Exercise-based rehabilitation on functionality and quality of life in head and neck cancer survivors. A systematic review and meta-analysis

Head and Neck Cancer (HNC) is a globally rare cancer that includes a variety of tumors affecting the upper aerodigestive tract. It presents with difficulty breathing or swallowing and is mainly treated with radiation therapy, chemotherapy, or surgery for tumors that have spread locally or throughout the body. Alternatively, exercise can be used during cancer treatment to improve function, including pain relief, increase range of motion and muscle strength, and reduce cancer-related fatigue, thereby enhancing quality of life. Although existing evidence suggests the adjunctive use of exercise in other cancer types, no previous studies have examined the effects on HNC survivors. The aim of this meta-analysis was to quantify the effect of exercise-based rehabilitation on functionality and quality of life in HNC survivors who underwent surgery and/or chemoradiotherapy. A systematic review and meta-analysis were carried out following PRISMA statement and registered in PROSPERO (CRD42023390300). The search was performed in MEDLINE (PubMED), Cochrane Library, CINAHL and Web of Science (WOS) databases from inception to 31st December 2022 using the terms “cancer”, “head and neck neoplasms”, “exercise”, “rehabilitation”, “complications”, “muscle contraction”, “muscle stretching exercises” combining with booleans “AND”/“OR”. PEDro scale, Cochrane Risk of Bias Tool and GRADE were used to assess methodological quality, risk of bias and grade of recommendation of included studies respectively. 18 studies (n = 1322) were finally included which 1039 (78.6%) were men and 283 (21.4%) were women. In patients who underwent radio-chemotherapy, overall pain [SMD = − 0.62 [− 4.07, 2.83] CI 95%, Z = 0.35, p = 0.72] and OP [SMD = − 0.07 [− 0.62, 0.48] CI 95%, Z = 0.25, p = 0.81] were slightly reduced with exercise in comparison to controls. Besides, lower limb muscle strength [SMD = − 0.10 [− 1.52, 1.32] CI 95%, Z = 0.14, p = 0.89] and fatigue [SMD = − 0.51 [− 0.97, − 0.057] CI 95%, Z = 2.15, p < 0.01] were also improved in those who receive radio-chemoradiation. In HNC survivors treated with neck dissection surgery, exercise was superior to controls in overall pain [SMD = − 1.04 [− 3.31, 1.23] CI 95%, Z = 0.90, p = 0.37] and, in mid-term, on shoulder pain SMD = − 2.81 [− 7.06, 1.43] CI 95%, Z = 1.76, p = 0.08]. No differences in quality of life were found at any of the follow-up periods. There is evidence of fair to good methodological quality, low to moderate risk of bias, and weak recommendations supporting the use of exercise-based rehabilitation to increase functionality. However, no evidence was found in favor of the use of this modality for improving the quality of life of HNC survivors who underwent chemoradiotherapy or surgery.

GBM), and in case of disagreement, a third author (NKY) was the responsible of resolving discrepancies. A standardized work template based on PICO question was used to extract and detail all the information related to authors, year and country of publication, study design, aim of the study, outcomes, participants (characteristic of disease, medical intervention, sample size, gender distribution, etc.), intervention and control details, results of measured outcomes and conclusions. The Cochrane Handbook for Systematic Reviews of Interventions-v.5.1.0 was used to develop these sections. The reliability of the table was tested using a representative sample of the studies to be reviewed.
Risk of bias analysis. Risk of bias analyzes of randomized clinical trials were independently performed by SMP using the Cochrane Risk of Bias Tool for Randomized Trials (RoB 2.0) 47 . The tool evaluates the methods researchers use in clinical trial design and individually rates the presence of the following biases: (1) randomization process, (2) deviations from the intended interventions, (3) missing outcome data, (4) measurement of the outcome and (5) selection of the reported result. Interpretation of the scores obtained considers the fact that a low risk of bias means that the bias committed is unlikely to significantly alter the results, whereas a high risk of bias indicates lower confidence in the results receive. Any disagreement of the authors was resolved by discussion, and in case of conflicting scores, the third reviewer (FRH) resolved to make the decision. Data synthesis. Meta-analyses were undertaken using Review Manager (RevMan v.5.3; Cochrane Collaboration, Oxford, UK) when more than two studies reported on the same outcome. In the pooled analysis of studies by duration, outcome data were organized into short-term (≤ 6 weeks), medium-term (7-23 weeks) and long-term (≥ 24 weeks) according to previous studies 49 . In cases where it is not possible to convert the units of measurement, the standardized mean difference is used. Data are presented as standard mean differences (SMD) and 95% CIs. The I 2 statistic is used to quantify statistical heterogeneity as follows: 0-40%, probably not important; 30-60%, moderate heterogeneity; 50-90%, significant heterogeneity; 75-100%, significant heterogeneity. Analysis was performed using a fixed-effects model, however, when statistical heterogeneity (I 2 > 40%) was found, a random-effects model was used for meta-analysis.

Results
Study selection. A total of 5,672 studies were detected and analyzed by performing the agreed searches in the detailed databases MEDLINE (PubMed) (n = 1875), Cochrane Library (n = 986), CINAHL (n = 1050) and Web of Science (WOS) (n = 1761). After removing duplicates (n = 812) and analyzing the titles and abstracts of the remaining articles (n = 4036), 824 full-text articles were potentially relevant studies according to the search strategies. Finally, 806 of these manuscripts were excluded because the studies did not meet our eligibility criteria as they had a different study design (n = 671), published in different languages (n = 5), carried out another intervention (n = 90) and have measured other outcomes of little interest for our research question (n = 40). Therefore, 18 studies were ultimately selected for this review for qualitative synthesis and meta-analysis ( Fig. 1). Methodological Quality assessment (PEDro Scale). The methodological quality of the studies included in our review was good (PEDro score = 6.38 out of 10, SD = 1.09). During methodological quality assessment, 14 studies of good quality [50][51][52][53][55][56][57][58][59][60]62,63,65,67 and 4 studies of acceptable methodological quality were identified 54,61,64,66 . Most of the studies made systematic errors in relation to the blinding of the participants, since of the total that were included, only 1 carried out the blinding of the participants 50 . Also, the blinding of the therapists who applied treatment was only carried out in a study 54 (Table 3).  (Table 4).

Grade of recommendation (GRADE). The Grade of recommendation is weak in favor of exercise to
improve functionality and quality of life in survivors with HNC (        (Fig. 4).        55 in patients with HNC underwent chemoradiation (Fig. 5).   (Fig. 8).

Effectiveness of exercise-based rehabilitation in
Functionality related to range of motion.

Discussion
Effectiveness of exercise-based rehabilitation in HNC survivors undergone chemo-radiotherapy. The aim of this meta-analysis was to quantify the effect of exercise-based rehabilitation on functionality and quality of life in HNC survivors who underwent surgery and/or chemoradiotherapy. This modality has a positive effect on overall pain in HNC survivors undergoing chemoradiation. Bragante et al. found that RET combined with joint mobility training and stretching reduce overall pain on long-term in patients undergoing CHT or RDT compared to controls 62 . Cancer pain use to be neuropathic and can be caused both in the Central Nervous System (CNS) and the Peripheral Nervous System (PNS). During tumor growths, compression, and even invasion, can occur triggering painful sensation. CHT and RDT contributes to this neuropathic cancer pain (NCP) 68 . RDT induced pain could be attributed to fibrosis or sensitization of peripheral nerves 69 . Although the underlying effects of the exercise on NCP are not fully understood different hypotheses are postulated. One of the most plausible hypotheses is that dynamic RET produces hypoalgesia by the activation of baroreceptors induced blood pressure changes which are linked to central pain inhibition pathway 70 . Other potential mechanism is that RET increases not only delayed-onset muscle soreness (DOMS), responsible of the increase muscle pain thresholds, but also a stimulation of low-threshold motor units. Both joint mobility and stretching exercises activate centers of descending inhibitory opioid dependent axes and others non-opioids pathways 71 . Although no positive results were found in favor of exercise in patients with HNC undergoing radio-chemotherapy, Pauli et al. 64 found that RET combined with joint mobility and stretching exercises reduce OP in www.nature.com/scientificreports/ the medium term 64 . OP is a localized pain disorder in face and jaw that causes moderate to severe deterioration of chewing 72 . OP uses to be the result of a combination of several joint, myofascial and/or neuropathic mechanisms 73 . In the study of Pauli et al. 64 pain can be attributed to radio-chemotherapy as well. In this sense RDT, when it acts in the environment of the temporomandibular joint (TMJ), provokes radiation induced fibrosis that arise NCP 69,74 . RET progressively stimulates a local anti-inflammatory response and activation of central microglial activity 75 that explains the amelioration in neuropathic OP as it is observed in Pauli et al. 64 .
Despite not positive effects of exercise on upper limb muscle strength were observed in patients who received radio-chemotherapy, Capozzi et al. found that the combination of RET with health education and behavioral therapy has a positive effect on the long-term enhancement of upper limb muscles in patients receiving CHT or RDT 30,55 . All these patients with HNC suffer from sarcopenia, a condition characterized by the skeletal muscle and weight loss resulting in a decrease of physical performance.
According to some authors the prevalence of sarcopenia in patient treated by HNC is 35.5-54.5%, which is mainly attributed to weight loss and malnutrition 76,77 . Among the main mechanisms of sarcopenia include an intensification of protein catabolism probably related to tumor growths 78 .
In addition to this, CHT and RDT exacerbate this malnutrition and weight loss 79 . In some cases, CHT has a direct influence on muscle catabolism leading to a loss of muscle strength 78 , in others, the loss of muscle mass can be explained by CHT adverse effects such as fatigue, loss of appetite, nausea, vomiting or diarrhea, as a result of several reduction of food intake and physical activity 78 . When toxicities such as xerostomia, dysphagia or oral mucositis occur, a reduction of oral intake may imply malnutrition, weight loss and subsequent sarcopenia 79 . Unlike CHT, the mechanism of RDT in the onset of sarcopenia remain poorly understood. Sarcopenia is thought to decrease survival and increase disease relapses or the likelihood of death of these patients 80 .
Exercise prevents sarcopenia because it increases muscle mass and function 77,81 . Strength exercises are one of the most effective modalities to promote the upper limb functionality. However, long time exposure to exercise is required to draw muscular and neural adaptations 82 . www.nature.com/scientificreports/ Given that CHT produces malnutrition due to a decreased intake and physical performance an improvement of nutritional status is mandatory 83 . Healthy lifestyle education in HNC patients encourages the acquisition of healthy feeding behavior in the long term. It is previously shown that doing exercise reduces carbohydrates and lipids intake leading to an indirect incorporation through daily meals of proteins, oligoelements and other minerals improving muscle health 84 .
Exercise-based rehabilitation increases lower limb strength in patients with HNC who had received CHT or RDT. In this sense, Lønbro et al. 58 found that RET have a positive effect on the potentiation of lower limb muscle strength in the medium term in patients receiving CHT or RDT. As describe above a reduction of muscle strength as a result of cancer-related sarcopenia or the administration of CHT and RDT agents affect muscles equally. Thus, a RET program in the lower limb promotes the knee extension peak of force through an increase of hamstring complex muscle mass in the medium term. In this case, muscle mass changes could be associated to the fact, on the one hand, that lower limb musculature allows higher force loads compared to upper limb, and on the other hand, the higher frequency and longer duration program carried out as it has been formerly published in other populations 85,86 .
Rogers et al. 53 also found that RET has a positive effect on the enhancement of lower limb muscle strength in the medium term in patients receiving CHT or RDT. For his part, Zhao et al. 56 found that in HNC patients receiving CHT or RDT, the combination of RET with AET has a positive effect on the enhancement of lower limb muscle strength in the medium term. Authors differ on the effects of RET combined with AET. For some of them, hypertrophy produced by AET depends on modality, frequency, and intensity so that the higher they are www.nature.com/scientificreports/ the more hypertrophy there is. For many others, when they are combined, the effects can be inhibited by each other 87,88 . Nevertheless, more clinical trials are needed to delve deeper into this question. Flexibility exercises within a multimodal program also impact on the increase of muscle maximum strength. In this sense, Lin et al. 50 found when combining RET with AET flexibility exercises has a positive effect on the enhancement of lower limb muscle strength in the medium term in patients receiving CHT and RDT. Despite the discrepancies between some studies, recent works have shown that chronic exposure to elasticity exercises can contribute to an increase in maximum contraction force due to changes in both musculotendinous stiffness and neural impulse. It is probably due to changes in the muscle tension-length relationship and the rate of sarcomere shortening. In the case of Lin et al. 50 active stretching exercises in combination with maximal strength exercises seem to be more effective than when they are performed with explosive force exercise, currently, not supported by evidence 89,90 .
Exercise-based rehabilitation was successful in reducing disease-related fatigue in patients with HNC. Samuel et al. found that the combination of RET with AET has a positive effect on short-term fatigue in patients receiving CHT and RDT 51 . Fatigue occurs both in the active phase of the disease and during the treatment phases producing a significant impairment of the functionality of the HNC survivor 91 . This disorder is defined as a distressing, persistent, and subjective feeling of physical, emotional, and/or cognitive tiredness or exhaustion related to cancer or cancer treatment non-proportional to activity that interferes with functioning 92 . The prevalence of fatigue in oncologic population ranges from 26.2 to 56.3% depending on the study 93 , but, in the HNC is moderate to severe only in 18% of patients 94 . www.nature.com/scientificreports/ Fatigue can be caused by a set of mechanism whose etiologies are not clearly elucidated. In general, it can be attributed to both factors related to the CNS (neuroinflammation, hypothalamic pituitary adrenal (HPA) axis, etc.) and the PNS (cachexia, alteration of energy metabolism, etc.). It is central when the patient is not able to perfom physical or mental task without major cognitive or motor impairment 92 . Fatigue is peripheral if the patient's muscles are not able to perform a task after being stimulated or there is a reduction in endurance 92 .
Both cancer and its treatment activate the inflammation cascade increasing cytokines releasing that act on the CNS. Other potential mechanism is the alteration of the HPA axis that is induced by cytokines as part of this neuroinflammation. Activation of the HPA axis increases cortisol releasing to limits systemic damages due to this inflammatory state. This will lead to physical fatigue, circadian cycle disturbances and lack of sleep 92 .
Fatigue in these patients is a result of loss of physical condition, combined with radiation and chemotherapy treatments, resulting in changes in anaerobic metabolism that lower lactate thresholds leading to fatigue and weakness. In addition, loss of appetite and the presence of nausea and vomiting can lead to caloric and nutrient deficiencies.
Fatigue-related functions may also occur at the peripheral level due to changes in energy metabolism. Not only from cachexia, but also from sarcoplasmic reticulum and/or mitochondrial damage caused by CHT and RDT, skeletal muscle may be compromised, increasing fatigue more than random or milky, depending on the primary fuel consumed 92 . The lactic acid system converts glucose into cellular energy nucleotides (AMP, ADP, ATP), which generate lactic acid as a by-product, and when lactic acid accumulates in the muscles, it produces fatigue and the patient manifests weakness, which is an illness with symptoms caused by CHT or RDT can exacerbate anorexia, nausea, or vomiting, ultimately reducing caloric intake and thus the availability of energy substrates. The lactate pathway was the predominant pathway activated when the various RETs of the program of Samuel et al. performed on HNC patients who had undergone CHT or RDT. This improvement could be explained by exercise shifting the lactate threshold to the right, leading to increased lactate tolerance in cancer patients undergoing CHT or RDT (whose threshold is generally lower) by preventing the onset of delayed fatigue 91 .
Fatigue-related functions may also occur at the peripheral level due to changes in energy metabolism. Not only from cachexia, but also from sarcoplasmic reticulum and/or mitochondrial damage caused by CHT and RDT, skeletal muscle may be compromised, increasing fatigue more than random or milky, depending on the primary fuel consumed 92 . The lactic system generates by-product lactic acid that when accumulates in the muscle generates fatigue that the patient manifests as weakness, a symptom that can worsen by CHT or RDT which lack of appetite, nausea or vomiting that end up reducing caloric intake and consequently decreasing the availability energy substrates. The lactate pathway was the predominant pathway activated when the various RETs of the program of Samuel et al. 51 performed on HNC patients who had undergone CHT or RDT. This improvement could be explained by exercise shifting the lactate threshold to the right, leading to increased lactate tolerance in cancer patients undergoing CHT or RDT (whose threshold is generally lower) by preventing the onset of delayed fatigue 91 .
Effectiveness of exercise-based rehabilitation in HNC survivors undergone surgery. Exercisebased rehabilitation has a positive effect on overall pain in HNC survivors undergoing surgery. In this sense, Loh et al. 65 found that relaxation exercises have a positive short-term effect on reducing overall pain in patients undergoing head and neck surgery. As noted, apart from CHT and RDT, HNC treatment also includes surgery which appear to be effective as in early as advanced stages of the disease 18 .
Among surgical techniques, modified radical neck dissection (MRND), selective dissection of the lymph nodes of the neck, with preservation of the accessory nerve or resection of cervical nerve root branches and subsequent flap reconstruction are the most commonly used procedures [95][96][97] .
Neuromuscular complications due to surgical treatment are more frequent in HNC survivors compared to those who only received CHT or RDT. When cancer invades laryngopharyngeal region, musculoskeletal structures related to speech and swallowing (i.e. dysphagia, hoarseness, etc.) need to be removed 97,98 . Although the surgical procedure is safe, some neuromuscular complications due to sacrifice of the accessory nerve have been reported in cases related to cancers or inadvertent injury to the nerve during surgery. As a consequence, an alteration of the strength and motor coordination of the cervico-scapular musculature could develop, which could alter the functionality of the upper limb [99][100][101] . Another serious post-operative complication is shoulder pain which depends on the type of surgical technique performed 102,103 .
In order to relieve pain, Loh et al. found that progressive relaxation exercises have a positive effect on the reduction of overall pain in the short-term 65 . Muscle relaxation techniques improve pain in HNC undergoing surgery by decreasing anxiety and stress 104 . The mechanism that could best explain these positive effects is decreasing serum cortisol levels that has been previously associated in with the onset of myofascial pain in these subgroup cancer patients 105 . However, there is not consensus among authors about this hypothesis regarding the role of relaxation exercise on pain releasing as Kim et al. do not find blood cortisol reductions in patients with colorectal cancer 106 . Further studies are required to identify the mechanisms of the effects of exercise on postoperative pain in HNC cancer.
Overall pain in patients undergoing surgery can be reduced by joint mobility exercise in the short-term. This kind of exercise consist of the displacement of joint segments in different dimensions of space with the aim of reaching the maximum range of joint play 107 . Behind these beneficial effects on overall pain are probably the improvement on somatosensory cortical representation of movement 108 , the activation of the endogenous inhibition analgesic system 109 , and the decreasing of psychological factors related to pain (fear-avoidance behavior, etc.) 110,111 .
The amelioration of functionality associated with overall pain produced by the program was probably due the fact that joint mobility exercises, combined with maximum active stretching to the limit of pain tolerance, Scientific Reports | (2023) 13:8523 | https://doi.org/10.1038/s41598-023-35503-y www.nature.com/scientificreports/ relying on through the counterirritation mechanism a phenomenon that, according to Wall and Melzack (1965), occurs as a result of introducing a noxious stimulus more intense than the base pain achieving a downward modulation of it 112 . Also exercise-based rehabilitation has a positive effect on shoulder pain in HNC survivors who undergo surgery. McGarvey et al. 57 found that RET, cervical mobilization exercises have a positive long-term effect on reducing shoulder pain in patients undergoing surgery. Shoulder pain is present in at least 70% of patients after non-selective dissection of the lymph nodes of the neck. Unlike radical neck resection, MRND removes all lymph nodes by radical dissection, but respects one or more of the non-lymphatic structures (spinal accessory nerve, jugular vein, or sternocleidomastoid muscle). In nodal selective dissection, non-lymphatic structures are preserved, and nodule removal is done according to the location of the metastases 113 .
The positive effects of the exercise program of McGarvey et al. 57 were probably attributed to that the included participants were undergone a surgical technique with nerve root, nerve, or its spinal branches preservation, and thus, reducing the possibility of neuropathic pain linked, not only to the mechanical damage of the nerve but also with the entrapment produced by post-surgical fibrosis at the interfaces of the nerve path 57 . The positive effects of RET program on shoulder pain depend, as discussed above for CHT patients, on the increase in muscle pain threshold. Its combination with cervical mobility exercises, which the subsequent activation of endogenous pain inhibition systems, and in the study of McNeely 59 , with muscle stretching, that decreases the activation of muscle nociceptors, result in a significant improvement of shoulder pain 59,71 .
Despite not having found effects in favor of exercise on fatigue in patients with post-surgical HNC, Valkenet et al. found that inspiratory RET has a positive short-term effect on reducing fatigue in those who were operated 60 .
As commented, fatigue-related functionality is the result of a peripheral mechanism accompanied by a reduction in energy metabolism achieved initiation of the anaerobic metabolism pathway leading to fatigue of the muscles involved in ventilation. This fatigue can also be explained by the increased ventilatory rate as part of the onset of systemic inflammatory response syndrome (SIRS) due to surgery in which it is aggravated by the previous state of cancer cachexia that occurs with loss of muscle mass and loss of function 114 .

Limitations
The methodology carried out to conduct this study implies several limitations. First, most of the RCTs include in this meta-analysis have significant heterogeneity and low sample sizes could be an important limitation for external validity of our results. Secondly, the good quality and the high risk of bias detected on the allocation concealment and the blinding of participants or therapist can produce an overestimation of the impact of exercise in the studied outcomes resulting in the possibility of misinformed conclusions. These limitations imply the need for further research into the effectiveness of standardized clinical trial protocols to refute the results and obtain more evidence to reach the level of clinical recommendation.

Conclusions
There is evidence of fair to good methodological quality, low to moderate risk of bias, and weak recommendation supporting the use of exercise-based rehabilitation to increase functionality. However, no evidence was found in favor of the use of this modality for improving the quality of life of HNC survivors who underwent chemoradiotherapy or surgery. The lack of standardization in the development of exercise programs, the diversity of randomized trials, and the heterogeneity of interventions and evaluations warrant further study.