1

Introduction

Numerous treatment options for obstructive sleep apnea (OSA) presently exist. These treatments range from noninvasive behavioral modifications to nightly use of continuous positive airway pressure (CPAP) devices to numerous pharyngeal expansion procedures that either alter airway anatomy or modify airway physiology. This chapter seeks to provide a framework to help guide physicians to a synthesis and evaluation of treatment options in this disease. It is important to note that few true practice guidelines exist to inform decision making; hence this chapter is designed to synthesize available evidence informing surgeons as to how best to treat their patients. A more detailed description of the individual components of the evaluation, diagnosis, and treatment of OSA patients is discussed in chapters fully dedicated to them.

2

Perspectives on Disease Outcomes

It has been almost two centuries since Charles Dickens published The Posthumous Papers of the Pickwick Club, wherein the appearance and health concerns of Mr. Pickwick, the protagonist, were described in such detail so as to leave little doubt in later years that he suffered from severe OSA. This is often quoted as the first instance of the disease being recognized in society, and consequently OSA is now also known by the eponym Pickwickian syndrome . Since Dickens’ time, understanding of this disease process has truly exploded. The prevalence of OSA in the adult population is currently estimated at a high 3% in women and 9% in men. OSA is recognized to lead to daytime somnolence, reduced concentration, headaches, and memory loss. Numerous cohort studies have demonstrated OSA to be a risk factor for cardiovascular/neurologic disease and mortality. OSA has deleterious effects on economic productivity and is known to be a significant source of fatal and nonfatal motor vehicle accidents. Treatment of OSA is proven to be associated with an improvement in cardiac status, a decrease in the incidence of stroke, and reduction in car accident rates.

Sleep-disordered breathing can be viewed as a continuum from simple snoring without apnea at one end to a patient with severe OSA and metabolic syndrome at the other. The level of disease in an individual patient can improve or worsen, moving the patient up or down this continuum both over time and on a night-to-night basis. For instance, a normal patient who gains a significant amount of weight may begin to snore, develop arousals during the night, and begin to move into a disease state. With additional weight gain, this patient may develop increased airway resistance or sleep apnea. A patient who consumes alcohol on a given night can also move on the scale in a similar fashion; as muscle tone decreases, obstruction increases. Similarly, a patient with sleep apnea who loses weight may develop less severe airway resistance, simple snoring, or even achieve a normal state of breathing at night. Other important parameters such as blood pressure or metrics of alertness can also fluctuate. Thus the numbers generated on a polysomnogram (PSG) must be viewed in the context of the patient’s overall disease and night-to-night variation of their sleep-disordered breathing. Sleep apnea is much more than the Apnea/Hypopnea Index (AHI) and O ₂ saturation data as measured on an overnight sleep study. The PSG is simply a single-night snapshot in time that may not reflect a patient’s level of disease over a longer period, and thus it is crucial that clinicians look not just at PSG data but also at patient-specific data when evaluating OSA severity and outcomes after treatment. Moreover, recent evidence has shown significant discordance between the levels of AHI used to denote outcomes of therapy and real-world clinical outcomes such as quality of life (QOL), patient perception of disease, cardiovascular measures, disease burden, and/or survival. There is a mountain of evidence showing how the AHI can vary from night to night, from laboratory to laboratory, from various nasal thermistor to pressure transducers, and based on the different definitions of hypopnea used in different laboratories and software. Hobson et al. recently showed in a creative study that differences even in the definition of AHI severity cutoff can greatly influence reported efficacy of surgery in patients with OSA. The contemporary reliance on AHI as generally the only outcome measure assessed in research programs is not in line with many other aspects of medicine that are becoming patient centered as opposed to test centered.

According to the surgical literature on OSA treatment, Sher’s success criteria of 50% reduction in AHI and an AHI less than 20 tend to be the benchmark for success. However, this archaic concept was based entirely on an arbitrary AHI number that did not stratify patients by likelihood of surgical success. These oft-quoted criteria should be abandoned as being both insufficient and out of date. Consider a patient with preoperative AHI of 95 who has a postoperative PSG showing an AHI of 21; this person would likely experience significant symptomatic clinical improvement with a huge decrease in disease burden even though they are not, by definition, considered a successful surgical outcome by the numerical AHI criteria. Whereas a patient with a preoperative AHI of 35 reduced postoperatively to under 14 is considered a successful AHI outcome even though the likelihood of clinical cardiovascular, disease burden, and/or QOL impact may be minimal. It is essential that OSA treatment has clinical significance rather than satisfy some numerical criteria.

The authors introduce the acronym SLEEP GOAL as a more comprehensive set of success parameters than just relying on AHI:

• 
  

  S – Snoring visual analog scale

  
  
• 
  

  L – Latency of sleep onset (PSG or Multiple Sleep Latency Test)

  
  
• 
  

  E – Epworth Sleepiness Scale—normalization to less than 10 (if it was abnormal pretreatment)

  
  
• 
  

  E – Execution time—improvement by more than 50%, using performance vigilance testing

  
  
• 
  

  P – Pressure (systolic blood pressure [SBP])—(a) reduction in mean blood pressure by 7 mm Hg, or (b) single reduction in either SBP or diastolic blood pressure (DBP) by 10 mm Hg, or (c) 5 mm Hg reduction in both SDP and DBP

  
  
• 
  

  G – Gross weight/body mass index (BMI)—loss of >10% gross weight and/or reduction in BMI from one category to another

  
  
• 
  

  O – Oxygenation—improvement of duration (minutes) of O ₂ <90% by at least half

  
  
• 
  

  A – AHI via sleep study—reduction by 50% and AHI <20

  
  
• 
  

  L – Life score—improvement in a validated OSA-related quality of life assessment tool

3

Threshold to Treat Obstructive Sleep Apnea

There are three principal manifestations of sleep-disordered breathing: physiologic , behavioral, and social . The physiologic manifestations are the most important and principally relate to cardiovascular consequences of OSA. Cardiovascular consequences include associations with stroke, hypertension, and myocardial infarction and are being studied rigorously in the Sleep Heart Health Study, which is longitudinally tracking a cohort of OSA patients with measures of OSA and medical outcomes. Patients with observed cardiovascular dysfunction that is associated with OSA should be treated. Perhaps the most typical example is that of a patient with resistant hypertension for which OSA is often underrecognized as a reason for the blood pressure problem.

The behavioral effects of sleep disruption are commonly evident. These may manifest as tiredness in the morning or daytime, falling asleep in permissive situations such as watching TV or reading a book, an increased incidence of motor vehicle accidents, and losses of concentration and productivity. A simple patient self-report scale, the Epworth Sleepiness Scale, can be used to quantify this level of tiredness, though there are also objective measures of alertness, such as the psychomotor vigilance test. The behavioral effects of OSA can vary among patients, where a patient with snoring may have significant tiredness, whereas a patient with an elevated AHI and greater nighttime obstruction may not display tiredness. This discrepancy appears to make behavioral measurement alone insufficient to fully characterize the disease or its treatment. The other main behavioral effect of untreated OSA is that of the newly recognized synergism with depression. Patients who are symptomatic with OSA or who have associated depression should be treated.

Lastly, social effects are principally related to snoring. This is the most common reason that patients with OSA present to their physician. Even though it is the least important physiologic issue, it is paradoxically often the most important reason the patient expresses in their desire to obtain treatment. Subjective measures are commonly used to measure this feature of the disease, such as bed partner scoring of snoring severity. Wearable technology or smartphone apps are now available that can also measure snoring severity. Although there is no mandatory need to treat simple snoring, if patients present with this, then other more severe manifestations of sleep-disordered breathing still should be ruled out.

In light of the variability in the manifestation of OSA in patients and the gaps in our knowledge as to which patients will suffer cardiovascular consequences of this disease, significant controversy exists with respect to selecting patients for OSA treatment. In a review article on OSA, Ward Flemons writes, “In the majority of patients without coexisting conditions … the primary reason to test for and treat sleep apnea is the potential to improve the quality of life.” Studies vary in determining the magnitude of effects of mild to moderate OSA on hypertension and cardiovascular consequences. However, ongoing studies of large cohorts of patients such as the Sleep Heart Health Study are yielding important data that will provide guidelines for treatment in the future.

Most physicians still look to PSG data as the best numerical evidence to inform treatment thresholds. As discussed earlier, it is the opinion of the authors that using PSG outcomes alone is insufficient to manage this patient population. Nevertheless, in an attempt to help delineate a treatment threshold from PSG data, many use the cutoff of an AHI of 30 as the point for intervention, even though rigorous evidence for treatment of patients with an AHI of 30 is lacking. Patients with AHI of 30, and especially those with clinically significant oxygen saturation changes on the PSG and/or with significant comorbidity or behavioral effects (e.g. tiredness, loss of concentration), should be considered candidates for treatment of OSA.