Exercise-induced asthma (EIA) describes a condition in which vigorous physical activity triggers acute airway obstruction in individuals with heightened airway reactivity. Up to 90% of those with chronic asthma, 3-10% of the general population, and 11–50% of elite athletes (26) may experience EIA (27). Significant numbers of asthmatics demonstrate signs and symptoms of EIA, including cough, wheeze, dyspnea, shortness of breath, chest tightness, and chest pain, normally beginning after a brief period of exercise (5-8 min) and resolving spontaneously within 30–60 min after stopping exercise (27). Exercise is a powerful trigger of asthma symptoms and may result in asthmatic patients avoiding physical activity resulting in detrimental consequences to their health.
The mechanism(s) unique to exercise that triggers EIA in asthmatic patients have been extensively investigated (3,18). It has been suggested that transient dehydration of the airways activates inflammatory mediators such as histamine, neuropeptides, and the arachidonic acid metabolites, leukotrienes, and prostanoids from airway cells, resulting in bronchial smooth muscle contraction (3). Alternatively, it has been suggested that rapid rewarming of the airways following exercise leads to reactive hyperemia resulting in vascular engorgement and perivascular edema (18), which would further contribute to airway narrowing caused by bronchoconstriction.
There is accumulating evidence that dietary modification has the potential to reduce the prevalence and incidence of EIA (19). Recent studies have supported a role for dietary salt as a modifier of airway narrowing following exercise in EIA subjects, as it has previously been shown that a high-salt diet (HSD) worsens and a low-salt diet (LSD) improves postexercise pulmonary function in EIA (9,20).
The mechanism(s) by which dietary salt may modify EIA is not known. However, possible mechanisms include a direct effect of the sodium (28) and/or chloride ion (20) on airway smooth muscle contractility, the release of bronchoconstrictor mediators from airway cells either directly or through changes in airway osmolarity as a result of dietary salt modification (21), and the potential influence of dietary salt on vascular volume and microvascular pressure, resulting in mucosal edema and narrowing of the airway lumen (8). How dietary salt is involved in any or some of these probable mechanisms remains to be determined.
Therefore, the main aim of this study was to demarcate a possible mechanism by which dietary salt modification may alter airway narrowing following exercise in EIA subjects. We sought to characterize airway inflammation directly via induced sputum by measuring differential cell counts and fluid phase mediators, and to compare inflammatory markers on different salt-loading regimens. In addition, we wished to determine whether intravascular volume expansion, as a result of dietary salt loading, may have an effect on the diffusion capacity of the lung (DLCO) and its subdivisions, and therefore be a causative factor in exercise-induced airway narrowing in asthmatic subjects. We hypothesized that DLCO and its subdivisions and the degree of airway inflammation would be dependent on the amount of dietary salt intake in EIA subjects.
Twenty-four subjects (15 male, 9 female; age (mean ± SD), 24 ± 1.8; height, 164 ± 1.5 cm; body mass, 71.4 ± 5.8 kg (normal salt diet (NSD)) with physician-diagnosed asthma and with documented EIA were recruited from a population of university students, the local community, and were recreationally active. All subjects had clinically treated mild persistent asthma, with FEV1 greater than 70% of predicted (Table 1). In addition, all subjects were atopic as assessed by skin prick tests for common allergens conducted by their physician. A group of nonasthmatic (control) subjects was not included in the present study as it has been shown repeatedly that dietary salt manipulation does not alter pre- and postexercise pulmonary function in this population (9,20). All subjects had a history of shortness of breath, chest tightness, and intermittent wheezing after exercise, relieved by bronchodilator therapy (N = 14, salbutamol; N = 10, terbutaline). All subjects were asked to discontinue their maintenance medication (N = 12, budesonide; N = 4, montelukast; N = 5; zileuton; N = 3, zafirlukast) during the course of the trial (informed consent was obtained from each subject and his or her physician). Subjects were excluded if they were pregnant, or if they had a history of hyperlipidemia, hypertension, diabetes, bleeding disorders, or delayed clotting time, and were taking aspirin medication. Short-acting β2-agonists were discontinued 12 h before exercise testing. The subjects were also asked to refrain from coffee/alcohol and physical exercise 8 and 24 h, respectively, before the exercise challenge. For female participants, stage of the menstrual cycle was recorded on exercise testing days. Each subject completed a health questionnaire and gave written informed consent before enrollment in the study, approved by the institutional research ethics committee.
Study design and protocol.
The study was conducted as a double-blind, randomized, crossover trial over five consecutive weeks. All subjects entered the study on their NSD (phase 1), after which they were randomly assigned to either an HSD or an LSD for 2 wk (phase 2). Thereafter, they followed a 1-wk washout period (phase 3) on their NSD and then switched to the alternative diet for 2 wk (phase 4). All subjects were required to consume a base diet of 1500 mg·d−1 (65 mmol·L−1·d−1) of sodium and approximately 2250 mg·d−1 (64 mmol·L−1·d−1) of chloride, which was provided by a menu plan, whether on the LSD or HSD. For the HSD, the base diet was supplemented with 10 1-g salt capsules per day comprising 4000 mg·d−1 (174 mmol·L−1·d−1) of sodium, and approximately 6000 mg·d−1 of chloride (169 mmol·L−1·d−1). For the LSD, the base diet was supplemented in the same manner, but with placebo (sucrose) tablets.
At the initial screening test on the NSD and at the end of each 2-wk treatment phase, pulmonary function (PF) and DLCO was assessed pre- and postexercise (PF: 1, 5, 20, 45, 75, 90, 105, and 120 min; DLCO: 10 and 25 min) (Fig. 1). The screening test was conducted to screen all subjects for the presence of EIA, as indicated by a more than 10% decrease in postexercise FEV1 compared with preexercise values. Each subject underwent collection of induced sputum 48 h before exercise to establish baseline values on the respective dietary salt regimen, and at 1, 6, and 24 h after each exercise challenge (Fig. 1). Twenty-four hours before each exercise challenge at the end of each 2-wk treatment period, subjects reported to the laboratory to have their plasma volume measured using the Evans blue (EB) dilution method (Fig. 1). Blood pressure measurements, using brachial artery sphygmomanometry, were taken at the beginning of the study and every third day of the treatment period. Blood pressure was also measured pre- and postexercise. To monitor dietary compliance, 24-h urine collections were made at the beginning of the study (NSD), at the end of each 2-wk treatment period, and at the end of the washout period (Fig. 1). Body mass was determined before exercise at screening and at the end of each treatment period. At the end of the 1-wk washout period, all subjects reported to the laboratory to void additional urine and undergo further pre- and postexercise PFT, to verify that urinary electrolyte balance and the degree of exercise-induced airway narrowing had returned to baseline levels established at the beginning of the study (NSD) (Fig. 1). Dietary cards were recorded throughout the study period. In addition, all subjects were asked to record bronchodilator use during each 2-wk treatment period.
Exercise challenge test.
Each subject was required to run on a motorized treadmill, which was elevated 1% per minute (modified Balke protocol), until volitional exhaustion. The speed/inclination of treadmill differed for each subject. However, the starting treadmill speed/inclination was kept constant between diets. Each subject wore a nose clip during the exercise bout to promote mouth breathing, as nasal breathing decreases the water loss from the airways. In addition, each subject inspired compressed dry air (relative humidity less than 3 mg H2O·L−1 air) at room temperature (22°C) collected in a 150-L Douglas bag (Hans Rudolph, Kansas City, MO) attached to the inspiratory port of a two-way breathing valve connected to a mouth piece. During the exercise test, HR was continuously monitored by ECG (Quinton Q710 Stress test Monitor; Quinton Instrument Company, Seattle, WA), and breath-by-breath analysis of expired gases was accomplished by indirect open circuit calorimetry (SensorMedics Vmax 22 metabolic cart; SensorMedics Corp., Yorba Linda, CA).
Spirometry was performed with the subject in the sitting position breathing room air, with the nose occluded by a clip. All testing was completed using a calibrated computerized pneumotachograph spirometer (SensorMedics Vmax 22, SensorMedics Corp., Yorba Linda, CA) according to American Thoracic Society recommendations (2). The pulmonary function technician and spirometer were the same throughout the study. The procedure for all spirometry tests was 1) three normal tidal volume breaths, 2) maximal inhalation, 3) forced maximal exhalation, and 4) maximal inhalation. The best of three consistent trials was recorded. If any pre- and postexercise challenge pulmonary function time point measurement was technically unacceptable, it was repeated. The maximum percentage fall in FEV1 from the baseline (preexercise) value was calculated using the following equation: (preexercise FEV1 - lowest postexercise FEV1)/(preexercise FEV1). In addition, the bronchoconstrictor response to exercise was also assessed as the area under the curve (AUC) of the percentage fall in postexercise FEV1 plotted against time for 120 min (AUC0-120). The AUC0-120 was calculated using trapezoidal integration.
Diffusion capacity of the lung and its subdivisions.
The diffusion capacity of the lung (DLCO, mL·min−1·mm Hg−1) measurements were performed according to the American Thoracic Society standard recommendations for the single-breath carbon monoxide diffusing capacity (transfer factor) (1), using a SensorMedics Vmax 22 (SensorMedics Corp.). Each subject wore a nose clip and remained seated throughout the DLCO procedure. Subjects exhaled to residual lung volume (RV); inhaled a gas mixture of 20.9% O2, 0.3% CO, and 0.3% CH4 balanced with N2 to total lung capacity; held their breath for 10 s; and then exhaled to RV. Between trials, 2 min were allowed to permit the washout of gases. DLCO measurements were accomplished in duplicate and averaged if within 10% or 3 mL·min−1·mm Hg−1. To calculate the intrinsic diffusion capacity of the alveolar-capillary membrane (DMCO) and pulmonary capillary blood volume (VC), a second DLCO test was then performed immediately in the same manner as the first test, with a gas mixture of 90% O2, 0.3% CO, and O.3% CH4 balanced with N2. The calculations of VC and DMCO were made using the formula 1/DLCO = 1/DMCO + 1/ØVC, where Ø is the reaction constant of hemoglobin (Hb) and oxygen. The value for Ø was determined from 1/Ø = 0.33 + 0.0057Pco2, where Pco2 is the partial pressure of oxygen in the pulmonary capillary. Pco2 was estimated as end-expiratory partial pressure of oxygen (Peo2) - 10, where Peo2 = barometric pressure - 47 × Feo2. Once the value of Ø was computed, 1/DLCO was plotted against I/Ø for both oxygen tensions. Pulmonary capillary blood volume (VC) was calculated as the reciprocal of the slope, and DMCO was calculated as the reciprocal of the y-intercept. At each diffusion measurement period, a fingertip sample of capillary blood was collected onto a microcuvette, and the Hb concentration was measured using a Hemocue® B-Hb photometer (Hemocue, Inc., Mission Viejo, CA) to correct DLCO measures and to calculate DMCO and VC.
Induced sputum production and processing.
Before the sputum induction, all subjects inhaled 200 μg salbutamol to minimize bronchoconstriction during induction procedure. Sputum was induced by inhalation of 3, 4, and 5% hypertonic saline in sequence for 5 min, generated by a DeVilbiss 65 ultrasonic nebulizer, and the subjects were encouraged to cough and expectorate sputum into sterile ampoules. FEV1 was measured after each nebulization, and subjects were asked to blow their noses and rinse their mouths with water before expectoration to minimize nasal contamination of the sample. Nebulizations were stopped if a decrease in FEV1 of >20% compared to baseline values occurred or if troublesome symptoms appeared (24).
Sputum was examined as previously described (6). Briefly, sputum free of salivary contamination was selected immediately after expectoration, weighed, and mixed with four times its volume of 0.1% dithiothreitol (DTT) saline solution (Sigma Chemical, St. Louis, MO) maintained at 4°C. After rocking on a bench for 15 min, the sample was further diluted with four volumes of Dulbecco’s phosphate-buffered solution (phosphate-buffered saline (PBS)). The suspension was filtered through a 48-μm gauze and centrifuged at 800 × g for 10 min. The cell-free supernatant was removed and stored at −80°C and until analysis. The cell pellets were resuspended in PBS (1%), and total cell count was measured in a hemocytometer (AO Scientific Instruments, Buffalo, NY) and viability determined by trypan blue exclusion. The cells were cytocentrifuged (Cytospin 2, Shandon Instruments, Runcorn, UK). The suspension was air-dried, stained with Romanovski stain, and a differential cell count obtained by counting at least 400 nonsquamous cells by a blinded observer. Differential counts were expressed as corrected percentages, after subtraction of squamous cells. Sputum eosinophilia was defined as a sputum differential eosinophils count greater than 2%. In order to ensure good cell viability, sputum was processed within 2 h of collection.
Inflammatory mediator measurements.
The concentrations of mediators in sputum supernatant were determined by competitive immunoassays for PGD2 (Cayman Chemical, Ann Arbor, MI), cysteinyl leukotrienes (LTC4-E4) (Cayman Chemical), leukotriene (LT) B4 (Neogen, Lansing, MI), and sandwich enzyme-linked immunosorbent assay for IL-1β (Cayman Chemical) and IL-8 (BD Biosciences Pharmingen, San Diego, CA). Because PGD2 is a relatively unstable compound, we measured PGD2-methoxime (MOX), a stable derivative of PGD2. The concentration of ECP in the sputum supernatant was measured by radioimmunosorbent assay using an ECP radioimmunoassay kit (Pharmacia Diagnostics, Uppsala, Sweden). An ELx405™ Automated Plate Washer (Bio-Tek® Instruments, Inc., Winooski, VT) was used to wash the 96-well microplates during the preassay preparation. All assays were performed using a Powerwave XS™ Spectrophotometer (Bio-Tek® Instruments, Inc., Winooski, VT). The sensitivity of the assays were PGD2, 3.2 × 10−3 ng·mL−1; LTC4-E4, 13 × 10−3 ng·mL−1; LTB4, 0.1 ng·mL−1; IL-1β, 0.8 × 10−3 ng·mL−1; IL-8, 0.8 × 10−3 ng·mL−1; and ECP, 0.125 ng·mL−1. The intraassay and interassay coefficients of variability were 5-10% and 3-15%, respectively, across the range of mediators measured. Specificity for the variety of mediators was as follows: PGD2-MOX assay kit: PGD2-MOX 100%, PGD2 0.2%; LTC4-E4 assay kit: LTC4 100%, LTD4 100%, LTE4 67%; LTB4 assay kit: LTB4 100%, 6-trans-LTB4 25%, LTB5 14.58%; IL-1β assay kit: IL-1β 100%, IL-1α and IL-2 <0.01%; IL-8 assay kit: no cross-reactivity (value greater than 2 pg·mL−1) was identified for IL-1α, IL-1β, and IL-2 through IL-15; ECP does not cross-react with eosinophil-derived neurotoxin.
Urinary analysis of electrolytes.
Urine samples were analyzed for sodium, potassium, and chloride concentrations using ion-specific electrodes (Beckman Astra Analyzer, Beckman Instruments, Inc., LaBrea, CA). Urinary creatinine concentration was determined by a modified Jaffe rate reaction, using the same instrument, to verify completeness of the 24-h urine samples.
Determination of plasma volume.
During all plasma volume measurements, using the EB dye dilution method, the subjects always sat in an upright posture (30 min) to stabilize body fluids during this procedure. Subjects were instructed to abstain from fluid and food 12 h before the experiment. Approximately 2.5 mL (5 mg·mL−1) EB (New World Trading Corporation, DeBary, FL) was injected through a venous catheter into the antecubital vein. The syringe was thereafter flushed with 15-20 mL of isotonic saline to wash out residual dye from the syringe and catheter. A baseline value in triplicate (3 × 3 mL) was drawn 1 min before injection. Ten, 20, and 30 min after injection, 3 mL of blood was collected from the central venous catheter and transferred to polyethylene tubes containing 15 IU heparin·mL−1 of blood. After centrifugation for 15 min at 2000 × g, the absorbencies of EB dye in plasma were read at a wavelength of 620 nm using a Beckman Coulter DU640 spectrophotometer (Beckman Coulter, Fullerton, CA), and plasma volume was calculated as described previously in detail (7), with the average value of the calculated plasma volume being computed.
Body mass (nude) was determined by using a Seca 883 digital weighing machine (Seca Corporation, Hanover, MD) to the nearest 0.1 kg. Height was measured using a Seca 216 wall-mounted Stadiometer (Seca Corporation) without socks and shoes.
Data were analyzed using the SPSS version 12 statistical software (SPSS Inc., Chicago, IL). The data were assessed for normality using the Kolmogorov-Smirnov test, and Levene’s test was used to test for homogeneity of variance between groups. A repeated-measures ANOVA was used to analyze the data, with both treatment and time as “within-subject” effects. Mauchly’s test was conducted to determine whether sphericity was violated. If sphericity was violated, the repeated-measures ANOVA was corrected using the Greenhouse-Geiser correction factor. Where a significant F-ratio was found (P < 0.05), Fisher’s protected least-square difference post hoc test was used to isolate differences in group means (P < 0.05). Induced sputum differential cell counts were analyzed nonparametrically using the Friedman repeated-measures ANOVA on ranks and described as median and interquartile range. Induced sputum supernatant mediator concentrations (corrected for sputum dilution and expressed as nanograms per milliliter) were normally distributed and were described as mean ± SD. Correlations between induced sputum cell counts, sputum supernatant mediator concentrations, and the severity of EIA were calculated using Spearman’s rank order correlation coefficients and Pearson’s product moment correlation coefficient. Data were analyzed for the presence of carryover effects between treatments using a 2 × 2 ANOVA. All reported P values were considered significant at the 0.05 level. Pulmonary function, lung diffusion data, and 24-h urinary excretion of electrolytes and plasma volume data were expressed as mean and their 95% confidence intervals (CI).
With 24 asthmatic EIA patients, the data had at least 80% power to detect statistical significance, and this calculation was based on our previous work (9,20). Gotshall et al. (9) and Mickleborough et al. (20) demonstrated effect sizes of 0.69 and 0.71, respectively, with sample sizes of eight for a study power of 80%. All the above power calculations assumed a two-tailed α of 0.05.
All subjects who entered the trial completed it. Bronchodilator use was significantly reduced on the LSD (12 puffs; 95% CI, 9.6-14.4 puffs), compared with the NSD (18 puffs; 95% CI, 16-20 puffs) and HSD (26 puffs; 95% CI, 23.2-28.8 puffs). A 2 × 2 ANOVA that was used to test for the presence of crossover (carryover effects) between treatments indicated that none was present (P > 0.05) for all variables measured; this was further supported by urinary electrolyte concentrations and pulmonary function values at the end of the washout period returning to baseline levels established at the beginning of the study (NSD). No subjects showed any abnormal increases in blood pressure at screening or during the course of the study. Compared with baseline (NSD, 69.4 kg; 95% CI, 67.3-71.5 kg) and the LSD (68.3 kg; 95% CI, 65.3-71.3 kg), body mass significantly increased (P < 0.05) on the HSD (72.6 kg; 95% CI 70.6-74.6 kg).
No significant differences (P > 0.05) were observed in preexercise (baseline) pulmonary function values among diets (Table 1). The differential effect of the percentage change in pre- to postexercise FEV1 is shown in Figure 2. Subjects demonstrated airway obstruction on the NSD and HSD with a significant (P < 0.05) percentage of decrease in FEV1 of 18.3% (−0.52 L; 95% CI, −0.21 to −0.83 L) and 27.4% (−0.74 L; 95% CI, −0.35 to −1.13 L), respectively, at 20 min postexercise. However, on the LSD, the percentage of decrease in FEV1 significantly decreased (P < 0.05) to 7.9% (−0.24 L; 95% CI, −0.10 to 0.38 L). Reductions in postexercise FEV1 in excess of 10% occurred for up to 75 and 120 min on the NSD and HSD, respectively (Fig. 2). In addition, similar changes as a result of diet were observed in forced vital capacity (FVC) and forced expiratory flow (FEF) at 25-75% of FVC (FEF25-75%). The severity of EIA as determined by the AUC0-120 was significantly greater (P < 0.05) on the HSD (2023.8; 95% CI, 1458-2432) compared with the NSD (1411.6; 95% CI, 954-1884) and LSD (535.3; 95% CI, 237-745). In addition, no relationship (r = 0.03, P = 0.89; 95% CI, −0.38 to 0.43) was evident between maximal ventilation achieved during exercise and the change in pre- to postexercise FEV1.
Diffusion capacity of the lung and its subdivisions.
Lung diffusion data are presented in Table 2. The preexercise (baseline) DLCO and carbon monoxide transfer coefficient (KCO; DLCO/VA) were not significantly different (P > 0.05) between diets. Postexercise DLCO and KCO values were not significantly changed (P < 0.05) compared with baseline values on the LSD; however, there was a significant reduction (P < 0.05) on the NSD and HSD in postexercise DLCO and KCO compared with baseline values. The DLCO at 10 and 25 min postexercise was significantly reduced (P < 0.05) on the HSD compared with the LSD by 25% (−8.0 mL·min−1 · mm Hg−1; 95% CI, 2.12-13.88 mL·min−1·mm Hg−1) and 27% (−8.4 mL·min-1·mm Hg−1; 95% CI, 2.52-14.28 mL·min−1·mm Hg−1, respectively. Comparable results in KCO, as a consequence of diet, were observed postexercise challenge. The preexercise (baseline) DMCO was not significantly different (P > 0.05) between diets. However, the DMCO was significantly reduced at 10 and 25 min postexercise on the HSD compared with LSD by 16% (−10.0 mL·min−1·mm Hg−1; 95% CI, 2.16-17.84 mL·min−1·mm Hg−1 and 17% (−10.1 mL·min−1·mm Hg−1; 95% CI, 3.21-16.63 mL·min−1·mm Hg−1, respectively. Preexercise (baseline) VC data revealed that subjects had significantly higher (P < 0.05) VC on the HSD compared with the LSD by 6% (3.9 mL; 95% CI, 1.55-6.3 mL). At 25 min postexercise, VC significantly increased (P < 0.05) by 6.3 mL (95% CI, 2.0-10.6 mL) and 9.6 mL (95% CI, 4.7-14.5 mL) on the NSD and HSD, respectively, compared with baseline values, with no significant changes (P > 0.05) observed on the LSD. Comparable changes in postexercise VC/VA were seen on the HSD and NSD. There was a significant correlation between DLCO and severity of EIA on the NSD (percentage of decrease in FEV1 (r = 0.64, P = 0.0084; 95% CI, 0.43-0.82), AUC0-120 (r = 0.72, P = 0.0056; 95% CI, 0.51-0.81)) and on the HSD (percentage of decrease in FEV1 (r = 0.74, P = 0.0049; 95% CI, 0.56-0.80), AUC0-120 (r = 0.78, P = 0.0042; 95% CI, 0.49-0.89)).
Induced sputum differential cell counts and inflammatory markers.
The total cell numbers and percentages of eosinophils, neutrophils, lymphocytes, macrophages, and bronchial epithelial cells in induced sputum were not significantly different (P > 0.05) at baseline between diets, as shown in Table 3. In addition, there was no significant difference (P > 0.05) in the total cell count and percentage of sputum lymphocytes and bronchial epithelial cells between diets following the exercise challenge. However, the percentages of sputum eosinophils and neutrophils were significant elevated (P < 0.05) and macrophages significantly suppressed on the HSD compared with the NSD and LSD. Sputum eosinophilia (≥2%) was found in all baseline samples.
The sputum supernatant mediator concentrations are shown in Figures 3–8. There were no significant differences (P > 0.05) in mediator concentrations at baseline between diets. However, significantly higher (P < 0.05) sputum ECP (Fig. 3), IL-1β (Fig. 4), IL-8 (Fig. 5), LTC4-E4 (Fig. 6), LTB4 (Fig. 7), and PGD2-MOX (Fig. 8) concentrations were observed at 1 and 6 h following exercise on the HSD compared with the NSD and LSD. At 6 h postexercise, ECP, IL-1β, IL-8, and PGD2-MOX concentrations peaked at 353.7 (95% CI, 336.5-370.9), 13.4 (95% CI, 9.5-18.5), 3.5 (95% CI, 2.7-3.9), and 0.36 (95% CI, 0.33-0.39) ng·mL−1, respectively, on the HSD, whereas on the salt-restricted diet, ECP, IL-1β, IL-8, and PGD2-MOX concentrations at 6 h postexercise were 189.4 (95% CI, 174.6-204.2), 7.2 (95% CI, 6.3-7.6), 1.7 (95% CI, 1.5-1.9), and 0.25 (95% CI, 0.23-0.27) ng·mL−1, respectively. Sputum LTC4-E4 concentration significantly increased (P < 0.05) by 72.1% (+10.6 ng·mL−1; 95% CI, 7.1-14.2 ng·mL−1) on the HSD, with a comparable increase in LTC4-E4 observed at 6 h postexercise. However, no significant increase (P > 0.05) in LTC4-E4 was observed at 1 and 6 h postexercise on the LSD. The HSD caused a significant increase (P < 0.05) in sputum LTB4 concentration of 62.2% (+10.4 ng·mL−1; 95% CI, 6.6-15.4 ng·mL−1) compared with baseline values, whereas on the LSD, no significant change (P > 0.05) in postexercise sputum LTB4 concentration was observed compared with baseline values.
The percentage of neutrophils in induced sputum postexercise on the NSD was correlated with the severity of EIA (percentage of decrease in FEV1 (r = 0.64, P = 0.0024; 95% CI, 0.32-0.83), AUC0-120 (r = 0.71, P = 0.0038; 95% CI, 0.43-0.87)) and for the HSD (percentage of decrease in FEV1 (r = 0.61, P = 0.0048; 95% CI, 0.27-0.81), AUC0-120 (r = 0.63, P = 0.0057; 95% CI, 0.30-0.82)). There was also a positive correlation between the percentage of induced sputum eosinophils and severity of EIA on the NSD (percentage of decrease in FEV1 (r = 0.69, P = 0.0034; 95% CI, 0.40-0.86), AUC0-120 (r = 0.74, P = 0.0029; 95% CI, 0.48-0.88)) and on the HSD (percentage of decrease in FEV1 (r = 0.76, P = 0.0041; 95% CI, 0.51-0.89), AUC0-120 (r = 0.65, P = 0.0057; 95% CI, 0.33-0.83)). Similar positive correlations were observed between sputum supernatant concentration of ECP and the severity of EIA on the NSD (percentage of decrease in FEV1 (r = 0.67, P = 0.0068; 95% CI, 0.37-0.83), AUC0-120 (r = 0.71, P = 0.0035; 95% CI, 0.41-0.86)) and HSD (percentage of decrease in FEV1 (r = 0.78, P = 0.0027; 95% CI, 0.55-0.90), AUC0-120 (r = 0.76, P = 0.0032; 95% CI, 0.54-0.86)).
Urine electrolytes and plasma volume.
Table 4 presents the 24-h urinary excretion data of electrolytes and plasma volume changes. Subjects demonstrated dietary compliance throughout the course of the study. A graded dose of dietary sodium was achieved in this study from 1446 (95% CI, 1148-1744) mg of sodium per day (62.8, 95% CI 49.9 to 75.8 mmol of sodium per day) for the LSD to 3537 (95% CI, 3198-3875) mg of sodium per day (153.8, 95% CI 139.0 to 168.4 mmol·L−1 of sodium per day) for NSD through to 9873 (95% CI, 9,460-10,285) mg of sodium per day (429.3, 95% CI 411.3 to 447.2 mmol·L−1 of sodium per day) for the HSD. The plasma volume on the LSD was significantly reduced (P < 0.05) by 11.1% (−409 mL; 95% CI, 232-624 mL) and significantly increased (P < 0.05) by 12.2% (+534 mL; 95% CI, 240-828 mL) on the HSD compared with the NSD (Table 4).
This double-blind, placebo-controlled trial in asthmatic patients has demonstrated for the first time that modifying dietary salt intake for 2 wk alters airway inflammation and DLCO and its subdivisions and confirms previous observations that altering dietary salt intake can modify the severity of EIA (9,20) with a concomitant reduction in bronchodilator drug use. We found that differential eosinophil and neutrophil cell counts, ECP, proinflammatory cytokines IL-1β and IL-8, and eicosanoids LTC4-E4, LTB4, and PGD2 in induced sputum were significantly higher following exercise on the NSD and HSD compared with the LSD. In addition, the NSD and HSD caused a significant reduction in DLCO, KCO, and DMCO and an increase in VC, whereas the LSD produced a significant increase in DLCO, KCO, and DMCO and a decrease in VC.
In the present study, increasing dietary salt intake resulted in an increase in plasma volume, weight gain, and urinary volume output in the asthmatic subjects. The increases in weight gain and plasma volume are good indices of the degree of increase in extracellular volume due to salt loading. It is possible that the substantial changes observed in pulmonary function following exercise, occurring as a result of altered dietary salt intake, might reflect altered central blood volume with potential downstream effects on pulmonary venous and microvascular pressure. Indeed, it has been demonstrated in asthmatic patients that significant decreases in lung mechanics and altered airway hyperresponsiveness to provocations have occurred when large amounts of saline have been infused to rapidly expand the volume of the bronchial circulation (8). Gilbert and coworkers (8) demonstrated that the administration of 2 L of normal saline within a 20-min time period in asthmatic patients results in profound decrements in airway function following isocapnic hyperventilation. Brown and coworkers (5) have shown in animals that saline infusion causes airway wall edema, as documented by high-resolution computed tomography. These authors (5) demonstrated in canine airways that acute volume loading with saline not only causes an increase in airway wall thickness and a decrease airway luminal area, but also increased airway responsiveness to an aerosol histamine challenge.
Dietary salt restriction in this study improved postexercise pulmonary function to below the diagnostic limit of a 10% postexercise fall in FEV1, and caused no change in postexercise DLCO and its subdivisions relative to baseline values. However, on the NSD and HSD, postexercise DLCO, KCO, and VA values were significantly reduced compared with baseline values, whereas postexercise VC was significantly increased compared with baseline values. These observed changes in DLCO and its subdivisions on the NSD and HSD were accompanied by postexercise reductions in pulmonary function. These data suggest that the airway obstruction may in part be related to increased blood flow through the pulmonary circulation, which could affect airway narrowing through capillary engorgement, airway wall edema, or direct mechanical compression (8,13), with an attendant decrease in forced expiratory flow rates. Similarly, parenchymal lung mechanics may be altered by increased pulmonary capillary blood flow leading to capillary leak and interstitial edema (8). It is feasible that major changes in bronchial blood flow may cause secondary changes in pulmonary capillary blood flow at the level of the peripheral airways because of the extensive anastomoses of the two circulations in this area. However, our protocol was not designed to test changes in bronchial blood flow and therefore cannot address the bronchial circulation issue directly. A model of airway wall thickening proposed by Hogg and coworkers (11) suggests that a small increase in thickness, such as that caused by edema or vascular congestion, might account for the increased airway resistance seen with bronchial provocations.
An observed decrease in DMCO and KCO following exercise implies that the thickness of the alveolar-capillary membrane is acutely augmented, thus impeding the passage of the respiratory gases. The alveolar-capillary interface is formed by a number of different physical layers, including the alveolar epithelium, interstitial fluid, capillary endothelium, plasma, and red blood cell membrane, any of which might potentially be affected by volume loading as a result of increasing salt intake. It is therefore reasonable to suggest that volume loading as a result of increasing dietary salt consumption in asthmatic subjects may cause subclinical interstitial pulmonary edema, which results in the reduction of DMCO and KCO. However, Robertson and coworkers (25) recently showed that although saline infusion in healthy subjects causes airway obstruction as documented by a reduction in FEV1, no changes in DLCO or DMCO were noted. A possible reason for the apparent discrepancy in DLCO and its subdivisions to volume loading may be due to the fact that asthmatics have a larger and leakier capillary bed in their airway walls, and is most likely secondary to their chronic inflammatory status (17). Therefore, asthmatics might be at an exaggerated risk for the development of airway obstruction due to vascular dilation with mucosal thickening following volume loading.
However, airway narrowing following exercise in asthmatics may be caused by a number of factors, including airway smooth muscle contraction, airway edema, mucus secretion, and reactive hyperemia of the airway wall. Reactive hyperemia has been suggested to occur as a result of a rapid increase in bronchial blood flow within the airway wall in response to the cooling and drying of the airway during exercise (18), leading to vascular engorgement and perivascular edema following exercise, both of which could contribute to narrowing of the airway (18). In the present study, dietary salt loading significantly elevated pre- and postexercise VC, which is indicative of pulmonary capillary engorgement. It is therefore possible that capillary engorgement due to salt loading may lead to an exaggerated hyperemic response to exercise, resulting in augmented capillary leakage, edema formation, and subsequent airway narrowing following exercise.
One of the aims of the present study was to compare the degree of airway inflammation by assessing the percentage of inflammatory cells and concentration of inflammatory mediators in induced sputum, while on the various dietary salt regimens. We have measured a wide spectrum of mediators with different functions representing the major effector and airway-damaging mediators in EIA. Dietary salt loading increased all the induced sputum supernatant mediators that were measured following exercise. Dietary salt loading increased the concentration of LTC4-E4 and LTB4 in induced sputum compared with the LSD following exercise. LTC4-E4 are produced by eosinophils and mast cells, are potent airway smooth muscle contractile agonists, increase mucus production and vascular permeability, and may directly increase eosinophilic airway inflammation (15). LTB4 is a potent chemoattractant and activator of neutrophils, without any significant effect on airway smooth muscle (4). The increase in PGD2 after exercise on the NSD and HSD is highly suggestive of mast cell activation (4). PGD2 has effects on airway smooth muscle similar to those of LTC4-E4, although less potent, and is primarily responsible for neutrophil activation and increasing vascular permeability (4). We chose to investigate mast cell activation using a sputum supernatant PGD2 assay because, unlike alternative measures such as sputum mast cell counts and markers of mast cell degranulation, this is a well-validated technique (23). Our group has previously shown in an animal model of EIA that urinary LTE4 release may be moderated by changes in dietary salt intake (21). However, a major limitation in this study is that urinary LTE4 excretion may not be reflective of airway cell activity, but may in fact be linked to the influence of dietary salt on the kidneys via changes in osmoreceptor activity or the rennin-angiotensin system during exercise, because the kidneys are rich sources of leukotrienes and prostaglandins. Given that leukotrienes and prostaglandins may not only cause airway smooth muscle contraction but also cause dilatation and leakage of the mucosal and submucosal capillary beds, resulting in airway edema (16), it is possible that variations in salt intake might lead to changes in airway narrowing following exercise in asthmatics by means of both mechanisms.
Dietary salt loading increased and dietary salt restriction reduced the levels of induced sputum supernatant IL-1β and IL-8 concentration following exercise. Both of these cytokines are associated with neutrophilic inflammation (10). In addition, hyperosmolarity has been shown to stimulate IL-8 production in human bronchial epithelial cells in vitro (10). Since the human airway mucosa is a semipermeable membrane across which osmotic equilibration occurs, this may represent a potential pathway by which dietary salt loading increases airway osmolarity and enhances the release of proinflammatory mediators.
In this study, a higher percentage of eosinophils, neutrophils, and levels of ECP were found in induced sputum following exercise on the NSD and HSD compared with the LSD. There was a significant correlation between the degree of eosinophilic and neutrophilic infiltration and activation in the asthmatic airways following exercise and bronchoconstrictor responses, such as the severity of EIA and the AUC of the percentage fall in FEV1 from preexercise values. In addition, there was a significant correlation between ECP (a marker of eosinophil activation) and the severity of EIA on the NSD and HSD. This study supports previous studies that eosinophilic airway inflammation is an important determinant of the bronchoconstrictor response to exercise in asthmatics (14,29. The mechanism by which dietary salt loading enhances eosinophilic airway inflammation is unknown. It is possible that an HSD triggers the release of leukotrienes from eosinophils in asthmatic airways, thereby causing airway smooth muscle contraction and microvascular leakage (29), and/or activates the release of leukotrienes from another source causing eosinophil migration to the airways. We observed sputum neutrophils at 1, 6, and 24 h after exercise on the NSD and HSD, and at 1 and 6 h postexercise on the LSD. The increase in percentage sputum neutrophils at 6 and 24 h postexercise is likely a result of the sputum induction procedure, as repeated induction of sputum has been shown to cause elevations in the percentage of sputum neutrophils (12). The decrease in the percentage of sputum macrophages following exercise, on all diets, may be a direct result of the observed neutrophils (22).
In conclusion, this study has shown for the first time that dietary salt restriction lessens and dietary salt loading enhances airway inflammation in asthmatics following exercise. In addition, this study has demonstrated for the first time that modifying dietary salt intake alters postexercise DLCO and its subdivisions in asthmatic subjects. Our findings indicate that small salt-dependent changes in vascular volume and microvascular pressure might have substantial effects on airway function following exercise in the face of mediator-induced increased vascular permeability.