumor necrosis factor alphapromoter increases the risk for severe sepsis after trauma. J Trauma2002;52:817-25.

[15] Watanabe E, Hirasawa H, Oda S, et al. Extremely high interleukin-6blood levels and outcome in the critically ill are associated with tumornecrosis factor- and interleukin-1–related gene polymorphisms. CritCare Med 2005;33:89-97.

[16] Barber RC, Aragaki CC, Rivera-Chavez FA, et al. TLR4 and TNF-αpolymorphisms are associated with an increased risk for severe sepsisfollowing burn injury. J Med Genet 2004;41:808-13.

[17] Reid CL, Perrey C, Pravica V, et al. Genetic variation in proin-flammatory and anti-inflammatory cytokine production in multipleorgan dysfunction syndrome. Crit Care Med 2002;30:2216-21.

[18] Dianliang Z, Lieshou L, Zhiwei J, et al. Association of plasma levels oftumor necrosis factor (TNF)-α and its soluble receptors, twopolymorphisms of the TNF gene with acute severe pancreatitis andearly septic shock due to it. Pancreas 2003;26:339-43.

[19] Powell JJ, Fearon KC, Siriwardena AK, et al. Evidence against a rolefor polymorphisms at tumor necrosis factor, interleukin-1 andinterleukin-1-receptor antagonist gene loci in the regulation of diseaseseverity in acute pancreatitis. Surgery 2001;129:633-40.

[20] Waterer GW, Wunderink RG. Science review: genetic variability inthe systemic inflammatory response. Crit Care 2003;7:308-14.

[21] Papachristou GI, Clermont G, Sharma A, et al. Risk and markers ofsevere acute pancreatitis. Gastroenterol Clin North Am 2007;36:277-96.

[22] Zhang DL, Zheng HM, Yu BJ, et al. Association of polymorphisms ofIL and CD14 genes with acute severe pancreatitis and septic shock.World J Gastroenterol 2005;11:4409-13.

[23] Sabelnikovs O, Nikitina-Zake L, Zhuravlova J, et al. Association ofIL-10 promoter polymorphism -1082 G/A with adverse outcome insevere sepsis and septic shock. Crit Care 2009;13(Suppl 1):P355(abstr).

[24] Stanilova SA, Miteva LD, Karakolev ZT, et al. Interleukin-10-1082promoter polymorphism in association with cytokine production andsepsis susceptibility. Intensive Care Med 2006;32:260-6.

[25] Schaaf BM, Boehmke F, Esnaashari H, et al. Pneumococcal septicshock is associated with the interleukin-10-1082 gene promoterpolymorphism. Am J Respir Crit Care Med 2003;168:476-80.

542.e8 M. Surbatovic et al.

[26] Shu Q, Fang X, Chen Q, et al. IL-10 polymorphism is associatedwith increased incidence of severe sepsis. Chin Med J 2003;116:1756-9.

[27] Gong MN, Thompson BT, Williams PL, et al. Interleukin-10polymorphism in position -1082 and acute respiratory distresssyndrome. Eur Respir J 2006;27:674-81.

[28] Sutherland AM, Walley KR, Russell JA. Polymorphisms in CD14,mannose-binding lectin, and Toll-like receptor 2 are associated withincreased prevalence of infection in critically ill adults. Crit Care Med2005;33:638-44.

[29] Yuan FF, Marks K, Wong M, et al. Clinical relevance of TLR2, TLR4,CD14 and FcγRIIA gene polymorphisms in Streptococcus pneumo-niae infection. Immunol Cell Biol 2008;86:268-70.

[30] Agnese DM, Calvano JE, Hahm SJ, et al. Human toll-like receptor 4mutations but not CD14 polymorphisms are associated with anincreased risk of Gram-negative infections. J Infect Dis 2002;186:1522-5.

[31] Hubacek JA, Stuber F, Frohlich D, et al. The common functional C(-159)T polymorphism within the promoter region of the lipopoly-

saccharide receptor CD14 is not associated with sepsis developmentor mortality. Genes Immun 2000;1:405-7.

[32] Nakada T, Hirasawa H, Oda S, et al. Influence of toll-like receptor 4,CD14, tumor necrosis factor and interleukin-10 gene polymorphismson clinical outcome in Japanese critically Ill patients. J Surg Res2005;129:322-8.

[33] Gu W, Dong H, Jiang DP, et al. Functional significance of CD14promoter polymorphisms and their clinical relevance in a Chinese Hanpopulation. Crit Care Med 2008;36:2274-80.

[34] Fang XM, Schroder S, Hoeft A, et al. Comparison of twopolymorphisms of the interleukin-1 gene family: interleukin-1receptor antagonist polymorphism contributes to susceptibility tosevere sepsis. Crit Care Med 1999;27:1330-4.

[35] Arnalich F, Lopez-Maderuelo D, Codoceo R, et al. Interleukin-1receptor antagonist gene polymorphism and mortality in patients withsevere sepsis. Clin Exp Immunol 2002;127:331-6.

[36] Ma P, Chen D, Pan J, et al. Genomic polymorphism within interleukin-1 family cytokines influences the outcome of septic patients. Crit CareMed 2002;30:1046-50.