Switch of Noninvasive Ventilation (NIV) to Continuous Positive Airway Pressure (CPAP) in Patients With Obesity Hypoventilation Syndrome

A Pilot Study

Sarah Orfanos; Dany Jaffuel; Christophe Perrin; Nicolas Molinari; Pascal Chanez; Alain Palot


BMC Pulm Med. 2017;17(50) 

In This Article


This prospective open trial, testing the feasibility of a protocol switching NIV to CPAP in stable patients in real life conditions, suggests the possibility of this switch in patients suffering from OHS with an AHI ≥ 15/h and without pulmonary obstructive disease. In these conditions, we reported no significant treatment effect differences between before and after the switch to CPAP in diurnal and nocturnal alveolar gas exchange, mean AHI, mean compliance, quality of sleep and quality of life. In addition, 80% of patients preferred CPAP over NIV.

The pathophysiology of OHS is complex and multifactorial. Eventually leading to chronic, diurnal hypercapnia resulting from an imbalance between apnea and inter-apnea periods and deficiency of the compensatory mechanisms to unload the carbon dioxide in excess.[17,18] This suggests the need to determine different phenotypes of OHS, and therefore adapt therapeutic intervention according to the phenotype. Therefore we have chosen in this trial, to exclude from the study, patients with an AHI < 15/h, speculating that CPAP mechanism to treat hypoventilation was mostly due to the correction of apneas and hypopneas. Indeed, if CPAP is not intrinsically the most efficient treatment of hypoventilation, it is still effective on most of the pathophysiological mechanisms leading to OHS by decreasing upper airways resistance, increasing central response to hypercapnia and hypoxemia,[19,20] increasing lung volumes, treating atelectasis, decreasing intrinsic PEEP.[21] This is the reason why CPAP is efficient in most patients with OHS associated with OSA.

Two prospective randomized trials have compared the efficacy of CPAP and NIV in patients with OHS.[5,11] These two prospective studies corroborate the fact that patients with OHS and concomitant OSA can be treated with CPAP. Piper and colleagues did not find a significant difference, at three months, in daytime paCO2 between the two groups NIV and CPAP.[11] However in this study, nine of the 45 patients initially selected were excluded from the trial, due to persisting nocturnal hypoxemia or nocturnal hypercapnia on the first night of CPAP titration. Treating hypoventilation in some phenotypes may take up to a few weeks,[22] excluding patients after one night of CPAP titration is debatable. Masa and colleagues confirmed the absence of a significant difference between NIV and CPAP, at two months, in diurnal paCO2.[5] However, only NIV and not CPAP significantly improved paCO2 compared to the control group. An improvement was only noted for patients using CPAP more than 4 h per night. Moreover, the authors found a significant difference, favoring NIV, on the 6 min' walk test and respiratory function tests.

The heterogeneity in responses to CPAP, led some studies to try to define phenotypes of OHS and predictor factors of CPAP failure. Therefore, patients exhibiting a higher BMI, a higher percentage of recording time below 90% saturation, a lower AHI, or a decreased forced vital capacity could be at higher risk of CPAP failure, even though these predictor factors are not constantly found in all studies.[8–10] In our study, we found that only BMI had a significant association with CPAP or NIV failure (Table 6). When using the two criteria defining failure in Salord and colleagues study[10] (more than 30% of recording time spent below 90% saturation or paCO2 > 45 mm Hg), in our study 8/15 patients would be categorized as NIV failure and 7/15 as CPAP failure. Interestingly, the seven patients failing CPAP were the same patients who were failing NIV before the switch. We do not think this failure can be attributed to NIV or CPAP settings in our study. Mean EPAP was 8 cm H2O, mean IPAP was 18 cm H2O, back-up RR was 14 cycles/min, all NIV were in ST mode. Concerning CPAP, mean EPAP was 9.8 cm H2O. These pressures are concordant with previous trials.[5,11]

Our study has some limitations. First of all it can be argued that the sample size in our study (15 patients) could be insufficient to detect a significant difference after the switch to CPAP, and that this absence of difference could be due to a lack of power. This study does not have the pretention to assert the absence of difference between NIV and CPAP at a larger scale, but describes a feasible protocol for most sleep clinic, to switch stable patients with OHS from NIV to CPAP. As suggested by our preliminary results, for some patients, the switch of NIV to CPAP can be deleterious emphasizing the need for a careful monitoring. A second limitation is that the results of our study can only be extended to patients with OHS and concomitant OSA with AHI ≥ 15/h, without obstructive pulmonary diseases, and cannot be used when treating patients in an acute state or with mild OSA. These drastic selection criteria aimed at targeting CPAP responders' phenotypes and the limited number of patients included in the study could partly explain the stable diurnal capnia after the switch to CPAP in all of our patients. The treatment of patients with OHS with pure hypoventilation and without OSA, is seldom discussed in scientific literature, and these patients still benefit from NIV in most teams.[23] For ethical reasons, we decided not to switch to CPAP, patients with mild OSA (AHI < 15/h), to not deprive these patients with previous acute exacerbation from an effective therapy.

To eliminate NIV persistent effect, we introduced in the protocol, a withdrawal from NIV for seven nights before starting CPAP, and reevaluated patients more than a month after starting CPAP. It is possible that not all persistent effects attributable to NIV were eliminated by this withdrawal.[24] An ideal protocol would have been to randomize patients into three groups: one group pursuing NIV, one group switching to CPAP and one group stopping ventilation; but this protocol did not seem ethical for patients withdrawn from ventilation, considering that a higher mortality rate has been reported in untreated patients with OHS.[4,6,7] In our study, 12 out of 15 patients were diagnosed with OHS and started on NIV, following an episode of exacerbation. This raises the question as to whether the diagnosis of OHS was made in excess in patients with normocapnic OSA undergoing an exacerbation. This assertion is however contradicted by the fact that nine out of the 12 patients diagnosed with OHS following an exacerbation, had high bicarbonates level (HCO3- > 28 mmol/l), in favor of chronic hypercapnia.[25] When measuring the AHI, we used the built-in software available in the NIV and the CPAP. Although the accuracy of these systems is debatable, recent studies are in favor of a good reliability of these built-in softwares.[26]