Differential Effects of Electroconvulsive Therapy (ECT) on Patients with Schizophrenia Versus Depressive Disorder: Clinical Distinction Between Antipsychotic and Antidepressant Effects of ECT

1. Introduction

Electroconvulsive therapy (ECT) has been widely utilized in the treatment of various psychiatric disorders, including major depressive disorder (MDD) [1,2,3,4], bipolar disorder (BD) [5,6], and schizophrenia (SCZ) [7,8,9]. Recent longitudinal magnetic resonance imaging studies have demonstrated ECT-induced changes in gray matter volume in patients with MDD [10,11,12] and SCZ [13,14,15]. Furthermore, differences in diagnosis-specific volume changes [16] and large-scale network interaction [17] between MDD and SCZ have been elucidated, although the mechanism of the antidepressant and antipsychotic actions of ECT remains incompletely understood. We are interested in whether the diagnosis-specific brain changes in response to ECT would be accompanied by clinical improvement. However, a direct comparison of ECT response between SCZ and MDD has not yet been performed using the same rating scale and measuring plasma BDNF levels.

BDNF is considered one of the potential biomarkers of ECT effects and clinical response [22]. BDNF induces sustained elevation of intracellular Ca2+ [24], and it might exhibit an anti-inflammatory effect through the inhibition of microglial activation [25,26]. These facts suggest that transdiagnostic or diagnosis-specific changes in BDNF could be associated with the clinical effects of ECT. Serum BDNF increases with long-term aerobic exercise, and it is thought that the neural activity associated with physical exercise leads to increased expression of BDNF. We are interested in whether a similar mechanism of increased BDNF expression, similar to the effects of aerobic exercise, occurs through a period of ECT in patients with severe psychiatric symptoms.

In this study, our aim is to estimate the comparative clinical efficacy and acceptability of ECT between groups of patients with SCZ and combined groups of patients with MDD and BD. The patients were evaluated using the Brief Psychiatric Rating Scale (BPRS), which is useful for measuring symptom severity and changes in symptom status in patients with MDD, BD, as well as SCZ [18,19]. We also investigated the relationship between BDNF levels and clinical outcomes before and after ECT.

2. Materials and Methods

All patients included in the study received ECT at the Hizen Psychiatric Center, Saga, Japan, between November 2017 and January 2020. Written informed consent for participation in the study was obtained from all patients and their relatives, and the procedure was approved by the ethics committee of the hospital.

Of the 24 participants, 17 were diagnosed as SCZ or schizoaffective disorder, 7 were diagnosed as MDD or BD based on DSM-IV guidelines. We have listed the patient profiles, dividing them into three groups: SC, MDD+BD, and the excluded group. The summary includes age, sex, duration of illness, number of ECT sessions, medications prescribed, and side effects. For the excluded patients, the reasons for their exclusion are also provided (Table 1). The results were compared between the SC and MDD+BD groups. Age and age of onset were significantly higher in the MDD+BD group, while the number of hospitalizations was significantly lower. Although the frequency of ECT varied depending on the patient's clinical condition, no significant differences were found in the number of ECT sessions between the two groups (Table 2).

The symptom and disease severity were assessed by experienced clinicians using the Brief Psychiatric Rating Scale (BPRS), the Clinical Global Impression scale (CGI), and the Global Assessment of Functioning (GAF). A six-factor model of the 24-item BPRS (Affect, Positive symptoms, Negative symptoms, Activation, Resistance, and Disorganization) was utilized [20].

ECT was administered with bilateral or unilateral electrodes. Anesthesia and muscle relaxation were induced intravenously using propofol (1–2 mg/kg body weight) and rocuronium (1 mg/kg), respectively.

After ECT, venous blood was collected in blood collection tubes containing EDTA 2Na and aprotinin, centrifuged at 3,500 g for 15 min at 4 °C, and the supernatant was used as a plasma sample. Plasma BDNF concentrations were measured using a commercially available ELISA kit (Mature BDNF ELISA Kit Wako 296-83201, Fujifilm Wako Pure Chemical Industries, Ltd., Osaka, Japan).

All samples were analyzed in duplicate in one session. The average values of plasma BDNF for pre-ECT (SCZ: 682±527 pg/mL and MDD+BD: 484±228 pg/mL) were close to the average value (747±193 pg/mL) and within the range (170–1126 pg/mL) reported previously [35]. All samples were anonymized, and no clinical parameters or patient data were available.

The JMP18 software (SAS Institute, Cary, NC, USA) was utilized for all statistical analyses. Analysis of variance (ANOVA) was used to compare BPRS, CGI, GAF, and BDNF between the groups, and the mean, standard deviation, F-value, and p-value were obtained. Cohen's d and 95% confidence intervals (CI) used in the forest plot were calculated from the sample size, mean, and standard deviation of each group [36]. Regression analysis was used to determine the correlation between BDNF blood concentration and BPRS sub-scores, yielding the correlation coefficient and p-value.

3. Results

There were statistically significant differences in the rating scales before ECT compared with those after ECT for each patient group; however, no significant change between the groups for each disorder was observed (Figure 1). In addition, there were no significant differences between the SCZ and MDD+BD groups in pre-ECT BPRS (p = 0.195), CGI (p = 0.938), GAF (p = 0.758), and BDNF (p = 0.736). Similarly, no significant differences were observed in post-ECT BPRS (p = 0.222), CGI (p = 0.116), GAF (p = 0.103), and BDNF (p = 1.000).

To estimate the difference between the SCZ and MDD plus BD groups, we used Cohen’s d effect size measure to reanalyze the 24 individual BPRS items. This involved subtracting the rating scales for depressive and bipolar patients from those for schizophrenia patients, thus negative values represent favoring of psychopathology in the schizophrenia group, and vice versa (Figure 2). At baseline, only depression showed negative values (p=0.029), indicating favoring of SCZ, while hallucinations (p<0.0001), unusual thought content (p=0.001), bizarre behavior (p=0.006), conceptual disorganization (p=0.025), and mannerisms and posturing (p=0.038) showed positive values, indicating favoring of MDD+BD group. During treatment, many items in addition to baseline showed significant differences between groups and positive values. Thus, the MDD+ BD groups favored compared with the SCZ in terms of hostility (p=0.007), grandiosity (p=0.011), suspiciousness (p<0.0001), hallucinations (p<0.0001), unusual thought content (p<0.0001), bizarre behavior (p<0.0001), self-neglect (p=0.018), disorientation (p=0.050), conceptual disorganization (p<0.0001), tension (p=0.044), excitement (p=0.012), distractibility (p=0.003), and mannerisms and posturing (p=0.001). Only depression showed negative values (p=0.014), indicating favoring of SCZ at post ECT.

There were no significant differences in plasma BDNF levels between the groups at baseline, and no significant variation occurred in any group during treatment (Figure 1).

Although there was a tendency for increased BPRS levels in patients with MDD+BD observed after ECT, the difference was not statistically significant. Therefore, correlations between BDNF levels and BPRS sub-scores were estimated, yielding the following results (Table 3): rating scales of hostility (p=0.0006), suspiciousness (p=0.0260), conceptual disorganization (p=0.0409), tension (p=0.0228), uncooperativeness (p=0.0006), excitement(p<0.0001), and mannerisms and posturing (p=0.0139), positively correlated with BDNF levels in MDD+BD patients after ECT. Only suspiciousness sub-score (p=0.0154), was correlated with BDNF levels in SCZ patients. These findings suggest that BDNF may be involved in the effects of ECT in patients with MDD and BD, but not in patients with SCZ.

4. Discussion

Although ECT was effective for patients with SCZ, the positive symptoms, activation, and disorganization responses to ECT were statistically lower in SCZ than in MDD+BD, with the exception of depressive behavior. The gray matter volume (GMV) has been reported to increase after ECT in both the MDD and SCZ groups, indicating that some aspects of the mechanism of ECT are common across diagnoses, while others differ between diagnoses. Specifically, alterations in GMV within the left pregenual anterior cingulate cortex (pACC) were identified as being unique to SCZ and significantly correlated with percent changes on BPRS [16]. In MDD, there was an increased SMN (somatomotor network) to VIN (visual network) connection, increased self-connection of DMN (default mode network), decreased connection from LIN (limbic network) to multiple networks, and from FPN (frontoparietal network) to SMN, as well as decreased self-connection of DAN (dorsal attention network) and LIN, compared to SCZ. The self-connection of DAN was positively associated with the Beck Depression Inventory revised version (BDI-II) scores. On the other hand, the self-connection within the DMN showed a positive association, whereas the self-connection within LIN exhibited a negative association, with the Positive and Negative Syndrome Scale (PANSS) positive symptom scores in SCZ. Furthermore, the SMN-to-VIN connection demonstrated a positive association, while the VIN-to-DMN connection and the self-connection of the DMN showed negative association with the PANSS general psychopathology scores in individuals with SCZ [17].

The diagnosis-specific effects of ECT found in this study might be related to GMV change or aberrant large-scale network interactions, although a direct explanation could not be possible. Thus, comparing post-ECT changes between the two diagnostic groups would provide differentiation between diagnosis-specific changes and transdiagnostic ones.

BDNF induces long-lasting Ca2+-activated K+ currents and sustained elevation of intracellular Ca2+ [24]. BDNF also impacts intracellular Ca (2+) signaling in microglial cells [25,26], which may be important for the regulation of inflammatory responses and could also be involved in the pathophysiology and/or treatment of psychiatric disorders. Additionally, BDNF induces a rapid increase in the total number of cell surface GABA(A) receptors through the activation of Trk B receptors [31], suggesting a relationship between GABA(A) receptor deficits and CNS disorders [33]. The pathophysiology of SCZ is related to inflammatory responses mediated by microglia and intracellular Ca (2+) signaling [34]. However, our results showed no significant change in BDNF levels in SCZ both before and after ECT. On the other hand, no association of pro BDNF and BDNF levels with depressive state was reported [35], although the present results elucidated positive correlations between BDNF levels and BPRS sub scores in MDD + BD patients after ECT. Our results might indicate a diagnosis-specific change in BDNF after ECT. The ECT-induced change in BDNF levels has been reported in patients with MDD [27] and SCZ [29], though the results were conflicting [28,30].

A 2019 Cochrane review revealed that ECT has a positive effect on medium-term clinical response for patients with treatment-resistant schizophrenia (TRS); however, there is no clear and convincing advantage or disadvantage for adding ECT to standard care for other outcomes [7]. On the other hand, ECT has been proven effective in patients with treatment-resistant depression (TRD) [1,21]. Searches of the PubMed database were conducted in June 2023 using the medical subject headings (MeSH) terms 'ECT', 'treatment-resistant depression (TRD)', and 'treatment-resistant schizophrenia (TRS)'. The search identified 7,994 TRD reports, including 761 ECT reports, and 2,857 TRS reports, including 202 ECT reports. Therefore, the percentage of 'ECT' in 'TRD' was larger than that in 'TRS' (9.9% vs. 7.1%). These identified publications suggest that the utilization rate of ECT may be lower for the treatment of SCZ than for MDD + BD.

Utilization and practice of ECT worldwide indicate that affective disorders (unipolar/bipolar depression) are the main diagnoses in Australia, New Zealand, USA, Europe, while SCZ is the main diagnosis in Asia overall [23]. In Western countries, the majority of patients are older women with MDD, whereas in Asian countries, younger men with schizophrenia constitute the majority [4].

If the brain sites of ECT's antipsychotic and antidepressant effects can be estimate, we may be able to predict the therapeutic efficacy of ECT using brain imaging after ECT. This might provide an effective means of treating treatment-resistant depression as well as treating treatment-resistant schizophrenia, including clozapine.

5. Conclusions

In conclusion, although ECT was effective for patients with both SCZ and MDD+BD, the positive symptoms, activation, and disorganization responses to ECT were statistically higher in MDD+BD than in SCZ, with the exception of depressive behavior. Additionally, the ECT effects for several clinical outcomes of MDD+BD were associated with plasma BDNF levels. These findings suggest that ECT may be even more effective for MDD than for SCZ. The findings and their implications should be discussed in the broadest context possible. Future research directions may also be highlighted.

Limitation: In this study, we grouped depression and bipolar affective disorder together. The breakdown is as follows: depression (F329) 3 cases, bipolar affective disorder (F319) 2 cases, bipolar affective disorder with severe depressive episode without psychotic symptoms (F314) 1 case, and bipolar affective disorder with severe depressive episode with psychotic symptoms (F315) 1 case. In general, manic episodes are last from two weeks to four to five months, while depressive episodes typically last around six months. Particularly in the elderly, depressive episodes can last for more than a year. The ages of the F319 cases are 57 and 63, which are elderly, and it is thought that their depressive states may continue for a relatively long time. Since F314 represents severe depression without psychotic symptoms, these three cases can be included in the depression group. The issue is with F315, but we have included this case in the group. The F323 exhibited severe psychiatric symptoms and was administered risperidone, so was added to the SC group. These points should be noted to draw attention to careful interpretation of the results.