Endotoxin inhibits heat shock protein 70 (HSP70) expression in peripheral blood mononuclear cells of patients with severe sepsis

Objective: To investigate the ex vivo endotoxin-inducible heat shock protein 70 (HSP70) expression in the peripheral blood mononuclear cells (PBMC) of patients with severe sepsis in order to assess the capacity of this potentially protective response during systemic inflammation. Design: Prospective observational study in consecutive patients with severe sepsis and healthy blood donors. Setting: Surgical intensive care unit in a university hospital. Patients and participants: Eleven patients with the diagnosis of severe sepsis, one patient who had recovered from severe sepsis and 13 healthy blood donors. Interventions: None. Measurements and results: We studied the inducibility of HSP70 expression in the PBMC of patients with severe sepsis and healthy blood donors ex vivo. Human whole blood was incubated with variable lipopolysaccharide (LPS from Salmonella minnesota Re 595) concentrations (0; 0.1; 10; 100 ng/ml) for different periods of time (0.5; 2; 4; 10 h). The PBMC were separated by Ficoll density gradient and then disrupted by hypotonic lysis. HSP70 was measured by means of enzyme-linked immunosorbent assay (ELISA). We found a LPS dose- and time-dependent inhibition of ex vivo HSP70 expression in the PBMC of both patients with severe sepsis and healthy individuals. However, the levels of HSP70 expression in patients were significantly lower compared to those of healthy individuals at all LPS concentrations and incubation times. On average, HSP70 expression in the PBMC of healthy controls was 2.8 (range 1.2–3.9) times higher than in patients. HSP70 expression was inducible by thermal heat shock in the PBMC of both patients and healthy individuals. Conclusions: Endotoxin inhibits HSP70 expression in PBMC ex vivo. In vivo, the suppression of HSP70 expression induced by endotoxin and high levels of proinflammatory cytokines may contribute to the cellular dysfunction of immunocompetent cells concerning antigen presentation, phagocytosis and antibody production associated with decreased resistance to infectious insults during severe sepsis. Received: 9 July 1998 Final revision received: 12 October 1998 Accepted: 29 October 1998     

Should we breathe quiet or noisy?

External noise is introduced by computer-generated random levels of pressure assistance during noisy pressure support ventilation (PSV). In patients, noisy PSV was associated with higher tidal volume variability but not improved cardio-pulmonary function compared with conventional PSV. The potential role of noisy PSV in the management of critically ill patients requiring ventilatory support has to be explored further.Although introduced as weaning techniques, modes providing mechanical support of spontaneous breathing have become standard in primary mechanical ventilator support in critically ill patients. In Critical Care, Spieth and colleagues [1] report for the first time the use of noisy pressure support ventilation (PSV) in patients with acute hypoxemic respiratory failure.Normal breathing shows considerable variation in tidal volume (V_T), flow rate, and respiratory rate, which is lost during mechanical ventilation (MV). In addition to elevation of intrathoracic and intrapulmonary pressures, MV causes a non-physiological uniform breathing pattern. Based on the concept that MV should mimic physiologically noisy breathing patterns, biologically variable or noisy MV was introduced, which attempted to mimic spontaneous breath-to-breath variability during volume-controlled MV [2]. Experimental and small clinical trials suggest that biologically variable or noisy MV may improve pulmonary gas exchange, compliance and dead space by preventing de-recruitment when compared to conventional MV [2,3]. These findings support the concept that alveolar recruitment achieved by large V_T exceeds the de-recruitment by small V_T. Other mechanisms claimed to explain improved lung function during biologically variable or noisy MV include stochastic resonance [4], increased respiratory sinus arrhythmia [5], endogenous surfactant release [6,7], and dynamic effects on the pressure-volume curve [7].Nowadays, modes providing assisted support of spontaneous breathing should not only assure adequate gas exchange and unloading of the patient’s work of breathing but should provide patient-ventilator synchrony, optimized diaphragmatic unloading, lung-protective ventilation, and the preservation of physiological respiratory patterns and variability. To accomplish all these tasks ventilatory assistance can no longer be constant but has to continuously adapt to changes in ventilatory demand and respiratory mechanics. To adapt to the continuously changing respiratory demand, pressure assistance is proportional to the instantaneous flow and volume, reflecting the inspiratory muscles’ pressure during proportional assist ventilation (PAV) [8] or proportional to the electrical activity of the diaphragm during neurally adjusted ventilator assist (NAVA) [9]. Thus, PAV and NAVA try to amplify the patient''s respiratory center output. Several experimental investigations and clinical trials in small groups of critically ill patients have demonstrated that PAV and NAVA, when compared to conventional PSV, enhance patient-ventilator interaction and synchrony, which translates into better comfort and sleep quality and preserves V_T variability, which has been associated with improvements in gas exchange and lung mechanics [10,11].In patients with mechanically assisted spontaneous breathing, the noise can be introduced externally or can come directly from the respiratory center. In contrast to PAV and NAVA, which amplify the noise coming from the respiratory center, external noise is introduced by computer-generated random levels of pressure assistance during noisy PSV [12]. Experimental investigations in induced lung injury showed that noisy PSV, when compared with conventional PSV and pressure-controlled MV, was associated with a significantly higher coefficient of variation of V_T and airway pressure, and resulted in better pulmonary gas exchange, reduced alveolar edema in overall lung as well as reduced inflammation in the nondependent parts of the lungs [13]. In a porcine model with induced lung injury an external source of noise (noisy PSV) and noise derived from amplification of the respiratory center (PAV) improved gas exchange and produced higher V_T variability, whereas the pulmonary inflammatory response and diffuse alveolar damage score did not differ when compared to conventional PSV lungs [14].Spieth and colleagues [1] investigated the short-term effects of conventional and noisy PSV in patients with acute hypoxemic respiratory failure. Noisy PSV was associated with higher V_T variability and a lower number of asynchrony events. In contrast to experimental findings, cardio-pulmonary function and spatial distribution of ventilation were comparable between conventional and noisy PSV [1]. During conventional and noisy PSV, however, V_T significantly higher than 8 ml/kg predicted body weight was frequently noticed. By design noisy PSV applies V_T as high as 16 ml/kg and as low as 1.6 ml/kg once every 20 to 30 minutes [12]. Although short-term experiments demonstrate that this mixture of ultra-protective and non-protective V_T during noisy PSV does not add to lung injury based on histological examinations, long-term investigations have to clarify the relevance of periodic non-protective V_T ventilation in critically ill patients.