The Association between Mortality Rates and Decreased Concentrations of Interleukin-10 and Interleukin-1 Receptor Antagonist in the Lung Fluids of Patients with the Adult Respiratory Distress Syndrome

Methods

We used Murray and colleagues' expanded definition of ARDS [21]. The extent of injury to study patients with trauma was assessed by using the injury severity score [22]. With regard to multiple organ failure, the extent of organ involvement and the severity of organ failure were quantified by using a modified Goris organ-failure score, as described elsewhere [23]. We enrolled 28 consecutive patients with ARDS from the intensive therapy units at the Royal Infirmary and Western General Hospitals, Edinburgh, United Kingdom. Informed consent was obtained from the relatives or guardians of the patients. The study was approved by the Lothian Health Board Ethics Committee.

The control group was composed of 9 patients with trauma who had degrees of trauma insult similar to those of the enrolled patients with ARDS (as assessed by injury severity scores) and who had been ventilated for a period similar to that of the enrolled patients. These controls had had severe head injuries but had no clinical, radiographic, or microbiological evidence of lung injury at the time of enrollment.

Within 24 hours of the clinical diagnosis of ARDS, a bronchoscopic and bronchoalveolar procedure was done. The fiberoptic bronchoscope was introduced through an indwelling endotracheal tube. The distal end of the bronchoscope was wedged into the right middle lobe. Three 60-mL aliquots of saline 0.9% NaCl solution were instilled and were immediately gently aspirated. On average, 59% of instilled fluid was recovered (range, 40% to 85%). All bronchoscopy procedures were done by the same bronchoscopist. Recovered fluid was stored at 4 °C until processing, which was done within 1 hour of the bronchoscopy procedure. Processing entailed first straining the lavage fluid through sterile gauze to remove mucus. The strained fluid was then centrifuged at 400 g at 4 °C for 10 minutes to recover cells. Total cell counts were obtained using a hemocytometer. Aliquots of cells were pelleted onto glass slides using a cytospin 2 (Shandon Scientific, Cheshire, United Kingdom) and were then stained with Diff-Quick (Merz-Dade AG, Dudingen, Switzerland), a modified Wright-Giemsa stain. Differential counts were determined by counting 500 cells under oil immersion (original magnification × 100). The lavage fluid supernatant was respun at 1000 g at 4 °C for 10 minutes to remove cellular debris and was then stored at − 70 °C until it was assayed for selected cytokines.

Cytokine Enzyme-Linked Immunosorbent Assay

Concentrations of antigenic tumor necrosis factor, interleukin-1 β, interleukin-6, interleukin-8, interleukin-10, and interleukin-1 receptor antagonist were quantified using a modification of a double-ligand method, as described elsewhere [8, 24, 25]. Flat-bottomed 96-well microtiter plates (Nunc Immuno-Plate I 96-F, Life Technologies, Paisley, United Kingdom) were coated with 50 µL of the appropriate polyclonal antibodies per well (1 ng/µL in 0.6 mmol/L of NaCl, 0.26 mmol/L of H₃BO₄, and 0.08 normal NaOH; pH, 9.6) for 24 hours at 4 °C and were then washed with phosphate-buffered saline (pH, 7.5; 0.05% Tween 20 as the wash buffer). Microtiter plate nonspecific binding sites were blocked with 2% bovine serum albumin in phosphate-buffered saline and incubated for 60 minutes at 37 °C. Plates were rinsed three times with wash buffer and diluted (neat and 1:10), and samples (50 µL/well) were added. This was followed by incubation for 1 hour at 37 °C. Plates were washed three times, and to each well we added 50 µL of biotinylated polyclonal rabbit or the appropriate antihuman tumor necrosis factor, interleukin-1 β, interleukin-6, interleukin-8, interleukin-10, or interleukin-1 receptor antagonist antibodies (3.5 ng/µL in phosphate-buffered saline [pH, 7.5]; 0.05% Tween 20; 2% fetal calf serum added). Plates were incubated for 45 minutes at 37 °C.

Plates were washed three times, streptavidin peroxidase conjugate (Bio-Rad Laboratories, Richmond, California) was added, and the plates were incubated for 30 minutes at 37 °C. Plates were washed three times, and chromogen substrate (Bio-Rad Laboratories) was added. The plates were incubated at room temperature to the desired extinction, and the reaction was terminated with 50 µL of 3 mmol/L of H₂SO₄ solution per well. Plates were read at 490 nm in an automated microplate reader (Bio-Tek Instruments, Inc., Winooski, Vermont). Standards were 0.5-log dilutions of recombinant cytokines from 100 ng to 1 pg/mL (50 µL/well). This method consistently detected specific cytokine concentrations of at least 50 pg/mL in a linear fashion.

Statistical Analysis

Patient group data and assay results are expressed as medians (ranges). Intergroup comparisons were made using nonparametric methods (the Mann-Whitney test or the Spearman rank correlation coefficient, as appropriate). Significance was defined as a P value less than 0.05. Calculations were done using the statistical software package Statview (Abacus Concepts, Berkeley, California) for the Macintosh computer.

Results

Twenty-eight patients with ARDS were enrolled in our study; 36% subsequently died. The initiating event was sepsis for 22 patients (8 of whom died) and multiple trauma for 6 patients (2 of whom died). For enrolled patients with sepsis, the focus of sepsis was distant to the lung (perforated bowel in 18 patients and pancreatitis in 4 patients). The patients with ARDS and trauma had a median injury severity score of 33 (range, 24 to 59); and a median age of 53 years (range, 18 to 77 years). Nine ventilated controls with trauma (6 were men) were also enrolled during the study period; their median age was 34 years (range, 22 to 86 years), and their median injury severity score was 27 (range, 15 to 38). The associated mortality rate for the controls was 44% (4 of the 9 died). The controls had been ventilated for a median of 36 hours (range, 18 to 56 hours). No significant difference was found between patients with ARDS and controls for duration of ventilation before study enrollment, age, or injury severity scores (P > 0.2 for all three comparisons). No significant difference was found between survivors and nonsurvivors in the time spent in an intensive therapy unit before study enrollment (P > 0.2). In patients with ARDS who died, the median time from study enrollment to death was 12 days (range, 3 to 32 days).

The median interleukin-10 concentrations in bronchoalveolar lavage fluid were 100 pg/mL (range, 0 to 1600 pg/mL) in patients with ARDS and 0 pg/mL (range, 0 to 50 pg/mL) in controls. The median interleukin-1 receptor antagonist concentrations in bronchoalveolar lavage fluid were 820 pg/mL (range, 0 to 18 900 pg/mL) in patients with ARDS and 50 pg/mL (range, 0 to 240 pg/mL) in controls. Concentrations of both interleukin-10 and interleukin-1 receptor antagonist were significantly higher in patients with ARDS than in controls (P = 0.0005 for interleukin-10 and P = 0.002 for interleukin-1 receptor antagonist). In patients with ARDS, a significant positive correlation was found between concentrations of interleukin-10 and interleukin-1 receptor antagonist in bronchoalveolar lavage fluid (P = 0.003).

Median interleukin-10 concentrations were 40 pg/mL (range, 0 to 110 pg/mL) in patients with ARDS who died and 120 pg/mL (range, 30 to 1600 pg/mL) in patients with ARDS who survived (Figure 1). Again, this difference was found to be highly significant (P = 0.003). The median interleukin-1 receptor antagonist concentrations were 90 pg/mL (range, 0 to 3400 pg/mL) in patients with ARDS who died and 1600 pg/mL (range, 80 to 18 900 pg/mL) in patients with ARDS who survived (Figure 2). A highly significant difference was seen between both patient groups (P = 0.009).

View larger version:

• In this window
• In a new window

Figure 1. Concentrations of interleukin-10 in bronchoalveolar lavage samples obtained from ventilated controls (n = 9), survivors of the adult respiratory distress syndrome (ARDS) (n = 18), and nonsurvivors of ARDS (n = 10).

View larger version:

• In this window
• In a new window

Figure 2. Concentrations of interleukin-1 receptor antagonist in bronchoalveolar lavage samples obtained from ventilated controls (n = 9), survivors of the adult respiratory distress syndrome (ARDS) (n = 18), and nonsurvivors of ARDS (n = 10).

To rule out the possibility that a dilutional phenomenon had contributed to our findings, we measured albumin concentrations in collected bronchoalveolar lavage samples (median, 550 µg/mL [range, 110 to 1400 µg/mL] in patients with ARDS; median, 28 µg/mL [range, 0.0 to 65 µg/mL] in controls). No significant difference was found in albumin concentrations in bronchoalveolar lavage fluid between survivors and nonsurvivors (P > 0.2).

In the 18 patients with ARDS who survived, a significant inverse relation was seen between a high bronchoalveolar interleukin-10 concentration and the lowest Pa O_2/FiO2 ratio subsequently attained in the intensive therapy unit (P = 0.008). In addition, a significant positive correlation was found between measured interleukin-10 concentrations and subsequent multiple organ failure scores (P = 0.04). No significant correlation was found between concentrations of this anti-inflammatory cytokine and time spent on a ventilator (P > 0.2).

Concentrations of proinflammatory cytokines (tumor necrosis factor, interleukin-1 β, interleukin-6, and interleukin-8) were also measured in bronchoalveolar lavage samples taken from the 22 patients with ARDS and the 8 controls. Significantly elevated concentrations of the proinflammatory cytokines tumor necrosis factor (P = 0.01) (median, 90 pg/mL [range, 0.0 to 2500 pg/mL] in patients with ARDS; median, 0.0 pg/mL [range, 0.0 to 118 pg/mL] in controls), interleukin-6 (P = 0.001) (median, 179 pg/mL [range, 0.0 to 2200 pg/mL] in patients with ARDS; median, 0.0 pg/mL [range, 0.0 to 80 pg/mL] in controls), and interleukin-8 (P = 0.0001) (median, 628 pg/mL [range, 0.0 to 4700 pg/mL] in patients with ARDS; median, 0.0 pg/mL [range, 0.0 to 278 pg/mL] in controls) were found in the alveolar air spaces of patients with ARDS compared with controls. However, no significant difference was found in the concentrations of these proinflammatory cytokines in bronchoalveolar lavage samples taken from survivors and nonsurvivors (Table 1).

Discussion

In the alveolar air spaces of patients with ARDS, we detected concentrations of interleukin-10 and of interleukin-1 receptor antagonist that exceeded those found in ventilated controls who did not have ARDS. However, in patients with ARDS, low concentrations of these agents—which have been associated with various anti-inflammatory properties—were significantly associated with higher subsequent patient mortality rates. These findings support the hypothesis that failure to mount a major intrapulmonary anti-inflammatory cytokine response contributes to a worse overall prognosis in patients with ARDS.

Interleukin-10 was originally identified as a product of murine T helper cell 2 clones and is a potent downregulator of macrophage and monocyte function. Specifically, interleukin-10 results in the down regulation of major histocompatibility complex class II antigens, changes in cell cytotoxicity, and the inhibition of proinflammatory cytokine production [16-19]. Interleukin-1 receptor antagonist is a naturally occurring inhibitor of the biological activity of interleukin-1 and has no known agonist action [15]. It is a 23- to 25-kd molecule whose inhibitory action results from competition with interleukin-1 for the type 1 interleukin-1 receptor; the binding of interleukin-1 receptor antagonist to this receptor does not result in signal transduction [26, 27]. Interleukin-1 receptor antagonist, when administered in vivo, has been shown to attenuate acute inflammatory processes by significantly decreasing inflammatory cell infiltration, edema, and tissue necrosis [28]. Both interleukin-10 and interleukin-1 receptor antagonist are believed to play key regulatory roles in limiting proinflammatory response. In animal models of acute lung injury, investigators have recently shown that blockage of endogenous interleukin-10 is associated with a heightened inflammatory response and substantially enhanced lung injury [20]. Our finding of an association between low concentrations of interleukin-10 and interleukin-1 receptor antagonist in the alveolar air spaces of patients who recently received a diagnosis of ARDS and subsequent patient mortality rates supports the results of recent work done both in vitro and in vivo. It also suggests that an inability to modulate and down regulate the inflammatory process at this stage of disease pathogenesis leads to a worse prognosis for patients with ARDS.

As have previous investigators [9, 29-31], we found increased proinflammatory cytokine concentrations in bronchoalveolar lavage samples obtained from patients with ARDS. Our finding of elevated proinflammatory cytokine concentrations in the alveolar air spaces of patients with ARDS and our finding that alveolar concentrations of these cytokines did not differ significantly between survivors and nonsurvivors of ARDS provide evidence of an enhanced inflammatory process in the alveolar air spaces of patients with ARDS, regardless of subsequent survival. Furthermore, our data were obtained shortly after ARDS was diagnosed; the median time from diagnosis to death was 12 days. This suggests that the inability to mount an appropriate localized anti-inflammatory response is not a terminal event before death but occurs earlier in the pathogenesis of ARDS.

In survivors of ARDS, we found a significant association between elevated interleukin-10 concentrations and lower Pa O_2/FiO2 ratio, which is a marker of the severity of lung injury. This finding suggests that the initial severity of lung injury did not prevent the local generation of these anti-inflammatory agents and may partly explain why it has been so difficult to effectively predict patient outcome on the basis of the initial degree of lung injury.

Clearly, in the complex situation of established ARDS, many factors relating to disease pathophysiology and management in the intensive therapy unit are likely to influence outcome. Nevertheless, we have shown, at the time of ARDS diagnosis, a significant association between low concentrations of anti-inflammatory cytokines in the lungs and subsequent patient mortality rate. To definitively address whether measurement of these concentrations in bronchoalveolar lavage would offer practical clinical prognostic benefit, a larger prospective study of sequential measurements is required. In addition, because we took samples at only one time, we cannot state whether patients with ARDS who died 1) were able to mount a substantial interleukin-10-interleukin-1 receptor antagonist response within the alveolar air spaces at an earlier time point [a response that was subsequently downregulated] or 2) ever had the ability to initiate an appropriate anti-inflammatory response during their illness. Serial measurements in a larger patient sample would allow us to more accurately delineate the intraalveolar anti-inflammatory cytokine profile over time. These issues are currently being addressed in ongoing research within our department.

In summary, we found elevated concentrations of the anti-inflammatory cytokines interleukin-10 and interleukin-1 receptor antagonist in the alveolar air spaces of patients with ARDS compared with ventilated controls. Unlike concentrations of important proinflammatory mediators, these elevated concentrations had prognostic significance with regard to subsequent rates of death from ARDS. Our data support the hypothesis that the failure to mount a localized anti-inflammatory response early in the pathogenesis of established ARDS leads to a worse overall prognosis.

Drs. Strieter, Burdick, and Kunkel: Departments of Medicine and Pathology, Division of Pulmonary and Critical Care Medicine, University of Michigan, Ann Arbor, MI 48109.

Dr. Armstrong: Department of Intensive Therapy Unit, Royal Infirmary Hospital, Edinburgh, Scotland.

Dr. Mackenzie: Department of Anaesthetics, Queen Margaret Hospital, Dunfermline, Scotland.