Abstract
BACKGROUND: In non-small-cell lung cancer patients, high peak oxygen uptake (peak V̇O2) predicts lower rates of postoperative complications and better long-term survival. Neoadjuvant chemotherapy (NAC) may negatively impact peak V̇O2.
METHODS: Cardiopulmonary exercise testing (CPET) was performed in 34 consecutive stage IIIA/IIIB non-small-cell lung cancer subjects scheduled for elective lung surgery. Using multivariate linear regression adjusted for potential confounders, we compared CPET results in subjects receiving or not receiving NAC (NAC+, n = 19; NAC−, n = 15).
RESULTS: Adjusted peak V̇O2 was lower in NAC + compared with NAC− subjects (−5.3 mL/min/kg [95% CI −8.3 to −2.2], P = .01). Likewise, oxygen pulse, maximal work load, and ventilatory threshold were also lower in NAC+ subjects, whereas peak heart rate and breathing reserve were similar. NAC+ subjects presented lower values of diffusion capacity for carbon monoxide (DLCO) (P = .035) and hemoglobin concentrations (P < .001). DLCO was strongly correlated with peak V̇O2 (r2 = 0.56). Adjustment for DLCO reduced the effect of NAC on peak V̇O2 without suppressing it.
CONCLUSIONS: NAC was associated with lower preoperative peak V̇O2 in subjects with non-small-cell lung cancer. This lower aerobic fitness may result from NAC-induced reduction in pulmonary gas exchange or heart toxicity. Since lower fitness is linked to poorer outcome, the decision for NAC may have to be balanced with its possible toxicity.
- chemotherapy
- peak oxygen uptake
- cardio-pulmonary exercise tests
- lung cancer
- lung surgery
- training
- post operative complications
Introduction
Patients with non-small-cell lung cancer generally have limited exercise capacity.1 Aerobic exercise capacity is a key variable for the decision to perform tumor resection. This process involves oxygen transport and delivery by the pulmonary and cardiovascular systems and oxygen consumption by the skeletal muscles.2 Peak oxygen uptake (peak V̇O2), as determined during cardiopulmonary exercise testing (CPET), is the reference method to evaluate exercise capacity and predicts early postoperative complications as well as long-term outcome.3–6 Besides aging, comorbidities such as COPD, diabetes mellitus, hypertension, peripheral artery disease, and heart failure are highly prevalent in these patients and may all limit aerobic exercise capacity.7 Moreover, sedentary behavior and the inflammatory component associated with cancer contribute to muscle wasting and decreased peripheral aerobic metabolic capacity. Finally, there can be deleterious effects of surgery, radiotherapy, and chemotherapy.1
Chemotherapeutic agents can affect the cardiovascular system and induce anemia, both impeding cardiovascular oxygen transport.8 Platinum-based chemotherapy may affect the pulmonary system, worsening pulmonary oxygen uptake.8 The effects of other classes of chemotherapeutic agents and the effects of chemotherapy in general on other components of oxygen delivery and consumption are still poorly explored.9 Despite such serious adverse effects, survival is improved with preoperative neoadjuvant chemotherapy (NAC) followed by surgery, compared with surgery alone.10–12 By reducing the tumor burden and downgrading the tumor stage, NAC facilitates surgical resection and may improve long-term outcome by eradicating micrometastases. But NAC may be a double-edged sword. There is emerging evidence that NAC may decrease physical fitness by affecting oxygen transport system components, although it is still unclear which NAC toxicity-related effects result in lowered aerobic capacity.13
To address this point, we compared preoperative CPET-derived parameters in a cohort of non-small-cell lung cancer subjects scheduled for curative lung resection and receiving or not receiving NAC. We hypothesized that subjects receiving NAC would have lower aerobic exercise capacity compared with subjects not treated with NAC.
QUICK LOOK
Current knowledge
In patients with non-small-cell lung cancer, the postoperative complication rate is associated with low peak oxygen uptake. Neoadjuvant chemotherapy has been shown to be effective in improving survival when compared with surgery alone in selected patients. However, toxicity of chemotherapeutic agents may affect heart and lung function beyond anemia, thus reducing peak oxygen uptake.
What this paper contributes to our knowledge
Neoadjuvant chemotherapy was associated with significantly lower preoperative oxygen uptake in subjects with non-small-cell lung cancer. This lower aerobic fitness may result from chemotherapy-induced reduction in diffusion capacity and heart or skeletal muscle toxicity.
Methods
Study Design and Setting
This prospective study took place over a period of 49 months, from January 2011 to February 2014, as part of an ongoing randomized controlled trial comparing the effect of short-term preoperative exercise training versus usual care in subjects with operable non-small-cell lung cancer.14 All subjects were recruited in the pulmonary divisions of 3 tertiary care centers located in the French-speaking part of Switzerland (University Hospitals of Geneva, University Hospital of Lausanne, and Hôpital du Valais, Sion). A cross-sectional analysis was carried out to assess preoperative exercise capacity (as measured by CPET) in non-small-cell lung cancer subjects with or without NAC scheduled for elective curative lung resection.
Participants
We included all subjects >18 y old with non-small-cell lung cancer IIIA and IIIB, documented by CT scan, positron emission tomography-CT, and pathological confirmation and judged eligible for tumor resection (open thoracotomy or video-assisted thoracoscopy), with or without NAC. The institutional research ethics board approved the study and informed written consent was obtained from each subject. Institutional review board approval (Protocol 09-263) was obtained from the University Hospitals of Geneva on August 10, 2010.
Based on preoperative investigations, subjects were staged according to the guidelines of the American Joint Committee on Cancer.15 Decisions to offer NAC were made by a multidisciplinary tumor board, based on guidelines, clinical judgment, and subjects' preferences. Exclusion criteria consisted of contraindications to perform CPET (uncontrolled heart disease, severe pulmonary hypertension) or functional limitations related to joint or psychiatric disease that precluded CPET.
Outcomes, Exposure Variables, and Potential Confounders
The main outcome variable was peak V̇O2 obtained from CPET. Secondary outcomes were other CPET results (maximal work load, ventilatory threshold, ventilatory equivalents [V̇E/V̇O2 and V̇E/V̇CO2], breathing reserve [estimated], peak heart rate, and oxygen pulse [V̇O2/heart rate]). Data were collected preoperatively after a median of 3.1 ± 0.5 cycles of NAC. In accordance with international guidelines, NAC regimens consisted of platinum-based doublet chemotherapy (supplementary Table 1; see the supplementary materials at http://www.rcjournal.com).16,17
Nineteen subjects with stage IIIA and IIIB non-small-cell lung cancer who received NAC as part of a multimodal treatment were considered as exposed (NAC+). Subjects with non-small-cell lung cancer stage IIIA and IIIB who did not receive NAC (n = 15) were considered as the control group (NAC−). Subjects with stage I to IIB operable lung cancer who did not receive NAC (n = 113) were used as additional controls in a sensitivity analysis. The following potential confounders or precision variables were taken into account in the statistical analysis: age, sex, weight, height, self-reported comorbidities (cardiovascular disease, hypertension), pulmonary function (FEV1, carbon monoxide diffusion capacity [DLCO]), venous blood hemoglobin concentration, and N-terminal prohormone of brain natriuretic peptide (NT-proBNP).
Pulmonary Function and CPET Procedures
All subjects performed pulmonary function testing according to American Thoracic Society/European Respiratory Society recommendations with measurement of post-bronchodilator FEV1 and carbon monoxide diffusion capacity (DLCO adjusted for hemoglobin according to the following formula: DLCO predicted for Hb = DLCO predicted × (1.7 Hb/10.22 + Hb).18,19 Symptom-limited CPET was performed on a cycle ergometer (Vmax Encore, SensorMedics, San Diego, California) with a maximal incremental work load protocol (10–15-W increments every minute, after a 3-min warm-up at 25 W), recording work load, breath-by-breath gas exchange, electrocardiogram, pulse oximetry, and noninvasive blood pressure. The flow meter was calibrated with a 3-L syringe, and the gas analyzers were calibrated with certified calibration gas mixtures before each test. The subjects exercised until exhaustion or until reaching criteria for exercise termination according to the American Thoracic Society/American College of Chest Physicians statement.20 CPET was considered objectively maximal if the respiratory quotient was >1.2 or peak heart rate was >80% predicted.21 It was considered as subjectively maximal if the subject perceived breathlessness of 9–10 on a 0–10 Borg scale. Ventilatory threshold was determined with the ventilatory equivalent method and the V-slope method.20 Breathing reserve was estimated with the formula: (measured maximal ventilation [MVV] predicted − MVV)/MVV predicted.22
Statistical Analysis
Categorical and continuous variables were compared between the NAC+ and NAC− groups with parametric Student test or chi-square test. Univariate and multivariate linear regressions were performed with CPET parameters as dependent variables and NAC exposure (dichotomous) as the main independent variable. Potential confounders were added hierarchically into the models.
Model 1 was unadjusted. Model 2 was adjusted for anthropometric variables (age, sex, height, and weight). Model 3 controlled for comorbidities (cardiovascular disease, hypertension, and FEV1 as proxy of chronic obstructive pulmonary disease). Model 4 was further adjusted for hemoglobin, whereas Model 5 was further adjusted for DLCO.
A supplementary analysis including as comparators all NAC− lung cancer subjects (stages I to III instead of stage III only) was performed. Statistical analyses were performed with STATA IC 11 (StataCorp, College Station, Texas).
Results
Between January 2011 and February 2014, we enrolled 34 non-small-cell lung cancer subjects from the University Hospitals of Geneva (n = 24) and Lausanne (n = 5) and the Valais Hospital (n = 5). Subjects' characteristics are shown in Table 1. NAC+ subjects had either stage IIIA non-small-cell lung cancer (n = 13) or stage IIIB non-small-cell lung cancer (n = 6). Overall, NAC+ subjects tended to be younger and had lower hemoglobin and DLCO but similar percent-of-predicted FEV1 and NT-proBNP compared with those without NAC.
Subjects' Baseline Characteristics
Chemotherapeutic regimens were platinum-based for all subjects (cisplatin n = 16; carboplatin n = 3) combined with docetaxel (n = 16), paclitaxel (n = 1), gemcitabin (n = 1), or vinorelbine (n = 1). Two subjects received a combined treatment of cisplatin-docetaxel and cetuximab. The median number of cycles was 3 (range 1–4). Mean time between first cycle of NAC and CPET was 76 d (95% CI 62–90).
Table 2 compares the main CPET results between non-small-cell lung cancer IIIA and IIIB subjects with and without NAC. Peak V̇O2, ventilatory threshold, and oxygen pulse at peak were lower in the NAC+ group as compared with control groups. Maximal heart rate reserve at peak (expressed in percent of maximal predicted heart rate) was similar in both groups (12% ± 14% in NAC−; 11% ± 11% in NAC+). Table 3 displays the differences in CPET parameters in subjects with versus without NAC as obtained from the multivariate analysis.
Cardiopulmonary Exercise Test Results
Adjusted Differences From Multivariate Analyses in Cardiopulmonary Exercise Test Parameters of Subjects With Stage IIIA/IIIB Non-Small-Cell Lung Cancer Receiving Neoadjuvant Chemotherapy Versus Subjects Not Receiving Neoadjuvant Chemotherapy
Peak V̇O2 was significantly lower in NAC+ subjects in multivariate linear regression models when compared with NAC− subjects. Lower peak V̇O2 associated with NAC was also found in Models 2 and 3 controlling for anthropometric variables and comorbidities, respectively (−5.3 mL/min/kg (95% CI −8.3 to −2.2), P = .001 and −5.2 mL/min/kg (95% CI −8.2 to −2.3), P = .001). The impact on peak V̇O2 was less pronounced in Models 4 and 5, which were additionally adjusted for hemoglobin and DLCO, respectively, suggesting a role for these 2 variables in exercise capacity. Maximal work load was lower in the NAC+ subjects in Models 2 and 3, and ventilatory threshold was lower in Models 1–3.
The relations between peak V̇O2 and DLCO, against peak heart rate, are depicted in supplementary Figures 1 and 2. We observed no statistically significant differences between groups for V̇E/V̇O2, V̇E/V̇CO2, breathing reserve, and heart rate. By contrast, the slope of the V̇O2/heart rate relationship was flatter in NAC+ subjects (peak V̇O2 = 9.3 ± 0.5, r2 = 0.09, P = .20) compared with the NAC− subjects (peak V̇O2 = −1.1 ± 0.16, r2 = 0.31, P = .035) (supplementary Fig. 2). Supplementary Figures 3 and 4 show similar relations between VO2 and DLCO against heart rate, based on the sensitivity analysis, including all cancer subjects as controls. We also noted a lower oxygen pulse at peak exercise in subjects with NAC in all multivariate models (see Table 3).
Discussion
As hypothesized, we found that subjects with non-small-cell lung cancer stage IIIA and IIIB who received NAC had a 25.8% lower preoperative peak V̇O2 (−5.3 mL/min/kg) compared with those not receiving NAC, after adjustment for anthropometric variables, comorbidities, and FEV1. NAC+ subjects also reached a lower maximal work load, had a lower oxygen pulse at peak exercise, and had a decreased ventilatory threshold, whereas breathing reserve, heart rate at peak exercise, and ventilatory equivalents at ventilatory threshold were similar in both groups. FEV1 and NT-proBNP were comparable in NAC+ and NAC− subjects, but DLCO and hemoglobin level were lower in the NAC+ group. DLCO and, to a somewhat lesser extent, hemoglobin were associated with exercise capacity. The finding of a flatter V̇O2/heart rate regression slope and a lower peak oxygen pulse in the NAC+ group compared with the control groups suggests a cardiac component to their lower aerobic capacity (see Table 2 and supplementary Fig. 2).
Three different mechanisms may explain the lower aerobic capacity in the non-small-cell lung cancer subjects receiving NAC. First the reduction in DLCO, although nonspecific, suggests impaired gas exchange limiting exercise capacity. Second, reduced blood hemoglobin concentration limits oxygen carrying capacity. Third, the data presented are consistent with a lower cardiac stroke volume and therefore cardiac output, limiting oxygen delivery. We cannot exclude a peripheral component involved in the decrease in peak V̇O2, such as muscle (mitochondrial) toxicity of NAC, but we have no data to support this hypothesis.
Diffusion Capacity
In our study, DLCO was 15 points lower in NAC+ compared with NAC− subjects. DLCO is an important determinant of aerobic capacity. For example, in COPD patients, DLCO is a strong predictor of peak V̇O2.23 A decrease in DLCO is a common finding when performing lung function tests during or after chemotherapy. Lakoski et al9 reviewed the effects of chemotherapeutic agents on components of oxygen transport and suggested that platinum-based chemotherapy impacts its pulmonary component, presumably by diminishing diffusion capacity. Pulmonary toxicity was also reported for taxanes and gemcitabine. Kreuter et al24 assessed lung function after 2 different adjuvant chemotherapeutic regimens for early stage non-small-cell lung cancer and found a decline of DLCO after a cisplatin-vinorelbine regimen (−8%) but not after cisplatin-pemetrexed (−0.4%). This is consistent with the results of Rivera et al25 who found an 8% DLCO decline after 3 courses of gemcitabine plus carboplatin, cisplatin, or paclitaxel for stage I and II non-small-cell lung cancer. Leo et al26 compared pre- and post-NAC (3 courses of cisplatin-gemcitabine) lung function in 30 subjects with stage IIIA-N2 non-small-cell lung cancer. They found an 11% DLCO reduction after chemotherapy. Takeda et al27 found a 21% reduction in DLCO after induction therapy (chemotherapy alone or chemo-radiotherapy) for stage IIB (Pancoast), IIIA, and IIIB non-small-cell lung cancer. Importantly, DLCO was an independent predictor of postoperative pulmonary morbidity.28
Oxygen Transport
From the Fick principle, we know that a reduction in blood hemoglobin content will result in a reduction of peak V̇O2.29 Virtually all chemotherapeutic agents may cause anemia through different pathways (deficient production or destruction of red blood cells, excessive blood loss). Anemia during chemotherapy occurs in 30–100% of patients.30 A correlation between hemoglobin levels and peak V̇O2 reduction was described for women receiving adjuvant chemotherapy for breast cancer.31 In our study, adjustment for hemoglobin marginally altered the association between NAC and peak V̇O2, suggesting that the effect of NAC on peak V̇O2 in our cohort was mainly mediated by other mechanisms.
Cardiotoxicity
Several chemotherapeutic agents are cardiotoxic. Anthracylines, trastuzumab, and tyrosine kinase inhibitors are frequently associated with left ventricular dysfunction and impairment in cardiovascular autonomic regulation.32,33 Antimicrotubule agents such as paclitaxel are reported to induce various cardiac problems, including bradyarrhythmias, atrioventricular conduction blocks, bundle branch blocks, and cardiac ischemia. Our observation of a flattening of the peak V̇O2/heart rate relationship and a lower peak oxygen pulse is compatible with a cardio-toxic component of NAC, although simple deconditioning cannot be excluded.
Lung Volumes
In our study, NAC was not associated with lower lung volumes. This is in line with published literature showing similar total lung capacity, FVC, and FEV1 in subjects exposed to different combinations of chemotherapeutic agents, mainly platinum-based doublet. Kreuter et al24 and Rivera et al25 reported no decrease of FEV1, FVC, or total lung capacity. Two studies even reported an improvement in lung volumes after NAC.24,26
Limitations
Several limitations should be taken into account when interpreting our results. First, direct toxicity on pulmonary, circulatory, neural, and muscular components cannot be distinguished from the indirect effects of chemotherapeutic agents such as deconditioning or changes in body weight and composition due to physical inactivity. Second, our study was not designed to determine the relationship between NAC and short- or long-term surgical therapeutic outcome, precluding any conclusions on this specific point. Finally, we did not perform CPET before starting NAC to measure longitudinally its impact on CPET results. Observations were made on the basis of the CPET post-NAC only. Our cross-sectional design thus limits any strong conclusions on causality, although collectively, the literature convincingly suggests such a causal relationship. We find that the potential limitation of peak V̇O2 after NAC should nevertheless be taken into account in the still ongoing debate of choosing preoperative versus postoperative chemotherapy. Research should specifically address the effects of NAC on postoperative complications and the role of CPET as a risk-stratifying tool.
Implications for Clinical Decision and Research
Clinicians in charge of multimodal therapy in non-small-cell lung cancer should carefully weigh the risks against the potential benefits of NAC, since it may contribute to decreasing fitness level and thus impact the short- and long-term outcome of primary tumor resection. Patients should be advised to maintain or even increase their level of fitness when started on NAC. In this perspective, randomized controlled trials suggests benefits from preoperative high intensity training in subjects with non-small-cell lung cancer bring new therapeutic modalities when addressing surgical risk for lung cancer patients.34,35 However, exercise training in patients with NAC may represent a significant challenge.
Conclusion
NAC was associated with a reduced exercise capacity in potentially resectable stage IIIA and IIIB non-small-cell lung cancer subjects. Impaired gas exchange (reduced DLCO), decreased oxygen transport (anemia), cardiotoxicity, and physical deconditioning are plausible mechanisms involved. Because peak V̇O2 is a major predictive factor for peri- and postoperative complications and of short-term survival, NAC could potentially influence these outcomes. Randomized controlled trials should be conducted to assess whether the reduced exercise capacity related to NAC influences outcome and whether it could be improved by concurrent physical rehabilitation.
Acknowledgments
We thank Elise Dupuis for statistical assistance.
Footnotes
- Correspondence: Isabelle Fresard, University Hospitals of Geneva, Division of Pulmonary Medicine, Hopital de La Tour, Av J.-D. Maillard 3, 1217 Meyrin, Switzerland. E-mail: isabelle.fresard{at}latour.ch.
This work was supported by the University of Geneva, the University Hospitals of Geneva, and the Ligue Genevoise contre le Cancer. The authors have disclosed no conflicts of interest.
Supplementary material related to this paper is available at http://www.rcjournal.com.
- Copyright © 2016 by Daedalus Enterprises
