. 64
( 68 .)


1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

Baseline assessment: a 31-year-old female
previously diagnosed as rapid cycling begins

the month with a hypomanic phase.

Hours Slept
In light of labs indicating valproate levels of
62 mmol/mg, treatment commenced with
increasing dose of valproic acid and
DVP 250 33 33333333 3344 44444444 444445 5 55 5
assessed weekly.
Lorazepam 111 11111 1111 3 33321 1 1 1 1 1111 1 22211

Month 1 with 7-day assessment intervals
Week 1 Week 2 Week 3 Week 4
Psychotic Sxs
Diagnostic impressions made serially based

on careful assessment of the preceding 7

days leaves much room for subjectivity.

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

Potential diagnoses:

Week 1: Depression, mania, mixed
Week 2: Depression, mania, recovering
Week 3: Depression, recovering
Hours Slept
Week 4: Depression, mania, recovering

Entire month = rapid cycling 1“2 phases
DVP 250 3 333 3333 3 333 44 4 4 444444 4 4 4 4 4 5 5 55 5
Lorazepam 1111 1 11 111 1 1 3 33 321 1 1 1 1 11111 222 11
CGI-I = 4 (no change)

Month 12 mood chart
Psychotic Sxs

Month 12 diagnostic assessments on based
serial weekly assessments made during

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

-2 cycles

Severe Assessment
2 cycles
Hours Slept
Anxiety Recovered
Entire monthly: recovered (0 cycles)
CGI-I = 1 (very much improved)
DVP 250 mg 66 6 64 8 66 4 8 6 66 66 66 66 6 666 66 6 666 6 6 6
Olanzapine 10 mg 1111 11 11 1 1 111 11 1 11 1 1 1 1 1111 1 111 1 1
22 1 1 11 22 2 2
Lorazepam 1 mg

Fig 17.4 Challenges in assessment of phases during course of rapid cycling. CGI-I, Clinical Global
Impression of Improvement.

phase-counting technique allows both continuous and categorical outcome defi-
nitions for rapid cycling. As a continuous measure, the 48-h rule limits the range of
phase changes per month to between zero and 15. Even so, it can be problematic to
reach consensus on the total number of phase changes. Consensus is much easier to
reach when the number of phases is collapsed to one of three categories: none, one, or
more than one. Thus, in the example shown in Figure 17.4, the last month would be
379 Investigational strategies

categorized as zero phase changes, because no episode lasted longer than 24 h. This
approach simplifies the measurement of treatment outcome and decreases the
variance attributable to individual raters by not requiring fine distinctions.
STEP-BD assesses clinical states prospectively using the Clinical Monitoring
Form (CMF), which also serves as the progress note in the patient™s medical record.
The CMF is a one-page assessment tool (available from www.manicdepressive.org;
Sachs et al., 2002a) consisting of nine parts, including modified versions of the
Structured Clinical Interview for DSM-IV (SCID) current mood modules; asso-
ciated symptoms, stressors, medical problems, and comorbid conditions; selected
mental status items; current medication compliance and adverse effects; labora-
tory data; summary scores (i.e., clinical status, Clinical Global Impressions Severity
of Illness and Global Assessment of Functioning) narrative, and treatment plan.
The CMF™s operational conventions for concise clinical record-keeping allow it
to serve as the source document as well as a key outcome measure in STEP-BD
(Sachs et al., 2002b). Central to the tracking of outcome is designation of the
current clinical status. Course of illness over time can be determined prospectively
by using the CMF clinical status designations made at every follow-up visit. The
eight mutually exclusive clinical states are outlined in Figure 17.5. This clinician-
rated technique makes use of, but does not rely on, self-report measures such as the
STEP-BD waiting-room self-report form and daily mood-charting. The CMF, its
companion waiting-room self-report form, daily mood charts, and full instruc-
tions for their use are available at www.manicdepressive.org. These methods do
not eliminate the problems encountered in tracking treatment response in patients
with rapid cycling. Figure 17.4 depicts the disparate assessments made by raters for
a rapid-cycling bipolar patient subject during the 1st and 12th month of valproic
acid treatment. As noted above, using the 48-h rule and simplifying the outcome
criteria helps to avoid potential discrepant ratings.

Rapid cycling: clinical trial design issues
Research on rapid cycling need not wait until perfect solutions are found for the
many daunting problems noted here. Considerable progress can be made by first
testing simple approaches to the easiest questions.
Sample selection and outcome measures are perhaps the most important aspects of
any trial. Although retrospective assessment of rapid cycling at study entry does
appear to give an indication of the propensity to cycle (Baldessarini et al., 2000),
studies are disadvantaged when subjects likely to remain well are randomized.
Limiting randomization to only those subjects with active cycling during a prospec-
tive assessment period can help reduce this problem. We recommend entering
retrospectively diagnosed subjects into a stabilization phase lasting 2“3 months
380 G. Sachs and M. Graves

’ Clinical status

• DSM-IV • Clinical status
Acute episode criteria Acute episode criteria
• Depression • Depression
• Hypomania • Hypomania
• Mania • Mania
• Mixed episode • Mixed episode
Continued symptomatic
Partial remission

Recovered Recovered


If DSM criteria for current episode are positive
Associated symptoms
Assigned status
of mania or depression
≥ 5 moderate
Major depression Depression
≥ 3 moderate
Mania Mania
≥ 3 moderate Hypomania
≥ 3 moderate for mania and
Major depression and mania Mixed
≥ 5 moderate for depression

If DSM criteria for current episode are negative

"Recovered" Associated symptoms
Assigned status
from last acute episode of mania or depression

≥ 3 moderate symptoms
No Continued symptomatic
¤ 2 moderate symptoms
No Recovering
Yes, if "recovering" ≥ 8 consecutive ¤ 2 moderate symptoms Recovered
≥ 3 moderate symptoms
Yes Roughening

Fig. 17.5 Mapping DSM-IV to eight operationally defined clinical states. DSM-IV, Diagnostic and
Statistical Manual of Mental Disorders, 4th edn.

and randomizing only those subjects who have at least one prospectively observed
phase change during any month. Subjects should be excluded who meet criteria for
secondary rapid cycling (see definitions and Table. 17.2).
The difficulties in determination of mood state and cycle rate are magnified as the
time period being assessed becomes briefer. This problem may be better managed by
avoiding fine but unreliable measurement of brief changes in mood. It is likely that
381 Investigational strategies

Table 17.2 Causes of rapid cycling

Brain injury

“ Mental retardation
“ Head trauma
“ Multiple sclerosis

“ Hypothyroidism
“ Reproductive hormones
Psychotropic drugs

“ Alcohol
“ Stimulants
“ Antidepressant drugs
Circadian rhythm abnormality

retrospective measures like the LIFE can make more meaningful assessments
when applied over short intervals. Standard operating procedures for assigning a
clinical status for each week using the CMF appear to offer a reasonably reliable
means of prospective assessment. Outpatient studies could employ either of
these techniques to establish episode pattern as well as measure phase changes.


. 64
( 68 .)