Sleep Breath DOI 10.1007/s11325-014-1108-4
ORIGINAL ARTICLE
Improving detection of obstructive sleep apnoea by overnight oximetry in children using pulse rate parameters Dg Zuraini Sahadan & Margot J. Davey & Rosemary S. C. Horne & Gillian M. Nixon
Received: 18 November 2014 / Accepted: 23 December 2014 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Overnight oximetry is a simple tool for investigation of obstructive sleep apnoea (OSA) in children, but only severe cases will be detected, and children with obstructive events resulting in arousal, but not desaturation, will have a normal (inconclusive) result. We hypothesised that pulse rate rises using pulse rate indices per hour (PRI) and pulse rate standard deviation (PR-SD) automatically calculated from commercially available software would improve oximetry as a diagnostic tool. Methods Children having home overnight oximetry for suspected OSA were identified over 12 months, and those with a normal result who went on to have polysomnography (PSG) were included. Oximetry, including PR-SD and PRI (rises of 8, 10 and 15 beats/min per hour), was analyzed using commercially available software. PR parameters were compared between those with OSA (obstructive apnoeahypopnoea index (OAHI) >1 event/h) and those without OSA. Results One hundred sixteen children had normal oximetry, of whom 93 (median age 4.5 years; 55 % M) had PSG. Fiftyseven of 93 (61 %) children had OSA (median OAHI 4.5 events/h, range 1.1–24). PR-SD was not different between the OSA and non-OSA groups (p=0.87). PRI tended to be
D. Z. Sahadan Paediatric Respiratory Unit, Serdang Hospital, Selangor Darul Ehsan, Malaysia M. J. Davey : G. M. Nixon Melbourne Children’s Sleep Centre, Monash Children’s Hospital, Monash Health, Melbourne, Victoria, Australia M. J. Davey : R. S. C. Horne : G. M. Nixon (*) The Ritchie Centre, MIMR-PHI Institute of Medical Research and Department of Paediatrics, Monash University, Melbourne, Victoria 3168, Australia e-mail: [email protected]
higher in those with OSA, but there was considerable overlap between the groups: PRI-8 (mean ± SD 58.5±29.0/h in OSA group vs 48.6±20.2/h in non-OSA group, p=0.07), PRI-10 (45.1±25.0 vs 36.2±16.7, p=0.06) and PRI-15 (24.4±14.5 vs 18.9±9.0, p=0.04). A PRI-15 threshold of >35/h had specificity of 97 % for OSA. Conclusion The PRI-15 shows promise as an indicator of OSA in children with normal oximetry. Keywords Oximetry . Pulse rate . Obstructive sleep apnoea . Child
Introduction Obstructive sleep apnoea (OSA) is a common disorder, affecting 1–5 % of children [1, 2]. OSA is a major health issue in childhood, with significant impacts on cognition, behaviour and cardiovascular health [3]. The cardinal symptom of OSA is snoring. Up to 28 % of children are reported to snore often or always [3], but as only about 10 % of snoring children (1– 5 % of the population) will have OSA, formally defining the presence of OSA in a snoring child requires polysomnography (PSG), a technically challenging and expensive test that is not widely available worldwide. Overnight oximetry involves the continuous measurement of oxygen saturation and pulse rate using a non-invasive probe that is attached to the finger or toe with tape. It can be easily used by parents at home and downloaded into analysis software the following day. Overnight oximetry has a high positive predictive value for the presence of OSA in a snoring child if it demonstrates clusters of dips in oxygen saturation (SpO2) as shown in Fig. 1 [4]. The frequency of dips in SpO2 below different saturation thresholds have been shown to be related to severity of OSA
Sleep Breath
Fig. 1 Overnight recording of SpO2 and pulse, showing typical clusters of desaturation (red, A) and recurrent surges in pulse rate (blue, B). SpO2 (red) and pulse rate (blue) are in separate panels, with time marked in hourly intervals along the x-axis
(The McGill score [5]). However, oximetry has limited value as a screening tool for OSA amongst snoring children because of its low sensitivity—in a referred population, about half of the children with a normal oximetry result have OSA [4, 5]. Oximetry therefore misses many children with significant OSA who have frequent arousal from sleep without desaturation. Also, oximetry is abnormal (McGill score 2–4) in only about 20 % of children with OSA [4, 5]. Therefore children with suspected OSA and normal oximetry (McGill score 1) require more detailed testing to confirm or exclude the diagnosis. The episodes of upper airway obstruction characteristic of OSA are usually terminated by an arousal from sleep or brief awakening, even in the absence of desaturation [6, 7]. This arousal is often associated with movement and a surge in heart rate and blood pressure, with rises in heart rate in the order of 15 beats/min [8, 9] as illustrated in Fig. 2. This finding has lead to an interest in analysis of heart rate patterns during sleep as a way of detecting arousals related to respiratory events that do not cause desaturation [10]. This study aimed to assess whether quantifying pulse rate surges per hour (pulse rate indices (PRI)) and pulse rate standard deviation (PR-SD) would detect children with OSA who have a normal oximetry, thereby potentially improving oximetry as a diagnostic tool for OSA in children.
for overnight oximetry performed at home from the beginning of June 2009 to the end of May 2010. From this group, only those referred for evaluation of possible OSA in whom the oximetry result was deemed to be normal based on the McGill oximetry score were included. Using that score, an abnormal oximetry is defined as the presence of ≥3 clusters of desaturation (five or more dips of 4 % in a 10–30-min period) and ≥3 dips in oxygen saturation below 90 % during the night [5]. Children with an abnormal oximetry were referred for treatment. Children with a normal oximetry went on to have a full attended PSG to determine the presence and severity of OSA and were included in the final analysis. Oximetry (Masimo Radical with 2-s averaging time, Masimo Corp., CA, USA) was analysed using Download 2001 software (Stowood Scientific Instruments, Oxford, UK). No manual modifications to the data were made, such as removal of artefact or periods of movement/wakefulness. PR parameters are automatically calculated by this software and details of how the pulse rate rises are calculated are not available. The extent of pulse rate rise that is quantified is selected by the user. For this study, we selected the following pulse rate parameters based on evidence of the magnitude of relevant pulse rate rises from the literature: &
Methods
&
A retrospective study was conducted, with inclusion of all patients referred to the Melbourne Children’s Sleep Centre
&
Pulse rate standard deviation for the study period (PR-SD) [11] Pulse rate increases of 8 bpm (PRI-8) to reflect baseline pulse rate fluctuations during quiet periods of sleep [12] Pulse rate increases of 10 bpm (PRI-10) to reflect average pulse rate difference between wake and sleep [13]
Sleep Breath
Fig. 2 An obstructive hypopnoea showing desaturation of less than 4 % (A) but an increase in heart rate (B) associated with the recovery period from the event
&
Pulse rate increases of 15 bpm (PRI-15) typical of arousal from sleep [8, 9]
These indices are reported by the Download 2001 software as pulse rate rises of the expressed amount per hour of recording. PSG was carried out in an attended laboratory setting using a commercially available PSG system (E-Series, Compumedics, Melbourne, Australia) and standard pediatric recording techniques [14]. Recorded parameters included electroencephalogram (Cz, C4-M1, C3-M2, O2-M1, O1-M2 (≥4 years); Cz, C4-M1, C3-M2 (1.0 event/h [15, 16]. Pulse rate parameters were compared between groups using Student’s t tests. Given the higher prevalence of OSA in pre-school children, we then repeated these analyses confined to those aged ≤5 years. Correlation analysis was performed between pulse rate indices and OAHI and total arousal index on PSG. Sensitivity, specificity and positive and negative predictive values for the prediction of the presence of OSA were calculated for a range of binary cut points of PRI-15.
Results We identified 177 overnight oximetry tests performed in the study period. Thirty-three children did not meet inclusion criteria (14 had known central sleep apnoea; eight were on supplemental oxygen, continuous positive airway pressure or non-invasive ventilation; seven had less than 4-h recording; two were for follow-up of previous results; and two were for post-adenotonsillectomy assessment). Of the remaining 144 children who had oximetry for testing for possible OSA, 28 children (19 %) had an abnormal oximetry with a McGill score of 2 (n=12), 3 (n=8) or 4 (n=8). These children did not have further testing except for one 12-month-old child with a McGill score of 2 who went on to have PSG (OAHI 4.3/h) but was excluded from further analysis given the abnormal oximetry result. A total of 116 patients had a normal oximetry result, of whom 93 went on to have PSG during the study period and formed the study population. The median age of the children was 4.5 years (range 0.8– 16.5 years), with 55 % male. Ninety-one children had no co-morbidity; one child had Down syndrome and one had Beckwith-Wiedemann syndrome. The median interval between oximetry and PSG was 1.2 months (range 1 week to 6 months). Fifty-six patients (60 %) were diagnosed with OSA (OAHI median 4.5 events/h, range 1.1–24 events/h). In the whole group (n=93), PR-SD was not different between the OSA and non-OSA groups (mean ± SD 9.1 ±1.7 in both groups, p=0.87). PR indices were consistently higher in those with OSA (Table 1), reaching statistical significance for PRI-15 (Fig. 3). There was however a large overlap in PRI values between groups. PRIs were not significantly correlated with the OAHI or total arousal index on PSG (strongest relationship for PRI-15 and total arousal index: r2 =0.04, p=0.06).
Sleep Breath Table. 1 Pulse rate parameters (mean ± SD) for all subjects without and with OSA
Table 2 Pulse rate parameters (mean ± SD) for children aged ≤5 years without and with OSA
PR metric
Non-OSA (n=36)
OSA (n=57)
p value
PR
Non-OSA (n=22)
OSA (n=31)
p value
PR-SD PRI-8 PRI-10 PRI-15
9.1±1.7 48.6±20.2 36.2±16.7 18.9±9.0
9.1±1.7 58.5±29.0 45.1±25.0 24.4±14.5
0.87 0.07 0.06 0.04
PR-SD PRI-8 PRI-10 PRI-15
9.5±1.6 53.6±15.9 39.8±12.8 20.3±6.3
9.8±1.5 69.9±28.8 54.9±25.3 30.0±14.9
0.49 0.02 0.01 0.006
When the PR parameters were analyzed in the children e5 years of age only, the differences in PR parameters between those with and without OSA were greater (Table 2). PR-SD still did not distinguish between groups with and without OSA; however, PRI-8, PRI-10 and PRI-15 were all significantly higher in the OSA group (p
ORIGINAL ARTICLE
Improving detection of obstructive sleep apnoea by overnight oximetry in children using pulse rate parameters Dg Zuraini Sahadan & Margot J. Davey & Rosemary S. C. Horne & Gillian M. Nixon
Received: 18 November 2014 / Accepted: 23 December 2014 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Overnight oximetry is a simple tool for investigation of obstructive sleep apnoea (OSA) in children, but only severe cases will be detected, and children with obstructive events resulting in arousal, but not desaturation, will have a normal (inconclusive) result. We hypothesised that pulse rate rises using pulse rate indices per hour (PRI) and pulse rate standard deviation (PR-SD) automatically calculated from commercially available software would improve oximetry as a diagnostic tool. Methods Children having home overnight oximetry for suspected OSA were identified over 12 months, and those with a normal result who went on to have polysomnography (PSG) were included. Oximetry, including PR-SD and PRI (rises of 8, 10 and 15 beats/min per hour), was analyzed using commercially available software. PR parameters were compared between those with OSA (obstructive apnoeahypopnoea index (OAHI) >1 event/h) and those without OSA. Results One hundred sixteen children had normal oximetry, of whom 93 (median age 4.5 years; 55 % M) had PSG. Fiftyseven of 93 (61 %) children had OSA (median OAHI 4.5 events/h, range 1.1–24). PR-SD was not different between the OSA and non-OSA groups (p=0.87). PRI tended to be
D. Z. Sahadan Paediatric Respiratory Unit, Serdang Hospital, Selangor Darul Ehsan, Malaysia M. J. Davey : G. M. Nixon Melbourne Children’s Sleep Centre, Monash Children’s Hospital, Monash Health, Melbourne, Victoria, Australia M. J. Davey : R. S. C. Horne : G. M. Nixon (*) The Ritchie Centre, MIMR-PHI Institute of Medical Research and Department of Paediatrics, Monash University, Melbourne, Victoria 3168, Australia e-mail: [email protected]
higher in those with OSA, but there was considerable overlap between the groups: PRI-8 (mean ± SD 58.5±29.0/h in OSA group vs 48.6±20.2/h in non-OSA group, p=0.07), PRI-10 (45.1±25.0 vs 36.2±16.7, p=0.06) and PRI-15 (24.4±14.5 vs 18.9±9.0, p=0.04). A PRI-15 threshold of >35/h had specificity of 97 % for OSA. Conclusion The PRI-15 shows promise as an indicator of OSA in children with normal oximetry. Keywords Oximetry . Pulse rate . Obstructive sleep apnoea . Child
Introduction Obstructive sleep apnoea (OSA) is a common disorder, affecting 1–5 % of children [1, 2]. OSA is a major health issue in childhood, with significant impacts on cognition, behaviour and cardiovascular health [3]. The cardinal symptom of OSA is snoring. Up to 28 % of children are reported to snore often or always [3], but as only about 10 % of snoring children (1– 5 % of the population) will have OSA, formally defining the presence of OSA in a snoring child requires polysomnography (PSG), a technically challenging and expensive test that is not widely available worldwide. Overnight oximetry involves the continuous measurement of oxygen saturation and pulse rate using a non-invasive probe that is attached to the finger or toe with tape. It can be easily used by parents at home and downloaded into analysis software the following day. Overnight oximetry has a high positive predictive value for the presence of OSA in a snoring child if it demonstrates clusters of dips in oxygen saturation (SpO2) as shown in Fig. 1 [4]. The frequency of dips in SpO2 below different saturation thresholds have been shown to be related to severity of OSA
Sleep Breath
Fig. 1 Overnight recording of SpO2 and pulse, showing typical clusters of desaturation (red, A) and recurrent surges in pulse rate (blue, B). SpO2 (red) and pulse rate (blue) are in separate panels, with time marked in hourly intervals along the x-axis
(The McGill score [5]). However, oximetry has limited value as a screening tool for OSA amongst snoring children because of its low sensitivity—in a referred population, about half of the children with a normal oximetry result have OSA [4, 5]. Oximetry therefore misses many children with significant OSA who have frequent arousal from sleep without desaturation. Also, oximetry is abnormal (McGill score 2–4) in only about 20 % of children with OSA [4, 5]. Therefore children with suspected OSA and normal oximetry (McGill score 1) require more detailed testing to confirm or exclude the diagnosis. The episodes of upper airway obstruction characteristic of OSA are usually terminated by an arousal from sleep or brief awakening, even in the absence of desaturation [6, 7]. This arousal is often associated with movement and a surge in heart rate and blood pressure, with rises in heart rate in the order of 15 beats/min [8, 9] as illustrated in Fig. 2. This finding has lead to an interest in analysis of heart rate patterns during sleep as a way of detecting arousals related to respiratory events that do not cause desaturation [10]. This study aimed to assess whether quantifying pulse rate surges per hour (pulse rate indices (PRI)) and pulse rate standard deviation (PR-SD) would detect children with OSA who have a normal oximetry, thereby potentially improving oximetry as a diagnostic tool for OSA in children.
for overnight oximetry performed at home from the beginning of June 2009 to the end of May 2010. From this group, only those referred for evaluation of possible OSA in whom the oximetry result was deemed to be normal based on the McGill oximetry score were included. Using that score, an abnormal oximetry is defined as the presence of ≥3 clusters of desaturation (five or more dips of 4 % in a 10–30-min period) and ≥3 dips in oxygen saturation below 90 % during the night [5]. Children with an abnormal oximetry were referred for treatment. Children with a normal oximetry went on to have a full attended PSG to determine the presence and severity of OSA and were included in the final analysis. Oximetry (Masimo Radical with 2-s averaging time, Masimo Corp., CA, USA) was analysed using Download 2001 software (Stowood Scientific Instruments, Oxford, UK). No manual modifications to the data were made, such as removal of artefact or periods of movement/wakefulness. PR parameters are automatically calculated by this software and details of how the pulse rate rises are calculated are not available. The extent of pulse rate rise that is quantified is selected by the user. For this study, we selected the following pulse rate parameters based on evidence of the magnitude of relevant pulse rate rises from the literature: &
Methods
&
A retrospective study was conducted, with inclusion of all patients referred to the Melbourne Children’s Sleep Centre
&
Pulse rate standard deviation for the study period (PR-SD) [11] Pulse rate increases of 8 bpm (PRI-8) to reflect baseline pulse rate fluctuations during quiet periods of sleep [12] Pulse rate increases of 10 bpm (PRI-10) to reflect average pulse rate difference between wake and sleep [13]
Sleep Breath
Fig. 2 An obstructive hypopnoea showing desaturation of less than 4 % (A) but an increase in heart rate (B) associated with the recovery period from the event
&
Pulse rate increases of 15 bpm (PRI-15) typical of arousal from sleep [8, 9]
These indices are reported by the Download 2001 software as pulse rate rises of the expressed amount per hour of recording. PSG was carried out in an attended laboratory setting using a commercially available PSG system (E-Series, Compumedics, Melbourne, Australia) and standard pediatric recording techniques [14]. Recorded parameters included electroencephalogram (Cz, C4-M1, C3-M2, O2-M1, O1-M2 (≥4 years); Cz, C4-M1, C3-M2 (1.0 event/h [15, 16]. Pulse rate parameters were compared between groups using Student’s t tests. Given the higher prevalence of OSA in pre-school children, we then repeated these analyses confined to those aged ≤5 years. Correlation analysis was performed between pulse rate indices and OAHI and total arousal index on PSG. Sensitivity, specificity and positive and negative predictive values for the prediction of the presence of OSA were calculated for a range of binary cut points of PRI-15.
Results We identified 177 overnight oximetry tests performed in the study period. Thirty-three children did not meet inclusion criteria (14 had known central sleep apnoea; eight were on supplemental oxygen, continuous positive airway pressure or non-invasive ventilation; seven had less than 4-h recording; two were for follow-up of previous results; and two were for post-adenotonsillectomy assessment). Of the remaining 144 children who had oximetry for testing for possible OSA, 28 children (19 %) had an abnormal oximetry with a McGill score of 2 (n=12), 3 (n=8) or 4 (n=8). These children did not have further testing except for one 12-month-old child with a McGill score of 2 who went on to have PSG (OAHI 4.3/h) but was excluded from further analysis given the abnormal oximetry result. A total of 116 patients had a normal oximetry result, of whom 93 went on to have PSG during the study period and formed the study population. The median age of the children was 4.5 years (range 0.8– 16.5 years), with 55 % male. Ninety-one children had no co-morbidity; one child had Down syndrome and one had Beckwith-Wiedemann syndrome. The median interval between oximetry and PSG was 1.2 months (range 1 week to 6 months). Fifty-six patients (60 %) were diagnosed with OSA (OAHI median 4.5 events/h, range 1.1–24 events/h). In the whole group (n=93), PR-SD was not different between the OSA and non-OSA groups (mean ± SD 9.1 ±1.7 in both groups, p=0.87). PR indices were consistently higher in those with OSA (Table 1), reaching statistical significance for PRI-15 (Fig. 3). There was however a large overlap in PRI values between groups. PRIs were not significantly correlated with the OAHI or total arousal index on PSG (strongest relationship for PRI-15 and total arousal index: r2 =0.04, p=0.06).
Sleep Breath Table. 1 Pulse rate parameters (mean ± SD) for all subjects without and with OSA
Table 2 Pulse rate parameters (mean ± SD) for children aged ≤5 years without and with OSA
PR metric
Non-OSA (n=36)
OSA (n=57)
p value
PR
Non-OSA (n=22)
OSA (n=31)
p value
PR-SD PRI-8 PRI-10 PRI-15
9.1±1.7 48.6±20.2 36.2±16.7 18.9±9.0
9.1±1.7 58.5±29.0 45.1±25.0 24.4±14.5
0.87 0.07 0.06 0.04
PR-SD PRI-8 PRI-10 PRI-15
9.5±1.6 53.6±15.9 39.8±12.8 20.3±6.3
9.8±1.5 69.9±28.8 54.9±25.3 30.0±14.9
0.49 0.02 0.01 0.006
When the PR parameters were analyzed in the children e5 years of age only, the differences in PR parameters between those with and without OSA were greater (Table 2). PR-SD still did not distinguish between groups with and without OSA; however, PRI-8, PRI-10 and PRI-15 were all significantly higher in the OSA group (p
