The Relationship between Exercise and Salivary Oxidative Stress: A Systematic Review

Salivary content has been reported as a potential biomarker for oxidative stress assessments especially in context of exercise-induced oxidative stress. This systematic review following PRISMA guidelines aimed to evaluate the effects of physical exercise and changes promoted in oxidative stress identified in saliva. Methods: Studies published up to May 2022 were searched in online databases (PubMed, Scopus, Web of Science, The Cochrane Library, LILACS, OpenGrey, and Google Scholar). Risk of bias evaluation were performed using the Quality Assessment Tool for Before-After (Pre-Post) Studies with No Control Group. Results: A total of 473 references were identified and 22 considered eligible. In this case 14 studies reported increase of antioxidant parameters in saliva while eight studies demonstrated increased lipid peroxidation after exercise. Regarding nitrite levels, two studies showed higher levels after exercise. The quality of evidence was very low due to high heterogeneity, inconsistency and indirectness among studies according Grading of Recommendations, Assessment, Development and Evaluation analysis. Conclusion: Increase of oxidative stress and antioxidant activity in saliva appears to be present after exercise, especially at moderate intensity. However, the wide variety of methods leads to divergent data. For precision in salivary assessments, new research with larger sample sizes and better participant matching are recommended.


Introduction
Physical activity and physical exercise are essential methodologies applied to improve physical capacities, prevent diseases, and strengthen social interactions with adaptations for the biopsychosocial well-being state. Both are beneficial to cardiovascular, metabolic [1], osteomyoarticular, and, recently, cognitive functions [2]. Physical exercise differs from physical activity as it considers individual planning through delimiting the volume, frequency, type of exercise, and intensity [3].
From a metabolic point of view, it is worth mentioning that the oxidative stress is a necessary and inevitable consequence of aerobic or anaerobic metabolism closely related to production of reactive oxygen species (ROS) and reactive nitrogen species (RNS) [4].
Consequently, it seems likely that cell membrane and mitochondria are both a source and a target of oxidizing agents [5]. Furthermore, the development of compensatory responses to oxidative stress induced by physical training [6,7] reflects the decrease of ROS/RNS generation markers, which translate into maintenance. This hypothesis has been supported by clinical and experimental studies, which indicate that a significant adaptive response to physical training involves a large increase in resistance capacity facilitated by increased O 2 consumption and mitochondrial biogenesis [7][8][9]. It has been demonstrated that exercise training prevented the downregulation of relevant mitochondrial functioning parameters, such as the membrane potential (∆ψ), CS (citrate synthase) and dehydrogenases activities [10]. However, despite the many benefits of exercise, we should consider that a frequent practice of high-intensity exercise may increases blood oxidative stress [7].
Considering the benefits and harmful effects of exercise-induced oxidative stress, studies report a dose-response pattern of oxidative and antioxidant parameters [11]. Prooxidant factors participate in the modulation of muscle strength; thus, a low level of ROS/RNS induces an adaptative effect in cell or in the organism, generating a beneficial process called hormesis by authors [7,11]. Hormesis is defined as the process in which a low dose of ROS/RNS, that is harmful in high doses, induces a beneficial adaptive control vasodilation by nitric oxide pathway, for example [12]. Even so, the possibility of oxidative stress induced by physical exercise causing damage mainly in the context of high-performance sports requires the need for methods of oxidative balance assessment that are efficient and of low cost, preferably [13].
Identifying exercise-induced oxidative stress biomarkers is still a challenge, especially in large-scale studies where the cost-effectiveness and non-invasive nature of the measures are crucial [14]. Although plasma has been the most widely utilized tissue to test reactions to physical exercise, in recent years, there are increasing evidences suggesting that saliva may also be employed as a tool to measure exercise-induced chemical changes and adaptations [15,16]. Authors reported a link between antioxidant markers in plasma and saliva in their study. Their research found that antioxidant capacity, when assessed in saliva, can be used to predict oral cavity disorders [17].
In this aspect, saliva has an essential role in the physiology of oral tissues [18]. Salivary defense functions of hard and soft tissues of the oral cavity are observed, and also includes functions of lubrication, swallowing, and digestion [19]. Furthermore, salivary content has been reported as a potential biomarker for several assessments in medical research [20] and sports [21].

Registration Protocol and Study Design
This systematic review was registered at Open Science Framework under the registration 10.17605/OSF.IO/ZHW9D and performed according to PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) guidelines [22] (Supplementary Table S1).

Eligibility Criteria
After developing a PICO statement, interventional studies published until March 2022 were searched in online databases. They should have been made in adult populations (P), having physical exercise as an intervention (I), pre-exercise assessment as a comparison (C), and parameters of salivary oxidative stress as an outcome (O).
Furthermore, the eligibility criteria did not restrict the year of publication and language, but only considered published manuscripts. Articles that did not follow our eligibility criteria, opinion articles, technical articles, guidelines, animal, and pilot studies were excluded.

Information Sources
Two reviewers (R.C.A. and R.O.F.) conducted an independent systematic search on six databases by March 2022: PubMed, Scopus, Web of Science, The Cochrane Library, LILACS, and Clinical Trials. In addition, the OpenGrey and Google Scholar were selected to perform searches on grey literature. Moreover, the reviewers conducted a hand search on the reference lists cited in the studies included in the systematic review.

Search Strategy
We employed a combination of pre-defined controlled MeSh (Medical Subject Headings) and free terms relating to physical exercise and oxidative stress. It was then coupled with terms "humans" or "adults" and "sports" or "athletic" or "exercise" or "athletic performance" and or "saliva" or "salivation" and "oxidative stress." Supplementary Table S2 describes the search approach for selected databases. To execute a combination of predefined terms, differences in database syntax rules were examined.

Selection Process
Two examiners (R.C.A. and R.O.F.) conducted all searches in databases. After that, all relevant citations were saved in a bibliographic reference manager (EndNote, x9 version, Thomson Reuters). The authors subsequently deleted duplicated studies independently, maintaining only one publication. Additionally, titles and abstracts that did not meet the set eligibility criteria were rejected. The remaining articles were examined and evaluated by reading their full texts. Finally, citations were searched from the reference lists of all previously selected remaining articles throughout the full-text analysis. A third examiner (D.R.F.) was consulted if there was any disagreement.

Data Collection Process
The data from the research was retrieved and tabulated separately by two reviewers (R.C.C.A. and R.O.F.). When there was a controversy, another two authors were considered. The researchers consider author information, region and year of publication, study design, sample source, sample sizes, participant age, physical training evaluation, salivary oxidative stress evaluation, statistical analysis, results, and outcomes. We attempted to contact the authors through e-mail in the absence of relevant information for data extraction or risk of bias evaluation. For up to five weeks in a row, we sent the authors a weekly e-mail.

Quality Assessment Analysis
Assessment of the methodological quality and the risk of bias was performed using the checklist of Quality Assessment Tool for Before-After (Pre-Post) Studies with No Control Group (National Institutes of Health, Bethesda, MD). In this case, 12 questions for quality assessment involve methodological issues considering the risk of potential for selection bias, information bias, measurement bias, or confounding (the mixture of exposures that one cannot tease out from each other). Answers are "YES," "NO," "CD" (cannot determine), "NA" (not applicable), or "NR" (not reported) for each question regarding methods. Two reviewers (R.O.F. and D.R.F.) performed the quality rating of selected articles, and in case of disagreement, a third reviewer analyzed the quality assessment.
If the absence of information compromised data extraction or the risk of bias evaluation, an attempt to contact the authors by e-mail was made. The contact consisted of sending a weekly e-mail for five consecutive weeks.

Risk of Bias
Guidance for risk of bias focus on the concepts underlying the questions in the quality assessment tool. For any box checked "no," reviewers should ask, "What is the potential risk of bias resulting from this flaw in study design or execution?" That is, does this factor lead to doubt about the results reported in the study or doubt about the ability of the study to accurately assess an association between the intervention or exposure and the outcome?
After analysis of all twelve questions, a study can be classified with a quality rate as good (low risk of bias), fair (moderate risk of bias), or poor (high risk of bias). If a study was classified as inferior, issues related to methods were described.

Statistical Synthesis
Because the majority of the studies selected had different methodologies and high heterogeneity, a statistical analysis of the data was not considered for this study [23]. As a result, we decided not to perform the meta-analysis in order to avoid reporting data with high heterogeneity or imprecise estimations regarding the effects of physical exercise on salivary oxidative stress [24].

Assessment of Quality of Evidence
The certainty of the evidence was evaluated following the GRADE approach and recommendations for when there is no single estimate of the effect. Observational studies start as moderate evidence, and the certainty of evidence decreases to low or very low quality. It happens if serious or very serious issues, related to risk of bias, inconsistency, indirectness, imprecision, and publication bias, are present. In addition, the quality of the evidence can be upgraded if the magnitude of effect is large or very large, or if the effect of all plausible confounding factors would be to reduce the effect, or suggest a spurious effect. In this way, the quality of the evidence can vary from very low to high.

Study Selection
Among the 473 references identified in searching databases, 93 duplicates were removed. After reading titles and abstracts, 351 references were excluded based on the eligibility criteria; thus, 29 were selected for full-text appraisal. After that, seven studies were excluded: 5 studies conducted their oxidative stress evaluation in the blood, not in saliva [25][26][27][28][29]; one study did not perform salivary evaluations during exercise intervals [30]; and 1 article did not analyse oxidative stress parameters [31].For reasons for excluding articles, please see Supplementary Table S3. Finally, 22 articles were eligible for qualitative assessment ( Figure 1).

Results of Individual Studies and Syntheses
Ten studies evaluated lipid peroxidation parameters in which six studies reported changes in lipid peroxidation parameters. Five of them related an increase of its levels [33,41,42,48,49], while one reported a decrease [46]. Among the 5 studies that reported an increase, two of them evaluated saliva in the context of anaerobic exercise [41,42] with fair [41] and low [42] risk of bias. The other three articles evaluated in an aerobic exercise context [33,48,49] with high [33] and fair [48,49] risk of bias. The authors who reported a decrease evaluated oxidative stress after aerobic exercise [46], but the study showed high risk of bias [46]. Four studies that evaluated lipid peroxidation parameters did not show the statistical difference [15,34,40,43], with low and fair [43] risk of bias.

Risk of Bias
The quality of the measurements reported in each article is shown in Table 2. Only two studies had a poor quality and high risk of bias [14,46]. That happened because their eligibility criteria for the study population were not prespecified and clearly described. Also, the participants in the study were not representative enough. Besides, they did not report if the eligible participants met the prespecified entry criteria enrolled in the study, and the sample size was not sufficiently large. Nine studies presented a fair quality and medium risk of bias [12,33,38,41,[43][44][45]48,49] because some of them did not clearly state the study question [12,44]. Moreover, those fair-quality studies had issues regarding representativeness and sample size. Finally, 11 articles presented a good quality and low risk of bias [15,[35][36][37]39,40,42,47,50,51], with minimum issues related to blindness and sample size calculation.

Certainty of Evidence
In general, and without considering the type of exercise implemented or the variations in the moment of the evaluations, the certainty of the evidence was rated as very low. The studies included in the syntheses presented methodological limitations that could have seriously affected the estimates reported. For the synthesized outcomes that included more than one study, the reported results were not entirely consistent regarding the direction and size of the effects. Clearly, the methodological heterogeneity among the studies explains the inconsistency of the results. Furthermore, since uncontrolled studies were included, it cannot be affirmed that exercise is really the factor that explains the variations in the levels of the oxidative stress parameters evaluated; therefore, it was judged that item indirectness was also seriously affected. Finally, even for the outcomes that included more studies, the number of individuals considered when synthesizing the information was small; thus, it was deemed that the imprecision item was seriously affected (Table 3).

Discussion
Physical exercise appears to produce changes in salivary oxidative stress, with an increase in lipid peroxidation and nitrite levels, as well as an increase in antioxidant activity, according to the selected studies. Due to the methodological differences and limitations of the studies, such as small sample sizes, the certainty of evidence is quite low.
Selected articles of this review demonstrated that salivary content evaluation is feasible to oxidative stress assessment [51]. However, with minimization of oral diseases consequences, saliva can be used for assessment of several biomarkers of lipid peroxidation, protein oxidation, and DNA damage caused by reactive oxygen species [21]. Free radicals and reactive oxygen/nitrogen species (ROS/RNS) in saliva, such as those in plasma and tissues, play a vital role in redox-dependent signaling and are required for physiological processes [52].
Excessive generation of free radicals, on the other hand, might result in oxidative stress [53]. Thus, the redox balance is changed in favor of oxidants. ROS/RNS have the potential to cause oxidative damage to biological components, which can have major pathophysiological effects [54]. Saliva, on the other hand, contains a variety of antioxidant mechanisms, including low molecular antioxidants such as glutathione, ascorbic, and uric acid, as well as melatonin [51]. Saliva also contains antioxidant enzymes such as superoxide dismutase, catalase, and glutathione peroxidase [21,51], which protect the oral cavity from the harmful effects of both endogenous and exogenous ROS/RNS.
Lipid peroxidation is directly linked to muscular contraction during exercise due to oxidation of polyunsaturated fatty acids of cell membranes [7]. Mitochondria, phospholipase A2, and NADPH oxidases are the three main producers of superoxide radicals in muscle fibers (NOX2 and NOX4). NADPH oxidases (NOX2 and NOX4) have been identified as substantial producers of superoxide radicals in the cytosol and sarcolemma, notwithstanding the importance of mitochondrial respiratory activity. An increase in malonaldehyde and nitrite levels is often linked to short-term, high-intensity activity, whether aerobic or not [7,11].
The Thiobarbituric Acid Reactive Substances (TBARS) and Griess reactions are commonly used in assays for evaluating exercise-related lipid and protein peroxidation [43], but the Schiff reaction is also used to evaluate malondialdehyde for peroxidation [12]. In the presence of high-intensity exercise, diene and triene conjugate levels in saliva are raised, indicating probable exercise-related cellular damage. The presence of these conjugates, however, does not always imply cell injury [21]. This is due to the fact that they can be formed from processes such as muscular fatigue, lactate present, and musculoskeletal injuries [55,56]. As a result of the TBARS assay's limited sensitivity and specificity, many contradictory results with regard to saliva were discovered in this review [35,47,48].
Authors report elevated levels of nitric oxide (NO) metabolites during and after exercise [12,36]. NO metabolites derived from reactions of 1-Arginine with Nitric Oxide Synthase isoforms (NOS) indicating protein oxidation [12,36]. Variations in NO levels are also discovered based on the intensity of training [11]. Furthermore, an increase in nitric oxide metabolites is associated with a decrease in salivary immunoglobulin A, potentially aggravating acute oral illnesses [12]. High NO salivary levels might be related to immunological reactions that possibly leads to salivary function reduction. Due to small number of studies evaluating NO metabolites, we suggest further evaluations that consider these parameters since exercise practitioners and athletes usually report hyposalivation [57].
Authors find higher blood levels of many enzymatic components such as superoxide dismutase (SOD), catalase, and peroxidase [15]; non-enzymatics such as reduced gluthatione (GSH) and uric acid; and indices such as total antioxidant capacity (TAC) when connecting exercise and antioxidant activity [7]. High levels of salivary antioxidant indices or enzymes are important to neutralize oxidative radicals in mouth [49]. In the oral cavity, an increase in SOD and catalase levels was detected, which was mostly related to the intensity of the activity, whether aerobic or anaerobic. In terms of non-enzymatic components, uric acid levels are higher in saliva, however the results are mixed among the writers. GSH, on the other hand, is shown to be enhanced in salivary levels, as measured by the TAC index [36,45,47,48]. Because antioxidant activity is dose-dependently related to physical exercise, these variations in antioxidant activity may be related to the establishment of an adaptation to the stress induced and a general condition of individual [7].
It is important to highlight that nine studies reported increased levels of uric acid associated with other antioxidant parameters. Uric acid is a by-product of purine metabolism and is related to peroxynitrite radicals scavenging [16]. In saliva, uric acid is responsible for 70-85% of salivary antioxidant status and may be related to protection of oral tissues against oxidative damage, especially when the periodontal disease, a common inflammatory of dental supporting tissues (gingiva, periodontal ligament, cement, and alveolar bone), affects individuals [16,50]. UA might also reflect systemic alterations regarding exercise since both aerobic and anaerobic exercise showed relationship with increased levels of UA. Despite the relationship between increased levels of salivary UA and exercise, only 2 of 9 studies evaluate UA and NO metabolites. Thus, further studies to evaluate UA and reduction of reactive nitrogen species are recommended.
Although exercise may be linked to changes in salivary biochemical balance, the studies reviewed have a number of flaws. The lack of control groups for matching operations is one of them. Many people object to studies without a control group since there is no way to know whether what happened to the participants would happen to a group that did not receive intervention [58]. To address some of these concerns, several researchers employ the before-and-after methodology, in which baseline data replaces data from a control group. Still, the data must be carefully evaluated, as the work with athletes and active and trained individuals, the biochemical changes of salivary oxidative stress in evaluated group may be related to internal and external validation issues [58].
Another issue that frequently arises in papers is the sample size [59]. Estimating the effects of exercise on salivary oxidative stress is easier when the number of participants is planned ahead of time based on the key outcomes [21]. Random error is common in studies with few participants, and systematic error is also typical, which is caused by difficulty in estimating a specific parameter, the influence of confounding factors, and other issues [58].
It should be highlighted that sports practice and training research presents numerous challenges [58]. The following stand out among them: Difficulties in obtaining participants who meet specific eligibility criteria, resulting in a reduction in the sample size; methods used in sports evaluation, such as VO 2 Max, but which are impractical in certain intervention methodologies, being costly for many researchers and "time-consuming" for athletes, coaches, and investors; and finally, the compliance and adherence of athletes as well as the technical team with regard to activities outside the training routine [60]. Such problems result in many conflicting data and low clinical relevance in the context of intervention in this population. Therefore, even if there is a lot of uncertainty about the evidence of the impact of physical training and the implications of salivary oxidative stress, the data presented by the studies should be interpreted with caution.
This review has limitations. The absence of a quantitative analysis limited the conclusions about precision of data and regarding salivary oxidative stress during physical exercise [23]. However, in the context where there is a heightened of heterogeneity between studies, a meta-analysis is not recommended [24]. Another limitation of this work comprises the evaluation of articles that are observational or quasi-experimental studies. These studies lack the methodological rigor of an RCT, especially in terms of randomization and participant allocation [61]. As a result, even when executing a "pool" of data, the outcomes tend to have a reduced weight of evidence, notably when the sample size is small and with low representativeness of the population [61]. However, due to the lack of articles that present judicious methodologies, this review aims to demonstrate the methodological issues so that further studies can reduce the uncertainty of this association between exercise and salivary oxidative stress. Therefore, to increase the level of evidence about the effects of exercise on salivary oxidative stress, more randomized clinical trials or observational studies with blind evaluations, adequate matching of participants, similar measures to evaluate outcomes, and adequate control of cofounding factors are required [62].

Conclusions
In the context of physical exercise and sports practice, an increase in oxidative stress and antioxidant activity in saliva appears to be present. However, the wide variety of study methodology leads to divergent data and a low level of confidence in the evidence. For precision in saliva findings, new research with larger sample sizes and approaches that promote better participant matching are recommended.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox11081489/s1, Table S1: Prisma 2020 checklist; Table S2: Terms used on databases searches. Table S3. Articles excluded after reading in full and reasons for exclusions; Table S4