Physical activity and lung function association in a healthy community-dwelling European population

The PneumoLaus study is part of the CoLaus|PsyCoLaus study (https://​www.​colaus-psycolaus.​ch), an ongoing prospective study aiming to assess the determinants of cardiovascular and psychiatric diseases using a population-based sample drawn from the city of Lausanne, Switzerland [11]. In June 2014, all participants of the CoLaus|PsyCoLaus study were invited to take part in the PneumoLaus study. Baseline examinations were conducted between June 2014 and August 2017, and the follow-up survey was conducted between June 2018 and February 2021. All ethnic groups were included as the general population was randomly invited to participate in the CoLaus|PsyCoLaus study.

The CoLaus|PsyCoLaus studiy were approved by the local Ethics Committee and participants provided written informed consent (https://​www.​cer-vd.​ch; project number PB_2018–00038, reference 239/09).

PneumoLaus methodology was already described [12]. Briefly, LF was assessed using a MasterScreen-PFT spirometer (Carefusion, Hoechberg, Germany), with Sentry Suite software (Version 2.17). Each manoeuvre was automatically assessed by computer, based upon acceptability and reproducibility criteria according to the 2005 American Thoracic Society– European Respiratory Society standards [13].

Reference values were applied according to the Global Lung Function Initiative (GLI) 2012, adjusting for the following ethnic origins: Caucasian, African, Northeast Asian, Southeast Asian and other [14]. If FEV₁/FVC or FVC was found to be below the lower limits of normal (LLN), spirometry was repeated 10–15 minutes after administration of 4 × 100 μg of salbutamol via a metered-dose inhaler and a spacer. Normal spirometry was defined by baseline FEV₁/FVC ratio and FVC above LLN, representing the lower 5th percentile (corresponding to a z-score of − 1.645) based on age, sex, height and ethnicity [15]. The maximal mid expiratory flow (MMEF) was defined by the mean forced expiratory flow between 25 and 75% of the FVC. An experienced respiratory technician and a consultant pulmonologist evaluated the quality of spirometry manoeuvres. Spirometry was included using recognised acceptability and reproducibility criteria [16].

PA was assessed using a wrist-worn triaxial accelerometer (GENEActiv, Activinsights Ltd., United Kingdom, https://​www.​activinsights.​com). The accelerometer has been validated against caliometry demonstrating excellent correlations against METs (Metabolic Equivalent of Task). The area under the Receiver Operating Characteristic (ROC) curves to discriminate sedentary activity ranged from 0.844–0.896, and for moderate to vigorous activity from 0.991–0.993 [17].

The devices were pre-programmed with a 50 Hz sampling frequency and subsequently attached to the participants’ right wrist. Participants were requested to wear the device continuously for 14 days in their free-living conditions. A valid day was defined as ≥10 hours of diurnal wear-time on weekdays and ≥ 8 hours on week-end days. At least 5 week days and 2 week-end days of valid data were required [18, 19].

PA was categorised into inactivity (< 85 m-g), light (85–180 m-g), moderate (181–437 m-g) and vigorous (> 437 m-g) PA, based on the thresholds defined by White et al. [20] and the average daily time spent within each category was utilised for analysis. MVPA was obtained by summing time spent in moderate and vigorous PA.

Educational level was categorised into high (university), middle (high school), and low (apprenticeship or mandatory) [21]. Nationality was categorised as born in Switzerland or not. Marital status was categorised as living with or without a partner. Smoking status was self-reported and categorised as never, former, or current smoker [22].

Body weight and height were measured with participants barefoot and in light indoor clothes. Body weight was measured in kilograms to the nearest 100 g using a Seca® scale (Hamburg, Germany). Height was measured to the nearest 5 mm using a Seca® (Hamburg, Germany) height gauge. Body mass index (BMI) was categorised as normal (< 25 kg/m²), overweight (≥25 and < 30 kg/m²) and obese (≥30 kg/m²) [23].

Participants from the PneumoLaus study (2014–17 and 2018–21) were considered as eligible. Those with possible restrictive ventilatory impairment defined as FVC < LLN before and after bronchodilation or obstructive ventilatory impairment defined as FEV₁/FVC < LLN or missing any information concerning PA and smoking status were excluded.

Of the initial 4881 participants of the baseline survey, 1910 (39.1% participation rate, 54.7% women, 62.0 ± 9.7 years) were included in the analysis. Of the initial 3751 participants of the follow-up survey, 1174 (participation rate 31.3, 53.4% women, 65.8 ± 9.5 years) were included. The reasons for exclusion and the characteristics of the included and excluded participants are summarised in Fig. 1 and Supplementary Table 1, respectively. In the baseline survey, included participants were younger, more frequently professionally active, of higher educational level and less frequently smokers or hypertensive than excluded participants. In the follow-up survey, included participants were older, less professionally active, more frequently of middle educational level, and less frequently smokers than excluded participants.

The results of bivariate and multivariable analyses for percentage of PV are summarised respectively in Tables 1 and 2, and the R² for bivariate associations are provided in Supplementary Table 2. In both studies and analyses, vigorous PA was significantly and positively associated with FEV₁ and FVC except in current smokers of the follow-up study. R² were low, ranging between 0 and 4.2%.

In the multivariable analysis of both studies, moderate PA was significantly and positively associated with FEV₁ and FVC except in current smokers of the follow-up. No significant association was found between any PA levels and MMEF, except for a positive association for moderate and vigorous PA in current smokers of the baseline study. Significant associations were found between light PA and both FVC and FEV₁ as % of PV, except for current smokers of the follow-up.

Table 3 shows the mixed method analysis, merging both follow-ups, and including repeated measures of participants who performed LF in both follow-ups. The results confirm those obtained separately for each follow-up. For example, an increase of 15 minutes in inactivity leads to a decrease in FEV₁ of 0.12%.

Similar findings were found upon LF in absolute values (Supplementary Tables 3, 4 and 5). In fact, only association between vigorous PA and both FEV₁ and FVC were significantly positive when adjusting for BMI and age. A visual depiction of the bivariate associations between spirometry values and vigorous PA, respectively inactivity, is provided in Supplementary Figs. 1 and 2.

The present study is consistent with the positive association found in three other cross-sectional studies which also used accelerometer-assessed PA [10, 30, 31]. In contrast to our study, the association was described as stronger for former and current smokers compared to non-smokers [10, 30] but this difference might only be due to lack of power of our small smoker group. On the contrary, Barboza et al. [32], who also used accelerometer-assessed PA, found no association between PA and LF, nor did Smith et al. [5]. Our distinct results may be attributed to the difference in definition of PA intensity levels, as several studies assessed PA based on questionnaires and others on accelerometry but without forcibly applying the same thresholds as we did. Similarly, there were differences between the studied populations. Ours was older: mean 62.0 ± 9.7 years at baseline and 65.8 ± 9.5 years at follow-up versus 47 ± 14 years (males) and 53 ± 8 years (females) in Barboza et al. [32] and 15.2 ± 0.25 years in Smith et al. [5]. Our population was also less overweight: mean BMI 26.3 ± 4.6 kg/m² at baseline and 26.2 ± 4.3 at follow-up versus 27.4 ± 4.5 kg/m² (males) and 29.8 ± 7.07 kg/m² (females) in Barboza et al. [32]. In fact, Barboza et al. [32] used both questionnaire and accelerometery-assessed PA but only show accelerometer-derived inactivity data. Also, the device was worn on the wrist in our study, versus the hip in the others [5, 32]. The waist location has indeed been shown to have better accuracy, followed by the left and then the right wrist, which was used in our study [33]. As inflammation increases with age, one hypothesis to explain our findings could be that PA counterbalances excessive inflammation, which is not present in youth [26]. Indeed, Garcia-Aymerich et al. found an association between PA and LF only in older people (> 40 years) [3], already starting with moderate PA. Our study seems to confirm the age-related association but not the association between moderate PA and spirometric indices. Moreover, as shown in Rose et al.s systematic review and metanalysis, the intensity of exercise did not influence chronic inflammatory response in general, even though sub-analyses suggest that higher training intensity may have more effect in middle-aged adults [34]. For ease of understanding, Table 4 summarises the results of the cited studies.

The strength of this study is the objectively assessed PA by accelerometery and LF by spirometry. As this study was conducted in a sample of apparently healthy, community-dwelling people, our results indicate that maintaining an adequate level of PA is associated with a better LF. Finally, our sample size is larger than most other studies conducted in healthy people (n = 62 for Barboza et al. [32], n = 322 for Luzak et al. [30], n = 895 for Smith et al. [5]), thus allowing a bigger statistical power.

Our study also has limitations. First, it was conducted in a single city with mainly subjects of European ancestry, hence, results might not be generalisable in other settings or other ethnicities, although we believe that such associations should hold in most cases. Second, the cross-sectional design does not allow to establish the cause-effect relationship between PA and LF or to assess whether a long-term relationship between PA and LF exists. A dedicated prospective analysis would be necessary to answer this. Third, we did not determine the correlation between PA and static lung volumes such as total lung capacity, functional residual capacity, and residual volume. Fourth, due to low R², clinical significance is not guaranteed despite statistically significative results, as shown in other studies. A meta-analysis would allow to draw stronger conclusions. Nevertheless, this small fraction of influence on spirometry metrics is similar or even higher than that obtained in polygenic risk scores for asthma, where 95% of R² values are below 6% [35] or in similar scores in psychiatry, where R² reach 2% for major depressive disorders, 5% for bipolar disorders or 6% for attention deficit hyperactivity disorders [36‐39]. Fifth, the accelerometer was positioned in the dominant (right) wrist, contrary to other studies that used the nondominant side. This was done to increase participation, as wearing the accelerometer on the left wrist would interfere with the wristwatch that most participants wore. A study [33] suggested that results obtained using the right wrist might be less reliable than using the left one, but the study was conducted during a very limited period of time (1 hour) and in a laboratory setting. Hence, the findings might not be applicable to free-living people. Nevertheless, it would be important that other studies be conducted using the accelerometer on the nondominant hand and with a sample size like ours.