This systematic review and meta-analysis followed PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [28]. The study protocol was registered with PROSPERO under registration number CRD42023472590.

A systematic literature search was conducted from the earliest available date to November 2024 across 11 databases, including 7 international sources (Web of Science, CINAHL, Cochrane Library, Scopus, PsycINFO, PubMed, and Embase) and 4 Chinese sources (SinoMed, CNKI, Wanfang, and VIP).

The search strategy included basic strings of Medical Subject Headings (MeSH) terms and free terms combined with Boolean operators. The search terms were “virtual reality,” “VR,” “virtual environment,” “video game*,” “virtual simulation,” “virtual medicine,” “mixed reality,” “commercial game*,” “virtual game*,” “exergam*,” “play-based therapy,” “augmented reality,” “virtual reality exposure therapy,” “x-box 360,” “Kinect,” “Wii,” “virtual world,” “head-mounted display,” “pulmonary disease, chronic obstructive,” “chronic obstructive pulmonary disease*,” “chronic obstructive airway disease,” “chronic obstructive lung disease,” “COAD,” “COPD,” “chronic airflow obstruction*,” “airflow obstruction, chronic,” and “airflow obstructions, chronic.” Multimedia Appendix 1 details the search strategies for each database. Our search included only Chinese and English sources and full-text articles from peer-reviewed journals. Additionally, we reviewed published reviews, reference lists of included studies, and similar articles. Before data synthesis, all databases were researched in December 2024 to capture newly published studies.

The Participant, Intervention, Comparator, Outcomes, and Study Design (PICOS) model was used to establish the inclusion criteria for each article (Textbox 1).

Endnote X9 (Clarivate Plc) was used to export and manage all search results and to identify and remove duplicate studies. The screening process consisted of 2 stages. First, the titles and abstracts of the remaining articles were reviewed, and any study that examined the relationship between VR and COPD was retained for further analysis. After irrelevant articles were removed, the remaining articles were downloaded and further reviewed to determine which studies should be included in the final analysis. All steps were independently screened and cross-checked by 2 researchers (YC and YZ) against the eligibility criteria. Disagreements between the 2 reviewers were resolved through a consensus process involving additional investigators (HT and JC).

For each eligible study, a predesigned Excel form (Microsoft Corporation) was used for data extraction by an author (YC) on the following subheadings: (1) publication details (first author’s surname, publication year, and country); (2) participant characteristics (sample size, mean age, sex, and disease severity); (3) intervention details (site, exercise intensity, comparator group, intervention group, VR content, and intervention format); (4) outcome measures; and (5) both pre- and postintervention data (eg, mean and SD) for the intervention and comparator groups. For studies reporting SEs or median and IQR instead of means and SD, these values were converted using standard conversion tools [32,33]. The extracted data were reviewed by the second author (YZ) for accuracy.

Two researchers (YC and XL) independently assessed the quality of the included studies. For RCTs, the Cochrane Risk of Bias tool [34] was used, which evaluates randomization sequence generation, allocation concealment, participant blinding, outcome blinding, incomplete outcome data, selective reporting, and other biases. Items were categorized as “low risk,” “high risk,” or “unclear risk.” Any discrepancies encountered during the review process were deliberated with other investigators (HT and JC) and ultimately reconciled through mutual agreement.

Statistical analysis was performed using Cochrane Review Manager 5.4 (Cochrane Collaboration) to assess the efficacy of VR-complemented PR in patients with COPD compared with comparators and to generate forest plots. Mean difference (MD) and 95% CI were calculated for continuous variables.

The I² statistics were used to assess heterogeneity for each comparison. A fixed-effects model was applied when P0.10 and I²≤50%, indicating statistical homogeneity. Conversely, a random-effects model was used when heterogeneity was high (P<.10 and I²>50%).

Subgroup analyses were conducted to compare the efficacy of VR-complemented PR based on 2 factors: (1) comparator types (active exercise controls vs nonactive exercise controls) and (2) intervention duration (eg, ≤4 weeks, 5-12 weeks, and >12 weeks). These analyses aimed to explore differences in effectiveness across varying baseline conditions and intervention durations.

A sensitivity analysis was performed to assess the robustness of the findings by consecutively omitting each study. For the overall effect, a P value of less than .05 was considered statistically significant.

The search of 11 databases and other sources identified 1045 potentially relevant articles. After removing 482 duplicates and reviewing 563 titles and abstracts, 45 articles were selected for full-text screening. Finally, 16 articles were deemed eligible for inclusion in the meta-analysis. The literature screening process, reasons for exclusion, and results are illustrated in Figure 1 and Multimedia Appendix 2.

The features of the 16 trials [35-50] are displayed in Tables 1 and 2. The articles were published from 2014 to 2024, and 11 (69%) papers [40-50] were published within the last 5 years (2020-2024). In total, 9 (56%) [36,37,42-45,48-50] of the included studies were conducted in China, 3 (19%) in Poland [39-41], 2 (13%) in Italy [35,46], and 1 (6%) each in Indonesia [38] and Turkey [47].

The trials recruited a total of 1052 people with COPD. Fifteen studies [36-50] reported the sex of the participants (1012 individuals), with 638 (63.04%) women and 374 (36.96%) men. Additionally, 1 study [35] did not provide information on sex. Twelve studies [35,37-44,46,47,49] provided data on the severity of COPD in the participants. According to the GOLD criteria, COPD severity was classified as stage I to III in 1 study [46] and as stage II to IV in 4 studies [42,44,47,49]. Additionally, 7 studies [35,37-41,43] reported the percentage of predicted forced expiratory volume in 1 second (FEV₁% pred), with values ranging from 39.2% to 86.5%. Based on both the GOLD stages and FEV₁% pred, the majority of patients in this review had COPD severity ranging from moderate to very severe.

A total of 6 studies [35,38,40,41,47,49] reported exercise intensity. Among them, 5 studies [38,40,41,47,49] monitored heart rate during exercise to ensure the intensity remained within the target range, while 1 study [35] assessed exercise intensity using the Borg dyspnea scale either during or immediately after the exercise session. Of the included studies, 11 [35-44,47] were categorized as active exercise controls with structured and supervised exercise training, while 5 [45,46,48-50] were classified as nonactive exercise controls, focusing on standard COPD management or low-intensity, nonstructured interventions. The VR technologies used in the studies varied, with 14 studies providing information on the types of VR technologies used: 2 (14%) [35,38] studies used the Nintendo Wii, 4 (29%) [39,40,42,44] used the Microsoft Xbox Kinect, 4 (29%) [36,37,43,48] used the BioMaster system, 2 (14%) [46,47] used the Oculus Quest 2, 1 (7%) [41] used the TierOne device (Stolgraf), and 1 (7%) [50] used the SUBOR A20 gaming console (Xiaobawang Company). Additionally, 2 studies [45,49] did not report the type of VR technology used. Interventions ranged from 2 weeks to 24 weeks, with frequency occurring 2-7 days per week. The duration of each intervention varied from 5 minutes to 1 hour.

Lung function was assessed using FEV₁ (%), FEV₁ (L), forced vital capacity (FVC), and FEV₁/FVC (%). Exercise capacity was evaluated using the 6MWD test. Dyspnea severity was measured using the modified British Medical Research Council (mMRC) scale. Health status was assessed with the COPD Assessment Test (CAT). Secondary outcomes included oxygenation status, assessed by peripheral capillary oxygen saturation (SpO₂), and patient-reported experience measures such as acceptability and engagement.

Figures 2 and 3 (see also [35-50]) illustrate the results of the Cochrane Risk of Bias Tool, which was applied to evaluate the quality of the RCTs included in this analysis. All 16 RCTs reported random sequence generation, and 6 (38%) studies provided details on allocation concealment. Because of the nature of the VR intervention, none of the studies were able to blind participants or personnel. However, 5 (31%) studies implemented blinding of outcome assessors, which reduced the risk of detection bias. A total of 14 (88%) studies reported complete outcome data, while 7 (44%) offered sufficient information to evaluate the risk of selective reporting. Additionally, all studies were considered to have a low risk of other biases.

This review included 16 eligible trials to summarize the evidence regarding the effects of VR-complemented PR compared with comparators in people with COPD. Compared with the previous meta-analysis by Chai et al [24], 5 additional studies were included, providing a more comprehensive and updated synthesis of evidence. The findings from this meta-analysis underscore the effectiveness of VR-complemented PR compared with comparators in improving a range of critical outcomes, including lung function (FEV₁ [L], FEV₁/FVC, FVC), exercise capacity (6MWD), dyspnea (mMRC), health status (CAT), oxygenation status (SpO₂), and patient-reported experience measures (acceptability and engagement). Furthermore, subgroup analyses revealed that VR-complemented PR had a significantly greater effect on FEV₁ (L) and 6MWD compared with nonactive exercise controls. Additionally, VR-complemented PR showed a greater improvement in FEV₁/FVC compared with active exercise controls. The effectiveness of VR-complemented PR was also influenced by intervention duration, with the most significant improvements observed in programs lasting 5-12 weeks. However, sensitivity analyses indicated more consistent results after excluding certain studies. Despite this, the findings should be interpreted with caution due to the small sample sizes and limited number of trials, which may introduce bias and reduce generalizability.

COPD is characterized by reduced lung function, which is strongly associated with an increased risk of exacerbations, hospitalization, and mortality [51,52]. Improving lung function is therefore a key goal of PR. In this meta-analysis, VR-complemented PR demonstrated significant improvements in FEV₁ (L), FEV₁/FVC, and FVC compared with comparator groups. Our results align with those reported by Chai et al [24], Liu et al [25], and Obrero-Gaitán et al [26], who similarly identified significant improvements in lung function following VR-based interventions compared with comparators. However, our meta-analysis expands upon these studies by incorporating a broader range of pulmonary outcomes and a larger data set, providing a more comprehensive synthesis of the evidence. Notably, the observed improvements in FEV₁ (L) (MD 0.2 L) exceeded the widely accepted minimally clinically important difference (MCID) of 0.1 L [53], underscoring the clinical relevance of these changes. While specific MCID thresholds for FEV₁/FVC and FVC are not well-established in the literature, the improvements observed in FEV₁/FVC (MD 6.1%) and FVC (MD 0.3 L) suggest meaningful enhancements in lung function and capacity. By contrast, FEV₁ (%) did not show a statistically significant improvement (MD 2.4%, P=.49), which may be attributed to the variability in baseline characteristics, treatment intensity, or measurement methods across studies [41]. Despite this, the improvements in other lung function measures (eg, FEV₁ [L], FEV₁/FVC, and FVC) suggest that VR-complemented PR may still have a beneficial effect on lung function, with further studies needed to better understand the specific impact on FEV₁ (%).

The observed improvements in pulmonary function are likely driven by the enhanced physical activity, adherence, and engagement facilitated by the immersive and interactive features of VR technology [15]. By reducing the monotony of traditional PR and creating a more engaging and motivating environment, VR encourages consistent participation, leading to improved respiratory muscle strength and overall physical conditioning [54]. Furthermore, VR’s real-time feedback and personalized intensity adjustments enable patients to train within optimal zones, maximizing therapeutic benefits and contributing to the observed improvements in lung function.

The subgroup analyses of FEV₁ (L) and FEV₁/FVC revealed contrasting results, emphasizing the complexity of evaluating lung function in patients with COPD. In the case of FEV₁ (L), VR-complemented PR showed a significantly greater effect in the nonactive exercise control group (MD 0.3 L, P=.005) compared with the active exercise control group (MD 0.2 L, P=.05). This may suggest that VR-complemented PR could be more beneficial when combined with less intensive rehabilitation or usual care, possibly due to greater room for improvement in patients with lower baseline rehabilitation. However, further research is needed to confirm this potential benefit. These findings are consistent with Donath et al [55], who also found more pronounced VR effects in inactive control conditions. The clinical relevance of these results lies in the potential for VR to enhance rehabilitation in patients who are typically less engaged or receiving lower-intensity interventions, thus improving outcomes for a wider range of patients with COPD.

By contrast, the analysis of FEV₁/FVC yielded different results. VR-complemented PR showed a significant improvement compared with the active exercise control group (MD 6.2%, P<.001), with no heterogeneity (I²=0%), indicating a robust effect. However, VR-complemented PR did not show a significant improvement compared with the nonactive exercise control group (MD 5.8%, P=.07), with high heterogeneity (I²=95%). These results may be due to the nature of FEV₁/FVC as a ratio that is more sensitive to lung mechanics than absolute measures such as FEV₁ (L) [56,57]. The lack of a significant result in the nonactive exercise control group may also reflect lower intervention intensity, limiting improvement.

Dysfunction and atrophy of skeletal muscle are common features of COPD, significantly impairing patients’ physical function and limiting their ability to perform daily activities. The 6MWD is a widely used measure of exercise capacity in COPD rehabilitation. In this meta-analysis, VR-complemented PR demonstrated a significant improvement in 6MWD compared with comparator groups, highlighting the effectiveness of VR in enhancing functional exercise capacity. Our findings align with those of Wang et al [22], Patsaki et al [23], Obrero-Gaitán et al [26], Chai et al [24], and Liu et al [25], who also reported significant improvements in 6MWD with VR-based interventions. The observed improvement in 6MWD (MD 23 m) approaches the MCID (mean 26 m, SD 2 m) for COPD [58], which suggests that while the improvement is statistically significant, it may not fully meet the threshold for clinical significance.

The subgroup analysis revealed that VR-complemented PR had a significantly greater effect (MD 41 m) on 6MWD compared with the nonactive exercise control group, a value that far exceeds the MCID [58]. This may indicate that VR interventions could be more beneficial when combined with less intensive rehabilitation or usual care, where there is more room for improvement. However, these findings are consistent with the results of Janhunen et al [59], but further research is needed to confirm this hypothesis. The engaging nature of VR likely enhances patient motivation and adherence, leading to more substantial gains in exercise capacity [60]. By contrast, VR-complemented PR showed a smaller effect size (MD 15 m) when compared with the active exercise control group, which is well below the MCID [58]. This suggests that the additional benefit of VR in settings with already structured exercise programs is more limited. The smaller effect may be due to the already high baseline exercise intensity in the active exercise control group, where the potential for further improvement is constrained.

A subgroup analysis based on intervention duration showed that VR-complemented PR is more effective with a 5-12-week duration compared with comparators. Programs of 5-12-week duration resulted in the largest improvements in 6MWD, while shorter interventions (≤4 weeks) showed no significant improvement. This is closer to the standard duration of PR (4-8 weeks) [29,30], suggesting that VR-complemented PR within this time frame may provide optimal improvements in exercise capacity. Shorter interventions (≤4 weeks) may not allow enough time for meaningful progress, while 5-12 weeks appears to be the most beneficial for improving physical function in patients with COPD.

The improvement may be attributed to VR’s interactive and engaging nature, which reduces the monotony of traditional PR, enhances adherence, and provides real-time feedback [61]. VR also creates an immersive environment that distracts patients from dyspnea and fatigue, enabling higher intensity and longer duration of exercise [62]. Moreover, VR’s personalized features allow tailored adjustments in exercise intensity and difficulty, making it adaptable to patients with varying physical capacities [63]. These factors collectively enhance skeletal muscle conditioning and overall physical function, directly supporting improved exercise capacity.

Dyspnea, a hallmark symptom of COPD, reflects an imbalance between ventilatory demand and capacity, significantly impacting patients’ quality of life and physical function [64]. In this meta-analysis, the mMRC scale, measured at rest, was used to evaluate dyspnea. The pooled analysis indicated a significant reduction in mMRC scores in the VR-complemented PR group compared with the comparator group, consistent with the findings of Chai et al [24]. However, our results differ from those of Patsaki et al [23], which included only 2 studies and did not show significant improvement. The broader inclusion criteria and larger data set in our analysis likely provide a more comprehensive understanding of VR-complemented PR’s effects on dyspnea.

The improvements in dyspnea can be attributed to VR’s immersive nature, which enhances exercise adherence by reducing boredom and distraction from breathlessness, enabling consistent training at higher intensities [65]. Physiologically, improved skeletal muscle function and exercise capacity lower ventilatory demand during activity, reducing dyspnea symptoms. Psychologically, VR’s engaging and gamified features alleviate anxiety and boost confidence, positively influencing breathlessness perception [66]. While the improvements in mMRC scores observed in our analysis are statistically significant, it is important to consider their clinical relevance. The modest reduction of 0.3 points, though notable, may have limited practical impact in isolation. However, when combined with significant gains in exercise capacity and health status, VR-complemented PR demonstrates a multifaceted approach to managing COPD symptoms, addressing both physical and psychological aspects of dyspnea.

Health status is a key outcome in COPD management, reflecting the disease’s impact on daily life and well-being. In this meta-analysis, VR-complemented PR significantly improved health status, as measured by CAT, with an MD of –3 compared with comparators. This reduction exceeds the MCID of 2 points, demonstrating both statistical and clinical relevance [67]. The included studies showed consistent results with no significant heterogeneity (I²=0%), reinforcing the reliability of these findings. Our results are in line with the review by Wang et al [22], which suggested that VR-complemented PR could improve CAT scores. However, the conclusions by Wang et al [22] were based on qualitative evidence, as they did not perform a meta-analysis and considered CAT as a measure of the quality of life. By contrast, our quantitative synthesis offers a more robust evaluation, strengthening the evidence base for the effectiveness of VR-complemented PR in improving health status.

The observed improvements can be attributed to the engaging and interactive nature of VR, which enhances adherence to PR programs by reducing monotony and providing real-time feedback. Better adherence leads to improved symptom management, reduced fatigue, and enhanced physical function [68]. Additionally, the immersive VR experience can alleviate anxiety and depression, common in patients with COPD, thereby contributing to a more positive perception of health [41]. Furthermore, VR’s adaptability allows for personalized rehabilitation tailored to individual needs and disease severity, ensuring that a broader range of patients can benefit. This comprehensive approach addresses both physical and psychological dimensions, underscoring the potential of VR-complemented PR as a valuable adjunct to conventional rehabilitation for improving health status in COPD.

SpO₂, a measure of oxygen saturation in the blood, is an important indicator of respiratory function, especially in patients with COPD, who often experience oxygen desaturation during physical exertion [69]. In this meta-analysis, SpO₂ was measured at rest, and our pooled analysis revealed a significant improvement in oxygenation status in the VR-complemented PR group compared with the comparator group (MD 1.35%). This finding suggests that VR-complemented PR has the potential to enhance oxygenation in patients with COPD, contributing to overall respiratory health. Similarly, Condon et al [70] explored the effectiveness of VR gaming and exercise-based games for patients with respiratory disease and reported a significant increase in SpO₂ compared with the control group (standardized MD 0.2%).

The immersive nature of VR likely enhances engagement and motivates patients to perform more sustained and intense physical activity, which can help improve respiratory efficiency [71]. By encouraging higher-intensity exercise, VR facilitates greater use of lung capacity, which in turn may reduce the frequency and severity of desaturation events during physical exertion [72]. Additionally, VR interventions provide real-time biofeedback, allowing patients to monitor their respiratory patterns and adjust breathing techniques. This feedback helps them maintain controlled, efficient breathing during exercise, reducing the risk of rapid, shallow breathing, a common cause of desaturation in patients with COPD [73]. Moreover, VR’s interactive environment provides a more engaging and motivating rehabilitation experience compared with traditional methods, which may lead to more consistent participation and, consequently, better exercise outcomes [74]. By enhancing patient motivation and adherence, VR can foster improvements in both physical capacity and respiratory efficiency, ultimately contributing to more stable oxygen saturation levels during activity [75].

Acceptability and engagement are crucial for the success of PR programs. In this review, VR-complemented PR demonstrated good acceptability, with most patients reporting positive experiences. While 1 study found no significant difference compared with traditional PR, another reported higher satisfaction in the VR group. The immersive and interactive nature of VR likely reduces boredom and increases motivation, making rehabilitation more appealing.

Engagement, assessed through adherence rates, was consistently higher in VR-complemented PR groups. Features such as real-time feedback, gamification, and personalized difficulty levels help sustain interest and participation [61]. These findings are consistent with previous systematic reviews, which also reported that participants enjoyed VR technology and demonstrated good adherence [22,23]. VR addresses common barriers to adherence, such as monotony and lack of motivation, by transforming exercise into an engaging experience. However, challenges such as joint pain and forgetfulness still impact adherence and should be addressed through tailored strategies [44].

This review has several notable strengths. First, it is one of the most comprehensive systematic reviews and meta-analyses on VR-complemented PR for COPD, incorporating a rigorous search strategy across multiple international and Chinese databases. The inclusion of recent studies offers a robust and up-to-date synthesis of evidence. Second, stringent inclusion and exclusion criteria were applied to ensure that only high-quality RCTs were analyzed, enhancing the reliability of the findings. Third, the use of subgroup analyses, which examined factors such as comparator type and intervention duration, provided a nuanced understanding of the conditions under which VR interventions are most effective, offering valuable insights for clinical practice. Additionally, sensitivity analyses were conducted to address heterogeneity, reinforcing the robustness and consistency of the results. Finally, the assessment of patient-reported experience measures, such as acceptability and engagement, offers a holistic perspective on the intervention’s feasibility and its impact on patient-centered outcomes.

Several limitations should be acknowledged. Blinding of participants and personnel was not feasible due to the nature of VR interventions, increasing the risk of performance bias. The inclusion of studies with PR durations shorter than standard recommendations [29,30] may limit comparability to conventional PR programs. While sensitivity analyses excluding these shorter-duration studies confirmed the robustness of the findings, this limitation highlights the need for future research adhering to standardized PR protocols. Additionally, the focus on hospital-based VR programs may reduce the generalizability of the findings to home-based settings, where factors such as environmental differences and the availability of supervision could influence outcomes. A potential limitation is the possible overlap of participant samples in Rutkowski et al [39] and Rutkowski et al [40], as both studies reported identical mean age and disease severity for their comparator groups. While the intervention groups and objectives differed, this overlap may have introduced bias. Sensitivity analyses excluding 1 study confirmed the robustness of the findings, but the possibility of partial overlap should be considered when interpreting the results. Finally, while the subgroup analyses provided valuable insights into the effectiveness of VR-complemented PR, their findings should be interpreted with caution. The subgroup sizes were relatively small, limiting statistical power and increasing the potential for type I or type II errors.

Future research should aim to address these limitations by adopting more rigorous and standardized study designs. Specifically, studies with larger sample sizes and longer follow-up periods are needed to evaluate the long-term effectiveness and sustainability of VR-complemented PR. Developing standardized VR protocols that define exercise intensity, frequency, duration, and content would help reduce heterogeneity and improve the comparability of results across studies. Expanding research to include home-based VR programs is also essential, as this could enhance accessibility, reduce the burden of travel, and improve adherence, particularly for patients with mobility issues. Additionally, future studies should explore the effects of VR-complemented PR on diverse patient subgroups to identify those who may benefit the most from this intervention. Comprehensive evaluation indicators, such as psychological function, cognitive function, and frailty status, should be included to assess the full range of potential benefits.

This systematic review and meta-analysis demonstrate that VR-complemented PR effectively improves lung function, exercise capacity, dyspnea, health status, and oxygenation status compared with comparators in patients with COPD. The engaging and immersive nature of VR enhances patient adherence and participation, addressing key limitations of traditional PR. Subgroup analyses revealed that VR-complemented PR had a significantly greater effect on FEV₁ (L) and 6MWD when compared with the nonactive exercise control groups. Additionally, VR-complemented PR showed a greater improvement in FEV₁/FVC compared with the active exercise control groups. The effectiveness of VR interventions in 6MWD also varied with intervention duration, with the most pronounced benefits observed in programs lasting 5-12 weeks. The findings underscore the potential of VR as an innovative, patient-centered adjunct to traditional PR. Future studies should focus on long-term outcomes, standardized protocols, and the applicability of VR in diverse and home-based settings to optimize its clinical implementation.