Effects of Preoperative Breathing Exercise on Postoperative Outcomes for Patients With Lung Cancer Undergoing Curative Intent Lung Resection: A Meta-analysis

Abstract

Objective:

To determine the benefits of preoperative breathing exercises on hospital length of stay (LOS), pneumonia, postoperative pulmonary complications (PPC), 6-minute walk distance (6MWD), forced expiratory volume in 1 second (FEV₁), and health-related quality of life (HRQOL) in patients undergoing surgical lung cancer resection.

Data Sources:

PubMed, EMBASE, Web of Science Core Collection, and Cochrane Central Register of Controlled Trials were comprehensively searched from inception to March 2021.

Study Selection:

Only studies including preoperative inspiratory muscle training (IMT) and/or breathing exercises compared with a nontraining control group were included. The meta-analysis was done using Cochrane software for multiple variables including LOS, pneumonia, PPC, 6MWD, FEV₁, mortality, and HRQOL.

Data Extraction:

Two authors extracted the data of the selected studies. The primary outcomes were LOS and PPC.

Data Synthesis:

A total of 10 studies were included in this meta-analysis, 8 of which had both IMT and aerobic exercise. Pooled data for patients who performed preoperative breathing exercises, compared with controls, demonstrated a decrease in LOS with a pooled mean difference of −3.44 days (95% confidence interval [CI], −4.14 to −2.75; P<.01). Subgroup analysis also demonstrated that LOS was further reduced when breathing exercises were combined with aerobic exercise (χ², 4.85; P=.03). Preoperative breathing exercises reduce pneumonia and PPCs with an odds ratio of 0.37 (95% CI, 0.18-0.75; P<.01) and 0.37 (95% CI, 0.21-0.65; P<.01), respectively. An increase in 6MWD of 20.2 meters was noted in those performing breathing exercises (95% CI, 9.12-31.21; P<.01). No significant differences were noted in FEV₁, mortality, or HRQOL.

Conclusions:

Preoperative breathing exercises reduced LOS, PPC, and pneumonia and potentially improved 6MWD in patients undergoing surgical lung cancer resection. Breathing exercises in combination with aerobic exercise yielded greater reductions in LOS. Randomized controlled trials are needed to test the feasibility of introducing a preoperative breathing exercise program in this patient population.

Lung cancer is the predominant cause of cancer-induced mortality worldwide and is the second most common cancer in men and women, representing approximately 15% of all new cancer cases.¹ As lung cancer screening becomes the standard for early detection, the epidemiology of this deadly cancer will continue to change. As a result of new and improved screening programs, it is expected that a greater proportion of lung cancers will be diagnosed at earlier, more treatable stages. Almost 85% of new lung cancer cases represent the histologic group of non–small cell lung cancer, and surgical resection is the treatment of choice for stages I-IIIa. Resection offers the highest potential for survival.² However, in this high-risk population, surgery is not without risk or consequence and is not always an option because of preexisting comorbidities, such as chronic obstructive lung disease (COPD).

It is estimated that postoperative complications occur in 38%-58% of patients, with 15%-25% directly related to respiratory health (eg, lung infection, pneumonia, atelectasis), impairments that contribute to hospital length of stay (LOS), functional performance, medical costs, cancer recurrence, 30-day readmission rates, and mortality.^1,3-6 Risk factors such as age, smoking status, cardiopulmonary exercise capacity (peak oxygen uptake), pulmonary disease (forced expiratory volume in 1 second [FEV₁]), and muscular weakness, including that of the respiratory muscles, can predict postoperative complications and survival.^7-10 Diaphragm weakness after lung resection surgery is also associated with a reduction in cardiopulmonary exercise capacity as well as survival^11,12 and will predict surgical complications up to 2 years after thoracic surgery,^7,13 negatively affecting long-term health-related quality of life (HRQOL).¹⁴ Rehabilitation programs targeting the respiratory system are beginning to demonstrate that they can reduce postoperative complications after lung resection surgery.¹⁵

Improving presurgical pulmonary health and aerobic capacity is becoming more common because research is beginning to demonstrate that exercise performed in the period leading up to or immediately after lung resection surgery may reduce postoperative morbidity and mortality.¹⁶ Unfortunately, there is not always time for a traditional cardiopulmonary rehabilitation program before surgery⁸ because there tends to be urgency to proceed with definitive therapy once a diagnosis of lung cancer is made. Instead, breathing exercises (eg, pursed-lip breathing, abdominal breathing, thoracic breathing, incentive spirometry, inspiratory muscle training [IMT]) are logistically easier to perform and will improve exercise performance, dyspnea, and HRQOL in patients with lung disease.^4,6,7,17,18 It is theorized that breathing exercises reduce postoperative complications by improving chest wall motion, arterial oxygen saturation, and dyspnea; preventing small airway collapse; and more evenly distributing ventilation throughout the lungs.

A recent meta-analysis in patients electing surgery for lung cancer demonstrated that breathing exercises, including IMT, performed before or immediately after surgery improved postoperative pulmonary function, LOS, and pulmonary complications.¹⁶ However, the included studies were limited in sample size, were primarily cohort studies with limited randomization schemes, involved numerous types of breathing exercises, and included interventions that did not distinguish between pre- and postoperative breathing exercises.^16,19 Therefore, we performed a meta-analysis including studies only in the preoperative setting that incorporated breathing exercises with or without aerobic exercise that were compared with a usual care control group. The purpose of this meta-analysis was to systematically review the published literature and derive a pooled estimate detailing the benefits associated with preoperative breathing exercises on postoperative outcomes. The primary outcomes were LOS and postoperative pulmonary complications (PPC), including pneumonia. Secondary outcomes included pulmonary function and HRQOL for patients electing curative intent resection for lung cancer.

Methods

This meta-analysis was performed in accordance to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.²⁰

Search strategy

We performed a computer-generated literature search using the following electronic databases: PubMed, EMBASE, Web of Science Core Collection, and Cochrane Central Register of Controlled Trials. The data bases were searched from inception up until March 15, 2021 by an experienced research librarian (M.L. Z.). Search terms used included: “lung” or “pulmonary” and “cancer,” “carcinoma” or “tumor” or “neoplasm” or “malignancy,” “preoperative,” “prehabilitation,” “inspiratory” or “respiratory” muscle training,” “breathing exercises or “retraining,” “length of stay,” “postoperative complications,” “pneumonia,” “exercise capacity,” “6-min walk distance,” “lung function,” “quality of life,” and/or “mortality” (see supplemental table S1 for the full search strategy, available online only at http://www.archives-pmr.org/). Two authors (C.P., H.B.) individually screened the studies against our inclusion criteria. The list of included studies then underwent full text review for eligibility.

Eligibility criteria

Inclusion criteria were randomized controlled trials comparing preoperative IMT or breathing exercises with standard of care in patients electing curative intent resection for non–small cell lung cancer. Lung resection included wedge resection, segmentectomy, lobectomy, bilobectomy, or pneumonectomy. Types of breathing retraining or IMT that were considered for inclusion were any type of IMT (“inspiratory muscle training,” “threshold loading,” “normocapnic hyperpnoea”) and breathing exercises that have a broad range of respiratory exercises, including abdominal breathing, incentive spirometry, and thoracic expansion. The IMT could be standalone or part of a complete pulmonary rehabilitation program.

We excluded studies involving pediatrics population (age <16y), non-English language articles, cardiac surgery, upper abdominal surgery, noncancer resection thoracic surgery, and studies that included IMT or breathing exercises as part of the usual care in the control group. We did not exclude studies combining IMT or breathing exercises with aerobic exercise because there were only 2 studies that performed IMT or breathing exercises alone.

Outcome of interest

Our primary outcomes of interest were LOS and PPC. Secondary outcomes included mortality; Clavien-Dindo classification for surgery complications; pneumonia; and change in FEV₁, diffusing capacity for carbon monoxide (DLCO), 6-minute walk distance (6MWD), and HRQOL. The predetermined minimal clinically important difference (MCID) for the following outcomes were: FEV₁ >0.1 L,²¹ 6MWD>26 meters,²² and HRQOL (European Organisation for Research and Treatment of Cance Core Quality of Life questionnaire-C30) score change≥5.^23,24

Data selection process

Two authors (C.P., H.B.) individually screened the available studies. Verification of eligibility was determined based on information from the title and abstract. If the information was insufficient, the full text was assessed. The extracted studies were then reevaluated to decide on their eligibility for our meta-analysis. Decisions by the 2 authors were compared and any discrepancies were resolved by a third author (M.J.M.).

Data extraction

Data extraction was performed by 2 authors (C.Y., H.B.). The following data were extracted: authors, publication year, journal, sample size, primary and secondary outcomes, type of IMT intervention, type of breathing exercise, type of aerobic exercise intervention, duration and frequency of rehabilitation program, and whether studies included patients with a diagnosis of COPD.

Pre- and postintervention change in FEV₁, DLCO, and 6MWD were measured at the time of diagnosis and immediately before lung resection surgery as this was the only data presented in the articles that we found.

Risk of bias

The risk of bias and quality of the included studies were assessed using the Cochrane risk of bias assessment tool. The tool addresses 7 specific domains of potential bias: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other biases. Risk of bias assessment was performed for all the included studies individually by 2 authors (C. Y., H.B.); a third author (M.J.M.) was available to resolve any disagreements (supplemental table S2, available online only at http://www.archives-pmr.org/).

Quality analysis

A summary of findings table was made to describe the primary outcomes included in the analysis. We also used the table to describe our assessment of the evidences according to the GRADE criteria.²⁵ The GRADE system (Grading of Recommendations Assessment, Development and Evaluation) is a framework for grading certainty of evidence and strength of recommendations.

Statistical analysis

Data were pooled into Cochrane Collaboration’s Review Manager (RevMan) version 5.3.^a RevMan was used for statistical analyses and generation of forest plots. Ninety-five percent confidence intervals (CIs) for pooled estimates were derived for each variable of interest when ≥2 studies were available. The random effects model was used in view of significant variability between studies. The DerSimonian-Laird method assumes that the included studies are estimating different but related intervention effects.²⁶ This is based on the inverse-variance approach. Data heterogeneity was defined as I²>50%.

Results

Study selection

A flow diagram representing the literature retrieval process is presented in figure 1. A total of 2161 references were imported for screening. After removing all duplicates, there were 1731 references remaining. After the initial screening of titles, another 1621 references were excluded. Of the 110 relevant studies identified for full text review, only 10 studies met the criteria for the current meta-analysis.^4,6,7,27-33

Fig 1.

Literature search to identify relevant IMT trials.

Study designs

All 10 studies included were randomized controlled trials (RCTs), as shown in table 1. One of the studies was performed in the United States,²⁷ 1 in Turkey,³¹ 1 in Italy,³² 1 in Serbia,³⁰ and the other 6 were conducted in China.^{4,6,7,28,29,33} The study by Benzo et al in 2011 conducted 2 RCTs; however, we are only including the later trial for this review because the data for the first trial was incomplete.²⁷

Table 1:

Characteristics of included studies

Study	Country	Sample size	Main Outcome*	Duration	IMT		COPD diagnosis	Aerobic Exercise
					Type and (intensity)	Frequency and timing		Type and (intensity)	Frequency and timing
Benzo 2011	USA	IMT: 10 Control: 9	Length of stay and postoperative pulmonary complications†	1 week	Threshold inspiratory muscle trainer -P-Flex valve (up to perceived exertion of “Some-what hard”)	15-20 min, twice a day	Not specified	Lower extremity (LE) endurance training -Treadmill or Nu-Step	20min, twice a day
								Strengthening exercises -Thera-band (intensity that is sustainable on light BORG scale)	10-12 repetitions, twice a day
Gao 2015	China	IMT: 71 Control: 71	Length of stay, postoperative complications† and average hospital costs	3-7days	Volumetric exerciser spirometer -Voldyne5000	20 min, 4 time a day	Not specified	Lower extremity endurance training: -Power bicycle (up to BORG score 5-7) -Stair climbing (stop if dyspneic)	15-20 min, twice a day
Huang 2017	China	IMT: 36 Control: 44	Postoperative pulmonary complications†	1 week	Abdominal breathing	20-30 breaths, twice a day	Not specified	Training device -Nust-T5XR, Nustep (resistance increase progressively, stop if dyspneic or exhaustion)	15-20 min, twice a day
					Volumetric exerciser spirometer -Voldyne5000	20 min, 4 times a day
Huang 2017a,b	China	IMT: 30 IMT+ aerobic exercise: 30 Control: 30	Primary: Postoperative pulmonary complications† Secondary: Length of stay and 6MWD distance, PEF, fatigue and dyspnea index, HRQoL score	1 week	Volumetric exerciser spirometer -Voldyne 5000 Abdominal breathing training	20 min, 4 times a day 15-20 min, 2-3 times a day	13-20% of patients	Training device -NuStep	20min, twice a day
Lai 2017	China	IMT: 51 Control: 50	Primary: Postoperative pulmonary complications† Secondary: Hospitalization cost, 6MWD distance, PEF, HRQoL score	1 week	Thoracic expansion and incentive spirometry exercises -HUDSON RCI 2500	20 breaths per session, 2 sessions per day	42-47% of patients	Training device -NuStep (stop if BORG>6, oxygen saturation <88%, pain or other discomfort)	30 min, daily
					Abdominal breathing	15-30 min, twice a day
Lai 2017a	China	IMT: 30 Control: 30	Postoperative pulmonary complications†, HRQoL score, 6MWD distance and PEF	1 week	Abdominal breathing training	15-20 min, twice a day	13-16% of patients	Training device -NuStep device (stop if dyspneic or exhaustion)	30 min, daily
					Volumetric exerciser spirometer -Voldyne5000	20 min, 3 times a day
Mujovic 2015	Serbia	IMT: 56 Control: 47	Changes in lung function and postoperative pulmonary complications†	2-4 weeks	Diaphragmatic breathing	45min sessions, 3 times a day, 5 days a week	All COPD	Not done
					Thoracic cage expansion (first week without load, second week with 1kg load)	10 repetitions, 2 series.
Pehlivan 2011	Turkey	IMT: 30 Control: 30	Primary: Length of stay	1 week	Diaphragmatic, pursed lip, segmental breathing exercise, usage of incentive spirometry, coughing exercise	Twice a day	Not specified	Walking exercise on treadmill (according to patient’s tolerance to speed and time)	3 times a day
Stefanelli 2013	Italy	IMT: 20 Control: 20	Primary: Physical performance after intervention (VO2) before surgery Secondary: Physical performance after intervention (VO2) after surgery	3 weeks	Respiratory exercises on bench, mattress pad and wall bars	15 3 hour sessions, Monday to Friday	All COPD	High intensity training of the upper limbs with rowing ergometer and the lower limbs treadmill and ergometric bicycle (work load at 70% of maximum score of CPET and increase by 10W until patient was able to tolerate the set for 30min).	15 3 hour sessions, Monday to Friday

^*Primary and secondary outcomes are specified if listed in the study

^†Postoperative pulmonary complications are not defined exactly the same between studies. They generally include pneumonia, acute respiratory distress syndrome, atelectasis requiring bronchoscopy, bronchopleural fistula, pleural effusion, venous thromboembolism, air leakage >7days, prolong chest tube>7days, prolong mechanical ventilation >24-48hours, and reoperation.

IMT= inspiratory muscle training.

Participants

The combined sample size from all 10 studies included 768 patients, with each study ranging between 19-142 patients. The mean age was 63.1±7.9 years (range, 54.7-72y). Men made up 60.6% of the total population, with a range between studies between 31.5%-87.4%. Only the Pehlivan 2011 study did not report sex.

Types of interventions

Most studies included breathing exercises using an incentive spirometry^4,6,7,28-32 with or without abdominal breathing. Only 1 study included a true IMT group.²⁷ Almost all of the studies also performed aerobic exercise, ^{4,6,27-29,31,32} except Mujovic 2015 and Huang et al 2017a).^7,30 Huang et al 2017a performed a 3-arm RCT consisting of IMT, IMT combined with aerobic exercise, and a usual care control group.⁷ The IMT and IMT with aerobic exercise results were combined for analysis but separated in the subgroup analysis comparing studies with breathing exercises only vs breathing exercises with aerobic exercise. Although Mujovic 2015 performed pre- and postoperative breathing exercises, postoperative breathing exercises were considered standard of care (incentive spirometry only), which does not affect the validity of its inclusion in the meta-analysis.³⁰ The preoperative intervention training period ranged between 1-4 weeks before surgery.

Outcomes

A reduction in PPCs was the primary outcome for the studies by Gao 2015, Huang 2017, Huang 2017a, Lai 2017, and Lai 2017a.^4,6,7,28,29 LOS and PPCs were the primary outcomes for the studies by Benzo 2011 and Gao 2015.^27,28 LOS was a primary outcome in the study by Pehlivan 2011 and a secondary outcome in the study by Huang 2017a.^7,31 Other secondary outcome measures reported include 6MWD, HRQoL, lung function, and measures of physical performance.

Risk of bias and quality

The risk of bias and quality of the included studies was assessed using the Cochrane risk of bias assessment tool. All studies had a high risk of performance bias because it is difficult to blind participants and personnel for the interventions. Moreover, risk of selection bias was unclear in several studies as random sequence generation was not described,^{6,27,28,30-32} and Mujovic 2015 had a high risk of bias because their patients were consecutively assigned to the intervention and control groups.³⁰ Four of the 10 studies did not specify any blinding of their assessment of outcomes.^29-32 The outcomes were all from RCTs deemed to be of moderate quality owing to suboptimal designs that were lacking allocation concealment and random sequence generation. The risk of bias assessment and grade and quality assessments for each article can be found in supplemental table S2 and S3, respectively (available online only at http://www.archives-pmr.org/).

Hospital LOS

Seven RCTs ^{4,6,7,27,28,30,31} included a measure of hospital LOS. Compared with the control groups, the intervention groups demonstrated a significantly shorter LOS with a pooled mean difference (PMD) of −3.4 days (95% CI, −4.1 to −2.8; P<.01; I²=0) (fig 2).

Fig 2.

Forest plot of LOS.

PPCs

The studies by Benzo 2011, Huang 2017, Lai 2017, Mujovic 2015, and Pehlivan 2011 all quantified PPCs with definitions that included pneumonia, acute respiratory distress syndrome, atelectasis requiring bronchoscopy, bronchopleural fistula, pleural effusion, venous thromboembolism, air leakage>7 days, prolonged chest tube>7 days, prolonged mechanical ventilation>24-48 hours, and reoperation.^4,27,29-31 From these studies, the intervention groups were less likely to experience PPCs with an odds ratio (OR) of 0.37 (95% CI, 0.21-0.65; P<.01; I²=0%; 5 RCTs) (fig 3). Pooled analysis of the 4 RCTs^4,7,29,33 that used Clavien-Dindo complication classification also demonstrated a significant difference (score, >2; OR, 0.22; 95% CI, 0.10-0.46; P<.01; I²=0%).

Fig 3.

Forest plot of PPC.

Postoperative pneumonia

There were 7 studies that included postoperative pneumonia as an outcome.^{4,7,27-29,31,33} Pooled rates of pneumonia demonstrate a significant improvement in the intervention groups with an OR of 0.37 (95% CI, 0.19-0.73; P<.01; I²=0%; 7 RCTs) (fig 4).

Fig 4.

Forest plot of postoperative pneumonia complication.

Mortality

Five studies included mortality as an outcome.^4,7,29,30,33 From these studies, the mortality OR was 0.65 (95% CI, 0.09-4.78; P=.7 and was not significantly different between the intervention and control groups (fig 5).

Fig 5.

Forest plot of mortality rate.

HRQOL

Of the 3 studies that evaluated quality of life,^4,6,7 all used the European Organisation for Research and Treatment of Cance Core Quality of Life questionnaire-30 as their HRQOL measure. The overall changes in HRQOL and the subcomponents were not significantly different as the global score PMD was 2.0 (95% CI, −2.7 to 6.7; P=.4), the physical function PMD was 0.99 (95% CI, −1.61 to 3.60; P=.4), the emotional function PMD was −0.38 (95% CI, −3.83 to 3.07; P=.8), and the dyspnea subcomponent was −4.47 (95% CI, −10.73 to 1.80; P=.2) (fig 6).

Fig 6.

Forest plot of HRQOL EORTC global score. Abbreviation: EORTC, European Organisation for Research and Treatment of Cancer.

6MWD

Studies by Huang 2017a, Lai 2017, Lai 2017a, and Mujovic 2015 included the 6MWD before and after the intervention.^4,6,7,30 The distance covered improved in the intervention groups compared with the control groups because the PMD was 20.2 m (95% CI, 9.1-31.2; P<.01) (fig 7). Although this is statistically significant, it is still below the MCID of 26 m for clinical significance.

Fig 7.

Forest plot of quality of 6MWD

Pulmonary function

The PMDs for peak expiratory flow rate,^4,6,7,31 FEV₁,^6,7,30-32 and DLCO^6,7,32 were 17.5 L/min (95% CI, −22.7 to 57.6; P=.39), 0.14 L (95% CI, −0.09 to 0.36; P=.23), and 0.02 mL/min/mmHg (95% CI, −1.76 to 1.79; P=.99), respectively. However, there were no significant differences. Comparison of forced vital capacity (FVC) was not performed because there were only 2 studies providing data on FVC.

Subgroup analysis

Two different subgroup analyses were performed. The first assessed the effects of breathing exercises alone compared with breathing exercise combined with aerobic exercise. The second analysis examined the confounding effects of a COPD diagnoses, which is described in more detail in the supplemental appendix, (available online only at http:www.archives-pmr.org/).

The first analysis compared the differences in outcomes in patients who performed breathing exercises alone^7,30 with those who performed breathing exercises in combination with aerobic exercise.^{4,6,7,27,28,30-33} The 3-arm study by Huang at al 2017a included a breathing exercise arm as well as a breathing exercise combined with aerobic exercise subgroup.⁷ Overall, patients performing breathing exercises plus aerobic exercise had a greater reduction in LOS as demonstrated by a PMD of −3.65 days (95% CI, −4.37 to −2.93; P<.0001) compared with −1.7 days (95% CI, −3.25 to −0.11; P=.04) for the breathing exercise-only arm. The differences between these 2 groups were measured by chi-square test (χ², 4.85; P=.03) (fig 8). However, the differences in the number of studies included and patients in each group were quite different for this analysis, as there were only 163 patients from 2 studies in the breathing group compared with 440 patients from 6 studies in the breathing combined with aerobic exercise group.

Fig 8.

Forest plot of LOS with IMT vs IMT+aerobic exercise subgroup.

We also compared differences in FEV₁, 6MWD, and mortality from the breathing-only group to the breathing combined with aerobic exercise subgroups. For the 6MWD, the breathing-only subgroup^7,30 had a PMD of 25.8 (95% CI, −14.83 to 66.41; P=.21; I²=1%) and the breathing combined with aerobic exercise group^4,6,7 had a PMD of 19.4 (95% CI, 8.1-30.8; P<.01; I²=0%). There was no subgroup difference with chi-square of 0.09 (P=.77; I²=0%) (fig 9).

Fig 9.

Forest plot of 6MWD with IMT vs IMT+aerobic exercise subgroup.

For FEV₁, the breathing exercises-only group^7,30 demonstrated a PMD of 0.15 (95% CI, −0.33 to 0.63; P=.54; I²=46%) compared with the breathing combined with aerobic exercise group,^6,7,32 which had a PMD of 0.1 (95% CI, −0.17 to 0.37; P=.46; I²=0%). The subgroup differences for FEV₁ was chi-square of 0.03 (P=.86; I²= 0%).

Mortality did not reach significance for either of the subgroups. The breathing exercises-only group^7,30 had an OR of 1.25 (95% CI, 0.10-16.11; P=.86; I²= 24%) compared with the breathing combined with aerobic exercise subgroup,^6,29,33 which had an OR of 0.32 (95% CI, 0.01-8.24; P=.49). Heterogeneity was not applicable in the studies by Huang 2017 and Liu 2020 because there were no deaths reported in either group.^29,33 The differences between the 2 groups were measured by chi-square test (χ², 0.42; P=.53; I²= 0%).

Sensitivity analysis

A sensitivity analysis was conducted to evaluate the stability of the meta-analysis. The heterogeneity I² was very low in all outcomes except FEV₁, with an I² of 66%. The heterogeneity in the LOS with subgroup analysis comparing breathing only vs breathing and aerobic exercise subgroups had an I² of 76%. We tried eliminating 1 of the other studies at a time; however, this led to an increase in the heterogeneity between the 2 subgroups. For FEV₁, the heterogeneity outcome in the breathing exercise-only subgroup was 46%, but we could not exclude any studies because this group included only 2 studies. The outcome was not significant for FEV₁ in our analysis, regardless of the heterogeneity. The following outcomes had an I² of 0%: 6MWD, peak expiratory flow rate, FVC, DLCO, LOS, PPC, pneumonia, and Clavien-Dindo score >2. Mortality outcome had an I² of 17%. There were only 2 studies^30,32 that provided postsurgical physiological outcomes (FEV₁, DLCO, 6MWD). Hence, all physiological outcomes were analyzed as changes pre- and postintervention before surgery.

Discussion

Preliminary work incorporating respiratory muscle training or breathing exercises before surgery has shown considerable promise to reduce LOS and PPC in patients undergoing surgery for lung cancer as well as other noncancer surgeries.^16,34 Our meta-analysis is one of the first reviews to include RCTs examining breathing exercises compared with a nontraining control group as a preoperative strategy to reduce postoperative complications after lung resection surgery. The current meta-analysis demonstrated that preoperative breathing exercises, compared with a nontraining control group, reduced LOS, PPC, and the rate of pneumonia and increased the distance covered during the 6-minute walk test. Although the 6MWD was statistically significant, it was below the MCID, with no effect on lung function, HRQOL, and mortality. Importantly, combining breathing exercises with aerobic exercise offers additional advantages, whereas a diagnosis of lung disease (COPD) did not influence the outcomes (see supplemental appendix, available online only at http://www.archives-pmr.org/).

Three other systematic reviews also investigated the effects of breathing exercises on postoperative outcomes in patients undergoing resection for lung cancer.^16,34,35 Our results are consistent with these reviews by demonstrating a decrease in postoperative LOS and PPC. However, the current review is the first to demonstrate a reduction in postoperative pneumonia.^16,18 The review by Cavalheri et al³⁴ was one of the first to describe a benefit of breathing exercises in the preoperative phase. They demonstrated a reduction in PPC with an OR of 0.33 (95% CI, 0.17-0.61; P>.01) when breathing exercises were performed before surgery. This Cochrane review³⁴ included 5 studies performed in the preoperative phase^{6,27,31,32,36,37} that were also included in the current analysis (excluding Morano 2013 and Morano 2014 because it lacked a true control group^36,37). We also included 7 additional RCTs not included in their work. The Wang et al meta-analysis in 2019 also focused on studies incorporating breathing exercises to reduce postoperative complications for surgically eligible lung cancer patients.¹⁶ In contrast to our study, this review included non-randomized studies during the preoperative, perioperative and postoperative phases of treatment. Even with more study heterogeneity, their work continued to demonstrate that breathing exercises before, during, and after lung resection surgery will reduce PPCs with an OR of 0.32 (95% CI, 0.21-0.49; P<.01).¹⁶ We also included many of the same studies as the Wang¹⁹ review, except for Morano 2014³⁷ and Sebio Garcia 2017³⁸ because they did not include similar outcome measures and Wang 2010 because it was only published in Chinese. However, we included studies by Gao 2015, Huang 2017, Lai 2017, Liu 2020, and Mujovic 2015 that were not included in these prior reviews. We suggest that the additional studies as well as our decision to include only RCTs adds power to the current meta-analysis and may also explain why we were the first to see a reduction in postoperative pneumonia. The third meta-analysis by Rosero et al³⁵ focused primarily on aerobic exercise before lung cancer resection surgery. This meta-analysis included 10 studies total, 6 of which incorporated aerobic exercise concurrently with IMT that were also included in our meta-analysis^{4,6,7,27,32,36} and 4 additional studies not included in our meta-analysis because they did not perform IMT, IMT was performed in both trial arms, or they were a non-English publication, Findings from Rosero³⁵ also demonstrate improvements in LOS, 6WMD, and postoperative pulmonary complications. Subgroup analysis also demonstrated no differences in 6MWD or HRQOL when IMT was performed concurrent with aerobic exercise and strength training.

The previous literature is conflicting regarding the ability of breathing exercise to increase the 6MWD in the preoperative setting. Cavalheri et al³⁴ demonstrated a benefit of preoperative breathing exercises on 6MWD with a PMD of 18.2 m (95% CI, 8.50-27.96). However, their analysis was based on 2 studies^6,36 with considerable variability in the distance covered during the 6MWD (SD, 18-27 m vs 118-184 m). In addition, if the Lai 2017a paper⁶ (SD, 118-184 m) was excluded from the analysis, we suggest the benefits would have been even less. In comparison, Wang et al¹⁶ did not show a significant benefit to the 6WMD, which could be related to the limited number of studies, the different types of interventions, or the timing of the interventions in their review. Consistent with Cavalheri,³⁴ our metanalysis demonstrated a significant PMD of +20.2 m, slightly below the MCID of 26 m in the breathing exercise group, suggesting that breathing exercises can improve preoperative function. Additional work is needed to investigate whether the benefits are real and, more importantly, if they affect postoperative function or outcomes.

We were unable to determine whether IMT could reduce the decline in lung function commonly seen after surgery because most studies only report outcomes (FEV₁, FVC, 6MWD) after the intervention and before surgery. This is a novel question and one that remains to be determined, especially because most prior IMT studies done for other reasons had longer training durations.³⁹ Importantly, 1-4 weeks of preoperative IMT has a positive effect of reducing PPCs. From the current meta-analysis, only Mujovic 2015³⁰ and Stefanelli 2013³² collected outcomes after surgery. These findings demonstrate that 1-4 weeks of preoperative IMT could minimize the decline in FEV₁ after surgery, even when time was limited before surgery. Consistently, 2 weeks of IMT improved oxygenation and physical activity after lung cancer resection surgery,⁴⁰ and only 12 sessions of IMT and expiratory muscle training (over 4wk) were needed to increase respiratory muscle strength 20%-30% and double submaximal exercise capacity in survivors of lung and breast cancer.⁴¹ Finally, 5 weeks of IMT altered muscle physiology in patients with COPD.⁴² Nonetheless, longer duration of IMT may be more beneficial and should be explored. However, lung surgery should not be delayed, and the current evidence supports IMT as an alternative to aerobic exercise, especially when time is short. Ideally future RCTs should perform physiological measurements before intervention and after intervention before and after surgery. 6MWD, respiratory muscle strength, pulmonary function, and possibly peak oxygen uptake during a maximal cardiopulmonary exercise test should be evaluated in the postoperative period. Mujovic 2015 and Stefanelli 2013 were the only studies that collected some of these data at these 3 time points.^30,32

Although preoperative breathing exercises have been shown to reduce LOS, PPC, and pneumonia, there is considerable heterogeneity in the types of breathing interventions used, which may lessen the overall effects of any meta-analysis. Many of the studies performing breathing exercises included incentive spirometers, with or without abdominal breathing exercise, as well as aerobic exercise. Only 1 study incorporated a threshold training device to specifically increase respiratory muscle strength. Threshold devices are known to be superior to incentive spirometers to improve inspiratory muscle strength,⁴³ whereas incentive spirometers are generally used to gain pulmonary volume as well as respiratory muscle endurance and do not represent the ideal strategy to improve respiratory muscle strength. Leith et al⁴⁴ was the first to demonstrate that resistance respiratory muscle training improves respiratory muscle strength, whereas respiratory muscle endurance training (eg, spirometry) is better at increasing respiratory muscle endurance, suggesting that the intervention chosen should be based on what mechanism is being hypothesized as the reason why breathing exercises reduce postoperative complications. Furthermore, heterogeneity associated with the different intervention strategies and study design makes it more challenging to distinguish how belly breathing may compare with resistance or endurance respiratory muscle training programs or spirometry. Because Benzo 2011²⁷ was the only intervention from our meta-analysis to include a threshold device it is impossible to complete subgroup comparisons based on training devices or strategies. Unfortunately, we excluded others³⁶ that incorporated threshold training before surgery because their control also performed the same breathing exercises as our defined treatment arm. Although this study demonstrated a reduction in PPC and LOS, it combined respiratory muscle training with strength and endurance training over a 1- to 4-week period before surgery, also making it difficult to determine which component of the program was responsible for the postoperative improvements. As a result, all breathing exercises appear to be beneficial, and we suggest that future analyses are needed to distinguish the benefits of each type of training to define the most efficient way to reduce postoperative complications. In particular, more studies on breathing exercises or IMT alone, without concurrent aerobic or resistance training, are urgently needed to better delineate the effects of this training on postoperative outcomes.

Our results are consistent with others describing added benefits when breathing exercises were combined with aerobic exercise to reduce LOS and PPC.^16,34 Although the addition of aerobic exercise added additional benefits, it is difficult to say for certain why there is a difference given the disproportionate number of studies included in each subgroup. Most of the studies on breathing plus aerobic exercises had an intervention duration of 1-4 weeks, which is considerably less than optimal if improvement in aerobic capacity is the desired outcome. Thus, the mechanism by which aerobic exercise improves postoperative outcome and the degree to which it is responsible for the improvements we observed requires further study.

Study limitations

The biggest limitation of this study is the small number of studies that incorporate vastly different intervention strategies with inadequate follow-up. Most studies compared outcomes pre- and postintervention (before surgery) but neglected to report how improvements in function may affect postoperative outcomes. Very few studies that performed IMT did so without some form of aerobic or resistance training. Despite this, the 2 studies that were IMT-only showed improvement in LOS. Ideally, we will need more IMT-only studies to better delineate the benefits of this training modality. Moreover, the lack of HRQOL outcomes after surgery limits the detection of any additional benefits from preoperative breathing exercises, especially because HRQOL initially declines after surgery. Descriptions of the specific RCT protocols were also missing, which leads to unclear assessments and increased risk for bias in many studies.

Conclusions

In conclusion, our meta-analysis demonstrated that preoperative breathing exercises in conjunction with aerobic exercise will reduce LOS, PPCs, and pneumonia. However, additional work is needed to determine the ideal type and timing of the intervention. Preoperative breathing exercises alone reduce LOS, but the overall effects of this modality used without concomitant aerobic exercise training require further study. Preoperative breathing exercises or aerobic exercise should be implemented for patients undergoing resection for lung cancer.

Supplier

a. RevMan, version 5.3; Cochrane Training.

List of abbreviations:

CI

confidence interval

COPD

chronic obstructive lung disease

DLCO

diffusing capacity for carbon monoxide

FEV₁

forced expiratory volume in 1 second

FVC

forced vital capacity

HRQOL

health-related quality of life

IMT

inspiratory muscle training

LOS

length of stay

MCID

minimal clinically important difference

OR

odds ratio

PMD

pooled mean difference

PPC

postoperative pulmonary complication

RCT

randomized controlled trial

6MWD

6-minute walk distance