Association between physical activity and risk of gastroesophageal reflux disease: A systematic review and meta-analysis

Highlights

• •Physical activity (PA) is associated with a lower risk (about 20% decrease) of gastroesophageal reflux disease (GERD).
• •PA exhibits a heightened protective effect against prevalent GERD in elder populations and smokers.
• •Participating in a minimum of recommended PA levels (150 minutes per week) can significantly reduce the prevalence of GERD.

Abstract

Background

Lifestyle plays an important role in preventing and managing gastroesophageal reflux disease (GERD). In response to the conflicting results in previous studies, we performed a systematic review and meta-analysis to investigate this association.

Methods

Relevant studies published until January 2023 were retrieved from 6 databases, and the prevalence of symptomatic gastroesophageal reflux (GER) or GERD was determined from the original studies. A random effects model was employed to meta-analyze the association by computing the pooled relative risk (RR) with 95% confidence intervals (95%CIs). Furthermore, subgroup and dose–response analyses were performed to explore subgroup differences and the association between cumulative physical activity (PA) time and GERD.

Results

This meta-analysis included 33 studies comprising 242,850 participants. A significant negative association was observed between PA and the prevalence of symptomatic GER (RR = 0.74, 95%CI: 0.66–0.83; p < 0.01) or GERD (RR = 0.80, 95%CI: 0.76–0.84; p < 0.01), suggesting that engaging in PA might confer a protective benefit against GERD. Subgroup analyses consistently indicated the presence of this association across nearly all subgroups, particularly among the older individuals (RR_{<40 years}:RR_{≥40 years} = 0.85:0.69, p < 0.01) and smokers (RR_smoker:RR_non-smoker = 0.67:0.82, p = 0.03). Furthermore, a dose–response analysis revealed that individuals who engaged in 150 min of PA per week had a 72.09% lower risk of developing GERD.

Conclusion

Maintaining high levels of PA decreased the risk of GERD, particularly among older adults and smokers. Meeting the recommended PA level of 150 min per week may significantly decrease the prevalence of GERD.

Graphical abstract

1. Introduction

Gastroesophageal reflux disease (GERD) is one of the most common gastrointestinal disorders, with an estimated overall prevalence of approximately 10%–15% of the global population.^1,2 Persistent GER not only diminishes patients’ quality of life but also elevates the risks of complications such as Barrett's esophagus and esophageal adenocarcinoma,³ thereby making GERD a significant health concern. Consequently, understanding the relationships between modifiable risk factors and GERD is critical for clinicians to develop more personalized interventions and treatment plans that meet the specific needs of patients and optimize their health outcomes.

Physical activity (PA) has increasingly been recognized over the years for its contribution to preventing and controlling chronic illnesses.^4,5 While prior evidence has suggested that PA can help reduce the risk of premature death and over 25 chronic diseases,⁶ the protective role of PA in GERD remains a topic of debate. Furthermore, although the Montreal Consensus states that the symptoms of GERD may be induced during PA,⁷ there is currently a lack of concrete evidence establishing a clear association between PA and long-term management of GERD. Different studies have reported conflicting findings in this regard. While a study indicated that PA might strengthen the diaphragm, leading to improved anti-reflux barrier function,⁸ another study found that exercising 1–3 times/week increased the risk of developing GERD.⁹ Additionally, other studies have shown that regular PA acts as a protective factor against GERD among individuals with obesity, but not among individuals with a normal body mass index (BMI).¹⁰ Considering these mixed observations, the role of PA in developing and managing GERD requires further investigation, taking into account individualized factors such as BMI and exercise frequency.

Despite numerous studies investigating the association between PA and the prevalence of GERD, there are limited systematic reviews and meta-analyses in this domain. Only a small fraction has mentioned the relationship between PA and GERD, highlighting the need for more targeted reviews and discussions on this topic.^11,12 Accordingly, we conducted this study to comprehensively summarize and pool the available evidence from observational studies to measure the relative risk (RR) between PA and the prevalence of GERD. Additionally, we conducted a dose–response meta-analysis to investigate the potential association between the duration of PA per week and the prevalence of GERD in the general population.

2. Methods

2.1. Protocol and registration

We formulated and adhered to a procedure for all steps of our study, which was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42022377255) (Supplementary Table 1). This study also adhered to the Meta-Analysis of Observational Studies in Epidemiology¹³ reporting guidelines.¹⁴

2.2. Literature search and inclusion criteria

The search strategy employed in this study included the following databases: PubMed, Embase, Web of Science, SPORTDiscus, Scopus, and the Cochrane Library. No restriction on language was applied, and the search covered online articles from inception to January 1, 2023. The search strategies and keywords used for each database are listed in Supplementary Table 2. In addition to thoroughly searching the databases, we reviewed the references of pertinent publications and systematic reviews published within the last 5 years. Although our search was not explicitly designed to identify PA, it included a broad range of studies pertaining to exercise.

The evaluation was performed using Rayyan (http://rayyan.qcri.org/), Microsoft Excel (Microsoft Corporation, Redmond, WA, USA), and Zotero software (Corporation for Digital Scholarship, Fairfax, VA, USA). Two reviewers (CY and TW) independently evaluated the titles and abstracts to determine inclusion criteria, and a detailed examination of the full text was performed when an abstract was deemed potentially relevant. Any conflicts or divergences between the reviewers were deliberated upon and determined through unanimity, with the involvement of a third reviewer (LX) if necessary. All 3 reviewers were professional investigators with systematic literature retrieval training. We employed the Patient, Intervention, Comparison, Outcome, and Study design principle to establish the eligibility criteria for study inclusion.^14,15 The following inclusion criteria were applied: (a) Participants: The participants in the study were adults (aged >18 years), including both healthy individuals and patients diagnosed with GERD or symptomatic GER. Patients with symptomatic GER were incorporated in the criteria to include as many relevant studies as possible; (b) Exposure: The exposure under investigation was PA, assessed through a self-report questionnaire or questions regarding participation in exercise, sports, and objective measures of PA (e.g., treadmill mileage); (c) Outcomes: Prevalence of GERD and symptomatic GER; (d) Study design: Observational studies such as retrospective or prospective cohort, case-control, and cross-sectional studies.

The exclusion criteria were as follows: (a) studies involving adolescents (≤18 years of age); (b) studies that did not include data regarding associations between the PA exposure and GERD in the exposure group, the non-exposure group, or both; (c) duplicate publications or sub-studies of included trials; (d) articles categorized as reviews, meta-analyses, editorials, comments, letters, guidelines, or news; and (e) studies with a sample size of less than 10 in both the exposure and non-exposure groups.

2.3. Definition

Among the included studies, the diagnosis of GERD had to be made through one of the following pathways. (a) Physician's diagnosis: GERD was diagnosed by a qualified physician based on clinical evaluation and diagnostic methods according to various guidelines or criteria; (b) Questionnaire-based diagnosis: The diagnosis of GERD was established using authoritative questionnaires such as gastroesophageal reflux disease questionnaire (GerdQ)^16,17 and Frequency Scale for the Symptoms of GERD;¹⁸ (c) Symptoms meeting the Montreal Definition and Classification,⁷ in which GERD was defined as a condition characterized by mild symptoms occurring on 2 or more days per week or moderate-to-severe symptoms occurring on more than 1 day per week. Symptomatic GER was defined as experiencing symptoms related to the reflux of gastric contents into the esophagus, such as heartburn and regurgitation, without necessarily meeting the specific diagnostic criteria for GERD. Any physical movement that encompasses the constriction of skeletal muscles and necessitates energy expenditure was characterized as PA.¹⁹

2.4. Data extraction

Relevant data were retrieved from the selected studies using a structured form. The form included detailed information about the authors, year of publication, country/region, study interval, sample size, sex, participants, and diagnostic method. Both reviewers (CY and TW) independently extracted data using a standardized form, and a third reviewer (LX) performed rigorous quality checks.

2.5. Study quality and bias assessment

Two reviewers (CY and TW) independently assessed the methodological quality of the included studies. The Agency for Healthcare Research and Quality scale was used for scoring the cross-sectional studies on a scale of 0–11. Scores falling within the ranges of 0–3, 4–7, and 8–11 were classified as indicative of low, moderate, and high quality, respectively.²⁰ The Newcastle-Ottawa Quality Assessment Scale was used to score case-control and cohort studies on a scale of 0–9. These scores were further categorized into 3 groups: 0–3, 4–6, and 7–9, corresponding to studies classified as having low, moderate, and high quality, respectively.^21,22 Any disagreements between the 2 reviewers (CY and TW) were resolved through discussion and consensus. Furthermore, the Grading of Recommendations Assessment, Development, and Evaluation methodology was employed to evaluate the total reliability of the evidence.²³

2.6. Data synthesis and statistical analysis

2.6.1. Meta-analysis statistics

The data extracted for the meta-analysis were analyzed using R software Version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria), with the significance threshold set at p < 0.05 for all analyses. Initially, data from all included studies were comprehensively pooled to compare the prevalence of GERD and symptomatic GER between the physical inactivity group (lowest level of PA) and the PA groups (characterized by the study authors as having increased frequency, intensity, volume, energetic expenditure, and other metrics). The relative risks (RRs) and their corresponding 95% confidence intervals (95%CIs) were extracted for meta-analysis. Heterogeneity was evaluated utilizing the I² statistic and Cochran's Q value, and a random effects model was employed for pooled analyses of GERD and symptomatic GER prevalence in the presence of significant heterogeneity (I² ≥ 50% or p < 0.05), while fixed-effects models were used otherwise. Heterogeneity among the studies was assessed using the I² statistic (0%–25% low heterogeneity, >25%–50% moderate heterogeneity, >50%–75% substantial heterogeneity, >75%–100% high heterogeneity).

2.6.2. Subgroup analysis

To ascertain subgroup variations and potential origins of the observed heterogeneity, we conducted a series of prespecified subgroup analyses, including examining the relationship between the year of publication (based on the most recent 5 years, before 2019 vs. in or after 2019), country/region income level (based on the World Bank's definition, high-income country/region vs. middle-income country/region), sample size (>2000 vs. ≤2000), diagnosis method (physician's diagnosis vs. authoritative questionnaire vs. Montreal definition), quality evaluation (median vs. high), and study design (cross-sectional study vs. case-control study).

Considering the impact of demographic and lifestyle factors on GERD prevalence, we explored potential influences, including demographics variables (sex, age, race, BMI, education level, and marital status) and lifestyle variables (smoking, daily habits such as midnight snacking, inadequate sleep, lack of breakfast, having dinner just before bedtime, and dietary habits including consumption of coffee, alcohol, chocolate, carbonated drinks, spicy foods, and salty foods).24, 25, 26 Subsequently, we systematically gathered pertinent information from the included literature, with factors reported in 10 or more articles selected to perform a subgroup analysis. Several factors were selected, including age, race, status, sex, BMI, smoking status, educational level, alcohol consumption, lack of breakfast, and coffee consumption. In cases where data were not reported in the study, we have denoted it as “none”. Statistical significance was set at p < 0.10 for differences in estimates between subgroups.

2.6.3. Dose–response analysis

To investigate the dose–response relationship, a 2-stage random effects meta-analysis was conducted using the dosresmeta package in R software (R Foundation for Statistical Computing).²⁷ The analysis included reports providing a distribution of the prevalence of GERD and the corresponding RR values with their corresponding 95%CIs across at least 3 distinct categories of PA time, which were standardized into 30 min per week for data analyses.

Both linear and non-linear analyses were employed to fit the following formulas:

• a)
  Linear trend:

  $$\text{log}\left(\text{RR}\right)={\beta }_0+{\beta }_1\times \text{Time category}$$
• b)
  Non-linear (quadratic) trend:

  $$\text{log}\left(\text{RR}\right)={\beta }_0+{\beta }_1\times \text{Time category}+{\beta }_2\times \text{Time categor}{\mathrm{y}}^2.$$

The restricted maximum likelihood estimation method was utilized throughout the analysis. To approximate the covariance, we utilized Greenland and Longnecker's method.²⁸ This approach ensured the attainment of a unique solution by maintaining consistent margins between pseudo-counts derived from multivariable-adjusted log RRs and the margins of unadjusted data.

2.6.4. Statistical analysis

The inter-reviewer agreement was evaluated using κ analysis. IBM SPSS Statistics for Windows, Version 26 (IBM Corp., Armonk, NY, USA) was used to calculate the RR and the corresponding 95%CI based on the prevalence data obtained from the exposed and non-exposed groups. To evaluate potential publication bias, funnel plots were visualized for asymmetry (if ≥10 studies were included) and both Egger's and Begg's tests were conducted. Sensitivity analyses were performed to evaluate the robustness of the pooled results by excluding 1 study at a time.

3. Results

3.1. Study selection

The initial search identified 14,418 potentially relevant studies. After eliminating 4235 duplicates, 9908 studies were further screened based on their titles and abstracts, leading to the exclusion of most studies. The κ coefficient for this selection process between reviewers was 0.770. A full-text review was conducted for the remaining 275 studies, resulting in the exclusion of 242 studies based on the predefined inclusion criteria. The κ coefficient for the study selection process between reviewers was 0.681. Finally, 33 studies were deemed eligible for data extraction and inclusion in this systematic review and meta-analysis.29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61 Among these, data from 22 articles conformed to the definition of GERD in the context of the current study.^{30,31,33,34,36,}40, 41, 42, 43, 44^,48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59 A flowchart illustrating the study selection process is shown in Fig. 1.

Fig. 1.

Literature flowchart and study selection.

3.2. Study characteristics

The characteristics of the included studies are summarized in Table 1. The included studies were drawn from 6 databases and were published between 2004 and 2022. Two studies adopted a case-control design,^34,51 4 studies adopted a cohort study design,^29,46,47,61 and the remaining studies were cross-sectional.30, 31, 32, 33^,35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45^,48, 49, 50^,52, 53, 54, 55, 56, 57, 58, 59, 60 The sample sizes of the individual studies ranged from 211 to 50,001, with a total of 66,098 cases of symptomatic GER (including 14,192 cases of GERD) and 167,855 participants across all included studies. In terms of GERD diagnosis, 4 studies relied on physician diagnosis,^34,40,50,51 5 studies used the Montreal definition,^{41,43,44,52,53} and 13 studies employed authoritative questionnaires for diagnosis.^{30,31,33,36,42,48,49,}54, 55, 56, 57, 58, 59 The results of the quality assessment analyses were presented in Supplementary Tables 3 and 4. Most studies were assessed as having a low-to-medium risk of bias. The overall quality of evidence for the GERD outcomes was categorized as low. A summary of the quality of evidence obtained using the GRADE framework is provided in Supplementary Table 5.²³

Table 1.

Main characteristics of studies included in this meta-analysis

Reference	Country/region	Design	Quality	Study interval	Sample size	Male/female (n)	Participant (disease/normal)	Diagnosis method	PA definition
Ahmed et al. (2020)60	Pakistan	Cross-sectional study	Median	2018	2000	824/1158	1000/1000	Questionnaire	More than 15 min per day
Alkhathami et al. (2017)57	Saudi Arabia	Cross-sectional study	Median	2016	2043	–	587/1456	Questionnaire	More than 30 min per week
Alrashed et al. (2019)56	Saudi Arabia	Cross-sectional study	Median	2018–2019	400	227/173	95/305	Questionnaire	More than 30 min per week
Al-Towairqi et al. (2020)58	Saudi Arabia	Cross-sectional study	High	2019–2020	256	0/256	75/181	Questionnaire	More than 5 times per week
Arivan et al. (2018)55	India	Cross-sectional study	Median	2015	358	188/170	18/340	Questionnaire	More than 5 times per week
Bert et al. (2021)54	Italy	Cross-sectional study	Median	2019	559	138/418	155/404	Questionnaire	PA
Cela et al. (2013)53	Albania	Cross-sectional study	High	2012	845	345/500	101/744	Questionnaire	PA
Jiang et al. (2010)52	China	Cross-sectional study	Median	2008	2615	1178/1437	220/2395	Questionnaire	PA
Dore et al. (2008)51	Italy	Case-control study	Median	–	500	169/331	300/200	Doctor's diagnosis	PA
Eslami et al. (2017)50	Iran	Cross-sectional study	High	2014–2015	505	156/349	285/220	Doctor's diagnosis	More than 2 h per week
Al Ghadeer et al. (2021)59	Saudi Arabia	Cross-sectional study	Median	2021	1517	625/892	312/1205	Questionnaire	More than 30 min per week
Gong et al. (2019)49	China	Cross-sectional study	Median	2009–2016	37,442	10,168/0	7449/29,993	Questionnaire	PA
Gunasinghe et al. (2020)48	Sri Lanka	Cross-sectional study	Median	2019	1166	488/626	580/534	Questionnaire	More than 30 min per day
Hägg et al. (2019)47	Sweden	Cohort study	High	2000–2010	4882	0/4882	605/4277	Questionnaire	More than 4 h per week
Islami et al. (2014)46	Iran	Cohort study	High	2004–2008	50,001	21,216/28,785	30,441/19,560	Questionnaire	Regular non-intense or regular/irregular intense
Karayaka et al. (2014)45	Turkey	Cross-sectional study	Low	2008–2009	500	213/287	204/296	Questionnaire	Average PA at leisure time
Karim et al. (2011)44	Pakistan	Cross-sectional study	Median	2009–2010	1875	1019/856	689/1186	Questionnaire	Routine post dinner: walking
Koul et al. (2018)43	Kashmir	Cross-sectional study	Median	2014–2015	2600	–	529/2071	Questionnaire	Non-sedentary
Kuddus et al. (2021)42	Saudi Arabia	Cross-sectional study	Median	2020	704	364/340	408/269	Questionnaire	More than once per week
Kumar et al. (2011)41	India	Cross-sectional study	Median	2007	905	399/506	169/736	Questionnaire	Non-sedentary
Mehta et al. (2021)61	America	Cohort study	Low	2005–2017	42,955	0/42955	9291/33,664	Questionnaire	More than 30 min moderate-to-vigorous activity per day
Murao et al. (2011)40	Japan	Cross-sectional study	Median	2004–2005	2853	1975/878	667/2186	Doctor's diagnosis	More than once per month
Nandurkar et al. (2004)39	America	Cross-sectional study	Median	–	211	90/121	95/116	Questionnaire	PA > 4 METs
Nilsson et al. (2004)38	Norway	Cross-sectional study	Median	1984–1997	43,363	20,369/22,994	3153/40,210	Questionnaire	More than 30 min per week
Nocon et al. (2006)37	Germany	Cross-sectional study	Median	–	6954	3356/3598	2964/4000	Questionnaire	Recreational sports only
Otayf et al. (2022)36	Saudi Arabia	Cross-sectional study	Median	2021	953	384/569	220/733	Questionnaire	More than 30 min per week
Pandeya et al. (2012)35	Australia	Cross-sectional study	Median	2002–2005	1580	1040/540	853/727	Questionnaire	PA
Sadiku et al. (2021)34	Albania	Case-control study	High	2013–2014	521	–	248/273	Doctor's diagnosis	PA
Sarwar et al. (2021)33	Pakistan	Cross-sectional study	Median	2018	340	0/340	218/122	Questionnaire	PA
Shaha et al. (2012)32	Bangladesh	Cross-sectional study	Median	2010–2011	2000	1064/936	110/1890	Questionnaire	PA
Sharma et al. (2018)31	India	Cross-sectional study	Median	2015	600	500/100	150/450	Questionnaire	More than 30 min per day
Yadegarfar et al. (2018)30	Iran	Cross-sectional study	Median	2014	1130	522/575	717/380	Questionnaire	PA
Zheng et al. (2007)29	Sweden	Cohort study	High	1967–2002	27,717	8847/10,075	3190/15,732	Questionnaire	Light PA at leisure time

Note: – means unmentioned in text.

Abbreviations: METs = metabolic equivalents; PA = physical activity.

3.3. Study outcomes

Upon investigating the association between PA and the risk of symptomatic GER, the random effects model indicated a significant negative association (RR = 0.74, 95%CI: 0.66–0.83; p < 0.01) (Fig. 2A). However, the heterogeneity among all studies was assessed as high (I² = 92.0%, p for heterogeneity < 0.05), indicating that less stringent definitions among the studies may have contributed to less conclusive results.

Fig. 2.

RRs for PA of (A) symptomatic GER and (B) GERD. 95%CI = 95% confidence interval; GER = gastroesophageal reflux; GERD = gastroesophageal reflux disease; PA = physical activity; RR = relative risk.

When investigating the association between PA and risk of GERD, the random effects model also indicated a significant negative association (RR = 0.80, 95%CI: 0.76–0.84; p < 0.01). Furthermore, the analysis revealed a moderate level of heterogeneity across included studies (I² = 36.2%, p for heterogeneity < 0.05) (Fig. 2B), which is significantly lower than the heterogeneity within the symptomatic GER group.

Among the studies that included patients with GERD, Funnel plots as well as Egger's and Begg's tests (p for Egger's test = 0.31; p for Begg's test = 0.21) (Supplementary Figs. 1–3) revealed no evidence of significant publication bias. Leave-one-out analysis was conducted to assess the influence of each study on the overall results (Supplementary Fig. 4). Exclusion of individual studies did not substantially alter the pooled estimate of the primary outcome, indicating that the results were robust.

3.4. Subgroup analysis

Subgroup analyses were performed based on various study characteristics to investigate the potential origins of heterogeneity in the association between PA and the prevalence of GERD. We observed consistent findings irrespective of the publication year, income level of the country/region, or sample size. The results of the subgroup analysis revealed significant differences between subgroups based on the diagnosis method (p = 0.02), quality evaluation (p < 0.01), and study type (p < 0.01) (Fig. 3 and Supplementary Figs. 6–10). Studies that enrolled patients based on authoritative questionnaires (RR = 0.84, 95%CI: 0.80–0.88), rated as median quality (RR = 0.82, 95%CI: 0.79–0.84), and those using cross-sectional study designs (RR = 0.82, 95%CI: 0.79–0.84) showed a weaker association compared to studies that included patients diagnosed by physician (RR = 0.72, 95%CI: 0.63–0.82) and those adhering to the Montreal definition (RR = 0.70, 95%CI: 0.59–0.83), as well as those rated as high quality (RR = 0.66, 95%CI: 0.58–0.77) and that have case-control study designs (RR = 0.65, 95%CI: 0.57–0.75).

Fig. 3.

Subgroup analyses of association between PA and GERD according to study characteristics. ^a Both the Authoritative questionnaire and the Montreal definition rely on Questionnaire based diagnosis. 95%CI = 95% confidence interval; GERD = gastroesophageal reflux disease; PA = physical activity; RR = relative risk.

In the subgroup analyses for potential confounders, statistically significant variations among subgroups were identified for age (p < 0.01) and smoking status (p = 0.03), indicating that these factors may have a statistically significant impact on modifying the relationship between PA and the prevalence of GERD (Fig. 4 and Supplementary Figs. 11–20). The overall association between PA and the prevalence of GERD was found to be stronger among participants aged ≥40 years (RR = 0.69, 95%CI: 0.61–0.77) and smokers (RR = 0.67, 95%CI: 0.58–0.84) as compared to participants aged <40 years (RR = 0.85, 95%CI: 0.80–0.91) and non-smokers (RR = 0.82, 95%CI: 0.79–0.84).

Fig. 4.

Subgroup analyses of association between PA and GERD according to adjustments. 95%CI = 95% confidence interval; BMI = body mass index; GERD = gastroesophageal reflux disease; PA = physical activity; RR = relative risk.

3.5. Dose–response analysis

Eight studies³⁰,^{36,42,51,54,56,57,59} with more than 2 categories of PA time were included in the dose–response analysis, comprising a total of 7806 participants (2794 cases with GERD). The analysis uncovered a linear relationship between PA duration and GERD prevalence (χ² test; p = 0.004, Akaike information criterion = −8.17). For every additional 30 min of weekly PA time, there was a 9.48% reduction in the risk of prevalence of GERD (RR = 0.92, 95%CI: 0.88–0.95) (Fig. 5A). Additionally, the analysis suggested a non-linear relationship between PA and GERD, which may better reflect the true association (χ² test; p = 0.004, Akaike information criterion = −26.93), indicating that the risk of GERD decreases with increasing exercise frequency (Fig. 5B). Our results revealed that individuals engaging in 150 min of PA per week had a 72.09% reduced risk of developing GERD compared with individuals with the lowest level of PA.

Fig. 5.

(A) The linear dose–response and (B) the non-linear dose–response relationship between PA and GERD according to adjustments. AIC = Akaike information criterion; GERD = gastroesophageal reflux disease; PA = physical activity; RR = relative risk.

4. Discussion

PA is widely recognized as a low-cost health-protective factor with significant potential for preventing various diseases.^62,63 To attain notable health advantages, existing guidelines recommend that adults engage in at least 150 min of moderate-intensity or 75 min of vigorous-intensity aerobic PA every week.^64,65 Consistently meeting and maintaining these recommended levels of PA can significantly reduce various risk factors for non-communicable chronic diseases, including metabolic, hemodynamic, functional, body composition, and epigenetic factors.⁶⁶

Despite several guidelines recommending PA to prevent certain gastrointestinal motility disorders such as constipation and irritable bowel syndrome,67, 68, 69 there is currently a lack of guidelines addressing the role of PA in the risk of GERD. This meta-analysis provides an overview of current research by examining the association between PA levels and both symptomatic GER and the prevalence of GERD. By synthesizing data from 33 studies, our meta-analysis provides robust evidence supporting the notion that higher PA levels are associated with a lower risk of developing symptomatic GER and GERD. Specifically, we found that compared with those with the lowest level of PA, individuals who engaged in PA had 26% and 20% lower risks of symptomatic GER and GERD, respectively. Considering the potential impact of excluding a study with notably large sample sizes on the estimation of RR, we also scrutinized the findings of the study conducted by Gong et al. ⁴⁹ Excluding this study did not cause any significant deviation from the RR calculated in our comprehensive meta-analysis.

Expanding on these findings, we investigated the relationship between varying PA levels and the prevalence of GERD. We found that for each additional 30-min increment of PA per week, there was an associated 8% reduction in the risk of developing GERD. Notably, our analysis also revealed a non-linear dose–response relationship, with individuals who engaged in 150 min of PA per week having a 72.09% reduced risk of developing GERD as against individuals who abstained from PA. These findings underscore the potential advantages of consistent PA in reducing the prevalence of GERD.

PA encompasses various exercise types and intensity levels that can affect disease prevention and rehabilitation.⁶³ These factors also play different roles in the risk of developing GERD and symptomatic GER. For instance, an increase of 1000 daily steps was found to be associated with a reduced prevalence of GERD,⁷⁰ while yoga showed promise as an effective approach for alleviating GER symptoms.⁷¹ Notably, a study by Clark et al.⁷² revealed that running was associated with a higher prevalence of GER, whereas cycling was associated with a lower occurrence. These studies indicate that low-intensity aerobic exercises, such as walking, cycling, and yoga, may offer protective benefits against GERD. In contrast, high-intensity aerobic and anaerobic exercises, such as running, may have deleterious effects.⁶³ These findings were consistent with the observations of Mendes-Filho et al.⁷³ However, it should be noted that these studies primarily involved a limited number of participants and focused on analyzing GER symptoms during exercise. Accordingly, additional research is warranted to elucidate the influence of long-term protective factors, and recommendations for suitable exercise modalities to prevent the onset of GERD in different populations should be formulated.

Several studies have investigated the influence of varying intensity levels of PA on GER symptoms. Nandurkar et al.³⁹ observed that, although there was no statistically significant association between the intensity of PA and GER symptoms, moderate-intensity PA displayed a protective effect compared to light-intensity PA. Conversely, vigorous-intensity PA was found to be associated with an increased risk of GER symptoms. However, Pandeya et al.³⁵ reported a different finding, indicating that a higher PA intensity was significantly associated with a reduced risk of weekly GER symptoms. A similar association was identified in studies by Hägg et al.⁴⁷ and Sadiku et al.,³⁴ who demonstrated the protective effect of moderate-intensity PA against GERD. Nevertheless, whether high-intensity exercise confers protection remains debatable. Additionally, inconsistencies in the definitions of low-to-high PA intensity across various articles may hinder our ability to identify a definitive exercise intensity level to recommend for GERD prevention.

In the present study, the associations between PA and the prevalence of GERD revealed a reduction in risk between 11% and 34% in the subgroup analyses, suggesting a robust relationship between PA and the prevalence of GERD. Subgroup analyses also revealed several noteworthy findings. Festi et al.⁷⁴ reported that the relationship between PA and the prevalence of GERD may be influenced by differences in the studied populations, methods used to evaluate exercise, assessment of PA, and diagnosis of the disease. In our subgroup analyses, we identified differences in age, with a stronger association between PA and the prevalence of GERD in the older adult population. This observation may be attributed to the fact that PA can exert a more protective influence against the associated decline in functional capacity among older individuals than in their younger counterparts.⁷⁵ Smokers also exhibited a stronger association between PA and the prevalence of GERD compared to non-smokers. However, due to the limited availability of research specifically focusing on smokers, additional studies are required to provide further evidence to test this hypothesis. Additionally, the heterogeneity of results may have been influenced by factors such as study quality and type. Specifically, higher-quality studies and case-control studies showed a stronger association between PA and prevalence of GERD.

Obesity is significantly associated with physical inactivity and increases the risk of developing GERD.^76,77 Excessive body weight was found to increase intra-abdominal pressure and reduce lower esophageal sphincter pressure, leading to GERD.⁷⁷ Therefore, it is important to consider the potential confounding effects of obesity when examining the relationship between PA and the prevalence of GERD. In our subgroup analysis, we found no significant differences in the association between PA and the prevalence of GERD between the overweight/obese and normal/underweight groups. This finding suggests that the potential protective effect of PA against the prevalence of GERD may extend across all populations, irrespective of weight status.

Although our meta-analysis supports the notion that PA has a protective effect against the onset of GERD, the exact mechanisms underlying this effect remain unclear. It is plausible that various biochemical and psychosocial factors contribute to this relationship. First, PA may have anti-inflammatory effects that lower the risk of GERD. A 10-year follow-up study revealed that regular PA lowers inflammatory markers, potentially preventing age-related pro-inflammatory conditions.⁷⁸ Nevertheless, we observed the release of muscle-derived cytokines during acute exercise, which could potentially lead to the development of GERD symptoms.⁷⁸ Second, regular PA confers benefits in terms of preventing reflux by enhancing the strength of the crural diaphragm, a critical element in anti-reflux mechanisms.⁷⁹ The association between diaphragmatic activation and PA was consistently demonstrated by Sinderby et al.⁸⁰ Third, it is noteworthy that PA is associated with a lower risk of developing depression and decreased levels of stress.^21,81 These psychological factors may play a role in reducing the prevalence of GERD through gut-brain peptides.⁸² However, our study did not consider this perspective.

Despite the novelty and significance of our meta-analysis, it is imperative to acknowledge its limitations. First, one-half (11/22) of the included studies relied on self-reported questionnaires to assess both the exposure factor (PA) and outcome (GERD). Although such questionnaires are widely used in studies of PA or GERD, they are subject to recall bias, which may have affected the accuracy of our findings. Second, we observed heterogeneity in the definition of PA across the included studies, which poses a challenge for interpreting the results. Although we conducted a dose–response analysis using studies that assessed weekly PA duration to recommend the optimal PA duration for reducing the odds of GERD prevalence, the optimal PA intensity remains unclear due to limited data availability within these studies. Third, the results of most subgroup analyses were not significant. This may be due to the relatively small number of studies included in these analyses, which led to low statistical power. Fourth, our findings may not be generalizable to professional athletes, and further investigation is warranted to elucidate the relationships between PA and GERD within these cohorts. Finally, the cross-sectional nature of the studies included in our analysis meant that we could not establish a causal relationship between PA and GERD. Moreover, because GERD and PA are both gradual processes, it may be challenging to define their temporal relationships. Future research should seek to establish a more rigorous and standardized baseline for the onset and duration of GERD and PA, which will enable further elucidation of the potential implications of tailored interventions.

5. Conclusion

Our study provides evidence supporting the potential protective effect of higher PA levels against GERD, especially in older adults and smokers. Meeting the recommended PA level (150 min per week) may significantly reduce the prevalence of GERD. These findings emphasize the need for further research on this relationship.