INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The hyperproduction of oxygen free radicals (OFR) and the weakening of natural scavenging mechanisms are implicated in endothelial cell damage, myocardial alterations, and multiple organ failure during septic shock. One of the most important sources of OFR is the reduction of altered tetravalent oxygen as a consequence of endotoxic, hypoxic, and acidotic conditions. Muscle cell damage and acidosis increase the quantity of free iron released from myoglobin and hemoglobin (1), and by the dangerous Fenton reaction. The monocyte and polymorphonuclear neutrophil (PMN) pool undergoes alterations that are suggestive of leukocyte activation as a response to stimulation by tumor necrosis factor (TNF) and interleukins (IL) (2). IL-6 magnifies by 10-fold the activation of these cells, resulting in superoxide generation (3). Final targets of OFR damage in septic shock include the lung, kidney, brain, liver, cardiovascular system, skeletal muscle (4), and complement and coagulation pathways (5).

Many authors have proposed the use of antioxidants such as sulfurated amino acids, vitamins, enzymes (superoxide dismutase, SOD) (6), dual cyclooxygenase and lipooxygenase inhibitors (ketoprofen) (7), glutathione (GSH), and N-acetylcysteine (NAC) (8) to decrease OFR damage in patients affected by septic shock.

The tripeptide GSH is one of the most important endogenous antioxidants. It plays the role of a sulfhydryl (SH) group provider for direct scavenging reactions. In doing so, it acts both as a substrate in the scavenging reaction catalyzed by glutathione peroxidase and as a scavenger of vitamin E and vitamin C radicals.

The scavenging capacity of the human body depends mainly on the pool of GSH and SH groups. New GSH may be recovered from the oxidized form (glutathione disulfide, GSSG) by glutathione reductase, with the consumption of nicotinamide adenine dinucleotide phosphate (NADPH). The amount of NADPH may be reduced during shock, contributing to a reduction in the effectiveness of mechanisms for recovering GSH.

NAC is a well-known artificial precursor of GSH. As a mucolytic drug it has been widely used for many years. Its antioxidant and antitoxic effects in animal models and humans have recently been demonstrated (9). NAC both increases GSH levels (10) and acts as a powerful OFR scavenger, yielding NAC-disulfide end products (11). Furthermore, NAC inhibits the release of TNF- (12), the activation of proinflammatory cytokines (13), and cellular apoptosis (14), and is currently indicated for use against acute overdosing with paracetamol (15).

NAC alone has been used with good results in patients with septic shock, who were identified as "responders" to this therapy (8). A pilot trial in our department (O. Ortolani, unpublished data) has shown a positive synergism of GSH with NAC in septic shock. No adverse effects of GSH or NAC administration were observed in that trial.

The present study was intended to confirm the capacity of GSH and NAC in high doses to cooperate in reducing lipoperoxidative damage in patients with early septic shock. We expected a favored membrane crossing by GSH, with an increased extracellular concentration and a further intracellular increase in GSH by adding NAC, provided that no adverse effects were described in the literature for GSH or NAC at doses up to 150 mg/kg/d.

Measurement of free radical production is very difficult, owing to the very short lives of these substances. We therefore chose four biochemical parameters to indirectly evaluate the entity of peroxidative stress: expired ethane and plasma malondialdehyde, which are directly related to the extent of lipoperoxidative events (16); erythrocyte membrane deformability, because the peroxidation of membrane phospholipids increases the stiffness of all cell membranes, including the erythrocyte membrane (17); and the activation of complement factor C5, on the assumption that the complement system activation in patients with septic shock is mainly a free radical-mediated event (18). We also monitored plasma GSH, but did not consider it an indicator for variations in the redox potential because the values registered during the trial might have been affected by the continuous administration of GSH and NAC.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Forty-five septic shock patients with an Acute Physiology and Chronic Health Evaluation (APACHE) II score (19) of 23 ± 7 (mean ± SD) and Logistic Organ Disfunction (LOD) score (20) of 10.7 ± 2.2 were studied within 24 h of diagnosis in the Intensive Care Departments of the Universities of Naples and Florence.

Septic shock was identified through the following criteria: (1) oliguria with a urine volume of less than 20 ml/h; (2) systolic arterial blood pressure (BP)  90 mmHg, a decrease in systolic BP from baseline exceeding 40 mm Hg or a requirement for inotropes for more than 1 h to maintain a systolic arterial blood pressure above 90 mm Hg; (3) evidence of a septic focus or positive blood culture, fever, or hypothermia (body temperature > 37.5° C or < 35.5° C), and leukocytosis or leukopenia (white blood cell count > 10,000/ml or < 4,000/ml).

Exclusion Criteria

The following patients were excluded from the trial:

Patients unable to maintain hemodynamic conditions that allowed optimal conventional resuscitation, and with a mean arterial pressure persistently under 70 mm Hg despite inotropic support.

Patients with hematocrit values below 30%, or receiving blood transfusions.

Patients unable to keep a Pa_O2 of 80 to 140 mm Hg and a CO₂ of 35 to 50 mm Hg, or requiring a fractional inspired oxygen concentration (FI_O2) of over 50%.

Patients with severe heart disease, or taking calcium channel antagonists, angiotensin converting enzyme inhibitors, corticosteroids, NAC, or other drugs with antioxidant activity.

Patients who died before Day 5 of the trial.

Treatment Protocol

Eligible patients were randomly admitted to each of the treatment groups, at up to 10 patients per group.

All patients received conventional therapy for septic shock, including parenteral nutrition with lipids and oligoelements, antibiotics, volume-expanding and inotropic agents, and other drugs, according to the patient's pathology. Drugs with known antioxidant properties were avoided. Group A (n = 10) had no additional treatment and served as a control group. Group B (n = 10) received 70 mg/kg/d of intravenous GSH. Group C (n = 10) received 70 mg/kg/d of GSH and 75 mg/kg/d of intravenous NAC.

GSH and NAC were administered in continuous, syringe pump-controlled saline infusions (200 ml/24 h). The trial lasted 5 d. Administration of GSH and NAC was continued until dismissal from the trial. The treatment protocol was chosen on the basis of unpublished previous, preliminary results.

The patients or their relatives were informed of the purpose of the trial, and gave their written informed consent. The trial was conducted according to the Helsinki principles and was approved by the local ethics committees of the Universities of Florence and Naples.

Biochemical Procedures

The following parameters were measured: expired ethane; plasma concentration of malondialdehyde (MDA); erythrocyte membrane stiffness (EMS); activation of complement factor C5; and plasma concentrations of GSH and GSSG.

Air and blood samples were collected immediately before enrollment in the trial (time = 0) and again on Days 3 and 5 of treatment. GSH and NAC administration was suspended 1 h before sample collection.

Quantitation of expired ethane was performed according to the method of Lawrence and coworkers (21), through the collection of 500 ml of human expirate into two 250-ml gas-tight syringes (Supelco Inc., Bellefonte, PA), which were kept at 37° C in thermostatic containers until processed. The expirate was then passed through a 100 × 3-mm coil column filled with Porasil-A (Supelco), and refrigerated at 76° C in a CO₂/acetone bath. The column was then eluted in a sand bath at 160° C for 10 min, and the eluate (5 ml average) was collected in a gastight syringe. The gas chromatograph used for quantitating ethane was a Pye Unicam (Cambridge, UK) Model 205, chromatograph equipped with a flame ionization detector. High-grade purified gases for gas chromatography were obtained from SON, Inc. (Naples, Italy). The ethane standard (99 ppm) was obtained from Supelco. The gas chromatography was performed with a 30 ft × 1/8 inch stainless steel column containing 23% SP 1700 coated on 80/100-mesh Chromosorb P AW (Supelco) at a constant temperature of 70° C, an injector temperature of 100° C, a flame ionization detector 8 × 10¹⁰ at a temperature of 110° C, and a He carrier gas flow of 25 ml/min. Two tests were performed on each sample (injection volume: 1 ml each) with a gastight syringe. A 0.5-ml gas sample containing 1 ppm of ethane was injected as a standard. The ethane concentration in the subjects' expired air was evaluated with an automatic Pye Unicam CDP1 integrator. The results were expressed in nanomoles.

Quantitation of plasma MDA was done with high-pressure liquid chromatography (HPLC) according to the method of Wong and colleagues (22). Blood samples were collected in 7 ml Vacutainer tubes (Becton Dickinson, Plymouth, UK) containing the dipotassium salt of ethylenediamine tetraacetic acid (K₂-EDTA). After 30 min of incubation at 100° C with thiobarbituric acid (TBA) as a reactant, the TBA adduct was injected into a 50-µl sampling loop device for HPLC. The HPLC apparatus (Beckman Instruments, Berkeley CA) was equipped with two pump systems (Model 110 A), a sample injector valve (Model 210 A), a UV detector (Model 160), and a microprocessor system controller (Model 420), all from Beckman, and an Omniscribe D-5000 strip-chart recorder (Bausch & Lomb, Austin, TX). The separation was performed with 0.44 M H₃PO₄/methanol 60/40 (vol/vol) on an Ultrasphere ODS, 5 µm, 250 × 4.6 mm column at room temperature, with a flow rate of 2 ml/min and detector wavelength of 532 nm. The TBA adduct from the patients was isographic with the tetraethoxypropane standard, showing an average retention time of 4.2 min. The results were reported in nanomoles.

Erythrocyte membrane deformability was measured according to the method of Teitel (23). Fifteen milliliters of blood were drawn into three siliconated Vacutainers containing K₂-EDTA. One tube was kept at 37° C and was processed within 15 min. The remaining two test tubes were centrifugated at low speed for 10 min to obtain the autologous plasma needed to dilute the sample to a hematocrit of 30%. One milliliter of the diluted blood was placed into a 1-ml plastic insulin syringe connected with a 13-mm, 5-µ-pore filter (Biorad, Segrate, Italy), and the filtration time at 37° C under 20 cm H₂O pressure was recorded. The system was checked before use on each occasion, with an air flow under 10 cm H₂O pressure to ensure proper filter assembly. The results were reported as seconds per milliliter of blood.

Complement activation (C5a) was evaluated as the percentage of leukocyte aggregation produced by the patient's plasma on a PMN suspension from a normal subject (24). The blood sample was collected in Vacutainers containing 7 ml K₂-EDTA and centrifuged. The leukocyte aggregation was evaluated with an optic aggregometer (Braun Melsungen A.G., Melsungen, Germany). Ninety percent transmission was assigned to a leukocyte suspension diluted to 50% and 10% transmission was assigned to the undiluted leukocyte suspension. A Bausch & Lomb paper recorder recorded the percent transmittance and generated an aggregation curve. The results were expressed as % leukocyte aggregation.

The aggregation potential of the leukocyte suspension was evaluated by adding 200 µl of plasma in which the complement was activated by zymosan.

Plasma total and oxidized GSH were evaluated according to the method of Stenzel and coworkers (25). The GSH concentration was calculated by subtracting the GSSG concentration from the total GSH.

Statistics

Results are reported as mean ± SD. Differences among groups were evaluated through analysis of variance followed by the Tukey's post hoc test. Differences within individual groups were evaluated with Student's t test. Statistical significance was accepted at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thirty of 45 patients, consisting of 16 males and 14 females aged 52 ± 12 yr (mean ± SD) and with an APACHE II score of 21 ± 4 were admitted to the trial. The three groups in the trial were similar with respect to patient demographics, diagnostic categories, comorbidities, age, sex, weight, and organ failure score. Table 1 shows the diagnostic categories and characteristics for each group. Table 2 shows the data for patients eliminated from the trial.

Figure 1 shows the values for expired ethane and plasma MDA, and Figure 2 shows erythrocyte filtration and complement C5 activation monitored in Groups A, B, and C on Days 0, 3, and 5. The basal values were high in all patients with septic shock as compared with the normal values recorded with the same methods, which were: 0.7 ± 0.3 nM for ethane; 5.0 ± 0.4 nM for MDA; 21% ± 3 for C5a (PMN aggregation); and 28 ± 4 s/ml for EMS (erythrocyte filtration time).

View larger version (20K): [in this window] [in a new window]   Figure 1.   Values (mean ± SD) of expired ethane and MDA in the three trial groups on Days 0, 3, and 5. Groups B and C showed a marked reduction of lipoperoxidative indexes (ethane in the expirate and plasma MDA). This reduction was already statistically significant by Day 3 in Group C (p < 0.05).

View larger version (24K): [in this window] [in a new window]   Figure 2.   Values (mean ± SD) of erythrocyte filtration time and percent granulocyte aggregation. Erythrocyte membrane stiffness and C5 complement activation were significantly decreased in Group C after 3 d of treatment (p < 0.05). The differences in Group B were significant by Day 5 (p < 0.01).

Group B and C patients showed a marked decrease in their lipoperoxidative indexes (ethane in the expirate and plasma MDA) on Day 5 as compared both with basal values (p < 0.01) and those of control Group A (p < 0.01). This reduction was already statistically significant (p < 0.01) by Day 3 in Group C, which was receiving both GSH and NAC. In this group, lipoperoxidative indexes were significantly decreased by the Day 5 as compared with Group B (p < 0.01). Complement activation (C5a) showed a similar trend, but with little difference between the two treated groups.

Erythrocyte membrane stiffness was also significantly decreased in Group C after 3 d of treatment as compared with basal values (p < 0.05). The differences in Group B were significant by Day 5 (p < 0.05). Differences among groups were more significant (p < 0.01).

Figure 3 shows values (mean ± SD) for plasma GSH and GSSG. Plasma GSH was significantly increased in Groups B and C, but it was impossible to determine whether this increase was due to the continuous administration of GSH and NAC or to the antioxidant effects of the therapy. GSSG values varied greatly, and differences within and between groups were therefore not considered.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Values (mean ± SD) of GSH and GSSG in the three groups on Days 0, 3, and 5. Plasma GSH was significantly increased in Groups B and C as a result of continuous administration of GSH. The standard deviation for GSSG values was too high for statistical evaluation of the results.

There were no significant changes in the mean arterial pressure, heart rate, or cardiac output of the treated patients during the study in comparison with those of the control group.

Mortality was similar in all the groups during the trial (of the 45 studied, 2, 3, and 2 patients died, in groups A, B, and C, respectively), but by Day 10 the mortality in group A was twice that in Groups B and C (4, 2, and 2 more patients of the 30 admitted patients died between Days 6 and 10) (p < 0.01). APACHE II and LOD scores (Table 3) did not show significant variations during the 5 d of the trial. These scores in the treatment groups were significantly reduced by Day 10 as compared both with the basal values (p < 0.01) and with the control group values (p < 0.05), and were also significantly decreased in Group C as compared with Group B (p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The protective action of GSH may be exerted in both extracellular and intracellular fluids. In order to cross the cell membrane, GSH is split into two components (cysteinyl glycine and gamma glutamyl amino acids) by -glutamyltranspeptidase. These amino acid components pass through the membrane by the action of -glutamyltransferase. The GSH is then reconstituted in the protoplasm of the cell by GSH-synthetase. A direct carrier-mediated transport of GSH across brain cells (26) and a protective role of GSH against oxidative stress (27) have been demonstrated.

The interesting finding in our trial was the increased protective effect of adding NAC to GSH against OFR damage in patients with septic shock, as demonstrated by the results in Group C.

A number of studies have demonstrated the antioxidant role of NAC. Thus, NAC supplementation was found to increase peroxyl radical scavenging capacity and to reduce the oxidative stress by improving the thiol redox status (28); NAC was found to increase GSH levels in brochoalveolar fluid (29); and NAC was found to inhibit human neutrophil and monocyte chemiotaxis and oxidative metabolism, to scavenge hypochlorous acid (HClO) and hydroxyl radicals (OH·) from neutrophil myeloperoxidase (30), and to reduce the oxidative damage from endotoxins (31). Zhang and associates (32) measured an increase in the oxygen extraction ratio secondary to NAC administration in endotoxic shock.

Very recently, Spies and colleagues (8) reported the results of a trial of high doses of NAC, given within 24 h of ICU admission, in humans with septic shock. The authors did not find positive protection in all patients, but only in a group that they defined as the group of "responders." This may have been caused by the treatment with GSH or NAC having been given too late in some cases. The hypothesis that the choice of patients with poor residual reactivity may explain the different results of other authors should be considered (33). Furthermore, we are conscious that pathologies with very different etiologic and prognostic values are grouped under the same definition of septic shock despite efforts to narrow the classification criteria for this condition. In the present trial, we attempted to keep narrow inclusion criteria in order to ensure the uniformity of the sample.

Red cell filtration time was indicative of OFR damage as expected. This seems consistent with the increase in plasma MDA in shock states regardless of their etiology, and with the increase in plasma MDA in red cells during septic shock (34). Our hypothesis is that better deformability of red cells in the treated groups was directly linked to the reduction of peroxidative stress, and therefore of plasma MDA. The marked reduction of complement activation (C5a) can be ascribed to the same mechanisms.

We chose not to correlate increased plasma GSH with the variation in redox potential in the treated groups, owing to the possible influence of the high quantity of infused thiols. We were unable to draw conclusions from the behavior of GSSG, which was very variable in all of the trial groups.

Mortality was not affected by the treatment during the 5 d of the trial, but it was quite different by Day 10. We do not expect that antioxidant therapy alone, as a primary treatment, will greatly improve the survival of patients with septic shock, because septic shock cannot be simply reduced to a free-radical pathology; however, we consider antioxidants to be useful components of multidrug therapies.

We found no adverse effects in the groups receiving high doses of GSH and NAC. Even though a depression in cardiovascular performance with NAC was reported in the literature (35), we suggest that this was due to the excessive dosage of more than 200 mg/kg in 24 h. Spies and colleagues (8) reported an increased cardiac index, stroke index, and left ventricular stroke work index in their group of responders to NAC treatment at 150 mg/kg/24 h. Galley and associates (33) reported similar results.

Conclusions

Our results confirm the protective role of GSH on some lipoperoxidative parameters in patients with septic shock, and are consistent with the hypothesis that NAC associated with GSH can increase the protection against free radical-mediated damage by modulating GSH metabolism, scavenging HClo and hydroxyl radical generated by activated human neutrophils, and increasing the extracellular pool of SH (11, 36). GSH and NAC given in high doses to patients with shock may act cooperatively in normalizing some important coagulative and metabolic pathways that are frequently involved in shock-related complications. The patients in our treated groups showed an improvement in biochemical indexes and clinical scores. The results were still better in the group treated with added NAC, which was characterized by a more rapid correction of the considered indexes and by a faster recovery.

In conclusion, the results of our study show that the administration of high doses of GSH (70 mg/kg/d) in association with NAC (75 mg/kg/d) in a continuous infusion adds further protection over that conferred by GSH alone, against the peroxidative stress of septic shock.