Systematic review: relationships between sleep and gastro-oesophageal reflux

Transcription

1 Alimentary Pharmacology and Therapeutics Systematic review: relationships between sleep and gastro-oesophageal reflux J. Dent*, R. H. Holloway* & P. R. Eastwood *Royal Adelaide Hospital and University of Adelaide, Adelaide, SA, Australia. West Australian Sleep Disorders Research Institute, Sir Charles Gairdner Hospital and University of Western Australia, Perth, WA, Australia. Correspondence to: Prof. J. Dent, Department of Gastroenterology and Hepatology, Royal Adelaide Hospital, North Terrace, Adelaide, SA 5000, Australia. Publication data Submitted 4 June 2013 First decision 3 July 2013 Resubmitted 17 July 2013 Accepted 18 July 2013 EV Pub Online 19 August 2013 This uncommissioned systematic review was subject to full peer-review. SUMMARY Background Gastro-oesophageal reflux disease (GERD) adversely impacts on sleep, but the mechanism remains unclear. Aim To review the literature concerning gastro-oesophageal reflux during the sleep period, with particular reference to the sleep/awake state at reflux onset. Methods Studies identified by systematic literature searches were. Results Overall patterns of reflux during the sleep period show consistently that oesophageal acid clearance is slower, and reflux frequency and oesophageal acid exposure are higher in patients with GERD than in healthy individuals. Of the 17 mechanistic studies identified by the searches, 15 reported that a minority of reflux occurred during stable sleep, but the prevailing sleep state at the onset of reflux in these studies remains unclear owing to insufficient temporal resolution of recording or analysis methods. Two studies, in healthy individuals and patients with GERD, analysed sleep and with adequate resolution for temporal alignment of sleep state and the onset of reflux: all 232 sleep period reflux evaluated occurred during arousals from sleep lasting less than 15 s or during longer duration awakenings. Six mechanistic studies found that transient lower oesophageal sphincter relaxations were the most common mechanism of sleep period reflux. Conclusions Contrary to the prevailing view, subjective impairment of sleep in GERD is unlikely to be due to the occurrence of reflux during stable sleep, but could result from slow clearance of acid reflux that occurs during arousals or awakenings from sleep. Definitive studies are needed on the sleep/awake state at reflux onset across the full GERD spectrum. Aliment Pharmacol Ther 2013; 38: doi: /apt.12445

2 J. Dent et al. INTRODUCTION Impairment of sleep quality by night-time reflux symptoms Impairment of sleep quality by night-time reflux symptoms is well documented by studies that have found that gastro-oesophageal reflux disease (GERD) is associated with perceived interruptions of sleep, fatigue and consequent impaired work productivity. 1 3 Results from a placebo-controlled trial of patients troubled by nocturnal heartburn and regurgitation have shown that there are significant improvements in patient- sleep quality in participants treated with proton pump inhibitors, suggesting that reflux, in particular acid exposure, contributes directly to impairment of sleep. 4 These findings provide strong logic for conducting mechanistic studies to evaluate the detailed patterning or architecture of distinct sleep stages, arousals and awake periods in patients with GERD, and to relate these sleep monitoring data to the onset of reflux, to symptoms and to the subsequent clearance of the refluxed acid. More accurate insights provided by such mechanistic studies could provide the basis for development of better targeted interventions on the events that lead to reflux-induced sleep fragmentation. Evaluations of the relationship between sleep state and occurrence of reflux Evaluations of the relationship of sleep state to occurrence of reflux agree that arousals and awakenings from sleep are closely associated with reflux. However, it is important to recognise that any such association could be a reflection of the large number of arousals that normally occur during sleep. For example, the frequency of arousals varies from 4 to 26 per hour, depending on the scoring criteria used, 5 and increases with age, from 11 per hour in individuals aged years to 22 per hour in those aged years. 6 Therefore, there exists a high probability, by chance, of a reflux event occurring in proximity to any given arousal or awakening. Precise timings and occurrences of arousals and awakenings can only be obtained by polysomnography, which requires concurrent recording of the electroencephalogram (EEG), the submental electromyogram (EMG) and the electro-oculogram (EOG). The EOG captures the speed and pattern of eye movements, allowing recognition of rapid eye movements (REMs), which, in combination with silence on the EMG tracing and low-amplitude, high-frequency EEG readings, indicates REM sleep. Nonrapid eye movement (NREM) sleep, the second major category of sleep, is subdivided into three or four depth stages mainly on the basis of the EEG pattern. Evaluation of sleep/awake patterns can also be obtained using actigraphy. An actigraph is a wrist-watch-like data logger device attached to the wrist of the nondominant arm that detects movements with accelerometers and records the timings. 7 No other variables, such as EEG or EMG, are recorded. Documented periods of stillness during the sleep period are taken to indicate sleep and the occurrence of body movements is taken to indicate interruptions of stable sleep. Actigraphy is a useful method for investigating group differences, gross sleep-pattern variations over time and the effects of behavioural or treatment interventions, 8, 9 as long as the recording period is long enough to provide reliable measures. 10, 11 However, it has significant limitations for the detailed assessment of the mechanistic relationship between sleep and reflux: wakefulness during the sleep period has been detected by actigraphy for only about 35 50% of polysomnographically detected awake time Arousals and awakenings are variably defined Arousals and awakenings have been variably defined. This presents a significant practical issue for review of their relationship with reflux events. Some reports use these terms interchangeably (i.e. an awakening having the same criteria as an arousal), whereas others define an awakening as a period of wakefulness that occupies at least half of the time of the epoch(s) in which it occurred (i.e. at least 15 or 30 s of the most commonly used 30- or 60-s ), an approach that will recognise only some awakenings and no arousals. The most accurate approach to the recognition and scoring of arousals and awakenings is to use a second-by-second analysis of the polysomnography recording. Using this method, an arousal is widely defined as a disturbance of stable sleep lasting between 3 and 15 s that has an alpha EEG rhythm. A 3-s minimum duration is a pragmatic cut-off, 18 19, 20 but Younes and coworkers defined even brief (>1.5 s) high-frequency EEG shifts as arousals because these appear to be relevant to the abrupt resolution of obstructive respiratory events and so may also be relevant to the onset of some reflux. An awakening is taken as a disturbance of stable sleep that lasts for more than 15 s. Hence, an awakening will always start as an arousal. Three major possible mechanistic scenarios Three major possible mechanistic scenarios could underlie fragmented sleep in GERD and the universally 658 Aliment Pharmacol Ther 2013; 38:

3 Systematic review: sleep and gastro-oesophageal reflux observed close temporal relationship between arousals/ awakenings and reflux onset. These are illustrated schematically in Figure 1 and are not necessarily mutually exclusive. The first and most widely accepted mechanistic possibility is that reflux occurs during sleep, causing oesophageal sensations that provoke arousals/awakenings, as illustrated in Figure 1a. It is possible, however, that some, maybe many, arousals/awakenings thought to be triggered by reflux could be unrelated to it and may occur by chance as part of the normal architecture of sleep. (a) AWAKENING (b) AWAKENING Swallows Swallows Time (min) Time (min) (c) AROUSAL AWAKENING Swallows Time (min) Figure 1 Schematic illustration of three possible scenarios for the timing of reflux onset in relation to arousals or awakenings. Swallows are shown, but for simplicity, oesophageal body pressures are not included. The sleep analysis summarised in the top row of each scenario represents a second-by-second scoring of sleep or arousals/awakenings. (a) Reflux during stable sleep. Reflux occurs shortly before sleep and is then interrupted by an awakening. In this scenario, it has been assumed that the awakening has been provoked by reflux-induced sensations. (b) Reflux during arousals/awakenings. In this scenario, it is proposed that arousal or awakening is the permissive event that allows occurrence of motor events that trigger reflux. In this case, the arousal/awakening preceded the occurrence of reflux onset, so it cannot be considered to have been triggered by reflux. Rather, the arousal is most likely to be consistent with the normal architecture of sleep. Some reflux are associated with arousals that last only a few seconds (see Figure 1c and Figure 2): in such cases, the motor events that lead to reflux may still be evolving when sleep is resumed, so that reflux onset may occur after return to sleep. (c) Prolonged oesophageal acidification leading to arousals/awakenings. This schematic encompasses a 20-min period, but only illustrates 10 min of recording, with the time between 5 and 15 min excised. During this excised period, there was no reflux or swallowing and stable sleep persisted throughout. Reflux occurs according to the scenario shown in Figure 1b, except that it happens during an arousal that lasts for just under 15 s. There is only one swallow following reflux before sleep resumes and, as a result, acid clearance fails because swallowing does not occur during the subsequent 16.5-min period of stable sleep. Oesophageal remains below 4 until it is cleared by swallowing during the next awakening. It is possible that this prolonged period of oesophageal acidification leads to heartburn or chest pain, which then triggers the awakening. Aliment Pharmacol Ther 2013; 38:

4 J. Dent et al. AWAKENING 8 Oesophageal 4 0 Pharyngeal pressure (mmhg) Oesophageal pressure (mmhg) LOS pressure (mmhg) Stomach pressure (mmhg) Time (min) Figure 2 Manometric and recording in a healthy individual made during polysomnography, which was scored on a second-by-second basis for sleep and arousals/awakenings, as shown in the top panel. The spikes in the pharyngeal pressure trace are generated by swallowing. LOS pressure was monitored with a sleeve sensor. The first recorded swallow occurred at the start of an awakening from sleep. During this awakening, there was abrupt acidification (dashed line) of the distal oesophagus during a typical transient LOS relaxation that started about 15 s after the second swallow and was not causally related to it. Offset of the transient relaxation is associated with a feeble synchronous oesophageal body contraction that has no discernible effect on oesophageal. Even though the recovery of LOS pressure makes it impossible for further reflux to occur, oesophageal remained well below 4. Subsequent swallow-induced peristalsis, still during the awakening, caused a minor degree of acid clearance, after which the subject returned to sleep; during this sleep interval (about 5 min of which is shown in the figure), there was a slow upward drift of. Beyond the time shown in the figure, the sleep interval lasted for just over 20 min, without any swallowing and with maintenance of the LOS pressure barrier: during this time, oesophageal remained below 4. The next arousal resulted in effective acid clearance because it was associated with several swallows and peristalsis. e that small, irregular pressure waves not related to respiration were recorded in the stomach pressure tracing during the awakening before and just after reflux onset. These were caused by nonrespiratory body movements and indicate that this particular awakening would have been recognised by actigraphy. Probably, actigraphy would have scored the last swallow of the awakening as having occurred during sleep, because awake stillness dominated this time period. Adapted from Dent et al. 21 with permission. LOS, lower oesophageal sphincter. The second possible scenario is that reflux occurs during arousals/awakenings (Figure 1b). In this circumstance, the change in sleep state cannot be a consequence of the reflux that follows: rather, it is possible that arousals/awakenings are permissive factors for the motor changes that lead to reflux. The third possible scenario (Figure 1c) is that sleep is disrupted by the development of heartburn minutes after 660 Aliment Pharmacol Ther 2013; 38:

5 Systematic review: sleep and gastro-oesophageal reflux the onset of reflux when there has been a failure of oesophageal acid clearance. In this situation, prolonged periods of oesophageal acidification occur when there is an early resumption of stable sleep when oesophageal is still low after a reflux episode: acid clearance does not occur with resumption of sleep because of sleep-related suppression of swallowing and salivation, the normal acid clearance mechanism Even in healthy individuals, this scenario can lead to very prolonged exposure of the oesophageal mucosa to noxious gastric content (Figure 2). Experimental infusions into the oesophagus during stable sleep support this third scenario of sleep-induced prolonged failure of oesophageal acid clearance as a plausible cause of fragmentation 22, 23 of sleep, especially in patients with reflux disease. Arousals occurred significantly more frequently and earlier with intra-oesophageal infusions of acid during stable sleep than with infusions of water. Patients with reflux disease also had a significantly shorter latency to acid-infusion-associated arousal than healthy volunteers, probably because of acid sensitivity of the oesophageal mucosa caused by high levels of daytime reflux. 23 A study by Orr and Johnson in healthy individuals showed that the lower the of the infusate used, the shorter the latency to arousal. 26 Currently, there is disagreement in the literature on whether arousals/awakenings are triggered by reflux or vice versa. This article reviews the literature that provides data on the overall patterning of sleep period reflux, the possible relationships between sleep state and the occurrence of reflux during the sleep period, and how reflux disease could cause fragmentation of sleep. Search string ( Sleep OR nocturnal OR nighttime OR night OR supine OR recumbent OR arousal ) AND ( gastroesophageal reflux OR gastro-oesophageal reflux )* PubMed n = 1148 Limits: exclude reviews, exclude ( infant OR child OR children OR pediatric OR paediatric ) Studies excluded based on manual review of titles, abstracts and/or full papers 17 studies on relationship between gastro-oesophageal reflux and sleep state Embase n = 1453 Combined, duplicates removed n = 1750 Reviews, editorials or commentaries n = 284 Studies not about gastro-oesophageal reflux and sleep state n = 1386 Studies not conducted in adults n = 63 Included studies *Search update also included the terms acid clearance OR esophageal acidification OR oesophageal acidification. Figure 3 Flow chart of literature searches. The systematic searches identified 17 studies about the relationship between reflux and sleep/awake state. METHODS Systematic literature searches were conducted in Pub- Med and Embase on 25 May 2010 and were updated on 4 January 2012 to identify studies on the relationship between sleep and measures of gastro-oesophageal reflux. The following search string was used: ( Sleep OR nocturnal OR nighttime OR night OR supine OR recumbent OR arousal ) AND ( gastroesophageal reflux OR gastro-oesophageal reflux ). The search update also included the terms acid clearance OR esophageal acidification OR oesophageal acidification. Search filters were applied to exclude reviews and studies conducted in children. The remaining studies were screened manually based on titles and abstracts, and the full article was examined when the relevance of the study was not clear from the abstract. Commentaries and editorials were excluded, as were studies in children and reviews that had not been eliminated by the search filters. Studies were included if they evaluated the temporal relationship between sleep state and reflux. Supporting studies reporting on the overall patterns of oesophageal acid exposure during awake and sleep periods were also included where relevant. A flow chart of the systematic searches is shown in Figure 3. RESULTS AND DISCUSSION PART I: OVERALL PATTERNING OF PERIOD REFLUX AND OESOPHAGEAL ACID CLEARANCE The systematic literature searches identified 17 publications that evaluated the temporal relationship between sleep state and individual reflux (Table 1). 2, 21, 24, These are presented and discussed in Part II of the Results and Discussion. For studies Aliment Pharmacol Ther 2013; 38:

6 J. Dent et al. Table 1 Mechanistic sleep/reflux studies identified by the literature review Reference Oesophageal reflux detection method Reflux onset related to sleep state? Temporal resolution of reflux recording Participants Proportion of reflux during sleep LOS function Sleep recording/ analysis methods Comments Poh et al., No Digitisation every 5 s Poh et al., No Probably digitisation every 5 s Di Marino et al., Dickman et al., stated stated Inadequate information Healthy: n = 9; GERD: n = 39 GERD: n = 39 GERD: n = 8; disturbed sleep, no GERD: n = 8 stated GERD: n = 15 (11 RO) Orr et al., Yes stated Poor sleep, no heartburn: n = 81; normal sleep, no heartburn: n = 39 Kuribayashi et al., manometry Yes <1 s OSA and RO: n = 8 OSA without RO: n = 9 Healthy: n = 8 No data 14.4% (see Comments) Some NR/53 13/39 (33%) 43/80 (54%) 37/82 (45%) 9/30 (30%) 10/45 (22%) 41% during TLOSRs 78% during TLOSRs 50% during TLOSRs Actigraphy, (?) (see Comments) Actigraphy, (?) (see Comments), and arousals and awakenings Only conscious awakenings (>2 min) evaluated. Actigraphic analysis criteria not stated No systematic analysis of sleep state and sleep period reflux. Actigraphic analysis criteria not clear All sleep period reflux associated with ( occurring 5 min before sleep event ) arousal/ awakening 19/21 asleep reflux followed by arousals or awakenings in next epoch No data on relationships between arousals/ awakenings and reflux Arousals/ awakenings <1 min before or after reflux in 46/67 asleep reflux. Arousals scored, but no systematic analysis given on their relationship to onset of all reflux 662 Aliment Pharmacol Ther 2013; 38:

7 Systematic review: sleep and gastro-oesophageal reflux Table 1 (Continued) Reference Oesophageal reflux detection method Reflux onset related to sleep state? Temporal resolution of reflux recording Participants Proportion of reflux during sleep LOS function Sleep recording/ analysis methods Comments Kuribayashi et al Kuribayashi et al., impedance manometry impedance manometry stated Yes Penzel et al., stated Ing et al., stated Cohen et al., stated Mello-Fujita et al Gagliardi et al., Rapid digitisation, rate not reported Rapid digitisation, rate not reported Healthy: n = 15 22/76 (29%)* GERD: n = 9 25%* (estimate) OSA without GERD: n = 6; OSA and GERD: n = 11; healthy: n = 15 2 s Suspected OSA: n = 15 (of whom n = 5 with reflux symptoms) stated Suspected OSA: n = 63; controls: n = 41 Digitisation 12 9 per second GERD: n = 11; controls: n = 7 Yes stated Asthma: n = 31; controls with sleep disturbance: n = 31 No stated Controls: n = 8; GERD: n = 15 13%* (estimate) 14%* (estimate) 33%* (estimate) 52/69 (75%), but 68/69 (99%) associated with arousal Unclear from reported data Asleep reflux all during TLOSRs (see Comments) Asleep reflux all during TLOSRs (see Comments) 54% Up to 43% Unclear from reported data, and arousals and awakenings. No data on arousals or awakenings relative to reflux, but also detailed assessment for arousals/ awakenings analysis not described Analysis concentrated on TLOSRs. All of the 76 impedancedetected reflux appear to have been during TLOSRs. explicitly stated that this was the only mechanism of reflux Analysis concentrated on TLOSRs and sleep apnoea. explicitly stated whether TLOSR was the only mechanism of reflux One reflux episode definitely occurred during stable sleep Data for temporal relationship between sleep state and reflux is unclear Limited details on analysis methods. Found no association of reflux with arousals. Data tabulation unclear Occurrence of reflux during sleep implied but detailed mechanistic data not provided Aliment Pharmacol Ther 2013; 38:

8 J. Dent et al. Table 1 (Continued) Reference Oesophageal reflux detection method Reflux onset related to sleep state? Temporal resolution of reflux recording Participants Proportion of reflux during sleep LOS function Sleep recording/ analysis methods Comments Shepherd et al manometry Suzuki et al., stated Freidin et al., manometry Dent et al., manometry Yes 1 s Patients with OSA and nocturnal GER 1 s Suspected OSAs: n = 61 (of whom n = 37 with GERD) Yes <1 s Healthy: n = 11; GERD (RO): n = 11 Yes <1 s Healthy: n = 10 (two consecutive days) 22% 40% of nocturnal GER due to TLOSRs >90% associated with arousals 2/6 (33%) No data 0/105 (0%) 42/105 (40%) scored as occurring during TLOSRs 0/130 (0%) 266/272 (98%) during TLOSRs initially, then secondby-second in GERD second-by-second Limited evaluation of temporal relationship between nocturnal GER events and arousals from sleep Analysis did not determine whether arousals preceded or followed onset of reflux Asleep reflux not re-evaluated by secondby-second sleep analysis 3/10 scored by the epochal analysis were not re-evaluated with secondby-second analysis, so have been subtracted from the total 105 acid reflux No numerical tabulation of sleep/awake state data in relation to onset of reflux. Data refer to an awake period of about 3 h, as well as the 8-h sleep period. Sleep was monitored only during the second night GER, gastro-oesophageal reflux; GERD, gastro-oesophageal reflux disease; LOS, lower oesophageal sphincter; NR, not reported; OSA, obstructive sleep apnoea; RO, reflux oesophagitis; TLOSR, transient lower oesophageal sphincter relaxation. Actigraphy was used in references 38 and 39; polysomnographic epoch-based analyses were used in references 2, and 40; at least predominantly second-by-second analyses of sleep and reflux were employed in references 21 and 24. * Impedance-detected reflux. 664 Aliment Pharmacol Ther 2013; 38:

9 Systematic review: sleep and gastro-oesophageal reflux identified as addressing other aspects of the relationship between reflux and sleep, only those judged to provide key information relevant to understanding of the mechanistic studies are cited and discussed below. Patterns of oesophageal acid exposure and numbers of reflux during awake and sleep periods Interpretation of the literature on overall patterns of day/ night reflux should take into account two major factors that are likely to influence them during the sleep period independently of physiological changes directly attributable to sleep. First, the sleep period is usually a prolonged time of fasting. Food intake, especially a large meal, is a potent stimulus for reflux. Secondly, it is possible that different recumbent body positions (i.e. supine, left lateral, right lateral or prone) could affect the risk of occurrence of reflux, but we are not aware of any data that investigate this possibility in adults during sleep. Furthermore, relative to upright, the recumbent body position has been shown to inhibit the occurrence of transient lower oesophageal sphincter (LOS) relaxations and to reduce numbers of reflux and consequent oesophageal acid exposure when food intake is taken into account. 41 These nonsleep influences complicate evaluations of the effects of sleep alone on reflux. 42 Data on oesophageal acid exposure for the sleep period vs. daytime can be safely compared among studies, as this measure is derived with a standard approach. This is not the case for the frequently reported data on reflux episode numbers or frequency. There are major differences in the and duration criteria used to score occurrence of a reflux episode among all monitoring studies and, in particular, those that deal with the sleep period. Consequently, valid comparisons of data on reflux frequency can only be made within a single study or across studies that use the same /duration criteria. The advent of impedance monitoring, which detects all reflux regardless of, has established that many reflux do not result in a recognisable change. 43 Healthy individuals. Under normal conditions of food intake and daytime physical activity, the literature is consistent in showing that acid exposure is substantially lower during the sleep period than in the awake 7, 25, 44 period. This effect is primarily due to a much lower frequency of reflux during the sleep period than prior to the first onset of sleep. 7, 25, 44, 45 Impedance monitoring has revealed that traditional acid reflux ( <4) account for a minority of all sleep period reflux and a median of 80% of these are weakly acidic (drops of that do not go below 4 43 ). 44 Patients with reflux disease. At night, most reflux in patients with GERD occurs during the first half of the sleep period. 24, 25, 46, 47 A within-study comparison found that, during the sleep period, patients with reflux oesophagitis had both an increased frequency of reflux overall and longer acid clearance times than healthy volunteers. 25 The single measure of sleep period reflux that best distinguishes reflux oesophagitis from non-erosive reflux disease appears to be the number of reflux with prolonged clearance (>5 min). 48 In their recent /impedance study, Fornari et al. 44 found that patients with GERD have three times more sleep period reflux of all types (i.e. regardless of ) than healthy individuals. Also, a larger proportion of reflux was acidic in patients with GERD compared with healthy individuals (median proportion with <4: 43% vs. 20% respectively). In patients with reflux disease, the proportion of weakly acidic reflux rose steadily from about 33% in the first recumbent hour to approximately 70% in the last recumbent hour. For individuals with non-erosive reflux disease and mild [Los Angeles (LA) grade A or B] reflux oesophagitis, acid exposure levels during sleep are relatively low compared with those occurring during the awake period, so that the 24-h pattern of acid exposure resembles that of healthy individuals, albeit at a higher level. 49 In contrast, in the small minority of patients with severe reflux oesophagitis (LA grade C or D), the day/night pattern of reflux differs substantially from that found in patients with less severe reflux disease and in healthy subjects. Patients with severe oesophagitis have very high levels of night-time acid exposure that may, in some cases, be similar to the very high daytime levels seen in this GERD subgroup Oesophageal acid clearance during the sleep period The combined suppression of salivation and swallowing during the sleep period can greatly reduce the rate at which the oesophageal luminal recovers to a non-injurious level above 4 following acid reflux (Figure 2). During arousals and awakenings, there is frequently a cluster of swallowing and oesophageal body peristalsis, so that provided the arousal/awakening is of sufficient duration, clearance of any oesophageal acidification is usually rapid (Figures 1a,b and 4). 34 Healthy individuals. Oesophageal acid clearance times in healthy subjects indicate that individual reflux Aliment Pharmacol Ther 2013; 38:

10 J. Dent et al. Apnoea Apnoea Apnoea Hypopnoea Oesophageal Swallow marker 5.1 Arousal GOR Arousal Swallow Arousal Oesophageal body pressure GOJ pressure Stomach pressure Sleep stage TLOSR Strain REM MVT S Time (s) Figure 4 Tracing of oesophageal motility (LOS pressure monitored with a sleeve sensor) and with superimposed sleep and apnoea analyses (polysomnographic raw tracing not shown) in a patient with obstructive sleep apnoea. This illustrates the potential for physiological and temporal complexity of sleep, arousals, oesophageal motility, apnoea and body movements around the time of a sleep period reflux episode. The sleep analysis at the bottom of the figure used the epochal approach. Three arousals occurred, as shown at the top of the figure, but they were too short to be reflected in the epochal sleep analysis. The transient LOS relaxation that leads to GER, although associated with a brief arousal, would have been scored by the epochal method as starting during stable REM sleep. Adapted from Kuribayashi et al. 34 with permission. GOJ, gastro-oesophageal junction pressure; GER, gastro-oesophageal reflux; LOS, lower oesophageal sphincter; MVT, movement; REM, rapid eye movement; S1, sleep stage 1; TLOSR, transient lower oesophageal relaxation. during the sleep period result in a very wide range of acid exposure times and a longer average time for recovery of oesophageal to 4 compared with during the daytime. 21, 25 Monitoring of swallowing, oesophageal body motility, sleep and oesophageal in healthy individuals has shown how the sleep-related absence of swallowing can lead to long oesophageal acidification times following reflux during the sleep period (Figure 2). 21 Patients with reflux disease. The study by Dodds et al., 25 in 20 patients with reflux oesophagitis, and 15 healthy control subjects, provides detailed comparative data on patterns and mechanics of oesophageal acid clearance. Consistent with other studies of clearance during wakefulness, acid clearance time is, on average, substantially longer in patients with GERD than in controls. Dodds et al. found that clearance of reflux during wakefulness was more than three times longer in patients with reflux oesophagitis than in healthy controls. 25 Unfortunately, because sleep was not monitored formally in this study, detailed data are not available on the interaction of resumed sleep after a reflux episode during the sleep period. However, for the night-time recording period, very prolonged periods of acidification were seen, so that oesophageal acid clearance to 4 took longer than 5 min in 15% of reflux in the patients with oesophagitis and in 9% of reflux in the control subjects. Even longer periods of acidification of at least 10 min accounted for 6% and 3% of the in the patients with oesophagitis and the controls respectively. The rate of swallowing during individual periods of acidification varied widely, but differed little overall between the two groups. For individual, swallow frequency correlated with acid clearance time, but the much higher rate of failure of swallow-induced peristalsis in patients with oesophagitis (27% vs. 7% in controls) is a plausible major contributor to the slow acid clearance in the patients with oesophagitis. Suppression of primary peristalsis because of resumption of sleep before acid clearance was completed is likely to have augmented the effect of failed primary peristalsis on acid clearance. A pronounced failure of triggering of secondary peristalsis in response to oesophageal distension has been documented in daytime 666 Aliment Pharmacol Ther 2013; 38:

11 Systematic review: sleep and gastro-oesophageal reflux studies in patients with GERD. 52 The finding of Dodds et al. 25 that, in patients with oesophagitis, secondary peristalsis was the first oesophageal body motor event following reflux in only 17% of, compared with 57% in the controls, suggests that failed triggering of secondary peristalsis also contributes to the slow acid clearance seen in GERD. Given the evidence that slow acid clearance is likely to play a significant role in provocation of sleep period reflux symptoms, it is of interest to note that bed head elevation or sleeping on a wedge has been shown consistently to be associated with a modest reduction in sleep period oesophageal acid exposure when compared with sleeping recumbent. This occurs because of reduction in the time taken for oesophageal acid clearance, presumably because of the effects of gravity. 53, 54 Unfortunately, this effect appears too small to result in a major reduction in sleep period symptoms when used on its own. None of the studies sought to determine whether sleep quality was improved by postural therapy. Polysomnography was used by Freidin et al., 24 in their study of patients with GERD, but minimal data are presented on events relevant to clearance. Studies that monitor and correlate sleep, oesophageal, occurrence of symptoms and oesophageal motor events in well-characterised patients with GERD and healthy controls are needed. These would be likely to advance substantially the mechanistic understanding of factors that contribute to the development of reflux-induced symptoms during the sleep period. Factors determining the development of refluxinduced symptoms during the sleep period in reflux disease Apart from the sleep period oesophageal acid infusion 22, 23, 25 studies of Orr and colleagues (see Introduction), our searches did not reveal any detailed analyses of the characteristics of sleep period reflux that are associated with reflux-induced symptoms, by comparison with the great majority of that are asymptomatic. The acid infusion studies indicate that both the duration of oesophageal acid exposure arising from a single acidification episode 22, 23 and the of the refluxate 26 are relevant. Studies of the association of symptoms with individual reflux in the awake state by Bredenoord et al. 55 are consistent with the findings of 22, 23, 26 Orr et al., in that a high cumulative level of prior acid exposure, larger drops of to lower nadirs, higher extent of flow of refluxate up the oesophagus and prolonged clearance times were associated with symptomatic reflux, compared with asymptomatic. We have found no such data in relation to the sleep period. This is an area of research that merits attention. RESULTS AND DISCUSSION PART II: STATE AND ANTIREFLUX FUNCTION AT THE TIME OF REFLUX Important methodological considerations Transitions from sleep to arousals and awakenings occur within 1 s (Figure 2), sometimes with considerable temporal complexity (Figure 4). 21, 34 It is important to document these events from second to second around the time of reflux onset because of the general agreement that arousals/awakenings are closely associated with the timing of reflux. Accurate resolution of the timings of both sleep state and reflux onset is required to determine whether arousals/awakenings could be triggered by reflux or vice versa. Factors influencing the accuracy of reflux episode timing measurements The time of onset of reflux is the pivot point for determining the relationship between sleep state and the occurrence of reflux. In seven of the 17 studies investigating this relationship (Table 1), the onset of reflux was used as the reference time for correlations with 21, 24, sleep. Three other studies appear to have taken the entire time that oesophageal was below 4 as the reference period In the remaining seven studies, the criteria used to judge the timing of reflux from the tracing were not stated, 2, thus data on the relationship between the reflux events and sleep cannot be adequately evaluated in these studies. The three mechanistic studies that used the total time the was below 4 to define reflux referred to this as the duration of reflux, 38, 39 or as a reflux event. 40 These terminologies fail to distinguish between the event of reflux and the impacts on oesophageal that persist after cessation of flow across the gastro-oesophageal junction. The true period of reverse flow of gastric contents into the oesophagus usually occurs over only a few seconds, being cut off by the prompt re-establishment of the gastro-oesophageal junction pressure barrier, which is normally contributed to by the LOS and diaphragmatic crura (Figure 2). Importantly, this cessation of flow usually occurs considerably earlier than the return of the to 4, meaning that the true onset of reflux cannot be accurately defined with such an approach. A more correct Aliment Pharmacol Ther 2013; 38:

12 J. Dent et al. description of the time taken for the to return to 4 is oesophageal clearance time or duration of oesophageal acidification. Another potential source of temporal inaccuracy when defining reflux onset arises from the use of early digital recording systems with slow sampling rates, often as low as one value every 5 s (Table 1). Such a sampling rate could delay recognition of the onset of a reflux episode by up to 5 s: this is a significant limitation when attempts are made to align reflux events temporally with other, frequently rapidly evolving, physiological events such as an arousal from sleep. Of the 14 2, 27 33, studies in which the trace was digitised, a 5-s digitisation interval was explicitly stated as being used in one publication, 38 and was probably also used in another report by the same group. 39 In another six studies, the stated digitisation rate was suited to mechanistic sleep studies as was sampled at least every 2 s (Table 1). 28, 29, 31, 32, 35, 36 In three studies, the analogue signal was recorded from the monitoring equipment, 21, 24, 34 capturing a true second-to-second record. Methods used for analysis of sleep state around the time of reflux onset in sleep and reflux studies The standard, pragmatic method used to gain an overview of sleep architecture over the entire sleep period for actigraphic and polysomnography recordings is a time block or epochal analysis. In the case of actigraphy, the only possible judgement to make is whether body movement (the proxy for awake) or stillness (the proxy for sleep) predominates within each epoch. In the case of polysomnography, a single stage of sleep or wakefulness is assigned to each epoch by identifying the sleep state that occupies the most time within each epoch. Polysomnographic analyses have used a range of epoch lengths, mainly from 20 to 60 s. Importantly, arousals and shorter periods of awakening are not recognised by this routine epoch-based reporting method (Figure 4). Thus, an epoch classified as sleep could contain significant unscored intervals of wakefulness or arousals. A reflux episode that occurs during such an epoch would, however, be scored as taking place during sleep even if its onset was during an arousal, or towards the end of an awakening. It follows that mechanistic studies of sleep and reflux using polysomnography require a special approach to sleep analysis that evaluates the sleep state on a second-by-second basis around the time of onset of each reflux episode. Only two of the 17 studies examining the relationship between reflux episode and sleep state used this approach to derive their main reported data. 21, 24 Five other studies, however, supplemented an epochal sleep analysis with a second-by-second approach (as shown in Figure 4) to define the timings of arousals/ awakenings, but did not report data derived with this alternative approach in adequate detail to determine 28, 29, sleep state reliably at the time of reflux onset. Mechanistic studies of sleep and reflux using actigraphy The two actigraphic studies discussed below evaluated patterns of reflux only around periods of awakening 38, 39 >2 min, which were called conscious awakenings. The rationale for such a classification was that these awakenings are able to be recalled at the end of the sleep period, while awakenings shorter than 2 min are not. 38, 39 This approach is unfortunately not supported by the publications cited to justify this subdivision of 56, 57 awakenings. Quite apart from the validity of the concept of conscious awakenings, it is illogical for this subdivision to determine which periods of a recording should be analysed for the relationships between sleep and reflux. Healthy individuals. We identified one /actigraphic study, by Poh et al., from which data were published on the relationship between sleep state and reflux in healthy volunteers. 38 Nine individuals were studied and the analysis was limited to conscious awakenings (see above). While such awakenings were detected in seven of the nine participants, no acid reflux were recorded during these times. Patients with reflux disease. A total of 39 patients with 38, 39 GERD were evaluated in two actigraphic studies. These papers, from the same research group, probably refer to the same patient population, although this is not explicitly stated. In one study by Poh et al., 39 data are reported on overall patterns of reflux frequency and acid exposure in relation to recumbency before and after the first episode of stable sleep, and it is stated that some reflux occurred during sleep. In their other study, Poh et al. found that acid reflux was associated with 54 of the 104 of these conscious awakenings, with 14.4% of these reflux occurring during sleep. 38 However, this conclusion is likely to be overstated, because the duration of reflux was defined as the time that oesophageal was below 4 (an analysis approach that does not evaluate sleep state at the time of reflux onset) and because actigraphy cannot be used to document the timings of intervals of stable sleep reliably. 668 Aliment Pharmacol Ther 2013; 38:

13 Systematic review: sleep and gastro-oesophageal reflux Mechanistic studies of sleep and reflux using polysomnography The use of polysomnography presents significant practical problems for mechanistic research because it is a cumbersome, relatively expensive and intrusive recording method. However, it remains the current gold standard with which to assess sleep comprehensively and is, therefore, the most accurate method for correlating sleep state with reflux. Our searches identified 15 mechanistic studies that used this method (Table 2, 21, 24, 27 37, 40 1). Overall comparisons of sleep architecture in patients with reflux disease vs. healthy volunteers There is a notable lack of within-study polysomnographic comparisons of detailed sleep architecture between patients with reflux disease and healthy volunteers, presumably because of the significant challenges of such a study and the complexity of this recording method. Epoch-based analyses of polysomnographic recordings and reflux We identified 13 studies that relied primarily or solely on epoch-based sleep analyses of polysomnographic 2, 27 37, 40 recordings. With this approach, these studies found that 13 75% of sleep period reflux occurred during actual sleep in healthy individuals, patients with sleep disorders and those with reflux disease (Table 1). Four of the 13 primarily epoch-based studies supplemented their analysis with a second-by-second scoring of arousals/awakenings around the time of onset of reflux, 28, 29, 34, 36 but incomplete or temporally imprecise data were reported for these secondary analyses. For reflux scored as occurring during sleep on the basis of epochal analysis, 68 99% were reported as being associated (defined variably) with arousals or awakenings. The illustration of both epoch-based and second-by-second sleep analyses around a sleep period reflux episode from Kuribayashi et al. 34 (Figure 4) reveals the inadequacies of assessing sleep state at the time of onset of reflux using an epoch-based approach. Apart from the study by Orr et al., 33 the primarily epoch-based studies evaluated relatively small numbers of individuals. Results from participants both with and without GERD are reported. Some studies included individuals with or without subjectively sleep disturbance or with obstructive sleep apnoea (OSA), with or without GERD. In several of these studies, a close association (defined variably see Table 1) was noted between of wakefulness and reflux, but this was usually interpreted as evidence that the wakefulness was caused by the reflux. Second-by-second analyses of polysomnographic recordings and reflux Healthy individuals. Only one study of healthy volunteers was identified in which polysomnographic recordings around the time of reflux onset were systematically analysed. 21 Over two consecutive nights (each night being a 12-h period encompassing an evening meal, an awake postprandial period and a sleep period of about 8 h), 272 reflux were recorded in 10 individuals. Data on reflux onset and sleep state were confined to the second study night because polysomnography was only conducted on that night. A detailed numerical analysis of the sleep/awake state at the onsets of the approximately 130 acid reflux recorded during sleep monitoring on the second night is not provided in that report, but the results section states that the onset of acid reflux occurred only during arousals or more prolonged periods of wakefulness. Figure 2 illustrates one of the reflux recorded in this study during polysomnography. Patients with reflux disease. Freidin et al. concluded from their study of 11 patients with GERD that wakefulness or arousals were necessary conditions for reflux to occur during the sleep period. 24 An initial epochal analysis found that 92 of the 105 reflux had their onset during wakefulness and 13 during sleep. The polysomnography recordings for the 10 scored as occurring during stage II NREM sleep were re-examined on a second-by-second basis, revealing that in all cases, reflux onset took place during brief arousals that were not recognised by the 60-s epoch scoring approach. No re-examination of the other three scored as occurring during other stages of sleep was mentioned. Due to the relatively small sample of 11 patients with reflux oesophagitis in the study by Freidin et al., 24 it was not possible to evaluate sleep state/reflux relationships according to clinically important subgroups of patients with different mixes of pathophysiological causes of GERD. 58 Although Freidin et al. did not describe the spectrum of severity of reflux oesophagitis in their patients with reflux disease, 24 it is likely that few would have had severe reflux oesophagitis and the associated very high levels of sleep period reflux. 49, 51 Patients with LA grade C or Dreflux oesophagitis typically have large hiatus hernias, extreme laxity of the diaphragmatic crura and a particularly low basal LOS pressure. 58 Such patients are most at Aliment Pharmacol Ther 2013; 38:

14 J. Dent et al. risk of occurrence of reflux during stable sleep and hence should be studied as a separate group. Patients with snoring and/or OSA reflux disease status unknown. Penzel et al. used a mixture of epochbased sleep analysis and second-by-second scoring of arousals for correlating sleep with reflux in 15 patients with OSA and/or snoring problems, but the data reported did not include a second-by-second correlation of arousals or awakenings with reflux onset. 29 Their report is, however, of particular interest as they provide an illustration of a polysomnographic and tracing during an atypically gradual onset of apparent oesophageal acidification that occurred during clearly documented stable sleep (Figure 5). There are two possible explanations for this event. First, the apparent oesophageal acidification could be a recording artefact due to downwards displacement of the electrode into the stomach, suggested by the unusual time pattern of the decrease, or secondly, it could be a true reflux event with mechanics different from those of the great majority of reflux. This was the only episode of apparent oesophageal acidification that Penzel et al. found to occur without associated arousal or awakening out of a total of 69 reflux. 29 If this was a true reflux episode, to our knowledge, it is the only one that has been clearly documented to occur in an adult during stable sleep. It remains unclear whether this individual had GERD. Mechanical function of the gastro-oesophageal junction associated with the onset of reflux in the sleep period Six studies investigated the mechanical factors underlying the abrupt occurrence of reflux during the sleep period. 21, 24, 28, In this regard, the patterns of motor activity of the LOS and diaphragm are crucial measures. To date, there has been little emphasis on distinguishing diaphragmatic function from that of the LOS in studies during the sleep period. The study by Kuribayashi et al. 35 is the only one that has been identified as commenting on diaphragmatic function during sleep period reflux. These authors reported that transient LOS relaxations associated with OSA were associated with manometric patterns consistent with potent crural diaphragmatic inhibition; no such inhibition was seen for apnoeic occurring independently of transient relaxations. 35 This illustrates the possibility that the diaphragmatic inhibitory component of transient LOS relaxations 58 could play a mechanical role in sleep period reflux. LOS function during sleep period reflux Six studies used or /impedance monitoring with oesophageal manometry to evaluate motor patterns at the time of sleep period reflux. 21, 24, 28, Transient LOS relaxations were judged to account for % of sleep period reflux in healthy individuals or 21, 28, patients with OSA. Similarly, in studies of patients with reflux disease, it was concluded that EEG EOG EMG Flow Thorax Abdomen 100 SaO 2 % 50 Snoring Time (min) Figure 5 Recording of the single published instance of reflux that had its onset during clearly demonstrated stable sleep, in this case associated with snoring. The polysomnography recording in the upper three panels shows no sign of arousal or awakening. Oesophageal acidification has an atypically gradual evolution compared with the abrupt acidifications seen in most sleep period reflux (see Figures 2 and 4). Adapted from Penzel et al. 29 with permission. EEG, electroencephalogram; EMG: electromyogram; EOG, electro-oculogram; Flow, oronasal airflow;, oesophageal ; SaO 2, arterial oxygen saturation. 670 Aliment Pharmacol Ther 2013; 38: