Obesity, dubbed the “global epidemic of the 21st century”, is a growing phenomenon in the developed world, and now affects one third of the adult population.1 One obesity-related disorder is obesity hypoventilation syndrome (OHS), characterized by chronic hypercapnic respiratory failure not secondary to other causes, episodes of sleep apnea and hypoventilation.2 Less is known about the repercussions of OHS than those of sleep apnea–hypopnea syndrome (SAHS), although pulmonary hypertension and metabolic syndrome seem to be more common. The greater degree of sleep hypoxia in OHS patients may generate a greater risk for other cardiovascular events and mortality compared with obese individuals with no respiratory failure.3 The best treatment is weight loss,4 as a return to normal weight reverses respiratory failure, pulmonary hypertension and sleep disturbances.5 However, these patients find it difficult to achieve and maintain significant weight loss. Bariatric surgery is a marginal option, but it is associated with greater mortality and is usually rejected by patients. Although continuous positive airway pressure (CPAP) corrects sleep apnea in OHS patients, it does not seem to normalize daytime PaCO2 in all cases.6 The role of apnea and hypopnea in the development of daytime hypercapnea, therefore, is not entirely clear, although it has been suggested that patients who respond to CPAP may present predominantly apneic episodes, rather than alveolar hypoventilation.6 Daytime hypercapnia in SAHS has been associated with the event:interevent duration ratio, which would mean that sleep and daytime hypercapnia result from the individual's failure to recover from the hypoventilation caused by episodes of apnea, particularly during the rapid eye movement (REM) stage of sleep.7 CPAP and non-invasive ventilation (NIV) are used extensively in the treatment of chronic OHS. The effectiveness of CPAP vs NIV has been studied in only 1 randomized trial,8 in which 36 patients were selected on the basis of a favorable response to a trial CPAP treatment session. The authors concluded that clinical and blood gas improvement at 3 months was similar for both therapies. Another randomized trial compared the effectiveness of NIV with “healthy lifestyle” counseling (i.e., sleep hygiene and diet) in patients with OHS selected on the basis of mild daytime hypercapnia.9 No randomized trials have been performed in non-selected patients to determine which of these treatments is most effective (NIV vs CPAP), or which is most effective compared to a control intervention (healthy lifestyle).

Likewise, no studies with a similar design have been undertaken to show whether the complications associated with OHS (arterial and pulmonary hypertension, cardiovascular events, hospitalization and mortality) decrease more or less with any of these treatments. A new study is needed to determine whether NIV is superior to CPAP or lifestyle changes in the medium- and long-term, and to validate the current widespread use of NIV in OHS patients.

We undertook a prospective, open-label, parallel group, randomized controlled trial in OHS patients to evaluate the medium- and long-term efficacy of NIV, CPAP and lifestyle changes (control). The primary outcome was PaCO2 and days of hospital stay, respectively. Operating variables were dropout for medical reasons (%) and mortality.

The Steering Committee of the Separ Comprehensive Research Study on Sleep (CRS on Sleep) appointed Dr. Juan Fernando Masa to head Pickwick, a project that would bring together clinical and basic research to find more comprehensive answers to the incognitos surrounding this increasingly widespread disease (see Addendum 2).

Patient selection in participating centers began in 2009 (a list of centers is given in Addendum 3).

The following precepts were established in advance: (a) monthly newsletters would be sent by the principal investigator to all co-investigators comparing the progress made by different trial centers, encouraging recruitment efforts, and reporting on the latest developments; (b) investigator meetings would be held to coincide with SEPAR conferences and the winter meeting; and (c) the authorship policy was defined, with an estimation of the number of articles to be published and the names of the authors, based on the number of patients contributed to the trial (Addendum 4).

The short-, medium-, and long-term primary and secondary objectives are summarized in Table 1.

• (a)

  Informed consent: patients were given written information on the nature and purpose of the trial. The project was assessed and approved by the ethics committees of participating hospitals, in accordance with the declaration of Helsinki.

• (b)

  During the trial, an authorized external monitoring agency has been given access to the periodic analyses of the primary outcome variable (days of hospitalization), and the number of dropouts due to medical reasons and to death. Their task is to determine whether the trial should be halted if any particular treatment was found to be significantly worse, irrespective of the trial cohort (SAHS or non-SAHS).

The trial has now reached the mid-way point. Recruitment for the SAHS group has finalized, and the intragroup (NIV, CPAP and control) efficacy analysis is currently underway in this cohort.

A total of 302 patients have been included in the trial: 221 in the SAHS group, and 81 in the non-SAHS group. In total, 49 patients have been excluded.12

To date, 70 (23%) patients have dropped out from the trial: 53 (23%) from the SAHS group, and 17 (21%) from the non-SAHS group. Three (0.01%) patients dropped out for medical reasons: 1 from the control group (prior to the second randomization) and 2 from the SAHS with NIV group. Another 7 (0.02%) patients have died: 2 from the SAHS with NIV group, 4 from the SAHS with CPAP group, and 1 from the non-SAHS control group. So far, no difference between groups in respect of dropouts for medical reasons or death has been observed.

This trial will conclusively show whether treatment with NIV is superior to CPAP or a sleep hygiene and diet regimen. These findings may validate the widespread, though scientifically unfounded, use of NIV in this disease. Efficacy will be evidenced through physiological and clinical variables, as well as other more specific variables such as days of hospital stay, cardiovascular events, pulmonary hypertension, or even the dropout and mortality rate. Considering that the prevalence of OHS is thought to be rising in parallel with an increase in obesity, it is of the utmost importance that the most effective treatment for this condition is defined.

The results will be released in a series of articles published at different stages of the trial:

• 1.

  Stage 1, after 2 months of follow-up: (a) comparative efficacy of the 3 study treatments in terms of symptoms, respiratory function, and echocardiogram and polysomnography findings; (b) importance of apneic episodes and leptin in the pathogenesis of daytime hypoventilation; (c) association of metabolic, biochemical and vascular endothelial dysfunction disorders with apnea–hypopnea (intermittent hypoxia hypothesis) or daytime hypoventilation (sustained hypoxia hypothesis).

• 2.

  Stage 2, after 3 years of follow-up: (a) comparative efficacy based on evolution of symptoms, respiratory function and complications (hospitalizations, cardiovascular events, arterial and pulmonary hypertension, dropouts and death); (b) comparative evaluation of the evolution of biological parameters associated with each treatment.

As mentioned above, daytime hypercapnia in SAHS has been associated with the event:interevent duration ratio, which would mean that sleep and daytime hypercapnia depend on the individual's failure to recover from the hypoventilation caused by episodes of apnea, particularly during the rapid eye movement (REM) stage of sleep.13 This trial will also show whether sleep apnea plays a pivotal role in the development of daytime hypercapnia.

A typical finding in patients with OHS is an increased alveolar-arterial oxygen gradient resulting from changes in respiratory mechanics caused by the build-up of fat in the chest wall and abdomen. To date, no randomized trials have evaluated the benefits of long-term oxygen therapy alone or in combination with a sleep hygiene and diet regimen in this patient population. Moreover, oxygen therapy has not yet been compared with CPAP or NIV in the context of OHS. In this trial, oxygen is prescribed at the discretion of the patient's doctor, a factor that enabled us to analyze subgroups that did or did not receive supplementary oxygen in each treatment group.

Leptin is a protein secreted by white adipose tissue. Its function is to inhibit hunger and increase energy expenditure. Obesity is thought to be associated with central resistance to leptin, leading to hyperleptinemia. Studies in animal models have shown that leptin is a powerful respiratory stimulant, and lack of this protein (or of its effect) could cause hypoventilation. Serum leptin levels normalize in SAHS patients treated with CPAP,14 although it is thought that apnea–hypopnea is the cause, rather than the result, of elevated leptin levels. Studies suggest that serum leptin levels are twice as high in OHS patients as in patients with a similar level of obesity and frequency of apnea–hypopnea episodes, but with no daytime hypercapnia.15 Leptin levels can be reduced with NIV therapy16 or weight loss. The foregoing has led researchers to hypothesize that leptin resistance could be the cause of OHS. This theory, however, has been called into question by the contradictory findings of a more recent study,17 which reported that leptin levels are lower in OHS patients with no major apnea–hypopnea than in obese patients with a similar number of obstructive episodes and no daytime hypercapnia, and that leptin levels increase with NIV. Our study will give further insight into this conundrum: we will show whether serum leptin levels decrease or increase to within normal ranges (for the level of obesity) in OHS patients in whom daytime hypercapnia has been reversed, particularly in the group with few or no episodes of apnea–hypopnea. Our findings will also show whether leptin levels in OHS patients with obstructive episodes and no reversal of hypercapnia following CPAP treatment are higher or lower than in those with reversal of hypercapnia following CPAP or NIV. In conclusion, we will show whether hyperleptinemia is the cause of hypoventilation or the result of hypoventilation or sleep apnea.

SAHS has been associated with metabolic changes and vascular endothelial dysfunction, which could be caused by either intermittent hypoxia or the increased sympathetic activity typical of these patients. Ultimately, these conditions can contribute to the development of arterial and pulmonary hypertension and cardiovascular events. Few studies have explored this phenomenon in OHS patients, but some inflammatory mediators, such as IL-6, could be more elevated in OHS patients than in SAHS patients with no daytime hypercapnia.18 Neither has the cause of this been investigated: whether OHS patients with no major apnea–hypopnea have elevated basal levels of these markers that could decrease with treatment, or whether these markers either increase or decrease due to the presence of apnea–hypopnea. Our study could provide answers to most of these questions, and will show whether inflammation at the molecular level is driven more by intermittent hypoxia (obstructive episodes) or sustained hypoxia (hypoventilation), and how this evolves with different treatments.

The dropout rate (23%) so far is below the estimated rate of 30%, but the 3-year follow-up is still ongoing. Dropout always leads to a loss of clinical trial data, as investigators collect information on days of hospital stay, type of treatment (CPAP, NIV, oxygen or lifestyle changes), and mortality after the dropout has occurred. If the dropout rate is ultimately higher in any particular treatment group (for example, higher in CPAP than in NIV), the per protocol analysis will determine whether this is related with treatment efficacy or other circumstances (for example, patient discomfort).

The difficulties involved in conducting this study include:

• (a)

  The need for a large sample size, calling for patient recruitment in several different centers.

• (b)

  Three years of follow-up, with 2 pre-inclusion studies.

• (c)

  Lack of information in OHS literature on the incidence of cardiovascular events per year or how these might be affected by treatment.

• (d)

  The parallel cohort study design.

• (e)

  Lack of information in the literature on the prevalence of severe SAHS (AHI>30) in OHS patients (see Addendum 8).

Instituto de Salud Carlos III (Fondo de Investigaciones Sanitarias, Ministry of Health and Consumer Affairs)PI050402, Sociedad Española de Neumología y Cirugía Torácica 2005 (SEPAR) and Air Liquide España.

The authors declare they have no conflicts of interest.