Risks of restrictive red blood cell transfusion strategies in patients with cardiovascular disease (CVD): a meta-analysis

AIM

To evaluate the risks of restrictive red blood cell transfusion strategies (haemoglobin 7-8 g dL^-1 ) in patients with and without known cardiovascular disease (CVD).

BACKGROUND

Recent guidelines recommend restrictive strategies for CVD patients hospitalised for non-CVD indications, patients without known CVD and patients hospitalised for CVD corrective procedures.

METHODS/MATERIALS

Database searches were conducted through December 2017 for randomised clinical trials that enrolled patients with and without known CVD, hospitalised either for CVD-corrective procedures or non-cardiac indications, comparing effects of liberal with restrictive strategies on major adverse coronary events (MACE) and death.

RESULTS

In CVD patients not undergoing cardiac interventions, a liberal strategy decreased (P = 0·01) the relative risk (95% CI) (RR) of MACE [0·50 (0·29-0·86)] (I²  = 0%). Among patients without known CVD, the incidence of MACE was lower (1·7 vs 3·9%), and the effect of a liberal strategy on MACE [0·79, (0·39-1·58)] was smaller and non-significant but not different from CVD patients (P = 0·30). Combining all CVD and non-CVD patients, a liberal strategy decreased MACE [0·59, (0·39-0·91); P = 0·02]. Conversely, among studies reporting mortality, a liberal strategy decreased mortality in CVD patients (11·7% vs·13·3%) but increased mortality (19·2% vs 18·0%) in patients without known CVD [interaction P = 0·05; ratio of RR 0·73, (0·53-1·00)]. A liberal strategy also did not benefit patients undergoing cardiac surgery; data were insufficient for percutaneous cardiac procedures.

CONCLUSIONS

In patients hospitalised for non-cardiac indications, liberal transfusion strategies are associated with a decreased risk of MACE in both those with and without known CVD. However, this only provides a survival benefit to CVD patients not admitted for CVD-corrective procedures.

The effects of steroids during sepsis depend on dose and severity of illness: an updated meta-analysis

A previous meta-analysis determined that the effects of steroids during sepsis were dose-dependent; since then, additional trials have been published. The current analysis updates our previous analysis examining the effects of steroids during sepsis. A literature search from 2004 to 2008 identified seven randomized controlled trials in adult patients; these were added to 14 previously identified trials. The effects of steroids on mortality were highly variable among the 21 trials (p <0.001, I(2) = 60%). In trials published before 1989, which involved short courses of high-dose steroids, steroids increased mortality (n = 8, I(2) = 14%, OR of death 1.39 (95% CI 1.04-1.86), p 0.03). In trials published after 1997, which involved longer courses of lower-dose steroids, steroids consistently improved shock reversal (n = 7, I(2) = 0%, OR of shock reversal 1.66 [95% CI 1.25-2.20), p <0.001), but demonstrated a more heterogeneous beneficial effect on mortality (n = 12, I(2) = 25%, OR of death 0.64 (95% CI 0.45-0.93), p 0.02). An inverse linear relationship between severity of illness and the effects of steroids on mortality was identified across all trials (p 0.03) and within the subgroup of trials published after 1997 (p 0.03); steroids were harmful in less severely ill patient populations and beneficial in more severely ill patient populations. There was no effect of response to adrenocorticotrophic hormone (ACTH) stimulation testing concerning the effects of steroids and no increase in steroid-associated adverse events. Low-dose steroids appear to improve mortality rates in patients with septic shock who are at high risk of death; however, additional trials in this subpopulation are necessary to definitively determine the role of low-dose steroids during sepsis.

Increasing the efficacy of anti-inflammatory agents used in the treatment of sepsis

Excessive production of inflammatory mediators during invasive infection plays a key role in the pathogenesis of septic shock. In an attempt to improve survival of patients with this lethal syndrome, agents were developed to selectively inhibit mediators in this inflammatory response. Despite promising preclinical results, several different mediator-specific anti-inflammatory agents failed to demonstrate significant benefit in patients. There was, however, a significant difference in mortality between preclinical and clinical trials. The median control mortality in preclinical trials, performed almost uniformly in highly lethal sepsis models, was 88%. In clinical trials however, the median control mortality rate was much lower, at 41%. A recent meta-regression analysis of these preclinical and clinical trials in combination with prospective confirmatory studies demonstrated that risk of death as assessed by control group mortality rate significantly altered the treatment effect of these agents in both humans and animals. While anti-inflammatory agents were very beneficial in groups with high control mortality rates, they were ineffective or harmful in groups with low control mortality rates. Thus, variation in the risk of death due to sepsis provides a basis for the marked difference in the efficacy of these anti-inflammatory agents in preclinical and clinical trials over the last decade. In contrast to mediator-specific anti-inflammatory agents, glucocorticoids and activated protein C have recently demonstrated significant beneficial effects in individual clinical trials. However, glucocorticoids were studied only in patients with vasopressor-dependent septic shock, which is associated with a high control mortality rate (i.e. 61%) similar to the level at which mediator-specific agents would have been expected to be markedly beneficial. Furthermore, consistent with earlier findings for mediator-specific anti-inflammatory agents, analysis of the activated protein C study also demonstrated a relationship between risk of death and effect of treatment. Developing better methods to define high-risk septic populations for treatment with anti-inflammatory agents will increase the efficacy of this therapeutic approach and minimize its potential for harm.

Acute G-CSF therapy is not protective during lethal E. coli sepsis

We investigated whether decreases in circulating polymorphonuclear neutrophils (PMN) during lethal Escherichia coli (E. coli) sepsis in canines are related to insufficient host granulocyte colony-stimulating factor (G-CSF). Two-year-old purpose-bred beagles had intraperitoneal E. coli-infected or -noninfected fibrin clots surgically placed. By 10 to 12 h following clot, both infected survivors and nonsurvivors had marked increases (P = 0.001) in serum G-CSF levels (mean peak G-CSF ng/ml +/- SE, 1,931 +/- 364 and 2,779 +/- 681, respectively) compared with noninfected controls (134 +/- 79), which decreased at 24 to 48 h. Despite increases in G-CSF, infected clot placement caused delayed (P = 0.06) increases in PMN (mean +/- SE change from baseline in cells x 10(3)/mm(3) at 24 and 48 h) in survivors (+3.9 +/- 3.9 and +13.8 +/- 3.6) compared with noninfected controls (+13.1 +/- 2.8 and +9.1 +/- 2.5). Furthermore, infected nonsurvivors had decreases in PMN (-1.4 +/- 1.0 and -1.1 +/- 2.3, P = 0.006 compared with the other groups). We next investigated whether administration of G-CSF immediately after clot placement and continued for 96 h to produce more rapid and prolonged high levels of G-CSF after infection would alter PMN levels. Although G-CSF caused large increases in PMN compared with control protein from 2 to 48 h following clot in noninfected controls, it caused much smaller increases in infected survivors and decreases in infected nonsurvivors (P = 0.03 for the ordered effect of G-CSF comparing the three groups). Thus insufficient host G-CSF is unlikely the cause of decreased circulating PMN in this canine model of sepsis. Other factors associated with sepsis either alone or in combination with G-CSF itself may reduce increases or cause decreases in circulating PMN.

ICAM-1 and CD11b inhibition worsen outcome in rats with E. coli pneumonia

We investigated whether inhibiting an endothelial adhesion molecule [intracellular adhesion molecule 1 (ICAM-1)] would alter outcome and lung injury in a similar fashion to inhibition of a leukocyte adhesion molecule (integrin CD11b) in a rat model of gram-negative pneumonia. Inhibition of ICAM-1 with monoclonal antibody (MAb) 1A29 (1 mg/kg sc or 0.2 or 2 mg/kg iv, q 12 h x 3) or of CD11b with MAb 1B6 (1 mg/kg sc, q 12 h x 3) were compared against similarly administered placebo proteins in rats challenged with intrabronchial Escherichia coli. After challenge, all animals were treated with antibiotics. ICAM-1 MAb (6 mg/kg, iv, total dose) increased mortality vs. control (P = 0.03). CD11b MAb (3 mg/kg, sc, total dose) did not significantly (P = 0.16) increase mortality rates, but this was not in a range of probability to exclude a harmful effect. All other doses of MAb had no significant effect on survival rates. ICAM-1 and CD11b MAbs had significantly different effects on the time course of lung injury, circulating white cells and lymphocytes, and lung lavage white cells and neutrophils (P = 0.04-0.003). CD11b MAb decreased, whereas ICAM-1 MAb increased these measures compared with control from 6 to 12 h after E. coli. However, from 144 to 168 h after E. coli both MAbs increased these measures compared with control rats but to a greater level with CD11b MAb. Thus both ICAM-1 and CD11b appear to be necessary for survival during E. coli pneumonia. Although these adhesion molecules may participate differently in early lung injury, with CD11b increasing and ICAM-1 decreasing inflammation and injury, both are important for the resolution of later injury. During gram-negative pneumonia the protective roles of ICAM-1 and CD11b may make their therapeutic inhibition difficult.

Neutrophil and endothelial cell interactions in sepsis. The role of adhesion molecules

Although adhesion molecules present on circulating neutrophils and endothelial cells are essential for normal host defense, generalized activation of these molecules has been implicated in the inflammatory tissue injury occurring during sepsis and septic shock. A review of both preclinical and clinical studies suggests, however, that although these molecules mediate tissue injury related to a variety of microbial and host inflammatory mediators, their predominant role during sepsis with infection is a protective one.

G-CSF during Escherichia coli versus Staphylococcus aureus pneumonia in rats has fundamentally different and opposite effects

We investigated if bacteria type alters outcome with prophylactic granulocyte colony stimulating factor (G-CSF) therapy during pneumonia. Rats received G-CSF or placebo daily for 6 d and after the third dose were intrabronchially inoculated with either Escherichia coli or Staphylococcus aureus. Without G-CSF, E. coli and S. aureus produced similar (p = NS) mortality rates (36 versus 38%) and serial changes in mean circulating neutrophil counts (CNC), but differing mean (+/- SE) tumor necrosis factor (TNF) levels (E. coli, 259 +/- 104 versus S. aureus, 51 +/- 17 pg/ml, p = 0.01). G-CSF prior to bacteria increased mean CNC more than six times compared with placebo (p = 0.001). However, with G-CSF in the first 6 h after E. coli, there was a greater than 20-fold decrease in mean (+/- SE) CNC (x 10(3)/ mm3) to below placebo (0.5 +/- 0.1 versus 0.8 +/- 0.1), whereas with G-CSF after S. aureus, there was only a fivefold decrease in mean CNC and CNC were greater than placebo (1.8 +/- 0.2 versus 0.8 +/- 0.1) (E. coli versus S. aureus decrease in CNC with G-CSF, p = 0.001). With E. coli, G-CSF worsened oxygenation and increased bacteremia and mortality, whereas with S. aureus, G-CSF improved oxygenation and decreased bacteremia and mortality (G-CSF therapy, E. coli versus S. aureus, p = 0.03, 0.05, and 0.001, respectively). Thus, during S. aureus pneumonia with low TNF levels, G-CSF increased CNC and bacterial clearance, resulting in less pulmonary injury and decreased death. During E. coli pneumonia with high TNF levels, G-CSF paradoxically decreased CNC, resulting in impaired bacterial clearance and worsened pulmonary injury and death. Bacterial species and the associated inflammatory mediator response can alter outcome with prophylactic G-CSF therapy during pneumonia.

Increasing doses of pentoxifylline as a continuous infusion in canine septic shock

We investigated effects of pentoxifylline during septic shock. Two-year-old (10-12 kg), purpose-bred beagles were infected i.p. with Escherichia coli 0111:B4 (1.2-1.5 x 10(9) colony-forming units per kilogram b.wt.) in a fibrin clot and then immediately treated with one of five doses of pentoxifylline (0.5-20 mg. kg-1. h-1 i.v.) as a 36-h continuous infusion or placebo. All animals received antibiotics and fluid resuscitation. Pentoxifylline levels increased in a dose-dependent manner during (p =.001) and were undetectable 12 h after stopping the infusion. During infusion of pentoxifylline at all doses, there were increases (p =.003), and once the infusion was stopped, there were decreases (p =.049) in endotoxin levels compared with controls. After clot implantation, at all pentoxifylline doses there was a significant increase in tumor necrosis factor levels, compared with controls (p =.025). The relative risk of death was significantly increased with pentoxifylline therapy in a dose-dependent fashion (20 >/= 10 >/= 5.0 >/= 1.0 >/= 0.5 mg. kg-1, p =.008). One hypothesis consistent with these data is that high pentoxifylline levels slowed endotoxin clearance, resulting in high levels of endotoxemia and increased proinflammatory mediator release and death. Pentoxifylline, used as a long-term continuous infusion as is commonly done clinically, can be harmful during Gram-negative septic shock.

Effects of L-NMMA and fluid loading on TNF-induced cardiovascular dysfunction in dogs

We investigated the effects of N(omega)-monomethyl-L-arginine (L-NMMA) and fluid loading on tumor necrosis factor (TNF)-induced cardiovascular dysfunction in awake dogs. L-NMMA (40 mg x kg(-1) given intravenously over a period of 10 min, and followed by dosing at 40 mg x kg(-1) x h(-1) for 6 h) and TNF (20 or 45 microg x kg(-1) given intravenously for 20 min), given alone or in combination, significantly decreased stroke volume, cardiac index, oxygen delivery, and left-ventricular (LV) function plots over a period of 6 h. Of note was that the cardiac-depressant effects of TNF and L-NMMA given together were significantly less than additive. Thus, the combination was beneficial (or significantly less harmful to cardiac performance than expected), possibly because L-NMMA augmented cardiac preload as shown by significant increases in both pulmonary capillary wedge pressure (PCWP) and central venous pressure (CVP). Fluid challenges at 6 h (Ringer's solution at 80 ml x kg(-1) given over a period of 30 min) also significantly increased PCWP and CVP, and abolished the beneficial preload effect of L-NMMA on cardiac performance. Thus, after fluid loading, the cardiac-depressant effects of TNF and L-NMMA given together became equal to the sum of those produced by TNF and L-NMMA given separately. Although L-NMMA significantly decreased serum nitrite/nitrate levels, TNF did not increase these end products of nitric oxide (NO) production relative to controls. Therefore, after preload abnormalities were eliminated with fluid loading, L-NMMA had no beneficial effect on TNF-induced cardiac depression, and TNF did not increase end products of NO production. These findings are not consistent with NO being the mechanism of TNF-induced acute cardiac depression.

Response of the septic vasculature to prolonged vasopressor therapy with N(omega)-monomethyl-L-arginine and epinephrine in canines

OBJECTIVE

To investigate the effect of blocking nitric oxide production on cardiovascular function and survival in canine septic shock treated with or without a conventional vasopressor.

DESIGN

Randomized, controlled trial.

SETTING

An animal research laboratory at the National Institutes of Health.

SUBJECTS

Sixty purpose-bred beagles.

INTERVENTIONS

Fibrin clots containing Escherichia coli were surgically placed into the peritoneal cavity. N(omega)-monomethyl-L-arginine (L-NMMA) 10 mg/kg followed by 0.5, 1.0, or 4.0 mg/kg/hr), epinephrine (1 microg/kg/min), both, or neither were infused for 24 hrs beginning 6 hrs after the onset of infection. All animals received fluid and antibiotic therapy.

MEASUREMENTS AND MAIN RESULTS

Serum nitric oxide metabolites, nitrite and nitrate, increased with infection (p = .024) and decreased with L-NMMA (p = .004, all doses combined). Myocardial nitric oxide synthase activity was ranked as follows: nonsurvivors > survivors > noninfected controls (p < .01). Other tissues examined showed the same pattern. L-NMMA produced sustained increases in systemic vascular resistance index and mean arterial pressure 9 and 24 hrs after the onset of infection (p < or = .04). Left ventricular ejection fraction was depressed by septic shock (p = .01) and further decreased by L-NMMA (p = .02). However, control and L-NMMA cardiac index values were similar (p > .4), perhaps because L-NMMA increased pulmonary artery occlusion pressure (p = .02). From 9 to 24 hrs, epinephrine, in the absence or presence of L-NMMA, blunted recovery of cardiac index (p < .02) and had a diminishing vasopressor effect (p = .05). Neither L-NMMA nor epinephrine, individually or combined, significantly altered survival rates at the doses investigated (p > or = .69).

CONCLUSIONS

The tested doses showed that nitric oxide production was inhibited by L-NMMA in canine septic shock, but mortality and myocardial depression were unaffected. These results suggest that if L-NMMA has a beneficial effect on survival rates in septic shock, it is small.

Cardiopulmonary effects of inhaled nitric oxide in normal dogs and during E. coli pneumonia and sepsis

We investigated the effect of inhaled nitric oxide (NO) at increasing fractional inspired O2 concentrations (FIO2) on hemodynamic and pulmonary function during Escherichia coli pneumonia. Thirty-eight conscious, spontaneously breathing, tracheotomized 2-yr-old beagles had intrabronchial inoculation with either 0.75 or 1.5 x 10(10) colony-forming units/kg of E. coli 0111:B4 (infected) or 0.9% saline (noninfected) in one or four pulmonary lobes. We found that neither the severity nor distribution (lobar vs. diffuse) of bacterial pneumonia altered the effects of NO. However, in infected animals, with increasing FIO2 (0.08, 0.21, 0.50, and 0.85), NO (80 parts/million) progressively increased arterial PO2 [-0.3 +/- 0.6, 3 +/- 1, 13 +/- 4, 10 +/- 9 (mean +/- SE) Torr, respectively] and decreased the mean arterial-alveolar O2 gradient (0.5 +/- 0.3, 4 +/- 2, -8 +/- 7, -10 +/- 9 Torr, respectively). In contrast, in noninfected animals, the effect of NO was significantly different and opposite; NO progressively decreased mean PO2 with increasing FIO2 (2 +/- 1, -5 +/- 3, -2 +/- 3, and -12 +/- 5 Torr, respectively; P < 0.05 compared with infected animals) and increased mean arterial-alveolar O2 gradient (0.3 +/- 0.04, 2 +/- 2, 1 +/- 3, 11 +/- 5 Torr; P < 0.05 compared with infected animals). In normal and infected animals alike, only at FIO2 < or = 0.21 did NO significantly lower mean pulmonary artery pressure, pulmonary artery occlusion pressure, and pulmonary vascular resistance index (all P < 0.01). However, inhaled NO had no significant effect on increases in mean pulmonary artery pressure associated with bacterial pneumonia. Thus, during bacterial pneumonia, inhaled NO had only modest effects on oxygenation dependent on high FIO2 and did not affect sepsis-induced pulmonary hypertension. These data do not support a role for inhaled NO in bacterial pneumonia. Further studies are necessary to determine whether, in combination with ventilatory support, NO may have more pronounced effects.

rG-CSF reduces endotoxemia and improves survival during E. coli pneumonia

We investigated the effects of recombinant granulocyte colony-stimulating factor (rG-CSF) during canine bacterial pneumonia. Beagles with chronic tracheostomies received daily subcutaneous rG-CSF (5 micrograms/kg body wt) or placebo for 14 days, beginning 9 days before intrabronchial inoculation with E. coli. Animals received antibiotics and fluid support; a subset received humidified oxygen (fractional inspired O2 0.40). Compared with controls, rG-CSF increased circulating neutrophil counts (57.4 vs. 11.0 x 10(3)/mm3, day 1 after infection; P = 0.0001), decreased plasma endotoxin (7.5 vs. 1.1 EU/ml at 8 h; P < 0.01) and serum tumor necrosis factor-alpha (3,402 vs. 729 pg/ml at 2 h; P = 0.01) levels, and prolonged survival (relative risk of death = 0.45, 95% confidence interval 0.21-0.97; P = 0.038). Also, rG-CSF attenuated sepsis-associated myocardial dysfunction (P < 0.001). rG-CSF had no effect on pulmonary function or on blood and lung bacteria counts (all P = not significant). Other animals challenged with endotoxin (4 mg/kg i.v.) after similar treatment with rG-CSF had lower serum endotoxin levels (7.62 vs. 5.81 log EU/ml at 6 h; P < 0.01) and less cardiovascular dysfunction (P < 0.05 to < 0.002) but similar tumor necrosis factor-alpha levels (P = not significant) compared with controls. Thus prophylactic rG-CSF sufficient to increase circulating neutrophils during bacterial pneumonia may improve cardiovascular function and survival by mechanisms that in part enhance the clearance of bacterial toxins but do not improve lung function.

Tyrphostin AG 556 improves survival and reduces multiorgan failure in canine Escherichia coli peritonitis

Tyrosine kinase-dependent cell signaling is postulated to be a pivotal control point in inflammatory responses initiated by bacterial products and TNF. Using a canine model of gram-negative septic shock, we investigated the effect of tyrosine kinase inhibitors (tyrphostins) on survival. Animals were infected intraperitoneally with Escherichia coli 0111: B4, and then, in a randomized, blinded fashion, were treated immediately with one of two tyrphostins, AG 556 (n = 40) or AG 126 (n = 10), or with control (n = 50), and followed for 28 d or until death. All animals received supplemental oxygen, fluids, and antibiotics. Tyrphostin AG 556 improved survival times when compared to controls (P = 0.05). During the first 48 h after infection, AG 556 also improved mean arterial pressure, left ventricular ejection fraction, cardiac output, oxygen delivery, and alveolar-arterial oxygen gradient compared to controls (all P < or = 0.05). These improvements in organ injury were significantly predictive of survival. Treatment with AG 556 had no effect on clearance of endotoxin or bacteria from the blood (both P = NS); however, AG 556 did significantly lower serum TNF levels (P = 0.03). These data are consistent with the conclusion that AG 556 prevented cytokine-induced multiorgan failure and death during septic shock by inhibiting cell-signaling pathways without impairing host defenses as determined by clearance of bacteria and endotoxin.

Serial measures of total body oxygen consumption in an awake canine model of septic shock

We examined serial changes in total body oxygen consumption (Vo2) in a permanently tracheotomized canine sepsis model. On Day 0, beagles had an Escherichia coli-infected (septic) or sterile (control) clot surgically placed in the peritoneum. During the 21-d study, 10 of the 16 septic animals and none of the six control animals died (p = 0.02). After clot placement septic versus control animals had decreased mean arterial blood pressure (mm Hg; Day 1: 106 versus 128, p = 0.055; Day 2: 95 versus 125, p = 0.004, respectively) and left ventricular ejection fraction (Day 1: 0.44 versus 0.69, p = 0.0006; Day 2: 0.33 versus 0.57, p = 0.0001, respectively). Despite significant lethality and cardiovascular dysfunction, in the septic group on Days 1 and 2, septic versus control animals had no significant differences in mean metabolic cart measured (Vo2DIR, ml/kg/min; Day 1: 11.9 versus 12.4, p = 0.81; Day 2: 14.2 versus 13.5, p = 0.72, respectively) and intravascular catheter calculated (Vo2INDIR, ml/kg/min; Day 1: 11.2 versus 11.2, p = 0.99; Day 2: 12.8 versus 15.4, p = 0.49, respectively). On Day 1 in septic and control animals, volume infusion produced increases (p < 0.001) in oxygen delivery (Do2). In septic and control animals these changes in Do2 were similar and were associated with similar increases in Vo2DIR (p = 0.001), and Vo2INDIR (p = 0.001). In fact, at all time points studied (baseline, Day 1, 2, and 21), both before and after volume infusion, levels of Do2, Vo2DIR, and Vo2INDIR did not differ between septic and control animals, nor did they differ between septic survivors and nonsurvivors. Because levels of Vo2DIR and Vo2INDIR were similar in both groups, we pooled data from septic and control animals. Throughout the study, Vo2 showed a moderate association with Vo2INDIR (r = 0.55, p = 0.003), but mean Vo2DIR was lower at baseline (p = 0.001) and on Day 21 (p = 0.07) and greater on Day 2 (p < 0.01). In summary, our techniques, which detected small changes in both Vo2DIR and Vo2INDIR occurring with volume infusion, did not demonstrate differences in these parameters comparing control and septic animals. These results in euvolemic septic animals suggest that total body Vo2 may not reflect pathogenetic mechanisms during sepsis and septic shock. Furthermore, these results suggest that although the level of total body Vo2 may reflect the effects of therapeutic interventions such as volume loading, it should not itself serve as a therapeutic target.

Controlled trials of rG-CSF and CD11b-directed MAb during hyperoxia and E. coli pneumonia in rats

We studied the effects of inhibiting and augmenting neutrophil function by using an immunocompetent rat model of infectious and hyperoxic lung injury. After intrabronchial Escherichia coli challenge at all fractional inspired O2 (FIO2) values studied (FIO2 = 0.21, 0.60, and 0.95) and after lethal O2 exposure alone (FIO2 = 0.90), lung injury, as measured by histological and physiological changes, was reduced by a CD11b/CD18-directed monoclonal antibody (MAb 1B6, P < 0.05 vs. controls) but was increased by recombinant granulocyte colony-stimulating factor (rG-CSF; P < 0.05 vs. control; MAb 1B6 vs. rG-CSF, P < 0.004). Pulmonary neutrophil counts were reduced by MAb 1B6 (P < 0.04) and increased by rG-CSF (P < 0.0004) compared with control animals. However, despite antibiotics, MAb 1B6 and rG-CSF both significantly increased the relative risk of death, independent of O2 concentration, during E. coli pneumonia (1.74 [symbol: see text] 1.20 and 2.39 [symbol: see text] 1.19, respectively, each P < 0.01). During lethal hyperoxia, MAb 1B6 increased the relative risk of death (1.76 [symbol: see text] 1.28, P < 0.16), whereas rG-CSF had no effect on survival (0.97 [symbol: see text] 1.28, P = 0.89). Thus inhibition of neutrophil function attenuated and enhancement worsened lung injury in response to infectious and hyperoxic challenges, supporting a pathophysiological role of the neutrophil in these processes. However, it is problematic that MAb 1B6 therapy, despite preventing lung damage, ultimately worsened host defenses and survival. Furthermore, rG-CSF also adversely affected survival during infectious lung injury, demonstrating the inherent risks of inhibiting or augmenting neutrophil function in an immunocompetent host during infection.

Cardiopulmonary effects of granulocyte colony-stimulating factor in a canine model of bacterial sepsis

We investigated the effects of recombinant granulocyte colony-stimulating factor (G-CSF) in a canine model of septic shock. Awake 2-yr-old beagles were studied before and after intraperitoneal placement of an Escherichia coli-infected clot. Nine days before and until 3 days after clot placement, animals received daily high-dose (G-CSF (5 microgram/kg body wt; n = 17), low-dose G-CSF (0.1 microgram/kg body wt; n = 17), or a control protein (5 micrograms/kg body wt; n = 20). Survival rate was greater (P < 0.04, Wilcoxon test) in the high-dose G-CSF group (14/17) than in the low-dose G-CSF (10/17) and control (12/20) groups. High-dose G-CSF improved cardiovascular function, as evidenced by increased left ventricular ejection fraction (day 1 after clot; P < 0.001) and mean arterial pressure (day 2; P < 0.02) compared with low-dose G-CSF and control groups. High-dose G-CSF increased (P < 0.001) mean peripheral neutrophils before (-3 days) and after (2 h to 4 days) clot and produced a more rapid (P < 0.001) rise (day 2) and fall (day 4) in mean alveolar neutrophil numbers compared with the low-dose G-CSF and control groups. High-dose G-CSF decreased mean serum endotoxin (2-8 h; P < 0.002) and tumor necrosis factor (2 h; P < 0.02) levels and lowered blood bacteria counts (2-6 h; P < 0.04) compared with the low-dose G-CSF and control groups. Thus, in this canine model, G-CSF sufficient to increase peripheral neutrophils before and during peritonitis and septic shock enhances host defense, reduces cytokine (tumor necrosis factor) levels, and improves cardiovascular function and survival.

Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis

PURPOSE

To review selected new therapies for septic shock designed to inhibit bacterial toxins or endogenous mediators of inflammation.

DATA SOURCES

Scientific journals, scientific meeting proceedings, and Food and Drug Administration advisory committee proceedings.

STUDY SELECTION AND EXTRACTION

Preclinical and clinical data from trials using core-directed antiendotoxin antibodies and anticytokine therapies for sepsis and studies in animal models of sepsis from our laboratory.

RESULTS OF DATA SYNTHESIS

Ten clinical trials using core-directed antiendotoxin antibodies produced inconsistent results and did not conclusively establish the safety or benefit of this approach. Both anti-interleukin-1 and anti-tumor necrosis factor (TNF) therapies have been beneficial in some animal models of sepsis but did not clearly improve survival in initial human trials, and one anti-TNF therapy actually produced harm. Neutrophils, another target for therapeutic intervention, protect the host from infection but may also contribute to the development of tissue injury during sepsis. In a canine model of septic shock, granulocyte colony-stimulating factor increased the number of circulating neutrophils and improved survival, but an anti-integrin (CD11/18) antibody that inhibits neutrophil function worsened outcome. Nitric oxide, a vasodilator produced by the host, causes hypotension during septic shock but may also protect the endothelium and maintain organ blood flow. In dogs challenged with endotoxin, the inhibition of nitric oxide production decreased cardiac index and did not improve survival.

CONCLUSIONS

No new therapy for sepsis has shown clinical efficacy. Perhaps more accurate clinical and laboratory predictors are needed to identify patients who may benefit from a given treatment strategy. On the other hand, the therapeutic premises may be flawed. Targeting a single microbial toxin such as endotoxin may not represent a viable strategy for treating a complex inflammatory response to diverse gram-negative bacteria. Similarly, the strategy of inhibiting the host inflammatory response may not be beneficial because immune cells and cytokines play both pathogenic and protective roles. Finally, our scientific knowledge of the complex timing of mediator release and balance during sepsis may be insufficient to develop successful therapeutic interventions for this syndrome.

The third component of complement protects against Escherichia coli endotoxin-induced shock and multiple organ failure

We investigated whether the third component of complement (C3) is involved in the pathophysiology of endotoxic shock, and if it is involved, whether it plays a protective role or whether it mediates shock and multiple organ failure. In a prospective, controlled investigation, six Brittany spaniels that were homozygous for a genetically determined deficiency of C3 (C3 deficient, < 0.003% of normal serum C3 levels) and six heterozygous littermates (controls, approximately 50% of mean normal serum C3 level) were given 2 mg/kg of reconstituted Escherichia coli 026:B6 acetone powder as a source of endotoxin, intravenously. All animals were given similar fluid and prophylactic antibiotic therapy, and had serial hemodynamic variables obtained. After E. coli endotoxin infusion, C3-deficient animals had higher peak levels of endotoxin and less of a rise in temperature than controls (P < 0.05). During the first 4 h after E. coli endotoxin infusion, C3-deficient animals had significantly greater decreases in mean central venous pressure and mean pulmonary artery pressure than controls (P < 0.02). During the first 48 h after E. coli endotoxin infusion, C3-deficient animals had significantly greater decreases in mean arterial pH, left ventricular ejection fraction, and mean pulmonary capillary wedge pressure, and greater increases in mean arterial lactate, arterial-alveolar O2 gradient, and transaminases (aspartate aminotransferase and alanine aminotransferase) than controls, (all P < 0.05). After E. coli endotoxin infusion, C3-deficient animals compared to controls had significantly less of a decrease in mean C5 levels (P < 0.01), but similar (P = NS) increases in circulating tumor necrosis factor levels, bronchoalveolar lavage neutrophils, and protein, and similar (P = NS) decreases in blood leukocytes and platelets. Two of six C3-deficient animals and two of six controls died. In summary, after intravenous infusion of E. coli endotoxin, canines with C3 deficiency have decreased endotoxin clearance and worse E. coli endotoxin-induced shock and organ damage. Thus, the third component of the complement system plays a beneficial role in the host defense against E. coli endotoxic shock.

Leukocyte CD18 monoclonal antibody worsens endotoxemia and cardiovascular injury in canines with septic shock

We investigated the effects of a murine monoclonal antibody directed against the canine leukocyte CD11/18 adhesion complex (MAb R15.7) in a canine model of septic shock. Awake 2-yr-old purpose-bred beagles were studied 7 days before and 1, 2, 4, and 10 days after intraperitoneal placement of an Escherichia coli-infected fibrin clot. Starting 12 h before clot placement, animals received 0.5-1 mg/kg iv every 12 h (4 doses total) of either MAb R15.7 (MAb group, n = 8) or, as controls, murine serum protein (n = 8). After infected clot placement, all animals received antibiotic (ceftriaxne, 100 mg.kg-1.day-1 for 4 days). Two of eight control animals and four of eight MAb animals died (P = 0.4). During the first 8 h after clot placement, MAb animals, compared with control animals, had greater (P < 0.06) increases in serum endotoxin levels and higher (P < 0.05) neutrophil counts. Day 1 after clot placement, MAb animals, compared with control animals, had decreased (P < 0.05) central venous pressure and arterial pH and increased (P < 0.05) arterial lactate. Day 2 after clot placement, MAb animals, compared with control animals, had decreased (P < 0.05) cardiac index and mean arterial pressure. In summary, MAb R15.7, although associated with increased neutrophil counts, worsened serum endotoxemia, acidosis, and cardiovascular function in this canine model of septic shock. These data suggest that in septic shock, antibody directed against this leukocyte membrane protein complex may be harmful, possibly via impairment of normal leukocyte function.

Therapeutic trial of lipid X in a canine model of septic shock

Three groups of dogs were given lipid X (0, 1, or 10 mg/kg) every 8 h for for seven doses, starting simultaneously with the intraperitoneal placement of Escherichia coli-containing fibrin clots. All animals developed bacteremia, hypotension, and a pattern of decreased left ventricular ejection fraction characteristic of septic shock (P = .01). Survival rates and survival times were not significantly different between treatment groups (P > .2). In a similar experiment, higher doses of lipid X resulted in a significantly decreased survival time compared with concurrent controls (P = .04). Animals receiving lipid X did not differ from controls in serial determinations of temperature, hemodynamic measurements, or laboratory parameters (except serum total protein). Although lipid X has antiendotoxin effects, no benefit could be demonstrated in this antibiotic-treated, gram-negative bacillary-infected model of septic shock. These data do not support a therapeutic role for lipid X in the treatment of gram-negative sepsis.

Acute hemorrhage in dogs: construction and validation of models to quantify blood loss

We examined the ability of commonly used clinical parameters to quantify acute hemorrhage in dogs. Eight animals were bled 40 ml/kg body wt over 100 min. Ten hemodynamic and 20 blood laboratory parameters were obtained every 10 min to construct, with use of linear regression analysis, models that quantify blood loss. During model construction, the best indicator of quantity of hemorrhage was arterial base deficit [ABD; coefficient of variation (CV) 35%]. This model was more accurate (P < 0.05) than 27 others (CV range 43 to 63%) and similar to systolic (CV 40%) and mean (CV 40%) arterial pressures. In validation studies in 10 additional animals, our best models based on ABD and systolic and mean arterial pressures each unexpectedly showed a significant (P < 0.05) decrease in accuracy (CV 86, 57, and 60%, respectively) attributable to large baseline (before hemorrhage) variability among animals. To eliminate this variability, models based on changes from baseline measurements were investigated. The best predictor of change in blood volume was change in ABD (CV 27%). This model was significantly (P < 0.05) more accurate than any of 27 others (CV range 36 to 65%) and similar to change in venous base deficit and venous pH (each CV 31%). When validated, acid-base models such as ABD, venous pH, and arterial bicarbonate were the best predictors of volume change (CV range 28 to 40%). With the use of multivariate analysis, pairwise combinations of single parameter models (n = 465) improved prediction errors only minimally. In summary, most commonly used hemodynamic and blood indexes could not be validated as accurate measurements in quantifying hemorrhage. In contrast, changes in acid-base parameters were validated as moderately accurate predictors of blood volume changes and therefore may have utility in the assessment of patients with ongoing hemorrhage.

Tumor necrosis factor challenges in canines: patterns of cardiovascular dysfunction

Three groups of conscious canines were given different intravenous doses of human recombinant tumor necrosis factor (TNF) over 1 h, and the resulting cardiovascular abnormalities were examined for 10 days. As TNF dose increased [0 (controls), 30, 60, and 120 micrograms/kg body wt], the number of deaths increased (P less than 0.025; 0 of 6, 1 of 8, 4 of 8, number of deaths in each group, respectively). In all three groups receiving TNF, the mean left ventricular ejection fraction (LVEF) at 2 h after infusion decreased (P less than 0.003) compared with controls. The group receiving the highest dose of TNF (120 micrograms/kg body wt) had the greatest decrease (P less than 0.05) in LVEF from 0 to 2 h. At 8 h, all three groups receiving TNF had similar LVEF. In these three groups, other multiple measures of LV function at 8 h showed significant and similar decreases in cardiac contractility compared with controls. From 24 to 240 h, however, the time required for cardiac performance (LVEF) to return to normal was dose dependent (30 less than 60 less than 120 micrograms/kg body wt; P less than 0.05). Canines receiving the lowest dose of TNF had near normal cardiac function (LVEF) at 24 h, whereas canines receiving the highest dose had persistent cardiac abnormalities at 240 h. Thus, at 8 h, the severity of cardiac dysfunction is independent of TNF dose, but the rate of onset and the duration of cardiac abnormality are markedly dependent of dose.(ABSTRACT TRUNCATED AT 250 WORDS)

Leukocyte CD11b/18 antigen-directed monoclonal antibody improves early survival and decreases hypoxemia in dogs challenged with tumor necrosis factor

This study examined the effect of monoclonal antibody (MAb) directed against leukocyte CD11b/18 glycoprotein complex (904MAb) on cardiopulmonary injury induced by tumor necrosis factor (TNF), and death. Eighteen 2-yr-old, purpose-bred beagles with chronic tracheostomies were challenged with TNF (60 micrograms/kg of body weight) intravenously. Nine of 18 animals were treated with 0.5 to 1.0 mg/kg of body weight 904MAb intravenously 45 min before and 12, 36, and 48 h after TNF infusion. Serial femoral and pulmonary arterial catheter hemodynamics, blood gas analysis, and radionuclide cineangiographic left ventricular ejection fractions (EF) were done before and after a fluid challenge. Serial bronchoalveolar lavages (BAL) with cell and protein analysis also were performed using the chronic tracheostomies. Compared with animals given TNF alone, animals treated with 904MAb did not differ in overall survival (TNF alone, 2/9; 904MAb, 3/9); however, the group of animals treated with 904MAb had significantly (p less than 0.01) fewer deaths within the first 30 h of TNF challenge. At 4 h after TNF challenge, all animals had significantly (p less than 0.05) reduced PaO2 after fluid challenge; however, animals given 904MAb (compared with animals given TNF alone) had significantly (p less than 0.05) smaller reductions in PaO2. Throughout the study, animals given 904MAb before TNF or TNF alone had similar changes in cardiac index, mean arterial pressure, EF, and BAL protein and neutrophil concentration. Thus, MAb directed against the leukocyte CD11b/18 glycoprotein complex prolonged survival and reduced the hypoxemia occurring after TNF challenge, but this antibody did not improve overall survival or cardiopulmonary function.(ABSTRACT TRUNCATED AT 250 WORDS)

TNF but not IL-1 in dogs causes lethal lung injury and multiple organ dysfunction similar to human sepsis

We compared the early and late pulmonary effects of human recombinant tumor necrosis factor (TNF) and interleukin 1 (IL-1) challenges in awake dogs with chronic tracheostomies. Serial blood gas analysis, bronchoalveolar lavage (BAL) with cell and protein analysis, intravascular catheter hemodynamics, and radionuclide left ventricular ejection fractions (LVEF) were determined before and after infusion of TNF (60 micrograms/kg body wt, n = 8), IL-1 (1,000 micrograms/kg body wt, n = 6), or heat-inactivated IL-1 (n = 6, controls). Controls given heat-inactivated IL-1 had no changes (P = NS) in any pulmonary parameter throughout the study. Animals given IL-1 had a transient increase (P less than 0.05) in BAL neutrophil concentration 1 day after infusion but no other changes (P = NS) in pulmonary function throughout the study. Animals given TNF had early (0-4 h) decreases (P less than 0.05) in arterial PO2, increases (P less than 0.05) in physiological shunt fraction and alveolar-to-arterial PO2 gradient, and a high mortality rate (50%). In TNF animals, volume challenges at 4 h were associated (P less than 0.05) with death and noncardiogenic pulmonary edema. In TNF survivors, hypoxemia persisted for 2-3 days and was associated with increases (P less than 0.05) in alveolar protein and neutrophil concentration on days 1 and 3, respectively, which in survivors returned to near normal over 6-21 days. Animals challenged with TNF and not IL-1 had reversible depression of LVEF similar in time course to abnormalities in arterial PO2. In this study, TNF (but not IL-1) challenges were lethal and produced acute pulmonary dysfunction sustained over days (reversible in survivors) that was similar to that seen in human septic shock. The ability of TNF to induce pulmonary injury similar to bacterial shock suggests that TNF is a key mediator of sepsis-induced lung injury. Furthermore, because TNF challenge induced both sustained pulmonary and cardiac injury, TNF may be a common pathway for the multiple organ dysfunction that occurs during septic shock.

Factors that determine the hemodynamic response to inhalation anesthetics

The hemodynamic response to inhalation anesthesia is influenced by three factors: 1) the specific drug, 2) the dose, and 3) individual characteristics of the subject. To investigate the importance of these factors on the cardiovascular response, we administered five doses [0, 0.5, 1.0, 1.5, and 2.0 minimum alveolar concentration (MAC)] of enflurane, halothane, and isoflurane to each of six dogs. Twelve hemodynamic variables were measured. For all variables, a change in the dose of each drug produced a consistent effect in each dog. Increases in dose resulted in significant decreases in seven variables [left ventricular ejection fraction, cardiac index (CI), stroke volume index (SVI), mean arterial pressure (MAP), mean pulmonary arterial pressure (MPAP), left ventricular stroke work index (LVSWI), and heart rate (HR)] and a significant increase in one variable [central venous pressure (CVP)]. In contrast, the response of individual dogs to different drugs was not consistent. For seven variables [MAP, MPAP, LVSWI, CVP, pulmonary capillary wedge pressure (PCWP), end-diastolic volume index (EDVI), and end-systolic volume index (ESVI)], a significant difference in the responses of a dog to two drugs was greater than zero, whereas a significant difference in the response of at least one other dog to the same two drugs was less than zero (discordant dog-drug interactions). Thus, in contrast to the consistency of the cardiovascular response to changes in dose, the hemodynamic response to different drugs was inconsistent among dogs. We also studied the effect of fluid challenge on hemodynamic response at 1.5 or 2.0 MAC of the three drugs given to each dog.(ABSTRACT TRUNCATED AT 250 WORDS)

The effects of heterologous platelet transfusion on pulmonary function during ARDS

Pulmonary platelet sequestration has been implicated in the pathophysiology of ARDS. Since HPTs are frequently given therapeutically to patients with ARDS, a serious concern clinically has been the possibility that platelet transfusions might worsen pulmonary function in such patients. To prospectively evaluate this issue, we assessed the effects of HPT on two parameters of pulmonary function: (1) Qs/Qt; and (2) PVR. Measurements were made immediately before and at 1 hour and 4 hours after HPT in 16 patients during 18 separate episodes of ARDS and in 11 control patients. The ARDS had been present for at least 48 hours in all patients at the time of study. Prior to HPT, patients with ARDS had significantly worse pulmonary function compared to control subjects. Although a small number of patients with ARDS had large changes in PVR and Qs/Qt after HPT, no group showed significant changes in any pulmonary parameter following transfusion. Neither individual changes in the platelet count after HPT nor the duration of ARDS before HPT was associated with changes in pulmonary function after platelet administration. We concluded that therapeutically required HPT did not significantly worsen pulmonary function at up to four hours after transfusion in this group of patients with well-defined and fully evolved (greater than 48-hour duration) ARDS. The reason for this absence of effect was not known, but the pathogenesis of ARDS might only be susceptible to alteration during the early developmental stage.

Radiographic measurement of total lung capacity in acute asthma

The thoracic cage appears to be large during attacks of asthma. Lung volume measurements by body plethysmography and helium dilution have suggested that total lung capacity may be increased during an acute attack of asthma, but doubt has been cast on the accuracy of these measurements in the presence of airflow obstruction. The change in total lung capacity has therefore been investigated during and after an acute attack of severe asthma in 32 patients by a radiographic technique. There was a small decrease (0.29 l) in mean total lung capacity between admission and follow up, though a quarter of the subjects showed a slight increase. There was no correlation between change in total lung capacity and change in expiratory flow rates, arterial carbon dioxide tension on admission, body mass index, and length of stay in hospital. Our findings agree with previous reports of a decrease in total lung capacity with improving airway obstruction, but the changes were small and inconsistent.

Endotoxin and tumor necrosis factor challenges in dogs simulate the cardiovascular profile of human septic shock

Survivors of both human and animal bacterial shock develop a characteristic pattern of progressive changes in cardiovascular function over a period of 7-10 d. In this present study, we examined whether endotoxin (a product of Gram-negative bacteria) or TNF (a cytokine released from macrophages) could reproduce the same complex cardiovascular changes observed in septic shock over a period of 7-10 d. To test this hypothesis, we implanted a thrombin-fibrin clot containing purified endotoxin from E. coli into the peritoneal cavity of eight dogs, and infused TNF into eight different dogs. Over the next 10 d, serial simultaneous heart scans and thermodilution cardiac outputs were performed in these awake nonsedated animals. By day 2 after challenge with either endotoxin or TNF, animals developed a decrease (p less than 0.05) in both mean arterial pressure and left ventricular ejection fraction. With fluid resuscitation, animals manifested left ventricular dilatation (increased [p less than 0.05] end diastolic volume index), increased or normal cardiac index, and decreased or normal systemic vascular resistance index. In surviving animals, these changes returned to normal with 7-10 d. The time course of these changes was concordant (p less than 0.05) with that previously described in a canine model of septic shock using viable bacteria. During the 10-d study, control animals receiving sterile clots or heat-inactivated TNF had not significant changes in hemodynamics. The results from this canine model demonstrate that either endotoxin or TNF alone can produce many of the same hemodynamic abnormalities seen in human septic shock and in a canine septic shock model induced by live bacteria. These findings support the hypothesis that the action of endogenous mediators (TNF) responding to bacterial products (endotoxin) is the common pathway that produces the serial cardiovascular changes found in septic shock.

Respiratory muscle dysfunction in hereditary motor sensory neuropathy, type I

Pulmonary function tests were performed on ten patients who were shown to have hereditary motor sensory neuropathy, type I. Mean values for spirometry, static lung volumes, and diffusion capacity were all greater than 80% of the predicted normal values for the group. In contrast, both inspiratory and expiratory muscle testing showed substantial reductions in function for the group. These abnormalities have not previously been reported, and they may be important in the management of these patients and in patients with other neuromuscular diseases.

Serial pulmonary function testing in patients with systemic lupus erythematosus

Previous studies have documented the pulmonary function abnormalities associated with systemic lupus erythematosus (SLE). There are very few data, however, regarding the progression of such changes. To study this question, we evaluated the pulmonary function of a group of 25 patients with SLE from two to seven years after a set of pulmonary function tests had been performed as part of their overall initial assessment. Reductions in diffusing capacity, FVC, and total lung capacity did not change significantly for the group over the period of our study. The mean FEF25-75%, which was initially low, and the mean FEV1/FVC ratio, which was initially normal, both decreased significantly. The observed abnormalities in airway function were not related to smoking history. Other aspects of lupus activity, as measured by serum creatinine levels and clinical activity, did not appear related to progression of lung disease.

Acute cardiogenic pulmonary edema: clinical and noninvasive evaluation

Left ventricular echocardiograms performed within ninety-six hours of admission were prospectively correlated with the clinical course in 87 consecutive patients admitted with acute pulmonary edema. Patients were stratified into four groups based on their two-dimensional echocardiogram: hyperdynamic, normal, mildly reduced, and severely reduced. Echocardiographic estimates of left ventricular function were compared with their ejection fraction measured by the gated radioisotope technique. The authors found that 48% of the patients were either normal or hyperdynamic (38% and 10% respectively). Patients in these two groups had a greater incidence of left ventricular hypertrophy (wall thickness greater than 13 mm) (66% vs 39%, p less than .05), hypertension on admission (BP greater than 160/100) (66% vs 41%, p = .05), and smaller end-diastolic dimension (p less than .05) than those with decreased left ventricular function. The authors conclude that echocardiography is a good screening test of left ventricular function in patients presenting with pulmonary edema. Patients with normal or increased left ventricular systolic function should be evaluated for correctable or treatable causes of acute pulmonary edema.

Methacholine bronchial reactivity testing in patients with chronic congestive heart failure

To assess the role of bronchial hyperresponsiveness in chronic congestive heart failure (CHF), we performed standard methacholine bronchial provocation testing on nine patients with chronic CHF in New York Heart Association functional class 4. Mean pulmonary capillary wedge pressure and mean stroke volume index for the group at the time of study were 27 +/- 3 mm Hg and 27.9 +/- 8.9 ml/m2, respectively. Mean MMEFR was 71 percent of normal and mean delta to FEV1% was 76.3 percent. Only one patient, who was a 50-pack-year smoker, had a greater than 20 percent decrease in FEV1 with methacholine challenge, and this was at the highest concentration (25 mg/ml). We conclude that, in stable CHF, bronchial reactivity as assessed by standard methacholine challenge is not increased.

Characterization of the anti-stem cell activity of anti-mouse brain serum

The anti-stem cell activity of a high-titer rabbit anti-mouse brain serum preparation has been further characterized. Following absorption with bone marrow and erythrocytes the antiserum had dose-dependent cytotoxicity against pluripotent stem cells. Rigorous absorptions with bone marrow, spleen, liver, erythrocytes, and thymus failed to remove the anti-stem cell activity of the serum. Adult brain, the immunogen, but not neonatal brain, removed a substantial amount of the activity against stem cells. Maximal cytotoxicity occurred both with and without complement and was maximal following only a 4 degrees C incubation of cells with serum. The anti-stem cell activity was present in the serum globulin fraction. No increase in the frequency of microcolonies or, with longer growth periods, in splenic macrocolonies was observed, suggesting that CFU-s were completely inactivated by exposure to the antiserum. Injected antiserum also reduced CFU-s in vivo.