ADVERTISEMENT

Issue Navigator

Volume 12 No. 04
Earn CME
Accepted Papers
Classifieds





Scientific Investigations

Respiratory Variability during Sleep in Methadone Maintenance Treatment Patients

Chinh D. Nguyen, PhD1,2; Jong Won Kim, PhD1; Ronald R. Grunstein, MD, PhD1,3; Cindy Thamrin, PhD1; David Wang, PhD1,3,4
1Woolcock Institute of Medical Research and Sydney Medical School, University of Sydney, Glebe, New South Wales, Australia; 2Neuroscience Research Australia (NeuRA), Randwick, New South Wales, Australia; 3Department of Respiratory and Sleep Medicine, Royal Prince Alfred Hospital, Sydney Local Health District, Central Clinical School, University of Sydney, Camperdown, New South Wales, Australia; 4Department of Respiratory and Sleep Disorders Medicine, Western Hospital, University of Melbourne, Victoria, Australia

ABSTRACT

Study Objectives:

Methadone maintenance treatment (MMT) patients have a high prevalence of central sleep apnea and ataxic breathing related to damage to central respiratory rhythm control. However, the quantification of sleep apnea indices requires laborious manual scoring, and ataxic breathing pattern is subjectively judged by visual pattern recognition. This study proposes a semi-automated technique to characterize respiratory variability in MMT patients.

Methods:

Polysomnography, blood, and functional outcomes of sleep questionnaire (FOSQ) from 50 MMT patients and 20 healthy subjects with matched age, sex, and body mass index, were analyzed. Inter-breath intervals (IBI) were extracted from the nasal cannula pressure signal. Variability of IBI over 100 breaths was quantified by standard deviation (SD), coefficient of variation (CV), and scaling exponent (α) from detrended fluctuation analysis. The relationships between these variability measures and blood methadone concentration, central sleep apnea index (CAI), apnea-hypopnea index (AHI), and clinical outcome (FOSQ), were then examined.

Results:

MMT patients had significantly higher SD and CV during all sleep stages. During NREM sleep, SD and CV were correlated with blood methadone concentration (Spearman R = 0.52 and 0.56, respectively; p < 0.01). SD and CV were also correlated with CAI (R = 0.63 and 0.71, p < 0.001, respectively), and AHI (R = 0.45 and 0.58, p < 0.01, respectively). Only α showed significant correlation with FOSQ (R = −0.33, p < 0.05).

Conclusions:

MMT patients have a higher respiratory variability during sleep than healthy controls. Semi-automated variability measures are related to apnea indices obtained by manual scoring and may provide a new approach to quantify opioid-related sleep-disordered breathing.

Citation:

Nguyen CD, Kim JW, Grunstein RR, Thamrin C, Wang D. Respiratory variability during sleep in methadone maintenance treatment patients. J Clin Sleep Med 2016;12(4):607–616.


INTRODUCTION

Preventing prescription opioid overdose deaths is a major public health priority in Western societies. Deaths from these medications exceed deaths from all illicit drugs combined.1 Methadone is a long-acting μ-opioid agonist with pharmacological properties similar to morphine. The primary uses of methadone are relief of chronic pain, treatment of opioid abstinence syndromes, and treatment for heroin addiction.2 However, methadone accounts for one-third of US prescription opioid overdosing deaths, despite contributing to only 5% of all distributed opioids.3

Deaths from opioids are nearly always due to respiratory arrest, and often occur during sleep when breathing is primarily regulated by autonomic neurochemical control.2 Acute usage of μ-opioids can cause dose-dependent depression of respiration,2 reduce ventilatory responses to hypercapnia and hypoxia,4,5 increase respiratory pauses, and develop irregular breathing patterns.6 With long-term usage, stable methadone maintenance treatment (MMT) patients have a high prevalence of central sleep apnea, which occurs in 30% of patients.7,8 The severity of central sleep apnea has been shown to correlate with increased blood methadone concentration,8 and its appearance is related to increased opioid dose.911

BRIEF SUMMARY

Current Knowledge/Study Rationale: Methadone maintenance treatment (MMT) patients have a high prevalence of central sleep apnea and ataxic breathing related to damage to central respiratory rhythm control. However, the quantification of sleep apnea indices requires laborious manual scoring, and ataxic breathing pattern is subjectively judged by visual pattern recognition.

Study Impact: This study showed that MMT patients have a higher respiratory variability during sleep than normal controls population. Semi-automated variability measures are related to apnea indices obtained by manual scoring and may provide a new approach to quantify opioid-related sleep-disordered breathing.

Opioids also predispose patients with chronic usage to the development of ataxic breathing or Biot's respiration.9,10 This is distinct from Cheyne-Stokes respiration and periodic breathing with central sleep apnea.12 Ataxic breathing is characterized by irregular respiratory rate, rhythm, and depth with or without brief respiratory pauses less than 10 seconds, or a repeating pattern of several breaths.10 Walker et al. studied 60 patients using chronic opioids and found that ataxic breathing was more prevalent in patients who chronically used opioids (up to 70%), compared to healthy controls (5%) with matched sex and age.10 Furthermore, ataxic breathing was observed to relate to the dosage of morphine.10

The quantification of sleep apnea indices requires laborious manual scoring, and ataxic breathing pattern is subjectively judged by visual pattern recognition. Therefore, it is desirable to develop a semi-automated technique to characterize breathing in chronic opioid users. Inter-breath interval (IBI) is the duration between two consecutive breaths extracted from the respiratory signal. Examining the IBI series over multiple breaths potentially provides important information about different physiological and pathological conditions. Previous studies have shown that IBI variability changes with sleep stage, age, and gender.1316 Opioid use is known to affect respiratory timing, and IBI has been recommended as a useful and reproducible measure to characterize the time course of opioid effects on respiration in anesthetized patients.17,18 IBI variability has also been used to characterize breathing patterns in both infants and adults.1921 The automated analysis of IBI variability, hence, might provide useful, objective information on the effect of opioid on respiratory rhythms, without need for laborious manual scoring of apnea.

The current study aims to quantify IBI variability and use this to characterize the effect of opioid on respiratory rhythms in MMT patients during sleep. We hypothesized that objective measures of IBI variability during sleep in MMT patients (1) are different from healthy controls and relate to blood metha-done concentration, (2) correlate with sleep apnea indices, and (3) are related to clinical outcomes.

METHODS

Dataset

Subjects and procedure of this study have been previously described.8 The MMT patients had to be on methadone for 2 months and received a stable dose of methadone. All subjects underwent a screening examination to exclude those with severe cardiac, respiratory, neurological, or liver diseases, as well as those with psychotic disorder and pregnancy. In the afternoon of each sleep study, venous blood samples were taken from MMT patients for blood toxicology tests to examine methadone plasma concentration as well as potential concomitant drugs. The blood toxicology tests were performed at Victorian Institute of Forensic Medicine. Each subject completed a functional outcomes of sleep questionnaire (FOSQ), which assesses the impact of excessive sleepiness on daily activities.22 The research protocol was approved by the Institutional Research and Ethics Committee. All subjects gave written informed consent prior to participation. Data from 50 MMT patients (25 male and 25 female) and 20 healthy subjects were analyzed.

Polysomnography

In-lab full polysomnography was performed overnight on all subjects. Sleep stage was scored according to standard R&K criteria.23 Respiratory events, apnea-hypopnea index (AHI) and central apnea index (CAI) were scored using Chicago criteria.24

Preprocessing

In this study, we only analyzed the respiratory signal during periods free of artifacts during (1) wakefulness, which were either immediately after lights off and before subjects' sleep onsets, or those ≥ 5 minutes of continuous wakening during sleep; and (2) each sleep stage, which included those ≥ 5 minutes of continuous sleep without wakening. The representative hypnograms of a healthy subject and a MMT patient are presented in Figure S1 (supplemental material). The signal was sampled at 50 Hz and filtered using a second-order Savitzky-Golay (SG) filter of length 0.5s.25 The S-G filter was designed to perform as a low pass filter with the cutoff frequency of 2 Hz in this work.26 The filtered signal was then processed to calculate inter-breath-interval (IBI) series, using an amplitude threshold algorithm.27 This algorithm identifies zero-crossing points, peak inspiratory and peak expiratory points of the respiratory signal, and then applies an amplitude threshold to reject false breath-detection. Finally, the time between successive breaths is calculated to generate the IBI signal (Figure 1). Statistical properties of these IBIs were analyzed during wakefulness, REM sleep, and NREM sleep. The preprocessing process and extraction of IBI from flow signal were performed by our own signal processing modules developed in MATLAB (version 8, MathWorks).

Extraction of inter-breath interval from flow signal.

The raw flow signal was filtered and breath detection algorithm was applied to extract the inter-breath interval (IBI) signal. Closed circles represent the starts of inspiration.

jcsm.12.4.607a.jpg

jcsm.12.4.607a.jpg
Figure 1

Extraction of inter-breath interval from flow signal.

(more ...)

Analysis of Respiratory Variability

Basic Variability Measures of IBI

The variability of the IBI signal was characterized using the standard deviation (SD) and coefficient of variation (CV). SD as well as the mean was calculated during wakefulness and each sleep stage for all MMT patients and healthy subjects. CV was calculated in non-overlapping windows of 100 breaths during wake and each sleep stage, and the mean CV of all windows was obtained.

IBI Variability over Time

Detrended fluctuation analysis (DFA) is a well-established technique to quantify long-range correlations in the signal. This is the strength of the correlation between fluctuations in the signal over different short-term to long-term time scales, quantified by a single quantitative parameter, the scaling exponent α. A signal with α > 0.5 indicates the presence of long-range correlations, as distinct to a random signal with no apparent “structure” or relationship between points over time. DFA has been used extensively to study physiological signals such as heart rate variability, due to its ability to cope with changing statistical properties of the signal, such as intrinsic oscillations or external fluctuations due to environmental noise.15,28 The technique has more recently been applied to characterize respiratory variability.14,15,1821,29,30 Navarro et al. compared the performance of different techniques to characterize the scaling exponent with different breathing patterns and suggested that DFA is a robust technique which can be used to analyze signals with periodic breathing or mixed patterns.19

To calculate the scaling exponent α, we applied a procedure that has been previously reported.30 For each sleep stage, the squared fluctuation F2(s) was calculated for each segment. Then the F2(s) values of all segments of each subject for that sleep stage were averaged. Finally, we calculated the square root and measured α as the slope of the line of best fit in the log-log plot of F(s) ∼ (s)α. In this study, α was calculated in the scale range of 5 to 200 breaths. We applied this technique to IBIs during wakefulness, NREM, and REM sleep stages, separately. Figure S2 (supplemental material), shows the log-log plots used to calculate DFA α in representative MMT subjects and healthy controls.

Statistical Analyses

Statistical analyses were performed with SPSS software, version 22. The differences in respiratory variability (mean, SD, and CV of IBI) between MMT patients and healthy controls for each sleep stages were assessed by Mann-Whitney unpaired t-test (two-tailed). The differences between sleep stages were also examined using a one-way analysis of variance (ANOVA).

The relationships between respiratory variability (SD, CV, α) and level of blood methadone concentration as well as apnea indices (CAI, AHI) were examined using Spearman correlation (two-tailed). The relationships between respiratory variability and clinical outcomes (FOSQ index) were examined using multiple regression models (adjusted for age and BMI) in healthy controls and MMT patients. Comparisons were also made with CAI and AHI. Since the distribution of CAI, AHI, and level of blood methadone concentration data were skewed, they were transformed in the logarithmic scale. A p value of < 0.05 was chosen to evaluate the statistical significance.

RESULTS

Data from 50 MMT patients and 20 healthy subjects with matched age, sex, and body mass index were analyzed. Fifty MMT patients (25 male and 25 female) had previous heroin addiction, and most had additional polydrug abuse. Twenty healthy controls well matched for age, sex, and BMI were recruited into the study. Data of one healthy subject were excluded, due to a corrupted signal. Table 1 and Figure S3 (supplemental material) show the demographic data and blood toxicology and drug use (where applicable) of the MMT patients and healthy controls. MMT patients had significant lower FOSQ (15.47 ± 3.19) than healthy controls (19.37 ± 0.47), p < 0.0001. MMT patients also had higher CAI and AHI than healthy controls (Table 1 and Figure S3). The raw data of CAI, AHI, and blood methadone concentration were not normally distributed, and log-transformation was applied to convert to a normal distribution. Separate statistical tests were also performed with raw data of these indices and similar results were obtained.

Demographic data and sleep measures of stable methadone maintenance treatment (MMT) patients and healthy controls.

jcsm.12.4.607.t01.jpg

table icon
Table 1

Demographic data and sleep measures of stable methadone maintenance treatment (MMT) patients and healthy controls.

(more ...)

Difference in Mean and Variability of IBI between MMT Patients and Healthy Controls

MMT patients had significantly higher mean duration of IBIs (mean of IBI) during wakefulness and NREM, compared to healthy controls (p < 0.01; Figure 2A), but no difference was found during REM (p = 0.84; Figure 2A). In contrast, basic variability measures SD and CV of IBI were higher in MMT patients in all sleep stages, compared to the controls, p < 0.01 (Figure 2B, 2C). However, no difference was found in α between the 2 groups in all sleep stages (p > 0.49; Figure 2D).

Mean and variability measures of inter-breath interval signals.

Mean and variability measures (standard deviation [SD], coefficient of variation [CV], and slope exponent [α]) of inter-breath interval (IBI) signals in stable methadone maintenance treatment (MMT) patients and healthy controls. Boxplots represent median and 5th–95th percentile. **p < 0.01, ***p < 0.001, ****p < 0.0001.

jcsm.12.4.607b.jpg

jcsm.12.4.607b.jpg
Figure 2

Mean and variability measures of inter-breath interval signals.

(more ...)

Difference in Mean and Variability of IBI between Wakefulness and Sleep Stages

Figure S4 (supplemental material) provides detailed statistical comparisons of mean and variability of IBI between wakefulness and sleep stages in both control and patient groups. In general, mean IBI was lower in REM than NREM in MMT group, p < 0.001 (Figure S1A). However, variability of IBI tended to be higher in REM than NREM—significant for SD in controls only (p < 0.01), and CV and α in both groups (p < 0.01; Figure S1BS1D).

Similarly, mean of IBI was again lower during wakefulness than NREM in healthy group, (p < 0.01; Figure S1A). In contrast, variability of IBI tended to be higher during wakefulness than NREM—significant for SD and CV in both groups (p < 0.0001; Figure S1BS1D) and α in MMT group (p < 0.01).

No differences in mean of IBI between wakefulness and REM were found in both groups (Figure S1A). However, variability of IBI tended to be higher during wakefulness than REM—significant for SD in both groups (p < 0.0001) and CV in healthy group only (p < 0.05; Figure S1BS1D).

Relationship between Variability of IBI and Blood Methadone Concentration

In MMT patients, basic variability of IBI showed correlations with blood methadone concentration during NREM (R = 0.52, p < 0.01 for SD, and R = 0.56 p < 0.001 for CV, Figure 3). No relationship was found between α and blood methadone concentration during NREM and REM sleep and during wakefulness. Apnea indices were significantly correlated with blood methadone concentration (R = 0.5, p < 0.01 for CAI and R = 0.44, p < 0.01 for AHI, Figure 3).

Relationship between blood methadone concentration, respiratory variability measures and apnea indices.

Relationship between blood methadone concentration and respiratory variability measures (standard deviation [SD], coefficient of variation [CV], and slope exponent [α]) calculated from inter-breath intervals (IBI) and apnea indices in methadone maintenance treatment (MMT) patients during NREM. Data of CAI, AHI, and blood methadone concentration were transformed using base-10 logarithm transformation.

jcsm.12.4.607c.jpg

jcsm.12.4.607c.jpg
Figure 3

Relationship between blood methadone concentration, respiratory variability measures and apnea indices.

(more ...)

Relationship between Variability of IBI and Apnea Indices

Basic variability of IBI significantly correlated with CAI (R = 0.63, p < 0.001 for SD and R = 0.71, p < 0.001 for CV), and with AHI (R = 0.45, p < 0.01 for SD and R = 0.58, p < 0.001 for CV) during NREM (Figure 4). No relationship was found between α and CAI or AHI during all sleep stages and wakefulness.

Relationship between respiratory variability measures and apnea indices.

Relationship between respiratory variability measures (standard deviation [SD], coefficient of variation [CV], and slope exponent [α]) calculated from inter-breath intervals (IBI) and apnea indices in methadone maintenance treatment (MMT) patients during NREM. Data of CAI and AHI were transformed using base-10 logarithm transformation.

jcsm.12.4.607d.jpg

jcsm.12.4.607d.jpg
Figure 4

Relationship between respiratory variability measures and apnea indices.

(more ...)

Relationship between Variability of IBI and Clinical Outcomes

In MMT patients, the scaling exponent α was the only variability measure of IBI that was significantly correlated with FOSQ index (R = −0.329, p = 0.021) during NREM (Figure 5). This relationship remained significant after adjusting for age and BMI using multiple regression models, with standardized beta coefficient −0.377 (95% CI −28.549 to −3.194), p = 0.015 (Table 2). None of the IBI variability predictors were correlated with FOSQ index during wakefulness and REM. Apnea indices (CAI and AHI) did not correlate with FOSQ (Table 2).

Relationship between the slope exponent α and FOSQ.

Relationship between the slope exponent α obtained by detrended fluctuation analysis and FOSQ index in healthy controls (closed circles) and methadone maintenance treatment (MMT) patients (open circles) during NREM.

jcsm.12.4.607e.jpg

jcsm.12.4.607e.jpg
Figure 5

Relationship between the slope exponent α and FOSQ.

(more ...)

Beta coefficients for the relationship between respiratory variability (predictors) and clinical outcome (FOSQ index) of methadone maintenance treatment patients during NREM, adjusted for age and BMI, using multiple linear regression.

jcsm.12.4.607.t02.jpg

table icon
Table 2

Beta coefficients for the relationship between respiratory variability (predictors) and clinical outcome (FOSQ index) of methadone maintenance treatment patients during NREM, adjusted for age and BMI, using multiple linear regression.

(more ...)

DISCUSSION

This study found that using the inter-breath interval signal, (1) basic measures of respiratory variability (SD, CV) during sleep were significantly higher in MMT patients than healthy controls, and were related to the level of blood methadone concentration; (2) those basic measures were also related to apnea indices (CAI and AHI); however, (3) only the variability measure over time (scaling exponent α) was negatively correlated with clinical outcomes (FOSQ index).

Effect of Methadone on Respiratory Variability

The higher respiratory variability in MMT patients suggests that respiratory rhythm in this clinical group is more irregular than in healthy subjects, and possibly demonstrates a depressant effect of methadone on respiration. The finding is consistent with the previous reports that MMT patients have a high prevalence of irregular breathing patterns, including central sleep apnea and ataxic breathing.10 It has also been reported that the presence of obstructive sleep apnea was also common in MMT patients with subjective sleep complaints.31

During NREM sleep, the respiratory variability measures were found to significantly correlate with the level of blood methadone concentration in MMT patients. This is consistent with the reported correlation between central apnea indices and blood methadone concentration in these patients.8 Walker et al. found irregular breathing patterns (ataxic breathing) were likely to relate to the dosage of morphine—92% of subjects with daily morphine dose of equivalent or greater than 200 mg were observed to have evidence of ataxic breathing.10 Likewise, Smart et al. have found the breathing interval to be increased by the opioid effect in anaesthetized patients and suggested that breathing interval provided a method to calculate the time course of opioid effects, which could guide dosage.18

Relationship between Respiratory Variability and Apnea Indices

Importantly, respiratory variability measures during NREM sleep were significantly related to standard apnea severity indices (CAI and AHI). These findings suggest that the analysis of IBI variability could provide a useful and semi-automated measure to assess respiratory rhythm in MMT patients during sleep, which is related to CAI and AHI. The significant correlation between respiratory variability measures and apnea indices only in NREM could be due to the differences in respiratory control between sleep stages of MMT patients. It has been reported that in MMT patients, the number of central sleep apneas, apnea duration, and severity of hypoxia are higher during NREM than REM sleep.810 Consistent with this, in the current study, we found that the mean duration of IBI in MMT patients was higher in NREM compared to REM. The significant higher AHI in methadone patients compared to healthy subjects has also been suggested to be primarily due to central sleep apneas occurring during NREM sleep.10 Furthermore, ataxic breathing patterns were prevalent during NREM sleep in these patients.9,10 Hence, the overall CAI and AHI in MMT patients could be attributed to central sleep apneas and ataxic breathing mainly occurring during NREM rather than REM.

The underlying mechanism may be because of the association between opioid-related central sleep apnea/respiratory instability and increased hypoxic ventilatory chemosensitivity/ controller gain, which would be naturally depressed during REM sleep.8,32 By a similar mechanism, heart failure-related Cheyne-Stokes respiration or central sleep apnea are rarely seen during REM sleep.33 Therefore, for opioid-related studies, quantification of IBI variability during NREM sleep might be more sensitive compared to during REM or wakefulness.

Relationship between Respiratory Variability and Clinical Outcomes

Despite the fact that the scaling exponent (α) was not different between MMT and control groups, which may be due to the large variances of α in both groups potentially masking the group differences (as evident in Figure 5), it was the only predictor that negatively correlated with daytime function outcomes (FOSQ index). In contrast to the basic variability measures of the breath intervals (SD and CV) quantifying the magnitudes of variability, the scaling exponent α characterizes a different aspect of variability, which is the correlation over time of the breath intervals. Higher α values of breath intervals indicate a long breath is more likely to be followed by another long breath, and similarly a short breath is more likely to be followed by another short breath, as opposed to a random or no relationship between breath intervals over time. Our finding suggests that in MMT patients, the likelihood to have frequent and long pauses/apneas that are repeated over time, which is consistent with central apnea and ataxic breathing pattern, is correlated with reduced FOSQ index (worse outcome). In the control group, with the same α values, the shorter breaths are likely followed by similarly short breaths, to the same extent (similar degree of correlation). Therefore, basic variability measures and the scaling exponent α provide two distinct and perhaps complementary sources of information, the latter of which seems to be more related to clinical outcomes.

The lack of correlation between α and apnea measures can be explained by the fact that apnea severity indices are strongly dependent on both the number of apneic events and the threshold by which the events are defined. In contrast, α quantifies the structure of inter-breath intervals over time, examining the relationship between short-term and long-term behavior rather than using thresholds or events.

We also speculate that α values might reflect an individual's increased cortical activity related to arousals or other measures of sleep fragmentation,3437 in turn resulting in poorer FOSQ index. In support of this, other studies have shown increased α values in respiratory parameters during wakefulness and REM sleep compared to NREM.14,30 It has been suggested that this is driven by increased cortical influence on the respiratory pattern during REM.14,30,38 Consistently, our study also found higher α values during REM; however, we were not able to find any relationship between α values and arousal index, or between α values and measures of power spectrum analysis of EEG (data not shown). More detailed investigation on the relationship between cortical activity and arousals on the scaling exponent α of respiratory inter-breath intervals and subjective sleep outcomes is required.

Limitations

Some limitations need to be noted. Firstly, it is almost impossible to recruit a pure MMT cohort without concurrent use of various medications such as benzodiazepine, antidepressant, or cannabinoids. However, we did account for these drugs in blood toxicology, and the presence of those concomitant drugs was not a significant predictor of CAI.8 It is reasonable to suspect that the concurrent usage of multiple drugs in conjunction with opioids may increase the severity of sleep disordered breathing. Conversely, to our knowledge there is a lack of studies systematically addressing this issue, and previous studies have found no relationship between multiple drug usage and chronic opioid-associated central sleep apnea.31,39,40 Secondly, we took into account both pauses in respiratory signal and scored sleep apneas to characterize respiratory variability. These events, which tend to have longer durations than normal breaths, would create spikes in the inter-breath interval signal. These spikes could introduce bias in the estimation of the scaling exponent.41 It is arguable whether this represents artifact or physiological information. However, the significant relationship between α and FOSQ was not apparent for SD and CV, suggesting that the relationship was not driven by spikes in the IBI signal per se. Finally, the periods of wakefulness, NREM and REM naturally alternate several times during sleep, which limits the length of each segment available to estimate α. We attempted to overcome this by removing all segments shorter than 5 minutes. We also compared our preprocessing approach to estimate α from all segments with an alternative approach42 and found that they were statistically equivalent (data not shown).

In conclusion, stable MMT patients have higher respiratory variability during sleep than healthy controls. The respiratory variability measures were related to blood methadone concentration and apnea indices obtained by manual scoring. Semi-automated quantification of respiratory variability may provide a practical biomarker to quantify opioid-related sleep disordered breathing.

DISCLOSURE STATEMENT

This was not an industry supported study. Data collection for this study was performed at Western Hospital, The University of Melbourne. Analysis was performed at Woolcock Institute of Medical Research and Sydney Medical School, University of Sydney, Glebe, 2037 NSW, Australia. Drs. Nguyen and Thamrin are supported by an Australian National Health and Medical Research Council (NHMRC) Project Grant (1065938) and Dr. Nguyen also by a NHRMC NeuroSleep Centre of Research Excellence Fellowship (1060992). Dr. Wang is supported by NHMRC Project Grant (1043633). Dr. Grunstein is supported by NHMRC Practitioner Fellowship (1022730). The authors have indicated no financial conflicts of interest.

ABBREVIATIONS

AHI

apnea-hypopnea index

ANOVA

analysis of variance

CAI

central apnea index

CPAP

continuous positive airway pressure

CV

coefficient of variation

FOSQ

functional outcomes of sleep questionnaire

IBI

inter-breath interval

MMT

methadone maintenance treatment

SD

standard deviation

REFERENCES

1 

Centers for Disease Control and Prevention (CDC). Addressing prescription drug abuse in the United States: current activities and future opportunities. 2013. Retrieved June 11, 2015 from http://www.cdc.gov/drugoverdose/pdf/hhs_prescription_drug_abuse_report_09.2013.pdf.

2 

Gutstein H, Akil H. Opioid analgesics. In: Hardman JG, Limbird LE, Gilman AG, editors. Opioid analgesics. New York, NY: McGraw-Hill, 2005.

3 

Webster LR, Cochella S, Dasgupta N, et al., authors. An analysis of the root causes for opioid-related overdose deaths in the United States. Pain Med. 2011;12 Suppl 2:S26–35. [PubMed]

4 

Shook JE, Watkins WD, Camporesi EM, authors. Differential roles of opioid receptors in respiration, respiratory-disease, and opiate-induced respiratory depression. Am Rev Respir Dis. 1990;142:895–909. [PubMed]

5 

Weil JV, Mccullough RE, Kline JS, Sodal IE, authors. Diminished Ventilatory response to hypoxia and hypercapnia after morphine in normal man. N Engl J Med. 1975;292:1103–6. [PubMed]

6 

Bailey P, Egan T, Stanley T. Intravenous opioid anesthetics. In: Miller RD, editor. Intravenous opioid anesthetics. Churchill Livingstone, 2000.

7 

Teichtahl H, Prodromidis A, Miller B, Cherry G, Kronborg I, authors. Sleep-disordered breathing in stable methadone programme patients: a pilot study. Addiction. 2001;96:395–403. [PubMed]

8 

Wang D, Teichtahl H, Drummer O, et al., authors. Central sleep apnea in stable methadone maintenance treatment patients. Chest. 2005;128:1348–56. [PubMed]

9 

Farney RJ, Walker JM, Cloward TV, Rhondeau S, authors. Sleep-disordered breathing associated with long-term opioid therapy. Chest. 2003;123:632–9. [PubMed]

10 

Walker JM, Farney RJ, Rhondeau SM, et al., authors. Chronic opioid use is a risk factor for the development of central sleep apnea and ataxic breathing. J Clin Sleep Med. 2007;3:455–61. [PubMed Central][PubMed]

11 

Berry RB, author. Central apnea during stage 3,4 sleep. J Clin Sleep Med. 2007;3:81–2. [PubMed]

12 

Biot C. Contribution a l'étude du phénomène respiratoire de Cheyne-Stokes. Riotor. 1876.

13 

Terrill PI, Wilson SJ, Suresh S, Cooper DM, Dakin C, authors. Attractor structure discriminates sleep states: recurrence plot analysis applied to infant breathing patterns. IEEE Trans Biomed Eng. 2010;57:1108–16. [PubMed]

14 

Kantelhardt JW, Penzel T, Rostig S, Becker HF, Havlin S, Bunde A, authors. Breathing during REM and non-REM sleep: correlated versus uncorrelated behaviour. Physica a-Statistical Mechanics and Its Applications. 2003;319:447–57.

15 

Peng CK, Mietus JE, Liu YH, et al., authors. Quantifying fractal dynamics of human respiration: age and gender effects. Ann Biomed Eng. 2002;30:683–92. [PubMed]

16 

Akay M, Moodie KL, Hoopes PJ, authors. Age related alterations in the complexity of respiratory patterns. J Integr Neurosci. 2003;2:165–78. [PubMed]

17 

Ferguson LM, Drummond GB, authors. Acute effects of fentanyl on breathing pattern in anaesthetized subjects. Br J Anaesth. 2006;96:384–90. [PubMed]

18 

Smart JA, Pallett EJ, Duthie DJ, authors. Breath interval as a measure of dynamic opioid effect. Br J Anaesth. 2000;84:735–38. [PubMed]

19 

Navarro X, Poree F, Beuchee A, Carrault G, authors. Performance analysis of Hurst exponent estimators using surrogate-data and fractional lognormal noise models: application to breathing signals from preterm infants. Digit Signal Process. 2013;23:1610–19.

20 

Frey U, Silverman M, Barabasi AL, Suki B, authors. Irregularities and power law distributions in the breathing pattern in preterm and term infants. J Appl Physiol. 1998;85:789–97. [PubMed]

21 

Tellez JP, Herrera S, Benito S, Giraldo BF, authors. Analysis of the breathing pattern in elderly patients using the hurst exponent applied to the respiratory flow signal. Conf Proc IEEE Eng Med Biol Soc. 2014;2014:3422–5. [PubMed]

22 

Weaver TE, Laizner AM, Evans LK, et al., authors. An instrument to measure functional status outcomes for disorders of excessive sleepiness. Sleep. 1997;20:835–43. [PubMed]

23 

Rechtschaffen A, Kales A, authors. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. US Government Printing Office, US Public Health Service. 1968.

24 

Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. The Report of an American Academy of Sleep Medicine Task Force. Sleep. 1999;22:667–89. [PubMed]

25 

Savitzky A, Golay MJ, authors. Smoothing and differentiation of data by simplified least squares procedures. Anal Chem. 1964;36:1627–39.

26 

Schafer RW, author. What is a Savitzky-Golay filter? IEEE Signal Process Mag. 2011;28:111–7.

27 

Schmidt M, Foitzik B, Wauer RR, Winkler F, Schmalisch G, authors. Comparative investigations of algorithms for the detection of breaths in newborns with disturbed respiratory signals. Comput Biomed Res. 1998;31:413–25. [PubMed]

28 

Peng CK, Havlin S, Stanley HE, Goldberger AL, authors. Quantification of scaling exponents and crossover phenomena in nonstationary heartbeat time-series. Chaos. 1995;5:82–87. [PubMed]

29 

Thamrin C, Zindel J, Nydegger R, et al., authors. Predicting future risk of asthma exacerbations using individual conditional probabilities. J Allergy Clin Immunol. 2011;127:1494–U261. [PubMed]

30 

Rostig S, Kantelhardt JW, Penzel T, et al., authors. Nonrandom variability of respiration during sleep in healthy humans. Sleep. 2005;28:411–7. [PubMed]

31 

Sharkey KM, Kurth ME, Anderson BJ, Corso RP, Millman RP, Stein MD, authors. Obstructive sleep apnea is more common than central sleep apnea in methadone maintenance patients with subjective sleep complaints. Drug Alcohol Depend. 2010;108:77–83. [PubMed Central][PubMed]

32 

Teichtahl H, Wang D, Cunnington D, et al., authors. Ventilatory responses to hypoxia and hypercapnia in stable methadone maintenance treatment patients. Chest. 2005;128:1339–47. [PubMed]

33 

Solin P, Roebuck T, Johns DP, Walters EH, Naughton MT, authors. Peripheral and central ventilatory responses in central sleep apnea with and without congestive heart failure. Am J Respir Crit Care Med. 2000;162:2194–200. [PubMed]

34 

Orr WC, Stahl ML, authors. Sleep patterns in human methadone addiction. Br J Addict. 1978;73:311–5.

35 

Staedt J, Wassmuth F, Stoppe G, et al., authors. Effects of chronic treatment with methadone and naltrexone on sleep in addicts. Eur Arch Psychiatry Clin Neurosci. 1996;246:305–9. [PubMed]

36 

Oyefeso A, Sedgwick P, Ghodse H, authors. Subjective sleep-wake parameters in treatment-seeking opiate addicts. Drug Alcohol Depend. 1997;48:9–16. [PubMed]

37 

Stein MD, Herman DS, Bishop S, et al., authors. Sleep disturbances among methadone maintained patients. J Subst Abuse Treat. 2004;26:175–80. [PubMed]

38 

Lu J, Sherman D, Devor M, Saper CB, authors. A putative flip-flop switch for control of REM sleep. Nature. 2006;441:589–94. [PubMed]

39 

Farney RJ, McDonald AM, Boyle KM, et al., authors. Sleep disordered breathing in patients receiving therapy with buprenorphine/naloxone. Eur Respir J. 2013;42:394–403. [PubMed]

40 

Mogri M, Desai H, Webster L, Grant BJ, Mador MJ, authors. Hypoxemia in patients on chronic opiate therapy with and without sleep apnea. Sleep Breath. 2009;13:49–57. [PubMed]

41 

Chen Z, Ivanov P, Hu K, Stanley HE, authors. Effect of nonstationarities on detrended fluctuation analysis. Phys Rev E Stat Nonlin Soft Matter Phys. 2002;65:041107.

42 

Kirchner M, Schubert P, Liebherr M, Haas CT, authors. Detrended fluctuation analysis and adaptive fractal analysis of stride time data in Parkinson's disease: stitching together short gait trials. PLoS One. 2014;9.


Supplemental Material

Supplemental Material

(.pdf | 264 KB)