The incidence of cancers in the global population is increasing.1 For selected patients, surgery with or without radiochemotherapy is the main treatment option. The goal of surgery is to obtain a clear resection margin and ultimately cure or prolong survival with an acceptable quality of life.2,3 However, despite the significant improvements in long-term survival over the recent years, the rate of postoperative morbidity remains high—increasing the length of hospital stay, reducing quality of life and contributing to a high treatment burden.

During the past 20 years, cardiopulmonary exercise test (CPET) was introduced during the preoperative period as an objective measure of functional capacity to evaluate the risk of adverse perioperative events and inform the perioperative management, particularly in high-risk patients undergoing high-risk surgery.4,5 Recently, CPET has gained popularity and is commonly used in high-risk patients undergoing cancer surgery in some surgical units. It is hypothesised that fitter patients, who were identified by using CPET, have greater physiological reserve to undergo surgery and recover sooner with fewer postoperative complications.6 This is extremely important for clinicians to inform decision-making, to better understand the postoperative course, and to guide postoperative management.

Several systematic reviews have explored the potential association between preoperative CPET variables and postoperative outcomes, demonstrating mixed results. While some systematic reviews have reported a significant positive association between preoperative CPET variables and postoperative complications, length of hospital stay, unplanned ICU admission, and 12-months survival,5,7,8 others have reported nonsignificant association.9,10,11 The prospective, multinational cohort (METS) study, for example, demonstrated an association between peak oxygen uptake (peak VO2) and noncardiac complications in a cohort of relatively well patients having noncardiac surgery and not limited to cancer surgery.12 Some limitations encountered within the previous systematic reviews include the absence of meta-analysis, inclusion of a mixed population (i.e., cancer and noncancer patients), outdated, or focused on a narrow cohort of patients.5,7,8,9,10,11 Better understanding of the potential association between preoperative CPET variables and postoperative outcomes in cancer patients is extremely important; this can guide preoperative interventions designed to improve patients preoperative physical status. This, in turn, has the potential to reduce postoperative morbidity.

As the number of publications are rapidly growing further analysis, taking into consideration the limitations of the previous systematic reviews, is warranted. This systematic review aims to determine whether the preoperative CPET variables peak VO2, anaerobic threshold (AT), and ventilatory equivalent for carbon dioxide (VE/VCO2), are associated with postoperative complication rates, length of hospital stay, and quality of life in patients undergoing cancer surgery.

Methods

Protocol and Registration

This systematic review was reported in accordance with the meta-analyses of Observational Studies in Epidemiology (MOOSE) checklist.13 The protocol for this systematic review was registered on the Open Science Framework website (https://osf.io/8ntvc/).

Information Sources and Search

A sensitive electronic search was performed via Ovid in MEDLINE, Embase, AMED, and Web of science via www.webofknowledge.com from inception to April 2020. An amalgamation of Medical Subject Headings (MeSH) terms and key words for “preoperative,” “cardiopulmonary exercise test,” and “neoplasm” was used in the search strategy (Supplementary Table 1). In addition, citation tracking of the included studies and relevant systematic reviews were conducted. The search was limited to humans with no date or language restrictions applied.

Study Selection

The screening process was conducted using Covidence (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org). The initial screening was completed by one review author removing clearly irrelevant studies (DS). Screening of titles and abstracts of potentially eligible studies was be performed by two independent review authors (DS and PRB) with full-text article assessed against the inclusion and exclusion criteria. Any disagreements over the eligibility of particular studies were resolved through discussion with a third review author (NP).

Data Collection Process

A standardized data form was used to extract data from eligible studies for assessment of the study quality and evidence synthesis. Two independent review authors extract the data independently (DS and PRB). Disagreements over the data extraction were resolved through discussion with a third reviewer (NP). The following information were extracted from each individual study: participant characteristics, study characteristics, CPET description and measures, postoperative outcome measures, and measures of association.

Eligibility Criteria

Longitudinal studies reporting on the association between preoperative CPET values and postoperative outcomes in adult patients aged ≥18 years old undergoing a cancer-related surgical procedure were included if they reported the following: (i) at least one of the CPET measures of interest: (a) Peak Oxygen uptake (peak VO2): defined as the highest VO2 attained on a rapid incremental test. (b) Anaerobic threshold (AT): a submaximal index of exercise capacity defined as the oxygen uptake (VO2) above which there is a metabolic transition to increased glycolysis and lactate begins to rise with an associated metabolic acidosis. (c) Ventilatory equivalent for carbon dioxide (VE/VCO2): defined as the ratio of minute ventilation to carbon dioxide production usually reported at the AT; (ii) reported at least one postoperative outcome measure, including complication rate, length of hospital stay, and/or quality of life; (iii) Reported data on the association between preoperative CPET and postoperative outcome or provide enough data for the association to be calculated by the review authors.

Studies were excluded if they presented the following: (i) reported on mixed populations (e.g., cancer and noncancer patients, where the noncancer population >5% of the investigated sample); (ii) the population of interest underwent open and close procedure (e.g., not completed as planned); (iii) abstracts of studies published on conference proceedings.

Risk of Bias Assessment

Risk of bias was assessed by using the Quality in Prognosis Studies (QUIPS) tool and was rated by two review authors (DS and PRB).14 Risk of bias was rated as “high”, “moderate”, or “low” risk according to the following domains: (i) Study participation; (ii) Study attrition; (iii) Outcome measurement; (vi) Statistical analysis and reporting. Due to the nature of this systematic review, the prognostic factor and study confounding domains were not judged as they were deemed not applicable. Disagreements over the risk of bias were resolved through discussion with a third review author (NP).

Strategy for Data Synthesis

For studies reporting on the association between preoperative CPET values and postoperative outcomes using continuous data, measures of central tendency (i.e., mean, median) and dispersion (i.e., standard deviation, 95% confidence intervals [CI]) were extracted. However, for the studies reporting on dichotomous data, the number of patients presenting high/low CPET values and presence or absence of postoperative outcomes were extracted. Whenever possible, mean values and standard deviation were estimated using previously published formulas in order to pool data.15 When raw data were available, mean difference and 95% CI (continuous) or odds ratios and 95% CI (dichotomous) were calculated. For homogeneous studies (e.g., presenting comparable measures of CPET and postoperative outcomes) reporting on the association between preoperative CPET and postoperative outcomes a meta-analysis using a random-effect model was conducted. Studies presenting high variability of data types and format were presented descriptively. A post-hoc subgroup analysis was performed to investigate the association between preoperative CPET variables and postoperative outcomes according to cancer type. Pooled estimates were obtained with Comprehensive Meta-Analysis Software V.3 (Biostat, Englewood, NJ).

Results

Study Selection

The electronic search yielded 843 potential studies after duplicates were removed. Of these, 212 full-text articles were considered for inclusion. A total of 58 published articles (including 52 unique cohorts) were included in this systematic review. The flow diagram of the inclusion process is presented in Fig. 1.

Fig. 1
figure 1

Included studies

Full size image

Study Characteristics

Of the 52 unique cohorts included, three included patients presenting with bladder cancer,16,17,18 5 colorectal,19,20,21,22,23 5 esophageal,24,25,26,27,28,29 3 liver,30,31,32 27 lung,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60 3 pancreatic,61,62,63 1 rectal,64 and 5 included mixed cancer populations.65,66,67,68,69,70 The sample size of the included studies ranged from 8 to 1684.51,60 Most of the preoperative CPETs were performed by using a cycle ergometer. Peak VO2 was assessed in most studies (88%), followed by AT (44%). All included studies reported postoperative complication as an outcome, whereas quality of life was not reported in any of the included studies. The characteristics of the individual studies are presented in Table 1.

Table 1 Characteristics of the included studies
Full size table

Risk of Bias

Overall, most studies were rated as having low risk of bias. Study participation was the domain with higher risk of bias (46% rated as moderate/high risk of bias), whereas study attrition (10% rated as moderate risk of bias), outcome measurement (10% rated as moderate risk of bias), and statistical analysis and reporting (13% rated as moderate risk of bias) were rated with lower risk of bias. The risk of bias assessment for each of the included study can be found in Table 2.

Table 2 Risk of bias assessment using the Quality in Prognosis Studies (QUIPS) tool
Full size table

Association between Preoperative CPET Values and Postoperative Complications

Peak VO2

The association between preoperative Peak VO2 and postoperative complications are presented in Fig. 2, Supplementary Fig. 1, and Tables 3 and 4. Our pooled analysis demonstrated that patients with no postoperative complication presented for surgery with a higher Peak VO2 (MD: 2.28; 95% CI: 1.26–3.29; I2 = 9%) compared with patients who had postoperative complications (Fig. 2). A subgroup analysis in lung cancer patients demonstrated similar association (MD: 2.40; 95% CI: 1.50-3.30) (Supplementary Fig. 2).

Fig. 2
figure 2

Forest plot of the association between preoperative peak oxygen uptake (peak VO2) in ml/kg/min and postoperative complication. Mean difference > 0 indicate higher preoperative peak VO2 in patients with no postoperative complications. CI=Confidence level

Full size image
Table 3 Association between preoperative cardiopulmonary exercise test variables and postoperative complication
Full size table
Table 4 Association between preoperative cardiopulmonary exercise test variables and postoperative complications
Full size table

In addition, patients with no postoperative pulmonary complications (MD: 1.47; 95% CI: 0.49–2.45; I2 = 0%), minor complications (MD: 2.01; 95% CI: 0.90–3.13; I2 = 27%), no cardiovascular complication (MD: 2.23; 95% CI: 0.30–4.15), or no in-hospital mortality (MD: 2.78; 95% CI: 1.12–4.43) compared with patients who presented with postoperative complications, presented for surgery with a significantly higher Peak VO2 (Fig. 2). No difference in Peak VO2 was found for patients with or without postoperative cardiopulmonary complications (Fig. 2). Other studies were not pooled in the meta-analysis due to high heterogeneity and reported mixed results (Tables 3 and 4).

AT

The association between preoperative AT and postoperative complications are presented in Fig. 3, Supplementary Fig. 1, and Tables 3 and 4. Our pooled analysis demonstrated no significant difference in preoperative AT values for patients with or without postoperative complications (MD: 0.15; 95% CI: −0.32 to 0.62) and cardiopulmonary complication (MD: 1.05; 95% CI: −0.17 to 2.26; I2 = 0%). Preoperative AT values were significantly higher in patients who presented minor complications compared with major complications (MD: 2.15; 95% CI: 1.29–3.00; I2 = 0%) and for no in-hospital mortality compared with in-hospital mortality (MD: 2.27; 95% CI: 1.03–3.51) (Fig. 3). Other studies were not pooled in the meta-analysis due to heterogeneity and reported mixed results (Tables 3 and 4). Similar results were found on our subgroup analysis according to cancer type (Supplementary Fig. 3).

Fig. 3
figure 3

Forest plot of the association between preoperative anaerobic threshold (AT) in ml\kg\min and postoperative complication. Mean difference>0 indicate higher preoperative AT in Patients with no postoperative complications. CI = Confidence level

Full size image

VE/VCO2

The association between preoperative VE/VCO2 and postoperative complications are presented in Fig. 4 and Tables 3 and 4. Our pooled analysis demonstrated that preoperative VE/VCO2 values were significant lower in patients with no pulmonary complication compared with patients with pulmonary complication (MD: 3.54; 95% CI: 1.82-5.25; I2 = 0%). No significant differences in preoperative VE/VCO2 values were observed for patients with or without postoperative complication (MD: 0.80; 95% CI: −0.95 to 2.54) and minor or major postoperative complication (MD: 0.93; 95% CI: −1.53 to 3.38) (Fig. 4). Other studies were not pooled in the meta-analysis due to heterogeneity and reported mixed results (Tables 3 and 4).

Fig. 4
figure 4

Forest plot of the association between preoperative ventilatory equivalent for carbon dioxide (VE/VCO2) and postoperative complications. Mean difference >0 indicate lower preoperative VE/VCO2 in patients with no postoperative complications. CI=Confidence level

Full size image

Association between Preoperative CPET Values and Length of Hospital Stay

The association between preoperative Peak VO2 (7 studies), AT (8 studies), and VE/VCVO2 (3 studies) and length of hospital stay is presented in Table 5. Results of individual studies provided mixed results. Some studies reported a positive association between CPET variables and length of hospital stay (i.e., patients presenting higher CPET values stayed shorter in hospital), and others reported no statistical differences. However, none of the studies reported a significant negative association (i.e., patients presenting lower CPET values stayed for shorter periods in hospital) (Table 5).

Table 5 Association between preoperative cardiopulmonary exercise test variables and length of hospital stay
Full size table

Association between Preoperative CPET Values and Postoperative Quality of Life

Currently, no study has investigated the association between preoperative CPET values and postoperative quality of life outcomes in patients undergoing cancer surgery.

Discussion

Statement of Principal Findings

This systematic review identified many studies investigating the potential association between preoperative CPET values and postoperative complications and length of hospital stay. Our meta-analysis demonstrated that higher preoperative Peak VO2, AT, and lower VE/VCO2 values were predominately significantly associated with absence of postoperative complications. Several individual studies were not included in the meta-analysis due to heterogeneity in the CPET values and outcomes or did not report appropriate values to be pooled. While the results of individual studies provided mixed results, it is important to note that none reported a negative association (i.e., superior preoperative CPET values associated with worst postoperative outcome). Similarly, the association between preoperative CPET values and length of hospital stay reported in individual studies provided mixed results; none reported a negative association. Interestingly, this review was not able to identify any study investigating the association between preoperative CPET values and postoperative quality of life outcomes.

Strengths and Weaknesses of the Study

The strengths of this systematic review and meta-analyses were the methodology employed, following recommendation from the Cochrane Prognosis Review Group, and were reported according to the MOOSE framework. In addition, we conducted a sensitive search on major medical databases, that was supported by a senior librarian. Our search was only limited by human subjects and included all the literature irrespective of language and publication year. Furthermore, we assessed risk of bias using a well stablished tool (QUIPS).

The limitation of our systematic review included the heterogeneity between the included studies. For many included studies, meta-analysis was not possible as the CPET variables and outcome measures were not standardised and prevented pooling of the data. Also, due to the population of interest (patients undergoing cancer surgery), peak VO2 and VO2 max were used in this review interchangeably.6 Because these patients are older and debilitated by their conditions, it is difficult to demonstrate that the plateau criterion for VO2 max has been met in response to exercise. Furthermore, none of the included studies investigated the association between preoperative CPET and postoperative quality of life, underpinning the lack of evidence for this important patient reported outcome. Lastly, while we included a large number of full-text manuscripts published in scientific journals, we excluded studies that were published as abstracts of conference proceedings.

Comparison with Other Studies

The association between preoperative CPET variables and postoperative complications and/or length of hospital stay has been investigated in previous systematic reviews, reporting mixed results. While there are few systematic reviews indicating a positive association between superior preoperative CPET values and absence of postoperative complications, others reported no significant association.5,7,8,10,11 This is somewhat in line with the results of the current review. Our meta-analysis showed that superior preoperative CPET values are significantly associated with the absence of most postoperative complications. However, results from studies that were not included in our meta-analysis are somewhat less favorable.

Despite this, there are some differences between the current and previous systematic reviews that are important to note. Previous systematic reviews included a smaller number of studies (ranging from 7–37), investigated postoperative complications as the main outcome measure, and included either a specific cancer population undergoing surgery (e.g., lung, esophageal), or mixed populations undergoing surgery for cancer and/or noncancer related conditions.5,7,8,9,10,11 Meta-analysis was attempted in only half of the previous published systematic reviews. Therefore, the mixed results encountered between the current review and previous reviews may be because the inclusion and exclusion criterion were different. The heterogeneity of the included cohorts, including the lack of consistency in reporting or standardisation of outcomes were highlighted in most of the previous systematic reviews. This has prevented the pooling of a larger number of studies and a stronger conclusion across the identified reviews.

Meaning of the Study

Despite the advances in the medical field and surgical approaches, postoperative complications following cancer surgery remain high, increasing the length of hospital stay and subsequently hospital costs. Therefore, identifying preoperative factors that accurately predict adverse postoperative outcomes would be of great benefit to inform potential optimization strategies, improve the processes of shared decision making, and informed consent in patients presenting for major cancer surgery. In a mixed group of cancer patients undergoing surgery, our systematic review and meta-analysis found that superior CPET values were associated with improved postoperative outcomes. Furthermore, results for individual studies, not included in the meta-analysis, also provided some positive trends. These also trends have been reported in other systematic reviews.5,7,8 Therefore, the assessment of functional capacity in the preoperative period should be used in conjunction with other clinical assessments to support clinicians, patients, and payers on optimization strategies and treatment decision making. This has the potential to provide the best possible outcome for patients and reduce the economic burden.

Unanswered Questions and Future Research

One of the goals of this study was to explore the association between preoperative CPET values and postoperative quality of life outcomes. Unfortunately, none of the included literature assessed this potential association. Future prospective cohort studies should include quality of life as one of the postoperative outcomes of interest. Our review not only focused on continuous measures of preoperative CPET, but also extracted dichotomous outcomes, or potential CPET cutoff points. This information was presented descriptively as the included studies presented a wide range of heterogeneity, especially using different cutoff points. Larger, prospective, cohort studies or perhaps a systematic review of individual patient data, should explore this further. Whenever possible, a subgroup analysis, involving specific groups of patients should be explored, to test whether different cutoff points for different patient cohorts provide more accurate predictive models. Future studies should attempt to use standardized CPET protocols and standardized definitions for postoperative outcomes.71 This would allow future systematic reviewers to pool data from a larger number of studies. Finally,, future clinical trials should investigate the most effective exercise regime to increase preoperative physical fitness. The measurement of peak VO2 and AT before and after the preoperative exercise regime would facilitate the investigation of this effect.

Conclusions

This systematic review and meta-analysis has demonstrated that superior preoperative CPET values, especially peak VO2, were significantly associated with improved postoperative outcomes in patients undergoing cancer surgery. The predictive value of preoperative CPET on length of hospital stay or quality of life outcomes was not able to be determined, due to the high heterogeneity or lack of studies, respectively. Results from individual studies not included in the meta-analysis also reported positive trends. Most importantly, none of the identified studies reported a negative association between preoperative CPET values and postoperative outcomes. The authors of this review recommend the use of preoperative CPET before cancer surgery to predict postoperative outcomes.