Chronic Opioid Use is a Risk Factor for the Development of Central Sleep Apnea and Ataxic Breathing

Abstract

Background:

Chronic opioid therapy for pain management has increased dramatically without adequate study of potential deleterious effects on breathing during sleep.

Methods:

A retrospective cohort study comparing 60 patients taking chronic opioids matched for age, sex, and body mass index with 60 patients not taking opioids was conducted to determine the effect of morphine dose equivalent on breathing patterns during sleep.

Results:

The apnea-hypopnea index was greater in the opioid group (43.5/h vs 30.2/h, p < .05) due to increased central apneas (12.8/h vs 2.1/h; p < .001). Arterial oxygen saturation (SpO₂) in the opioid group was significantly lower during both wakefulness (difference 2.1%, p < .001) and non-rapid eye movement (NREM) sleep (difference 2.2%, p < .001) but not during rapid eye movement (REM) sleep (difference 1.2%) than in the nonopioid group. Within the opioid group, and after controlling for body mass index, age, and sex, there was a dose-response relationship between morphine dose equivalent and apnea-hypopnea (p < .001), obstructive apnea (p < .001), hypopnea (p < .001), and central apnea indexes (p < .001). Body mass index was inversely related to apnea-hypopnea index severity in the opioid group. Ataxic or irregular breathing during NREM sleep was also more prevalent in patients who chronically used opioids (70% vs 5.0%, p < .001) and more frequent (92%) at a morphine dose equivalent of 200 mg or higher (odds ratio = 15.4, p = .017).

Conclusions:

There is a dose-dependent relationship between chronic opioid use and the development of a peculiar pattern of respiration consisting of central sleep apneas and ataxic breathing. Although potentially significant, the clinical relevance of these observations remains to be established.

Citation:

Walker JM; Farney RJ; Rhondeau SM; Boyle KM; Cloward TV; Shilling KC. Chronic opioid use is a risk factor for the development of central sleep apnea and ataxic breathing. J Clin Sleep Med 2007;3(5):455-461.

Over the last decade, there has been a dramatic change in the way chronic pain has been managed. Prior to the 1990s, opioids were used sparingly and generally reserved for use in patients with severe advanced forms of chronic pain.¹ However, without the benefit of prospective randomized studies,^2,3 the American Academy of Pain Medicine and the American Pain Society issued a joint position statement in 1997 stating, “It is now accepted by practitioners of the specialty of pain medicine that respiratory depression induced by opioids tends to be a short-lived phenomenon, generally occurs only in the opioid-naive patient, and is antagonized by pain. Therefore, withholding the appropriate use of opioids from a patient who is experiencing pain on the basis of respiratory concerns is unwarranted.”⁴ This declaration coincided with 3 seminal events—the requirement by the Joint Commission for hospitals to be more aggressive and comprehensive in pain management, the renewed and expanded use of methadone in chronic pain, and the introduction of new sustained-release opioids. Over the following 10 years, opioid use in the United States escalated on an unprecedented scale.^5,6 Retail distribution of 2 opioids, methadone and oxycodone, increased 824% and 660%, respectively, between 1997 and 2003.⁷

A central tenet of the above recommendations rested upon the belief that long-term use of opioids led to tolerance and diminished side effects of sedation and, more importantly, respiratory depression.⁶ Even though that might be the case for wakefulness, there is little scientific basis that this is also true during sleep. We have recently reported distinct respiratory patterns during sleep in a small number of patients receiving chronic opioid therapy for pain management.⁸ In addition, central sleep apnea has been shown in a small pilot study (6 out of 10 patients)⁹ and more recently in 30% of a larger group of patients (n = 50) in stable-dose methadone maintenance treatment programs.¹⁰

Because we have consistently observed certain types of distinctive breathing patterns in those patients in our clinic population who have been taking opioid medications chronically, we conducted a retrospective cohort study of patients referred for evaluation of sleep apnea who were on chronic opioid therapy, as compared with patients not using opioids. We hypothesized that chronic opioid use was an independent risk factor for sleep-disordered breathing in patients referred for assessment of sleep apnea.

METHODS

Subjects

Subjects were referred to the Intermountain Sleep Disorders Center at LDS Hospital, Salt Lake City, Utah (elevation 1500 m) for diagnostic testing of suspected sleep apnea. Out of 1762 diagnostic studies conducted between December 2002 and March 2005, 104 consecutive patients taking opioid medications for chronic pain were identified by chart review. Of those, 60 qualified for study participation. Exclusion criteria, based on interview or questionnaire information, were age younger than 18, congestive heart failure, coronary artery disease, stroke, primary neurologic disease, prior use of nocturnal oxygen, and inadequate opioid dosage information or less than 6 months of opioid use. Questionnaires consisted of the Epworth Sleepiness Scale and a general sleep survey that included symptoms of sleep apnea, type, dose, and frequency of medications. Oral morphine dose equivalent in milligrams was computed¹¹ based upon chronic use and no crossover tolerance. The morphine dose equivalent for methadone and fentanyl was calculated based upon more conservative literature recommendations.^12,13 Opioid subjects were matched for sex, age (± 5 years), and body mass index (BMI) (± 5 kg/m²) with 60 consecutive subjects who were not using opioids, using the same exclusion criteria. Comorbidities of sleep apnea and medications used in major drug classifications were determined from patient questionnaire and physician interview. The Institutional Review Board of the hospital approved the study protocol and waived the patient-consent requirement.

Polysomnography

Attended 17-channel polysomnography (Cadwell Laboratories, Inc, Kenniwick, Wash) was performed, consisting of central (C3/A2 and C4/A1) and occipital (O1/A2 and O2/A1) electroencephalogram, right and left electrooculogram, and submentalis electromyogram. Airflow was detected by air pressure transducers, and respiratory effort was determined by measurement of chest and abdomen motion with piezoelectric transducers. Arterial oxygen saturation (SpO₂) was measured by an internal Cadwell oximeter set in the 4-second averaging mode. Sleep was manually scored using standard criteria.¹⁴ Apneas were scored on the basis of absence of airflow for 10 or more seconds. Obstructive apneas were defined by the presence of respiratory effort, and central apneas by the absence of respiratory effort. Hypopneas were defined by reduction in airflow for 10 seconds or more associated with at least a 3% decrease in SpO₂.¹⁵ Apnea-hypopnea as well as obstructive apnea, central apnea, and hypopnea indexes were computed as the total of respiratory events divided by the total sleep time, in hours. These same indexes were also computed for non-rapid eye movement (NREM) and rapid eye movement (REM) sleep. Sleep and respiratory scoring was done blindly, without knowledge of the hypothesis or condition. Similarly, respiratory patterns in the 2 groups were jointly evaluated for ataxic breathing or Biot's respiration in NREM sleep by 2 diplomates in sleep medicine (JMW and RJF) without knowledge of condition. Specifically, ataxic breathing, as first described by Biot¹⁶ in 1876, is characterized by an irregular respiratory rate, rhythm, and depth with or without brief respiratory pauses less than 10 seconds, or a repeating pattern of several breaths. The irregular respiratory patterns characteristic during REM sleep were excluded from consideration.

Statistical Analysis

The primary outcome measure was the apnea-hypopnea index (AHI), whereas secondary outcome measures were the components of the AHI; obstructive apnea, central apnea, and hypopnea indexes by sleep time or states (NREM or REM sleep), mean SpO₂ (awake, NREM, and REM sleep), and the prevalence of ataxic breathing during sleep. Paired-samples t-tests were performed to test the null hypothesis of no difference between the groups for the various AHI and SpO₂ measures. The Fisher exact test was used to test for differences in the prevalence of ataxic breathing between the 2 groups. The 2 groups were considered significantly (2-tailed test) different in the primary outcome (apnea-hypopneas index) and the prevalence of ataxic breathing if p values were less than .05. Significant difference was concluded for the components of the AHI if p values were less than .017 (a Bonferroni-corrected p-value=.05/3) to control for probability of a Type I error. For similar reasons, significant differences were concluded for by sleep stages if p values were less than .025 and, for SpO₂, if p values were less than .017.

Poisson regression analysis was used for variables of apnea-hypopnea, hypopneas, and obstructive and central apneas, with the outcome variable as the dependent variable and age, sex, and BMI as independent variables. The morphine dose equivalent was an additional independent variable used in the chronic opioid-use group. Continuous independent factors that were not linearly related to the log rate were categorized as appropriate. Rate ratios compared to reference values, along with 95% confidence intervals, are reported.

Presence of ataxic breathing in the chronic opioid-use group was modeled using backward stepwise multivariable logistic regression. The presence of ataxic breathing was the dependent factor, and age, sex, BMI, and morphine dose equivalent were independent factors. Continuous independent factors were categorized only if they were not linearly related to the log odds. Odds ratios, along with 95% confidence intervals, are reported.

RESULTS

Of the 60 patients taking opioids, only 2% indicated that the primary purpose of use was for management of malignant pain (cancer). In the remaining 98% of patients, the primary reason for opioid therapy was for nonmalignant pain classified as orthopedic (55%) or neuropathic (15%), pain associated with fibromyalgia (20%), and other (8%). Opioid medications used are shown in Table 1. Mean morphine dose equivalent was 143.9 mg, median was 79.4 mg, and 25^th and 75^th percentiles were, 30.0 mg and 213.8 mg, respectively. The range was 7.5 to 750 mg. Patients had been taking opioid medication for a minimum of 6 months and for as long as 15 years (mean 3.2 years). In terms of nonopioid medications, a significantly higher proportion of those subjects who were chronically taking opioids, as compared with those not on opioids, were also taking benzodiazepines (25% vs 7%, p < .05), zolpidem (20% vs 3%, p < .01), and gabapentin (28% vs 0.5%, p < .001). Conversely, those not taking opioids had a higher proportion using ibuprofen (17% vs 2%, p < .01) and statins (7% vs 22%, p < .05). Both groups were similar in symptoms of Epworth Sleepiness Scale scores and prevalence of comorbidities associated with sleep apnea (Table 2).

Table 1.

Number of Patients Using Opioid Medications and Doses^a

Opioid Medication	No.	Dose	Morphine Equivalent
Hydrocodone	26	22.0 (5–60)	22.0 (5–60)
Methadone	20	37.5 (10–120	140.6 (37.5–450)
Oxycodone	24	82.1 (10–240)	123.1 (15–360)
Morphine	3	230.0 (30–600)	230.0 (30–600)
Fentanyl transdermal	6	64.2 (50–100)	143.3 (120–240)
Propoxyphene	1	100.0	30.0

^aDoses are presented as mean (range) per day in milligrams.

Table 2.

Comorbid Conditions With Sleep Apnea in the Opioid and Control Groups

Comorbid Condition	Opioid Group	Control Group
Hypertension	32	35
Arrhythmias	17	13
GERD	28	28
Hypothyroidism	22	20
Depression	52	43
Diabetes	17	12

Data are presented as percentage. GERD refers to gastroesophageal reflux disorder.

Descriptive statistics, comparisons of baseline characteristics, and outcome measures for both sleep-disordered breathing groups are displayed in Table 3. The BMI of the group not using opioids was slightly higher than that of the group chronically using opioids; but, the difference was extremely small. Both groups exhibited sleep apnea; however, the AHI (43.5/h vs 30.2/h) and central apnea index (12.8/h vs 2.1/h) was significantly higher in the chronic opioid users than in those not using opioids (p = .027 and .001, respectively). The significant difference in the AHI is a reflection of inclusion of central apneas, since neither obstructive apneas nor hypopneas differed between groups. It is interesting to note that the AHI increased during NREM sleep (p = .024) but not during REM sleep (p = .860). Because there was a higher proportion of patients on hypnotics in the opioid group, a comparison of respiratory events was conducted eliminating patients taking opioids who were also taking zolpidem or benzodiazepines. Patients in the opioid group still had a significantly greater AHI, as compared with controls (50.9/h vs 25.6/h, p < .01), indicating that hypnotics did not appear to increase the risk of apneas and hypopneas.

Table 3.

Comparison of Outcome Measures and Baseline Characteristics of Patients Chronically Using Opioids With Sleep-Disordered Breathing and Patients Not Using Opioids

Demographic Factors	No.	Chronically Using Opioids	Not Using Opioids	p Value
Age	60	52.7 ± 13.1	52.9 ± 13.0	0.551
Women	60	40 (66.7%)	40 (66.7%)	1.00
Body Mass Index, kg/m2	60	31.8 ± 7.7	32.5 ± 7.6	0.12
Epworth Sleepiness Scale, score	45a	11.6 ± 5.6	10.2 ± 5.8	0.263
Respiratory Outcomes
Obstructive Apnea Index, no./h	60	16.8 ± 24.0	14.2 ± 18.8	0.508
Central Apnea Index, no./h	60	12.8 ± 22.4	2.1 ± 4.1	<0.01
Hypopnea Index, no./h	60	15.3 ± 14.1	13.7 ± 13.0	0.466
Apnea-Hypopnea Index, no./h	60b	43.5 ± 35.2	30.2 ± 26.2	0.27
Total
During NREM sleep	60	43.6 ± 38.8	28.6 ± 27.2	0.24
During REM sleep	43c	37.4 ± 24.5	37.6 ± 30.4	0.977
Mean SaO2, %
During wake	60	91.0 ± 2.6	93.1 ± 1.8	<0.01
During NREM sleep	60	89.7 ± 3.3	91.9 ± 2.0	<0.01
During REM sleep	43	89.3 ± 7.0	90.5 ± 3.9	0.259
Sleep Outcomes
Recording time, min	50d	451.0 ± 42. 2	456.0 ± 36.1	0.510
Sleep time (min)	50	336.0 ± 78.5	348.5 ± 77.5	0.351
Sleep stage, %
1	50	8.0 ± 11.7	19.2 ±11.6	<0.01
2	50	78.0 ± 10.8	64.6 ±13.5	<0.01
3	50	1.2 ± 3.0	1.4 ±3.2	0.783
4	50	0.1 ± 0.6	0.2 ± 0.6	0.626
REM	50	12.0 ± 8.5	13.3 ±7.0	0.399

Data are presented as mean ± SD, except for women, which are presented as number (%). NREM refers to non-rapid eye movement.

^aApneas and hypopneas were calculated for all 60 pairs of patients, but continuous positive airway pressure (CPAP) or oxygen therapy was implemented partway through the study in 8 patients in the opioid group and 3 in the control group, and only the room-air portion of study was included in the calculation.

^bNot all patients completed the Epworth Sleepiness Scale.

^cNot all patient pairs exhibited rapid eye movement (REM) sleep when breathing room air.

^dSleep measures were not included in the analysis for patient pairs in which 1 or both received CPAP or oxygen partway through the night.

After adjusting for age, sex, and BMI, an increase in morphine dose equivalent was strongly associated with an increasing number of apneas, hypopneas, or both (all p values <.001), with the exception of AHI during REM sleep (Table 4). Figure 1 shows the rate ratio for increasing morphine dose equivalent for obstructive apneas, central apneas, and hypopneas during sleep, as well as during NREM and REM sleep.

Table 4.

Poisson regression analyses for apnea-hypopnea, obstructive apnea, hypopnea, and central apnea indices.

CHRONIC OPIOID	AHI (n 60)	OAI (n 60)	HI (n 60)	CAI (n 60)	NREM AHI (n 60)	REM AHI (n 47)
Morphine dose equivalent (100 mg increase+)	1.2 (1.1–1.2)***	1.1 (1.1–1.2)***	1.2 (1.2–1.3)***	1.3 (1.3–1.4)***	1.1 (1.1–1.2)***	1.0 (0.9–1.0)
Age < 40 years (vs 50–84 years+)	0.7 (0.6–0.8)***	0.2 (0.1–0.3)***	0.9 (0.7–1.1)	1.1 (0.9–1.3)	0.7 (0.7–0.9)***	0.9 (0.8–1.1)
Age 40–49 years (vs 50–84 years+)	1.1 (1.0–1.3)*	0.9 (0.7–1.1)	1.3 (1.1–1.5)**	1.4 (1.1–1.8)**	1.4 (1.3–1.5)***	1.2 (1.1–1.4)**
Gender= Male (vs Female+)	1.9 (1.7–2.1)***	2.4 (2.0–2.7)***	0.9 (0.8–1.1)	2.6 (2.2–3.1)***	1.9 (1.7–2.1)***	1.2 (1.1–1.3)**
Body mass index 16–28 (vs32–55+)	1.4 (1.3–1.6)***	2.0 (1.7–2.3)***	0.6 (0.5–0.7)***	3.7 (2.9–4.7)***	1.4 (1.3–1.5)***	0.8 (0.7–0.9)***
Body mass index 29–31 (vs32–55+)	1.1 (0.9–1.2)	1.0 (0.8–1.2)	0.6 (0.5–0.7)***	2.6 (2.0–3.4)***	1.0 (0.9–1.1)	0.9 (0.8–1.0)
NON-OPIOID	AHI (n 60)	OAI (n 60)	HI (n 60)	CAI (n 60)	NREM AHI (n 60)	REM AHI (n 57)
Age < 40 years (vs 50–84 years+)	0.5 (0.4–0.6)***	0.3 (0.2–0.4)***	0.8 (0.6–1.0)	0.4 (0.2–0.8)*	0.5 (0.4–0.6)***	0.7 (0.6–0.8)***
Age 40-49 years (vs 50–84 years+)	1.1 (0.9–1.2)	0.7 (0.6–0.9)***	1.5 (1.3–1.8)***	0.7 (0.4–1.1)	1.1 (1.0–1.3)	1.2 (1.0–1.3)**
Gender = Male ( vs Female+)	1.8 (1.6–2.0)***	2.5 (2.2–2.9)***	1.2 (1.0–1.3)	2.3 (1.6–3.3)***	2.0 (1.8–2.2)***	1.2 (1.1–1.3)***
Body mass index 16–28 (vs32–55+)	0.5 (0.5–0.6)***	0.5 (0.4–0.6)***	0.4 (0.4–0.5)***	1.6 (1.1–2.3)*	0.5 (0.5–0.6)***	0.6 (0.5–0.6)***
Body mass index 29–31 (vs32–55+)	0.5 (0.5–0.6)***	0.6 (0.5–0.7)***	0.5 (0.4–0.6)***	0.7 (0.4–1.2)	0.6 (0.5–0.6)***	0.5 (0.4–0.5)***

Data are presented as rate ratios (95% confidence intervals). AHI refers to apnea-hypopnea index; OAI refers to obstructive apnea index; HI refers to hypopnea index; CAI refers to central apnea index; NREM AHI refers to apnea-hypopnea index during non-rapid eye movement sleep; REM AHI refers to apnea-hypopnea index during rapid eye movement sleep. In the opioid group, comparisons are made to reference values adjusted for morphine dose equivalent, age, sex, and body mass index. In the nonopioid group, comparisons are made to reference values adjusted for age, sex, and body mass index.

+Reference value

* p < .05

**p < .01

***p < .001

Figure 1.

Rate ratio for morphine dose equivalent and obstructive apnea, central apnea, and hypopnea indexes; all adjusted for weight, sex, and age. REM refers to rapid eye movement sleep; NREM, non-rapid eye movement sleep.

Increasing age appears to be a risk factor associated with increasing apneas or hypopneas in both groups. Men tended to have higher apnea indexes than women in both groups. However, hypopneas were similar for men and women in both groups. Patients taking opioids and in the lowest weight category had the highest rate of apneas and hypopneas (p < .001), whereas those in the same category not taking opioids had the lowest rate (p < .001), compared with the reference group with a BMI of 32 to 55 kg/m² (Figure 2).

Figure 2.

Rate ratios from the Poisson regression analyses for apnea-hypopnea index for body mass index (kg/m²) (BMI) adjusted for age and sex, as well as morphine dose equivalent (mg), in the opioid group.

Ataxic breathing, for which we assessed only during NREM sleep, since REM sleep is characterized by an erratic breathing pattern, was more prevalent in patients who used chronic opioids than in those not on chronic opioids (70.0% vs 5.0%, p < .001), and there was a significant dose response. A morphine dose equivalent of 200 mg or higher was found to be associated with the presence of ataxic breathing in patients with sleep-disordered breathing who used chronic opioids (odds ratio = 15.4, 95% confidence interval = 1.6 to 145.5, p=.017). The proportion of patients with ataxic breathing patterns was 92% of those with a morphine dose equivalent of 200 mg or higher, compared to 61% in those with morphine dose equivalent less than 200 mg.

An example of ataxic breathing in a patient chronically taking opioids is shown in Figure 3 and compared to that of an age- and weight-matched control patient. Note the variation in depth and rate, as well as frequent respiratory pauses, in the patient on opioids, compared with the stable breathing in the control patient. Because of to our elevation, we have considered the possibility that ataxic breathing could be facilitated by hypoxia, but we have not observed resolution of ataxic breathing in our clinic patients who were receiving supplemental oxygen. There was an insufficient number of cases in this study to systematically evaluate the effect of supplemental oxygen on breathing. However, Figure 4 from a patient from this study shows that this type of breathing pattern seems to persist with oxygen administration.

Figure 3.

A 120-second period from a 32-year-old female (A) opioid patient, body mass index 22 kg/m², morphine dose equivalent 375 mg, and apnea-hypopnea index 67/hour (A) compared with a 32-year-old female (B) control subject with a body mass index 23 kg/m² and apnea-hypopnea index 5/hour.

Figure 4.

Respiratory patterns during 300 seconds of non-rapid eye movement (NREM) stage 2 sleep in a 32-year-old woman with a body mass index of 22 kg/m² and a morphine dose equivalent of 375 mg, breathing room air and 2 L/min oxygen. Note the ataxic respiration and brief respiratory pauses lasting less than 10 seconds under both breathing conditions.

The SpO₂ in the opioid group was significantly lower during both wakefulness (difference 2.1%, p < .001) and NREM sleep (difference 2.2%, p < .001) but not during REM sleep (difference 1.2%), as compared with the group not using opioids.

DISCUSSION

The importance of this study is that the chronic use of opioid medication (ie, longer than 6 months) predisposes patients to the development of an irregular breathing pattern with central apneas during sleep, best characterized as Biot's respiration, ataxic breathing, or both. It is distinct from periodic breathing with central apneas or Cheyne-Stokes respiration. The presence of these ventilatory patterns is rare in other patients seen in our clinic at this elevation and who have been referred for evaluation of sleep apnea but who have not been taking opioids.

In 1876, Biot¹⁶ observed the presence of an irregular breathing pattern with pauses, distinct from Cheyne-Stokes respiration, in a 16-year-old male patient with tuberculous meningitis. His description has been translated as follows, “This irregularity of the respiratory movements is not periodic, sometimes slow, sometimes rapid, sometimes superficial, sometimes deep, but without any constant relation of succession between the two types, with pauses following irregular intervals, preceded and often followed by a sigh more or less prolonged.”¹⁷ The irregularity of the respiratory pattern is reminiscent of the cardiac irregularity seen with atrial fibrillation. Although the term, Biot's respiration, has come to refer to a type of breathing pattern in which there are groups of similar sized breaths alternating with regular periods of central apnea, it is clear from inspection of Biot's original illustration (Figure 5) that he was actually describing what has also been referred to as ataxic breathing. Whether or not these 2 patterns are distinct is unclear because, in our experience, they can be seen together in the same patient. This type of breathing is clearly distinct however from Cheyne-Stokes breathing in which there is a very regular waxing and waning pattern of the tidal volume occurring with very obvious periodicity. Until we previously reported on irregular breathing and central apneas being associated with chronic opioid therapy,⁸ ataxic breathing or Biot's respiration was previously associated with primary diseases of the central nervous system such as meningitis and cerebral hemorrhage. Irregular nonperiodic breathing has also been reported in patients experiencing high-altitude pulmonary edema (2850 m), likely related to coexistent cerebral edema and increased intracranial pressure.¹⁸

Figure 5.

Adaptation of part 1 and 2 of a pneumograph of the original depiction of Biot's respiration (Contribution a 1'ètude de phènomène respiratoire de Cheyne-Stokes. Lyon Mèd; 1876;). Note respiratory variability in frequency and depth.

The ataxic breathing pattern; the occurrence of central apneas, particularly during NREM sleep when metabolic control systems are usually the most stable; and the inverse relationship between BMI and respiratory disturbances point to a fundamental alteration of respiratory control feedback systems. Ataxic respiration during NREM sleep is unusual and, to our knowledge, has only been reported in case studies.⁸ The presence of central apneas is also uncommon except in patients with cardiovascular or neurologic compromise.¹⁹ Finally, and contrary to previous findings,²⁰ when age and sex were controlled, patients with the highest rate of apneas and hypopneas in the opioid group were those with lower BMIs. In the opioid group, patients with a BMI of 16 to 28 had a 42% increase in rate of apneas and hypopneas, compared with patients with a BMI greater than 32 kg/m². However, the opposite is the case in the group that was not using opioids; those with a BMI of 16–28 kg/m² had a 47% reduction in the rate of apneas and hypopneas. Therefore, patients not taking opioids and in the lowest weight category have the least risk for sleep-disordered breathing, whereas those in the same weight category but who were taking opioids have the highest risk. All of these factors point to an elementary difference in the regulation of breathing in these patients on chronic opioids. Because of the retrospective nature of this study and the various confounding factors, the pathophysiologic mechanisms underlying these apparent unique breathing patterns cannot be determined from our data.

Based upon these data, it appears that chronic opioid use is a dose-related and independent risk factor for the development of irregular breathing with central apneas and hypopneas in this group of patients referred for the evaluation of sleep apnea. Ataxic respiration was apparent in 92% of those taking a morphine dose equivalent of 200 mg or higher, 61 % in those taking less than a 200-mg morphine dose equivalent, and 5% of patients not taking opioids. Apneas were also dose related. Each 100-mg morphine dose equivalent increased the rate of apneas by 14.4% and of central apneas by 29.2%, adjusted for factors of weight, age, and sex.

The long-held tenet that the chronic administration of opioids is safe because of the development of tolerance to respiratory-depressive effects of opioids may not be valid and is possibly explained by the shift in hypoxic ventilatory regulation with sleep. It has been well established that the normal sleep state has a significant effect on respiratory regulatory patterns, with alterations in sensitivity to hypercarbia, hypoxia, and respiratory rhythmicity.^21,22 One of these changes involves the contribution of peripheral carotid chemoreceptors in sensing hypoxia and augmenting respiratory drive in the presence of hypoxia during sleep.^21,22 Failure to detect adverse effects of chronic opioid therapy on respiratory control in previous studies is a consequence of testing subjects in the awake state during which the principal respiratory controller is neural or cerebral drive. However, during NREM sleep, the peripheral chemoreceptors in the carotid body are primarily responsible for the detection of airway occlusion, hypoxia, and breath-by-breath control of ventilatory pattern.^23,24 This raises a potential explanation for the destabilization of breathing patterns seen in our patients. Individuals on long-term opioids have been observed to develop tolerance to the central effects of opioids (eg, sedation, hypercapnic response, inspiratory loading) but not to their peripheral effects.^25,26 Chronic constipation is a well-described problem, and, regardless of the duration of administration, these patients experience reduced gastric motility and chronic constipation requiring therapeutic interventions.^27,28 Although it is not possible to exclude opioid effects on the central pattern generator, we suspect that sleep-disordered breathing develops in patients on chronic opioids by a similar peripheral effect. Mu-receptor inhibition of the carotid body and possibly other peripheral hypoxia-sensing chemoreceptors may modulate the peripheral signaling that influences central hypoxic ventilatory drive. However, a recent study showed that, during wakefulness, there is an elevated peripheral and a blunted central chemoreceptor response, as measured by hypoxic and hypercapnic responses, respectively.²⁹

There are several limitations of the present study in terms of generality of findings: First, bias may be introduced due to the retrospective analysis of data involving 2 groups of patients who may be fundamentally different or who were referred to a center for the evaluation of sleep apnea. The 2 groups were matched in the primary parameters (age, sex, and BMI). There were no differences in Epworth Sleepiness Score scores or comorbidities (Table 2). Opioid medications were used almost exclusively for control of nonmalignant pain and no subjects had associated conditions that would predictably alter respiratory control, such as congestive heart failure or neurologic disorders.¹⁹ We could not control the matching for medications; however, when we compared the groups after eliminating those subjects taking benzodiazepines, the difference in the AHI was greater (50.9/h vs 25.6/h; p < .01). With respect to patients referred for the evaluation of sleep apnea, our results are similar to those of 2 other studies^9,10 in which subjects were enrolled in a methadone maintenance treatment program and were not presenting for evaluation of sleep apnea. One study showed that 6 out of 10 participants in the program had central sleep apnea, whereas it was absent in 9 control subjects. The second study examined the respiration in 50 subjects undergoing methadone maintenance treatment, compared with matched controls. The subjects in the methadone maintenance treatment showed significantly higher AHIs, primarily due to central sleep apneas occurring during NREM sleep. This phenomenon was present in 30% of the methadone patients and none of the control patients. Therefore it is not surprising that the recent revision of International Classification of Diseases-9 (ICD-9) codes now has a category specifically identified as “Central Sleep Apnea due to Drug or Substance ICD-9 327.29.” ³⁰ Presumably this would encompass opioids and does not differentiate between short- and long-term use. Based upon the above considerations, it does not appear that our findings can be accounted for by disproportionate comorbidities or because patients had been referred for sleep apnea.

A second limitation is that patients receiving opioids are generally taking multiple medications with possible and unpredictable interactions. Similar to the present study in which the central apnea index was related to morphine dose equivalent in patients using opioids for chronic pain, Wang et al¹⁰ found that methadone blood concentration was the only statistically significant variable associated with central apnea in patients enrolled in a methadone maintenance program. Taken together, these 2 studies would suggest that the development of central apnea is related to the chronic administration of opioids, whether or not pain is present.

The present study was conducted at an elevation of 1500 m (barometric pressure 647 mm Hg), which can raise the question that hypoxia may have induced periodic breathing. We discount this possibility on the basis of the following arguments. Periodic breathing of high altitude typically occurs at an elevation of at least 2500 meters (barometric pressure 575 mm Hg). It is typically regular, is consistent with a Cheyne-Stokes-type pattern with relatively short cycle lengths of 12 to 34 seconds, and resolves with acclimatization except at very high elevations.^31,32 An irregular nonperiodic breathing pattern has also been reported in some individuals susceptible to high-altitude pulmonary edema at 2850 m,¹⁸ but this condition is likely a function of coexistent cerebral edema. In any case, periodic breathing during sleep at altitude occurs immediately after ascent, whereas high-altitude pulmonary edema and high-altitude cerebral edema usually occur within the first 2 to 4 days in subjects who have rapidly ascended to at least moderate altitudes. None of these conditions occur in residents at our elevation of 1500 meters. The previous two studies cited^9,10 showing central apnea in patients in a methadone treatment program were conducted at an elevation of 140 meters. Furthermore, Wang et al¹⁰ described “non-periodic breathing type of central sleep apnea” (reference 10, figure 3, page 1352) in which the pattern appears to correspond with Biot's respiration or ataxic breathing. The breathing pattern in this figure is characterized by a variable tidal volume, irregular respiratory rate, and central pauses of variable duration. In another study conducted at 77 meters' elevation, 4 of 8 patients on opioid medication and presenting for evaluation of sleep apnea showed both central apneas and Biot's or ataxic respiration.³³ With respect to our present study, all of our patients were obviously tested at the same elevation. Irregular periodic breathing or Biot's respiration was seen in 92% of patients taking a morphine dose equivalent of 200 mg or higher and in only 3 (5%) of patients not taking opioids in our control group. Finally, we have treated with oxygen numerous patients taking opioids but have not observed any consistent effect on the ataxic breathing patterns present in those cases (see our Figure 4 and figure 2 in reference 8). The possibility remains that an interaction between modest hypoxemia at elevation and other factors, such as medications, cannot be excluded.

Although the above limitations of this study are relevant concerns, our findings should raise awareness about the potential adverse consequences of chronic administration of sustained-release opioid medications. A recent report indicates a 3-fold increase in mortality rate associated with rising prescription rates for nonillicit use of methadone and oxycodone.³⁴ Based upon a retrospective analysis of these data, we that found decedents were discovered dead in the morning or dead in bed during the day.³⁵ Teichtahl et al⁹ proposed that sleep-disordered breathing may play a role in unexplained excess mortality in patients treated with methadone. The high prevalence rate (92%) of ataxic or Biot's respiration in our patients receiving a morphine dose equivalent at least 200 mg is of particular interest. Recent studies investigating the respiratory-depressant properties of μ-agonistic opioid infusions have found that respiratory variability or ataxic respiration is the most sensitive predictor of impending respiratory depression and apneas in adults and children.^36,37 Bouillon et al³⁶ concluded from their data that irregular breathing heralded respiratory arrest, and they stated, “Clinically apparent irregular breathing appears to be a very ominous sign signalling severe respiratory depression… Opioids apparently not only change the set point for P_aCO₂, but also impair the function of respiratory centres involved in rhythm generation.” Whether or not the present data provide additional crucial pathophysiologic evidence for the increased mortality rate associated with opioid therapy, the findings from this study are at the very least provocative. We believe that patients, independent of the usual risk factors such as snoring and obesity, who are receiving chronic opioid therapy of greater than a 200-mg morphine dose equivalent should be objectively evaluated with polysomnography for sleep-disordered breathing. In view of the marked increase in the prescription rate of opioids for chronic nonmalignant pain, it seems inevitable that adverse respiratory effects of these medications will become increasingly evident if more patients are objectively evaluated. The relationship between chronic opioid use and breathing disturbances during sleep needs to be further explored.