Sepsis is a devastating illness that accounts for more than 210,000 deaths annually in the United States (1). Previously believed to be a syndrome of uncontrolled inflammation, sepsis is now recognized to initiate both proinflammatory and anti-inflammatory immune mechanisms that, in many cases, result in prolonged periods of immunosuppression (2–5). Sepsis-induced derangements of innate and adaptive immunity prevent pathogen clearance and predispose to secondary infections, as evidenced by autopsy studies that have demonstrated ongoing foci of infections and functional defects in immune effector cells in patients who have died of sepsis (6, 7). This immune dysfunction is a primary cause of late mortality in sepsis (3, 8).
An important feature of sepsis-induced immunosuppression is apoptosis-related loss of immune cells, including CD4 and CD8 T, B, and follicular dendritic cells (3, 9). Clinical studies have demonstrated that circulating levels of lymphocytes fall during the onset of sepsis and can remain depressed for up to 28 days (10–16). Antiapoptotic therapies that prevent lymphocyte death have successfully improved survival in animal models of sepsis (17, 18), and immunomodulatory therapy may be a promising area of future clinical research (19, 20). However, human phenotypic variability in the immune response to infection suggests that the broad application of immunomodulatory therapy across a broad cohort of patients with sepsis is unlikely to show benefit (21). Ideally, immune-stimulating therapy would be applied only to those patients who are known to be persistently immunosuppressed. Therefore, strategies to better identify these patients in the clinical setting should be pursued.
Prolonged lymphopenia is a candidate marker of persistent immunosuppression in patients with sepsis, and absolute lymphocyte counts are easily measured during routine care. Previous studies have shown that low absolute lymphocyte counts are predictive of postoperative sepsis and are a better predictor of bacteremia than conventional infection markers in the emergency care unit (22, 23). Studies have also shown that septic nonsurvivors have sustained reductions in specific T and B cell populations during the first week after sepsis diagnosis (11–16). However, data are limited, and these studies did not often evaluate the absolute lymphocyte count. We therefore aimed to determine whether persistent lymphopenia over the first 4 days following the diagnosis of sepsis predicts mortality. This time point was chosen, in part, to avoid including patients in the nonsurvivor group who died during the initial hyperinflammatory stage of sepsis rather than during the immunosuppressive phase. Also, studies of patients with sepsis have shown significantly decreased levels of lymphocyte subpopulations in adult nonsurvivors compared with survivors at days 3 and 7, suggesting that day 4 absolute lymphocyte counts could potentially discriminate between these two groups (11–13). Finally, previous literature has demonstrated that persistent lymphopenia on the fourth day after presentation was predictive of death in trauma patients, another group of critically ill patients at risk for developing persistent immunosuppression (24). We hypothesized that patients with persistently low lymphocyte counts would be more likely to die and to develop secondary infections compared with patients whose lymphocyte counts recovered to normal by the fourth day after sepsis diagnosis.
MATERIALS AND METHODS
We conducted a retrospective observational cohort study and reported its results in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (25). Data collection and analysis were approved by the Human Research Protection Office at our institution with waiver of informed consent.
Study setting and population
This study was conducted at a 1,167-bed university-affiliated adult tertiary care hospital between January 1, 2010, and July 31, 2012. All patients admitted to the hospital with sepsis and blood cultures positive for bacteria and/or fungal organisms within 5 days after hospital admission were evaluated for inclusion. Sepsis was defined according to consensus criteria (26). For the purpose of this study, the diagnosis of sepsis required the presence of at least two systemic inflammatory response syndrome criteria within 24 h of the time the positive blood culture was collected. Furthermore, patients were included only if they survived until at least the fourth day after the positive blood culture was drawn and also had a complete blood cell count with differential (including lymphocyte, neutrophil, and monocyte counts) collected on that day. Each patient was included in the study only once. For patients who had multiple hospital admissions with positive blood cultures during the study period, only data from the most recent hospitalization were included. For patients with multiple positive blood cultures during a single admission, data collection began at the time of the first positive culture.
Exclusion criteria included age younger than 18 years, blood cultures positive only for common contaminates (e.g., coagulase-negative Staphylococcus species, Propionibacterium species, Corynebacterium species), diagnosis of hematological or immunological disease, and treatment with chemotherapy agents or corticosteroids within 6 months prior to or during the hospitalization.
A computer query was used to identify all adult patients with positive blood cultures admitted during the study period. Detailed chart review was then performed to exclude patients based on inclusion and exclusion criteria and to collect demographic, microbiology, and outcome data. Leukocyte counts were collected by a separate research assistant blinded to the study hypothesis, the patients’ baseline characteristics, and outcomes.
The first 24-h time period following the culture collection time was considered to be day 1; the next 24-h period, day 2; etc. If multiple leukocyte counts were collected within any 24-h period, the mean value was documented for that period. Lymphopenia was defined as an absolute lymphocyte count less than 1.2 × 103 cells/μL, which is the lower limit of normal at our institution. Severe lymphopenia was defined as an absolute lymphocyte count less than 0.6 × 103 cells/μL, which is half of the lower limit of normal. Severe sepsis and septic shock were defined per consensus criteria (26).
Baseline demographics included age, sex, source of bacteremia, presence of comorbidities, and modified Acute Physiology and Chronic Health Evaluation (APACHE) II score (excluding the Glasgow Coma Score) (27, 28). The source of bacteremia was determined by the presence of concurrent cultures growing the same organism from another site or from documentation in the medical record by the treating physician. Requirement for vasopressors, requirement for mechanical ventilation, and surgical procedures prior to or during the 4-day study period were documented. Microbiology and antibiotic susceptibility data were collected for each positive blood culture. Antibiotic administration times were recorded, and length of time to appropriate antibiotic coverage was calculated.
For every patient, the white blood cell (WBC) count, neutrophil count, lymphocyte count, and monocyte count were collected for the first 4 days following the culture collection date.
The primary outcome was 28-day mortality. Secondary outcomes included the development of secondary infections, 1-year mortality, hospital length of stay, and intensive care unit (ICU) length of stay. Secondary infections were defined as culture-positive infections identified greater than 48 h after the primary bacteremia and arising from a secondary source. Identification of secondary infections was based on culture growth of a new organism at any site and documentation of a new infection by the treating physician in the medical record.
Descriptive statistics, including mean (SD), median (interquartile range [IQR] defined as the 25th and 75th percentile), and frequency distributions were used to describe the patient cohort. Normality was assessed using histograms and the Kolmogorov-Smirnov test. Variables that failed tests of normality were presented and analyzed in a nonparametric fashion. Comparison of baseline characteristics between 28-day survivors and nonsurvivors was done using unpaired t tests, Mann-Whitney U tests, or χ2 tests, as appropriate.
Because not every patient had complete blood cell counts collected on days 1 through 3, longitudinal analyses of the blood cell counts (WBC count, absolute neutrophil count, absolute lymphocyte count, and absolute monocyte count) were based on a repeated-measures mixed-model analysis of variance with an autoregressive covariance structure as implemented by the MIXED procedure in SAS (SAS Institute Inc, Cary, NC). The primary focus of these analyses was on the significance of the interactions that tested hypotheses regarding the equality of changes over time in the two groups. Within the framework of the mixed-model analysis and when the interaction was significant, the appropriate statistical contrasts were used to test the null hypotheses that (a) changes between days 1 and 4 in survivors were equal to corresponding changes in nonsurvivors; (b) day 1 values were equal for survivors and nonsurvivors; and (c) day 4 values were equal for survivors and nonsurvivors. P values were not adjusted for multiple comparisons. Data were log transformed prior to analysis.
To determine the independent effect of the day 4 absolute lymphocyte count on mortality after accounting for significant confounders, multivariable logistic regression with a forced entry method was used to model the odds of death at 28 days and 1 year. All variables that were significant at P = 0.05 in univariable analysis were included in the model. For all continuous variables (except absolute lymphocyte count), the odds ratios (ORs) reflect the increased odds of death for a 1-unit increase in the baseline variable. For absolute lymphocyte count, the OR indicates the increased odds of death for every increase in the absolute lymphocyte count by 0.5 × 103 cells/μL. This value was chosen to improve clinical interpretation of the OR given that a normal absolute lymphocyte count is only 1.2 × 103 cells/μL. For categorical variables, the reference category for the OR was absence of the condition. Colinearity diagnostics were evaluated to ensure variable independence. Multivariable models report ORs adjusted for all variables in the model.
Based on previous literature (11–13, 24), day 4 absolute lymphocyte count was selected a priori to be an independent variable in analysis of the primary and secondary outcomes. However, to determine whether earlier absolute lymphocyte counts could also predict 28-day and 1-year mortality, we performed additional post hoc univariable and multivariable logistic regression analyses using the day 3 absolute lymphocyte count as an independent variable.
Patients were divided into three groups based on their day 4 absolute lymphocyte count. Group 1 consisted of patients with normal lymphocyte counts (≥1.2 × 103 cells/μL) by day 4; group 2 included patients with moderate persistent lymphopenia (day 4 count 0.7–1.1 × 103 cells/μL); and group 3 consisted of patients with severe persistent lymphopenia (day 4 count ≤0.6 × 103 cells/μL). The reason for separating persistently lymphopenic patients into moderate and severe groupings was to facilitate observation of graduated changes in outcomes associated with increasingly severe lymphopenia. Kruskal-Wallis one-way analysis of variance and χ2 tests were used for comparisons of baseline characteristics between the three lymphocyte groups. Kaplan-Meier survival curves were created and compared using the log-rank test. In addition, univariable logistic regression was used to model the odds of developing secondary infections using group assignment as the independent variable. The reference category for the reported ORs was absence of persistent lymphopenia.
A receiver operating characteristic (ROC) plot was created to illustrate the ability of day 1 and day 4 absolute lymphocyte counts to predict 28-day mortality. Lower values of absolute lymphocyte counts were chosen to indicate a more positive test, such that higher areas under the curve (AUCs) indicated increased discriminatory ability. Confidence intervals (CIs) on AUCs were calculated using nonparametric assumptions.
Except where otherwise stated, all statistical tests were carried out using SPSS 21.0 (SPSS Inc, Chicago, Ill). All tests were two-tailed, and P < 0.05 was considered statistically significant.
A total of 335 patients were included in the study (Fig. 1). Seventy-six patients (22.7%) died within 28 days. Table 1 reports the baseline characteristics for the entire patient cohort. Nonsurvivors were older and had more severe illness, longer time to antibiotic administration, and higher incidences of congestive heart failure, chronic renal insufficiency, and liver disease.
Box plots demonstrating the distributions of the WBC count, absolute lymphocyte count, absolute monocyte count, and absolute neutrophil count in survivors and nonsurvivors for the first 4 days after the diagnosis of sepsis are shown in Figure 2. On day 1, the absolute lymphocyte counts were well below the lower limit of normal and not significantly different between the survivors (median, 0.7 × 103 cells/μL [IQR, 0.4–1.1 × 103 cells/μL]) and nonsurvivors (median, 0.6 × 103 cells/μL [IQR, 0.4–1.1 × 103 cells/μL]) (P = 0.35). By day 4, the absolute lymphocyte count was significantly higher in the survivors (median, 1.1 × 103 cells/μL [IQR, 0.7–1.5 × 103 cells/μL]) compared with the nonsurvivors (median, 0.7 × 103 cells/μL [IQR, 0.5–1.0 × 103 cells/μL]) (P < 0.0001), and the overall increase in absolute lymphocyte count over the 4-day period was significantly greater in the survivors (P = 0.0004). There was no significant difference in the overall change in the WBC count (P = 0.06) or absolute monocyte count (P = 0.11) over the 4-day period between the two groups. The absolute neutrophil count at day 4 was significantly lower in the survivors (median, 8.2 × 103 cells/μL [IQR, 5.9–13.1 × 103 cells/μL]) compared with the nonsurvivors (median, 9.2 × 103 cells/μL [IQR, 7.2–14.1 × 103 cells/μL]) (P = 0.04), and the overall decrease in absolute neutrophil count over the 4-day period was significantly greater in survivors (P = 0.03).
Table 2 shows the multivariable analysis of factors associated with 28-day mortality. Day 4 absolute lymphocyte count was significantly associated with 28-day mortality (P = 0.009) with an adjusted OR of 0.68 (95% CI, 0.51–0.91). Day 4 absolute lymphocyte count was also associated with 1-year mortality (P = 0.008) with an adjusted OR of 0.74 (95% CI, 0.59–0.93). In the subset of patients with septic shock, day 4 absolute lymphocyte count continued to be a significant predictor of both 28-day (OR, 0.69 [95% CI, 0.48–0.98]; P = 0.04) and 1-year mortality (OR, 0.71 [95% CI, 0.51–0.97]; P = 0.03).
Mortality at 28 days and 1 year was significantly higher in patients with severe persistent lymphopenia (39.5%, 57.9%) and moderate persistent lymphopenia (24.6%, 40.3%) compared with those without persistent lymphopenia (10.4%, 28.8%). The Kaplan-Meier survival analyses comparing patients based on day 4 lymphocyte groups are shown in Figure 3.
In the 305 patients who had complete blood cell counts drawn on day 3, univariable analysis demonstrated that day 3 absolute lymphocyte count was also a significant predictor of 28-day mortality (OR, 0.62 [95% CI, 0.46–0.83]; P = 0.001) and 1-year mortality (OR, 0.71 [95% CI, 0.57–0.89]; P = 0.002). However, the adjusted ORs for day 3 absolute lymphocyte count were not statistically significant for death at 28 days or 1 year when other predictors of mortality were included in multivariable analyses.
Table 3 summarizes the baseline characteristics and outcomes of patients stratified by day 4 absolute lymphocyte count. Patients with persistent lymphopenia had higher APACHE II scores, whereas patients without persistent lymphopenia had a lower incidence of diabetes. There were no differences between the groups in hospital or ICU lengths of stay. Table 4 shows the occurrence of secondary infections based on the presence of persistent lymphopenia. The presence of severe persistent lymphopenia (OR, 2.11 [95% CI, 1.02–4.39]; P = 0.04), but not moderate persistent lymphopenia (OR, 1.60 [95% CI, 0.83–3.11]; P = 0.16), was significantly associated with development of secondary infections compared with those without persistent lymphopenia.
The discriminatory ability of day 1 and day 4 absolute lymphocyte counts to predict 28-day mortality is demonstrated in Figure 4 using ROC plots. Day 4 absolute lymphocyte count, but not day 1 absolute lymphocyte count, demonstrated a significantly higher AUC compared with a reference of 0.5. The performance of specific day 4 absolute lymphocyte count thresholds to predict 28-day mortality is indicated on the graph. The most accurate discriminatory threshold was calculated to be 1.0 × 103 cells/μL, which demonstrated a sensitivity of 76% and a specificity of 56%.
Information regarding immune phenotype in patients with sepsis is vital before any consideration of immunomodulatory therapeutic interventions can occur. The current study provides new findings in this domain and offers potential insight into areas for further research. Our findings have several implications.
After initial predominance of a proinflammatory cytokine-driven response, many patients with sepsis develop persistent and profound immunosuppression prior to succumbing to the syndrome (3, 7, 8, 15). The results of this study demonstrate that a persistently low level of circulating lymphocytes on the fourth day following the diagnosis of sepsis independently predicts short- and long-term survival and may serve as a biomarker for sepsis-induced immunosuppression. The immune response to sepsis is extremely variable and can change dramatically as the syndrome progresses. Patients who die early do so as a result of overwhelming hyperinflammation marked by cardiovascular collapse and multiple organ failure (4, 29). Many of the patients who survive this phase, however, go on to develop a compensatory anti-inflammatory response characterized by increased inhibitory receptors on T cells and antigen-presenting cells, decreased production of proinflammatory cytokines, expansion of myeloid-derived suppressor cells, and apoptosis-related loss of lymphocytes and dendritic cells (7, 9, 30–32). In the current study, only 42 patients (who otherwise would have met inclusion criteria) died prior to the fourth day after the diagnosis of sepsis. This accounted for only 35.6% of the total deaths due to sepsis, signifying that the majority of patients who ultimately died were at risk for sepsis-induced immunosuppression.
Previous studies have demonstrated persistently decreased levels of specific subpopulations of B and T cells in adult septic nonsurvivors compared with survivors but have not shown significant differences in the overall absolute lymphocyte count (11–13, 16). Our study, which included a large number of septic nonsurvivors, demonstrated that whereas absolute lymphocyte counts decrease to similarly low levels in survivors and nonsurvivors at the onset of sepsis, nonsurvivors’ absolute lymphocyte counts remain persistently low, while survivors experience lymphocyte recovery. We theorize that the initial fall in circulating lymphocytes at the onset of infection in patients with sepsis reflects two separate processes. First, lymphocytes are recruited out of the peripheral circulation to areas of infection and inflammation, and second, sepsis induces a number of stimuli that trigger lymphocyte apoptosis (18). Furthermore, we postulate that the persistent lymphopenia seen in nonsurvivors is most likely due to ongoing sepsis-induced lymphocyte apoptosis secondary to continued release of proapoptotic stimuli during the immunosuppressive phase of sepsis (12, 32–34). These findings are especially striking because the optimal host response to infection should be to increase lymphocyte proliferation and thereby augment the number of lymphocyte effector cells.
Another significant finding of our study is that patients with severe persistent lymphopenia had a significantly higher incidence of secondary infections compared with patients whose absolute lymphocyte counts had recovered to normal by day 4, and there was also a strong trend toward increased secondary infections in the group with moderate persistent lymphopenia. These results suggest that increased susceptibility to new infections may be contributing to the higher mortality seen in persistently lymphopenic patients.
Similar to prior studies (10), we found no differences in the absolute monocyte counts of septic survivors compared with nonsurvivors; the median monocyte counts of the entire cohort of patients with sepsis were within the reference range throughout the 4-day study period. It is well known, however, that monocytes show severe functional deficits, which contribute to the immunosuppressive state in patients with sepsis (34–36). Laboratory procedures designed to identify monocytic defects (such as measurement of monocyte HLA (human lymphocyte antigen)–DR expression or ex vivo lipopolysaccharide-induced tumor necrosis factor α levels) have successfully been shown to predict survival in patients with sepsis (36–39). These tests have also been used to guide therapy in randomized controlled trials of granulocyte-macrophage colony-stimulating factor, a cytokine that enhances monocyte proliferation and phagocytosis in adult and pediatric patients with sepsis (19, 20). However, these tests are not widely available in typical clinical settings and therefore cannot be routinely used by clinicians to assess the prognosis or immune status of patients with sepsis. In contrast, absolute lymphocyte counts are easily measured in most clinical laboratories and are already frequently obtained in infected patients. It is a routine test that could be used both clinically and in future trials of sepsis therapies to identify the patients at highest risk for immunosuppression and death.
For this study, we chose to focus on day 4 absolute lymphocyte counts based on previously published evidence that indicated that lymphocyte counts on this day would discriminate between survivors and nonsurvivors (11–13, 24). The time point at which patients with sepsis transition from a predominantly proinflammatory state to one of hypoinflammation is unknown and is likely extremely variable because of host-, pathogen-, and disease-related factors. Therefore, we also analyzed day 3 lymphocyte counts in the subset of patients who had complete blood counts measured on that day to determine whether persistent lymphopenia at an earlier time point could also predict death. We found that day 3 absolute lymphocyte counts were not associated with short- or long-term mortality after accounting for other covariates. Despite this, determining the time at which continued lymphopenia becomes most clinically relevant is difficult to ascertain, and there are likely subsets of patients (e.g., elderly patients) in whom the lymphocyte pattern diverges between survivors and nonsurvivors at an earlier or later time point than that seen in this study. Similarly, lymphopenia was defined a priori as an absolute lymphocyte count less than 1.2 × 103 cells/μL for this study because this was the lower limit of normal specified by the laboratory at our institution. Based on our ROC analysis, however, a day 4 absolute lymphocyte count less than 1.0 × 103 cells/μL may be a more accurate discriminatory threshold to optimally predict 28-day mortality. In the future, larger studies that could stratify patients by age or disease severity may clarify whether different absolute lymphocyte cutoff values would be more useful in specific subpopulations of patients with sepsis.
By including only culture-positive patients with bacteremia, we aimed to limit heterogeneity in the cohort by exclusively studying known patients with sepsis. Patients without positive blood cultures were excluded to avoid including patients incorrectly diagnosed with sepsis based on growth of colonizing bacteria in respiratory, urine, or other cultures. In addition, we limited our study population to patients whose positive cultures were drawn within 5 days of admission to the hospital to avoid biasing our study with chronically hospitalized patients who might be more likely to have underlying immunological dysfunction as well as a greater risk of death. To improve generalizability, we did not restrict our population to only patients with severe sepsis or septic shock. Despite this, approximately 93% of our patients were identified as having severe sepsis, suggesting that the patient cohort was fairly homogenous in terms of physiological perturbation. To further validate our results, we performed multivariable analyses to identify predictors of 28-day and 1-year mortality in the subset of patients with septic shock, and these yielded similar ORs for absolute lymphocyte count as in the multivariable analyses that included the entire cohort.
This study has several limitations. As a retrospective study, it was prone to limitations inherent in this study design, such as imbalance between the study groups. In contrast to previous studies of lymphopenia in patients with sepsis, though, we attempted to account for major care-related and patient-specific determinants of survival by adjusting for these factors in a multivariable model. In fact, an important aspect of this study was that we included a sufficient number of nonsurvivors to show that persistent lymphopenia increased the risk of death at 28 days and 1 year even after accounting for other known predictors of mortality such as age, APACHE II score, time until appropriate antibiotic coverage, and comorbidities. As a retrospective study, data collection was limited to variables that were obtained during the usual clinical care of the patients, so we were unable to correlate persistent lymphopenia with other known biomarkers of sepsis-induced immunosuppression such as decreased monocyte HLA-DR expression, overexpression of programmed cell death 1 receptor on T cells, or decreased ex vivo lipopolysaccharide-induced tumor necrosis factor α levels. Instead, we used the development of secondary infections as a clinically relevant surrogate marker of immunosuppression. Going forward, future clinical trials in this area should measure not only clinical outcomes, but also biomarkers of immunosuppression to further establish causation.
Another limitation of this study is that the exact time of sepsis onset could not be precisely determined. For the purpose of this study, we defined the onset of sepsis to be the time that the first positive culture was ordered by the treating physician. The majority of included patients, however, had blood cultures drawn in the emergency department, so potential delays in presentation to the hospital further complicate interpretation of timing in this study. Furthermore, we excluded patients who died or were discharged prior to day 4. Although this may limit some of the generalizability of our results, it allowed us to focus on sepsis-induced immunosuppression by eliminating patients experiencing early death (which is typically due to cardiovascular collapse and not immunosuppression) or early clinical recovery.
Finally, the main limitation of this study is that the results do not allow us to conclude whether persistent lymphopenia directly contributes to mortality in patients with sepsis or whether it is simply a marker of disease severity. We are also not able to presume that immune-stimulatory therapy aimed at reversing lymphopenia would alter patient outcomes. We postulate that ongoing lymphocyte apoptosis caused by the continued release of proapoptotic stimuli contributes to morbidity and mortality in patients with sepsis by inhibiting clearance of primary infections and increasing susceptibility to secondary infections. This hypothesis is supported by multiple animal studies, which have suggested that lymphocyte apoptosis is a key pathogenic mechanism in sepsis and that prevention of lymphocyte apoptosis improves survival (40–43). Still, prospective studies are needed to establish lymphopenia as a causative factor in late mortality in sepsis.
The clinical value of persistent lymphopenia is yet to be elucidated. Because of the enormous overlap in lymphocyte counts among survivors and nonsurvivors, day 4 absolute lymphocyte count would most likely be useful as a predictor of mortality in groups of patients rather than an accurate predictor of death in any individual patient. More important, perhaps, is the potential value of day 4 absolute lymphocyte count in predicting an increased risk of secondary infection and sepsis-induced immunosuppression. Currently, in the clinical setting, a persistently low absolute lymphocyte count in patients with sepsis should prompt clinicians to reevaluate their patients’ response to therapy and assess for the presence of new or untreated infections. In the future, sepsis therapies could be tailored to individuals’ immunological phenotypes. Potential immunotherapeutic agents, such as interleukin 7 and anti–programmed cell death 1 receptor antibody, act to increase CD4 and CD8 T-cell production, block lymphocyte apoptosis, and prevent T-cell exhaustion. These therapies may be most effective if administered selectively to patients with evidence of lymphocyte dysfunction or loss. Identification of patients with persistent lymphopenia will select not only for those at high risk for short- and long-term mortality, but also for those with the specific immunological derangements that these therapies aim to improve.
The authors thank Karen Steger-May, with the Division of Biostatistics at Washington University in St Louis, for her assistance with statistical analysis.
1. Lagu T, Rothberg MB, Shieh MS, Pekow PS, Steingrub JS, Lindenauer PK: Hospitalizations, costs, and outcomes of severe sepsis in the United States 2003 to 2007. Crit Care Med 40 (3): 754–761, 2012.
2. Munford RS, Pugin J: Normal responses to injury prevent systemic inflammation and can be immunosuppressive. Am J Respir Crit Care Med 163 (2): 316–321, 2001.
3. Hotchkiss RS, Karl IE: The pathophysiology and treatment of sepsis. N Engl J Med 348 (2): 138–150, 2003.
4. Angus DC, van der Poll T: Severe sepsis and septic shock. N Engl J Med 369 (9): 840–851, 2013.
5. Ward PA: Immunosuppression
in sepsis. JAMA 306 (23): 2618–2619, 2011.
6. Torgersen C, Moser P, Luckner G, Mayr V, Jochberger S, Hasibeder WR, Dunser MW: Macroscopic postmortem findings in 235 surgical intensive care patients with sepsis. Anest Analg 108 (6): 1841–1847, 2009.
7. Boomer JS, To K, Chang KC, Takasu O, Osborne DF, Walton AH, Bricker TL, Jarman SD 2nd, Kreisel D, Krupnick AS, et al.: Immunosuppression
in patients who die of sepsis and multiple organ failure. JAMA 306 (23): 2594–2605, 2011.
8. Schefold JC, Hasper D, Volk HD, Reinke P: Sepsis: time has come to focus on the later stages. Med Hypotheses 71 (2): 203–208, 2008.
9. Hotchkiss RS, Swanson PE, Freeman BD, Tinsley KW, Cobb JP, Matuschak GM, Buchman TG, Karl IE: Apoptotic cell death in patients with sepsis, shock, and multiple organ dysfunction. Crit Care Med 27 (7): 1230–1251, 2009.
10. Venet F, Davin F, Guignant C, Larue A, Cazalis MA, Darbon R, Allombert C, Mougin B, Malcus C, Poitevin-Later F, et al.: Early assessment of leukocyte alterations at diagnosis of septic shock. Shock 34 (4): 358–363, 2010.
11. Monserrat J, de Pablo R, Reyes E, Diaz D, Barcenilla H, Zapata MR, de la Hera A, Prieto A, Alvarez-Mon M: Clinical relevance of the severe abnormalities of the T cell compartment in septic shock patients. Crit Care 13 (1): R26, 2009.
12. Hein F, Massin F, Cravoisy-Popovic A, Barraud D, Levy B, Bollsert PE, Gibot S: The relationship between CD4+
regulatory T cells and inflammatory response and outcome during shock states. Crit Care 14 (1): R19, 2010.
13. Monserrat J, de Pablo R, Diaz-Martin D, Rodriguez-Zapata M, de la Hera A, Prieto A, Alvaraez-Mon M: Early alterations of B cells in patients with septic shock. Crit Care 17 (3): R105, 2013.
14. Inoue S, Suzuki-Utsunomiya K, Okada Y, Taira T, Iida Y, Miura N, Tsuji T, Yamagiwa T, Morita S, Chiba T, et al.: Reduction of immunocompetent T cells followed by prolonged lymphopenia in severe sepsis in the elderly. Crit Care 41 (3): 810–819, 2013.
15. Felmet KA, Hall MW, Clark RS, Jaffe R, Carcillo JA: Prolonged lymphopenia, lymphoid depletion, and hypoprolactinemia in children with nosocomial sepsis and multiple organ failure. J Immunol 174 (6): 3765–3772, 2005.
16. Cheadle WG, Pemberton RM, Robinson D, Livingston DH, Rodriguez JL, Polk HC Jr: Lymphocyte subset responses to trauma and sepsis. J Trauma 35 (6): 844–849, 1993.
17. Oberholzer C, Oberholzer A, Bahjat FR, Minter RM, Tannahill CL, Abouhamze A, LaFace D, Hutchins B, Clare-Salzler MJ, Moldawer LL: Target adenovirus-induced expression of IL-10 decreases thymic apoptosis and improves survival
in murine sepsis. Proc Natl Acad Sci U S A 96 (20): 14541–14546, 2001.
18. Hotchkiss RS, Tinsley KW, Swanson PE, Change KC, Cobb JP, Buchman TG, Korsmeyer SJ, Karl IE: Prevention of lymphocyte cell death in sepsis improves survival
in mice. Proc Natl Acad Sci U S A 96 (25): 14541–14546, 1999.
19. Meisel C, Shefold JC, Pschowski R, Baumann T, Hetzger K, Gregor J, Weber-Carstens S, Hasper D, Keh D, Zuckermann H, et al.: Granulocyte-macrophage colony-stimulating factor to reverse sepsis-associated immunosuppression
: a double-blind, randomized, placebo-controlled multicenter trial. Am J Respir Crit Care Med 180 (7): 640–648, 2009.
20. Hall MW, Knatz NL, Vetterly C, Tomarello S, Wewers MD, Vold HD, Carcillo JA: Immunoparalysis and nosocomial infection in children with multiple organ dysfunction syndrome. Intensive Care Med 37 (3): 525–532, 2011.
21. Angus DC: The search for effective therapy for sepsis: back to the drawing board? JAMA 306 (23): 2614–2615, 2011.
22. Lewis RT, Klein H: Risk factors in postoperative sepsis: significance of preoperative lymphocytopenia. J Surg Res 26 (4): 365–371, 1979.
23. de Jager C, van Wijk P, Mathoera RB, de Jongh-Leuvenink J, van der Poll T, Wever PC: Lymphocytopenia and neutrophil-lymphocyte count ratio predict bacteremia
better than conventional infection markers in an emergency care unit. Critical Care 14 (5): R192, 2010.
24. Heffernan DS, Monaghan SF, Thakkar RK, Machan JT, Cioffi WG, Ayala A: Failure to normalize lymphopenia following trauma is associated with increased mortality, independent of the leukocytosis pattern. Crit Care 16 (1): R12, 2012.
25. Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; for the STROBE Initiative: The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Int Med 147 (8): 573–577, 2007.
26. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL, Ramsay G: International Sepsis Definitions Conference: 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Intensive Care Med 29 (4): 530–538, 2003.
27. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM, Sprung CL, Colardyn F, Blecher S: Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Crit Care Med 26 (11): 1793–1800, 1998.
28. Vincent JL, Moreno R, Takala J, Willatts S, de Mendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG: The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. Intensive Care Med 22 (7): 707–710, 1996.
29. for the Early Goal-Directed Therapy Collaborative Group: Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 345 (19): 1368–1377, 2001.
30. Venet F, Chung CS, Monneret G, Huang X, Garber M, Ayala A: Regulatory T cell populations in sepsis and trauma. J Leukoc Biol 83 (3): 523–535, 2008.
31. Ertel W, Kremer JP, Kenney J, Steckholzer U, Jarrar D, Trentz O, Schildberg FW: Downregulation of proinflammatory cytokine release from whole blood from septic patients. Blood 85 (5): 1341–1347, 1995.
32. Weighardt H, Heidecke CD, Emmanuilidis K, Maier S, Bartels H, Siewert JR, Holzmann B: Sepsis after major visceral surgery is associated with sustained and interferon-gamma–resistant defects of monocyte cytokine production. Surgery 127 (3): 309–315, 2000.
33. Rigato O, Salomao R: Impaired production of interferon-gamma and tumor necrosis factor-alpha but not of interleukin 10 in whole blood of patients with sepsis. Shock 19 (2): 113–116, 2003.
34. Sinistro A, Almerighi C, Ciaprini, Natoli S, Sussarello E, Di Fino S, Calò-Carducci F, Rocchi G, Bergamini A: Downregulation of CD40 ligand response in monocytes from sepsis patients. Clin Vaccine Immunol 15 (12): 1851–1858, 2008.
35. Munoz C, Carlet J, Fitting C, Misset B, Blériot JP, Cavaillon JM: Dysregulation of in vitro
cytokine production by monocytes during sepsis. J Clin Invest 88 (5): 1747–1754, 1991.
36. Monneret G, Lepape A, Voirin N, Bohé J, Venet F, Debard AL, Thizy H, Bienvenu J, Gueyffier F, Vanhems P: Persisting low monocyte human leukocyte antigen-DR expression predicts mortality in septic shock. Intensive Care Med 32 (8): 1175–1183, 2006.
37. Monneret G, Elmenkouri N, Bohé J, Debard AL, Gutowski MC, Bienvenu J, Lepape A: Analytical requirements for measuring monocytic human lymphocyte antigen DR by flow cytometry: application to monitoring of patients with septic shock. Clin Chem 48 (9): 1589–1592, 2002.
38. Volk HD, Reinke P, Krausch D, Zuckermann H, Asadulla K, Müller JM, Dōcke WD, Kox WJ: Monocyte deactivation—rationale for a new therapeutic strategy in sepsis. Intensive Care Med 22 (Suppl 4): S474–S481, 1996.
39. Dōcke WD, Randow F, Syrbe U, Krausch D, Asadulla K, Reinke P, Volk HD, Kox W: Monocyte deactivation in septic patients: restoration by IFN-gamma treatment. Nat Med 3 (6): 678–681, 1997.
40. Unsinger J, McGlynn M, Kasten KR, Hoekzema AS, Watanabe E, Muenzer JT, McDonough JS, Tschoep J, Ferguson TA, McDunn JE, et al.: IL-7 promotes T cell viability, trafficking, and functionality and improves survival
in sepsis. J Immunol 184 (7): 3768–3779, 2010.
41. Unsinger J, Burnham CA, McDonough J, Morre M, Prakash PS, Caldwell CC, Dunne WM Jr, Hotchkiss RS: Interleukin-7 ameliorates immune dysfunction and improves survival
in a 2 hit model of fungal sepsis. J Infect Dis 206 (4): 606–616, 2012.
42. Brahmamdam P, Inoue S, Unsinger J, Chang KC, McDunn JE, Hotchkiss RS: Delayed administration of anti–PD-1 antibody reverses immune dysfunction and improves survival
in sepsis. J Leukoc Biol 88 (2): 233–240, 2010.
43. Zhang Y, Zhou Y, Lou J, Li J, Bo L, Zhu K, Wan X, Deng X, Cai Z: PD-L1 blockade improves survival
in experimental sepsis by inhibiting lymphocyte apoptosis and reversing monocyte dysfunction. Crit Care 14 (6): R220, 2010.