Maximal Inspiratory Pressure: Results
The sample was 69 men and 22 women with moderate to severe COPD (Table 1). Most patients were taking standard pharmacologic therapies, including inhaled p-agonists, oral (3-agonists, and methylxan-thines. Nine were using home oxygen. Ten were currently smoking, one had never smoked, and the remainder had quit smoking. The sample was middle class as indicated by the Hollingshead Index.
Performance gradually improved as demonstrated by the mean (SD) Pimax for each test session: first — 53 (24) cm H20, second -57 (25) cm H20, third -60 (25) cm H20, and fourth -62 (25) cm H20. Analysis of variance for repeated measures identified a significant increase in Pimax with a significant linear effect and no significant curvilinear effect over time (p<0.001). Statistically, performance did not plateau; however, between the third and fourth tests, the net increase in Pimax was so small that it was not clinically significant and performance appeared to be plateauing (Table 2). The frequency distribution for net differences in Pimax is presented in Figure 1. A weak negative correlation between Pimax and age was strengthened from the first (r= —19; p>0.05) to the fourth test session (r= —0.30; p<0.01).
Sample characteristics were examined to determine if selected characteristics influenced the magnitude of improvement observed with four test sessions. Age correlated negatively (r= —0.33; df= 89) with net differences in the first and fourth measures of Pimax (Pimax4 minus Pimaxl). However, the correlation with age decreased when net difference scores were calculated either as a percent of the initial Pimax (r = — 0.19; df = 89) or as a percent of the fourth Pimax (r= —0.16; df= 89). No significant correlations were observed between improvement in Pimax and spirometry or body mass index. Mean net differences and mean absolute differences among repeated measures of Pimax were similar for men and women.
This finding was further demonstrated by grouping patients based on the magnitude of improvement from the first to fourth measure of Pimax. From the first to the fourth measure, 61 patients increased Pimax by <10 cm H20 and 30 patients increased by >10 cm H20. At the first test session, Pimax was similar for both groups with means of — 53 (SD = 25) cm H20 for the group that improved by ^10 cm H20 and —54 (SD = 22) cm H20 for the group that improved by > 10 cm H20. At the fourth test session, mean Pimax was -56 (SD = 25) and -74 (SD = 22) cm H20, respectively, for the same groups. Sample characteristics for these groups were compared using independent t tests. The groups were significantly different with respect to age (p<0.01) and not significantly different for any other variables (Table 3).
The test-retest reliability was examined by comparing the third and fourth test of Pimax. Third and fourth tests of Pimax were highly reliable as reflected by a mean absolute difference of less than 5 cm H20 (Table 2). A cumulative frequency distribution of absolute differences in Pimax illustrated that 79 percent of subjects were within 5 cm H20 and 93 percent of subjects were within 10 cm H20 on third and fourth tests (Fig 2). Test-retest reliability coefficient was r = 0.97 (df = 89) for the third and fourth test of Pimax.
Data were further examined to determine iffive trials were sufficient to attain peak performance at each test session. Thirty-four percent of the sample generated their best Pimax at the first trial. Ten percent generated their best Pimax at the fifth trial.
Table 1—Sample Characteristics (N=91)
Mean (SD) | |
Age, yr | 65 (7) |
FEV,, % predicted | 40 (13) |
FVC, % predicted | 68 (16) |
FEV./FVC | 41 (9) |
MW,* % predicted | 40 (22) |
Weight, kg | 72.45 (16.2) |
Height, cm | 170 (10) |
Body mass index | 25 (4) |
Hollingshead Index | 40 (12) |
Table 2—Differences in Pimax (—cm НгО) as Measured at Four Test Sessions (N=91)
Comparison of Tests | Net Difference Betw een Testsf | 95% Cl Net Difference | Absolute Difference Between Testsf | 95% Cl Absolute Difference | Range of Differences Between Testsf |
2 and 1 | 4 (8) | 2- 5 | 7 (6) | 6- 8 | -20 to 30 |
3 and 2 | 3 (8) | 1- 4 | 5 (6) | 4- 7 | -10 to 35 |
4 and 3 | 2 (6) | 1- 4 | 4 (5) | 3- 5 | -25 to 15 |
3 and 1 | 7 (10) | 5- 9 | 9 (9) | 7-10 | -20 to 35 |
4 and 1 | 9 (10) | 7-11 | 11 (8) | 9-12 | -15 to 30 |
4 and 2 | 5 (8) | 4- 7 | 7 (6) | 6- 8 | -15 to 30 |
Table 3 —Comparison of Patients Grouped According to the Magnitude of the Learning Effect From the First to the Fourth Measure of Pimax
<10 cm H20(n = 61) | > 10 cm H20(n = 30) | |
Age, vr | 66 (7) | 61 (6)t |
FEV,, % predicted | 39 (13) | 41 (14) |
FVC, predicted | 67 (17) | 70 (15) |
FEV./FVC | 41 (9) | 42 (10) |
MW, % predicted | 39 (20) | 43 (25) |
Weight, kg | 71.55 (16.65) | 74.25 (14.85) |
Height, cm | 170 (10) | 170 (7.5) |
Body mass index | 24 (4) | 25 (4) |
Pimax, first measure, cm II20 | -53 (25) | -54 (22) |
Figure 1. Frequency distribution for net differences in maximal inspiratory pressure (Pimax) as measured at the (1) first and second, (2) second and third, and (3) third and fourth test sessions. Differences were calculated as the earlier test subtracted from the later test so that positive numbers reflect improvements in Pimax.
Figure 2. Cumulative frequency distribution for absolute differences in maximal inspiratory pressure (Pimax) measured at third and fourth test sessions. The X axis represents the absolute difference between the third and fourth Pimax. The Y axis represents the cumulative percentage of patients. Each point on the graph reflects the cumulative percentage of patients with an absolute difference in Pimax less than or equal to the value on the X axis.