Background: Currently, consensus on the effectiveness of incentive spirometry (IS) following cardiac, thoracic, and upper abdominal surgery has been based on randomized controlled trials (RCTs) and systematic reviews of lower methodological quality. To improve the quality of the research and to account for the effects of IS following thoracic surgery, in addition to cardiac and upper abdominal surgery, we performed a meta-analysis with thorough application of the Grading of Recommendations Assessment, Development and Evaluation scoring system and extensive reference to the Cochrane Handbook for Systematic Reviews of Interventions. Objective: The objective of this study was to determine, with rigorous methodology, whether IS for adult patients (18 years of age or older) undergoing cardiac, thoracic, or upper abdominal surgery significantly reduces30-day post-operative pulmonary complications (PPCs), 30-day mortality, and length of hospital stay (LHS) when compared to other rehabilitation strategies. Methods: The literature was searched using Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, CINAHL, and Web of Science for RCTs between the databases’ inception and March 2019. A random-effect model was selected to calculate risk ratios (RRs) with 95% confidence intervals (CIs). Results: Thirty-one RCTs involving 3,776 adults undergoing cardiac, thoracic, or upper abdominal surgery were included. By comparing the use of IS to other chest rehabilitation strategies, we found that IS alone did not significantly reduce 30-day PPCs (RR = 1.00, 95% CI: 0.88–1.13) or 30-day mortality (RR = 0.73, 95% CI: 0.42–1.25). Likewise, there was no difference in LHS (mean difference = −0.17,95% CI: −0.65 to 0.30) between IS and the other rehabilitation strategies. None of the included trials significantly impacted the sensitivity analysis and publication bias was not detected. Conclusions: This meta-analysis showed that IS alone likely results in little to no reduction in the number of adult patients with PPCs, in mortality, or in the LHS, following cardiac, thoracic, and upper abdominal surgery.

Over 230 million major operations occur annually, citing cardiac, thoracic, and upper abdominal surgeries as the most frequent major surgical procedures performed globally [1]. Following cardiac, thoracic, and upper abdominal surgeries, patients are at risk of experiencing adverse events known as post-operative pulmonary complications (PPCs) [2, 3]. These complications are broadly defined as respiratory tract infections, pneumonia, atelectasis, pleural effusion, pneumothorax, bronchospasm, aspiration pneumonitis, and respiratory failure requiring invasive or non-invasive mechanical ventilation [3]. Such PPCs are the leading cause of mortality following the above-mentioned surgeries [4], with incidence ranging from 17 to 88% [5, 6]. The wide variation in incidence of PPCs may be attributed to the lack of consensus to which conditions constitute PPCs [7], in addition to the variability in surgeon-level technical skills [8].

Anaesthesia during surgery can have profound effects on the respiratory system that increase the risk of developing PPCs [9]. A notable effect is the significant decrease in functional residual capacity (FRC) caused by the increased abdominal pressure from patients lying supine and the altered function of respiratory muscles [10, 11]. Moreover, post-operative pain leads to shallow breathing (splinting), which further reduces FRC and inspiratory volumes, causing atelectasis and pneumonia. Impaired movement of the diaphragm and chest wall in addition to reduced pulmonary compliance during anaesthesia also hinders inspiratory volumes. Many physiotherapy strategies are designed to prevent PPCs by enhancing inspiration through increases in post-operative FRC. Bartlett and colleagues developed an alternative respiratory intervention known as incentive spirometry (IS), which targeted (maximal) inspiratory lung volumes rather than resting lung volumes through the recording and visual feedback of successful breathing manoeuvres [12].

Best practice guidelines have recommended that early mobilization and respiratory interventions should be used to minimize risk of PPCs [13]. However, guidelines released by the American Association for Respiratory Care (AARC) specifically recommended against the use of post-operative IS alone following upper abdominal or cardiac surgery [14]. Although important, these guidelines were based on systematic reviews lacking rigorous methodologies; the guidelines did not account for thoracic surgery, which involves a highly susceptible population to PPCs, and they failed to address some patient-important outcomes, such as length of hospital stay (LHS). Specifically, there has been 4 systematic reviews [15‒18] that have assessed the effects of post-operative IS in reducing PPCs. The review performed by do Nascimento Junior et al. [17] was an update of the original systematic review by Guimarães et al. [18]; therefore, for the remainder of this systematic review, references to “do Nascimento Junior et al. [17]” will also account for Guimarães et al. [18]. Both Overend et al. [15] and do Nascimento Junior et al. [17] restricted their reviews to only include cardiac and upper abdominal surgery. Carvalho et al. [16] did include thoracic surgery; however, they only found 3 studies relating to thoracic surgery at the time. These systematic reviews, though informative in their time, are outdated and possess methodological limitations. For these reasons, it is timely to review the preventive effects of IS following cardiac, thoracic, and upper abdominal surgery. As such, we aimed to contribute to the AARC guidelines by conducting an updated critical analysis of the available literature, which addressed the limitations from previous systematic reviews and accurately followed the Cochrane Handbook for Systematic Reviews of Interventions [19]. The objective of this systematic review was to determine the effects of post-operative IS in comparison to either no rehabilitation or various other rehabilitation strategies in reducing 30-day PPCs, 30-day mortality, and LHS for adult patient undergoing thoracic, cardiac, or upper abdominal surgery.

Our systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [20] with a corresponding flow diagram and checklist found in Figure 1; online supplementary Appendix 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000517012), respectively. Moreover, our systematic review was registered in PROSPERO under ID: CRD42019137540.

Fig. 1.

Study flow diagram. As per PRISMA guidelines. ERAS, enhanced recovery after surgery; ICTRP, International Clinical Trials Registry Platform; RCT, randomized controlled trial.

Fig. 1.

Study flow diagram. As per PRISMA guidelines. ERAS, enhanced recovery after surgery; ICTRP, International Clinical Trials Registry Platform; RCT, randomized controlled trial.

Close modal

Available evidence was assessed from randomized controlled trials (RCTs) that evaluated the post-surgical introduction of IS in patients aged 18 years or older who underwent thoracic, cardiac, or upper abdominal (performed through an incision above or extending above the umbilicus) [21] surgery. There were no exclusion criteria for surgical approach. Therefore, our systematic review included studies that utilized minimally invasive and/or open techniques during surgery. Included studies involved the comparison of IS to either no rehabilitation or various other rehabilitation strategies, such as positive-pressure breathing techniques (i.e., bi-level positive airway pressure, continuous positive airway pressure, and intermittent positive pressure breathing), deep breathing exercises, and other chest physiotherapies. Patients were required to start their rehabilitation therapies within the first week following surgery. Trials where participants were a part of an enhanced recovery after surgery (ERAS) program that involved IS as a component of a multimodal approach or trials in which IS was combined with one of the rehabilitation strategies stated above (without an active comparator group of the other rehabilitation strategy on its own) were excluded. The rationale for exclusion was the high risk of co-intervention with ERAS programs and combined therapies. Results would not be attributable to IS alone; therefore, we would not be able to contribute to the AARC guidelines for IS as an independent rehabilitation therapy. Trials in which participants engaged in pulmonary exercises prior to surgery were also excluded.

Three outcomes of interest were sought among the trials which are as follows:

  • Total number of patients with a PPC (within 30 days of surgery): as defined by do Nascimento Junior et al. [17], which includes atelectasis as diagnosed radiographically, tomographically, or bronchoscopically; respiratory failure requiring invasive or non-invasive mechanical ventilation; and tracheobronchial infection or pneumonia

  • 30-day mortality: measured as any cause of death between the day of surgery and 30 days afterwards

  • LHS (days): measured from the day of surgery to hospital discharge

We did not assess compliance with IS as an outcome because a recent systematic review dedicated its entire purpose to analyse the level of compliance for IS following cardiac, thoracic, and abdominal surgery [22].

We searched the Cochrane Central Register of Controlled Trials (CENTRAL; inception to The Cochrane Library 2019, Issue 3), Ovid MEDLINE® and Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily (1946 to March 11, 2019), EMBASE (1974 to 2019 Week 11), CINAHL (1981 to March 2019), and Web of Science (all databases; 1926 to March 2019). Databases were not limited to any publication date. There were also no restrictions placed on language or publication status. Current ongoing trials were identified using ClinicalTrials.gov (until March 24, 2019) and the International Clinical Trials Registry Platform through the World Health Organization registry (until March 24, 2019). We reviewed reference lists of relevant journal articles, contacted trialists for additional or preliminary data, and searched the grey literature through OpenGrey (until March 24, 2019). A detailed description of the search strategies can be found in online supplementary Appendix 2.

Two review authors (K.S. and I.C.) independently screened titles and abstracts of the identified trials for possible inclusion. Full-text screening and data abstraction were also conducted independently by each reviewer (K.S. and I.C.). Data were extracted for study characteristics, participants, intervention, comparators, follow-up, and the outcomes listed above. Risk of bias was independently assessed using the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions with assessments made on an outcome-to-outcome basis [19]. Disagreements during full-text screening, data abstraction, and risk of bias were resolved through discussion, and when agreement could not be achieved, a third reviewer (D.H.) resolved any conflicts.

During risk of bias assessment, domains were rated as either “high risk,” “low risk,” or “unclear risk” based on Section 8.5d of the Cochrane Handbook for Systematic Reviews of Interventions (shown in online suppl. Appendix 3) [19]. To measure the treatment effect of our outcomes, we used the generic inverse variance method. Dichotomous outcomes were pooled as risk ratios (RRs) with a 95% confidence interval (CI). The unit of analysis was the number of patients with the outcomes. For LHS (i.e., a continuous outcome), we presented the results as a mean difference (MD) with a 95% CI. Trials reporting results as medians and interquartile ranges were converted to means and standard deviations by applying estimation theories [23‒25]. Studies that contributed multiple comparator groups had the groups combined using a pooled weighted mean to create a single pair-wise comparison to the IS group.

When the included trials were missing data, we contacted the authors. If the authors were unreachable, we assessed the reasoning for the missing data and lost to follow-up. Provided reasonings were similar among intervention groups, we considered <20% of the total number of patients’ data missing as a complete case since this would still prevent Type I error rates [26]. If >20% of data were missing, we considered this to be a high risk of bias.

Heterogeneity between pooled trials was assessed using a visual inspection of the forest plots, the χ2 test (p < 0.10), and the I2 statistic. The I2 value was evaluated based on the guide proposed by Higgins and Green [19]. Publication bias was assessed per outcome using funnel plots.

Results of the trials were pooled using Review Manager 5.3 (RevMan 5.3) [27]. We applied the random-effect model as heterogeneity was expected in the studies due to the different surgery types. A subgroup analysis of the studies was performed based on these different surgery types. However, an additional subgroup analysis for different rehabilitation strategies was not included within this review because if studies were multi-armed (≥3 arms), they would cause a redundancy in assessing the IS arm to different comparator groups. For outcomes that had an I2 > 29%, we conducted a sensitivity analysis to explore the cause of heterogeneity and the robustness of the pooled results by removing trials with high risk of bias in at least one domain, specifically randomization and allocation concealment.

The quality of evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach, where we determined the level of quality (high, moderate, low, or very low) for each outcome. We downgraded the evidence from “high quality” by one level for serious problems or by 2 for very serious problems when appraising the GRADE domains [28]. GRADE profiler (GRADEPRO) [29] was used to import data from RevMan 5.3 to create the Summary of Findings table.

We screened a total of 6,013 records. The extent of agreement between the 2 independent assessors (K.S. and I.C.) during title and abstract screening was 98.98% (κ = 0.9). Of these records, 31 RCTs were included for qualitative synthesis [30‒60] and 28 were combined in a meta-analysis [30‒40, 42‒52, 54, 56‒60]. Sixty-three studies were excluded during full-text screening, while 12 studies were awaiting classification, with 2 considered ongoing studies and the remaining 10 pending full-report access and/or production (shown in Fig. 1). As shown in Table 1, the 31 included RCTs involved n = 3,776 adults undergoing cardiac (n = 9), thoracic (n = 6), or upper abdominal surgeries (n = 16). From these RCTs, IS was mostly compared to other chest physiotherapies (n = 13), while 2 two-armed studies specifically compared IS to no intervention. Most RCTs were conducted in Europe (n = 13) or North America (n = 9). For the meta-analysis, 26 studies analyzed the impact of IS on PPCs (shown in Fig. 2a) [30‒37, 39, 40, 42‒52, 54, 56‒58, 60], 9 studies investigated the effects of IS on mortality (shown in Fig. 2b) [30, 32, 39, 40, 42, 46, 48, 59, 60], and 15 studies examined the use of IS in decreasing LHS (shown in Fig. 2c) [30, 32, 33, 35, 37‒40, 46, 48, 51, 54, 59, 60].

Table 1.

Demographics of included studies

 Demographics of included studies
 Demographics of included studies
Fig. 2.

Forest plot of IS compared to other rehabilitation strategies (i.e., the comparator) in reducing (a) post-operative pulmonary complications, (b) mortality, and (c) LHS. Studies were subcategorized into surgery type. RR and MD were the units of measurement for dichotomous and continuous outcomes, respectively. A random-effect model was used. Notable heterogeneity was present if I2 > 40% or p < 0.10. CI, confidence interval; IS, incentive spirometry; LHS, length of hospital stay; RR, risk ratio; MD, mean difference.

Fig. 2.

Forest plot of IS compared to other rehabilitation strategies (i.e., the comparator) in reducing (a) post-operative pulmonary complications, (b) mortality, and (c) LHS. Studies were subcategorized into surgery type. RR and MD were the units of measurement for dichotomous and continuous outcomes, respectively. A random-effect model was used. Notable heterogeneity was present if I2 > 40% or p < 0.10. CI, confidence interval; IS, incentive spirometry; LHS, length of hospital stay; RR, risk ratio; MD, mean difference.

Close modal

The risk of bias for each included RCT is shown in Figure 3a-b; online supplementary Appendix 4. Only 4 RCTs had a low risk of bias in both randomization and allocation concealment [39, 40, 48, 54]. The remaining 27 RCTs had an unclear risk in at least one of these domains. The domain with the highest judged bias was blinding of participants and personnel for LHS. Meanwhile, the blinding of participants and personnel for mortality was judged to have the lowest bias. Most studies were found to be low risk of bias for blinding of outcome assessor when it came to PPCs and 30-day mortality. Although some of the studies had loss to follow-up, the risk of bias for incomplete outcome data was not considered high as none of the RCTs had >20% of data missing. Finally, 4 studies were found to have high selective reporting bias [30, 31, 48, 52], and 6 studies were found to have “other” high risk of bias for blocked randomization without blinding [59] or imbalanced baseline characteristics [30, 37, 44, 46, 53].

Fig. 3.

Risk of bias by (a) domain, and (b) study. Domains were outlined as directed by the Cochrane Handbook for Systematic Reviews of Interventions. Green, yellow, and red circles denote low, unclear, and high risk of bias, respectively. LHS, length of hospital stay; PPC, post-operative pulmonary complication.

Fig. 3.

Risk of bias by (a) domain, and (b) study. Domains were outlined as directed by the Cochrane Handbook for Systematic Reviews of Interventions. Green, yellow, and red circles denote low, unclear, and high risk of bias, respectively. LHS, length of hospital stay; PPC, post-operative pulmonary complication.

Close modal

Data on the incidence of PPC were pooled for meta-analysis from 26 studies (3,444 participants; shown in Fig. 2a) [30‒37, 39, 40, 42‒52, 54, 56‒58, 60] and an RR of 1.00 (95% CI 0.88–1.13; p = 0.96) was calculated. When assessed using GRADE, the number of events exceeded 300 (i.e., 657 events); therefore, the results were not imprecise [61]. Additionally, inconsistency was not considered to be serious as little to no heterogeneity was found in the effect estimates and their corresponding CIs. The only opportunity for indirectness may have arisen from the majority of included studies solely reporting PPCs developed in hospital. However, despite most studies not reporting for the entirety of 30 days, previous literature showed that majority of PPCs occur between 1 and 3 days following surgery [62]. Therefore, we did not have justification to suggest that the indirectness would change the relative effect of IS on the incidence of PPCs. The only domain that was rated down for PPC was the risk of bias. Allocation concealment was unclear in 25 out of the 29 studies [30‒37, 41‒47, 49‒53, 55‒58, 60], which introduced the opportunity for selection bias. Furthermore, only 9 out of the 29 studies had low risk of bias in the random sequence generation (shown in Fig. 3a-b; online suppl. Appendix 4) [30, 33, 35, 39, 40, 44, 48, 54, 58]. As such, there is moderate quality of evidence that IS does not reduce the incidence of PPCs when compared to other rehabilitation strategies (shown in Table 2). Celli et al. [33], Lim et al. [45], and Lunardi et al. [46] were multi-armed studies for upper abdominal surgery in which one arm received no rehabilitation. Additionally, the study by Schwieger et al. [56] was a two-armed study for upper abdominal surgery, which compared IS to no rehabilitation. Only Celli et al. [33] demonstrated a statistically significant (p < 0.05) reduction in PPCs for IS when compared to no rehabilitation. The remaining 3 studies saw no statistical difference in the number of PPCs between IS and no rehabilitation [45, 46, 56]. Three studies were not included in the meta-analysis as they either did not provide the total number of patients that developed a PPC [53, 55] or they did not include the groups’ sample sizes [41]. Oulton et al. [53] tested 2 types of IS devices (Triflo and Spirocare) against other chest physiotherapy post-CABG surgery. They detected the degree of atelectasis in each group at 5 different time points. For each time point, the Spirocare group had a statistically significant reduction in atelectasis compared to the 2 other groups. Savci et al. [55] compared IS to active breathing following CABG surgery. They reported the incidence of atelectasis on post-operative days 1, 3, and 5, which had an RR of 1.04 (95% CI: 0.78–1.40), 1.06 (95% CI: 0.71–1.57), and 0.90 (95% CI: 0.43–1.90), respectively. Thus, it showed that there were no significant differences among rehabilitation strategies in developing atelectasis. In contrast, Hsiao [41] utilized IS following thoracic surgery and reported one PPC in the IS group, while 10 PPCs occurred in the control group. These results were significantly different (X2 = 8.39, p = 0.004).

Table 2.

Summary of findings table for all outcomes described in this systematic review using GRADE

 Summary of findings table for all outcomes described in this systematic review using GRADE
 Summary of findings table for all outcomes described in this systematic review using GRADE

Nine studies reported 30-day mortality (2,169 participants; shown in Fig. 2b) [30, 32, 39, 40, 42, 46, 48, 59, 60], whereby the pooled analysis showed an RR of 0.73 (95% CI 0.42–1.25; p = 0.25). Using GRADE, imprecision was rated as serious since the collective sample size of the pooled studies did not have sufficient power to detect a statistical difference in mortality (a rare event) between IS and the other rehabilitation strategies [61]. The study by Hall et al. [39] was a low-risk study and comprised more than half the weight (65.7%) of studies included for the mortality meta-analysis. Nonetheless, we rated down risk of bias by one level given that allocation concealment remained unclear in most studies, and there was high risk of bias in the domains of “Other bias” [30, 46, 59], “Selective reporting” [30, 48], and “Random sequence generation” (shown in Fig. 3a-b; online suppl. Appendix 4) [42]. Inconsistency and indirectness were not considered serious problems as there was little to no heterogeneity in the relative point effect estimates and their corresponding CI (i.e., all CIs overlapped). Nor was there heterogeneity in the test for subgroup differences (I2 = 0%, p = 0.67). For these above-mentioned reasons, we have low quality of evidence for mortality, in which IS does not significantly reduce mortality when compared to other rehabilitation strategies (shown in Table 2). The multi-armed abdominal surgery study by Lunardi et al. [46] demonstrated that a similar number of patients died in both the IS and no rehabilitation groups. Comparatively, the study by Tyson et al. [59] was a two-armed study for upper abdominal surgery, which compared IS to no rehabilitation. They found a statistically significant (p = 0.02) reduction in mortality for IS when compared to no rehabilitation.

Fifteen studies were pooled for meta-analysis as it pertained to LHS (2,629 participants; shown in Fig. 2c) [30, 32, 33, 35, 37‒40, 46, 48, 51, 54, 57, 59, 60] and showed an MD of 0.17 fewer days (95% CI: 0.65 fewer to 0.30 more days; p = 0.48). LHS was not viewed as imprecise because it surpassed the threshold of a sample size greater than 400 (i.e., n = 2,629) [61]. Despite 8 out of the 15 studies displaying low risk for random sequence generation [30, 33, 35, 39, 40, 48, 54, 59], risk of bias was rated as serious because allocation concealment demonstrated a possible risk of selection bias with most studies unclear in this domain. Additionally, high risk of performance bias was observed in most studies (73%) as we felt that LHS was an outcome that could be affected by lack of patient blinding (shown in Fig. 3a-b; online suppl. Appendix 4). Lastly, inconsistency was downrated one level due to the unexplained heterogeneity (I2 = 42%, p = 0.04) observed in the relative effect estimates. Specifically, surgery-type subgroup analysis did not account for the heterogeneity, and the CI of Celli et al. [33] did not overlap with that of Lunardi et al. [46]. In an attempt to explain the heterogeneity, a sensitivity analysis was performed to remove studies with high risk of bias, and it was found that the overall heterogeneity increased from an I2 of 42% to 46%. As such, the risk of bias did not explain the heterogeneity observed amongst RCTs for LHS and the quality of evidence was low in which IS does not significantly reduce LHS when compared to other rehabilitation strategies (shown in Table 2). The multi-armed studies by Celli et al. [33] and Lunardi et al. [46] and the two-armed study by Tyson et al. [59] all compared LHS between IS to no rehabilitation. Only Celli et al. [33] demonstrated a statistically significant (p < 0.05) reduction in LHS for IS when compared to no rehabilitation. Lunardi et al. [46] and Tyson et al. [59] saw no statistical difference in LHS between IS and no rehabilitation. From a narrative perspective, one study – Ali et al. [31], was unable to be pooled in the meta-analysis due the LHS data being dichotomized. Ali et al. [31] compared IS with diaphragmic breathing following upper abdominal surgery and found that 92.5% of patients were discharged after >3 days in both groups, thus showing no statistical difference.

Given the comprehensive search strategy that included trial registries and no restrictions on publication date or language, this systematic review did not detect any publication bias (shown in Fig. 4a–c).

Fig. 4.

Funnel plots to assess publication bias of studies reporting (a) PPCs, (b) mortality, and (c) LHS, as an outcome. MD, mean difference; RR, risk ratio; SE, standard error; LHS, length of hospital stay; PPCs, post-operative pulmonary complications.

Fig. 4.

Funnel plots to assess publication bias of studies reporting (a) PPCs, (b) mortality, and (c) LHS, as an outcome. MD, mean difference; RR, risk ratio; SE, standard error; LHS, length of hospital stay; PPCs, post-operative pulmonary complications.

Close modal

This systematic review met its objective in using rigorous methodology to identify the effects of IS alone compared to other rehabilitation strategies in reducing PPCs, mortality, and LHS following cardiac, thoracic, and upper abdominal surgery. Our results show similar trends to previous systematic reviews in that the inclusion of additional RCTs continues to indicate that IS alone following cardiac, thoracic, and upper abdominal surgery likely does not significantly reduce the incidence of PPC, when compared to other rehabilitation strategies. Only 4 studies assessing PPCs had no rehabilitation as one of their treatment arms and all of these studies were for upper abdominal surgery. The majority and more recent of these studies suggest no statistical difference in the number of PPCs between IS and no rehabilitation, which puts into question the effectiveness of IS as a rehabilitation therapy, following upper abdominal surgeries. In terms of thoracic and cardiac surgeries, the comparison between no rehabilitation and IS may differ but remains unknown as all of these studies implemented some form of post-operative chest physiotherapy.

The 2 other outcomes, LHS and mortality, have been shown in previous literature to increase with the development of a PPC [5]. As such, in addition to being patient-important outcomes, the effects of IS on LHS and mortality were of interest due to their relations with PPCs. Although the current best evidence is low quality, to the best of our knowledge, this is the first systematic review to demonstrate through GRADE and meta-analyses that the use of IS alone following these surgeries does not significantly reduce LHS nor mortality, when compared to the other rehabilitation strategies. Both outcomes were downrated for risk of bias, while LHS also exhibited unexplained heterogeneity and mortality was imprecise due to the low number of events. Only 2 studies with no rehabilitation as a treatment arm assessed mortality, whereas 3 assessed LHS. Similar to PPCs, all studies were for abdominal surgery. The studies provided mixed results in whether IS was more effective than no rehabilitation for reducing mortality and LHS, following upper abdominal surgery.

All 3 outcomes from this systematic review contradict the original optimism healthcare professionals had in using IS to ameliorate maximal inspiration post-surgery. The device was invented with the belief that PPCs would be reduced as a result of its assumed improvements towards alveolar aeration, oxygenation, and pulmonary compliance [12]. Realistically over the last 40 years, there has been an accumulation of RCTs and systematic reviews that have put into question the beneficial effects of IS alone. However, a reoccurring criticism of the previous systematic reviews has been the paucity in rigorous methodology and partial adherence to the Cochrane Handbook guidelines [19]. Although our systematic review had similar findings to these previous systematic reviews, we were able to contribute to the existing literature in multiple ways. Firstly, our meta-analysis was enriched by the inclusion of more studies. This was made possible through the assessment of 3 types of surgical procedures and the reduced stringency of our search strategy (e.g., omitting outcome terms within the search strategy and accepting studies written in any language allowed us to identify a missed Korean RCT [45]). We were also able to conduct a comprehensive risk of bias assessment using the Cochrane Handbook’s most updated guidelines. Specifically, the Cochrane Handbook now recommends assessing performance bias on an outcome-to-outcome basis rather than automatically mark all RCTs as high risk of performance bias if blinding of personnel or patient is not feasible [19]. With this new take on assessment, the incidence of mortality and PPC could be viewed as low risk of performance bias, given their severity, objectivity, and physiological nature. The application of rigorous methodology enabled our systematic review to prove for the first time with moderate quality of evidence that IS alone does not significantly reduce the number of PPCs compared to other rehabilitation strategies, following cardiac, thoracic, and upper abdominal surgery. Moreover, we are the first systematic review of this topic to pool data for mortality and LHS within a meta-analysis and assess these outcomes with GRADE. Given PPCs are often correlated with prolonged LHS and heightened risk of mortality, our novel findings of no significant difference in LHS or mortality between IS alone and other rehabilitation strategies further reinforce that IS may not be an effective post-operative therapy on its own.

An apparent limitation with the included RCTs was the unclear risk of bias observed in the domains of random sequence generation and allocation concealment. Most studies did not adequately explain the processes they used in order to ensure randomization and allocation concealment. However, with the development of the CONSORT guidelines [63], it is anticipated that future RCTs will improve in these domains. Another component of IS that was not accessed in this systematic review was compliance. Compliance with the intervention could have been an issue for why the possible beneficial effects of IS alone were not observed in this systematic review; however, most of the included RCTs did not report on compliance, which is in line with the results from a recent systematic review [22]. Nonetheless, we believe the findings of this systematic review are more generalizable to surgical populations because it is at the discretion of the patient in how often they use the IS device. Imprecision was a limitation for the outcome of mortality because the cumulative sample size of the pooled studies for this outcome was too small to identify statistical differences between IS and its comparators. To mitigate the issue of rare events, RCTs with greater sample sizes would be required in order to accurately assess, with sufficient statistical power, the effects of IS on mortality. LHS differed from the other outcomes in that it displayed considerable heterogeneity overall, I2 = 42% (p = 0.04), which was not explained by our surgical-type subgroup analysis. Another issue regarding LHS involved the variability of its reporting as a median or mean. To resolve this issue, we used the estimation theories, which relied on the sample size, median, and range/interquartile range to estimate the mean and standard deviation [23‒25]. Although it was a useful technique in converting all data to one measure of central tendency, the conversions were estimations that may have contributed to some of the heterogeneity.

As shown in our systematic review, there is a cumulative amount of evidence to suggest that IS alone may not be a beneficial post-operative therapy. However, it may have an important role within a multimodal rehabilitation strategy. I COUGH is a rehabilitation program that combines the use of post-operative IS, patient education, coughing, deep breathing, early mobilization, head-of-bead elevation, and oral care. In 2013, results from the program’s clinical trial demonstrated that both incidence of post-operative pneumonia and unplanned intubation were reduced by participating in this program [64]. I COUGH along with other ERAS programs are being developed and seem to show promise as an effective strategy to reduce PPCs, mortality, and LHS [65‒67]. Therefore, future research should continue to explore the role of IS as part of one of these multimodal rehabilitation strategies.

Our research findings further validate the guidelines set by the AARC [10], but they also provide additional outcomes of IS for the use in thoracic surgery, which was not originally included within these guidelines. IS alone compared to other rehabilitation strategies results in little to no difference in PPCs, mortality, and LHS, following cardiac, thoracic, and upper abdominal surgeries. Future systematic reviews should assess whether there is any benefit of combining IS with other rehabilitation strategies post-operatively (e.g., the I COUGH program), in order to expand upon the recommendations set by the AARC [14]. Future RCTs should also improve upon reporting domains such as randomization sequence and allocation concealment. Lastly, RCTs with greater sample sizes are needed to increase the statistical power in assessing rare outcomes, such as mortality, and how rehabilitation strategies affect these outcomes following cardiac, thoracic, and upper abdominal surgery.

We would like to thank the teaching staff, from McMaster University’s HRM 743 – Systematic Review Methods course, for their guidance and counsel in developing this systematic review.

This was a systematic review and meta-analysis with studies acting as the unit of analysis instead of patients. As such, written consent and ethics approval were not required. However, this systematic review did comply with the reporting guidelines established by PRISMA to ensure transparency. The systematic review was also registered and accepted into PROSPERO with the reference ID: CRD42019137540.

There are no conflicts of interest.

The authors did not receive any funding.

Kerrie A. Sullivan and Isabella F. Churchill collaborated together to develop most components of this systematic review from inception to manuscript. Kerrie A. Sullivan also formatted the manuscript and provided the edits necessary to fit the specific guidelines set by Respiration. Danielle A. Hylton acted as the third reviewer during data acquisition and made significant contributions to the completion of this systematic review and its manuscript. Dr. Waël C. Hanna oversaw and contributed significantly to the production of this systematic review with full access to the data. Therefore, he along with Kerrie A. Sullivan and Isabella F. Churchill take full responsibility for the integrity and accuracy of the data, its analysis, and interpretation.

1.
Weiser
TG
,
Regenbogen
SE
,
Thompson
KD
,
Haynes
AB
,
Lipsitz
SR
,
Berry
WR
,
.
An estimation of the global volume of surgery: a modelling strategy based on available data
.
Lancet
.
2008
;
372
:
139
44
.
2.
Hemmes
SNT
,
Gama de Abreu
M
,
Pelosi
P
,
Schultz
MJ
.
High versus low positive end-expiratory pressure during general anaesthesia for open abdominal surgery (PROVHILO trial): a multicentre randomised controlled trial
.
Lancet
.
2014
;
384
:
495
503
.
3.
Miskovic
A
,
Lumb
AB
.
Postoperative pulmonary complications
.
Br J Anaesth
.
2017
;
118
:
317
34
.
4.
de Paisani
DM
,
Lunardi
AC
,
da Silva
CC
,
Porras
DC
,
Tanaka
C
,
Carvalho
CR
.
Volume rather than flow ncentive spirometry is effective in improving chest wall expansion and abdominal displacement using optoelectronic plethysmography
.
Respir Care
.
2013
;
58
(
8
):
1360
6
.
5.
Patel
RL
,
Townsend
ER
,
Fountain
SW
.
Elective pneumonectomy: factors associated with morbidity and operative mortality
.
Ann Thorac Surg
.
1992
;
54
:
84
8
.
6.
Stéphan
F
,
Boucheseiche
S
,
Hollande
J
,
Flahault
A
,
Cheffi
A
,
Bazelly
B
,
.
Pulmonary complications following lung resection: a comprehensive analysis of incidence and possible risk factors
.
Chest
.
2000
;
118
:
1263
70
.
7.
Kelkar
KV
.
Post-operative pulmonary complications after non-cardiothoracic surgery
.
Indian J Anaesth
.
2015
;
59
:
599
605
.
8.
Stulberg
JJ
,
Huang
R
,
Kreutzer
L
,
Ban
K
,
Champagne
BJ
,
Steele
SE
,
.
Association between surgeon technical skills and patient outcomes
.
JAMA Surg
.
2020
;
155
:
960
8
.
9.
Rehder
K
.
Anaesthesia and the respiratory system
.
Can Anaesth Soc J
.
1979
;
26
(
6
):
451
62
.
10.
Saraswat
V
.
Effects of anaesthesia techniques and drugs on pulmonary function
.
Indian J Anaesth
.
2015
;
59
:
557
64
.
11.
Fernandes
A
,
Rodrigues
J
,
Lages
P
,
Lança
S
,
Mendes
P
,
Antunes
L
,
.
Root causes and outcomes of postoperative pulmonary complications after abdominal surgery: a retrospective observational cohort study
.
Patient Saf Surg
.
2019
;
13
:
40
9
.
12.
Bartlett
RH
,
Gazzaniga
AB
,
Geraghty
TR
.
Respiratory maneuvers to prevent postoperative pulmonary complications: a critical review
.
JAMA
.
1973
;
224
:
1017
21
.
13.
Hanekom
S
,
Louw
QA
,
Coetzee
AR
.
Implementation of a protocol facilitates evidence-based physiotherapy practice in intensive care units
.
Physiotherapy
.
2013
;
99
:
139
45
.
14.
Restrepo
RD
,
Wettstein
R
,
Wittnebel
L
,
Tracy
M
.
Incentive spirometry: 2011
.
Respir Care
.
2011
;
56
:
1600
4
.
15.
Overend
TJ
,
Anderson
CM
,
Lucy
SD
,
Bhatia
C
,
Jonsson
BI
,
Timmermans
C
.
The effect of incentive spirometry on postoperative pulmonary complications: a systematic review
.
Chest
.
2001
;
120
:
971
8
.
16.
Carvalho
CR
,
Paisani
DM
,
Lunardi
AC
.
Incentive spirometry in major surgeries: a systematic review
.
Rev Bras Fisioter
.
2011
;
15
:
343
50
.
17.
do Nascimento Junior
P
,
Módolo
NS
,
Andrade
S
,
Guimarães
MM
,
Braz
LG
,
El Dib
R
.
Incentive spirometry for prevention of postoperative pulmonary complications in upper abdominal surgery
.
Cochrane Database Syst Rev
.
2014
;
2014
:
CD006058
.
18.
Guimarães
MM
,
El Dib
R
,
Smith
AF
,
Matos
D
,
Gomes
B
.
Incentive spirometry for prevention of postoperative pulmonary complications in upper abdominal surgery
.
Cochrane Database Syst Rev
.
2009
;
2009
(
3
):
CD006058
.
19.
Higgins
JPT
,
Green
S
, editors.
Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]
.
The Cochrane Collaboration
;
2011
.
20.
Moher
D
,
Liberati
A
,
Tetzlaff
J
,
Altman
DG
.
Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement
.
PLoS Med
.
2009
;
6
:
e1000097
.
21.
Colucci
DB
,
Fiore
JF
,
Paisani
DM
,
Risso
TT
,
Colucci
M
,
Chiavegato
LD
,
.
Cough impairment and risk of postoperative pulmonary complications after open upper abdominal surgery
.
Respir Care
.
2015
;
60
:
673
8
.
22.
Narayanan
AL
,
Hamid
SR
,
Supriyanto
E
.
Evidence regarding patient compliance with incentive spirometry interventions after cardiac, thoracic and abdominal surgeries: a systematic literature review
.
Can J Respir Ther
.
2016
;
52
:
17
26
.
23.
Hozo
SP
,
Djulbegovic
B
,
Hozo
I
.
Estimating the mean and variance from the median, range, and the size of a sample
.
BMC Med Res Methodol
.
2005
;
5
:
13
.
24.
Wan
X
,
Wang
W
,
Liu
J
,
Tong
T
.
Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range
.
BMC Med Res Methodol
.
2014
;
14
:
135
.
25.
Luo
D
,
Wan
X
,
Liu
J
,
Tong
T
.
Optimally estimating the sample mean from the sample size, median, mid-range, and/or mid-quartile range
.
Stat Methods Med Res
.
2018
;
27
:
1785
805
.
26.
McNeish
D
.
Missing data methods for arbitrary missingness with small samples
.
J Appl Stat
.
2017
;
44
(
1
):
24
39
.
27.
The Nordic Cochrane Centre
.
Review Manager (RevMan)
.
Cochrane Collaboration
;
2014
.
28.
Schünemann
H
,
Brożek
J
,
Guyatt
G
,
Andrew
O
, editors.
GRADE handbook for grading quality of evidence and strength of recommendations [Internet]
.
The GRADE Working Group
;
2013
. Available from: https://gdt.gradepro.org/app/handbook/handbook.html.
29.
The Cochrane Collaboration
.
GRADEpro GDT
.
Cochrane Community
;
2017
.
30.
Agostini
P
,
Naidu
B
,
Cieslik
H
,
Steyn
R
,
Rajesh
PB
,
Bishay
E
,
.
Effectiveness of incentive spirometry in patients following thoracotomy and lung resection including those at high risk for developing pulmonary complications
.
Thorax
.
2013
;
68
:
580
5
.
31.
Ali
SB
,
Gaad-Elmoula Shabaan
E
,
Mohammed Diab
T
,
Fehro
A
,
Eid
RM
.
Synergistic effect of flow incentive spirometer and diaphragmatic breathing exercise for patients with upper abdominal surgery
.
IOSR J Nurs Heal Sci
.
2018
;
7
:
1
11
.
32.
Al-Mutairi
FH
,
Fallows
SJ
,
Abukhudair
WA
,
Islam
BB
,
Morris
MM
.
Difference between continuous positive airway pressure via mask therapy and incentive spirometry to treat or prevent post-surgical atelectasis
.
Saudi Med J
.
2012
;
33
:
1190
5
.
33.
Celli
BR
,
Rodriguez
KS
,
Snider
GL
.
A controlled trial of intermittent positive pressure breathing, incentive spirometry, and deep breathing exercises in preventing pulmonary complications after abdominal surgery
.
Am Rev Respir Dis
.
1984
;
130
:
12
5
.
34.
Craven
JL
,
Evans
GA
,
Davenport
PJ
,
Williams
RH
.
The evaluation of the incentive spirometer in the management of postoperative pulmonary complications
.
Br J Surg
.
1974
;
61
:
793
7
.
35.
Crowe
JM
,
Bradley
CA
.
The effectiveness of incentive spirometry with physical therapy for high-risk patients after coronary artery bypass surgery
.
Phys Ther
.
1997
;
77
:
260
8
.
36.
Gale
GD
,
Sanders
DE
.
Incentive spirometry: its value after cardiac surgery
.
Can Anaesth Soc J
.
1980
;
27
:
475
80
.
37.
Gosselink
R
,
Schrever
K
,
Cops
P
,
Witvrouwen
H
,
De Leyn
P
,
Troosters
T
,
.
Incentive spirometry does not enhance recovery after thoracic surgery
.
Crit Care Med
.
2000
;
28
:
679
83
.
38.
Gunay
S
,
Eser
I
,
Ozbey
M
,
Agar
M
,
Koruk
I
,
Kurkcuoglu
IC
.
Evaluation of two different respiratory physiotherapy methods after thoracoscopy with regard to arterial blood gas, respiratory function test, number of days until discharge, cost analysis, comfort and pain control
.
Niger J Clin Pract
.
2016
;
19
:
353
8
.
39.
Hall
JC
,
Tarala
R
,
Harris
J
,
Tapper
J
,
Christiansen
K
.
Incentive spirometry versus routine chest physiotherapy for prevention of pulmonary complications after abdominal surgery
.
Lancet
.
1991
;
337
:
953
6
.
40.
Hall
JC
,
Tarala
RA
,
Tapper
J
,
Hall
JL
.
Prevention of respiratory complications after abdominal surgery: a randomised clinical trial
.
BMJ
.
1996
;
312
:
148
3
.
41.
Hsiao
WL
.
An early rehabilitation intervention for enhancing oxygenation from lung cancer surgery. IASLC 19th world conference on lung cancer
.
J Thoracic Oncol
.
2018
:
S328
9
.
42.
Iverson
LI
,
Ecker
RR
,
Fox
HE
,
May
IA
.
A comparative study of IPPB, the incentive spirometer, and blow bottles: the prevention of atelectasis following cardiac surgery
.
Ann Thorac Surg
.
1978
;
25
:
197
200
.
43.
Jenkins
SC
,
Soutar
SA
,
Loukota
JM
,
Johnson
LC
,
Moxham
J
.
A comparison of breathing exercises, incentive spirometry and mobilisation after coronary artery surgery
.
Physiother Theory Pract
.
1990
;
6
(
3
):
117
26
.
44.
Jung
R
,
Wight
J
,
Nusser
R
,
Rosoff
L
.
Comparison of three methods of respiratory care following upper abdominal surgery
.
Chest
.
1980
;
78
:
31
5
.
45.
Lim
C
,
Cho
H
,
Chang
SH
.
The effect of deep breathing exercise and incentive spirometry to prevent postoperative pulmonary complications after abdominal surgery in geriatric patients
.
Korean J Anesthesiol
.
1997
;
33
(
6
):
1185
91
.
46.
Lunardi
AC
,
Paisani
DM
,
Silva
CCBMD
,
Cano
DP
,
Tanaka
C
,
Carvalho
CRF
.
Comparison of lung expansion techniques on thoracoabdominal mechanics and incidence of pulmonary complications after upper abdominal surgery: a randomized and controlled trial
.
Chest
.
2015
;
148
:
1003
10
.
47.
Lyager
S
,
Wernberg
M
,
Rajani
N
,
Bøggild-Madsen
B
,
Nielsen
L
,
Nielsen
HC
,
.
Can postoperative pulmonary conditions be improved by treatment with the Bartlett-Edwards incentive spirometer after upper abdominal surgery
.
Acta Anaesthesiol Scand
.
1979
;
23
:
312
9
.
48.
Malik
PRA
,
Fahim
C
,
Vernon
J
,
Thomas
P
,
Schieman
C
,
Finley
CJ
,
.
Incentive spirometry after lung resection: a randomized controlled trial
.
Ann Thorac Surg
.
2018
;
106
:
340
5
.
49.
Matte
P
,
Jacquet
L
,
Van Dyck
M
,
Goenen
M
.
Effects of conventional physiotherapy, continuous positive airway pressure and non-invasive ventilatory support with bilevel positive airway pressure after coronary artery bypass grafting
.
Acta Anaesthesiol Scand
.
2000
;
44
:
75
81
.
50.
Minschaert
M
,
Vincent
JL
,
Ros
AM
,
Kahn
RJ
.
Influence of incentive spirometry on pulmonary volumes after laparotomy
.
Acta Anaesthesiol Belg
.
1982
;
33
:
203
9
.
51.
O’Connor
M
,
Tattersall
MP
,
Carter
JA
.
An evaluation of the incentive spirometer to improve lung function after cholecystectomy
.
Anaesthesia
.
1988
;
43
:
785
7
.
52.
Oikkonen
M
,
Karjalainen
K
,
Kähärä
V
,
Kuosa
R
,
Schavikin
L
.
Comparison of incentive spirometry and intermittent positive pressure breathing after coronary artery bypass graft
.
Chest
.
1991
;
99
:
60
5
.
53.
Oulton
JL
,
Hobbs
GM
,
Hicken
P
.
Incentive breathing devices and chest physiopathy: a controlled therapy
.
Can J Surg
.
1981
;
24
:
638
40
.
54.
Rowley
DD
,
Malinowski
TP
,
Di Peppe
JL
,
Sharkey
RM
,
Gochenour
DU
,
Enfield
KB
.
A randomized controlled trial comparing two lung expansion therapies after upper abdominal surgery
.
Respir Care
.
2019
;
64
:
1181
92
.
55.
Savci
S
,
Sakinç
S
,
Ince
DI
,
Arikan
H
,
Can
Z
,
Buran
Y
,
.
Active cycle of breathing techniques and incentive spirometer in coronary artery bypass graft surgery
.
Fiz Rehabil
.
2006
;
17
:
61
9
.
56.
Schwieger
I
,
Gamulin
Z
,
Forster
A
,
Meyer
P
,
Gemperle
M
,
Suter
PM
.
Absence of benefit of incentive spirometry in low-risk patients undergoing elective cholecystectomy. A controlled randomized study
.
Chest
.
1986
;
89
:
652
6
.
57.
Sleszynski
SL
,
Kelso
AF
.
Comparison of thoracic manipulation with incentive spirometry in preventing postoperative atelectasis
.
J Am Osteopath Assoc
.
1993
;
93
:
834
8
, 843–5.
58.
Stock
MC
,
Downs
JB
,
Gauer
PK
,
Alster
JM
,
Imrey
PB
.
Prevention of postoperative pulmonary complications with CPAP, incentive spirometry, and conservative therapy
.
Chest
.
1985
;
87
:
151
7
.
59.
Tyson
AF
,
Kendig
CE
,
Mabedi
C
,
Cairns
BA
,
Charles
AG
.
The effect of incentive spirometry on postoperative pulmonary function following laparotomy: a randomized clinical trial
.
JAMA Surg
.
2015
;
150
:
229
36
.
60.
Vilaplana
J
,
Sabaté
A
,
Ramon
R
,
Gasolibe
V
,
Villalonga
R
.
[Ineffectiveness of incentive spirometry as coadjuvant of conventional physiotherapy for the prevention of postoperative respiratory complications after thoracic and esophageal surgery]
.
Rev Esp Anestesiol Reanim
.
1990
;
37
:
321
5
.
61.
Guyatt
GH
,
Oxman
AD
,
Kunz
R
,
Brozek
J
,
Alonso-Coello
P
,
Rind
D
,
.
GRADE guidelines 6. Rating the quality of evidence-imprecision
.
J Clin Epidemiol
.
2011
;
64
:
1283
93
.
62.
Thompson
JS
,
Baxter
BT
,
Allison
JG
,
Johnson
FE
,
Lee
KK
,
Park
WY
,
.
Temporal patterns of postoperative complications
.
Arch Surg
.
2003
;
138
:
596
3
.
63.
Schulz
KF
,
Altman
DG
,
Moher
D
;
the CONSORT Group
.
CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials (Chinese version)
.
BMC Med
.
2010
;
340
:
c332
.
64.
Cassidy
MR
,
Rosenkranz
P
,
McCabe
K
,
Rosen
JE
,
McAneny
DI
.
COUGH: reducing postoperative pulmonary complications with a multidisciplinary patient care program
.
JAMA Surg
.
2013
;
148
:
740
5
.
65.
Visioni
A
,
Shah
R
,
Gabriel
E
,
Attwood
K
,
Kukar
M
,
Nurkin
S
.
Enhanced recovery after surgery for noncolorectal surgery? A systematic review and meta-analysis of major abdominal surgery
.
Ann Surg
.
2018
;
267
:
57
65
.
66.
Bagnall
NM
,
Malietzis
G
,
Kennedy
RH
,
Athanasiou
T
,
Faiz
O
,
Darzi
A
.
A systematic review of enhanced recovery care after colorectal surgery in elderly patients
.
Colorectal Dis
.
2014
;
16
:
947
56
.
67.
McGinigle
KL
,
Eldrup-Jorgensen
J
,
McCall
R
,
Freeman
NL
,
Pascarella
L
,
Farber
MA
,
.
A systematic review of enhanced recovery after surgery for vascular operations
.
J Vasc Surg
.
2019
;
70
:
629
40.e1
.