The “Talk Test” is a practical tool to determine aerobic training zones in overweight and obese patients.

Background: to validate the traditional talk test (TTT) and an alternative talk test (ATT; using a visual analog scale) in overweight/obese (OW-OB) patients and to establish its accuracy in determining the aerobic training zones. Methods: We recruited 19 subjects aged 34.9 ± 6.7 years, diagnosed with overweight/obesity (BMI 31.8 ± 5.7). Every subject underwent incremental cycloergometric tests for maximal oxygen consumption, and TT in a randomized order. At the end of each stage during the TT each subject read out loud a 40 words text and then had to identify the comfort to talk in two modalities: TTT which consisted in answering “Yes”, “I don’t know” or “No” to the question Was talking comfortable?, or ATT through a 1 to 10 numeric perception scale (visual scale analog: VAS). The magnitude of differences was interpreted in comparison to the smallest worthwhile change (SWC) and was used to determine agreement. Results: Agreement between the power output at the VAS 2-3 of ATT and the power output at the ventilatory threshold 1 (very likely equivalent; mean difference -1.3 W, 90 % CL (-8.2; 5.6), % chances for higher/similar/lower values of 0.7/99.1/0.2 %). Also, there was an agreement between the power output at the VAS 6-7 of ATT and the power output at the ventilatory threshold 2 (very likely equivalent; mean difference 11.1 W, 90 % CL (2.8; 19.2), % chances for higher/similar/lower values of 0.0/97.6/2.4 %). Conclusions: ATT is a tool to determine exercise intensity and to establish aerobic training zones for exercise prescription in OW-OB patients.


Background
Worldwide, low levels of physical fitness are associated with an increased risk of all-cause and cardiovascular disease [1], and physical inactivity is responsible for a substantial economic burden [2]. Also, poor cardiorespiratory fitness is an independent risk factor for the development of noncommunicable diseases (NCDs) and cardiovascular disease [3,4]. Noteworthily, regular physical activity reduces the all-cause mortality risk in ~ 14% and is one of the leading non-pharmacologic strategies for the prevention and treatment of obesity (OB) [1]. However, exercise prescription is complicated, depending on duration, frequency, intensity, and type of exercise [5]. Moreover, under the same standardized training prescription, there are evident individual variations in post-training adaptations [6,7]. Possibly, these individual training responses variations are dose-dependent.
Therefore, the control of the physical training load is essential to maximize the benefits associated with health-related exercise [8].
There are several ways of prescribing exercise intensity, and some of these guidelines rely on objective criteria, such as percentages of absolute values of the heart rate reserve (%HRR) or the maximal oxygen uptake (VO2max). However, there is a convincing body of evidence suggesting that the relative distribution of training intensity is regulated more effectively based on the individual metabolic response to training [9,10]. Exercise intensity determined by the individual physiologic and metabolic response (e.g., ventilatory threshold 1 [VT1], ventilatory threshold 2 [VT2], or lactate threshold [LT]) induces homogeneous training response and adaptations to training programs [11].
Besides, is beneficial to identify these thresholds since it allows establishing the three classical training zones: Zone 1, intensity < VT1; Zone 2, intensities between VT1 and VT2; and Zone 3, intensity > VT2 [12], favoring the control of external and internal training load [13,14]. Despite the usefulness of laboratory evaluations individualizing exercise prescription, access to these measurements at the healthcare level is limited due to the cost of its implementation [15]. Therefore, research needs to develop low cost and easy to apply training load control methods to improve access to individualized training programs in a specific population.
Low-cost methods, such as the Talk Test (TT) and the Rating of Perceived Exertion (RPE), have been demonstrated to be of value relative to both performance diagnostics and prescription. Within the last years, the TT has been suggested as a useful surrogate of gas exchange thresholds in a variety of populations [16]. The TT is easy to apply and low-cost tool for intensity monitoring [17], and involves an individual reading a similar text during exercise and then being asked if he or she can still speak comfortably [16]. However, the traditional talk test (TTT) only has three options for the question "was talking comfortably?" and lacks quantitative psychometric properties, being questioned as a substitute for objective physiological measures for prescribing individual training exercise [18].
The main objective of the study was to validate the TTT and an alternative TT (ATT; using a visual analog scale) in overweight/obese (OW-OB) patients and to establish its accuracy in determining the aerobic training zones, previously described. We hypothesized that ATT is a valid tool to establish aerobic training zones in OW-OB patients.

Participants characteristics
A total of 19 subjects with a nutritional diagnosis of OW-OB according to BMI ≥ 25 kg/m 2 , physically inactive according to the World Health Organization (WHO) classification [19], and no diagnosis of NCDs participate in the study. Before the evaluations, the subjects signed an informed consent approved by the Scientific Ethics Committee of the Universidad Finis Terrae (resolution Nº21/2017).
All procedures were performed in compliance at Helsinki Declaration principles for human experiments.

Study Design
The participants visited the laboratory in five separate days. Subjects arrived between 08:00 and 10:00 am and were evaluated in a randomized order for the following procedures: body composition, cardiorespiratory fitness test, TTT, and ATT.

Body Composition
Fat, lean, and fat-free body mass were measured by double energy X-ray absorptiometry (DEXA) using manufacturer-supplied algorithms (Total Body Analysis, version 3.6; Lunar, Madison, WI, United States). The general characterization of subjects is presented in Table 1.

Cardiorespiratory Fitness
After a 5-min warm-up at 50 watts and a constant cadence of 55 ± 5 rpm, the participants performed a maximal incremental test on an electronic automatized cycle-ergometer (Cyclus2, Germany). An initial workload of 40 watts (W) was used, with increments of 15 W (women) and 20 W (men) every 1 min until exhaustion. The test was performed with a constant cadence of 55 ± 5 revolutions per minute (rpm). Gas exchange was recorded continuously with a portable breath-to-breath gas analyzer (Cortex Metalyzer 3B, Leipzig, Germany) and was calibrated according to the manufacturer's instructions before each trial. Pulmonary ventilation (VE), oxygen uptake (VO2), expired carbon dioxide (VCO2), and respiratory exchange ratio (RER) were averaged over 10 s in the mixing chamber mode, with the highest 30 s value (i.e., three consecutive 10 s averages) used in the analysis.
VO2max was determined according to previously established criteria [20]: (i) plateau in VO2 (i.e., increase < 150 ml·min − 1), (ii) RER > 1.1, and (iii) ≥ 90% of theoretical maximal heart rate. The VO2max was expressed both as absolute values (L·min − 1) and relative to body mass (ml·kg − 1·min − 1). The power output at VO2max (pVO2max) was determined as the minimum workload at which VO2max was reached. Ventilatory threshold 1 (VT1) and ventilatory threshold 2 (VT2) were identified separately by three researchers according to the following criteria (24): an increase in VE/VO2 and PETO2 without a concomitant increase in VE/VCO2 for VT1, and an increase in VE/VO2 and VE/VCO2 and a decrease in PETCO2 for VT2. The cardiorespiratory fitness and ventilatory threshold are shown in Table 2.

Talk Test
After a 10-minute warm-up, subjects performed an incremental test on an electronic automatized cycle-ergometer (Cyclus2, Germany). The protocol considered load (W) increments every 3 minutes, the time necessary to stabilize ventilation, primary variable for voice production [21,22]. During the last 30 seconds of each stage, out loud reading of 40 words from the text "Lectura del Abuelo" was requested. Two methods evaluated the ability to converse during exercise: (i) traditional talk test (TTT) by answering "yes", "no", or "I do not know" to the question "was talking comfortably?", and (ii) alternative talk test (ATT) using a 1 to 10 visual analog scale (VAS) [23]. Both, text "Lectura del Abuelo" and VAS are shown in Fig. 1.

Statistical Analyses
Data in the text and figures are presented as mean ± SD and 90% confidence limit/interval (CL/CI). All data were first log-transformed to reduce bias arising from nonuniformity error. The magnitude of differences was interpreted in comparison to the smallest worthwhile change (SWC) (0.6 x pre-tests between-subjects SDs) [24]. This SWC was setting as the equivalence region, representing about one stage of difference during the incremental test and was used to determine agreement. The probability of any substantial difference or realistic equivalence relative to the predefined target values was interpreted using the following scale: <0.5%, most unlikely; 0.5-5%, very unlikely; 5-25%, unlikely; 25-75%, possibly; 75-95%, likely; 95-99.5%, very likely; >99.5%, most likely [25]. Effects were declared relevant if the outcome probability was likely (≥ 75%) (i.e., methods were considered in agreement and, therefore, interchangeable). Statistical analysis was performed with the "mbir" package of the R software [26].

Results
We recruited 34 participants, of which six were excluded because they did not meet the criteria for entering the study, and nine did not complete all testing procedures. The final analysis, therefore, included 19 patients who completed the evaluations.

Agreement Between Traditional Talk Test And Ventilatory Thresholds
Results of the equivalence tests between TTT and ventilatory thresholds are presented in Fig. 2 and Table 3. Evidence for an agreement was observed between the power output at the "first no" (FN) and the power output at the ventilatory threshold 2 (most likely equivalent; mean difference − 2.9 W, 90% CL (-10.9; 5.1), % chances for higher/similar/lower values of 0.1/99.9/0.0%). There was no agreement between the power output at the "last yes" (LY) and the power output at the ventilatory   Figure 2B, there was an agreement between the power output at the VAS 6-7 of ATT and the power output at the ventilatory threshold 2 (very likely equivalent; mean There was an agreement between the power output at the VAS 6-7 of ATT and the power output at the "first no" of TTT (very likely equivalent; mean difference -13.9 W, 90 % CL (-18.7; -9.1), % chances for higher/similar/lower values of 1.9/98.1/0.0 %).

Discussion
The regulation and control of exercise intensity are some of the most challenging parts of exercise prescription. There are several ways of prescribing exercise intensity, and some of these recommendations are based on objective criteria, such as percentages of absolute values of the heart rate reserve (%HRR) or the maximal oxygen uptake (VO2max). However, recent investigations propose a more individualized exercise prescription to personalize a training regime based on individual metabolic responses and, therefore, enhance the potential benefits of regular physical activity [11]. Therefore, the goal of this investigation was to prove the usefulness of the TTT and/or ATT as a low-cost tool to determine exercise intensity and establish aerobic training zones for exercise prescription in OW-OB patients.
Our main finding shows that the three aerobic training zones delimited by VT1 and VT2 could be established through the TT, primarily through the ATT. Regarding the TTT, previous studies have shown an association between VT1 and the last stage of the TT where talking was comfortable in healthy subjects [27,28,16] and between VT2 with TT stages where comfort to talk is lost in patients with heart diseases [29,30]. However, these previous studies used the VO2 values to compare intensities between the TTT and ventilatory thresholds. This methodology does not allow obtaining external load values (e.g., watts) that can be used for prescribing and controlling aerobic training.
In our study, the TTT failed to determine the transition from zone 1 to zone 2 because we found no agreement between the power output of the different answers related to the TTT and the power output at VT1. The transition threshold between zone 2 and zone 3 could be established with the power output at the first stage where the answer was "no," which was most likely equivalent to the power output at VT2. This lack of consistency with previous results could be related to the statistical analysis. The previous studies used correlation analysis (e.g., Pearson correlation) [16,27], which focuses on the association of changes in two outcomes that often measure quite different constructs [31]. Our study used agreement analysis, which measures the degree of concordance in the results between two or more assessments of the variable of interest and assumes that the variables measure the same construct [31], being agreement analysis better to assess if methods are interchangeable [32]. To our knowledge, there is only one study in well-trained cyclists that measure agreement between workload at VTs and TT. Rodriguez et al. found agreement between the power output at the first stage where the answer was "I do not know" and the power output associated with VT1 [33].
Also, they found an agreement between the power output at the first stage where the answer was "no" and the power output associated with VT2, results that partially disagree with our recent findings in OW-OB patients.
Regarding the ATT, the absence of psychometric properties of the TTT may induce an under or overestimation of the degree of talking comfort during physical exercise in physically inactive persons. The ATT would allow identifying the "difficult to talk" with numeric magnitudes by giving a quantitative variable to the TT [34,35,18]. Speech production during exercise is associated with changes in physiological variables related to exercise, being the consequence of the need to adapt the breathing pattern compatible with speech production. Accordingly, Rotstein et al. found a significant association between VO2, HR, and VE responses and the ratings of perceived speech production difficulty. Our results agree with these previous findings, showing that the intensity (power output) associated with VT1 is very likely equivalent to the last stage where talking was "very easy" (VAS 2-3), allowing to determine the threshold to zone 1. The power output at the first stage where talking was "hard" (VAS 6-7) was very likely equivalent to the power output at VT2 (Table 3), which allows determining the threshold to zone 3. The power output where talking was "somewhat hard" (VAS 4-5) was most likely higher than the power output at VT1 and most likely lower than the power output at VT2, thus, representing the intensity related to zone 2 (Fig. 2B). Taken together, these results showed that the ATT could be used to determine exercise intensity and establish aerobic training zones for exercise prescription in OW-OB patients. The main limitation of clinical context is to have a low-cost tool to prescribe physical exercise. TT is a simple tool that could be applied to large populations due to low-cost and easy application. However, further research is needed to determine the effect of endurance training controlled with TT on obese people.

Conclusions
ATT is a low-cost and easy to apply tool to determine exercise intensity and to establish aerobic training zones for exercise prescription in OW-OB patients. The TTT could under or overestimate the physical effort in patients diagnosed with OW-OB, specifically at the training zone 1.

Ethical Approval and Consent to participate
The Scientific Council of University Finis Terrae approved the research (Nº21/2017). 58076/14-11-2018). All the participants gave written consent.

Consent for publication
All the gave written consent for publication of data.

Availability of data and materials
Please contact author for data requests.

Funding
This research received no external funding.

Authors' contributions
IOC has carried out the conceptualization, methodology, recruitment of participants, data collection, as well as the statistical study, writing -original draft. HCK has been the reviewer of the work helping in the statistical analysis and perfection of the methodological aspects. CBJ, RMV and RGP has carried out data collection and data curation. CS has collaborated in the project administration, methodology, writing and guiding for its edition, and in the perfection of the methodological aspects. All authors have contributed to proof-reading of the manuscript and have approved the final article.

Figure 1
Text "Lectura del Abuelo" and visual scale analog.