If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Respiratory and Critical Care Medicine and Ludwig Boltzmann Institute for COPD, Otto-Wagner Hospital, Vienna, AustriaMedical University of Vienna, Austria
COPD patients who remain hypercapnic after acute respiratory failure requiring mechanical ventilation have a poor prognosis. Long-term nocturnal non-invasive ventilation (NIV) may be beneficial for these patients. We hypothesized that stable patients on long-term NIV would experience clinical worsening after withdrawal of NIV.
Methods
We included 26 consecutive COPD patients (63 ± 6 years, 58% male, FEV1 31 ± 14% predicted) who remained hypercapnic after acute respiratory failure requiring mechanical ventilation. After a six month run-in period, during which all patients received NIV, they were randomised to either continue (ventilation group, n = 13) or to stop NIV (withdrawal group, n = 13). The primary endpoint was time to clinical worsening defined as an escalation of mechanical ventilation.
Results
All patients remained stable during the run-in period. After randomisation the withdrawal group had a higher probability of clinical worsening compared to the ventilation group (p = 0.0018). After 12 months, ten patients (77%) in the withdrawal group, but only two patients (15%) in the ventilation group, experienced clinical worsening (p = 0.0048). Six-minute walking distance increased in the ventilation group.
Conclusion
COPD patients who remain hypercapnic after acute respiratory failure requiring mechanical ventilation may benefit from long-term NIV.
The benefit of non-invasive ventilation (NIV) for the treatment of acute respiratory failure in COPD has been well documented in multiple studies and meta analyses;
To date, one sufficiently powered study suggests that NIV conveys a small survival benefit. There are conflicting results, however, about the effects of NIV on clinical stability as assessed by hospitalisation rate.
In addition, those COPD patients who are hypercapnic on hospital discharge have a higher mortality rate than those who are normocapnic. Therefore, we hypothesised that NIV would be beneficial for COPD patients who remain hypercapnic after an episode of acute respiratory failure requiring mechanical ventilation. If this hypothesis were true, the withdrawal of long-term nocturnal NIV should induce clinical worsening in stable COPD patients. Hence, the objective of this study was to determine whether the withdrawal of long-term NIV causes clinical worsening in stable COPD patients who remained hypercapnic after an episode of acute respiratory failure requiring mechanical ventilation.
Methods
This trial was approved by the Institutional ethics committee, and written informed consent was obtained from all participants.
Participants
The study was conducted in the Department of Respiratory and Critical Care Medicine at the Otto Wagner Hospital, Pulmonary Centre, Vienna. Patients were recruited on the respiratory care unit, which cares for patients with acute and chronic respiratory failure and patients with prolonged weaning, including those who require long-term invasive or non-invasive ventilation. Inclusion criteria for the study were as follows: (1) COPD according to standard criteria;
(2) requiring invasive or non-invasive mechanical ventilation during an episode of acute respiratory failure due to exacerbation of COPD, according to standard criteria;
(3) clinically stable and hypercapnic (PaCO2 > 45 mmHg measured immediately after awakening from a night without mechanical ventilation) after six months on home nocturnal NIV and (4) patient approval.
Exclusion criteria were (1) severe psychiatric disorder likely to impair the ability to comply with NIV; (2) other severe pulmonary diseases except COPD; (3) other severe non-pulmonary disease limiting prognosis (e.g., metastatic cancer, congestive heart failure, liver cirrhosis); (4) non-compliance with NIV; (5) women of childbearing age and (6) polysomnographic evidence of sleep apnoea (>15 apnoeas plus hypopnoeas per hour of sleep).
Conduct of the trial
Treatment of acute respiratory failure
Acute respiratory failure was treated according to international standards.
Mechanical ventilation was applied non-invasively using individual interfaces and/or invasively via an endotracheal tube or tracheal cannuala. NIV was used as the preferred primary method and was also used to facilitate weaning.
After primary stabilisation from acute respiratory failure and upon enrolment, all patients were free of an invasive airway.
Initiation of NIV
COPD patients in our department are routinely evaluated for nocturnal NIV if they remain hypercapnic with symptoms of chronic respiratory failure after an episode of acute respiratory failure requiring mechanical ventilation. Evaluation was started before discharge from the ICU or immediately after transfer to regular wards. Patients who started nocturnal NIV were transferred to the respiratory care unit for three to five days and trained to use NIV according to a standardised protocol. Various types of patient-triggered bi-level positive pressure ventilators were used. The level of inspiratory aid was increasingly raised from 10 cm H20 to approximately 20 cm H20, with a level of expiratory positive airway pressure around 5 cm H20. The inspiratory time was limited to a maximum of 1.3 s to avoid leak-induced prolongation of inspiration. The fraction of inspired oxygen was adjusted to correct for nocturnal hypoxaemia, as ascertained by nocturnal oximetry. The choice of nasal versus full face mask and humidification was determined by the respiratory physiotherapist to maximise patient comfort. NIV efficiency was ascertained by nocturnal oximetry and serial arterial blood gases.
Six month run-in phase
All patients meeting the study criteria who were discharged from the respiratory care unit with home NIV were enrolled in the run-in phase. At that time, patients were informed about the trial, and informed consent was obtained. The run-in phase lasted six months, during which patients needed to remain stable on home nocturnal NIV. After three months, they were re-admitted to the respiratory care unit for one night for re-assessment of the inclusion and exclusion criteria. Therefore, patients were excluded if they had become normocapnic or had stopped home NIV for any reason. Appropriate compliance with NIV was assured using the hour meter values on the home ventilator. Optimal medical treatment according to the GOLD guidelines was ascertained in all patients.
After the six months run-in phase, patients were re-admitted to the respiratory care unit and randomly assigned with the use of a block randomisation scheme to either stop NIV (withdrawal group) or continue NIV (ventilation group). Randomisation was done with equal allocation to groups. Patients from whom NIV was withdrawn were informed that they had the option to voluntarily resume NIV if dyspnoea increased severely.
Primary outcome measure
The primary endpoint was time to clinical worsening defined as an escalation of mechanical ventilation. In both groups, escalation of mechanical ventilation was defined as intensive care unit (ICU) admission for treatment of acute-on-chronic respiratory failure. In the withdrawal group, escalation of mechanical ventilation additionally comprised (1) voluntary resumption of NIV due to severe dyspnoea or (2) re-institution of NIV due to progressive hypercapnia defined by an increase of PaCO2 ≥ 10% compared to the prior visit. These two endpoints were included for ethical concerns, as a successfully established therapy was to be withdrawn from stable patients. It was felt that they must have access to NIV in case their subjective shortness of breath became unbearable. On the other hand, an increase in PaCO2 at a control visit was used as an objective criterion to resume NIV to detect those patients who were about to develop acute-on-chronic respiratory failure in a timely manner. These safety considerations were confirmed by the Institutional ethics committee.
Follow-up and secondary outcome measures
Patients were followed for twelve months with control visits every three months. At each control visit, patients were admitted to the respiratory care unit for two nights. Irrespective of group allocation, all patients slept one night without NIV. Supine resting arterial blood gases were taken prior to sleep and on first waking on both occasions with the patient breathing oxygen via nasal cannulae at their prescribed flow rate. Polysomnography was repeated during the second night. To obtain exercise PaCO2, a cycle ergometer test was performed using a 2-min incremental cycle exercise protocol. Patients started with a 2-min period of unloaded pedalling at 60 cycles/min, followed by 20-Watt increments/2 min. Arterial blood gases were collected after 120 s of cycling at 20 Watts. Functional exercise capacity was additionally assessed using the 6-min walking distance.
using a standard pulmonary function test unit (SensorMedics Vmax 22, Viasys Healthcare). Blood gases were determined using the AVL Compact 3 Blood Gas Analyzer (Roche Diagnostics, Graz, Austria). Maximum inspiratory pressure was measured using an esophageal balloon catheter (Bicore Monitoring Systems, Irvine, CA, USA). Disease-specific quality of life was assessed by the Saint Georges Respiratory Questionnaire (SGRQ, 0 points = best result, 100 points = worst result).
Incidences of antibiotic and oral steroid therapy were recorded.
Sample size calculation
We estimated that a clinical relevant difference in median time to clinical worsening would be 100 days. From our clinical experience we estimated the time to clinical worsening to be 60 days in the withdrawal group and hence 160 days in the ventilation group. Based on this difference, an alpha error probability of 0.05 and a beta error probability of 0.9, the necessary total sample size was determined to be 26. Sample size was calculated using ‘PS Power and Sample Size Calculations Version 2.1.30’
and confirmed using STATA version 8. Interim analyses and stopping rules were not used in the trial.
Statistical methods
Statistical analyses were performed using SPSS version 15.0 software (SPSS Inc., Chicago, Illinois). The Mann–Whitney U test was used for comparison of continuous variables. Fisher’s exact test was used to compare categorical variables between the two groups. Time to clinical worsening was depicted using a Kaplan–Meier plot, and the log-rank test was used to compare the two groups. To compare serial PaCO2 measurements between the two groups, we used a generalised estimating equation assuming an exchangeable correlation matrix for repeated observations within one patient.
To compare the number of days using antibiotics or steroids between the two groups, we used a generalised estimating equation assuming an exchangeable correlation matrix for repeated observations within one patient.
Data are presented as mean ± SD, median (1st–3rd quartiles), or a proportion as appropriate. Differences between groups are reported as median or mean and 95% confidence interval. A two-sided p-value < 0.05 was considered statistically significant.
Results
All 998 patients treated at the respiratory care unit from April 1, 2003, to February 28, 2007, were screened for eligibility; 38 patients (4%) met the study criteria and were enrolled in the run-in phase. During the run-in phase, 12 patients were excluded for various reasons (Fig. 1). After the run-in phase, 26 patients still fulfilled the study criteria and were randomised. Baseline demographic and clinical characteristics of the withdrawal group and of the ventilation group are shown in Table 1. With respect to co-morbidities arterial hypertension tended to be more common in the withdrawal group (p = 0.11).
Figure 1Study flow chart of the NIVEX trial. NIV, non-invasive ventilation.
Enrolment was stopped after 26 patients had been randomised. All randomised patients remained in their allocated group and could be followed over the subsequent 12 months.
During the 12-month follow-up period, the withdrawal group had a significantly higher probability of clinical worsening compared to the ventilation group (p = 0.0018) (Fig. 2). Mean time to clinical worsening in the withdrawal group and in the ventilation group was 162 ± 40 days and 391 ± 36 days, respectively.
Figure 2Kaplan–Meier plot of the probability of clinical worsening in patients after withdrawal (withdrawal group) or continuation of NIV (ventilation group). Log-rank test statistic: p = 0.0018.
At the end of the 12-month follow-up period, ten patients (77%) in the withdrawal group, but only two patients (15%) in the ventilation group, experienced clinical worsening (p = 0.0048). In the withdrawal group, three patients were admitted to the ICU for acute-on-chronic respiratory failure; four patients needed to resume NIV due to severe dyspnoea, and in three patients NIV was recommenced due to progressive hypercapnia. In the ventilation group, two patients were admitted to the ICU for acute-on-chronic respiratory failure.
Three months after randomisation, 6-min walking distance increased by 43 ± 79% in the ventilation group but decreased by 11 ± 25% in the withdrawal group, p = 0.0425 (Fig. 3). Three months after randomisation, the SGRQ total score did not differ between the ventilation group [57 (44–67)] and the withdrawal group [53 (45–63); median (1st–3rd quartile), p = 0.80).
Figure 3Six-minute walking distance three months after randomisation to withdrawal (withdrawal group, n = 10) or continuation of NIV (ventilation group, n = 12). Mann–Whitney U test statistic: p = 0.0425.
Withdrawal of NIV had no significant effect on follow-up on daytime resting PaCO2 (Table 2). However, in the withdrawal group resting daytime pH was slightly lower (Table 2 and Fig. 4). At randomisation, PaCO2 during exercise was stable in both groups; the median (1st–3rd quartile) change in PaCO2 between rest and 20 Watts was −0.4 (−3.3 to 4.0) mmHg in the withdrawal group and 1.0 (−3.0 to 4.3) mmHg in the ventilation group (p = 0.60 for between group comparison). Three months after randomisation, the median (1st to 3rd quartile) change in PaCO2 showed a trend toward an increase in the withdrawal group [3.7 (−0.3 to 7.5)] but remained unchanged in the ventilation group [−0.9 (−3.6 to 3.0)]. This difference, however, did not reach statistical significance (p = 0.0825 for between group comparison).
Table 2Blood gas data and FEV1 during the course of the study. The p values refer to the between-groups comparisons.
Daytime resting PaO2, litres of oxygen administered, arterial bicarbonate and FEV1%predicted did not differ between the two groups during follow-up (Table 2). At all follow-up visits, sleep studies performed in all patients did not show evidence of newly acquired obstructive sleep apnoea.
The incidence density rate of antibiotic therapy due to exacerbation was 0 (0–1.02) and 0.23 (0–2.48) days per 100 patient days in the withdrawal group and in the ventilation group respectively, p = 0.35. The incidence density rate of oral steroid therapy due to exacerbation was 0 (0–0) and 0 (0–1.79) days per 100 patient days in the withdrawal group and in the ventilation group respectively, p = 0.59; data are medians and 1st to 3rd quartiles.
Discussion
We show that the withdrawal of NIV causes clinical worsening in COPD patients with chronic hypercapnic respiratory failure. Patients who had previously required mechanical ventilation for treatment of acute respiratory failure and who remain hypercapnic thereafter may benefit from long-term NIV.
The rationale and main novelty of our study were both the idea that the most severely sick subgroup of COPD patients is the most likely to benefit from NIV. Therefore, our study is the first to include only patients who have already required mechanical ventilation for an episode of acute respiratory failure. Our positive results confirm the recent findings of McEvoy et al,
who showed a small survival benefit of NIV in stable COPD patients after adjustment of baseline gas exchange and symptoms. Earlier studies on this topic were significantly underpowered to examine mortality.
Our finding of sustained clinical stability due to NIV is in line with the findings of Casanova et al. who observed a reduction in the number of hospital admissions after three months.
Additionally, Tuggey et al. showed that domiciliary NIV in COPD patients with recurrent admissions for acute-on-chronic respiratory failure requiring NIV reduced hospital admissions.
In contrast, studies by Clini et al. and McEvoy et al. showed no reduction in hospital admission in the NIV group as compared to the group on long-term oxygen therapy. The clearly positive results of our study can likely be explained by our selection of patients with a high probability of recurrent respiratory failure.
We chose the time to clinical worsening as the primary endpoint because we considered it to be one of the most relevant clinical outcome variables in this patient population. Moreover, the statistical analysis of the time to an outcome is more robust and independent of assumptions compared to the analysis of repeated outcomes such as exacerbation rate or hospitalisation rate.
In this trial, by definition the achievable endpoints depended on the group allocation of the patient: after withdrawal of NIV, ventilator support could be escalated by either resumption of NIV or by ICU admission for acute-on-chronic respiratory failure. In patients who already were on nocturnal NIV, however, only the latter endpoint was applicable.
The criterion of a 10% increase in resting morning PaCO2 for re-institution of NIV was arbitrarily chosen after consultation with various experts in the field. It allows for larger variations in severely hypercapnic patients but is very sensitive when hypercapnia is mild. McEvoy et al. found that PaCO2 remained almost completely unchanged during one year in stable hypercapnic COPD patients with our without NIV.
Therefore the criterion of a 10% increase of PaCO2 probably represents ventilatory deterioration. However, the amount of natural fluctuations of PaCO2 is unpredictable within an individual patient. Therefore it cannot be excluded that some of the increases of PaCO2 which were regarded as an endpoint were actually due to the natural variation of alveolar ventilation.
Although dyspnoea is a subjective measure and varies between patients, we chose to use it as an endpoint, when it became so severe that the patient deliberately resumed NIV. NIV is an intervention, which may also bring along subjective discomfort during sleep. We therefore reasoned, that voluntary resumption of a potentially disturbing therapy to alleviate dyspnoea actually represents clinical worsening and strongly suggests that the patient actually benefits from NIV.
Our results have some limitations. While basically both groups were well matched with regard to baseline characteristics arterial hypertension tended to be more common in the withdrawal group. Since blood pressure was checked and treated in all patients during all visits and since no episode of uncontrolled hypertension or heart failure occurred, this difference is unlikely to be relevant for the main finding.
The sample size calculation was based on an arbitrary number of days with clinical stability. Since this specific subtype of COPD patients after mechanical ventilation due to acute-on-chronic respiratory failure has not been studied in a comparable way so far, we were not able to find a reasonable estimate from the literature. We reasoned that 100 days spent at home without clinical worsening would be considered relevant by most patients and the health care providers.
The exact mechanisms of the physiologic effects of long-term mechanical ventilation in COPD are unclear.
It can be speculated that nocturnal NIV allows maintaining the balance between load and capacity of the ventilatory pump. The observed trend towards CO2 retention during exercise indicates that ventilatory failure may be responsible for exercise limitation after withdrawal of NIV.
The effect of the withdrawal of NIV on self perceived well-being using the SGRQ could only be assessed until three months after randomisation. Potential effects on long-term quality of life, however, could not be analysed owing to the small remaining sample size after many patients had experienced endpoints.
Our findings indicate that a very specific patient population benefits from NIV. COPD patients who require mechanical ventilation due to an exacerbation and remain hypercapnic thereafter should be evaluated for this therapy in specialised respiratory care centres. During this time patients are trained, informed and motivated, and ventilatory settings are optimised. After primary initiation of home NIV, patients should have regular follow-up visits at the hospital every three to six months. It should be clearly stated, that our findings only relate to the most severely ill COPD patients, who have already survived one episode of acute respiratory failure requiring mechanical ventilation. It is not suggested that all COPD patients with stable hypercapnia benefit from long-term NIV.
Future research should be aimed at early identification of those patients who are most likely to benefit from NIV. Candidates may be not only patients who have already required mechanical ventilation, but also patients who have been discharged hypercapnic after a severe exacerbation.
Aggravation of hypercapnia during exercise may be a tool for the identification of patients at risk for acute-on-chronic ventilatory failure.
In conclusion, discontinuation of NIV causes clinical worsening in stable COPD patients who remain hypercapnic after an episode of acute respiratory failure requiring mechanical ventilation. Therefore, these patients should be evaluated for NIV in a specialised centre as they may benefit from long-term NIV.
Acknowledgements
The authors would like to thank the physicians and nurses of the respiratory care unit for supporting this study; Nicole Dornhofer for excellent assistance in managing and cleaning the data; and Kora Geyer who performed the pulmonary function tests.
Conflicts of interests
The authors declare that they have no conflicts of interests in relation the subject of the paper.
Authors’ contributions
GCF performed the statistical analysis and wrote the manuscript. MKB participated in the design of the study and performed patient interviews. OCB helped to draft the manuscript. EK conceived of the study, participated in its design and coordination and helped to draft the manuscript. KK participated in the design of the study, created the questionnaires and performed patient interviews. RK helped to draft the manuscript. IS participated in the design of the study and performed 6-min walking tests. SH conceived of the study, participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Abbreviation list
COPD
Chronic obstructive pulmonary disease
NIV
Non-invasive ventilation
PaCO2
Partial pressure of carbon dioxide in arterial blood
ICU
Intensive care unit
SGRQ
Saint Georges Respiratory Questionnaire
FEV1%predicted
Predicted forced expiratory volume in 1 s
FVC
Forced vital capacity
PaO2/FiO2
Ratio of the partial pressure of oxygen in arterial blood to fraction of oxygen in inspired air
References
Ram F.S.
Picot J.
Lightowler J.
Wedzicha J.A.
Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease.
00391-4&id=gr1.jpg){kind=link}
00391-4&id=gr2.jpg){kind=link}
00391-4&id=gr3.jpg){kind=link}
00391-4&id=gr4.jpg){kind=link}