Search results
The literature search was conducted in accordance with PRISMA FLOW, and a summary of the search results is presented in Fig. 2. The total number of studies found in each database was 215. After excluding 101 studies due to duplication, 114 studies were left. Afterwards, 83 studies were excluded due to animal studies, non-original grey literature, non-English literature, and literature with unsuitable indications and treatments according to the inclusion and exclusion criteria. 31 documents were judged by the full text review whether they were suitable for our research purposes and finally, 11 literatures were suitable for identifying the effects and safety of rTMS on peripartum depression and detailed Excluded literature and the reasons are presented in Appendix 2. The original text was requested through the author's e-mail if only abstract was present. Finally, there were 11 studies suitable for qualitative synthesis and 5 studies suitable for quantitative synthesis.
Characteristics of Selected Studies
There were a total of 11 selected literature related to the efficacy and safety of rTMS for PPD, of which 2 RCT studies [59, 60], 4 NRS [61–64], and 5 case studies [64–68]. The total number of participants are 100, 83 of whom received activee rTMS treatment. During pregnancy, 65 patients were treated with rTMS and most of them consisted of the second and third trimester of pregnancy. 17 patients were treated for postpartum depression and 2 patients were treated during pregnancy and after childbirth. The people included in the study were suffering from severe depression with HDRS scores of 17 or more point, two of which showed biopolar disorder. 20 participants were treated antidepressants [61, 63, 66, 69] along with rTMS, 6 participants were treated on psychotropic [59, 64]and 1 participant was treated on clonazepam for insomnia [62].
Table 1
Characteristics of subjects included studies
Study | Study design | Subjects | Age | Gestational age | Psychiatric diagnosis | Simultaneous treatment |
D. R. Kim et al., 2019 [59] | Randomized controlled trial | Active 11 | 30.13 ± 5.78 | 22.19 ± 7.11 (Weeks) | MDD | Free |
Sham 11 | 26.41 ± 5.11 | 25.62 ± 7.61 (Weeks) |
Myczkowski et al., 2012 [60] | Double-blind Randomized controlled trial | Active 8 | 29.63 ± 6.37 | 4.13 ± 2.85 (Month) | MDD | Free |
Sham 6 | 26.67 ± 7.15 | 3.50 ± 2.74 (Month) |
D. R. Kim et al., 2011 [61] | Non Randomized controlled trial | 10 | 31.2 ± 5.6 | 25.8 ± 5.16 (weeks) | MDD | 4 patients treated on antidepressant |
Hizli Sayar et al., 2014 [63] | Non-Randomized controlled trial | 29 | 32.69 ± 3.69 | 14.26 ± 8.25 (weeks) | MDD | 12 patients treated on antidepressant |
Garcia et al., 2010 [62] | Non-Randomized controlled trial | 7 | 34.11 ± 6.05 | After birth 30 days to 1 year | MDD | Free |
Tarhan et al., 2012 [70] | Non-Randomized controlled trial | 7 | * | * | MDD | * |
Zhang et al., 2010 [65] | Case report | 1 | 28 | 14 (weeks) | MDD | Free |
Tan et al., 2008 [64] | Case report | 1 | 30 | From 0 to postpartum period | MDD | Free |
Ferra˜o et al., 2018 [66] | Case report | 3 (Left) | 35.7 ± 2.05 | 6.67 ± 3.06 (weeks) | MDD | 2 patients treated on antidepressant |
1 (Right) | 36 | 8 (weeks) | 1 patients treated on antidepressant |
Cohen et al., 2008 [67] | Case report | 1 | 30 | 20 (weeks) | MDD | Free |
Monika Klírová et al [68] | Case report | 1 (Left) | 30 | 16 (weeks) | MDD | Treated on antidepressant |
1 (Right) | 30 | 31 (weeks) | MDD | Treated on antidepressant |
Because protocols for rTMS treatment for peripartum depression have not yet been established, protocols for each study have all been different. 22 patients were stimulated the right DLPFC [59, 61, 66, 67], 79 patients were stimulated the left DLPFC [60, 62–66, 70], and Burton et al [69] attempted to treat using combination of right and left DLPFC stimulation, and Ferra˜o et al [66] assigned different protocol and stimulation sites depending on the symptom of patients. The right DLPFC was stimulated using 1 Hz and the left DLPFC was treated with relatively high frequencies of 1 to 25 Hz. In the process of treatment, some rTMS parameters such as number of pulses, and number of session were different in all studies, but researcher consider interval time and stimulus duration to minimize adverse event like a seizure.
Table 2
Characteristics of Treatment included studies
Study | Motor threshold | Site of Stimulation | Frequency | Number of pulses | Inter-event interval | Number of Session |
D. R. Kim et al., 2019 [59] | 100% | Right DLPFC* | 1-Hz | 900 | 60 s on 60 s off | 20 |
Myczkowski et al., 2012 [60] | 120% | Left DLPFC | 5-Hz | 1250 | 10 s on 20 s off | 25 |
D. R. Kim et al., 2011 [61] | 100% | Right DLPFC | 1-Hz | 300 | 60 s on 60 s off | 20 |
Hizli Sayar et al., 2014 [63] | 100% | Left DLPFC | 25-Hz | 1000 | 2 s on 30 s off | 18 |
Garcia et al., 2010 [62] | 120% | Left DLPFC | 10-Hz | 150 | 4 s on 26 s off | 20 |
Tarhan et al., 2012 [70] | 100% | Left DLPFC | 25-Hz | 1000 | 2 s on 30 s off | 18 |
Zhang et al., 2010 [65] | 90% | Left DLPFC | 1-Hz | 1200 | 20 s off | 42 |
Tan et al., 2008 [64] | 110% | Left DLPFC | 25-Hz | 1000 | 2 s on 28 s off | 77 |
Ferra˜o et al., 2018 [66] | 120% | Left DLPFC | 10-Hz | 3000 | * | 42.67 |
Right DLPRC | 1-Hz | 1800 | 20 |
Cohen et al., 2008 [67] | 110% | Right DLPFC | 1-Hz | 1600 | * | 1 |
Monika Klírová et al [68] | 100% | Left DLPFC | 20-Hz | 2000 | 2 s on 30 s off | 15 |
Right DLPRC | 1-Hz | 300 | 60 s on 60 s off | 15 |
* DLPFC: Dorsolateral prefrontal cortex |
The average improvement in depression showed a decrease rate of 59% for 76 patients, except for a study by Tarhan et al [70], which did not disclose specific HDRS scores of patients. 37% participants showed remission of the depression (HDRS-17 ≤ 8, HDRS-21 ≤ 7, HDRS-24 ≤ 8) and 66% showed responded to rTMS (HDRS score reduced more than 50%). Among the two random clinical trial studies, D. R. Kim et al [67] showed 45.45% response rate for the control group, while response rate of the experimental group is 81% (p = 0.088) and remission rate was 18.18% for the control group whereas remission rate of the experimental group is 27.25% (p = 0.613), showing a high therapeutic effect. Additionally, another Myczkowski et al [60] showed 7% decrease of HDRS scores in the placebo group, but in the experimetal group, they showed a more than 30% deccrease rate (p = 0.020), similar to antidepressants. In NRS, the response rate of the 56 participants was 33% and the mission rate was 59%. In Garcia et al., 2010 [62], HDRS scores of all participants fall below eight and without 1 participants, all of them experienced their HDRS scores drop by more than 50% and other NRS also revealed significant clinical results. In the case study, nine participants responded (rating scale reduction rate ≥ 50%).
Table 3
Characteristics of result included studies
Study | Instrument | Pre-TMS | Post-TMS | Remission | Response | Side effect (mother) | Infants |
D. R. Kim et al., 2019 [59] | HDRS-17 | 23.18 ± 3.54 | 9.27 ± 6.05 | 3 | 9 | 1 Patients had headache | 3 pre term births 1 shoulder dystocia |
22.27 ± 2.65 | 13.18 ± 8.00 | 2 | 5 |
Myczkowski et al., 2012 [60] | HDRS-17 | 29.13 ± 5.64 | 18.50 ± 9.83 | * | * | 2 Patients had mild headache | * |
26.67 ± 5.68 | 24.83 ± 7.60 |
D. R. Kim et al., 2011 [61] | HDRS-17 | 24.4 ± 5.6 | 9.7 ± 6.1 | 3 | 7 | 4 Patients had mild headache. 1 patients had an episode of supine hypotension | All infants were well-baby nursery |
Hizli Sayar et al., 2014 [63] | HDRS-17 | 26.67 ± 5.58 | 13.03 ± 6.93 | 6 | 12 | None | None of baby showed any Abnormalities. |
Garcia et al., 2010 [62] | HDRS-24 | 22.67 ± 6.44 | 2.14 ± 3.19 | 8 | 9 | Headache, site pain | * |
Tarhan et al., 2012 [70] | HDRS-17 | * | * | 2 | 5 | None | All gave healthy babies |
Zhang et al., 2010 [65] | HDRS-24 | 35 | 8 | 1 | 1 | None | Healthy boy |
Tan et al., 2008 [64] | HDRS-17 | 38 | 4 | 1 | 1 | None | No diagnosis of disease |
Ferra˜o et al., 2018 [66] | HDRS-21 | 24.33 ± 5.24 | 7.33 ± 4.03 | 2 | 3 | 2 patients had discomfort at the application site. | Twins were preterm |
12 | 6 | 1 | 1 | discomfort at the application site and sore throat |
Cohen et al., 2008 [67] | HDRS-17 | 18 | 6 | 1 | 1 | * | The infant had a normal neurologic development |
Monika Klírová et al., 2008 [68] | MADRS | 33 | 2 | * | * | None | Healthy baby |
BDI | 29 | 12 | * | * |
Meta-analysis
Therapeutic effect
All of the included studies used HDRS to identify degree of depression. The correlation coefficient was 0.5 [71] and because heterogeneity was p < 0.001, I 2= 71.933 the random effects model was applied [72]. rTMS have an effect on mitigating depression with SMD = 1.806, 95% CI: 0.920 – 2.692 and the difference was statistically significant. (Z=6.079, p <0.01) [73].
Abbreviations: BAI: Beck Anxiety Inventory; BDI: Beck depression Inventory; CGI-S: Clinical Global
impression scale; EPDS: Edinburgh Postnatal Depression Scale; HDRS: Hamilton Depression
Rating Scale; GAS, Global Assessment Scale; SF-36-V and SF-36-MH: 36-item Quality of Life
Health Survey, Vitality and Mental Health scores; IDS-SR: Inventory of Depressive
Symptomatology-Self-Report.
Figure 3. Forest plot of therapeutic effect
Side effect
The rate of occurrence of all side effects from the included studies was determined. The heterogeneity of the studies was p = 0.112, I 2= 46.631 so fixed effect model was applied [72]. The probability of side effects was small (event rate = 0.346, Z= -2.696, p =0.007) [73].
Figure 4. Forest plot of side effect
Sensitivity analysis
Except for Garcia’s study, which was too high therapeutic effect compared to other studies, the rTMS
Of SMD = 1.074 (95% CI: 0.689 – 1.459, Z=5.468, P <0.001) showed significant therapeutic effect
for PPD (Figure5).
Abbreviations: BAI: Beck Anxiety Inventory; BDI: Beck depression Inventory; CGI-S: Clinical Global
impression scale; EPDS: Edinburgh Postnatal Depression Scale; HDRS: Hamilton Depression
Rating Scale; GAS, Global Assessment Scale; SF-36-V and SF-36-MH: 36-item Quality of Life
Health Survey, Vitality and Mental Health scores.
Figure 5. Forest plot of therapeutic effect without Garcia study
Quality of the included studies
According to experimental design, RCT was evaluated ROB2, NRS was ROBIN-I, case study was Methodological quality tool and the detailed domain for the risk of bias assessment is presented in Appendix 3. Figure 7 is funnel plot for the Treatment effect of rTMS in treatment of PPD. The occurrence of asymmetric can be caused by publication bias or other causes.
(A) RCT risk of bias graph; (B) NRS risk of bias graph; (C) case study risk of bias graph
Figure 6. risk of bias graph