Asthma in Competitive Cross-Country Skiers: A Systematic Review and Meta-analysis

Original articles from PubMed, EBSCO Academic Search Premier, Web of Science, Scopus, Cochrane Library and clinicaltrials.gov were searched. The search date was August 29, 2019. No language or time filters were used. The details of the search strategy are presented in Supplementary File 1. After study selection, the references from these articles were screened for further references.

Three authors (RM-H, JK, LL) independently screened the titles and abstracts of the studies. The studies had to meet the following criteria to be included: (1) The study participants were active, competitive cross-country skiers, biathletes, Nordic combined athletes or ski-orienteers, i.e., athletes who compete using cross-country skis. (2) The study produced new original data on the prevalence, incidence or pathophysiology of asthma, physiological phenomena related to asthma or asthma-related symptoms or asthma medication.

The exclusion criteria for our review were as follows: (1) the study population included skiers, but their results were not reported separately from those of other athletes, and (2) the subjects were recreational skiers not competitive skiers.

From each included study, we extracted the first author, country of the study, publication year, title, study population and characteristics, test protocols, prevalence of self-reported or physician-diagnosed asthma, asthma-related symptoms and environmental factors, allergies and other respiratory diseases, study funding and conflicts of interest. One reviewer (RM-H) extracted the data, and two other authors (JK and LL) verified the correctness of the collected data.

Three authors (RM-H, JK, LL) reviewed the risk of bias in the included studies in collaboration using the Cochrane risk of bias tool. Meta-analyses were conducted by calculating Freeman–Tukey-transformed [62] proportions and summary estimates with a random effects model using a restricted maximum likelihood (REML)-approach (transformed prevalence of self-reported physician-diagnosed asthma, asthma in skiers with either a self-reported physician diagnosis or a diagnosis based on lung function measurements, and the use of asthma medication). Inverse variance weighting was used to calculate the weights for individual studies. After performing the meta-analyses, individual proportions and summary estimates were back-transformed according to Miller [63]. Forest plots were then drawn using the back-transformed data. The analyses were conducted using R version 3.4.3 [17] and R-package Metafor [18]. Other outcomes are presented narratively.

Asthma-related symptoms include chest tightness, shortness of breath, coughing and wheezing. The prevalence of these symptoms has been studied in different studies using structured questionnaires. In a study by Heir and Oseid [21], 84% of 153 skiers had at least one symptom. Sue-Chu et al. [22] found that 46% of the skiers studied in Norway and 51% of those studied in Sweden had wheezing and breathlessness or chest tightness. In a screening study by Turmel and others, 50% of 44 skiers and biathletes had exercise-induced symptoms [25]. Norqvist et al. [26] found that 22% of the skiers (n = 238) had asthma-related symptoms. In the study by Rundell and others [46], 62% of the population of cross-country skiers (n = 21) reported at least one symptom.

There was a significant difference in the prevalence of asthma symptoms between sexes among 15–19-year-old skiers (10% in males vs. 30% in females, p = 0.001) but not in the 20–34-year-old group (24% in males vs. 29% in females, p = 0.582) [26]. In a study by Eklund and colleagues [27], 16% of high-school aged skiers had asthma-related symptoms. They also found a significant difference in the prevalence of symptoms between sexes (9% in males vs. 23% in females, p = 0.005) [27].

Studies investigating the prevalence of asthma were divided into three categories based on how asthma was defined and diagnosed: (1) self-reported physician-diagnosed asthma; (2) asthma diagnosed based on lung function measures as part of the study; (3) either self-reported physician-diagnosed asthma or asthma diagnosed based on lung function tests in the study.

Several studies have revealed the possibility of asthma being underdiagnosed in skiers. Previously healthy skiers without a prior asthma diagnosis or medication use were found to fulfill diagnostic criteria of asthma according to lung function tests. There seems to be a decreasing trend in the prevalence of undiagnosed asthma in previously healthy skiers, as it decreased from 55 to 20% from 1993–2010, excluding the data from the most recent and relatively small study of 13 subjects [19, 29, 32, 33, 35, 36].

Larsson et al. [19] conducted a study in 1993 in 42 cross-country skiers, where asthma was diagnosed based on positive methacholine challenge test results and the presence of at least two asthma symptoms. Thirty-four percent of the previously healthy skiers fulfilled the diagnostic criteria used. In a later study conducted by Sue-Chu et al. in 1999 [28], 40% (12/30) of previously healthy skiers were considered to have ski asthma. In this study, the diagnosis was also defined as positive methacholine challenge test results and the presence of asthma-like symptoms.

A field exercise challenge was used as the diagnostic test by Ogston and Butcher [32]. They found that 31% (20/91) of the unmedicated high school skiers had exercise-induced bronchoconstriction [32]. A Finnish study by Pohjantähti and others [33] reported that 35% (7/20) of the previously healthy skiers had a decrease in FEV1 by ≥ 10%, a decrease in MMEF (maximal mid-expiratory flow) by ≥ 20% or a decrease in both (2 FEV1, 7 MMEF) with a similar challenge.

In 2010, Sue-Chu et al. [35] compared different diagnostic tests and observed that a total of 25% (12/48) of the previously healthy skiers had bronchial hyperresponsiveness in either the methacholine challenge test or EVH (eucapnic voluntary hyperpnoea) test and fulfilled the criteria for therapeutic use exemption (TUE) at that time.

A British multisport study conducted by Dickinson et al. [36] showed that 62% (8/13) of the biathletes included were previously healthy but had bronchial hyperresponsiveness in the EVH test.

There is only one study assessing possible risk factors of developing asthma among competitive skiers. Eklund and colleagues reported among 244 skiers from a high school population that family history of asthma and nasal allergy were significant risk factors for asthma both among competitive skiers and non-skiers [27]. It seemed that allergy was not as significant risk factor among skiers as it is among non-skiers.

The age of asthma onset in skiers has been studied in Norway and Sweden in four different studies [19, 21, 26, 27]. The onset age varied from early adolescence to early adulthood, and the onset occurred at a later age in the skiers than in the controls.

Eklund et al. [27] compared junior elite cross-country skiers, biathletes and ski-orienteers to controls in upper secondary schools in Sweden. The median age at onset of asthma in the skier group was significantly higher than that in the controls (12 vs. 8 years, p < 0.001). the onset age of asthma was distributed evenly from birth into adolescence in the control group and was concentrated to the 10–15-year-old age range in the skier group. The mean age of the skiers was 16.8 ± 1.2 years.

In another Swedish study conducted by Norqvist et al. [26], the onset age of asthma in both skiers in the 15–19-year-old group and those in the 20–34-year-old group was mostly in early adolescence.

In a study by Heir and Oseid [21], 16 of 22 skiers with self-reported symptoms recalled their onset age of asthma. Fifteen skiers reported an onset age in late adolescence or early adulthood. Only one skier reported the onset of asthma in early childhood. Moreover, Larsson et al. [19] reported that none of the 42 skiers with asthma in their study recalled their onset age of asthma to be during childhood.

Six studies reported the use of asthma medication among skiers. Asthma medication use was defined as the use of one or more of the following medications: inhaled bronchodilators (β₂-agonists or anticholinergic agents), inhaled anti-inflammatories (corticosteroids, cromoglycates), oral theophylline or corticosteroids (Fig. 4, Table 4). The prevalence of asthma medication use in skiers was on average 23% (CI 95% 19–26%) across six studies with 1146 subjects.

There were two studies investigating the effect of bronchodilators on lung function and exercise performance in skiers [43, 44] and one assessing the effect of anti-inflammatory treatment [30].

Sandsund and others 1998 [43] studied salbutamol exclusively in cross-country skiers. Eight male skiers completed two exercise tests in a climatic chamber set to − 15 °C after the inhalation of salbutamol (3 × 400 μg) or a placebo. Three skiers were using anti-asthmatic medication but had not been diagnosed with asthma. The withdrawal of medication before the test was not reported. FEV₁ increased significantly after the inhalation of salbutamol before the test and was higher during the test compared to the levels with the placebo at all six time points. No significant changes were observed in maximal oxygen uptake or blood lactate levels, and salbutamol did not have performance-enhancing effects.

In 1999, Sue-Chu et al. [44] investigated the effect of salmeterol on physical performance on a treadmill in a climatic chamber in eight healthy male cross-country skiers at − 15 °C. The inhalation of 50 μg salmeterol significantly improved the FEV1 level before, during, and after the exercise test but did not have an effect on the time to exhaustion (392.5 s with salmeterol vs. 395.6 s with the placebo, p = 0.84).

Sue-Chu et al. [30] studied the budesonide treatment over a 22-week period in 25 cross-country skiers who had’ski asthma’ (i.e., two or more asthma-like symptoms, including wheezing and abnormal breathlessness or chest tightness upon exertion, at rest, or upon exposure to irritants and BHR to methacholine) and had not used anti-asthmatic medication. They demonstrated no significant improvement in lung function, airway inflammation or tenascin expression.

There were three studies comparing skiers’ bronchial reactivity in different diagnostic tests [31, 34, 35], one study comparing self-reported symptoms to bronchial hyperresponsiveness [56] and one study reporting the longitudinal follow-up study in three cross-country skiers [45]. The studies are represented in Table 6.

In short, three studies [31, 34, 35] assessed the prevalence of BHR among skiers using different tests. There was marked variation in the proportion of subjects having positive test results using different protocols. Stenfors reported that the self-reported symptoms have poor diagnostic accuracy in predicting BHR [56]. The longitudinal case study by Verges and others showed variable signs of airway obstruction in three skiers [45].

The risk of bias in the selected articles was assessed by the Cochrane risk of bias tool. Among the 33 articles that were included in the synthesis, 19 articles were considered to have a low risk of bias, 9 articles were considered to have a moderate risk of bias, and 3 articles were considered to have a high risk of bias. Two articles were excluded from the review process due to a very high risk of bias and unclear reporting [65, 66]. See electronic Supplementary File 2 for details.

Among cross-country skiers, the mean prevalence of self-reported physician-diagnosed asthma was 21%, and for a combination of self-reported physician-diagnosed asthma and asthma based on lung function measures was 28%, and that for objectively verified asthma was 20% [19‐27]. These values are considerably higher than the prevalence of asthma in the general population, which is approximately 10% [4, 5]. Three of the studies also included controls who did not engage in competitive sports, and the prevalence of asthma was significantly higher in cross-country skiers [20, 21, 27].

About 10% of people in the general population are diagnosed with asthma, and it has been estimated that approximately another 5–10% have asthma-like symptoms [6]. According to the current review, the prevalence of at least one asthma-like symptom among competitive cross-country skiers was high and varied between 22 and 84%. However, it is important to note that asthma-like symptoms as such are often not related to asthma [56]. In studies where at least two symptoms were required to be considered asthma-like, the highest prevalence was 51% [22]. In all three studies that compared the difference between cross-country skiers and controls, the prevalence of asthma-like symptoms was higher in skiers [20, 21, 27].

The high prevalence of asthma among competitive skiers is probably not only related to the sport but also cold air exposure and high ventilatory demand, since the prevalence of asthma varies between sports with high and low ventilatory demand [8].

The relative frequency of diagnosed asthma to that of asthmatic symptoms without a formal diagnosis is partly related to the responsibility of health care professionals to suspect asthma and conduct diagnostic testing in symptomatic subjects. If there is a high threshold for suspecting and testing for asthma, asthma may be underdiagnosed, and a high proportion of subjects with asthma suffer from symptoms without proper a diagnosis and medication. The prevalence of asthma-related symptoms relative to the prevalence of asthma differed across the studies included in this review. In three studies, the number of symptomatic skiers was higher than that of asthmatic skiers [21, 22, 25], but in two more recent studies, there were more asthmatic skiers than skiers with asthma-related symptoms without a diagnosis of asthma [26, 27]. Furthermore, in many studies, a large share of previously healthy skiers had variable airway obstruction compatible with asthma, and this proportion was higher in earlier studies and lower in more recent studies [19, 32, 33, 35, 36, 44]. Together, these results suggest that asthma has been underdiagnosed but that this problem has diminished according to the latest studies, possibly due to increased general awareness of asthma among skiers. In most of the studies, the criteria for asthma used with the different lung function tests were based on international guidelines or common practice, but not all criteria were based on these guidelines or common practice (e.g., changes in MMEF after exercise). There are no studies discussing possible overdiagnosis of asthma in cross-country skiers. It is important to note that symptoms as such are not reliable predictors of airway hyperresponsiveness and asthma [56]. In studies reporting self-reported physician-diagnosed asthma it is not clear if objective measures had been used to diagnose asthma or if the diagnosis was based on symptoms only.

The prevalence of asthma medication use was 23% in 1146 skiers across six studies. The proportion of competitive skiers reporting the use of asthma medication is in accordance with the reported prevalence of asthma in this population. Based on these results, it can be cautiously stated that there is no evidence for the misuse of asthma medication in skiers. However, based on all the studies included, it cannot be concluded whether the subjects with a diagnosis of asthma and the subjects using asthma medication were the same subjects. In addition, the use of asthma medication is based on self-reports only. Of the studies in this review, only Heir and Oseid [21] reported the use of asthma medication among skiers without a diagnosis of asthma, and they found that nine percent of athletes with no asthma diagnosis used asthma medication.

The use of asthma medication among competitive skiers has been a controversial topic and widely discussed in media. The use of corticosteroids is now allowed and also the use of several bronchodilators is allowed in up to relatively high doses (salbutamol 1600 ug over 24 h period and 800 ug over 12 h period, formoterol 54 ug over 24 h, salmeterol 200 ug over 24 h) [60]. We identified two studies in which the effects of a relatively high dose of a short-acting β₂-agonist (salbutamol) or a normal therapeutic dose of a long-acting β₂-agonist (salmeterol) on lung function and exercise capacity were studied in healthy skiers or skiers with treated asthma [43, 44]. Interestingly, exercise capacity was not affected by these bronchodilators in either of the studies, although lung function was mildly improved, as expected. These results suggest that occasional use of bronchodilators in normal therapeutic doses does not improve exercise capacity, but it is beneficial in preventing bronchoconstriction in skiers with asthma. However, high oral doses of salbutamol or a combination of inhaled β₂-agonists improved sprinting capacity or maximal strength in two studies [47, 48] acutely and in longitudinal use. It remains to be studied whether these kinds of improvements in strength may benefit healthy cross-country skiers to improve, for example, their final sprint performance. Moreover, there is a lack of research investigating the longitudinal effects of bronchodilators on exercise performance apart from previously mentioned study by Hostrup and others [47].

There are only a few studies on the cellular and inflammatory mechanisms of asthma among skiers. In general, skiers had higher levels of inflammatory cells in their airways than did healthy non-skiers. Interestingly, inflammatory changes were observed in the airways of both asthmatic and non-asthmatic skiers, and the level of eosinophils was higher in non-skiers with asthma than in skiers with asthma. The exhaled nitric oxide concentration, an inflammatory biomarker associated with eosinophilic inflammation in individuals with asthma, was not different between skiers and healthy controls and was lower in the skiers than in asthmatic controls. Taken together, these results suggest that the distribution of inflammatory endotypes may differ between skiers and non-skiers with asthma that eosinophilic inflammation may not be as prevalent in skiers with asthma and that skiing even in the absence of asthma may trigger noneosinophilic inflammation.

Asthma is usually associated with chronic airway inflammation [1]. Currently, asthma is divided into different endotypes based on the underlying inflammatory mechanisms. The most frequent inflammatory endotypes include allergic eosinophilic asthma, non-allergic eosinophilic asthma, neutrophilic asthma and pauci-granulocytic asthma [49]. In all subjects, whether they have asthma or not, an increase in minute ventilation and dryness of the inhaled air increases the loss of water from epithelial lining fluid (ELF) covering the airway mucosa, making it hypertonic. Mast cells and other inflammatory cells are present in the airways of people with asthma, and they are activated to excrete inflammatory mediators, such as histamine and leukotrienes, in a hypertonic environment. This so-called osmotic mechanism is thought to trigger airway obstruction in asthmatic people during or after exercise [13, 50].

Neutrophilic inflammation in asthma is associated with different irritating stimuli [51], and exposure to cold air is reported to increase airway neutrophilia in people with asthma [52]. Since skiers compete and train in cold conditions and cold air has a very low absolute humidity, their airways are repeatedly exposed to a loss of water and increased osmolality of the epithelial lining fluid. This repeated exposure to high ventilation rates in cold air may, in addition to triggering airway obstruction in people with established asthma, cause irritation and low-grade chronic inflammation and thereby induce asthma in skiers. This idea is also supported by the findings that bronchial hyperresponsiveness and the number of inflammatory cells in airways increase among skiers during the winter [38, 40]. It is important to note that airway infections increased bronchial responsiveness in skiers contrary to healthy controls [41].

The age at onset of asthma is associated with the inflammatory mechanism underlying asthma. Allergic eosinophilic asthma often begins in childhood or early adulthood, while non-allergic eosinophilic and noneosinophilic endotypes more often begin in adulthood [49]. As asthma is a relatively common disease in the general population, some skiers with asthma have likely developed asthma even if they had never started skiing, but some of the skiers may have true skiing-induced asthma. As the onset age of asthma was higher among skiers than among non-skiers with asthma [27] and in one study, none of the skiers reported the onset of asthma in early childhood [19], the prevalence of allergic eosinophilic asthma may be lower among skiers than among non-skiers with asthma. This concept is logical, since children who are diagnosed with allergic asthma during childhood are probably less likely to engage in sports at the competitive level and start an active career as a skier. On the other hand, once young people start actively training and competing in cross-country skiing, they may be at a higher risk of developing asthma due to repeated airway irritation with cold air exposure. The fact that allergic asthma is eosinophilic and asthma induced by irritation is more often neutrophilic can partly explain why asthma in skiers in the existing studies was associated more strongly with neutrophilic than eosinophilic inflammation. The relatively higher neutrophilic activity in skiers with asthma may also have treatment-related consequences, since anti-inflammatory treatment with inhaled glucocorticoids is most effective in people with asthma and eosinophilic inflammation [53‐55]. In fact, this concept is in line with the results of a study that assessed the effect of inhaled budesonide on ski asthma and found no improvement in lung function or airway inflammation [30].

Due to the high prevalence of asthma among competitive skiers, asthma should be actively screened. Since self-reported symptoms are not a very reliable predictor of asthma-like reversible or variable airway obstruction among skiers [56], objective measures of lung function are needed. In general, tests of bronchial hyperresponsiveness seem to be more sensitive tools in detecting asthma than spirometry with bronchodilation. Based on two studies [34, 35], the tests used to detect hyperresponsiveness in order of the highest to the lowest level of sensitivity are as follows: the methacholine challenge, exercise test, EVH test, AMP challenge and mannitol challenge. Of these tests, the methacholine test acts directly on muscarinic receptors in bronchial smooth muscle, whereas the other tests are indirect and cause airway obstruction by increasing the tonicity of epithelial lining fluid, which triggers inflammatory cells to release mediators that activate smooth muscle [2]. It is understandable that direct tests of hyperresponsiveness, such as the methacholine challenge, are more sensitive than indirect tests, because the latter tests are dependent on the presence and activity of inflammatory cells in airways. However, it is important to note that the sensitivity and specificity of each of these tests is dependent on the threshold used (e.g., the provocative dose of methacholine or the level of decrease in FEV₁ in exercise challenge). EVH test is usually considered more sensitive than the exercise test in detecting hyperresponsiveness [57]. The opposite finding in skiers in one study may be related to the fact that top athletes may be able to achieve even higher minute ventilation in their own sport than in the EVH test with a predefined level of minute ventilation.

The onset and severity of asthma-related symptoms may progress gradually, and athletes may not be able to differentiate asthma-related symptoms from normal breathlessness and exhaustion. Therefore, it may be useful to conduct regular screenings of asthma to promote athletes’ health. This would decrease the probability of athletes training and competing with asthma that they are not aware of. On the other hand, regular screening of asthma requires resources and there is no universal agreement over the correct criteria of asthma in athletes. A single positive test result in a direct bronchial challenge with no asthma-related symptoms may not indicate asthma and needs further verification in indirect challenges, since healthy people, especially cross-country skiers, may have increased bronchial responsiveness during the winter season, as shown in the study by Heir [40]. A positive test result in an indirect bronchial challenge, such as the exercise challenge, with no asthma-related symptoms may, however, be indicative of asthma, because people have a measurable level of obstruction in real-life conditions, which affect exercise performance. Regardless of whether the lung function test is used, it is worth considering that due to the lack of a gold standard in asthma diagnostics, all the diagnostic cut-off values are somewhat controversial and are based on studies in non-athletes. Whether different cut-off values are needed for skiers or other competitive athletes is not known.

Exercise-induced laryngeal obstruction (EILO) is a transient, reversible and inappropriate narrowing of the larynx in response to exercise [58]. In contrast to exercise-induced asthma, the symptoms in EILO peak during exercise, whereas the symptoms in exercise-induced asthma may worsen after exercise. There is one study assessing the prevalence of EILO in a retrospective cohort by Nielsen and others [61]. EILO was found in 35.2% (n = 31) of the studied athletes and asthma was concurrently present in 38.7% (n = 12) in athletes with EILO. Since asthma and laryngeal obstruction may coexist [59, 61], it is important to remember that exercise-induced symptoms in an individual treated for asthma may not be due to bronchial obstruction but rather due to untreated laryngeal problems.

This review is based on a systematic literature search focusing on all aspects of asthma in competitive cross-country skiers, and all the major databases were searched without restrictions. The number of studies identified and included was relatively small, and each topic was often covered by only a few studies with considerable differences in aspects of the studies, such as the methods, definitions, and cut-off levels. One common problem among skiers and a possible confounder in all asthma-related research is repeated respiratory infections. Intensive training in cold conditions, a high level of stress frequent travelling and mass gatherings in sports increase the risk for infection and symptoms suggestive of asthma. In fact, one of the studies that assessed bronchial responsiveness among skiers reported increased bronchial responsiveness after an airway infection in skiers but not in controls [41].

Studies included in the review were published between 1993 and 2018. During this time, the understanding of asthma pathophysiology has increased, and the diagnostic methods and criteria have been standardized. Moreover, the treatment of asthma and doping regulations have evolved based on new knowledge. Thus, the methods and research questions have also evolved during this time period.

In our risk of bias assessment, 10% of the articles were considered to have moderate or high risk of bias, and this result should also be taken into account.

The meta-analysis was based on the eight available publications on the prevalence of asthma and use of asthma medication. The relatively small number of published studies decreases the reliability of the results and makes it challenging to use formal publication bias assessment. The methods assessing possible publication bias usually require more than ten articles to be included in the meta-analysis due to heterogeneity and low power [64]. We used random effects model in the meta-analysis. However, the number of studies was small and this may affect the reliability of significance testing, as the small number of studies undermines the accuracy of the estimate of the between-study variance, standard error of the estimated prevalences and thereby ultimately the significance testing.