Do Functional Tests Estimate Physical Activity in COPD?

Abstract The association between physical activity in daily life (PADL) and simple functional tests is not yet clearly understood in subjects with chronic obstructive pulmonary disease (COPD). Therefore, the aim of this study was to investigate the association of two functional tests (Sit-to-Stand test [STS] and the 4-Metre Gait Speed test [4MGS]) with PADL, as well as to identify whether these tests can discriminate those subjects who are physically inactive. In this cross-sectional study, 28 subjects with COPD performed the five repetitions Sit-to-Stand (STS5r), the 4MGS and used the DynaPort activity monitor for 7 days in order to assess PADL. Walking time, movement intensity while walking (MI) and Physical Activity Level index (PAL) were considered as PADL outcomes. STS5r and 4MGS, respectively, were significantly associated with walking time (R2 = 0.16; p = 0.024 and R2 = 0.25; p = 0.006) and PAL index (R2 = 0.17; p = 0.002 and R2 = 0.30; p = 0.003), whereas movement intensity was associated only with the 4MGS (R2 = 0.23; p = 0.009). Additionally, both tests were able to discriminate physically inactive subjects (cutoffs: STS5r = 11.48s [AUC = 0.73]; 4MGS = 1.09m/s [AUC = 0.88]). In conclusion, STS5r and 4MGS can predict up to 30% of PADL in subjects with COPD. Both tests are related to PADL duration (e.g. time spent walking), while only the 4MGS reflects movement intensity. Both tests presented discriminative capacity to identify subjects with worse PADL pattern.


Introduction
Field tests are a widely used alternative to assess functional capacity in patients with chronic respiratory diseases, considered to be low cost, fast and requiring smaller spaces, for example the 4 m Gait Speed (4MGS) and five repetitions Sit-to-Stand (STS5r) [1,2].Despite their simplicity, the clinical relevance of these tests has been increasingly recognized in the scientific literature.The 30-second and 1-minute Sit to Stand test protocols have been used as a valid and reliable tool to assess lower limb muscle performance in patients with COPD and have been shown to be responsive after an 8-week pulmonary rehabilitation program [1].
Furthermore, slower STS5r time was associated with worse health-related quality of life, dyspnea in daily life, as well as worsening prognosis [2].On the other hand, gait speed tests have been promoted as predictors of essential information such as mortality and exacerbation on subjects with COPD [3].The latter also reflects many of the multisystemic effects of the disease severity in this population and is an excellent screening measure for exercise capacity and frailty [3].In addition, both STS5r and 4MGS were recommended as potentially useful markers to classify low performance in the 6MWT (<350 m) in this population [4].
Due to dyspnea and the multisystemic characteristics of the disease, subjects with COPD generally also have a low level of physical activity in daily life (PADL) [5,6].Physical inactivity can be objectively quantified with devices used to detect body movement; however, a detailed measurement of PADL with technologically advanced motion sensors can be expensive [6].In addition, it is known that both PADL and functional tests are associated with important outcomes such as exacerbation rates, mortality and prognosis in subjects with COPD [7,8].For instance, the number of steps per day and the physical activity level (PAL) may reflect two or more exacerbations in the preceding year even when adjusted for confounders (r 2 = −0.34 and 0.26, respectively) [9].Furthermore, Puhan et al. identified that 2-year mortality in subjects with COPD can be predicted by the STS test (AUC 0.78) as well as by the ADO index (age, dyspnea and airflow obstruction) (AUC 0.80), which was specifically developed for this purpose and showed a similar magnitude of discrimination than the STS test [8].More recently, a multi-centre UK prospective observational study identified that the STS was associated with risk of hospital admissions and length of hospital stay due to acute exacerbation of COPD [10].The same was observed with the 4MGS which independently predicts the risk of older patients hospitalized for acute exacerbation of COPD [11] as well as predicts mortality in stable COPD [11].
Functional tests and daily physical activity also present some similar movements; however, only a few studies investigated the associations between PADL and functional tests [12] and there is no evidence to date whether the 4MGS or the 5-repetition STS can predict or discriminate PADL in patients with COPD.Therefore, the aim of this study was to verify the association between these functional tests and PADL in subjects with COPD, as well as to identify whether these tests can discriminate less active subjects from those who are more physically active.

Methods
In this cross sectional study, we analyzed retrospective data from a baseline assessment available in a database of a previous study [13], which was approved by the ethics committee of the State University of Londrina, Brazil (123/09), and all subjects signed a written informed consent.Subjects with diagnosis of COPD who were referred to our institutional exercise training program were recruited (outpatient setting).We selected for the present study subjects from that database who performed lung function assessment, STS5r and 4MGS tests as well as wore the physical activity monitor for 24 h/day during seven days.The inclusion criteria of the original study were: diagnosis of COPD according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [14]; absence of exacerbations in the previous 3 months; and not taking part in regular physical exercise programs in the last year.Exclusion criteria were any comorbidity that could impair performance in the functional capacity tests such as severe cardiovascular, orthopedic or neuromuscular conditions.
At the first visit, patients performed lung function assessment and the two functional tests (STS5r and 4MGS), additionally, they received the activity monitor for PADL assessment.After 8 days, they returned for the second visit and performed the 6MWT.
Lung function was assessed according to the American Thoracic Society/European Respiratory Society (ATS/ERS) [15] using a portable spirometer (Spiropalm, COSMED, Italy).Predicted values were calculated using a specific equation for the Brazilian population [16].The STS test was performed using the protocol of five repetitions, in which the subjects had to stand up from the sitting position and sit down repeatedly 5 times as fast as possible while the rater measured the time [2].Regarding the 4MGS test, patients were instructed to walk at their usual pace in a 4 m corridor signalized by cones, and the time to perform this was measured by a stopwatch [17].Two tests were performed and the fastest was considered for the analysis.Anthropometric characteristics, such as height, weight, and subsequent calculation of the body mass index, were objectively assessed.
The 6MWT was performed twice according to international recommendations [18] and Brazilian reference values were used [19].PADL was assessed with the DynaPort Move Monitor (DMM-McRoberts BV, the Netherlands) which was worn on the subject's waist 24 h/day for 7 consecutive days.A valid assessment was considered with at least 4 days of wearing time.The Dynaport is a physical activity monitor containing a triaxial accelerometer, validated for subjects with COPD [20][21][22].PADL variables analyzed were: number of steps, time per day spent walking, movement intensity and the physical activity level index (PAL), defined as total energy expenditure divided by basal metabolic rate (TEE/BMR).
All analyses were performed using SPSS version 20.0 (Statistical Packages for the Social Sciences Inc., Chicago, Illions, USA) and GhaphPad Prism 6.0 (GhaphPad Prism software Inc., San Diego, California, USA).The distribution of continuous variables was analyzed by the Shapiro Wilk test.Associations between functional tests and PADL variables were analyzed with the Pearson or Spearman correlation coefficients and univariate linear regression.In addition, the analysis of the receiver operating characteristics (ROC) curve, which takes into account all sensitivity and specificity values for each test variable, were carried out to assess whether functional tests can discriminate subjects who are less physically active; for this, the subjects were dichotomized using the value of PAL index for cutoff point of 1.4 [23].Statistical significance was set as p < 0.05.

Results
Twenty-eight subjects with COPD were included, there were no exclusions although there were missing data of the 4MGS in three subjects due to logistical problems during the test.Table 1 shows the characteristics of the sample, which present an equal proportion of men and women with moderate to severe COPD.
As an additional analysis, the 6MWT presented significant correlations with time per day spent walking (r = 0.70; p < 0.0001), PAL (r = 0.58; p = 0.002), movement intensity (r = 0.40; p = 0.047) and number of steps (r = 0.40; p = 0.042).Table 2 shows that both STS and 4MGS were able to discriminate subjects who were less physically active, presenting satisfactory AUC values.Moreover, to compare discrimination ability between these two functional tests and the 6MWT, we also provided the 6MWT results of AUC in Table 2.

Discussion
This study found moderate associations between two simple functional tests and objectively measured PADL variables, as well as identified that the STS and 4MGS are able to discriminate less physically active patients with COPD from those physically more active.Noteworthy, only the 4MGS was associated with movement intensity in daily life.
This study found no significant correlation of these two functional tests with steps/day; in contrast, Van Gestel et al. found that although the 1-minute STS test was associated with number of steps/day, it cannot be used to identify subjects with an extremely inactive lifestyle (PAL < 1.4; AUC < 0.31) [24].In the present study, both the STS and 4MGS were associated with PAL index in univariate analysis, as well as both tests were able to identify subjects with extremely inactive lifestyle (AUC > 0.73 and 0.88, respectively).To the best of our knowledge, this is the first study to show the association between the 5-repetition STS test and PAL index, whereas the 4MGS test had been previously shown to correlate with this index [25].
The fact that only the 4MGS was associated with movement intensity might have happened because of the similar characteristics of the two assessments (both are measured while  walking).In spite of not surprising, this result was not shown in the scientific literature yet.In a previous study, Pitta et al. [6] found moderate to strong correlations between movement intensity and walking speed during fast walking (r = 0.72; p < 0.05) and also with an increase in walking speed (r = 0.81; p < 0.01).However, they did not find significant correlations between movement intensity and cycling at any speed (r ≤ 0.25).Therefore, simple walking tests such as the 4MGS may be useful to reflect movement intensity during walking in daily life.
A previous study has shown that the 5-repetition STS test is correlated with quadriceps strength (r = −0.38;p < 0.001) [2], and the later is related to walking time (r = 0.45; p < 0.01) [5].This may explain the moderate correlation between STS and time per day spent walking in the present study.Unfortunately, peripheral muscle strength was not measure in the present study; however, we can hypothesize that those subjects with better muscle strength who are able to spend more time walking are also able to perform the STS faster.Another study found significant correlations between 5-repetition STS and the time per day spent in standing position and locomotion activities (r = 0.35 and 0.37, respectively) [26].These findings corroborate with our results and are reasonable to suggest that the STS test reflects well the time spent in different postures or activities which are lower limb dependent.
A recent study [26] showed that the STS test was able to identify subjects with COPD who presented low and preserved exercise capacity according to the 6MWT; however, it was not possible to discriminate physically active from inactive subjects.Notewhorthy, in that study subjects were classified as physically active or inactive according to achieving or not 30 min per day in activities of at least moderate intensity.In contrast, in the present study, subjects were classified as less (or more) active according to being below (or above) the well-known PAL cutoff of 1.4 [23,24].Interestingly, the 6MWT presented worse ability to discriminate physically active from inactive subjects in comparison to the STS and 4MGS (AUC = 0.64).
The STS and 4MGS tests are commonly used to identify other outcomes; for example, the failure to complete the 5-repetition STS test suggests significant impairment in exercise capacity and significant quadriceps muscle weakness [2].A poor health-related quality of life can also be found in subjects with COPD who perform a walking speed less than 0.8 m/s in 4MGS [2].The present study complements these findings by showing that patients with slower 4MGS (<1.09 m/s) had a less active PADL pattern in comparison to those with preserved 4MGS.Interestingly, this finding is also in accordance with a previous study which suggested to use the 4MGS as a screening test to predict worsening daily physical activity, since it was associated with PAL in subjects with chronic respiratory diseases [27].In that study, although the sample was composed by patients with interstitial lung disease and COPD, there was a very close cutoff value suggested for the 4MGS (1.07 m/s) [27].
An objective measure of PADL, as provided by the use of activity monitors, can be expensive and may not be feasible in every setting.Because of that, researchers and clinicians may find more suitable to investigate this outcome through subjective methods such as questionnaires; however, these instruments present limited validity and reliability.On the other hand, functional tests require simple and low cost equipment, besides being fast and easy to perform since they require common daily living movements [6].Because of these characteristics and the associations that they have with important outcomes, such as muscle strength, exercise capacity, PADL and others, functional tests can be considered as excellent assessment options for subjects with COPD.Although these tests may reflect aspects of physical activity, the objective assessment of PADL should not be replaced by functional tests.
Despite all efforts, this study has some limitations.Variables of sedentary behavior, such as time spent per day in sitting and lying positions during awake time were not analyzed.Furthermore, we recognize that the sample size is relatively small.However, secondary results of the present study showed correlation values between 6MWT and PADL which are very similar to the previous literature [22], which may indicate that the sample was not under represented in this sense.Moreover, correlation coefficient is independent of sample size; therefore, we believe that the clinical messages of the study are relevant and not influenced by sample size.Of note, multivariate analysis adjusted for potential confounders associated with physical activity (such as sex, age and lung function) were also significant and did not change the clinical message of the study; however, these results were not reported due to the minimum sample size required for this specific analysis.We encourage the development of new studies with larger samples to confirm and expand on these findings.

Conclusion
In summary, both the 4MGS and the STS5r tests explained from 16-30% of PADL in subjects with COPD and are associated with variables which reflect activity duration such as time per day spent walking and the PAL index.In contrast, only the 4MGS was able to predict the movement intensity during walking in daily life.Furthermore, both tests were able to discriminate those patients with COPD who are less physically active.

Declaration of interest
The authors report no conflict of interest.

Figure 1 .
Figure 1.Correlation of the five-repetition sit-to-stand (sts) test with (a) walking time spent per day; (b) walking movement intensity and (c) physical activity level (Pal) index.

Figure 2 .
Figure 2. Correlations of the 4-metre gait speed (4mGs) test with (a) walking time spent per day; (b) walking movement intensity and (c) physical activity level (Pal) index.

Funding
NSC, JMO, LPS and AAM are supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior e Programa de Suporte

Table 1 .
Characteristics of the patients included in the study.
Data expressed in absolute and percentage or mean ± standard.bmi = body mass index; fev 1 = forced expiratory volume in first second; fvC = forced vital capacity; GOlD = Global Obstructive lung Disease; 6mWt = six-minute walk test; PaDl = Physical activity in daily life; sts = five-repetition sit-to-stand test; 4mGs = 4-meter Gait speed test.a n = 25.