The Nuts and Bolts of Scoring Apneas and Hypopneas

AIRFLOW AND THORACO/ABDOMINAL EFFORTS
It is important to realize that the labels ascribed to the airflow and effort channels of a polysomnogram are misleading. What we call airflow is typically measured by temperature variations detected with a thermocouple, a thermistor, or airflow temperature sensors. (A thermocouple produces a change in voltage associated with a change in temperature of the thermocouple sensor.) In contrast, a thermistor (a cross between the terms thermometer and resistor) requires a voltage source passing a constant current through the thermistor junction of two dissimilar metals. Temperature variations, produced by exhaled and inhaled air, produce changes in the resistance of the thermistor junction. These changes in resistance are roughly proportional to the amplitude and frequency variations of respiration. Thus, a thermocouple produces a change in voltage and requires no power source. In contrast, a thermistor produces a change in resistance, does require a power source (usually a battery), and requires special circuitry to measure the resistance changes. The important thing to realize is that what we label as airflow on a polysomnogram is really not airflow, but simply variations in air temperature detected with a thermistor, thermocouple, or similar temperature-transducing device.

Individuals who have been scoring polysomnograms for many years may have occasionally seen central apneas, in which there is essentially no detectable effort in the abdominal or thoracic effort channels, but in which there is an apparent (but false) signal in the thermocouple or thermistor channel. In such circumstances, the apparent variations in the airflow tracing obviously must be deemed as artifact due to some sort of temperature fluctuations detected by the transducer. Thus, the airflow signal is really only a correlate of airflow, but not actual airflow. Actual airflow would be measured in units of liters per minute, milliliters per second, or some sort of combination of volume and time. We do not measure airflow traditionally, except with a pneumotachograph, which involves placing a mask over both the nose and the mouth and actually measuring airflow in terms of liters per minute.

In a similar line of reasoning, the thoracic and abdominal effort signals are really not measures of effort at all. In most laboratories nowadays, these effort signals merely measure the changes in the expansion and contraction of the thoracic and abdominal cavities. Placing an elastic belt in series with a piezo-electric sensor does this. One is placed at the level of the upper thorax, generally right underneath the arms to pick up rib cage movements, and the other is placed at the level of the umbilicus to pick up abdominal movements. In fact, neither one of these effort signals actually measures respiratory effort. Efforts are measured in units of force (dynes in the CGS system) or Newtons in the MKS system. We are not measuring a physical force. Instead, during the inhalation signal, we are measuring the result of air being sucked into the lungs, thus causing the thoracic and abdominal cavities to expand.

Therefore, as most experienced sleep technicians have noticed, sometimes a very good indicator of a hypopnea is not so much a reduction in the airflow signal, but a reduction in the thoracic or abdominal effort signals. All this means is that because of an upper airway obstruction, or because of actually reduced effort, less air is filling up the lungs. This causes the thoracic and abdominal cavities to expand less, thus producing a smaller amplitude signal on the two effort channels. This is why the most recent hypopnea definitions stress that a hypopnea is to be defined, in part, as a visibly discernible reduction in the amplitude of either the airflow or the effort signals.

APNEAS, HYPOPNEAS, AND HYPERPNEAS
Respiratory events are divided into two phases: the apnea or hypopnea phase, in which airflow is substantially or partially reduced, and the subsequent hyperpnea phase, during which there is an increase, usually in both airflow (temperature variations) and effort (respiratory movements). The arousal typically occurs coincident with the onset or slightly before the first hyperpneic breath begins. However, on occasion, the arousal may occur during the second or later breath of the hyperpneic phase of respiration. In contrast, the oxygen desaturation usually occurs 10 to 30 seconds later, owing to the blood circulation delay time.

While apneas have enjoyed a relatively solid consensus definition for many years, hypopneas have not. An apnea has traditionally been defined as “a cessation of airflow at the level of the nostrils and mouth lasting at least 10 seconds.”¹ Many courses in sleep disorders medicine and many computerized sleep systems quantify the word “cessation” as meaning an 80% to 100% reduction in the airflow signal for 10 seconds or longer. However, in order for an apnea to be considered clinically significant, it must be followed by either an arousal of some sort and/or an oxygen desaturation.

In contrast, hypopneas have not been defined so precisely.^2,3 Various authorities have defined hypopneas either conservatively or liberally. A conservative definition of a hypopnea (one that yields fewer hypopneas than a liberal definition) was defined in a paper by Gould et al.⁴ Their conservative definition was that a hypopnea was defined as a 50% reduction in the sum of thoracoabdominal excursion signal obtained from inductive plethysmography, lasting for >10 seconds. In contrast, Guilleminault1 initially proposed a liberal definition of a hypopnea as “a reduction—but not complete cessation—of air exchange.” More recent liberal definitions of a hypopnea have ranged all the way up to any visually discernible change in the appearance in the airflow signal or effort signals. This change may be a decrease in the amplitude or change in the shape of either the airflow signal (airflow flattening) or one of the two effort signals (abdominal or thoracic movement signals). Authorities² agree that in order for a hypopnea to be clinically significant, it must be followed by either an arousal of some sort and/or an oxygen desaturation.

A recent slide show presentation produced by the American Academy of Sleep Medicine (and presumably approved by the AASM Board of Directors)⁵defines a hypopnea as any visually discernible reduction in the amplitude of either the airflow or the effort signals. This decrease in airflow must last for 10 seconds or longer. Also, in order to be counted as clinically significant, either an arousal of some sort and/or an oxygen desaturation of three points or more must follow the hypopnea.

However, the issue of an apnea occurring within a hypopnea (an event that could be called an apnea/hypopnea) has not yet been addressed. Most experienced sleep technicians have seen an event that could be scored as a hypopnea based on the majority of the event but contains an apnea as well. That occurs at the beginning and/or the end of the event with airflow signals that are less than an 80% reduction (typically only about a 50% reduction) in airflow. However, somewhere within the apnea/hypopnea event (beginning, middle, or end), there exists a period of 10 seconds or more of 80% to 100% airflow reduction that is great enough to be called an apnea.

The question arises whether such an apnea/hypopnea event should be scored as an apnea for the entire duration of the event (making the apnea longer than it actually is) or scored as a hypopnea for the entire duration of the event (thus ignoring the apnea component within). While arguments can be marshaled for both cases, I have recently changed my mind on this issue. I used to abide by the strict definition of an apnea argument (an apnea had to be an 80% to 100% reduction in airflow for the entire duration of the event). I have changed my mind to a more liberal interpretation of an apnea as being an event that contains the requisite 80% to 100% reduction in airflow but sometimes only during a portion of the event. The reason why I changed my mind is that nowhere in any of the definitions of an apnea does it specifically say that the entire apnea has to be an 80% to 100% reduction in airflow. They just say that there has to be an 80% to 100% reduction in nasal/oral airflow for 10 seconds or longer. Those 10 seconds could be anywhere within an event that would otherwise be counted as a hypopnea. Second, the shape and duration of the oxygen saturation level change parallel the duration of the entire apnea/hypopnea complex, not just the apnea portion of the event. Third, given the current Medicare rule, which requires that a sleep study have 30 apneas within 6 hours of sleep in order to qualify for continuous positive airway pressure (CPAP) therapy, it is to the patient’s advantage to call an event an apnea whenever we can legitimately do so. At the present time Medicare has a frustrating restriction on CPAP usage for patients who have less than 30 apneas even though they may have hundreds of hypopneas and oxygen desaturations down to 50%.

Apneas are classified as three types: central, obstructive, and mixed.

Central Apnea. Traditionally, a central apnea has been defined as a nearly complete (80% to 100%) reduction in both the thoracic and abdominal effort channels and, of course, the airflow signal, since logically there should be no airflow if there is no respiratory effort. There should be no paradoxical breathing before or after a purely central apnea. The abdominal and thoracic movement signals should be entirely in phase before and after a central apnea. However, there are some exceptions mentioned earlier, in which temperature variations produce artifacts, which mimic airflow.

Rather frequently, a false signal or ballistocardiogram artifact will appear in one or both of the effort channels. This can be identified as small to medium sized oscillations in the thoracic or abdominal movement channels. The cause of the ballistocardiogram is the force of the heart beat causing the chest to expand and contract slightly but at a rate exactly equal to the ongoing heart rate. Noting the almost perfect correlation between the oscillations in the effort channels and the ongoing electrocardiography (ECG) signal allows one to identify such ballistocardiogram artifact. The main point to recognize is that there is little or no movement in either the abdominal or thoracic movement channels for 10 seconds or longer. Again, such an event is considered by some to be clinically significant only if either an arousal of some sort and/or an oxygen desaturation follows it. Central apneas are often a significant cause of insomnia.⁶

Indicators of Upper Airway Obstruction. A topic that is appropriate to discuss here is the list of indicators that identify upper airway obstruction. One way is to examine the relationship between the thoracic and the abdominal movements. Normal, unobstructed breathing is usually associated with in-phase movements of the thorax and abdomen moving in and out more or less together, outward at the same time during inhalation, and inward at the same time during exhalation, tracking each other synchronously. During periods of upper airway obstruction, there is a compensatory increase in negative airway pressure to compensate for the obstruction. If excessive, this increase in negative airway pressure causes the chest to partially collapse during inhalation, just as it causes the upper airway to collapse just prior to inhalation. This results in phase shifted or, if completely out-of-phase thoracic and abdominal movement signals, paradoxical breathing. This refers to a situation in which the thoracic and abdominal cavities move opposite to each other in a seesaw fashion. During true paradoxical breathing during inhalation, the chest caves in while the belly moves out. Similarly, during exhalation, the chest then expands and the abdominal cavity moves inward. There are various gray zones between perfectly in-phase and perfectly out-of-phase breathing, such as partial phase shifts between abdominal and thoracic movements. These should be interpreted as partial upper airway obstructions. Also, if the patient does not develop enough negative airway pressure to cause a slight collapse of the chest cavity, then the patient can still be having an obstructive apnea or hypopnea without paradoxical movements.

A second way of identifying upper airway obstruction is to examine the shape of the airflow signal. Airflow signal flattening or the presence of airflow signals with “shoulders” on them are two indications of upper airway resistance.⁷

A third way of identifying upper airway obstruction is to note the presence of snoring. Snoring is almost always considered to be an indication of partial upper airway obstruction, since the obstruction causes an increased airflow speed (to compensate for the obstruction) and this causes structures in the upper airway to vibrate. Snoring occurs only during hypopneas, and not during complete apneas (when airflow is totally obstructed), because if airflow is reduced too much, there is generally not enough airflow to cause the uvula or the base of the tongue to vibrate, therefore, no snoring sounds can be produced.

Obstructive Apnea. An obstructive apnea is an event defined as an 80% to 100% reduction in the airflow signal accompanied by some identifiable respiratory effort. Usually the thoracic and abdominal movement channels indicate either paradoxical breathing (out-of-phase movements of the thorax and abdominal cavities), or at least phase-shifted movements (the peaks of the indicators of the chest and abdomen do not line up exactly). Often, the beginning of an obstructive apnea is the peak of the last in-phase, large breath. Similarly, the ending of an obstructive apnea is often the beginning of the next in-phase, large breath. However, an obstructive apnea does not always have to have paradoxical breathing associated with it. Sometimes, the movement channels indicate in-phase breathing. This is because not enough negative airway pressure has developed to partially compress the chest cavity. As before, either an arousal of some sort and/or an oxygen desaturation must follow the event in order for it to be considered clinically significant. If there is no oxygen desaturation and no electroencephalographic (EEG) arousal or if the apnea occurs entirely during a waking EEG, the apnea may (at the judgment of the sleep laboratory medical director) not be counted as clinically significant.

Mixed Apnea. The most difficult type of apnea to define is a mixed apnea. In the textbooks on sleep medicine, mixed apneas are classified for categorization and pathophysiology purposes in the same category as obstructive apneas (obstructive apneas and mixed apneas count together as indicators of upper airway obstruction as the underlying pathology). However, experienced technicians have probably seen an event in which most of the event (maybe 90%) appears to be central, but there is evidence of paradoxical breathing either at the very beginning of the apnea or near the very end. Such events are classified as mixed apneas. The rationale for this is that paradoxical breathing is an indication of upper airway obstruction caused by the increased negative airway pressure causing the chest cavity to partially collapse. Most times when there is evidence of paradoxical breathing, one must assume that an upper airway obstruction has led to this consequence. However, reduced chest wall compliance can also cause paradoxical breathing.

Hypopneas. Some authorities in the field have also categorized hypopneas as central, obstructive, or mixed.^1,5 While not acknowledging the existence of “mixed hypopneas,” the American Academy of Sleep Medicine Task Force (1999) does recognize and distinguishes between obstructive apnea/hypopnea events and central apnea/hypopnea events. Other authorities have only recognized obstructive hypopneas. Some individuals have attempted to distinguish central from obstructive hypopneas based on airflow (usually temperature) and effort (or movement) signals.

Central Hypopnea. Central hypopneas are associated with reductions of purely in-phase thoracic and abdominal effort or movement signals, followed by an increase in chest and belly movements at the end. There is no evidence of phase shifting or paradoxical breathing, no airflow flattening, and no snoring throughout the entire central hypopnea.

Obstructive Hypopnea. In contrast, obstructive hypopneas should be scored when there is evidence of even a slight degree of paradoxical breathing, including even slight phase shifting and/or snoring and/or airflow flattening in the airflow signal. This is particularly observable using a nasal pressure transducer, which is very readily capable of picking up the shoulders or flattenings in the airflow signal. A nasal pressure transducer detects the variations in negative airway pressures at the nares using a cannula inserted into the nose. A tube is connected to an air pressure transducer. Such systems have been shown to be much more sensitive in terms of detecting airflow flattening than thermistors or thermocouples. Also, nasal pressure transducers have been shown to detect far more hypopneas than either thermistors or thermocouples.⁸ Shoulders or flattening of slopes in the airflow signal are also an indication of upper airway obstruction, as is paradoxical breathing and/or snoring.^7,9 However, it is possible for a hypopnea to be obstructive and still exhibit in-phase thoracic and abdominal movements (like a central hypopnea), as long as not enough negative airway pressure is created to cause paradoxical or even phase shifting of the thoracic and abdominal movement signals. In such cases there must be other evidence of upper airway obstruction such as airflow flattening or snoring in order to score an event as an obstructive hypopnea.

CLINICAL CONSEQUENCES
The one consistent requirement that seems to have pervaded most definitions of all respiratory disturbances is the idea of a clinically significant consequence. Two clinically significant consequences have enjoyed a nearly universal recognition. These are either an arousal of some sort and/or an oxygen desaturation.

Before we proceed further, we need to distinguish between two types of arousals: EEG arousals and cardiac arousals, either one of which will serve to meet the arousal requirement making an apnea or hypopnea clinically significant.

EEG Arousals. The American Sleep Disorders Association (ASDA), in one of its position papers, defined an EEG arousal as “an abrupt shift in EEG frequency, which may include theta, alpha, and/or frequencies greater than 16 Hz but not spindles, subject to the following rules and conditions.” The ASDA task force then delineated 11 conditions that further defined its definition of an EEG arousal.¹⁰ These EEG arousals usually occur at the conclusion of respiratory disturbances (apneas or hypopneas), usually (but not always) just before or at the beginning of the first hyperpneic breath. However, sometimes the EEG arousal begins at the second or later hyperpneic breath. Different authorities define arousals differently.¹¹

There exist the occasional occurrences of questionable events in which at the very beginning of the hyperpnea phase there is no EEG arousal, and there is no oxygen desaturation following the hyperpnea phase within 30 seconds. Instead, somewhat rarely, an EEG arousal will occur during the second or later breath of the hyperpnea phase. My view is that such events should be scored as apneas or hypopneas as appropriate, ie, scored as apneas if there is an 80% to 100% reduction in airflow or scored as hypopneas if there is a visibly discernible reduction in either the airflow or effort signals. The rationale for this is that the EEG arousal probably would not have occurred had the respiratory disturbance not occurred. Just because the EEG arousal does not begin (as it usually does) coincident with or slightly before the onset of the first hyperpneic breath does not mean that it is not a consequence of the sleep-disordered breathing.

Cardiac Arousals. The second type of physiological arousal is called the cardiac arousal. Cardiac arousals were defined in an excellent paper by Martin et al¹² as “recurrent induced arousals detected by transient blood pressure and/or heart rate increases in normal subjects.” These cardiac arousals produced “small but significant changes in daytime sleepiness in response to invisible sleep fragmentation.” Tones were used as arousal stimuli. The intensity ranged from 38 to 65 dB and the duration from 0.25 to 4 seconds. They used a plethysmograph to measure average increases in blood pressure of 8 mm Hg. These cardiac arousals caused decreases in the multiple sleep latency test scores from 11 to 7 minutes and decreases in maintenance of wakefulness test scores from 34 to 24 minutes, on average. For the purposes of this paper, a cardiac arousal can be of two types: an increase in blood pressure measured in mm Hg and an increase in heart rate measured in beats per minute. The heart rate increase can be of two levels: a slight increase in heart rate to a rate of less than 100 beats per minute but higher than during the apnea or hypopnea, and a more substantial increase in heart rate to over 100 beats per minute or more, an event given the term tachycardia.

Oxygen Desaturations. Oxygen desaturations have been defined either conservatively (a four point or greater drop in oxygen saturation levels) or liberally (only a two point or greater drop in oxygen saturation levels). There has been no mention as to what nadir value an oxygen desaturation must reach to be clinically significant. Some physicians require that an oxygen desaturation go down to at least 90% Sao2 or lower in order to be significant, but this has not been formalized. At a normal blood pH of about 7.35 to 7.45, an oxygen saturation level of 90% corresponds to a Po2 level of about 60 mm Hg. This is usually considered to be a clinically significant drop in the blood oxygen level.

There are some rare occurrences of apneas or hypopneas that produce no oxygen desaturations and no EEG arousals but do produce marked cardiac arousals. Such events should be counted as they probably play a major role in contributing to the deleterious effects of sleep apnea upon heart health. The repeated deceleration of heart rate during the apneas or hypopneas sometimes reaches the level of bradycardia with a heart rate of less than 50 beats per minute. The subsequent repeated accelerations in heart rate, during the following periods of hyperpneic breathing, usually are quite visually noticeable but rarely reach the level of being called tachycardia with a heart rate of greater than 100 beats per minute. Such rapid and repeated deceleration and accelerations in heart rate put a strain on the heart and in fact the entire cardiopulmonary-pulmonary system.

A rough analogy could be made to driving. Assume you had been driving at a constant speed for some time (analogous to a steady heart rate during a period of the night when no apneas or hypopneas are occurring). Then, for some reason that has yet to be determined, the apneas and hypopneas begin to occur (sometimes out of the blue). This would be analogous to repeatedly and alternately first stepping on the brake (heart rate slowing down possibly to the level of bradycardia), then stepping on the accelerator pedal (heart rate speeding, rarely to the level of tachycardia). Doing so repeatedly for several hours or more is likely to put more of a strain on the mechanisms of a car than driving at a constant cruising speed. In a like manner, the repeated deceleration and accelerations of heart rate (in the extreme, known as bradytachycardia) put a strain on the mechanisms of the cardiopulmonary system. This stress leads to high blood pressure (systemic hypertension), pulmonary hypertension, and occasionally right-sided heart failure. Symptoms of right-sided heart failure include chronic fatigue, dyspnea on exertion, and/or swollen legs, ankles, or feet.

It is important to address the rare occurrences of apneas and hypopneas without apparent consequence. Just before writing this article, I came across a record during which the patient had severe obstructive sleep apnea/hypopnea syndrome during the diagnostic portion of a split-night study. Once CPAP was applied, the patient continued to have central apneas and central hypopneas, sometimes very clearly demarcated, with no visible signal in either the airflow or effort channels, but without clinical consequence. In other words, during the hyperpneic phase of respiration, more often than not, there was no evidence of either an EEG arousal or an oxygen desaturation whatsoever. Such hypopneas (and perhaps apneas too, if the laboratory director concurs) should not be counted unless they produce a noticeable degree of heart rate slowing during the event and heart rate acceleration during the hyperpnea phase (a cardiac arousal).

John Zimmerman, PhD, is laboratory director of Mountain Medical Sleep Disorders Center, Carson City, Nev. He is also a site visitor on the accreditation committee for the AASM.

REFERENCES
1. Guilleminault C. Sleeping and Waking Disorders: Indications and Techniques. Menlo Park, Calif: Addison-Wesley Publishing Co; 1982.
2. Moser NJ, Phillips BA, Berry DTR, Harbison L. What is hypopnea, anyway? Chest. 1994;105:426-428.
3. Redline S, Sanders M. Hypopnea, a floating metric: implications for prevalence, morbidity estimates, and case finding. Sleep. 1997;20:1209-1217.
4. Gould GA, Whyte KF, Rhind GB, et al. The sleep hypopnea syndrome. Am Rev Respir Dis. 1988;137:895-898.
5. American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep. 1999;22:667-689.
6. Frederickson PA, Krueger BR. Insomnia associated with specific polysomnographic findings. In: Kryger MH, Roth T, Dement WC, eds. Principles and Practice of Sleep Medicine. 2nd ed. Philadelphia: W.B. Saunders & Company; 1994:523-534.
7. Condos R, Norman RG, Krishnasamy I, et al. Flow limitation as a noninvasive assessment of residual upper-airway resistance during continuous positive airway pressure therapy of obstructive sleep apnea. Am J Respir Crit Care Med. 1994;150:475-480.
8. Norman RG, Ahmed MM, Walsleben JA, Rapoport DM. Detection of respiratory events during NPSG: nasal cannula/pressure sensor versus thermistor. Sleep. 1997;20:1175-1184.
9. Montserrat JM, Ballester E, Olivi H, et al. Time course of stepwise CPAP titration. Behavior of respiratory and neurological variables. Am J Respir Crit Care Med. 1995;152:1854-1859.
10. ASDA Task Force on EEG Arousals. EEG arousals: scoring rules and examples. A preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep. 1992;15:174-184.
11. Martin SE, Englman HM, Deary IJ, Douglas NJ. Comparison of microarousal definitions in patients with sleep apnea [abstract]. Am J Respir Crit Care Med. 1995;151:A154.
12. Martin SE, Wraith PK, Deary IJ, Douglas NJ. The effect of invisible sleep fragmentation on daytime function. Am J Respir Crit Care Med. 1997;155:1596-1601.