Clinical Chemistry 54:6 10001007 (2008)

Molecular Diagnostics and Genetics

Cell-Free Plasma DNA as a Predictor of Outcome in Severe Sepsis and Septic Shock
Katri Saukkonen,1* Pa ivi Lakkisto,2,4 Ville Pettila ,3 Marjut Varpula,1 Sari Karlsson,5 Esko Ruokonen,6 2 Kari Pulkki, for the Finnsepsis Study Group

BACKGROUND: Increased concentrations of cell-free DNA have been found in plasma of septic and critically ill patients. We investigated the value of plasma DNA for the prediction of intensive care unit (ICU) and hospital mortality and its association with the degree of organ dysfunction and disease severity in patients with severe sepsis. METHODS:

that decreases its clinical value in severe sepsis and septic shock.
2008 American Association for Clinical Chemistry

We studied 255 patients with severe sepsis or septic shock. We obtained blood samples on the day of study inclusion and 72 h later and measured cell-free plasma DNA by real-time quantitative PCR assay for the -globin gene. Cell-free plasma DNA concentrations were higher at admission in ICU nonsurvivors than in survivors (median 15 904 vs 7522 genome equivalents [GE]/mL, P 0.001) and 72 h later (median 15 176 GE/mL vs 6758 GE/mL, P 0.004). Plasma DNA values were also higher in hospital nonsurvivors than in survivors (P 0.008 to 0.009). By ROC analysis, plasma DNA concentrations had moderate discriminative power for ICU mortality (AUC 0.70 0.71). In multiple regression analysis, first-day plasma DNA was an independent predictor for ICU mortality (P 0.005) but not for hospital mortality. Maximum lactate value and Sequential Organ Failure Assessment score correlated independently with the first-day plasma DNA in linear regression analysis.

RESULTS:

CONCLUSIONS: Cell-free plasma DNA concentrations were significantly higher in ICU and hospital nonsurvivors than in survivors and showed a moderate discriminative power regarding ICU mortality. Plasma DNA concentration was an independent predictor for ICU mortality, but not for hospital mortality, a finding

Severe sepsis and septic shock has remained a challenge in intensive care despite improvements in modern treatment. It represents the major cause of death in patients admitted to the intensive care unit (ICU),7 with a mortality rate varying between 30% and 60% in different studies (1, 2 ). Several biomarkers have been evaluated for predicting mortality in patients with sepsis and its sequelae, severe sepsis and septic shock, but none has proven entirely useful in clinical practice. Increased concentrations of cell-free plasma DNA have been found in various clinical conditions, including trauma, cancer, stroke, and myocardial infarction (37 ). According to current evidence, the DNA is released into the circulation from apoptotic and necrotic cells (8 11 ), although the exact mechanism is not clear. Information from sex-mismatched transplantation models suggests that most of the plasma DNA is of hematopoietic origin in healthy transplant recipients (12, 13 ). Apoptosis plays a major role in the pathophysiological process in sepsis (14 ), and circulating DNA has been detected in the plasma of septic patients (15 ). Furthermore, increased plasma levels of nucleosomes, in which fragmented DNA is packed during apoptosis, have been found in patients with severe sepsis and septic shock (16 ). Preliminary data from ICU patients suggest that admission plasma DNA concentrations may be higher in nonsurvivors than in survivors (17, 18 ). We have recently shown that the maximum cell-free plasma DNA concentration in the first days of intensive care is independently associated with hospital

Departments of Medicine and Emergency Care, Helsinki University Central Hospital, Helsinki, Finland; 2 Department of Clinical Chemistry, Helsinki University Central Hospital, Helsinki, Finland; 3 Department of Surgery, Intensive Care Unit, Helsinki University Central Hospital, Helsinki, Finland; 4 Minerva Research Institute, Helsinki, Finland; 5 Department of Intensive Care Medicine, Tampere University Hospital, Tampere, Finland; 6 Department of Anesthesiology and Intensive Care Medicine, Kuopio University Hospital, Kuopio, Finland. * Address correspondence to this author at: Ilmarinkatu 8 A 18, 00100 Helsinki,

Finland. e-mail katri.saukkonen@helsinki.fi. Received November 26, 2007; accepted March 5, 2008. Previously published online at DOI: 10.1373/clinchem.2007.101030 7 Nonstandard abbreviations: ICU, intensive care unit; SAPS, Simplified Acute Physiology Score; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; GE, genome equivalent; IQR, interquartile range; AUC, area under the curve.

1000

Cell-Free Plasma DNA in Severe Sepsis and Septic Shock

mortality in critically ill patients (19 ). Moreover, in our earlier study, plasma DNA concentrations were higher in patients with infection. We therefore set out to investigate cell-free plasma DNA concentrations in a more homogeneous group of patients with severe sepsis or septic shock. The aim of this study was to evaluate the value of cell-free plasma DNA for predicting ICU and hospital mortality. The second aim was to evaluate the association of cell-free plasma DNA with degree of organ failure and disease severity. Materials and Methods

the nature of the study, no standardized protocol for treatment was used.
BLOOD SAMPLING

Blood samples were drawn at study inclusion and 72 h thereafter into lithium heparin tubes after obtaining written informed consent. The plasma fraction was separated by centrifugation as soon as possible, and samples were stored at 20 C or colder at the enrolling site before being sent to the Helsinki University Hospital, where samples were stored at 80 C.
PREPARATION OF PLASMA DNA AND REAL-TIME PCR

PATIENTS AND DEFINITIONS

This study was a part of a prospective observational cohort study about the incidence and prognosis of severe sepsis in Finland (the Finnsepsis study). The study was conducted in a 4-month period (November 1, 2004, to February 28, 2005) in 24 intensive care units in Finland. Approval for this study was granted by the ethics committees at each hospital. The design, methodology, and results of the Finnsepsis study are published elsewhere (20 ). All patients 18 years old consecutively admitted to participating ICUs during the study period were screened daily for severe sepsis and septic shock according to specific criteria. Severe sepsis at admission or during ICU treatment was defined according to American College of Chest Physicians (ACCP)/Society of Critical Care Medicine (SCCM) criteria (21 ): suspected or confirmed infection, 2 or more manifestations of systemic inflammatory response syndrome, and at least 1 sepsis-induced acute organ dysfunction. All patients meeting these 3 criteria were included in the Finnsepsis study.
DATA COLLECTION

Using the Finnish intensive care quality consortiums database (Intensium), we collected and stored the following patient information: demographic data; diagnosis by International Classification of Diseases, 10th edition; Simplified Acute Physiology Score (SAPS) II (22 ); Acute Physiology and Chronic Health Evaluation (APACHE) II score (23 ); and ICU and hospital mortality. We obtained 1-year mortality data from Statistics Finland. Basic hemodynamics and laboratory test data and the use of vasoactive medication during the ICU stay, as well as the amount of fluids administered in the first 24 h, were recorded on a daily basis. We calculated estimated creatinine clearance using the Cockcroft-Gault formula (24 ). We recorded data concerning the source of infection, use of antibiotic treatment, and the course of various organ dysfunctions and supportive treatments. The Sequential Organ Failure Assessment (SOFA) score (25 ) was calculated daily to assess the severity of organ dysfunction. Owing to

We prepared and measured plasma DNA by the same methodology we have used previously (19 ). Plasma samples were centrifuged at 16 000g for 10 min before DNA extraction to remove any residual cells (26 ). DNA was extracted from 200-L plasma samples using the QIAamp DNA Blood Mini Kit (Qiagen) according to the blood and body fluid protocol recommended by the manufacturer. We measured plasma DNA in duplicate samples by real-time quantitative PCR assay for the -globin gene (27 ) using the ABI PRISM 7000 sequence detection system (Applied Biosystems). The sequences were as follows: forward primer 5-GCA CCT GAC TCC TGA GGA GAA-3, reverse primer 5CAC CAA CTT CAT CCA CGT TCA-3, and a singlelabeled fluorescent MGB-probe 5-FAM-TCT GCC GTT ACT GCC CT-MGB-NFQ, where MGB is a minor groove binding molecule and NFQ a nonfluorescent quencher molecule. We used a 10-fold serial dilution of human genomic DNA (Roche) as a standard curve. Results are expressed as genome equivalents (GE)/mL; 1 GE equals 6.6 pg DNA. To determine the precision of the real-time quantitative PCR method, we performed PCR runs 8 times in duplicate with 2 samples and standard curve. The interassay CV for the threshold cycle values of PCR analyses was 0.5% to 2.1% over the whole dynamic range from 2 102 to 2 106 GE/mL, with the highest imprecision in the range from 2 102 to 2 103 GE/ mL. The investigators performing the measurements were not aware of the patients clinical course.
STATISTICAL ANALYSIS

Data are presented as median values and interquartile ranges (IQRs, 25th to 75th percentiles) or as absolute values and percentages. We compared differences in continuous variables with the nonparametric MannWhitney test and used 2 and Fisher exact test for categorical variables. We determinded bivariate correlations using nonparametric Spearman . Variables associated significantly with the plasma DNA concentration at baseline were then tested multivariately by linear regression analysis using log-transformed
Clinical Chemistry 54:6 (2008) 1001

Table 1. Comparison of the cell-free plasma DNA study group to the rest of the Finnsepsis study group.
Cell-free plasma DNA study group Rest of Finnsepsis study group

No. patients Age, years Male sex APACHE II score SAPS II score SOFA score on day 1 ICU mortality Hospital mortality
Data are median (IQR) or n (%).

255 60 (4972) 176 (69) 23 (1829) 42 (3254) 8 (611) 34 (13) 67 (26)

215 60 (5272) 139 (65) 24 (1830) 43 (3557) 8 (611) 39 (18) 65 (30)

0.72 0.20 0.42 0.15 0.82 0.09 0.19

plasma DNA concentration as the dependent variable. To determine the discriminative power of the cell-free plasma DNA for survival, we constructed ROC curves and calculated areas under the curve (AUCs) with 95% CIs. The best predictive cutoff values maximizing the sum of sensitivity and specificity were defined, and we calculated sensitivity, specificity, and positive likelihood ratios with 95% CIs using GraphROC for Windows (28 ). We performed a stepwise multiple logistic regression analysis with forward variable selection to identify factors that had independent predictive value for hospital and ICU mortality. In all tests, P 0.05 was considered significant. All statistical procedures used SPSS 12.0 statistical software. Results We obtained blood samples from 255 of 470 patients (54%) comprising the whole Finnsepsis study population. The reason for exclusion was inability to obtain consent for laboratory measurements. The study group did not differ from the rest of the study cohort concerning demographics, disease severity, or mortality and can therefore be considered as a representative sample of the severe sepsis population (Table 1). Of 255 patients, we obtained blood samples from 252 patients at study inclusion and from 220 patients 72 h later. ICU and hospital mortality rates were 13% (34 of 255) and 26% (67 of 255), respectively. Baseline characteristics of hospital survivors and nonsurvivors are presented in Table 2. The median cell-free plasma DNA concentration was 8070 GE/mL (IQR 388318 934 GE/ mL) at admission and 7457 GE/mL (IQR 3668 16 311 GE/mL) 72 h later. The cell-free plasma DNA concentrations were higher in ICU nonsurvivors than in survivors at admission (median 15 904 GE/mL vs 7522 GE/mL, P 0.001) and 72 h later (median 15 176 GE/mL vs 6758 GE/mL, P 0.004). Plasma DNA con1002 Clinical Chemistry 54:6 (2008)

centrations were also higher in hospital nonsurvivors than in survivors at admission (median 12 386 GE/mL vs 7678 GE/mL, P 0.009) and 72 h later (median 11 428 GE/mL vs 6414 GE/mL, P 0.008). The differences between the admission and 72-h measurements of plasma DNA were not found to be associated with ICU or hospital mortality (P 0.42 to 0.93). Cell-free plasma DNA concentration at inclusion correlated significantly with the first 24-h SOFA score (r 0.29, P 0.001), the maximum SOFA score during the first 7 days in the ICU (r 0.30, P 0.001) (Fig. 1), the APACHE II score (r 0.18, P 0.005), the SAPS II score (r 0.22, P 0.001), shorter height (r 0.13, P 0.04), higher body mass index (r 0.13, P 0.05), and several first-day laboratory parameters: maximum lactate concentration (r 0.40, P 0.001), plasma creatinine concentration (r 0.15, P 0.018), lowest estimated creatinine clearance (r 0.16, P 0.013), lowest platelet count (r 0.24, P 0.001), lower urine output (r 0.24, P 0.001), and urea concentration (r 0.21, P 0.005). Neither age nor sex correlated with the DNA concentration (P 0.32 to 0.77 and 0.06 to 0.77, respectively). Multivariate linear regression analysis revealed that the maximum lactate value (P 0.003) and the SOFA score during the first day of intensive care (P 0.015) were independently associated with plasma DNA concentration at baseline. ROC curve AUCs with 95% CIs were produced for cell-free DNA concentrations as used in predicting ICU and hospital mortality. For the prediction of ICU mortality, the AUC was 0.71 (95% CI 0.62 0.80) for the DNA concentration at inclusion and 0.70 (95% CI 0.57 0.82) for the DNA concentration 72 h later (Fig. 2). AUCs in predicting ICU mortality were slightly higher for the first-day maximum lactate value and SAPS II scores (0.77, 95% CI 0.68 0.85, and 0.75, 95% CI 0.65 0.84, respectively), whereas those for the

Cell-Free Plasma DNA in Severe Sepsis and Septic Shock

Table 2. Baseline characteristics of the study hospital survivors and nonsurvivors (n 255).
Survivors Nonsurvivors P

n Age, years Male sex Cardiovascular comorbidity Length of ICU stay, days Positive blood culture Primary site of infection Lung Intra-abdominal Urinary tract Central nervous system Soft tissue Trauma Hospital-acquired infection Ventilatory treatment Fluids in the first 24 h in ICU, mL Vasoactive medication on day 1 Maximum lactate in the first 24 h, mmol/L Creatinine clearance, mL/min APACHE II score SAPS II score SOFA score on day 1 Max SOFA score
Data are median (IQR) or n (%).

188 58 (4770) 136 (72) 39 (21) 5.8 (3.111.0) 54 (29) 76 (40) 59 (31) 9 (5) 6 (3) 21 (11) 4 (2) 58 (31) 131 (70) 4550 (25806890) 138 (73) 2.0 (1.33.0) 70 (44111) 22 (1728) 40 (3050) 8 (610) 8 (611)

67 65 (5676) 40 (60) 27 (40) 6.1 (3.112.8) 17 (25) 30 (45) 23 (34) 2 (3) 0 (0) 4 (6) 2 (3) 27 (40) 55 (82) 4410 (33307200) 55 (82) 3.7 (2.05.6) 45 (2390) 28 (2236) 54 (4064) 10 (713) 11 (815)

0.001 0.06 0.002 0.42 0.6 0.5 0.7 0.7 0.3 0.3 0.7 0.16 0.05 0.46 0.16 0.001 0.001 0.001 0.001 0.001 0.001

first-day SOFA score were slightly lower (AUC 0.69, 95% CI 0.58 0.80). For the prediction of hospital mortality, plasma DNA had less discriminative power: the AUC for the baseline DNA concentration was 0.61 (95% CI 0.53 0.69) and that for the DNA concentration at 72 h was 0.63 (95% CI 0.53 0.72). Because the prognostic value of plasma DNA differed for ICU and hospital mortality, we investigated the differences between these patient groups. ICU nonsurvivors (n 34) had higher APACHE II and maximum SOFA scores, higher lactate concentrations, and higher first-day plasma DNA concentrations than patients who died later in the hospital (n 33) (P 0.003, 0.018, 0.025, and 0.005, respectively). The groups of patients who died in the ICU or in the hospital were not found to differ in their age, sex, or treatment with vasoactive drugs (P 0.62 to 1.00). The best cutoff value of plasma DNA at baseline for ICU mortality was 12 000 GE/mL, with a sensitivity of 67% (95% CI 51% 80%), specificity of 67% (95% CI 62%72%), positive likelihood ratio of 2.03 (95%

CI 1.49 2.76), and correct classification rate of 67%. The best cutoff value of plasma DNA after 72 h was 12 500 GE/mL, with a sensitivity of 60% (95% CI 39% 78%), specificity of 70% (95% CI 64%75%), positive likelihood ratio of 2.00 (95% CI 1.323.03), and correct classification rate of 69%. The 1-year mortality rate in the study population was 40% (102 of 255). A Kaplan-Meier survival curve using the best cutoff value of plasma DNA at study inclusion is presented in Fig. 3. We performed multiple logistic regression analysis to identify factors having independent predictive value for mortality. All variables significantly associated with hospital mortality (SAPS II score minus SAPS II age score; SOFA score on day 1; age; maximum first-day lactate value; cardiovascular comorbidity; cell-free plasma DNA concentration at inclusion and 72 h later) and ICU mortality (SAPS II score minus SAPS II age score; SOFA score on day 1; age; maximum first-day lactate value; cell-free plasma DNA concentration at inclusion and 72 h later) were included in separate
Clinical Chemistry 54:6 (2008) 1003

Fig. 1. First-day plasma DNA concentrations in maximum SOFA score quartiles. Plasma DNA presented as median values (line), IQR (boxes), and 5th to 95th percentiles (whiskers). Differences in plasma DNA concentrations between the SOFA quartiles were significant between the first and fourth quartiles (P 0.001), second and fourth quartiles (P 0.004), and third and fourth quartiles (P 0.028) with Bonferroni-Dunn test.

Fig. 3. Kaplan-Meier survival curve according to the plasma DNA concentration at baseline. The selected cutoff limit is the best for predicting ICU mortality.

analyses. To assess the independent effect of age on mortality and to avoid duplication of information, we used the SAPS II score from which the SAPS II age score

had been subtracted. SAPS II score (P 0.001) and cardiovascular comorbidity (P 0.015) were independent predictors of hospital mortality, and first-day plasma DNA concentration (P 0.005) and SAPS II score (P 0.008) were independent predictors of ICU mortality (Table 3). Discussion To the best of our knowledge, this is the first large clinical study to evaluate the prognostic value of cell-free plasma DNA in patients with severe sepsis or septic shock. In the present study, the plasma DNA concentrations were significantly higher in ICU and hospital nonsurvivors than in survivors, and the first-day plasma DNA was an independent predictor of ICU mortality. However, plasma DNA concentration was not an independent predictor of hospital mortality, a preferred outcome measure in critical care medicine. The AUCs for plasma DNA concentration at inclusion and at 72 h showed a moderate discriminative power regarding ICU mortality, no better than that of maximum lactate value or SAPS II scores. Cell-free plasma DNA concentration at baseline had a moderate independent correlation with lactate value and first-day SOFA score. In accordance with our previous results with a nonhomogeneous group of critically ill patients (19 ), the cell-free plasma DNA concentration in the

Fig. 2. ROC curves for the plasma DNA concentrations at baseline (AUC 0.71, 95% CI 0.62 0.80) and 72 h later (AUC 0.70, 95% CI 0.57 0.82) in predicting intensive care unit mortality.

1004 Clinical Chemistry 54:6 (2008)

Cell-Free Plasma DNA in Severe Sepsis and Septic Shock

Table 3. Results of the multiple logistic regression analyses: independent predictors of hospital and ICU mortality.
Variable Odds ratio (95% CI) P

Hospital mortality SAPS II score, 1 point increase Cardiovascular comorbidity ICU mortality First-day plasma DNA concentration SAPS II score, 1 point increase 1.00 (1.001.00) 1.05 (1.011.09) 0.005 0.008 1.05 (1.021.08) 2.93 (1.246.94) 0.001 0.015

current study reflected the degree of present and developing organ failure. Our study was the first to demonstrate that rising or falling concentrations of plasma DNA were not related to the outcomes of patients with severe sepsis. Our results showing that nonsurvivors had higher plasma DNA concentrations than survivors are in agreement with earlier studies performed in intensive care patients (1719 ). Rhodes et al. (18 ) demonstrated in a critically ill patient population that patients who developed severe sepsis or septic shock had higher plasma DNA concentrations. In their study, plasma DNA also had a good discriminative power by ROC analysis for ICU and hospital mortality. The sample size was relatively small, however, and the study was not designed to evaluate the performance or predictive power of plasma DNA in sepsis. In addition, the number of deaths was not sufficient to evaluate the independent effect of plasma DNA on mortality. In our previous study of a critically ill intensive care population, in which one-fourth of the study patients had infection as their primary diagnosis in the ICU (19 ), the maximum plasma DNA concentrations of hospital nonsurvivors were lower (median 9366 GE/mL) than the baseline concentrations of severe sepsis or septic shock hospital nonsurvivors measured in the present study (median 12 386 GE/mL). Also the critically ill hospital survivors in our previous study had lower plasma DNA concentrations (median 6506 GE/mL) than the survivors in the present study (median 7678 GE/mL), though the difference was not as large (19 ). In our present study, the median concentration of cell-free plasma DNA on admission was 1.5-fold higher than in a previous study among the critically ill (17 ), but one-third lower than in another study of a similar patient population (18 ). These differences may arise from either selection of a different patient population or inconsistency in the preanalytical and PCRbased methodology. Without a standardized quantitative plasma DNA analysis protocol, it is always possible that the plasma DNA results are affected by differences

in methodology. With the method we used in the present study, the results are obtained within 3 h, which makes the method applicable in the clinical setting. By using novel fast real-time PCR systems, the turnaround time could be cut to 2 h. Several mechanisms may lead to hyperlactatemia, apoptosis, and organ failure in sepsis, although the main mechanism in early sepsis is hypoperfusion and tissue hypoxia. The persistent imbalance between oxygen delivery and consumption results in anaerobic glycolysis and lactate production. Oxygen deprivation may induce apoptotic cell death (29 ). We found that the degree of elevation of lactate concentrations correlated independently with admission plasma DNA concentration, which may reflect the effect of septic shock induced tissue hypoxia on apoptotic or necrotic cell death. The exact characteristics of plasma DNA kinetics and clearance have remained uncertain. It has been shown in an earlier study that fetal DNA is rapidly cleared in maternal plasma after delivery, with a mean half-life of only 16.3 min (30 ). Recently, increased cellfree plasma DNA concentrations after hemodialysis session were found to return to normal predialysis concentrations 30 min after finishing the session (31 ). It has been suggested that the liver and kidneys play a role in eliminating circulating plasma DNA. In mice, nucleotides are predominantly metabolized in the liver (32 ). Botezatu et al. (33 ) showed that approximately 0.5%2% of cell-free plasma DNA crosses the kidney barrier and is excreted in the urine. No clinical study has evaluated the impact of liver and renal failure on cell-free plasma DNA concentrations. The previous study investigating the effect of hemodialysis on plasma DNA levels found no significant difference between the predialysis plasma DNA concentrations of patients with chronic renal insufficiency and the healthy control group (31 ). In parallel with that observation, we found neither the serum creatinine nor the estimated creatinine clearance to have an independent effect on plasma DNA concentration in the
Clinical Chemistry 54:6 (2008) 1005

present study. However, the effect of impaired renal and hepatic function on circulating DNA levels needs further studies in large, well-characterized patient populations. The present study is the largest on this topic performed in an intensive care setting, with an unselected representative sample of severe sepsis patients. Owing to the 4-month study period, we cannot exclude any seasonal variation in sepsis epidemiology or outcome. Some limitations have to be considered. First, in addition to hemolysis, contamination of cell-free circulating plasma DNA measurements by residual white blood cells or platelets may provide a source of error. However, high-speed centrifugation at 16 000g after storage, as used in this study, completely eliminates cellular contamination in these assays (26 ). The choice of anticoagulants has also been shown to influence quantitative plasma DNA analysis (34 ). According to the current knowledge, EDTA is the anticoagulant of choice in delayed blood processing, but EDTA, heparin, or citrate as the anticoagulant produces similar quantitative plasma DNA results within 6 h of phlebotomy (34 ). In this multicenter study, we collected blood samples in heparin tubes and recommended that the plasma fraction be separated by centrifugation as soon as possible. We cannot exclude the possibility that these preanalytical factors could have had an effect on our results. It is evident that for better data comparability of different plasma DNA studies, standardized preanalytical protocols are required. Second, the PCR-based

method is time-consuming and therefore expensive in routine use. Automated sample preparation and analysis in a 24-h service are needed before these methods can be put into routine clinical use. The exact characteristics of plasma DNA kinetics also remain uncertain. Consequently, the timing of obtaining the blood samples may be a potential source of error, although in our study the blood samples were drawn quite uniformly at baseline and 72 h later. In addition, the correlations between the first-day plasma DNA concentration and lactate or SOFA score, although statistically significant, were, unfortunately, not particularly large. In our study, the cell-free plasma DNA had better discriminative power for ICU mortality than hospital mortality, and it was an independent predictor of ICU, but not hospital, mortality. This can be explained by the fact that plasma DNA is increased in patients with very advanced sepsis with established organ failure and, thus, higher early mortality. Although we did hypothesize the use of cell-free plasma DNA as a single marker, it is unlikely that any one marker could predict the outcome of a patient in a complicated disease like sepsis. Grant/Funding Support: Supported, in part, by EVO grant TYH 6235 from Helsinki University Central Hospital, Helsinki, Finland and a grant from Pa ivikki and Sakari Sohlberg Foundation. Financial Disclosures: None declared. Acknowledgments: We thank all participating investigators and study nurses of the Finnsepsis study.

References
1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States. Crit Care Med 2001; 29:130310. 2. Alberti C, Brun-Buisson C, Burchardi H, Martin C, Goodman S, Artigas A, et al. Epidemiology of sepsis and infection in ICU patients from an international multicentre cohort study. Intensive Care Med 2002;28:108 21. 3. Lo YMD, Rainer TH, Chan LYS, Hjelm NM, Cooks RA. Plasma DNA as a prognostic marker in trauma patients. Clin Chem 2000;46:319 23. 4. Rainer TH, Wong LKS, Lam W, Yuen E, Lam NYL, Metreweli C, Lo YMD. Prognostic use of circulating plasma nucleic acid concentration in patients with acute stroke. Clin Chem 2003;49:5629. 5. Chang CP-Y, Chia R-H, Wu T-L, Tsao K-C, Sun C-F, Wu JT. Elevated cell-free serum DNA detected in patients with myocardial infarction. Clin Chim Acta 2003;327:95101. 6. Rainer TH, Lam NY, Man CY, Chiu RW, Woo KS, Lo YM. Plasma -globin DNA as a prognostic marker in chest pain patients. Clin Chim Acta 2006;368:110 3. 7. Tong Y-K, Lo YMD. Diagnostic developments involving cell-free (circulating) nucleic acids. Clin Chim Acta 2006;363:18796. 8. Fournie GJ, Martres F, Pourrat JP, Alary C, Rumeau M. Plasma DNA as cell death marker in elderly patients. Gerontology 1993;39:21521. Fournie GJ, Courtin JP, Laval F, Chale JJ, Pourrat JP, Pujazon M-C, Lauque D, Carles P. Plasma DNA as a marker of cancerous cell death. Investigations in patients suffering from lung cancer and in nude mice bearing human tumours. Cancer Lett 1995;91:2217. Jiang N, Pisetsky DS. The effect of inflammation on the generation of plasma DNA from dead and dying cells in the peritoneum. J Leukoc Biol 2005; 77:17. Jahr S, Hentze H, English S, Hardt D, Fackelmayer FO, Hesch RD, Knippers R. DNA fragments in the blood plasma of cancer patients: quantitations and evidence for their origin from apoptotic and necrotic cells. Cancer Res 2001;61:1659 65. Lui YY, Chik KW, Chiu RW, Ho CY, Lam CW, Lo YMD. Predominant hematopoietic origin of cellfree DNA in plasma and serum after sex-mismatched bone marrow transplantation. Clin Chem 2002;48:4217. Lui YY, Woo KS, Wang AY, Yeung CK, Li PK, Chau E, Ruygrok P, Lo YM. Origin of plasma cell-free DNA after solid organ transplantation. Clin Chem 2003;49:495 6. Hotchkiss RS, Nicholson DW. Apoptosis and caspases regulate death and inflammation in sepsis. Nat Rev Immunol 2006;6:81322. 15. Martins GA, Kawamura MT, Da Costa Carvalho MDG. Detecting of DNA in the plasma of septic patients. Ann N Y Acad Sci 2000;906:134 40. 16. Zeerleder S, Zwart B, Wuillemin WA, Aarden LA, Groeneveld AB, Caliezi C, et al. Elevated nucleosome levels in systemic inflammation and sepsis. Crit Care Med 2003;31:194751. 17. Wijeratne S, Butt A, Burns S, Sherwood K, Boyd O, Swaminathan R. Cell-free plasma DNA as a prognostic marker in intensive treatment unit patients. Ann N Y Acad Sci 2004;1022:232 8. 18. Rhodes A, Wort SJ, Thomas H, Collinson P, Bennett ED. Plasma DNA concentration as a predictor of mortality and sepsis in critically ill patients. Crit Care 2006;10:R60. 19. Saukkonen K, Lakkisto P, Varpula M, Varpula T, Voipio-Pulkki L-M, Pettila V, Pulkki K. Association of cell-free plasma DNA with hospital mortality and organ dysfunction in intensive care unit patients. Intensive Care Med 2007;33:1624 7. 20. Karlsson S, Varpula M, Ruokonen E, Pettila V, Parviainen I, Ala-Kokko TI, et al. Incidence, treatment, and outcome of severe sepsis in ICUtreated adults in Finland: the Finnsepsis study. Intensive Care Med 2007;33:435 43. 21. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definition for sepsis and

9.

10.

11.

12.

13.

14.

1006 Clinical Chemistry 54:6 (2008)

Cell-Free Plasma DNA in Severe Sepsis and Septic Shock

22.

23.

24.

25.

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. Le Gall JR, Lemeshow S, Saumer F. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA 1993;270:2957 63. Knaus WA, Drapper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med 1985;13:818 29. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16:31 41. Vincent J-L, De Mendonc a A, Cantraine F, Moreno R, Takala J, Suter PM, Sprung CL, Colardyn F, Blecher S, on behalf of the working group on sepsis-related problems of the European Society of Intensive Care Medicine. Use of the SOFA score to assess the incidence of organ

26.

27.

28.

29.

30.

dysfunction/failure in intensive care units: results of a multicenter, prospective study. Crit Care Med 1998;26:1793 800. Swinkels DW, Wiegerinck E, Steegers EAP, de Kok JB. Effects of blood-processing protocols on cellfree DNA quantification in plasma. Clin Chem 2003;49:525 6. Lo YM, Tein MS, Lau TK, Haines CJ, Leung TN, Poon PM, et al. Quantitative analysis of fetal DNA in maternal plasma and serum: implications for noninvasive prenatal diagnosis. Am J Hum Genet 1998;62:768 75. Kairisto V, Poola A. Software for illustrative presentation of basic clinical characteristics of laboratory tests: GraphROC for Windows. Scand J Clin Lab Invest Suppl 1995;222:43 60. Brunelle JK, Chandel NS. Oxygen deprivation induced cell death: an update. Apoptosis 2002;7: 475 82. Lo YMD, Zhang J, Leung TN, Lau TK, Chang AMZ, Hjelm NM. Rapid clearance of fetal DNA from

31.

32.

33.

34.

maternal plasma. Am J Hum Genet 1999;64:218 24. Moreira VG, de la Cera Martinez T, Gonzales EG, Garcia BP, Alvarez Menendez FV. Increase in and clearance of cell-free plasma DNA in hemodialysis quantified by real-time PCR. Clin Chem Lab Med 2006;44:1410 5. Gauthier VJ, Tyler LN, Mannik M. Blood clearance kinetics and liver uptake of mononucleosomes in mice. J Immunol 1996;156:1151 6. Botezatu I, Serdyuk O, Potapova G, Shelepov V, Alechina R, Molyaka Y, et al. Genetic analysis of DNA excreted in urine: a new approach for detecting specific genomic DNA sequences from cells dying in an organism. Clin Chem 2000;46: 1078 84. Lam NYL, Rainer TH, Chiu RWK, Lo YMD. EDTA is a better anticoagulant than heparin or citrate for delayed blood processing for plasma DNA analysis. Clin Chem 2004;50:256 7.

Clinical Chemistry 54:6 (2008) 1007