This study has been conducted according to the principles expressed in the Declaration of Helsinki and after obtaining local Ethics Committee approval. Signed informed consent was obtained from all patients or relatives and healthy volunteers. Oxidative stress data were not used in patients’ management and did not interfere with patient care.

Consecutive patients at ICU admission were included in a prospective and observational study with diagnosis of bacterial community-acquired pneumonia (BCAP) according to ATS/IDSA Guidelines21 or confirmed 2009 pandemic Influenza A/H1N1 pneumonia (VCAP) as reported elsewhere.22 In brief, BCAP was defined as an acute lower respiratory tract infection characterized by (a) an acute pulmonary infiltrate evident on chest radiography and consistent with pneumonia; (b) confirmatory findings of a clinical examination; (c) acquisition of the infection outside a hospital, long-term care facility, or nursing home23 associated with positive bacteriologic data or with a favorable outcome following antibiotic therapy.24 An organism was considered to be a “definite” etiologic agent only if it could be isolated from samples of blood or pleural fluid or if serological tests revealed a four-fold increase in antibody levels. Isolation of Pneumocystis jiroveci or culture of Legionella pneumophila or Mycobacterium tuberculosis from any of the samples was considered to be the basis of a “definite” diagnosis. Other microorganisms isolated from sputum, tracheal aspirate, protected specimen brush (PSB), or bronchoalveolar lavage (BAL) fluid samples were considered to be “probable” pathogens. A urinary antigen test result positive for Legionella pneumophila or Streptococcus pneumoniae was considered to provide a “probable” aetiology.25

The 2009 A/H1N1 primary viral pneumonia (VCAP) was defined in patients presenting during the acute phase of influenza virus illness with acute respiratory disease and unequivocal alveolar opacification involving two or more lobes with negative respiratory and blood bacterial cultures.22 The 2009 A/H1N1 infection was confirmed by means of real-time reverse-transcription-polymerase chain reaction (RT-PCR) on either nasopharyngeal swab samples or tracheal secretions (bronchoaspirate). RT-PCR methods and further details are described elsewhere.22,23 A confirmed case was defined as an acute respiratory illness with laboratory-confirmed An/H1N1 and included in the study.

Severe CAP was defined as patients with CAP requiring ICU admission. Patients with HIV infection, neoplasia, those taking cytotoxic drugs or long-term oral steroid therapy, such as daily dose of 20mg of prednisolone or the equivalent for >2 weeks were considered immune-compromised and were excluded.

Two groups were considered for comparison:

• a)

  Control group (CG): Ten healthy volunteers were included; they were matched with patients according to age (±5 years).

• b)

  Non-infectious group: Twelve patients with non-complicated acute myocardial infarction (AMI) were included. AMI was considered as a good group for comparison because it is an acute pathology of non-infectious origin, which induces a systemic inflammatory response with production of oxygen free radical and tissue damage. This non-infectious group of patients were matched by age (±5 years) and onset time of acute pathology (24h) with the study group patients. AMI was defined as a clinical (or pathologic) event caused by myocardial ischemia in which there is evidence of myocardial injury or necrosis.26,27 Criteria are met when there is a rise and/or fall of cardiac biomarkers, along with supportive evidence in the form of typical symptoms, suggestive electrocardiographic changes, or imaging evidence of new loss of viable myocardium or new regional wall motion abnormality.

In all patients, clinical, demographics and biochemical variables were recorded to determine severity of illness by measuring Acute Physiology and Chronic Health Evaluation (APACHE II score),28 Sequential Organ Failure Assessment (SOFA score)29 and Systemic Inflammatory Response Syndrome (SIRS score).30

Subjects were distributed into 3 groups: (1) CAP patients (study group), (2) healthy volunteers (control Group) and (3) AMI (non-infectious group) for comparison. Subsequently, we performed a comparison within the CAP group differentiating bacterial (BCAP) vs. 2009 A/H1N1 viral (VCAP) etiology. In addition, BCAP and VCAP patients were compared with healthy volunteers group.

After obtaining informs consent from the patient or her/his legal representative and healthy volunteers, blood samples were obtained from all patients on ICU admission, day 3 and ICU discharge. In healthy volunteers only one blood sample was collected.

They were collected in lithium–heparin as well as EDTA tubes. The haemoglobin (Hb) concentration was determined by a spectrophotometer (Perkin Elmer Lambda 2) at 540nm. The blood was centrifuged at 2000rpm (850×g) for 15min to obtain plasmatic and erythrocytary fractions. Plasma was separated and stored at −80°C until analysis. The erythrocytes were washed twice with saline solution, processed and stored at −80°C until analysis. Index of oxidative injury (lipoperoxidative damage) was determined by MDA (Malondialdehyde). Antioxidant capacity was determined by copper–zinc superoxide dismutase (SOD), CAT and glutathione redox system (GSSG, GSH, GR, GPx) activities. GSSG/GSH ratio was calculated as redox buffer index.

MDA concentration was estimated by measurement of thiobarbituric acid reactant substances (TBARS) by a modified fluorescence method31. Briefly, one volume of sample was mixed with 2vol of a solution of 15% (w/v) trichloroacetic acid, 0.375% (w/v) thiobarbituric acid and 0.25N hydrochloric acid, and the mixture heated for 15min in a boiling water bath. After cooling, the precipitate was removed by centrifugation at 3000rpm (1900×g) for 10min. Absorbance was measured at 548nm. Plasma and erythrocytes MDA are expressed as nmol/ml and nmol/g Hb, respectively.

SOD activity in erythrocytes was measured by the epinephrine auto-oxidation according to the method described by Mishra et al.32 One unit of SOD activity was defined as the amount of enzyme required to inhibit the rate of epinephrine auto-oxidation by 50%. The SOD activity results are expressed in UI/g Hb. CAT activity in erythrocytes was measured by the rate of decrease in the hydrogen peroxide (H2O2) absorbency at 240nm.33 Results of CAT activity are expressed in mmol/min/gHb. GPx activity in erythrocytes was determined by the decrease of NADPH at 340nm as described Wheeler CR et al.34. The result is expressed as μmol/min/g Hb. GR activity was determined by the increase of NADP+ at 340nm according to the method of Goldberg et al.35 GR activity is expressed in μmol/min/g Hb.

GSH and GSSG concentrations were determined according to Hissin et al.36 The reduced and oxidized thiols were determined by reaction with O-phtataldehyde for color development detected at 350 and 420nm by a spectrofluorimeter Perkin Elmer LS 50B. Plasma levels are expressed as nmol/ml and erythrocytes levels are expressed as μmol/g Hb.

Discrete variables were expressed as counts (%) and continuous variables as mean and standard deviation (SD), unless stated otherwise; all statistical tests were two-sided. p<0.05 was considered significant. Differences in categorical variables were calculated using two-sided likelihood ratio Chi-square test or Fisher exact test and the Mann–Whitney U test or Kruskal–Wallis test were used for continuous variables, when appropriate. To evaluate the predictive ability of the different biomarkers in regard with mortality, we calculated a receiver operator characteristic (ROC) curve and the area under the ROC (AUROC) curve. We calculate the AUROC because the area does not depend only on a particular portion of the plot, such as the point closest to the diagonal or the sensibility at 90%, but also on the entire plot, and provides a comprehensive picture of the ability of a test to make a distinction over all decision thresholds.

Data analysis was performed using SPSS for Windows 13.0.0 (SPSS, Chicago, IL, US).

Fifteen patients with severe CAP were included and 450 measurements of OS biomarkers were performed. Among the study group, 12 (80%) patients had bacterial CAP (BCAP) and 3 patients (20%) had 2009 A/H1N1 viral pneumonia without bacterial co-infection (VCAP). Clinical characteristics of patients are shown in Table 1. BCAP patients were older and had higher level of organic dysfunction (SOFA score) than patients with VCAP. However, APACHE II score, SIRS score, comorbidities and mortality were not different between groups.

At baseline, day-3 and ICU discharge patients with VCAP showed lower plasmatic TBARS levels than patients with BCAP. Erythrocyte TBARS levels were also lower in patients with VCAP during entire study period. However, this difference only achieved significance at ICU admission (Table 2).

Throughout the study period, erythrocyte GR activity was significantly higher in patients with VCAP in respect of BCAP patients. No other significant differences were observed when different biomarkers of the enzymatic system were compared (Table 2). Within the biomarkers of non-enzymatic system, only plasmatic GSSG levels were significantly higher in patients with VCAP compared to BCAP patients during the entire study period (Table 2). No other significant differences were observed between groups, except for plasmatic GSSG/GSH ratio at ICU discharge.

Etiologic diagnosis was established in all cases of BCAP. A definite diagnosis was provided by blood culture in only 1 (8.3%) case. A probable diagnosis was provided by tests for urinary antigen detection in 4 cases (33.3%) and by bronchoscopic specimen or tracheal aspirate specimens with quantitative cultures in 8 (66.6%) patients. No diagnosis was made by sputum specimen culture. The most frequent microorganism isolated in BCAP was Streptococcus pneumoniae (n=5; 33.3%) followed by Legionella pneumophila (n=3; 20%), Staphylococcus aureus (n=2; 13.3%), Enterococcus faecalis (n=1; 6.7%) and Proteus mirabilis (n=1; 6.7%). No significant differences were observed when comparing oxidative stress biomarkers between pneumococcal and non-pneumococcal pneumonia. However, 2009 A/H1N1 viral aetiology was associated with lower levels of lipoperoxidation (Table 3).

Ten healthy volunteers formed the control group. Plasmatic but not erythrocyte TBARS were higher in CAP patients. No differences were observed within enzymatic system between groups. Among non-enzymatic system, erythrocyte GSSG and GSSG/GSH ratio were higher in CAP patients. However, erythrocyte GSH and plasmatic GSSG were higher in HV (Table 4). CAP patients were older than control group patients due to the influence of patients with BCAP. This difference may explain the higher level of plasmatic TBARS observed. To avoid this possible confounding factor, we matched control group subjects with patients according to age. Two age-matched groups of healthy volunteers were compared with BCAP and VCAP patients respectively. Surprisingly, VCAP patients showed a lower level of lipoperoxidation with respect to controls. Significant differences were not observed when considering patients with BCAP (Table 4). Among enzymatic system, no significant differences were observed in all comparisons. However, GRE activity was higher in VCAP patients than control patients, although the difference was near to the significance (p=0.08). Lower levels of erythrocyte GSH and higher activity of GSSG E were observed in both BCAP and VCAP patients than in control group. In contrast, only BCAP patients showed lower levels of plasmatic GSSG than in control patients (Table 4).

Twelve patients with acute myocardial infarction (AMI) were included in non-infectious patients group. Clinical characteristics of study and AMI group are shown in Table 5. Age, sex and comorbid conditions were not different between groups. However, as expected, the severity of illness measured by the APACHEII, SOFA and SIRS scores was higher in patients with CAP.

Although patients with CAP had higher levels of plasmatic and erythrocyte TBARS than patients with AMI, this difference did not reach statistical significance in any of the moments of the study (Table 6).

At baseline and day-3, erythrocyte SOD activity and SOD/CAT ratio were lower in CAP patients than in AMI patients. No significant differences were observed in other biomarkers of enzymatic system analyzed in different moments of the study. Within the biomarkers of non-enzymatic system, only the plasmatic GSH levels at ICU admission and ICU discharge were significantly lower in patients with CAP than in AMI patients (Table 6).

ICU mortality of CAP was 26.6% (4/15). Three (25%) BCAP patients and 1 (33.3%) VCAP patient died. We found no significant differences in oxidative stress biomarkers between survivors and non-survivors at baseline and day-3 of study period (Table 7).

Discrimination for mortality of different biomarkers and severity of illness scores at baseline was assessed using ROC. The area AUROC showed no consistent mortality discrimination for plasmatic TBARS (AUROC=0.32), GSH (AUROC=0.42), GSSG (AUROC=0.52), erythrocyte TBARS (AUROC=0.44), GSH (AUROC=0.58), GSSG (AUROC=0.42), CAT E (AUROC=0.52); GRE (AUROC=0.36), GPX (AUROC=0.44) and SOFA score (AUROC=0.68). Only SOD E (AUROC=0.72; 95%CI 0.46–0.98) and APACHE II score (AUROC=0.90; 95%CI 0.40–1.0) showed adequate mortality discrimination.