Skip to main content
NCBI home page
Preventive Medicine Reports logoLink to Preventive Medicine Reports
. 2023 Feb 8;32:102130. doi: 10.1016/j.pmedr.2023.102130

Baseline physical activity and the risk of severe illness and mortality from COVID-19: A dose–response meta-analysis

Junjie Liu a,1, Zhiguang Guo b,1, Songtao Lu a,⁎
PMCID: PMC9905049  PMID: 36778629

Abstract

To provide a scientific basis for improved exercise-based immunity, a meta-analysis was used to explore the dose–response relationship between physical activity (PA) and the risk of severe illness and mortality related to COVID-19 (coronavirus disease 2019). To this end, we searched PubMed, Web of Science databases from January 2020 through April 2022. 14 observational studies met the criteria for inclusion in the meta-analysis, including 2840 cases of severe illness and death from COVID-19. Categorical dose–relationship analysis showed that the risks of severe illness and mortality from COVID-19 were, respectively, 46% (risk ratio (RR): 0.54; confidence intervals (CIs): 0.41–0.68) and 59% (RR = 0.41; 95%CI: 0.23–0.58) lower for the highest dose of PA compared with the lowest dose of PA. The results of the continuous dose–response analysis show an inverse nonlinear relationship (Pnon-linearity < 0.05) between PA and both the risk of severe illness and mortality from COVID-19. For PA below 10 MET-h/week (MET-h/week: metabolic equivalent of task-hours/week), an increase of 4 MET-h/week (1 h of moderate-intensity or 0.5 h of high-intensity PA) was associated with 8% and 11% reductions in the risk of severe illness and mortality from COVID-19. PA above 10 MET-h/week lowered the risk of severe illness and mortality from COVID-19 by 7% and 9%, respectively, for each 4 MET-h/week increase. Doses of WHO-recommended PA levels (10 MET-h/week) may be required for more substantial reductions in the risk of severe illness and mortality from COVID-19.

Keywords: Physical activity, COVID-19, Mortality, Severe illness, Meta-analysis, Dose–response

1. Introduction

The coronavirus disease 2019 (COVID-19) continues to spread worldwide. As of July 15, 2022, COVID-19 had caused 6,356,812 deaths worldwide (World Health Organization, 2022a, World Health Organization, 2022b). Governments and medical institutions worldwide are committed to controlling, limiting, and ending this pandemic, via, for example, vaccination and antiviral treatments based on blood products and antibodies. However, the emergence of >5000 mutations of COVID-19 and the second and subsequent waves of the infection in various countries make the current situation extremely complex (HAAs et al., 2021). Baseline characteristics of patients with COVID-19, such as old age, obesity, heavy smoking status, and potential comorbidities (such as hypertension, respiratory disease, cardiovascular disease, and cancer), are associated with a higher risk of COVID-19-related mortality (Lee et al., 2022). Moreover, whether baseline physical activity (PA) is an essential factor that can change the symptoms of COVID-19 illness is a question that has attracted extensive attention from scholars.

Physical activity (PA) had been proved as one of the main factors in promoting health (Ekelund et al., 2019, Piercy et al., 2020, Chastin et al., 2020). The lack of adequate PA is associated with 6% to 10% of the major chronic diseases and a 9% higher risk of premature mortality worldwide (Lee et al., 2012). As we all know, adequate PA can significantly reduce all-cause and disease-specific mortality (such as death from pneumonia) (Kyu et al., 2016, Lu et al., 2022). Studies have shown that PA significantly improves human immunity and protects against severe and infectious respiratory tract infections. (Piercy et al., 2020, Chastin et al., 2020). In addition, PA may positively impact inflammation and is a potential causal factor for chronic diseases such as cardiovascular disease and cancer. Higher PA doses have a more positive link with disease inflammation impact (Zbinden‐Foncea et al., 2020). Bizuti et al., 2022 assume the regular practice of moderate-intensity physical activity is responsible for promoting a reduction in the concentrations of pro-inflammatory cytokines (IL-6, TNF-α and IL-1β), as well as triggering the increase in the production of anti-inflammatory cytokines (IL-4 and IL-10) (Bizuti et al., 2022).

However, previous studies have investigated the potential link between PA and COVID-19, with inconsistent results. Recent studies have found that physical activity is not associated with SARS-CoV-2 infectivity and COVID-19 infection but is significantly associated with COVID-19 severity and mortality (Rowlands et al., 2021). Pinto et al. (2021) found that baseline PA of hospitalized COVID-19 patients was not independently associated with the length of stay or any other clinically relevant outcome (severe illness and mortality). The authors concluded that the underlying diseases and ages of COVID-19 patients had attenuated the protective effect of PA against the risk of COVID-19 symptoms (Pinto et al., 2021). Research on this topic is just emerging, and the impact of PA on the risk of COVID-19 is still unclear, requiring systematic reviews and meta-analyses of relevant research data.

Rahmati et al. (2022) conducted a meta-analysis of studies on this topic. Although the effect of PA was calculated on the basis of the original data in this study, the actual cause of the impact of PA on severe COVID-19 and mortality was difficult to analyze from the results because confounding factors such as age, body weight, and related underlying diseases were not included in the analysis (Rahmati et al., 2022). In addition, this study took cardiopulmonary function as one measure of physical activity level, which made the results inconclusive. Finally, this study did not address the controversy regarding the protective effects of different levels of physical activity (Rahmati et al., 2022). Furthermore, A recent systematic review assumed the regular practice of adequate intensity is suggested as an auxiliary tool in strengthening and preparing the immune system for COVID-19 (da Silveira et al., 2021). However, no systematic review of the above empirical studies has been known to quantify the preventive effects of various doses of physical activity against severe illness and mortality from COVID-19. Therefore, to take into account the current international guidelines for physical activity that specifically aim to reduce the risk of COVID-19 illness burden, in our study, we conducted a dose–response meta-analysis to quantify the relationship between different physical exercise doses and the risk of severe illness and mortality from COVID-19.

For this study, baseline physical activity refers to the physical activity status of COVID-19 patients prior to infection. Metabolic equivalent of task (MET) used as dose in this study is a physiological indicator describing the body's metabolic equivalent of PA energy, defined as energy consumption per kilogram of body weight per hour: 1 MET = 1 kcal/kg ∗ h.

2. Methods

This study was conducted following the structured reporting procedures prescribed by PRISMA (International Prospective Register of Systematic Reviews Database Management Organization) (Page et al., 2021) and registered with PROSPERO; the number is CRD42022344701.

2.1. Eligibility criteria

The criteria for inclusion and each article determined for inclusion were discussed by the three authors. The inclusion criteria were as follows: (1) The studies were published as epidemiological observational cohort studies, case controls, and cross-sectional design investigation studies. (2) The studies provide the odds ratio, the relative risk (RR or HR), and the 95% confidence interval (95%CI) of physical activity levels related to the risk of severe illness and mortality from COVID-19（or provide raw data to calculate these indicators). Duplicates were excluded. Only the latest studies were selected if they were studied at different time points in the same cohort. In addition, if multiple articles were published in the same group, we chose papers that were conducted for a longer period of time or with a larger sample size. First, two authors decided to include the literature independently. Then, the three authors discussed the inconsistencies to determine whether to include the literature.

2.2. Search strategy

We searched PubMed and the Web of Science database (from January 2020 through April 2022) for any literature on the relationship between physical activity and COVID-19. The terms “exercise or physical activity or sport or walking or motor activity” and “COVID-19 or SARS-CoV-2″ and ”severe or ICU or mortality or die or death survival“ were used to search for the relevant literature. The last of the search was conducted in April 2022, and there was no language limit. The reference lists of selected and related review articles were screened step by step to identify potentially relevant studies. All searches were conducted independently by two authors, and differences were resolved by a group discussion. When the same population or cohort data were included in several works of literature, only the most recent literature or literature using the largest population was used in the meta-analysis.

2.3. Quality of evidence

The Newcastle–Ottawa Scale (NOS) was used to evaluate literature quality, and scores of 0 ∼ 3, 4 ∼ 6, and 7 ∼ 9 were determined as low, medium, and high quality, respectively (Wells et al., 2009). Each article was evaluated independently by two authors and cross checked. In the group meeting, the results were publicized and the reasons for the score of each item were specified. If the evaluation of literature quality was inconsistent, the group focused on solving the issue and identifying a final score in terms of its quality.

2.4. Data synthesis and analysis

Stata16.0 software was used for the meta-analysis. P values<0.05 were considered statistically significant, and all tests were double-sided. The effect sizes (adjusted or not adjusted), the risk ratios (RRs), and the 95% confidence intervals (CIs) of the group with the highest dose of PA compared with the group with the lowest dose of PA were combined and the combined effect values were calculated using a random effect model. Heterogeneity was assessed and described by I2 statistics as the percentage of variation in the study; I2 values of 25%, 50%, and 75% indicate low, moderate, and high levels of heterogeneity, respectively (Higgins and Green, 2011). Egger and Begg tests were used to determine any publication bias. During sensitivity analysis, each study was deleted one by one to check whether the combined effect of the remaining studies had changed (Higgins and Green, 2011). Subgroup meta-analysis was conducted by PA dose categories, gender, age, study area, study quality, and adjustment for confounding factors, and meta-regression was used to examine the heterogeneity among studies.

The categorical dose–response relationships were divided into dichotomous and multi-categorical doses, shown in the study. The combined effect value RR was generated by comparing the highest and lowest doses. For the analysis of the continuous dose–response relationship, we calculated the total weekly dose of PA for each effect value RR based on the PA intensity, duration, and weekly frequency of the baseline survey provided in the literature. We assumed that their dose of PA remained at this level in the follow-up survey. As for determining the exposure value of the included dose, we set the median as its determined dose. If the development interval was < 0.5, we set it as 0.25. If the upper open interval was greater than or equal to 1, the difference between the intermediate dose interval was 0.25, so the exposure value was set as 1.25 (Zhang et al., 2015). These are combined absolute indices of intensity, duration, and frequency and are also used to calculate exposures to MET units that are not directly reported in the literature. To address the differences in PA units in different studies, we adopted the classification method by Ainsworth et al.(2011), classifying PA into light PA (3 of MET-h, such as walking exercise), moderate PA (4 MET-h), and vigorous PA (8 MET-h). We then converted the duration of a particular PA intensity (h/week) to MET-h/week in combination with the frequency of the week (Ainsworth et al., 2011).

We used robust error meta-regression (REMR) (Xu and Doi, 2018) for model fitting. The REMR approach is based on a “one-stage” framework, treating each study as a cluster and fitting the revised regression to the average PA dose across the entire dataset. In addition, the method uses the inverse variance method to weight each dose–specific effect in the data and balances heteroscedasticity in the REMR model to ensure the unbiased estimation of parameters. Finally, we used restricted cubic splines as connection functions to fit linear and nonlinear dose–response models. On the basis of dose-centralization treatment, the independent variable PA dose of the model was set as three nodes (0, 6.75, and 21), including two regression splines. The χ2 test was used to test the hypothesis that the regression coefficient of the second regression spline is significant (p < 0.05), indicated by a linear or nonlinear dose–response relationship. A dose–response relationship curve was drawn using the Stata software XBLC command (Zhang et al., 2015).

3. Results

3.1. Description of included studies

Preliminarily, 270 items from the literature were detected. According to the literature inclusion and exclusion criteria formulated in this study, 8 cohort studies, 4 case–control studies, and 2 cross-sectional studies were finally included, involving a total of 560,078 subjects and about 2864 cases of severe illness and mortality related to COVID-19 (there were studies but no reported cases). Fig. 1 displays the steps for retrieval and inclusion. Table 1 shows literature features. Of the 14 works of literature, 2 are from North America (Brandenburg et al., 2021, Sallis et al., 2021), 6 are from Asia (Baynouna AlKetbi et al., 2021, Cho et al., 2021, Halabchi et al., 2021, Lee et al., 2022, Tavakol et al., 2021, Yuan et al., 2021), 5 are from Europe (Ahmadi et al., 2021, Ekblom-Bak et al., 2021, Hamrouni et al., 2021, Katsoulis et al., 2021, Salgado-Aranda et al., 2021), and 1 is from Oceania (Steenkamp et al., 2022). The NOS scale was used to score the included literature. The most problematic domains of study quality were selection for ascertainment of exposure, Comparability for outcome assessment, and follow-up rate. 4 studies did not report ascertainment of exposure, 3 did not report Confounding factors, and 3 cohort studies did not report follow-up rates, respectively. In all, 9 works of literature were of high quality, or equal to 7 points, and the others were of medium quality (see Supplementary A Table 1).

Fig. 1.

Fig. 1

Open in a new tab

Flow chart of study selection.

Table 1.

Summary of Included Studies.

Study
(year)		 Country Research
type		 Case\total Age Gender
(Female%)		 Dose of PA Adjusted factor*





















 Outcome Research
Quality
Ahmadi(2021) UK Cohort 387\468569 56.5 ± 8.1 54.6% 3 doses:High, medium, and low 1–15 Mortality 8
Baynouna(2021) AUH Cohort 135、16\641 average 44 36% 5 doses:exercise times NP* Mortality, Severe 6
Brandenburg(2021) CA Case-control NP\263 86%<65 57% 4 dose:High, medium, low, and no 2 5 6 8 9 12 15 Severe 7
Cho(2021) KR Case-control 75\ 125,780 50.7 ± 14.3 60.9% 5 doses:MET-min/week 1–3 9 14 16 17 Mortality 9
Ekblom-Bak(2021) SWE Case-control 172、138/407131
 average 49.9 30% 3 doses:exercise times 1 2 5 16 Mortality, Severe 7
Hamrouni(2021) UK cohort 397\ 259 397 37–73 55.1% 3 doses:High, medium, and low 1 2 4 5 16 17 Mortality 8
Halabchi(2021) Iran. Cross-section 60\79\4694 36.45 ± 9.77 55% 2 dose:sedentary and physically active NP Mortality, Severe 6
Katsoulis(2021) UK cohort NP\ 85,308 18–69 NP 3 doses:High, medium, and low 1 2 7 9 Severe 7
Lee (2021) KR cohort 277\118 768 NP 51.2% 4 dose:High, medium, low, and no 1 2 5–9 14 16 Mortality, Severe 9
Salgado(2021) ESP cohort 45\552 18–70 NP 2 dose:sedentary and physically NP Mortality 7
Sallis(2021) USA cohort 1199、771\2970 47 ± 16.97 61.9% 3 doses:High, medium, and low 1–15 Mortality, Severe 9
Steenkamp(2021) ZA cohort NP\65361 41 ± 12.1 48.2% 3 doses:High, medium, and low 1 2 6 8 9 14 15 Mortality, Severe 8
Tavakol(2021) Iran Cross-section 64\188 18–75 52.7% 2 dose:sedentary and physically NP Severe 6
Yuan(2021) CN Case-control 29\164 61.8 ± 13.6 48.8% 2 dose:sedentary and physically NP Mortality, Severe 6
Open in a new tab

*Case\total: number of cases and total sample size; Age characteristics: A single value indicates the average age, and the others are age ranges. * Correction for confounding factors: 1 age, 2 gender, 3 socioeconomic status, 4 race, 5BMI, 6 cardiovascular disease, 7 cancer, 8 diabetes, 9 hypertension, 10 use of antihypertensive drugs, 11 use of corticosteroids, 12 chronic lung/respiratory disease, 13 liver disease, 14 HIV, 15 end-stage kidney and immune diseases, 16 smoking, 17 alcohol consumption. NP*: Not reported. ZA: South Africa, UK: United Kingdom, AUH: United Arab Emirates, CA: Canada, SWE: Switzerland, Iran, KR: South Korea, ESP: Spain, USA, ZA: New Zealand, CN: China。.

3.2. Categorical dose–response relationship between PA and the risk of severe COVID-19 illness

Compared with the lowest dose of PA, the highest dose of PA can reduce the risk of severe COVID-19 by 46% (RR = 0.54; 95%CI: 0.41–0.68), and the heterogeneity test result is I2 = 69.03% (p < 0.01), showing that there is significant heterogeneity in the research results (as shown in Fig. 2). On excluding each study-one by one, no significant change was found in the results of the combined effect. Published bias analysis with the Begg test (p = 0.54 > 0.05), the Egger test (p = 0.70 > 0.05), and the funnel diagram also showed no significant published bias (see Supplementary B Fig. 1).

Fig. 2.

Fig. 2

Open in a new tab

The relationships of PA and Covid-19 illness severe.

As for the source of heterogeneity, between-groups heterogeneity appeared in dose categories, adjusted confounding BMI factors indicating that these factors affected the overall effect value and that different levels of PA had significantly different effects on the risk of severe COVID-19 illness (as shown in Table 2). The results of heterogeneity within subgroups showed that there was significant heterogeneity within the subgroups of overall effect value, high vS low dose comparison, moderate vS low dose comparison, PA Questionnaire for measurement, case–control group, high-quality research papers, European population, adjusted confounding factors (age, sex, and BMI), and unadjusted comorbidities hypertension(NO), diabetes, Cancer indicating that the research results of these subgroups may be affected by other factors.

Table 2.

Results of the subgroup analysis.

subgroup Severe risk
 Mortality risk
N RR(95% CI) I2 (%) pa* Pb* N RR(95% CI) I2 (%) pa* Pb*
total(highest vs lowest) 14 0.54(0.41,0.68) 69 0.01> 10 0.41(0.23,0.58) 83 0.01>
Dose Categories
 binary * 4 0.36(0.07,0.66) 35 0.28 0.01> 3 0.13(-0.02,0.28) 0 1 0.01
H vS L * 10 0.59(0.44,0.73) 70 0.01> 7 0.47(0.27,0.67) 84 0.01>
M vS L * 15 0.75(0.63.0.87) 29 0.01 7 0.73(0.63,0.83) 16 0.14
Research type Cohort 9 0.50(0.41,0.59) 28 0.6 0.61 6 0.38(0.14,0.62) 90 0.01> 0.60
Case-control 3 0.66(0.11,1.21) 85 0.01> 3 0.52(0.23,0.82) 44 0.23
Crosss-section 2 0.31(-0.14,0.76) 4 0.58 1 0.11(-0.79,1.00)
Research quality 》7 9 0.59(0.43,0.74) 75 0.01> 0.14 7 0.47(0.28,0.67) 85 0.01> 0.01
7< 5 0.36(0.09,0.63) 25 0.43 3 0.09(-0.10,0.28) 0 1
Measurement of PA Questionnaire 12 0.58(0.42,0.74) 54 0.01 0.09 9 0.40(0.20,0.60) 75 0.01> 0.80
Objective 2 0.44(0.38,0.49) 2 0.64 1 0.43(0.36,0.49)
Different continent Europe 5 0.71(0.45,0.96) 62 0.01 0.17 4 0.56(0.24,0.88) 86 0.01> 0.01>
Asia 5 0.41(0.17,0.65) 31 0.16 4 0.12(-0.06,0.29) 0 0.9
Others 4 0.46(0.35,0.56) 30 0.58 2 0.43(0.36,0.49) 0 0.83
Confounding factor
Age Yes 8 0.59(0.42,0.76) 80 0.00 0.21 6 0.54(0.36,0.73) 74 0.01> 0.01>
No 6 0.40(0.15,0.64) 21 0.43 4 0.12(-0.01,0.24) 0 0.9
Sex Yes 9 0.59(0.43,0.74) 75 0.01> 0.14 6 0.54(0.36,0.73) 74 0.01> 0.01>
No 5 0.36(0.09,0.63) 22 0.43 4 0.12(-0.01,0.24) 0 0.9
BMI Yes 4 0.76(0.49,1.02) 66 0.02 0.03 4 0.64(0.44,0.85) 57 0.07 0.01>
No 10 0.43(0.31,0.55) 40 0.68 6 0.25(0.11,0.38) 51 0.01>
comorbidities
hypertension Yes 8 0.49(0.40,0.58) 29 0.58 0.60 4 0.50(0.24,0.76) 82 0.01 0.34
No 6 0.58(0.25,0.92) 63 0.01> 6 0.33(0.09,0.56) 71 0.01>
Diabetes Yes 9 0.57(0.35,0.79) 59 0.01> 0.54 3 0.56(0.27,0.85) 85 0.01> 0.17
No 5 0.49(0.38,0.60) 40 0.33 7 0.31(0.11,0.52) 65 0.01>
angiocardiopathy Yes 5 0.49(0.38,0.60) 40 0.33 0.54 5 0.51(0.31,0.72) 76 0.01 0.15
No 9 0.57(0.35,0.60) 59 0.01> 5 0.26(0.00,0.53) 70 0.01
Cancer Yes 5 0.57(0.43,0.70) 0 0.84 0.93 2 0.64(0.20,1.08) 84 0.01> 0.21
No 9 0.53(0.31,0.74) 75 0.01> 8 0.34(0.16,0.51) 76 0.01>
Open in a new tab

* Note: binary is generally expressed as exercise and non-exercise, or physical activity and inactivity; * H vS L represent high vs low amounts of PA when multiple categories include the high, moderate, low or more categorical doses; M vS L represent the various subcategories of multiple categories: moderate vs low. Pa and Pb represent heterogeneity within and between subgroups, respectively.

3.3. Categorical dose–response relationship between PA and the risk of COVID-19 mortality

Compared with the lowest-dose PA, the highest-dose PA can reduce the risk of COVID-19 mortality by 59% (RR = 0.41; 95%CI: 0.23–0.58) and the heterogeneity test result is I2 = 82.43% (p < 0.01), showing that the research results have significant heterogeneity (as shown in Fig. 3). In the sensitivity analysis, on excluding each study-one by one, no significant change was found in the results of the combined effect quantity. Published bias analysis with the Begg test (p = 0.65 > 0.05) and the Egger test (p = 1.1420 > 0.05) indicated no significant release bias (see Supplementary B Fig. 2).

Fig. 3.

Fig. 3

Open in a new tab

The relationships of PA and Covid-19 mortality.

As for the source of heterogeneity, between-groups heterogeneity appeared in Multiple dose comparison, Research quality, Different continent, adjusted confounding factors (age, sex, and BMI), which means that these factors may affect the overall effect value and that different levels of PA had significantly different effects on the risk of COVID-19 mortality (as shown in Table 2). The results of heterogeneity within subgroups showed that there was significant heterogeneity in total effect values, high vS low dose comparison, Questionnaire, Europe cohort studies, high-quality research papers, European population subgroups, adjusted confounding factor (age, gender, and BMI(no)), and unadjusted comorbidities (hypertension, diabetes,and cardiovascular disease), indicating that the research results of these subgroups may be affected by other factors.

3.4. Continuous dose–response relationship between PA and severe COVID-19 and the risk of mortality

Fig. 4 shows a negative nonlinear response relationship between PA and the risk of severe COVID-19 (pnon-linearity < 0.05). When PA was lower than 10 MET-h/week (the lowest physical activity dose recommended by the WHO), the effect size RR of severe COVID-19 lowered by 8% for each increase of 4 MET-h/week (medium intensity for 1 h or high intensity for 0.5 h) (p < 0.01; RR = 0.92; 95%CI: 0.88–0.97). When PA was higher than 10 MET-h/week, the RR of severe COVID-19 lowered by 7% for each increase of 4 MET-h/week (p < 0.01; RR = 0.93; 95%CI: 0.90–0.96).

Fig. 4.

Fig. 4

Open in a new tab

The continuous dose response relationship between PA and the risk of COVID-19 severe illness.

Fig. 5 shows a negative nonlinear response relationship between PA and the risk of COVID-19 mortality (pnon linearity < 0.05). When PA was lower than 10 MET-h/week, the risk of COVID-19 mortality lowered by 11% for each increase of 4 MET-h/week (medium intensity for 1 h or high intensity for 0.5 h) (p < 0.01; RR = 0.89; 95%CI: 0.88–0.90). When PA was higher than 10 MET-h/week, the risk of COVID-19 mortality lowered by 9% for each increase of 4 MET-h/week (p < 0.01; RR = 0.91; 95%CI: 0.90–0.92).

Fig. 5.

Fig. 5

Open in a new tab

The continuous dose response relationship between PA and the risk of COVID-19 mortality.

4. Discussion

This study is the first dose–response meta-analysis on the relationship between PA and the risk of severe illness and mortality from COVID-19. Included are the cohort, case–control, and cross-sectional studies on the relationship between PA and the risk of severe illness and mortality from COVID-19. Through the categorical dose meta-analysis, our main conclusion is that with the highest dose of PA, the risk of severe illness and mortality from COVID-19 is reduced by 46% and 59%, respectively, compared with the lowest dose of PA. The significant heterogeneity in the different dose groups (dichotomous and multiclass) indicates that the protective effects of different levels of PA against the risk of severe illness and mortality from COVID-19 are significantly different. Furthermore, we assume that the dose–response relationship between PA and the risk of severe illness and mortality from COVID-19 is nonlinear and negative for continuous dose–response analysis.

The dose of PA involves complex factors, such as intensity, time, and frequency, so we analyze the dose–response relationship according to classification and continuous dose. The two-classification dose–response relationship explains the difference between exercise and nonexercise, and the multi-classification dose–response relationship explains the dose effect of increasing PA relative to low-dose PA. This result is the same as that by Lu et al. (2022), by Kunutsor et al. (2022) on the mortality risk of general pneumonia, and by Rahmati et al. (2022) on the severity and mortality analysis of COVID-19 (Lu et al., 2022, Kunutsor et al., 2022, Rahmati et al., 2022). However, this study confirmed that the protective effect of different doses of PA against the risk of severe illness and mortality in patients with COVID-19 is significantly different. The result of this study is different from the conclusion of Rahmati et al. (2022) related to meta-analysis (Rahmati et al., 2022). In addition, this study included more literature, and the conclusion is more robust. The fitting of the continuous dose also makes the classified dose further confirm the results of categorical dose analysis. The higher level of PA, the stronger the protective effect against the risk of severe mortality in patients with COVID-19. There is a negative nonlinear relationship between PA and the risk of severe mortality in patients with COVID-19. The continuous dose–response analysis adds the intermediate dose to the multi-category dose for comparative analysis, which ensures that the conclusion of this research topic is more specific and systematic. A cohort study found a curvilinear association between healthy and CVRF (cardiovascular risk factors) individuals with a steep risk reduction at low to moderate MVPA volumes (Bakker et al., 2021). Moreover, a recent meta-analysis also confirmed a non-linear association with mortality risk reductions even for low levels of activity and a flattening of the curve at higher levels of post-diagnosis PA in adults with noncommunicable diseases (Geidl et al., 2020). Our study has the same result: low levels of PA have higher benefits for the risk of severe illness and mortality in patients with COVID-19. As shown in Fig. 5, 10 MET-h/week is almost nearby the inflection point. Furthermore, a previous dose–response meta-analysis of reduction in all-cause mortality from walking and cycling found the inflection point is about 11.25 MET-h/week.

The WHO global recommendation on the health benefits of PA points out that adults should perform at least 150 min of moderate-intensity aerobic PA per week, at least 75 min of high-intensity aerobic PA per week, or a combination of moderate- and high-intensity activities. These doses are equivalent to 10 MET-h/week (World Health Organization, 2022a, World Health Organization, 2022b). Analysis of the continuous dose–response relationship in this study shows that PA and exercise of about 10 MET-h/week are the most obvious in reducing the risk of severe illness and mortality from COVID-19. The practical significance is that the minimum dose of 10 MET-h/week to prevent severe illness and mortality from COVID-19 is equivalent to the dose of physical activity recommended by the WHO (World Health Organization, 2022a, World Health Organization, 2022b). At the same time, when the PA level exceeds 10 MET-h/week, the degree of increase in the protective effect against the risk of severe illness and mortality from COVID-19 is significantly reduced.

However, additional protective effects are still obtained, which is also consistent with the recommendations of the WHO physical activity guidelines (World Health Organization, 2022a, World Health Organization, 2022b). A dose–response meta-analysis by Geidl et al. (2020) revealed that each 10 MET-h/week increase in PA was associated with a 22% lower mortality rate in adults with noncommunicable diseases (Geidl et al., 2020). However, 10 MET-h/week increase in PA is not practical. Our study confirmed that the practical significance of increasing the dose of PA by 4 MET-h/week per week is that exercising more than once a week (1 h of moderate PA or 0.5 h of vigorous PA) will reduce the risk of severe illness and mortality from COVID-19 by 7%-11%. Research by Williams et al. (2014) assumed that the mechanism of PA reducing mortality from general pneumonia might not only involve the root cause of the impact on pneumonia but also affect whether pneumonia leads to fatal consequences from other underlying diseases (Williams et al., 2014). For example, because pneumonia itself causes heart problems (Benson et al., 1970), such as left ventricular dysfunction (Corrales-Medina et al., 2011), or increases arrhythmias (Corrales-Medina et al., 2012), people with pneumonia and cardiac complications are more likely to die (Nieman et al., 1990). Therefore, PA may have more significant health benefits for patients with heart disease. It can enhance the mucosal immune function of patients and reduce the risk of mortality in patients with pneumonia and cardiac complications (Kimura et al., 2006). At the same time, the heterogeneity between diabetes subgroups in this study is close to significant, showing that PA may have a greater effect on the remission of diabetes in patients. PA can effectively regulate insulin and glucose metabolism, which is the protective mechanism of its health-promoting effect (Kunutsor et al., 2022). In addition, moderate PA, such as walking, enhances immune function by increasing the activity of macrophages, natural killer cells (Shephard et al., 1995), and neutrophils and regulating cytokines. These effects on systemic health may be the mechanism by which PA protects against severe illness and mortality from COVID-19 (Shephard and Shek, 1995).

The COVID-19 epidemic has led to a continuous increase in the number of severe cases and mortality, posing a threat to personal health and a heavy disease burden on all aspects of regions and countries. We found that increasing physical activity can significantly reduce the disease risk of COVID-19 and physical activity should be a positive factor in reducing the disease burden of COVID-19. However, the global burden of disease (GBD) 2019 ranked low physical activity 19th among the 20 most risky lifestyles leading to chronic diseases, lower than 10th place in the same report published in 2010 (Collaborators, 2020, Stamatakis et al., 2021). In addition, according to a recent meta-analysis (Wunsch et al., 2022), due to the COVID-19 pandemic, PA in all age groups of the general population has lowered. Facing the global spread of COVID-19, we must accept the vital role of physical activity in reducing the burden of the disease.

Finally, our study is the first meta-analysis study on the dose–response relationship between PA and severe illness and mortality from COVID-19. The results of this study are based on a large-sample cohort study and a case–control group study, as well as the advantages of years of follow-up investigation, so the results are relatively robust. However, this study may have the following limitations: (1) The literature we included may be insufficient. The possible reason is that we set strict inclusion criteria. In addition, cohort studies require long-term follow-up surveys and the requirements of extensive sample data, so there are relatively few cohort studies on related topics. Therefore, we can only set three nodes in the restrictive cubic bar to produce two regression splines to judge whether the overall study is linear or nonlinear according to the second regression spline. Such treatment may not verify the specific linear or nonlinear conclusion of PA dose after 10 MET-h/week, and we need to explain this conclusion carefully. Therefore, more in-depth studies on the protective factors of high-intensity PA against mortality from COVID-19 are required for the future. (2) The methods of PA evaluation included in the literature of this study are subjective measurements, and the dose may be inaccurate. In addition, the habit of exercise or sport is based on the assumption that there is no change in the long-term follow-up. The dose of PA is assumed to remain unchanged in the long-term observational study. This will make the results inaccurate, but some cohort studies have not been corrected or classified, which should be paid attention to in future related research.

5. Conclusion

There was a significant nonlinear negative dose–response relationship between the level of PA and the risk of severe illness and mortality from COVID-19. Various doses of PA were protective factors against severe illness and mortality from COVID-19. The higher physical activity dose significantly reduces the risk of severe illness and mortality from COVID-19. The degree of risk reduction is more significant when the total physical activity dose is lower than 10 MET-h/week, and the degree of enhancement is weakened when PA is higher than 10 MET-h/week. Doses of WHO-recommended PA levels may be required for more substantial reductions in severe illness and mortality from COVID-19.

6. Ethical compliance

The underlying work is based on systematic reviews of published data and thus does not require ethical review approval.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pmedr.2023.102130.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary figure 1.

Supplementary figure 1

Open in a new tab

Supplementary figure 2.

Supplementary figure 2

Open in a new tab
Supplementary data 1
mmc1.docx (20.1KB, docx)

Data availability

Data will be made available on request.

References

  1. Ahmadi M.N., Huang B.-H., Inan-Eroglu E., et al. Lifestyle risk factors and infectious disease mortality, including COVID-19, among middle aged and older adults: Evidence from a community-based cohort study in the United Kingdom. Brain Behav. Immun. 2021;96:18–27. doi: 10.1016/j.bbi.2021.04.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ainsworth B.E., Haskell W.L., Herrmann S.D., et al. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med. Sci. Sports Exerc. 2011;43(8):1575–1581. doi: 10.1249/MSS.0b013e31821ece12. [DOI] [PubMed] [Google Scholar]
  3. Bakker E.A., Lee D.C., Hopman M.T.E., Oymans E.J., Watson P.M., Thompson P.D., Eijsvogels T.M.H. Dose-response association between moderate to vigorous physical activity and incident morbidity and mortality for individuals with a different cardiovascular health status: A cohort study among 142,493 adults from the Netherlands. PLoS Med. 2021;18(12) doi: 10.1371/journal.pmed.1003845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baynouna AlKetbi L.M., Nagelkerke N., Abdelbaqi H., et al. Risk Factors for SARS-CoV-2 Infection Severity in Abu Dhabi. J Epidemiol Glob Health. 2021;11(4):344–353. doi: 10.1007/s44197-021-00006-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Benson H., Akbarian M., Adler L.N., et al. Hemodynamic effects of pneumonia. I. Normal and hypodynamic responses. J. Clin. Invest. 1970;49(4):791–798. doi: 10.1172/JCI106292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bizuti M.R., Starck É., da Silva Fagundes K.K., et al. Influence of exercise and vitamin D on the immune system against Covid-19: an integrative review of current literature. Mol. Cell. Biochem. 2022;477(6):1725–1737. doi: 10.1007/s11010-022-04402-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brandenburg J.P., Lesser I.A., Thomson C.J., et al. Does higher self-reported cardiorespiratory fitness reduce the odds of hospitalization from COVID-19? J. Phys. Act. Health. 2021;18(7):782–788. doi: 10.1123/jpah.2020-0817. [DOI] [PubMed] [Google Scholar]
  8. Chastin S., Abaraogu U., Bourgois J., et al. Physical activity, immune function and risk of community acquired infectious disease in the general population: systematic review and meta-analysis. SSRN Electron. J. 2020 doi: 10.2139/ssrn.3673184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cho D.-H., Lee S.J., Jae S.Y., et al. Physical activity and the risk of COVID-19 infection and mortality: a nationwide population-based case-control study. J. Clin. Med. 2021;10(7):1539. doi: 10.3390/jcm10071539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Collaborators G.R.F. Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet (London, England) 2020;396(10258):1223–1249. doi: 10.1016/S0140-6736(20)30752-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Corrales-Medina V.F., Suh K.N., Rose G., et al. Cardiac complications in patients with community-acquired pneumonia: a systematic review and meta-analysis of observational studies. PLoS Med. 2011;8(6):e1001048. doi: 10.1371/journal.pmed.1001048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Corrales-Medina V.F., Musher D.M., Wells G.A., et al. Cardiac complications in patients with community-acquired pneumonia: incidence, timing, risk factors, and association with short-term mortality. Circulation. 2012;125(6):773–781. doi: 10.1161/CIRCULATIONAHA.111.040766. [DOI] [PubMed] [Google Scholar]
  13. da Silveira M.P., da Silva Fagundes K.K., Bizuti M.R., et al. Physical exercise as a tool to help the immune system against COVID-19: an integrative review of the current literature. Clin. Exp. Med. 2021;21(1):15–28. doi: 10.1007/s10238-020-00650-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ekblom-Bak E., Väisänen D., Ekblom B., Blom V., Kallings L.V., Hemmingsson E., Andersson G., Wallin P., Salier Eriksson J., Holmlund T., Lindwall M., Stenling A., Lönn A. Cardiorespiratory fitness and lifestyle on severe COVID-19 risk in 279,455 adults: a case control study. Int. J. Behav. Nutr. Phys. Act. 2021;18(1) doi: 10.1186/s12966-021-01198-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ekelund U., Tarp J., Steene-Johannessen J., et al. Dose-response associations between accelerometry measured physical activity and sedentary time and all cause mortality: systematic review and harmonised meta-analysis. BMJ. 2019;366 doi: 10.1136/bmj.l4570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Geidl W., Schlesinger S., Mino E., Miranda L., Pfeifer K. Dose-response relationship between physical activity and mortality in adults with noncommunicable diseases: a systematic review and meta-analysis of prospective observational studies. Int. J. Behav. Nutr. Phys. Act. 2020;17(1):109. doi: 10.1186/s12966-020-01007-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Haas E.J., Angulo F.J., McLaughlin J.M., et al. Impact and effectiveness of mRNA BNT162b2 vaccine against SARS-CoV-2 infections and COVID-19 cases, hospitalizations, and deaths following a nationwide vaccination campaign in Israel: an observational study using national surveillance data. Lancet. 2021;397(10287):1819–1829. doi: 10.1016/s0140-6736(21)00947-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Halabchi F., Mazaheri R., Sabeti K., Yunesian M., Alizadeh Z., Ahmadinejad Z., Aghili S.M., Tavakol Z. Regular sports participation as a potential predictor of better clinical outcome in adult patients with COVID-19: A large cross-sectional study. J. Phys. Act. Health. 2021;18(1):8–12. doi: 10.1123/jpah.2020-0392. [DOI] [PubMed] [Google Scholar]
  19. Hamrouni M., Roberts M.J., Thackray A., Stensel D.J., Bishop N. Associations of obesity, physical activity level, inflammation and cardiometabolic health with COVID-19 mortality: a prospective analysis of the UK Biobank cohort. BMJ Open. 2021;11(11):e055003. doi: 10.1136/bmjopen-2021-055003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Higgins P.T., Green S. New Jerse: A Version 5.1.0. The Johu Wiley Sons, Ltd Publication; 2011. Cochrane handbook for systematic reviews of interventions. [Google Scholar]
  21. Katsoulis M., Pasea L., Lai A.G., Dobson R.J.B., Denaxas S., Hemingway H., Banerjee A. Obesity during the COVID-19 pandemic: both cause of high risk and potential effect of lockdown? A population-based electronic health record study. Public Health. 2021;191:41–47. doi: 10.1016/j.puhe.2020.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kimura F., Shimizu K., Akama T., Akimoto T., Kuno S., Kono I. The effects of walking exercise training on immune response in elderly subjects. Internat. J. Sport Health Sci. 2006;4:508–514. [Google Scholar]
  23. Kunutsor S.K., Seidu S., Laukkanen J.A. Physical activity reduces the risk of pneumonia: systematic review and meta-analysis of 10 prospective studies involving 1,044,492 participants. Geroscience. 2022;44(1):519–532. doi: 10.1007/s11357-021-00491-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kyu H.H., Bachman V.F., Alexander L.T., et al. Physical activity and risk of breast cancer, colon cancer, diabetes, ischemic heart disease, and ischemic stroke events: systematic review and dose-response meta-analysis for the Global Burden of Disease Study 2013. BMJ. 2016;354 doi: 10.1136/bmj.i3857. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lee S.W., Lee J., Moon S.Y., Jin H.Y., Yang J.M., Ogino S., Song M., Hong S.H., Abou Ghayda R., Kronbichler A., Koyanagi A.i., Jacob L., Dragioti E., Smith L., Giovannucci E., Lee I.-M., Lee D.H., Lee K.H., Shin Y.H., Kim S.Y., Kim M.S., Won H.-H., Ekelund U., Shin J.I., Yon D.K. Physical activity and the risk of SARS-CoV-2 infection, severe COVID-19 illness and COVID-19 related mortality in South Korea: a nationwide cohort study. Br. J. Sports Med. 2022;56(16):901–912. doi: 10.1136/bjsports-2021-104203. [DOI] [PubMed] [Google Scholar]
  26. Lee I.-M., Shiroma E.J., Lobelo F., Puska P., Blair S.N., Katzmarzyk P.T. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838):219–229. doi: 10.1016/S0140-6736(12)61031-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lu S., Liu W., Li R., et al. Leisure time physical activity and risk of pneumonia mortality: a dose-response meta-analysis. J. Sports Med. Phys. Fitness. 2022;62(4):547–553. doi: 10.23736/s0022-4707.21.12017-1. [DOI] [PubMed] [Google Scholar]
  28. Nieman D., Nehlsen-Cannarella S., Markoff P., Balk-Lamberton A., Yang H., Chritton D., Lee J., Arabatzis K. The effects of moderate exercise training on natural killer cells and acute upper respiratory tract infections. Int. J. Sports Med. 1990;11(06):467–473. doi: 10.1055/s-2007-1024839. [DOI] [PubMed] [Google Scholar]
  29. Page M.J., McKenzie J.E., Bossuyt P.M., et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. J. Clin. Epidemiol. 2021;134:178–189. doi: 10.1016/j.jclinepi.2021.03.001. [DOI] [PubMed] [Google Scholar]
  30. Piercy K.L., Troiano R.P., Ballard R.M., et al. The physical activity guidelines for Americans. J. Am. Med. Assoc. 2020;320(19):2020. doi: 10.1001/jama.2018.14854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pinto A.J., Goessler K.F., Fernandes A.L., Murai I.H., Sales L.P., Reis B.Z., Santos M.D., Roschel H., Pereira R.M.R., Gualano B. No independent associations between physical activity and clinical outcomes among hospitalized patients with moderate to severe COVID-19. J. Sport Health Sci. 2021;10(6):690–696. doi: 10.1016/j.jshs.2021.08.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rahmati M., Shamsi M.M., Khoramipour K., Malakoutinia F., Woo W., Park S., Yon D.K., Lee S.W., Shin J.I., Smith L. Baseline physical activity is associated with reduced mortality and disease outcomes in COVID-19: A systematic review and meta-analysis. Rev. Med. Virol. 2022;32(5) doi: 10.1002/rmv.2349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Rowlands A.V., Kloecker D.E., Chudasama Y., et al. Association of timing and balance of physical activity and rest/sleep with risk of COVID-19: A UK Biobank Study. Mayo Clin. Proc. 2021;96(1):156–164. doi: 10.1016/j.mayocp.2020.10.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Salgado-Aranda R., Pérez-Castellano N., Núñez-Gil I., et al. Influence of Baseline Physical Activity as a Modifying Factor on COVID-19 Mortality: A Single-Center. Retrospective Study. Infect Dis Ther. 2021;10(2):801–814. doi: 10.1007/s40121-021-00418-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sallis R., Young D.R., Tartof S.Y., et al. Physical inactivity is associated with a higher risk for severe COVID-19 outcomes: a study in 48 440 adult patients. Br. J. Sports Med. 2021;55(19):1099–1105. doi: 10.1136/bjsports-2021-104080. [DOI] [PubMed] [Google Scholar]
  36. Shephard R., Rhind S., Shek P. The impact of exercise on the immune system: NK cells, interleukins 1 and 2, and related responses. Exerc. Sport Sci. Rev. 1995;23:215–241. [PubMed] [Google Scholar]
  37. Shephard R.J., Shek P.N. Cancer, immune function, and physical activity. Can. J. Appl. Physiol. = Revue canadienne de physiologie appliquee. 1995;20(1):1–25. doi: 10.1139/h95-001. [DOI] [PubMed] [Google Scholar]
  38. Stamatakis E., Ding D., Ekelund U., Bauman A.E. Sliding down the risk factor rankings: reasons for and consequences of the dramatic downgrading of physical activity in the Global Burden of Disease 2019. Br. J. Sports Med. 2021;55(21):1222–1223. doi: 10.1136/bjsports-2021-104064. [DOI] [PubMed] [Google Scholar]
  39. Steenkamp L., Saggers R.T., Bandini R., et al. Small steps, strong shield: directly measured, moderate physical activity in 65 361 adults is associated with significant protective effects from severe COVID-19 outcomes. Br. J. Sports Med. 2022;56(10):568–576. doi: 10.1136/bjsports-2021-105159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Tavakol Z., Ghannadi S., Tabesh M.R., Halabchi F., Noormohammadpour P., Akbarpour S., Alizadeh Z., Nezhad M.H., Reyhan S.K. Relationship between physical activity, healthy lifestyle and COVID-19 disease severity; a cross-sectional study. Z Gesundh Wiss. 2021;31(2):267–275. doi: 10.1007/s10389-020-01468-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wells GA, Peterson J, Welch V, et al., 2009. The Newcastle-Ottawa scale (NOS) for assessing the quality in nonrandomized studies in meta-analyses, Ottawa, University of Ottawa. Available from: http://www. ohri. ca/ programs/ clinical_ epidemiology/ oxford. Asp.
  42. Williams P. Dose-response relationship between exercise and respiratory disease mortality. Med. Sci. Sports Exerc. 2014;46(4):711–717. doi: 10.1249/mss.0000000000000142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. World Health Organization, 2022. Global Recommendations on Physical Activity for Health. [2022-07-15]. https://www.who.int/publications/i/item/9789241599979. [PubMed]
  44. World Health Organization, 2022. Coronavirus disease (COVID-19) Numbers at a glance[R/OL]. [2022-07-15]. https://www.who.int/emergencies/diseases/novel-coronavirus-2019.
  45. Wunsch K., Kienberger K., Niessner C. Changes in Physical Activity Patterns Due to the Covid-19 Pandemic: A Systematic Review and Meta-Analysis. Int. J. Environ. Res. Public Health. 2022;19(4) doi: 10.3390/ijerph19042250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Xu C., Doi S. The robust error meta-regression method for dose-response meta-analysis. Int. J. Evid. Based Healthc. 2018;16(3):138–144. doi: 10.1097/xeb.0000000000000132. [DOI] [PubMed] [Google Scholar]
  47. Yuan Q., Huang H.-Y., Chen X.-L., Chen R.-H., Zhang Y., Pan X.-B., Chen J.-N., Liu N., Du H. Does pre-existent physical inactivity have a role in the severity of COVID-19? Ther. Adv. Respir. Dis. 2021;15 doi: 10.1177/17534666211025221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zbinden‐Foncea H., Francaux M., Deldicque L., Hawley J.A. Does high cardiorespiratory fitness confer some protection against proinflammatory responses after infection by SARS‐CoV‐2? Obesity. 2020;28(8):1378–1381. doi: 10.1002/oby.22849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Zhang T., Dong S., Zhou Z., et al. The advanced meta-analysis method-based on stata. Fudan Univ. Press. 2015:220–232. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data 1
mmc1.docx (20.1KB, docx)

Data Availability Statement

Data will be made available on request.

Articles from Preventive Medicine Reports are provided here courtesy of Elsevier

RESOURCES