Abstract
Objective. To determine factors that influence 6-minute walk distance (6MWD) in patients with scleroderma (systemic sclerosis, SSc)-interstitial lung disease (ILD), SSc-pulmonary hypertension (PH), and idiopathic pulmonary fibrosis (IPF).
Methods. We retrospectively evaluated all patients with SSc or IPF who performed a 6-minute walk test (6MWT) at a university hospital between 1999 and 2003. Chi-square, ANOVA, simple linear regression, and backwards elimination multivariable regressions were performed.
Results. Forty-eight consecutive IPF patients with 6MWT were compared to 33 patients with SSc-ILD, 13 with SSc-PH, 19 with both SSc-ILD and SSc-PH (SSc-Both), and 15 with SSc without ILD or PH (SSc-Neither). Mean 6MWD did not differ between groups. Limitations to 6MWT trended toward dyspnea in IPF and lower extremity pain in SSc. SSc-Both had dyspnea limitation more than other SSc subgroups (p = 0.017). Percentage predicted forced vital capacity (FVC%) and percentage predicted carbon monoxide diffusing capacity (DLCO%) were more strongly predictive of 6MWD in IPF than in SSc; however, exclusion of SSc subjects with pain limitation improved the predictive value. Significant correlates of 6MWD in multivariable analysis differed between subgroups.
Conclusion. Pain limitations confound the utility of the 6MWT, particularly in SSc. Pain may cause failure to reach a dyspnea limitation during 6MWT, especially in SSc patients without both ILD and PH. Correlates of 6MWD differ between SSc subgroups and IPF; therefore, the 6MWT distance is not always reflective of the same physiological process. 6MWT interpretation should include consideration of vascular, pulmonary, and musculoskeletal exercise limitations.
Pulmonary disease is the most common cause of death in patients with scleroderma (systemic sclerosis, SSc)1. The 2 most frequent pulmonary complications in patients with scleroderma are interstitial lung disease (ILD) and pulmonary hypertension (PH)2. The importance of pulmonary complications and their associated morbidity and mortality in this patient population has led to an increase in the numbers of clinical trials conducted to evaluate treatments for both PH and ILD in SSc.
Exercise testing is a common endpoint used in clinical trials evaluating treatments for pulmonary diseases. The 6-minute walk test (6MWT) is a simple way to measure submaximal exercise tolerance3 that has been used as a primary endpoint in clinical trials for treatment and prognosis of idiopathic pulmonary fibrosis (IPF)4, and is the most commonly used endpoint in clinical trials for pulmonary arterial hypertension (PAH)5–8. Six-minute walk distance (6MWD) has been shown to correlate with mortality in IPF9 and idiopathic PAH10, but may have less value as a surrogate endpoint in patients with systemic diseases such as SSc. The musculoskeletal manifestations of SSc are well described, and include joint pain and inflammation11, muscle pain and weakness12, and peripheral vascular disease13,14. Although patients with SSc lung disease have been shown to have decreased exercise capacity15,16, systemic manifestations of SSc may influence the walk distance and complicate interpretation of the test.
Recently, Villalba and colleagues17 found the 6MWD in SSc to correlate with age, dyspnea index, fibrosis on the chest radiograph, systolic pulmonary arterial pressure (sPAP), and oxygen desaturation, but not with percentage predicted forced vital capacity (FVC%) or findings on chest computed tomography (CT). Oxygen desaturation was found to be a better surrogate endpoint than 6MWD, correlating with dyspnea index, FVC%, fibrosis on chest radiograph, ground-glass or reticular opacities on chest CT, and sPAP. Another study found the 6MWD had test-retest reliability, but only a weak correlation with the Borg Dyspnea Index and FVC%, and no correlation with percentage predicted carbon monoxide diffusing capacity (DLCO%)18. These results suggest that the 6MWD may be influenced by factors other than lung disease in patients with SSc.
Disagreement concerning acceptable endpoints for clinical trials in patients with SSc continues due to the multisystem expression of the disease and the inadequacy of traditional outcome measures, resulting in debate concerning the best endpoints. Progress has been made in the development of a core set of measures to be used in clinical trials, but further study is still required19–21.
Our objectives were to evaluate the limitations to 6MWT in patients with SSc compared to IPF, a disease whose predominant feature is ILD, and to assess differences in physiologic measures between these 2 groups. We further hypothesized that the utility of the 6MWD will differ between SSc subgroups. Differences in symptom limitations to 6MWT and physiologic variables may further define exercise limitations and inform the design of future clinical trials.
MATERIALS AND METHODS
Patient population
We retrospectively evaluated all adult patients with American College of Rheumatology-defined SSc or American Thoracic Society (ATS)-defined IPF seen at the Medical University of South Carolina from 1998 to 2003 who performed a 6MWT. The SSc group (All SSc) was prospectively subdivided into SSc with ILD (SSc-ILD), SSc with PH (SSc-PH), SSc with both ILD and PH (SSc-Both), and SSc with neither ILD nor PH (SSc-Neither). The 6MWT was performed in accord with ATS guidelines22. There were no requirements concerning gender, limited versus diffuse SSc, overlap with other connective tissue diseases, or disease severity or duration. To be diagnosed with SSc-associated PH, a subject had elevated pulmonary artery pressures by echocardiogram (sPAP > 40 mm Hg) or right-heart catheterization (mean PAP > 25 mm Hg). IPF and SScassociated ILD were diagnosed by any ground-glass or reticular opacities on high resolution chest CT with slice thickness 0.6–1.0 mm. The 6MWT was oxygen supplemented in some patients, although the 6MWT was not stopped at any level of oxygen desaturation. Each subject’s 6MWD, prewalk and post-walk Borg dyspnea scores, resting oxygen saturation (SpO2), and minimum SpO2 during 6MWT were recorded. After the 6MWT, each subject was asked, “What prevented you from walking farther?” and the answer was recorded as the patient-reported limitation to the 6MWT. Answers were characterized into one of 6 categories: (1) dyspnea, (2) lower extremity pain, (3) generalized fatigue, (4) chest pain or tightness, (5) lightheadedness or dizziness, or (6) no limitation. In addition, FEV1/FVC, FVC%, and DLCO% within 3 months of 6MWT were recorded, performed by ATS standards23 and assessed by standards of Crapo, et al24 with hemoglobin adjustment according to Cotes, et al25.
Statistical analysis
Categorical data were compared using chi-square tests and continuous data were compared using analysis of variance (ANOVA) and Tukey’s multiple comparison test. Simple linear regression was used to determine whether physiologic variables correlated with 6MWD. Backwards elimination multivariable regression was used to determine the patient characteristics and physiologic variables correlated with 6MWD. We limited each model to one predictor per 10 subjects. Variables included age, gender, pre- and post-walk Borg dyspnea score, resting SpO2, minimum SpO2 during 6MWT, FVC%, DLCO%, FEV1/FVC, and reported limitation to 6MWT, and sPAP. P value ≤ 0.05 was considered significant. Continuous data are reported as mean and 95% confidence interval (95% CI) or 25%–75% range as specified. Categorical data are reported by frequency.
RESULTS
Between 1998 and 2003, 80 patients with SSc and 48 patients with IPF underwent 6MWT. Baseline characteristics of the IPF and SSc groups, with SSc broken down by pulmonary diagnosis, are shown in Table 1. These characteristics agree with the known demographics of SSc, a disease that tends to occur in younger women, and IPF, a disease that occurs more often in older individuals. Of the 80 subjects with SSc, missing data included 5 subjects who had not been asked their limitation to the 6MWT and one subject whose walk distance was not recorded. Seven subjects with IPF and 13 with SSc had no spirometry recorded within 3 months of the 6MWT. Three subjects in each group who underwent spirometry did not have a concurrently recorded DLCO%.
Baseline subject characteristics for idiopathic pulmonary fibrosis (IPF) and scleroderma (SSc), then with SSc broken down by pulmonary disease groups. Continuous data are reported as mean (25%–75% range), categorical data are reported as frequency (%).
Limitations to 6MWT
There was no significant difference in mean 6MWD between the IPF and the SSc subgroups (p = 0.14; Table 2). Dyspnea was reported as an exercise limitation more often in subjects with IPF (67%) than SSc (57%), although this was not statistically significant (p = 0.30; Figure 1). Dyspnea comparisons between IPF (67%) and SSc-ILD (48%) were not statistically different (p = 0.11). Lower extremity pain was the primary limitation to walk distance for 20% of subjects with SSc compared to 15% of subjects with IPF (p = 0.40; Figure 1). Pain was the limitation in 23% of the SSc-ILD subjects, but this was also not significantly different from the IPF group (p = 0.51). The SSc subgroups differed on the primary limitation to 6MWT distance. Almost all subjects with both ILD and PH reported a dyspnea limitation to 6MWD, whereas fewer than half in the subgroup with neither ILD nor PH had dyspnea (p = 0.017; Figure 1). SSc subjects with both ILD and PH never reported lower extremity pain as their primary limitation, but lower extremity pain was reported as a limitation in patients in each of the SSc-ILD, SSc-PH, and SSc-Neither subgroups (p = 0.014; Figure 1).
Patient-reported limitations to 6-minute walk test (6MWT) in idiopathic pulmonary fibrosis (IPF) and scleroderma (SSc) subgroups. ILD: interstitial lung disease; PAH: pulmonary arterial hypertension.
Mean 6-minute walk distance (6MWD), resting, minimum, and drop in oxygen saturation (SpO2), percentage predicted forced vital capacity (FVC%), and percentage predicted carbon monoxide diffusing capacity (DLCO%) compared by ANOVA between idiopathic pulmonary fibrosis (IPF) and scleroderma (SSc)-interstitial lung disease (ILD) (Table 2A); and between SSc subgroups (Table 2B). All data reported as mean (95% CI).
Although sample sizes were too small to make meaningful statistical comparisons, the distribution of reported dyspnea limitation did not vary by gender in IPF, SSc, or any of the SSc subgroups. Women tended to report “other” limitations including chest tightness or dizziness more often than men (p = 0.29). Pain appeared to be a limitation more often in older subjects across IPF and all SSc subgroups, especially in the SSc-Neither subgroup, when age was analyzed as a dichotomous variable, > 55 versus ≤ 55 years old (p = 0.07).
Predictors of 6MWT distance
The correlation coefficients between both FVC% and DLCO% and 6MWD were significant in IPF but not in SSc (Figure 2A, 2B). Association of DLCO% but not FVC% with 6MWD in SSc became weakly significant with exclusion of patients who reported either lower extremity pain or “no limitation” as their limitation to the 6MWT (Figure 2C, 2D). The correlation between 6MWD and FVC% was much stronger for SSc subjects with severe lung disease (FVC% ≤ 60%, N = 17; R2 = 0.205, p = 0.03) than subjects with less severe disease (FVC% > 60%, N = 49; R2 < 0.001, p = 0.86). 6MWD was not correlated with sPAP in subjects with sPAP < 45 (N = 19, R2 = 0.069, p = 0.227), but was significantly correlated with sPAP ≥ 45 (N = 14; R2 = 0.383, p = 0.018). 6MWD was not correlated with DLCO for those with DLCO ≤ 60% (N = 49; R2 = 0.0146, p = 0.41) or DLCO > 60% (N = 14; inversely related). Significant multivariable correlates of 6MWD differed between subjects with IPF and those with SSc (Table 3). The correlates of 6MWD in IPF were DLCO% and gender; the correlates in all SSc patients were pre-walk Borg score, minimum SpO2, resting SpO2, and gender; the correlates in SSc-ILD were DLCO%, minimum SpO2, and prewalk Borg score; the strongest correlate in SSc-PH was post-walk Borg score; the strongest correlate in SSc-Both and SSc-Neither was minimum SpO2. Exclusion of SSc subjects with pain limitation improved model fit, increasing the percentage of explained variability from 26% in the full group to 36% in the limited group (Table 3).When only age, gender, DLCO%, and FVC% were included in the model, 47% of the variability in IPF (n = 37) and 41% of the variability in SSc-ILD (n = 26) was explained.
A, B. Percentage predicted forced vital capacity (FVC%) and percentage predicted carbon monoxide diffusing capacity (DLCO%) compared with 6-minute walk distance (6MWD) in subjects with idiopathic pulmonary fibrosis (IPF) and scleroderma (SSc). (A: IPF n = 41, SSc n = 66; B: IPF n = 38, SSc n = 63) C, D. Comparison between FVC% (n = 45) or DLCO% (n = 43) and 6MWD in SSc subjects who reported a limitation to walk distance other than pain or “no limitation.”
Predictors of 6-minute walk distance (6MWD) by multivariable regression.
Physiologic measures
Physiologic measures such as minimum SpO2 during 6MWT, FVC%, and DLCO% may indicate the severity of pulmonary interstitial or vascular disease. The 6MWD was statistically equal between the IPF and SSc-ILD groups; however, the IPF group had a significantly lower resting SpO2, lower minimum SpO2, and greater drop in SpO2. In contrast, the FVC% was lower in the SSc-ILD group compared to the IPF group.
DISCUSSION
Our goal was to better inform future decisions concerning preferred endpoints in clinical trials. The absence of a clearly superior primary endpoint for clinical trials in patients with SSc has led to controversy concerning the best surrogate outcome measures in this patient population19. Although pulmonary disease is the leading cause of death in SSc, physiologic endpoints such as the 6MWT that are sensitive to disease in multiple organ systems may fail to specifically evaluate lung disease.
We found that the causes of limitation to the 6MWT differ between SSc subgroups and IPF. Walk distance alone is unable to distinguish between subjects with ILD or PH and subjects without clinically evident pulmonary disease. The 6MWT is therefore sensitive to some other disease manifestations of SSc that limit the 6MWD.
The correlates of 6MWD vary between SSc subgroups and between SSc-ILD and IPF. In IPF, a single predictor, DLCO%, explained a large percentage (37%) of the variability in 6MWT distance. Similarly in the SSc-ILD cohort, DLCO% was a strong predictor of 6MWD, but other variables such as minimum SpO2 were significant predictors. However, in the entire SSc population, the picture was much more mixed, with several predictors contributing to the model, with measured variables accounting for a smaller percentage of the variability in 6MWD. Interestingly, the model was improved with exclusion of subjects who reported pain as a limitation to the walk test.
Subjects with SSc have varying degrees of musculoskeletal and peripheral vascular involvement, including arthralgia, myopathy, and claudication11–14. In addition to myositis, alterations in small blood vessels in the skeletal muscles may influence cellular oxygen delivery and contribute to an early transition to anaerobic metabolism. Therefore, removing SSc patients who reported pain allowed stronger correlations between measures of lung disease and the 6MWD. The extent to which this improvement is due to joint, muscle, and/or circulatory limitations remains unknown.
Sensitivity of the 6MWD to pain does not necessarily make the endpoint inadequate for all clinical trials. If the goal of a study is to improve the overall well-being of a patient, without the need to distinguish between pulmonary and musculoskeletal disease, the 6MWT may be adequate. However, in SSc, where pulmonary disease is the main cause of mortality, clinical trials for medical therapies would be amiss to ignore the importance of evaluating pulmonary disease severity. If the musculoskeletal disease and peripheral vascular involvement remain stable, changes in pulmonary disease may be detected. Future studies of SSc pulmonary disease should exclude patients with unstable lower extremity pain limitation if 6MWT is used.
The 6MWD is not reflective of the same physiological process in each disease or each SSc subgroup. Although each group had statistically the same mean 6MWD, the ILD subgroups clearly had more severe lung disease as assessed by FVC%, DLCO%, and minimum SpO2. The association between FVC% and DLCO% with 6MWD is improved with exclusion of pain, but pain may not be the only confounding limitation in SSc. The interesting group of SSc patients with neither PH nor ILD on routine testing likely sheds some light on additional 6MWT limitations in SSc. After excluding patients with pain, this group had a mean 6MWD of 340 m. These patients had similar degrees of O2 desaturation, FVC%, and DLCO% abnormalities compared to the group with SSc-PH. These similarities suggest a vascular limitation to exercise that could be due to small-vessel disease in the muscles or to exercise-induced PH, a diagnosis that would require exercise echocardiography or exercise right-heart catheterization to detect.
Less severe pulmonary disease may be missed by the 6MWD. Detection and evaluation of early stages of pulmonary disease will become more important as better medical therapies for ILD and PH are developed. For this reason, other measures of lung disease should be considered as endpoints in studies of subjects with early pulmonary disease. For example, measuring oxygen desaturation may improve the 6MWT utility17,26. In our study, minimum SpO2 is one of the strongest correlates of 6MWD in almost every subgroup of SSc and SSc as a whole.
Our SSc results are similar but not identical to those of Villalba and colleagues17. In univariate analysis we found a correlation between 6MWD and diminished SpO2 (p = 0.033) and pre-walk Borg dyspnea score (p = 0.041), with no correlation to gender or FVC%. In contrast, we did not find 6MWD correlated with age, either as a continuous variable or a dichotomous variable. Similarly to Villalba, et al17 we found SpO2 decline correlated with 6MWD and FVC% (p = 0.027) but not gender. In contrast to the previous study, the drop in SpO2 was not correlated with Borg dyspnea score or age. The differences between the studies are likely due to different age distribution and populations. Further, Villalba, et al17 did not specify whether the dyspnea score was pre- or post-walk, which may explain the discrepancy with our results.
Limitations of our study include missing data, a problem common to all retrospective studies. Subjects in the PH subgroup were diagnosed by either right-heart catheterization or echocardiogram with sPAP > 40 mm Hg. Because some subjects did not have right-heart catheterization close to the time of the 6MWT, some of these individuals may not have World Health Organization Group I PAH. Further, we defined SSc-ILD very sensitively, with any ground-glass opacity on chest CT regardless of FVC%, a categorization that is different from some studies. The 6MWD was measured on supplemental oxygen for some patients, which may have altered our ability to measure distance or desaturation. Some SSc groups had very small numbers, limiting our analytic abilities. Therefore, our results should be validated in a larger prospective study. A strength of our study was the comparison of subjects with SSc-pulmonary disease to those with IPF. A limitation was the lack of an idiopathic PAH comparator group due to the small number of patients with idiopathic PAH at our center. A future study may provide more insight into mechanisms of exercise intolerance in SSc. Unfortunately, our database did not contain other disease features such as skin extent, disease duration, arthritis, or myositis. Information on these factors would be extremely valuable in a future study to elucidate the specific variables that confound the association between 6MWD and pulmonary disease measures. Finally, few of the IPF subjects underwent echocardiogram or right-heart catheterization, so these patients could have undetected cor pulmonale, which could confound the results.
Our study reveals that the 6MWT is influenced by many factors that affect the specificity for pulmonary disease; however, these factors may not affect the sensitivity of the test. Two important future studies remain to be done in this area. The first would be to determine the responsiveness of the 6MWD and desaturation to changes in disease state for the different SSc lung disease phenotypes. An important longterm project would define the prognostic value of a low 6MWD between subgroups.
Limitations due to pain confound the utility of the 6MWT, particularly in SSc. 6MWT interpretation should include consideration of vascular, pulmonary, and musculoskeletal limitations to exercise. Sensitivity of the 6MWD to these various organ-system disease processes may necessitate exclusion of subjects with pain or the use of multiple endpoints to fully improve sensitivity of the outcome measures to pulmonary disease19. Continued investigation of how to use the 6MWT in the clinical care and research of patients with SSc is warranted.
Footnotes
- Accepted for publication September 23, 2008.
