قدرت دلتا و اختلاف افزایش در اندازه گیری های خواب عینی و ذهنی در اختلال شخصیت مرزی

Abstract Background Previous studies have shown depression-like sleep abnormalities in borderline personality disorder (BPD). However, findings in BPD are not unequivocal for REM dysregulation, as well as for a decrement of slow wave sleep and sleep continuity disturbances. Earlier findings in sleep EEG abnormalities in BPD may have been confounded by concomitant depressive symptoms. Methods Twenty unmedicated female BPD patients without current comorbid major depression and 20 sex- and age-matched control subjects entered the study. Conventional polysomnographic parameters and for the first time sleep EEG spectral power analysis was performed on two sleep laboratory nights. Subjective sleep parameters were collected by sleep questionnaires in order to assess the relationship between objective and subjective sleep measurements. Results BPD patients showed a tendency for shortened REM latency and significantly decreased NonREM sleep (stage 2). Spectral EEG analysis showed increased delta power in total NREM sleep as well as in REM sleep in BPD patients. Subjective ratings documented drastically impaired sleep quality in BPD patients for the two weeks before the study and during the two laboratory nights. Conclusion Not-depressed BPD patients only showed tendencies for depression-like REM sleep abnormalities. Surprisingly, BPD patients displayed higher levels of delta power in the sleep EEG in NREM sleep than healthy control subjects. There was a marked discrepancy between objective and subjective sleep measurements, which indicates an altered perception of sleep in BPD. The underlying psychological and neurobiological mechanisms of these alterations are still unclear and need to be clarified in future studies including interventions on a pharmacological and cognitive-behavioral level.

. Introduction Borderline personality disorder (BPD) is a frequently diagnosed personality disorder with a prevalence of 1–2% in the general population (Gunderson and Zanarini, 1987) encompassing a variety of symptoms such as aversive inner tension, self-injurious behavior and affective dysregulation. Furthermore, disturbances of sleep continuity and nightmares are frequently encountered in BPD (Asaad et al., 2002). Up to now 10 sleep-EEG studies have been published in BPD. Some of the earlier studies did not explicitly control for concomitant major depression in BPD. Therefore, in these studies findings of depression-like sleep abnormalities such as shortened REM latency (Bell et al., 1983, McNamara et al., 1984, Reynolds et al., 1985 and Lahmeyer et al., 1988) as well as sleep continuity disturbances (Bell et al., 1983) and increased REM density (Bell et al., 1983 and McNamara et al., 1984) might be interpreted as a consequence of the concomitant depression. Six studies explicitly measured and reported concomitant depression and four of these studies reported reduced REM latency (Akiskal et al., 1985, Asaad et al., 2002, Battaglia et al., 1993 and Benson et al., 1990) suggesting a common biological origin for affective disorders and BPD. In one study (Battaglia et al., 1999), young never-depressed BPD patients had higher REM density during the first REM period extending the view that REM density in the first REM period could be a marker of liablitiy to affective disorders (Modell et al., 2002). In one recent study REM latency did not differentiate BPD and MDD and normal controls (De la Fuente et al., 2001). To our knowledge, there is no study which investigated the microstructure of sleep in BPD. Recently, higher theta and delta power in stage 4 have been found in male antisocial patients with BPD comorbidity compared to controls (Lindberg et al., 2003a). This is in contrast to almost all other studies of spectral sleep EEG power in psychiatric populations (with the exception of patients with panic disorders), where mainly reductions of delta power have been described (especially in depressed patients; for overview see Riemann et al., 2001). One sleep study reported more concomitant sleep complaints in BPD than in healty controls (Asaad et al., 2002). However, earlier polysomnographic studies in BPD did not provide data on questionnaires validated to indicate the presence of sleep disturbances (e.g., Pittsburgh Sleep Quality Index, PSQI, (Buysse et al., 1989)). In the present study, we aimed to assess whether the subjective ratings of sleep disturbances in patients with BPD are associated with objective polysomnographic findings and wether alterations of the sleep EEG spectral power can be observed in nondepressed female BPD patients when compared with healthy controls.

Results There was no significant difference between age in the patient group and healthy volunteers (BPD 28.6 ± 7.9 years vs. controls 28.6 ± 7.1 years). Although BPD patients did not fulfill criteria for MDD according to SCID, the BDI values in BPD patients were significantly higher than in controls (BPD: 24.1 ± 11.8 vs. controls: 3.0 ± 4.3; F(1; 27) = 29.2, p = 0.000). 3.1. Objective sleep parameters Data for sleep variables from conventional PSG including results of statistical testing are shown in Table 2. Table 2. Sleep parameters measured using conventional polysomnography in 20 female patients with BPD and 20 female control subjects Adaptation night Baseline night Diagnosis Night Night ∗ Diagnosis Controls Borderline Controls Borderline F p F p F p Multivariate test (Wilk’s lambda) 1.2 0.321 3.2 0.006 1.5 0.184 Sleep onset latency (min) 45.8 ± 53.6 32.8 ± 34.9 25.7 ± 27.1 28.2 ± 30.9 0.3 0.593 7.3 0.010 2.0 0.163 Sleep period time (min) 450.8 ± 17.9 437.9 ± 51.5 463.5 ± 12.7 450.1 ± 33.1 1.9 0.171 6.7 0.014 0.0 0.962 Total sleep time (min) 413.8 ± 53.2 396.3 ± 76.0 444.9 ± 29.6 420.3 ± 50.1 1.5 0.221 15.1 0.000 0.1 0.773 Sleep efficiency (%) 86.4 ± 11.2 83.3 ± 16.1 92.6 ± 6.1 87.8 ± 10.1 1.2 0.272 12.5 0.001 0.1 0.737 Arousal frequency 21.0 ± 18.3 21.9 ± 11.5 20.5 ± 12.5 24.4 ± 11.6 0.4 0.519 0.2 0.684 0.4 0.554 Arousal index (per h TST) 14.9 ± 7.9 15.8 ± 5.8 11.7 ± 4.8 13.3 ± 6.1 0.4 0.532 9.1 0.005 0.1 0.821 Wake Time (% SPT) 8.4 ± 10.1 9.9 ± 11.2 4.1 ± 5.0 6.6 ± 8.4 0.5 0.464 7.8 0.008 0.0 0.850 Stage 1 (% SPT) 6.8 ± 3.4 7.4 ± 4.6 5.3 ± 2.4 6.5 ± 4.2 0.5 0.506 9.0 0.005 0.0 0.864 Stage 2 (% SPT) 52.8 ± 6.2 46.3 ± 7.7 55.2 ± 6.8 47.1 ± 9.3 10.9 0.002 2.5 0.125 0.3 0.581 SWS (% SPT) 10.0 ± 7.3 14.0 ± 9.4 11.8 ± 7.2 14.0 ± 9.8 1.6 0.212 3.3 0.077 1.4 0.248 REM sleep (% SPT) 21.6 ± 5.8 21.8 ± 6.8 23.0 ± 4.6 25.5 ± 4.4 0.9 0.359 6.8 0.013 1.6 0.214 REM latency (min) 93.4 ± 64.8 67.5 ± 16.5 73.6 ± 44.1 57.2 ± 10.1 3.7 0.062 4.2 0.047 0.5 0.504 REM density (Total %) 22.7 ± 6.3 23.6 ± 7.8 25.0 ± 7.5 24.2 ± 7.6 0.0 0.881 2.0 0.164 2.7 0.112 Abbreviations. TST, total sleep time; SWS, slow wave sleep; SPT, sleep period time. Table options The multivariate repeated measures factor NIGHT was highly significant: In the second night, BPD patients and controls showed significantly enhanced sleep efficiency, decreased sleep latency and a reduced number of arousals during sleep. Stages Wake and 1 (% SPT) were significantly reduced whereas SWS and REM (% SPT) were increased in night 2 compared to the first night in the laboratory. REM latency was significantly decreased in the second night. The multivariate factor GROUP was not significant for polysomnographic parameters. With univariate testing, BPD patients had a lower proportion of stage 2 sleep (% SPT) compared with controls (F(1; 37) = 11.05, p = 0.002 (see Fig. 1). There was a tendency for shortened REM latency in BPD patients in both nights (F(1; 37) = 3.71, p = 0.062). None of the other sleep variables, including REM density, differed between the two groups, none of the interaction effects (Night x Group) was significant. Stage 2 (%) SPT in 20 females with BPD and 20 controls measured using PSG. Fig. 1. Stage 2 (%) SPT in 20 females with BPD and 20 controls measured using PSG. Figure options In an additional analysis only encompassing the second laboratory night and comparing control subjects and BPD patients for conventional sleep parameters, these results were confirmed (data not shown). Also for the second night alone, stage 2 sleep% was significantly decreased and the analysis revealed a diminished tendency for a decreased REM latency in BPD patients. Data for the spectral sleep EEG parameters of total NonREM sleep and REM sleep are depicted in Table 3. Table 3. Sleep parameters measured using power spectral analysis in total NonREM sleep in 20 female patients with BPD and 20 female healthy control subjects (spectral power unit log μV2 Adaptation night Baseline night Diagnosis Night Night ∗ Diagnosis Controls Borderline Controls Borderline F p F p F p Multivariate test (Wilk’s lambda) 2.41 0.029 1.34 0.259 1.39 0.233 Delta 5.36 ± 0.65 5.78 ± 0.51 5.45 ± 0.43 5.69 ± 0.42 5.08 0.030 0.00 0.956 1.24 0.273 Theta 3.57 ± 0.63 3.80 ± 0.40 3.65 ± 0.37 3.80 ± 0.39 2.18 0.148 1.27 0.267 0.20 0.655 Alpha 2.50 ± 0.65 2.72 ± 0.54 2.60 ± 0.49 2.75 ± 0.60 1.40 0.244 1.80 0.188 0.01 0.932 Sigma 1.90 ± 0.66 1.90 ± 0.51 2.00 ± 0.36 1.87 ± 0.52 0.11 0.747 0.34 0.562 0.82 0.372 Beta1 0.57 ± 0.55 0.72 ± 0.43 0.65 ± 0.27 0.62 ± 0.40 0.24 0.626 0.07 0.801 1.83 0.184 Beta2 −0.49 ± 0.45 −0.17 ± 0.77 −0.38 ± 0.29 −0.38 ± 0.67 0.98 0.329 0.43 0.514 3.33 0.076 Gamma −1.17 ± 0.51 −0.72 ± 1.27 −0.97 ± 0.50 −1.09 ± 1.06 0.41 0.528 0.32 0.574 4.77 0.035 RDelta 3.80 ± 0.58 4.15 ± 0.61 3.91 ± 0.28 4.05 ± 0.47 2.94 0.095 0.00 0.980 1.49 0.229 Rtheta 2.92 ± 0.61 3.25 ± 0.46 3.04 ± 0.32 3.22 ± 0.44 3.68 0.063 0.84 0.366 1.25 0.272 Ralpha 1.92 ± 0.63 2.18 ± 0.49 2.04 ± 0.46 2.13 ± 0.50 1.32 0.259 0.97 0.331 1.51 0.227 Rsigma 0.85 ± 0.73 0.91 ± 0.45 1.00 ± 0.50 0.86 ± 0.40 0.03 0.864 0.69 0.411 2.07 0.159 Rbeta1 0.75 ± 0.72 0.95 ± 0.42 0.87 ± 0.42 0.87 ± 0.45 0.44 0.514 0.22 0.645 2.33 0.136 Rbeta2 −0.33 ± 0.69 0.12 ± 0.64 −0.16 ± 0.36 −0.06 ± 0.64 2.66 0.112 0.00 0.992 4.25 0.047 Rgamma −1.24 ± 0.67 −0.76 ± 1.10 −0.92 ± 0.43 −1.10 ± 0.89 0.47 0.499 0.01 0.934 6.62 0.014 Abbreviations. Delta, delta power in total NonREM sleep; RDelta, delta power in REM sleep; corresponding to the other spectra. Table options No significant NIGHT effects were observed. However, for spectral power measures, the multivariate factor GROUP was significant. BPD patients had significantly higher delta power in total NonREM sleep (see Fig. 2) and across all NREM sleep cycles (see Fig. 3) as well as tendencies for higher delta and theta power in REM sleep. NREM sleep EEG delta band power (derivation C3-A2, logarithmic to base e) in 20 ... Fig. 2. NREM sleep EEG delta band power (derivation C3-A2, logarithmic to base e) in 20 females with BPD and 20 controls. Delta power is increased in BPD patients. Figure options NREM sleep EEG delta band power across all NREM sleep cycles. Delta power is ... Fig. 3. NREM sleep EEG delta band power across all NREM sleep cycles. Delta power is increased in all NREM sleep cycles in BPD patients. Figure options If looking only at the second night, our finding of increased delta power in BPD was still present but slightly less significant (p = 0.076). When calculating delta ratio, i.e., the quotient of delta power in the first to the second NREM cycle (see also Fig. 3), an explorative analysis did not reveal any significant difference for delta ratio between BPD patients and healthy controls. For interaction effects (NIGHT ∗ Group) the multivariate test was not significant for any of the investigated variables. 3.2. Subjective sleep parameters For the PSQI the multivariate factor GROUP was highly significant (F(14; 24) = 4.79, p = 0.000). BPD patients rated their total sleep quality during the two weeks prior to investigation significantly worse than control subjects (PSQI sum score; Controls: 3.8 ± 1.7 vs. BPD:, 11.2 ± 4.4; F(1; 37) = 49.91, p = 0.000). For this period sleep efficiency (%) and total sleep time (min) were also estimated significantly worse by BPD patients (Sleep efficiency: BPD: 71.5 ± 18.4% vs. Controls: 91.5 ± 5.9%; F(1; 37) = 19.76, p = 0.000; Total Sleep Time: BPD: 360.0 ± 99.3 min vs. Controls: 460 ± 70 5 min; F(1; 37) = 17.92, p = 0.000, respectively). For the SF-A (data see Table 4), the multivariate factor GROUP was highly significant whereas NIGHT and NIGHT ∗ GROUP were only tendentially significant. Table 4. Subjective sleep parameters in 20 female patients with BPD and 20 female healthy control subjects Adaptation night Baseline night Diagnosis Night Night ∗ Diagnosis Controls Borderline Controls Borderline F p F p F p Multivariate test (Wilk’s lambda) 10.1 0.000 2.2 0.051 2.0 0.076 SF-SQ 3.1 ± 0.9 3.0 ± 0.7 3.9 ± 0.6 3.2 ± 0.73 5.2 0.029 8.8 0.006 3.4 0.075 SF-GES 3.6 ± 0.9 2.6 ± 0.9 3.6 ± 0.8 2.6 ± 0.98 17.7 0.000 0.0 0.967 0.1 0.713 SF-PSYAA 4.1 ± 0.6 2.1 ± 0.9 4.0 ± 0.7 2.5 ± 0.71 96.3 0.000 2.2 0.146 5.0 0.032 SF-PSYEA 2.6 ± 0.9 3.7 ± 0.8 2.4 ± 1.0 3.4 ± 0.70 14.0 0.001 4.6 0.040 0.8 0.367 SF-PSS 1.6 ± 0.6 2.0 ± 0.6 1.4 ± 0.5 1.7 ± 0.57 5.8 0.022 3.8 0.060 0.0 0.830 SF-sz-min 426.6 ± 31.7 402.2 ± 53.5 447.8 ± 37.3 439.4 ± 49.01 2.1 0.152 9.2 0.005 0.1 0.765 SF-SEI 89.8 ± 7.1 87.9 ± 8.8 93.9 ± 3.6 92.9 ± 4.80 0.5 0.479 7.5 0.010 0.2 0.653 SF-WACHZ 18.0 ± 25.1 29.0 ± 35.7 8.7 ± 12.5 13.7 ± 16.66 1.5 0.225 3.9 0.056 0.0 0.958 SF-ESL 30.7 ± 19.3 24.0 ± 16.1 20.4 ± 12.9 18.8 ± 16.58 0.4 0.512 7.1 0.011 1.6 0.215 Abbreviations. SF-SQ, total sleep quality; SF-GES, restorative value of sleep; SF-PSYAA, evening mood; SF-PSYEA, fatigue in the evening; SF-PSS, psychosomatic symptoms during sleep onset; SF-sz-min, estimated total sleep time; SF-SEI, sleep efficiency; SF-WACHZ, nocturnal wake time; SF-ESL, subjective estimates of sleep latency. Table options BPD patients reported significantly worse total sleep quality (SF-SQ) in the two study nights than controls, felt significantly more exhausted in the evening and had significantly worse evening mood. Additionally, in the BPD group significantly more “psychosomatic” symptoms during sleep onset such as palpitations, myalgia, etc. were reported and the restorative value of sleep was rated significantly lower by the patients. However, BPD patients reported ameliorated evening mood in the baseline night whereas controls did not improve. In contrast, subjective ratings for the two nights for total sleep time (min), sleep efficiency (%) as well as sleep latency (min) and nocturnal wake time (min) did not reveal differences between BPD patients and controls. Both groups reported significant improvements of these parameters in the second night. In the baseline night six BPD patients reported dreaming. Five of them rated their dreams as unpleasant, while one rated her dreams as neutral. Meanwhile, only two controls remembered dreaming (both unpleasant and pleasant content). Exploratory analysis of sleep parameter dependence upon reported dreaming (SF-A; three-level factor “can’t remember having dreamt”, have dreamt but can’t remember contents”, “have dreamt and remember contents”) revealed a significant dependence of subjective wake time on dream recall (F(2; 35) = 3.88, p = 0.03) and tendentially between “psychosomatic” symptoms during sleep onset and dream recall (F(2; 35) = 3.22, p = 0.052). No other correlations between subjective and objective (conventional and spectrally analyzed) sleep parameters were observed (data not shown).