Subjects will start at 20mg bid days 1-2, then 40mg bid days 3-4, them 60mg bid for days 5-6 then flexible dosing starting on day 7 between 60-80mg bid (60 mg bid or 80mg bid) for the remainder of the 6 week trial

Subjects will start on placebo and remain on placebo for the remainder of the 6 week trial

Participants with greater than or equal to 50 percent decrease from baseline in HAMD-17 total score responded yes; others responded no. Endpoint is LOCF endpoint among Week 1 - 6; Total score is first 17 items of HAM-D 25: measures range of depressive symptoms; scale 0-2 or 0-4 with higher scores being more severe. Total possible score 0 - 52.

Change is observed value at endpoint minus baseline value. Endpoint is Last Observation Carried Forward (LOCF) endpoint among Week 1 - 6; GAF is used to assess global psychological, social, & occupational functioning; 100=normal and 0=greatest abnormality

Response was Yes if MADRS Total Score was less than, equal to 12, if not, response was no. Endpoint is LOCF endpoint among Week 1 through 6; MADRS is 10-item instrument measuring depression; scale range 0(Normal) and 6(most abnormal)

Response was yes when HAM-D 17 total score was less than or equal to 7 , if not, response was no. Endpoint is LOCF endpoint among Week 1 through Week 6. Total score is first 17 items of the HAM-D 25,which measures the range of depressive symptoms; scale 0-2 or 0-4 with higher scores being more severe. Total possible score 0 - 52.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. Total score is first 17 items of the HAM-D 25, which measures the range of depressive symptoms patient currently experiencing; scale 0-2 or 0-4; 0=absent or not depressed, 2 or 4=most severe or extreme. Total possible score 0 - 52.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. HAM-A is a 14-item scale: rates intensity of psychic anxiety and somatic anxiety on a 5-point severity scale (range: 0=not present to 4=very severe). Total possible score is 0 - 56.

Change is observed value at each visit minus baseline value. Overall is average response of Weeks 1 - 6. YMRS: 11 item instrument with scale 0 to 4 for 7 items and 0 to 8 for 4 items; 0=normal; 4 or 8=most abnormal. Total possible score is 0 - 60.

Change is observed value at each visit minus baseline value. Overall is average response of Weeks 1 - 6. CGI-S measures severity of patient's mental illness. Scale range: 0 = not assessed, 1 = normal, 7 = among most extremely ill

Change is observed value at each visit minus baseline value. Overall is average response of Weeks 1 - 6. CGI-I is an instrument for Global assessment of improvement in patient's condition. Scale range: 0=not assessed, 1=very much improved, 7=very much worse

Change: observed value at each visit minus baseline value. Endpoint is Last Observation Carried Forward (LOCF) endpoint among Week 1 through Week 6. HAM-D 25: measures the range of depressive symptoms experienced. 25 Items with Scale range:0-2 or 0-4; 0=absent or not depressed, 2 or 4=most severe or extreme.Total possible score is 0 - 72.

Response each week was yes if CGI-I score less than or equal to 2 (much or very much improved), if not, response was no; Endpoint is LOCF endpoint among Week 1 through Week 6. CGI-I is a Global assessment of improvement in patient's condition. Scale range: 0=not assessed, 1=very much improved, 7=very much worse

Observed value each visit minus baseline value. Endpoint is LOCF Week 1 through Week 6. SDS: patient rated measure of disability and impairment in 3 items: work/school, social life, family life/home responsibilities:0(no disruption)- 10(extreme disruption). Total possible is 30.

Change is observed value at each visit minus baseline value. LOCF endpoint among Week 1 through Week 6. Q-LES-Q: 16-item instrument for patients assessment of his/her quality of life; overall level of satisfaction scale 1=very poor to 5=Very good (1 item re medication can be blank). Total possible score 15 - 80

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. Bech Melancholia is the sum of Scores on 6 Items pertaining to melancholia within Hamilton Depression Rating Scale (HAM-D). Scale 0 to 4, higher scores reflecting greater severity;Total possible 0 - 24.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This instrument = sum of Scores on 6 Items measuring anxiety/somatization within Hamilton Depression Rating Scale (HAM-D). Scale range is 0 to 4, higher scores reflecting greater severity. Total possible 0 - 24.

Change is observed value at each visit minus baseline value. This instrument = sum of Scores of 4 items on retardation within Hamilton Depression Rating Scale (HAM-D). Scale range is 0 to 4 with higher scores reflecting greater severity. Total possible is 0 - 16.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This instrument = sum of Scores of 3 items on sleep disturbance within Hamilton Depression Rating Scale (HAM-D). Scale range 0 to 4 with higher scores reflecting greater severity. Total possible is 0 - 12.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This test is part of Brief Assessment of Cognition and measures recall of 15 words repeated 5 times. Range 0-75 words, higher number reflects better recall.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition measuring immediate recall of 15 emotional words; higher number of words is better recall.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This instrument measures immediate recall of 15 non-emotional words (List 1); higher number of words is better recall.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This test is in Brief Assessment of Cognition and measures immediate recall of 15 emotional words (List 2); higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures immediate recall of 15 non-emotional words (List 2); higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures immediate recall of 15 emotional words (List 3); higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures immediate recall of 15 non-emotional words (List 3); higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures immediate recall (Cued) of 15 non-emotional words; higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures immediate recall (Cued) of 15 emotional words; higher number of words is better recall

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition in which patient sequences digits from lowest to highest. Range of number of correct responses (0-28); higher numbers show better digit sequencing and greater cognition.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition in which a patient places as many of 100 tokens (2 at a time) into a container as they can within 60 seconds. The higher number of tokens placed = patient is better at motor tasks

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition.Patients are given 60 seconds to name as many words as possible within a given category. The more words named=better cognition.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition. Patients given 60 seconds to generate as many words as possible that begin with a given letter; better verbal fluency = more words

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition. For 90 seconds, Patient writes numerals 1-9 as matched to symbols. Range 0 to 110 with higher totals = better cognition.

Change is observed value at each visit minus baseline value. Endpoint: LOCF endpoint among Week 1 through Week 6. Brief Assessment of Cognition: subjects asked to arrange balls in 2 pictures so they are identical and give the total number of ball movements to reach this arrangment. Range: 0-22; more correct = better cognition.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. Test in Brief Assessment of Cognition in which the number of correct emotional words in delayed recognition is measured. Range 0-75 with higher numbers showing better cognition.

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. Test in Brief Assessment of Cognition which measures number of correct emotional word's false alarms (during delayed recognition). Higher number = better cognition

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures number of Non-Emotional Words (Delayed Recognition); higher number of words = better cognition

Change is observed value at each visit minus baseline value. Endpoint is LOCF endpoint among Week 1 through Week 6. This is a test in Brief Assessment of Cognition which measures number of correct non-eEmotional word's false alarms(at delayed recognition). Higher number of words = greater cognition

Subjects were assigned to placebo

Week 2

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 2

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 3

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 3

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 4

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 4

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 5

Hochberg's adjustment used for multiple comparisons adjustment

Hochberg's adjustment used for multiple comparisons adjustment

Week 5

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 6 An estimated sample size of 180 was needed per treatment arm in order to achieve 85% power to detect a treatment difference of 3.5 in the mean change from baseline to Week 6 in MADRS total score with a two-sided t-test at the 0.05 significance level. The common standard deviation was estimated as 11.0.

The null hypotheses for the primary parameter is equality of mean change from baseline to Week 6 in MADRS total score between ziprasidone and placebo groups.

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Week 6 An estimated sample size of 180 was needed per treatment arm in order to achieve 85% power to detect a treatment difference of 3.5 in the mean change from baseline to Week 6 in MADRS total score with a two-sided t-test at the 0.05 significance level. The common standard deviation was estimated as 11.0.

The null hypotheses for the primary parameter is equality of mean change from baseline to Week 6 in MADRS total score between ziprasidone and placebo groups.

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Overall

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment

Overall

Fixed categorical: treatment, rapid cycling, center, visit, prior hosp & treatment-by-visit interaction & fixed continuous effect of baseline value

Hochberg's adjustment used for multiple comparisons adjustment