RESULTS
Since the two anxious subgroups did not significantly diffe r from each other with respect to any of the variables under baseline conditions, their MSLT, SSS, and HARS data have been pooled, and this body of data will be referred to as "baseline" for further comparisons.
Clinical Anxiety
The baseline HARS scores are reported in Table 2. As a whole, the generalized anxiety disorder patients showed a moderate to severe anxious mood (mean score on a 0-4 point scale = 2.8), a severe inner tension (mean = 3.0), moderate 
fears and insomnia (mean = 1.7), and mild to moderate somatic anxiety and depressed mood (mean = 1.6). Although there was general complaining about some difficulty with nighttime sleep, under basal conditions subjects showed almost no tendency to fall asleep during the daytime (Table 3). A slight drop in objective alertness was recorded at the time of the third nap (2 p.m.), about 90 min after the main meal that was served in the hospital. The effect of this slight postprandial dip, however, was either statistically not significant (Friedman AN-OVA for time of day effects: 3.64; p = n.s.) or subjectively recognized, since the baseline SSS profile did not show any consistent difference across naps (Friedman ANOVA: 1.83; p = n.s.). The effects of CDDZ and APZ administration on clinical anxiety are summarized in Table 4. Both drugs were equally effective in reducing total mean HARS scores and improving nighttime sleep after 2 and 4 weeks of therapy. Consequently, only the comparisons between baseline and a single "drug" conditions are listed in Table 4. No significant difference between drugs or time of observation and no severe side effects or adverse reactions were reported at the clinical examinations, except for excessive sleepiness and residual anxiety that occurred early in some subjects and did not subside after the first week of treatment. Generally, CDDZ dosage had to be reduced for excessive sedation whereas that of APZ increased for excessive residual anxiety. In fact, the mean CDDZ daily dose after 14 days was 1.4 mg whereas the mean APZ was 1.75 mg. This regime was successfully maintained until study completion.
Sleep Latencies and Subjective Sleepiness
As compared to a selected group of patients with PPI, our anxious patients had a significantly higher daily mean MSLT (18.7 vs. 15.0; p < 0.05). The effects of CDDZ and APZ administration on the mean MSLT profiles are shown in Table 3. Patients given CDDZ were significantly sleepier than at baseline only 
at the chronic MSLT session (p < 0.05), with no significant change from day 1 to day 30. Patients given APZ were sleepier than baseline on both MSLT sessions, with a significant change from the first to the 30th day of treatment (p < 0.01). Intergroup comparisons showed that alprazolam produced more sleepiness than CDDZ on acute administration (p < 0.01), but not after chronic use. The analysis for time-of-day effects showed that on the last nap of the acute condition, the APZ group still showed some objective tendency to fall asleep as compared to the baseline and to the CDDZ group (Kruskall-Wallis T: 8.4; p < 0.01). Mean SSS scores after acute ingestion were increased by drugs (Table 5), but returned to baseline at the chronic control, no differences between drugs having been observed.
Plasma Levels
Parent drugs mean plasma levels are shown in Table 6. Mean kinetics on acute administration were affected by the different administration schedules, with an early peak for CDDZ (at the first MSLT) and a late peak for APZ (at the last MSLT). Blood levels at the time of the first and the last naps on day 30 demonstrated a significant drug accumulation, the latter being five to seven times higher than at the first day, and a persistence of daytime fluctuation. No significant relationships were found between acute mean plasma levels, mean MSLT, and SSS scores.