CC BY-NC-ND
0Egyptian Journal of Chest Diseases and Tuberculosis xxx (2016) xxx-xxx
HOSTED BY
Contents lists available at ScienceDirect
Egyptian Journal of Chest Diseases and Tuberculosis
journal homepage: www.sciencedirect.com
Optimal level of continuous positive airway pressure: Auto-CPAP titration versus predictive formulas
Nashwa Abdel Wahaba'*, Yasser Noeman Ahmed b
a Chest Diseases Department, Faculty of Medicine, University of Alexandria, Egypt b Critical Care Medicine Department, Faculty of Medicine, University of Alexandria, Egypt
ARTICLE INFO
ABSTRACT
Article history: Received 22 October 2016 Accepted 13 November 2016 Available online xxxx
Continuous positive airway pressure (CPAP) is the most effective treatment of obstructive sleep apnea hypopnea syndrome (OSAHS). The therapeutic level of CPAP is achieved by manual titration or an auto CPAP device, but an alternative way involves the use of mathematical equations.
Aim of the work: to compare between CPAP pressure obtained by autoCPAP titration and that calculated from five different mathematical formulas.
Methods: We enrolled 30 patients with OSAHS. In addition to standard examination, anthropometric measurements and investigations, all patients underwent an over-night polysomnography (PSG) AutoCPAP (level II) to diagnose OSAHS. Manual scoring of polysomnographic data was performed according to criteria established by the American Academy of Sleep Medicine in 2012. All patients used AutoCPAP with pressure ranging from 4 to 20 cm H2O. Information recorded was downloaded and the following data were retrieved: 1) P90% (90th percentile pressure delivered by the autoCPAP device and eliminated snoring, flow limitation and apnea), this pressure was chosen as the therapeutic pressure, 2) peak pressure level, 3) mean pressure level, 4) the estimated residual apnea hypopnea index (AHI), 5) P90%/mean pressure level ratio, which is an index of pressure variability, was calculated. Five predictive formulas were retrieved from published literature and calculated for all patients. Results: The mean ± SD of AHI was 47.94 ± 27.21 events/h. OSAHS was mild in 5 patients (16.67%), moderate in seven patients (23.33%), and severe in 18 patients (60%). AutoCPAP P90% was 10.59 ±2.66 cm H2O and showed significantly direct correlation with each of Epworth sleepiness scale (p = 0.000), BMI (p = 0.001), neck circumference (p = 0.002), neck to height ratio (p = 0.000), and AHI (p = 0.000). Also it was found to show statistically significant inverse correlation with each of minimal (p = 0.000), and average minimal oxygen saturation (p = 0.032). The pressure calculated by Hoffstein et al. equation was significantly lower than autoCPAP P90%. There was no significant difference with pressure calculated by the other four formulas and P90%. The following model of predictive equation was derived from the studied sample: Predicted therapeutic pressure = 4.740 + 68.575 X (N/H) - 0.153 X (Minimal SpO2). Conclusions: Predictive formulas might be useful as an alternative to autoCPAP. The model of predictive formula derived from the present small sample of Egyptian patients with OSAHS should be validated on a larger sample size.
© 2016 Production and hosting by Elsevier B.V. on behalf of The Egyptian Society of Chest Diseases and Tuberculosis. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
Introduction
Continuous positive airway pressure (CPAP) is considered the standard, safe, and efficacious treatment for patients with obstructive sleep apnea hypopnea syndrome (OSAHS). OSAHS is a common disorder with established detriment to quality of life and
Peer review under responsibility of The Egyptian Society of Chest Diseases and Tuberculosis.
* Corresponding author. E-mail address: nashwahassan65@yahoo.com (N. Abdel Wahab).
adverse consequences for cardiovascular health. The effective pressure level (Peff) is the one that abolishes obstructive breathing disorders including inspiratory flow limitation and snoring in every sleep stage and body position [1]. According to the American Academy of Sleep Medicine (AASM) CPAP Titration Task Force guidelines, an appropriate pressure is determined by manual titration in a laboratory with full polysomnographic (PSG) monitoring [1].
The clinician is usually faced with the challenge of starting patients on positive airway pressure (PAP) treatment. In some health care systems PSG titration is not available. Newer automatic positive airway pressure (APAP) devices are available that can
http://dx.doi.org/10.1016/j.ejcdt.2016.11.004
0422-7638/® 2016 Production and hosting by Elsevier B.V. on behalf of The Egyptian Society of Chest Diseases and Tuberculosis. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
respond to patient-derived physiologic signals to automatically titrate a pressure for long-term use rather than use PSG for PAP titration. One can perform unattended auto-titration at home using an autoCPAP device for several nights. The 90th or 95th percentile pressure obtained from the device information download is chosen as the level of CPAP for long term treatment [2].
Another method is to calculate therapeutic pressure by predictive equations. There are different equations but all are based on anthropometric data, such as body mass index (BMI) or neck circumference (NC), and severity of disease according to oxygen desaturation index (ODI) or apnea hypopnea index (AHI). Some studies have investigated the differences between these equations and manual or autoCPAP titration and how these differences could influence clinical approach [3,4].
Aim of the work
The aim of this study was to compare between CPAP pressure obtained by autoCPAP titration and that calculated from five different mathematical formulas, in 30 Egyptian patients with OSAHS. Also, to derive a preliminary predictive equation from the derived data to be validated later on a larger representative sample.
Study population and subjects
The present study is a prospective study including thirty patients with OSAHS. Patients with obesity hypoventilation syndrome, overlap syndrome, chronic lung diseases, central apnea, heart failure, and neuromuscular diseases were excluded. The study protocol was approved by the local ethics committee, and informed consents were obtained from the studied patients.
Study measurements
All subjects were subjected to the following:
(1) Full history taking including assessment of daytime hyper-somnolence by Epworth Sleepiness Scale [5].
(2) Anthropometric measurements including weight, height, body mass index (BMI) [6], neck circumference (NC), and neck to height ratio (N/H) were measured for each patient.
OSAHS diagnosis was confirmed by full Polysomnography (PSG) (somnomedic) Level II sleep study [7]. Sleep stages and arousals, periodic limb movement, and nocturnal blood pressure were assessed in addition to respiratory events.
Respiratory events
According to the recommendations of the AASM [8] monitoring of airflow was performed using both thermistor and nasal cannula connected to pressure transducer. Respiratory effort was assessed using thoracic and abdominal belts while oxygen saturation was assessed by finger pulse oximeter. Detection of snoring was done via microphone. Manual scoring was performed for all patients. Apneas and hypopneas were defined using criteria established by the AASM [9]. Apnea was considered when there is drop in the flow excursion by p90% of baseline for at least ten seconds and hypopnea was defined when there is drop by p30% of baseline for at least 10 s and associated with p3% desaturation from pre-event baseline and/or arousal. Utilizing these definitions, an apnea-hypopnea index (AHI) was calculated by dividing the total number of apneas and hypopneas by the total sleep time (TST).
AutoCPAP titration: All patients used autoCPAP REMstar with pressure ranging from 4 to 20 cm H2O. The software version is Encore Pro 2 (Respironics, Murrysville, PA, USA). The auto-CPAP was used with the smart card in place and subjects underwent three to seven consecutive unattended home autoCPAP titration nights. The device measure changes in air flow by an internal pneu-motachograph to identify respiratory events. Sensors in the airway circuit of PAP devices measure airflow, vibration, and flattening of the airflow profile. Auto-adjusting PAP devices use this feedback to make online adjustments in pressure to maintain upper airway patency. At the end of the last titration night, information recorded was downloaded and the following data were retrieved: 1) P90% which is the 90th percentile pressure delivered by the autoCPAP device and eliminated snoring, flow limitation and apnea; this pressure was chosen as the therapeutic pressure, 2) peak pressure level, 3) mean pressure level, and 4) the estimated residual AHI. P90%/mean pressure level ratio, which is an index of pressure variability, was calculated.
The recording was considered to be acceptable if: (i) the duration per night was at least 3 h; (ii) excessive leaks were not present for more than 20% of the recording time and (iii) the AHI was 65.
Prior to titration, patients were educated regarding machine operation and mask placement and subjected to a 30-min period of autoCPAP exposure for pressure acclimatization. The mask was chosen according to need and comfort of the patient either nasal or oronasal mask.
Fixed CPAP pressure was adjusted for every patient according to his P90% and patients were followed up.
Predictive formula for CPAP pressure: Five predictive formulas were retrieved from the published literature and calculated for all patients:
1) CPAP pred = 0.16(BM1) + 0.13 (NC) + 0.04 (AHI) - 5.12 (Hoffstein et al.) [10].
2) CPAPpred = 30.8 + RD1 x 0.03 - nadir saturation x 0.05 - mean saturation x 0.2. (Loredo et al.) [11].
3) CPAPpred = 0:048 (ODI)+ 0:128 (NC)+ 2.1 (Stradling et al.) [12].
4) CPAPpred = 6.656 + 0.156 (BM1) - 0.071 (minimal SpO2[%]) + 0.041 (respiratory disturbance index) + 0.094 (score of Epworth Sleepiness Scale) (Lee et al.) [13].
5) CPAPpred = 0.193(BM1) +0.077 (NC) + 0.02 (AHI) - 0.611 (Series et al.) [14].
Where CPAPpred: CPAP predicted pressure (cm H2O), BM1: body mass index (kg/m2), NC: neck circumference (cm), AH1: apnea hypopnea index, RD1: respiratory disturbance index, OD1: oxygen desaturation index.
Statistics methodology [15]
After data input in a specially designed sheet using Microsoft Excel version 10 software, a print out of the data was thoroughly revised and entry mistakes were corrected. Data were transferred to SPSS (Statistical Package for Social Science version 21) [16] format and Kolmogorov-Smirnov test of normality was carried out and revealed no significant difference (p > 0.05) (i.e. normally distribution data) [17]. A 0.05 level of significance was considered.
- Exploration of the data: This yielded complete descriptive statistics including the minimum and maximum, range, mean, standard deviation, median and inter-quartile range for each variable.
- Data were described using minimum, maximum, mean and
standard deviation.
- Intra-class correlation (ICC) [18] (using SPSS) and Bland-Altman plot [19] (using MedCalc version 12.2.1.0) were done for agreement.
In the present study an alpha level was set to 5% with a significance level of 95%, and a beta error accepted up to 20% with a power of study of 80%.
Table 2
Respiratory polysomnography data of thirty patients with OSA.
Polysomnographie data Min-Max Mean ± SD
Apnea Hypopnea index (A + H/hour) 10.. 4-91.7 47.94 ±27.21
Baseline O2 saturation (%) 84- -96 92.47 ±3.21.5
Mean O2 saturation (%) 82- -96 89.93 ± 3.28
Minimal O2 saturation (%) 54- -91 77.23 ±8.06
Average minimal O2 saturation (%) 79- 95 85.4 ± 6.2
Results
The present study was carried out on thirty adult patients with OSAHS. Table 1 shows different characteristics and anthropometric data of the studied patients.
Table 2 shows the results (in form of range and mean ± SD) of respiratory polysomnographic data of the studied patients with OSA. Fig. 1 represents boxplots for baseline, mean, minimal and average minimal oxygen saturation. The thick line in the middle of the box represents the median, the box represents the interquartile range (from 25th to 75th percentiles), the whiskers represents the minimum and maximum values after excluding outliers (black-filled circle) and extremes (asterisk). Using the OSAHS severity classification which is defined as mild for AHI p 5/h and <15/h, moderate for AHI p15/h and 630/h, severe for AHI >30/h [9], five patients (16.67%) had mild OSAHS apnea, seven patients (23.33%) had moderate OSAHS, and 18 patients (60%) had severe OSAHS. According to exclusion criteria, none of the included patients had central sleep apnea or obesity hypoventilation syndrome.
All patients were subjected to titration using autoCPAP. Measurements retrieved from autoCPAP (in form of average of accepted nights) included: P90% which is considered as the therapeutic pressure, peak pressure, mean pressure and the ratio of P90% to the mean pressure as an index of pressure variability which was very small (1.17 ±0.09). Other data to assess the effectiveness of the pressure and accuracy of the results included residual AHI and amount of leak. The residual AHI was statistically significantly lower than the baseline AHI measured by PSG (Fig. 2). The residual events were obstructive apneas in 9 patients, hypopneas in 20 patients, RERA in 9 patients, flow limitation in 9 patients, and vibrating snoring in 16 patients. A predictive CPAP pressure was calculated for each patient using the five studied equations. Table 3 shows the results of the data retrieved from autoCPAP device after exclusion of unacceptable nights and values of predictive pressure calculated by different equations.
To examine the effect of leak on P90%, a correlation analysis between P90% and each of leak average, percentage of night in large leak and minutes in large leak was performed and no statistically significant effect was found. Then, a multiple correlation analysis was performed using block method and also there was no significant effect in the model (F = 0.887, p = 0.461 NS)
Paired samples correlation between P90% and value of pressure calculated by each equation was performed and revealed positive correlations (Table 4). Then the data was tested using paired sam-
Table 1
Different characteristics and anthropometric data of thirty patients with OSA.
Variables
Min-Max
Mean ± SD
Gender
Male No. 16 53.33%
Female No. 14 46.66%
Age (year) 40-87 61.03 ±13.66
Epworth sleepiness scale (/24) 7-18 13.67 ±2.69
Body Mass Index (kg/m) 27.46-62.43 37.45 ± 9.5
Neck Circumference (cm) 36-47 42.85 ± 2.8
Neck/Height ratio 0.22-0.303 0.258 ± 0.021
Fig. 1. Boxplots represents the values of baseline, mean, minimal and average minimal oxygen saturation.
Baseline AHI by PSG Residual AHI by autoCPAP
Fig. 2. oxplots compares the baseline AHI and residual AHI on autoCPAP .
ple t test which revealed that the pressure calculated by Hoffstein et al. equation was significantly lower than autoCPAP P90%. There was no significant difference between pressure calculated by the other four formulas and P90% (Table 4 and Fig. 3).
When we used Intra-class correlation (ICC) and Bland-Altman plot, it showed statistically significant agreement between auto-CPAP P90% and the value of pressure calculated by Hoffstein et al. equation (Fig. 4).
Linear correlation between autoCPAP P90% as a therapeutic pressure and different variables to detect predictors of therapeutic pressure was carried out. Therapeutic pressure was found to correlate directly with each of the following: Epworth sleepiness scale (p = 0.000, Fig. 5), BMI (p = 0.001, Fig. 6), neck circumference
Table 3
AutoCPAP device data and values of predictive pressure calculated by different equations.
Variable Min-Max Mean ± SD
P90% (cm H2O) 4.8-14.9 10.59 ±2.66
Peak P (cm H2O) 5.5-19 12.35 ±2.95
Mean P (cm H2O) 4.2-13.1 9.1 ± 2.39
P90%/mean 1.01-1.35 1.17 ±0.09
Residual AHI (events/h) 0-4.4 2.34 ±1.38
% of night in large leak (%) 0-21 1.42 ±4.24
Minutes in large leak (minutes) 0-23 3.6 ± 7.29
Average leak 12-44 28.94 ± 8.89
Ppred (Hoffstein et al.) 5.55-13.8 8.52 ± 2.03
Ppred (Loredo et al.) 7.58-13.8 10.38 ±1.55
Ppred (Stradling et al.) 7.34-11.9 9.89 ±1.51
Ppred (Lee et al.) 6.94-15.87 10.21 ±2.45
Ppred (Sériés) 8.37-18.83 10.87 ±2.09
(p = 0.002, Fig. 7), neck to height ratio (p = 0.000, Fig. 8), and AHI (p = 0.000, Fig. 9). Also therapeutic pressure was found to show statistically significant inverse correlation with each of minimal (p = 0.000), and average minimal oxygen saturation (p = 0.032).
Then to construct an equation that predicts the therapeutic pressure from correlated variables, a multiple linear regression analysis, stepwise method was carried out and revealed that neck to height ratio (N/H) is the highest predictor followed by minimal SpO2 which increased the predictive value by its addition (Table 5).
The equation
Yi = (b0 + b1X1 + b2X2 + bnXn) + e1
Avg P90% = 4.740 + 68.575 x (N/H) - 0.153 x (Minimal SpO2) The following model of predictive equation was derived to be validated on a larger sample size
Predicted Therapeutic pressure = 4.740 + 68.575 x (N/H) -0.153 x (Minimal SpO2)
Discussion
OSAHS is the most common form of sleep related breathing disorders and its prevalence is similar to both asthma and diabetes combined [20]. Although it is estimated that 75% of severe sleep related breathing disorders cases remain undiagnosed, over the past 5-10 years, the sleep industry's focus has shifted from diagnosing sleep apnea to managing the patient's sleep apnea treatment [20]. Untreated OSAHS is associated with an increased cardiovascular and metabolic consequences, health care cost, mortality risks, and traffic accidents [20]. There are multiple methods to treat sleep apnea and CPAP is considered the most common method currently used to treat sleep apnea. As OSAHS occurs when the muscles in the back of the throat fail to keep the airway open despite efforts to breathe, positive airway pressure (PAP) treatment uses mild air pressure to keep the airways open [21]. Although the standard approach to achieve the therapeutic pressure is the manual PAP titration using PSG, there are some assumed or potential limitations including the cost and inconvenience of repeating PSG due to incomplete titrations. Also the possibility of
prescribing pressures that are not suitable due to the limited sampling that occurred when titration takes place over only one or one-half night in case of split-night study [22]. The desire to simplify pressure titration has inspired the development of autotitrating positive airway pressure (APAP) devices and developing updated practice parameters for their use [23].
In the present study we compared the therapeutic pressure obtained from autoCPAP titration with that predicted from 5 different mathematical equations to investigate if any of them could replace autoCPAP usage in our Egyptian sample to diminish the cost and accelerate the start of the treatment.
Thirty patients diagnosed to have OSAHS used autoCPAP for a period of 3-7 days. The mean and standard deviation of the P90%, which is considered as the therapeutic pressure, were 10.59 and 2.65 cm H2O respectively.
Data from randomized controlled trials have shown that auto-CPAP titration is as effective as manual titration among selected patients with OSA without significant comorbidities, is less costly, and requires fewer health care resources [24]. Also, auto-titration allows the patient to experience PAP at home for a short time before CPAP treatment is initiated. Important clinical information about PAP efficacy (device interrogation) and side effects experienced by the patient during the short trial may result in interventions at the time the patient is started on longterm CPAP treatment that improve outcomes [25].
In the present study, we selected patients by excluding those with heart failure, central sleep apnea or hypoventilation due to any cause. We supplied the patients with autoCPAP for at least three days after intensive education of its use. After downloading data we excluded nights with large leaks, less than three hours usage, or if the AHI was >5/h.
Although many researchers reported equivalent clinical outcomes from APAP titration compared with traditional methods [2,25]. McArdle et al. [24] found significant residual respiratory events (upper quartile residual AHI = 14 events/h) after one month on a fixed pressured determined by APAP titration. The reason is unclear but persistent central apneas and hypopneas that may occur with cardiovascular disease, is one possibility. Although significant cardiovascular comordibities are carefully excluded in most of research studies, many patients have at least mild comor-bidity (e.g., hypertension). A recent practice review recommended further research to clarify which patients are appropriate candidates for APAP titration with particular attention to the role of comorbidities [23].
Current APAP devices provide data about residual respiratory events while using the device. In the present study the mean (SD) of residual AHI was 2.34 (1.38)/h, taking into consideration exclusion of nights with high residual AHI. The accuracy of this downloaded information in the clinical setting was studied by many investigators [2,27-31].
Haung et al. [2] studied a large cohort of consecutive patients with OSA undergoing APAP titration to determine clinical or other readily available predictors of poor physiologic control during APAP titration. They found that optimal or good titration may not be achieved among patients with severe disease, particularly when sleep is highly fragmented (i.e., a high arousal index). Cau-
Table 4
Paired samples correlations and paired samples t test differences between P90% and mathematical equation.
Correlation Sig. Paired difference t test Sig (2 tailed)
Pair 1 P90% & Ppred (Hoffstein et al.) 30 0.801 0.000 7.107 0.000
Pair 2 P90% & Ppred (Loredo et al.) 30 0.622 0.000 0.527 0.603
Pair 3 P90% & Ppred (Lee et al.) 30 0.833 0.000 1.406 0.170
Pair 4 P90% & Ppred (Stradling et al.) 30 0.679 0.000 1.956 0.060
Pair 5 P90% & Ppred (Sériés) 30 0.712 0.000 0.831 0.413
N. Abdel Wahab, Y.N. Ahmed/Egyptian Journal of Chest Diseases and Tuberculosis xxx (2016) xxx-xxx
Auto CMP P90(%) Ftred(Hoffstehetal) Ppred (Loredoetsl) l^red (Leeetal) Ppred (Stradhig etal) Ppred (Swle«)
Fig. 3. Boxplots comparing the values of pressure obtained by autoCPAP, Ppred (Hoffstein et al.), Ppred (Loredo et al.), Ppred (Stradling et al.), Ppred (Lee et al.), and Ppred (Series et al.).
■o 3
o) 0 -i -2 -3
+ 1.96 SD
o o 5.2
- O o 0 ° 0 OO o o o o Mean
- o o 0 0 0 ° o ° o 2.1 o
o -1.96 SD
-1 . 1.1,1, 1 ■ 1 . 1
4 6 8 10 12 14 16
Mean of Avg P90% and Hoffest Pred P
Fig. 4. Bland-Altman plot for autoCPAP P90% and the value of pressure calculated by Hoffstein et al. equation.
Fig. 6. Correlation between body mass index (BMI) and autoCPAP P90%.
Fig. 7. Correlation between Neck circumference and autoCPAP P90%.
Fig. 5. Correlation between Epworth Sleepiness scale and autoCPAP P90%.
tion is also needed when there is a history of cardiac disease, but cardiovascular risk factors alone, such as hypertension and dia-
Fig. 8. Correlation between Neck to Height ratio and autoCPAP P90%.
betes, did not appear to increase the risk of poor APAP titration. They suggested that clinicians cannot rely on downloaded data alone to determine if the patient's OSAHS is well controlled during
15.000" o oo
12.500- O
^ 10.000" 0 o
7.500-
I I I I I !
0.000 20.000 40.000 60.000 80.000 1 00.000
PSGAHI before titration (event/hour)
Fig. 9. Correlation between baseline Apnea Hypopnea index and autoCPAP P90%.
Table 5
Data obtained from multiple linear regression analysis, stepwise method to construct a predictive equation.
B SE B ß
Step 1
Constant -11.223 4.572
N/H 84.631 17.681 0.671*
Step 2
Constant 4.740 5.506
N/H 68.575 14.934 0.544*
Minimal SpO2 -0.153 0.039 -0.465*
Note R = 0.450 for Step 1, R = 0.200 for Step 2, (ps < 0.01), *p <0.001.
the APAP titration and close follow-up is needed. They recommended formal laboratory PSG assessment in case of suboptimal clinical response.
They [2] also investigated the accuracy of the device download AHI against the gold standard AHI determined from simultaneous PSG and factors that might affect this. They found that device AHI and PSG AHI are highly correlated and patients who did not achieve optimal titration often have a high device AHI. However, there is poor case by- case agreement between these two measures. They found that patients with a high AHI and a high hypop-nea index during their diagnostic study and those with lower% predicted FVC are more likely to have either inadequate or unacceptable OSA control and are also less likely to have an accurate autoCPAP AHI. They explained that as a lower% predicted FVC reflects the influence of obesity, which is likely to predispose to greater oxygen desaturation for any given event and due to the important role of oxygen desaturation in identifying hypopneas. Their findings are consistent with those of Prasad and colleagues [26], who found a poor correlation between the PSG and device measures of hypopnea at therapeutic pressure, using a different device.
Also, Woodson et al. [27] found that auto-CPAP overestimated AHI by an average of 1.4 events per hour when compared to PSG-AHI and that both BMI and lowest SO2 can be used as good predictors for PSG-AHI when the value of 5 is used as a cutoff for PSG-AHI score. Higher values of BMI and lowest SO2 seemed to indicate high risk for higher values of PSG-AHI.
Although Ueno et al. [28] found a strong correlation between the PSG-AHI and the autoCPAP-AHI, they showed that autoCPAP overestimated the AHI when compared with the PSG-AHI (9.9 vs. 4.2). They suggested that higher 95th percentile leakage and higher
AHI on diagnostic PSG were the factors associated with greater difference between the autoCPAP-AHI and the PSG-AHI.
On the other hand, Desai et al. [29] studied the accuracy of auto-CPAP in estimating the residual AHI in 99 patients with OSA who underwent a repeated sleep study using auto-CPAP. The estimated AHI from auto-CPAP was compared with the AHI from an overnight PSG on auto-CPAP and they used a PSG AHI cutoff of 5 events per hour to differentiate patients optimally treated with auto-CPAP from those with residual OSA on therapy. They concluded that auto-CPAP estimate of AHI can be used to estimate residual AHI in patients with OSA of varying severity treated with auto-CPAP.
Also, Cilli et al. [30] confirmed the accuracy of autotitrating CPAP-determined residual AHI in a selected population consisting of 137 patients with OSAHS.
In the present study, the auto-CPAP device used by all patients was the REMstar Auto with software version Encore Pro 2. The device identifies and stores mask on time and the presence of snoring, obstructive apneas, central apneas, hypopneas, respiratory effort related arousals (RERAs) and total air leak. The device use algorithm to detect whether the airway is obstructed or clear (i.e., open) during device-detected apneas. Li et al. [31] demonstrated that device-detected airway status agrees closely with PSG respiratory event scoring but should not be equated with a specific type of respiratory event. Device-detected RERAs are weakly correlated with manually scored RERAs. Clinicians also need to be aware that a device-detected apnea identified by the device as having a clear airway is not always a central apnea and a device-detected apnea identified by the device as having an obstructed airway is not necessarily an obstructive apnea. This is because of some factors including the different criteria for respiratory events used by the PAP device and on PSG, the inability of the PAP device to detect respiratory effort, and the inability of PSG to determine airway status during central apneas. However, they suggested that a preponderance of events with clear airway on PAP download strongly suggests the presence of central apneas. They concluded that a device-detected AHI (AHIFlow) <10 events/h on a PAP device is strong evidence of good treatment efficacy. When interpreted appropriately, PAP device reports provide important information needed for adequate management of patients on PAP treatment. The reliability of pressure recommendations of different Automatic CPAP machines was studied by many investigators [3234]. Although auto-CPAP is not currently recommended to diagnose OSA [23], its diagnostic value was tested in many studies [26,35,36] in patients undergoing simultaneous PSG. Many studies suggested that APAP treatment of patients diagnosed with moderate to severe OSA can result in equivalent PAP adherence and improvement in subjective sleepiness and quality of life compared to an approach using a PAP PSG titration followed by CPAP treatment if patients are carefully selected [23,25,37,38]. Teschler et al. [39] concluded that residual AHI during 2 months of home autoadjusting nasal CPAP is comparable to that during conventionally titrated fixed pressure CPAP and that leak, while common, did not importantly affect residual AHI. Peak auto pressure was typically 2 cm H2O higher than manual; mean auto "recommended" pressure was 0.7 cm H2O higher than manual, and varied by 1.0 cm H2O SD from night to night. Compliance was essentially maximal in both arms of the study [39].
In the present study, correlation analysis was performed to evaluate the effect of leak on P90%. No significant correlations were found between each of percent of night in leak, average leak, and minutes in large leak with the P90%. This could be due to exclusion of nights with large leaks in the present study.
With increasing healthcare costs and long waiting lists that delay the start of the treatment, it is currently difficult to perform PAP titrations in every OSA patient. Methods have been explored such as using mathematical equations that can be used to predict
effective CPAP pressures. Researchers have used multiple linear regressions analyses on variables such as age, AHI, BMI, height, lowest oxygen saturation, mean oxygen saturation, NC, oxygen desaturation index (ODI), race, RDI, sex (male versus female) and sleepiness [40]. These multiple linear regressions analyses have been used to formulate predictive mathematical equations as compared to the effective CPAP pressures determined during PAP titration studies. The equations have subsequently been used either for estimating starting pressures or for use during in-lab PAP titration studies. In the present study we calculated the predicted pressure according to five different mathematical equations for all the patients and compared it with P90%. The earliest published study was by Hoffstein et al. [10] in 1993 and the most recent study was published in 2013 [13]. In the present study we chose equations from different races and that used variables which are easy to be measured by clinicians. There were more recent equations [41,42] but they used anatomical variables that are not applicable.
In the present study, the pressure estimated using Hoffstein et al. equation (8.52 ± 2.03 cm H2O) was statistically significantly lower (p = 0.000 using Paired Samples t test, Table 4 and Fig. 3) than P90% (10.59 ± 2.66 cm H2O).
This is in consistence with findings of our previous work [43] where P predicted using Hoffstein et al. equation was significantly lower than that obtained by manual titration. Also, Lee et al. [13] concluded that the Hoffstein formula significantly underestimated CPAP in Asian patients with OSA. Although their equation [13] was somewhat better to predict optimal CPAP level in Asian subjects than the Hoffstein equation, they suggested that CPAP prediction equation did not accurately predict the prescribed CPAP level and its usefulness is limited in some clinical settings.
Lacedonia et al. [44] investigated the differences between predicted pressure by three equations [10,12,14] and that obtained by manual or ACPAP titration. Similar to our results, a positive correlation was found between titrated CPAP pressure and the value of pressure calculated by each equation but there were significant differences in the pressure calculated by Hoffstein et al. equation, which was lower than titrated pressure. By using the Bland-Alt-man test, in contrast to our findings, they ascertain that there were more differences between the pressure values calculated by equations and real titrated pressure. In the present study, we used Intra-class correlation (ICC) and Bland-Altman plot to test the agreement between autoCPAP P90% and the value of pressure calculated by Hoffstein et al. equation, it showed a statistically significant agreement.
The variables used by Hoffestein et al. formula [10] were BMI, neck circumference and AHI while those used by Lee et al. [13] were BMI, minimal SpO2, RDI and score of ESS. Sériés et al. [14] used the same variables as Hoffestein et al. but the estimated pressure using their formula (10.87 ± 2.09 cm H2O) was statistically not different than P90%. The difference is that Sériés et al. [14] used a small constant to be subtracted from the formula than that used by Hoffestein et al. [10] (0.611 vs 5.12 respectively). Stradling et al. [12] used only oxygen desaturation index and NC as variable and add a constant (2.1) to the formula instead of subtraction. The pressure estimated by Stradling et al. formula was (9.89 ± 1.51 cm H2O). Loredo et al. [11] formula RDI, minimal saturation and mean saturation as variables and add a constant to them (30.8). The pressure estimated by Loredo et al. formula was (10.39 ±1.55 cm H2O).
In the present study we studied the predictive effect of some variable on P90% .These variables were; score of ESS, BMI, NC, N/ H, AHI, baseline oxygen saturation, mean oxygen saturation, minimal oxygen saturation and average minimal oxygen saturation. First we did a correlation analysis, and choose the predictors which showed significant correlations (p < 0.05). All except baseline oxygen saturation and mean oxygen saturation were correlated signif-
icantly with P90%. After that, a multiple linear regression analysis, stepwise method was carried out and it was found that both N/H and minimal oxygen saturation were the highest predictors of P90%.
In the present study, according to statistical analysis, the following model of predictive equation was derived to be tested on a sample of larger size: Predicted Therapeutic pressure = 4.740 + 68.575 x (N/H) - 0.153 x (Minimal SpO2).
However, a larger sample size may enable inclusion of more variables in the equation.
Recently Ho et al. [45] demonstrated that N/R is a better predictive tool in adults, corroborating the long-standing use of absolute NC measurements in the clinical evaluation of potential OSA.
Given the fact that the lowest oxygen saturation generally lasts seconds per event, it makes sense that the mean oxygen saturation would end up having more weight in the formulas given that mean oxygen saturation takes into account the oxygen saturation throughout the entire night. However, the study by Akashiba et al. [46] reported a correlation between the mean oxygen saturation during sleep and optimal CPAP pressure, demonstrating that a large percentage of OSA patients in the study had hypercapnia and underlying obesity hypoventilation syndrome (OHVS) which is a 24-h condition, and this fact can confound the results. In the present study, patients with OHVS were excluded and this gives an explanation for the nonsignificant correlation between mean oxygen saturation and therapeutic CPAP pressure.
Camacho et al. [40] conducted a systematic review on mathematical equations to predict positive airway pressures. After detailed review of the fifty studies, they came to a consensus for 26 studies which presented equations that met their inclusion and exclusion criteria. They found that BMI was the most common variable in the mathematical equations, being included in 18 studies and one study also included percent of the ideal body weight. As BMI increases, excessive fat deposits around the neck and within the pharynx and this can affect the collapsibility of upper airway. They found that in the Asian studies, the mean value for the coefficient was 0.16871, while, in the non-Asian studies, the mean value for the coefficient was 0.1003, demonstrating that this variable emerged as a more heavily weighted variable than other variables during the multiple linear regressions analyses [40].
They found also that AHI was a variable in mathematical equations for 17 and RDI for 4 studies. The mean overall coefficient for AHI was 0.0442 for all studies and 0.03963 for Asian and 0.04878 for non-Asian studies [40]. In the mathematical equation by Loredo et al. [11] RDI was a variable in the mathematical equation and there was overall good prediction which demonstrates that respiratory effort related arousals (RERAs) could be a possible factor that may influence the derived mathematical equations. For the studies that included oxygen desaturation index, mean oxygen saturation and lowest oxygen saturation the factor that had the largest influence when used was the mean oxygen saturation but presence of OHVS may interfere with the accuracy [40].
They suggested that race/ethnicity and gender may affect PAP pressures. Factors affecting the predictive pressures for PAP between males and females include that males generally have a higher AHI and a longer soft palate [46]. In the three studies reporting separate mathematical equations for men versus women, the coefficient was higher for men [40].
They concluded that although the mathematical equations have helped improve PAP titration study success, the formulas are not completely generalizable secondary to physical, behavioral, comorbidity, and PSG differences in OSA patients [40].
Choi et al. [47] concluded that when compared to full-night manual titration as the standard method, auto-adjusting titration appeared to be more reliable than using a predictive equation for determining the optimal CPAP level in patients with OSAS.
Also, Lacedonia et al. [44] concluded that manual or auto CPAP titration remains the best way to define the appropriate CPAP. However, predictive formulas can be useful if used with caution and always after verifying the real efficacy, particularly for patients needing higher pressure.
According to Camacho et al. [40], ethnicity and gender have an important impact on the level of CPAP therapeutic pressure. So we need to test and validate our model of predictive equation [Therapeutic pressure = 4.740 + 68.575 x (N/H) - 0.153 x (Minimal SpO2)]. We are planning to conduct another study that will include a larger sample representative of the Egyptian patients with OSAHS of both gender and considering the effect of menopause on females to validate the applicability of this model and upgrade it if needed.
Conclusion
Predictive formulas might be useful as an alternative to auto-CPAP. The model of predictive formula derived from the present small sample of Egyptian patients with OSAHS should be validated on a larger sample size.
Disclosure statement
The authors have indicated no conflicts of interest.
References
[1] C.A. Kushida, A.D. Chediak, R.B. Berry, et al., Clinical guidelines for the manual titration of positive airway pressure in patients with obstructive sleep apnea, J. Clin. Sleep Med. 4 (2008) 157-171.
[2] H.C.C. Huang, D.R. Hillman, N. McArdle, Control of OSA during automatic positive airway pressure titration in a clinical case series: predictors and accuracy of device download data, Sleep 35 (9) (2012) 1277-1283.
[3] O. Marrone, A. Salvaggio, S. Romano, G. Insalaco, Automatic titration and calculation by predictive equations for the determination of therapeutic continuous positive airway pressure for obstructive sleep apnea, Chest 133 (3) (2008) 670-676.
[4] K.B. Hertegonne, J. Volna, S. Portier, R. De Pauw, G. Van Maele, D.A. Pevernagie, Titration procedures for nasal CPAP: automatic CPAP or prediction formula?, Sleep Med 9 (7) (2008) 732-738.
[5] M.W. Johns, A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale, Sleep 14 (1991) 540-545.
[6] E. Jequier, Energy, obesity, and body weight standards, Am. J. Clin. Nutr. 45 (1987) 1035-1047.
[7] C.A. Kushida, M.R. Littner, T. Morgenthaler, et al., Practice parameters for the indications for polysomnography and related procedures: an update for 2005, Sleep 28 (4) (2005) 499-521.
[8] C. IBer, S. Ancoli-Israel, A.L. Chesson, S.F. Quan, American Academy of Sleep Medicine, Westchester IL, The AASM manual for the scoring of sleep and associated events, Rules, terminology, and technical specifications, 2007.
[9] R.B. Berry, R. Budhiraja, D.J. Gottlieb, D. Gozal, C. Iber, V.K. Kapur, C.L. Marcus, R. Mehra, S. Parthasarathy, S.F. Quan, S. Redline, K.P. Strohl, S.L.D. Ward, M.M. Tangredi, Rules for scoring respiratory events in sleep: update of the 2007 AASM manual for the scoring of sleep and associated events, J. Clin. Sleep Med. 8 (5) (2012) 597-619.
[10] V. Hoffstein, S. Mateika, Predicting nasal continuous positive airway pressure, Am. J. Respir. Crit. Care Med. 150 (2) (1994) 486-488.
[11] J.S. Loredo, C. Berry, R.A. Nelesen, J.E. Dimsdale, Prediction of continuous positive airway pressure in obstructive sleep apnea, Sleep Breath. 11 (1) (2007 Mar) 45-51.
[12] J.R. Stradling, M. Hardinge, J. Paxton, D.M. Smith, Relative accuracy of algorithm-based prescription of nasal CPAP in OSA, Respir. Med. 98 (1994) 152-154.
[13] G.-H. Lee, M.J. Kim, E.M. Lee, C.S. Kim, S.A. Lee, Prediction of optimal CPAP pressure and validation of an equation for Asian patients with Obstructive Sleep Apnea, Respir. Care 58 (5) (2013) 810-815.
[14] F. Sériés, Accuracy of an unattended home CPAP titration in the treatment of obstructive sleep apnea, Am. J. Respir. Crit. Care Med. 162 (2000) 94-97.
[15] A. Field (Ed.), Discovering Statistics Using IBM SPSS, 4th ed., SAGE Publications Ltd, London, California, New Delhi, 2013.
[16] IBM Corp, IBM SPSS Statistics for Windows, Version 21.0, IBM Corp, Armonk, NY, 2012.
[17] K.D.S. Young, Bayesian diagnostics for checking assumptions of normality, J. Stat. Comput. Simul. 47 (3-4) (1993) 167-180.
[18] A. Cicchetti, V. Domenic, Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology, Psychol. Assess. 6 (4) (1994) 284-290, http://dx.doi.org/10.1037/1040-3590.6.4.284.
[19] D.G. Altman, J.M. Bland, Measurement in medicine: the analysis of method comparison studies, Statistician 32 (1983) 307-317, http://dx.doi.org/ 10.2307/2987937.
[20] M. Knauert, S. Naik, M.B. Gillespie, M. Kryger, Clinical consequences and economic costs of untreated obstructive sleep apnea syndrome, World J. Otorhinolaryngol. 1 (1) (2015) 17-27.
[21] A. Qaseem, J.E. Holty, D.K. Owens, P. Dallas, M. Starkey, P. Shekelle, Clinical Guidelines Committee of the American College of Physicians, Management of obstructive sleep apnea in adults: a clinical practice guideline from the American College of Physicians, Ann. Intern. Med. 159 (7) (2013) 471-483.
[22] T.E. Weaver, A.M. Sawyer, Adherence to continuous positive airway pressure treatment for obstructive sleep apnea: implications for future interventions, Indian J. Med. Res. 131 (2010) 245-258.
[23] T.I. Morgenthaler, R.N. Aurora, T. Brown, R. Zak, C. Alessi, B. Boehlecke, A.L. Chesson, L. Friedman, V. Kapur, R. Maganti, J. Owens, J. Pancer, T.J. Swick, Standards of Practice Committee of the AASM. Practice parameters for the use of autotitrating continuous positive airway pressure devices for titrating pressures and treating adult patients with obstructive sleep apnea syndrome: An update for 2007, Sleep 31 (1) (2008) 141-147.
[24] N. McArdle, B. Singh, M. Murphy, K.R. Gain, C. Maguire, S. Mutch, D.R. Hillman, Continuous positive airway pressure titration for obstructive sleep apnoea: Automatic versus manual titration, Thorax (2010) 606-611.
[25] R.B. Berry, P. Sriram, Auto-adjusting positive airway pressure treatment for sleep apnea diagnosed by home sleep testing, J. Clin. Sleep Med. 10 (12) (2014) 1269-1275.
[26] B. Prasad, D.W. Carley, J.J. Herdegen, Continuous positive airway pressure device-based automated detection of obstructive sleep apnea compared to standard laboratory polysomnography, Sleep Breath 14 (2010) 101-107.
[27] B.T. Woodson, A. Saurejan, L.T. Brusky, J.K. Han, Nonattended home automated continuous positive airway pressure titration: comparison with polysomnography, Otolaryngol. Head Neck Surg. 128 (2003) 353-357.
[28] K. Ueno, T. Kasai, G. Brewer, H. Takaya, K. Maeno, S. Kasagi, F. Kawana, S. Ishiwata, K. Narui, Evaluation of the apnea- hypopnea index determined by the S8 auto-CPAP, a continuous positive airway pressure device, in patients with obstructive sleep apnea-hypopnea syndrome, J. Clin. Sleep Med. 6 (2010) 146-151.
[29] H. Desai, A. Patel, P. Patel, B.J. Grant, M.J. Mador, Accuracy of autotitrating CPAP to estimate the residual Apnea-Hypopnea Index in patients with obstructive sleep apnea on treatment with autotitrating CPAP, Sleep Breath. 13 (4) (2009 Nov) 383-390, http://dx.doi.org/10.1007/s11325-009-0258-2, Epub 2009 May 1.
[30] A. Cilli, R. Uzun, U. Bilge, The accuracy of autotitrating CPAP-determined residual apnea-hypopnea index, Sleep Breath 17 (1) (2013) 189-193, http:// dx.doi.org/10.1007/s11325-012-0670-x.
[31] Q.Y. Li, R.B. Berry, M.G. Goetting, B. Staley, H. Soto-Calderon, S.C. Tsai, J.G. Jasko, A.I. Pack, S.T. Kuna, Detection of upper airway status and respiratory events by a current generation positive airway pressure device, Sleep 38 (4) (2015) 597605, http://dx.doi.org/10.5665/sleep.4578.
[32] M.F. Damiani, V.N. Quaranta, E. Tedeschi, R. Drigo, T. Ranieri, P. Carratù, O. Resta, Titration effectiveness of two autoadjustable continuous positive airway pressure devices driven by different algorithms in patients with obstructive sleep apnoea, Respirology 18 (2013) 968-973, http://dx.doi.org/10.1111/ resp.12098.
[33] F. Sériés, J. Plante, Y. Lacasse, Reliability of home CPAP titration with different automatic CPAP devices, Respir. Res. 9 (2008) 56, http://dx.doi.org/10.1186/ 1465-9921-9-56.
[34] V. Isetta, D. Navajas, J.M. Montserrat, R. Farré, Comparative assessment of several automatic CPAP devices' responses: a bench test study, ERJ Open Res. 1 (1) (2015), pii: 00031-2015.
[35] K. Rees, P.K. Wraith, M. Berthon-Jones, N.J. Douglas, Detection of apnoeas, hypopnoeas and arousals by the Auto-Set in the sleep apnoea/hypopnoea syndrome, Eur. Respir. J. 12 (1998) 764-769.
[36] P. Mayer, J.C. Meurice, F. Philip-Joet, A. Cornette, D. Rakotonanahary, N. Meislier, J.L. Pepin, P. Levy, D. Veale, Simultaneous laboratory-based comparison of ResMed Autoset with polysomnography in the diagnosis of sleep apnoea/hypopnoea syndrome, Eur. Respir. J. 12 (1998) 770-775.
[37] B. Hevener, W. Hevener, Continuous positive airway pressure therapy for obstructive sleep apnea maximizing adherence including using novel information technology-based systems, Sleep Med. Clin. 11 (2016) 323-329, http://dx.doi.org/10.1016/j.jsmc.2016.04.004.
[38] T. Xu, T. Li, D. Wei, Y. Feng, L. Xian, H. Wu, J. Xu, Effect of automatic versus fixed continuous positive airway pressure for the treatment of obstructive sleep apnea: an up-to-date meta-analysis, Sleep Breath. 16 (4) (2012) 1017-1026, http://dx.doi.org/10.1007/s11325-011-0626-6, Epub 2011 Dec 3.
[39] H. Teschler, T.E. Wessendorf, A.A. Farhat, N. Konietzko, M. Berthon-Jones, Two months auto-adjusting versus conventional nCPAP for obstructive sleep apnoea syndrome, in: Eur. Respir. J. 15 (2000) 990-995.
[40] M. Camacho, M. Riaz, A. Tahoori, V. Certal, C.A. Kushida, Mathematical equations to predict positive airway pressures for obstructive sleep apnea A systematic review, Sleep Disord. (2015) 293868, http://dx.doi.org/10.1155/ 2015/293868, Epub 2015 Jul 30.
[41] C.-C. Lai, M. Friedman, H.-C. Lin, et al., Clinical predictors of effective continuous positive airway pressure in patients with obstructive sleep apnea/hypopnea syndrome, Laryngoscope (2015).
[42] E. Ito, S. Tsuiki, K. Namba, Y. Takise, Y. Inoue, Upper airway anatomical balance contributes to optimal continuous positive airway pressure for Japanese
patients with obstructive sleep apnea syndrome, J. Clin. Sleep Med. 10 (2) (2014)137-142.
[43] N.H. Abdel Wahab, Evaluation of Effects of Continuous Positive Airway Pressure on Cardiopulmonary Performance in Obstructive Sleep Apnea MD Thesis, Faculty of Medicine, Alexandria University, 2005.
[44] D. Lacedonia, R. Sabato, G.E. Carpagnano, P. Carratu, A. Falcone, F. Gadaleta, O. Resta, M.P. Foschino Barbaro, Predictive equations for CPAP titration in OSAS patients, Sleep Breath. 16 (1) (2012 Mar) 95-100, http://dx.doi.org/10.1007/ s11325-010-0461-1, Epub 2011 Jan 6.
[45] A.W. Ho, D.E. Moul, J. Krishna, Neck circumference-height ratio as a predictor of sleep related breathing disorder in children and adults, J. Clin. Sleep Med. 12 (3) (2016) 311-317, http://dx.doi.org/10.5664/jcsm.5572.
[46] T. Akashiba, N. Kosaka, H. Yamamoto, D. Ito, O. Saito, T.T. Horie, Optimal continuous positive airway pressure in patients with obstructive sleep apnoea: role of craniofacial structure, Respir. Med. 95 (5) (2001) 393-397.
[47] J.H. Choi, Y.J. Jun, J.I. Oh, J.Y. Jung, G.H. Hwang, S.Y. Kwon, H.M. Lee, T.H. Kim, S. H. Lee, S.H. Lee, Optimal level of continuous positive airway pressure: auto-adjusting titration versus titration with a predictive equation, Ann. Otol. Rhinol. Laryngol. 122 (5) (2013 May) 339-343.
