A comprehensive search was conducted of multiple databases from inception to December 2021 for studies limited to the English language and excluding animal studies. The databases included Ovid MEDLINE(R) and Epub Ahead of Print; In-Process & Other Non-Indexed Citations and Daily; Ovid Embase; Ovid Cochrane Central Register of Controlled Trials; Ovid Cochrane Database of Systematic Reviews; and Scopus.

The search strategy was designed and conducted by an experienced librarian with input from the study’s principal investigator. Controlled vocabulary supplemented with keywords was used to search for studies describing the effects of IMT in CABG. The actual strategy listing all search terms used and how they were combined is available in Supplementary Table 1 online.

Studies were included if they met the following criteria: randomized controlled trials published in English regardless of the publication date with populations of all ages and either sex with a history of CABG and receiving center-based, home-based, or community-based IMT. In the analyses, IMT before CABG, after CABG, and both before and after CABG was compared with usual care and sham IMT.

Nonhuman studies, nonrandomized controlled trials, abstracts, books, dissertations, study protocols, and case reports were excluded. Subjects who were scheduled for CABG or who only stated that they were undergoing cardiac surgery were excluded. IMT combined with other exercise interventions such as aerobic/resistance training or respiratory muscle strength training was also excluded.

The main outcomes of interest were pulmonary function, respiratory muscle strength and endurance, exercise capacity, PPCs, and LOS.

An experienced librarian searched for articles relevant to this review, and then two authors independently reviewed all titles and abstracts. The reference lists were also screened for other potentially eligible studies. These two authors assessed the full texts of all selected articles to verify if they met the inclusion criteria for the review. Disagreements between them were resolved by another senior author until the final decision was reached by consensus.

Data extraction was undertaken using a standard form adapted from the Cochrane Collaboration by two authors [15]. The form included the basic characteristics of each study: publication date, country, sample size, average age, sex, intervention period (preoperative, postoperative, and preoperative and postoperative), outcomes, and intervention characteristics (device, intensity, training time, frequency and duration, supervision).

The quality assessment of all eligible articles was conducted by two independent authors using the physiotherapy evidence database (PEDro) scale, which is a typical tool for evaluating the quality of physical therapy and rehabilitation studies [16]. The PEDro scale consists of 11 items based on a 9-item Delphi list [17]. The score range is 0 to 10, as 1 item on the PEDro scale (eligibility criteria) is related to external validity and is not used to calculate the score [18]. A PEDro score of 6–10 indicates high quality, a score of 4–5 indicates fair quality, and a score of 4 indicates poor quality [19].

Review Manager 5.1 software (RevMan 2011, Cochrane Collaboration, Copenhagen, Denmark) was used for the data analysis [20]. Pooled-effect estimates were obtained by comparing the least square mean percentage change from baseline to study end for each group. The effect size was expressed as the mean difference (MD) with a 95% confidence interval (CI), due to the diversity of methodologies. An MD higher than 0.5 indicated a moderate effect size, while an MD higher than 0.8 indicated a large effect size [21]. We used either a fixed-effect model or a random-effect model depending on the heterogeneity. Four comparisons were made: combined exercise with IMT versus the exercise group, preoperative IMT versus the control group, postoperative IMT versus the control group, and both preoperative and postoperative IMT versus the control group. A p-value of $\le$0.10 was regarded as significant. Cochran’s Q-test and the inconsistency statistic I${}^2$ were used to assess heterogeneity. I${}^2$ greater than 50% was considered indicative of high heterogeneity [22].

We identified 351 articles from the database without additional resources. After the removal of duplicates, 339 articles were screened out. The remaining 26 full texts were retrieved and assessed for potential eligibility; of these, 12 RCTs were included in the meta-analysis [23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34]. Four studies performed preoperative IMT [23, 24, 25, 34], five were in the postoperative period [26, 27, 28, 30, 31], and three compared preoperative and postoperative IMT with a control group [29, 32, 33]. The specific study selection flow chart is shown in Fig. 1.

Fig. 1.

PRISMA flow chart.

The included studies provided data on 918 patients, with sample sizes ranging from 20 to 279. According to the IMT period, 624 patients were allocated to preoperative IMT, 185 to postoperative IMT, and 109 to preoperative and postoperative IMT (Table 1).

Table 2 presents the characteristics of the interventions included. For the IMT, most of the studies used a threshold IMT device except for two: one of these used an expiratory positive airway pressure (EPAP) mask [26], and the other used a DHD respiratory trainer [27]. Patients started the training with a load from 15% to 40% of maximal inspiratory pressure (MIP). IMT frequency varied between studies from twice daily to 7 days per week. The duration of training also varied from 15 minutes to 30 minutes, and the total training length varied depending on whether IMT was administered in the preoperative or the postoperative period. It is common to provide supervision during the process of training.

We used the PEDro scale to assess the study methodologies, and the total scores varied from 4 to 8 points. Five studies were of high quality (PEDro score higher than 5) [24, 25, 26, 29, 32], while seven were of moderate quality (5 or 4) [23, 27, 28, 30, 31, 33, 34]. Most studies did not report a concealed allocation process (N = 9) [24, 27, 28, 30, 31, 32, 33, 34], and none of the included studies reported the blinding of subjects. The PEDro scores for the included studies are presented in Table 3.

For patients who underwent CABG, decreased inspiratory muscle strength represented as lower MIP and PImax was common and contributed to the incidence of PPCs after CABG [36, 37]. Respiratory muscles, similar to skeletal muscles, can be trained for strength and endurance [38, 39, 40]. In our findings regarding preoperative IMT, all included studies reported improved inspiratory strength or endurance expressed as Pmpeak/PImax. There were also improvements in MIP from pre- and postoperative IMT, although the heterogeneity was high (I${}^2$ = 66%) due to the different protocols adapted in the intervention groups [29, 32, 33].

For postoperative IMT studies [26, 27, 28, 30, 31], benefits were shown for both MIP and MEP. Although IMT did not focus on expiratory muscles, postoperative training still affected MEP; the explanation for this was that strong inspiratory muscles were helpful in the elastic recoil of the lungs and chest wall by expanding the position of the thorax. Improved expiratory muscle strength benefits cough strength and effective airway clearance, which therefore delays the occurrence of PPCs [7].

CABG causes diaphragmatic dysfunction that leads to restrictive lung volumes, impaired ventilatory mechanics, and decreased lung compliance [12].

In our meta-analysis, only two studies investigated FEV1 and FVC, and both reported positive effects [23, 25]. After CABG, the pulmonary function decreased within two weeks. Previous studies demonstrated that pulmonary function decreased by 40% after CABG accompanied by a 25% decrease in PaO${}_2$ [41, 42]. The results in our study agree with the effects of IMT on pulmonary function in other populations: FVC and FEV1 are not directly affected by expiratory or inspiratory muscle strength but rather by lung size and expiratory airflow, which helps explain this study’s outcomes.

The analysis of exercise capacity included studies measuring 6MWT, and these showed benefits from postoperative IMT. IMT is established as an effective intervention that attenuates inspiratory muscle weakness and has been applied in the context of cardiovascular rehabilitation for heart failure [43]. Sinoway et al. [44] showed a minimum 6MWT increase of 25 m in patients with coronary artery disease who underwent IMT. Our study results show a 6MWT increase of 86.09 m, positive evidence that postoperative IMT intervention could improve exercise capacity [26, 31]. These results can be explained by the reduction in diaphragmatic overload associated with subsequently increased oxygen delivery to the limbs, which leads to increased exercise capacity [45, 46]. Inspiratory muscle weakness leads to a decrease in lung volume that contributes to exercise intolerance. Thus, improvements in inspiratory muscle strength and pulmonary function after IMT could also help increase exercise capacity.

It is known that prolonged hospital stays are associated with PPCs, morbidity, and mortality [47, 48]. These complications in turn increase the rate of hospital readmission, creating a vicious circle. Pooled data from the included studies show that inspiratory muscle training during the preoperative period reduced LOS (MD: 1.96 days) and PPCs (RR: 0.42) [23, 24, 25, 34]. The same effect was also observed during the postoperative period (MD: 1.55 days) [28, 30, 31]. For most of the related studies, IMT could effectively improve inspiratory muscle strength and pulmonary function, further helping to reduce the incidence of PPCs. Our findings are in accordance with previous meta-analysis findings showing that IMT can prevent PPCs such as atelectasis and pneumonia [12, 35].

Some outcomes help to explain how IMT reduces the risk of LOS and PPCs. The improvement in inspiratory muscle strength and pulmonary function can help with recovery from surgery, contributing to reducing the length of the hospital stay. In addition, forceful respiratory muscles assist with lung expansion and increase vital capacity and tidal volume, which, in turn, reduce the risk of pulmonary complications. The use of IMT for patients who undergo CABG can reduce the length of the hospital stay and pulmonary complications, in turn reducing the burden on the whole medical care system.