Optimizing Ice Slurry Ingestion for Endurance Performance in the Heat : A Meta-Analysis

Introduction Environmental heat has been well documented to impact negatively athletes during high-intensity, long duration competitions. Exercising in high temperature and/or high humidity decreases time to exhaustion, increases heat storage, and reduces endurance performance (Nybo, Rasmussen, & Sawka, 2014). For elite endurance athletes competing at the highest level, performance in the heat has been shown to be impaired by ~3% compared to thermoneutral competition conditions (Guy, Deakin, Edwards, Miller, & Pyne, 2015). Considering the smallest worthwhile variation in elite endurance runners is less than 2.5% (Hopkins & Hewson, 2001), strategies that could eff ectively mitigate the negative infl uences of heat are highly relevant to the medal perspective for elite athletes competing in hot and humid environments. Not surprisingly, there is an ever-increasing interest in active cooling strategies for high-intensity sports that are of a long duration in challenging environments. Successful interventions include cold water immersion, cooling vests, cold/ ice drinks that can be applied as pre-cooling and/or during exercise (Bongers, Th ijssen, Veltmeijer, Hopman, & Eijsvogels, 2015). Active cooling strategies are recognized by both practitioners and sports scientists for the successful participation in endurance exercise in the heat. Interestingly, ice slurry ingestion has been reported to be the most prevalent cooling strategy amongst elite athletes preparing for the 2015 World Athletics Championships (Periard et al., 2017). While the specifi c reason for the popularity of ice slurry ingestion amongst elite athletes is not known, one of the possible reasons could be that ice slurry ingestion represents an eff ective yet practical intervention for fi eld applications. When employed as a pre-cooling intervention, ice slurry ingestion could eff ectively reduce core temperature at the commencement of exercise and, such eff ect could be carried over during the early stage of exercise, thus acting as a heat sink in challenging environments (Naito, Iribe, & Ogaki, 2017; Siegel et al., 2010). It has been reported Abstract


Introduction
Environmental heat has been well documented to impact negatively athletes during high-intensity, long duration competitions.Exercising in high temperature and/or high humidity decreases time to exhaustion, increases heat storage, and reduces endurance performance (Nybo, Rasmussen, & Sawka, 2014).For elite endurance athletes competing at the highest level, performance in the heat has been shown to be impaired by ~3% compared to thermoneutral competition conditions (Guy, Deakin, Edwards, Miller, & Pyne, 2015).Considering the smallest worthwhile variation in elite endurance runners is less than 2.5% (Hopkins & Hewson, 2001), strategies that could eff ectively mitigate the negative infl uences of heat are highly relevant to the medal perspective for elite athletes competing in hot and humid environments.
Not surprisingly, there is an ever-increasing interest in active cooling strategies for high-intensity sports that are of a long duration in challenging environments.Successful in-terventions include cold water immersion, cooling vests, cold/ ice drinks that can be applied as pre-cooling and/or during exercise (Bongers, Th ijssen, Veltmeijer, Hopman, & Eijsvogels, 2015).Active cooling strategies are recognized by both practitioners and sports scientists for the successful participation in endurance exercise in the heat.Interestingly, ice slurry ingestion has been reported to be the most prevalent cooling strategy amongst elite athletes preparing for the 2015 World Athletics Championships (Periard et al., 2017).While the specifi c reason for the popularity of ice slurry ingestion amongst elite athletes is not known, one of the possible reasons could be that ice slurry ingestion represents an eff ective yet practical intervention for fi eld applications.When employed as a pre-cooling intervention, ice slurry ingestion could eff ectively reduce core temperature at the commencement of exercise and, such eff ect could be carried over during the early stage of exercise, thus acting as a heat sink in challenging environments (Naito, Iribe, & Ogaki, 2017;Siegel et al., 2010).It has been reported

Abstract
Ice slurry ingestion is a simple cooling intervention purported to improve endurance performance.Despite its popularity in the fi eld, a recent meta-analysis suggested this intervention has no performance eff ect.The aim of the present meta-analysis was to determine the eff ect of ice slurry ingestion on endurance performance in the heat.Data for this meta-analysis were retrieved from the PubMed.Eff ect sizes were calculated as the standardized mean diff erence (Hedges' g), and meta-analyses were completed using a random-eff ects model.A method-of-moments meta-regression was used to determined confounding factors.Sixteen studies using randomized controlled trials with a total of 152 subjects were included.Improvement in endurance performance in the heat was moderate: g=0.54 (95% confi dence interval, 0.30-0.77,p<0.001).There was a signifi cant dose eff ect associated with the endurance performance (p=0.024);moreover, the performance eff ects of ice slurry ingestion were not infl uenced by the timing of ingestion or environmental conditions.These data support the ingestion of ice slurry during endurance events in the heat.To optimize this simple cooling strategy in the fi eld, it is recommended to ingest no more than 10 g • kg −1 before or during exercise.
that manipulating core temperature by lowering 0.4°C at the commencement of exercise reduced the sweat rate by 20.3% and subsequently increased endurance capacity by 6.8% in a thermoneutral environment (Hessemer, Langusch, Bruck, Bodeker, & Breidenbach, 1984).Increasing the body's heat storage capacity is a primary mechanism that enables ice slurry ingestion to improve endurance performance in the heat.As importantly, logistical issues associated with active cooling strategies prior to and/or during elite competition in the fi eld also favor the adoption of this simple intervention.
Given that environmental heat and humidity can severely limit performance, hence medal chances, in elite competition, optimizing ice slurry ingestion is of interests to elite athletes participating in endurance events.Currently, there is no consensus regarding the dose-response relationship, timing strategy, and infl uences of diff erent environmental temperature and humidity conditions (Bongers et al., 2015;Choo, Nosaka, Peiff er, Ihsan, & Abbiss, 2018).Accordingly, this meta-analysis explored up-to-date evidence on how eff ective ice slurry ingestion could enhance endurance performance in the heat, and in particular, highlighted key intervention strategies that could infl uence its eff ect.

Literature search
A computer-based systematic search (EndNote, ver.X9, Clarivate Analytics, USA) was performed in the PubMed database to identify potentially relevant studies.Th e keywords used were Ice slu* (slurry, slushy, slush) ingestion, which yielded 201 initial records.Searches were made in December 2018 and last updated in December 2018.

Data extraction
Using a standardized sheet, the following information was extracted from each of the included studies: total number of subjects, demographics, environmental conditions, intervention strategies, exercise protocols, and comparisons in outcome measures of endurance performance test.Th e ice dose was standardized as ice slurry ingestion relative to the body mass (i.e., g•kg-1).Four studies did not report this value directly and the doses were estimated using the total fl uid consumption divided by the mean body mass reported in each study (Maunder et al., 2017;Riera et al., 2014;Stanley et al., 2010;Tran Trong et al., 2015).For one study (Riera et al., 2014), the ice dose was determined by combining the ice/fl uid ingestion during warm-up, immediately at the beginning of the performance trial, at the 1st, 2nd and 3rd 5-km of the 20-km time trial.For one study (Tran Trong et al., 2015), the ice dose was determined based on during one block of 4-km cycling and 1.5-km running.Studies that employed ice slurry ingestion prior to the endurance performance test only, were coded for subgroup analysis.
Th e assessment of risk of bias for each study was completed by using the Physiotherapy Evidence-Based Database Scale (PEDro) (Moseley, Herbert, Sherrington, & Maher, 2002).Th e scale assesses bias for each study from 11 evidence-based criteria.However, due to blinding of the ice slurry ingestion was not practical, the highest score that could be obtained from the scale was adjusted to 8.

Meta-analysis
Since a post-analyses of the results suggest the temperature gradient of the ice and fl uid ingestion does not aff ect the eff ect size, the trial results based on two diff erent control fl uid temperatures (Tran Trong et al., 2015) were combined (Julian PT Higgins & Green, 2008) to avoid the unit-of-analysis error.Studies that investigated the temporal eff ect of ice slurry ingestion were treated as independent data points (Naito et al., 2017;Takeshima et al., 2017).For one study (Yeo et al., 2012), the environmental condition was estimated (ambient temperature as 32°C, and relative humidity as 55%) based on general information from that article.Sensitivity analysis checked how this imputation of missing data would have infl uenced the precision of the results.
Due to the nature of sports science studies, the included studies were based on small sample size.Hence, the Hedges' g values were calculated as eff ect size in this meta-analysis (Comprehensive Meta-Analysis, ver.3.3, Biostat, USA) and interpreted as small eff ect (g=0.2),moderate eff ect (g=0.5), or large eff ect (g=0.8).Considering populations and experimental protocols vary across studies, a random-eff ects model was fi tted for the meta-analysis.A positive g indicates improved endurance performance following ice slurry ingestion, whereas a negative g indicates null eff ects following ice slurry ingestion.Confidence intervals (CIs, 95%) not overlapping the null were considered a statistically signifi cant eff ect.
Th e inter-study heterogeneity was quantifi ed based on the Cochran's Q statistic, with a cutoff signifi cance level of 10% (J.P. Higgins, Th ompson, Deeks, & Altman, 2003).A method-of-moments meta-regression was performed to identify covariates for the dispersion of the main eff ect size.Th e publication bias was visually inspected based on the funnel plot method.

Results
A detailed summary of the main characteristics of the 16 studies is presented in Table 1.Th is meta-analysis represents data synthesized from 152 subjects, varying in sample size from 7 to 12.All subjects included in these studies were defi ned as healthy, physically active, recreational athletes or moderately trained athletes.Of the 16 studies, 11 studies investigated the eff ects of pre-exercise cooling, 3 studies investigated pre-and mid-exercise cooling, and 2 studies investigated mid-exercise cooling.Th e risk of bias of the studies assessed as PEDro scores ranged from 6-8 out of a possible 8 points, qualifying as high quality.
Figure 3 shows meta-regression investigating the dispersion of the eff ect size that could be explained by the study's characteristics.Th ere was a signifi cant association of ice dose with the eff ect size (p=0.024).Th ere were no signifi cant association between either time from ice slurry ingestion leading to the endurance performance (p=0.751) or environmental temperature and humidity (p=0.505), and the eff ect size.By removing the one study with imputed ambient temperature and humidity from the meta-regression, the regression coeffi cient remained insignifi cant (QModel=1.70,p=0.428), suggesting there is minimal impact of imputation on the results.

Figure 1.
Forest plot of eff ect size from 16 studies that assessed the eff ect of ice slurry ingestion on endurance performance in the heat.A circle represents the eff ect size of a study, and the circle size is proportional to the study's weighting in a random-eff ects meta-analysis.The horizontal line depicts the 95% confi dence interval for each eff ect.Open circle represents insignifi cant eff ect while fi lled circle represents signifi cant eff ect.
Th e eff ect size of the 16 studies was plotted against the inverse of the standard errors (Figure 3).Th ere was little evidence of asymmetry in the funnel plot, indicating a considerably equal distribution of the studies along the horizontal line, hence low probability of publication bias.

Discussion
Th is meta-analysis focused on the practical aspects of ice slurry ingestion and the major fi ndings show that ice slurry ingestion does improve endurance performance in hot environments, and this benefi t is moderate in magnitude (g=0.54).Results from available studies demonstrated that increasing dose of ice slurry ingestion does not add further improvement on endurance performance and the benefi t could be diminished with higher dose.Dispersion in the eff ect size was not signifi cantly associated with timing of ice slurry ingestion or environmental conditions.Th ese fi ndings therefore extend and add new information to the present knowledge regarding ice slurry ingestion on endurance performance in the heat.
Th is meta-analysis found a comparable eff ect on endurance performance between ice slurry ingestion (g =0.54) and other cooling strategies.Th ere are a number of choices of pre-cooling interventions for exercise in the heat: palm cooling, foot cooling, neck cooling, forearm immersion, torso cooling via cooling vests, whole-body cold water immersion, iced towel, fan cooling with mist, cold room, cold water ingestion, and more recently L-menthol application.Previous meta-analytic review has summarized common cooling strategies on exercise performance in the heat: mixed method cooling g=0.72, cold water immersion g=0.49, cold water/ice slurry ingestion g=0.40, cooling packs g=0.40, and cooling vests g=0.19 (Bongers et al., 2015).Cooling strategies during elite competitions are practically limited to cold water/ice slurry ingestion and mist cooling.Ice slurry ingestion is eff ective as a performance enhancement in the heat, and also is very practical.As would be expected, elite athletes consider ice slurry ingestion superior to other cooling strategies (Periard et al., 2017).
Th e meta-regression of dose eff ects revealed a signifi cant association with the eff ect size for endurance performance.Th e total amount of heat transfer is moderated by the ice volume ingested.Th eoretically, a higher dose of pre-cooling ice slurry ingestion yields a higher cooling capacity, which could lead to lower core temperature at the commencement of exercise.None of these 16 studies investigated dose eff ects and the mechanisms for the present fi nding cannot be explained.Interestingly, Ihsan and colleagues suggested the maximal tolerable amount of ice consumable within a 30-min period was 6.8 g kg−1 (Ihsan et al., 2010).A review of fl uid temperature on palatability and fl uid ingestion suggests 10-20°C is the preferred temperature range for beverages used for fl uid replacement during endurance exercise in temperate to warm environments (Burdon, Johnson, Chapman, & O'Connor, 2012).Taken together, this might explain a slower time trial performance, when ice (~10.5 g kg−1 or 0.81 kg of fl uid) was ingested ad Libitum (Maunder et al., 2017).Nonetheless this meta-analysis summarizing the available data suggests that the ergogenic eff ect of ice slurry ingestion on endurance performance is not proportional to higher doses.
Given there is a temporally delayed pattern from external cooling to body cooling (Arngrimsson, Petitt, Stueck, Jorgensen, & Cureton, 2004), it is reasonable to assess whether temporal eff ects associated with ice slurry ingestion could infl uence the ergogenic eff ect.Th e present analysis indicates that the internal cooling eff ect on endurance performance was not time dependent.Ingestion of ice 15 min prior to the exercise eff ectively improved (g=0.75) endurance capacity in the heat (Takeshima et al., 2017).As such, ice slurry ingestion can be recommended for cooling an athlete immediately before exercise and during exercise.
Another novel fi nding of the present review is that the performance eff ect of ice slurry ingestion is not infl uenced by the environmental eff ects.Previous reports have suggested that exercise capacity is progressively impaired as the relative humidity increases (Maughan, Otani, & Watson, 2012).Interestingly, ice slurry ingestion improved (g=0.30) performance in a hot-dry environment (33°C, 25% relative humidity) (Stevens et al., 2013) and the eff ect does not diminish in a hot-humid environment (g=0.42/0.75;30°C, 80% relative humidity) (Takeshima et al., 2017).High ambient temperature and relative humidity are of great importance when considering athletic performance.Th e upcoming Tokyo 2020 Summer Olympics is expected to have a higher wet bulb global temperature than any of previous three Olympics (Kakamu, Wada, Smith, Endo, & Fukushima, 2017) and poses a real challenge to the health and performance of athletes.Ice slurry ingestion can be recommended to elite athletes participating in endurance events, or spinal cord injured athletes with compromised sweating capacity (Price, 2006) as an eff ective countermeasure in the likely hot-humid environments during Tokyo 2020.
In conclusion, this meta-analysis has provided several novel fi ndings.First, it does not support a recent review (Choo et al., 2018) on this topic.Ice slurry ingestion provides a signifi cant benefi t for endurance performance in the heat, and this eff ect is moderate in magnitude (g=0.54).Second, a specifi c new insight from the current literature is increasing ingestion of ice slurry may not further improve the performance eff ect.A dose as low as 1.25 g kg−1 prior to the exercise could be eff ective.Th ird, the performance eff ect is not time dependent and ice slurry can be ingested before and during exercise.Finally, the performance eff ect is not associated with environmental temperature and humidity.Ice slurry ingestion off ers a comparable performance eff ect compared to other more complex cooling strategies and should be highly recommended for endurance events occurring in the heat.

Figure 2 .
Figure 2. Meta-regression analyses exploring potential heterogeneity as a result of: A. dose of ice slurry ingestion, B. time from ice slurry ingestion to the commencement of performance test, C. ambient temperature, and D. relative humidity (rH).The bubbles are drawn with sizes proportional to the contribution of individual studies towards the method-of-moments linear prediction.The solid line represents linear predictions for the eff ect size while the curved lines are 95% confi dence bands.QModel is the measure of variation explained by the covariates.An alpha level of 0.10 indicates signifi cant variation in the eff ect size was moderated by the covariate.

Figure 3 .
Figure 3. Funnel plot displaying the eff ect size and the inverse of standard errors in a random-eff ects meta-analysis.The vertical line marks the weighted eff ect size and the dashed lines represent 2 and 3 sigma intervals.

Table 1 .
Characteristics of the 16 Studies Examining the Eff ect of Ice Slurry Ingestion on Endurance Performance Citation