Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Experimental Pain and Opioid Analgesia in Volunteers at High Risk for Obstructive Sleep Apnea

  • Anthony G. Doufas ,

    Affiliations Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University School of Medicine, Stanford, California, United States of America, Outcomes Research Consortium, the Cleveland Clinic, Cleveland, Ohio, United States of America

  • Lu Tian,

    Affiliation Health Research and Policy, Stanford University School of Medicine, Stanford, California, United States of America

  • Kevin A. Padrez,

    Affiliation School of Medicine, University of California San Francisco, San Francisco, California, United States of America

  • Puntarica Suwanprathes,

    Current address: Department of Physiology, Faculty of Medicine, Mahidol University, Bangkok, Thailand

    Affiliation Stanford Sleep Medicine Center, Stanford University School of Medicine, Stanford, California, United States of America

  • James A. Cardell,

    Affiliation Stanford Sleep Medicine Center, Stanford University School of Medicine, Stanford, California, United States of America

  • Holden T. Maecker,

    Affiliation Department of Microbiology and Immunology, and Human Immune Monitoring Center, Stanford University School of Medicine, Stanford, California, United States of America

  • Periklis Panousis

    Affiliation Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University School of Medicine, Stanford, California, United States of America

Experimental Pain and Opioid Analgesia in Volunteers at High Risk for Obstructive Sleep Apnea

  • Anthony G. Doufas, 
  • Lu Tian, 
  • Kevin A. Padrez, 
  • Puntarica Suwanprathes, 
  • James A. Cardell, 
  • Holden T. Maecker, 
  • Periklis Panousis



Obstructive sleep apnea (OSA) is characterized by recurrent nocturnal hypoxia and sleep disruption. Sleep fragmentation caused hyperalgesia in volunteers, while nocturnal hypoxemia enhanced morphine analgesic potency in children with OSA. This evidence directly relates to surgical OSA patients who are at risk for airway compromise due to postoperative use of opioids. Using accepted experimental pain models, we characterized pain processing and opioid analgesia in male volunteers recruited based on their risk for OSA.


After approval from the Intitutional Review Board and informed consent, we assessed heat and cold pain thresholds and tolerances in volunteers after overnight polysomnography (PSG). Three pro-inflammatory and 3 hypoxia markers were determined in the serum. Pain tests were performed at baseline, placebo, and two effect site concentrations of remifentanil (1 and 2 µg/ml), an μ-opioid agonist. Linear mixed effects regression models were employed to evaluate the association of 3 PSG descriptors [wake after sleep onset, number of sleep stage shifts, and lowest oxyhemoglobin saturation (SaO2) during sleep] and all serum markers with pain thresholds and tolerances at baseline, as well as their changes under remifentanil.


Forty-three volunteers (12 normal and 31 with a PSG-based diagnosis of OSA) were included in the analysis. The lower nadir SaO2 and higher insulin growth factor binding protein-1 (IGFBP-1) were associated with higher analgesic sensitivity to remifentanil (SaO2, P = 0.0440; IGFBP-1, P = 0.0013). Other pro-inflammatory mediators like interleukin-1β and tumor necrosis factor-α (TNF-α) were associated with an enhanced sensitivity to the opioid analgesic effect (IL-1β, P = 0.0218; TNF-α, P = 0.0276).


Nocturnal hypoxemia in subjects at high risk for OSA was associated with an increased potency of opioid analgesia. A serum hypoxia marker (IGFBP-1) was associated with hypoalgesia and increased potency to opioid analgesia; other pro-inflammatory mediators also predicted an enhanced opioid potency.

Trial Registration:



Obstructive sleep apnea (OSA) is a common health condition [1], [2] characterized by cyclic cessations of airflow due to intermittent complete or partial airway obstruction, leading to episodic hypoxia and repeated arousals from sleep [3]. It is estimated that approximately 20% of the general [2] and surgical [4] population suffer from OSA with a fraction up to 80% of these individuals lacking a formal diagnosis [4], [5]. Sleep-disordered breathing has been linked to cardiovascular [6], [7] and metabolic [8] morbidity, while accumulating evidence suggests that OSA might increase the risk for respiratory complications [9][14] in the postoperative period due to drug-induced airway compromise, especially when μ-opioid receptor agonists are administered to treat pain [15][17]. It is thus important to delineate the relationship between sleep-disordered breathing, pain behavior, and opioid analgesia because this would facilitate opioid titration and thus safer postoperative pain management in these patients.

Important phenotypic components of sleep-disordered breathing, like sleep disruption, recurrent nocturnal hypoxemia, and systemic inflammation [18], have been linked to altered pain processing. Both experimental [19][21] and clinical [22], [23] evidence suggest that inadequate and/or disrupted sleep could enhance pain sensitivity in humans, while prolonged deprivation of sleep promoted the expression [24] and release of major sleep-regulating cytokines, including tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6 [25][30], which are known to exert hyperalgesic effects in various experimental models [31], [32]. Consistent with these observations, treatment of OSA with continuous positive airway pressure (CPAP), presumably enhancing sleep continuity, decreased the sensitivity to painful stimuli in adults [33]. On the other hand, children suffering from OSA demonstrated an increased sensitivity to the postoperative analgesic effect of morphine; interestingly, morphine dose requirement for postoperative pain was inversely proportional to the degree of nocturnal hypoxemia preoperatively [34], [35].

Herein, we evaluated experimental pain processing and opioid analgesia in male volunteers suffering from sleep-disordered breathing. We examined the effect of sleep disruption and nocturnal hypoxemia on pain behavior and opioid analgesia. We hypothesized that nocturnal hypoxemia would be associated with a decreased sensitivity to painful stimuli and increased potency to opioid analgesia, whereas sleep disruption would enhance sensitivity to pain. In this context we also investigated the role of specific inflammatory and hypoxia markers in predicting sensitivity to pain and opioid analgesic effect.


Ethics Statement

The Stanford Research Compliance Office (Human Subjects Research and IRB: approved the study. All participants gave written informed consent, while all procedures were conducted according to the principles expressed in the Declaration of Helsinki.

Subjects Selection

Epidemiologic evidence shows that male gender [1] and habitual snoring [1], [36] are among the known risk factors for the presence of obstructive sleep apnea (OSA) in the general population. Thus, from January 2008 till March 2010, we invited male volunteers (18–55 years old) with a history of habitual snoring and/or a formal diagnosis of OSA to participate in a study evaluating sleep and experimental pain processing.

We excluded volunteers with a body mass index (BMI) equal or higher than 35 Kg/m2, history of diabetes, pulmonary or heart conditions that could affect oxygenation status, chronic pain condition, psychiatric disease, moderate to severe depression (based on Beck depression inventory), untreated thyroid disease, and those who were taking analgesics or other drugs affecting the central nervous system, including antidepressants. Also, we excluded volunteers who had been treated for their OSA with continuous positive airway pressure (CPAP). The trial was registered at the under the title “study of pain processing in subjects suffering from obstructive sleep apnea” (NCT00672737, PI: AG Doufas).


The study was conducted in two separate days, approximately one week apart. In the first day, volunteers underwent overnight polysomnography (PSG) at home or in a special research facility of the Sleep Clinic at Stanford Center for Human Sleep Research. Approximately one week later, they were subjected to quantitative sensory testing at the Human Pain Laboratory in the Department of Anesthesiology, Perioperative and Pain Medicine at Stanford. The volunteers were not aware of the study hypothesis and they were informed the results of the PSG assessment only after the end of the trial.

On the second study day, the volunteers fasted for at least 6 hours and arrived at the laboratory at 8:30 AM. After undergoing a basic physical examination and completing a series of questionnaires (see Measurements), a 20-g catheter was inserted at the antecubital fossa of the left arm and a 10-mL blood sample was obtained to measure serum glucose, glycated hemoglobin (HbA1c), 3 pro-inflammatory cytokines, and 3 hypoxia markers.

Standard cardio-respiratory monitors including oscillometric blood pressure, electrocardiogram, and pulse oximetry (SaO2) were applied to the volunteers. Normothermia was maintained with forced-air warming, while volunteers breathed supplemental oxygen via nasal prongs to maintain a SaO2 greater than 90%.

After a short training period, quantitative sensory testing for heat and cold was performed at four different phases: baseline, placebo, and at two different effect site concentrations (Ce) of remifentanil. Placebo was always administered first, while the order of remifentanil Ce was randomized between 2 and 4 µg/ml in the first 9 volunteers and between 1 and 2 µg/ml in the rest. The reason for this change in the testing remifentanil concentrations was that in several subjects the predefined maximum values for heat (52°C) and cold (180 sec) painful stimuli (see below) were reached when a Ce = 4 µg/ml was targeted.

The drug delivery system consisted of a Harvard 22 (Harvard Apparatus, Holliston, MA) electronic syringe pump and the STANPUMP software driver (freely available by Steve L. Shafer, MD), based on a previously developed [37] and validated [38] pharmacokinetic-pharmacodynamic model for remifentanil. Volunteers were blinded to the pump function, during all phases of the experiment, while the investigator performing the sensory testing and sedation assessment was blinded to the remifentanil Ce and PSG recordings. During the placebo and the two remifentanil Ce phases of the experiment volunteers were told that they may or may not be administered an active drug.

Quantitative sensory testing for heat and cold with a 10-min-long resting interval in-between, begun approximately 10 minutes after the initiation of the simulated or actual remifentanil in order to make sure that a pseudo-steady state has been achieved. At the end of each of the four phases vital signs and sedation, using the responsiveness component of the observer’s assessment alertness/sedation (OAA/S) scale [39], were recorded.

At the end of the experiment the remifentanil infusion was stopped and volunteers left to recover completely before they were discharged home.

A concise study flow diagram is presented in Figure 1.

Figure 1. Study flow scheme.

On the first study day, the volunteers underwent home-based or in-hospital overnight polysomnography (PSG). Approximately a week later, they underwent experimental pain testing after a blood sample was withdrawn for determination of pro-inflammatory and hypoxia markers. Threshold and tolerances were determined for both heat and cold painful stimuli at baseline (no drug), placebo, and two (low and high) effect site concentrations (Ce) of remifentanil (1 and 2 µg/ml, or 2 and 4 µg/ml) that were targeted using an electronic syringe pump connected to a laptop equiped with a special software driver (STANPUMP, written by Steve L. Shafer, MD) and a pharmacokinetic-pharmacodynamic model for remifentanil. The order of presentation for the two remifentanil Ce was randomized, using special software and sealed envelopes that were opened on the day of trial.

Protocol deviations.

We recruited volunteers with a BMI greater than the originally set upper limit of 29 kg/m2, due to difficulty in recruiting overweight (BMI<30 kg/m2) subjects that would also demonstrate the phenotypic characteristics or have a formal diagnosis of obstructive sleep apnea. Consequently, we have recruited 9 volunteers with a BMI>29 kg/m2; all were lower than or equal to 32.66 kg/m2.

Several originally planned assessments and/or procedures were not included in the analysis or reported results, either because of insufficient data collection or because the evaluated outcomes were irrelevant to the aim of the present project. These assessments include heart rate variability, instantaneous sleepiness using the Karolinska sleepiness scale (KSS), and assessment of vigilance using the psychomotor vigilance test (PVT).


The volunteers’ demographic and morphometric characteristics were recorded. Heart rate from the electrocardiogram (ECG), blood pressure, respiratory rate via ECG and direct observation, and arterial oxyhemoglobin saturation (pulse oximetry, SaO2) were continuously monitored and recorded at the end of each of the four phases of the experiment, along with the responsiveness component of the OAA/S score.


Volunteers underwent an overnight polysomnography either at their home or in a sleep research facility, depending on the availability of the latter. In the first case, a type II [40] portable sleep recording device (Crystal Monitor 20-B, CleveMed, Cleveland Medical Devices, Inc., Cleveland, OH) was used to perform unattended home-based polysomnography (PSG). To ensure that maximal sleep was attained, approximately 2 hours prior to their usual bedtime, the volunteers arrived at the study center where they were connected to the PSG monitor by a certified polysomnography technologist. Physiological signals for a comprehensive PSG assessment were acquired through 14 different channels, including electroencephalogram (EEG; A1, A2 C3, O2), electrocardiogram (ECG), chin and leg electromyogram (EMG), electro-oculogram (EOG), nasal air flow, snore, thoracic and abdominal respiratory effort, and pulse oximetry (SaO2). Subjects returned to their home with the device firmly connected and were instructed to turn on the device as they lay down for bed and turn off the device after they were out of bed in the morning. The data was recorded on a removable SD memory card. Subjects returned with the device the following morning where it was safely removed by the technician.

Alternatively, volunteers arrived at the sleep research facility approximately 2.5 hours prior to their usual bedtime in order to allow time for adequate preparation. A certified polysomnography technologist connected the volunteers to a laboratory-based PSG equipment using the same channels and as the home-based PSG device. Signals were monitored and acquired using Alice software (Alice Sleepware 2.8.78, Philips Respironics, Murrysville, PA). More specifically, a standard diagnostic PSG study was conducted according to practice guidelines published by the American Academy of Sleep Medicine (AASM) [41], with recording electrodes and equipment attached to the volunteer to enable monitoring of the EEG (C3-A2 and C4-A1, O2-A1 and O1-A2), EOG (ROC-A1, LOC-A2), chin and anterior tibialis EMG, heart rate by two-lead ECG, snoring intensity (anterior neck microphone), nasal pressure (nasal cannula), oral airflow (thermocouple), thoracic and abdominal movement (inductance plethysmography bands), and SaO2.

Polysomnography recordings were scored in 30-sec epochs. Sleep staging was based on the Rechtschaffen and Kales criteria [42], while EEG arousals were scored according to published rules [43]. Apnea and hypopnea events were determined using the criteria for clinical research proposed by the AASM [44]. More specifically, apnea was defined by a more than 90% decrease from baseline in the amplitude of the nasal pressure signal lasting ≥10 sec. Hypopneas were determined by a clear reduction (>50% but ≤90%) in the amplitude of the signal from its baseline value, or if the decrease in the amplitude of the signal was <50%, but it was associated with either an oxygen desaturation >3% or an arousal, and the event duration was ≥10 seconds.

Polysomnographic descriptors of sleep included the total sleep period (TSP, time in bed with lights off), total sleep time (TST), sleep efficiency expressed in % [total time in bed sleeping, SE = (TST/TSP) * 100], sleep onset latency (SOL; time between lights off and falling asleep), sleep stages 1, 2, and 3 plus 4 (combined; all as percentages of TST), rapid eye movement (REM) sleep stage, number of REM periods, number of sleep stage shifts, wake after sleep onset (WASO; time spent awake after falling asleep) number of awakenings, apnea/hypopnea index (AHI; the sum of all apneic and hypopneic events divided by the hours of TST), nadir arterial oxyhemoglobin saturation (SaO2, pulse oximetry), and time spent with an SaO2<90% (in % of TST). PSG studies with sleep efficiency lower than 50% were excluded from the analysis.

All PSG studies were scored manually using Sandman software (Sandman Elite 8.0, Covidien, Boulder, CO) according to the criteria for clinical research published by the AASM [44], by a sleep physiologist who was blinded to the volunteers’ medical history. Polysomnography always preceded quantitative sensory testing by approximately one week. When volunteers fulfilled the criteria for diagnosis of sleep-disordered breathing (SDB), they were informed accordingly at the end of the study, so that they could follow up on their condition.

Scales and questionnaires.

The presence of signs and symptoms that are compatible with the diagnosis of obstructive sleep apnea (OSA) [45], [46], including snoring, unintentional sleep during wakefulness, daytime sleepiness, non-restorative sleep, fatigue, waking up breath holding, gasping, or choking, and/or having witnessed apneas, was evaluated using the questionnaire (10 yes/no questions) for exploring obstructive sleep apnea hypopnea syndrome (OSAHS) symptoms, proposed by the American Academy of Sleep Medicine [47]. Although clinical screening can increase confidence in the diagnosis of OSA (especially when 5≤ AHI<15), the absence of these findings cannot be used to exclude such a diagnosis.

Habitual sleepiness was assessed using the Epworth sleepiness scale (ESS) [48], which measures the general level of daytime sleepiness by rating the tendency of subjects to doze off or fall asleep (0 for “no chance of dozing” to 3 for “high chance of dozing”) when in 8 different situations commonly encountered in daily life.

Mood was assessed using the Beck depression inventory (BDI) [49]. According to BDI, 21 different statements were scored between 0 (indicating no depression) and 3 (indicating severe depression), with a total score up to 20 indicating none to mild depression.

Sedation level was evaluated at the end of each of the four experimental phases, using the responsiveness component of the observer’s assessment alertness/sedation (OAA/S) scale: (5) when “Responds readily to name spoken in normal tone”, (4) for “lethargic response to name spoken in normal tone”, (3) when “responds only after name is spoken loudly and/or repeatedly”, (2) when “responds only after mild prodding or shaking”, and (1) when “does not respond to mild prodding or shaking” [39].

Inflammation and hypoxia markers.

On the second study day and before the initiation of quantitative sensory testing, we determined in the volunteers’ serum the concentration of 9 pro-inflammatory (granulocyte macrophage colony-stimulating factor, GM-CSF; interferon-gamma, IFN-γ; interleukin-10, IL-10; IL-12p70; IL-1β; IL-2; IL-6; IL-8; tumor necrosis factor-α, TNF-α) and 3 hypoxia (erythropoietin, EPO; insulin growth factor binding protein-1, IGFBP-1; vascular endothelial growth factor, VEGF) markers.

Assays kits were purchased from Meso Scale Discovery (Gaithersburg, MD) and the protocol was followed as recommended. Briefly, antibody-coated 96-well plates were pre-incubated for 30 minutes with 25 ml/well of Diluent 2. Sample or 7-point diluted Calibrators were added and plates were sealed and incubated at room temperature for 2.5 h with shaking. Plates were aspirated and washed three times with PBST (PBS +0.05% Tween-20), while 25 ml of Sulfo-Tag secondary antibody in Diluent 3 were then added and the plates incubated for 2.5 h with shaking. Plates were washed 3 times in PBST as above and 150 ml of 2X Reading Buffer were added. Plates were read immediately in a Sector Imager 2400 (MesoScale Discovery). Results were analyzed with Masterplex software (MiraiBio Group, Hitashi Corporation).

For the 3-plex hypoxia kit, the following changes apply: Plates were pre-blocked with 150 ml of Blocker C for 1 h at room temperature with shaking, followed by PBST washes as above. Twenty-five milliliters per well of Diluent 7 were added followed by either sample or Calibrators. The Sulfo-Tag antibody was diluted in Diluent 8 and the assay was completed as outlined above except that 1X Reading Buffer was used.

Standards were fitted to 5-parameter logistic regression curves, from which concentrations of test samples were estimated. R2 values for the fitted curves were >0.98 for all markers. Lower limits of detection were as follows (in pg/mL): GM-CSF, 0.16; IFN-γ, 0.28; IL-10, 0.09; IL-12p70, 0.65; IL-1β, 0.01; IL-2, 0.50; IL-6, 0.14; IL-8, 0.01; TNF-α, 0.54; IGFBP-1, 20.32; EPO, 2.19; VEGF, 12.99. Spike-recovery experiments (not shown) using the S6 Calibrator yielded recovery of 79–112% for all cytokines diluted in serum versus standard buffer.

Quantitative sensory testing.

Experimental sensory testing for heat and old was performed in a 25-°C silent room, appropriate for psychophysical assessment, with the volunteers comfortably seated on a reclining chair. Both Aδ and C nociceptive fibers are responsible for encoding and relaying centrally the noxious cold and noxious heat stimuli [50]. The respective nociceptors demonstrate response characteristics that correlate well with psychophysical pain ratings during the application of noxious thermal stimuli [51][53]. Heat pain and cold pressor tests were performed in the same order with a 10-min-long resting interval in-between, at each of the four phases of the experiment (baseline, placebo, and two remifentanil concentrations). A short training period consisted of 2–3 assessments for each of the heat and cold pain parameters preceded the actual testing phase of the experiment.

Heat pain test.

A thermal sensory analyzer (TSA II Neurosensory Analyzer; Medoc Advanced Medical Systems, Durham, NC) was used to administer heat stimuli via a 30×30 mm thermode in contact with skin of the volar left forearm. According to the method of limits [54], a pressurized air-driven vortex stimulator raised the temperature of the thermode at a rate of 1°C/sec from a baseline of 35°C until the patient pushed the button of a hand-held device (computer “mouse”) to indicate that the perception changed from very hot to painful. This procedure was repeated at least 3 times and until the difference among individual threshold values was less than 0.5°C. All the obtained temperature values were recorded for the analysis. A similar procedure was used to determine the thermode temperature causing maximum tolerable pain (tolerance). After each individual testing session the location of the thermode was slightly changed to avoid sensitization of the skin to heat. The inter-stimulus interval was 30 seconds. The maximum thermode temperature was limited to 52°C to prevent tissue damage.

Cold pressor test.

A standard commonly used procedure was followed to determine pain sensitivity to cold stimuli [55]. Subjects were asked to immerse one hand and forearm into a container filled with 12 L of 0–1-°C ice water. To maintain testing consistency among subjects, care was taken to assure that the whole palm stayed in full contact with the bottom of the container throughout the immersion. A pump steadily circulated the ice water, while the temperature was continuously monitored using an electronic thermometer. Volunteers were asked to verbally indicate when they first detected pain and remove their hand when the pain could no longer be tolerated. The time elapsed between hand immersion to pain detection (threshold) and hand withdrawal (tolerance) was recorded. Cold pain threshold and tolerance were measured at least twice, or till two measurements were within a 20-%-range. The inter-stimulus interval was 5 min. The maximum time the hand was allowed immersed in ice water was 180 sec to avoid any cold-related skin injury.

Data Analysis

All data were presented as medians (ranges), or number (%) of volunteers. Age, body mass index (BMI), fasting morning glucose, glycated hemoglobin (HbA1c), Epworth sleepiness scale (ESS) and Beck depression inventory (BDI) scores, sleep descriptors (PSG), and cytokine serum levels were compared between subjects with an apnea/hypopnea index (AHI) ≥5 events per hour of sleep and those with AHI<5, two-sample Mann-Whitney rank tests. The chi-square test, whose null distribution was generated via permutation, was used to compare the distribution of race between the two groups.

Mixed effects linear regression model.

For each of the four pain-related variables (heat pain threshold, heat pain tolerance, cold pain threshold, and cold pain tolerance), we fit a linear mixed-effects regression model:where Yij is the response for the ith patient in the jth experiment, CeREMI ij is the targeted remifentanil effect site concentration level [i.e., 0 (baseline and placebo), 1, 2, and 4 µg/ml remifentanil Ce) used in the ith patient in jth experiment, Sleepi represents PSG descriptors used in the analysis (i.e., wake after sleep onset time, number of stage shifts, and nadir SaO2) for the ith patient, Ti is a binary covariate indicating if the ith patient received intervention (i.e., any drug treatment including placebo) in the jth experiment and Zi represents confounding factors adjusted in the regression analysis (see below). The subject-specific random intercept, remifentanil effect, and placebo effect (ei, fi, gi), were assumed normally distributed and used to account for the potential correlations between repeated pain-related measurements for the same volunteer. In this model, β2 and β3 measure the effect of sleep (PSG) descriptors on the pain response and on the remifentanil effect on pain response (i.e., elevation of pain threshold/tolerance), respectively. In the first set of analyses, we adjusted the model for the BMI, age, and a binary variable indicating the type (home-based vs. in-laboratory) of PSG study (Zi), as potential confounders, while in the second set of analyses we additionally adjusted for the inflammation and hypoxia markers measured by IL-6, IL-1β, TNF-α, and IGFBP-1, VEGF, EPO, respectively. The purpose of adjusting for systemic inflammation and hypoxia-related markers was to demonstrate if a potential effect of sleep fragmentation or recurrent hypoxemia on pain behavior and/or opioid analgesia was mediated through changes in the inflammatory or hypoxia markers, respectively. We selected the specific sleep and pro-inflammatory descriptors based on prior literature evidence [19], [21], [24], [34], [35], [56][58] and the results of non-paired comparisons between OSA and non-OSA subjects, described above. The statistical significance level was set at 0.05.

All statistical analyses were performed using R version 2–10-1 (The R Foundation for Statistical Computing;


Figure 2 presents the recruitment and inclusion scheme of the volunteers to the various study procedures. A total of 49 volunteers underwent both a polysomnography (PSG) study and experimental pain testing while in 44 of them blood was also withdrawn for determining pro-inflammatory and hypoxia markers. After excluding one volunteer who had an inadequate PSG study, we included data from 48 subjects in the final analysis.

Figure 2. Volunteers and study procedures.

A total of 167 volunteers were screened via telephone or a personal interview. Fifty six volunteers signed an informed concent; 3 of them withdrew before participation in any of the study procedures. All remaining 53 underwent an overnight polysomnography, but only 49 came for the pain test on the second study day. Blood was not drawn from 3 volunteers due to technical difficulties, while 2 more blood samples were lost before analysis was done; a total of 44 blood samples provided data on inflammatory and hypoxia markers. The first 9 volunteers were tested on remifentanil Ce of 2 and 4 µg/ml and all the remaining on 1 and 2 µg/ml. Two volunteers experienced intence nausea; one at the end and the other at the beginning of the first remifentanil Ce (2 µg/ml); the experiment was stopped in both cases, but pain-related measurements were acquired only in the first case. A pump malfunction prohibited the administration of remifentanil in 1 subject. One volunteer was excluded from the analysis due to an inadequate sleep study. Finally, data from 48 volunteers were included in the analysis.

Thirteen (7 with an AHI<5 events/h) out of the 48 subjects underwent a home-based, unattended PSG study. With the exception of 2 subjects, all participants (N = 31) with an AHI≥5 responded positively to at least 3 (median No of “yes” = 7) out of the 10 questions exploring OSAHS, thus raising the probability for diagnosis of OSA.

During all the phases of the experiment cardiorespiratory parameters remained essentially unchanged from their baseline levels. Besides an OAA/S score of 3 (responds only after name is spoken loudly and/or repeatedly) that was encountered in only 2 occasions under remifentanil concentration of Ce = 4 µg/ml, volunteers were able to respond either readily or lethargically to their name spoken in normal tone (i.e., an OAA/S score of 5 or 4, respectively). Baseline comparisons between volunteers with (AHI≥5 events/h) and without (AHI<5) a PSG-based diagnosis of OSA revealed no significant differences in the morphometrics and demographics, metabolic, mood, sleepiness, and pro-inflammatory markers (Table 1). Non-paired comparisons of sleep descriptors between OSA and non-OSA volunteers revealed significantly “lighter” and more discontinuous sleep with a greater amount of stage 1 sleep (P = 0.019), higher number of stage shifts (P = 0.006), more WASO time (P = 0.006), and higher number of awakenings (P = 0.005) in the former, compared to the latter. In addition, volunteers with an AHI≥5 events/hour or sleep presented with significantly more arterial oxyhemoglobin desaturation with a lower nadir SaO2 (P<0.0001) and a longer period of total sleep time spent at an SaO2<90% (P<0.0001), compared to those with AHI<5 (Table 2).

Table 1. Baseline morphometrics and demographics, metabolic, sleepiness, mood, and pro-inflammatory markers.

Table 2. Sleep architecture and breathing based on overnight polysomnography.

Mixed effects linear regression models

In this analysis we included only the 43 volunteers with a complete set of data (i.e., polysomnography, pain testing, and pro-inflammatory markers).

Insulin growth factor binding protein-1 (IGFBP-1) in the serum was a significant predictor for heat pain threshold at baseline (P = 0.0148), while the IL-1β was positively associated with the change in the heat tolerance per 1-µg/ml-increase (analgesic potency) of remifentanil (P = 0.0256, Table 3).

Table 3. Linear regression results for the effect of sleep, inflammation, and hypoxia markers on heat pain at baseline and under remifentanil.

Nadir SaO2 was negatively associated with the analgesic potency of remifentanil for the cold pain threshold (P = 0.0440). Serum TNF-α was positively associated with the analgesic potency of remifentanil for the cold pain threshold (P = 0.0276), while IL-1β was positively associated with the analgesic potency of remifentanil for cold pain tolerance (P = 0.0218). From the hypoxia markers, IGFBP-1 was positively associated with both the baseline cold pain threshold (P = 0.0013) and the analgesic potency of remifentanil (P = 0.0018, Table 4).

Table 4. Linear regression results for the effect of sleep, inflammatory, and hypoxia-related markers on cold-induced pain at baseline and under remifentanil.


We found that nocturnal hypoxemia, quantified by the lowest arterial oxyhemoglobin saturation (nadir SaO2) during sleep, was inversely proportional to the elevation of cold pain threshold under remifentanil infusion. Nocturnal hypoxemia thus potentiated the analgesic effect of an μ-opioid agonist. In corroboration of this finding, increased serum IGFBP-1, a hypoxia marker with an extensive research record in perinatal medicine [59][62], was predictive of decreased pain sensitivity and an enhanced potency for opioid analgesia. These findings confirm previous evidence in children suffering from obstructive sleep apnea (OSA), where the analgesic requirement for postoperative morphine was inversely proportional to the magnitude of nocturnal hypoxemia (lowest SaO2) experienced preoperatively [34], [35]. Furthermore, both children suffering from OSA [63] and rats treated during development with chronic intermittent hypoxia [64] were more sensitive to the respiratory depressant effect of fentanyl. Although the mechanism for this effect of recurrent hypoxia on pain and analgesia systems is vague, in vivo evidence suggests that an up-regulation of μ-opioid receptors might be responsible for the observed increase in the potency of opioids [65][67].

Nocturnal hypoxemia in OSA frequently results in, and is associated with disruption of sleep. Thus, polysomnographic (PSG) descriptors like apnea/hypopnea index, desaturation index, or arousal index may reflect physiological events that are associated with both sleep disruption and nocturnal hypoxemia. In our effort to dissect these two phenotypic components of OSA, we selected PSG indices like the number of sleep stage shifts and “wake after sleep onset” (WASO), which, although reflecting sleep disruption, were not correlated with arterial desaturation expressed by nadir SaO2. The use of nadir SaO2 as indicator of nocturnal hypoxemia was selected based on prior clinical evidence regarding its effect on the postoperative morphine requirement in children with OSA [34], [35].

Although sleep continuity was disturbed in subjects with a PSG-based diagnosis of OSA (i.e., those with AHI≥5, Table 2), none of the examined parameters (Wake time After Sleep Onset and number of Sleep Stage Shifts) was significantly related to either pain behavior or opioid analgesic effect, even after adjusting our regression model for a home-based vs. in-laboratory PSG study. Previous evidence in healthy volunteers supports a hyperalgesic effect for sleep deprivation and sleep fragmentation [19][21], [23], [57], [68], [69]. Although this finding has been replicated in subjects suffering from primary insomnia [70], experimental testing demonstrated decreased sensitivity to pain in patients with OSA [58]. It may be that the two phenotypic components of OSA (i.e., sleep disruption and nocturnal hypoxemia) exert opposite effects on pain behavior; i.e., sleep disruption enhances, whereas nocturnal hypoxemia ameliorates pain. Therefore, depending on the severity of nocturnal hypoxemia, the effect of the latter could either mask the hyperalgesia produced by sleep disruption, or even result in hypoalgesia. Based on the presumed evidence that inadequate or fragmented sleep could promote systemic inflammation [21], [24], [30], [57], [71], another reason for the lack of an effect of sleep disruption on pain behavior and/or analgesia might have been the adjustment of the model for the presence of systemic inflammation, using IL-6, IL-1β, and TNF-α; however, the unadjusted model did not reveal any effect neither. Another potential confounder for the hyperalgesic effect of sleep disruption may be inquired in the evidence that physiological sleepiness (i.e., determined objectively by the Multiple Sleep Latency Test, MSLT) in otherwise healthy pain-free volunteers has been associated with an enhanced sensitivity to experimental heat pain [72] and decreased opioid potency for analgesia [73]. In the present experiment, we employed the Epworth sleepiness scale (ESS), a subjective instrument with a nonetheless excellent correlation with MSLT [74], to evaluate habitual sleepiness in our subjects. Median ESS scores were close to a generally accepted normal level (ESS = 6) and did not differ between subjects with and without OSA. In spite of the fact that certain subjects from both groups presented with excessive sleepiness scores (ESS>10), a further adjustment of our model for ESS (results not shown) did not change our basic findings. Finally, the relatively small number of subjects suffering from moderate (15≤ AHI<30, N = 7) and severe (AHI≥30, N = 6) OSA might have been responsible for our inability to detect associations between sleep disruption and pain behavior. Lastly, we need to recognize the fact that most of the study participants, possibly due to a first night effect [75], experienced a decreased amount of deep (stage 3 and 4) sleep, which might have also hindered our ability to detect an association between OSA-related inadequate/fragmented sleep and pain behavior.

Based on prior evidence for their involvement in the regulation of sleep function [25], [26] and in pro-inflammatory physiology of sleep-disordered breathing [18], [76], [77], we selected TNF-α, IL-1β, and IL-6, to include as the basic explanatory inflammatory markers in our model. We found that increased IL-1β in the serum significantly predicted enhanced sensitivity to heat pain (i.e., decreased pain threshold), while both IL-1β and TNF-α were positively associated with an augmented potency of remifentanil analgesia for heat and cold-induced pain. These findings are in agreement with molecular and clinical evidence that inflammation-related changes may result in hyperalgesia and an increased sensitivity to the analgesic effect of opioids [78][82].

Both TNF-α and IL-1β are up-regulated in the brain during prolonged wakefulness [83], while injection of exogenous TNF-α or IL-1β imitates, to a certain extent, the physiology of sleep loss by increasing propensity to sleep [25], [26], [84]. Both chronic intermittent hypoxia and sleep fragmentation selectively and independently activate major inflammatory pathways with increased TNF-α, and IL-6 in the serum of OSA patients [74], [77], [85]. Expression of the same inflammatory paths in volunteers, as result of total or partial sleep deprivation [24], [71] was associated with the development of spontaneous pain [21], [57], while TNF-α, IL-1β, and IL-6 have been implicated in the development of neuropathic pain through modification of peripheral receptor channels and/or induction of changes in the synaptic function of CNS neurons [31], [32], [82], [86][88]. Furthermore, molecular and in vivo evidence suggest that changes occurring in the milieu of afferent sensory neurons during inflammation, may augment the potency of μ-agonists either by up-regulation of μ-opioid receptors or by increasing the affinity of the latter for their exogenous or endogenous ligands [78], [81], [89].

Heat and cold pain-related measurements that were censored at 52°C and 180 sec, respectively, were included in the analysis without any special modeling. This statistical treatment consists a conservative approach, thus should preserve the validity of the reported findings. Considering the dimension of covariates in the fully adjusted regression model, the sample size of the current study is limited and the analysis results might be sensitive to specific model assumptions. We recognize the preliminary character of our investigation, which is potentially underpowered to detect some associations of interest. A larger study is needed to verify our findings as well as identify smaller yet still clinically meaningful and/or scientifically interesting associations.

In conclusion, we found that nocturnal hypoxemia in volunteers at high risk for obstructive sleep apnea was associated with an increased potency for opioid analgesia. A hypoxia marker like IGFBP-1 was associated with a decreased sensitivity to pain and increased potency to opioid analgesia, whereas pro-inflammatory markers also predicted an enhanced sensitivity to opioid analgesia.


The authors would like to thank Professor C. Guilleminault (Stanford Sleep Medicine Center) for lending the portable sleep monitoring device, and also appreciate the contributions of Yael Rosenberg-Hasson, PhD from the Human Immune Monitoring Center, and Martha Tingle RN, CCRC, Julia McNeal RA, Martin Angst MD, and Steve L. Shafer MD, all from the Department of Anesthesiology, Perioperative, and Pain Medicine at Stanford University School of Medicine.

Author Contributions

Conceived and designed the experiments: AGD. Performed the experiments: AGD KAP JAC PP. Analyzed the data: AGD LT PS HTM. Contributed reagents/materials/analysis tools: LT PS HTM. Wrote the paper: AGD LT PS HTM.


  1. 1. Young T, Palta M, Dempsey J, Skatrud J, Weber S, et al. (1993) The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 328: 1230–1235.
  2. 2. Young T, Peppard PE, Gottlieb DJ (2002) Epidemiology of obstructive sleep apnea: a population health perspective. Am J Respir Crit Care Med 165: 1217–1239.
  3. 3. Dempsey JA, Veasey SC, Morgan BJ, O’Donnell CP (2010) Pathophysiology of sleep apnea. Physiol Rev 90: 47–112.
  4. 4. Finkel KJ, Searleman AC, Tymkew H, Tanaka CY, Saager L, et al. (2009) Prevalence of undiagnosed obstructive sleep apnea among adult surgical patients in an academic medical center. Sleep Med 10: 753–758.
  5. 5. Young T, Evans L, Finn L, Palta M (1997) Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middle-aged men and women. Sleep 20: 705–706.
  6. 6. Peppard PE, Young T, Palta M, Skatrud J (2000) Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med 342: 1378–1384.
  7. 7. Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM, et al. (2005) Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med 353: 2034–2041.
  8. 8. Punjabi NM, Beamer BA (2009) Alterations in Glucose Disposal in Sleep-disordered Breathing. Am J Respir Crit Care Med 179: 235–240.
  9. 9. Brown KA, Morin I, Hickey C, Manoukian JJ, Nixon GM, et al. (2003) Urgent adenotonsillectomy: an analysis of risk factors associated with postoperative respiratory morbidity. Anesthesiology 99: 586–595.
  10. 10. Ankichetty S, Chung F (2011) Considerations for patients with obstructive sleep apnea undergoing ambulatory surgery. Curr Opin Anaesthesiol 24: 605–611.
  11. 11. Gali B, Whalen FX Jr, Gay PC, Olson EJ, Schroeder DR, et al. (2007) Management plan to reduce risks in perioperative care of patients with presumed obstructive sleep apnea syndrome. J Clin Sleep Med 3: 582–588.
  12. 12. Gupta RM, Parvizi J, Hanssen AD, Gay PC (2001) Postoperative complications in patients with obstructive sleep apnea syndrome undergoing hip or knee replacement: a case-control study. Mayo Clin Proc 76: 897–905.
  13. 13. Blake DW, Chia PH, Donnan G, Williams DL (2008) Preoperative assessment for obstructive sleep apnoea and the prediction of postoperative respiratory obstruction and hypoxaemia. Anaesth Intensive Care 36: 379–384.
  14. 14. Chung F, Yegneswaran B, Liao P, Chung SA, Vairavanathan S, et al. (2008) Validation of the Berlin questionnaire and American Society of Anesthesiologists checklist as screening tools for obstructive sleep apnea in surgical patients. Anesthesiology 108: 822–830.
  15. 15. Eastwood PR, Szollosi I, Platt PR, Hillman DR (2002) Comparison of upper airway collapse during general anaesthesia and sleep. Lancet 359: 1207–1209.
  16. 16. Ankichetty S, Wong J, Chung F (2011) A systematic review of the effects of sedatives and anesthetics in patients with obstructive sleep apnea. J Anaesthesiol Clin Pharmacol 27: 447–458.
  17. 17. Gali B, Whalen FX, Schroeder DR, Gay PC, Plevak DJ (2009) Identification of patients at risk for postoperative respiratory complications using a preoperative obstructive sleep apnea screening tool and postanesthesia care assessment. Anesthesiology 110: 869–877.
  18. 18. Mills PJ, Dimsdale JE (2004) Sleep apnea: a model for studying cytokines, sleep, and sleep disruption. Brain Behav Immun 18: 298–303.
  19. 19. Smith MT, Edwards RR, McCann UD, Haythornthwaite JA (2007) The effects of sleep deprivation on pain inhibition and spontaneous pain in women. Sleep 30: 494–505.
  20. 20. Roehrs T, Hyde M, Blaisdell B, Greenwald M, Roth T (2006) Sleep loss and REM sleep loss are hyperalgesic. Sleep 29: 145–151.
  21. 21. Haack M, Lee E, Cohen DA, Mullington JM (2009) Activation of the prostaglandin system in response to sleep loss in healthy humans: potential mediator of increased spontaneous pain. Pain 145: 136–141.
  22. 22. Smith MT, Klick B, Kozachik S, Edwards RE, Holavanahalli R, et al. (2008) Sleep onset insomnia symptoms during hospitalization for major burn injury predict chronic pain. Pain 138: 497–506.
  23. 23. Edwards RR, Almeida DM, Klick B, Haythornthwaite JA, Smith MT (2008) Duration of sleep contributes to next-day pain report in the general population. Pain 137: 202–207.
  24. 24. Irwin MR, Wang M, Campomayor CO, Collado-Hidalgo A, Cole S (2006) Sleep deprivation and activation of morning levels of cellular and genomic markers of inflammation. Arch Intern Med 166: 1756–1762.
  25. 25. Krueger JM (2008) The role of cytokines in sleep regulation. Curr Pharm Des 14: 3408–3416.
  26. 26. Krueger JM, Clinton JM, Winters BD, Zielinski MR, Taishi P, et al. (2011) Involvement of cytokines in slow wave sleep. Prog Brain Res 193: 39–47.
  27. 27. Pollmacher T, Schuld A, Kraus T, Haack M, Hinze-Selch D, et al. (2000) Experimental immunomodulation, sleep, and sleepiness in humans. Ann N Y Acad Sci 917: 488–499.
  28. 28. Chennaoui M, Sauvet F, Drogou C, Van Beers P, Langrume C, et al. (2011) Effect of one night of sleep loss on changes in tumor necrosis factor alpha (TNF-alpha) levels in healthy men. Cytokine 56: 318–324.
  29. 29. Frey DJ, Fleshner M, Wright KP Jr (2007) The effects of 40 hours of total sleep deprivation on inflammatory markers in healthy young adults. Brain Behav Immun 21: 1050–1057.
  30. 30. Vgontzas AN, Zoumakis E, Bixler EO, Lin HM, Follett H, et al. (2004) Adverse effects of modest sleep restriction on sleepiness, performance, and inflammatory cytokines. J Clin Endocrinol Metab 89: 2119–2126.
  31. 31. Kawasaki Y, Zhang L, Cheng JK, Ji RR (2008) Cytokine mechanisms of central sensitization: distinct and overlapping role of interleukin-1beta, interleukin-6, and tumor necrosis factor-alpha in regulating synaptic and neuronal activity in the superficial spinal cord. J Neurosci 28: 5189–5194.
  32. 32. Ren K, Dubner R (2010) Interactions between the immune and nervous systems in pain. Nat Med 16: 1267–1276.
  33. 33. Khalid I, Roehrs TA, Hudgel DW, Roth T (2011) Continuous positive airway pressure in severe obstructive sleep apnea reduces pain sensitivity. Sleep 34: 1687–1691.
  34. 34. Brown KA, Laferriere A, Lakheeram I, Moss IR (2006) Recurrent hypoxemia in children is associated with increased analgesic sensitivity to opiates. Anesthesiology 105: 665–669.
  35. 35. Brown KA, Laferriere A, Moss IR (2004) Recurrent hypoxemia in young children with obstructive sleep apnea is associated with reduced opioid requirement for analgesia. Anesthesiology 100: 806–810.
  36. 36. Olson LG, King MT, Hensley MJ, Saunders NA (1995) A community study of snoring and sleep-disordered breathing. Health outcomes. Am J Respir Crit Care Med 152: 717–720.
  37. 37. Minto CF, Schnider TW, Egan TD, Youngs E, Lemmens HJ, et al. (1997) Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology 86: 10–23.
  38. 38. Minto CF, Schnider TW, Shafer SL (1997) Pharmacokinetics and pharmacodynamics of remifentanil. II. Model application. Anesthesiology 86: 24–33.
  39. 39. Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, et al. (1990) Validity and reliability of the Observer’s Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 10: 244–251.
  40. 40. Chesson AL Jr, Berry RB, Pack A (2003) Practice parameters for the use of portable monitoring devices in the investigation of suspected obstructive sleep apnea in adults. Sleep 26: 907–913.
  41. 41. Kushida CA, Littner MR, Morgenthaler T, Alessi CA, Bailey D, et al. (2005) Practice parameters for the indications for polysomnography and related procedures: an update for 2005. Sleep 28: 499–521.
  42. 42. Rechtschaffen A, Kales A, editors (1968) A manual of standardized terminology, techniques and scoring system of sleep stages in human subjects. Los Angeles: Brain Information Service/Brain Research Institute, University of California.
  43. 43. American Sleep Disorders Association (1992) EEG arousals: scoring rules and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 15: 173–184.
  44. 44. American Academy of Sleep Medicine (1999) 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 22: 667–689.
  45. 45. American Academy of Sleep Medicine (2005) International classification of sleep disorders: Diagnostic and coding manual. Westchester, IL: American Academy of Sleep Medicine.
  46. 46. Epstein LJ, Kristo D, Strollo PJ Jr, Friedman N, Malhotra A, et al. (2009) Clinical guideline for the evaluation, management and long-term care of obstructive sleep apnea in adults. J Clin Sleep Med 5: 263–276.
  47. 47. Meoli AL, Rosen CL, Kristo D, Kohrman M, Gooneratne N, et al. (2003) Upper airway management of the adult patient with obstructive sleep apnea in the perioperative period–avoiding complications. Sleep 26: 1060–1065.
  48. 48. Johns MW (1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14: 540–545.
  49. 49. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) An inventory for measuring depression. Arch Gen Psychiatry 4: 561–571.
  50. 50. Bessou P, Perl ER (1969) Response of cutaneous sensory units with unmyelinated fibers to noxious stimuli. J Neurophysiol 32: 1025–1043.
  51. 51. Handwerker HO, Kobal G (1993) Psychophysiology of experimentally induced pain. Physiol Rev 73: 639–671.
  52. 52. Plaghki L, Decruynaere C, Van Dooren P, Le Bars D (2010) The fine tuning of pain thresholds: a sophisticated double alarm system. PLoS One 5: e10269.
  53. 53. Schepers RJ, Ringkamp M (2009) Thermoreceptors and thermosensitive afferents. Neurosci Biobehav Rev 33: 205–212.
  54. 54. Yarnitsky D (1997) Quantitative sensory testing. Muscle Nerve 20: 198–204.
  55. 55. Wolff B (1984) Methods of testing pain mechanisms in normal man. In: Wall PD, Melzack R, editors. Textbook of Pain. Edinburgh: Churchill Livingstone. 186–194.
  56. 56. Edwards RR, Grace E, Peterson S, Klick B, Haythornthwaite JA, et al. (2009) Sleep continuity and architecture: Associations with pain-inhibitory processes in patients with temporomandibular joint disorder. Eur J Pain 13: 1043–1047.
  57. 57. Haack M, Sanchez E, Mullington JM (2007) Elevated inflammatory markers in response to prolonged sleep restriction are associated with increased pain experience in healthy volunteers. Sleep 30: 1145–1152.
  58. 58. Smith MT, Wickwire EM, Grace EG, Edwards RR, Buenaver LF, et al. (2009) Sleep disorders and their association with laboratory pain sensitivity in temporomandibular joint disorder. Sleep 32: 779–790.
  59. 59. Huang ST, Vo KC, Lyell DJ, Faessen GH, Tulac S, et al. (2004) Developmental response to hypoxia. FASEB J 18: 1348–1365.
  60. 60. Kajimura S, Aida K, Duan C (2005) Insulin-like growth factor-binding protein-1 (IGFBP-1) mediates hypoxia-induced embryonic growth and developmental retardation. Proc Natl Acad Sci U S A 102: 1240–1245.
  61. 61. Krampl E, Kametas NA, McAuliffe F, Cacho-Zegarra AM, Nicolaides KH (2002) Maternal serum insulin-like growth factor binding protein-1 in pregnancy at high altitude. Obstet Gynecol 99: 594–598.
  62. 62. Tazuke SI, Mazure NM, Sugawara J, Carland G, Faessen GH, et al. (1998) Hypoxia stimulates insulin-like growth factor binding protein 1 (IGFBP-1) gene expression in HepG2 cells: a possible model for IGFBP-1 expression in fetal hypoxia. Proc Natl Acad Sci U S A 95: 10188–10193.
  63. 63. Waters KA, McBrien F, Stewart P, Hinder M, Wharton S (2002) Effects of OSA, inhalational anesthesia, and fentanyl on the airway and ventilation of children. J Appl Physiol 92: 1987–1994.
  64. 64. Moss IR, Brown KA, Laferriere A (2006) Recurrent hypoxia in rats during development increases subsequent respiratory sensitivity to fentanyl. Anesthesiology 105: 715–718.
  65. 65. Brown KA (2009) Intermittent hypoxia and the practice of anesthesia. Anesthesiology 110: 922–927.
  66. 66. Laferriere A, Liu JK, Moss IR (2003) Neurokinin-1 versus mu-opioid receptor binding in rat nucleus tractus solitarius after single and recurrent intermittent hypoxia. Brain Res Bull 59: 307–313.
  67. 67. Lerman J (2006) Unraveling the mysteries of sleep-disordered breathing in children. Anesthesiology 105: 645–647.
  68. 68. Kundermann B, Spernal J, Huber MT, Krieg JC, Lautenbacher S (2004) Sleep deprivation affects thermal pain thresholds but not somatosensory thresholds in healthy volunteers. Psychosom Med 66: 932–937.
  69. 69. Onen SH, Alloui A, Gross A, Eschallier A, Dubray C (2001) The effects of total sleep deprivation, selective sleep interruption and sleep recovery on pain tolerance thresholds in healthy subjects. J Sleep Res 10: 35–42.
  70. 70. Haack M, Scott-Sutherland J, Santangelo G, Simpson NS, Sethna N, et al. (2012) Pain sensitivity and modulation in primary insomnia. Eur J Pain 16: 522–533.
  71. 71. Irwin MR, Wang M, Ribeiro D, Cho HJ, Olmstead R, et al. (2008) Sleep loss activates cellular inflammatory signaling. Biol Psychiatry 64: 538–540.
  72. 72. Chhangani BS, Roehrs TA, Harris EJ, Hyde M, Drake C, et al. (2009) Pain sensitivity in sleepy pain-free normals. Sleep 32: 1011–1017.
  73. 73. Steinmiller CL, Roehrs TA, Harris E, Hyde M, Greenwald MK, et al. (2010) Differential effect of codeine on thermal nociceptive sensitivity in sleepy versus nonsleepy healthy subjects. Exp Clin Psychopharmacol 18: 277–283.
  74. 74. Bravo Mde L, Serpero LD, Barcelo A, Barbe F, Agusti A, et al. (2007) Inflammatory proteins in patients with obstructive sleep apnea with and without daytime sleepiness. Sleep Breath 11: 177–185.
  75. 75. Agnew HW Jr, Webb WB, Williams RL (1966) The first night effect: an EEG study of sleep. Psychophysiology 2: 263–266.
  76. 76. Vgontzas AN, Zoumakis E, Lin HM, Bixler EO, Trakada G, et al. (2004) Marked decrease in sleepiness in patients with sleep apnea by etanercept, a tumor necrosis factor-alpha antagonist. J Clin Endocrinol Metab 89: 4409–4413.
  77. 77. Arnardottir ES, Mackiewicz M, Gislason T, Teff KL, Pack AI (2009) Molecular signatures of obstructive sleep apnea in adults: a review and perspective. Sleep 32: 447–470.
  78. 78. Mohan S, Davis RL, DeSilva U, Stevens CW (2010) Dual regulation of mu opioid receptors in SK-N-SH neuroblastoma cells by morphine and interleukin-1beta: evidence for opioid-immune crosstalk. J Neuroimmunol 227: 26–34.
  79. 79. Sachs D, Coelho FM, Costa VV, Lopes F, Pinho V, et al. (2011) Cooperative role of tumour necrosis factor-alpha, interleukin-1beta and neutrophils in a novel behavioural model that concomitantly demonstrates articular inflammation and hypernociception in mice. Br J Pharmacol 162: 72–83.
  80. 80. Niikura K, Narita M, Butelman ER, Kreek MJ, Suzuki T (2010) Neuropathic and chronic pain stimuli downregulate central mu-opioid and dopaminergic transmission. Trends Pharmacol Sci 31: 299–305.
  81. 81. Przewlocki R, Przewlocka B (2001) Opioids in chronic pain. Eur J Pharmacol 429: 79–91.
  82. 82. Zhang JM, An J (2007) Cytokines, inflammation, and pain. Int Anesthesiol Clin 45: 27–37.
  83. 83. Lue FA, Bail M, Jephthah-Ochola J, Carayanniotis K, Gorczynski R, et al. (1988) Sleep and cerebrospinal fluid interleukin-1-like activity in the cat. Int J Neurosci 42: 179–183.
  84. 84. Vgontzas AN, Papanicolaou DA, Bixler EO, Lotsikas A, Zachman K, et al. (1999) Circadian interleukin-6 secretion and quantity and depth of sleep. J Clin Endocrinol Metab 84: 2603–2607.
  85. 85. Ryan S, Taylor CT, McNicholas WT (2005) Selective activation of inflammatory pathways by intermittent hypoxia in obstructive sleep apnea syndrome. Circulation 112: 2660–2667.
  86. 86. Schafers M, Sorkin L (2008) Effect of cytokines on neuronal excitability. Neurosci Lett 437: 188–193.
  87. 87. Kawasaki Y, Xu ZZ, Wang X, Park JY, Zhuang ZY, et al. (2008) Distinct roles of matrix metalloproteases in the early- and late-phase development of neuropathic pain. Nat Med 14: 331–336.
  88. 88. Sorkin LS, Xiao WH, Wagner R, Myers RR (1997) Tumour necrosis factor-alpha induces ectopic activity in nociceptive primary afferent fibres. Neuroscience 81: 255–262.
  89. 89. Hua S, Cabot PJ (2010) Mechanisms of peripheral immune-cell-mediated analgesia in inflammation: clinical and therapeutic implications. Trends Pharmacol Sci 31: 427–433.