© 2002 American Thoracic Society
Granulocyte Macrophage-Colony Stimulating Factor Augmentation Therapy in SepsisIs There a Role?Children's Hospital Medical Center Cincinnati, Ohio
Sepsis is an infection-induced syndrome (1) characterized by a systemic inflammatory response syndrome (SIRS) (2) and a clinical outcome that can be self-limited or as severe as septic shock. The annual incidence and mortality of sepsis in the United States are estimated at 751,000 and 215,000, respectively, with an associated annual cost of approximately $16.7 billion (3). Antibiotics are necessary but not sufficient for treatment of severe sepsis because a number of large, well-designed trials have not shown improvement in mortality despite vast improvements in the antimicrobial armamentarium (4). However, such studies have dramatically expanded our knowledge of the pathophysiology of sepsis (5) and its therapeutically useful division into sequential stages: high risk of infection Granulocyte macrophage-colony stimulating factor (GM-CSF) is a hematologic growth factor now known to be required for alveolar macrophage function, lung host defense, and surfactant homeostasis (reviewed in [8]). GM-CSF deficiency in mice (generated by gene-targeting) and in humans (due to neutralizing anti–GM-CSF autoantibodies) results in defects in alveolar macrophage function. GM-CSF functional deficiency in these settings leads to development of pulmonary alveolar proteinosis and defects in lung host defense. Of relevance to the article by Presneill and coworkers in this issue of AJRCCM (pp. 138–143), GM-CSF restores immunologic functions in monocytes from septic patients (9) and can prevent immune paralysis in endotoxin-desensitized mice. Presneill and coworkers have conducted a randomized, double-blind, placebo-controlled phase II trial of GM-CSF augmentation therapy for the treatment of severe sepsis with respiratory dysfunction (10). This study included 18 patients (10 treatment, 8 placebo) and tested the hypothesis that low-dose GM-CSF infusion (3 mcg/kg/day intravenously for 5 consecutive days) in individuals with severe sepsis and sepsis-related pulmonary dysfunction would improve leukocyte function without exacerbating lung or other organ dysfunction. Treatment and control groups were well-matched for clinical parameters, APACHE II and SOFA scores, and the proportion of subjects with ARDS at enrollment. The treatment was well-tolerated, but did not improve mortality. Notwithstanding this lack of effect on mortality, GM-CSF significantly improved oxygenation, modulated the numbers and functions of blood and alveolar neutrophils, and tended to reduce the prevalence of ARDS. One interesting observation by Presneill and coworkers is that GM-CSF improved oxygenation in patients with sepsis (10). No mechanism was identified for this improvement. The results, however, could not be explained by changes in several factors known to influence pulmonary gas exchange during mechanical ventilation, including continuous positive airway pressure, alveolar PO2, pulmonary vascular hemodynamics, and cardiovascular resuscitation. Improved oxygenation was associated with a decrease in the proportion of patients developing ARDS in the GM-CSF treatment group (40% to 20%), whereas this proportion increased in the placebo group (37.5% to 62.5%). A link between GM-CSF and reduction in ARDS was previously suggested by a report from this group showing that an elevated GM-CSF level in the lungs early in the course of ARDS was associated with patient survival (11). The power of the present study, however, was not adequate to reach statistical significance. The concept that GM-CSF may modulate function(s) of the respiratory gas exchange surface was also previously suggested by studies in transgenic mice in which overexpression of GM-CSF in the lung produces marked type II alveolar epithelial cell hyperplasia and lung growth. Further studies are clearly needed to assess the role of GM-CSF on lung structure and function and to determine the relationship of GM-CSF and ARDS in severe sepsis with pulmonary dysfunction. The observation by Presneill and colleagues that GM-CSF stimulated both the number and function of blood neutrophils is consistent with the expected effects of GM-CSF. The lack of statistical significance for the increase in blood neutrophils, however, suggests that GM-CSF dosing was suboptimal to stimulate production of these cells. The reduction in lung neutrophils in the treatment group is more difficult to explain, and again no mechanism was identified. Although the pharmacokinetics and distribution of endogenous murine GM-CSF in mice and of synthetic recombinant human GM-CSF in humans could well be different, data from transgenic mice suggest that GM-CSF is highly compartmentalized between blood and lungs (8). Additional studies should thus explore GM-CSF dose optimization with regard to effects on the function of blood and alveolar leukocyte during sepsis. Finally, it is disappointing that GM-CSF therapy did not improve mortality of sepsis with pulmonary dysfunction. There are several possible reasons for the failure, including the small number of patients studied, uncontrolled use of glucocorticoids, and the lack of patient stratification according to the degree of monocyte deactivation at the time of enrollment. For example, it is plausible that the observed treatment effect was limited to a subgroup of individuals in which immune paralysis was present or more significant. If true, this suggests that stratification of subjects on the basis of monocyte dysfunction might be useful in identifying individuals who could benefit from GM-CSF therapy during sepsis. A recent randomized, double-blind, placebo-controlled trial of GM-CSF to reduce nosocomial infections in very low birth weight neonates also showed no treatment effect (12). Although this study included 264 individuals and used higher GM-CSF doses (8 µg/kg/day) for a longer period of time, it should be also noted that in this study, use of glucocorticoids was not controlled and patients were not stratified with respect to monocyte dysfunction. Based on the observations that (1) individuals in the post-hyperinflammatory phase of sepsis appear to be immunocompromised, (2) late mortality in sepsis is associated with infection, and (3) GM-CSF prevents the immune paralysis in animal models of sepsis, further studies are warranted. From the previous discussion, one could envision that such studies might explore GM-CSF dose optimization and patient stratification based on immunophenotyping on enrollment. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
infection 
(TNF-
