Apneas were few and their number and duration did not increase after benzodiazepine treatment (Table 3). The patient with an abnormal number of sleep apneas, excluded from the final analysis, had 29 apneas per hour on placebo and 38 and 33 per hour on nitrazepam and flunitrazepam, respectively. The mean duration of his apneas was 17, 19 and 20 s and mean SaOz fell by 2.5, 2.6 and 4.4 percent, respectively. Link
The fall in mean Sa02 from wakefulness to sleep (average for all stages) was small and did not differ significantly between the study nights (Table 3). We also analyzed the fall in mean Sa02 for each sleep stage but found no differences among the study nights (Fig 2). The lowest SaO£ (usually occurring during REM sleep) did not differ between the study nights. There was, however, a small but significant difference in the sleep-induced increase in tcPco2 among the study nights (Table 3).

Sleep Variables
When compared with healthy volunteers studied in our sleep laboratory, our patients had a lower total sleeping time, especially during the placebo night which, for safety reasons, was always the first. The first night in a sleep laboratory usually is associated with a decrease in REM sleep and an increase in total wake time as compared with consecutive nights. The importance of this effect has, however, been questioned. When patients sleep at home between study nights, as in our study, the first night effect seems to be small. We, therefore, believe that a first night effect does not explain all differences in sleep quality between treatment and placebo nights in our study. It also should be emphasized that there is no first night effect for the respiratory variables.
Impaired sleep quality previously has been described in hypoxemic COPD patients. This may be due to frequent arousals caused by respiratory stimuli. Attenuation of the arousal response to such stimuli may thus contribute to the increased sleeping time after benzodiazepines. The reduction of REM sleep seen after benzodiazepine administration in healthy individuals did not occur in our COPD patients.
Table 3—Respiratory Variables (Means and SD)

Sa02 Fall with Sleep (All Stages) (%) Lowest Sa02 During Sleep Max tcPco2 Increase with Sleep (kPa) No. of Apneas
Placebo 1.3± 1.9 81.7±6.0 0.6±0.2 8± 12
Nitrazepam 1.4± 1.6 79.7 ±7.9 0.8±0.2 1 ±2
Flunitrazepam 1.9±2.4 80.3 ±8.5 ©1+o


5 ±10
p value* NS NS p<0.05* NS


Figure 2. Fall in mean Sa02 for each sleep stage after placebo and active treatment.

Figure 2. Fall in mean Sa02 for each sleep stage after placebo and active treatment.