The Glittre-ADL Test Cut-Off Point to Discriminate Abnormal Functional Capacity in Patients with COPD

ABSTRACT

The study objective was to determine a cut-off point for the Glittre activities of daily living (ADL)test (TGlittre) to discriminate patients with normal and abnormal functional capacity. Fifty-nine patients with moderate to very severe COPD (45 males; 65 ± 8.84 years; BMI: 26 ± 4.78 kg/m²; FEV₁: 35.3 ± 13.4% pred) were evaluated for spirometry, TGlittre, 6-minute walk test (6 MWT), physical ADL, modified Medical Research Council scale (mMRC), BODE index, Saint George's Respiratory Questionnaire (SGRQ), and COPD Assessment Test (CAT). The receiver operating characteristic (ROC) curve was used to determine the cut-off point for TGlittre in order to discriminate patients with 6 MWT < 82% pred. The ROC curve indicated a cut-off point of 3.5 minutes for the TGlittre (sensitivity = 92%, specificity = 83%, and area under the ROC curve = 0.95 [95% CI: 0.89–0.99]). Patients with abnormal functional capacity had higher mMRC (median difference 1 point), CAT (mean difference: 4.5 points), SGRQ (mean difference: 12.1 points), and BODE (1.37 points) scores, longer time of physical activity <1.5 metabolic equivalent of task (mean difference: 47.9 minutes) and in sitting position (mean difference: 59.4 minutes) and smaller number of steps (mean difference: 1,549 minutes); p < 0.05 for all. In conclusion, the cut-off point of 3.5 minutes in the TGlittre is sensitive and specific to distinguish COPD patients with abnormal and normal functional capacity.

KEYWORDS:

• Activities of daily living
• outcome evaluation
• outcome assessment (health care)
• functional capacity

Introduction

Patients with chronic obstructive pulmonary disease (COPD) often have limitations to perform activities of daily living (ADL) (1). This impairment of functional status is directly related to the number of exacerbations and the risk of death (2), and, for this reason, the evaluation of this outcome is an essential component of pulmonary rehabilitation programs (1). Among the methods used for its evaluation are field tests, capable of objectively simulating ADL.

The 6-minute walk test (6 MWT) stands out as the most used to evaluate the functional capacity of COPD patients; it is a valid, reliable, simple, and low cost test (3). Furthermore, it has several criteria that allow the interpretation of results, such as reference values (4), minimal important difference (3), and cut-off points related to the prognosis of COPD (5, 6). When using the percentage of predicted values, a distance traveled below 82% may be considered abnormal (7).

Despite all its advantages, the 6 MWT only measures the walking activity, making it impossible to evaluate the limitation in other activities that are problematic for patients with COPD. Multiple task tests may better reflect how ADL are performed (8). The ADL-Glittre test (TGlittre) was developed with the objective of evaluating the ADL reported as those more limited by COPD patients and includes four tasks: sitting down and getting up from a chair, walking on the flat, walking up and down 25 stairs and moving objects from one shelf to another (9). The test has already proved to be valid (9), reliable (10), capable of reflecting the perception of functional limitation (11), physical ADL (PADL) (12), and health status (11) of COPD patients and is responsive to a pulmonary rehabilitation program in this population (9). However, there are still no criteria for interpretability, such as a cut-off point to identify patients with impaired functional capacity.

Therefore, the objective of the present study was to determine a cut-off point for the TGlittre to discriminate patients with normal and abnormal functional capacities, based on the cut-off point of 82% pred of the 6 MWT (7).

Methods

COPD patients referred to the Núcleo de Assistência, Ensino e Pesquisa em Reabilitação Pulmonar (NuReab) participated in the study. The study included patients with a clinical diagnosis of COPD confirmed by spirometry (post-bronchodilator FEV₁/FVC < 0.70) with a disease severity of moderate to very severe (13) aged 40–80 years; with clinical stability in the month prior to the protocol; and optimized medication. The following exclusion criteria were adopted: presence of associated diseases such as cardiomyopathies, incapacitating neuromuscular and orthopedic diseases and other respiratory conditions; cessation of smoking less than six months ago; exacerbation of COPD during the protocol; hospital admission in the 12 weeks prior to the protocol; and participation in the pulmonary rehabilitation program within the last six months. This was a cross-sectional study approved by the Human Research Ethics Committee of the Universidade do Estado de Santa Catarina (UDESC)—Florianópolis (SC), Brazil (Protocol no. 222/2011).

Protocol

The protocol consisted of three days of evaluation. On the first day, data on pulmonary function, the COPD Assessment Test (CAT), the Saint George's Hospital Respiratory Questionnaire (SGRQ), and the modified Medical Research Council scale (mMRC) were collected. The 6 MWT and the TGlittre were performed in the subsequent days (days 2 and 3). The BODE index was determined (14).

Pulmonary function by spirometry using an EasyOne portable spirometer (NDD Medical Technologies®, Zurich, Switzerland) was assessed, and its calibration was checked before each evaluation, according to manufacturer's recommendation. The methods and criteria adopted followed the American Thoracic Society/European Respiratory Society (ATS/ERS) recommendations (15). The predicted values were calculated through equations proposed for the Brazilian population (16).

CAT was completed to assess the impact of COPD on health status (17). The final score was used for analyses; higher scores indicate worse health status. The final score was also categorized as: “low impact”, “medium impact”, “high impact”, and “very high impact” (18).

The mMRC scale (19) was applied to assess dyspnea. A score between zero and four was used for the analysis; the higher the score, the greater the dyspnea (19). The SGRQ (20) was applied to evaluate the quality of life of COPD patients. The score was calculated as a percentage, as previously described (20), and the total scores of each of its three domains (symptoms, activity, and impact) were used for analyses. A higher score indicates a worse quality of life (20).

To quantify the level of PADL, patients used a triaxial accelerometer (DynaPort MiniMod, McRoberts BV®, The Hague, The Netherlands) for 12 hours on two consecutive working days (21). All patients received a manual with instructions. The time spent in the following positions and tasks were used to analyses: sitting, lying, standing, walking, and number of steps. In addition to these, the time spent in sedentary behavior was also computed, considering sedentarism the time of ≥8.5 hours (22) spent in physical activities with energy expenditure <1.5 metabolic equivalent of task (MET) (23). In order to categorize patients regarding their level of physical activity/inactivity, the cut-off point of 80 minutes day⁻¹ was used in moderate to vigorous activities (≥3 METs); patients spending ≥80 minutes day⁻¹ were considered physically active and those spending <80 minutes day⁻¹, inactive (24). The MiRA2 software was used to read and process the data provided by the accelerometer.

Six-minute walk test (6 MWT)

The 6 MWT was performed as recommended by the ATS (3). Two tests were performed in a 20-meter corridor, with a 30-minute interval between them or until the return of signs and symptoms to baseline values. The performance in the test was interpreted based on the distance traveled in meters and %pred, as proposed by Britto et al. (4) The best performance test was used for the analyses. In addition, patients were categorized into two groups: abnormal (distance walked <82% pred) and normal functional capacity (distance walked ≥ 82% pred) (7).

ADL-Glittre test

The TGlittre was performed as previously described by Skumlien et al. (9) It comprises a circuit of functional activities in a 10-meter corridor to be performed by the individual within the shortest time possible. The patient begins the test in the sitting position and, from this position, stands up and walks on the flat; in the middle of the circuit, goes up and down two steps (17 cm high × 27 cm wide) and walks again on the flat until reaching the end of the circuit, where there is a shelf. At this point, the patient moves three objects weighing 1 kg each, from the highest (which corresponds to the height of the patient's shoulders), one by one, to the lowest shelf (at waist level) and then to the floor; then the objects must be placed back on the lowest shelf, and finally on the highest shelf. Then the individual returns, making the course backward. Right after that, the individual restarts another turn, traveling the same circuit of ADL. The time spent to complete five turns represents the outcome variable. During the test, the patient carries a backpack containing 2.5 kg for women and 5.0 kg for men (9). Two tests were performed with a minimum 30-minute interval between them or until the return of signs and symptoms to baseline values. The fastest test was used for the analyses.

Statistical analysis

Data distribution^{Table 1} was evaluated using the Kolmogorov–Sminorv test. The Spearman's correlation coefficient was used to test the correlation between the time spent on TGlittre and the performance at the 6 MWT in percentage of the predicted (6MWT %pred). The following classification was used to interpret the correlation coefficient values found: weak correlation (r ≥ 0.30), moderate correlation (r ≥ 0.50), strong correlation (r ≥ 0.70), and perfect correlation (r = 1)(25). The Mann–Whitney U test was used to compare TGlittre performance between patients with 6 MWT < 82% pred and 6 MWT ≥ 82% pred.

The receiver operating characteristic (ROC) curve was used to determine the cut-off point with the most appropriate values of specificity and sensitivity in the TGlittre to discriminate patients with normal and abnormal functional capacity based on the 6MWT %pred and physical activity level. The data point closest to the upper left corner, which indicates the data with highest sensitivity and specificity represents the cut-off point. Based on ROC curve results, patients were divided into two groups and compared regarding PADL and scores of the mMRC, the SGRQ, CAT, and BODE index. For the latter analysis and to compare the same variables between the groups 6 MWT < and ≥ 82% pred, the independent t-test or Mann–Whitney U test were used. The Chi-square test was applied to compare the frequency of the genders between the classification based on the TGlittre and 6MWT %pred and to test the association between these classifications and severity of COPD according to GOLD, the PADL classification (sedentary and physical activity levels), and the quartile of the BODE index. The Cramer's V coefficient showed the strength of these associations. The level of significance adopted was 5% (p < 0.05).

Results

Eighty-three patients were initially assessed for eligibility for this study, among which 68 were potentially eligible. Of these, seven were excluded due to exacerbation of COPD during the protocol and two patients were excluded because they could not complete TGlittre due to dyspnea. Thus, 59 patients (45 males, 76.3%) completed the protocol. Of these, 10 presented moderate airflow obstruction (GOLD II); 28, severe airflow obstruction (GOLD III); and 21, very severe airflow obstruction (GOLD IV).

According to the BODE index score, 23 patients were in the first quartile, 21 in the second, 11 in the third, and 4 in the fourth. Thirty-six patients had walking distance at the 6 MWT below 82% pred, while 23 had equal or greater than 82% pred. Anthropometric characteristic of pulmonary function, dyspnea, quality of life, and health and functional status of the sample are listed in Table 1.

Table 1. Anthropometric characteristics, pulmonary function, dyspnea, quality of life, health status, functional status, and physical activities of daily life of the sample.

Variables	Mean ± SD	6 MWT < 82% pred, Mean ± SD (n = 36)	6 MWT > 82% pred, Mean ± SD (n = 23)	p
Sex (male/female)	45/14^b	27/9^b	18/5^b	0.77
LTOT	5^b	4^b	1^b	—
Age (years)	65.5 ± 8.84	67 ± 9.05	62.3 ± 8.09	0.02
Body weight (kg)	72.8 ± 15.32	74.7 ± 15.7	70.4 ± 14.7	0.41
Height (m)	1.66 ± 0.1	1.68 ± 0.10	1.67 ± 0.10	0.89
BMI (kg/m^2)	26 ± 4.78	26.6 ± 4.68	25.4 ± 5.18	0.42
Pack-years	55.6 ± 32.9	62.6 ± 35.5	45.0 ± 25.8	0.03
FEV1/FVC	0.44 ± 0.09	0.42 ± 0.10	0.49 ± 0.08	0.006
FEV1 (L)	1.08 ± 0.46	0.96 ± 0.44	1.26 ± 0.46	0.008
FEV1 (%pred)	35.3 ± 13.4	32.0 ± 13.1	40.2 ± 13.3	0.01
FVC (L)	2.35 ± 0.76	2.24 ± 0.68	2.52 ± 0.90	0.14
FVC (%pred)	60.8 ± 16.0	57.9 ± 15.9	65.1 ± 16.2	0.09
TGlittre (min)	4.56 ± 2.15	5.52 ± 2.36	3.14 ± 0.37	<0.001
6 MWT (m)	437 ± 95.4	380 ± 62.5	526 ± 67.4	<0.001
6 MWT (%pred)	77.9 ± 14.7	69.0 ± 10.6	91.8 ± 8.62	<0.001
mMRC	1 [1]^a	2 [2]^a	1 [0]^a	<0.001
CAT (total)	16.9 ± 7.96	19.8 ± 7.61	12.6 ± 6.57	<0.001
SGRQsymptoms	37.7 ± 20.8	46.0 ± 20.0	26.3 ± 16.3	<0.001
SGRQactivity	65.7 ± 20.8	72.4 ± 19.9	56.1 ± 19.1	0.003
SGRQimpact	36.7 ± 21.0	42.9 ± 22.4	28.7 ± 15.5	0.02
SGRQtotal	45.7 ± 19.1	52.5 ± 19.3	36.6 ± 14.4	0.001
BODE index (score)	3.4 ± 1.82	4.06 ± 1.97	2.50 ± 1.06	0.002
Time sitting (min)	380.5 ± 106.6	394 ± 112	360 ± 96.0	0.24
Time lying (min)	107.3 ± 106.8	107 ± 113	108 ± 98.3	0.89
Time standing (min)	140.8 ± 56.5	137 ± 56.5	146 ± 58.2	0.58
Time walking (min)	68.5 ± 39.0	58.6 ± 30.5	84.4 ± 46.0	0.03
Number of steps taken	5,314 ± 2,766	4,534 ± 2,422	6,545 ± 2,876	0.006
Time in activities < 1.5 MET (min)	583.7 ± 68.6	601 ± 50.9	557 ± 84.4	0.05
Time in activities ≥ 3 METs (min)	84.3 ± 41.9	74.5 ± 33.0	100 ± 50.1	0.02

BMI, body mass index; CAT, COPD assessment test; FEV₁, forced expiratory volume in 1 s; FVC, forced vital capacity; kg, kilograms; L, liter; LTOT, long-term oxygen therapy; m, meters; MET, metabolic equivalent of task; min, minutes; mMRC, modified Medical Research Council scale; 6 MWT, six-minute walk test; %pred, percentage of predicted; SD, standard deviation; SGRQ, Saint George's Respiratory Questionnaire; and TGlittre, ADL-Glittre test.

^a Results are presented as median (interquartile range).

^b Results are presented as frequency.

The time spent in the TGlittre presented a strong negative correlation with the 6MWT %pred (r = −0.85; p < 0.001). When comparing the variables between patients with normal and abnormal functional capacity according to performance at the 6 MWT, it was observed that patients with abnormal functional capacity took an average of 2.29 extra minutes in TGlittre (95% CI: 1.48–3.10 minutes) than patients with preserved functional capacity (5.46 ± 2.35 and 3.16 ± 0.37, respectively, p < 0.001). Besides that, patients with 6 MWT < 82% pred presented worse pulmonary function, health-related quality of life, health status and dyspnea, as well as lower time walking, time in activities ≥3 METs and number of steps, compared to those with a 6 MWT ≥ 82% pred (Table 1).

Table 2. Anthropometric characteristics, pulmonary function, dyspnea, quality of life, health status, functional status, and physical activities of daily life of the patients with TGlittre ≤3.5 minutes and TGlittre >3.5 minutes.

	TGlittre	TGlittre
	≤3.5 min,	>3.5 min,
	Mean ± SD	Mean ± SD
Variables	(n = 22)	(n = 37)	p
Sex (male/female)	15/7^b	30/7^b	0.26
LTOT	1^b	4^b	—
Age (years)	60.3 ± 7.23	68.5 ± 8.34	<0.001
Body weight (kg)	68.3 ± 14.4	75.5 ± 15.4	0.082
Height (m)	1.65 ± 0.10	1.68 ± 0.09	0.33
BMI (kg/m^2)	25.0 ± 5.02	26.7 ± 4.6	0.19
Pack-years	46.4 ± 25.7	61.0 ± 35.7	0.16
FEV1/FVC	0.47 ± 0.10	0.43 ± 0.10	0.08
FEV1 (L)	1.22 ± 0.48	1.00 ± 0.43	0.07
FEV1 (%pred)	39.7 ± 13.5	32.7 ± 12.9	0.06
FVC (L)	2.50 ± 0.90	2.26 ± 0.67	0.27
FVC (%pred)	65.2 ± 16.2	58.2 ± 15.4	0.10
TGlittre (min)	3.07 ± 0.30	5.45 ± 2.30	<0.001
6 MWT (m)	520 ± 73.7	388 ± 69	<0.001
6 MWT (%pred)	90.0 ± 10.4	70.7 ± 11.9	<0.001
mMRC	1 [0]^a	2 [2]^a	0.001
CAT (total)	14.0 ± 6.60	18.5 ± 8.30	0.03
SGRQsymptoms	28.4 ± 17.1	43.3 ± 21.0	0.007
SGRQactivity	57.1 ± 17.6	70.8 ± 21.1	0.01
SGRQimpact	30.5 ± 15.7	40.4 ± 23.0	0.06
SGRQtotal	38.1 ± 14.0	50.2 ± 20.4	0.009
BODE index (score)	2.55 ± 1.18	3.92 ± 1.95	0.001
Time sitting (min)	342.9 ± 105.2	402.4 ± 102.6	0.004
Time lying (min)	123.9 ± 104.6	97.6 ± 109.3	0.42
Time standing (min)	149.6 ± 59.7	135.7 ± 54.7	0.37
Time walking (min)	82.5 ± 46.8	60.4 ± 31.5	0.08
Number of steps taken	6,292 ± 2,925	4,743 ± 2,537	0.004
Time in activities < 1.5 MET (min)	553.4 ± 86.0	601.4 ± 49.4	0.04
Time in activities ≥ 3 METs (min)	99.5 ± 50.9	75.5 ± 33.4	0.07

BMI, body mass index; CAT, COPD assessment test; FEV₁, forced expiratory volume in 1 s; FVC, forced vital capacity; kg, kilograms; L, liter; LTOT, long-term oxygen therapy; m, meters; MET, metabolic equivalent of task; min, minutes; mMRC, modified Medical Research Council scale; 6 MWT, six-minute walk test; %pred, percentage of predicted; SD, standard deviation; SGRQ, Saint George Respiratory Questionnaire; and TGlittre, ADL-Glittre test.

^a Results are presented as median (interquartile range).

^b Results are presented as frequency.

The ROC curve indicated a cut-off point of 3.5 minutes (3 minutes and 30 seconds; sensitivity = 92%; specificity = 83%; and area under the ROC curve = 0.95 [95% CI: 0.89–0.99]) to discriminate patients with abnormal and normal functional capacity according to the 6MWT %pred (Figure 1)^{Table 1}.

Figure 1. Receiver operating characteristic (ROC) curve for the cut-off point of time spent on the TGlittre to discriminate patients with abnormal (distance walked on the 6-minute walk test < 82% pred) and normal (distance walked on the 6-minute walk test ≥ 82% pred) functional capacity. Cut-off point = 3.5 minutes; sensitivity = 92%; specificity = 83%; and area under the ROC curve = 132 meters, 95% CI: 94.1–170 meters.

Patients who spent time ≤3.5 minutes on the TGlittre had lower scores on the mMRC, CAT, and SGRQ (total score and in the symptom and activity domains), and BODE index, when compared to patients who spent more than 3.5 minutes in the TGlittre. There was no statistically significant difference (p = 0.06) in FEV₁% pred between groups. The comparisons of pulmonary function, distance at the 6 MWT, and scores in the mMRC, CAT, SGRQ, and PADL between the TGlittre ≤3.5 minutes and TGlittre >3.5 minutes groups are described in Table 2.

Regarding to PADL, patients who spent ≤3.5 minutes in the TGlittre presented shorter times in activities <1.5 MET (mean difference: 47.9 minutes, 95% CI: 12.1–83.8 minutes) and in the sitting position (mean difference: 59.4 minutes, 95% CI: 2.49–116.4 minutes) and greater number of steps (mean difference: 1,549, 95% CI: 71.8–3,027) than those who spent >3.5 minutes in the TGlittre. Furthermore, patients who spent ≤ 3.5 minutes in the TGlittre had greater distances in the 6 MWT (mean difference: 132 m, 95% CI: 94.1–170 m) than those who spent >3.5 minutes in the TGlittre.

The classification of normal and abnormal functional capacity based on the TGlittre cut-off point showed a weak association with BODE index quartiles (Cramer's V = 0.41, p < 0.05) and with the classification of sedentary behavior (time < 1.5 MET, Cramer's V = 0.33; p < 0.05). It was not associated, therefore, with the GOLD classification (Cramer's V = 0.22; p > 0.05) and with the classification of physical activity/inactivity (time ≥ 3 METs, Cramer's V = 0.13, p > 0.05).

Discussion

The main result of this study was that the cut-off point of 3.5 minutes (3 minutes and 30 seconds) in the TGlittre, based on the 6 MWT (<82% pred),(7) is sensitive and specific to differentiate patients with abnormal functional capacity from those with preserved functional capacity. Patients with TGlittre time higher than 3.5 minutes had worse dyspnea, quality of life and health status, smaller number of steps and longer time in sedentary behavior, as well as a higher risk of mortality according to BODE index, when compared to patients with normal functional capacity.

The 6 MWT was chosen as reference to determine the cut-off point of the TGlittre because, in the validation study of this test, Skumlien et al. (9) demonstrated a strong and negative correlation between the time spent to perform the TGlittre and the distance walked on the 6 MWT (r = −0.82; p < 0.05). These results were later confirmed in the study by Karloh et al. (26) and also in the present study. However, it is known that no study has yet investigated in detail a cut-off point capable of determining a reduction in functional capacity.

When identifying factors related to the performance of healthy elderly subjects at the 6 MWT and based on values predicted through a reference equation developed in the study, Troosters et al. (7) reported that a distance less than 82% pred can be considered abnormal. Although the authors did not detail the method used to identify this value, this seems to be the only way available to interpret whether or not the functional capacity is reduced, based on the 6 MWT.

Thus, using this criterion, the 3.5-minute cut-off point in the TGlittre found in the present study showed an area under the ROC curve very close to one, with high values of sensitivity and specificity. In addition, this cut-off point appears to have adequate discriminatory capacity, since differences were found between those who spent a time greater and less than/equal to 3.5 minutes in relation to relevant outcomes of the disease.

Although the cut-off point found for TGlittre was not associated with the severity of COPD and did not statistically differentiate patients in relation to lung function, there was a tendency for patients with abnormal functional capacity to have lower lung function. It was demonstrated previously that severity of lung disease is weakly correlated with functional capacity (27). Furthermore, the cut-off point found for TGlittre differentiated patients with regard to the BODE index. This is a multiscaled system that encompasses pulmonary function (FEV₁% pred), as well as other important clinical outcomes of COPD (6 MWT, body mass index, and mMRC score) (14). The BODE index was primarily intended to predict mortality risk, but was also able to predict the risk of hospitalizations for exacerbations (28) and response to pulmonary rehabilitation programs (29). It was also correlated with PADL (−0.32 ≤ r ≤ −0.47; p < 0.05) (30), performance on functional capacity tests (−0.66 ≤ r ≤ −0.86; p < 0.05) (31) and variables of maximal effort (−0.33 ≤ r ≤ −0.51; p < 0.05) (32) in COPD patients.

Another factor that reinforces the utility of this cut-off point in clinical practice and in future studies involving the TGlittre is its ability to differentiate dyspnea from exertion, health status, respiratory disease-related quality of life, and level of physical activity of patients. All these outcomes complement each other and are, therefore, of fundamental importance in the evaluation of COPD patients, since they are largely related to the disease prognosis (2, 33–37).

Among the outcomes that the proposed cut-off point was able to differentiate, we can highlight the level of PADL. This is important because the level of PADL is considered one of the best predictors of mortality in COPD (38). Besides that, in the present study, we observed an association between the classification of normal and abnormal functional capacity by the cut-off point of TGlittre and the presence of sedentarism. The sedentary behavior frequently adopted by COPD patients is associated with an increased risk of cardiovascular and metabolic diseases, as well as a higher risk of death in the general population (39). In addition, COPD patients who spend more than 8.5 hours of their day on activities <1.5 MET have about fourfold higher risk of death, and every extra hour spent in these activities during the day increases the risk of death by 42% (22).

Although no association between the classification based on the TGlittre cut-off point and physical activity/inactivity was observed, the results of the present study showed that patients with abnormal functional capacity spent less than 80 minutes on average in activities > 3 METs. In addition, it is notable that patients with ≤3.5 minutes in the TGlittre performed, on average, 1,549 more steps per day than patients with >3.5 minutes. This value is more than double the value considered as the minimal important difference for the level of physical activity in COPD patients (40) and the improvement in this outcome has been a major challenge for pulmonary rehabilitation programs (1).

Differences between the two groups in relation to the other variables also showed higher values than the minimal important difference for outcomes such as health status (41), quality of life related to respiratory disease (42), and functional capacity (3). The minimal important difference is the threshold of change after an intervention that represents some clinical significance for the patient (43). Thus, it is possible to say that the proposed cut-off point for TGlittre has clinically relevant discriminatory power.

The present study may have had some limitations. The absence of difference in the comparison analysis of lung function and some PADL variables among patients with TGlittre ≤3.5 minutes and TGlittre >3.5 minutes may have been caused by a type II error, since the p was borderline. However, the main objective of this study was to determine a cut-off point for TGlittre based on the classifications of normal and abnormal functional capacity according to the 6 MWT and, for this, we observed a power greater than 95% (44) for the correlation between the time spent in the TGlittre and the reference variable (distance walked in the 6MWT %pred) and an area of 0.95 in the ROC curve, with sensitivity of 92% and specificity of 83%. Another limitation of the present study was that the 6 MWT cut-off point used to determinate the TGlittre cut-off point has not been validated. However, in the present study, this cut-off point of 82% pred in 6 MWT was able to discriminate patients with COPD regarding clinically relevant outcomes.

This was the first study to identify a cut-off point to aid the interpretability of the TGlittre, so the findings are relevant to clinical practice. In this way, this cut-off point can be useful in future studies with TGlittre, especially in those focusing on evaluating patients with higher functional impairment, higher dyspnea, poorer health status, health-related quality of life, lower level of physical activity, and higher risk of death. We suggest the development of new studies to identify cut-off points for risk of exacerbations, hospitalizations, and mortality.

Conclusion

The cut-off point of 3.5 minutes (3 minutes and 30 seconds) in the time spent on the TGlittre is sensitive and specific to distinguish COPD patients who have abnormal functional capacity from those who are normal. This cut-off point is able to differentiate patients in relation to dyspnea sensation, health status, quality of life, PADL, and risk of death.

Declaration of interest statement

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article, and also received no financial support for the research, authorship, and/or publication of this article.