Same exposure, different outcome. Inorganic nitrate / beetroot is the identical exposure appraised on blood pressure in Dietary Nitrate and Blood Pressure — this page is the endurance-performance outcome, a distinction, not a tension: the BP effect and the ergogenic effect are separate claims on separate endpoints and must not be conflated (a BP-lowering finding is no evidence of an ergogenic one, or the reverse). Both plausibly share the nitrate-nitrite-NO mechanism, but each carries its own evidence.
Scope note. Athletic performance is peripheral on the wiki’s outcome menu (the core stratum is a reasonably-healthy non-athlete; longevity/healthspan/function are central). Performance endpoints are patient-important-adjacent for a competing endurance athlete and are graded here on the same evidence bar as any exposure, but the Layer-1 ranking keeps this a small, low-priority lever for most people. The same-exposure/different-outcome distinction and the peripheral-scope placement are the wiki’s own framing, applying the outcome-menu and Layer-1 rules to this source.
The evidence
Gao 2021 is a systematic review and meta-analysis of 73 placebo-controlled RCTs (n = 1061) of dietary nitrate supplementation vs placebo in adults doing endurance-based exercise, random-effects pooling, GRADE certainty (Gao et al., 2021). All trials were placebo-controlled and single-centre; participants ranged from sedentary to elite, the majority recreationally active healthy adults (Gao et al., 2021). The supplements were commercial nitrate products (blends of nitrate-rich foods/extracts), not whole foods — «our results should be interpreted in con- text of commercial nitrate supplementation, rather than ingestion of natural foods» (Gao et al., 2021), so the result does not transport cleanly to eating beetroot or leafy greens.
Effect by outcome — mean difference (95% CI), heterogeneity, GRADE certainty
The pattern that carries the decision: nitrate moves open-loop lab tasks (constant-load time to exhaustion; power) but not the closed-loop race-simulation endpoint (time trial), and does not raise maximal capacity (VO2max). Every estimate is low or very-low certainty (all downgraded for risk of bias, most for further reasons). (Gao et al., 2021)
| Outcome | Studies | MD (95% CI) | P | I2 | GRADE | State |
|---|---|---|---|---|---|---|
| Power output (W) | 28 | +4.59 [2.6, 6.58] | <0.0001 | 0% | low | benefit (small) |
| Time to exhaustion (s) | 20 | +25.27 [12.69, 37.84] | <0.00001 | 38% | low | benefit |
| Distance travelled (m) | 2 | +163.73 [18.4, 309.1] | 0.03 | 0% | very low | benefit (fragile — 2 trials) |
| Time trial performance (s) | 28 | -1.98 [-4.37, 0.41] | 0.10 | 8% | low | no significant effect |
| Work done (kJ) | 4 | +0.02 [0.0, 0.03] | 0.09 | 0% | low | no significant effect |
| VO2 / O2 cost (L/min) | 42 | -0.04 [-0.05, -0.02] | <0.0001 | 0% | low | reduced (efficiency surrogate) |
| VO2max (L/min) | 10 | +0.04 [-0.02, 0.10] | 0.23 | 0% | very low | no effect |
| RPE (Borg) | 20 | -0.11 [-0.34, 0.12] | 0.36 | 62% | very low | no effect |
| Blood lactate (mM) | 23 | -0.08 [-0.21, 0.05] | 0.22 | 12% | very low | no effect |
Read the four evidence states apart [INFERRED, from the per-outcome CIs above]:
- Benefit (CI excludes null): power, time to exhaustion, distance travelled.
- No meaningful effect (reasonably-powered null, CI around zero): time trial performance (28 trials) and VO2max (10 trials) — these are evidenced nulls, not silence. Work done points null too (CI [0.0, 0.03] kJ) but rests on only 4 trials, so it is a weak null.
- Insufficient / fragile: distance travelled rests on only 2 trials (very-serious imprecision, hence very-low certainty despite significance — a wide CI [18, 309] read off few studies).
- The RPE and lactate nulls are very-low certainty, so closer to insufficient than a firm null.
The decision-relevant gap: the endpoint that matters least is the one that moved most. Time to exhaustion — ride/run to failure at a fixed load — is a lab construct, easy to shift and not how races are won; the authors themselves flag it as a nuanced outcome sensitive to test design (Gao et al., 2021). Time-trial performance — complete a set distance as fast as possible, the closest proxy to a real race — showed no significant effect across 28 trials. A reader optimizing for competition outcomes should weight the TT null heavily and the TTE benefit lightly; the streetlight here shines on the wrong endpoint. [INFERRED, from the TTE vs TT contrast in the table]
Mechanism — efficiency, not capacity
Nitrate is reduced to nitrite (oral bacteria) then to NO (acidic stomach), supplementing endogenous NO; NO promotes vasodilation and improves type-II fibre and mitochondrial efficiency, lowering the oxygen cost of a given workload (Gao et al., 2021). The measured VO2 reduction (−0.04 L/min) is the fingerprint of this mechanism; the unchanged VO2max says the effect is on economy, not maximal aerobic capacity — nitrate makes a given effort cheaper, it does not raise the ceiling. Plasma nitrate/nitrite peak a few hours after ingestion and return to baseline in ~24 h (Gao et al., 2021). Both VO2 and lactate are surrogates (Surrogate Outcomes) — the VO2 drop corroborates the mechanism but is not itself a performance benefit.
Trained vs recreational — the one effect-modification signal (weak, surrogate-only)
Subgroups were attempted by dosing length, athletic level, nitrate source, age, co-supplementation, and risk of bias; daily-dose data were insufficient to subgroup. Only one interaction reached significance: athletic level × treatment on VO2 (interaction p = 0.005) — a significant O2-cost reduction in recreational athletes (MD −0.05 [−0.07, −0.03]) but null in elite (0.01 [−0.02, 0.04]) and sedentary (0.06 [−0.16, 0.29]) athletes (Gao et al., 2021).
This is a route-(b) effect-modification claim, and it is held weakly:
- It is positive interaction evidence (not mere mechanism-inference), and its direction matches the standing expectation that nitrate’s ergogenic effect attenuates in highly-trained athletes — the authors reason elite muscle-fibre efficiency «may already be optimized» (Gao et al., 2021).
- But it is on the VO2 surrogate, not a performance endpoint; it is one significant interaction out of many tested, unadjusted for multiplicity; and the authors caution «we cannot draw con- clusive inferences from this potential interaction as sub- group analyses are prone to type II errors, especially in a meta-analysis where subgroup analyses are not adjusted for multiplicity» (Gao et al., 2021).
So: the trained-attenuation prior gets partial, low-certainty support — enough to note that a highly trained athlete should expect less than a recreational one, not enough to assert it as established. [INFERRED, from the single unadjusted surrogate interaction]
Dose, timing, and harms
- Dose/timing not resolvable from this MA — wide variability in doses, routines and sources; no dose subgroup possible (Gao et al., 2021). A common protocol from the included literature and the co-author athlete’s account: ~500 mL beetroot juice 2-3 h before an event; multiple-day loading is also used (Gao et al., 2021). This is implementation lore, not a dose-response finding.
- Harm — GI distress. The Olympic-marathoner co-author reports beetroot-juice GI distress that scaled with race distance (tolerable at 10 km, problematic at the marathon) and was his reason to stop (Gao et al., 2021) — a single-practitioner, self-reported account, but a concrete tolerability signal that the effective volume can itself impair a long event. Higher-concentration low-volume commercial shots exist partly to mitigate this.
Decision relevance — a small, low-certainty, peripheral lever
For a recreational endurance athlete (the realistic user): nitrate offers a small, low-certainty efficiency gain — lower O2 cost, longer time to exhaustion, more power — but no demonstrated benefit on actual time-trial/race performance and no VO2max gain. Judged against the realistic alternative (their existing training plus other listed ergogenic aids — World Athletics names caffeine and nitrate as the two most-endorsed for endurance (Gao et al., 2021)), it is a marginal add-on, not a big rock. Tolerability (GI) and formulation are the practical constraints.
For a highly-trained athlete, expect less (the surrogate interaction points that way), though the certainty is low. For the wiki’s core reasonably-healthy non-athlete, performance is peripheral; this lever ranks low and thin by construction.
Certainty ceiling. The entire result is low-to-very-low GRADE (risk of bias throughout, publication bias suspected on power/TTE/lactate, imprecision on several). The abstract’s headline «The available evidence suggests that dietary nitrate supplementation benefits performance-related outcomes for endurance sports» (Gao et al., 2021) is true of the pooled averages but overstates what a competitor should expect on the race-relevant endpoint.
Limits
- Single source, single MA. Confidence is low: one gold MA whose own evidence is low/very-low certainty. — a second gold SR/MA to harden the effect states, the TTE-vs-TT split, and the trained-attenuation interaction (directional G-gap; no paper identified).
- Commercial-supplement, not food — does not transport to whole-food nitrate intake.
- Small, select, well-monitored trial populations limit external validity; heterogeneous exercise tests (constant-load vs graded) blur the pooled performance endpoints.
- No nitrate-industry funding for the review; senior authors report unrelated cardiac-device/pharma grants (not nitrate) (Gao et al., 2021).