This analysis was conducted on a subgroup of older depressed participants enrolled in a randomized placebo-controlled clinical trial comparing the efficacy of maintenance paroxetine, monthly IPT, and their combination as a maintenance therapy for LLD (Reynolds et al., 2006
). This analysis is limited to data from the subgroup of participants not receiving maintenance paroxetine (but on pill placebo) in order to examine the effects of monthly maintenance IPT (versus supportive clinical management) in the context of varying levels of cognitive function. This study took place in a university geropsychiatry clinic specializing in late-life depression. The 52 participants whose data were analyzed for this report (17 men and 35 women ≥ 70 years of age) met SCID/DSM-IV requirements for current major depression and had scores of ≥ 15 on the 17-item Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960
) and ≥ 17 on the Mini-Mental State Examination (Folstein et al., 1975
). Written informed consent was obtained after a complete description of the study was given to subjects and caregivers.
In the acute phase of the study, the entire study group (n = 195) received open treatment with paroxetine (10–40ml/day) and weekly IPT sessions. Subjects who achieved HRSD-17 scores of ≤ 10 for three consecutive weeks entered into a 16-week continuation phase (n = 116) in which paroxetine was continued and IPT sessions were reduced to bimonthly. Patients who remained stable throughout continuation treatment were eligible to enter maintenance treatment, in which they were randomly assigned to paroxetine and monthly maintenance IPT, paroxetine and clinical management (CM), monthly maintenance IPT and placebo, or CM and placebo.
During maintenance treatment subjects were seen monthly by the same clinician as during acute and continuation treatment in order to avoid an effect due to withdrawal of therapist. Clinicians conducted both IPT and supportive clinical management (CM) sessions, depending upon randomized treatment assignment. They were blinded as to whether participants were receiving pill placebo or paroxetine (in the parent study). Sessions were audiotaped and evaluated for treatment fidelity by an independent rater blind to treatment assignment. IPT sessions lasted 45 min while CM sessions lasted 30 min. CM sessions contained no specific psychotherapy; instead, patients were encouraged to report any symptoms and adverse (side) effects. Subjects remained in the maintenance phase for 2 years or until a recurrence of a SCID-defined major depressive episode, whichever occurred first. As noted in the report of the parent study (Reynolds et al., 2006
), the four maintenance treatment groups did not differ in sociodemographic, and clinical measures, Mattis scores, or extent of coexisting medical burden.
The Dementia Rating Scale (DRS; Mattis, 1988
), which was used as the primary cognitive measure, is an extended screening instrument designed to assess cognitive functioning across five separate cognitive domains (Attention, Conceptualization, Construction, Initiation/Perseveration, and Memory) in dementia. The test consists of 36 items including: repeating digit strings, following one and to-step commands, counting target letters embedding in a random array of letters, generating abstract concepts common to series of two items presented verbally and three items presented visually, copying designs, name writing, naming supermarket items, repeating series of rhymes, performing double-alternating hand movements, copying rows of alternating symbols, answering orientation items, delayed recall of two sentences and recognition memory for series of word pairs and design pairs. The study used DRS age- and education-corrected Total Scaled Scores which have a mean of 10 and Standard Deviation of 3. We defined cognitive impairment as a DRS Total Scaled Score ≤ 7. (As shown in , the two groups (IPT, CM) did not differ in total or domain scores on the Mattis DRS, either raw or scaled). A cut-off scaled score of 7 is one standard deviation below the mean, which translates into the 17th percentile of subjects from the norm group of similar age and with similar levels of education.
Mattis Dementia Rating Scale Scores
Time to recurrence was examined with a Cox proportional hazard model using treatment (IPT/CM) and scaled DRS score at randomization as covariates. We specifically limited the analysis to the IPT + placebo and CM + placebo groups. A significant interaction would indicate the moderating effect of cognition on treatment. To illustrate the survival curve of the model, trajectory curves for each treatment at two levels of cognition were generated, using scores of 10 and 7 (1 SD below the mean; Jurica et al., 2001
) for normal and impaired cognition, respectively. Actuarial recurrence rates were calculated for subjects grouped by level of cognition. The Breslow-Day test for homogeneity of the odds ratio was used to compare the rates of recurrence across cognitive level.