Subjects and methods

Subjects were selected from among inpatients consecutively admitted from June 1994 to June 1997 to the Rehabilitation Center Clinica Santa Lucia after a recent stroke. During those years, both the clinical aspects of poststroke depression and the rehabilitation process had been assessed with valid and reliable measures during research programs aiming to investigate: (a) the phenomenological aspects12 13 and the anatomical correlates14 of poststroke depression, and (b) the variables associated with a positive outcome of rehabilitation.9

Furthermore, some patients had been treated and others had not (after the initial assessment of the variables under study) according to the opinions of their respective physicians. The records of all patients who had undergone a standard assessment of poststroke depression and a rehabilitation programme in that period were taken into account. The patients included in the study met the following criteria: (a) Interval between stroke and admission to the clinic of 1 to 4 months; (b) monohemispheric stroke established by a single CT; (c) no history of strokes and depressive disorders or other important psychiatric disturbances; (d) a decision to treat poststroke depression or not (if a significant level of depression has been detected) based on the general a priori opinion of their physician and not on the severity of depression. Records of 290 patients were examined and 64 patients were included in this study.

ASSESSMENT OF POSTSTROKE DEPRESSION

The assessment of poststroke depression included a structured psychiatric interview and the administration of the Hamilton depression rating scale (HDRS), which is a well known general depression scale.15 Forty nine patients who on the structured psychiatric interview and on the depression rating scale had met the DSM-III-R16 diagnostic criteria of (major or minor) depression, were considered as depressed. The remaining 15 patients were classified as non-depressed.

PHARMACOLOGICAL TREATMENT OF POSTSTROKE DEPRESSION

During the period considered for our research, some physicians of the Clinic Santa Lucia tended to treat poststroke depression with the serotonine reuptake inhibitor fluoxetine or with other antidepressant drugs, whereas other physicians were reluctant to use antidepressant drugs in patients with stroke. Twenty four of the depressed patients with stroke had been treated in our sample, whereas the other 25 had not received antidepressant drugs. Twenty three out of 24 patients received fluoxetine in monotherapy, with dosages ranging between 20 and 40 mg/day and the remaining patient underwent treatment with amitryptiline.

ASSESSMENT OF THE OUTCOME OF THE REHABILITATION PROGRAMME

The outcome of rehabilitation was evaluated by computing scores obtained by our three groups of patients with stroke before, during, and at the end of the rehabilitation programme with the following impairment and disability scales: (1) the Barthel index,17 the scores of which range between 0 and 100—the top scores implying complete functional independence in daily life activities; (2) the revised and validated version of the Canadian neurological scale,18 the higher scores of which indicate better neurological performances; (3) the Rivermead mobility index,19 which assesses the patient's ability in 15 common daily movements with the highest score corresponding to complete motor ability.

The baseline assessment made at the entry into the clinic, the intermediate assessment made after 1 month of rehabilitation, and the final assessment made at the end of the rehabilitation programme were used to evaluate the effect of the main variables (groups and repeated measures) and their interaction with the results obtained with the impairment and disability scales (Canadian neurological scale, Barthel index, and Rivermead motility index). Data were subjected to a two way multivariate analysis of variance (MANOVA) repeated measures analysis. Post hoc comparisons were by Tukey test for unequal sample size when the effect of the main factors or the interaction reached significance.

Results

The characteristics of the 49 depressed (24 treated and 25 non-treated) and the 15 non-depressed patients with stroke who entered the study are reported in table 1.

Age, sex, education level, hemispheric side of lesion, and interval between the onset of stroke and the beginning of the rehabilitation programme were similar among groups. On the contrary, severity of depression and scores obtained at the entry of the study with the Canadian neurological scale, Barthel index, and Rivermead mobility index were worse in depressed than in non-depressed patients with stroke. On the other hand, neither the severity of baseline depression nor baseline scores obtained with impairment and disability scales were different in treated and untreated depressed patients. The fact that scores obtained on the HDRS were similar in treated and untreated depressed patients with stroke, supports our assumptions that the decision to treat or not was based on the a priori opinion of their physicians and not on the severity of depression.

A one way ANOVA repeated measures comparing the HDRS for the treated and non-treated depressed patients with stroke, showed a better recovery from depression in the treated than in the untreated group (p<0.01)

RELATION BETWEEN PRESENCE OF POSTSTROKE DEPRESSION, ADMINISTRATION OF ANTIDEPRESSANT DRUGS, AND OUTCOME OF THE REHABILITATION

Figure 1 shows the scores obtained with the Canadian neurological scale, Barthel index, and Rivermead mobility index at entry into the study, during, and at the end of the rehabilitation programme by non-depressed and by treated and non-treated depressed patients with stroke. The main scores obtained by the three groups of patients on the same functional scales at the entry and end of the rehabilitation programme, are also shown in table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Mean scores on the Barthel index, Canadian neurological scale, and Rivermead mobility index obtained by the three groups of patients at the times studied.

For all three measures a non-significant difference was found in the group factor; a significant result was obtained with the repeated measure variables and a significant (groups×repeated measures) interaction was found (Rao's R (12,112) 3.07; p<0.001). This interaction was due to the fact that improvement after rehabilitation was higher in non-depressed and in depressed but treated patients with stroke, whereas non-treated depressed patients showed a much lower level of improvement. A specific effect analysis showed that this intergroup difference was particularly significant on the Rivermead mobility index (F (2, 122) 8.2; p < 0.0001) and less significant on the Canadian neurological scale (F (2, 122) 3.7; p<0.01) and on the Barthel index (F (2, 122) 2.8; p<0.03).

The post hoc comparison between treated and non-treated depressed patients with stroke showed a significant group×treatment interaction with the Rivermead mobility index (p<0.001) and Barthel index (p<0.02), but not with the Canadian neurological scale (p<0.16).

Discussion

The aim was to evaluate the influence of poststroke depression and antidepressant treatment on functional recovery in patients with stroke undergoing rehabilitation therapy. In particular, we intended to see if the negative effect of poststroke depression on functional recovery could be counterbalanced by the administration of antidepressant drugs. The results, showing that non-treated depressed patients present the lowest rate of functional recovery, whereas depressed but treated patients with stroke have a recovery rate similar to the non-depressed ones, clearly support this hypothesis. This claim is strengthened by two aspects of our study which suggest that the particularly modest results obtained by the untreated depressed patients were due to the harmful influence of psychological factors, and not to uncontrolled sampling bias between treated and non-treated patients. The first refers to the fact that at the beginning of the study, the two groups of depressed patients were indistinguishable in severity of depression and measure of impairment. The second refers to the fact that a different level of interaction between groups and treatment was found on the different outcome measures considered. This interaction was significant on the Rivermead mobility index and on the Barthel index (which are disability scales that examine behavioural aspects and are, therefore, more sensitive to the influence of psychological factors) and non-significant on the Canadian neurological scale (which examines more basic neurological impairment and is, therefore, less sensitive to the effect of these factors). The effects of antidepressant drugs on mood and functional recovery could be, at least partly, independent. Dam et al 11 have shown that the noradrenaline (norepinephrine) reuptake blocker maprotiline and the serotonine reuptake inhibitor fluoxetine have a similar effect on mood, but a different one on functional recovery; both improve the HDRS scores, but only fluoxetine actually improves functional recovery, whereas maprotiline hinders it. Future investigation is needed to clarify if the association between functional improvement and reduction of the depressive symptomatology found in our study with fluoxetine is seen with other antidepressant drugs. Clarification of this issue could help establish the most favourable drug for the treatment of poststroke depression.