Exercise and cancer: a position statement from the Spanish Society of Medical Oncology

Regular and adequate physical activity (PA) is associated with key benefits to human health, such as improvements in weight control, muscular and cardiorespiratory fitness, bone and functional health and a reduced risk of falls and several noncommunicable diseases, including diabetes, cardiovascular disease, depression and some cancers [1].

Due to improvements in the management of cancer, the number of cancer survivors and survival time are increasing. Consequently, interest in healthy behaviors, such as PA, and their potential impact on cancer- and non-cancer-related morbidity in these individuals has rapidly increased [2].

Commissioned by the Spanish Society of Medical Oncology (SEOM), in this review, we sought to distill the most recent evidence on this topic, focusing on the mechanisms that underpin the effects of PA on cancer, describing the role of PA in cancer prevention and prognosis, and providing practical recommendations to clinicians on managing PA counseling.

Of note, PA is any bodily movement produced by skeletal muscles that requires energy expenditure, whereas exercise is a subset of PA that is planned, structured and repetitive and that has a final or an intermediate objective of improving or maintaining physical fitness. The epidemiological evidence regarding the risk of some cancers mainly refers to regular PA (usually self-reported, i.e., through questionnaires). Thus, regular exercise or “exercise training” is a proxy, but not a perfect surrogate, for PA and is thought to induce more profound molecular adaptations than PA.

There is growing evidence from preclinical research that regular exercise can influence cancer development or the rate of tumor growth once malignancy has initiated. For instance, a recent meta-analysis of 28 preclinical studies in breast tumors (n = 2085 animals) found large favorable effects for exercise training on proliferation and apoptosis [3]. Exercise is also emerging as a potential coadjuvant treatment; when combined with cyclophosphamide, exercise delays murine breast tumor growth versus chemotherapy alone [4], and similar findings have been reported for exercise combined with anthracyclines [5‐7]. There is, however, heterogeneity among studies in the tumor models used, ranging from tumor transplant (where there is also substantial variability, e.g., syngeneic versus xenograft models) to carcinogen-induced or genetically engineered mouse models [8]. The type of exercise training to which mice are typically subjected to before/upon tumor inoculation also varies between studies (forced treadmill, forced swimming, voluntary wheel running). The duration of exercise in the trials typically consists of several weeks (~ 4 to 10 weeks), which can be translated to human “years.”

Exercise can have an impact on tumor development, growth or dissemination through several mechanisms. First, exercise might help to prevent cancer by reducing the circulating levels of several mediators, such as insulin growth factor-1 (IGF-1) [9‐14], a mitogen that triggers cell proliferation [15]. Exercise can also reduce the levels of hyperphosphorylated retinoblastoma protein (Rb) in a chemically induced rat model of mammary carcinogenesis [16, 17], increase ß-catenin phosphorylation in colon polyps [18, 19], and reduce the levels of micro-RNA 21 [20].

Exercise can upregulate tumor suppressors, such as the tumor suppressor programmed cell death protein 4 in a murine model of estrogen receptor-positive breast cancer (BC) [21]. In addition, exercise-induced catecholamines might reduce BC development through activation of the Hippo tumor suppressor pathway [22] and exercise-induced increased p53 activation, leading to tumor prevention, as shown in mouse models of skin [13, 23] and lung [24] cancer.

Exercise training can stimulate apoptosis, as shown in xenograft models of lung adenocarcinoma [24] and human pancreatic and prostate cancers [25, 26] and in murine models of skin tumorigenesis [12] and mammary carcinoma [4, 16, 27]. Exercise also exerts proapoptotic effects on cultured prostate cancer cells [28], estrogen receptor-positive BC cells [29, 30] and lymph node metastases of prostate tumor cells [25]. Exercise additionally reduces the levels of the antiapoptotic protein B cell lymphoma 2 [16, 31] and stimulates the proapoptotic proteins Bax and Bak [4, 16, 24] and the protein kinase AMPK [32, 33].

Hypoxia and poor blood supply promote an aggressive cancer phenotype and contribute to ineffective systemic therapy [34]. In this respect, exercise may promote a shift toward a more “normalized” tumor microenvironment by improving intratumoral perfusion/vascularization, at least in orthotopic murine models of human BC [4, 35] and prostate cancer [36‐38] and in xenografts of different tumors (melanoma, pancreas) [39].

Exercise might also attenuate the development of metastases. Mouse exercise training can decrease catenin while increasing E-cadherin inside tumors [19, 40]. Importantly, cadherins act as glue between epithelial cells, and their loss can favor malignancy by allowing the disaggregation of cells, which can then invade locally or metastasize [40]. Moderate-intensity mouse swimming can suppress liver cancer metastases via boosting the activity of dopamine receptor 2 [40]. Exercise may also modulate blood–brain barrier integrity by maintaining the expression levels of occludin or claudin-5 proteins [41], thereby preventing metastatic progression to the brain [42]. On the other hand, inflammatory cells within the tumor microenvironment supply bioactive molecules that sustain cancer hallmarks [43‐45]. In this context, mouse exercise training decreases macrophage infiltration in allogeneic lymphoid tumors [46], Ehrlich tumor cells [47] and colon polyps [18].

One major potential “anticancer” effect of exercise lies in an enhancement of immune function [43, 48]. At moderate intensities, exercise can stimulate the innate immune system, especially natural killer (NK) cells [49, 50]. A 6-week mouse wheel running program had preventive effects against the development of several types of tumors (melanoma, liver and lung cancers), and the effect was mediated by improved NK cell infiltration into the tumors, which in turn was mediated by the enhanced tumor expression of ligands for several NK cell-activating receptors [51]. A previous study showed that exercise training increased the cytolytic capacity of resident peritoneal macrophages against mastocytoma cells [52]. Mouse exercise training could also polarize the immunological response toward an efficient “antitumor” macrophage profile 1, which is linked to the production of T-helper 1 cytokines [52‐56]. Short-term (6-day) moderate exercise before the injection of melanoma cells into mice decreased their metastatic spread, which was partially mediated by increased antitumor macrophage cytotoxicity [57]. Preliminary data from mice [53, 58‐60] and cancer patients suggest that exercise training may help to reduce the immunosuppressive effects of T regulatory lymphocytes [61]. Finally, regular exercise can increase alveolar macrophage antitumor cytotoxicity in vitro, which would mediate a protective effect against mouse lung metastases [62, 63].

Importantly, skeletal muscle, especially during contractions, releases molecules collectively known as “myokines” into the bloodstream, which act systemically to induce a myriad of health-promoting effects, such as decreased inflammation and reduced insulin resistance [64]. Some myokines might also induce direct anticancer effects (via the stimulation of apoptosis in tumor cells), such as oncostatin M in hormone-sensitive BC [30] or secreted protein acidic and rich in cysteine (SPARC, also known as osteonectin) in colon cancer [65]. The aforementioned exercise-induced infiltration of NK cells into tumors seems to be mediated by the release of interleukin 6 by muscle into the bloodstream [49, 51, 66].

According to the World Health Organization (WHO), up to 31% of the adult population worldwide and 35% in Europe are physically inactive [67]. PA is difficult to measure for the following reasons: (1) there are at least four domains: occupational, household, transportation and leisure time; (2) PA questionnaires on past and current activity are subject to recall bias; and (3) objective methods (pedometers or accelerometers) can only be used in prospective studies for short time periods, and they may not always represent overall PA. Fortunately, smartphones and other devices now allow PA monitoring and will hopefully provide more accurate measures in the future.

Body mass is related to PA and cancer risk, acting as a confounder. However, the prevention of adiposity may mediate the relationship between PA and cancer, and controlling for adiposity could lead to underestimating the real effect of PA [68].

Of note, 1 metabolic equivalent (MET) is the rate of energy expenditure while resting or 3.5 ml O₂/kg body weight/min on average. Moderate PA (e.g., brisk walking) usually requires an energy expenditure of 3–6 MET, whereas vigorous PA (e.g., jogging) requires an energy expenditure above 6 MET. The WHO recommends that adults engage in ≥ 150 min/week of moderate PA or ≥ 75 min/week of vigorous PA or a combination thereof. If a person does an ~ 3 MET activity (e.g., brisk walking on a level surface) for 1 h, he or she has done 3 MET-hours of PA. If this person does this same PA on every day of the week, he/she has done 21 (= 3 × 7) MET-hours/week. If a person does an ~ 8 MET activity (e.g., jogging) for 1 h on each day of the week, he/she has done 56 (= 8 × 7) MET-hours/week.

The World Cancer Research Fund and the American Institute for Cancer Research periodically publish the conclusions of a panel reviewing evidence linking food, nutrition and PA with cancer risk [69]. The evidence is classified as follows: (1) convincing: available results support a causal relationship; (2) probable: evidence supports a probable causal relationship; and (3) limited: results are not considered sufficient to rate the relationship as convincing or probable. In the last category, a distinction is made between limited-suggestive evidence when an effect is reported but there were methodological problems and limited-not conclusive evidence when there were insufficient data and/or the results were too heterogeneous. The panel concluded that regular, sustained PA protects against several types of cancer independent of body fat [69]. This evidence comes from high-income countries and is mainly based on leisure-time PA. The three tumors with the most solid results are colon, postmenopausal BC and endometrial.

Several reviews and meta-analyses of observational studies have suggested the benefit of PA on cancer outcomes. Most of the studies included breast cancer (Tables 1, 2) and colon cancer survivors (Table 3). A few studies have been conducted on patients with other types of neoplasms, such as prostate (Table 4), esophageal, lung and kidney cancer (Table 5). In these studies, PA is reported as lifetime PA in the latest years before or after diagnosis. The outcomes reported are usually overall survival, cancer-related survival, cancer recurrence and quality of life (QoL).

Epidemiologic and observational studies show a decrease in the risk of cancer recurrence and all-cause mortality in patients who practice regular PA [121‐123]. A systematic review of studies published through June 2013 concluded that PA performed before or after cancer diagnosis is associated with a reduced mortality risk among BC and colorectal cancer survivors [124]. Mortality in adult survivors of childhood cancer was inferior in those patients who practiced vigorous exercise after diagnosis in a large multicentric observational study [125]. In 2015, Lahart et al. [97] published a meta-analysis of 22 studies analyzing the impact of PA on BC outcomes. A literature search was performed using PubMed, EMBASE and CENTRAL databases from 1995 to October 2014. In 40% of the observational studies, the risk of relapse and death in BC survivors decreased in most physically active women. Most studies included an analysis of leisure-time PA and only a few of interventional PA programs. The majority of the studies did not perform a multivariable analysis to exclude the effect of known confounding factors, and less than half included clinical prognostic factors, such as stage, nodal status, age or type of treatment. These studies found a positive impact on all-cause and BC mortality in patients who practiced moderate or intense lifetime PA before the diagnosis of BC and in recent years before diagnosis. However, the authors recommend interpreting these results with caution due to the large heterogeneity of the studies. A post-diagnosis activity of at least 10 MET-hours/week was associated with a decrease in all-cause, and BC mortality and was not influenced by the heterogeneity; however, not all the studies could corroborate a decrease in recurrence risk. The “After Breast Cancer Pooling Project” included more than 13,000 women from four prospective cohorts of BC survivors in the USA and Shanghai and analyzed the association between PA at 18–48 months after diagnosis and risk of all-cause and BC-specific mortality and BC recurrence [126]. BC mortality was reduced in patients who achieved 18.7 or more MET-hours/week, and no association was found between PA and BC recurrence [126]. A comprehensive review of sixty-three interventional studies on women after BC adjuvant therapy concluded that PA interventions might have certain beneficial effects on QoL, cardiorespiratory fitness and psychological and social functions, but conclusions about BC recurrence, BC mortality and all-cause mortality could not be made [98].

Several prospective observational studies and meta-analyses in patients with colorectal cancer have suggested the benefit of PA before and after diagnosis in terms of improvements on all-cause and cancer-specific mortality after controlling for other confounding factors, such as BMI, sex, number of positive lymph nodes, age, baseline performance status (PS), adjuvant chemotherapy regimen or recurrence-free survival period [110, 111, 127, 128]. Similarly, three observational prospective studies in prostate cancer found a strong inverse relationship between exercise and the risk of cancer progression regardless of other known prognostic factors [115‐117]. A Chinese study in patients who underwent esophagectomy for esophageal cancer supported the benefit of PA (> 9 MET-hours/week) on recurrence risk and all-cause mortality [120]. Data from a prospective observational study in kidney cancer survivors investigating PA and diet changes suggested a decrease in the recurrence rate in patients who did any PA compared with those that were totally inactive [118]. The only study in lung cancer survivors showed better overall survival in patients who met > 9 MET-hours/week, but no difference in the recurrence rate was observed [119].

In conclusion, the real impact of PA on the risk of relapse and cancer mortality is not well-defined. PA may contribute to reduced cancer-related mortality and all-cause mortality in cancer survivors by modifying fat accumulation and improving cardiovascular and skeletal muscle function [129]. Numerous prospective observational studies consistently showed the benefit of PA on cancer outcomes; however, most of these studies were based on measures from self-reported questionnaires, including heterogeneous populations, and only a few performed a multivariable analysis to exclude the contribution of other confounding factors. Interventional studies with reliable and objective measures of PA in homogeneous populations are needed to confirm the data from observational studies and to evaluate the real effect of exercise on cancer prognosis.

Exercise-oncology is a new field of cancer care with the goal of the appropriate and rationale introduction of exercise programs into the overall management of cancer patients to take advantage of the numerous benefits associated with PA. Several major comprehensive cancer centers have created exercise-oncology units to implement these programs in a timely and organized manner. A collaborative work among rehab specialists, physiotherapists and exercise physiologists, as well as oncologists and radio-oncology specialists, is developed in these units.

Exercise has demonstrated numerous benefits on the QoL of patients with cancer throughout the history of the disease, ameliorating the negative impact of cancer on physical and psychological health and having a positive impact on patient survival [130‐133].

Despite these benefits, many questions about PA/exercise in cancer patients remain, as it is particularly challenging to elucidate how much exercise is needed to achieve patient improvements and how the exercise should be recommended and monitored by clinicians.

Regular PA is associated with major benefits to human health, including a reduced risk of some cancers.

The mechanisms through which exercise exerts its antitumor activity are still poorly understood but might be related to a direct effect on tumor cells (inhibition of tumor cell proliferation, induction of apoptosis, upregulation of tumor suppressor genes, anti-inflammatory effects) or to an enhancement of immune function.

There is convincing evidence that regular PA reduces the risk of colorectal cancer, while the reduction in postmenopausal BC and endometrial cancer risk is judged as probable. The effect of PA on the risk of other tumors is less evident but still possible.

Several epidemiological studies have suggested an association of regular PA with reduced cancer-related and all-cause mortality in some tumor types, particularly BC and colorectal cancer. The minimum amount of PA needed to achieve such a benefit is still unknown, although the US recommendations suggest that a minimum 10 MET-hours/week (equivalent to ≥ 150 min of moderate-intensity PA) is needed.

Exercise-oncology is a field of cancer care in which the goal is the introduction of exercise programs into the overall management of cancer patients. The first exercise guidelines for cancer patients were published in 2010 by the ACSM. These guidelines, which are mainly based on general WHO guidelines to the general population, consider that regular PA in cancer patients is safe and exerts positive effects in patients at multiple levels, particularly QoL. Exercise programs in cancer patients are feasible along the course of the disease, including the presurgical period, during adjuvant antitumor medical treatment (including chemotherapy) and in cancer survivors; a summary of these recommendations is shown in Table 7. However, the experience with regular exercise in metastatic cancer patients is limited.

The introduction of exercise programs into the global management of cancer patients remains a challenge due to conceptual and logistic issues. The most effective behavioral interventions to achieve long-term changes in a patient’s lifestyle must be defined. New technologies, such as mobile health applications and wrist and watch bands (the so-called “mHealth”), can be of great help to monitor the compliance to these programs. The optimal intensity and duration of PA should be defined with more precision in future studies. Regarding logistics, the intervention of both exercise-oncology specialists and trained clinicians is probably necessary at different time points to provide the best care. Several major comprehensive cancer centers have created exercise-oncology units to implement these programs in a timely and organized manner, and these models could serve as a reference for other institutions.