3.1. Normality Test
The aerobic exercise THR data obtained from MetS patients using standardized methods and three maximum heart rate prediction equations, as well as HRVT1 data collected during the CPET, all showed a normal distribution (P>0.05). The results of the Kolmogorov‒Smirnov test are presented in Table 1.
3.2. Comparison of the Standardized Methods THR and HRVT1
The THR determined by the standardized %HRR method (r>0.8, P<0.001) and HRVT1 were strongly correlated. The THR obtained by combining 35% HRR with any one of the three HRmax equations was not significantly different from HRVT1. That is, the difference between 35% HRR (113.56±8.20) calculated with the Fox equation and HRVT1 (113.21±9.16) was not statistically significant (P=0.600), the 35% HRR (114.72±6.80) calculated with the Miller equation was similar HRVT1 (P=0.051), and the 35% HRR (114.12±7.28) calculated with the Tanaka equation was similar to HRVT1 (P =0.200) (Table 2).
The THR determined by the standardized %HRmax method (0.8>r>0.6, P<0.001) and HRVT1 were relatively highly correlated. Only the THR obtained by combining 65% HRmax with the two HRmax equations was not significantly different from HRVT1. The 65% HRmax (114.20±6.90) calculated with the Fox equation was not significantly different from HRVT1 (P=0.326). The 65% HRmax (115.04±4.83) calculated with the Tanaka equation was not significantly different from HRVT1 (P=0.075) (Table 3).
3.3. Bland‒Altman Consistency Testing
According to the standardized reserve heart rate method, the mean difference between the 35% HRR determined by the Fox equation and HRVT1 was -0.350, which was close to the 0th line (Table 4); 95.83% of the patients (46/48) had values within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 7.90 (Figure 1A). The mean difference in the test results for 35% HRR according to the Miller equation and HRVT1 was -1.509, which was relatively close to the 0th line (Table 4); 95.83% of the patients (46/48) had values within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 9.82 (Figure 1B). The mean difference in the test results for 35% HRR according to the Tanaka equation and HRVT1 was -0.913, which was quite close to the 0th line (Table 4); 95.83% (46/48) of the points were within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 8.85 (Figure 1C).
Therefore, among the three HRmax prediction equations, the 35% HRR determined by the Fox equation was highly consistent with HRVT1, the 35% HRR determined by the Tanaka equation was fairly consistent with HRVT1, and the 35% HRR determined by the Miller equation was the least consistent with HRVT1. Compared to the other two equations, the Miller equation for 35% HRR does not meet the conditions for practical application even though it is not significantly different from HRVT1.
According to the standardized maximum heart rate method, the mean difference in the test results for the 65% HRmax determined by the Fox equation and HRVT1 was -0.989, close to the 0th line (Table 5); 91.67% of the patients (44/48) had values within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 7.80 (Figure 2A). The mean difference in the test results for 65% of the HRmax determined by the Miller equation and HRVT1 was -2.97, which was relatively close to the 0th line (Table 5); 93.75% of the patients (45/48) had values within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 11.23 (Figure 2B). The mean difference in the test results for 65% of the HRmax values determined by the Tanaka equation and HRVT1 was -1.829, which was relatively close to the 0th line (Table 5); 91.67% of the patients (44/48) had values within the 95% limits of agreement, and the absolute value of the maximum difference within the limits of agreement was 9.36 (Figure 2C).
Therefore, among the three HRmax predictions, the 65% HRmax determined by the Fox equation was highly consistent with HRVT1, the 65% HRmax determined by the Tanaka equation was less consistent with HRVT1, and the 65% HRmax determined by the Miller equation was the least consistent with HRVT1.
3.4. ICC Reliability Testing
Like the Bland‒Altman test, the ICC test showed that among the three standardized HRRs, only 35% HRR combined with the three HRmax equations showed high consistency with HRVT1 (113.21±9.16), and among the three equations, the 35% HRR calculated with the Fox equation exhibited the highest consistency with HRVT1 (Table 6). The ICC for 35% HRR (113.56±8.20) determined by the Fox equation and HRVT1 was 0.862 (P<0.001), and the 95% confidence interval was 0.767-0.920, indicating high consistency and good accuracy. The ICC for 35% HRR (114.72±6.80) determined by the Miller equation and HRVT1 was 0.780 (P<0.001), and the 95% confidence interval was 0.636-0.871, indicating relatively high consistency and comparatively good accuracy. The ICC for 35% HRR (114.12±7.28) determined by the Tanaka equation and HRVT1 was 0.825 (P<0.001), and the 95% confidence interval was 0.709-0.898, indicating high consistency and good accuracy (Table 6).
Among the three standardized HRmax percentages, only the 65% HRmax calculated with the Fox and Tanaka equations showed moderate consistency with HRVT1 (113.21±9.16) (Table 7). The ICC for the 65% HRmax (114.20±6.90) determined by the Fox equation and HRVT1 was 0.639 (95% confidence interval 0.437-0.780) (P<0.001), with moderate consistency and normal accuracy. The ICC for the 65% HRmax (115.04±4.83) determined by the Tanaka equation and HRVT1 was 0.537 (95% confidence interval 0.304-0.710) (P<0.001), with moderate consistency and normal accuracy (Table 7).