OSA Endotypes and Phenotypes: Toward Personalized OSA Care

Obstructive sleep apnea (OSA) contributes a major health burden to society due to its high prevalence and substantial neurocognitive and cardiovascular consequences. Estimates suggest that at least 10% of adults in North America are afflicted with OSA, making it probably the most common respiratory disease in the developed world (Peppard et al. Am J Epidemiol. 2013;177[9]:1006). Nasal CPAP is a highly efficacious therapy that has been shown to improve neurocognitive and cardiovascular outcomes. However, CPAP is not always well tolerated. Alternative therapies, such as oral appliances and upper airway surgery, have highly variable efficacy, and evidence of important clinical benefits are uncertain. Therefore, efforts are ongoing to determine optimal alternative strategies for therapy.

In order to treat any condition optimally, one needs to be able to predict who is at highest risk of developing the condition, then to assess the consequences if left untreated, and finally to be able to predict response to various treatment options. Currently, the OSA field is still in its early stages of our understanding. Clinically, we are often faced with patients who have varying presentations and manifestations, but, for reasons that are unclear. For instance, two individuals with the same body mass index may have very different clinical manifestations, one with severe OSA and one without any OSA. Similarly, two individuals with an apnea hypopnea index of 40 events per hour (ie, severe OSA) may have very different symptoms attributable to OSA, eg, one could be asymptomatic and the other could be debilitated from sleepiness. We and others have been making efforts to determine why these phenomenon occur. At present, the techniques to define mechanisms underlying OSA are labor-intensive, requiring one or two overnight experiments to gather meaningful data. Although we are gathering new insights based on these techniques, efforts are ongoing to simplify these approaches and to make assessment of pathophysiologic characteristics more accessible to the clinician (Orr et al. Am J Respir Crit Care Med. 2017 Nov 30. doi: 10.1164/rccm.201707-1357LE. [Epub ahead of print]).

We ultimately believe that a thorough analysis of a sleep recording combined with demographic data and other readily available clinical data (perhaps plasma biomarkers) may yield sufficient information for us to know why OSA is occurring and what interventions might be helpful for an individual patient. Currently, our use of the polysomnogram to derive only an apnea hypopnea index does not take full advantage of the available data. An apnea hypopnea index can be readily obtained from home sleep testing and does not truly provide much insight into why a given individual has OSA, what symptoms are attributable to OSA, and what interventions might be considered for the afflicted individual. By analogy, if the only useful data derived from an ECG were a heart rate, the test would rapidly become obsolete. Along these lines, if the only role for the sleep clinician was to prescribe CPAP to everyone with an AHI greater than 5/h, there would be little need or interest in specialized training. In contrast, we suggest that rich insights regarding pathophysiology and mechanisms should be gathered and may influence clinical management of patients afflicted with OSA. Thus, we encourage more thorough analyses of available data to maximize information gleaned and, ultimately, to optimize clinical outcomes.

Recent studies suggest that sleep apnea occurs for varying reasons, a concept that is now thought to be clinically important (Jordan et al. Lancet. 2014;383[9918]:736). We draw a crucial distinction between endotypes (mechanisms underlying disease) and phenotypes (clinical expression of disease). Important endotypes include compromised upper airway anatomy, dysfunction in pharyngeal dilator muscles, unstable ventilatory control (high loop gain), and low arousal threshold (wake up easily), among others. Important phenotypes of sleep apnea are emerging and still evolving to include minimally symptomatic OSA, OSA with daytime sleepiness, and OSA with major cardiometabolic risk, among others. Several important concepts have emerged regarding different OSA endotypes and phenotypes:

1 The mechanism underlying OSA may predict potential response to therapeutic interventions. For instance, the endotype of OSA with unstable ventilatory control (high loop gain) may respond to agents such as oxygen and acetazolamide, which serve to stabilize control of breathing. In patients with anatomical compromise at the level of the velopharynx, uvulopalatopharyngoplasty may be an effective intervention. For patients with multiple pathophysiologic abnormalities, combination therapy may be required to alleviate OSA (Edwards et al. Sleep. 2016;9[11]:1973).

2 Given that OSA has many underlying etiologies, efforts are underway to determine whether individuals with different risk factors for OSA develop their disease based on varying mechanisms. As an example, people with posttraumatic stress disorder (PTSD) may be at increased risk of OSA perhaps on the basis of a low threshold for arousal (Orr et al. JCSM. 2017, 13[1]: 57-63). Another example would be patients with neuromuscular disease who may be at risk of OSA primarily based on impaired pharyngeal dilator muscle function.

3 A new concept is emerging whereby endotypes of OSA may actually predict differing OSA phenotypes. In theory, loop gain-driven OSA may have different consequences from OSA driven by compromise of pharyngeal anatomy. To this point, data suggest that OSA in the elderly may not have as many consequences as OSA in younger people matched on severity of illness. OSA in the elderly has lower loop gain than OSA in younger people and is associated with less negative intrathoracic pressure at the time of arousal as compared with younger individuals with OSA (Kobayashi et al. Chest. 2010; 137[6]:1310). As such, the endotype of OSA in the elderly may explain why the clinical consequences are fewer than in the younger OSA counterparts.

4 The mechanism underlying OSA may be important in determining response to clinical interventions, such as nasal CPAP. Patients with a low arousal threshold may be prone to insomnia when placed on CPAP and could theoretically be poorly tolerant of therapy based on disrupted sleep architecture. Such patients may benefit from non-myorelaxant hypnotic therapy to consolidate sleep and improve CPAP adherence. In addition, patients with high loop gain (unstable ventilatory control) may be prone to develop central apneas when placed on CPAP therapy (Stanchina et al. Ann Am Thorac Soc. 2015;12[9]:1351). These patients may benefit from newer technologies, eg, auto or adaptive servo ventilation - ASV. High loop gain has also been shown to predict failure of upper airway surgery as a treatment for OSA by several groups (Li et al. JCSM. 2017;13[9]:1029). Such patients should, perhaps, undergo nonsurgical therapies for OSA.

We emphasize that some of the points being made are somewhat speculative and, thus, encourage further basic and clinical research to test our assumptions. Robust, multicenter clinical trials assessing hard outcomes will ultimately be required to change the current standard of care. Nonetheless, we believe that a more thorough understanding of OSA pathogenesis can help guide clinical care today and will be critical to the optimal treatment of afflicted individuals tomorrow.

Dr. Owens is Assistant Clinical Professor of Medicine; Dr. Deacon is a Post-Doctoral Research Scholar; and Dr. Malhotra is Kenneth M. Moser Professor of Medicine and Chief, Division of Pulmonary, Critical Care and Sleep Medicine, University of California San Diego.

Obstructive sleep apnea (OSA) contributes a major health burden to society due to its high prevalence and substantial neurocognitive and cardiovascular consequences. Estimates suggest that at least 10% of adults in North America are afflicted with OSA, making it probably the most common respiratory disease in the developed world (Peppard et al. Am J Epidemiol. 2013;177[9]:1006). Nasal CPAP is a highly efficacious therapy that has been shown to improve neurocognitive and cardiovascular outcomes. However, CPAP is not always well tolerated. Alternative therapies, such as oral appliances and upper airway surgery, have highly variable efficacy, and evidence of important clinical benefits are uncertain. Therefore, efforts are ongoing to determine optimal alternative strategies for therapy.

In order to treat any condition optimally, one needs to be able to predict who is at highest risk of developing the condition, then to assess the consequences if left untreated, and finally to be able to predict response to various treatment options. Currently, the OSA field is still in its early stages of our understanding. Clinically, we are often faced with patients who have varying presentations and manifestations, but, for reasons that are unclear. For instance, two individuals with the same body mass index may have very different clinical manifestations, one with severe OSA and one without any OSA. Similarly, two individuals with an apnea hypopnea index of 40 events per hour (ie, severe OSA) may have very different symptoms attributable to OSA, eg, one could be asymptomatic and the other could be debilitated from sleepiness. We and others have been making efforts to determine why these phenomenon occur. At present, the techniques to define mechanisms underlying OSA are labor-intensive, requiring one or two overnight experiments to gather meaningful data. Although we are gathering new insights based on these techniques, efforts are ongoing to simplify these approaches and to make assessment of pathophysiologic characteristics more accessible to the clinician (Orr et al. Am J Respir Crit Care Med. 2017 Nov 30. doi: 10.1164/rccm.201707-1357LE. [Epub ahead of print]).

We ultimately believe that a thorough analysis of a sleep recording combined with demographic data and other readily available clinical data (perhaps plasma biomarkers) may yield sufficient information for us to know why OSA is occurring and what interventions might be helpful for an individual patient. Currently, our use of the polysomnogram to derive only an apnea hypopnea index does not take full advantage of the available data. An apnea hypopnea index can be readily obtained from home sleep testing and does not truly provide much insight into why a given individual has OSA, what symptoms are attributable to OSA, and what interventions might be considered for the afflicted individual. By analogy, if the only useful data derived from an ECG were a heart rate, the test would rapidly become obsolete. Along these lines, if the only role for the sleep clinician was to prescribe CPAP to everyone with an AHI greater than 5/h, there would be little need or interest in specialized training. In contrast, we suggest that rich insights regarding pathophysiology and mechanisms should be gathered and may influence clinical management of patients afflicted with OSA. Thus, we encourage more thorough analyses of available data to maximize information gleaned and, ultimately, to optimize clinical outcomes.

Recent studies suggest that sleep apnea occurs for varying reasons, a concept that is now thought to be clinically important (Jordan et al. Lancet. 2014;383[9918]:736). We draw a crucial distinction between endotypes (mechanisms underlying disease) and phenotypes (clinical expression of disease). Important endotypes include compromised upper airway anatomy, dysfunction in pharyngeal dilator muscles, unstable ventilatory control (high loop gain), and low arousal threshold (wake up easily), among others. Important phenotypes of sleep apnea are emerging and still evolving to include minimally symptomatic OSA, OSA with daytime sleepiness, and OSA with major cardiometabolic risk, among others. Several important concepts have emerged regarding different OSA endotypes and phenotypes:

1 The mechanism underlying OSA may predict potential response to therapeutic interventions. For instance, the endotype of OSA with unstable ventilatory control (high loop gain) may respond to agents such as oxygen and acetazolamide, which serve to stabilize control of breathing. In patients with anatomical compromise at the level of the velopharynx, uvulopalatopharyngoplasty may be an effective intervention. For patients with multiple pathophysiologic abnormalities, combination therapy may be required to alleviate OSA (Edwards et al. Sleep. 2016;9[11]:1973).

2 Given that OSA has many underlying etiologies, efforts are underway to determine whether individuals with different risk factors for OSA develop their disease based on varying mechanisms. As an example, people with posttraumatic stress disorder (PTSD) may be at increased risk of OSA perhaps on the basis of a low threshold for arousal (Orr et al. JCSM. 2017, 13[1]: 57-63). Another example would be patients with neuromuscular disease who may be at risk of OSA primarily based on impaired pharyngeal dilator muscle function.

3 A new concept is emerging whereby endotypes of OSA may actually predict differing OSA phenotypes. In theory, loop gain-driven OSA may have different consequences from OSA driven by compromise of pharyngeal anatomy. To this point, data suggest that OSA in the elderly may not have as many consequences as OSA in younger people matched on severity of illness. OSA in the elderly has lower loop gain than OSA in younger people and is associated with less negative intrathoracic pressure at the time of arousal as compared with younger individuals with OSA (Kobayashi et al. Chest. 2010; 137[6]:1310). As such, the endotype of OSA in the elderly may explain why the clinical consequences are fewer than in the younger OSA counterparts.

4 The mechanism underlying OSA may be important in determining response to clinical interventions, such as nasal CPAP. Patients with a low arousal threshold may be prone to insomnia when placed on CPAP and could theoretically be poorly tolerant of therapy based on disrupted sleep architecture. Such patients may benefit from non-myorelaxant hypnotic therapy to consolidate sleep and improve CPAP adherence. In addition, patients with high loop gain (unstable ventilatory control) may be prone to develop central apneas when placed on CPAP therapy (Stanchina et al. Ann Am Thorac Soc. 2015;12[9]:1351). These patients may benefit from newer technologies, eg, auto or adaptive servo ventilation - ASV. High loop gain has also been shown to predict failure of upper airway surgery as a treatment for OSA by several groups (Li et al. JCSM. 2017;13[9]:1029). Such patients should, perhaps, undergo nonsurgical therapies for OSA.

We emphasize that some of the points being made are somewhat speculative and, thus, encourage further basic and clinical research to test our assumptions. Robust, multicenter clinical trials assessing hard outcomes will ultimately be required to change the current standard of care. Nonetheless, we believe that a more thorough understanding of OSA pathogenesis can help guide clinical care today and will be critical to the optimal treatment of afflicted individuals tomorrow.

Dr. Owens is Assistant Clinical Professor of Medicine; Dr. Deacon is a Post-Doctoral Research Scholar; and Dr. Malhotra is Kenneth M. Moser Professor of Medicine and Chief, Division of Pulmonary, Critical Care and Sleep Medicine, University of California San Diego.

Obstructive sleep apnea (OSA) contributes a major health burden to society due to its high prevalence and substantial neurocognitive and cardiovascular consequences. Estimates suggest that at least 10% of adults in North America are afflicted with OSA, making it probably the most common respiratory disease in the developed world (Peppard et al. Am J Epidemiol. 2013;177[9]:1006). Nasal CPAP is a highly efficacious therapy that has been shown to improve neurocognitive and cardiovascular outcomes. However, CPAP is not always well tolerated. Alternative therapies, such as oral appliances and upper airway surgery, have highly variable efficacy, and evidence of important clinical benefits are uncertain. Therefore, efforts are ongoing to determine optimal alternative strategies for therapy.

In order to treat any condition optimally, one needs to be able to predict who is at highest risk of developing the condition, then to assess the consequences if left untreated, and finally to be able to predict response to various treatment options. Currently, the OSA field is still in its early stages of our understanding. Clinically, we are often faced with patients who have varying presentations and manifestations, but, for reasons that are unclear. For instance, two individuals with the same body mass index may have very different clinical manifestations, one with severe OSA and one without any OSA. Similarly, two individuals with an apnea hypopnea index of 40 events per hour (ie, severe OSA) may have very different symptoms attributable to OSA, eg, one could be asymptomatic and the other could be debilitated from sleepiness. We and others have been making efforts to determine why these phenomenon occur. At present, the techniques to define mechanisms underlying OSA are labor-intensive, requiring one or two overnight experiments to gather meaningful data. Although we are gathering new insights based on these techniques, efforts are ongoing to simplify these approaches and to make assessment of pathophysiologic characteristics more accessible to the clinician (Orr et al. Am J Respir Crit Care Med. 2017 Nov 30. doi: 10.1164/rccm.201707-1357LE. [Epub ahead of print]).

We ultimately believe that a thorough analysis of a sleep recording combined with demographic data and other readily available clinical data (perhaps plasma biomarkers) may yield sufficient information for us to know why OSA is occurring and what interventions might be helpful for an individual patient. Currently, our use of the polysomnogram to derive only an apnea hypopnea index does not take full advantage of the available data. An apnea hypopnea index can be readily obtained from home sleep testing and does not truly provide much insight into why a given individual has OSA, what symptoms are attributable to OSA, and what interventions might be considered for the afflicted individual. By analogy, if the only useful data derived from an ECG were a heart rate, the test would rapidly become obsolete. Along these lines, if the only role for the sleep clinician was to prescribe CPAP to everyone with an AHI greater than 5/h, there would be little need or interest in specialized training. In contrast, we suggest that rich insights regarding pathophysiology and mechanisms should be gathered and may influence clinical management of patients afflicted with OSA. Thus, we encourage more thorough analyses of available data to maximize information gleaned and, ultimately, to optimize clinical outcomes.

Recent studies suggest that sleep apnea occurs for varying reasons, a concept that is now thought to be clinically important (Jordan et al. Lancet. 2014;383[9918]:736). We draw a crucial distinction between endotypes (mechanisms underlying disease) and phenotypes (clinical expression of disease). Important endotypes include compromised upper airway anatomy, dysfunction in pharyngeal dilator muscles, unstable ventilatory control (high loop gain), and low arousal threshold (wake up easily), among others. Important phenotypes of sleep apnea are emerging and still evolving to include minimally symptomatic OSA, OSA with daytime sleepiness, and OSA with major cardiometabolic risk, among others. Several important concepts have emerged regarding different OSA endotypes and phenotypes:

1 The mechanism underlying OSA may predict potential response to therapeutic interventions. For instance, the endotype of OSA with unstable ventilatory control (high loop gain) may respond to agents such as oxygen and acetazolamide, which serve to stabilize control of breathing. In patients with anatomical compromise at the level of the velopharynx, uvulopalatopharyngoplasty may be an effective intervention. For patients with multiple pathophysiologic abnormalities, combination therapy may be required to alleviate OSA (Edwards et al. Sleep. 2016;9[11]:1973).

2 Given that OSA has many underlying etiologies, efforts are underway to determine whether individuals with different risk factors for OSA develop their disease based on varying mechanisms. As an example, people with posttraumatic stress disorder (PTSD) may be at increased risk of OSA perhaps on the basis of a low threshold for arousal (Orr et al. JCSM. 2017, 13[1]: 57-63). Another example would be patients with neuromuscular disease who may be at risk of OSA primarily based on impaired pharyngeal dilator muscle function.

3 A new concept is emerging whereby endotypes of OSA may actually predict differing OSA phenotypes. In theory, loop gain-driven OSA may have different consequences from OSA driven by compromise of pharyngeal anatomy. To this point, data suggest that OSA in the elderly may not have as many consequences as OSA in younger people matched on severity of illness. OSA in the elderly has lower loop gain than OSA in younger people and is associated with less negative intrathoracic pressure at the time of arousal as compared with younger individuals with OSA (Kobayashi et al. Chest. 2010; 137[6]:1310). As such, the endotype of OSA in the elderly may explain why the clinical consequences are fewer than in the younger OSA counterparts.

4 The mechanism underlying OSA may be important in determining response to clinical interventions, such as nasal CPAP. Patients with a low arousal threshold may be prone to insomnia when placed on CPAP and could theoretically be poorly tolerant of therapy based on disrupted sleep architecture. Such patients may benefit from non-myorelaxant hypnotic therapy to consolidate sleep and improve CPAP adherence. In addition, patients with high loop gain (unstable ventilatory control) may be prone to develop central apneas when placed on CPAP therapy (Stanchina et al. Ann Am Thorac Soc. 2015;12[9]:1351). These patients may benefit from newer technologies, eg, auto or adaptive servo ventilation - ASV. High loop gain has also been shown to predict failure of upper airway surgery as a treatment for OSA by several groups (Li et al. JCSM. 2017;13[9]:1029). Such patients should, perhaps, undergo nonsurgical therapies for OSA.

We emphasize that some of the points being made are somewhat speculative and, thus, encourage further basic and clinical research to test our assumptions. Robust, multicenter clinical trials assessing hard outcomes will ultimately be required to change the current standard of care. Nonetheless, we believe that a more thorough understanding of OSA pathogenesis can help guide clinical care today and will be critical to the optimal treatment of afflicted individuals tomorrow.

Dr. Owens is Assistant Clinical Professor of Medicine; Dr. Deacon is a Post-Doctoral Research Scholar; and Dr. Malhotra is Kenneth M. Moser Professor of Medicine and Chief, Division of Pulmonary, Critical Care and Sleep Medicine, University of California San Diego.