Lack of night-to-night variability of sleep-disordered breathing measured during home monitoring. (Original Article).

Abstract

The apnea-hypopnea index (AHI) is an important objective measure used in the diagnosis of sleep-disordered breathing. In affected patients, the AHI has been reported to vary across successive nights. We conducted a multi-channel home sleep study on 44 patients with sleep-disordered breathing to determine whether the AHI does indeed vary and, if so, to quantify the degree of night-to-night variability. Of this group, 23 patients were tested for 3 consecutive nights and 21 were tested for 2 consecutive nights. Among the group as a whole, we found no statistically significant change in AHI across nights, although we did identify variations among individual patients.

Introduction

Researchers who have studied the reputed variability of sleep-disordered breathing across nights have consistently found no significant differences in measurements among groups as a whole (table 1). (1-10) On an individual basis, however, these data have shown that there are increases in the severity of sleep-disordered breathing in some patients, decreases in others, and no change in still others. In these studies, the percentage of subjects whose classification changed across nights according to various cutoff thresholds ranged from 6 to 54.5%.

One hypothesis that has been proposed to explain the variability between first- and second-night recordings is the impact of the "first-night effect." (11-13) The first night of polysomnography recording is often considered to be an "adaptation night" because the subject's adjustment to sleeping in a novel environment while attached to monitoring equipment might lead to a decrease in sleep quality and quantity. If such changes in habitual sleep patterns do have an effect on measurements of sleep-disordered breathing, they might result in an increase in variability across nights until the subject becomes acclimated to sleeping in a laboratory environment. Therefore, in some research studies, first-night data are not used in the final analysis. In clinical studies, however, it is usually not possible to disregard first-night data because a single night of recording is often all that is practically and economically feasible to obtain. Home sleep recordings are becoming more frequently used, but little is known a bout the night-tonight variability in sleep-disordered breathing as measured by such recorders. (14)

In early studies, researchers traditionally determined the variability of sleep-disordered breathing by calculating the apnea-hypopnea index (AHI) and the apnea index (AI). In more recent studies, researchers have used other parameters, such as oxygen saturation and average apnea duration. (5) We conducted a sleep study to investigate the nightly variability in sleep-disordered breathing as determined by the AHI and AI in a home environment.

Patients and methods

In 1999, we sought to enroll in a study patients who had come to our Sleep Disorders Clinic with symptoms of sleep-disordered breathing (primarily snoring) and who had been recommended and approved for sleep testing. Participation was asked of all patients except for minors, pregnant women, and patients with dementia. Each candidate was asked if he or she would undergo an institutional-review-board-approved multichannel home sleep study for 3 consecutive nights.

Of the 59 eligible patients who were invited into the study, 44 (74.6%) agreed to participate. Those who declined did so for reasons of inconvenience or a lack of time. The study group had a mean age of 48.0 years ([+ or -]12.4), a mean neck circumference of 40.8 cm ([+ or -]4.7), and a mean body mass index of 30.1 ([+ or -]8.5). Study participants were provided with an AutoSet Portable H Plus multichannel home sleep-test device (currently distributed as the Embletta; ResMed; Poway, Calif.) to record overnight respiration patterns.

We defined an episode of apnea as a decrease of more than 75% in nasal ventilation for at least 10 seconds. Hypopnea was defined as a decrease of 50 to 75% in nasal ventilation for at least 8 seconds. The AHI and AI were calculated by an automated scoring device. The denominator for these calculations was the total recording time. We used repeated analysis of variance (ANOVA) measurements to determine changes in respiratory variables across nights, using night as the independent variable and respiration as the dependent variable. Results were compiled from data obtained during the first 2 nights from all 44 patients and during 3 nights from those patients who completed the entire study. A linear relationship was fit to the data across 2 nights, and linear and quadratic relationships were fit to the 3-night data.

We estimated the probability that a diagnostic classification would change across 2 nights of recording, as determined by a particular AHI cutoff point. The frequency that such changes occurred was calculated on the basis of AHI cutoffs of 15, 20, and 30. At each cutoff point, we calculated the percentage of patients who were below the cutoff point on night 1 and above the cutoff point on night 2 and vice versa. Finally, we calculated the magnitude of change in sleep-disordered breathing between nights 1 and 2 by subtracting the absolute AHI value on night 1 from the absolute AHI value on night 2.

Results

All 44 patients participated in the first 2 nights of the study. Only 23 took part on the third night; the remaining 21 declined to continue, primarily because of the inconvenience.

Overall, there was no statistically significant difference in AHI or AI between nights 1 and 2, suggesting that there was no significant respiratory first-night effect (table 2). Based on an AHI cutoff point of 15 or greater, three patients (6.8%) who were below the cutoff point on night 1 were above the cutoff on night 2. No patient who was above this cutoff on night 1 dropped below it on night 2. At a cutoff point of 20 or more, four patients (9.1%) who were below the cutoff point on night 1 were above it on night 2; four other patients (9.1%) experienced the opposite pattern. At a cutoff point of 30 or more, five patients (11.4%) who were below the cutoff point on night 1 were above it on night 2, while four others (9.1%) experienced the opposite pattern.

The mean magnitude of change in AHI between nights 1 and 2 was 7.1 ([+ or -]8.8). The change in AHI was significantly correlated only with age, as older patients tended to experience greater variability (r = 0.38; p<0.05).

A Bland-Altman plot (15) of AHI recordings revealed consistent patterns across nights--that is, there was a decrease in indices from night 1 to night 2 (figure). Less of a change was observed between nights 2 and 3; some variables reflected a small increase and others a decrease.

Discussion

Our study represents the first investigation and documentation of night-to-night variability of sleep-disordered breathing as recorded by a multichannel home sleep device. The pattern of correlations observed in our study is comparable to those seen in previous studies, as we found no statistically significant changes across nights. The lack of changes might be attributable to the variability in the direction of change in that those patients who experienced an increase in the severity of sleep-disordered breathing cancelled out those who experienced a decrease in severity. A few patients did experience some variability across nights, particularly one who experienced a change in AHI of 40 across 2 nights.

The diagnostic implications of the variability of sleep-disordered breathing are clinically meaningful. On average, we found a change in AHI of approximately 7 events per hour between nights 1 and 2. Based on AHI cutoff points of 15, 20, and 30, changes in classification were seen in 6.8, 18.2, and 20.5% of patients, respectively.

The results of our study suggest that during unattended home sleep recordings, some individual night-to-night variability in sleep-disordered breathing across nights does occur. Given that the first-night effect should be less intense during home testing than during laboratory testing, our results do not support the hypothesis that the variability seen across nights in sleep-disordered breathing is the result of acclimation to the recording equipment. Instead, the variability probably reflects just the nature of sleep and sleep-disordered breathing.

The use of the AHI and other indices of the degree of sleep-disordered breathing is predicated on the assumption that the severity of this disorder can be quantified in a single value. The test/retest variability of a single measure has important implications when investigative or clinical decisions are based on this value.

Table 1

Summary of research on sleep-disordered breathing

Study                      Subjects

Lee and Giblin, 1982 (1)   14 healthy men



Bliwise et al, 1983 (2)    66 healthy adults




Wittig et al, 1984 (3)     22 men with sleep-
                           disordered breathing



Kramer, 1988 (4)           200 patients with sleep-
                           disordered breathing

Mosko et al, 1988 (5)      46 community-dwelling
                           elderly

Aber et al, 1989 (6)       14 healthy older men




Lord et al, 1991 (7)       30 older adults




Meyer et al, 1993 (8)      11 patients with sleep-
                           disordered breathing

Mendelson, 1994 (9)        50 patients with sleep-
                           disordered breathing


Masaquel et al, 1997 (10)  60 healthy older adults




Davidson et al, 2003 +     44 patients with sleep-
                           disordered breathing


Study                      Age (yr)

Lee and Giblin, 1982 (1)   Range: 49 to 72



Bliwise et al, 1983 (2)    Range: 44 to 88;
                           mean: 67.2 ([+ or -]11.2



Wittig et al, 1984 (3)     Mean: 49.7 ([+ or -]9.8)




Kramer, 1988 (4)           N/A


Mosko et al, 1988 (5)      Range: 60 to 95;
                           mean: 68.7 ([+ or -]6.7)

Aber et al, 1989 (6)       Range: 61 to 83;
                           mean: 66.1 ([+ or -]5.7)



Lord et al, 1991 (7)       Mean: 78 ([+ or -]5)




Meyer et al, 1993 (8)      Mean: 44.9 ([+ or -]4.2)


Mendelson, 1994 (9)        Mean: 50.2 ([+ or -]2.3)



Masaquel et al, 1997 (10)  Range: 65 to 92;
                           mean: 74.2 ([+ or -]6.6)



Davidson et al, 2003 +     48.0 ([+ or -]12.4)



Study                      Results

Lee and Giblin, 1982 (1)   Overall decrease in AHI; 28.5%
                           crossed cutoff of 5 over 3
                           nights

Bliwise et al, 1983 (2)    Overall nonsignificant
                           increase in RDI *; 18.2%
                           crossed cutoff of 5 over 2
                           nights

Wittig et al, 1984 (3)     Correlation of 0.88 between
                           nights 1 and 2; more
                           variability for those with
                           fewer apneas

Kramer, 1988 (4)           Between 12.5 and 17.5%
                           crossed cutoffs from 5 to 20

Mosko et al, 1988 (5)      43% crossed cutoff of 5 over 3
                           nights

Aber et al, 1989 (6)       Reliability of sleep-
                           disordered breathing indices
                           from 0.63 to 0.89 between
                           nights 1 and 2

Lord et al, 1991 (7)       14.3% crossed cutoff of 15
                           over 2 nights; more
                           variability for those with
                           high AHIs

Meyer et al, 1993 (8)      54.5% crossed cutoff of 5 over
                           2 nights

Mendelson, 1994 (9)        6% crossed cutoff of 5 over 2
                           nights; 12% crossed cutoff of
                           10 over 2 nights

Masaquel et al, 1997 (10)  Reliability of 0.92 for RDI
                           between 2 nights; 14, 20, and
                           22% crossed cutoffs of 5, 10,
                           and 20, respectively

Davidson et al, 2003 +     6.8, 18.2, and 20.5% crossed
                           cutoffs of 15, 20, and 30,
                           respectively

* Respiratory disturbance index.

+ Present study.

Table 2

Mean AHI and AI

             Night 1              Night 2              Night 3
Index       (n = 44)             (n = 44)             (n = 23)

AHI    19.8 ([+ or -]15.0)  19.5 ([+ or -]15.1)  17.3 ([+ or -]13.0)
AI     10.6 ([+ or -]14.6)  10.1 ([+ or -]13.8)   8.8 ([+ or -]12.3)

References

(1.) Lee K, Giblin E. Reliability of a one-night diagnostic study for sleep apnea [abstract]. Sleep Research 1982;l1:155.

(2.) Bliwise DL, Carey E, Dement WC. Nightly variation in sleep-related respiratory disturbance in older adults. Exp Aging Res 1983;9:77-81.

(3.) Wittig RM, Romaker A, Zorick FJ, et al. Night-to-night consistency of apneas during sleep. Am Rev Respir Dis 1984;129:244-6.

(4.) Kramer M. Obstructive sleep apnea--one night is not enough [abstract]. Sleep Research 1988;17:205.

(5.) Mosko SS, Dickel MJ, Ashurst J. Night-to-night variability in sleep apnea and sleep-related periodic leg movements in the elderly. Sleep 1988;11:340-8.

(6.) Aber WR, Block AJ, Hellard DW, Webb WB. Consistency of respiratory measurements from night to night during the sleep of elderly men. Chest 1989;96:747-5l.

(7.) Lord S. Sawyer B, O'Connell D, et al. Night-to-night variability of disturbed breathing during sleep in an elderly community sample. Sleep 1991;14:252-8.

(8.) Meyer TJ, Eveloff SE, Kline LR, Millman RP. One negative polysomnogram does not exclude obstructive sleep apnea. Chest 1993;103:756-60.

(9.) Mendelson WB. Use of the sleep laboratory in suspected sleep apnea syndrome: Is one night enough? Cleve Clin J Med 1994; 61:299-303.

(10.) Masaquel A, Srepnowsky C, Estline E, et al. Night-to-night variability in sleep disordered breathing in elderly African-Americans recorded at home [abstract]. Sleep Research 1997;26:675.

(11.) Agnew HW, Jr., Webb WB, Williams RL. The first night effect: An EEG study of sleep. Psychophysiology 1966;2:263-6.

(12.) Schmidt HS, Kaelbling R. The differential laboratory adaptation of sleep parameters. Biol Psychiatry 1971;3:33-45.

(13.) Webb WB, Campbell SS. The first night effect revisited with age as a variable. Waking Sleeping 1979;3:319-24.

(14.) Davidson TM, Do KL, Justus S. The use of ENT-prescribed home sleep studies for patients with suspected obstructive sleep apnea. Ear Nose Throat J 1999;78:754, 757, 760-2.

(15.) Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;l(8476):307-l0.

From the Department of Otolaryngology-Head and Neck Surgery, University of California, San Diego.

Reprint requests: Terence M. Davidson, MD, Professor of Otolaryngology-Head and Neck Surgery, VA San Diego Healthcare System (112C), 3350 La Jolla Village Dr., San Diego, CA 92161. Phone: [phone], ext. 3405; fax: [phone]; e-mail: [e-mail]

The study described in this article was supported by a grant from the San Diego Foundation's Farrell Fund.