In the last decade, there has been a substantial increase in research dedicated to determining the relationship between exercise and immune function, particularly upper respiratory tract infection (URTI). An early finding by Mackinnon and colleagues of an inverse relationship between secretory immunoglobulin A (s-IgA) and URTI (17) led to the prevailing belief that s-IgA is the immune variable most closely associated with URTI (10,19). s-IgA has the capacity to inhibit the colonization of pathogens, (4) bind antigens for transport across the epithelial barrier, and neutralize viruses (3,15), thus representing one of the body’s first lines of defense against infections related to URTI.
In spite of the amount of research being done in this area, it is difficult to make comparisons across studies. Recently, researchers have noted that the discrepant results in the findings related to s-IgA may be due to the various means used in expressing s-IgA results. In most of the earlier studies, results were expressed simply as the concentration of s-IgA (11,25,28). Later researchers attempted to account for the amount of s-IgA available on the mucosal surfaces by expressing either the secretion rate of s-IgA (6,9,14,18,20–24,27) or the ratio of s-IgA to protein (8,11,18,21,30). In the last 5 yr, other researchers have also suggested that IgA relative to osmolality is the preferred method for determining the impact of exercise on mucosal IgA (2,31).
All of the above expressions represent attempts to account for the amount of s-IgA actually available on the mucosal surfaces, mostly during measures of prepost exercise interventions. To date, no attempt has been made to account for the amount of s-IgA available on mucosal surfaces in longitudinal studies of athletes, and thus the effects of a season of competitive athletic training on mucosal immune variables other than the absolute concentration of s-IgA remains unknown. Although several studies observed a decrease in mucosal s-IgA throughout a training season, none of these studies have been able to link this decrease in s-IgA to incidence of URTI (10,11,26,29). Recent studies are attempting to make this link. Gleeson and colleagues (12) conducted a study on 26 elite swimmers and 12 controls over the course of a 7-month training period. Using fitted regression models, they found significant relationships between pretraining s-IgA levels and gender, months of training, and number of infections. In addition, Klentrou and colleagues (16) used Pearson correlation analysis and found a significant relationship between both the concentration of s-IgA and ratio of s-IgA to albumin and the number of days subjects reported both sickness and influenza symptoms but not to the number of days that subjects reported cold symptoms during 12 wk of moderate exercise in young adults. Neither of these studies examined saliva flow rate, the secretion rate of s-IgA, salivary protein, osmolality, or the ratios of s-IgA:protein or s-IgA:osmolality, and thus their role in URTI is undetermined.
Therefore, the purpose of this study was to determine (a) the effects of a season of competitive training for American football athletes on various mucosal immune variables, and (b) which standard s-IgA measure, either alone or in combination, serves best to predict the incidence of URTI.
METHODS
Subjects.
The subject pool consisted of 100 males. Seventy-five subjects (FB) were members of a large midwestern university football team (20.5 ± 1.5 yr, 96.5 ± 18.6 kg, and 1.7 ± 0.02 m). Twenty-five additional male subjects were used as controls (C) (20.5 ± 1.6 yr, 79.5 ± 12.2 kg, and 1.7 ± 0.03 m). The controls were nonvarsity university students who reported being physically active (defined as performing some type of exercise that induces sweat for at least 30 min 3× wk−1) and being full-time (a minimum of 12 semester hours credit) students. To determine activity levels, FB and C completed the Baecke physical activity questionnaire (1). All subjects were given a complete explanation of the research project procedures, and each provided written informed consent before being allowed to participate. The institutional review board at the university approved the study procedures.
Training protocol.
The subjects for this study were members of an American university football team; the controls were 25 nonvarsity athletes. Data collection took place across a time period of 12 months; thus, all seasons and types of training were incorporated into the study. The sample collection period and the level of training for each part of the season are as follows:
1. Baseline values as subjects reported to training camp in August for the NCAA official beginning of fall season.
2. Data point in mid-September. This data point represents immune variables present after the team has completed 6 wk of the most extensive, intensive training they undertake during the fall season.
3. Data point in early December. This data point represents immune variables present while the team is “in season.” At this point, they have completed 10 wk of intercollegiate competition. Practices consist of drills and strategies designed for competition on the weekend. No conditioning or strenuous training occurs during the season.
4. Data point in early February. This data point represents the athlete’s return to normal activity levels. There is no requirement for training or practice during this time period. The sample was taken after a 6-wk period that is officially scheduled as a time of rest by the university.
5. Data point in mid-February. This is very similar to data point 4, but was taken because it officially represents the beginning of spring football season at the university, and thus is used as the PRE sample for the spring season. During the spring football season, players compete against each other for the opportunity to earn a starting spot and/or a scholarship in the fall.
6. Data point in late March. This data point represents immune variables present after the team has completed 6 wk of the most extensive, intensive training they undertake during the spring season. It is identical to the training for data point 2.
7. Data point in early June. This data point represents immune variables present while the team is “in season.” At this point, they have completed 10 wk of intersquad competition. It is similar to data point 3.
8. Data point in mid-July. This data represents immune variables present 6 wk after the conclusion of the season that is officially scheduled as a period of rest.
Signs and symptoms.
Subjects were required to complete a weekly log in which they documented any sign or symptom consistent with URTI (including cough, runny nose, and nasal congestion) as well as the number of days that symptoms occurred. They were instructed to code the severity of the infection from 1 to 3, where 1 indicated that the infection had no impact on their daily activity, 2 indicated that the infection had some impact on their daily activity, and 3 indicated that the infection lead to a significant decrease in daily activities. The coach collected the logs weekly from FB and the PI collected them from the controls. The PI checked them immediately to make sure the URTI were being classified correctly and to make referrals to the team physician as soon as possible after incident reporting.
The entire following criterion had to be present in order for the infection to be classified as an URTI: the infection lasted three or more days, and all three symptoms of cough, runny nose, and nasal congestion were present for the entire three days. If the subject met the previous criteria, their logs were tagged for further consultation. They were also tagged if the infection impacted daily activities. All subjects whose logs were tagged were subsequently interviewed by the principal investigator and questioned regarding their history of allergies, other symptoms including fever, aches, and malaise, and how the illness lead to a decrease in their ability to follow their exercise plan and maintain normal daily activities. If the subject reported no allergies or allergy symptoms different than their current illness, no impact on daily activities, and no additional symptoms, the infection was coded as a URTI. If the subject reported other symptoms or a decrease in daily activity, they were referred to a physician who then made the diagnosis of URTI or “other illness.” If the PI was uncertain as to the nature of the illness, the physician made the diagnosis. This method of classifying URTI is consistent with studies similar to the current one (5,12,13,26). Other illnesses that subjects encountered throughout the year included mononucleosis and influenza.
Saliva collection.
All saliva samples were collected between 1200 and 1400 in the afternoon. Subjects reported at the same time for each collection period after fasting for 2 h and refraining from any strenuous physical activity for 20 h. After thoroughly rinsing their mouths with water, unstimulated saliva samples were collected for 4 min into 15-mL polypropylene tubes. Saliva was measured for volume and then stored at −70°C until analysis.
Saliva analysis.
Immunoglobulin concentrations were measured for s-IgA by enzyme-linked immunosorbent assay (ELISA) as described in Mackinnon et al. (21). Briefly, microplates were precoated with unconjugated goat antihuman IgA (Sigma) diluted 1:2000 in a carbonate buffer and left refrigerated overnight at 4°C. After being brought to room temperature (RT), they were washed with phosphate-buffered saline (PBS)-T (Sigma), and the plates were blocked with 1% bovine serum albumin (BSA) (Sigma) PBS-T and incubated at RT for 30 min. Saliva was spun at 10,000 rpm for 5 min, diluted 1:1500 with 0.1% BSA PBS-T, added to the plates, and incubated at RT for 60 min. The plates were washed with PBS-T after which biotin conjugated antihuman IgA diluted 1:2000 with 0.1% BSA PBS-T was added to each well and allowed to incubate at RT for 60 min. Horseradish streptavidin peroxidase (HRP-SA) (Sigma) diluted with 0.1% BSA PBS-T was added to each well and allowed to incubate at RT for 30 min. The plates were washed, PBS and K blue substrate were added, and the reaction was immediately stopped with 3% HCL. The plates were read on a microplate reader (Spectra Maxplus, Molecular Devices Corp., Sunnydale, CA) at 450 nm. Standards of known concentrations of purified IgA were included on each microplate, and absolute concentrations (μg·mL−1) were calculated from the standard curve. To avoid interassay variability, all samples from each subject were assayed on the same microplate. The coefficient of variation for IgA was 10.4%. Total salivary protein was measured using a kit based on Bradford’s standard assay (Bradford; Bio-rad Total Protein Assay Kit, Bio-rad Laboratories, Hercules, CA). The coefficient of variation for protein was 2.3%. Osmolality was measured using a freezing point depression osmometer (Advanced DigiMatic Osmometer 3DII, Advanced Instruments, Needham Heights, MA). The coefficient of variation for osmolality was 1.6%.
Calculations.
The concentrations of s-IgA were expressed in three ways (19):
1. The secretion rate of s-IgA (μg·min−1) or the total amount of s-IgA appearing on the mucosal surface per unit time. s-IgA secretion rate was calculated by multiplying absolute s-IgA concentration (μg·mL−1) times saliva flow rate (mL·min−1); this latter value was calculated by dividing the total volume of saliva obtained in each sample (mL) by the time taken to produce each sample (4 min).
2. The concentration of s-IgA relative to protein (μg·mg protein−1).
3. The concentration of s-IgA relative to osmolality (mg·mOsmol−1).
Statistical analysis.
The dependent variables of salivary s-IgA, saliva flow rate, the secretion rate of salivary s-IgA, protein, osmolality, IgA:protein, IgA: osmolality, percent of URTI, and duration of URTI were analyzed separately using a two groups (FB vs C) × eight times (time 1 to time 8) ANOVA with repeated measures on the last factor. Due to the number of analyses, a Bonferroni correction factor was applied. All statistical decisions were based on α = 0.007, and follow-up analyses on main effects were performed using Bonferroni’s post hoc procedure. The Baecke questionnaire for the measurement of physical activity was analyzed using independent samples t-test. The total score, work score, leisure score, and sport score were analyzed separately.
For FB, eight separate stepwise multiple regression analyses were conducted to predict the dependent variable “number of colds” by the independent variables, s-IgA, saliva flow rate, secretion rate of s-IgA, protein, s-IgA:protein, osmolality, and s-IgA:osmolality at each data collection point. Significance was set at P ≤ 0.05. The statistical package used to run all analysis was SPSS (version 11.5), Chicago, IL.
RESULTS
The sample population consisted of 75 male football players (FB) (20.5 ± 1.5 yr, 96.5 ± 18.6 kg, and 1.7 ± 0.02 m) and 25 additional male controls (C) (20.5 ± 1.6 yr, 79.5 ± 12.2 kg, and 1.7 ± 0.03 m).
Analysis of s-IgA data revealed main effects for both group, F(1,99) = 11.71, P = 0.001, and time factors F(1,99) = 4.288, P = 0.005, as well as a significant group × time interaction F(1,99) = P = 0.004. Post hoc analysis revealed the group main effect was the result of significantly lower s-IgA values in FB compared with C at time points 2, 3, 6, and 7. Post hoc analysis of the time factor revealed decreased s-IgA for FB at time points 2 and 3 compared with time points 1 and 4, as well as 6 and 7 compared with 5 and 8. A simple main effects analysis of the interaction revealed that groups differed with respect to values at time points 2, 3, 6, and 7 (see Fig. 1).
Analysis of saliva flow data revealed no main effect of group F(1,99) = 1.485, P = 0.226, time F(1,99) = 695, P = 0.627, or group × time interaction F(1,99) = 1.519, P = 0.182 (Tables 1 and 2).
;)
TABLE 1: Saliva analysis for variables with no significant differences during the fall season in college football players (mean ± SEM).
TABLE 2: Saliva analysis for variables with no significant differences during the spring season in college football players (mean ± SEM).
Analysis of the secretion rate of s-IgA data revealed main effects for both group, F(1,99) = 6.208, P = 0.000, and time factors F(1,99) = 4.525, P = 0.001, as well as a significant group × time interaction F(1,99) = 4.451, P = 0.004. Post hoc analysis revealed the group main effect was the result of significantly lower secretion rate of s-IgA values in FB compared with C at time points 2, 3, 6, and 7. Post hoc analysis of the time factor revealed decreased secretion rate of s-IgA for FB at time points 2 and 3 compared with time points 1 and 4 as well as 6 and 7 compared with 5 and 8. A simple main effects analysis of the interaction revealed that groups differed with respect to values at time points 2, 3, 6, and 7 (see Fig. 2).
Analysis of protein data revealed no main effect of group F(1,99) = 0.025, P = 0.875, time F(1,99) = 1.332, P = 0.251, or group × time interaction F(1,99) = 2.150, P = 0.146 (Tables 1 and 2). Analysis of osmolality data revealed no main effect of group F(1,99) = 2.421, P = 0.035, time F(1,99) = 2.150, P = 0.146, or group × time interaction F(1,99) = 1.141, P = 0.99 (Tables 1 and 2). Analysis of IgA:Protein data revealed no main effect of group F(1,99) = 1.974, P = 0.163, time F(1,99) = 0.826, P = 0.532, or group × time interaction F(1,99) = 1.367, P = 0.254 (Tables 1 and 2). Analysis of IgA:osmolality data revealed no main effect of group F(1,99) = 1.251, P = 0.203, time F(1,99) = 0.559, P = 0.678, or group × time interaction F(1,99) = 1.529, P = 0.194 (Tables 1 and 2).
Analysis of the duration of URTI revealed no main effect of group F(1,99) = 0.251, P = 0.323, time F(1,99) = 0.629, P = 0.572, or group × time interaction F(1,99) = 1.031, P = 0.224 (Tables 1 and 2). Analysis of percent of URTI revealed main effects for both group, F(1,99) = 7.911, P = 0.001, and time factors F(1,99) = 9.624, P = 0.000, as well as a significant group × time interaction F(1,99) = 8.768, P = 0.000. Post hoc analysis revealed the group main effect was the result of significantly higher percentage of colds in FB compared with C at time points 2, 3, 6, and 7. Post hoc analysis of the time factor revealed increased percentage of colds for FB at time points 2 and 3 compared with time points 1 and 4, as well as 6 and 7 compared with 5 and 8. A simple main effects analysis of the interaction revealed that groups differed with respect to percentage of colds at time points 2, 3, 6, and 7. (Table 3).
TABLE 3: Percent of FB and C with URTI at each time point.
Analysis of the Baecke physical activity questionnaire revealed significant difference between FB and C in the total score, t = 4.4, P = 0.003, and the measure of sport, t = 6.4, P = 0.001. There were no significant differences in the measure of work, t = 0.6, P = 0.782, or leisure, t = 0.6, P = 0.642 (Table 4).
TABLE 4: Baecke physical activity questionnaire (mean ± SEM).
The regression equation to determine which salivary immune measures are most likely linked to URTI was conducted on FB only. Predictors that contributed no unique variance to the model were excluded. For the number of colds, only the secretion rate of IgA made a significant contribution to the variance. The amount of the contribution varied from 12 to 42% (Table 5). The percent of the team experiencing URTI varied between 3 and 54 (Table 3). The secretion rate of s-IgA accounted for most of the variance when a greater percentage of the team was experiencing URTI.
TABLE 5: Independent variable significantly associated with number of URTI.
Each regression equation generated a unique prediction equation, which was then used to predict the incidence of URTI. Predicted URTI were compared with actual incidence of URTI. At all time points, the model slightly over predicted the number of athletes that should catch an URTI (Table 6). A comparison between the model’s prediction and the athletes URTI report revealed that although the model was accurate at predicting which athletes would not catch an URTI, it overestimated the number of athletes who would catch one by a minimum of four and a maximum of eight at each time point.
TABLE 6: URTI vs predicted URTI.
DISCUSSION
The most important finding of this research is that the secretion rate of s-IgA is a unique predictor of URTI throughout the course of an American competitive football season of 12 months’ duration. The response patterns were the same whether subjects were in the fall or spring training period implying that regardless of the season, the harder the training schedule, the greater the reductions in s-IgA and incidence of URTI. Because the secretion rate of s-IgA represents the actual amount of s-IgA available on the mucosal surfaces for protection against pathogens, (19) it makes intuitive sense that it would be linked to URTI, but until now there has been no research that demonstrated the clear link.
In addition, prediction equations generated from the regression model used in this study were accurate at predicting URTI in the majority of cases. That a small percentage of athletes did not catch an URTI when predicted from the model is understandable. Given the number of variables that contribute to infection, such as virulence and pathogenicity of the virus, exposure to the virus, and the ability of the individual to mount an immune response to counter the rhinovirus, (7) it is understandable that a small percentage of athletes did not catch URTI when predicted. Because the football players were in constant contact with each other, they had an increased opportunity to be exposed to viral infections that other team members were experiencing. This might explain the high percentage of URTI experienced at different points in the season. Other studies report a similar percentage among athletes competing on the same team. Pyne et al. (26) report infection rates higher than 50% and Bury et al. (5) report rates as high as 80%. The results of this study seem to indicate that the athlete is at an increased risk for infection when the secretion rate of s-IgA drops below 40 μg·min−1. Gleeson and colleagues (13) found 40 mg·mL−1 of IgA is associated with increased URTI through out a training season, but without a secretion rate, the two values cannot be compared.
A second major finding of this study is that the only mucosal immune variables that responded to a season of American football training are s-IgA and the secretion rate of s-IgA, both decreasing during periods of intense conditioning and competition. This indicates that measures of salivary protein and osmolality that further add to the time and cost of evaluation are superfluous. To date, the work done in this area is inconclusive, but there is some evidence that the mucosal s-IgA response to training is intensity related. There is a rich body of literature that demonstrates intensity-related postexercise decreases in s-IgA (8,17,20,21,23,30). However, the relationship between s-IgA and training is not as clear in the longitudinal literature. Three studies report an increase s-IgA over the course of a training season. Klentrou and colleagues (16) report an increase in s-IgA after 12 wk of moderate aerobic training in previously sedentary young adults, Fahlman and colleagues (9) report an increase in s-IgA levels after 16 wk of moderate intensity training in elderly subjects, and Gleeson and colleagues (12) report an increase in s-IgA after a 12-wk training period in elite swimmers. All of these training periods were of moderate intensity and relatively short duration. Conversely, intense training and training of a longer duration seems to have a suppressive effect on the mucosal s-IgA response. Tharp and Barnes (29) report a decrease in s-IgA across a 3-month training period in swimmers, and Gleeson and colleagues (11) report a downward trend in s-IgA levels across a 7-month training period in elite swimmers.
The findings of the current study are most in line with a previous study by Gleeson and colleagues (13). They examined a cohort of elite swimmers across a 7-month period and found significantly decreased s-IgA over the training season. Additionally, using fitted regression models, they found significant relationships between pretraining s-IgA levels and gender, months of training, and number of infections. Unfortunately, they did not examine the secretion rate of s-IgA or enter it into their prediction equation so a comparison of the two studies can only be made with respect to s-IgA. Nonetheless, both of these studies add support for the theory that decreased s-IgA is associated with URTI, and the combination of the current study with the previous contributes to the strength of the argument in favor of monitoring s-IgA and the secretion rate of s-IgA of athletes as potential indicators of impending URTI.
One limitation of this study is that it is not possible to quantify the intensity of the training sessions or the intensity of the actual competition. Therefore, it cannot be definitively stated that the suppression of both s-IgA and the secretion rate of s-IgA is due to a high intensity of training as is evident in the longitudinal studies of swimmers and the prepost acute exercise intervention literature. However, because the levels of both variables decreased during training and competition regardless of the time of year, and because the levels did not change for the controls, it is likely that the decrease was related to the football training and competition itself.
Additionally, it is possible that the high rates of infection experienced by the football players is the result of their increased exposure to other ill athletes, a unique, unidentified stress posed by being a varsity athlete or some other unidentified factor related to varsity participation. Finally, although every attempt was made to correctly classify URTI, viral load was not measured, and the possibility exists that some infections may have been misclassified.
In summary, the present study demonstrates that although s-IgA decreases across a training period in American football players, the immune variable most linked with URTI in these athletes over the course of a training season is the secretion rate of s-IgA. The decrease was related to training and competition, and because levels of s-IgA and the secretion rate of s-IgA and URTI returned to normal during periods of rest, athletes who are able to monitor these values throughout the season may be able to stem off infection by decreasing exercise when the secretion rate of s-IgA decreases below 40 μg·min−1. Finally, the present study provides evidence that from a methodological viewpoint, monitoring s-IgA and the secretion rate of s-IgA is adequate without the need to perform additional salivary protein or osmolality analysis.
REFERENCES
1. Baecke, J. A. H., J. Burema, and J. E. R. Frijters. A short questionnaire for the measurement of habitual physical activity in epidemiological studies.
Am. J. Clin. Nutr. 36:936–942, 1982.
2. Blannin, A. K., P. J. Robson, N. P. Walsh, A. M. Clark, L. Glennon, and M. Gleeson. The effect of exercising to exhaustion at different intensities on saliva immunoglobulin A, protein and electrolyte secretion.
Int. J. Sports Med. 19:547–552, 1998.
3. Brandtzaeg, P., S. Baekkevold, I. N. Farstad, et al. Regional specialization in the mucosal immune system: what happens on the microcompartments?
Immunol. Today 20:141–151, 1999.
4. Brandtzaeg, P., F. L. Jahnsen, and I. N. Farstad. Immune functions and immunopathology of the mucosa of the upper respiratory pathways.
Acta Otolaryngol. 116:149–159, 1996.
5. Bury, T., R. Marechal, P. Mahieu, and F. Pirnay. Immunological status of competitive football players during the training season.
Int. J. Sports Med. 19:364–368, 1998.
6. Carroll, K. K., M. G. Flynn, P. G. Brolinson, et al. Natural cell-mediated cytotoxicity and lymphocyte proliferation following a 24 hour road race.
Med. Sci. Sports Exerc. 26:S34, 1994.
7. Duncan, D. F.
Epidemiology. New York: MacMillan, 1988, pp. 17–39.
8. Fahlman, M. M., H. J. Engels, A. L. Morgan, and I. Kolokouri. Mucosal IgA response to repeated Wingate tests in females.
Int. J. Sports Med. 22:127–131, 2001.
9. Fahlman, M. M., A. L. Morgan, N. McNevin, D. J. Boardley, and R. Topp. The effects of different training modalities on s-IgA levels and upper respiratory tract infections in the elderly.
J. Aging Physical Activity (in press).
10. Gleeson, M. Mucosal immunity and respiratory
illness in
athletes.
Int. J. Sports Med. 21:s33–s43, 2000.
11. Gleeson, M., W. A. McDonald, A. W. Cripps, D. B. Pyne, R. L. Clancy, and P. A. Fricker. The effect on immunity of long-term intensive training in elite swimmers.
Clin. Exp. Immunol. 102:210–216, 1995.
12. Gleeson, M., W. A. McDonald, D. B. Pyne, et al. Immune status and respiratory
illness for elite swimmers during a 12 week training cycle.
Int. J. Sports Med. 21:302–307, 2000.
13. Gleeson, M., W. A. McDonald, D. B. Pyne, et al. Salivary IgA levels and
infection risk in elite swimmers.
Med. Sci. Sports Exerc. 31:67–73, 1999.
14. Housh, T. J., G. O. Johnson, D. J. Housh, S. L. Evans, and G. T. Tharp. The effect of exercise at various temperatures on salivary levels of immunoglobulin A.
Int. J. Sports Med. 12:498–500, 1991.
15. Kaetzel, C. S., J. K. Robinson, K. R. Chintalacharuvu, J. P. Vaerman, and M. E. Lamm. The polymeric immunoglobulin receptor (secretory component) mediates transport of immune complexes across epithelial cells: a local defense function for IgA.
Proc. Natl. Acad. Sci. 88:8796–8800, 1991.
16. Klentrou, P., T. Cieslak, M. MacNeil, A. Vintinner, and M. Olyley. Effect of moderate exercise on salivary immunoglobulin A and
infection risk in humans.
Eur. J. Appl Physiol. 87:153–158, 2002.
17. Mackinnon, L. T., T. W. Chick, A. V. As, and T. B. Tomasi. The effect of exercise on secretory and natural immunity.
Adv. Exp. Med. Biol. 216A:869–876, 1987.
18. Mackinnon, L. T., E. Ginn, and G. J. Seymour. Decreased salivary immunoglobulin A secretion rate after intense interval exercise in elite kayakers.
Eur. J. Appl. Physiol. 67:180–184, 1993.
19. Mackinnon, L. T., E. M. Ginn, and G. J. Seymour. Temporal relationship between decreased salivary IgA and upper respiratory tract
infection in elite
athletes.
Aust. J. Sci. Med. Sport 25:94–99, 1993.
20. Mackinnon, L. T., and S. Hooper. Mucosal (secretory) immune system responses to exercise of varying intensity and during overtraining.
Int. J. Sports Med. 15:S179–S183, 1994.
21. Mackinnon, L. T., and D. G. Jenkins. Decreased salivary immunoglobulins after intense interval exercise before and after training.
Med. Sci. Sports Exerc. 25:678–683, 1992.
22. McDowell, S. L., K. Chaloa, T. J. Housh, G. D. Tharp, and G. O. Johnson. The effect of exercise intensity and duration on salivary immunoglobulin A.
Eur. J. Appl. Physiol. 63:108–111, 1991.
23. McDowell, S. L., R. A. Hughes, R. J. Hughes, D. J. Housh, T. J. Housh, and G. O. Johnson. The effects of exhaustive exercise on salivary immunoglobulin A.
J. Sports Med. Phys. Fitness 32:412–415, 1992.
24. McDowell, S. L., J. P. Weir, J. M. Eckerson, L. L. Wagner, T. J. Housh, and G. O. Johnson. A preliminary investigation of the effect of weight training on salivary immunoglobulin A.
Res. Q. Exerc. Sport 64:348–351, 1993.
25. Pyne, D. B., and M. Gleeson. Effects of intensive exercise training on immunity in
athletes.
Int. J. Sports Med. 19:s183–s194, 1998.
26. Pyne, D. B., W. A. McDonald, M. Gleeson, A. Flanagan, R. L. Clancy, and P. A. Fricker. Mucosal immunity, respiratory
illness, and competitive performance in elite swimmers.
Med. Sci. Sports Exerc. 33:348–353, 2000.
27. Schouten, W. J., R. Verschuur, and H. C. G. Kemper. Habitual physical activity, strenuous exercise, and salivary immunoglobulin A levels in young adults: the Amsterdam growth and health study.
Int. J. Sports Med. 9:289–293, 1988.
28. Tharp, G. D. Basketball exercise and secretory immunoglobulin A.
Eur. J. Appl. Physiol. 63:312–314, 1991.
29. Tharp, G. D., and M. W. Barnes. Reduction of saliva immunoglobulin levels by swim training.
Eur. J. Appl. Physiol. 60:61–64, 1990.
30. Tomasi, T. B., F. B. Trudeau, and D. Czerwinski. Immune parameters in
athletes before and after strenuous exercise.
J. Clin. Immunol. 2:173–178, 1982.
31. Walsh, N. P., A. K. Blannin, A. M. Clark, L. Cook, P. J. Robson, and M. Gleeson. The effects of high-intensity intermittent exercise on saliva IgA, total protein and alpha-amylase.
J. Sports Sci. 17:129–134, 1999.
