Table 3.

Summary of studies of evidence-based performance supplements used singly and in combination

Study	Subjects and designa	Dose	Performance measure	Performance benefit	Summary
Bicarbonate and beta-alanine
Tobias et al. [28]	Well-trained judo and jiu-jitsu athletes (n = 37 M) Parallel group design (n = 9–10) for placebo, bicarbonate, beta-alanine, and combined	7 days @ 500 mg/kg/day sodium bicarbonate split into 4 doses) and/or 28 days @ 6.4 g/day beta-alanine: Total = 179 g	Combat sports simulation  4 × 30 s upper body Wingate tests with 3 min recovery Tests conducted pre and post 28 days of beta-alanine/placebo	Bicarbonate: Yes Beta-alanine: Yes Interaction: related mechanism with additive effects	Compared to placebo, beta-alanine (+7% p = 0.003) and bicarbonate (+8% p = 0.002) enhanced total work combined bicarbonate and beta-alanine resulted in the greatest gains (+14%, p < 0.0001), and significantly reduced perceived exertion. Authors noted that upper body exercise (smaller muscle mass) produces greater H+ disturbances than lower body exercise accounting for clear and additive effects of both buffering systems
Ducker et al. [29]	Competitive team sport athletes (n = 24 M) Parallel group design (n = 6 M) for placebo, bicarbonate, beta-alanine, and combined	300 mg/kg sodium bicarbonate @ 60 min pre-exercise and/or 28 days @ 80 mg/kg BM/day beta-alanine: Total ~ 168 g	Team sport simulation  3 sets @ 6 × 20 m sprints on 25 s with 45-s recovery Tests conducted pre and post 28 days of beta-alanine/placebo	Bicarbonate: Perhaps Beta-alanine: No Interaction: related mechanism but possible counteractive effect	Effect size comparison revealed that beta-alanine supplementation by itself produced trivial effects on total sprint time, mean sprint time, first sprint and fastest sprint. Whereas bicarbonate alone was associated with “very likely” improvements in these factors, combined beta-alanine and bicarbonate reduced the size of this improvement to “likely”. Small sample sizes were noted as a limitation of this study
De Salles Painelli et al. [30]	Well-trained junior swimmers (n = 6 M, 7 F) Crossover design (bicarbonate or placebo) undertaken post parallel group design (beta-alanine or placebo)	300 mg sodium bicarbonate @ 90 min pre-exercise (first swim) and/or 1 w @ 3.2 g/day + 3.5 w @ 6.4 g/day beta-alanine: Total = 202 g	Swimming  100 m  200 m 30-min recovery Tests conducted pre and post 28 days of beta-alanine/placebo. Competition elements built into research design (warm-up, racing against others etc.)	200 m Bicarbonate: Yes Beta-alanine: Yes 100 m Bicarbonate: Yes Beta-alanine: Possibly Interaction: related mechanism with possibly additive effects	Magnitude based inference analysis and conventional statistics showed for 200 m improvements of 2.3, 1.5, and 2.1% for bicarbonate, beta-alanine and combined (p < 0.05) from baseline with no change in placebo group. Although combined was NS different to bicarbonate alone, probability of positive effect was 78% 100-m swim showed significant improvement with bicarbonate and combined only. However, probability of beta-alanine having positive effect was 65% alone and 72% when combined. Effects more apparent in longer swim (greater acidosis?)
Mero et al. [31]	International and national swimmers (n = 13 M) Crossover design (bicarbonate) before and after all subjects supplemented with beta-alanine	300 mg/kg sodium bicarbonate @ 60 min pre-exercise (first swim) and 28 days @ 4.8 g/day beta-alanine: Total = 134 g	Swimming  2 × 100 m 12 min recovery Tests conducted pre and post 28 days of beta-alanine Competition elements built into research design (warm-up, racing against others etc.)	Bicarbonate: Possibly Beta-alanine: No Interaction: related mechanism but no additive effect	Prior to beta-alanine supplementation, there was no effect of bicarbonate supplementation on swim 1, but there was less attenuation of performance in swim 2 such that performance was ~1.5 s or 2.4% faster (p < 0.05). Following beta-alanine supplementation, bicarbonate failed to provide any change in performance of swim 1 or 2. Note that beta-alanine supplementation undertaken without placebo control and with order effect
Hobson et al. [32]	Well-trained male rowers (N = 20 M) Crossover design (bicarbonate or placebo) undertaken post parallel group design (beta-alanine or placebo)	300 mg/kg sodium bicarbonate: 200 mg/kg @ 4 h pre-exercise + 100 mg/kg BM @ 120 min pre-exercise and/or 30 days @ 6.4 g/day beta-alanine: Total = 192 g	Rowing  2000 m ergometer TT Tests conducted pre and post 28 days of beta-alanine/placebo	Bicarbonate: Possibly Beta-alanine: Probably Interaction: related mechanism with possible additive effect	Magnitude based inference analysis showed that bicarbonate had a likely benefit on rowing performance (3.2 ± 8.8 s improvement > placebo), whereas beta-alanine was very likely to be beneficial to 2000-m rowing performance (6.4 ± 8.1 s vs. with placebo) and there was a small (1.1 ± 5.6 s) but possibly beneficial additional effect with combined beta-alanine + bicarbonate supplementation vs. with beta-alanine alone
Bellinger et al. [33]	Highly trained cyclists (N = 14 M) Crossover design (bicarbonate or placebo) undertaken post parallel group design (beta-alanine or placebo)	300 mg/kg sodium bicarbonate @ 90 min pre-exercise and/or 28d @ 65 mg/kg/day beta- alanine: Total = 129 g	Cycling  4 min ergometer TT Tests conducted pre and post 28 day of beta-alanine/placebo	Bicarbonate: Yes Beta-alanine: No Interaction: related mechanism but no additive effect	Average power output was significantly (p < 0.05) increased compared to baseline for bicarbonate (+3.1 ± 1.8%) and combined beta-alanine + bicarbonate (3.3% ± 3.0%), but not for beta-alanine alone (1.6% ± 1.7%) or placebo. Magnitude based inferences noted a 37% probability of benefits from beta-alanine. Some subjects experienced mild paresthesia (side effect of beta-alanine) which may have compromised study blinding
Bicarbonate and caffeine
Felippe et al. [34]	Regional and national level judo players (n = 10 M) Crossover design to produce bicarbonate, caffeine, combined, and placebo trials	300 mg/kg sodium bicarbonate @ 60-120 min pre-exercise and/or 6 mg/kg caffeine @ 60 m pre-exercise	Judo  3 × JSFT interspersed by 5 min	Bicarbonate: Perhaps Caffeine: Perhaps Interaction: independent mechanism with additive effects	Compared with placebo trial, caffeine and bicarbonate trials both produced a small but non-significant increase in the total number of throws performed across all three JSFT tests. However, the combined trial showed a further increase in throws, which reached significance (72.7 ± 3.1 vs. 68.8 ± 4.2 throws, for combined and placebo respectively, p = 0.003; ES = 1.05)
Christensen et al. [35]	International level rowers (n = 11 M, 1 F; 6 lightweight and 6 heavyweight) Crossover design to produce bicarbonate, caffeine, combined, and placebo trials	300 mg/kg sodium bicarbonate @ 75 min pre-exercise and/or 3 mg/kg caffeine @ 45 m pre-exercise	Rowing  6 min ergometer TT	Bicarbonate: No Caffeine: Yes Interaction: independent mechanism with no additive effect	Compared with placebo trial (1865 ± 104 m), a greater distance was covered with caffeine (1878 ± 97, p < 0.05) and combined caffeine/bicarbonate (1877 ± 97, p < 0.05) but not for bicarbonate alone (1860 ± 96, p = 0.1). Therefore, although bicarbonate was not associated with performance enhancement, neither did it negate the effectiveness of caffeine, as seen in some other studies. Effects more detectable in lightweight rowers
Kilding et al. [36]	Well-trained cyclists (n = 10 M) Crossover design to produce bicarbonate, caffeine, combined, and placebo trials	300 mg/kg sodium bicarbonate @ 90–120 min pre-exercise and/or 3 mg/kg caffeine @ 60 min pre-trial	Cycling  3 km ergometer TT	Bicarbonate: Yes Caffeine: Yes Interaction: independent mechanism with no additive effect	Compared with placebo trial (373 ± 41 W), a greater mean power output was seen with caffeine (381 ± 67, p < 0.05), bicarbonate (383 ± 44, p < 0.05), and combined caffeine/bicarbonate (382 ± 39, p < 0.05). Study unable to discern mechanism of benefit with combined use and whether one supplement negated the effect of the other
Carr et al. [37]	Well-trained rowers (n = 6 M, 2F) Crossover design to produce bicarbonate, caffeine, combined, and placebo trials	300 mg/kg sodium bicarbonate @ 90 min pre-exercise and/or 6 mg/kg caffeine @ 30 min pre-exercise	Rowing  2000 m ergometer TT Overnight fasted	Bicarbonate: No Caffeine: Yes Interaction: independent mechanism with counteractive effect (indirect)	Performance was substantially (~2%) enhanced by caffeine (6:40.8 ± 22 min:s) compared with placebo (6:43.8 ± 23) but differences between bicarbonate (6:44.4 ± 23 min:s), combined (6:42.6 ± 22 min:s) and placebo were unclear. GI symptoms associated with bicarbonate caused failure of expected performance due to enhanced buffering, but also counteracted the benefits of caffeine. Note that protocol did not make use of strategies to reduce GI effects of bicarbonate.
Pruscino et al. [38]	Highly trained swimmers (n = 6 M) Crossover design to produce bicarbonate, caffeine, combined, and placebo trials	300 mg/kg sodium bicarbonate spread @ 30–120 min pre-exercise and/or 6 mg/kg caffeine @ 45 min pre-exercise	Swimming  2 × 200 m 30 min recovery Competition elements built into research design (warm-up, racing against others etc.)	Bicarbonate: Perhaps Caffeine: No, perhaps harm Interaction: independent mechanism with unclear additive effect (indirect)	Differences between trials for absolute times failed to reach statistical significance. However, magnitude based inference analysis showed generally trivial effects apart from caffeine where effects ranged from trivial effect to large harm. bicarbonate showed a smaller reduction in swimming time from Swim 1 to Swim 2 (0.3 ± 0.7% faster) than caffeine or combined trials (p < 0.05). Effect was less evident for a single effort. Majority of athletes recorded fastest TT for single and repeat performance from the combination of bicarbonate and caffeine. Authors suggested that caffeine might have benefit for first swim but this had negative carry-over effect on second swim that was overcome by bicarbonate
Caffeine and nitrate/beetroot juice
Glaister et al. [39]	Well-trained cyclists (n = 14 F) Crossover design to produce BJ, caffeine, combined, and placebo trials	7.3 mmol/day nitrate in BJ @ 2.5 h pre-exercise and/or 5 mg/kg BM caffeine @ 1 h pre-exercise	Cycling  20 km ergometer TT	Nitrate: No Caffeine: Yes Interaction: independent mechanism with no additive effect	Caffeine alone and combined resulted in improved ~3% power output (p < 0.05) compared with placebo however differences between these trials were not significant. Time to complete the 43.83 km distance was reduced to a similar extent (1.3%, p < 0.05) for both the combined and caffeine trials. BJ had no significant positive or negative effect on time to complete the TT. Using an inference based statistical approach. Caffeine would likely (89%) and very likely (99%) enhance TT performance for caffeine and combined, respectively. BJ was unlikely (7%)/very unlikely (1%) to have any positive effect on performance for BJ and combined, respectively
Lane et al. [40]	Well-trained cyclists (n = 12 M) Crossover design to produce BJ, caffeine, combined, and placebo trials	2-day pre-load (8.4 mmol/day nitrate in BJ) + acute dose 8.4 mmol nitrate @ 2 h pre-exercise and/or 3 mg/kg BM caffeinated gum @ 40 min pre-exercise	Cycling  43.83 ergometer TT simulating Olympic course Nutritional conditions simulating race practices (carbohydrate pre-event as mouth rinse during TT)	Nitrate: No Caffeine: Yes Interaction: independent mechanism with no additive effect	Caffeine alone improved power output compared with BJ and placebo (205 ± 21 W vs. 194 ± 22 W and 194 ± 25 W, p < 0.05), leading to a time improvement of ~45 s. However BJ had minimal effect on performance alone or in combination, and may have slightly reduced the effect of caffeine since the outcome of the combined trial was neither significantly different to placebo or to caffeine
Lane et al. [40]	Well-trained cyclists (n = 12 F) Crossover design to produce BJ, caffeine, combined, and placebo trials	2-day pre-load (8.4 mmol/day nitrate in BJ) + acute dose 8.4 mmol nitrate @ 2 h pre-exercise and/or 3 mg/kg BM caffeinated gum @ 40 min pre-exercise	Cycling  29.35 km ergometer TT simulating Olympic course Nutritional conditions simulating race practices (carbohydrate pre-event as mouth rinse during TT)	Nitrate: No Caffeine: Yes Interaction: independent mechanism with no additive effect	Caffeine alone and combined resulted in improved ~3% power output (p < 0.05) compared with placebo however there was no significant difference between these two conditions. Time to complete the 29.35-km distance was faster by 0.9 and 1.6% (p < 0.05) for the combined and caffeine trials. respectively. BJ had no significant positive or negative effect on time to complete TT. Using an inference based statistical approach, caffeine would likely (88%) improve TT performance and very likely (99%) enhance power output (caffeine vs. lacebo), with a small reduction in this effect when combined with BJ. BJ was very unlikely (0%) to have any positive effect on performance
Nitrate/beetroot Juice and phosphate
Buck et al. [41]	Recreationally trained team sport athletes (n = 13 F) Crossover design to produce BJ, phosphate, combined, and placebo trials	6 mmol nitrate in BJ @ 2 h pre-exercise and/or 6 days @ 50 mg/kg/day FFM sodium phosphate in split doses Washout of 17 days between trials	Team sport  6× 20 s sprints  Undertaken @ 0, 30 and 60 around a 4 × 15 min simulated team game circuit	Nitrate: No Phosphate: Yes Interaction: independent mechanism with possibly counteractive effect	Compared with placebo and BJ, phosphate trial produced faster total sprint time for first and second set of sprints and overall sprints (~5% improvement, p < 0.05). Phosphate produced fastest sprint in set 1 and overall compared with placebo (~6% faster). Trend to better sprint times with combined compared with placebo (~2% improvement) suggesting some lessening of the effect of phosphate when combined with nitrate. Effects of phosphate seen when fresh as well as fatigued
Caffeine and phosphate
Kopec et al. [42]	Trained team sport athletes (n = 11 M) Crossover design to produce caffeine, phosphate, combined, and placebo trials	6 mg/kg caffeine @ 60 min pre-exercise and/or 6 days @ 50 mg/kg/dayFFM sodium phosphate in split doses Washout of 17 days between trials	Team sport  6 × 20 s sprints  Undertaken @ 0, 30, and 60 around a 2 × 30 min simulated team game circuit	Caffeine: No Phosphate: Perhaps Interaction: independent mechanism with possibly counteractive effect	Although results failed to reach statistical significance, effect size and magnitude based analysis revealed that, compared with placebo, phosphate resulted in the fastest times for all sprints with moderate-large effect sizes and “likely” to “very likely” benefits. The beneficial effects of combined were smaller, and the effects of caffeine alone were minimal
Buck et al. [43]	Recreationally trained team sport athletes (n = 12 F) Crossover design to produce caffeine), phosphate, combined, and placebo trials	6 mg/kg caffeine @ 60 min pre-exercise and/or 6 days @ 50 mg/kg/day FFM sodium phosphate in split doses Washout of 17 days between trials	Team sport  6 × 20 s sprints  Undertaken @ 0, 30, and 60 around a 4 × 15 min simulated team game circuit	Caffeine: No Phosphate: Yes Interaction: independent mechanism with no additive effect	Overall results showed that combined and phosphate alone improved sprint times when fresh (set 1) and fatiguing (set 2 and 3) compared with placebo. Caffeine alone had small effects

BJ beetroot juice, BM body mass, ES effect size, FFM fat free mass, GI gastrointestinal, JSPT judo specific performance test, M male, F female, TT time trial

^aRandomized, double-blind, placebo-controlled unless noted