Training-Related Risk of Common Illnesses in Elite Swimmers over a 4-yr Period : Medicine & Science in Sports & Exercise

Journal Logo


Training-Related Risk of Common Illnesses in Elite Swimmers over a 4-yr Period


Author Information
Medicine & Science in Sports & Exercise: April 2015 - Volume 47 - Issue 4 - p 698-707
doi: 10.1249/MSS.0000000000000461
  • Free


High-level sport requires many years of daily training and competition, often with high to very high intensities and duration. High-level swimmers in particular compete 10 to 30 times per year and generally train 3 to 6 h·d−1, typically with two in-water sessions and one dryland session of resistance, plyometric or flexibility training (1).

The training and competition schedule of national and international swimmers is often undertaken in cold winter or hot and humid summer conditions or at altitude. These environments, occasionally associated with high concentrations of pathogens, as well as heavy travel schedules and changes in time zone, can disrupt daily living routines (e.g., sleep and eating habits) (18,22,31). Many athletes and swimmers also juggle educational, professional, and personal obligations, which only add to the physiological and psychological stress (21,31,40). Despite the practical interest in these issues, there are very few studies (7,8,18) that have systematically described the relations between training loads and clinical outcomes in highly trained swimmers.

The research community for the most part acknowledges that the cumulative effects of repeated, intense physiological, and psychological demands can disrupt immune system function (24,38,40). For example, grueling exercise or training sessions (e.g., >90% of the individual anaerobic threshold for >30 min or competitive events lasting >1 h 30 min) can induce muscle metabolic imbalances (24,28,39). Other outcomes can include a systemic physiological stress response characterized by increases in sympathetic adrenergic neuroendocrine and hypothalamic pituitary adrenal regulation (5) and elevated pro- and anti-inflammatory cytokines, reactive oxygen species, and other stress-related factors (18,28,30). These physiological responses to intensive and/or prolonged exercise are likely to reflect temporary suppression of immune function that last up to 72 h (10,11,18).

When several training sessions per day and/or competitive events are repeated over many days, the accumulation and interaction of physiological and psychological stresses intensify the long-lasting and acute effects of each exercise (11,12,14). This pattern of training or competition can amplify the duration of immune disturbances, especially when the recovery time between sessions is short (12,22,33). In contrast, low-intensity training and recovery periods such as tapering have been described as restoring and/or enhancing immune function (24,40).

Although most studies have shown that athletes are not immune deficient during the most intensive training periods, it is likely that the combined effect of slight declines in several functions and immune parameters is enough to compromise resistance to common minor illnesses such as upper respiratory tract infections (URTI) (12,22,40). However, although the findings from studies on immune responses to intensive and long-term training are consistent, the clinical consequences of prolonged intensive training in terms of increased risk of infectious pathology have yielded discrepant results (11,16,30).

Indeed, several researchers have shown an increase in the incidence of illness during intensive training phases (13,14,16) and around competition periods (25,29), whereas others have observed no systematic relation (8,30,35).

Several methodological limitations may explain these discrepancies, including observation periods that differ in length, differences in the seasonality of surveillance periods, categorization of athletic level, small sample sizes, and not taking into account intersubject effects (e.g., individual load thresholds above which illness risk increases) (7,8). However, none of these studies has examined the effects of multiple stressors (different types of loads and competitions, seasonality, past medical incidents, and individual susceptibility) using a multivariate approach. An early study of these relations in distance running was conducted over 20 yr ago (20), but there has been little work since. Very few studies have used physician-verified URTI in the study design, and most have instead used self-reports of a limited number of common symptoms and signs, which may have overestimated the incidence of URTI in elite athletes.

The aim of this 4-yr study was to investigate the short-term association of training variables, reported using different approaches (training loads, or training periods) and adjusted for potential confounders (sex, seasonality, performance level, or past medical incidents), with the incidence of physician-verified symptoms of upper respiratory tract and pulmonary illness, muscular affections (MA), and all-type pathologies (AP) in elite swimmers.



A single-group, prospective, longitudinal study of 28 elite swimmers (14 males and 14 females; 8 international and 20 national) age 16–30 yr was conducted over four training and competitive seasons. The study was conducted in two Olympic preparation centers. Both centers had a medical service located 500 m from the pool, and the same two physicians followed the swimmers during this study. Out of the initial population of 44 swimmers, only those volunteers who met the inclusion criteria were selected for the study. The minimum age was 16 yr for women and 18 yr for men. Swimmers had to have been selected in the previous year for the elite National Championships. The minimum number of training sessions per week was nine (including dryland conditioning), and the swimmers had to have demonstrated high motivation in the past 6 months. Swimmers were excluded if they had a chronic pathology that affected their immune status or if they had a debilitating acute or chronic pathology that required medical treatment and suspension of training for at least 3 wk in the 2 months preceding the start of the study. Swimmers were not eligible if they were taking medication known to affect immune function or inflammation, nor could they smoke cigarettes or consume alcohol immoderately. Swimmers who underwent changes concerning the inclusion criteria, such as changes in training location, serious disease, or unexpected personal or academic imperatives, were excluded from the study. In these cases, only the data before exclusion were considered. In the 6 months before the study and throughout the study period, the athletes’ food intake, nutritional supplementation, sleep quality, and psychological status were monitored by physicians, nutritionists, and psychologists, minimizing the risk of these factors having a potential influence on their immune status. The study was approved by the institutional review board of the host site, and written informed consent was obtained from all subjects.

Training-related measures

Intensity levels for quantifying swim workouts were determined as detailed in Mujika et al. (23). An incremental test to exhaustion was performed at the beginning of each season (repeated and adjusted four times per season) to determine the relation between blood lactate concentration and swimming speed. Each subject swam 6 × 200 m at progressively higher percentages of their personal best competition time over this distance, culminating in a maximal effort on the sixth and final swim. Lactate concentration was measured in capillary blood samples collected from the fingertip during each 1-min recovery periods separating the 200-m swims (23). All swimming sessions were divided into five intensity levels according to the individual results: 1) below ∼2 mmol·L−1; 2) at ∼4 mmol·L−1, the onset of blood lactate accumulation; 3) just above ∼6 mmol·L−1; 4) at ∼10 mmol·L−1; and 5) at maximal swimming speed. In-water workouts were quantified in meters per week at each intensity level. Strength training included dryland workouts at maximal strength (1–6 repetitions, 80%–100% of 1 repetition maximum (RM)), dryland workouts of resistance training, (6–15 repetitions, 60%–80% of 1 RM), and general conditioning (involving cycling, running, cross-country skiing, team sports, etc.). Strength training was quantified in minutes of active exercise per week (1). For each subject, weekly training volumes at each intensity level were scaled as percentages of the maximal volume measured at the same intensity level for comparative purposes (1).

In-water workouts performed at continuous intensity levels are usually highly correlated; thus, intensity levels were synthesized as follows: the weekly low-intensity training load was the mean training volume (in percentage) of intensity levels 1 to 3; the weekly high-intensity training load was the mean training volume (in percentage) of intensity levels 4 and 5; in-water and dryland intensity levels were summarized via the weekly total training load (TTL) of the mean training volume (in percentage) of dryland and in-water workouts (1).

Training periods

Six training periods were distinguished from the TTL and competition dates. With reference to Foster’s work (7), intensive training load periods were defined as the weeks characterized by mean training volumes of dryland and in-water workouts individually normalized (expressed as a percentage of the individual maximum) greater than or equal to 60%: TTL ≥60%. Taper periods were defined as weeks 1–4 before competition periods, during which TTL <60%. Moderate training load periods were defined as the weeks characterized by TTL <60%, except taper periods. The competition periods comprised four or more days of competition. The postcompetitive periods comprised the 2 wk after each competition period. Last, break periods (approximately 4–6 wk between two sport seasons) corresponded to holiday periods without follow-up. Break periods were excluded from the analyses. Figure 1 illustrates the five training periods over one season for an international female.

Weekly TTL over one sport season for an international female swimmer, expressed as the percentage of the maximal TTL for this swimmer. Moderate training load periods are shown in light gray, intensive training load periods are shown in dark gray, taper periods are shown in white, competition periods are shown in black, and postcompetition periods are shown with shading lines.

Illnesses and affections

The questionnaire used to document all URTI, pulmonary, gastrointestinal, and urogenital infections and all neurological symptoms and MA was derived from the WURSS-44 and the questionnaire of Fricker et al. (8).

The two study physicians checked each subject on a weekly basis, documenting the following symptoms: sneezing, stuffy nose, runny nose, hoarseness, sore throat, sinus pain, sinus pressure, sinus drainage, swollen glands, ear infection, earache, plugged ear, ear discomfort, watery eyes, eye discomfort, headaches, sweats, chills, feeling feverish, feeling tired, irritability, and feeling depressed (for upper respiratory infections); and chest congestion, chest tightness, cough, and sputum (for pulmonary infections). The term MA referred to the following presentations: muscle injury (violent pain requiring exercise cessation), pulled muscles (muscle pain during exercise insufficient to stop exercise), tendinopathies (tendon pain with swelling and altered function), delayed-onset muscle soreness persisting 24 h after training, shoulder pain, and knee pain.

At the appearance of any signs or symptoms, the swimmers were immediately (within 1–6 h) seen by one of the two physicians for a complete clinical examination. Once a diagnosis was made, the swimmer was asked to complete the questionnaire daily during the period of illness. For each illness episode, the following metrics were recorded: type, duration (number of days), and peak severity (mild, no change in training program; moderate, training program modified; severe, complete cessation of training). An illness was defined and taken into account for the study when a swimmer reported one or more signs or symptoms on two or more consecutive days or when the severity was rated as either moderate or severe (8).

Upper respiratory tract and pulmonary infections (URTPI), MA, and AP covered by the questionnaire (gastrointestinal and urogenital infections, and all neurological symptoms, in addition to URTPI and MA) were only recorded if the swimmer showed signs or symptoms for 48 h, required medication, and missed at least one training session as a result of the symptoms. A recurring illness was defined as URTPI, MA, or AP occurring within 1 wk of a previously recorded episode (8,35). Only the first episode of a recurring illness was retained for analysis. For statistical reasons (i.e., rare events), the pathologies were grouped into three responses of primary interest for physicians and trainers: URTPI, MA, and AP.

Statistical analysis

Mixed-effects logistic regression was used to examine associations between the presence or absence of URTPI, MA, and AP as dependent variables (in three separate analyses) and training measures immediately preceding the illness measures as independent variables. To control for confounding, models were adjusted for time: sport seasons, semiseasons (winter, from September to March, and summer, from April to August), and individual characteristics such as age, sex, training center, and competition level. We added a subject-specific random effect to account for the potential correlation among the multiple measurements per swimmer (37). Short-term residual autocorrelations might have been observed for the subjects because of past infections or affections with an explicit influence on the present observation. Thus, we also adjusted for the recent past medical episodes.

We employed two coding strategies of the training variables to assess the different aspects of the training illnesses risk relations. In the first strategy, we used several continuous covariates: low-load, high-load, resistance, maximal strength, and general conditioning training. In the second strategy, we used training period indicator variables, indicating whether the swimmer was in a moderate, intensive, taper, competition, or postcompetition period in the week preceding the outcome. To compare the two coding strategies of the training variables, we used subject-specific random effect ANOVA.

Mixed-effects models were fitted using the NLMIXED procedure of SAS version 9.1 (Cary, NC). Coefficients were interpreted in terms of the odds ratio (OR) of an average subject (37). Models were analyzed separately for the three outcomes of interest and for the two coding strategies of the training variables.

Model selection relied on Akaike’s information criterion (AIC). This criterion assesses which model is more likely to be correct by balancing change in goodness-of-fit versus the difference in number of parameters. The criterion can be used to compare either nested or nonnested models, and we therefore compared all possible models. Furthermore, receiver operating characteristic curves and the associated area under the curve (AUC) with its 95% CI were calculated for each final model to assess their discriminating power. An AUC of 1.0 indicates perfect discrimination, whereas an AUC of 0.5 represents random guessing. Last, we conducted an influence analysis. To assess whether a predisposition to a specific illness in certain subjects influenced the parameters and the significance of the models, a subject-by-subject study was performed by calculating the similarity distance and Cook distance (37).


The cohort comprised 28 swimmers who were observed over a total of 2112 person-weeks. After excluding the break periods, 1827 person-weeks were used in the analyses, with 38 to 100 wk observed per subject. Training periods were distributed as follows: 601 wk of moderate training (33% of the 1827 person-weeks), 584 wk of intensive training (32%), 309 wk of tapering (17%), 208 wk of competition (11%), and 125 wk of postcompetition training (7%). During these weeks, the health events observed were as follows: 135 URTPI events (7%), 178 MA events (10%), and 319 AP events (17%). Table 1 shows the incidence of health episodes by sex, competition level, season, and semiseason.

Number of URTPI, MA, and AP events by sex, competition level semiseason, and season for the 1827 wk of follow-up of 28 swimmers.

Figure 2 shows the distribution of the five intensity training loads by training periods (means ± SD). ANOVA accounting for subject heterogeneity showed that low-load, high-load, resistance, maximal strength, and general conditioning training were significantly higher during intensive training periods than during moderate, taper, competition, or postcompetition training periods (all P < 0.0001).

Weekly training volumes at each intensity level expressed as the percentage of the maximal load (performed by each subject, at each intensity level, and over the 4-yr study period) by training period. Values are means + SD. ANOVA accounting for subject heterogeneity showed significantly higher volumes of low-load (P < 0.0001), high-load (P < 0.0001), resistance (P < 0.0001), maximal strength (P < 0.0001), and general conditioning (P < 0.0001) training during the intensive training period than during moderate, taper, competition, or postcompetition training periods.

Parameter estimates based on multivariate analysis using mixed-effects logistic regression models for each of the three outcomes are shown in Table 2 (first coding of the training variables strategy, in which training is reported using low-load, high-load, resistance, maximal strength, and general conditioning training) and 3 (second strategy, in which training is reported using the training periods). Only variables selected in at least one final model are displayed. The sex, age, and training center effects were not significant and were not selected in any final model. To account for heterogeneity among swimmers, random intercepts appeared to be sufficient. The significant random intercepts represent the combined effect of all unmeasured subject-specific covariates that cause some subjects to be more prone to the illness than others. The risks of MA, AP, and especially URTPI were higher in winter. For example, when using the second modeling strategy (Table 3), the odds of MA, AP, and URTPI were 1.83 (1.17–2.87), 2.18 (1.59–2.99), and 2.62 (1.63–4.21) times higher in winter months than that in summer months, respectively.

Estimated OR, 95% CI, and P values for training variables and adjustment factors retained in the final models, according to the AIC as fixed effects; estimated variance of subject-specific effects retained in the final models according to the AIC and corresponding P values; estimated AUC (95% CI) for the 1827 wk of follow-up of 28 swimmers.
Estimated OR, 95% CI, and P values for training variables and adjustment factors retained in the final models, according to AIC as fixed effects; estimated variance of subject-specific effects retained in the final models according to the AIC and corresponding P values; estimated AUC (95% CI) for the 1827 wk of follow-up of 28 swimmers.

In some models, the adjustment for year appeared to be relevant, perhaps in part because of the variation in seasonal distribution of illness and in training programs with respect to major 2- or 4-yr competition events (for example, the World Championships or Olympic Games). Parameter values corresponding to variables not selected in a particular model are indicated by a line. The risk of URTPI was significantly higher for national swimmers: OR = 1.40 (1.06–1.88) when training is reported using the five intensity-level training loads. The risk of MA or AP did not appear to differ between national and international swimmers.

The presence of illness 1 or 2 wk before the current illness was significantly associated with an increased risk of URTPI, AP, and especially MA. For a given week, the odds of having an URTPI was more than 2.4 times higher for swimmers having an URTPI the precedent week, and the odds of having an AP was more than 3.8 times higher for swimmers having an AP the precedent week compared with no event the preceding week. In addition, the odds of having MA was more than 7.1 times higher for swimmers having MA the precedent week compared with no event the preceding week, and more than 3.0 times higher for swimmers having MA 2 wk earlier.

The risk of URTPI was slightly but significantly increased with resistance (OR (95% CI) = 1.08 (1.01–1.16)) and high-load (OR (95% CI) = 1.10 (1.01–1.19)) training (Table 2). Ten percent increases in general conditioning and high-load training were associated with 19% and 14% higher OR of AP, respectively. Last, the odds of MA increased by 1.49 (1.14–1.96) for each 10% increase in high-load training and by 1.63 (1.20–2.21) for each 10% increase in general conditioning training.

Table 3 shows results from the multivariate analyses when training is reported using designated training periods. In general, intensive periods were at higher risk of illnesses, and then this training period is presented as the reference period. Thus, we observed that the odds of having an URTPI was 70% lower during periods of taper and 50% lower during competition periods than during periods of intensive training. The odds of having any of the pathologies was 66% lower during taper, 54% lower during competition, and 65% lower during postcompetition periods compared with intensive training periods. On the other hand, the odds of having MA was 54% lower during taper, 47% lower during competition, and 73% lower during postcompetition periods compared with periods of intensive training.

All models showed good discriminatory power (estimated AUC ranging from 80 to 90); that is, given a training load value or period, a competition level, a semiseason, a season and past outcomes, the models have the ability to highly correctly identify if a pathology was observed or not. Results from the influence analysis indicate that three swimmers had a major influence on the parameter estimates related to the sport season, although no influence was observed on the other parameter estimates.


The risks for URTPI, MA, and AP symptoms were higher for swimmers training in winter, those competing at a national rather than international level (for URTPI), and those with an episode of illness in the preceding weeks (notably for MA). Similarly, an increase in in-water high-load, dryland resistance training or general conditioning load was also associated with increased risk of illness. Moreover, the risk of URTPI and AP was higher in the periods of intensive training than that in the taper, competition, and postcompetition periods. Coaches and practitioners who work with swimmers presenting frequent illness should review training loads.

The frequency of illness observed in this study (319 episodes over 1827 wk of observation or 17%) agrees with earlier studies of athletes exposed to prolonged (several months and/or seasons) and intense periods of training, such as training overload and competition (8,16,35). Only two of the 28 swimmers did not fall ill during the observation period. In contrast, two other swimmers contracted a very high number of illnesses (18 and 25 episodes per year). URTI and MA were the most frequent pathologies observed in this study, with a mean number of episodes per year of 2.7 for URTPI and 4.4 for MA. In comparison, other studies report an average between 1.5 and 4 episodes per year of URTI, depending on the sport and the performance level, which is similar to the two to four episodes per year observed in the general adult population (2,8,40). An 11-yr retrospective study of illness in competitive swimmers showed that at least 15% of the swimmers experienced at least four infectious episodes per year (8), with the incidence highest in winter. Gleeson et al. (14) observed 22 elite swimmers for 12 wk of summer training in 1999 and diagnosed at least one infection in 10 of the swimmers (45% of the study subjects). The results of our study confirm that the total incidence of illness episodes in swimmers is similar to that in other populations of elite athletes and in sedentary individuals, and that this incidence is concentrated in the winter and during the heaviest training periods in terms of volume and intensity.

Illness risk is higher in winter than that in summer, as recently confirmed by several studies on athletes (19,40) and the general population (2). In Europe, this long period of winter extends roughly from the beginning of November to the end of March and is characterized by a surge in viral outbreaks that increase the risk of infection, particularly in cases of lowered immune defense (2). Several studies have reported an increase in fat and cholesterol intake associated with a lower intake of fruits and vegetables in winter (34), as well as lower mean concentrations of vitamin D and selected hormones that regulate immune function (19). As April ends this period with national qualifying championships for international competition, February and March are characterized by high training loads in both volume and intensity, which probably accentuates the susceptibility to infection. A similar pattern is evident in the Southern Hemisphere with swimmers preparing in winter months for major international competitions typically held in the Northern Hemisphere summer (8,35).

The lower risk of pathology in international swimmers as opposed to national swimmers is striking, given the similarities in age and training. A possible explanation is that international swimmers may have underlying genetic predispositions to better resistance or lower proinflammatory responses to infection (4,6,36). It is also possible that the individual threshold (training) load, above which the risk of illness increases (7), is lower in national swimmers than that in international swimmers. From this perspective, prescribing the same training loads to all members of a training team would increase illness risk in national swimmers. It is also possible that the athletic lifestyle may be better managed by older typically more mature international swimmers (i.e., management of academic, personal, and professional schedules; emphasis on sleep quality and nutrition; and immunonutrition supplementation) (26,31,40). Coaches should review training loads at the level of the whole squad, subgroups (e.g., international, national, and junior), and individual swimmers.

The two categories of exercise that contributed most to increasing the risk of URTPI were high workouts at maximal speed (blood lactate ≥10 mmol·L−1) and dryland resistance training. A consensus opinion (40) is that the immune system disturbances following acute exercise depend directly on the intensity of that exercise. Many studies have reported lymphocytopenia, a marked decline in Con A-stimulated lymphocyte proliferative response, and diminished natural killer cell activity with higher intensity exercise (18,24,39). These immune disturbances seem to be mediated by several mechanisms acting in concert, including a rise in stress hormones (10,39), increased plasma concentrations of substances known to modulate leukocyte function, such as inflammatory cytokines (12,28), and a shift in neuroregulation toward orthosympathetic predominance (5,21). Moreover, the main hypotheses to explain the powerful impact of high-load exercise center on the so-called open window theory, which holds that the weakened body is more susceptible to contracting an infection (12,14,24).

Intense muscular exercise can induce immune system disturbances similar to those observed during high-endurance exercise (27,28). For example, Nieman et al. (27) studied the immune responses induced by dryland resistance training, which consisted of leg squats up to muscular failure (10 repetitions at 65% of 1 RM), and observed leukocytosis, lymphocytosis, and lymphocytopenia and a 40% decrease in natural killer cell cytotoxic activity despite lower metabolic and hormonal responses. These immune modifications observed after brief, intense exercise seem to have been induced by muscle cell injury associated with the release of the filament fragments implicated in the initiation of the inflammatory cytokine cascade (28) and disturbance of systemic immune function (27).

In the present study, high-load and general conditioning exercise increased the risk of muscle affections such as pulled muscles, tendinopathies, delayed-onset muscle soreness, shoulder pain syndrome, and knee pain syndrome. There are three likely explanations for the deleterious effects of excessive high load on muscle function. First, the repetition of muscular actions over thousands of stroke cycles at high intensity causes muscle–tendon inflammation, which, associated with chronic glycogen depletion (3), induces chronic muscle fatigue. Second, muscle oxidative damage can occur after high-load training sessions (12,39). Third, sympathetic overtraining syndrome associated with excessive intensive exercise is characterized by the complete abolition of sympathetic nervous system control of skeletal muscle, as manifested by losses in strength and muscle power and delayed muscle recovery (9,21). Results from the present study suggest that high-load and general conditioning training should not always be scheduled together. General conditioning undertaken in the early stages of the season should be reduced during intensive and postcompetition periods.

This study is among the few that have carefully categorized and quantified the training cycles to examine how they are related to the risk of illness. Our analysis indicates a substantially higher risk of illness in intensive training periods, with OR two to three times higher compared with the taper, competition, and postcompetition periods.

In this study, the intensive training periods were characterized by significantly increased loads in all training categories. In these full-load periods, three daily sessions were usually scheduled (two in-water sessions and one dryland session) separated only by a few hours of recovery. An insufficient recovery time between these high-intensity sessions probably raised the risk of immune system disturbance, as has been observed by other researchers (22,32,33).

Our results confirm the findings of other studies that have also reported an increase in URTI symptoms during training periods characterized by high loads imposed continuously over several weeks (16,17,30). These studies indicate that accumulation of elevated training loads without recovery is associated with a chronic depletion of cellular and mucosal parameters, which may lower resistance to potential viral (15,17) and nonviral pathogens (40), thereby partially explaining the higher incidence of URTI (11,15,16,32). Along these lines, several studies on competitive runners (20,25,29) have reported that the incidence of URTI contracted following long events such as the marathon was further elevated when weekly and yearly training volumes were very high.

In the multivariate models, the presence of pathologies in the two previous weeks was the strongest risk factor, evidenced by the highest OR. The increase in the incidence of illness in cases of recent past infection confirms the major finding of Ekblom et al. (6). In a cohort of 1694 marathoners, a significantly higher incidence of infectious episodes in the 3 wk postrace was observed in runners who had had at least one infectious episode in the 3 wk prerace, as compared with runners who had not been ill before the marathon (33% vs 16%). An illness is more likely to last longer or reoccur if intensive training is resumed too quickly in subsequent weeks (12,40). Even though URTPI lasts an average of 2 to 7 d, depending on the nature of the infection (inflammatory allergic, bacterial, or viral) (40), training should be discontinued in the case of fever. Some commentators suggest that the recovery time before the athlete returns to prepathology workout levels should be at least as long as the duration of the pathology (30,40). A substantial decline in exercise capacity has been observed (impaired cell function and performance-related enzymes in skeletal muscle and myocardium, suppressed defense mechanisms, and greater increases in inflammatory responses during exercise) in the infectious and postinfectious phases. During the recovery period, athletes should gradually increase the training frequency and volume and participate only in low-load sessions (less than or equal to the aerobic threshold).

There are several practical implications of this study for reducing the risk of illness in elite swimmers. Winter with its high-load training presents particular risks, and it is essential to start this period in good health with strong immune defenses. During this period, high-load sessions (in-water sessions with the intensity above the anaerobic threshold and dryland sessions of resistance strength and maximal strength training) should be individualized and well spaced (30). Each high-load session should end with a recovery period to eliminate muscle metabolites, reduce inflammation, and restore neurohumoral balance (via active recovery, cryotherapy, and massage) (40). Adequate energy intake, especially protein and carbohydrates, before and during intensive sessions should be scheduled. During periods of intensive training, the number and duration of competitions should be reduced as well as, as far as possible, academic and professional obligations (31). Swimmers would probably benefit from training camps in temperate climates and employing strategies to reduce the risk of contamination by pathogens.

It was not possible to measure the mucosal and general immune system parameters and inflammatory markers during the study, given the study length, the large number of subjects, and the high cost of laboratory investigations. Without these measures, our interpretation of the underlying immunoregulation remains speculative. Although the swimmers were examined systematically by a physician for all symptoms of illness, pathology analyses to determine the cause of illness were not conducted to distinguish pathogens from inflammatory causes. Although all the swimmers met the same inclusion criteria, three of them showed a higher incidence of pathology. An analysis of the frequency of cytokine gene polymorphisms would have been useful to confirm genetic predispositions. As highlighted above, although the number of follow-up weeks was high, the number of subjects was relatively low, resulting in low statistical power to explain the lack of significance of factors stable over time, such as age and training center effects.

In conclusion, the risk of illness in elite swimmers is influenced by multiple factors including a history of prior illness, winter conditions, level of competitive performance (national or international), and the cumulative loads of training. Pool-based and dryland high loads contributed most to raising the risk of infection. Training periods with a high total load in volume and intensity between 60% and 100% of the maximum individual load showed the highest risk compared with the taper, competition, and postcompetition periods. Measures should be taken to organize training, prophylaxis, and recovery to reduce the risk of illness in elite swimmers.

The authors have received funding for research on which this article is based from the French National Institute of Sport, Expertise and Performance, Paris, France.

The authors report no conflict of interest.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. Avalos M, Hellard P, Chatard JC. Modeling the training–performance relationship using a mixed model in elite swimmers. Med Sci Sports Exerc. 2003; 35 (5): 838–46.
2. Baker CJ, Kimberlin DW, Long SS. Infectious diseases epidemiology and control. In: Pickering LK, editor. Red Book Atlas of Pediatric Infectious Diseases. Elk Grove Village (IL): American Academy of Pediatrics, Committee on Infectious Diseases; 2009. pp. 131–38.
3. Costill DL, Maglischo EW, Richardson AB. Swimming. Oxford: Blackwell Scientific Publications; 1992. p. 214.
4. Cox AJ, Gleeson M, Pyne D, Callister R, Fricker PA, Scott RJ. Cytokine gene polymorphisms and risk for upper respiratory symptoms in elite athletes. Exerc Immunol Rev. 2010; 16: 8–21.
5. Downing JE, Miyan JA. Neural immunoregulation: emerging roles for nerves in immune homeostasis and disease. Immunol Today. 2000; 21 (6): 281–9.
6. Ekblom B, Ekblom Ö, Malm C. Infectious episodes before and after a marathon race. Scand J Med Sci Sports. 2006; 16: 287–293.
7. Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc. 1998; 30 (7): 1164–68.
8. Fricker PA, Pyne DB, Saunders PU, Cox AJ, Gleeson M, Telford RD. Influence of training loads on patterns of illness in elite distance runners. Clin J Sport Med. 2005; 15 (4): 244–50.
9. Fry AC, Schilling BK, Weiss LW, Chiu LZF. β2-Adrenergic receptor down regulation and performance decrements during high-intensity resistance exercise overtraining. J Appl Physiol. 2006; 101: 1664–72.
10. Gabriel H, Kindermann W. The acute immune response to exercise: what does it mean? Int J Sports Med. 1997; 18 (1 Suppl): 28–45.
11. Gleeson M. Mucosal immune responses and risk of respiratory illness in elite athletes. Exerc Immunol Rev. 2000; 6: 5–42.
12. Gleeson M. Immune function in sport and exercise. J Appl Physiol. 2007; 103: 693–99.
13. Gleeson M, Bishop N, Oliveira M, Tauler P. Influence of training load on upper respiratory tract infection incidence and antigen-stimulated cytokine production. Scand J Med Sci Sports. 2013; 23 (4): 451–7.
14. Gleeson M, McDonald WA, Cripps AW, Pyne DB, Clancy RL, Fricker PA. The effect on immunity of long-term intensive training in elite swimmers. Clin Exp Immunol. 1995; 102 (1): 210–16.
15. Gleeson M, McDonald WA, Pyne DB, et al. Salivary IgA levels and infection risk in elite swimmers. Med Sci Sports Exerc. 1999; 31 (1): 67–73.
16. Gleeson M, McDonald WA, Pyne DB, et al. Immune status and respiratory illness for elite swimmers during a 12-week training cycle. Int J Sports Med. 2000; 21 (4): 302–7.
17. Gleeson M, Pyne DB, Austin JP, et al. Epstein–Barr virus reactivation and upper-respiratory illness in elite swimmers. Med Sci Sports Exerc. 2002; 34 (3): 411–7.
18. Gleeson M, Walsh NP. The bases expert statement on exercise, immunity, and infection. J Sports Sci. 2012; 30 (3): 321–24.
19. He CS, Handzlik M, Fraser WD, et al. Influence of vitamin D status on respiratory infection incidence and immune function during 4 months of winter training in endurance sport athletes. Exerc Immunol Rev. 2013; 19: 86–101.
20. Heath GW, Ford ES, Craven TE, Macera CA, Jackson KL, Pate RR. Exercise and the incidence of upper respiratory tract infections. Med Sci Sports Exerc. 1991; 23 (2): 152–7.
21. Hellard P, Guimaraes F, Avalos M, Houel N, Hausswirth C, Toussaint JF. Modeling the association between HR variability and illness in elite swimmers. Med Sci Sports Exerc. 2011; 43 (6): 1062–70.
22. MacKinnon LT. Special feature for the Olympics: effects of exercise on the immune system: overtraining effects on immunity and performance in athletes. Immunol Cell Biol. 2000 32 (7): 369–76.
23. Mujika I, Busso T, Lacoste L, Barale F, Geyssant A, Chatard JC. Modeled responses to training and taper in competitive swimmers. Med Sci Sports Exerc. 1996; 28 (2): 251–8.
24. Nieman DC. Is infection risk linked to exercise workload? Med Sci Sports Exerc. 2000; 32 (7): 406–11.
25. Nieman DC. Marathon training and immune function. Sports Med. 2007; 37 (4–5): 412–5
26. Nieman DC. Immunonutrition support for athletes. Nutr Rev. 2008; 66 (6): 310–20.
27. Nieman DC, Henson DA, Sampson CS, et al. The acute immune response to exhaustive resistance exercise. Int J Sports Med. 1995; 16 (5): 322–8.
28. Pedersen BK, Hoffman-Goetz L. Exercise and the immune system: regulation, integration and adaptation. Physiol Rev. 2000; 80: 1055–81.
29. Peters EM, Bateman ED. Ultramarathon running and upper respiratory tract infections. An epidemiological survey. S Afr Med J. 1983; 64 (15): 582–4.
30. Pyne DB. Effects of an intensive 12-wk training program by elite swimmers on neutrophil oxidative activity. Med Sci Sports Exerc. 1995; 27 (4): 535–42.
31. Pyne DB, Gleeson M, McDonald WA, Clancy RL, Perry C Jr, Fricker PA. Training strategies to maintain immunocompetence in athletes. Int J Sports Med. 2000; 21 (1): 51–60.
32. Rama L, Teixeira AM, Matos A, et al. Changes in natural killer cell subpopulations over a winter training season in elite swimmers. Eur J Appl Physiol Occup Physiol. 2013; 113 (4): 859–68.
33. Ronsen O, Pedersen BK, Øritsland TR, Bahr R, Kjeldsen-Kragh J. Leukocyte counts and lymphocyte responsiveness associated with repeated bouts of strenuous endurance exercise. J Appl Physiol (1985). 2001; 91 (1): 425–34.
34. Shahar DR, Froom P, Harari G, Yerushalmi N, Lubin F, Kristal-Boneh E. Changes in dietary intake account for seasonal changes in cardiovascular disease risk factors. Eur J Clin Nutr. 1999; 53 (5): 395–400.
35. Spence L, Brown WJ, Pyne DB, et al. Incidence, etiology, and symptomatology of upper respiratory illness in elite athletes. Med Sci Sports Exerc. 2007; 39 (4): 577–86.
36. Trammell RA, Toth LA. Genetic susceptibility and resistance to influenza infection and disease in humans and mice. Expert Rev Mol Diagn. 2008; 8 (4): 515–29.
37. Verbeke G, Molenberghs G. Models for Discrete Longitudinal Data. New York: Springer; 2006. pp. 265–277. pp. 290–293.
38. Verde TJ, Thomas SG, Moore RW, Shek P, Shephard RJ. Immune responses and increased training of the elite athlete. J Appl Physiol. 1992; 73 (4): 1494–99.
39. Walsh NP, Gleeson M, Shephard RJ, et al. Position Statement: part one: immune function and exercise. Exerc Immunol Rev. 2011; 17: 6–63.
40. Walsh NP, Gleeson M, Pyne DB, et al. Position Statement: part two: maintaining immune health. Exerc Immunol Rev. 2011; 17: 64–103.


© 2015 American College of Sports Medicine