Increased ventilation with NiIPPV does not necessarily improve exercise capacity in COPD

Subjects

Subjects were recruited from the population of patients with COPD attending the present authors' centre. The inclusion criteria were airflow obstruction (defined as a forced expiratory volume in one second (FEV₁) <70% predicted 13 and an FEV₁/forced vital capacity (FVC) ratio <70% pred) and a history of impaired exercise tolerance limited by breathlessness. This was further examined by steady-state treadmill exercise as detailed below (see Experimental protocol day 1). Subjects were excluded if they could walk for >10 min on the treadmill at a speed found to be their maximum on a shuttle walking test 14. Subjects were excluded if they had recorded reversibility to steroids, another pulmonary disorder in addition to COPD, another medical condition likely to limit exercise capacity (such as cardiovascular or neuromuscular disease), or any change of symptoms or drug therapy in the 4 weeks prior to the study. All subjects gave informed consent.

Ventilators

Three different ventilators were compared: Bipap S/T 30 (Respironics Inc., Murrayville, PA, USA), Nippy2 (B+D Electrical Ltd, Stratford upon Avon, UK) and Vpap II ST (Resmed Ltd, Abingdon, UK). NiIPPV was applied with each subject wearing noseclips and breathing via a mouthpiece. The triggered/timed mode and the minimum back-up rate were used with each machine, to ensure that all ventilator breaths were triggered by the subject. The minimum expiratory airway pressure (EPAP) was used throughout. Maximum inspiratory airway pressure (IPAP) and inspiratory times (T_i) were set for each machine determined by patient comfort at rest and were not altered during exercise. The ventilators were attached to the mouthpiece using identical circuits incorporating a Whisper swivel II expiratory valve (Respironics Inc.). A Fleisch No.3 pneumotachograph (Phipps+Bird, Richmond, VA, USA) and a Vyggo pressure transducer (Vygon Ltd, East Rutherford, NJ, USA) with a range of 200 cmH₂O were inserted in the circuit between the mouthpiece and the expiratory valve to record the expiratory volumes (V_T) and pressure within the circuit.

Experimental protocol

An outline of the protocol is shown in figure 1⇓. Subjects were asked to perform walking tests on a treadmill to compare their exercise capacity under three different conditions. These were as follows: unencumbered breathing via a mouthpiece with noseclips, and breathing via a mouthpiece with noseclips and attached to one of the ventilators. For all walking tests, pulse oximetry (S_p,O2) using a finger probe (Ohmeda, Hatfield, UK), respiratory rate (RR) and cardiac frequency were documented before and after the test. For walks using the mouthpiece (with and without the ventilators), expiratory V_T, RR, T_i and expiratory time were recorded continuously. IPAP was also recorded during the ventilator walks. These were stored on a CARDAS data logging system (Oxcams Medical Sciences Ltd, Oxford, UK) for subsequent analysis. In addition, during these walks, each subject was also connected to a three lead electrocardiogram (ECG) and the finger probe was used throughout.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.—

Outline of experimental protocol. S_p,O2: peripheral oxygen saturation. Walks are classified as follows. Walks A–D mouthpiece alone, mouthpiece with Bipap, mouthpiece with Nippy2 and mouthpiece with Vpap II ST in random order. ^#: 30‐min rest between walks; ^¶: performed in random order and with 30‐min rest between walks.

All treadmill walks were carried out at constant speed and were terminated by breathlessness. The speed was initially set at 50% of the maximum speed attained during a previous shuttle-walking test. The speed was then titrated upwards to a pace each subject felt was equivalent to a brisk walk (and kept constant for subsequent tests). To complete the protocol, subjects were asked to attend on three occasions at the same time of day within a maximum period of 8 days. Subjects were asked to avoid food, caffeinated drinks and any bronchodilator medication in the 2 h prior to attendance. Otherwise subjects continued their normal medication throughout the study period.

Statistics

In all tests, a p‐value <0.05 was considered significant. Data are presented as mean±sd. A post-hoc power analysis was performed on the effect of time order on walking distance for the unencumbered walks.

Subjects

Twelve subjects were screened for the study. Four were excluded, as they were able to walk for >10 min on the treadmill at the maximum speed achieved on the shuttle-walking test. Eight subjects (two female), with a mean age 66±8 yrs, completed the study protocol. The demographic details are given in table 1⇓. The mean FEV₁ was 1.0±0.4 L (36% pred) and the mean FVC was 2.8±0.8 L (85% pred). The mean FEV₁/FVC ratio was 34±13%. Three subjects were already using nocturnal home ventilation; two subjects had used this treatment previously and one subject was using nocturnal continuous positive airway pressure (CPAP) via a nasal mask for obstructive sleep apnoea. The mean shuttle-walking distance was 235±83 m. The mean treadmill speed was 3.7±1.1 km·h⁻¹, which was 87±12% of the maximum speed determined during the shuttle-walking test. The mean preset IPAP during the ventilator walks was 12.2±2.2 cmH₂O.

Distances

Using analysis of variance for repeated measures (ANOVA), no significant difference in walking distance was found within the four unencumbered walks and they were combined for further analysis. There was a trend towards time-ordered improvement, although this was not statistically significant (p=0.3). However, the study protocol gave an estimated observed power of 50% for this and therefore the lack of significance cannot be confirmed. There were no differences within the three ventilator walks (fig. 2⇓) and the data were therefore combined. Mixed-effects ANOVA was used to compare the distances walked according to type (fig. 3⇓). The mean distance walked during the unencumbered walks was 259±123 m. The addition of monitoring equipment and the mouthpiece reduced the mean distance to 211±96 m and a further fall to 145±76 m was seen following the application of NiIPPV. The difference between the groups was significant (p=0.02). Post-hoc analysis with the Scheffe test showed that only the comparison between the unencumbered walks and the ventilator walks was significant (p<0.01).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.—

Distance walked for a) unencumbered walks, b) ventilator walks and c) time-ordered ventilator walks. Data are presented as individual data points and group means (indicated by horizontal lines). a) p=0.3, b) p=0.07 and p=0.15.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.—

Distances walked according to type. Data are presented as individual data points (▪) and group means (○). p=0.02.

Exercise physiology

Mixed-effects ANOVA was also used to compare the degree of desaturation, cardiovascular response and RR measured immediately postexercise between the three types of walk (table 2⇓). No significant differences were observed between the groups.

For the walks using the mouthpiece, measured parameters were compared for the last 20 s of exercise. Using ANOVA for repeated measures, significant differences (p<0.05) were seen between ventilators for T_i and IPAP. Mean values for T_i were as follows: Bipap 0.85±0.14, Nippy2 1.0±0.13 and Vpap 0.89±0.14 s. Mean values for IPAP were: Bipap 12±2, Nippy2 15±3, and Vpap 13±2 cmH₂O. No significant differences were seen between ventilators for IPAP measured before the start of exercise: Bipap 11±2, Nippy2 13±3, and Vpap 13±2 cmH₂O. For all other parameters there were no significant differences between ventilators. Mixed-effects ANOVA was used to distinguish differences between the ventilators and mouthpiece walks (table 3⇓); significant increases in V_T and minute ventilation (V'_E) were seen during the ventilator walks.

Discussion of results

NiIPPV did not improve exercise capacity in subjects with COPD and there was no difference between the brands of ventilator tested, despite the documented differences in their performance characteristics 12. There was no benefit from using NiIPPV, although there were significant increases in V_T, V'_E and V_T/vital capacity compared with the mouthpiece walks. In previous studies, Dolmage and Goldstein 10 showed an increase in V'_E and exercise endurance using PAV, while Bianchi et al. 9 found a significant increase in exercise endurance using PSV (and PAV), which was not accompanied by an increase in V'_E. An increase in maximal ventilation alone is, therefore, neither necessary nor sufficient to improve exercise performance.

It may be that the increase seen in the V'_E/MVV ratio in the patients in this study increased the sensation of dyspnoea and terminated exercise prematurely when using PSV. It is also possible that the cost of an increase in V'_E was greater respiratory effort. Patient work is expended in triggering the ventilator to inspiration and expiration, and will be increased by any incoordination between subject and ventilator. Incoordination was particularly obvious with the Nippy2, for which the T_i is preset. The rise in IPAP seen during exercise was due to the subjects expiring before the ventilator T_i was complete. For each of the machines tested, EPAP represented an additional resistance to expiratory flow. Expiratory airflow is severely compromised in COPD patients and EPAP may result in active expiration with abdominal muscle recruitment 15 contributing to breathlessness.

In contrast to the current findings, other authors have shown increased exercise capacity using PSV without positive end-expiratory pressure (PEEP) 8, PAV with CPAP 10, and PAV, PSV and CPAP in decreasing order of effectiveness 9. However, there are methodological problems with each of these studies. In two, the investigators did not include an unencumbered control exercise test 9. While, in the other, a control walk was performed but the results were not compared directly with the ventilator walks 8. The use of monitoring and breathing equipment can impair performance 16 and this will have been particularly marked in the study where a nasal mask was used 9. Nasal breathing is common at rest but oronasal breathing appears to be universal during exercise and so this is not a realistic exercise condition 18. In the current study, post-hoc analysis showed that only the difference in walking distance between the ventilator walks and unencumbered walks was significant, demonstrating the need to include all test results in the statistical analysis.

In two of the earlier studies, the order of the exercise tests was not fully randomised and a cycle ergometer was used as the exercise condition 9. Randomisation is important, as exercise tests are subject to learning effects 14. In the current results this is illustrated by the observed (though statistically insignificant) learning effect seen in the unencumbered walks. Since four unencumbered walks were compared with three ventilator walks, the learning effect may have exaggerated the difference between the walk types. However, this learning effect was small compared with the difference between the walk types. A treadmill was used, as this is more similar to normal daily activities than cycle exercise. Cycle and treadmill exercise are not interchangeable in COPD, as cycling leads to a greater rise in lactate at a comparable workload 21 with increased ventilation and breathlessness 22, which may limit performance.

Limitations of the present study

The number of subjects that were recruited was small but comparable with other studies with positive results 8. In agreement with previous authors 17, the current data show that exercise capacity was impaired by the use of a mouthpiece in patients with COPD. The mouthpiece has a much smaller deadspace compared with a facemask, but may impose an important increase in resistance to airflow. It also prevents purse-lip respiration and the subject may, therefore, lose control over the degree of intrinsic PEEP. With the whisper swivel valve, which was used in this study, up to 60% of the expired air may remain in the ventilator circuit at the end of expiration 24. To maintain equivalent blood gases, V'_E must increase 25. Other studies of assisted ventilation and exercise have used circuits and expiratory valves that lead to less carbon dioxide rebreathing 8–10.

COPD patients develop pulmonary hypertension during exercise due to increased pulmonary vascular resistance 26. NiIPPV increases pulmonary artery pressure at rest and may reduce cardiac output 27. In subjects with cardiac failure, CPAP may improve 28 or impair 29 cardiac output. The greatest improvements are seen in patients with high pulmonary capillary wedge pressure 29 and left ventricular compliance 28. In common with previous authors 8–10, left ventricular function was not determined. Diastolic dysfunction, which might be worsened by PEEP, is a possible confounding factor to explain some of the differences seen between the present results and those of Keilty et al. 8.

Practical implications

From the present results, it can be stated that bilevel PSV delivered via a mouthpiece with a whisper swivel valve will not increase exercise capacity and has no role in pulmonary rehabilitation exercise programmes. The large difference seen between the ventilator condition and the unencumbered walks questions previous positive trials of other modes of ventilatory support during exercise that did not make a comparison with an unencumbered condition. Further studies are required to make these comparisons and to investigate the interactions between the ventilator and cardiac function.

Conclusions

These results do not show any benefit from bilevel pressure support ventilation with any of the three different ventilators, despite the increase seen in expiratory volume. Possible explanations for these negative results include: a failure to detect a real difference due to insufficient subject numbers, an increase in work of breathing due to positive end-expiratory pressure, incoordination during expiration, carbon dioxide rebreathing, and a fall in cardiac output due to cardiac dysfunction or pulmonary hypertension. There are methodological weaknesses also seen in previous studies of assisted ventilation and exercise, and, in particular, the lack of comparison to an unencumbered baseline. The value of noninvasive-assisted ventilation to increase exercise capacity in chronic obstructive pulmonary disease patients remains uncertain.