Reference Equation for the Six-Minute Walk Test in Brazilian Patients with Obesity

Obesity is a multifactorial condition that causes chronic deleterious implications on health and affects one’s quality of life. It is recognized as an important cardiovascular risk factor which contributes to the reduction of cardiopulmonary functional capacity and exercise tolerance not only because of decreased skeletal muscle strength and elevated metabolic cost, but also because of progressive gait inefficiency.1,2 In Brazil, an increase in the prevalence of overweight and obese individuals has been observed in the last decade.1,3 In a recent report published by the Ministry of Health, VIGITEL Brasil 2016,4 it was stated that more than half of the population is overweight and that 18.9% of Brazilians are obese. These statistics show an increase of 60% in obesity taxes over the last 10 years, which contributes to a growth in hypertension (14.2%) and diabetes (61.8%) in the same period. The six-minute walk test (6MWT) is a useful tool in the assessment of functional capacity and of the response to the demands required in daily life activities. In the last two decades, especially after the American Thoracic Society (ATS) published a guideline for the 6MWT, reference 307

were standardized and given every minute. The walk could be interrupted at any moment by the patient or the examiner in the presence of chest pain, intolerable dyspnea, dizziness, pallor, sweating, cramps, imbalance, and physical exhaustion. Before and after the test, the patients' HR and peripheral oxygen saturation were recorded with a pulse oximeter (Onix II, Nonin Medical Inc., Plymouth, Minnesota). Their BP was measured with a cuff diameter corresponding to the group of patients under study, 17 and their subjective perception of effort was graded according to the Borg scale. 18 The distance reached at 6 minutes of walking (D6MW) was recorded and compared to predicted values based on three reference equations employed in other populations.

Statistical Analysis
Statistical analysis was performed with the usage of the IBM-SPSS Statistic 22 program. Statistical significance was set at p<0.05. Data were presented as mean ± standard deviation. Categorical variables were described using absolute and percentage values. To compare the means, an unpaired T-Test was used. From an initial universe of 510 patients, the Kolmogorov-Smirnov test identified a distribution pattern with discrete asymmetry. To determine the explanatory variables (demographic and anthropometric attributes) that are significantly and independently best associated with the result variable (D6MC), Pearson's correlation was conducted.
The bivariate association between age and BMI with D6MC respectively, as well as the 95% confidence interval (CI) lines for individual predicted values, are shown in Figures 1 and 2. The participants selected for our sample were between the lines of 95% CI. Therefore, 460 obese patients (306 women) remained with a BMI ≥ 30 kg/m 2 , the majority (71%) with grade III obesity (BMI> 40 kg/m 2 ). 19 Seeking to build a multiple linear regression model, and understanding the need to obey the assumptions for its use, it can be concluded that, despite the discreet asymmetry, the residues of our model have an approximately normal distribution, as the sample size is large enough to accept the premise of normality. 20

Results
The general characteristics of the 460 obese patients, 306 women (67%), are presented in Table 1. A significant difference was observed between men and women for age and D6MC, with p<0.05, as well as for weight, height, equations for the distance predicted in the test have been indicated for different populations, both for healthy individuals and for patients with several chronic diseases. [5][6][7][8] In 2009, Iwama et al., 9 proposed a reference equation for healthy Brazilian subjects. In 2011, Soares 10 also presented a reference equation for healthy adults in Brazil, taking into account the body mass index (BMI) in their composition. In 2013, based on a multicentric study, Britto et al. 11 published reference equations for the Brazilian population, which illustrates the growing interest in the 6MWT in our context.
The use of the 6MWT for the evaluation of obese patients has recently been observed, as shown in different already published studies. [12][13][14][15] Thus, this study aims to establish a reference equation for the distance predicted in 6MWT in obese Brazilian patients.

Methods
The sample size was defined by convenience, patients were recruited according to their admission to the Multidisciplinary  Patients with unstable angina, myocardial infarction, cardiac arrhythmias, heart rate (HR) greater than 120 beats per minute at rest, blood presure (BP) ≥ 160/100 mmhg at rest, peripheral oxygen saturation lower than 95% at rest, neurological and /or orthopedic conditions capable of restricting gait were excluded.
Using a standardized technique, (16) the patients' weight (w) was obtained with a Filizola calibrated balance (0.1 kg of precision) when they were barefoot and wearing light clothes. The patients' height (h) was measured with a stadiometer (0.5 cm accuracy). Afterwards, their body mass index (BMI) was calculated using the following formula: BMI = w (kg) / h (m 2 ).
The 6MWT was applied to all participants in accordance with the "ATS Statement: Guidelines for the Six-minute Walk Test". 16 Patients were advised to walk as fast as possible in a quiet and unobstructed hallway, 30 meters long, with each meter marked. Encouragement instructions and BMI, with p <0.001. Taking the entire group into consideration, the D6MW had a significant correlation with sex (r = 0.102, p < 0.029), age (r = -0.387, p < 0.001), weight (r = -0.312, p < 0.001), height (r = 0.163, p < 0.001), and BMI (r = -0.464, p <0.001). Therefore, the D6MW is explained 1% due to sex, 14.9% to age, 9.7% to weight, 2.6% to height, and 21.5% to BMI.
For multiple linear regression, the best model adjusted for the D6MW predictor was that using the variables of gender, age, and BMI, which together explain 48.7% of the total D6MW variance ( Table 2) (Figure 3).

Discussion
Functional capacity to exercise is an important marker of mortality and morbidity. In obese patients, walking requires greater metabolic and cardiopulmonary demand. Moreover, the additional musculoskeletal dysfunction due to joint overload and osteoarthrosis (knees, ankles and hip), associated with eventual loss of muscle mass, can be factors that contribute to a lower performance in walking. 2,6,21 As 6MWT is easy to apply, has a low cost, is safe, and is easily correlated with the activities of daily life, mostly at the submaximal level, it represents a relevant element in the evaluation of obese patients' functional capacity. Consequently, D6MW becomes a significant prognostic indicator. 6,22 Previous studies have already described age, sex, weight, height, and BMI as variables capable of influencing D6MW. 5,9,12,13,23,24 For obese individuals, especially those with BMI ≥ 40 kg / m2, as already mentioned, musculoskeletal dysfunctions and loss of muscle mass are factors that have a considerable influence on D6MW, particularly with advancing age 2,25 (Figure 1). Obesity Grade I, n (%) 18 (12) 52 (17) Obesity Grade II, n (%) 18 (12) 44 (14) Obesity Grade III, n (%)  The importance of BMI in the D6MW variance in a population of obese patients has been described in medical literature, 12,15,21,26 despite the significant difference in BMI between the genders (p <0.001) found in our patient group, where the BMI was able to explain 32.6% for men and 20% for women of the D6MW results . It was observed that, for a growing value of BMI, there is a reduction of D6MC in both sexes (Figure 2).
In our study, for the whole group, age was able to explain 14.9% and BMI 21.5% of the results. Data slightly different from those previously obtained by Capodaglio with obese patients, 13 where age explained 12.9% and BMI 24% of the D6MW results. This difference occurs because our population is older and with a smaller number of male subjects, who are proportionally heavier. However, our results were similar to Capodaglio's in relation to the total variance of the D6MW, explained by the three chosen variables (gender, age, and BMI), that is, 48% in Capodaglio`s study and 48.7% in our investigation. A greater number of variables introduced in the equation, as already pointed out by other authors, (14,20) could provide a better evaluation from a purely mathematical point of view. On the other hand, this study would lose in practicality, since a greater number of variables would lead to a greater complexity in the collection of data in an outpatient environment of basic health care.
Our study, does, however have some limitations. The sample was chosen for convenience, allowing for some selection bias. This study consists of patients with a predominance of grade III obesity, most of whom were female, who sought out bariatric surgery to treat their obesity. In addition, about 50% of the D6MC variance remains unexplained by our regression model. Caution should be taken when using the proposed equation for patients with different characteristics from our cohort. This study should also be repeated with a larger number of participants under the age of 40 and over 60, and with grade I and II obesity.

Conclusion
The reference equation proposed in this study shows a significant adjustment, when compared to the equations already described in the literature, and allows for a better estimate of the submaximal functional capacity of Brazilian patients with obesity, which provides better subsidies for monitoring the therapeutic interventions performed in this population.

Potential Conflict of Interest
No potential conflict of interest relevant to this article was reported.

Sources of Funding
There were no external funding sources for this study.

Study Association
This article is part of the thesis of Doctoral submitted by José Carlos do Vale Quaresma, from Universidade Federal do Rio de Janeiro.