Over the past decade, there has been considerable interest in the link between upper respiratory tract infections (URTI) and immune suppression in elite athletes. Although it has been reported that highly conditioned athletes have a higher incidence of URTI than control groups during and after events or training (2,4,15,32), the link with immune suppression has not been clearly established (28,33). Review of the literature by several groups (16,17,19,21,29-31) has suggested that the effects of exercise on the immune system are dependent on the level of fitness of the subjects, the degree of intensity of exercise, and the duration of the exercise. Whereas exercise in sedentary or moderately exercising subjects has been shown to boost immune parameters (1,27), in elite athletes the data suggest that immune suppression follows intensive exercise (5,7,12,20,31,37). However, some studies have not supported this view (6,27).
Previous studies of mucosal immunity have focused on the acute changes in salivary immunoglobulin levels after individual sessions of exercise (10,20,22-26,36,37). The consistent finding in these studies was a reduction in salivary IgA levels after strenuous exercise in cross-country skiers (37), cyclists (25,26), swimmers (10,22,36), kayakers (23,24), athletes in a national hockey team (23), and an Olympic squad training for a variety of events (20). Our own studies (10) have extended these findings to show a chronic suppression of salivary IgA over a 7-month training season in elite swimmers. This chronic suppression of salivary IgA has previously been reported in swimmers undertaking a 3-month training program (36) and may explain the lower salivary IgA levels reported in cross-country skiers even before competition (37).
In this study, we have examined the influence of exercise intensity, volume, and duration of training and psychological stress on mucosal humoral immunity in a cohort of elite swimmers undertaking a 7-month training program in preparation for national championships. The aim of the study was to determine whether there was a relationship between salivary IgA concentrations and the incidence of upper respiratory tract infection in athletes.
Subjects. The study was conducted with the informed consent of the Australian Institute of Sport (AIS) Swimming Team, which consisted of 26 elite swimmers (15 male, 11 female) aged 16-24 yr. The athletes undertook 20-25 h·wk−1 of pool training and 5 h of dry-land training involving flexibility, resistance work, and general calisthenics in preparation for the national swimming championships. Twelve AIS staff (7 male, 5 female) aged 19-41 yr, who were involved in regular but moderate exercise programs (up to 4 h·wk−1), were chosen to act as controls. AIS staff were selected as controls to ensure the same exposure to environmental conditions, including seasonal variations in climatic conditions and potential exposure to infections as the athletes resident at the AIS. The study had ethics approval from the Australian Sports Commission, and written informed consent was obtained from the subjects.
Study design. The athletes and controls were studied during routine training sessions at monthly intervals from April to October (during the Southern Hemisphere's fall to spring). The training sessions were selected to allow testing after a rest period of 24 h to exclude potential residual effects of the previous training session. The tests were conducted in the first week of each month to minimize possible hormonal influences in the female subjects. All testing sessions for the swimmers were conducted at 1600 h in an indoor 50-m pool with controlled ambient (24°C) and water (27°C) temperature. Saliva was collected from swimmers and controls before each monthly test session and again immediately after the session and assayed for secretory IgA concentration. A psychological assessment for anxiety/stress was also administered before each of the test training sessions. Daily infection logs were recorded, and episodes of infection were investigated by one of the sports physicians (W.McD.) and included a throat swab for culture (11).
Exercise gradings. The distances swum during each 24-h period were recorded daily by each athlete over the 7-month study period. The exercise gradings and phases of training were recorded by the AIS swim team physiologist (DBP). Training sessions were graded into four levels of physical intensity (34): low-intensity aerobic, anaerobic threshold, maximal aerobic, and maximal lactic. Based on terminology commonly used in international level swimming, the weekly phase of training (microcycle) was classified as either endurance (high volume, low intensity), quality (moderate volume, high intensity), or taper (precompetition work of reducing volume). Figure 1 describes the monthly training program undertaken during this study and the mean distances swum per week during each month. The distances swum, by each athlete, in the studied training session and during the previous week (7 d) and previous month (4 wk) were included for purposes of statistical comparison.
Salivary IgA. One milliliter of unstimulated whole mixed saliva was collected directly into plastic tubes and stored at −20°C until assayed. IgA was measured by electroimmunodiffusion (8) using commercially prepared IgA specific antisera (Tago, Burlingame, CA). The assays were calibrated with Standard Human Serum referenced against WHO 67/95 (Behringwerke, Marburg, Germany). The conversion factor to the new International Reference material CRM 470 for IgA is 0.83. The average between-run CV for the assay was 10% over the study period and within-run CV was 8%.
Psychological assessment. The Spielberger State-Trait Anxiety Inventory STAI Form Y self-evaluation questionnaire was used to determine the psychological anxiety/stress score (35). The age-related reference range for the anxiety stress score was 26-46.
Statistical procedures. Random effects regression models were used to examine the relationship of pretraining IgA, posttraining IgA, and session change in IgA with training volumes, rating of training intensity, psychological stress, age, gender, and number of infections over the training season. This method takes into account the repeated nature the data and is robust when there are missing data. IgA values were log-transformed for the purposes of analysis. Initially simple regressions were performed, then multiple regression models were established using a backward stepwise approach. Second-order effects were also investigated (i.e., interactions and variables to the second power). Where appropriate the mean or median value and 95% confidence interval (CI) has been reported for comparisons. These values have been summarized over the study period.
There were strong correlations between session volume, per week volume, per month volume, training month, and physical intensity. It was noted that physical intensity and training volume were two relatively fixed patterns over the training period according to gender (or more particularly, according to coach). This meant that for any given time point, training volume and physical intensity were constant for each gender and that there was little variation in training volume/physical intensity combinations. It is thus not possible to separate gender and training (coach) effects when interpreting the results from this data.
Two methods were used to assess the ability of the training and immunology variables to predict infection. Logistic regression was used to model the outcome infection, as at least one infection versus no infections. In the first analysis all data were used ignoring the repeated nature of the data, and in the second analysis individual means over the training period were used. Regular linear regression was used to model the total number of infections incurred during the training period. In an attempt to use more of the information in the data, the trends in salivary IgA levels over time (months of training) were summarized by calculating individual slopes of pretraining IgA and posttraining IgA for each swimmer. These slopes were used in linear regression models to predict the number of infections at a given IgA level.
Training, psychological, and infection variables. The median distances swum during test training sessions, weekly and monthly, by the male and female swimmers are presented in Table 1 along with the median Spielberger anxiety scores and number of infections for swimmers and controls. Training volume, physical intensity, and gender were highly correlated and coach dependent, as coaches exclusively trained either male or female swimmers. The Spielberger anxiety score showed little variation over the season for individual athletes or controls. The number of infections per individual were not statistically significantly different between swimmers and control subjects (P = 0.26) and was consistent for both genders.
Ignoring the repeated measures nature of the study design, we found no associations between the concentrations of pretraining salivary IgA in individual swimmers and training volumes measured as session volume (N = 140; ρs = 0.08; ρ = 0.33), previous weekly volume (N = 140; ρs = 0.15; P = 0.08), or previous monthly volume (N = 138; ρs = 0.01; P = 0.90). Similarly, there was no association between the individual stress/anxiety scores and the corresponding months pretraining salivary IgA concentration in either swimmers (N = 122; ρs = −0.005; P = 0.96) or control subjects (N = 53; ρs = 0.06; P = 0.67). There was a significant correlation between pretraining salivary IgA concentrations and the number of infections in both swimmers (N = 143; ρs = −0.22; P < 0.01) and control subjects (N = 67; ρs = −0.32; P < 0.01). The negative correlations indicated that the lower the pretraining salivary IgA concentration the higher the number of infections.
Pre- and post-training salivary IgA. The mean pre- and post-training salivary IgA concentrations for athletes and controls are presented in Table 2. The pretraining salivary IgA levels were significantly lower in female swimmers compared with male swimmers (P = 0.02), and a similar trend was observed for posttraining salivary IgA levels (P = 0.07). There were no statistically significant differences between swimmers and controls for either gender for pretraining (male: P = 0.10; female: P = 0.92) or posttraining (male: P = 0.62; female: P = 0.84) salivary IgA levels.
Trends in salivary IgA over time. The mean pretraining and posttraining salivary IgA concentrations for male and female swimmers and controls (Table 3) fell over the training season (Fig. 2). There was a trend for swimmers to have higher pretraining salivary IgA levels than controls (Table 3), but this was not statistically significant (P = 0.07). The decline in pretraining salivary IgA concentrations over the months of training was significant for both swimmers and controls (P = 0.002). Similar trends were observed for posttraining salivary IgA concentrations.
Associations with variables over time. The fitted regression models indicated that the mean pretraining salivary IgA levels for swimmers were significantly associated with gender, months of training, and the number of infections (Table 4). There were no significant relationships detected for measurement of physical intensity of training, training volume measures, psychological stress, and age. The pretraining salivary IgA levels were on average 12.8 mg·L−1 (17.9%) higher for male swimmers compared with female swimmers. The pretraining salivary IgA levels fell an average by 4.1% for each additional month of training and were 5.8% lower for each additional infection. Infection and months of training were significantly associated with the mean pretraining salivary IgA levels in the control subjects (Table 4). The regression models confirmed that significantly lower pretraining IgA levels were associated with an increase in the number of infections and increasing months of training for both swimmers and controls (Fig. 2).
The fitted regression models for the mean posttraining salivary IgA levels indicated significant associations with gender, months of training, and session volume for the swimmers (Table 4). There were no significant relationships detected for infection rate, physical intensity, weekly or monthly session volumes, psychological stress, and age. The posttraining salivary IgA levels were on average 10.9 mg·L−1 (27.4%) higher in male swimmers compared with female swimmers. The posttraining salivary IgA levels were on average 7.0% lower for each additional month of training and 8.5% lower for each additional 1 km swum during the training session. The mean posttraining salivary IgA levels in control subjects were significantly associated with numbers of infections and months of training (Table 4). The relationships showed significantly lower posttraining IgA levels with increasing months of training for both swimmers and controls (Fig. 2).
Predictors of infection. With infections modeled as the outcome, the mean pretraining salivary IgA concentration over the season for each swimmer was predictive of the number of infections in the elite swimmers (R2adj = 0.24, P = 0.006) (Fig. 3). The initial preseason salivary IgA concentrations for each individual swimmer were also predictive of the number of reported infections during the season (R2adj = 0.12, P = 0.048). The predicted mean number of infections from this model at selected salivary IgA levels is indicated in Table 5. The small sample size and low number of infections in the control subjects made the data unsuitable for regression modelling in the controls.
The trends in salivary IgA concentrations over the 7-month training season were summarized by calculating the individual slopes for pre- and post-training IgA levels for each swimmer. The range of the individual slopes was from −28% to +10% change in IgA concentration per month (Fig. 4). The linear regression model predicted an additional infection for each 10% drop in percent decrease (slope) of pretraining salivary IgA level over time (per month) (R2adj = 0.15, P = 0.031) (Table 5).
This study provides evidence that suppression of mucosal humoral immunity, reflected by low levels of salivary IgA, is associated with an increase in the number of episodes of upper respiratory infection. The lower the levels of salivary IgA, the higher the incidence of infection in both elite swimmers and moderately exercising control subjects. In elite swimmers, the number of infections was predicted by the IgA level when assessed by three different methods: by the preseason salivary IgA levels, the mean pretraining session IgA levels, and the rate of decrease in pretraining IgA levels over the 7-month training season. The results of this study suggest that monitoring the pretraining salivary IgA level over the season may be of benefit to athletes and coaches for the assessing an athletes' risk of infection.
Secretory IgA plays a major role in immune protection at mucosal surfaces by providing specific antibodies in response to pathogens, a transport mechanism for elimination of antigens in the submucosa and an exclusion barrier at the mucosal surface to prevent antigen entry. The lack of nonspecific secretory IgA at mucosal surfaces or an inability to produce specific IgA antibodies can lead to an increased risk of infection, as in the case of some IgA-deficient subjects (14). Recently, it has also been reported that transient absences of salivary IgA during the first year of life are associated with subsequent bronchial hyperreactivity that may have been the result of a respiratory infection at the time of the loss of protection at mucosal surfaces (9). The results of this study in elite swimmers confirm that there is a critical concentration of secretory IgA required at mucosal surfaces to provide adequate protection from respiratory infections. The data suggest that below this threshold, athletes who train over long periods of time at an elite or even a moderate level are susceptible to an increased risk of infection, and the lower the concentration of salivary IgA the higher the risk of infection.
This study, designed to assess the impact of exercising at an intensive level on immune function and susceptibility to infection, has clearly demonstrated an association between the degree of immune suppression and number of respiratory infection. The mean pretraining salivary IgA levels for the athletes over the season were predictive of the number of infections, and the data were used to model the mean salivary IgA concentration associated with a predicted number of infections (Table 2). However, this model would not be useful for predicting an individual athlete's susceptibility to infection, as the range of mean pretraining salivary IgA values for those with no infections spanned almost the entire range of all the mean pretraining IgA values (Fig. 3).
For individual athletes, the preseason salivary IgA concentration was predictive of the number of infections during the season and is a more useful indicator of risk of infection. The regression model found that, on average, lower preseason IgA values were predictive of an increased number of infection episodes during the training season. Though the precision of the model limits its usefulness for individual predictions, it suggests that athletes with preseason salivary IgA values below 40 mg·L−1 may warrant further attention, with either medical or training reviews. This hypothesis awaits verification in a prospective study.
The data also indicate that monitoring the change in salivary IgA concentrations in individuals at regular intervals during a training program and calculating the rate of decrease in IgA levels predicted the number of infections in athletes. This parameter may allow athletes and coaches to modify training programs to minimize immune suppression and potential infection risks at critical periods. Modifying training schedules may be effective, as this study has demonstrated that the concentration of salivary IgA was related to the months of training and training volume in the elite swimmers. The results of this study and previous investigations (10,11,22-26,36,37) suggest that monitoring should occur before the athlete commencing a training session, as individual sessions of intensive exercise result in a decrease in salivary IgA levels, and in this study the posttraining salivary IgA levels were not significantly associated with infection. One-off testing would only be informative if the salivary IgA concentration was already below the threshold the individual athlete for predicting infection risk.
The effect of psychological stress was included in this multivariate analysis, as previous studies have reported a decrease in salivary IgA levels in a variety of situations associated with increased stress (13,18), and for stress to be associated with an increased susceptibility to infections (3). A previous study of exercise training in a swimming squad (36) failed to show any association between the global Profile of Mood States (POMS) score and the significant decreases in salivary IgA after high-intensity training sessions. However, the tension-anxiety subscale did identify a significant association with the changes in salivary IgA (36). In our cohort, we assessed the anxiety level in the athletes and controls using the Spielberger questionnaire (35). In this longitudinal study, the anxiety-stress score showed no associations with changes in salivary IgA levels or infection rates in either the elite swimmers or moderately exercising controls.
In summary, this study has demonstrated a relationship between suppression of mucosal IgA levels and increase in the number of respiratory infections in both elite swimmers and moderately exercising controls. As there were no sedentary controls included in the study design, the possibility of seasonal effects cannot be excluded. Further studies are required to determine whether there is a true exercise effect and whether the observed changes in mucosal immunity and increase in respiratory infections are applicable to any exercising population. The monitoring of salivary IgA levels over the training season may identify athletes who are at risk of infection and allow appropriate intervention to prevent illness and subsequent loss of performance.
1. Brahmi, Z., J. E. Thomas, M. Park, P. Melvin, and I. R. G. Dowdeswell. The effect of acute exercise on natural killer-cell activity of trained and sedentary human subjects. J. Clin. Immunol.
2. Brenner, I. K. M., P. N. Shek, and R. J. Shephard. Acute infections and exercise. Sports Med.
3. Cohen, S., D. A. J. Tyrrell, and A. P. Smith. Psychological stress and susceptibility to the common cold. N. Engl. J. Med.
4. Douglas, D. J., and P. G. Hanson. Upper respiratory infections in the conditioned athlete. Med. Sci. Sports Exerc.
5. Eskola, J., O. Ruuskanen, E. Soppi, et al. Effect of sport stress on lymphocyte transformation and antibody formation. Clin. Exp. Immunol.
6. Field, C. J., R. Gougeon, and E. B. Marliss. Circulating mononuclear cell numbers and function during intense exercise and recovery. J. Appl. Physiol.
7. Fry, R. W., A. R. Morton, G. P. M. Crawford, and D. Keast. Cell numbers and in vitro responses of leucocytes and lymphocyte subpopulations following maximal exercise and interval training sessions of different intensities. Eur. J. Appl. Physiol.
8. Gleeson, M., A. W. Cripps, R. L. Clancy, A. J. Husband, M. J. Hensley, and S. P. Leeder. Ontogeny of the secretory immune system in man. Aust. N. Z. J. Med.
9. Gleeson, M., R. L. Clancy, M. J. Hensley, et al. Development of bronchial hyperreactivity following transient absence of salivary IgA. Am. J. Resp. Crit. Care
10. Gleeson, M., W. McDonald, A. Cripps, et al. The effect of intensive training on systemic and mucosal immunity in elite swimmers. Clin. Exp. Immunol.
11. Gleeson, M., W. A. McDonald, A. W. Cripps, et al. Exercise, stress and mucosal immunity in elite swimmers. In: Advances in Mucosal Immunology,
J. Mestecky, M. W. Russell, S. Jackson, S. M. Michalek, J. Tlaskalova-Hogenova, and J. Sterzl (Eds.). New York: Plenum Press, 1995, pp. 571-574.
12. Gmünder, F. K., P. W. Joller, H. I. Joller-Jemelka, et al. Effect of a herbal yeast food supplement and long-distance running on immunological parameters. Br. J. Sports Med.
13. Graham, N. M. H., G. R. Chiron, A. Bartholomeusz, N. Taboonpong, and J. T. La Brooy. Does anxiety reduce the secretion rate of secretory IgA in saliva? Med. J. Aust.
14. Hanson, L. Ä., J. Björkander, and V. A. Oxelius. Selective IgA deficiency. In: Primary and Secondary Immunodeficiency Disorders,
R. K. Chandra (Ed.). Edinburgh: Churchill Livingstone, 1983, pp. 62-64.
15. Heath, G. W., E. S. Ford, T. E. Craven, C. A. Macera, K. L. Jackson, and R. R. Pate. Exercise and the incidence of upper respiratory tract infections. Med. Sci. Sports Exerc.
16. Hickson, R. C., and J. B. Boone. Physical exercise and immunity. In: Stress and Immunity,
N. Plotnikoo, A. Murgo, R. Faith, and J. Wyburn (Eds.). Boca Raton, FL: CRC Press, 1991, pp. 211-234.
17. Hoffman-Goetz, L., and B. K. Pedersen. Exercise and the immune system: a model of the stress response? Immunol. Today
18. Jemmott, J. B., J. Z. Borysenko, M. Borysenko, et al. Academic stress, power motivation and decrease in secretion rate of salivary secretory immunoglobulin A. Lancet
19. Keast, D., K. Cameron, and A. R. Morton. Exercise and the immune response. Sports Med.
20. Levando, V. A., S. Suzdal'nitskii, B. B. Pershin, and M. P. Zykov. Study of secretory and antiviral immunity in sportsmen. Sports Training Med. Rehabil.
21. Mackinnon, L. T. Immunoglobulin, antibody and exercise. Exerc. Immunol. Rev.
22. Mackinnon, L. T., and S. Hooper. Mucosal (secretory) immune system responses to exercise of varying intensity and during overtraining. Int. J. Sports. Med.
23. Mackinnon, L. T., E. Ginn, and G. Seymour. Effects of exercise during sports training and competition on salivary IgA levels. In: Behaviour and Immunity,
A. J. Husband (Ed.). Boca Raton, FL: CRC Press, 1992, pp. 169-177.
24. 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.
25. Mackinnon, L. T., T. W. Chick, A. Van As, and T. B. Tomasi. Decreased secretory immunoglobulins following intense endurance exercise. Sports Training Med. Rehabil.
26. Mackinnon, L. T., T. W. Chick, A. Van As, and T. B. Tomasi. The effect of exercise on secretory and natural immunity. Adv. Exp. Med. Biol.
27. Nehlsen-Cannarella, S. L., D. C. Nieman, A. J. Balk-Lamberton, et al. The effects of moderate exercise training on immune response. Med. Sci. Sports. Exerc.
28. Nieman, D. C. Exercise, infection and immunity. Int. J. Sports Med.
29. Nieman, D. C., and S. L Nehlsen-Cannarella. The effects of acute and chronic exercise on immunoglobulins. Sports Med.
30. Pedersen, B. K., and H. B. Nielsen. Acute exercise and the immune system. In: Exercise Immunology,
B. K. Pedersen (Ed.). Heidelberg: Springer-Verlag, 1997, pp. 5-38.
31. Pedersen, B. K., M. Kappel, M. Klokker, H. B. Nielsen, and N. H. Secher. The immune system during exposure to extreme physiologic conditions. Int. J. Sports. Med.
32. Peters, E. M., and E. D. Bateman. Ultramarathon running and upper respiratory tract infections: an epidemiological survey. S. Afr. Med. J.
33. Peters-Futre, E. M. Vitamin C, neutrophil function, and upper respiratory tract infection risk in distance runners: the missing link. Exerc. Immunol. Rev.
34. Pyne, D. B., and R. D. Telford. Classification of swimming
training sessions by blood lactate and heart rate responses. Excel
35. Spielberger, C. D. Manual for State-Trait Anxiety Inventory.
Palo Alto, CA: Consulting Psychologists Press Inc., 1983.
36. Tharp, G. D., and M. W. Barnes. Reduction of saliva immunoglobulin levels by swim training. Eur. J. Appl. Physiol.
37. Tomasi, T. B., F. B. Trudeau, D. Czerwinski, and S. Erredge. Immune parameters in athletes before and after strenuous exercise. J. Clin. Immunol.