The effectiveness of non-invasive positive pressure ventilation combined with rehabilitation training in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis

Abstract

Background:

Pulmonary rehabilitation (PR) is a cornerstone of chronic obstructive pulmonary disease (COPD) management, but exercise intolerance often limits its effectiveness. Non-invasive positive pressure ventilation (NPPV) during PR may enhance training tolerance and outcomes, yet the overall evidence remains uncertain.

Objectives:

To evaluate the effects of adding NPPV to PR on exercise capacity, dyspnea, and respiratory muscle strength in patients with COPD.

Design:

Systematic review and meta-analysis of randomized controlled trials (RCTs).

Data sources and methods:

We searched PubMed, Web of Science, Embase, the Cochrane Library, and CINAHL from inception to December 2024 for RCTs evaluating NPPV combined with PR in patients with COPD. Two reviewers independently assessed risk of bias using the Cochrane Risk of Bias Assessment Tool for Randomized Trials (RoB 2), extracted data, and performed analyses using RevMan 5.3.

Results:

A total of 17 RCTs (489 participants; predominantly GOLD stage III–IV) were included. NPPV + PR significantly improved 6-minute walk distance (MD = 29.1 m, 95% CI: 3.6–54.6), incremental shuttle walk test distance (MD = 21.8 m, 95% CI: 5.0–38.7), peak oxygen uptake (SMD = 0.53, 95% CI: 0.23–0.83), maximal inspiratory pressure (Pimax; MD = 5.8 cmH₂O, 95% CI: 1.0–10.7), and maximal expiratory pressure (Pemax; MD = 14.9 cmH₂O, 95% CI:4.1–25.7). Significant reductions were observed in blood lactate levels (MD = −0.49 mmol/L, 95% CI:−0.79 to −0.19), BORG dyspnea score (MD = −1.1, 95% CI: −1.7 to −0.6), and mMRC scale (MD = –0.3, 95% CI: −0.5 to −0.1). No significant effect was found on quality of life.

Conclusion:

Adding NPPV to exercise-based PR provides clinically meaningful improvements in exercise capacity, dyspnea, and respiratory muscle strength in patients with COPD who have significant exercise limitation. NPPV may be a valuable adjunct to optimize PR outcomes in this population.

Trial registration:

This review was prospectively registered in PROSPERO (CRD42023486598).

Introduction

Chronic obstructive pulmonary disease (COPD) is a major global public health challenge, with an estimated 480 million cases worldwide, and ranks as the third leading cause of global mortality.^1,2 This disease is characterized by persistent airflow obstruction and is often accompanied by systemic complications, among which muscle atrophy is particularly prevalent. The pathogenesis of muscle atrophy in COPD involves mechanisms such as systemic chronic inflammation, oxidative stress, hypoxia, and disuse, which in turn exacerbate the condition by reducing muscle strength, diminishing exercise capacity, and worsening dyspnea, forming a vicious cycle that increases re-hospitalization and mortality.^3,4

Pulmonary rehabilitation (PR) is a cornerstone non-pharmacological management strategy for COPD, with exercise training as its core component. It is recognized by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) as a first-line intervention.^5,6 However, many patients with COPD, particularly those with severe disease, are unable to achieve adequate exercise intensity and duration due to expiratory airflow limitation and dynamic pulmonary hyperinflation during exertion. This leads to increased dyspnea, oxygen desaturation, and reduced exercise tolerance.^7–9

To overcome these limitations, various forms of respiratory support, including non-invasive positive pressure ventilation (NPPV), have been integrated into exercise training. Theoretically, NPPV can alleviate the work of breathing, reduce dynamic hyperinflation, and improve gas exchange, thereby potentially enhancing exercise performance and endurance.^10,11 While several studies have investigated the combination of PR and NPPV in COPD patients, the findings regarding its effects on exercise capacity, dyspnea, and quality of life have been inconsistent.^12–15 Therefore, this systematic review and meta-analysis aimed to synthesize the existing evidence to evaluate the efficacy of combining NPPV with pulmonary rehabilitation training in patients with COPD.

Method

The reporting of this systematic review and meta-analysis conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. ¹⁶

Search strategy

We searched five electronic databases (PubMed, Embase, CINAHL, Web of Science, and Cochrane Library from their inception to December 2024. The search strategy combined free text and medical subject heading terms related to the PICOS components. Boolean logical operators were used to connect search terms, including:

• Population: “chronic obstructive pulmonary disease” OR “chronic obstructive airway disease” OR “chronic airflow obstruction” OR “COPD.”

• Intervention: “non-invasive positive pressure ventilation OR mechanical ventilation OR positive pressure ventilation OR non-invasive mechanical ventilation OR non-invasive ventilation.”

• Comparison: “pulmonary rehabilitation OR respiratory rehabilitation OR rehabilitation training OR exercise therapy OR exercise training OR respiratory therapy OR pulmonary exercise.”

• Study design: “randomized controlled trial” [Publication Type]

Relevant synonyms and variations of the terms were flexibly used to ensure the comprehensiveness of literature retrieval. Only articles published in English or Chinese were included. Detailed search strategies and the completed PRISMA 2020 checklist are provided in the Supplemental Material.

Criteria for study inclusion and exclusion

The inclusion criteria for studies were as follows: (1) Study design: Randomized controlled trials (RCTs) assessing the effectiveness of NPPV during pulmonary rehabilitation exercise training. (2) Participants: Patients diagnosed with COPD according to the Global Initiative for Chronic Obstructive Lung Disease guidelines. (3) Intervention: The intervention group received NPPV during supervised pulmonary rehabilitation exercise sessions, while the control group received either room air or nasal cannula. Pairwise comparisons between interventions were also conducted. (4) Outcomes: Studies were required to report at least one of the following outcomes: exercise capacity (defined as peak exercise capacity, endurance exercise capacity or functional exercise capacity measured post exercise training), health-related quality of life (measured using disease-specific or generic HRQL instruments), physiological changes related to exercise training (e.g., peak VO2, blood lactate levels, minute ventilation), and dyspnea index (e.g., Borg score, modified Medical Research Council).

The exclusion criteria for studies were as follows: (1) Participants with non-COPD respiratory disease or participants with concomitant neuromuscular disease, a restrictive thoracic disorder, significant cardiac failure or cardiac disease, if data from participants with COPD could not be analyzed separately. (2) Only abstracts were available without full texts, and efficacy was not reported. (3) Studies with incomplete data, such as protocols, were also excluded.

Selection and data collection

Literature screening and data extraction were conducted independently by two researchers trained in evidence-based nursing. NoteExpress software (Beijing Aegean Sea Lezhi Technology Co., Ltd., Beijing, China) was used to remove duplicates and non-conforming articles. The title and abstract were read for preliminary screening. Then, the full texts were studied for further consideration. Two researchers independently assessed the full texts based on inclusion or exclusion criteria and extracted data, including participants, sample size, study design, rehabilitation training characteristics, intervention measures of the experimental group and control group, outcome indicators. In case of disagreement, discussion and analysis will be conducted, and if necessary, a third researcher will be consulted to reach a consensus.

Quality assessment

Two investigators independently evaluated the risk of bias in the included randomized controlled trials (RCTs) using the updated Cochrane Risk of Bias Tool for Randomized Trials (RoB 2). This tool assesses bias across five domains: ① Bias arising from the randomization process; ② Bias due to deviations from intended interventions; ③ Bias due to missing outcome data; ④ Bias in measurement of the outcome; ⑤ Bias in selection of the reported result. For each domain, the risk of bias was judged as “low risk,” some concerns,” or high risk.” The overall risk of bias for each study was determined based on the judgments across all domains. Disagreements between the investigators were resolved through discussion or by consulting a third reviewer if necessary.

Statistical analysis

Meta-analysis was performed using Review Manager 5.3. standardized mean difference (SMD) or mean difference (MD) were used as effect analysis statistics for continuous variables. χ² test and I² were used to determine the heterogeneity of the included studies. When p < 0.01 and I² > 50%, the test and statistics showed that the heterogeneity was high, so the random effect model was selected; otherwise, the fixed-effect model was adopted. Potential sources of clinical heterogeneity in the included studies were investigated by subgroup analyses. p < 0.05 was considered statistically significant. Sensitivity analyses were performed to assess the robustness of the pooled effect estimates by sequentially excluding studies that used supplemental oxygen or sham NIV as the control intervention.

Results

Selection process and study characteristics

A total of 1288 literature were obtained through preliminary search, and 17 literature were finally included, including 489 patients. 460 out of the 489 patients were at COPD patients at GOLD stage III or IV with forced expiratory volume in 1 second (FEV₁ ) less than 50% predicted. The remaining 29 patients from one study ¹⁷ had FEV₁ < 60% predicted. The process of literature screening is shown in Figure 1. The basic characteristics of the included studies are shown in Table 1.

Figure 1.

PRISMA flow diagram of study selection.

Table 1.

Characteristics of included studies.

Study	Country	Design	Sample	Participants ① Age, years ②male/Female	Rehabilitation training characteristics ① Type of exercise training ②Total duration of training ③Number of sessions per week ④Duration of each session	Experimental group	Control group
Schneeberger et al. (2023) 18	Germany	RCT	13/13	①E:59 ± 5,C:63 ± 7 ②E:4/9, C:7/6	①Not reported ②3 weeks ③5 days per week ④30 min each session	PR+HI-NPPV IPAP:26.0 ± 3.5cmH2O EPAP:5.7 ± 0.8cmH2O	PR+Room Air
Elmorshidy et al. (2023) 19	Egypt	RCT	22/21	①E:60.88 ± 8.74,C:60.56 ± 9.07 ②E:not reported	①Aerobic exercise, Stationary bicycle training, Upper limb exercise training, Respiratory training ②8 weeks ③5 days per week ④30–45 min (Aerobic exercise), 15 min(Stationary bicycle), 30 min (Upper limb exercise), 5–10 min (breathing exercises )	PR+NPPV IPAP: Initial 6 cm H2O EPAP: Initial 3 cm H2O	PR+Room Air
Deniz et al. (2023) 15	Turkey	RCT	12/9	①E:64.2 ± 7.3,C:66.1 ± 5.4 ②E:11/1,C:9/0	①Cycle ergometer, treadmill, lower and upper extremity strength training, Breathing exercises, Relaxation exercises ②8 weeks ③twice weekly ④for total 60-80 min/day, cycle ergometer (15 min), treadmill (15 min), lower and upper extremity strength training (5–10 min), breathing and relaxation therapies (15–20 min)	PR+NPPV IPAP: Initial 10 cmH2O EPAP: Initial 4 cmH2O	PR+Room Air
Labeix et al. (2022) 20	France	RCT	13/7	①64 ± 7 ②13/7	①Cycling exercise with NIV, upper limbs reinforcement (bench press, arm cycleergometer or gymnastics exercises), and strengthening of the lower limbs (seated press or gymnastics). ②6 weeks ③cycling exercise three sessions per week, upper limbs reinforcement one session per week, and strengthening of the lower limbs once a week ④75 min for each session	PR+NPPV-IPS IPAP: Initial 10 cmH2O EPAP: Initial 4 cmH2O	PR+NIV-sham
Marrara et al. (2018) 21	Brazil	RCT	22/21	①E:67.8 ± 8.9,C:68.2 ± 8.5 ②E:17/4, C:18/4	①Stretching for the cervical muscles and the upper and lower limbs ②6weeks ③3 times per week ④5 min of stretching for the cervical muscles and the upper and lower limbs	PR+NPPV IPAP: Initial 6 cm H2O EPAP: Initial 3 cm H2O	PR+Room Air
Gad et al. (2015) 22	Egypt	RCT	15/15	①E:65.70 ± 10,C:66.41 ± 9 ②E: 9/6, C:11/4	①breathing exercises, shoulder wheel exercise, walks or riding a stationary bicycle ②12 weeks ③2 times a week ④Each training session lasted approximately 30–35 min	PR+NIV IPAP: 15–20 cmH2O EPAP: 3–6 cmH2O	PR+Room Air
Márquez-Martín Eduaro et al. (2014) 23	Spain	RCT	14/14	not reported	①combination of resistance and strength training ②12 weeks ③3 times per week ④sessions lasting for 40-min plus a 10-min warm-up and 10-min cool down	PR+NPPV IPAP: Initial 10 cmH2O EPAP: Initial 4 cmH2O	PR+Room Air
Duiverman et al. 2011( 24 )	The Netherlands	RCT	15/20	①E:63 ± 10 C:61 ± 8 ②E:16/ 8, C:17/15	①cycling exercises, walking, and inspiratory muscle training ②12 weeks ③1–2 times per week ④30 min for each session	PR+NPPV IPAP: increased up to maximal tolerated pressure	PR+Room Air
Borghi-Silva et al. (2010) 25	Brazil	RCT	12/12	E:68 ± 9, C:67 ± 7 E:9/5, C:9/5	①treadmill ②6 weeks ③3 times per week ④60 min	PR+NPPV IPAP: Initial 6 cm H2O EPAP: Initial 3 cmH2O	PR+O2
Duiverman et al. (2008) 26	The Netherlands	RCT	24/32	①E:63 ± 10,C:61 ± 7 ②E:18/13,C:17/18	strength training, cycling, walking, inspiratory muscle training 12 weeks 3 times per week 60 min	PR+NIV IPAP: increased up to the maximal tolerated pressure	PR+Room Air
Toledo et al. (2007) 17	Brazil	RCT	9/9	E:66.8 ± 10.6, C:67.6 ± 8.5 not reported	①treadmill ②12 weeks ③3 times per week ④30 min	PR+NPPV IPAP:10–15 cmH2O EPAP:4–6 cmH2O	PR+Room Air
van et al. (2006) 27	The Netherlands	RCT	15/14	①E:70 ± 5, C:71 ± 4 ②E:10/4, C: 14/1	①cycle ergometry ②8 weeks ③3 times per week ④45 min	PR+NPPV: IPAP: 10 cmH2O EPAP: not reported	PR+NIV-sham
Reuveny et al. (2005) 28	Israel	RCT	9/10	①E:64 ± 9,C:63 ± 9 ②not reported	①training: treadmill ②8 weeks ③2 times per week ④45 min	PR+NPPV+O2 IPAP:7–10 cmH2 EPAP: 2 cmH2O d	PR+O2
Johnson et al. (2002) 29	United States of America	RCT	11/11	E:69 ±9,C:67 ± 8 E:8/3,C: 7/4	①training: treadmill ②6 weeks ③2 times per week ④20 min	PR+NPPV(BiPAP): IPAP: 8-12 cmH2O EPAP:2 cmH2O	PR+Medical Air
Bianchi et al. (2002) 30	Italy	RCT	9/10	①E:64 ±5,C:65 ± 7 ②Both male	①training: cycle ergometry ②6 weeks ③3 times per week ④30 min	PR+NPPV proportional assist ventilation: flow assist 3.5 (1.6) cmH2O/L/s, volume assist 6.6 (2.2) cmH2O/L with EPAP 2 cmH2O	PR+Room Air
Hawkins et al. (2002) 31	United Kingdom	RCT	10/9	①E:68 ± 9.1,C:66 ± 6.8 ②not reported	①training: cycle ergometry ②6 weeks ③3 times per week ④30 min	PR+NPPV(BiPAP-PAV) proportional assist ventilation: flow assist 3.6 (0.7) cmH2O/L/s, volume assist 12.7 (1.5) cmH2O/L during exercise training	PR+Room Air
Garrod et al. 2000 32	United Kingdom	RCT	17/20	①E:62 ±12,C: 67 ± 6 ②not reported	①upper and lower limb ②8 weeks ③2 times per week ④60 min	PR+NPPV(night) IPAP:13–24 cmH2O EPAP:4–6 cmH2O	PR+Room Air

Risk of bias of included studies

Most studies demonstrated a low risk of bias in the randomization process(15/17), deviations from intended interventions (12/17), missing outcome data (14/17), measurement of the outcome (17/17), and selection of the reported results (17/17). However, one study was rated high risk in deviation from intended interventions, and one study was rated high risk in missing outcome data. Overall, nine studies were judged to have a low risk of bias, while eight studies raised some concerns, primarily due to issues in deviations from intended interventions or missing data. Detailed results are presented in Figure 2.

Figure 2.

Risk of bias assessment of included studies.

Meta-analysis results

Exercise capacity

The 6-minute walk distance (6MWD) data were available for nine studies. Due to the heterogeneity (I² = 56%, p = 0.02), the random-effects model was selected. The meta-analysis showed a statistically significant improvement in 6MWD (MD = 29.12, 95% CI (3.62, 54.63), p = 0.03) for patients in the PR+NPPV group compared to the control group (Figure 3(a)). Three studies reported the Incremental Shuttle Walk Test (ISWT). A fixed-effects model was used due to low heterogeneity among studies (I² = 41%, p = 0.18). The meta-analysis showed that the ISWT distance of the PR+NPPV group was significantly higher than that of the PR group [MD = 21.81, 95% CI (4.96, 38.66), p = 0.01] (Figure 3(b)). For work rate (Figure 3(c)) and work time (Figure 3(d)) in the exercise test, the NPPV intervention was both better than pulmonary rehabilitation alone (MD = 1.06, 95% CI (0.38, 1.75), p < 0.01; MD = 1.91, 95% CI (0.20, 3.62), p = 0.03).

Figure 3.

Forest plot for analysis of exercise capacity. (a) 6-min Walk test distance. (b) Incremental shuttle walk test. (c) Work rate in the exercise test. (d) work time in the exercise.

Physiological outcomes

A total of six studies reported Pimax, and four studies reported Pemax, which are indicators of the maximum contractility of the respiratory muscles. A fixed-effect model was used in the meta-analysis, Both pimax and pemax in the PR+NPPV group were significantly higher than those in PR group (MD = 5.84, 95% CI (0.97, 10.71), p = 0.02; MD = 14.87, 95% CI (4.06, 25.68), p < 0.01), indicating that NPPV intervention may improve respiratory muscle strength (Figure 4(a) and (b)).

Figure 4.

Forest plot for analysis of physiological outcomes. (a) Pimax, (b) Pemax, (c) VO2Peak, (d) VE, and (e) Blood lactate.

VO₂Peak (peak oxygen uptake) was reported in seven studies. A fixed-effects model was used due to low heterogeneity (I² = 4%, p = 0.40). The meta-analysis showed a statistically significant difference in VO₂peak (SMD = 0.53, 95% CI (0.23, 0.83), p < 0.01) (Figure 4(c)).

Ventilatory equivalent (VE) was reported in seven studies. A fixed-effects model was selected due to low heterogeneity (I² = 0%, p = 0.71). The meta-analysis showed a statistically significant improvement in VE (MD = 1.44, 95% CI (0.19, 2.69), p = 0.02) for patients in the PR+NPPV group compared to patients in the PR group (Figure 4(d)).

Blood lactate levels were reported in five studies, A fixed-effect model was chosen due to low heterogeneity (I² = 16%, p = 0.32). The results showed that the blood lactate levels in the PR+NPPV group were significantly lower than those in the PR group after exercise (MD = −0.49, 95%CI (−0.79, 0.19), p < 0.01) (Figure 4(e)).

Dyspnea index

A total of seven studies used the BORG scale to assess dyspnea. A fixed-effects model was used for meta-analysis due to low heterogeneity (I² = 39%, p = 0.13). The results showed that the dyspnea index exhibited a greater reduction in the PR+NPPV group compared to the PR group before and after exercise [MD = −1.12, 95%CI (−1.65, −0.59), p < 0.01]. Four studies reported dyspnea using the modified Medical Research Council (mMRC) scale. A fixed-effects model was applied due to low heterogeneity (I² = 0%, p = 0.77). The results demonstrated a smaller increase in dyspnea severity in the PR+NPPV group than in the PR group (MD = −0.29, 95%CI (−0.51, −0.06), p = 0.01) Figure 5).

Figure 5.

Forest plot for analysis of dyspnea. (a) BORG scale and (b) mMRC scale.

Quality of life

Quality of life was reported in 12 studies, but different measurement tools were used. We grouped the scales into two categories: those with higher scores indicating better quality of life and those with higher scores indicating worse quality of life. Subgroup analysis was performed for different questionnaires. The results showed no statistically significant changes in quality of life between the experimental and control groups (MD = 0.3, 95%CI (−0.12, 0.72), p = 0.16; MD = −0.83, 95%CI (−2.79, 1.13), p = 0.41). Subgroup analysis of different scales also showed no statistically significant difference (Figure 6).

Figure 6.

Forest plot for analysis of quality of life. (a) higher scores for better quality of life and (b) higher scores for worse quality of life.

Sensitivity analysis

To evaluate the potential impact of heterogeneity in control interventions (e.g., room air, supplemental oxygen, or sham NIV), sensitivity analyses were conducted by excluding studies that used supplemental oxygen^25,28 and, separately, those that used sham NIV.^20,27 The pooled effect estimates for key outcomes—including 6-minute walk distance, BORG dyspnea score, and peak oxygen uptake (VO₂ peak)—showed minimal changes in magnitude and direction, and the overall conclusions remained unchanged.

Discussion

The 6MWD is a well-established prognostic marker in COPD, strongly associated with mortality and hospitalization risk.^33,34 The observed mean improvement of 29.1 m with PR+NPPV exceeds the widely accepted minimal clinically important difference (MCID) of 25 m for COPD, ³⁵ indicating a clinically meaningful enhancement in exercise capacity. In contrast, the ISWT—which better reflects the intermittent nature of daily life—showed a significant but modest improvement of 21.8 m, falling short of the established MCID of 48 m for chronic respiratory diseases. ³⁵ This discrepancy may stem from differences in test sensitivity or the specific mechanisms by which NPPV augments exercise performance. Notably, NPPV also significantly improved both work rate and endurance time, further supporting its beneficial role in enhancing exercise tolerance.

We observed that patients in the NPPV combined with PR group demonstrated significant improvements in several key physiological indicators, including Pimax, Pemax, VO₂peak, VE, and blood lactate levels. While these improvements were statistically significant, their clinical relevance remains uncertain due to the absence of established MCID values for these parameters. Further research is needed to determine the clinical significance of these findings and to establish standardized MCID thresholds that can better inform the interpretation of such outcomes in future studies.

Controlling dyspnea has always been one of the important goals of disease management in COPD patients. ³⁶ While Pulmonary rehabilitation significantly reduces dyspnea in COPD patients, it may be insufficient for those with severe respiratory failure or acute exacerbations.^37,38 Consequently, many studies have focused on combining NPPV with PR, demonstrating that this strategy yields more positive outcomes in alleviating dyspnea.^18,19,20,21 Meta-analysis revealed that the BORG scale scores in the PR+NPPV group showed a greater decline, exceeding the MCID of 1 point, ⁴¹ indicating both statistical and clinical significance. This suggests that NPPV may alleviate breathing difficulties during physical exertion. In contrast, the mMRC scale demonstrated a smaller increase in dyspnea severity in the PR+NPPV group compared to the PR group. Although statistically significant, this change did not reach the MCID of 1 point for the mMRC scale, ⁴² suggesting limited clinical relevance. The discrepancy between the two scales may be attributed to the differing contexts of dyspnea assessment: exercise versus daily activities. This suggests that NPPV may be particularly effective in addressing exertion-related symptoms.

Although we found that NPPV combined with PR showed significant improvements in exercise endurance, dyspnea, and some indicators of physiological function, surprisingly, these improvements did not translate directly into improvements in quality of life. The reasons may be explained as follows. First, quality of life is a complex concept that includes personal values, satisfaction, living conditions, functionality, cultural contexts, and spirituality. ⁴³ PR program is a comprehensive intervention that includes not only exercise training, but also disease self-management, medication, nutrition, health-related behavior change, and emotional support. ⁴⁴ NPPV is just a singular component affecting the exercise performance, not the whole PR program. Symptom relief in dyspnea or fatigue was observed in some included studies,^15,23,26,32 while the total scores in the quality of life questionnaires did not report an increase in the exercise training + NPPV group when compared with control groups.^{15,21,23,24,26,30} The influence of NPPV on increasing the quality of life within PR groups is limited and should not be overlooked. Therefore, even if the physical function is improved by the support from NPPV, that does not mean the NPPV can easily increase one’s whole status of well-being. In addition, quality of life is an index reflecting patient’s long-term condition, changes may not be seen during a short time following-up. The run-in period of the included studies was very short, ranging from 3 weeks to 12 weeks. Changes in some domains of quality of life, such as social support or emotion, cannot be expected to be seen in such a short time. Longitude cohort studies are required to further explore how to optimize the integration of non-invasive ventilation and pulmonary rehabilitation, and combine psychological support and behavioral intervention to comprehensively improve patients’ long-term quality of life.

Our group recently reported a network meta-analysis showing NIV+PR as the most effective respiratory support strategy in COPD, ⁴⁵ that study focused on relative rankings among multiple interventions. In contrast, this pairwise meta-analysis quantifies the absolute benefit of NIV+PR versus PR alone—demonstrating clinically meaningful improvements in exercise capacity, dyspnea, and respiratory muscle strength, with effect sizes directly applicable to clinical practice and trial design. Our results suggest that the benefit of combining NPPV with rehabilitation stems from a dual action: it simultaneously permits a greater volume of exercise and enhances its metabolic quality. By easing the burden of breathing, NPPV enables patients to achieve a higher training intensity, increasing the exercise dose. Concurrently, the observed improvements in peak oxygen uptake and blood lactate levels indicate that NPPV also creates a more favorable metabolic environment. This likely results from improved oxygenation and a shift in blood flow to the limb muscles, meaning the exercise performed is not just longer, but also physiologically more efficient.

This review has several limitations. First, considerable heterogeneity arose from variations in study design, including differences in pulmonary rehabilitation protocols (exercise modality, frequency, and duration), sample sizes, and outcome assessment methods, which may affect the internal validity of the pooled estimates. Although our analysis focused on within-group pre–post changes—thereby mitigating the impact of baseline imbalances—the diversity in intervention delivery limits definitive conclusions regarding optimal training prescriptions. Second, heterogeneity existed in both intervention and control protocols: while most trials applied NIV during exercise, a few used nocturnal NIV alone—approaches with distinct physiological mechanisms; meanwhile, control groups varied (e.g., room air, supplemental oxygen, or sham NIV), which may exert differing physiological effects despite lacking active ventilatory support. Future trials should clearly differentiate NIV delivery strategies. Subgroup analyses were not feasible due to sparse data, but sensitivity analyses excluding heterogeneous studies yielded consistent results, supporting the robustness of our findings. Finally, key subgroup analyses—by COPD severity, NIV pressure settings, or specific pulmonary rehabilitation components—were not possible. The overwhelming majority of participants had GOLD stage III–IV disease, precluding comparisons across severity levels; IPAP settings varied widely (6–26 cmH₂O) and were largely individualized; and all rehabilitation programs were multimodal, with no single exercise modality tested consistently across studies. Collectively, these gaps limit generalizability to milder COPD and highlight the need for large-scale, standardized trials to refine patient selection, optimize training protocols, and clarify the role of NPPV across the full spectrum of COPD severity.

Conclusion

The combination of NPPV and PR interventions offers benefits for COPD patients, particularly in improving exercise endurance, key physiological indicators, and dyspnea. Although no significant advantage was found in improving quality of life, this does not negate the overall positive effect. Future studies should focus on exploring the optimal timing, duration, and mode of combining NPPV with pulmonary rehabilitation to establish the most effective implementation strategy. In addition, high-quality RCTs with larger sample sizes are needed to strengthen the evidence base and provide more definitive conclusions regarding the clinical benefits of this combined approach.