Effects of Caffeine Supplementation on Physical Performance of Soccer Players: Systematic Review and Meta-Analysis

Abstract

Context:

Caffeine is 1 of the most popular supplements consumed by athletes, and the evidence for improving soccer performance remains limited.

Objective:

To investigate and update the effects (benefits and harms) of caffeine to improve performance on soccer players.

Data Sources:

Electronic search in Medline (via PubMed), CENTRAL, Embase, SPORTDiscus, and LILACS, from inception to March 28, 2020.

Study Selection:

Randomized clinical trials (RCTs) assessing the effects of caffeine on the performance of soccer players.

Study Design:

Systematic review with meta-analysis.

Level of Evidence:

Level 1.

Data Extraction:

Data extraction was conducted independently by 2 authors using a piloted form. We assessed methodological quality (Cochrane risk-of-bias [RoB] table) and the certainty of the evidence (GRADE [Grading of Recommendations Assessment, Development and Evaluation] approach).

Results:

Sixteen RCTs were included. Overall methodological quality was classified as unclear to low risk of bias. When assessing aerobic endurance, meta-analyses did not demonstrate the differences between caffeine and placebo (mean difference [MD], 44.9 m; 95% confidence interval [CI], −77.7 to 167.6). Similarly, no difference was observed during time to fatigue test (MD, 169.8 seconds; 95% CI, −71.8 to 411.6). Considering anaerobic power, meta-analyses also did not find differences for vertical jump (MD, 1.01 cm; 95% CI, −0.68 to 2.69) and repeated sprint tests (MD, −0.02 seconds; 95% CI, −0.09 to 0.04), as well as reaction time agility test (MD, 0.02 seconds; 95% CI, −0.01 to 0.04) and rating of perceived exertion (MD, 0.16 points; 95% CI, −0.55 to 0.87). Regarding safety, a few minor adverse events were reported. Based on the GRADE approach, the certainty of this evidence was classified as very low to low.

Conclusion:

We found no significant improvement in soccer-related performance with caffeine compared with placebo or no intervention. However, caffeine appears to be safe.

Caffeine (1,3,7-trimethylxanthine) is 1 of the most popular supplements consumed by athletes at all levels of sport, in an attempt to enhance performance because of its ergogenic and stimulant properties.^{8,13,15,18,40} Furthermore, it is easily available in different forms such as capsules, powder, caffeinated beverages, energy drinks, and gums. Numerous studies^{9,11,13,16,21} have been published suggesting the physiological mechanism of caffeine, which acts as an adenosine receptor antagonist, delaying fatigue, and providing support for intense training. However, some factors could influence its efficacy such as the dose consumed, timing of intake (before or after physical activity), and the type of performance task. ⁴⁰ Low-to-moderate doses, between 0.5 and 4 mg/kg body mass, seems to present satisfactory results without adverse events.^34,47

Regarding the nature of the performance activity, the ergogenic potential of caffeine has been more easily analyzed and reported in individual sports such as running or cycling than in team sports, which requires a higher demand for different skills from the athletes. ⁴³ Looking specifically at soccer, successful performance requires a high degree of technical and cognitive skills involving with a combination of aerobic and anaerobic endurance, running speed, jumping, change directions, muscle strength, and agility.^8,35

Given the influence of soccer on sports sciences and cultural economic fields, numerous clinical trials have been conducted evaluating the benefits and harms of caffeine consumption to improve performance on soccer-specific abilities. This fact implies the need to map, critically assess, and synthesize the results of these studies to provide more reliable information. Only 1 systematic review was found on this topic, which reported that a moderate dose of caffeine consumed before a soccer match seems to improve some athlete skills and physical performance with no report of adverse effects. ³⁵ However, this review did not estimate the magnitude of the caffeine effect through meta-analyses nor assessed the certainty of the overall evidence by tools such as the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) approach. ²⁰ Considering the popularity of caffeine and the interest of health professionals, athletes, and coaches for information based on quality evidence, the update of this topic is essential for the decision-making process. Therefore, the purpose of our systematic review was to investigate and update the evidence on the effects (benefits and harms) of caffeine to improve physical performance on soccer players.

Methods

This systematic review was conducted following the Cochrane Handbook of Systematic Reviews, ²³ and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. ³⁶ The protocol of this review was not registered in PROSPERO since this platform does not accept systematic reviews assessing sports performance as an outcome.

Criteria for Eligibility

We included randomized clinical trials (RCTs), with parallel or crossover design, assessing the effects of caffeine supplementation on the physical performance of soccer players. We considered any dose and type of caffeine intake (capsules, caffeinated beverages, chewing gum, powder, nasal sprays, bars, and gel) before or during performance tests, compared with placebo or no supplementation. We excluded studies that (1) evaluated only hematological and biochemical outcomes or (2) analyzed multiple interventions so that the isolated effects of caffeine could not be measured. The following outcomes were analyzed:

• Primary outcomes

• Aerobic endurance, measured by yo-yo test, maximal running distance, among other physical tests

• Anaerobic power, measured by sprint tests, maximal jump height, among other physical tests

• Skill performance and agility, measured by soccer-specific abilities such as change of directions, the Loughborough Soccer Passing Test (LSPT), among other tests

• Any adverse events resulting from caffeine ingestion such as sleep disturbance, headache, and gastrointestinal problems.

• Secondary outcomes

• Rating of perceived exertion (RPE), measured by a validated scale

• Individual satisfaction.

Search Strategies

We conducted a sensitive search without language, date, or publication status restrictions on Medline (via PubMed), Cochrane Central Register of Controlled Trials (CENTRAL), Embase (via Elsevier), SPORTDiscus (via EBSCO), and LILACS (via BVS) from inception to March 28, 2020, using a sensitive search strategy for each database (Appendix 1 available in the online version of this article). Ongoing studies were searched in the trial registry ClinicalTrials.gov (www.clinicaltrials.gov) and World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch). Gray literature was searched on OpenGrey (www.opengrey.eu). A hand search was also performed by screening the reference lists of relevant studies on this topic.

Selection of Studies and Data Extraction

Two authors independently screened all titles and abstracts identified in the search using the Rayyan platform (rayyan.qcri.org). ³⁷ Studies that potentially meet the inclusion criteria were analyzed in full text. Disagreements were resolved by consensus. Data extraction was conducted independently by 2 authors using a piloted data extraction form with the following items: study design, characteristics of participants, caffeine supplementation, comparator, physical performance outcomes, and numerical data from results. During this process, any discrepancies were also solved by consensus. Trial authors were contacted for additional information.

Assessment of the Risk of Bias of Included Studies

We assessed the methodological quality of the included studies using the Cochrane risk-of-bias (RoB) table. ²³ This tool consists of 7 domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcomes assessors, incomplete outcome data, selective reporting of outcomes, and other potential sources of bias. Two independent authors assessed the RoB and a third author was consulted in case of disagreement.

Data Synthesis and Analysis

The unit of analysis was the individual. For the estimated effects of caffeine, we calculated mean differences (MDs) for continuous outcomes and risk ratios for dichotomous outcomes (considering a 95% confidence interval). When possible, estimated effects were combined in meta-analysis with a random-effects model, using the Review Manager 5.4 software. ⁴² We planned to perform (1) subgroup analysis, comparing the results from participants with different levels of competition; (2) sensitivity analysis, excluding studies with a high RoB from the meta-analysis; and (3) publication bias analysis if a meta-analysis included more than 10 studies. However, it was possible to perform only 1 subgroup analysis.

Certainty of the Evidence

We assessed the overall certainty of the evidence for all outcomes of the main comparison (caffeine capsule versus placebo) using the GRADE approach. ²⁰ This assessment is composed of 5 domains: RoB, inconsistency, imprecision, indirectness, and publication bias. We detailed the reasons for downgrading the certainty of the evidence and created a summary of findings table using the GRADEpro GDT software. ¹⁷

Results

Search Results

Search strategies retrieved 208 references. After removing 43 duplicates, 165 references were analyzed by title and abstract and 140 did not meet the inclusion criteria and were excluded. Twenty-five potentially eligible studies were read by full-text and 7 of them were excluded: 5 assessed only hematological and biochemical outcomes,^6,30-33 and 2 were not randomized trials.^28,46 We found 1 ongoing study (NCT04181619). Therefore, 16 RCTs (reported in 17 references) were included in this review.^{1-5,7,10,12,14,19,22,25-27,29,38,41} Figure 1 details the study selection process.

Figure 1.

Flowchart of the study selection process.

Characteristics of the Included Studies

The 16 included studies enrolled 241 participants (28 amateur, 43 elite young, 15 collegiate, 47 semiprofessionals, and 108 professional players). Fifteen studies included RCTs with a crossover design and 2 with parallel design. The characteristics of the included studies are detailed in Table 1.

Table 1.

Main characteristics of the included studies

Author(s)/Year/Country	Study Design	Participants	Intervention	Comparator(s)	Study Period/Washout	Performance Tests	Funding Source
Andrade-Souza et al/2015 1 /Brazil	RCT crossover	n = 11 amateur players Age = 25.4 ±2.3 y BM = 83.7 ±14.6 kg Height = 179 ± 0.08 cm VO2 max = 43.9 ± 2.3 mL/kg/min	Caffeine capsule 6 mg/kg in distilled water; 3 times during tests (n = 11)	Placebo distilled water (n = 11)	1 day each trial/7 days apart	• RST 5 × 30 m • LSPT • CMJ • RPE	NR
Apostolidis et al/2018 3 /Cyprus	RCT crossover	n = 11 male professional players (high responders to caffeine) Age = 22 ± 4 y BM = 74.1 ± 8.1 kg Height = 179 ± 0.07 cm VO2 max = 61.8 ± 4.6 mL/kg/min	Caffeine capsule 6 mg/kg; 60 min before tests (n = 11)	Placebo composition NR (same scheme) (n = 11)	1 day each trial/4 days apart	• Run to fatigue • CMJ • RPE	NR
Astorino et al/2012 4 /USA	RCT crossover	n = 15 female collegiate players Age = 19.5 ± 1.1 y BM = 63.4 ± 6.1 kg Height = 168.4 ± 5.8 cm VO2 max = NR	Caffeinated drink 255 mL, 1.3 mg/kg caffeine; 60 min before tests (n = 15)	Placebo 190 mL caffeine- and taurine-free drink (same scheme) (n = 15)	1 day each trial/3 to 4 days apart	• RST 3 × 8 trials • Adverse effects • RPE	NR
Bassini et al/2013 5 /Brazil	RCT parallel	n = 14 male professional players Age = NR BM = NR Height = NR VO2 max = NR	Caffeine capsule 5 mg/kg; 60 min before tests (n = 6)	Placebo lactose capsule (n = 8)	1 day, both trials	Yo-Yo IR2	No funding
Bello et al/2019 7 /USA	RCT crossover	n = 24 male/female professional players Age = 20.96 ± 2.05 y BM = Male, 74.7 ± 8.67 kg; female, 65.76 ± 8.01 kg Height = male, 178.6 ± 6.88 cm; female, 165.2 ± 9.78 cm Male VO2 max = 55.3 ± 3.3 mL/kg/min; female VO2 max = 50.9 ± 3.9 mL/kg/min	Caffeine capsule 275 mg; 30 min before tests (n = 24)	Placebo 275 mg cellulose (same scheme) (n = 24)	1 day each trial/2 days apart	• Run to fatigue • Reaction time • RPE • CRT	Compound Solutions, Inc
Del Coso et al/ 2012 10 /Spain	RCT crossover	n = 19 male semiprofessional players Age = 21 ± 2 y BM = 67 ± 2 kg Height = 173 ± 6 cm VO2 max = NR	Caffeinated drink 630 ± 52 mL; 3 mg/kg caffeine; 60 min before tests (n = 19)	Control decaffeinated cola soft drink (same scheme) (n = 19)	1 day each trial/7 days apart	• RST 7 × 30m • Jump test • Total distance − simulated soccer game 2 × 40 min	No funding
Ellis et al/ 2018 12 /UK	RCT crossover	n = 15 male elite young players Age = 16 ± 1 y BM = 66.9 ± 7.9 kg Height = 177.3 ± 4.8 cm VO2 max = NR	Caffeine capsule 1, 2, and 3 mg/kg; 45 min before tests (n = 15)	Placebo (2 mg/kg maltodextrin; same scheme) (n = 15)	1 day each trial/7 days apart	• RST 3 × 20 m • CMJ • Yo-Yo IR1 • Change of direction	NR
Foskett et al/ 2009 14 / New Zealand	RCT crossover	n = 12 male professional players Age = 23.8 ± 4.5 y BM = 71.4 ± 7.4 kg Height = NR VO2 max = 56 ± 4 mL/kg/min	Caffeine capsule 6 mg/kg; 60 min before tests (n = 12)	Placebo (composition NR; same scheme) (n = 12)	1 day each trial/7 days apart	• LIST • LSPT • CMJ • RPE	NR
Guerra et al/ 2018 19 / Brazil	RCT crossover	n = 12 male professional players Age = 23.8 ± 5 y BM = 79.5 ± 9.1 kg Height = NR VO2 max = NR	Caffeinated drink5 mg/kg caffeine in 250 mL of artificially sweetened water; 60 min before tests (n = 12)	Placebo 250 mL of artificially sweetened water (same scheme) (n = 12)	1 day each trial/NR	CMJ	No funding
Hernández-Camacho et al/ 2017 22 /Spain	RCT crossover	n = 17 male amateur players Age = 17.6 ± 0.4 y BM = 69.08 ± 7.5 kg Height = NR VO2 max = NR	Caffeine capsule 4 mg/kg; 90 min before test (n = 9)	Placebo (microcrystalline cellulose; same scheme) (n = 8)	1 day each trial/7 days apart	• CMJ • Adverse events • RPE	NR
Jordan et al/ 2014 25 /USA	RCT crossover	n = 17 male elite young players Age = 14.1 ± 0.5 y BM = 61.1 ± 5.2 kg Height = 172.0 ± 4.8 cm VO2 max = NR	Caffeine capsule 6 mg/kg; 60 min before tests (n = 17)	Placebo (maltodextrin; same scheme) (n = 17)	1 day each trial/2 days apart	• RST 5-10 × 10 m • Reactive agility tests • RPE	NR
Kingsley et al/ 2017 26 / Australia	RCT crossover	n = 10 well-trained male soccer players Age = 22.0 ± 3.8 y BM = 75.8 ± 16.0 kg Height = NR VO2 max = NR	Caffeine capsule 3 mg/kg; 60 min before tests (n = 10)	Placebo (composition NR; same scheme) (n = 10)	NR	• Sprint performance	NR
Lara et al/ 2014 27 /Spain	RCT crossover	n = 18 female semiprofessional players Age = 21 ± 2 y BM = 57.8 ± 7.7 kg Height = 161 ± 6 cm VO2 max = NR	Caffeinated drink3 mg/kg caffeine in 250 mL water; 60 min before tests (n = 18)	Placebo (0 mg/kg caffeine; same scheme) (n = 18)	1 day each trial/7 days apart	• RST 7 × 30 m • CMJ • Total distance − simulated soccer game 2 × 40 min • Adverse events • RPE	Camilo Jose Cela University
Machado et al/ 2009 29 /Brazil	RCT parallel	n = 15 male professional players Age = 19 ± 1 y BM = 72 ± 2 kg Height = 177 ± 4 cm VO2 max = NR	Caffeine capsule 5.5 mg/kg; 60 min before tests (n = 8)	Placebo (500 mg cellulose capsule; same scheme) (n = 7)	1 day, both trials	• Yo-Yo IE1	NR
Pereira et al/ 2011 38 /Brazil	RCT crossover	n = 11 male elite young players Age = 15 ± 1.5 y BM = 58.8 ± 9.1 kg Height = 169 ± 0.07 cm VO2 max = NR	Caffeine capsule 6 mg/kg; 60 min before tests (n = 11)	Placebo (pharmaceutical powder; same scheme) (n = 11)	1 day each trial/3 days apart	• RST 6 × 40 m	NR
Pettersen et al/ 2014 39 /Norway	RCT crossover	n = 20 male professional players Age = 17.6 ± 1.1 y BM = 71.7 ± 6.9 kg Height = NR VO2 max = NR	Caffeine capsule 6 mg/kg; 65 min before tests (n = 20)	Placebo (dextrose; same scheme) (n = 20)	1 day each trial/7 days apart	• Yo-Yo IR2 • Sprinting distance	NR
Ranchordas et al/ 2015 41 /UK	RCT crossover	n = 10 male semiprofessional players Age = 19 ± 1 y BM = 75.5 ± 4.8 kg Height = 180 ± 0.1 cm VO2 max = NR	Caffeinated gum 200 mg, 2.7 ± 0.2 mg/kg, chewed for 5 min before tests (n = 10)	Placebo (gum without caffeine; same scheme) (n = 10)	1 day each trial/14 days apart	• CMJ • 20 m sprint test • Yo-Yo IR1 • Adverse events	Sheffield Hallam University

BM, body mass; CHO, carbohydrate; CMJ, countermovement jump; CRT, choice reaction time; LIST, Loughborough Intermittent Shuttle-Running Test; LSPT, Loughborough Soccer Passing Test; NR, not reported; RCT, randomized clinical trial; RPE, rating of perceived exertion; RST, repeated sprint test; VO₂max, maximum rate of oxygen consumption; Yo-Yo IR1, Yo-Yo intermittent recovery test level 1; Yo-Yo IE1, Yo-Yo intermittent endurance test level 1; Yo-Yo IR2, Yo-Yo intermittent recovery test level 2.

Most of the trial participants had a previous soccer experience of at least 5 years and trained between 2 and 5 days per week at the time of the study. In most of the studies, participants were instructed to refrain from intense exercise and consuming alcohol and caffeine-containing substances 24 to 72 hours before each protocol.

Caffeine doses were administered before each trial (around 30-90 minutes) and varied between 1 and 6 mg/kg with 9 studies investigated high doses (higher than 5 mg/kg, in different forms: in capsules, caffeinated drink, and chewing gum. Habitual caffeine intake varied from 0 to 350 mg/day between participants, based on self-reported daily consumption, with the majority coming from common caffeine sources such as coffee drink, tea, and chocolate.

Three studies^19,27,39 excluded data from the 2 goalkeepers of the analysis because of the difference in their movement compared with the outfield players. Three studies^7,14,41 excluded participants who reported any adverse effects from caffeine ingestion or a history of caffeine sensitivity.

Methodological Quality of Included Studies

Figure 2 presents a summary of the review authors’ judgments on each RoB item for each included study. The reasons for each judgment are presented in Appendix 2 (available online). Six trial authors^{4,5,12,22,29,41} provided additional information after email contact. Regarding selection bias, 4 studies^5,25,29,41 reported an adequate method to generate the randomization sequence of the participants and were judged as low RoB. All the other studies did not provide sufficient information for judgment and were classified as unclear RoB for this domain. Additionally, 2 studies^4,41 detailed the allocation concealment process and was also considered with low RoB.

Figure 2.

Risk of bias summary: review authors’ judgments about each risk of bias item for each included study. (+) = low risk of bias; (?) = unclear risk of bias; (−) = high risk of bias.

One study ¹⁰ was judged as a high RoB for “blinding of participants and personnel,” as the drinks offered to patients (caffeinated and control) had different flavors and the participants were able to differentiate to which group they were allocated. Nine studies^{1,4,5,12,19,22,27,29,41} were classified as low RoB due to the adequate method to blind participants. Regarding the “blinding of outcome assessors” item, 2 single-blinded studies^12,15 had a high RoB, and 6 studies^{5,10,22,27,29,41} reported the blinding of outcome assessors adequately and were judged as low RoB. All the others did not provide sufficient information and were considered as unclear RoB.

Most of the included studies did not have substantial losses of participants. Only 1 study ²⁷ was classified as a high RoB for the incomplete outcome domain due to a high rate of losses (18%). Trial register protocols were not available for any of the included studies, and lastly, no other sources of bias were detected in the included studies.

Effects of Interventions

Comparison 1: Caffeine Versus Placebo

Aerobic endurance

Five RCTs^{5,12,29,39,41} assessed aerobic endurance outcomes, the results from meta-analyses did not present an important difference between groups on both yo-yo tests measured (m): intermittent recovery test level 1 (IR1) (MD, 44.9 m; 95% confidence interval [CI] −77.7 to 167.6; 2 studies, 25 participants, I² = 0%) and intermittent recovery test level 2 (IR2) (MD, −10.0 m; 95% CI, −183.8 to 163.7; 2 studies, 33 participants, I² = 0%). Results from an individual study ²⁹ (14 participants) also showed a no difference on intermittent endurance test level 1 (IE1) (MD, −3.5 m; 95% CI, −25.4 to 21.7) (Figure 3).

Figure 3.

Meta-analyses of caffeine capsules compared with placebo on aerobic endurance tests (yo-yo tests).

Two studies^3,7 assessed the time to fatigue measured in seconds during a simulating soccer-game protocol, and there was no difference between caffeine and placebo (MD, 169.8 seconds; 95% CI, −71.8 to 411.6; 2 studies, 35 participants). However, 1 of these studies ³ distinguished the athletes as high and low caffeine responders, and as we decided to consider the data from higher responders, it could explain the substantial heterogeneity identified between studies in the meta-analysis (I² = 81%). Individual results from this RCT showed that time-to-fatigue improved with caffeine relative to placebo (MD, 310 seconds; 95% CI, 116.73 to 503.27; 11 participants), but the sample is small and confidence interval of this analysis is wide, suggesting that an important difference on effect cannot be ruled out. Last, only 1 study ¹⁴ (12 participants) assessed the aerobic endurance through the Loughborough Intermittent Shuttle-Running Test (LIST) and also did not present a relevant difference in this task when comparing caffeine to placebo (MD, −3.8 seconds; 95% CI, −9.52 to 1.92).

Anaerobic power: countermovement jump

Seven studies^{1,3,12,14,19,22,41} assessed countermovement jump as total height (in centimeters). Two studies^1,19 provided only graph results and it was not clear if they reported standard deviation, standard error, or confidence intervals. Therefore, they were not included in the meta-analysis. The meta-analysis did not find an important difference between the caffeine and placebo (MD, 1.01 cm; 95% CI, −0.68 to 2.69; 5 studies, 65 participants, I² = 3%) (Figure 4).

Figure 4.

Meta-analyses of caffeine capsules compared with placebo on the countermovement jump test.

Anaerobic power: repeated sprint test

Eight studies assessed this outcome.^{1,26,12,14,25,38,39,41} Six studies reported the time (in seconds) to complete 10-m sprint, ²⁵ 15-m sprint, ¹⁴ 20-m sprint,^12,41 30-m sprint, ¹ and 40-m sprint, ³⁸ and 2 studies^26,39 assessed the peak speed. Andrade-Souza et al ¹ (11 participants) only reported the results in a graphical presentation, and it was not possible to derive the estimates to perform the analysis. These authors showed no important effects of caffeine supplementation for the best sprint and total sprint time from the repeated-sprint test.

The results are presented in Figure 5, and no important difference was found between groups in any assessment: 10-m (MD, 0 seconds; 95% CI, −0.04 to 0.04; 1 study, 17 participants), 15-m (MD, −0.02 seconds; 95% CI, −0.12 to 0.08; 1 study, 12 participants), 20-m (MD, −0.02 seconds; 95% CI, −0.09 to 0.04, 2 studies; 25 participants; I² = 0%), 40-m (MD, −0.02 seconds; 95% CI, −0.20 to 0.16; 1 study, 11 participants), and peak speed (0.07 m/s; 95% CI, −0.12 to 0.27; 2 studies, 29 participants, I² = 0%).

Figure 5.

Meta-analyses of caffeine capsules compared with placebo on repeated-sprint tests.

Skill performance and agility

Two studies^1,14 tested soccer skills performance with the LSPT. Andrade-Souza et al ¹ (11 participants) only reported the results graphically and it was not possible to include it in the analysis. These authors reported no difference between caffeine and placebo on skill performance and agility. The results based on the study by Foskett et al ¹⁴ (12 participants) showed no important difference between groups (MD, −3.8 seconds; 95% CI, -9.52 to 1.92).

One study ¹² (15 participants) evaluated arrowhead agility with a change of direction movements to the left and right. There was no difference between caffeine and placebo for both changes of direction to the right (MD, −0.12 seconds; 95% CI, −0.30 to 0.06) and to the left (MD, −0.08 seconds; 95% CI, −0.22 to 0.06).

Simple reaction time was tested by 2 studies^7,25 after trial completion, to measure both the cognitive and physical aspects of agility. Results from a meta-analysis presented no difference between groups (MD, 0.02 seconds; 95% CI, −0.01 to 0.04; 2 studies, 34 participants, I² = 0%). Bello et al ⁷ (24 participants) also tested choice reaction time and the estimated effect showed no difference between caffeine and placebo (MD, −0.02 seconds; 95% CI, –0.06 to 0.01).

Adverse events

The adverse events resulting from caffeine ingestion were analyzed by 2 studies.^22,41 One (18 participants) described nonsignificant difference between caffeine and placebo regarding the report of headache (5.5% vs 5.9%,P = 1.00), abdominal discomfort (5.9% vs 17.6%, P = 0.60), and insomnia (11.8% vs 17.6%, P = 1.00). ²² The other study ⁴¹ (10 participants) reported no adverse events.

Rating of perceived exertion

Six studies measured this outcome using the Borg 6- to 20-point scale^3,7,14,22,25 and 15-point scale ¹ after test completion. Three of them^1,7,14 did not provide numerical data to pool in a meta-analysis, but individual studies reported that there were no differences between trials for participants’ RPE. Pooled results from the other 3 studies^3,22,25 also showed no important difference between caffeine and placebo on RPE (MD, 0.16 points; 95% CI, −0.55 to 0.87; 3 studies, 29 participants, I² = 0%) (Figure 6). The subgroup analysis, removing 1 study ²² that assessed amateur players, also showed no difference in the results (MD, −0.02 points; 95% CI, −1.21 to 1.17, 2 studies, 20 participants).

Figure 6.

Meta-analyses of caffeine capsules compared with placebo on the rating of perceived exertion.

Comparison 2: Caffeinated Drink Versus Placebo

Aerobic endurance

Two studies^10,27 assessed this outcome, testing the total distance running in a simulated soccer game. No important difference between groups was observed in the meta-analysis (MD, 436.3 m; 95% CI, −49.4 to 922.2; 2 studies, 37 participants, I² = 0%).

Anaerobic power: countermovement jump

Two studies^10,27 assessed the mean vertical jump height (in cm), during the 15-second maximal jump test, and results from the meta-analysis showed a nonsignificant difference between groups (MD, 0.93 cm; 95% CI, −1.22 to 3.08; 2 studies, 37 participants, I² = 0%).

Anaerobic power: repeated sprint test

Two studies^10,27 reported the mean peak running speed during 7 sets of 30-m sprint test and showed no difference between caffeinated drink and placebo (MD, 0.15 km/h; 95% CI, −0.83 to 1.13; 2 studies, 37 participants, I² = 29%). One study (15 participants) assessed the mean sprint time in seconds during 3 sets of 8 trials of the agility t test, and there was no difference between groups (MD, −0.13 seconds; 95% CI, −0.54 to 0.28).

Adverse events

Only 1 study ⁴ (15 participants) reported stomachache and mild tremor in 2 participants of the caffeinated drink group and none in the placebo group. Lara et al ²⁷ (18 participants) showed that more participants in the caffeinated drink presented insomnia (33% vs 0%) and headache episodes were more prevalent in the placebo group (22% vs 0%). However, there was no important difference between groups.

Rating of perceived exertion

Lara et al ²⁷ (18 participants) measured RPE after tests with a 1- to 10-point scale and showed no difference between caffeinated drink and placebo groups (MD, −1.00 points; 95% CI, −2.31 to 0.31). Astorino et al ⁴ (15 participants) also reported no effect of caffeine on RPE.

Certainty of the Evidence

Based on the GRADE approach, the certainty of the body of the evidence was classified as very low to low about the effect of caffeine (capsule) versus placebo on aerobic endurance, anaerobic power, skill performance, and agility, as well as on the occurrence of adverse events resulting from caffeine ingestion. We downgraded the evidence in 1 level for methodological limitations due to overall unclear RoB for random sequence generation, allocation concealment and blinding, and 2 levels for imprecision (small sample sizes and wide confidence intervals). The summary of findings table presenting the assessment details is available in Appendix 3 (available online).

Discussion

The effect of caffeine supplementation on sports performance is still a topic that has been widely investigated due to some aspects that remain uncertain, especially for team sports athletes, such as soccer players. This systematic review and meta-analysis found no important improvement in soccer-related performance outcomes after caffeine ingestion (mean, 5 mg/kg), when compared with placebo. Regarding safety, only a few minor adverse events resulting from caffeine intake have been reported. Based on the GRADE approach, the certainty of this evidence was classified as very low to low, which indicates that we have little confidence in the effect estimate, and the true effect is likely to be different from the observed in these analyses. Therefore, it is important to interpret these results carefully.

An important point to consider is the particular technical and metabolic demands of soccer players. It involves drills and skills, which are focused on individual techniques, tactics, and physical resources, considered in combination when assessing performance. Aerobic and anaerobic endurance, strength, power, and agility share importance in soccer players’ performance, contributing simultaneously to running, sprinting, jumping, and changing direction. ²⁴ Hence, the effects of any ergogenic aids for these athletes cannot be extrapolated to other team sports. Chia et al ⁸ examined the effects of caffeine on ball-game-specific performance measures and concluded that the efficacy of caffeine consumption varied according to the nature of the game, physical capacities, and habitual use of caffeine.

A systematic review of Salinero et al ⁴³ also assessed studies that analyzed the effects of caffeine on all team sports players. An overall meta-analysis presented a small positive improvement in jump performance, running distance, and the number of sprints, but did not affect the RPE. When analyzing separate data for soccer players, our results showed different estimated effects, except for RPE, which was expected since the nature of the exercise determines the response of RPE. We found only 1 previous systematic review that assessed the use of caffeine on soccer performance, ³⁵ reporting that caffeine may improve soccer skills. Such results do not corroborate our findings, probably because of some differences in the eligibility criteria, methodological aspects, and the addition of a recently published trial.

Although the caffeine doses varied between included studies (1 and 6 mg/kg), most of them used about 5 mg/kg 30 to 60 minutes before tasks, which is routinely administered and consistent with previous investigations.^8,18,35,43 In fact, this range of doses seems to be safe, as few adverse effects were reported, but the possible optimal ergogenic dose of caffeine on soccer players performance still need to be investigated by future and well-designed RCTs. Moreover, other aspects also could modify these results and need better understanding, such as (1) a caffeinated energy drink, which contains other potentially ergogenic substances, such as taurine, that can influence performance results; (2) the usual intake of coffee before training; and (3) participants’ characteristics, for example, competitive level, soccer field position, age, and physiological features as body weight and VO₂ max.

It is reasonable to emphasize the impact that nonblinded clinical trials may have on the estimated effect. Most of the studies included in the present review reported insufficient information about the blinding of participants, personnel, and outcome assessors. Given the subjective outcomes assessed and considering that the use of caffeine raises individuals’ expectancy, the knowledge of intervention allocation could modify, consciously or unconsciously, how the athletes perform the tests and how researchers evaluate the outcomes. ⁴⁴

The current systematic review conducted a comprehensive search in the literature, and strictly followed the proposed methodological criteria for conducting a systematic review of intervention.^23,36 We decided to select clinical trials in which caffeine was compared with placebo or no supplementation, precluding the possible influence of another type of supplement in the results. We also consider as relevant performance outcomes those related to physical tests and not to laboratory conditions, to replicate the results of caffeine ingestion in the specific soccer skills.

As limitations, we highlight the heterogeneous variety of performance tests that made it impossible to combine some data in a meta-analysis, which implies the need for standardization of outcomes assessment measures in future clinical trials. Another limitation of this review was the analysis of different athletes’ age categories and competitive levels. Given the lower number of studies pooled in this meta-analysis, it was possible to perform a subgroup analysis for 2 outcomes (jump performance and RPE) by removing studies that assessed amateur players, and there were no differences in the results.

Conclusion

Based on a very low to low certainty of evidence, this systematic review showed no significant improvement in soccer-related performance with caffeine supplementation compared with placebo or no intervention. However, caffeine seems to be safe since only a few minor adverse events were reported. Well-designed clinical trials are still required, with methodological rigor (mainly in the randomization and blinding process) according to the CONSORT (Consolidated Standards of Reporting Trials) statement. ⁴⁵ Future clinical trials needed to focus on standardizing the dose and form of caffeine administration, and on reliable tests to assess performance results.