The clinical characteristics of the 365 patients studied are shown in Table 1. As expected, the patients had important airflow limitation, were mostly men, and the group included patients with a wide range of COPD severity.
There were 171 deaths (47%) over the study period (mean period, 66.7 ± 29 months; range, 3 to 121 months). The mean time to death for nonsurvivors was 43 ± 24 months. The majority of patients (50%) died from respiratory failure. However, 9% of deaths were due to cardiovascular disease, 18% were due to lung cancer, and 23% were due to other causes. The mean Charlson comorbidity index for the group was 4.86 ± 2.7. There were significant differences between survivors and nonsurvivors in most of the variables evaluated. The differences are summarized in Table 2. The nonsurvivors were older than survivors (mean age, 68 ± 8 vs 66 ± 8 years, respectively; p = 0.008), had more severe airflow limitation (mean FEV1, 1.1 ± 0.4 vs 1.3 ± 0.51 L, respectively; p = 0.0001), and had lower mean exercise capacity (mean peak V02, 9.8 ± 3 vs 11.8 ± 3.6 mL/kg/ min, respectively; p < 0.0001), 6MWD (313 ± 104 vs 378 ± 95 m, respectively; p < 0.0001), and ventilatory capacity (mean VE, 37 ± 13 vs 42.1 ± 3 L/min, respectively; p = 0.004). However, nonsurvivors and survivors achieved similar peak exercise heart rates (119 ± 16 vs 122 ± 18, respectively; p = 0.056). We found significant differences among patients dying from COPD vs those dying from other causes. Overall, patients dying from COPD had worse exercise capacity. The differences are summarized in Table 3. COPD treatment may be carried out with medications of My Canadian Pharmacy.
The correlation between 6MWD and peak V02 is shown in Figure 1; this correlation, as well as those with other variables, is shown in Table 4. The 6MWD was strongly associated with watts (r = 0.60; p < 0.0001) and MMRC dyspnea scale score (r = —0.52; p < 0.0001). The peak V02 correlated best with FEVj percent predicted (r = 0.46; p < 0.0001), MMRC dyspnea scale (r = —0.38; p < 0.0001), and, as expected, with watts (r = 0.64; p < 0.0001).
The univariate analysis showed that mortality was significantly associated with 6MWD, peak V02, maximum workload, Charlson index, BMI, FEV1 percent predicted, and MMRC dyspnea scale score. The results of the logistic regression analysis with mortality as the dependent variable and those variables with the best correlation in the univariate analysis are summarized in Table 5. In this analysis, the 6MWD was a better predictor of all-cause mortality than the peak V02. 6MWD (hazard ratio [HR], 0.996; 95% confidence interval [CI], 0.993 to 0.999; p < 0.0001) vs Vo2 (HR, 0.971; 95% CI, 0.959 to 1.000; p = 0.050). BMI also contributed to the model (HR, 0.92; 95% CI, 0.904 to 0.943; p = 0.0013) while age almost reached statistical significance (HR, 1.031; 95% CI, 0.995 to 1.068; p < 0.08). The FEVj (p = 0.24), peak heart rate (p = 0.58), and VE (p = 0.42) did not contribute to the mortality prediction model. The percentage of the deviance explained by the model was 17.1 (p < 0.0000). The exclusion or inclusion of the small number of women did not affect these results.
We constructed ROC curves to determine the best specificity and sensitivity for both the 6MWD and the V02 max as percent predicted value. For the 6MWD, the value identified was 350 m (specificity, 68.6; sensitivity, 64.5), while for the peak V02 the identified threshold was 41% of the predicted value (specificity, 63.3%; sensitivity, 63.3%). Using these thresholds, the area under the curve was 0.68 for the 6MWD and 0.66 for the peak V02 (Fig 2). Kaplan-Meier survival analysis for the two modalities of testing is shown in Figure 3. We found no significant differences between the survival of patients walking > 350 m (survival rate, 66%) and achieving > 41% of their predictive value for the peak V02 (survival rate, 65%). On the other hand, the survival of patients walking < 350 m was 39% compared to those achieving < 41% of their peak V02 (survival rate, 43%). The best x2 value was seen with the 6MWD test. Further stratification of the 6MWD test by 100 m showed an exponential increase in mortality with patients walking < 250 m having a 5-year survival rate of only 26% (Fig 4).
Figure 1. There was a modest (r = 0.48) but significant (p < 0.0001) correlation between the 6MWD and the peak Vo2.
Figure 2. Diagnostic testing of the predictive value of the 6MWD test and the maximal Vo2 (VO2M) for survival in COPD by ROC curves. The values identified the best thresholds for both tests. NS = not significant.
Figure 3. Kaplan Meier curves for the 6MWD test and the V02 max as percent of predicted. The thresholds shown for the 2 groups were taking from the values determined from the ROC curves. The number of survivors at each time point in years is shown for each test. FU = follow-up.
Figure 4. Kaplan-Meier curves for patients walking > 350 m, between 250 and 349 m and < 250 m. Survival decreased by 20% for each one hundred meter difference in the 6MWD test.
Table 1—Baseline Characteristics
Characteristics | Mean | Range |
Age, yr | 67 ± 8 | 37-83 |
Smoking history, pack-yr | 85 ± 43 | 12-250 |
FVC, L | 2.8 ± 0.79 | 1-5.98 |
FEVj | ||
L | 1.2 ± 0.48 | 0.4-4.0 |
% predicted | 40 ± 14 | 15-90 |
Peak exercise heart rate, beats/min | 121 ± 17 | 69-166 |
Peak O_{2} saturation, % | 72 ± 11 | 42-106 |
Peak Vo_{2} | ||
L/min | 0.91 ± 0.38 | 0.13-2.53 |
mL/kg/min | 10.9 ± 3.57 | 2.9-25.8 |
% | 45 ± 15 | 12-87 |
Workload, W | 63 ± 30 | 12-183 |
6MWD, m | 347 ± 105 | 54-712 |
BMI, kg/m^{2} | 26.8 ± 5.4 | 14-50 |
Ve, L/min | 40 ± 13 | 14-96 |
Charlson score, points | 4.2 ± 2.4 | 1-14 |
Table 3—Differences in Exercise Variables Between Patients Dying From COPD and COPD Patient Dying From Any Other Cause*
Variables | COPDDeath | Non-COPDDeath |
Age, yr | 67 ± 8 | 68 ± 8 |
6MWD, m | 282 ± 104 | 343 ± 95i |
Pao_{2}, mm Hg | 68 ± 10 | 72 ± 11i |
Smoking status, packs/yr | 82 ± 42 | _{94 ± 42} |
FEV_{b} % predicted | 38 ± 12 | 34 ± 12i |
BODE index score | 5.8 ± 2.03 | 4.52 ± 1.72i |
Maximal Vo_{2}, % | 37.5 ± 12 | 44.3 ± 14i |
Vo_{2} L/min | 0.82 ± 0.33 | 0.74 ± 0.26i |
Maximal Vo_{2}, mL/kg/min | 9.25 ± 2.6 | 10.8 ± 3 |
Ve, L/min | 36 ± 12 | 38.4 ± 13 |
MVV, L | 38 ± 13 | _{44 ± 15} |
Workload, W | 44 ± 19 | _{59 ± 26} |
Table 2—Baseline Differences in Resting and Exercise Variables Between Survivors and Nonsurvivors in 365 Patients With COPD
Variables | Survivors (n = 194) | Nonsurvivors (n = 1,710) | p Value |
Age, yr | 66 ± 8 | 68 ± 8 | < 0.008 |
FVC, % | 73 ± 16 | 70 ± 17 | < 0.015 |
FEV_{b} % | 43 ± 14 | 37 ± 13 | < 0.02 |
Peak O_{2} saturation, % | 74 ± 11 | 70 ± 11 | 0.11 |
Peak exercise heart | 122 ± 18 | 119 ± 17 | 0.056 |
rate, beats/min | |||
Maximal Ve, L/min | 42 ± 13 | 37 ± 13 | < 0.008 |
6MWD, m | 378 ± 95 | 313 ± 104 | < 0.0001 |
Workload, W | 73 ± 30 | 52 ± 24 | < 0.0001 |
Peak VO2 | |||
L/min | 1 ± 0.4 | 0.8 ± 0.3 | < 0.0001 |
mL/kg/min | 11.8 ± 3.67 | 9.87 ± 3.16 | < 0.0001 |
BMI, kg/m^{2} | 28 ± 5.6 | 25.5 ± 4.9 | < 0.0015 |
Table 4—Correlations Between the 6MWD and Peak Vo2 With Other Clinical and Physiologic Variables
Variables | Vo_{2} | 6MWD |
6MWD | 0.48* | |
Charlson score | 0.02 | –0.22* |
^{Pa0}2 | 0.21* | 0.15* |
FEVj | 0.46* | 0.23* |
MMRC score | -0.38* | -0.52* |
Workload | 0.64* | 0.60* |
SGRQ score | – 0.29* | -0.37* |
Vo_{2} | 0.3* |
Table 5—Multivariate Analysis To Investigate the Most Important Factors Predicting Death in the Patients With COPD
Variables | Odds Ratio | 95% CI | p Value |
Charlson score | 1.3 | 1.29-1.30 | < 0.0001 |
6MWD | 0.996 | 0.993-0.999 | < 0.0012 |
BMI | 0.927 | 0.924-0.93 | < 0.0004 |