*2.2. Data Extraction*

Two reviewers (N.V.-S. and G.S.-S.) independently extracted the following data from each study using a predefined Microsoft Excel data extraction form including the number of participants within each group, participant characteristics, racquet sport discipline, and supplementation intervention characteristics, end points, measurement methods, and results in order to produce an overview table of all eligible studies.

#### *2.3. Study Selection*

Studies were eligible for inclusion if they met each of the following criteria: (a) not using any doping substances established by the World Anti-Doping Agency (WADA), (b) using a randomized controlled trial (RCT) design that included one group taking supplementation and 1+ groups receiving a placebo or not taking supplementation, (c) not including any ergogenic aids classified within group A by the Australian Sports Commission (AIS) because of their high evidence grade [22], (d) not presenting supplementation as a source of nutrients, such as bars, gels, or drinks rich in carbohydrates and electrolytes, and (e) not being gray literature (abstracts, conference proceedings, or editorials) or reviews.

#### *2.4. Quality Assessment and Publication Bias*

Characteristics of the retrieved RCTs were evaluated using the 'risk-of-bias' assessment tool following the recommendations by the Cochrane Handbook for Systematic Reviews of Interventions [23,24]. This evaluation was carried out by two reviewers (N.V.S. and G.S.S.) working independently in order to present bias comprehensively. The following criteria were analyzed: randomized treatment order and carry-over effect (selection bias), blinding of participants and research staff to group allocation (performance bias), blinding of outcome assessor (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias (it was assessed if there was controlled diet, exercise use of supplements or drugs, and sport stratification when a mixture of disciplines was analyzed). Then, the retrieved RCTs were classified as being of "high", "unclear", or "low" risk of bias. Effect size was calculated using Cohen´s d test.

#### **3. Results**

#### *3.1. Included Studies*

A total of 438 studies were screened by title and abstract, and 377 were assessed for eligibility criteria (full-text screening). From the retrieved articles, twenty-one met all the inclusion criteria and were included in the systematic review (Figure 1, Tables 2 and 3). Thirteen RTCs were found

in the Medline database (eleven for tennis and two for badminton), seven were found in the Scopus database (three for tennis, one for badminton, two for squash, and one for paddle) where one article was not available despite requesting it from its main author; and none were retrieved from the EBSCO database (because all those found there were repeated). Additionally, one article that was not found through the initial search but was found in a review published in the Medline database was added for full-text analysis. The PRISMA flowchart was applied to illustrate the step-by-step exclusion of unrelated/duplicate retrieved records, leading to the final selection of twenty-one RCTs that met the predefined inclusion criteria (Figure 1).

**Figure 1.** Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart [21] of the study selection process.


**Table 2.** Included studies on nutritional ergogenic aids in tennis. BCAAs:

Branched-chain

 amino acids; FFA: Blood free fatty acids; Glu: Blood glucose; Gly: Blood



**Table2.***Cont.*


**Table 2.***Cont.*

144

#### *Nutrients* **2020**, *12*, 2842



training intensity


#### *3.2. Risk of Bias and Quality Assessment of Studies*

The risk of bias of the included studies is illustrated in Table 4. The RCTs of Pluim; 2006 [33] and Hartono; 2017 [42] were parallel group trials, so the criteria of random sequence generation and allocation concealment were used instead of randomized treatment order and carry-over effect respectively. Most of the trials assessed showed an unclear level in the criteria of selection bias, both in the randomized treatment order and in the evaluation of the carry-over effect. Only the trials by Vergauwen; 1998, [26] Lopez-Samanes; 2020 [36] and Abian; 2015 [40] suitably described the tools used for randomization treatment, while trials by Wu; 2010 [34], Yang; 2017 [38], Abian; 2015 [40] and Muller; 2019 [45] used tests to check if the washout time between conditions was suitable. Moreover, most of the studies also showed a high risk of detection bias except for trials by Gallo-Salazar; 2015 [30] and Abian; 2015 [40], which specifically indicated that blinding was kept until the statistical analysis was performed. Five studies [27,38,42,44,45] were at a high risk of performance bias due to incomplete blinding or a lack of blinding.

**Table 4.** Quality assessment of the included studies. Cross-over studies where A = Randomized treatment order; B = Carry-over effect; C = Performance bias; D = Detection bias; E = Attrition bias; F = Reporting bias; G = Other bias. \* Parallel studies where A = Random sequence generation and B = Allocation concealment.



#### *3.3. Participants*

Age in all the examined studies ranged from 16.4 to 51.0 years old, so that included from junior to master players. Level ranged from university to professional level in both sexes, with a majority of players being males (*n* = 266) as compared to females (*n* = 27). Tennis was the racquet sport about which more studies on NEAs were checked (*n* = 15), followed by badminton (*n* = 3), squash (*n* = 2), and paddle (*n* = 1), but no study was found for table tennis. In the case of NEAs, caffeine was the most evaluated supplement (*n* = 10) followed by creatine monohydrate (*n* = 4), plasma buffers (*n* = 3), nitric oxide (NO) precursors (*n* = 3), and hydration agents (*n* = 1).

#### *3.4. Nutritional Ergogenic Aids and Intervention Characteristics in Tennis*

Caffeine was the most tested NEA with seven studies (Table 2). All trials had a duration of 1 day with variations in concentrations and timing. Most studies selected used a caffeine dosage of 3–6 mg/kg 30–60 min before the tests [26–28,30,31], with improvements in specific tennis skills such as accuracy serve, backhand stroke, serve velocity in last sets, total number of successful shots, handgrip force, and number of sprints compared with control groups. Other protocols with continued administration during tests but a smaller dosage (0.2–0.25 mg/kg) [25] or the same quantity of caffeine given to each player (80 mg) [29] only show an increase in urine epinephrine or no changes compared with control groups respectively.

Regarding creatine monohydrate, only two studies evaluated its efficiency. Neither a high dosage for five days (20 g/day) [32] nor a load period of six days (0.3 g/kg) followed by a maintenance period (0.03 g/kg) until completing five weeks [33] offered advantages compared to control groups.

NEAs related to plasma buffer function were evaluated by two one-day duration studies. A load of 0.3 g/kg sodium bicarbonate 70 min before test and continuous intake of 0.1 g/kg during test showed maintenance of serve and stroke consistency (number of balls landed within the singles court on the designated side) compared to the control group [34], while 0.5 g/kg sodium citrate 120 min before test increased stroke consistency [35]. Both NEAs increased plasma lactate significantly, but only sodium citrate was accompanied by an increase in blood pH.

Three studies about NO precursors were found. The intake of 70 mL beetroot juice 3 h before the test did not show differences compared to control [36]. On the other hand, only citrulline malate supplementation (80 g 60 min before the trial) [37] or together with arginine and BCAAs (0.05 g/kg 80 min before test) [38] showed improvements compared with control. Citrulline malate improved handgrip strength and relative peak power, and citrulline + arginine + BCAAs avoided the decrease of stroke accuracy and kept stroke consistency and stroke velocity.

Lastly, regarding hydration agents, glycerol was the only NEA found in just one study [39]. The consumption of 1 g/kg glycerol 150 min before the trial and 0.5 g/kg 15 min after it increased body weight, plasma osmolality, and plasma volume, and decreased urine volume.

#### *3.5. Nutritional Ergogenic Aids and Intervention Characteristics in Badminton, Squash, and Paddle*

Caffeine was tested in badminton male players 60 min before exercise protocol with a dosage ranging from 3 to 4 mg/kg [40,41]. It showed improvements in jumps and the number of impacts accompanied with a decrease in errors in anticipation, reaction time, and time of sprints (Table 3). In addition, in paddle, caffeine showed ergogenic effects. The intake of 6 mg/kg caffeine 30 min before the exercise protocol in twelve amateur male players increased the percentage of correct hits, diminishing errors [45].

Different plasma buffers were evaluated by one study in badminton male players [42]. A load of 0.3 g/kg sodium bicarbonate or 0.3 g/kg sodium citrate 90 min before the test showed an increase in time to exhaustion with both supplements (51.3 and 44.4% respectively). Both NEAs increased plasma lactate, but only sodium bicarbonate showed an increase in plasma pH.

Finally, the effect of creatine was evaluated in squash players by two studies, one of them with a load of 0.3 g/kg/day for 5 days before test [43], and the other with acute supplementation of a mixed product composed by 1 g creatine + 1.5 g guarana + 133 mg caffeine [44]. The creatine loading protocol showed a decrease in sprint time, while acute supplementation with guarana and caffeine increased peak power and decreased fatigue, reaction time under pressure, and time visual response.

#### **4. Discussion**

#### *4.1. E*ff*ects of Ca*ff*eine in Racquet Sports*

Caffeine has shown to be an effective ergogenic aid for aerobic and anaerobic exercise with improvements in performance and the perceptions of exertion and muscle pain with dosage ranging from 2.35 to 5 mg/kg [46,47]. A similar dosage range was used in most of the racquet sports studies that showed positive effects and a low risk of bias [26,30,40,41] with the exception of Hornery; 2007 [27] due to its methodology of randomization and blinding. Even using higher doses (6 mg/kg), the positive effects are verified [28,31,45] but with a moderate risk of bias due to aspects of randomization or blinding.

In tennis, caffeine improved power skills such as backhand stroke, serve velocity, handgrip force, and the number and velocity of sprints, as well as mental aspects such as the accuracy serve or total number of successful shots. Lower dosages such as 0.2–0.25 mg/kg before and during a tennis match or approximately 1 mg/kg 30 min before a serve test only increased epinephrine levels in urine but they have not shown any performance improvements [25,29]. Moreover, both studies have a moderate risk of bias since the control of randomization, the carry-over effect, the differences in caffeine dosage between sexes, and the lack of certain control groups could be affecting the results. In another study [26], it was shown that the use of a carbohydrate drink with or without caffeine showed improvements in sprints and serve quality compared with the placebo group. As there were no differences between both conditions, it is not possible to evaluate the real effect of caffeine in this study. Furthermore, although an increase in sweat rate has been observed with low caffeine dosages (3 mg/kg) in junior tennis players [30], several studies have disproved a dehydration risk [14].

On the other hand, handgrip force was not affected in badminton and paddle [40,45], but squat and counter jump height or power were significantly better than in the placebo group in badminton [40], offering a specific advantage in this discipline due to the net height in this sport.

In short resistance training, positive results have been observed regarding caffeine consumption in the reduction of perceived exercise exertion [47]. However, no changes were observed in racquet sports (long duration intermittent sports) with the only exception of using two intakes of 4 mg/kg caffeine before and after the first half of a badminton specific test and combining them with carbohydrates [41]. Despite these null effects, the number of total successful shots (with medium effect size (d = 0.57)) and volley precision (high effect size (d = 0.86)) were improved in tennis and paddle respectively using a high caffeine dosage (6 mg/kg) [28,45].

Therefore, the use of an acute ergogenic dosage of caffeine (3–6 mg/kg) 30–60 min before a match is better than the intake of smaller concentrations, despite its continuous use during the match. Due to the large seasons with accumulative long duration matches such as tennis, caffeine consumption could be a useful aid for all competitive levels, since it may maintain physical and mental conditions. More studies with a high caffeine dosage during long periods of intermittent exercise and in combination with carbohydrates are needed in order to prove caffeine capacity to elicit high accuracy and synergetic effects.

#### *4.2. E*ff*ects of Creatine Monohydrate in Racquet Sports*

Commonly, creatine monohydrate supplementation has been used as a strategy to increase muscle mass and strength during training, but it has been also reported to improve power and anaerobic capacity [48–50]. Thus, the use of creatine in intermittent sports such as racquet sports is of high interest since about 75% of top 100 rank tennis players take it [17]. However, up to the present, there is no evidence for recommendation.

Neither specific tennis skills, such serve or stroke, nor general physical aptitudes, such as sprints or strength (typical short-duration high-intensity movements), were improved with different protocols involving only a creatine load (20 g/day for 5 days) or load and maintenance (0.3 g/day for 6 days and 0.03 g/kg for 28 days) [32,33]. Both creatine protocols were used in two studies that had a low to moderate risk of bias.

On the contrary, in one study on squash, the intake of 0.3 g/kg creatine for 5 days was capable of improving the sprint time in a specific test on court [43]. Due to the heterogeneity of sprint protocols, it is not possible to reach a solid conclusion regarding its effect and the effect of moderate bias due to lack of information about the randomization method, carry-over effect and missing information about placebo composition. Moreover, the combination of 1 g of creatine + 1.5 g of guarana + 133 mg of caffeine in an acute dosage improved several physical and alertness aptitudes in squash players, but it has not been ruled out that the stimulant effect of guarana and caffeine were behind them [44]. Therefore, this fact—together with the lack of blinded groups—led to a high risk of bias.

Further studies should evaluate the sprint capacity or service and stroke skills with a high dosage (16 g/day or 0.3 g/kg/day) for a longer period (at least 14 days), as has been shown in previous works [49,50]. Further studies should also consider protocols that emulate long games. In addition, despite not showing any clear improvements in specific tennis skills, creatine consumption during the pre-season could be beneficial for the maintenance or increase of lean mass [50].

#### *4.3. E*ff*ects of Bu*ff*ering Supplements in Racquet Sports*

High-intensity intermittent exercise tends to accumulate acid (H<sup>+</sup>) and carbon dioxide (CO2) in the muscle and blood. Bicarbonate coming from CO2 acts as the primary mechanism to counteract plasma acidification. The efficiency of acute sodium bicarbonate supplementation is influenced by exercise duration. Specifically, extended duration (>4 min) sports have shown diverse results, with sodium bicarbonate improving performance in running and cycling, but not in rowing, rugby, water polo, or basketball [51].

An acute dosage (0.3 g/kg) and a continuous intake for a >4 min specific tennis test (0.1 g/kg) did not improve accuracy or perceptual exercise exertion but kept serve (small effect size (d = 0.42)) and stroke consistency (small effect size (d = 0.09)), which decreased in placebo condition [34] with a low risk of bias. On the other hand, in badminton players, only an acute dosage of 0.3 g/kg increased time to exhaustion in a treadmill test, but not in a specific test in a high risk of bias

study, since the randomized method, allocation concealment, blinding method and control of diet, other supplementation consumption, and exercise load were poorly controlled [42]. Although the blood lactate level was higher than in the placebo groups in both studies, this could be due to the carboxylate co-transporter, which extracts lactate and H<sup>+</sup> from working muscle cell to circulation after an increase in extracellular pH [52] and an increase of glycolytic activity. Despite no changes observed in extracellular pH between placebo and sodium bicarbonate before and after tests in tennis players [34], changes between pre- and post-tests with supplementation may be enough to activate lactate extrusion due to an enhance of glycolytic metabolism.

Other buffer supplements such as sodium citrate, used for causing less gastrointestinal distress than other supplements, also showed significant high values of blood lactate compared to placebo after an acute dosage (0.3–0.5 g/kg) 90–120 min before exercise [35,42]. Nevertheless, an increase in extracellular pH was observed in tennis players, [35] which decreased in badminton players [42], but there are contradictions about it, since the text of the study indicates otherwise. On the other hand, sodium citrate was able to increase stroke consistency (high effect size (d = 1.41)) in junior tennis players, just as sodium bicarbonate did, but it did not present effects in accuracy and perceptual exercise exertion in protocols of >4 min duration [35]. With a non-specific badminton test, sodium citrate was able to improve the time to exhaustion in a treadmill test [42]. Both studies have a moderate to high risk of bias due mainly to the blinding methodology and the control of the intake of other supplementation.

More studies with a higher number of subjects would be needed with the aim of achieving strong evidence about improvements in tennis skills as well as evidencing possible synergies between different buffers (for example, beta-alanine) and other NEAs.

#### *4.4. E*ff*ects of Nitric Oxide (NO) Precursors in Racquet Sports*

It is well known that NO plays a relevant role as a second messenger. Its production is also related to an increase in blood flow, which enhances nutrient and hormone delivery. NO also has a favorable impact on resistance and endurance training adaptations [53,54]. Recent systematic reviews and meta-analysis about NO synthase-independent pathway supplements showed that potassium nitrate and sodium nitrate were less effective than beetroot juice on endurance exercise. The use of beetroot juice supplementation containing 12–6 mmol nitrate displayed significant improvements in time to exhaustion in a cycling race of 5–30 min duration but slightly non-significant improvements in time trial or graded-exercise performance [55,56].

In intermittent sports such as tennis, beetroot juice containing 6.4 mmol nitrate did not show any improvements in either explosive movements (serve velocity, jump, sprint, handgrip force) or perceptual exertion in high-level tennis players [36] with a low risk of bias. These results are similar to the ones found in recent studies in which short and high-intensity movements (such as countermovement jump, isometric strength, or muscular movement concentric velocity) were evaluated after the consumption of beetroot juice containing 6.4–17.7 mmol nitrate [57–59]. It seems that the effect of beetroot juice could be beneficial in endurance performance due to nitrate conversion to NO, affecting improvement in aerobic adenosine triphosphate (ATP) synthesis due to a reduction of VO2. In intermittent and short-term exercise, where the anaerobic alactic system is the main source of energy, the effects are less clear. Only one-third of the studies evaluated in a recent systematic review of intermittent exercise protocols [60] showed significant results in different variables of power compared with the placebo group during repeated-sprint tests.

On the other hand, NO synthase-dependent pathway supplements, such as arginine or citrulline, have shown different results. While arginine supplementation has demonstrated improvements in both aerobic and anaerobic performance with acute (0.15 g/kg) or chronic (1.5–2.0 g/day for 4–7 weeks or 10–12 g/day for 8 weeks) protocols [61], acute protocols of citrulline supplementation (3–6 g) showed a small effect size (0.2) on high-intensity strength and power performance in resistance exercise [62]. In master female tennis players (51.0 ± 9.0 years), acute protocol with 8 g of citrulline improved handgrip strength and power peak in a specific anaerobic test, but not the capacity of sustained

power or jump power [37]. Due to the lack of a washing time between conditions and control of the consumption of other stimulant substances, the risk of bias is moderate. Further studies are necessary to analyze the role of citrulline supplementation in the performance of younger racquet sports players.

Yang et al. (2017) [38] showed improvements regarding the prevention of a decrease in stroke accuracy and keeping stroke consistency and velocity (as opposed to a worsening in the placebo group) using 0.05/kg citrulline +0.05 g/kg arginine +0.17 g/kg branched-chain amino acids (BCAAs). The study presented a low risk of bias. Additionally, perceived exertion after the test decreased significantly. These results appear to be due to a lower plasma tryptophan/BCAAs ratio than placebo, since theoretically, BCAAs compete for the same tryptophan transporter across the blood–brain barrier, avoiding serotonin formation and, consequently, central fatigue instauration [63]. It is common to use a mixture of several NEAs in one product with the objective to obtain a synergic effect, but further studies are necessary in order to verify the true effects of citrulline or arginine by themselves, without the presence of the BCAAs being able to distort them.

#### *4.5. E*ff*ects of Glycerol Supplementation in Racquet Sports*

Finally, glycerol is a naturally occurring metabolite that acts as a plasma expander and could help athletes prevent dehydration and improve thermoregulatory and cardiovascular changes [14]. Until 2018, the World Anti-Doping Agency (WADA) considered glycerol a banned substance, since it was hypothesized that it may alter athlete biological passport [64]. In any case, the results of its supplementation are mixed both in endurance and anaerobic disciplines [14]. In intermittent sports such as tennis, 1.0 g/kg glycerol before followed by 0.5 g/kg after 75 min of simulated match, in environmental conditions in the range of 29–38 ◦C and 50–90% relative humidity (emulating conditions of important tennis tournaments such as The Australian Open Grand Slam or Miami ATP Masters 1000), was not capable of improving accuracy in serves or strokes, sprint velocity, or agility, in spite of its effect increasing pre- and post-exercise plasma volume and osmolality [39]. This study has a moderate risk of bias, since its randomized method, carry-over effect, blinding method and control of diet, and other supplementation and drug consumption were poorly controlled. More research is needed to determine glycerol's supposed potential efficacy in racquet sports during more time-prolonged matches or during several matches on the same day or on consecutive days in hot conditions.

#### **5. Conclusions**

Caffeine is the NEA showing clearer evidence of benefits for racquet sport players. Acute dosages (3–6 mg/kg) 30–60 min before a match may improve specific skills and accuracy but may not contribute to improve perceived exertion. Even though some evidence concludes that other NEAs, such as creatine, sodium bicarbonate, sodium citrate, beetroot juice, citrulline and glycerol, could play an interesting role in improving performance, more studies are needed to strengthen the evidence (Table 5).

**Table 5.** NEA recommendations from current evidence. Green: High level of recommendation due to the high number and quality of studies and the effects produced; Orange: Low level of recommendation due to the low number and/or quality of studies and the effects produced; Red: Not recommended due to the low number and quality of studies and contradictory or low effects.


**Author Contributions:** Conceptualization, N.V.-S. and G.S.-S.; methodology, N.V.-S. and G.S.-S.; protocol drafting, N.V.-S. and G.S.-S.; risk of bias, N.V.-S. and G.S.-S.; quality assessment, N.V.-S. and G.S.-S.; data extraction, N.V.-S. and G.S.-S.; literature search, N.V.-S. and G.S.-S.; search flowchart, N.V.-S.; writing—original draft preparation, N.V.-S. and E.R.; writing—review and editing, N.V.-S., G.S.-S. and E.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the official funding agency for biomedical research of the Spanish government, Institute of Health Carlos III (ISCIII) through CIBEROBN CB12/03/30038), which is co-funded by the European Regional Development Fund.

**Acknowledgments:** CIBEROBN is an initiative of Instituto de Salud Carlos III, Spain.

**Conflicts of Interest:** The authors declare no conflict of interest.
