Reliability of a wearable sweat rate monitor and routine sweat analysis techniques under heat stress in females

Abstract

Introduction: The aim of the study was to evaluate the reliability of five different sweat analysis techniques which measure; whole body sweat rate [WBSR], local sweat rate [LSR] (via technical absorbent [TA] method and KuduSmart® monitor), sweat conductivity [SC] and sweat gland activation [SGA] in a female population when exercising moderately under heat stress. Methods: Fourteen females (age; 26 ± 7 years, body mass; 66.5 ± 7.6 kg, height; 167.1 ± 6.4 cm) completed a preliminary threshold walking test (to determine exercise intensity) and two main trials, separated by 2 days. Main trials consisted of 30-minutes seated rest in the environmental chamber (35°C, 50% relative humidity) in an upper body sauna-suit, before its removal, and walking at a moderate intensity (4 metabolic equivalents) for 30-minutes (speeds ranged from 4.8-6.5 km.hr-1). WBSR was measured via nude mass pre and post exercise. The TA and Tegaderm patches (for sweat sodium chloride) were placed on the back, forearm and chest for the entire 60-minutes, replicated for all participants for both trials. SGA was assessed following the 60-minute trial and the KuduSmart® monitor was placed on the left arm for the 30-minutes of exercise. Results: WBSR, LSR methods and SC demonstrated no difference between trials (p >0.05), good agreement (within limits), strong correlations (r ≥ 0.88) and low typical error of measurements [TEM] (<0.04L.min-1 , 0.13 mg.min-1 .cm-2 and 8 mmol.L-1 , respectively). SGA method showed moderate intra-class correlation (r=0.80), with high TEM (5 glands) and large limits of agreement. Conclusion: Sudomotor function is reliable, as demonstrated by good reliability, small TEM and strong correlations. The use of these sweat techniques is appropriate and practical in females who are exercising at moderate intensity under heat stress, and so, may aid future interventions. SGA shows larger variation and should be used with caution.

Validity of a wearable sweat rate monitor and routine sweat analysis techniques using heat acclimation

“…Numerous wearable devices provide precise measurements that demonstrate sufficient reliability for experimental and clinical purposes. This was recently reinforced for the KuduSmart® (Relf et al 2019), demonstrating an error of 0.08 mg.min -1 .cm -2 , coefficient of variation of 13.5% and intra-class correlation of 0.88.…”

“…Both Tegaderm and TA patches were placed on the dorsal surface of the three sites (lower back, chest and forearm) and standardised for all participants. Locations of both TA and Tegaderm patches are described in full detail by Relf et al (2019).…”

“…Although an array of other sweat analysis techniques have been deemed reliable alongside the KuduSmart® monitor (Relf et al, 2019), limited evidence exists of their validity (Lemon et al, 1986;Boisvert and Candas, 1994). Whilst reliability of a measurement is important in research designs, it cannot guarantee valid conclusions (Krippendorff, 2008), and therefore the occurrence of Type II errors (Hayden et al, 2004).…”

“…Sweat rate, and therefore, heat loss capacity is highly individualised, due to the quantity and size of sweat glands among different people (Shibasaki et al, 2006;Tansey and Johnson, 2015). Thus, with the error of each technique and reliability now established for a range of different sweat analysis markers (Relf et al, 2019), enhanced data surrounding sudomotor alterations via HA, specifically LSR is warranted to provide validation in the literature.…”

“…It was hypothesised that sweat rates will increase and sodium chloride concentrations will decrease, demonstrating sensitivity to sudomotor adaptation, following STHA and acceptable validity of techniques. It was also hypothesised that each sweat analysis technique would exceed the technical error of measurement previously acknowledged by Relf et al (2019); WBSR > 0.04L/hr -1 , KuduSmart® > 0.08 mg.min -1 .cm -2 , LSR > 0.13 mg.min -1 .cm -2 , SC > 8 mmol.L -1 and be significantly different (p < 0.05) between trials.…”