Mortality from sepsis has increased >90% over the last 20 years, at a hospital cost of nearly $17 billion a year (1,2). Each year, 660,000 to 750,000 persons become septic in the United States (1,3). Despite therapy being initiated in nearly all patients with sepsis, 120,000 to 210,000 people die of the disease annually (1,3).
Antibiotics are a mainstay of sepsis treatment, and have been repeatedly demonstrated to improve survival in both human studies and animal research models of sepsis (4–8). However, many animal studies of sepsis use a single strain of inbred mice with a single injury (7,9–11), and therefore do not assess whether antibiotics have a similar efficacy across a spectrum of injury severity. In contrast, human studies, by definition, use a heterogeneous patient population, and generally have broad entry criteria making it difficult to separate effects of genetics, age, disease severity, and antibiotic efficacy on survival (12–14).
Although antibiotics are frequently beneficial, they are not universally effective, and many patients with sepsis die despite seemingly appropriate antimicrobial therapy. Although prognostic scoring systems exist that can accurately predict outcomes of large patient populations, they are not useful in predicting survival of individual patients (15,16). One marker that has been demonstrated to be associated with sepsis mortality is interleukin (IL)-6. IL-6 levels drawn 6 h after an identical septic insult predict outcome in inbred mice with very high specificity (5,17). Elevated IL-6 levels are also associated with poor outcome in patients with sepsis (18–20).
Whether antibiotics have similar effects regardless of injury severity and whether antibiotics change survival in animals predicted to die based upon high IL-6 levels is unknown. Therefore, we used cecal ligation and puncture (CLP), a well-accepted model of intra-abdominal sepsis that yields a reproducible, titratable mortality depending on the gauge of needle used (21) to see whether changing injury severity impacts the ability of antibiotics to confer a survival advantage, whether we could identify an IL-6 level that predicts 100% mortality regardless of the injury used, and whether antibiotic therapy instituted after the onset of sepsis improves survival in animals predicted to die based upon elevated IL-6 levels.
MATERIALS AND METHODS
CLP was performed on 6-to 8-week-old male ND4 mice (Harlan, Indianapolis, IN) by the methods of Baker et al. (7,22) and as described previously. Briefly, anesthesia was induced with 5% halothane and was maintained with 2.5% halothane. The cecum was exteriorized via a small abdominal incision, and was then ligated immediately distal to the ileocecal valve without causing intestinal obstruction. The ligated cecum was punctured with a variable-sized hollow-bore needle (gauge varied with each experiment and is outlined in results below), and was then gently squeezed to extrude stool. The cecum was then replaced, and the abdomen was closed in layers. All mice received 1 mL of 0.9% NaCl via subcutaneous injection to compensate for insensible fluid loss. Sham mice were treated identically, except that the cecum was neither ligated nor punctured. All animals were acclimatized for 7 days before manipulation and were maintained on 12-h light-dark cycles with free access to food and water at all times. Experiments were conducted in accordance with the National Institutes of Health guidelines for the use of laboratory animals and approval from the Washington University Animal Studies Committee.
Animals were randomized to receive imipenem (25 mg/kg; Merck, West Point, PA) or an equivalent volume of 0.9% NaCl administered subcutaneously. To mimic the clinical situation in which there is a delay between the onset of sepsis and the initiation of antimicrobial therapy, antibiotics were begun 12 h after CLP. We based this upon consensus opinion that the abdomen must be exposed to infectious material for at least 12 h for an intra-abdominal infection to become established (4) and the fact that patients were enrolled up to 24 h after identification of an infectious source in a recent study improving survival in sepsis (14). Injections were repeated every 12 h and continued for 5 days or until death. Animals were followed 10 days postoperatively for survival.
Animals had 150 mL of blood harvested from their tail vein 6 h after CLP. Blood was drawn into a microhematocrit tube, transferred to a 0.5-mL microcentrifuge tube, and centrifuged at 3300 g for 5 min to separate plasma. IL-6 was measured by enzyme-linked immunosorbent assay (ELISA) using commercially available kits (R&D Systems, Minneapolis, MN) according to manufacturer specifications. IL-6 levels reported represent the averages of two ELISA experiments with duplicates of each sample run in both experiments.
Differences in group survival were analyzed by log-rank analysis. Cytokine levels were analyzed using nonparametric (Mann-Whitney) statistics. Data analysis was performed using GraphPad Prism 3.0 (GraphPad Software, San Diego, CA). P values ≤ 0.05 were considered to be statistically significant.
Antibiotic efficacy and injury severity
Mice (n = 54–64/group) were subjected to double-puncture CLP with either a 19-, 21-, or 23-gauge needle and were randomized to receive a 5-day course of 0.9% NaCl or imipenem, beginning 12 h postoperatively. Mortality increased with size of needle used, with animals given 0.9% NaCl having 16%, 26%, and 52% survival, respectively (Figs. 1 and 2). Anti-microbial therapy with the broad-spectrum antibiotic imipenem resulted in a significant improvement in survival at all injury levels (Fig. 1). The absolute efficacy of imipenem therapy was independent of injury severity because antibiotic treatment rescued 17% to 23% of mice, regardless of whether the different-sized needle resulted in moderate, severe, or near-lethal sepsis (Fig. 2).
IL-6 levels predict mortality
A different cohort of mice received either a mild septic insult with single-puncture 21-gauge CLP (n = 20) or a severe septic insult with double-puncture (n = 17) 21-gauge CLP. Six hours later, blood was drawn to measure IL-6 levels. Mice were then followed for 10 days post-CLP to assess survival (Fig. 3).
After survival curves had been generated, IL-6 levels were correlated to animal outcome in a post hoc analysis to see whether they predicted mortality as has previously been shown in BALB/c (17) and FVB/N (5) mice. Animals that died of either mild or severe sepsis had higher IL-6 levels than those that survived (Fig. 4). IL-6 levels were 13,819 ± 1,621 pg/mL in animals that subsequently died compared with 7,235 ± 541 pg/mL in animals that lived (P < 0.05). Retrospective analysis of the scatter plot of cytokine levels demonstrated that all animals (n = 7) with IL-6 levels >14,000 pg/mL died (Fig. 4).
Antibiotics do not improve survival in animals predicted to die by elevated IL-6 levels
A third cohort of mice (n = 15-17/group) had IL-6 levels drawn 6 h after being randomized to double-puncture CLP with a 21-, 23-, or 25-gauge needle to vary injury severity. Animals were further randomized to receive imipenem or 0.9% NaCl 12 h postoperatively and were followed for survival to see if antibiotics could rescue animals predicted to have a 100% mortality by having an IL-6 level >14,000 pg/mL. Blood for IL-6 levels was drawn 6 h before animals were given imipenem or 0.9% NaCl, so this allowed an analysis of whether antibiotics change outcome in animals that appear destined to die by elevated IL-6 levels if not treated with antimicrobial therapy. As in prior experiments, mortality was increased with larger needle size, and antibiotics conferred a survival advantage at all injury levels (P < 0.05 for all injuries, data not shown).
Animals that died had higher IL-6 level than surviving mice, regardless of whether they received antibiotics (Fig. 5). Pooled results from all three levels of injury severity showed that IL-6 levels were 14,030 ± 2,383 pg/mL in animals that received imipenem and subsequently died compared with 7,935 ± 557 pg/mL in animals treated with antibiotics that lived (P < 0.05, Fig 5). Similarly, IL-6 levels were 10,902 ± 1,147 pg/mL in animals that received 0.9%NaCl and subsequently died compared with 7,015 ± 1,125 pg/mL in untreated animals that lived (P = 0.057, Fig. 5).
All animals (n = 13) with IL-6 levels >14,000 pg/mL died, regardless of whether they received antibiotics (Fig. 5). Mortality was also independent of injury severity because all animals with IL-6 levels above >14,000 pg/mL died, regardless of whether they received a mild, moderate, or severe septic insult with a 25-, 23-, or 21-gauge needle, respectively. Pooled results from all experiments in which cytokine levels were drawn at 6 h (n = 131) demonstrate that IL-6 levels greater than 14,000 pg/mL predicted death with a specificity of 100%, a sensitivity of 31% and a positive predictive value of 100%.
This study demonstrates that antibiotics confer a similar absolute reduction in risk of death in murine intra-abdominal sepsis, regardless of the severity of insult. Elevated IL-6 levels drawn 6 h after the onset of sepsis were predictive of worse outcome, and levels >14,000 pg/mL were associated with 100% mortality. Although antimicrobial therapy started 12 h after CLP improved survival in animals regardless of severity of sepsis, antibiotics were unable to rescue mice that appeared to be destined to die based upon IL-6 levels > 14,000 pg/mL.
These results extend our understanding of the role of antibiotics in sepsis in a number of ways. We have previously shown that antibiotic efficacy is age dependent, with young (4-month-old) C57B1/6 mice receiving a substantial benefit in a double-puncture CLP with 25-gauge needle, whereas old (2-year-old) mice had no improvement in survival after antibiotic therapy (7). Our current results show that absolute risk reduction was similar in moderate, severe, and near-lethal sepsis, with a 17% to 23% reduction in each. However, relative risk reduction was substantially greater in animals with larger insults. Mice with near-lethal sepsis that received a double-puncture CLP with a 19-gauge needle had their relative risk of death decrease greater than 100%, with survival more than doubling from 16% in the untreated group to 34% in mice treated with imipenem. In contrast, mice with moderate sepsis from a 23-gauge needle had a similar absolute risk reduction (18% vs. 17% respectively), but their relative risk of death decreased only 33%, with survival improving from 52% in the untreated group to 69% in animals that received imipenem. The age-dependent and injury severity-dependent effects of antibiotics on survival may be important in designing and interpreting future sepsis trials. It has been proposed that part of the survival discrepancy between animal studies and human studies of anti-inflammatory agents in sepsis is due to the fact that preclinical trials used models of sepsis with higher mortality than subsequent clinical trials (23). Our results suggest that injury severity may be important in assessing the relative benefit of antibiotic therapy in future trials and highlights the potential importance of using different groups with different expected mortalities when analyzing the efficacy of antisepsis agents.
Our results also illustrate the importance of IL-6 levels as a predictor of outcome in critical illness. Remick et al. (17) have previously shown that IL-6 levels drawn 6 h after CLP predict mortality in female BALB/c mice. Similarly, we have shown that IL-6 levels predict mortality in FVB/N mice subjected to Pseudomonas aeruginosa pneumonia (5). Our data extend these results to male mice from the outbred ND4 strain. However, a substantially different critical cytokine level was predictive of death in each murine study. IL-6 levels >2000 pg/mL in BALB/c mice predicted 3-day CLP mortality with a specificity of 97%, whereas IL-6 levels >3600 pg/mL in FVB/N mice had 100% mortality in pneumonia. In contrast, nearly all ND4 mice in this study had IL-6 levels >3600 pg/mL, regardless of whether they survived. In addition, IL-6 levels are similar 6, 10, and 24 h after CLP in C57B1/6 and A/J mice even though mortality is higher in C57B1/6 mice although IL-6 levels were not specifically used as a predictor of mortality in this study (24). Taken together, these data suggest that the disparate contributions of genetic background and/or injury type to IL-6 levels would make it extremely difficult to use this test as a specific predictor of outcome in human sepsis.
Although IL-6 appears to be physiologically important in the pathophysiology of sepsis (25–27), the results presented herein use the cytokine only as a marker to predict mortality and do not address the functional role of IL-6 and the immune status of the animals subjected to injuries of varying severity. IL-6 has been demonstrated to have both pro- and anti-inflammatory effects (17,25–30). The functional significance of elevated IL-6 levels is unclear. Andrejko et al. (28) demonstrated that elevated serum IL-6 levels do not correlate with intrahepatic IL-6 activity, which is decreased after CLP. IL-6 knockout mice have no difference in mortality when subjected to CLP (29), and anti-IL-6 antibody does not effect survival in endotoxemia (30). However, anti-IL-6 antibody increases survival after CLP, potentially through reduced expression of C5aR levels (25). Interestingly, both high-dose and low-dose anti-IL-6 antibody are less effective in CLP, suggesting that excessive and insufficient IL-6 levels may be detrimental in sepsis (25). It is unknown if anti-IL-6 antibody would improve survival in ND4 mice with IL-6 levels greater than 14,000 pg/mL, although this experiment could potentially delineate the functional role this cytokine plays in CLP mortality.
The fact that antibiotics were unable to rescue animals that appeared to be destined to die if left untreated based upon IL-6 levels >14,000 pg/mL was unexpected. In the study examining IL-6 levels with BALB/c mice, animals were given antibiotics 2 h after CLP but 4 h before IL-6 levels were drawn (personal communication from D.G. Remick). In contrast, we gave antibiotics 12 h after CLP in this study, 6 h after IL-6 levels were drawn. We believe this mimics the clinical situation where a delay commonly occurs between onset of sepsis and initiation of antimicrobial therapy. Although antibiotics improved survival in mild, moderate, severe, and near-lethal sepsis in this manuscript, all animals with IL-6 levels >14,000 6 h after CLP died, regardless of whether they received antibiotics and regardless of the gauge of needle used for CLP. Therefore, even if the initial injury appears identical, a subset of animals are destined to die, and this can be identified early by IL-6 levels above a critical threshold of 14,000 pg/mL. Because antibiotics do not improve survival in this subset of animals, this suggests that the underlying physiology of these animals is different from those with lower IL-6 levels at this time point and that these mice either require antibiotics very rapidly after the onset of sepsis to receive a benefit from antimicrobial therapy or that they will die no matter when treatment is initiated.
Although this study provides new insights into the relationship between antibiotics and survival in murine intra-abdominal sepsis, it has a number of limitations. Although animals that received CLP with each of the different needle gauges had IL-6 levels above 14,000 pg/mL (data not shown), only five animals that received antibiotics were identified that had IL-6 levels that surpassed this “death threshold.” Although the data in Figures 4 and 5 (n = 20 total) are convincing that IL-6 levels >14,000 pg/mL predict death, it is possible that the relatively low number of animals that received antibiotics and had markedly elevated IL-6 levels led to an overly broad conclusion about the inability of antimicrobial therapy given 12 h after sepsis to improve survival. Furthermore, our efforts to make our study independent of injury severity by grouping together disparate CLP injuries may be obscuring an important variable that might have been identified if a greater number of animals was used for each injury. We also only used male animals because of studies showing gender is an independent predictor of mortality in critical illness in humans and animal models (31–33). Because previous studies using IL-6 levels as a predictor of survival have used not only different genetic strains but also either female mice (17) or male and female mice (5) (without correlating cytokine levels to gender), the role of gender in antibiotic efficacy and IL-6 levels needs to be more fully addressed. Another limitation is the difference in mortality in double-puncture CLP injury with a 21- or 23-gauge needle between animals that had a blood draw of 150 mL 6 h after CLP (source data for Fig. 5) and those that did not (Fig. 1). Although survival for animals receiving antibiotics was similar regardless of whether blood was drawn for IL-6 levels (49% vs. 56% in 21-gauge and 69% vs. 67% in 23-gauge), survival was lower in animals that did not receive antibiotics but had blood drawn at 6 h (52% vs. 26% in 21-gauge and 27% vs. 13% in 23-gauge). The impact of removing 10% to 15% of a mouse's blood volume is clearly an added stressor, but this appeared to impact mortality in animals that had a higher likelihood of dying (i.e., those that did not receive antibiotic therapy) while not affecting those with higher survival (i.e., those treated with antibiotics). The relationship between tail blood harvest and mortality likely reflects interactions between volume withdrawn, injury severity, and mortality. Preliminary data further examining this issue from our laboratory does not reproducibly show increased mortality with a 150-mL blood draw at 6 h (unpublished observations, I. R. Turnbull, T. G. Buchman), and this relationship merits further study.
One final limitation is that our results are based upon the assumption that waiting 12 h after CLP to begin antibiotics appropriately mimics the clinical situation in patients. Although a 12-h delay in initiation of antimicrobial therapy is common in people, this does not mean that a similar delay is appropriate in mice unless one believes that a 25-g mouse acts similar to a 70-kg patient, and the septic insults in both are similar. In fact, the rapidly fatal, high mortality models used in this manuscript may not be describing an outcome that is directly relevant to human sepsis (9). Although antimicrobial therapy is used to mimic the human condition, many support modalities used in patients—such as mechanical ventilation, adequate source control of infection, continued volume status monitoring and fluid resuscitation, and pressors—are not initiated in mice. Mice may also become septic quicker than the patients we are attempting to model as evidenced by the fact that mice subjected to CLP typically die quicker than the more chronic timeframe seen in the intensive care unit (9). Although some of these problems are inherent in using young mice with a single infectious insult to model older patients with diverse septic etiologies, it is possible that the 12-h delay after CLP before initiation of antibiotics is simply too long given the severity of murine sepsis. Future experiments drawing IL-6 levels at 6 h and then immediately starting antibiotics would help to define whether a window exists where antibiotics can rescue severely septic mice or if a subset of mice are destined to die, regardless of when therapy is initiated.
Despite these limitations, these data expand our understanding of the complex relationship between injury severity, antibiotics, and IL-6 levels as a predictor of outcome in sepsis. Our results demonstrate that although antibiotics are beneficial in groups of animals with varied injury severity, they have no benefit when started 12 h after the onset of sepsis in individual animals predicted to die by elevated IL-6 levels. The distinction between a survival benefit conferred to a group of genetically similar mice and the lack of survival benefit in a subset of individual mice that can be identified by elevated IL-6 levels is likely due to the individual animal's host response to sepsis. Further study is needed to elucidate the mechanisms underlying this discrepancy.
1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR: Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 29:1303–1310, 2001.
2. Murphy SL: Deaths: final data for 1998. Natl Vital Stat Rep 48:1–105, 2000.
3. Martin GS, Mannino DM, Eaton S, Moss M: The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med 348:1546–1554, 2003.
4. Mazuski JE, Sawyer RG, Nathens AB, DiPiro JT, Schein M, Kudsk KA, Yowler C: The Surgical Infection Society guidelines on antimicrobial therapy for intra-abdominal infections: an executive summary. Surg Infect 3:161–173, 2002.
5. Coopersmith CM, Amiot DM, Stromberg PE, Dunne WM, Davis CG, Osborne DF, Husain KD, Turnbull IR, Karl IE, Hotchkiss RS, Buchman TG: Antibiotics improve survival and alter the inflammatory profile in a murine model of sepsis from Pseudomonas aeruginosa
pneumonia. Shock 19:408–414, 2003.
6. Newcomb D, Bolgos G, Green L, Remick DG: Antibiotic treatment influences outcome in murine sepsis: mediators of increased morbidity. Shock 10:110–117, 1998.
7. Turnbull IR, Wlzorek JJ, Osborne D, Hotchkiss RS, Coopersmith CM, Buchman TG: Effects of age on mortality and antibiotic efficacy in cecal ligation and puncture. Shock 19:310–313, 2003.
8. Iregui M, Ward S, Sherman G, Fraser VJ, Kollef MH: Clinical importance of delays in the initiation of appropriate antibiotic treatment for ventilator-associated pneumonia. Chest 122:262–268, 2002.
9. Deitch EA: Animal models of sepsis and shock: a review and lessons learned. Shock 9:1–11, 1998.
10. Remick DG, Newcomb DE, Bolgos GL, Call DR: Comparison of the mortality and inflammatory response of two models of sepsis: lipopolysaccharide vs. cecal ligation and puncture. Shock 13:110–116, 2000.
11. Coopersmith CM, Chang KC, Swanson PE, Tinsley KW, Stromberg PE, Buchman TG, Karl IE, Hotchkiss RS: Overexpression of Bcl-2 in the intestinal epithelium improves survival in septic mice. Crit Care Med 30:195–201, 2002.
12. Angus DC, Crowther MA: Unraveling severe sepsis: why did OPTIMIST fail and what's next? JAMA 290:256–258, 2003.
13. Nasraway SA: The problems and challenges of immunotherapy in sepsis. Chest 123(Supp 5):451S–459S, 2003.
14. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, Fisher CJ: Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 344:699–709, 2001.
15. 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. Working Group on “Sepsis-Related Problems” of the European Society of Intensive Care Medicine. Crit Care Med 26:1793–1800, 1998.
16. Zimmerman JE, Wagner DP, Draper EA, Wright L, Alzola C, Knaus WA: Evaluation of Acute Physiology and Chronic Health Evaluation III predictions of hospital mortality in an independent database. Crit Care Med 26:1317–1326, 1998.
17. Remick DG, Bolgos GR, Siddiqui J, Shin J, Nemzek JA: Six at six: interleukin-6 measured 6 h after the initiation of sepsis predicts mortality over 3 days. Shock 17:463–467, 2002.
18. Oberhoffer M, Karzai W, Meier-Hellmann A, Bogel D, Fassbinder J, Reinhart K: Sensitivity and specificity of various markers of inflammation for the prediction of tumor necrosis factor-α and interleukin-6 in patients with sepsis. Crit Care Med 27:1814–1818, 1999.
19. Groeneveld AB, Tacx AN, Bossink AW, van Mierlo GJ, Hack CE: Circulating inflammatory mediators predict shock and mortality in febrile patients with microbial infection. Clin Immunol 106:106–115, 2003.
20. Spittler A, Razenberger M, Kupper H, Kaul M, Hackl W, Boltz-Nitulescu G, Fugger R, Roth E: Relationship between interleukin-6 plasma concentration in patients with sepsis, monocyte phenotype, monocyte phagocytic properties, and cytokine production. Clin Infect Dis 31:1338–1342, 2000.
21. Ebong S, Call D, Nemzek J, Bolgos G, Newcomb D, Remick D: Immunopathologic alterations in murine models of sepsis of increasing severity. Infect Immun 67:6603–6610, 1999.
22. Baker CC, Chaudry IH, Gaines HO, Baue AE: Evaluation of factors affecting mortality rate after sepsis in a murine cecal ligation and puncture model. Surgery 94:331–335, 1983.
23. Eichacker PQ, Parent C, Kalil A, Esposito C, Cui X, Banks SM, Gerstenberger EP, Fitz Y, Danner RL, Natanson C: Risk and the efficacy of anti-inflammatory agents: retrospective and confirmatory studies of sepsis. Am J Respir Crit Care Med 166:1197–1205, 2002.
24. Stewart D, Fulton WB, Wilson C, Monitto CL, Paidas CN, Reeves RH, De Maio A: Genetic contribution to the septic response in a mouse model. Shock 18:342–347, 2002.
25. Riedemann NC, Neff TA, Guo RF, Bernacki KD, Laudes IJ, Sarma JV, Lambris JD, Ward PA: Protective effects of IL-6 blockade in sepsis are linked to reduced C5a receptor expression. J Immunol 170:503–507, 2003.
26. Gennari R, Alexander JW: Anti-interleukin-6 antibody treatment improves survival during gut-derived sepsis in a time-dependent manner by enhancing host defense. Crit Care Med 23:1945–1953, 1995.
27. Dalrymple SA, Slattery R, Aud DM, Krishna M, Lucian LA, Murray R: Interleukin-6 is required for a protective immune response to systemic Escherichia coli
infection. Infect Immun 64:3231–3235, 1996.
28. Anrejko KM, Chen J, Deutschman CD: Intrahepatic STAT-3 activation and acute phase gene expression predict outcome after CLP sepsis in the rat. Am J Physiol 275:G1423–G1429, 1998.
29. Leon LR, White AA, Kluger MJ: Role of IL-6 and TNF in thermoregulation and survival during sepsis in mice. Am J Physiol 275:R269–R277, 1998.
30. Libert C, Vink A, Coulie P, Brouckaert P, Everaerdt B, Van Snick J, Fiers W: Limited involvement of interleukin-6 in the pathogenesis of lethal septic shock as revealed by the effect of monoclonal antibodies against interleukin-6 or its receptor in various murine models. Eur J Immunol 22:2625–2630, 1992.
31. George RL, McGwin G, Metzger J, Chaudry IH, Rue III: LW The association between gender and mortality among trauma patients as modified by age. J Trauma 54:464–471, 2003.
32. Angele MK, Schwacha MG, Ayala A, Chaudry IH: Effect of gender and sex hormones on immune responses following shock. Shock 14:81–90, 2000.
33. Zellweger R, Wichmann MW, Ayala A, Stein S, DeMaso CM, Chaudry IH: Females in proestrus state maintain splenic immune functions and tolerate sepsis better than males. Crit Care Med 25:106–110, 1997.