Aplicação e propriedades de medida do teste da fala em pacientes cardiopulmonares: revisão sistemática

Abstract

Background:

The talk test (TT) evaluates the exercise intensity by measuring speech comfort level during aerobic exercise. There are several application protocols available to assess individuals with cardiopulmonary diseases. However, the measurement properties of the TT were not systematically reviewed yet.

Methods:

A systematic review was developed, registered (CRD420181068930), and reported according to PRISMA Statement. Randomized clinical trials, cross-sectional studies, or series cases were identified through multiple databases and were selected if they presented concomitant speech provocation and an exercise test. Included studies were evaluated based on methodological quality (adapted New Castle-Ottawa Scale), descriptive quality (STROBE Statement), and risk of bias (COSMIN bias risk scale).

Results:

Ten studies were included. Seven studies presented moderate to high quality and the majority presented good scores according to the STROBE statement. Four hundred and fourteen subjects performed the TT, the majority being patients with coronary artery disease. The test validity was supported by the included studies. Talk Test reliability was considered satisfactory, although only one study presented an adequate reliability analysis. The studies found a correlation between the last positive stage of the TT with the first ventilatory threshold. Workload, oxygen uptake, and heart rate in the last positive stage of the TT were not different from the same parameters related to the first ventilatory threshold.

Conclusions:

The evidence indicates that the TT is suitable as an alternative tool for the assessment and prescription of exercise in individuals with cardiovascular diseases. The stage when the individual is still able to speak comfortably is suggested as the intensity for aerobic exercise prescription. As there is still no well-defined and fully explored TT protocol, caution is required when interpreting the TT results.

1. Introduction

Exercise training is recommended for cardiovascular disease (CVD) patients aiming to restore their maximum level of activity and promote their cardiovascular adaptation [¹, ²]. Usually, the exercise intensity is prescribed based on parameters from the cardiopulmonary exercise test (CPX), which is recommended on a regular basis for evaluation of adults with chronic diseases [³]. The CPX should be applied whenever there is clinical indication and availability [⁴]. However, some more accessible tools, at lower cost, have been used for prescription and monitorization of the exercise training such as the Borg Scale [⁵], and the Talk Test [⁶]. These alternative tools are important especially when the CPX is not available, potentially providing useful parameters to exercise prescription, facilitating the applicability and dissemination of the aerobic exercise intervention. They can also be useful for aerobic exercise prescription to cardiopulmonary patients in a home-based setting or when social distancing is necessary, as it has been experienced in the COVID-19 pandemic. Besides that, these prescription resources may be applied for the large part of the global population that is sedentary and presents a low level of physical activity.

The Talk Test (TT) is a non-invasive procedure that assesses metabolic stress through speech provocation, indicating the ideal intensity of aerobic exercise [⁶, ⁷]. Methodologically, the TT is divided into progressive stages with speech provocation at the end of each stage. The subject is asked to recite a paragraph and to answer the question: “Are you able to speak comfortably?”. There are three answer options: “Yes”, for positive stages; “More or less”, for equivocal or uncertain stages; and “No”, which corresponds to the negative stage, a condition used to finish the test [⁸, ⁹]. Another strategy for speech-provoking, which is less commonly used, is counting as a continuous measure of ventilatory stress. The subject is required to count out loud at their normal rhythm, with the number of counts attained in one breath used as a measure [⁷, ¹⁰].

There is a conflict between metabolic and phonetic functions, caused by increased gas exchange and reduced expiratory time during exercise [¹¹]. Studies have observed that the first ventilatory threshold identified during CPX (e.g., the moment when pulmonary ventilation starts to increase disproportionally in relation to oxygen consumption), may be associated with the moment when this conflict is insurmountable. This usually happens during the uncertain or negative stage of the TT [¹², ¹³]. Furthermore, the relationship between exercise intensity and physiological thresholds in the TT seems to be maintained even with different strategies for speech provocation [¹²].

The application of the TT has been explored since 2002 in different populations, such as athletes [¹³, ¹⁴, ¹⁵], men with prostate cancer [¹⁶], overweight and obese patients [¹⁷], and patients with CVD [¹⁸]. It is presently a recommended test in the current guidelines for CVD patients even without a clear definition of the most appropriate protocol [¹⁹, ²⁰]. In addition, its applicability to patients with pulmonary disease is still unknown. This increased interest and usefulness of the TT as an evaluation and prescription tool for patients with cardiopulmonary diseases calls for further knowledge about its applicability, protocols, and properties. Therefore, the present study aims to synthesize the application methods and measurement properties of the TT in individuals with cardiopulmonary diseases.

2. Materials and Methods

This systematic review was registered in the Prospective Register of Systematic Reviews (PROSPERO) under code CRD42018106893 and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA statement) [²¹].

2.1 Data Sources and Searches

The search was performed in CINAHL, EMBASE, LILACS, Pubmed, and Scopus databases (1995–2022), based on PECO strategy: Population, individuals with chronic cardiopulmonary diseases; Exposure, TT or similar; Comparison, CPX, or another available exercise test; Outcome, the application methods of the TT as a primary outcome (such as the ergometer used, the speech provocation method, stages duration and exercise protocol) and the measurement properties as secondary outcomes. Additionally, authors screened the bibliography of the selected manuscripts for full reading and performed a search on Research Gate and Google Scholar. The full search strategy is available in Supplemental Material (Supplementary Table 1).

2.2 Studies’ Selection

After performing the literature search, retrieves were organized using the reference manager software Mendeley® (Mendeley Desktop, version 1.19.8, Oxford, UK). After removing duplicate titles, the selection process was performed independently by three blinded researchers. The first stage consisted of titles and abstracts screening. At the end of this step, the independent spreadsheets were joined, and the answers were compared. A senior researcher was consulted to resolve disagreement. Next, the full reading of the selected scientific articles was done following the same process to incorporate the manuscripts that met the inclusion criteria.

Therefore, the scientific articles that met the following criteria were included: samples comprised of individuals with chronic cardiovascular and/or pulmonary diseases, of both genders, aged 18 years or older; manuscripts characterized as randomized clinical trials, cross-sectional studies, or case series, without language restriction, indexed in the databases of interest or retrieved from hand search. Scientific articles that did not present speech provocation, literature reviews, protocols, and study designs were excluded.

2.3 Data Extraction and Quality Assessment

Characteristics and main findings of the included studies, as well as the characteristics of the TT application protocols and their measurement properties, were extracted and organized in standardized spreadsheets. Measures of validity (TT stages versus CPX variables), reliability (test-retest or intraclass correlation coefficient), measurement error, and responsiveness were the variables of interest, based on the COSMIN Statement [²²].

The methodological quality was analyzed using the adapted New Castle-Ottawa Scale (NOS). This scale consists of seven items, subdivided into three domains which are rated by stars [²³, ²⁴]: selection (5-stars), comparability (2-stars), and outcome (3-stars). Studies were scored as low (5 stars or less), moderate (6 or 7 stars), or high-quality (8 or more stars) [²⁴].

The descriptive quality was verified applying the STROBE Statement, which contains 22 items divided into six categories [²⁵, ²⁶]. One point was assigned to each item present in the study. Besides, the COSMIN Risk of Bias Scale [²²] was used. This scale is composed of 10 boxes that can be chosen separately according to the evaluation needs and the properties explored by the studies. In this systematic review, for the bias risk analysis, we selected boxes 3 and 8 (validity) and box 6 (reliability). Each box has up to five response options in descending order: very good, adequate, doubtful, inappropriate, and not applicable. The scientific articles’ rating was based on the lowest response for each evaluated box.

3. Results

3.1 Selected Studies

Fig. 1 shows the process of manuscript selection. The searches retrieved 913 references and five were retrieved from grey literature. Researchers identified 22 references from the main search for a full-text assessment, plus the five references from hand search. The main reason for excluding studies after the full-text assessment was the population (e.g., not cardiopulmonary patients). The study design was also a common reason, as there was no original data on the application of the TT, but a review of the literature for instance. After the full-text assessment, ten studies met the inclusion criteria and were analyzed.

3.2 Studies Characteristics

Nine studies were classified as observational and one as a validation study (Table 1, Ref. [¹⁸, ²⁷, ²⁸, ²⁹, ³⁰, ³¹, ³², ³³, ³⁴, ³⁵]). Four hundred and twenty-six individuals underwent the TT. From those, the vast majority (more than 99%) were CVD patients, with a predominance of coronary artery disease. Only one study included subjects with chronic pulmonary disease in their sample [³²]. The sample size varied from 10 to 85 patients included in each study (mean 42.6 ± 25.18 patients).

3. Results (cont.)

Overall, the TT protocol used was a reproduction of the maximal exercise test (Table 2, Ref. [¹⁸, ²⁷, ²⁸, ²⁹, ³⁰, ³¹, ³², ³³, ³⁴, ³⁵]). Seven studies applied a single protocol to assess all subjects, without adjusting or individualizing the load increment. The tests were performed on a cycle ergometer (60%) [²⁹, ³¹, ³², ³³, ³⁴, ³⁵], or on a treadmill (20%) [¹⁸, ²⁸]. In one study (10%) individuals could choose the ergometer test based on their preference and exercise capacity [³⁰], while another study (10%) applied the TT on both a treadmill and an indoor track [²⁷].

3. Results

The exercise test protocol consisted of one-minute stages [²⁹, ³¹, ³², ³³, ³⁴], two-minute stages [¹⁸, ²⁸, ³⁰], and three-minute stages [³⁵]. The study that applied two different protocols, had one with two-minute stages and the other with three-minute stages [²⁷]. The speech provocation was done within the last 10 to 30 seconds of each stage and reciting a standard paragraph was the strategy applied in all studies. Only one study applied an additional challenge method [²⁷].

The speech provocation was based either on paragraphs that were well-known to the individuals, such as a 30-word Danish text passage [²⁹, ³¹, ³²], the “Pledge of Allegiance” (a 31-word paragraph widely known within the US population) [¹⁸, ²⁷, ²⁸, ³⁰] and the 19th article on the religious freedom of the Italian Constitution [³⁵]. Two studies did not detail the paragraph, just mentioned that had 30 words [³³, ³⁴].

As can be observed in Table 2, most studies used the question, “Can you still speak comfortably?” to evaluate the TT stages, giving three answers as options (Yes, Uncertain, or No). Brawner et al. [²⁷] used previously recorded questions as a method of speech provocation in the application of the TT in an indoor track. The individuals listened to the questions through a portable music player and answered these to understand their speech comfort level. Another method applied was evaluating whether the individual could read the paragraph in 10 seconds at a constant pace, without looking breathless. If the individual could not complete the reading requiring further inspiration, the TT was defined as negative and the test was finished [³³].

3. Results (cont.)

3.3 Measurement Properties

The Talk Test validity has been investigated in only four studies [¹⁸, ²⁷, ³⁰, ³⁵]. In general, it was identified that the workload, the oxygen uptake (VO2), or the heart rate in the last positive stage (TT+) was not different from the same parameters related to the first ventilatory threshold (VT1). When subjects were on TT+ or the first uncertain stage of the TT (TT±) they were on or below VT1, and when participants reached first negative stage (TT-), they were surely above their VT1. The respiratory compensation point, also known as second ventilatory threshold (VT2), was not explored.

From these findings, the studies proposed the TT+ stage as a prescription parameter and identified its usefulness in most of the individuals evaluated [¹⁸, ²⁷, ³⁵]. Sorensen et al. [³⁴] found correlations between the TT stages and VT1 from 0.37 to 0.60. Also, it was observed that individuals are unlikely to show electrocardiographic signals of myocardial ischemia at the load corresponding to the time they can still speak comfortably (TT+) [²⁸].

The concurrent validity was analyzed from the correlation between VT1 and the stages of TT (Table 2) [¹⁸, ³⁵]. The structural validity was classified as doubtful [³⁵] and as inappropriate [¹⁸], while the criterion validity was rated very good in both studies (Supplementary Table 3). The reliability was evaluated in four studies [²⁹, ³¹, ³³, ³⁵] and the values found were high (intraclass correlation coefficient; ICC >0.8).

Nielsen et al. [³¹] found good reliability for both the individuals and the physiotherapists’ evaluations after an incremental cycle ergometer test associated with TT. Similar results were observed in individuals with lacunar strokes [²⁹], and myocardial revascularization [³⁵]. Although an acceptable ICC value was found, Petersen et al. [³³] observed low reliability among evaluators when the TT was assessed only by physiotherapists and not by participants. The reliability values were classified as adequate [³³] in only one study, two studies presented doubtful [²⁹, ³¹] reliability, and one was classified as inadequate [³⁵] (Supplementary Table 3).

3.4 Quality Assessment

According to NOS, the methodological quality was considered high in three studies [²⁷, ²⁹, ³²], moderate in four [²⁸, ³¹, ³³, ³⁵], and low in other three studies [¹⁸, ³⁰, ³⁴] (Table 1 and Supplementary Table 2). The descriptive quality scores ranged from 15 to 22, with 22 being the maximum value achieved (STROBE Statement checklist, Table 1). Study size and funding were the items less described. All studies met the requirements for the categories from item 2 (background) to 9 (bias), and only one study [³⁴] did not complete the four discussion categories.

4. Discussion

To our knowledge, this is the first review to systematically synthesize the application methods and the measurement properties of the TT in individuals with chronic cardiopulmonary diseases. The present review included ten studies that showed the feasibility of the TT application and its usefulness in individuals with chronic cardiopulmonary diseases, mainly in coronary artery disease patients. Most of the studies included a small sample size and replicated the maximal exercise test to the TT protocol. The TT+ was related to the intensity of aerobic exercise prescription corresponding to the VT1. However, the studies present a divergence between the protocols of the TT application and the findings regarding the validity and reliability of the test.

There are several studies in the literature that support the relationship between the TT stages and ventilatory thresholds [¹⁸, ³⁶, ³⁷]. However, this relationship may change depending on the population. When applied to healthy individuals, the TT± stage has been correlated with the VT1 [³⁷]. Quinn & Coons found that when exercising at the TT+ level, individuals were within the recommended training zone based on %VO2max, % of maximum heart rate and subjective rating of perceived exertion [³⁸].

For both CVD patients and healthy individuals, there are discrepancies in the responses due to the different forms of the TT application. The method of speech provocation, mainly the paragraph length, can influence the response [¹²]. Schroeder et al. [³⁹] reported that long paragraphs may result in lower VT1, for instance, but with lower prediction errors. This is aligned with the studies included in this review, which did not show much variation in this respect as most studies used a 30-word paragraph. The duration of each TT stage, when the load is increased and the subject answers the speech provocation, is also a component of the TT protocol that can affect subjects’ response. Stages with a longer duration produce less prediction error. The load reached by the individuals in the points equivalent to the ventilatory thresholds significantly decreases when the stages duration increase [⁴⁰].

Despite the diversity in the TT protocols and the resulting difficulty in arriving at conclusions, the reliability of the TT was satisfactory [²⁹], both for the assessment by the individual and the physiotherapist [³¹]. Good results regarding the TT reliability in the context of cardiovascular rehabilitation (CR) were found [⁴¹, ⁴²]. However, Petersen et al. [³³] found poor reliability when the test was administered by more than one evaluator, even with acceptable ICC values.

Nevertheless, when assessed by the COSMIN scale [²²], only one study achieved the “adequate” rating for its reliability assessment design [³³]. It is noteworthy that the reliability can be increased by standardizing the TT protocols (e.g.,: ergometer, load increase pattern, speak provocation method, stage duration,…), which would reduce the variation of the test parameters. Among the measurement properties, it is noted that the validity and reliability of a test are critical to ensure that the same can be replicated and used to relate to clinical outcomes [⁴³].

Although systematic reviews are not available on the present topic, literature and narrative reviews shed light on the gaps and strengths around the TT. In a narrative review of major contributions to the literature, Foster et al. [¹²] presented the TT applicability for exercise assessment and prescription. The review showed strong evidence regarding the heterogeneity of TT protocols, also presented here in individuals with cardiopulmonary diseases. In agreement with other review articles [⁶, ⁴⁴, ⁴²], Foster et al. [¹²] demonstrated the physiological mechanisms involved in the relationship between the TT stages and ventilatory thresholds, reinforcing that TT can allow aerobic exercise practice at optimal intensities when the intensity is adjusted to the highest level of comfortable speech.

The TT was well tolerated by individuals with ischemic heart disease, and its application is safe and effective [³⁵] with minimal measurement errors [³¹]. Its good clinical applicability is related to its easy administration and low cost [⁶]. This can potentially impact the CR context, in which the exercise component is very beneficial to patients with CVD [¹, ²⁰]. Although there are still barriers to CR implementation, its importance has been well consolidated in the literature, showing a positive impact on several risk factors and on reducing the risk of adverse health events [¹, ¹⁹, ²⁰]. Among the parameters used for exercise prescription (frequency, intensity, time, type), the intensity seems to be the one that most requires attention so that the benefits of the exercise component of CR may be acquired. The TT can be an alternative tool that has the potential to facilitate exercise prescription and self-monitoring in the recommended target intensity [⁶, ¹⁸, ⁴²]. As highlighted by Saini et al. [⁴²], the TT can be useful especially in contexts like India, where low-cost and more accessible tools are necessary to overcome some causes of the CR underutilization.

Investigation of the TT in future research may include establishing standards for the application protocol in individuals with cardiovascular and pulmonary disease, as well as assessing responsiveness after rehabilitation programs. Because of the exploratory nature of this study, future research is warranted. Furthermore, it is necessary to evaluate the tolerance and validity of the test when applied exclusively to individuals with pulmonary diseases, since these individuals may present some peculiarities related to dyspnea resulting from dynamic hyperinflation [⁴⁵].

The present study has some limitations. Although it was not part of the study planning and purpose, it was not possible to meta-analyze the results due to the lack of similarity in the analysis and the presentation of the findings by the included studies. Besides, the tool used to evaluate the quality of studies may not have been sensitive enough, underestimating the classification of most studies concerning their methodological quality (low and moderate). Although the population included by the studies covered in this review is very representative of cardiovascular rehabilitation patients, the present study could not address the impact of functional limitations and various disease stages, affecting the validity of the TT. Evidence is missing regarding the applicability of the talk test in a more heterogeneous population, or as cited previously, pulmonary disease patients that present different physiological responses to exercise. As mentioned previously, future research is warranted, especially to explore the validation of an individualized talk test protocol.

5. Conclusions

Our findings reveal important implications for clinical practice and research. Until now, there is no well-defined and in-depth explored TT protocol. The most common protocol is the graded talk test, replicating the maximal exercise test. This protocol is applied on a cycle ergometer, with one-minute stages and asking the subject to read a standardized paragraph to evaluate the speech comfort level. Because of the heterogeneity in protocols and findings, caution is required when applying and interpreting it. The evidence indicates that the TT is suitable, safe, and feasible for assessment and prescription of exercise in individuals with CVD, showing no case of electrocardiographic evidence of myocardial ischemia. The TT+ and TT± stages can be indicated to prescribe exercise when targeting the VT1. The TT is a valid and low-cost tool that has the potential to be applied in clinical practice.