Continuous Positive Airway Pressure Reduces ... - ATS Journals

Continuous Positive Airway Pressure Reduces Postprandial Lipidemia in Obstructive Sleep Apnea A Randomized, Placebo-Controlled Crossover Trial Craig L. Phillips1,2,3, Brendon J. Yee1,2,4, Nathaniel S. Marshall1,2, Peter Y. Liu1,2,4, David R. Sullivan5, and Ronald R. Grunstein1,2,4 1

Sleep and Circadian Research Group, Woolcock Institute of Medical Research, Sydney, New South Wales, Australia; 2Discipline of Sleep Medicine, Sydney Medical School, University of Sydney, Sydney, New South Wales, Australia; 3Department of Respiratory and Sleep Medicine, Royal North Shore Hospital, St. Leonards, Sydney, New South Wales, Australia; 4Department of Respiratory and Sleep Medicine, and 5Biochemistry Department, Royal Prince Alfred Hospital, Camperdown, Sydney, New South Wales, Australia

Rationale: Dyslipidemia is common in Obstructive Sleep Apnea (OSA). Postprandial lipidemia (PPL) is a strong marker of cardiovascular risk. Evidence that OSA treatment improves PPL is lacking. Objectives: To investigate the effect of continuous positive airway pressure (CPAP) treatment on postprandial lipidemia (PPL) in patients with obstructive sleep apnea (OSA) in the upper moderate or severe range. Methods: In this randomized, placebo-controlled crossover trial, we compared the effects of 2 months each of therapeutic and placebo CPAP on PPL. Measurements and Main Results: PPL was determined from the area under the 24-hour triglyceride concentration curve (TAGAUC24) using seven blood samples drawn across both the wake and sleep periods. Secondary outcomes were the difference in other 24-hour lipid profiles. Thirty-eight eligible patients were randomly assigned to a treatment order and 29 patients completed the trial. CPAP reduced PPL compared with placebo with a mean TAGAUC24 difference of 2357 mmol/L/d (95% confidence interval [CI], 2687.3 to 226.8; P ¼ 0.035). During both the CPAP and placebo studies, TAG levels peaked during both wakefulness (2:00 P.M.) and sleep (3:00 A.M.). Both peaks were lower during CPAP than placebo: 2:00 P.M., 20.49 mmol/L (95% CI, 20.74 to 20.24; P , 0.0005) and 3:00 A.M., 20.40 mmol/L (95% CI, 20.65 to 20.15; P ¼ 0.002). Moreover, mean 24-hour total cholesterol was 20.19 mmol/L lower on CPAP (95% CI, 20.27 to 20.11; P , 0.00001). Conclusions: This randomized trial demonstrated that treatment of severe OSA with CPAP improves postprandial TAG and total cholesterol levels. These effects may reduce the risk for cardiovascular events. The results imply that the association between OSA and cardiovascular disease may, in part, be caused by direct effects on dyslipidemia. Clinical trial registered with the Australian and New Zealand Clinical Trials Registry at www.anzctr.org.au (ACTRN 12605000066684). Keywords: non-fasting; triglycerides; cholesterol; sham CPAP; cardiovascular risk

(Received in original form February 21, 2011; accepted in final form April 18, 2011) Supported by Australian National Health and Medical Research Council project grant 301936 (R.R.G.). Contributors: All authors equally contributed to the concept and design of the trial. R.R.G., P.Y.L., and C.L.P. were responsible for obtaining funding. C.L.P. and N.S.M. did the statistical analyses. C.L.P., B.J.Y., and R.R.G. supervised the study. D.R.S. provided access to all biochemistry analyses. All authors participated in the overall analysis and interpretation of the data and revision of the manuscript and provided final approval of the submitted version. Correspondence and requests for reprints should be addressed to Craig L. Phillips, Ph.D., Department of Respiratory and Sleep Medicine, Royal North Shore Hospital, St. Leonards, New South Wales 2065 Australia. E-mail: [email protected] Am J Respir Crit Care Med Vol 184. pp 355–361, 2011 Originally Published in Press as DOI: 10.1164/rccm.201102-0316OC on April 28, 2011 Internet address: www.atsjournals.org

AT A GLANCE COMMENTARY Scientific Knowledge on the Subject

Obstructive Sleep Apnea (OSA) has been associated with dyslipidemia, which can lead to cardiovascular morbidity and mortality. Postprandial levels of triglycerides (TAGs) or postprandial lipidemia are stronger markers of cardiovascular risk than fasting levels. Demonstrating that OSA treatment improves postprandial TAG metabolism would be an important finding that affects a large population. What This Study Adds to the Field

This randomized trial showed that CPAP reduced postprandial levels of TAGs and total cholesterol. These changes may have modest effects in reducing cardiovascular risk. Further research is needed to establish whether CPAP alone or combined with other lipid-lowering medication is useful in improving postprandial lipidemia in patients with OSA and dyslipidemia.

Obstructive sleep apnea (OSA) is an emerging risk factor for incident all-cause and cardiovascular mortality (1–4) and for incident myocardial infarction, heart failure, and stroke (4–6). Long-term treatment of OSA with nasal continuous positive airway pressure (CPAP) has been associated with a significant reduction in cardiovascular events and mortality to levels found in otherwise healthy individuals (4). Although these outcome improvements may be linked to a lowering of blood pressure by CPAP (7), it is still unclear whether improvements in other established risk factors also play a role. Dyslipidemia is common in OSA, but no observational studies have reported OSA as a preceding risk factor. Furthermore, there are only a small number of randomized trials that have examined the effect of CPAP on fasting lipid profiles (8–11), and none have examined lipids as a primary outcome. Although some data show that CPAP improves the fasting level of total cholesterol (11), the overall findings do not provide any clear evidence that effective treatment of OSA improves lipid metabolism. Although the prognostic usefulness of most lipids depends on their assessment in the fasted state, there is emerging evidence to suggest that this may not apply for triglycerides (TAG). In the fasted state, the association between TAGs and cardiovascular risk tends to be markedly attenuated after adjustment for highdensity lipoprotein (HDL) cholesterol and other established risk factors (12). However, two prospective community cohort studies have recently found that postprandial (nonfasting) TAGs are strongly associated with incident cardiovascular-related disease,

356

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

events, and death independent of other risk factors (13, 14). This has prompted debate as to whether an oral (high fat) TAG tolerance test (similar to the oral glucose tolerance test) could be used to improve risk stratification in individuals with normal fasting levels (15). Apart from insulin resistance, there is some evidence that TAG metabolism may be strongly influenced by sleep and circadian processes with a marked elevation of levels during sleep (16). Given that OSA is characterized by sleep disturbance and is commonly associated with insulin resistance, we aimed to investigate whether CPAP treatment would improve postprandial triglyceridemia over a full 24 hours that encompassed both wake and sleep periods. Instead of testing patients under a single (supraphysiological) high-fat meal condition over several hours during the day, TAG levels were assessed with provision of isocaloric western-style meals of predefined nutritional content. The results from this study have been previously reported in the form of abstracts (17–19).

METHODS Design Overview This is a randomized crossover trial comparing the effects of 2 months of CPAP to sham-CPAP (i.e., placebo) on lipid metabolism in patients with OSA (Figure 1).

Setting and Participants From October 2006 to September 2009 we recruited patients from three tertiary referral clinics to participate in the study: the Royal Prince Alfred and St. Vincent’s Hospitals and the Woolcock Institute of Medical Research (University of Sydney), Sydney, Australia. Adult patients (aged . 21 yr) with a diagnosis of OSA with severity in the upper moderate or severe range (apnea-hypopnea index [AHI] > 25/h of sleep) and/or associated with a significant component of hypoxia (oxygen desaturation index > 20 per hour; desaturation > 3% of baseline) by overnight in-laboratory polysomnography were recruited. Main exclusion criteria were a body mass index greater than 35 kg/m2, fasting TAGs greater than or equal to 4 mmol/L, use of fibrate medication, previous CPAP use, uncontrolled type II diabetes, and any clinically significant comorbidity (e.g., cardiovascular, pulmonary, renal, or psychiatric disease). This latter criterion precluded patients with heart failure or lung disease from participating in the trial. Additional minor criteria are listed in the Australian and New Zealand Clinical Trials Registry (ACTRN 12605000066684; available at http://www.anzctr.org.au).

VOL 184

2011

Interventions and Procedures CPAP. The real and placebo CPAP devices (Remstar Auto; Philips Respironics, Murrysville, PA) used in this trial are identical in appearance and are currently being used in other large-scale randomized controlled trials (20, 21). Before treatment each patient was fitted with a CPAP mask and was given instruction on device operation. Patients also received cognitive behavior therapy aimed at maximizing compliance with both devices. Before each treatment arm, each patient first underwent either a real auto-titrating CPAP or mock pressure determination study at home to determine individual therapeutic pressures at which to set their real CPAP device (Figure 1). In the mock study, patients slept with the device set at 0.5 cm H2O pressure. An adequate study on each device required greater than or equal to 4.5 hours of usage in a night. During real CPAP auto-titration we also required an average mask leak of less than 0.4 L/s with a mean AHI of less than or equal to 10/h before determining the optimum pressure at which to set the treatment device. The optimum pressure was usually set to the 90th centile pressure calculated by the device that controlled most sleep apnea, as routinely used in clinical practice. However, if there was evidence of residual OSA on CPAP, then the pressure was increased by up to 2 cm H2O above the 90th centile pressure. 24-hour Laboratory Studies. Baseline and outcome data were assessed in a 24-hour in-laboratory protocol (Figure 1) at the Woolcock Institute. Blood samples for lipid analyses were drawn at seven time points across the 24-hour sleep/wake period. Patients were briefly awoken during the sleep period for the midnight and 3:00 A.M. blood draws. At predetermined times patients were fed breakfast, lunch, and dinner as well as a snack in the midmorning and midafternoon. The meals were representative of a standard western diet with a total energy of 9433 kJ (30.5% fat, 54.6% carbohydrate, 14.9% protein). Throughout the studies, physical activity was kept to a minimum. After arriving at the laboratory patients were asked to empty their bladder before collecting all subsequent urine for catecholamine analysis during the remainder of their waking day. A separate bottle was provided immediately before sleep for collections during the night and for the first post-awakening void. During the night, patients were asked to use each of their treatment devices as they had been doing during the 2 months at home.

Outcome Measures The primary outcome was the area under the 24-hour TAG concentration curve (TAG-AUC24). Secondary outcomes included the seven time point 24-hour lipid profiles in TAG, HDL, total cholesterol, free fatty acids, and urinary catecholamines. Because estimated nonfasting LDL cholesterol is unreliable (using the Friedewald equation), we instead calculated non-HDL cholesterol determined from total cholesterol minus HDL cholesterol (22). All assays were performed by the biochemistry department at Royal Prince Alfred Hospital using standard techniques. At each visit patients also completed the Epworth Sleepiness Scale and the Functional Outcomes of Sleep Questionnaire, which are subjective measures of sleepiness (23, 24). At baseline we assessed Apolipoprotein E (APOE) genotype to identify the proportion of patients with the E2:E2 alleles, which have been associated with delayed remnant lipoprotein clearance (25).

Randomization, Allocation Concealment, and Blinding

Figure 1. Schematic of the trial protocol. PD: at-home pressure determination study (1–4 nights). Laboratory study: 24-hour outcomes collection. End-of-treatment studies were performed with continuous positive airway pressure (CPAP) or placebo according to at-home usage.

We used a computer program to produce the random treatment sequence using random block sizes of two, four, and six. These were stored in sequentially numbered opaque envelopes. The project manager was responsible for the allocation consignment and had no contact with any patient before or during the trial. Randomization occurred after the research nurse determined patient eligibility, obtained informed consent, and enrolled the patient. We attempted to blind outcomes assessors by limiting knowledge of treatment allocation to two people—the project manager and the trial physician—neither of whom interacted with the patients during the trial. The exception to this was when the trial physician was involved in withdrawing a patient. The project manager was responsible for configuring and monitoring compliance with all CPAP devices, including determining the therapeutic pressure at which to set the real CPAP

Phillips, Yee, Marshall, et al.: OSA Treatment with CPAP Reduces Postprandial Lipidemia

device. When unintentional unblinding occurred in an outcome assessor, they were no longer permitted to have contact with the participant for the remainder of the trial. The CPAP therapist charged with resolving patient–CPAP interface problems was often unblinded by clinical interaction with patients. However, she played no role in the collection or processing of any outcome measurements. We attempted to blind patients by telling them that we were testing two CPAP machines that “deliver pressurized air in a different way.” During a post-study debriefing interview patients were informed that the low-pressure machine was a placebo.

Statistical Analysis The trial was powered to detect a TAG-AUC24 difference between CPAP and placebo based on a previous observed day-long difference of 4.1 mmol/h/L (SD, 7.35) between mildly obese and lean participants (26). It was determined that 28 patients completing a crossover study would provide 80% power at the 5% significance level to detect this difference. We estimated that 25% of patients would drop out of this intensive and invasive trial, requiring 38 randomizations. The primary outcome (TAG-AUC24) was analyzed using a paired t test. Our analysis included all 29 patients who completed the study regardless of compliance with treatment. Among these patients there were only four missed TAG samples. These missing data were imputed with the mean of the preceding and subsequent samples from that day. We also performed an intention-to-treat analysis on the primary outcome. All serum lipids were then analyzed across 24 hours using mixed models. The treatment, order, time point, and treatment by time point

357

interaction were included as fixed effects, and the patient was a random effect in all models. PASW statistics version 17 and SAS v9.2 were used by two independent data analysts (C.L.P. and N.S.M.).

RESULTS The flow of patients through the study is shown in Figure 2. Nine patients withdrew or were withdrawn post randomization, including three before the baseline study (see Figure 2 for reasons). One of these patients had hypertriglyceridemia that was not detected before randomization because of an unexpected delay in analyzing the screening blood sample. This patient was subsequently withdrawn by the investigators for failing to meet the inclusion criteria. The baseline characteristics of the randomized patients are shown in Table 1. Severe OSA (AHI > 30) was present in 24 of the 29 patients who completed the trial. No randomized patients had the APOE E2:E2 allele, which is known to delay remnant lipoprotein clearance (25). Table 2 compares the effects of CPAP and placebo on sleep apnea, urinary catecholamines, and lipid outcomes. The primary outcome TAG-AUC24 was lower after therapeutic CPAP than after placebo, and a similar effect was seen using intention-to-treat analysis (results not shown). Importantly, there was no treatment order effect, and no randomized patient commenced any new medication or changed existing medication doses during the trial. Figure 3

Figure 2. Flow diagram of patients in the trial. CPAP ¼ continuous positive airway pressure; OSA ¼ obstructive sleep apnea; TAG ¼ triglyceride.

358


TABLE 1. CHARACTERISTICS OF ALL RANDOMIZED PATIENTS Characteristic Demographics Age, yr Male/female, n BMI, kg/m2 Medical history and medication, n Hypertension Type II diabetes Hypercholesterolemia Antihypertensives Statins Sleep apnea AHI, events/h ODI, events/h SaO2-T90, %TST Min SaO2, % ESS FOSQ APOE genotype, % E2:E3 E2:E4 E3:E3 E3:E4 Urinary catecholamines, nmol/d Adrenaline, (0–80) 24-h Wake Sleep Noradrenaline, (100–420) 24-h Wake Sleep Biochemistry assays, mmol/L Triglycerides, (< 2.5) HDL cholesterol, (1–2.5) LDL cholesterol, (< 3.5) Total cholesterol (< 5.2) Free fatty acids Glucose, (3.0–7.7) HbA1c, (3.5–6.0)

Measure

9 2 10 9 7 41.2 32.7 6.90 76.9 11.2 15.2

(23.9) (22.5) (10.9) (17.2) (4.90) (3.13)

20.6 5.9 52.9 20.6

24.8 (16.6) 29.5 (23.5) 16.4 (16.9) 339.7 (156.9) 359.2 (190.5) 305.8 (142.1) 1.92 1.17 3.13 5.11 522 5.68 5.78

(0.98) (0.28) (0.99) (1.03) (195) (1.49) (0.70)

Definition of abbreviations: AHI ¼ apnea-hypopnea index; BMI ¼ body mass index; ESS ¼ Epworth Sleepiness Scale (maximum sleepiness score ¼ 20); FOSQ ¼ Functional Outcomes of Sleep Questionnaire; HbA1c ¼ hemoglobin A1c; HDL ¼ high-density lipoprotein; LDL ¼ low-density lipoprotein; Min SaO2, % ¼ minimum arterial oxygen saturation; ODI ¼ oxygen desaturation index; SaO2-T90, %TST ¼ percentage of total sleep time spent with arterial oxygen saturation less than 90%, APOE ¼ Apolipoprotein E. Data are presented as mean (SD) unless otherwise noted. Biochemistry assays were obtained from the first fasting sample at the baseline study. Urinary catecholamine assays were performed during the baseline study. Note: Urinary and biochemistry values in parentheses represent the normal range or upper limit of normal.

plots lipid profiles across 24 hours on CPAP and placebo. The main treatment effect over 24 hours, after placebo adjustment, is summarized in Table 3. Body weight did not change during the trial (P ¼ 0.507). Triglyceride levels peaked once during both the wake and sleep periods and both peaks were reduced by CPAP. In addition, total cholesterol levels were significantly lower on CPAP than placebo, particularly in the evening and during the sleep period. The reduction in total cholesterol on CPAP was reflected by a small but clinically irrelevant decrease in HDL cholesterol.

DISCUSSION In this randomized placebo-controlled trial, treating obstructive sleep apnea with CPAP reduced postprandial triglyceridemia (TAG-AUC24). Time-profile analyses revealed two pronounced

2011

TABLE 2. CONTINUOUS POSITIVE AIRWAY PRESSURE VERSUS PLACEBO COMPARISONS Characteristic

49 (13) 35/3 32.1 (4.30)

VOL 184

Postprandial TAG TAG-AUC24, mmol/L/d Urinary catecholamines Adrenaline, nmol/d 24 h Awake Asleep Noradrenaline, nmol/d 24 h Awake Asleep Sleep Apnea AHI, events/h ODI, events/h SaO2-T90, % TST Min SaO2, % Treatment compliance, h/night ESS FOSQ

CPAP

Placebo

P Value

3,688 6 294

4,045 6 329

0.035

25.2 6 2.7 30.3 6 3.8 15.9 6 3.7

25.0 6 2.7 31.2 6 3.8 16.2 6 3.8

0.95 0.86 0.96

285.9 6 23.2 306.5 6 26.6 259.5 6 29.0

347.5 6 23.5 398.2 6 27.0 302.4 6 29.9

0.008 0.01 0.23

6 6 6 6 6 6 6

,0.00001 ,0.00001 ,0.00001 ,0.00001 ,0.05 0.004 0.056

6.8 5.0 1.0 90 4.4 8.1 16.0

6 6 6 6 6 6 6

3.7 3.2 1.6 1.5 2.2 0.8 0.53

40.7 38.1 9.4 79 3.4 9.6 16.7

3.7 3.1 1.5 1.5 2.3 0.8 0.52

Definition of abbreviations: AHI ¼ apnea-hypopnea index; CPAP ¼ continuous positive airway pressure; ESS ¼ Epworth Sleepiness Scale (maximum sleepiness score ¼ 20); FOSQ ¼ Functional Outcomes of Sleep Questionnaire; ODI ¼ oxygen desaturation index; SaO2-T90, %TST ¼ percentage of total sleep time spent with arterial oxygen saturation , 90%.Data are presented as mean 6 SEM. CPAP and placebo measurements were determined during the 24-hour studies at the end of each treatment arm.

peaks in TAG concentrations across the 24-hour period. The first peak occurred during wakefulness at approximately 2:00 P.M. before the lunch meal and approximately 5 to 6 hours after ingestion of breakfast. The second peak occurred in the middle of the sleep period at approximately 3:00 A.M., some 6 to 7 hours after the evening meal. CPAP treatment significantly reduced the magnitude of both peaks. Both OSA and postprandial triglyceridemia are strong emerging risk factors for cardiovascular disease (CVD) (13, 14). This trial provides the first direct evidence that sleep apnea causes disturbed lipid metabolism and that CPAP treatment reduces this important CVD risk factor. The significance of daytime postprandial TAGs as powerful predictors of cardiovascular risk has recently been established in two large cohort studies in the United States (the Women’s Health Study) (13) and Denmark (The Copenhagen City Heart Study and the Danish Population Study) (14). Daytime nonfasted TAGs, measured in more than one-half of the 40,000 participants, were independently predictive of incident cardiovascular events over follow-up periods of approximately 11 and 26 years, respectively. The American study then stratified analyses by time since last meal to show that the greatest risk prediction occurred 2 to 4 hours after a meal (13). The Danish study indicated that the 4-hour post-meal time point coincides with a peak in TAG concentration and that this peak is temporally aligned with peak levels of remnant (atherogenic) lipoprotein cholesterol (14). Together these cohorts suggest that 4-hour post-meal TAG peaks provide the most powerful prediction for CVD. The extent and duration of the increase in postprandial TAG is dependent on a number of exogenous factors, including meal size and composition (27). In addition, obesity (28) and comorbid diseases, including dyslipidemia, metabolic syndrome, and diabetes, can exaggerate this process (16). In healthy participants from the Copenhagen study, the consumption of either a high-fat test meal or normal food resulted in 4-hour post-meal TAG peaks of 2.3 mmol/L and 1.6 mmol/L, respectively. In contrast, in our study we observed that TAG peaks during both

Phillips, Yee, Marshall, et al.: OSA Treatment with CPAP Reduces Postprandial Lipidemia

359

appreciable cardiovascular risk in our patients. Furthermore, because our sampling was only done every 3 to 4 hours, we may have underestimated the true size of the peaks, if those peaks occurred between our measures. Nevertheless, our data demonstrate that the magnitude of peak daytime and nighttime TAGs were up to 0.5 mmol/L lower after CPAP treatment, compared with placebo. In the Women’s Health study the difference between the highest and lowest tertiles in 2- to 4-hour post-meal TAGs was around 0.75 mmol/L, corresponding to an increased risk of cardiovascular events (hazard ratio of 4.48). The data for men in the Copenhagen study implies that an increase from 2.5 to 3.5 mmol/L in daytime post-meal TAGs would be associated with an increased rate of myocardial infarction (MI) of around 50%. Hence, the daytime lowering of post-meal TAGs by 0.5 mmol/L with CPAP in our study may potentially reduce MI risk by around 25% based on the Copenhagen estimate and perhaps by more based on the U.S. study. In addition, the Copenhagen data suggest effect modification by age such that high TAGs cause substantively higher MI risk in people aged less than 55 years. Our patients were 49 years old on average. In addition to TAGs, total cholesterol was significantly reduced by CPAP therapy. A previous study has reported a decrease in fasting total cholesterol with CPAP by 0.28 mmol/L, which the authors estimated represented a 15% reduction in cardiovascular disease risk (11). Our final fasting total cholesterol measurement was 0.22 mmol/L lower on CPAP than placebo, which may confer similar risk reduction. As LDL cholesterol (which was originally a predefined secondary outcome) was mostly noncalculable due to elevated TAGs, we instead assessed the effect of CPAP treatment on non-HDL cholesterol and found it to be lower on CPAP than placebo. Non-HDL cholesterol is more predictive of coronary heart disease risk than LDL cholesterol when TAG levels are elevated because it represents a total measure of all atherogenic lipoproteins (22). These additional findings further support the hypothesis that CPAP treatment will reduce cardiovascular risk. Figure 3. Temporal changes in lipids across the wake and sleep periods during continuous positive airway pressure (CPAP) and placebo end-of-treatment studies Intra–time point differences between CPAP (open diamonds) and placebo (solid squares). The shaded area represents the sleep period during which patients were treated with a CPAP or placebo device. Data are means 6 SEM. P value for CPAP versus placebo: CP ¼ 0.052, *P , 0.05, **P ¼ 0.01, zP , 0.005, zzP , 0.0005. FFA ¼ free fatty acids; HDL ¼ high-density lipoprotein cholesterol; TAG ¼ triglycerides; TChol ¼ total cholesterol.

the day and night were delayed beyond 4 hours and levels during the placebo arm exceeded 3.3 mmol/L. This occurred despite ingesting meals with normal fat loads and likely reflects

Strengths and Limitations of This Study

Placebo control, assessment of lipid metabolism during both wake and sleep, and standardized normal food intake in both treatment studies were important experimental design strengths. However, it is also important to highlight several potential limitations of our study. (1) We did not repeat the baseline study after washout and so we have no reproducibility data for the TAG-AUC24 response after meals. However, we did control for the effect of meal composition and timing by providing identical meals at fixed times during both treatment arm studies. The results clearly demonstrate a double peak in 24-hour TAG profiles

TABLE 3. CONTINUOUS POSITIVE AIRWAY PRESSURE VERSUS PLACEBO EFFECT SIZES Main Effects, CPAP – Placebo Lipid, mmol/L TAG TChol HDL Non-HDL Chol FFA Catecholamine, nmol/d Adrenaline Noradrenaline

20.22 20.19 20.03 20.16 210.5 20.6 267.1

Lower 95% CI 20.31 20.27 20.05 20.24 237.6 27.6 2117.7

Upper 95% CI 20.12 20.11 20.01 20.08 16.6 6.4 216.4

P Value ,0.00001 ,0.00001 ,0.0005 ,0.0001 0.5 0.87 0.01

Definition of abbreviations: FFA ¼ free fatty acids; HDL ¼ high-density lipoprotein; TAG ¼ triglycerides; TChol ¼ total cholesterol.

360


that was temporally aligned (Figure 3) during both CPAP and placebo studies. Therefore, we believe that there is a consistent TAG response to meals and this is further supported by a study showing good reproducibility for the TAG-AUC measurement (29). (2) Unlike 24-hour ambulatory blood pressure, which improves cardiovascular risk stratification, we do not know the extent to which 24-hour TAG levels predict cardiovascular disease. However, given the evidence linking postprandial TAGs (particularly at 4 h post meal) to incident cardiovascular events and death, the post-meal peaks observed in our study likely reflect significant risk that is markedly attenuated with CPAP. The identification of a nocturnal peak during sleep may have additional cardiovascular importance. Nevertheless, unlike the 75-g oral glucose tolerance test, there is currently no equivalent oral fat tolerance test on which to base risk despite recent calls for the development of such a test (15). Ultimately, our data demonstrate that postprandial lipidemia was lower after CPAP than placebo, both during the day and at night, which is suggestive of a lowering of cardiovascular risk. However, it remains to be established whether CPAP reduces overall cardiovascular-related events and death. This evidence must come from large long-term clinical trials with hard endpoints, such as the ongoing SAVE (Sleep Apnea cardioVascular Endpoints) study (NCT00738179). (3) Although we believe that placebo control strengthened our study, a potential limitation was that compliance on sham CPAP was lower than therapeutic CPAP, albeit only by 1 hour. This was mainly influenced by a minority of patients who were good compliers with therapeutic CPAP but poor compliers with sham CPAP. In general, compliance on the two devices correlated (r ¼ 0.55, P , 0.005), which suggests that compliance could be a preexisting patient trait. (4) It is also important to recognize that the interpretation of these results is limited to patients with predominantly severe OSA. Future studies will need to determine whether the results are reproducible in patients with less severe OSA. (5) Finally, the mechanisms by which OSA causes PPL remains to be studied, as does the potential for sex to modify the effect. The influence of sex is particularly important given the higher cardiovascular risk associated with postprandial lipidemia in females (13). Although it is well recognized that statin treatment for dyslipidemia is highly effective, no studies have shown whether these pharmacotherapeutic approaches are equally efficacious in patients with OSA. In addition, although statins have proven benefits for cardiovascular events and mortality in secondary prevention trials (30, 31), similar benefits in primary prevention trials are still unproven (32, 33). It is also noteworthy that CPAP has additional benefits related to neurocognitive and sleepiness improvements that we are not aware any lipid-lowering medication could provide. The mechanism by which OSA may exacerbate dyslipidemia is unclear. The hallmarks of OSA are frequent arousals from sleep coupled with intermittent hypoxia. Protracted exposure to intermittent hypoxia in lean mice activates biosynthesis pathways for both TAGs and total cholesterol (34). The abolition of intermittent hypoxia with CPAP in humans may reduce TAG synthesis in the liver; however, it is unclear whether TAG absorption in the intestine would be similarly affected. Circulating TAGs are highly dependent on food ingestion, and it is therefore puzzling that we did not observe a post-lunch peak in the evening in conjunction with our observed post-breakfast and post-dinner peaks. This may be explained by recent evidence of strong circadian and sleep/wake effects on TAG metabolism (16). During experiments in which sleep is denied and small hourly meals are consumed, two peaks in circulating TAGs over the 24-hour cycle have been identified. The timing of these peaks coincided

VOL 184

2011

exactly with the two peaks that occurred in our patients and suggests that apart from meals, TAG metabolism is likely to be strongly influenced by circadian and sleep/wake influences. Our data suggest that CPAP treatment of OSA may influence these endogenous effects. However they do not appear to depend on sympathetically driven lipolysis (35), because a marked reduction in daytime sympathetic activity with CPAP therapy (as indicated by a reduction in urinary noradrenaline) did not alter circulating free fatty acids. Conclusions

When compared with placebo, 2-month CPAP treatment of patients with predominantly severe OSA resulted in a lowering of post-meal TAG, total cholesterol, non-HDL cholesterol, and daytime urinary noradrenaline. The TAG effects were particularly pronounced for peak concentrations observed during both wakefulness and sleep. Given that peak TAGs have the greatest prognostic significance for future cardiovascular events, CPAP may impart at least a modest benefit in lowering risk. However, the role for CPAP in improving postprandial lipid metabolism and reducing cardiovascular-related events and death must await long-term randomized trials. Author Disclosure : The institution of C.L.P., B.J.Y., N.S.M., P.Y.L., D.R.S., and R.R.G. received support from Philips Respironics, the CPAP manufacturer that provided the active and sham (placebo) CPAP machines for use in the study. D.R.S. has received reimbursement for serving on the advisory board, consultation, and postgraduate lecturing for Merck Sharp, Dohme/Schering Plough, Pfizer, AstraZenica, and Abbott. R.R.G.’s institution received payment from Fisher-Paykel Healthcare for validation of new CPAP technology. Acknowledgment : The authors thank a number of research staff at the Woolcock Institute of Medical Research. They include Pam Gee, Josie Dungan, Dr. Keith Wong, Angela Denotti, Kerri Melehan, Gislaine Gauthier, Assoc Prof Delwyn Bartlett, Dianne Richards, Nicole Lai, Marilyn Yee, and George Dungan. They also thank their patients for their valuable time, without which this study would not have been completed.

References 1. Marshall NS, Wong KK, Liu PY, Cullen SR, Knuiman MW, Grunstein RR. Sleep apnea as an independent risk factor for all-cause mortality: the Busselton Health Study. Sleep 2008;31:1079–1085. 2. Young T, Finn L, Peppard P, Szklo-Coxe M, Austin D, Nieto FJ, Stubbs R, Hla KM. Sleep disordered breathing and mortality: eighteen-year follow-up of the Wisconsin sleep cohort. Sleep 2008;31:1071–1078. 3. Punjabi NM, Caffo BS, Goodwin JL, Gottlieb DJ, Newman AB, O’Connor GT, Rapoport DM, Redline S, Resnick HE, Robbins JA, et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med 2009;6:e1000132. 4. Marin JM, Carrizo SJ, Vicente E, Agusti AG, et al. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet 2005;365:1046–1053. 5. Gottlieb DJ, Yenokyan G, Newman AB, O’Connor GT, Punjabi NM, Quan SF, Redline S, Resnick HE, Tong EK, Diener-West M, et al. Prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure. The Sleep Heart Health Study. Circulation 2010;122:352–360. 6. Redline S, Yenokyan G, Gottlieb DJ, Shahar E, O’Connor GT, Resnick HE, Diener-West M, Sanders MH, Wolf PA, Geraghty EM, et al. Obstructive sleep apnea hypopnea and incident stroke: The Sleep Heart Health Study. Am J Respir Crit Care Med 2010;182:269–277. 7. Bazzano LA, Khan Z, Reynolds K, He J. Effect of nocturnal nasal continuous positive airway pressure on blood pressure in obstructive sleep apnea. Hypertension 2007;50:417–423. 8. Comondore VR, Cheema R, Fox J, Butt A, John Mancini GB, Fleetham JA, Ryan CF, Chan S, Ayas NT. The impact of CPAP on cardiovascular biomarkers in minimally symptomatic patients with obstructive sleep apnea: a pilot feasibility randomized crossover trial. Lung 2009; 187:17–22.

Phillips, Yee, Marshall, et al.: OSA Treatment with CPAP Reduces Postprandial Lipidemia 9. Coughlin SR, Mawdsley L, Mugarza JA, Wilding JPH, Calverley PMA. Cardiovascular and metabolic effects of CPAP in obese men with OSA. Eur Respir J 2007;29:720–727. 10. Drager LF, Bortolotto LA, Figueiredo AC, Krieger EM, Lorenzi GF. Effects of continuous positive airway pressure on early signs of atherosclerosis in obstructive sleep apnea. Am J Respir Crit Care Med 2007;176:706–712. 11. Robinson GV, Pepperell JCT, Segal HC, Davies RJO, Stradling JR. Circulating cardiovascular risk factors in obstructive sleep apnoea: data from randomised controlled trials. Thorax 2004;59:777–782. 12. Sarwar N, Danesh J, Eiriksdottir G, Sigurdsson G, Wareham N, Bingham S, Boekholdt SM, Khaw KT, Gudnason V. Triglycerides and the risk of coronary heart disease: 10,158 incident cases among 262,525 participants in 29 Western prospective studies. Circulation 2007;115:450–458. 13. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM. Fasting compared with nonfasting triglycerides and risk of cardiovascular events in women. JAMA 2007;298:309–316. 14. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA 2007;298:299–308. 15. Ridker PM. Fasting versus nonfasting triglycerides and the prediction of cardiovascular risk: do we need to revisit the oral triglyceride tolerance test? Clin Chem 2008;54:11–13. 16. Rajaratnam SM, Arendt J. Health in a 24-h society. Lancet 2001;358: 999–1005. 17. Phillips C, Yee B, Marshall N, Liu P, Sullivan D, Grunstein R. Treatment of obstructive sleep apnea (OSA) with nasal continuous positive airway pressure (CPAP) reduces post-prandial lipidemia (PPL) [abstract]. Presented at the 24th annual meeting of the Associated Professional Sleep Societies (APSS). June 11–15, 2010, San Antonio, Texas. 18. Phillips CL, Yee BJ, Marshall NS, Liu PY, Sullivan DR, Grunstein RR. Treatment of obstructive sleep apnoea reduces post-prandial lipidemia: evidence from a randomized, placebo-controlled trial of continuous positive airway pressure [abstract]. J Sleep Res 2010;19: 513. 19. Phillips C, Yee B, Marshall N, Liu P, Sullivan D, Grunstein R. Treatment of obstructive sleep apnea (OSA) with nasal continuous positive airway pressure (CPAP) reduces postprandial lipidemia (PPL). A randomised placebo-controlled crossover study [abstract]. Sleep Biol Rhythms 2010;8:A18. 20. Kushida CA, Nichols DA, Quan SF, Goodwin JL, White DP, Gottlieb DJ, Walsh JK, Schweitzer PK, Guilleminault C, Simon RD, et al. The Apnea Positive Pressure Long-term Efficacy Study (APPLES): rationale, design, methods, and procedures. J Clin Sleep Med 2006;2: 288–300. 21. Rodway GW, Weaver TE, Mancini C, Cater J, Maislin G, Staley B, Ferguson KA, George CF, Schulman DA, Greenberg H, et al.

22.

23. 24.

25. 26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

361

Evaluation of sham-CPAP as a placebo in CPAP intervention studies. Sleep 2010;33:260–266. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002;106:3143–3421. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991;14:540–545. Weaver TE, Laizner AM, Evans LK, Maislin G, Chugh DK, Lyon K, Smith PL, Schwartz AR, Redline S, Pack AI, et al. An instrument to measure functional status outcomes for disorders of excessive sleepiness. Sleep 1997;20:835–843. Rubin J, Berglund L. Apolipoprotein E and diets: a case of gene-nutrient interaction? Curr Opin Lipidol 2002;13:25–32. Halkes CJ, Van Wijk JP, Ribalta J, Masana L, Castro Cabezas M. Diurnal triglyceridaemia and insulin resistance in mildly obese subjects with normal fasting plasma lipids. J Intern Med 2004;255:74–81. Lairon D, Lopez-Miranda J, Williams C. Methodology for studying postprandial lipid metabolism. Eur J Clin Nutr 2007;61:1145–1161. Blackburn P, Lamarche B, Couillard C, Pascot A, Tremblay A, Bergeron J, Lemieux I, Despre´s JP. Contribution of visceral adiposity to the exaggerated postprandial lipemia of men with impaired glucose tolerance. Diabetes Care 2003;26:3303–3309. Weiss EP, Fields DA, Mittendorfer B, Haverkort MAD, Klein S. Reproducibility of postprandial lipemia tests and validity of an abbreviated 4-hour test. Metabolism 2008;57:1479–1485. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383–1389. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. N Engl J Med 1998;339:1349–1357. Ray KK, Seshasai SRK, Erqou S, Sever P, Jukema JW, Ford I, Sattar N. Statins and all-cause mortality in high-risk primary prevention: a meta-analysis of 11 randomized controlled trials involving 65 229 participants. Arch Intern Med 2010;170:1024–1031. de Lorgeril M, Salen P, Abramson J, Dodin S, Hamazaki T, Kostucki W, Okuyama H, Pavy B, Rabaeus M. Cholesterol lowering, cardiovascular diseases, and the rosuvastatin-JUPITER controversy: a critical reappraisal. Arch Intern Med 2010;170:1032–1036. Li J, Thorne LN, Punjabi NM, Sun C-K, Schwartz AR, Smith PL, Marino RL, Rodriguez A, Hubbard WC, O’Donnell CP, et al. Intermittent hypoxia induces hyperlipidemia in lean mice. Circ Res 2005;97:698–706. Hucking K, Hamilton-Wessler M, Ellmerer M, Bergman RN. Burst-like control of lipolysis by the sympathetic nervous system in vivo. J Clin Invest 2003;111:257–264.