Snoring and Sleep Architecture 1 , 2

v.

HOFFSTEIN, J. H. MATEIKA, and S. MATEIKA

Introduction

Snoring is a common phenomenon, affecting on the average about 20070 of the adult population, and as much as 60070 of men older than 40 yr of age (1-3). Although it has been traditionally regarded as a social nuisance, recent reports indicate that snoring may have considerably more graveconsequences than purely social ones (4-13). Furthermore, snoring is one of the cardinal symptoms of sleep apnea; even the nonapneic snorers share several common clinical and pathologic features such as daytime symptoms and abnormal upper airway mechanics with apneic snorers. Morning tiredness, fatigue, and lack of refreshing sleep is a common complaint frequently voicedby heavy snorers. It raises a possibility that sleep architecture in these patients may be abnormal, and these abnormalities may at least in part account for their daytime complaints. However,no comparative studies of sleep architecture combined with continuous and quantitative sound level measurements in snorers and nonsnorers have been performed to date. The present study examines sleep architecture and its relationship to the frequency and loudness of snoring in a group of self-confessed heavy snorers and a group of nonsnoring control subjects. Methods We studied a group of 14patients (nine men, five women). Eight of them were self-confessed heavy snorers, and six denied any history of heavy snoring. All of them had full nocturnal polysomnography, which included the monitoring of the following variables: electroencephalogram (C3/A2 and/or C4/A1 leads), right and left electrooculogram, submental and anterior tibial electromyograms, single-lead electrocardiogram, chest wall and abdominal movements using Respitraces (Ambulatory Monitoring, Inc., Ardsley, NY), oronasal flow using thermocouples, oxygen saturation using an ear oximeter (Biox III; Ohmeda Corp., Boulder, CO), sound intensity, and snoring frequency. All of the above variables were displayed continuously on a polygraph recorder (Grass Instruments, Quincy, MA) with a paper speed of 10 mm/s. 92

SUMMARY The purpose of this study was to examine whether snoring adversely affects sleep architecture and sleep efficiency, and thus may account for the frequent complaints of daytime tiredness and fatigue expressed by heavy snorers. Werecruited eight self-confessed heavy snorers and six self-confessed nonsnorers. All SUbjectshad full nocturnal polysomnography, including continuous monitoring of snoring, which was quantified by counting the number of snores per hour of sleep (snoring Index), the number of snores per minute of snoring time (snoring frequency), maximal and mean nocturnal sound intensity (dBmax and dBmean, respectively). We found that even the self-confessed nonsnorers snored lightly, with significantly smaller frequency and index than the heavy snorers. Sleep architecture was similar in both groups. Distribution of snoring among the sleep stages differed for light and heavy snorers: light snorers snored uniformly throughout all sleep stages, whereas heavy snorers tended to snore more during slow-wave and REM sleep. Snoring frequency and snoring index were similar during all sleep stages In light snorers, but they were higher during slow-wave sleep In heavy snorers. Wakefulness time after sleep onset and sleep efficiency correlated significantly with the snoring Index. Weconclude that although snoring does not affect sleep architecture In general, It Influences sleep efficiency and wakefuln£'~ time after sleep onset; this may have an adverse effect on the daytime function of heavy snorers. AM REV RESPIR DIS 1991; 143:92-96

Snoring was measured using a calibrated microphone-sound meter system in the manner described previously (5, 6), except for placement of the microphone. In our previous study (5, 6) the microphone was taped to the neck, whereas in the present study we attached it to the Respitrace chest band over the third intercostal spacein the right parasternalline; this was done in order to eliminate the artifacts caused by noise arising from the friction between the skin and the microphone. With this standard placement, quiet breathing registered lower than 45 dB; consequently, any spikes in sound intensity greater than 45 dB were counted as snores (figure 1). The technologist, who was able to hear the subjects breathing, also perceivedthis sound level as snoring rather than as regular breathing. All sleep tracings were analyzed by dividing them into 60-s epochs. Each epoch was scored with respect to sleep stage, number of apneas and hypopneas, snoring frequency, and sound intensity. Apneas were defined as episodes of cessation of airflow lasting more than 10 s and accompanied by oxygen desaturation of more than 4070 from the baseline value. Hypopneas were defined in a similar manner, except that instead of complete cessation of airflow, there wasa greater than 50% reduction in tidal volume.Snores weredefined as spikes in sound intensity greater than 45 dB returning always to the baseline level of quiet breathing. The number of apneas and hypopneas per hour of sleep was termed the apnea/hypopnea index (AHI) and the number of snores per hour of sleep was termed the snoring index (SI). For each epoch we also computed snoring frequency, defined as the

number of snores per epoch. Epochs during which there was at least one snore weretermed snoring epochs, whereas the epochs during which there was no snoring weretermed nonsnoring epochs. This definition of the snoring epochs is independent of the number of breaths associated with snoring; in fact, as may be seen in figure 1, in heavy snorers approximately 50% of the breaths wereassociated with snoring, particularly during Stage II and slow-wave sleep. All epochs were summed to obtain the total number of apneas/hypopneas and the total time spent in apnea/hypopnea. The snoring epochs were summed to obtain the total number of snores per sleep stage and the total time spent snoring. The AHI was calculated as the total number of apneas and hypopneas divided by the total sleep time. SI was calculated as the total number of S':1o: es divided by the total sleep time; it represents the average distribution of snoring throughout the night. Snoring frequency was calculated as the total number of snores divided by the number of snoring epochs; it represents the average density of snoring during the night. (Received in original form November 4, 1989 and in revised form May 29, 1990) 1 From the Department of Medicine, St. Michael's Hospital, University of Toronto, Toronto, Ontario, Canada. 2 Correspondence and requests for reprints should be addressed to Dr. V. Hoffstein, St. Michael's Hospital, 30 Bond Street, Toronto,Ontario, Canada M5B lW8.

93

SNORING AND SLEEP ARCHITECTURE

Fig, 1, A representative section of sleep tracing illustrating the relationship between snoring, normal breathing, and sleep stages.

The sleep efficiency was calculated in the usual manner as the ratio of the total sleep time to total recording time. Analysis of variance was used to compare the sleep architecture and snoring data for the snorers and nonsnorers for different sleep stages. We used a general linear model with two class variables: GROUP (two levels; snorers and nonsnorers), and STAGE (four levels: Stage I, Stage II, SWS, REM). This model permitted us to examine separately the influence of GROUP and STAGE on snoring frequency and intensity.Wealso used multiple, stepwise, linear regression analysis to examine whether snoring and apneas were significant determinants of sleepefficiency,arousals, and wakefulness time after sleep onset. All of the statistical analysis wascarried out using SAS statistical software (SAS Institute, Gary, IN) release 6.03. The levelof significance was assumed to be < 0.05.

Results

The subjects ranged from 24 to 60 yr of age, were nonobese, and had no significant sleep apnea. Although Subject 12 had an AHI of 22, his apnea index was only 7, and since he was the oldest subject in the group, it is very likely that the increased number of hypopneas in this subject was age-related. All of the subjects snored, although the "nonsnorers" snored significantly less than did the snorers (table 1). Because even the selfconfessed nonsnorers werefound to snore lightly, in all subsequent discussion we shall designate the groups as "light snorers" (rather than "nonsnorers") and "heavy snorers." There was no significant difference between the light and heavy snorers with respect to age, body mass index (BMI), AHI, or the number of

arousals. Sleep architecture (table 2) revealed no significant difference between the two groups. Differences in snoring between the two groups and the effect of sleep stages on snoring frequency and loudness is shown in figures 2 to 4, which summarize the distribution of snoring time, snoring frequency, snoring index, and mean and maximal sound intensity over the sleep stages for light and heavy snorers. We note that light snorers snored uniformly little throughout all sleep stages, whereas heavy snorers tended to snore more

in slow-wave and REM sleep (figure 2, top panel) than in the other sleep stages, although the effect of sleep stage was not statistically significant. Snoring frequency was similar in all sleep stages in light snorers, but was significantly higher in slow-wave sleep in heavy snorers (figure 2, bottom panel). Snoring index in heavy snorers also tended to be higher during slow-wave sleep, although this difference was not statistically significant (figure 3). Mean and highest nocturnal sound intensity was lowest in Stage I in light snorers, and similar across the sleep stages in heavy snorers (figure 4). A summary of the statistical analysis of the data shown in figures 2 to 4 is provided in table 3, which givesthe results of the general linear model procedure (analysis of variance); it shows the probabilities that indicate whether or not GROUP, STAGE, or the interaction between them cause significant differences between snoring variables. We note that the total time spent snoring (but not when expressed as a percentage of stage time), frequencyof snoring, and mean and highest sound intensity were significantly different depending on the patient group and the stage of sleep. On the other hand, snoring index for a given stage, which represents the average frequency of snoring over the sleep stage time, was not significantly different between the stages, but it was different for light and heavy snorers. Stepwise, forward, multiple linear re-

TABLE 1 ANTHROPOMETRIC AND RESPIRATORY SLEEP DATA

Subject No. light snorers 1 2 3 4 5 6 Mean SO Heavy snorers 7 8 9 10 11 12 13 14 Mean SO P Value

Apena + Hypopnea Index

Age

Weight

Body Mass Index

Sex

(yr)

(kg)

(kglm2)

M F M F F F

29 28 26 30 27 24 27 2

82 60 97 53 80 74 74 16

25 21 26 22 26 28 25 3

0 0 0 1 0 0 0 1

1 3 4 2 1 2 2 1

M F M M M M M M

58

77 56 82 93 87

25 21 25 30 26 25 28 27 26 3 NS

2 0 0 3 5 7 0 0 2 3 NS

3 2 2 8 13 22 3 2 7 7 NS

27 27 36 54

60

73

27 32 40 15 NS

82 75 78 11 NS

Apnea Index

Snoring Index

6 9 39 5 0 0 10 15 62 79 152 168 198 215 216 639 216 181 < 0.01

• Total number of arousals; the numbers in parentheses refer to arousals associated With respiratory events.

Arousals'

39 30 88 28 30 36 42 23

(0) (2) (13) (0) (0) (0) (3) (5)

82 (1) 14 (0) 24 (2) 33 (7) 178 (59) 65 (37) 34 (4) 23 (2) 57 (14) 54 (22) NS(NS)

94

HOFFSTEIN, MATEIKA, AND MATEIKA

TABLE 2 SLEEP ARCHITECTURE DATA Subject No. light snorers 1 2 3 4 5 6 Mean SO Heavy snorers 7 8 9 10 11 12 13 14 Mean SO P Value

TSPT

TST

SE

SE1

TW%

ST1%

ST2%

SWSOfo

REM%

313 332 406 363 358 316 348 35

241 326 280 322 305 197 279 51

70 98 61 88

23 2 34 11 15 38 20 13

13 6 8 10 12 9 3

44 60 37 49 40 29 43 11

10 13 13 18 24

50 75 18

58 92 54 78 78 40 67 19

10 19 8 12 16 3 11 6

340 339 320 285 374 275 289 330 319 34 NS

256 238 272 225 336 224 283 284 265 38 NS

73 66 72 75 89 73 98 86 79 11 NS

52 58 70 66 73 68 92 76 60 12 NS

25 30 15 18 12 19 2 14

22 9 3 10 16 6 6 10 10 6 NS

34 42 66 49 56 44 64 38 49 12 NS

10 10 6 11 6 16 14 29 13

84

17

8 NS

5

17

16 5

7

NS

10 10 10 9 13 15 14 8 11 3 NS

Definitionof abbreviations: TSPT = total sleep period time; TST = total sleep time; SE = sleep efficiency; SE 1 "" sleep efficiency excluding Stage I; TW% = total wakefulness time after sleep onset expressed as a percentage of the total sleep time; TSTl %, TST2%, SWS%, REM% = percent of sleep time spent in Stages I, II, slow-wave, and REM sleep, respectively.

gressionanalysis showedthat among heavy snorers wakefulness time after sleep onset correlated significantly with snoring index (r2 = 0.64, p < 0.02); sleep efficiency (excludingStage I) showed a signi-

ficant negative correlation with the AHI and snoring index (r 2 = 0.74, p < 0.05). Discussion

This study shows that snoring does not

Snoring Time Stage I

100 80

20

Stage II

SWS

REM

affect sleep architecture, but the distribution of snoring among the sleep stages is different for the light and the heavy snorers. Furthermore, in heavy snorers sleep efficiency and wakefulness time after sleep onset are adversely affected by apneas and snoring. Clearly, the results of this study must be considered preliminary since the number of subjects studied was relatively small. The strength of the study, however,liesin the type of measurements performed in each subject. Complete nocturnal polysomnography was carried out, and a large amount of data was accumulated for each subject. The results of apnea, snoring, and sleep architecture summarize the analysis of many epochs per subject; these results are very robust. Previous studies have pointed out that snoring is not uniformly distributed across sleep stages. Perez-Padilla and coworkers (14)found that snoring occurred only in Stage II and in slow-wave sleep. Our findings are in general agreement with this observation. However, we found a difference between the light and heavy snorers in that light snorers spend similar amounts of time snoring in all sleep stages,whereas heavy snorerssnore longer in Stage II, probably because this is the longest sleep stage. When snoring time was expressed as a percent of stage time, there was no longer a significant difference in the distribution of snoring across the sleep stages. Snoring frequency is perhaps a better parameter to describe snoring than is the total snoring time since it is less dependent on the total stage time. When expressed in this manner, heavy snorers were found to snore predominantly in slow-wave sleep. This finding is perhaps not surprising considering the recent observations of Skatrud and Dempsey (15), who showed that total pulmonary resistance is state-dependent, and snorers demonstrate an increase in resistancewith progression to the deeper stages of sleep. This increase in resistance implies narrowing of the upper airways, which when coupled with reduction in upper airway muscle activity and increased "floppiness" of the pharynx may predispose to snoring. Recent results of Bellia and coworkers (16) indicate that lower airway resistance, including subglottic airway and the trachea, is also highest during slow-wave sleep. Many snorers complain of nonrefreshing sleep. There is objective evidence suggesting that snorers have impaired cognitive function (17). A possible explanation

95

SNORING AND SLEEP ARCHITECTURE

Snort"llindex

REM

SWS

Stage II

Stege I

800

....n Sound InteMlty Stage I

sws

Stage II

REM

700

50

600

40

500

o

LGT

HVY

300

LGT

HVY

LGT

HVY

LGT

HVY

Mulrnum Sound 1nt8n8lty Stage I

sws

Stage II

REM

70

200

~.

60

...

50

.:.

..

100

40

LOT

HVY

LOT

HVY

LOT

HVY

LOT

HVY

LOT

Fig. 3. Snoring index in different sleep stages for light (LGT)and for heavy (HVY) snorers.

of this phenomenon may be intermittent nocturnal hypoxia frequently associated with snoring (18, 19). However, administration of oxygen throughout the night does not appear to improve their neuropsychiatric function (20, 21). This raises a possibility that disturbed sleep architecture, rather than hypoxia, may be responsible for the daytime symptoms

HVY

LOT

HVY

LOT

HVY

LOT

HVY

Fig. 4. Mean (top pane~ and maximal (bottom pane~ nocturnal sound intensity during different sleep stages for light (LGT) and for heavy (HVY) snorers.

of nonapneic snorers. We have not demonstrated any differences in sleep architecture between light and heavy snorers, probably because the number of subjects studied here was quite small, and, furthermore, none of them had significant daytime symptoms. However, wedid find that snoring is a significant determinant of sleep efficiency and wakefulness af-

TABLE 3 PROBABILITIES ASSOCIATED WITH THE DIFFERENCES IN SNORING VARIABLES FOR LIGHT AND FOR HEAVY SNORERS DURING DIFFERENT SLEEP STAGES Variable

Stage

Group

Stage*Group

Model

SNORTIME SNORPERC SISTAGE SNORFREQ MEANDB HIGHDB

0.0010 0.4752 0.2140 0.0238 0.0036 0.0041

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

0.0076 0.9432 0.4297 0.2849 0.0936 0.0926

0.0001 0.0008 0.0022 0.0001 0.0001 0.0001

Definitionof abbreviations:Stage .. I, II, slow-wave, REM; Group = light snorers or heavy snorers; Stage. Group = interaction term describing whether difference in snoring variables depends on a particular combination of Stage and Group; Model .. p value of the complete model; SNbRTIME = total snoring time per stage; SNORPERC .. snoring time as a percentage of stage time; SISTAGE .. snoring index per stage; SNORFREQ = snoring frequency per stage; MEANDB = mean sound intensity per stage; HIGHDB .. maximal sound intensity per stage.

ter sleep onset. To the extent that daytime symptoms may be related to poor sleep efficiency and reduction in sleep time, it is possible that snoring may be partially responsible for morning tiredness and fatigue frequently commented upon by heavy snorers. However, since we have not performed any daytime neuropsychiatric measurements, we cannot make any definitive statement as to whether heavy snorers have any daytime sequelae of snoring, reduced sleep efficiency, and increased wakefulness time. Similar to previous investigationsdealing with the health effects of snoring, our study once again points out the necessity of accurate characterization of snoring, much as it has been done for sleep apnea. Despite the association between snoring and various disorders, no standard, objective, all-night measurements of snoring were performed in most studies. In fact, we have no clear definition of snoring, and usually accept a selfreport, a subjective impression of the

96

bedmate, or the impression of the sleep technologists performing polysomnographic evaluation. The necessity of accurate definition of snoring is emphasized by the recent studies dealing with snoring and asthma (16, 22, 23), which raise a question of distinguishing between snoring and other nocturnal sounds, e.g., wheezes. Perhaps simple sound intensity monitoring is not sufficient, and more complex techniques of spectral analysis need to be employed to properly define snoring (24, 25). Once we agree on the definition of snoring, there stilI remains a question of the pathologic classification: what constitutes abnormally heavy snoring as opposed to light snoring? In viewof the fact that even the self-confessed nonsnorers also snore, although less than the selfconfessed heavy snorers, is there a cutoff in snoring frequency or snoring index above which the snoring should be considered pathologic? Perhaps snoring should be considered as a continuum, and the pathologic consequences of snoring such as hypertension and hypoxia, if such are indeed demonstrated, are related to the frequency and intensity of snoring. The entire topic dealing with the pathologic consequences of snoring, which has received a great deal of attention recently (5, 8-13~ 26) needs to be reexamined since most of the conclusions regarding the relationship between snoring and adverse health effects werebased on studies where snoring was not measured objectively; the validity of these conclusions has recently been called into question (4). Precise definition of snoring is also important since snoring may be a precursor to sleep apnea; in fact, the

HOFFSTEIN, MATEIKA, AND MATElkA

proposal of Lugaresi and coworkers (27) regarding staging of the heavy snoring disease cannot be accomplished without a standard and accurate measurement of snoring. Once this is accomplished, rigorous and quantitative studies linking snoring with hypertension, hypoxemia, or adverse neuropsychiatric function may be performed. References 1. Lugaresi E, Cirignotta F, Coccagna G. Piana C. Some epidemiological data on snoring and cardiorespiratory disturbances. Sleep 1980; 3:221-4. 2. Norton PG, Dunn E"Y. Snoring as a risk factor for disease: an epidemiological survey. Br Med J 1985; 291:630-2. 3. Bloom JW, Kaltenbom WT, Quan SF. Risk factors in a general population for snoring: importance of cigarette smoking and obesity. Chest 1988; 93:678-k3. 4. Walier PC, Bhopal RS. Is snoring a cause of vascular disease? An epidemiologic review. Lancet 1989; 1:143-6. 5. Hoffstein V, Rubinstein I, Mateika S, Slutsky AS. Determinants of blood pressure in snorers. lancet 1988; 2:992-4. 6. Hoffstein V, Chaban R, Cole P, Rubinstein I. Snoring and upper airway properties. Chest 1988; 94:87-9. 7. Erkinjuntti T, Partinen M, Sulkava R, Palomaki H, Tilvis R. Snoring and dementia. Age Ageing 1987; 16:305-10. 8. Telakivi T, Partinen M, Koskenvuo M, Kaprio J. Snoring and cardiovascular disease. Compr Ther 1987; 13:53-7. 9. Koskenvuo M, Kaprio J, Telakivi T, Partinen M, Haikkilla K, Sarna S. Snoring as a risk factor for ischemic heart disease and stroke in men. Br Med J 1987; 294:16-9. 10. Partinen M, Palomaki H. Snoring and cerebral infarction. Lancet 1985; 2:1325-6. 11. Koskenvuo M, Kaprio J, Partinen M, Langinvainio H, Sarna S, Heikkila K. Snoring as a risk factor for hypertension and angina pectoris. Lancet 1985; 1:893-6. 12. Koskenvuo M, Partinen M, Kaprio J. Snoring and disease. Ann Clin Res 1985; 17:247-51.

13. Block AJ. Is snoring a risk factor? Chest 1981; 80:525-6. 14. Perez-Padilla JR, West P, Kryger M. Snoring in normal young adults: prevalence in sleep stages and associated changes in oxygen saturation, heart rate, and breathing pattern. Sleep 1987; 1O:249~53. 15. Skatrud lB. Dempsey lA. Airway resistance and respiratory muscle function in snorers during non-REM sleep. J Appl Physiol1985; 59:328-35. 16. Bellia V, Cuttitta G, Insalaco G, Visconti A, Bonsignore G. Relationship of nocturnal bronchoconstriction to sleep stages. Am Rev Respir Dis 1989; 140:363-7. 17. Telakivi T, Kajaste S, Partinen M, Koskenvuo M, Salmi T, Kaprio J. Cognitive function in middle aged snorers and controls: role of excessivedaytime somnolence and sleep-related hypoxic events. Sleep 1988; 11:454-62. 18. Berry or, Webb WB, Block AJ, Bauer RM, Switzer OA. Nocturnal hypoxia and neuropsychological variables. 1 Clin Exp Neuropsychol 1986; 8:229-38. 19. Block AJ, Berry D, Webb W. Nocturnal hypoxemia and neuropsychological deficits in men who snore. Eur J Respir Dis 1986; 146:405-8. 20. Block AJ, Hellard OW, Switzer OA. Nocturnal oxygen therapy does not improve snorers intelligence. Chest 1989; 95:274-8. 21. Block AJ, Hellard DW, Cicale MJ. Snoring, nocturnal hypoxemia, and the effect of oxygen inhalation. Chest 1987; 92:411-7. 22. Guilleminault C, Quera-Salva MA, Powell N, et al. Nocturnal asthma: snoring, small pharynx, and nasal CPAP. Eur J Respir 1988; 1:902-7. 23. Chan CS, Woolcock AJ, Sullivan CEo Nocturnal asthma: role of snoring and obstructive sleep apnea. Am Rev Respir Dis 1988; 137:1502-4. 24. Wilson K, Mulrooney T, Gawtry RR. Snoring: an acoustic monitoring technique. laryngoscope 1985; 95:1174-7. 25. Leiberman A, Cohen A, Tal A. Digital signal processing of stridor and snoring in children. Int J Pediatr Otorhinolaryngol 1986; 12:173-85. 26. Gislason T, Aberg H, Taube A. Snoring and systemic hypertension: an epidemiological study. Acta Med Scand 1987; 222:415-21. 27. Lugaresi E, Mondini S, Zucconi M, Montagna P, Cirignotta F. Staging of heavy snorers disease: a proposal. Bull Eur Physiopathol Respir 1983; 19:590-4.