Human Skeletal Muscle has Large Capacity to Increase Carnosine Content in Response to Beta-Alanine Supplementation. A Systematic Review with Bayesian Individual and Aggregate Data E-Max Model and Meta-Analysis

Beta-alanine (BA) supplementation increases muscle carnosine content (MCarn), and is ergogenic in many situations. Currently, many questions on the nature of the Mcarn response to supplementation are open, and the response to these has considerable potential to enhance the efficacy and applications of this supplementation strategy. Objective To conduct a Bayesian analysis of available data on the Mcarn response to BA supplementation. Methods A systematic review with meta-analysis of individual and published aggregate data using a dose response (Emax) model was conducted. The protocol was designed according to PRISMA guidelines. A three-step screening strategy was undertaken to identify studies that measured the Mcarn response to BA supplementation. In addition, individual data from 5 separate studies conducted in the authors’ laboratory were analysed. Data were extracted from all controlled and uncontrolled supplementation studies conducted on healthy humans. Meta-regression was used to consider the influence of potential moderators (including dose, sex, age, baseline Mcarn and analysis method used) on the primary outcome. Results and Conclusion The Emax model indicated that human skeletal muscle has large capacity for non-linear Mcarn accumulation, and that commonly used BA supplementation protocols may not come close to saturating muscle carnosine content. Neither baseline values, nor sex, appear to influence subsequent response to supplementation. Analysis of individual data indicated that Mcarn is relatively stable in the absence of intervention, and effectually all participants respond to BA supplementation (99.3% response [95%CrI: 96.2 – 100]).


68
Carnosine is a dipeptide formed from the amino acids β-alanine and L-histidine, and is present in high concentrations in human skeletal muscle 69 (approximately 20 -30 mmol·kg -1 dry muscle). Its purported roles include: proton buffering [1]; anti-oxidation [2]; anti-glycation; metal chelation [3] and 70 influencing calcium sensitivity [4,5], and hence muscle contractility. Although in vitro evidence supports carnosine's capacity to contribute to each of these 71 processes, the strongest line of in vivo and in vitro evidence supports an important role for carnosine in intracellular skeletal muscle buffering [3,6]. The 72 pKa of carnosine's imidazole ring (6.83 [7]) renders it ideally placed to aid in pH regulation within the physiological range of skeletal muscle (which 73 decreases from approximately 7.1 at rest to 6.5 after exhaustive exercise [8]). This mechanistic action is particularly relevant in a sporting context, given 74 that sustained high-intensity efforts are largely fuelled by anaerobic bioenergetic pathways, which lead to hydrogen ion accumulation and an acidic 75 environment. Acidosis is known to contribute to fatigue and limit performance via a wide range of mechanisms [9]. As such, the presence of intracellular 76 pH buffers such as carnosine are essential to maintain high intensity muscle contraction, and hence to sustain performance. was the limiting factor in intramuscular carnosine synthesis, and that supplementation with this amino acid could substantially increase muscle carnosine 81 content (MCarn). Shortly after, the same group reported that BA supplementation (mean of 5.2 g·day -1 for 4 weeks) was ergogenic to high-intensity exercise 82 performance [11]. Since then, the ergogenic benefits of this supplementation strategy have been tested and proven in a wide range of models and recently, 83 our group published a meta-analysis showing that BA supplementation is most ergogenic in capacity-based exercise tests that last between 30 seconds 84 and 10 minutes [12]. This strong evidence supporting the ergogenic potential of BA supplementation has earned it its place as one of the world's most 85 popular sports supplements, and it is one of just five ergogenic supplements endorsed by the International Olympic Committee [13].

87
It seems that substantial amounts of BA are required to increase MCarn content, with most studies using doses of approximately 3.2 -6.4 g·day -1 , for 88 periods ranging from 4 -24 weeks. But many questions about the nature of the muscle carnosine response to BA supplementation remain open, and 89 substantial research efforts are being made to refine BA dosing strategies in order to optimize its efficacy and applicability [14]. For example, inter-90 individual variation in response to supplementation is high [12], yet little is known about what factors underpin this [15]. What is the capacity of the muscle 91 to uptake BA and increase MCarn? What is the individual proportion of response to BA supplementation, and do baseline levels dictate the extent of this?

92
Do sex or age influence response to supplementation? To address these questions, we employed a comprehensive analysis that included individual 93 participant data from multiple studies conducted within our laboratory; and combined these findings with summary published data using a frequently 94 used dose-response model (Emax).

97
The protocol for this study was designed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. The

98
Population, Intervention, Comparator, Outcomes and Study Design (PICOS) approach was used to guide the inclusion and exclusion of studies for this 99 review and are described in Table 1.

111
Data Analysis

112
The present study comprised both individual and aggregate data meta-analyses from a Bayesian perspective. Individual data were pooled using Bayesian 113 mixed effects multilevel models. Analyses were performed on the outcome variable MCarn (absolute value) to quantify the effects of beta-alanine 114 supplementation and random noise due to biological variation and measurement error. Additionally, proportion of response was estimated across 115 controlled studies by calculating interindividual difference in response to supplementation and comparing this to a non-zero increase in MCarn [17].

116
Bayesian estimates of the standard deviation in observed change from active and placebo groups were used to obtain the intervention response standard 117 deviation (̂) describing interindividual difference in response. Aggregate data meta-analyses were performed using published pre-and post-intervention 118 mean and standard deviation values. Values were transformed into standardized mean differences (SMD) and sampling variance calculated using methods 119 described previously [18]. Three-level mixed effects models were used to quantify the effects of supplementation dose. Insufficient data were available to 120 allow investigation of the interaction between daily dose and intervention duration and so the total cumulative dose ingested was selected as the primary 121 outcome, which previous research has identified as being more influential than either daily dose or intervention duration [16,19,20]. Subset analyses using 122 study covariates were used to assess the effects of sex, age or measurement method on the main effect of BA supplementation. Finally, a model-based 123 approach was employed to investigate the dose-response relationship between cumulative BA supplementation and the SMD. A standard four parameter 124 sigmoid predicted maximum effect (Emax) model was estimated with: Where is the effect size (SMD), 0 is the baseline effect, is the maximum effect, 50 is the cumulative dose that provides 50% of the maximum 127 effect, is the input (cumulative dose) and is the Hill coefficient controlling the slope of the sigmoid response. Inferences from all models were

184
The predicted maximum effect of BA supplementation (Emax) was 3.0 (50%CrI 2.2 -3.7) and the estimated total cumulative dose (g) required to achieve 185 50% of this maximum effect (ED50) was 377g (50%CrI 210 -494). A density plot with the Emax curve generated from median parameter values is provided 186 in Figure 4. An extrapolation of posterior samples from the Emax model was performed to estimate probabilities that percentage of maximum effect could 187 be achieved with cumulative doses ranging from 1000 to 1500g (see Table 2). These results estimated, for example, that the probability of obtaining at

201
The purpose of this study was to conduct a comprehensive analysis with various modelling techniques to synthesize existing knowledge about the MCarn 202 response to BA supplementation. Collectively, our findings and models employed indicated that human skeletal muscle has large capacity for MCarn 203 accumulation, and that commonly used protocols (e.g., 4 weeks at 6.4 g·day -1 ) may not come close to saturating muscle carnosine content. Baseline values 204 do not appear to influence subsequent response to supplementation and the non-linear response to supplementation was not influenced by sex. Analysis

209
[41]. Their model describes changes in MCarn over time with BA supplementation, and was based on three studies that used a slow release BA formulation.

210
The authors described absolute increases in MCarn as a product of the rates of synthesis and decay, with carnosine synthesis considered to be constant in 211 relation to time and first order to daily BA dose. Similarly, carnosine decay was also considered to be first order, but in relation to total MCarn content. As 212 such, carnosine decay increases when absolute content is higher and so the rate of MCarn accumulation due to BA induced elevations in synthesis will 213 slow. Tissue saturation represents the point at which the rates of synthesis match decay, and so content will remain constant despite continued 214 supplementation. The lack of data at higher BA doses limits precision regarding the point at which human skeletal muscle saturation may occur; however, 215 our analyses suggest that humans have very large capacity to accumulate MCarn and that naturally occurring baseline levels are far below that which the 216 muscle is capable of maintaining.

218
Our analyses indicate that the nature of individual MCarn response to BA supplementation differs from other commonly used supplements, such as 219 creatine. Human skeletal muscle appears to reach creatine saturation at approximately 140 -160 mmol . kgDM -1 [42] and this can be achieved within 5 days 220 of high-dose supplementation. Response to creatine supplementation is largest in those with lowest baseline levels, whereas individuals whose creatine 221 content is habitually closer to this saturation point gain smaller benefit from supplementation [42]. In contrast, we observed no evidence of an influence 222 of baseline carnosine content on the subsequent response to supplementation. This makes sense when considered in relation to our predictive model, as 223 it appears that we have very large capacity to accumulate MCarn -far greater than is achieved with commonly used protocols (e.g., 179.2 grams provided 224 as 6.4g · day -1 for 4 weeks). This may be because baseline MCarn contents (approximately 25 mmol · kgDM -1 ) are substantially lower than those observed for 225 creatine, with many individuals having baseline contents close to the proposed creatine saturation limit of 140 -160 mmol · kgDM -1 . These data, in turn, 226 raise another interesting question. Does human skeletal muscle have a largely uniform saturation point, after which no further increases can be attained 227 (as seems to be the case with creatine)? Or does capacity to accumulate MCarn vary widely between individuals, with each having their own upper limit?

228
Currently, insufficient data using very high BA protocols on MCarn precludes the answering of this question, but one thing that is clear is that human 229 skeletal muscle has large capacity to uptake BA and to increase MCarn above naturally occurring, non-supplemented levels.

231
Our data indicated that humans have large capacity for MCarn accumulation, and this, in turn, raises other important questions, e.g., is maximal MCarn 232 accumulation necessary, or even desirable? Theoretically, the greater the increase in MCarn content, the greater its ability to buffer, and to contribute to 233 other processes, and so intuitively, attaining the largest increases possible seems desirable. Two studies did report that larger MCarn increases were 12 associated with greater performance effects [11,12], but meta-analytic data does not support this, and the total dose ingested was previously reported 235 not to influence the effect on exercise performance [18]. It would be counterintuitive to believe that performance benefits would continue to linearly 236 increase with ever-increasing MCarn, given that numerous factors, apart from acidosis, contribute to fatigue [43], and so it makes sense that at some 237 point, performance benefits would plateau. Numerous studies have reported that approximately 179g of BA can be ergogenic, and so it seems MCarn 238 saturation is not essential for BA supplementation to be ergogenic, although it does remain to be seen whether greater increases can elicit greater effects 239 on exercise performance. A very important avenue for future research is the identification of the lowest MCarn increase necessary to elicit an ergogenic 240 effect, along with the point after which no further benefits can be obtained. This information could have large potential to enhance the applicability and 241 efficacy of BA supplementation strategies. For example, it seems that the largest gains in MCarn are attained in the earlier phases of supplementation (see 242 Figure 4). It would be of interest to identify if strategies such as meal co-ingestion [33],intake in proximity to training [39] or intake in slow-release capsules

243
[35] can influence the early response to supplementation [14] and whether this, in turn, meaningfully impacts exercise performance.

245
The current analysis also brought to light some interesting points about the nature of carnosine response to supplementation, which has implications for 246 future study design. It seems that in the absence of intervention carnosine is relatively stable in the muscle, likely due to low intramuscular carnosinase those for which resources are limited) it may be prudent to direct resources toward the intervention group, in order to reduce within study variance. It is 256 important to note that this recommendation applies only to studies on the MCarn response to BA supplementation. The influence of BA supplementation 257 on exercise performance is far less well-characterized and subject to substantially more sources of internal and external variability and so control groups 258 are essential in studies for which exercise is the primary outcome of interest.

260
In addition to characterizing the nature of MCarn response to BA supplementation, we also considered the influence of various potential moderators on 261 this response. In relation to the method of assessment, it seems that lower effect estimates are generally observed when MCarn is measured using the 262 HR-MRS technique when compared to those obtained using HPLC analysis of muscle biopsies. When considering the influence of non-modifiable factors 263 on the MCarn response to supplementation (namely age and sex), we were unable to conduct analyses on the influence of age, as most studies were 264 conducted in young men, and insufficient data in older and younger groups was available. Women have previously been reported to have lower MCarn 265 than men [45,46], but our data indicate that both men and women have a similar response to BA supplementation. This implies that the lower values 266 previously reported in women are unlikely to relate to an inherent gender dysmorphism in the biological factors that underpin carnosine metabolism.

13
In conclusion, our findings indicate that human skeletal muscle has large capacity to accumulate carnosine. MCarn remains stable in the absence of 269 intervention and neither low baseline MCarn levels, nor sex, influence the subsequent response to BA supplementation. In turn, these findings lead to 270 other questions, the response to which may have large implications for future practice. From the point of view of athletic performance, key questions 271 include: what is the absolute MCarn increase required to elicit an ergogenic effect, along with the point after which no further benefits are attained? It is 272 clear that 4 weeks of BA supplementation can be ergogenic, but can this be achieved earlier? Can strategies to enhance the early response to BA 273 supplementation meaningfully impact the subsequent ergogenic benefits? The response to these questions may progress practical application of this 274 supplementation strategy, with potential benefit to many athletic and clinical populations.

276
Conflict of Interest Statement: Bryan Saunders has previously received a scholarship from Natural Alternatives International (NAI), San Marcos, California 277 for a study unrelated to this one. NAI has also partially supported an original study conducted within our laboratory. This company has not had any input 278 (financial, intellectual or otherwise) into this review. The authors have no other potential conflicts of interest to disclose.