RESULTS AND COMMENTS
Baseline profiles were derived from the HSCL-90 self-rating scale, and the SUMMY variables for this sample were first compared with the same baseline measures derived from the first 80 Pisa Center subjects enrolled in the Worldwide Upjohn Phase II Study for Panic Disorder (9). The Upjohn study patients were selected utilizing stricter criteria in two respects, compared with the CLO-IMI study; the Upjohn study required at least one panic attack each week during the month before evaluation, whereas the CLO-IMI study required panic attacks during the past month without specifying disorder, or other anxiety disorder, whereas the CLO-IMI study did not. Even though the 152 subjects participating in the IMI or CLO treatments were selected according to broader, less restrictive criteria, they did not differ significantly from the Upjohn sample in any of the major SUMMY variables (Table 2). Contrary to expectation, the depressive mood item from the HDS was not significantly higher in the more leniently selected open study sample. Obviously, owing to selection criteria, the Upjohn study patients would be expected to have more spontaneous and situational panic attacks, but even these differences were small and nonsignificant. The variance in monthly number of situational panic attacks was higher in the Upjohn sample (p < 0.05), indicating
that relatively few cases accounted for the mean difference of two attacks per month between the two samples. The open study patients had slightly higher ratings on phobic avoidance, anticipatory and baseline anxiety, and impaired social adjustment (i.e., greater disability). A comparative analysis of the HSCL-90 factors also confirmed nearly super-imposable psychopathological profiles for the two groups (Table 2). The open study sample was slightly to somewhat higher in mean scores on all the psycho-pathology factors, but only the obsessive-compulsive factor mean was significantly higher, whereas the depression factor was nearly so. The actual rank ordering of the 9 HSCL factors by their weighted mean item scores was practically identical for the two samples. These baseline comparisons of the two study samples suggest a particular stability of the panic-phobic diagnostic categories. Although the operational criteria for the disorders were less strictly followed in the open trial, the typical panic-phobic syndromes are clearly present in that sample, and the overall profiles of psychopathology are remarkably similar. In the CLO-IMI study, the two treatment groups were comparable in the proportion of the three DSM-III-R diagnoses and in age, sex, and age at onset. These characteristics were also not related to dropout rates. The attrition rate to date has included 31 IMI and 29 CLO patients. None of these cases returned after the first visit and then were lost to follow-up. The attrition rate did not differ for the two groups, but patients lost from the IMI group had significantly more situational panic attacks per month than those who discontinued CLO treatment (Table 3). Dropouts from the two treatments did not differ significantly in the other syndrome features shown in Tables 3 and 4. At present 33 patients are in progress in the study but have completed less than 10 weeks of treatment. The 33 IMI and 26 CLO cases who have completed the 10th week did not differ significantly at baseline in any of the major features shown in Table 3.
Thus, the two groups completing 10 weeks of treatments were comparable at baseline. Both drugs administered appeared to show equivalent and significant rates of improvement on nearly all measures of symptomatology (Table 4). In general, significant improvement occurred in both groups during the 10 weeks. After 10 weeks of treatment, the number of panic attacks dropped to a very low level with both medicines. The panic blocking effect of both drugs appeared to be very rapid and consistent. Both drugs also appeared to have affected anticipatory anxiety and phobic avoidance, with about equally significant improvement in both groups. The HDS ratings of depressive mood and psychic anxiety indicated both drugs were significantly active. The significant drug X time interaction found for the psychic anxiety item from the HDS indicated faster action by CLO; the CLO advantage was in the first 2 weeks. The other significant drug difference also favored CLO, which showed an advantage on somatic anxiety ratings (also from the HDS) throughout the 10-week period. Work and social disability were about equally affected by the two drugs. There also were no treatment differences in side effects, CGI severity, or CGI improvement.