Red Spinach Extract Increases Ventilatory Threshold during Graded Exercise Testing

Background: We examined the acute effect of a red spinach extract (RSE) (1000 mg dose; ~90 mg nitrate (NO3−)) on performance markers during graded exercise testing (GXT). Methods: For this randomized, double-blind, placebo (PBO)-controlled, crossover study, 15 recreationally-active participants (aged 23.1 ± 3.3 years; BMI: 27.2 ± 3.7 kg/m2) reported >2 h post-prandial and performed GXT 65–75 min post-RSE or PBO ingestion. Blood samples were collected at baseline (BL), pre-GXT (65–75 min post-ingestion; PRE), and immediately post-GXT (POST). GXT commenced with continuous analysis of expired gases. Results: Plasma concentrations of NO3− increased PRE (+447 ± 294%; p < 0.001) and POST (+378 ± 179%; p < 0.001) GXT with RSE, but not with PBO (+3 ± 26%, −8 ± 24%, respectively; p > 0.05). No effect on circulating nitrite (NO2−) was observed with RSE (+3.3 ± 7.5%, +7.7 ± 11.8% PRE and POST, respectively; p > 0.05) or PBO (−0.5 ± 7.9%, −0.2 ± 8.1% PRE and POST, respectively; p > 0.05). When compared to PBO, there was a moderate effect of RSE on plasma NO2− at PRE (g = 0.50 [−0.26, 1.24] and POST g = 0.71 [−0.05, 1.48]). During GXT, VO2 at the ventilatory threshold was significantly higher with RSE compared to PBO (+6.1 ± 7.3%; p < 0.05), though time-to-exhaustion (−4.0 ± 7.7%; p > 0.05) and maximal aerobic power (i.e., VO2 peak; −0.8 ± 5.6%; p > 0.05) were non-significantly lower with RSE. Conclusions: RSE as a nutritional supplement may elicit an ergogenic response by delaying the ventilatory threshold.


Introduction
Acute and/or sub-chronic exogenous nitrate (NO − 3 ) ingestion has demonstrated potential for improved sub-maximal exercise time-to-exhaustion (TTE) [1], time-trial performance [2,3], and graded exercise testing (GXT) performance [4]. However, others report NO − 3 ingestion has limited or no effect on performance outcomes [5][6][7]. Specificity of study designs with varying time-course supplementation (i.e., acute vs. chronic), NO − 3 sources, diverse bioactive phytochemicals, or enzymatic changes from enterosalivary circulation may contribute to variable bioavailability and production of nitric oxide (NO) and, thus, influence the observed outcomes [8]. Wylie and colleagues [9] recently characterized responses to acute ingestion of 4.2, 8.4, and 16.8 mmol of NO − 3 from beetroot juice. In that study, plasma (all instances of plasma nitrate and nitrite noted in brackets) [NO − 3 ] and [NO − 2 ] increased dose-dependently with peaks observed 2-4 h post-ingestion with the 6.8 mmol dose compared to 1-2 h post-ingestion with the lower doses. Moreover, arterial blood pressure dose-dependently decreased post-ingestion, but a significant effect was observed for the lowest dose (4.2 mmol) at only the 1-h post-ingestion time point, further illustrating a more transient response to lower doses of NO − 3 . Thus, the timing of NO − 3 -mediated effects may depend upon the dose delivered. An acute 2000 mg dose of red spinach extract (RSE) delivering~180 mg (~2.9 mmol) of NO − 3 has been shown to increase plasma [NO − 3 ] and [NO − 2 ] to peaks of 253 µmol/L (ca. a four-fold change) and 0.56 µmol/L (~1.8-fold change), respectively [10]. Interestingly, these peak concentrations are similar to, or exceed, those observed with acute ingestion of relatively higher NO − 3 doses from beetroot juice [9,11]. Moreover, with acute RSE ingestion plasma [NO − 3 ] peaked at 1 h, although plasma [NO − 2 ] was highly variable between consecutive 15 min time points with the largest spikes between 30-180 min post-ingestion [10]. Collectively, the data suggest differential pharmacokinetics from those observed with beetroot. The majority of inorganic nitrate supplementation studies regarding exercise performance have utilized beetroot. Remarkably, limited studies have evaluated NO − 3 rich leafy vegetable and, more specifically, RSE on exercise performance [12]. Importantly, not only is RSE rich in NO − 3 , potassium (>10% by weight), and anti-oxidant polyphenols (e.g., amaranthine), but it is also devoid of sugar and oxalates.
Given and affect GXT performance. We hypothesized the acute medial dose of RSE would increase exercise economy and improve oxygen utilization during GXT.

Participants
Fifteen (N = 15; males = 8, females = 7) participants were recruited through a community advertisement for this crossover, double-blinded study. Participants included were 19-35 years of age, free from metabolic disease and hypertension (SBP < 140 mmHg and DBP < 90 mmHg), recreationally active (>30 min of physical activity at least 3 days/week), and not consuming any 'supplements' for >1 month. All procedures were approved by the Auburn University Institutional Review Board and conformed to the standards set by the latest revision of the Declaration of Helsinki.

Study Design
Participants reported for two visits separated by >72 h, before the primary testing session. For each visit, participants were instructed to abstain from exercise and alcohol for 24 h and caffeine for 12 h. Participants were also asked to adhere to a low NO − 3 diet for >48 h adapted from the National Heart, Blood, and Lung Institute [13] to reduce the effects of diet on circulating nitrate and nitrate concentrations and to abstain from mouthwash due to its potential to inhibit bio-activation of NO − 3 [14] for 24 h prior to visits. Finally, participants were instructed to replicate their 24 h diet prior to visits and report >2 h post-prandial. In order to control for potential diurnal variation, all participants reported for their visits at the same time of day. Some of these data (i.e., participant characteristics) have been reported previously [15].
Upon reporting for each visit, adherence to guidelines was verbally confirmed. Thereafter, participants' height and weight were measured, a venipuncture was performed (BL), and a dose of RSE or PBO was ingested. At one occasion 65-75-min post-RSE/PBO ingestion, venipuncture was performed (PRE) after which subjects were immediately prepared for GXT. Maximal GXT commenced using the Bruce protocol with analysis of expired gases. At the conclusion of the GXT, a final venipuncture was performed (POST).

Blood Collection and Humoral Nitrate/Nitrite Concentrations
For each venipuncture (BL, PRE and POST),~4 mL of venous blood was collected from the antecubital space in tubes containing ethylenediaminetetraacetic acid (EDTA) and immediately centrifuged at 4000 g at 4 • C for 5 min. Plasma was then separated in cryotubes and immediately placed in an −80 • C freezer for later batch-processing. Following a thaw at room temperature, plasma samples were deproteinized using a cold ethanol precipitation. Briefly, the ethanol was pre-cooled to 0 • C and 500 µL of each plasma sample was added to 1 mL of the cold ethanol and allowed to let stand for 30 min at 0 • C. Thereafter, samples were vortexed and centrifuged at 14,000 rpm for 5 min. The resultant supernatant was collected for assessment of total plasma [

Supplementation: Red Spinach Extract and Placebo
A single oral dose of 1000 mg RSE powder (i.e., Oxystorm ® ) or PBO (maltodextrin) in gelatin capsules was consumed with bottled water. Per manufacturer, the RSE is derived from only pure Amaranthus dubius, has an herb to extract ratio of 50:1, and contains >9% NO − 3 (by HPLC) and >13% potassium (by ICP-MS). The expected ingested dose of nitrate in the RSE was~90 mg.
At the first visit, ingestion of RSE or PBO was determined at random (flip of coin) and the alternate supplement was ingested at the second visit. Investigators and subjects were un-blinded after completion of data collection and analysis.

Graded Exercise Testing
Participants underwent maximal GXT according to the Bruce protocol, which includes progressive increases in treadmill belt speed and deck incline every 3 min. Expired gases were collected with the subjects standing stationary on the treadmill until stable. Thereafter, participants were allowed to warm up by walking at 2.7 km/h at a 0% incline for 3 min followed immediately by initiation of the protocol (Table 1). GXT continued until participants reached volitional fatigue. Expired gases were continuously analyzed using a TrueMax 2400 metabolic measurement system (ParvoMedics, Sandy, UT, USA) and averaged in 20 s intervals.
Peak oxygen consumption rates (VO 2 peak) were determined using the highest 20 s average observed during the GXT and TTE was recorded as the time at which participants indicated volitional fatigue. The ventilatory threshold (VT) was identified by a computer program (WinBreak 3.7, Epistemic Mindworks, Ames, IA, USA) using multiple parallel methods of determination; excess carbon dioxide (ExCO 2 ) [18], ventilatory equivalent ((VEQ), [19]), and modified v-slope [20] methods. ExCO 2 was calculated as ((VCO 2 2 /VO 2 )-VCO 2 ) and plotted over time with the point at which ExCO 2 began to increase disproportionately identified as the VT [18]. For the VEQ method, the VT was identified as the first point at which the ventilatory equivalent for O 2 increased without a concurrent increase in the ventilatory equivalent for CO 2 . Finally, for the modified v-slope method, (1) a plot of VCO 2 vs. VO 2 was divided into two regions with each segment being fit with linear regression; (2) the location of the intersection between the two regression lines was calculated where the point dividing the two regions best fit the data by maximizing the ratio of the greatest distance of the intersection point from the single regression line of the data to the mean square error of regression; and (3) it was ensured that the slope of the first regression line was >0.6 and the change in slope from the first regression to the second was >0.1 [21]. For the ExCO 2 and VEQ methods, a standard algorithm for identifying the break point of two lines was used [22] and for the v-slope method the algorithm proposed by Beaver, Wasserman, and Whipp [20] was used. An illustration of the VT identification methods is presented in Figure 1. An average of the three methods was calculated as VT and the CV between the methods was found to be 3.35%. identified as the first point at which the ventilatory equivalent for O2 increased without a concurrent increase in the ventilatory equivalent for CO2. Finally, for the modified v-slope method, (1) a plot of VCO2 vs. VO2 was divided into two regions with each segment being fit with linear regression; (2) the location of the intersection between the two regression lines was calculated where the point dividing the two regions best fit the data by maximizing the ratio of the greatest distance of the intersection point from the single regression line of the data to the mean square error of regression; and (3) it was ensured that the slope of the first regression line was >0.6 and the change in slope from the first regression to the second was >0.1 [21]. For the ExCO2 and VEQ methods, a standard algorithm for identifying the break point of two lines was used [22] and for the v-slope method the algorithm proposed by Beaver, Wasserman, and Whipp [20] was used. An illustration of the VT identification methods is presented in Figure 1. An average of the three methods was calculated as VT and the CV between the methods was found to be 3.35%.

Statistical Analysis
All data were tested for normal distribution using the Shapiro-Wilk test. An alpha level of 0.05 was required for statistical significance. Independent t-tests were performed to compare subject characteristics. For plasma [NO − 3 ] and [NO − 2 ], a repeated measures two-way ANOVA was employed. When a significant treatment-by-time interaction was observed, within-and between-treatment comparisons were performed using Student's paired t-tests. Hedges' g was calculated to determine effect sizes with values of 0.5, and 0.8 representing moderate, and large effects, respectively [23]. For plasma [NO − 3 ] and [NO − 2 ] at PRE and POST, g was calculated using the ratio of the respective mean difference from BL in the RSE and PBO conditions to the pooled standard deviation of the differences. For GXT parameters, g was calculated using the ratio of the mean difference between treatment conditions to the standard deviation of the differences. Statistical analyses were performed using IBM SPSS Statistics 24 for Windows (Chicago, IL, USA). Data are reported as mean ± SD and effect sizes as mean (95% confidence interval (lower limit, upper limit)).

Participants
One participant was excluded due to the discovery of non-compliance with dietary guidelines ([NO − 3 ] >3-fold higher than study average at baseline). Participant characteristics are presented in Table 2. There were no significant differences between male and female participants for age, height, weight, BMI, and VO 2 peak. Moreover, sex and weight were not found to be significant covariates for any variable analyzed. Data are presented as mean ± SD. BMI, body mass index.

Graded Exercise Testing Performance
Treatment means and individual values for GXT performance parameters are presented in Figure 3. TTE and VO2 peak were not different between conditions (p > 0.100; Figure 3a,b) and no moderate or large effect sizes were observed (d < 0.50). VO2 at the VT was significantly higher during the RSE condition relative to PBO (p = 0.012; Figure 3c) and there was a large effect size (g = 0.73 [0.65, 0.80]). The time at which the VT was observed was not different between conditions (518 ± 110 and 519 ± 97 s for RSE and PBO, respectively; p = 0.991), though the percent of VO2 peak at which VT was observed was significantly higher with RSE (66.8 ± 10.7 and 63.1 ± 7.6% for RSE and PBO, respectively; p = 0.035).

Graded Exercise Testing Performance
Treatment means and individual values for GXT performance parameters are presented in Figure 3. TTE and VO 2 peak were not different between conditions (p > 0.100; Figure 3a,b) and no moderate or large effect sizes were observed (d < 0.50). VO 2 at the VT was significantly higher during the RSE condition relative to PBO (p = 0.012; Figure 3c) and there was a large effect size (g = 0.73 [0.65, 0.80]). The time at which the VT was observed was not different between conditions (518 ± 110 and 519 ± 97 s for RSE and PBO, respectively; p = 0.991), though the percent of VO 2 peak at which VT was observed was significantly higher with RSE (66.8 ± 10.7 and 63.1 ± 7.6% for RSE and PBO, respectively; p = 0.035).

Graded Exercise Testing Performance
Treatment means and individual values for GXT performance parameters are presented in Figure 3. TTE and VO2 peak were not different between conditions (p > 0.100; Figure 3a,b) and no moderate or large effect sizes were observed (d < 0.50). VO2 at the VT was significantly higher during the RSE condition relative to PBO (p = 0.012; Figure 3c) and there was a large effect size (g = 0.73 [0.65, 0.80]). The time at which the VT was observed was not different between conditions (518 ± 110 and 519 ± 97 s for RSE and PBO, respectively; p = 0.991), though the percent of VO2 peak at which VT was observed was significantly higher with RSE (66.8 ± 10.7 and 63.1 ± 7.6% for RSE and PBO, respectively; p = 0.035).

Discussion
The primary findings are that a single 1000 mg dose of RSE compared to PBO (1)  and [NO − 2 ] measurement using Griess-based assays [24]. Indeed, the peak plasma [NO − 3 ] response to a single 1000 mg dose of RSE herein was ~73% of that observed in another study utilizing a single 2000 mg dose of RSE [10]. Thus, given the greater sensitivity of the ozone-based chemiluminescence method [17] and a dose-response that appears to be consistent with previous RSE literature, from a quantitative standpoint our current findings are likely more accurate than those previously reported.
Despite utilizing methods with greater sensitivity for detecting plasma [NO − 2 ], no statistically significant effect of RSE ingestion on plasma [NO − 2 ] was observed. Moreover, the moderate effect sizes suggesting an increase in [NO − 2 ] at both the PRE and POST time points with RSE ingestion were not greater than their respective 95% confidence interval. The lack of a significant alteration in plasma [NO − 2 ] with RSE ingestion could be due the ingested NO − 3 dose. Following un-blinding, 1000 mg of RSE was diluted in nitrate-free water and tested for [NO − 3 ] using the ozone-based chemiluminescence methods described to better characterize the ingested dose. We found ~11.5% NO3 − by weight, which . Condition means ± SD, as well as individual values, are presented for (a) time-to-exhaustion (in minutes), (b) relative VO 2 peak (in mL/kg/min), and (c) observed ventilatory threshold (in L O 2 /min) during the Bruce protocol GXT. Student's paired t-tests were performed to compare performance between conditions. * Significantly different between conditions (p < 0.05).