Transcript Hypnotics and OSA: Rumors and Facts

Hypnotics and OSA:
Rumors and Facts
DOUGLAS KIRSCH, MD
CLINICAL INSTRUCTOR, HARVARD
MEDICAL SCHOOL
REGIONAL MEDICAL DIRECTOR,
SLEEP HEALTHCENTERS
Purpose of This Talk
 To examine the question of whether hypnotics are
dangerous in patients who have untreated sleepdisordered breathing
 To review information about whether some
hypnotics may be beneficial in the treatment of
OSA
To Cover:
 A Case
 What are Hypnotics Used For?
 What are Hypnotics?
 Do Hypnotics Affect the Airway?
 Can Hypnotics be of Benefit in OSA Treatment?
 Close
A Case
An Early Case (1981)
 A 38-year-old man with a long-standing history of
insomnia and daytime sleepiness was evaluated.
 He was found to have 7-18 primarily obstructive apneas
per night on four baseline recordings.
 On the first two nights on which he received 30 mg of the
benzodiazepine hypnotic flurazepam, there were 22 and
100 apneas, and during the daytime he became extremely
sleepy.
Mendelson WB, Garnett D, Gillin JC. J Nerv Ment Dis. 1981 Apr;169(4):261-4
An Early Case, Part II
 Upon cessation of medication, his clinical condition
improved, and the number of apneas decreased to 11 and
6 on withdrawal nights 4 and 6.
 Although respiratory depression is neither invariable nor
unique to flurazepam, this case suggests that it may be a
clinically significant problem with recommended oral
doses in some individuals
Mendelson WB, Garnett D, Gillin JC. J Nerv Ment Dis. 1981 Apr;169(4):261-4
Why Use a Hypnotic?
What is Insomnia?
 Insomnia is a symptom:

Difficulty in sleep initiation
 “ I can’t fall asleep”

Difficulty in sleep maintenance
 “I keep waking up all night”

Complaint of non-restorative sleep
 “I just don’t feel rested in the morning”
Epidemiology of Insomnia
NSF Poll 2005
 54% reported that, within the past year, they have experienced at
least one symptom of insomnia at least a few nights a week
 33% said they have experienced at least one symptom every night or
almost every night.
 Symptoms (at least a few times per week)
Waking up feeling unrefreshed (38%)
 Waking up a lot during the night (32%)
 Difficulty falling asleep (21%)
 Waking up too early and not able to get back to sleep (21%)

Hypnotics
 Benzodiazepines
 Lorazepam, Diazepam, Clonazepam, etc.
 Non-Benzodiazepine Receptor Agonists
 Zolpidem, Eszopiclone, Zaleplon
 Others:
 Alcohol
 Antidepressants: trazadone
 Anti-histamines: diphenhydramine
 Anti-psychotics: quietapine
 Sodium Oxybate
http://belaray.com/blog/wp-content/uploads/2008/06/medication.jpg
Insomnia Medications
Wilson and Nutt, Clin Med 2005
Sleeping Medications: General Rules
 Treatment with medications should:
 begin
with the lowest possible effective dose
 be short-term, if used nightly
 be intermittent, if used long-term
 be used only in combination with good sleep practices
and/or behavioral approaches
Practice Parameters, AASM
Hypnotics: A Survey of 130 Patients
Lu et al. JCSM 2007
Prescription Sedatives in OSA Pts.
 Sleep specialists tend to
prescribe sedatives at a
lower rate for patients
with OSA than nonsleep specialists
Lu et al. JCSM 2007
Benzodiazepines
 The “older” class of sleeping aids
 They are considered “CNS depressants”
 Enhancement of the effect of the
neurotransmitter GABA
 Also used to treat anxiety, insomnia,
agitation, seizures, muscle spasms, alcohol
withdrawal
 100 million scripts written in 1999 (DEA)
http://en.wikipedia.org/wiki/Benzodiazepine
http://library.thinkquest.org/C0115926/drugs/sedative2.htm
Changes in Sleep with Benzos: Meta-analysis
Nowell et al.,
JAMA, 278:24, 1997
Another Benefit of Hypnotics?
 Sedatives and Falls

In 34,163 nursing home residents


(76% women, mean age 84 +/- 8 y)
Evaluated Hypnotic use

Hypnotics as defined by CMS included aprobarbital, flurazepam, quazepam,
triazolam, pentobarbital, ethchlorvynol, estazolam, temazepam and
secobarbital; others such as NBZRAs (e.g., zolpidem) were also included
 Hypnotic use did not predict falls
 (adjusted odds ratio (AOR) 1.13, 95% confidence interval (CI) 0.98, 1.30).

In contrast, insomnia did predict future falls

(AOR 1.52, 95% CI 1.38, 1.66).
J Am Geriatr Soc 53:955–962, 2005
A Hypnotic Risk: MVAs in OSA
Lu et al. JCSM 2007
Early Studies with Benzodiazepines and Sleep
http://thebrain.mcgill.ca/flash/i/i_04/i_04_m/i_04_m_peu/i_04_m_peu.html
COPD
 Chronic Obstructive Pulmonary Disease
 Chronic obstructive bronchitis and emphysema

a pair of two commonly co-existing diseases of
the lungs in which the airways become narrowed

Associated with symptoms such as dyspnea,
cough and sputum production.

COPD is caused by noxious particles or gas,
most commonly from tobacco smoking, which
triggers an abnormal inflammatory response in
the lung
http://en.wikipedia.org/wiki/COPD
COPD and Benzodiazepines
 In 1972, Gaddie et al
 Nitrazepam
10mg may cause hypoventilation in 6 pts.
with COPD
 Ex: 1 Pt: PaO2 fell from 48 to 35 mm Hg while the
PaCO2 tension rose from 59.5 to 68 mm Hg
 Clark et al (1971) and Model (1973)
 Reported
serious benzodiazepine-induced respiratory
depression in patients with COPD
March 2, 1990 The American Journal of Medicine Volume 88 (suppl 3A)
What is Hypoventilation?
 Hypoventilation is too shallow or too slow breathing, which
does not meet the needs of the body. It may also refer to
reduced lung function.
 If a person hypoventilates, the body's carbon dioxide level
rises, which results in too little oxygen in the blood.
CO2
O2
Shea SA, White DP. Disorders of ventilatory control. In: Goldman L, Ausiello D, eds. Cecil Medicine. 23rd ed. Philadelphia, Pa: Saunders; 2007:chap 86
http://www.nlm.nih.gov/medlineplus/ency/article/002377.htm
http://www.southdartmoor.devon.sch.uk/pe/1127505986415.internal_lungs260.gif
What is Hypoventilation (PSG)?
CO2 >44 & O2 < 88
http://img.medscape.com/fullsize/migrated/491/438/sin491438.fig2.gif
Why Does Hypoventilation Occur?
 Benzodiazepines may depress the arousal response
to hypoxia and hypercapnia during sleep and
reduce genioglossal muscle tone
 So what does that mean for an OSA patient?
Hedemark LL, Kronenberg RS. Flurazepam attenuates the arousal response to co2 during sleep in normal subjects. Am Rev Respir Dis 1983;128:980-3.
Normal System for an OSA patient
Airway Collapse
↑ Carbon Dioxide
Levels
↓ Oxygen levels
Arousal
System with Benzodiazepines
Airway
Collapse
+
↓ Oxygen levels
Benzos
Arousal
↑ Carbon Dioxide
Levels
Hypnotics and The Airway
http://content.revolutionhealth.com/contentimages/n1573.jpg
Diazepam Injection in a Cat
150
100
(% control)
Peak Integrated activity
Phrenic Nerve Output
Hypoglossal Nerve Output
50
Diazepam Injection
0
0 5
15
30
Minutes after injection
60
Sanders MH. In: Principles and Practice of Sleep Medicine. Philadelphia: W.B. Saunders Company, 1994.
Benzos and the Airway
 Benzos may reduce genioglossal muscle tone
http://www.pwsdots.org/uploads/osa.jpg
Summary: Pathophysiology of OSA
Awake: Small airway + neuromuscular compensation
Loss of neuromuscular
compensation
Sleep Onset
+
Decreased pharyngeal
muscle activity
Airway collapses
Hyperventilate:
connect hypoxia &
hypercapnia
Airway opens
Benzos
Pharyngeal muscle
activity restored
Apnea
Hypoxia &
Hypercapnia
Increased
ventilatory effort
Arousal from sleep
Alcohol, Apnea, and the Airway
Alcohol as a Sleep Aid?
• Alcohol is also used as a sleep aid
• 28% of insomniacs indicated that they had used alcohol to help
them fall asleep
• Occasional insomniacs used alcohol for an average of 3.6
nights/month
• Chronic insomniacs used alcohol for an average of 6.8
nights/month.
• An equal number of occasional insomniacs and chronic
insomniacs (67%) described alcohol as an effective or very
effective method to induce sleep.
Ancoli-Israel S, Roth T. Sleep. 1999;22(suppl 2):S347-S353.
Risk Factor: Alcohol
Before Alcohol
Phrenic
Hypoglossal
Blood Alcohol = 83 mg/dl
Phrenic
Hypoglossal
Blood Alcohol = 134 mg/dl
Phrenic
Hypoglossal
Bonara M et al. Am Rev Respir Dis 1984;130 © American Lung Association.
A Fun Study? (i.e., 1982 Studies)
 Evaluating the effect of alcohol on sleep-disordered
breathing.
 On the night after the control study, (6.00-9.00 pm) the
subject drank wine or beer under supervision, to an amount
equivalent to the maximum he would drink on social
occasions.
Issa and Sullivan, JNNP 1982
http://www.brainandspinalcord.org/blog/wp-content/uploads/2009/10/alcohol.jpg
OSA Patient, Control Night; ex: Hour 1
Diaphragmatic EMG
Issa and Sullivan, JNNP 1982
OSA Patient, Alcohol night; ex. Hr 1
Prolonged
Apneas and
worse Oxygen
Desaturations
with Alcohol
Issa and Sullivan, JNNP 1982
OSA Patient
A
B
 A) OSA patient, Control night
 B) Same patient, Alcohol night
Issa and Sullivan, JNNP 1982
What about non-OSA?: A Snoring Patient
A
B
 A) Snoring patient, Control night
 B) Same patient, Alcohol night
Issa and Sullivan, JNNP 1982
Results of this Alcohol Study
 In all 7 pts. studied, alcohol exacerbated the sleep-induced breathing
abnormalities, and variably caused worsening of SaO2 in sleep.
 The effects of alcohol were:


Dose related
Occurred during the first 1-2 hr of sleep following alcohol intake.
 The finding was that alcohol intake can induce OSA in subjects with
“benign chronic snoring”.
 1) Alcohol clearly increased the duration of apneic episodes
 2) It promotes upper airway occlusion during sleep
Issa and Sullivan, JNNP 1982
Does Alcohol Cause Other Problems?
 Healthy Elderly Subjects
 0.6mg/kg ethyl alcohol (whiskey) 1h before bed
 In case of alcohol:
 Those patients with initial AHI of 5-10 -> increased apneas
 1 patient had increased PVCs with apneas
Guilleminault et al, Journal of Gerontology 1984 39(6):655-661
Alcohol and CPAP
 Ten obese male subjects undergoing CPAP
 1st night – CPAP titration
 2nd night – Control night at correct PAP pressure

3rd night - subjects ingested either 1.5 (A) or 2 (B) ml/kg of 50 percent
ethanol (100 proof vodka) over one half-hour starting 1 h before bedtime.
If using CPAP  No Change
with Alcohol
Chest 1991: 99:339-43)
Review: Effects of Alcohol
 The genioglossus and geniohyoid muscles undergo a decrease in
tone during REM sleep
 Alcohol, which depresses the CNS, significantly decreases the
activity of the genioglossus muscle during sleep and may be a factor
in snoring (partial obstruction of the upper airway) or OSA
(complete upper airway obstruction).
 However, alcohol with PAP in place does not have as much of an
effect on AHI
Guilleminault, The American Journal of Medicine Volume 88 (suppl 3A); March 2, 1990
So, What Happens with Benzos and OSA?
http://www.americasleeps.com/_borders/Snoring1.jpg
Effect of 30 mg of Flurazepam, 1982
 Double-blind, placebo-controlled, randomized study
 20 patient, 17 men and 3 women
 In controls, SDB minimally worse with 30 mg Flurazepam
FLURAZEPAM AND NOCTURNAL OXYGEN DESATURATION-DOLLY AND BLOCK, AJM 1982
Flurazepam in Controls
 In these patients, flurazepam had the same effect as
alcohol, i.e., a complete airway obstruction was noted
compared with baseline.
http://www.21stcenturydental.com/smith/sleepapena_tapappliance.htm
Berry, 1995: Triazolam and OSA
 Assessment of the effect of triazolam (0.25 mg) on apnea
duration and the arousal response to airway occlusion
during sleep in patients with severe OSA.
 12 male subjects were studied on two nights
 Mean age of 46.6 +/- 14.1 yr
 Mean weight of 260.8 +/- 55.9 lb
 They ingested triazolam (0.25 mg) or placebo 0.5h before
bedtime in a randomized double-blind crossover manner.
Berry RB, Kouchi K, Bower J, Prosise G, Light RW Am J Respir Crit Care Med. 1995 Feb;151(2 Pt 1):450-4.
Berry, 1995: Triazolam and OSA
 In non-rapid-eye-movement (NREM) sleep
 Mean duration of event was slightly increased with drug:
 Seconds:

Mean nadir in SaO2 lower on drug nights
%

26.8 vs 23.8, p < 0.02
Saturation: 80.1 vs 84.2, p < 0.001
In NREM sleep, the deflections in esophageal pressure prior to apnea
termination were higher on triazolam nights
 Pes:
53.3 vs 44.5 cm H2O, p < 0.001
 Triazolam increases the arousal threshold to airway
occlusion

This results in only modest prolongation of event duration and
increased desaturation at a dose of 0.25 mg in a group of OSA pts.
Berry RB, Kouchi K, Bower J, Prosise G, Light RW Am J Respir Crit Care Med. 1995 Feb;151(2 Pt 1):450-4.
Nitrazepam in OSA patients
 14 consecutive patients (12 males and 2 females), found to have mild
to moderate OSA (60–180 apneas/6 h of self-reported sleep time)
 The principal finding of this study was that NIT had no consistent
effect on the severity of sleep-disordered breathing in patients with
mild to moderate SA.
Eur Respir J, 1994,
Berry, 1992: Triazolam and Arousals
 6 men, mean age 28.1 +/- 7.1 yr , had their arousal response tested
by occluding a mask covering the nose with the mouth sealed.
 They ingested triazolam (0.25 mg) or placebo one-half hour before
bedtime in a randomized double-blind crossover manner.
 Mask occlusion was performed 1-4 h after triazolam/placebo
ingestion while the subjects breathed air /O2 mix -> SaO2 of 98%.
 Results
The time to arousal was significantly longer on triazolam nights
(32.0 +/- 5.2 s versus 22.6 +/- 3.2 s, p < 0.01).
 Conclusion: triazolam prolongs the time to arousal following
airway occlusion by increasing the arousal threshold.

Berry RB, McCasland CR, Light RW. Am Rev Respir Dis. 1992 Nov;146(5 Pt 1):1256-60.
Newer Hypnotics: Are They Better?
http://www.nytimes.com/2004/11/14/business/yourmoney/14drug.html?_r=1
Kryger, 2007: Ramelteon and OSA
 Ramelteon is a selective MT(1)/MT(2)-receptor agonist indicated for
insomnia
 Double-blind, randomized, crossover study
 26 adults with mild to moderate OSA received ramelteon 16 mg and
placebo for one night each, administered 30 min before habitual bedtime.
 AHI was similar: 11.4 vs 11.1, P = 0.812

Ramelteon – no effect on # of central, obstructive, or mixed apneas.
 No significant differences were observed in SaO(2) for the entire night

(95.1 vs 94.7%); P = 0.070
 Ramelteon did not statistically affect sleep when evaluated by
polysomnography and post-sleep questionnaire.
Kryger M, Wang-Weigand S, Roth T. Sleep Breath. 2007 Sep;11(3):159-64..
Rosenberg, 2007: Eszopiclone and OSA
 This double-blind, randomized crossover study
 Patients (35–64 yrs) with mild-to-moderate OSAS [AHI 10-40].
 Patients received eszopiclone 3 mg or placebo for two consecutive nights
 Results
 Mean total AHI, was similar to placebo - 16.5 (plac) and 16.7 (esz)

No significant differences in respiratory arousals, duration of respiratory
episodes, or oxygen saturation were noted.

Significant differences in:
Sleep efficiency (85.1% and 88.4%)
Wake time after sleep onset (61.8 and 48.1 min)
Wake time during sleep (55.9 and 43.2 min).



Rosenberg R, Roach JM, Scharf M, Amato DA. Sleep Med. 2007 Aug;8(5):464-70. 2007
Zaleplon and OSA on PAP
 Placebo controlled cross-over design: 15 mild to moderate OSA patients for the
presence of worsening apnea with home-monitoring
 Administering zaleplon (10 mg) or Placebo over a period of five consecutive
nights then cross over
 Results: No statistically significant treatment differences between zaleplon and
placebo were observed
 AHI (ZN=7.2 vs PL=7.5; p=0.602)
 Mean SpO2 (ZN=94.6 vs PL=94.7; p=0.859).
 Small difference: ZN (79.2±1.3) and PL (82.1±0.9) for nadir SpO2 (p=0.008).
 These data support the hypothesis that short-acting, non-benzodiazepines may
be used safely in middle-aged patients with mild to moderate OSA while
receiving CPAP therapy in the home environment.
Coyle et al. JCSM 2005
Effect of Zolpidem vs. Flurazepam
 In this 1988 study, Zolpidem was associated with slightly higher AHI
and lower oxygen saturations than Flurazepam or placebo.
 However, the n was 12 patients
Pharmacol Biochem Behav. 1988
Zolpidem and OSA on PAP
 Obese adult patients who had been undergoing treatment
of severe OSA (AHI > 30/hour) with CPAP therapy for
least 6 months. All patients were compliant with their
CPAP therapy
 14 men and 2 women
 3 nights: Titration, and then the patient slept in the lab
with PAP and one night of placebo and one night of
zolpidem 10 mg in a randomized order
Berry, Sleep 2006
Berry, Sleep 2006: Respiration
There was no significant effect of zolpidem on any respiratory variable
Berry, Sleep 2006: AHI
The AHI overall, AHI during
REM sleep, and the AHI
during supine sleep did not
differ between placebo and
zolpidem nights.
In summary, in a study of 16
patients with severe OSA,
there was no significant
worsening in the AHI or in
any index of arterial oxygen
desaturation during CPAP
treatment with the acute use
of zolpidem, 10 mg.
Patients Often Feel Trapped By PAP
Can Hypnotics Help This process?
http://uashome.alaska.edu/~jndfg20/website/youngfrankenstein.gif
Benefits to Sedatives in Sleep Apnea?
 Tolerance to CPAP can be problematic for patients
 Ranges of 50-70% of continued use over time
 The only consistently reliable predictor of long-term
adherence has been the use of CPAP during the initial
treatment period
 Long term adherence patterns may be determined within
the first few days of therapy.
 Therefore, strategies aimed at improving adherence with
therapy should focus on the initial experience with CPAP
Methods to Improve CPAP tolerance
 Adjustment through continued use
 For those experiencing difficulty:
 Early
Education
 Alterations in mask
 Changing pressures / type of pressure delivery
 Humidification
 Sedatives?
Can Hypnotics Help with PAP?
 So, CPAP is hard to use, particularly if poorly
tolerated early.
 Can the addition of a hypnotic medication help
improve PAP tolerance?
Bradshaw, Chest 2006
 Evaluation of Zolpidem 10mg vs. Placebo during
initiation of PAP
 Will it help improve PAP use over 28 days?
Bradshaw, Chest 2006: Study Flow
CHEST 2006; 130:1369–1376
Bradshaw, Chest 2006: Results in CPAP
CHEST 2006; 130:1369–1376
Use of Eszopiclone in Polysomnography
 Prospective, double-blinded, randomized, placebo-
controlled trial assessing the effect of eszopiclone 3 mg on
the quality of polysomnography
 3 study arms: diagnostic polysomnography, split-night
polysomnography, and CPAP titration polysomnography

79 diagnostic studies, 67 split-night studies, and 80 CPAP titration
studies
 Enrolled 226 subjects: 113 received eszopiclone and 113
received placebo
Letteri, Sleep 2008
Letteri, Sleep 2008
 Non-usable polysomnograms were defined as studies with less than
120 min of total sleep time (does not meet criteria for a diagnostic
study) or complete CPAP intolerance.
 Poor quality polysomnograms were defined as studies with less than
120 min of TST, sleep efficiency less than or equal to 70%, or an
incomplete CPAP titration.
 Defined incomplete CPAP titrations as those with a residual AHI ≥ 5
on the highest level of CPAP achieved, or complete CPAP intolerance.
 CPAP intolerance was defined as the patient’s complete inability to
sleep on CPAP or their request to end the study prematurely due to
CPAP discomfort.
Letteri, Sleep 2008
Individuals were not studied
both with and without this
agent. Therefore, the study
does not directly address the
question of whether or not
eszopiclone has these effects
in individual patients.
Letteri, Sleep 2008
Letteri, Annals Int Med 2009
 Evaluation of Eszopiclone to improve CPAP use for 14
days with open label afterward
Letteri, Annals Int Med 2009: AEs
Letteri, Annals Int Med 2009: Effect on PAP
Finally, One Last Point
 We’ve reviewed that data that hypnotics may help
with PAP use, but…
 Can “hypnotic” medications actually treat OSA?
Sodium Oxybate and OSA
 OSAS pts. (n=48) off treatment received 2-week SXB or
placebo (PBO) treatment with PSG at baseline and day 14.
 SXB led to a reduction in mean AHI with SXB and
significantly increased slow wave sleep duration (5.2± 25.0
min vs. 29.4±37.0 min; p=0.0038).
George et al., Sleep Breathing Jan 2010
Summary
 Data is MIXED
 Data suggests that alcohol and benzodiazepines
may worsen sleep disordered breathing in patients
who already are at risk for it
 Some studies demonstrate that use of NBZRAs
may be useful in CPAP titration and early home
PAP use