Prospective evaluation of short-term, high-volume isovolemic hemofiltration on the hemodynamic course and outcome in patients with intractable circulatory failure resulting from septic shock

Patrick M. Honore, MD; Jean Jamez, MD; Michel Wauthier, MD; Patrice A. Lee, PhD;

Thierry Dugernier, MD; Bruno Pirenne, MD; Genevieve Hanique, MD; James R. Matson, MD

Objective: To evaluate the effects of short-term, high-volume hemofiltration (STHVH) on hemodynamic and metabolic status and 28-day survival in patients with refractory septic shock.

Design: Prospective, interventional.

Setting: Intensive care unit (ICU), tertiary institution.

Patients: Twenty patients with intractable cardiocirculatory failure complicating septic shock, who had failed to respond to conventional therapy.

Interventions: STHVH, followed by conventional continuous venovenous hemofiltration. STHVH consisted of a 4-hr period during which 35 L of ultrafiltrate is removed and neutral fluid balance is maintained. Subsequent conventional continuous venovenous hemofiltration continued for at least 4 days.

Measurements and Main Results: Cardiac index, systemic vascular resistance, pulmonary vascular resistance, oxygen delivery, mixed venous oxygen saturation, arterial pH, and lactate were measured serially. Fluid and inotropic support were managed by protocol. Therapeutic endpoints were as follows during STHVH: a) by 2 hrs, a 50% increase in cardiac index; b) by 2 hrs, a25% increase in mixed venous saturation; c) by 4 hrs, an increase in arterial pH to >7.3; d) by 4 hrs, a 50% reduction in epinephrine dose. Patients who attained all four goals (11 of 20) were considered hemodynamic "responders"; patients who did not (9 of 20) were considered hemodynamic "nonresponders." There were no differences in baseline hemodynamic, metabolic, and Acute Physiology and Chronic Health Evaluation and Simplified Acute Physiology Scores between responders and nonresponders. Survival to 28 days was better among responders (9 of 11 patients) than among nonresponders (0 of 9). Factors associated with survival were hemodynamic-metabolic response status, time interval from ICU admission to initiation of STHVH, and body weight.

Conclusions: These data suggest that STHVH may be of major therapeutic value in the treatment of intractable cardiocirculatory failure complicating septic shock. Early initiation of therapy and adequate dose may improve hemodynamic and metabolic responses and 28-day survival. (Crit Care Med 2000; 28:3581-3587)

key words: sepsis; circulatory failure; high-volume hemofiltration; hemodynamics; 28-day survival; hemofiltration dose

Sepsis and septic shock remain the most common cause of death in the intensive care unit (ICU). The mortality rate of septic shock ranges from 40% to 80% and is >80% with refractory septic shock (1). Conventional therapeutic strategy for septic shock has two primary components: hemodynamic stabilization with restoration of blood pressure and optimization of oxygen delivery (2); and eradication of infection by administration of antibiotics and removal of any septic source. Evidence indicates a significant association among sepsis severity, mortality rate and serum concentrations of inflammatory mediators, including tumor necrosis factor and interleukins 1, 6, and 8 (3, 4). Various monoclonal antibodies and specific antagonists have been used as single agents to neutralize per ceived single, pivotal inflammatory mediators; none have succeeded in clinical trials (5, 6).

Clinical observations of hemofiltration in septic patients with acute renal failure indicate that hemofiltration controls not only azotemia and fluid balance but also removes excess inflammatory mediators (7) and improves cardiopulmonary function independent of fluid balance (8, 9). Animal studies have expanded these observations. Gomez et al. (10) studied the effects of low-volume hemofiltration (750-1000 mL/hr) in canine endotoxin shock and observed increased stroke volume and cardiac output. Grootendorst et al. (11) studied high-volume hemofiltration (6 L/hr) in porcine endotoxin shock and showed improved arterial blood pressure and cardiac output. Rogiers et al. (12) also studied high-volume hemofiltration (6 L/hr) in canine endotoxin shock; hemofiltration resulted in increases in blood pressure, cardiac output, and stroke index and improved hepatic arterial blood flow.

We sought to describe the effects of short-term, high-volume hemofiltration on specific hemodynamic and metabolic variables in patients with hypodynamic septic shock, of varied cause, who had failed to respond adequately to conventional fluid and cardiotonic treatment. We chose to perform an initial short-term (4 hrs), very high-volume (35) L of ultra-filtrate removed neutral balance hemofiltration (STHVH), followed by conventional hemofiltration for two reasons.

First, an initial intense period of hemofiltration may provide rapid abatement of excessive inflammatory activity. Second, a short period of high-volume hemofiltration may be more easily and safely managed in the clinical setting than a long period. After the 4-hr period of STHVH, the filter was changed and conventional hemofiltration continued with a new filter.

MATERIALS AND METHODS

Study Population

A prospective, interventional study was conducted in the ICU of the St-Pierre Hospital of Ottignies. This is a community hospital; the ICU has 15 beds and >850 admissions annually. The hospital serves a population of 150,000 and is an important regional trauma center.

Twenty patients with refractory septic shock were admitted to the study between February 1, 1996, and March 1, 1997. Sepsis was diagnosed according to the criteria of Bone (13) with clinical evidence of infection. Our local ethics committee approved this study, and informed consent was obtained from patients or family members. Conventional treatment was performed per protocol. Empirical antibiotics were administered initially and changed as culture sensitivities indicated. Percutaneous or surgical drainage of a septic source was done as indicated.

Volume resuscitation was considered adequate, and no further fluid boluses were administered when infusion of a fluid bolus of 7 mL/kg iv over 20 mins did not increase the cardiac index by at least 20%. If fluid resuscitation had resulted in a cardiac index of <2.5 L/min/m² and a mean systemic blood pressure of <70 mm Hg, then the vasopressor-inotrope protocol of Martin et al. (14) was used to maintain mean arterial pressure >70 mm Hg. The initial drug, dopamine, was titrated between 2.5 and 20 mcg/kg/min. If needed, nor-epinephrine (titrated from 0.2 to 2 mcg/kg/min) , and dobutamine (titrated from 2.5 to 20 mcg/ kg/min) were used to raise the cardiac index to >2.5 L/min/m². Finally, if a mean arterial pressure of >65 mm Hg and a cardiac index of >2.5 L/min/m² were not achieved, epinephrine (titrated from 0.2 to 1 mcg/kg/min) was added. Patients who failed to respond to this protocol were tnen treated with short-term, high-volume hemofiltration. During STHVH, no additional fluid challenges were given. After STHVH, responders required no further fluid challenge whereas nonresponders received additional fluid challenge before increasing vasopressor therapy. Patients were weighed at hospital admission before any volume resuscitation was given.

Hemodynamic and Metabolic Variables

Hemodynamic monitoring was done with a thermodilution pulmonary artery catheter with fiberoptic continuous monitoring of mixed venous oxygen saturation and continuous cardiac output monitoring (Vigilance, Baxter, McGaw Park, IL). A radial or a femoral artery catheter was used to measure blood pressure. Serial measurements of heart rate, mean arterial pressure, mean central venous pressure, mean pulmonary arterial pressure, and pulmonary artery occlusion pressure were taken. The oxygen delivery index and the oxygen consumption index were calculated using standard formulas. pH, PCO2, and PO2 were measured on a clinical blood gas analyzer (Ciba Corning, Halsted, Essex, UK). Lactate was measured in plasma obtained from iced arterial blood samples, using an enzymatic colorimetric method. The anion gap was calculated using the standard formula: Na + K - Cl + HC03.

Hemofiltration Technique

For vascular access, a double coaxial lumen 14-Fr catheter (Medcomp, Harleyville, Pa) was inserted percutaneously in either the right internal jugular or the femoral vein using the Seldinger technique. A conventional hemofiltration circuit (Gambro, Lund, Sweden) with a polysulfone hemofilter (1.6 m² surface area, 35-kD limit; Fresenius Medical Care, Lexington, MA) was used in a peristaltic blood pump. Hemofiltration was initiated with STHVH (4-hr period, 35-L ultrafiltrate volume, neutral balance). Blood flow was ~450 mL/min; transmembrane pressure was maintained between 300 and 500 mm Hg. The anticoagulant was heparin; patient-activated clotting time was adjusted to 60-70 secs. The ultrafiltrate was replaced by bicarbonate-buffered (40 mmol/L) hemofiltration fluid (Clearflex Bieffe Medital, Lugano, Switzerland); a strictly neutral balance was maintained using a digital balance system (Gambro). After STHVH, the patient was treated with conventional hemofiltration using a Prisma pump and a polyacrylonitrile AN 69 hemofilter, exchanging 24-L per day without dialysis using the same replacement fluid.

Elapsed time is expressed in hours or minutes from the beginning of STHVH as a T value: TO is the beginning of STHVH at zero minutes; T4 is 240 mins, the end of STHVH; T24 is 24 hrs after TO. The delay time is defined as the time from ICU admission to initiation of STHVH. This value was recorded for each patient. Ultrafiltrate dose is defined as the volume of ultrafiltrate removed per kilogram (kg) of body weight.

Therapeutic Protocol

Inclusion Criteria. Patients were treated with STHVH if hemodynamic endpoints of the resuscitation protocol were not achieved. Hemodynamic inclusion criteria were as follow mean arterial pressure, <55 mm Hg; cardiac index, <2.5 L/min/m²; maximum dopamine, dobutamine, and norepinephrine doses and epinephrine dose, 0.5 mcg/kg/min for >2 hrs pulmonary artery occlusion pressure, >16 mm Hg; and mixed venous oxygen saturation <55%. Metabolic and respiratory inclusion criteria were as follows: arterial pH, <7.15, serum lactate, >5 mmol/L; PaO2/FiO2 <100 with mechanical ventilation. Renal function was not used to include or exclude palients/

Exclusion criteria. Patients were excluded from the study protocol if they had previous cardiac disease (determined by transesophageal echocardiography), sepsis complicating pancreatitis, burns, cirrhosis, or cancer, an undrained source of surgical sepsis or if they were >80 yrs of age, had been admitted for >36 hrs, or had a life expectancy of <3 months.

Severity of illness and predicted mortality were assessed in all patients using the Acute Physiology and Chronic Health Evaluation (APACHE) II (15) and the Simplified Acute Physiology Score (SAPS) II (16).

Data Analysis

Hemodynamic, metabolic, and epidemiologic differences between responders and nonresponders at baseline were analyzed using the rank test (Mann-Whitney). Differences of serially measured variables between responders and nonresponders were analyzed using analysis of variance for repeated measurements. Values of skewness and kurtosis ranging from 0 to 2.5 assessed the normality of the distributions. The relationship between hemodynamic response status and 28-day survival was analyzed by the Fisher's exact test. Comparison between expected (APACHE II score and SAPS II) and observed mortality was made by the Poisson distribution. A p < .05 was considered significant. Results are reported as mean ± sd.

RESULTS

A patient was considered to be a hemodynamic responder if all of the following four therapeutic goals were achieved. First, after 2 hrs, a >50% increase in cardiac index. Second, after 2 hrs, a 25% increase in mixed venous saturation. Third, after 4 hrs, an increase in arterial pH to >7.3. Fourth, after 4 hrs, a s50% reduction in epinephrine dose. Eleven patients (responders) attained all therapeutic goals; nine patients did not attain all goals (hemodynamic nonresponder).

Before initiation of STHVH (TO hrs), there were no significant differences between the responder and nonresponder groups with respect to age, APACHE II score, predicted risk of death, SAPS II, baseline physiologic variables, or epinephrine requirements. Retrospective assessments were patient body weight at admission to the ICU and ultra-filtrate dose. Responders (66.2 ± 8.4 kg; median, 66 kg) weighed significantly less than nonresponders (82.6 ± 13.4 kg; median, 81 kg; p < .0031). STHVH ultrafiltrate dose was calculated by indexing the 35-L exchange volume to individual patient body weight. Ultrafiltrate dose in responders (0.53 ± 0.07 L/kg) was significantly higher then in nonresponders (0.43 ± 0.07 L/kg; p < .0031).

Hemodynamic Changes. Volume resuscitation before TO for responders (38 ± 8.16 mL/k'g of body weight) was not significantly different from that for nonresponders (41.6 ± 7.3 mL/kg). In the responder group from TO to T2, cardiac index, mean arterial pressure, left ventricular stroke work index, and mixed venous oxygen saturation increased significantly, pulmonary vascular resistance and epinephrine dose decreased significantly, systemic vascular resistance tended to increase to T2, then declined significantly by T4, and the pulmonary artery occlusion pressure did not change. In the nonresponder group, no changes were observed and the epinephrine dose could not be reduced from TO to T4. In the responder group, from T2 to T4, cardiac index, mean arterial pressure, left ventricular stroke work index, and mixed venous oxygen saturation were significantly greater than, and pulmonary vascular resistance and the epinephrine dose were significantly less than, concurrent values in the nonresponder group. In responders by T12, mean arterial pressure increased to >70 mm Hg; by T36, cardiac index increased to >3 L/min/m²; epinephrine and norepinephrine/dobutamine had been weaned and discontinued. Nonresponders showed no changes in these variables, and all nonresponders died by T24.

Metabolic Changes. In the responder group, the oxygen delivery index and the oxygen consumption index increased significantly by T2 and arterial pH increased significantly by T4. In the nonresponder group, no changes were observed. In the responder group, from T2 to T4, the oxygen delivery index and the oxygen con sumption index were significantly greater than the concurrent values in the nonresponder group. In the responder group, at T4, arterial pH was significantly greater than the concurrent value in the nonresponder group. No changes were observed in serum lactate in either group. Anion gap at TO was not different between responders (29.2 ± 3.96) and nonresponders (29.5% ± 3.7%); by T4, the difference was significant (responders, 24.3 ± 3.25; nonresponders, 28.5 ± 2.3; p <.01).

Twenty-Eight-Day Survival. Nine of 11 patients in the responder group survived >28 days, and all nine nonresponders died by T24. APACHE II scores (survivors, 30.6 ± 4.6; nonsurvivors, 31.5 ± 4.1) and SAPS II (survivors, 69.0 ± 7.8; nonsurvivors, 66.0 ± 8) were not different between survivors and nonsurvivors. Predicted mortality for the 20 study patients was 79% (APACHE II and SAPS II), and the observed mortality of 55% was significantly less than predicted (p < .05, Poisson).

Responder status was associated with delay time, body weight, and ultrafiltrate dose. Delay time for responders was 6.5 hrs (range, 3.25-12 hrs) and for nonresponders it was 13.8 hrs (range, 9.6 to 17.5 hrs). This difference was significant with a p < .01. Delay times for the two responders that died were 16.5 and 12 hrs, similar to that observed in the non-responder group. Ultrafiltrate dose, calculated retrospectively, revealed that the larger nonresponders received a significantly smaller dose of ultrafiltration (0.43 ± 0.07 L/kg; median, 0.43) than the smaller responders (0.53 ± 0.07 L/kg; median, 0.53; p < .0031, Mann Whitney test).

DISCUSSION

Sepsis and septic shock are the most common causes of death in the ICU. In the United States, these conditions consume 1% of the gross domestic product in supportive medical care. Proinflammatory cytokines and other inflammatory mediators are essential for an adequate immune response to infection and injury. However, when stimulated by severe injury or infection, inflammatory mediator production may become excessive and cause tissue injury and destruction, leading to vital organ dysfunction and failure. Clinically, this condition is recognized as multiple organ dysfunction syndrome. In addition to proinflammatory cytokines (tumor necrosis factor-a, interleukin IL-l, and IL-6) (17, 18), excess anti-inflammatory cytokines (IL-4, IL-10, and IL-1 receptor antagonist and soluble tumor necrosis factor receptor-I and soluble tumor necrosis factor receptor-II) accumulate in an excessive counter-regulatory response, reducing immune responsiveness and increasing the risk of secondary infection (19). Therapeutic efforts have focused on monoclonal antibodies and specific antagonists designed to block some perceived "key" mediators. Several have been tested as therapy for multiple organ dysfunction syndrome, A none have succeeded (20).

Hemofiltration is reported to improve cardiopulmonary function and possibly survival (21, 22) by removal of inflammatory mediators from the circulation through filtration or adsorption (23). These observations are unproven (24) and require further development. In our institution, we have observed that short periods (<4 hrs) of very high-volume hemofiltration appeared to improve hemodynamic and metabolic status promptly (25). Similar findings have been reported by Gotloib et al. (26). In the current study, we sought to confirm prospectively and to extend these observations.

Patients with advanced hypodynamic shock and a high risk of mortality, who did not responded adequately to conventional therapy, were selected for study. STHVH was used as the intervention for the following reasons. First, these patients had a very high risk of death and STHVH seemed ethically sound because it offered a plausible rescue therapy. Second, the extreme hemodynamic and metabolic abnormalities of these patients increased the likelihood of observing meaningful changes in a small study group. Third, the extreme physiologic abnormalities, theoretically, would be associated with extremely dysregulated inflammatory mediators and the cytokine system. Thus, a short, intense period of hemofiltration may promptly and effectively restore balance to the dysregulated immune system. Finally, the high fluid flux used in STHVH required very close surveillanance. Logistically, this can be more reliably maintained over a short treatment interval rather than a long one. Renal failure was not a criterion for this study because the intervention of STHVH was specific for sepsis-related multiple organ dysfunction synd'rome.

In septic patients with hypodynamic shock, fluid and inotrope resuscitation to achieve "supranormal" values for cardiac index, oxygen delivery, and oxygen consumption is associated with increased survival. Failure to attain these targets is associated with increased mortality (27). This effect may result from the presence of a useful cardiac reserve in survivors, which does not exist in those who do not respond and usually die (28). In our study, all patients failed to respond to a protocol of fluid and inotropic therapy, similar to those used to achieve "supranormal" cardiac index, oxygen delivery, and oxygen consumption (29). This failure to respond suggested a lack of cardiac reserve and a very poor prognosis. In 11 of 20 patients, STHVH significantly increased the cardiac index, oxygen delivery, and oxygen consumption. Nine of these patients survived. These results suggest that STHVH restored cardiac reserve in the responder group (30-32).

The mechanism of restoration or cardiac reserve is unknown but may relate to removal by filtration of myocardial depressant substances. Coraim et al. (9, 33) described increased oxygen delivery and oxygen consumption, reduced intrapulmonary shunt fraction, and inotropic drug dose after 2 hrs of hemofiltration (900-1200 mL/hr ultrafiltrate flow) in patients with postoperative cardiopulmonary failure; survival was not reported in this study. Coraim et al. (9, 33), Lefer (34), Reilly et al. (35), Parrillo et al. (36), and Grootendorst et al. (37) each reported on myocardial depressant substances in the blood of septic patients with a molecular weight of 10-30 kd, e.g., below the filtration cutoff of the hemofilter. Myocardial depressant substances may reduce ventricular compliance, contributing to low cardiac output in hypodynamic septic shock. In the present study, fluid resuscitation was the same for both groups before STHVH. After completing STHVH, nonresponders continued to require fluid support whereas the responders did not. This suggests that enhanced diastolic function was developed in responders. By T4, serum lactate showed no change in either patient group and pH had increased in responders. The delayed resolution of metabolic acidosis indicates that cardiovascular improvements were the result of the removal of toxic substances and not to neutralization of acid. High lactate levels are poorly reflected by pH values and the anion gap (38). In future studies, strong ion differences will be evaluated. The very close temporal association of STHVH with meaningful improvements in cardiovascular function and subsequent improvements in survival supports an important effect of hemofiltration in the removal of cardiotoxic substances and the improvement of outcome (36).

In this study the failure of hemodynamic response to fluid and inotropic resuscitation, high APACHE II score and SAPS II (predicted mortality, 79%), lactate levels of >9 mmol/L (predicted mortality, 80%) (39), and the presence of more than three vital organ failures in each patient (predicted mortality rate, 95%) (40) resulted in a combined predjcted mortaljty rate of >79%_for the entire group of 20 patients. The observed . mortality rate of 55% was significantly less than predicted. Survival was assocjated with hemodynamic response, delay time to initiation of STHVH, and dose of' hemofiltration.

Delay time (interval from ICU admission to TO) was significantly shorter for responders (6.5 hrs; range, 3.25-12 hrs) than nonresponders (13.8 hrs; range, 9.6-17.5 hrs). This is consistent with the therapeutic principal that the earliest intervention provides the best outcome. Two responders (delay times of 16.5 and 12 hrs) died despite improved hemodynamic status, suggesting that even though cardiovascular reserve may be present, delay in treatment may permit lethal injury to develop.

Measures of dose are rarely used in studies of human or animal hemofiltration and initially were not used in this study. In retrospect, patient body weighty and the related variable of ultrafiltration dose were found to be significantly associated with response status. Storck et al. (41) described an association of increased ultrafiltrate volume with increased survival but did not index to body size. Journois (42) indexed ultrafiltrate volume to total blood volume in hemofiltration done in association with rewarming after cardiopulmonary bypass; the ultrafiltrate volume to total blood volume correlated with the increase in mean arterial pressure during rewarming. Hemodialysis is related to hemofiltration. For many years, hemodialysis treatment was not indexed to body weight. Retrospective analysis revealed higher death rates in larger patients (43). Subsequently, adjustment of dialysis prescription to body size has been shown to improve survival (44, 45).

The ultrafiltration rate is the accepted measure of the intensity of hemofiltration. However, nonquantitative descriptors such as "high volume" and "conventional" lack precision, preventing meaningful comparison of published reports. Failure to quantitatively index the ultrafiltration rate to body size and delay time until the start of hemofiltration therapy invites several problems. These include inadequate treatment of larger patients, the risk of missing untoward effects in smaller patients, variability in results, and compromise of comparison and interpretation of studies. In convective hemofiltration, we recommend quantifying treatment with respect to patient body size and ultrafiltrate flow rate, e.g. milliliters of ultrafiltrate per kg of body weight per hour.

SUMMARY

This prospective, interventional study suggests that STHVH may significantly improve survival in hypodynamic septic shock. Delays to initiation of therapy of <6-8 hrs were associated with survival. Retrospective evaluation of the dose of hemofiltration was significantly correlated with survival. In patients with hypodynamic septic shock, initiating dose-adjusted STHVH within 6-8 hrs of diagnosis may improve hemodynamic response and 28-day survival.

ACKNOWLEDGMENTS

We thank all the nursing staff of both intensive care and nephrology departments for major contributions in the daily care of these patients. We also thank N. Michiels and V. Van Meensel for help in the preparation of the manuscript.

REFERENCES

1. Parrillo JE: Pathogenetic mechanisms of septic shock. NEnglJMed 1993; 328:1471-1477

2. Edwards JD, Brown GC, Nightingale P, tt al:

Use of survivors' cardiorespiratory values as therapeutic goals in septic shock. Crit Care Med 1989; 17:1098-1103

3. Calandra T, Baumgartner JD, Grau GE, et al:

Prognostic value of tumor necrosis factor, interleukin 1, interferon-alpha, and interfer-on-gamma in the serum of patients with septic shock. J Infect Dis 1990; 161:982-987

4. Bone RC: The pathogenesis of sepsis. Am Intern Med 1991; 115:457-469

5. Ziegler EJ, Fisher CJ Jr, Sprung CL, et al. Treatment of gram-negative bacteremia and septic shock with HA-1A human monoclonal antibody against endotoxin: A randomized, double-blind placebo controlled trial. NEngI JMed 1991; 324:429-436

6. Natanson C: NIH conference—Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis. Ann Intern Med 1994; 120:771-783

7. Silvester W, Honore PM, Sieffert E, et al. Interleukins 6 and 8, tumor necrosis factor alpha and complement D clearance by poly-acrylonitrile and polysulfone membranes during hemofiltration in critically ill pa-tiente Abstr. Blood Purif 1997; 15:S127

^BelKmo R, Tipping P, Boyce N: Continuous veno-venous hemofiltration with dialysis removes cytokines from the circulation of septic patients. Crit Care Med 1993; 21:522-526

9. Coraim F, Fasol, Stellwag RF, et al: Continuous Arteriovenous Hemofiltration (CAVH). Basel, Karger, 1985, pp 116-124

10. Gomez A, Wang R, Unruh H, et al: Hemofiltration reverses left ventricular dysfunction during sepsis in dogs. Anesthesiology 1990;

73:671-685

11. Grootendorst AF, van Bommel EF, van der Hoven B, et al: High volume hemofiltration improves hemodynamics of endotoxin induced shock in the pig. Int Care Med 1992;

18:235-240

12. Rogiers P, Zhang H, Smail N, et al: High volume hemofiltration improves hemodynamics in experimental endotoxic shock. Abstr. Int Care Med 1996; 22:S396

13. Bone RC: Let's agree on terminology: Definitions of sepsis. Crit Care Med 1991; 19:

973-976

14. Martin C, Saux P, Eon B, et al: Septic shock:A goal directed therapy using volume loading, dobutamine and/or norepinephrine. Acta Anaesthesiol Scand 1990; 34:413-417

15. Knaus WA, Draper EA, Wagner DP, et al:Apache II: A severity of disease classification system. Crit Care Med 1985; 13:818-829

16. Le Gall JR, Lemeshow S, Saulnier F: A new simplified acute physiology score (SAPS II) based on a European/North American multi-center study. JAMA 1993; 270:2957-2963

17. Dinarello CA: The proinflammatory cytokines interleukin-1 and tumor necrosis factor and treatment of the septic shock syndrome. J Infect Dzs 1991; 63:1177-1184

18. Taneira da Silva AM, Koulbach HC, Chuidan FS, et al: Shock and multiple organ dysfunction after self-administration of salmonella endotoxin. N Engi J Med 1993; 328: 1457-1460

19. Bone RC: Immunologic dissonance: A continuing evolution in our understanding of the systemic inflammatory response syndrome (SIRS) and the multiple organ dysfunction syndrome (MODS). Ann Intern Med 1996; 125:680-687

20. Suffredini AF: Current prospects for treatment of clinical sepsis. Crit Care Med 1994; 22:12-21

21. Oudemans-van Straaten HM, Bosnian R, Zandstra DF: Outcome in critically ill patients treated with intermittent high voli venovenous hemofiltration (HV-WH). Continuous Hemofiltration Therapies in I.C. U.—Proceedings. Paris, Didier Joumi 1996, p 29

22. Weksler N, Chomi I, Gunnan G, et al: 1 proved survival with continuous ver venous hemofiltration in nonoliguric sep patients. In: Continuous Hemofiltrati Therapies in the I.C. U.—Proceedings. Pai Didier Joumois,1996, p 39

23. Fujimori A, Naito H, Miyazaki T; Adsorpti of complement, cytokines, and proteins different dialysis membrane materials: Ev. uation by confocal laser scanning fluon cence microscopy. Artif Organs 1998; 2 1014-1017

24. Schetz M, Ferdinande P, Van den Berghe et al: Removal of proinflammatory cytokin with renal replacement therapy: sense i non-sense. Int Care Med 1995; 21:169-17(

25. Honore PM, Jamez J, Wittebole X, et al: Ii fluence of high volume hemofiltration on Q haemodynamic course and outcome of p:

tients with refractory septic shock: Retn spective study of 15 consecutive cases. Abst Blood Pun'f 1997; 15:135-136

26. Gotloib L, Shostak A, Lev A, et al: Treatmer of surgical and non-surgical septic multi organ failure with bicarbonate hemodialysi and seQuential hemofiltration. Int Care Mei 1995; 21:104-111

27. Shoemaker WC, Appel PL, Kram HB: Se quence of physiological patterns in surgica septic shock. Crit Care Med 1993; 21 1876-1889

28. Timmins AC, Hayes M, Yau E, et al: The relationship between cardiac reserve and survival in critically ill patients receiving treatment aimed at achieving supranormal oxygen delivery and consumption. Postgrad Med J 1992; 68;S34-S40

29. Hayes MA, Yau EH, Timmins AC, et al: Response of critically ill patients to treatment aimed at achieving supranormal oxygen delivery and consumption. Chest 1993; 103: 886-895

30. Heering P, Morgera S, Schmitz FJ, et al: Cytokine removal and cardiovascular hemodynamics in septic patients with continuous veno-venous hemofiltration. Int Care Med 1997; 23:288-296

31. Odnopozov VA, Laterre PF, Traystman RJ, et al: Effects of haemocarboperfusion on organ blood flow and survival in porcine endotoxic shock. Crit Care Med 1996; 24:2021-2026

32. Hoffmann JN, Harti WH, Deppisch R, et al: Effects of hemofiltration on hemodynamics and systemic concentrations of anaphylatox-ins and cytokines in human sepsis. Int Care Med 1996; 22:1360-1367

33. Coraim FJ, Coraim HP, Ebermann R, et al: Acute respiratory failure after cardiac surgery: Clinical experience with the application of continuous arteriovenous hemofiltration. Crit Care Med 1986; 14:714-718

34. Lefer AM: Blood-borne humoral factors in the pathophysiology of circulatory shock. CircRes 1973; 32:129-139 Reilly JM, Cunnion RE, Burch-Whitman C, et al: A circulating myocardial depressant substance is associated with cardiac dysfunction and peripheral hypoperfusion (lactic aci-demia) in patients with septic shock. Chest 1989; 95:1072-1080

36. Parrillo JE, Burch C, Shclhamer, et al: A circulating myocardial depressant substance in humans with septic shock. J din Invest 1985; 76:1539-1553

37. Grootendorst AF, van Bommel EF, van Leengoed LA, et al: Infusion of ultrafiltrate from endotoxemic pigs depresses myocardial performance in normal pigs. J Crit Care 1993; 8:161-169

38. Cohen R, Woods HF: Disturbances of acid-base homeostasis. In: Oxford Texbook of Medicine. Vol 9. Weatherall DJ, Ledingham JGG, Warrel DA (Eds). Oxford, Oxford University Press, 1987, pp 164-175

39. Mehta K, Kruse JA, Carison RW: The relationship between anion gap and elevated lac-tate. Crit Care Med 1986; 14:405-414

40. Maher ER, Robinson KN, Scoble JE, et al: Prognosis of critically-ill patients with acute renal failure: Apache II score and other predictive factors. 0 J Med 1989; 72:857-866

41. Storck M, Hart! WH, Zimmerer E, et al: Comparison of pump-driven and spontaneous continuous haemofiltration in postoperative acute renal failure. Lancet 1991; 337: 452-455

42. Joumois D, Pouard P, Greeley WJ, et al: Hemofiltration during cardiopulmonary bypass in pediatric cardiac surgery: Effects on hemostasis, cytokines and complement components. Anesthesiology 1994; 81: 1181-1189

43. Henderson LW, Leypoldt JK, Lysaght MJ, et al: Death on dialysis and the time/flux tradeoff. B/oorf Pur;/'1997; 15:1-14

44. dark WR, Rocco MV, Collins AJ: Quantification of hemodialysis: Analysis of methods and the relevance to patient outcome. Blood Purif 1997; 15:92-111

45. Collins AJ, Ma JZ, Andy Umen A, et al: Urea index and other predictors of hemodialysis patient survival. Am J Kidney Dis 1994; 23: 272-282