Obesity hypoventilation syndrome (OHS) is defined by obesity, hypercapnia during wakefulness, and sleep-disordered breathing, without other causes of hypoventilation.1 Roughly 90% of patients with OHS have concomitant obstructive sleep apnea (OSA),2 with 73% having severe OSA.3 Compared with eucapnic OSA4,5 and eucapnic obese,6,7 patients with OHS have a higher prevalence of cardiovascular and respiratory morbidity resulting in increased hospitalization risk, greater healthcare resource utilization, and augmented mortality.7–16

Ambulatory patients with OHS are typically treated with nocturnal positive airway pressure (PAP) therapy during sleep such as noninvasive ventilation (NIV), usually with bilevel pressure settings, or continuous positive airway pressure (CPAP). The effectiveness of NIV has been evaluated in several long-term, observational studies;4,6,7,16–23 and medium-term randomized trials.3,24,25 CPAP prevents upper airway obstructive events, but it is not the treatment of choice for non-obstructive sleep hypoventilation.21 Several medium-term clinical trials3,26,27 and one long-term28 randomized clinical trial have reported similar outcome between NIV and CPAP. CPAP is simpler to implement and is less costly than NIV,28 therefore, CPAP has been proposed as first-line treatment in OHS and concomitant severe OSA.28,29 In the largest randomized clinical trial to date (Pickwick study), the longitudinal improvement in arterial blood gases (ABG) were similar between NIV and CPAP.28

There is no consensus on how to define OHS severity. Classification of OHS severity based on PaCO2, OSA severity, or associated comorbidities have been proposed.30–32 However, there are no randomized clinical trials or prospective longitudinal studies assessing the effectiveness of PAP treatment (NIV and CPAP) based on any of the previously proposed OHS severity classifications. Higher PaCO2 has been related to mortality17,33,34 and reflects the consequences of potential mechanisms underlying chronic hypercapnic respiratory failure in OHS.35–37 We hypothesized that in ambulatory patients with OHS and concomitant severe OSA, long-term NIV therapy will lead to a greater improvement in awake PaCO2 and PaO2 than CPAP in the subgroup with more severe daytime hypercapnia at baseline. To test this hypothesis, we performed a post hoc analysis of the Pickwick study.3,25,28,38–44 We examined the longitudinal changes in arterial blood gases (PaCO2 and PaO2) in patients with OHS and concomitant severe OSA undergoing long-term therapy with NIV or CPAP. We classified patients into two OHS severity subgroups using baseline daytime PaCO2: PaCO2 of 45–49.9 or ≥50mmHg.

This post hoc analysis of the largest multi-center clinical trial with the longest period of follow-up in adult patients with OHS is the first to examine long-term outcomes based on the severity of OHS. The study included ambulatory patients with OHS and concomitant severe OSA who were in stable chronic hypercapnic respiratory failure. Our analysis shows that there were no significant long-term differences in the improvement of daytime PaCO2 and PaO2 between NIV and CPAP based on whether the baseline PaCO2 was between 45–49.9 or ≥50mmHg.

Conceptually, CPAP is not a treatment for nocturnal hypoventilation that is not a result of upper airway obstructive events.21,24,48 Consequently, the Pickwick study was designed to compare CPAP with NIV in OHS patients with severe OSA (AHI≥30) because CPAP can be less effective than NIV in patients with an AHI<30. This approach has been endorsed by the American Thoracic Society clinical practice guidelines.29

In addition to the severity of OSA, several other factors may influence clinicians to select NIV over CPAP in stable ambulatory patients with OHS and severe OSA. Some factors that may favor the prescription of NIV include higher baseline PaCO2 and persistent hypoxemia or hypercapnia during CPAP titration (i.e. CPAP titration failure). In this context, NIV would seem more appropriate for patients with more severe hypercapnia. Although several clinical trials have shown no significant difference in outcomes between CPAP and NIV,28,49 it remains unclear whether there is a subset of patients who would benefit more from NIV than CPAP. This is clinically relevant because an in-laboratory CPAP titration failure does not necessarily mean long-term CPAP treatment failure, suggesting that even an in-laboratory PAP titration study may not ultimately be able to distinguish patients who will finally fail long-term CPAP therapy.50 Higher PaCO2 has been associated with higher mortality.17,33–37 In OHS patients with severe OSA, awake baseline PaCO2 seems to be a reasonable marker of OHS severity (Table 1). We chose longitudinal ABG parameters as the outcome because in clinical practice, ABG is the most frequent and available objective test for medium and long-term assessment.

An observational retrospective study compared the effectiveness of PAP according to severity groups using baseline daytime PaCO2 in OHS.31 CPAP responders (who achieved oxygen saturation during sleep ≥90%, together with resolution of obstructive events) during polysomnographic CPAP titration were treated with CPAP and the remaining with NIV. CPAP responders were more frequently in the lowest severity group of baseline PaCO2 compared with the highest severity group, suggesting that clinicians may use baseline PaCO2 or the degree of persistent hypoxemia during CPAP titration as reasons to prescribe NIV. However, CPAP titration failure does not necessarily mean long-term CPAP treatment failure50 and results from several randomized clinical trials of unselected patients do not support this approach.3,27,28 Although some studies have observed a medium-term delay in CPAP effectiveness in comparison with NIV,3,27,28,38,43 long-term outcomes do not differ between CPAP and NIV in the subset of patients with OHS who also have co-existent severe OSA. The present study provides evidence that in patients with OHS and concomitant severe OSA (nearly 70% of all patients with OHS), a higher level of baseline hypercapnia is not a justification to recommend NIV over CPAP.

Short-term studies have demonstrated a significant association between CPAP or NIV adherence and improvement in daytime hypercapnia.3,51 Higher level of adherence to either NIV or CPAP (>4h per day) has been associated with long-term outcomes as lower hospital resource utilization and mortality in the Pickwick cohort.28 For this reason we included adherence in our longitudinal model despite the fact that adherence was similar between CPAP and NIV in the two severity subgroups (Table E2).

In the long-term Pickwick trial,28 body weight was reduced significantly during follow-up without differences between CPAP and NIV. We have included in the present analysis the longitudinal change in body weight in the two severity subgroups given that weight loss (or weight gain) may impact our outcome of PaCO2 and PaO2. Since there was no significant difference in the degree of weight change between CPAP and NIV, the potential influence of weight reduction on ABG parameters should be similar for CPAP and NIV.

Our study has several limitations. First and foremost, our results represent a post hoc analysis of the original randomized controlled trial and as such, the current analysis was a non-specified subgroup analysis. Importantly, by creating 2 subgroups, the statistical power is reduced. We chose a cut-off point of 50mmHg for the baseline PaCO2 to create two subgroups of OHS severity because it is a value frequently used in clinical practice and because in another study, this cut-off point discriminated between groups with higher and lower oxygenation level during sleep.31 However, it is possible that other cutoff points may lead to different conclusions. As such, our results should be interpreted with caution. Our cohort was limited to ambulatory OHS patients who also have concomitant severe OSA. Since all patients in this study had severe OSA, defined as an AHI≥30, we could not reliably classify patients into subgroups of OSA severity because of an AHI floor effect. As such, it is important to note that our findings do not apply to the minority of patients with OHS who do not have concomitant severe OSA (∼30% of patients with OHS). Our external validity may be limited because the study was performed exclusively in Spain and patients from other ethnic/racial backgrounds may have different characteristics, which may impact the results. Finally, we did not use transcutaneous CO2 as a direct measure of hypoventilation during polysomnographic PAP titration, although we increased the inspiratory positive airway pressure (IPAP) during NIV titration to achieve the best oxygenation level as a surrogate marker of hypoventilation.3

In conclusion, in patients with stable OHS and concomitant severe OSA, NIV and CPAP have similar long-term effectiveness in improving PaCO2 and PaO2 irrespective of OHS severity based on baseline hypercapnia. Therefore, clinicians can initiate CPAP treatment regardless of PaCO2 severity in this patient population. A randomized clinical trial focused on using various definitions of OHS severity is needed to determine in which patients NIV would lead to better long-term outcomes.

The sponsors and funders of the study had no involvement or any influence in study design, in the collection, analysis, and interpretation of data, in writing of the manuscript, and in the decision to submit the manuscript for publication. The corresponding author (JFM) confirms that he had full access to all the data and he had the final responsibility for the decision to submit for publication.

Additional related documents such as study protocol, statistical analysis plan, and informed consent form will be available upon request from the Pickwick Project principal investigator (Dr. Juan Fernando Masa). Deidentified patients’ data can be requested by researchers for use in independent scientific research and will be provided following review and approval of the research proposal (including statistical analysis plan) and completion of a data sharing agreement with the Pickwick Project Publications Committee. Investigator Data requests can be made anytime from 1 to 2 years after the publication of this trial. Requests should be sent to the corresponding author (Dr. Juan Fernando Masa – fmasa@separ.es).

1. Substantial contributions to study conception and design, acquisition of data, or analysis and interpretation of data: Juan F. Masa, MD, PhD, M-Ángeles Sánchez-Quiroga, MD, Francisco Javier Gomez de Terreros, MD, PhD, Iván Benítez, BSc(Stat), Francisco-José Vázquez-Polo PhD; Miguel A. Negrín PhD; María Martel-Escobar PhD; Ferran Barbe, MD, PhD, Auxiliadora Romero, MD, Candela Caballero-Eraso MD, PhD, Teresa Gomez-Garcia, MD, Mónica González, MD, PhD, Soledad López-Martín, MD; José M Marin, MD, PhD; Odile Romero, MD; Trinidad Díaz-Cambriles, MD, Eusebi Chiner, MD, PhD, Carlos Egea, MD, PhD, Babak Mokhlesi, MD, MSc, Jaime Corral, MD, Galo Fernandez, MD, Javier Barca, MD, Estrella Ordax, MD, Nicolás González-Mangado, MD, PhD, Maria F Troncoso, MD, Maria-Ángeles Martinez-Martinez, MD, Elena Ojeda-Castillejo, MD; Daniel López Padilla MD, Santiago J. Carrizo, MD, PhD, Begoña Gallego, MD, PhD, Mercedes Pallero, MD, Odile Romero MD,Sergi Martí, MD, PhD, Maria Antonia Ramón, PT, MSc, Eva Arias, MD, Jesús Muñoz-Méndez, MD, PhD, Cristina Senent, MD; Jose N. Sancho-Chust, MD, Nieves Belén Navarro Soriano, MD, Emilia Barrot MD, PhD, José M. Benítez, MD, Jesús Sanchez-Gómez, MD, Rafael Golpe, MD, PhD, Rocio Gallego, MD, Juan A Riesco, MD, Ana Santiago-Recuerda, MD, PhD; Silvia Gomez, MD, and Mónica Bengoa, MD.

2. Drafting the article or revising the article critically for important intellectual content: Juan F. Masa, MD, PhD, Babak Mokhlesi, MD, MSc, Iván Benítez, BSc(Stat), Francisco Javier Gomez de Terreros, MD, PhD, Maria Ángeles Sánchez-Quiroga, MD, Auxiliadora Romero, MD, Candela Caballero-Eraso MD, PhD, Maria F. Troncoso, MD, Mónica González, MD, PhD, Soledad López-Martín, MD, José M. Marin, MD, PhD, Sergi Martí, MD, PhD, Trinidad Díaz-Cambriles, MD, Eusebi Chiner, MD, PhD, Carlos Egea, MD, PhD, Francisco-José Vázquez-Polo PhD; Miguel A. Negrín PhD; María Martel-Escobar PhD; Ferran Barbe, MD, PhD and Jaime Corral, MD.

3. Performing of the version to be published: Juan F. Masa, MD, PhD, Babak Mokhlesi, MD, MSc, Ivan Benitez, BSc(Stat) Francisco-José Vázquez-Polo PhD; Miguel A. Negrín PhD; and Jaime Corral, MD.

Instituto de Salud Carlos III (Fondo de Investigaciones Sanitarias, Ministerio de Sanidad y Consumo, Madrid, Spain) PI050402, Spanish Respiratory Foundation 2005 (FEPAR) and Air Liquide Spain.

Clinicaltrial.gov identifier: NCT01405976.

No conflicts exist for the authors.