Reverse First Night Effect in Insomnia

Transcription

1 Sleep 12(2):97-105, Raven Press, Ltd., New York 1989 Association of Professional Sleep Societies Reverse First Night Effect in Insomnia Peter J. Hauri and *Elaine M. Olmstead Sleep Disorders Center, The Mayo Clinic, Rochester, Minnesota; and *Department of Medicine, Dartmouth-Hitchcock Medical Center, Hanover. New Hampshire. U.S.A. Summary: This study evaluates the reverse first night effect (FNE) in insomniacs. All insomniacs evaluated at the Dartmouth Sleep Disorders Center between 1975 and 1980 were studied if they met specific criteria (n == 89). First night effects were assessed using the following four variables: sleep efficiency, sleep latency, percentage of rapid eye movement (REM) sleep, and REM latency. The 20 subjects who showed the strongest normal FNE were compared with the 20 showing the strongest reverse FNE. The two groups were similar in age, sex, and most aspects of reported home sleep. They differed in their sleep on night 1, but on nights 2 and 3 their sleep was quite similar. Both groups overestimated their sleep latencies on night 1 (by subjective reports) but unlike the normal FNE group, the reverse FNE patients very accurately assessed their sleep latency on nights 2 and 3. Reverse FNE patients had significantly higher scores than normal FNE patients on the Minnesota Multiphasic Personality Inventory (MMPI) K (defensiveness, guardedness) and Pa (paranoia) scales, and they were less depressed, marginally less anxious, and somewhat more sensation-seeking and more susceptible to boredom. This study concluded that evaluations using only 1 night in the laboratory may be missing the larger picture of a patient's insomnia. When data from only 1 laboratory night are available for an insomniac, care should be taken with the label of "subjective complaint without objective findings" in patients who are defensive, guarded, and sensation-seeking. Key Words: First night effect-reverse first night effect-sleep disorders classification-subjective insomnia Personality traits in insomnia. During their first night in a sleep laboratory, most persons sleep somewhat worse than they do during later laboratory nights. This is called the "first night effect" (FNE) (1-5). Commonly noted differences include a longer sleep latency, lower sleep efficiency, a longer REM latency, and a lower percentage of stage REM sleep. Although the effect may extend beyond night 1 (4,6), studies generally conclude that the most significant effects occur on the first laboratory night. Some investigators argue that the FNE is not universal. They suggest that a pleasant environment and a warm and friendly staff may all but erase the FNE (7,8). Prior Accepted for publication August Address correspondence and reprint requests to Peter J. Hauri, Ph.D., Sleep Disorders Center, The Mayo Clinic, Rochester, MN 55905, U.S.A. 97

2 98 P. 1. HAURI AND E. M. OLMSTEAD familiarization with the lab environment may also attenuate the effect (8) although there is some disagreement about this (9). Contrary to expectations, some insomniacs sleep better, rather than worse, on their first night in the laboratory (10,11). This is called a reverse first night effect. Two explanations of this reverse FNE have been proposed, one by de la Pefia (11) and one by Hauri (10). De la Pefia proposes that for each person a specific level of waking sensory stimulation is optimal for induction of good sleep. For most people, the excessive sensory stimulation caused by sleeping in the laboratory for the first time may create a state of heightened arousal interfering with sleep. This would cause the normal FNE. However, for some insomniacs "information underload" may cause poor sleep. Such insomniacs may benefit from the additional sensory stimulation provided by sleeping in the laboratory and may therefore sleep better on the first laboratory night, i.e., show a reverse FNE. De la Pena indicates that members of this second group may rate themselves as being more bored and as being more sensation-seeking than average. They may also rate their first night in the laboratory as being better than their home sleep. Hauri's hypothesis involves learning principles. It proposes that a learned, maladaptive association has been established between some insomniacs' attempts to sleep and arousal. As reported first by Bootzin (12), some patients, unable to sleep well for a time because of other stresses, may have developed a conditioned arousal reaction to their own bedroom environment. This conditioned arousal hampers sleep at home but does not generalize to the novel surroundings of the sleep laboratory. Alternatively, some insomniacs may not try as desperately to fall asleep in the laboratory as they do at home, either because they do not expect to sleep at all in the laboratory and therefore have given up or because they see the laboratory as their chance to document their insomnia for all those who previously doubted it. Not trying so desperately to sleep, they may fall asleep easier in the laboratory than at home. According to the nosology developed by the Association of Sleep Disorders Centers (ASDC) (13), persistent psychophysiologic insomnia (PPI) is presumably the type of insomnia in which the learned maladaptive habits described above are most evident. However, in a recent study (14) our laboratory did not find a consistent reverse FNE in that type of insomnia. Thus, the question remains: what types of insomniacs do show a reverse FNE? The question appears to be important in our consistent efforts to illuminate the factors that drive and maintain chronic insomnia. It is the focus of the research we present here. METHODS To examine best the differences between insomniacs with reverse FNE and those with normal FNE, two groups were established showing either a strong reverse FNE or a strong normal FNE based on four sleep variables (sleep efficiency, sleep latency, the percentage of stage REM (rapid eye movement), and REM latency). Subject Selection All insomniacs evaluated at the Dartmouth Sleep Disorders Center between 1975 and 1980 were included in this study if they met the following criteria (for more information concerning this patient population, see ref. 10): 1. They had been referred by their own physicians for an evaluation of chronic and

3 REVERSE FNE IN INSOMNIA 99 serious insomnia. Therefore, few patients in this study had insomnia that was obviously secondary to psychopathology (e.g., major affective disorders, psychosis) or obviously secondary to medical problems (e.g., pain), because physicians referred those patients directly to the appropriate specialists, not to a sleep disorders center. 2. They had slept in the laboratory for 2 or 3 consecutive nights each (63% had 3 laboratory nights, the remaining 37% had 2 laboratory nights each). Besides the Rechtschaffen and Kales (15) channels, each insomniac had the following monitored for at least 1 night (almost always night 2 or 3): air flow (thermistor), bilateral anterior tibialis muscles, and electrocardiogram (ECG). Patients were excluded from this study if they showed abnormal recordings in any of these latter 4 channels, e.g., sleep apnea or periodic movement of the legs during sleep. 3. They had undergone an intensive (1-3 h) interview with one ofthe authors (P. H.) and had completed a battery of questionnaires including the Minnesota Multiphasic Personality Inventory (MMPI), Zung's Depression Inventory, Zuckerman and Lubin's Multiple Affect Adjective Check List (MAACL), selected subscales of the Zuckerman Sensation Seeking Scale, and the Cornell Medical Index (CMI). Procedures Each patient reported to the laboratory -1 h before bedtime. Each slept in a windowless, sound-attenuated bedroom. Recording and scoring were done according to the Rechtschaffen and Kales criteria (15). Based on the total available evidence (sleep in the laboratory, medical charts, psychological interviews, tests, etc.), each insomniac was then assigned an ASDC (13) classification (Table O. Assessment of FNE The first night in the laboratory primarily affects sleep efficiency, sleep latency, the percentage of stage REM, and REM latency. Each subject's FNE with respect to each of these four variables was calculated by subtracting the patient's score for the first night in the sleep laboratory from the patient's average score on the second and third nights. Thus, in this study, both the normal FNE and the reverse FNE are defined as Category Persistent psychophysiologic insomnia Associated with personality disorders Associated with affective disorders a Childhood onset insomnia Insomnia complaint without objective findings Other TABLE 1. ASDC diagnostic categories Overall pool % (n = 89) Selected for Normal FNE % (n = 20) Selected for Reverse FNE % (n = 20) The ASDC diagnoses are based on the sum total of all available information: three nights in the laboratory, psychological interviews, and tests. However, Table I cannot be used to compute prevalence data for the different categories because of the highly selective intake process described in the Methods section. a Mainly patients with dysthymia. Patients with other mood disorders were typically excluded by our referral process.

4 100 P. J. HAURI AND E. M. OLMSTEAD comparisons of laboratory night 1 versus laboratory nights 2 and 3. They are not defined as comparisons between sleep at home and laboratory sleep. Of the 89 insomniacs included in this study, only 16 showed a normal FNE on all four variables, and only 2 showed a reverse FNE on all four variables. More typical were mixed reactions. To deal with this problem of mixed FNE, we assigned each insomniac an overall FNE score as follows. First, we changed the plus or minus sign on two ofthe four FNE variables (sleep efficiency and percentage of stage REM) so that they all showed a uniform expected direction (minus equalling a normal FNE). The FNEs on each of the four variables were then normalized across all 89 subjects (Z scores; mean = 0, SD = 1). These four normalized FNE scores for each insomniac were added to obtain each insomniac's overall FNE score. All 89 insomniacs were then rank ordered according to the magnitude of their overall FNE. Two groups were then selected for study. One consisted of the 20 insomniacs with the strongest normal FNE. The other group contained the 20 insomniacs with the strongest reverse FNE. These two groups were then statistically compared. Nominal data (male versus female) was compared using the chi-square test. The Mann-Whitney test was used to test the difference between medians, and t tests were used to test the difference between means. Alpha was set at To explore further the data matrix, we considered alpha levels between 0.05 and 0.10 as "marginally significant," generating new hypotheses rather than confirming older ones. RESULTS Table 2 indicates that the insomniacs with a normal FNE were quite similar to those with a reverse FNE in age, sex, and reports of home sleep, except that the insomniacs who later showed a normal FNE in the laboratory had reported that they slept ~ 1 Y2 h less at home than those who later showed a reverse FNE in the lab. Table 2 atso shows that the p values for the four FNE variables used to select the two groups are statis- TABLE 2. Description of sample Normal variable Reverse FNE Normal FNE p Value n Age <0.22 Sex (M/F) 6/ <0.11 Typical sleep at home (mean of 2-wk sleep log) Total time in bed (min.s) <0.63 Sleep latency (min.s) <0.35 Total time asleep (min.s) <0.02 Awake after onset (min.s) <0.99 First night effect in laboratory [(average of nights 2 + 3) - night 1] Selection criteria a Sleep efficiency (%) <0.001 Sleep latency <0.001 REM (%) <0.001 REM latency <0.001 FNE, first night effect. a A normal FNE results in a positive sleep efficiency, a negative sleep latency, a positive REM %, and a negative REM latency. Values were obtained before the inversion of polarity discussed in the Methods section was done; p was assessed by two-tailed t test for all variables except sex (chi-square).

5 REVERSE FNE IN INSOMNIA 101 tically significant. This indicates simply that we were successful in creating two groups that were clearly different on the four FNE variables. Table 3 reveals first that the two groups slept quite differently on their first night in the laboratory. This was true for sleep efficiency, sleep latency, percentage of stage REM sleep, REM latency, and also for total time asleep, a measure that we had not included in our selection criteria. On nights 2 and 3 (data averaged), however, the normal FNE and reverse FNE groups slept nearly the same. Only percentage of stage REM showed a significant difference: insomniacs with a reverse FNE showed less REM sleep on nights 2 and 3 than did their counterparts with a normal FNE. Subjective reports of sleep latency in the laboratory show that both groups overestimated their sleep latencies on night 1 (as compared with polygraphic data in the same table). On nights 2 and 3, however, although the normal FNE group still overestimated sleep latency considerably, the subjective estimates of the reverse FNE group were within 1 min of the polygraphic findings. When patients compared the subjective quality of their lab sleep to the quality of their home sleep (Table 3), those with a normal FNE reported their first night clearly as "worse than average" (4.7 on a 14-point scale) as compared with their home sleep, and nights 2 and 3 as "average." This change in ratings from "worse than average" to "average" accurately reflects the improvement in lab sleep from night 1 to nights 2 and 3 for those with a normal FNE. The patients with a reverse FNE, on the other hand, rated all 3 nights as about "average" (7.4 and 7.0 on a 14-point scale). They were little aware that their sleep on night 1 had been better than their sleep on nights 2 and 3. Table 4 characterizes questionnaire responses by the two groups. Overall, both groups do show the typical insomniac MMPI patterns previously described by Kales' group and by others (16-18): elevated MMPI scales 1,2,3,4, and 7 and a high Welch TABLE 3. Sleep in laboratory Variable Normal FNE Reverse FNE p Value Assessed by polygraph Sleep on first laboratory night Sleep efficiency <0.001 Sleep latency (first spindle) <0.002 REM(%) <0.001 REM latency <0.01 Total time asleep <0.001 Mean of nights 2 and 3 Sleep efficiency <0.92 Sleep latency <0.70 REM (%) <0.03 REM latency <0.64 Total time asleep <0.11 Subjectively reported (morning questionnaire) Sleep on first night Sleep latency (min.s) <0.08 Comparison to home sleepq <0.004 Sleep on nights 2 and 3 Sleep latency (min.s) <0.18 Comparison to home sleepq <0.86 p was assessed by two-tailed t tests or Mann-Whitney test (comparison to home sleep). Q Compared with my home sleep, my lab sleep was: 2 = much worse than my average home sleep; 8 = average for me; 14 = the best in weeks.

6 102 P. J. HAURI AND E. M. OLMSTEAD TABLE 4. Psycholo/?ical and physical health questionnaires by group Scales Normal FNE Reverse FNE p Value MMPI F <0.72 K <0.02 Scale 1: Hs <0.32 Scale 2: D <0.54 Scale 3: Hy <0.47 Scale 4: Pd <0.63 Scale 6: Pa <0.05 Scale 7: Pt <0.70 Scale 8: Sc <0.34 Scale 9: Ma <0.98 Scale 10: Si <0.34 Welsh A <0.20 Welsh R <0.24 MAACL Anxiety <0.07 Depression <0.14 Hostility <0.21 Sensation seeking General <0.09 Boredom <0.06 Zung Depression Index <0.04 Cornell Medical Index: Total "yes" <0.16 p assessed by two-tailed t tests. R. However, within this overall pattern typical for insomnia, insomniacs with a normal FNE do differ from those with a reverse FNE on the following dimensions. Reverse FNE patients showed a higher K (validity) scale on the MMPI, as well as a higher Pa scale. They are less depressed according to the Zung Depression Inventory. They are also marginally (p = 0.07) less anxious on the MAACL, marginally more sensation-seeking (p = 0.09), and more susceptible to boredom (p = 0.06). However, these latter three differences did not reach our criteria of alpha set at They will have to be replicated before they can be accepted as fact. For clinical purposes, each insomniac received an ASDC diagnosis after all data for that case had been collected, but before any computations concerning the FNE had been done. Table 1 shows these ASDC diagnoses for the two FNE groups: A reverse FNE obviously is not exclusively associated with anyone AS DC diagnosis. Persistent psychophysiologic insomnia and childhood onset insomnia may be somewhat overrepresented in the reverse FNE group, whereas the psychiatric diagnoses may be somewhat overrepresented in the group with a normal FNE. However, there are too few data to evaluate this trend statistically. DISCUSSION The reaction to sleeping in the laboratory for the first time is clearly more complicated for insomniacs than it is for normal subjects. Using the relatively stringent criterion of three of the four FNE variables clearly pointing in one direction, only 37% of our insomniacs showed a normal FNE. More than half of all insomniacs show either no FNE or an ambiguous one, whereas 9% show a clear reverse FNE. This is clearly different from normal sleepers, most of whom show a normal FNE. The first night in Sleep. Vol. 12. No

7 REVERSE FNE IN INSOMNIA 103 the laboratory may be somewhat stressful for most normal sleepers whereas some insomniacs either find it no more stressful than home sleep or even easier than home sleep. Our findings are relevant because insomniacs' first night in the sleep laboratory is often their only one. Sleep disorders centers often can study insomniacs only for 1 night, because of the high costs of performing clinical somnograms. Our study suggests that using a I-night procedure to evaluate insomnia may result in -9% of all insomniacs being misclassified as having "subjective insomnia without objective findings" when later nights would reveal their "true" insomnia. This adds further weight to Trinder's (19) recent review (1988) in which he counseled extreme caution when classifying someone as a "subjective insomniac." Nevertheless, 17% of our patients had subjective insomnia (Table 1), whereas Trinder (19) doubts that such a category actually exists. A single night in the sleep laboratory is also often used to screen insomniacs for research studies. Our study suggests that such a procedure prevents evaluation of an entire class of insomniacs by research protocols because insomniacs who do not show certain sleep criteria on their first night in the laboratory are usually eliminated from such studies. The most important finding of our study may well be the conclusion that insomniacs who show a reverse FNE often appear to be more suspicious, guarded, and defensive than those who have a normal FNE. This statement is based primarily on elevated scores on K and Pa in the MMPI. The K scale was originally developed to detect guardedness and defensiveness in test taking. Patients who score high on this scale are said to deny psychological problems. The Pa scale was originally developed to measure suspiciousness and paranoia. Although in practice the Pa scale does not measure overt classical paranoia well, high scorers on this scale consistently are more suspicious, more defensive, more guarded, and more prone to deny emotional problems. Other data collected in this study support the above idea that increased defensiveness and a denial of emotional problems is typical for insomniacs who show a reverse FNE. Patients who show a reverse FNE tend to under-report their sleep problems. In our study, such patients claimed that at home they usually slept -378 min/night (Table 2), whereas they slept only 348 min on nights 2 and 3 in the laboratory (Table 3). This is contrasted with the typical exaggeration of sleep difficulties in insomniacs who show a normal FNE: They claimed that they slept an average of 294 min/night at home (Table 2) whereas in the laboratory they were found to sleep an average of 377 min on nights 2 and 3 (Table 3). Similarly, patients with a reverse FNE appear to report their sleep latencies accurately, at least on nights 2 and 3, whereas those with a normal FNE vastly exaggerate their sleep difficulties. Although not statistically significant, the Welsh Anxiety and Repression scores on the MMPI and the total number of "yes" answers on the Cornell Medical Index also tend to describe patients with a reverse FNE more as repressors and those with a normal FNE more as sensitizors. Superficially, these results seem surprising. Common sense would make one expect patients who are repressors and who are overly suspicious and guarded to sleep even more poorly in a strange environment since they remain tense, repress their emotions, and may worry suspiciously about all the things that might happen to them. However, both de la Pefia's (11) and Hauri's theories can incorporate these findings well. De la Pefia showed that information underload often leads to paranoid thinking, as patients try to increase their levels of stimulation by interpreting random events as meaningful

8 104 P. J. HAURI AND E. M. OLMSTEAD (11); de la Pefia also predicted more sensation seeking in patients with a reverse FNE. In our study, the scores in sensation-seeking scales (Table 4) barely missed significance, confirming de la Pefia's ideas at least marginally. According to Rauri's theory (20), a reverse FNE simply indicates a lack of generalization of learned, maladaptive habits from the home environment to the laboratory situation. The insomniac, typically sleep deprived, will then sleep better than average on the first night when the maladaptive sleep habits do not playa role. The two most obvious of these maladaptive sleep habits are (a) trying too hard to sleep (20), and (b) associating the bedroom with anxiety and frustration (12). These habits may be learned during a variety of sleep disturbing situations such as environmental noise, depression, medical illness, or stress but, according to Rauri's theory, they become chronic only in insomniacs whose sleep is inherently somewhat weaker than normal. One might expect such insomniacs to be guarded and defensive concerning their emotional health because, until quite recently, all insomnia was commonly assumed to have a psychiatric/ psychological etiology. Thus, whenever these patients complained about their poor sleep, they were accused (wrongly) of being excessively anxious, tense, or depressed. Table 1 supports this assessment, indicating that the psychiatric diagnoses may be more prevalent in insomniacs with a normal FNE and rarer in insomniacs with a reverse FNE. Other findings in this study tend to support Hauri's contention that patients with a reverse FNE are relatively more normal, both psychologically and medically, than insomniacs with a normal FNE. Those with a reverse FNE seemed less disturbed in their mood (MAACL Anxiety, Depression, and Hostility, and Zung Depression). They also had fewer medical complaints (Cornell Medical Index). They did not exaggerate their sleep problems as almost all other insomniacs typically seem to do. (However, many of these differences did not reach statistical significance in our sample.) On the other hand, Hauri's theory (20) does not predict that patients with a reverse FNE will be more sensation seeking, as was found, and neither theory predicted that patients with a reverse FNE typically have less REM sleep than patients with a normal FNE. The present study suggests that an insomniac's reaction to sleeping in the laboratory for the first time contains important diagnostic information. Thus, if understanding the patient is the goal, first nights in the laboratory should not routinely be discarded as "adaptation nights." Furthermore, descriptions of insomnia given by patients with a guarded, defensive lifestyle apparently can be taken much more literally than the exaggerated sleep descriptions of insomniacs who are more depressed, anxious, and accepting of emotional problems. Finally, if only 1 night of sleep is available for diagnostic purposes, care should be taken with the label of subjective complaint without objective findings in patients who appear to be somewhat defensive, guarded, and sensation seeking. Acknowledgment: Many thanks to Gerald T. O'Connor, D.Sc., for his insightful and constructive comments on the manuscript. REFERENCES 1. Rechtschaffen A, Verdone P. Amount of dreaming: effect of incentive, adaptation to laboratory, and individual differences. Percep Mot Skills 1964;19: Agnew HW Jr, Webb WB, Williams RL. The first night effect: an EEG study of sleep. Psychophysiology 1966;2:

9 REVERSE FNE IN INSOMNIA Mendels J, Hawkins DR. Sleep laboratory adaptation in normal subjects and depressed patients ("first night effect"). Electroencephalogr Clin Neurophysiol 1967;22: Schmidt HS, Kaelbling R. The differential laboratory adaptation of sleep parameters. Bioi Psychiatry 1971 ;3: Webb WB, Campbell SS. The first night effect revisited with age as a variable. Waking Sleeping 1979;3: Hartmann E. Adaptation to the sleep laboratory and placebo effect. Psychophysiology 1968;4: Coble P, McPartland RJ, Silva WJ, Kupfer DJ. Is there a first night effect? (a revisit). Bioi Psychiatry 1974;9: Browman CP, Cartwright RD. The first-night effect on sleep and dreams. Bioi Psychiatry 1980;15: Toner BB, Cairns J, Knowles JB, MacLean AW. Can the 'first night effect' be reduced? In: Chase MH, Stem WC, Walter PL, eds. Sleep Research, vol. 4. Los Angeles: Brain Information Service/Brain Research Institute, 1975: Hauri PJ. A cluster analysis of insomnia. Sleep 1983;6: de la Peiia A. Toward a psychophysiologic conceptualization of insomnia. In: Williams RL, Karacan I, eds. Sleep disorders-diagnosis and treatment. New York: John Wiley & Sons, 1978: Bootzin R. A stimulus control treatment for insomnia. Proc Am Psychol Assoc 1972;7: Association of Sleep Disorders Centers. Diagnostic classification of sleep and arousal disorders, first edition, prepared by the Sleep Disorders Classification Committee, HP Roffwarg, chairman. Sleep 1979;2: Hauri P, Fisher J. Persistent psychophysiologic (learned) insomnia. Sleep 1986;9: Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system sleep stages of human subjects. Los Angeles: Brain Information Service/Brain Research Institute, Kales A, Caldwell AB, Preston T, Healey S, Kales JD. Personality patterns in insomnia-theoretical implications. Arch Gen Psychiatry 1976;33: Zorick F, Kribbs N, Roehrs T, Roth T. Polysomnographic and MMPI characteristics of patients with insomnia. In: Hindmarch I, Ott H, Roth T, eds. Psychopharmacology supplementum I. Berlin: Springer Verlag, 1984: Levin D, Bertelson AD, Lacks P. MMPI differences among mild and severe insomniacs and good sleepers. J Personality Assessment 1984;48: Trinder J. Subjective insomnia without objective findings: a pseudo diagnostic classification? Psychol Bull 1988;103: Hauri P. The sleep disorders. Kalamazoo: The Upjohn Company, 1982:25-8.