Transcranial magnetic stimulation for geriatric depression: Promises and pitfalls

Table 1 Meta-analytic studies of the antidepressant efficacy of transcranial magnetic stimulation published on or after 2003

Ref.	No. of trials	Mean age	Effect size
Allan et al[29] 2011	25	Not presented	NNT = 5
Berlim et al[22] (bilateral) 2013	7	49.3 + 5.7	NNT = 6 (res); 7 (rem)
Berlim et al[47] (HF) 2013	29	47.6 + 7.1	NNT = 6 (res); 8 (rem)
Berlim et al[21] (LF) 2013	8	49.39 ± 7.0	NNT = 5 (res/rem)
Schutter[25] 2013 (add-on)	6	44.47 ± 7.55	NNT = 7
Couturier[91] 2005	6	Not presented	WMD = 1.1
Gaynes et al[76] 2014	18 (TRD)	Not presented	NNT = 9 (res); 5 (rem)
Gross et al[71] 2007	5	44.7 + 4.2	d = 0.76
Herrmann et al[28] 2006	33	49.14 (subgroups split at age 50)	d = 0.65
Lam et al[92] 2008	24	Not presented	NNT = 6
Lepping et al[16] 2014	22 (nTRD)	Not presented	d = 0.63 (nTRD)
	10 (TRD)		d = 0.74 (TRD)
Martin et al[20] 2003	14	41.8 - 60.9	d = 0.35
Schutter[25] 2009 (HF)	30	49.5 + 7.8	d = 0.39
Schutter[26] 2010 (LF)	9	50.0 + 6.3	d = 0.63
Slotema et al[13] 2010	40	Not presented	d = 0.55
Kedzior et al[19] 2014	14	27.0-53.0	d = 0.42

d: Cohen’s d effect size; LF: Low frequency; TRD: Treatment resistant depression; nTRD: No treatment resistance; NNT: Number needed to treat; WMD: Weighted mean difference; res: Response; rem: Remission.

Table 2 Randomised Controlled Trials investigating antidepressant effect of transcranial magnetic stimulation in older subjects (mean age of sample > 60)

Trial	Age range	TMS parameters	Sample size	Methodology	Results
Manes et al[30] 2001	60.7 ± 9.8	20 Hz; 80%MT; left DLPFC; 800 pulses/session; 5 sessions; no. of pulses = 4000	20 (sham = 10)	Double blind RCT with handle as sham in subjects with one antidepressant failure aged 50-70, withdrawn from antidepressants for 5 d	No sham vs active differences; in each group 30% responded. Non-responders had reduced frontal volume. No drop-outs due to adverse effects
Mosimann et al[31] 2004	62 + 12	20 Hz; 100%MT; left DLPFC; 1600 pulses/session; 10 sessions; no. of pulses = 16000	24 (sham = 9)	Double blind RCT with tilted-sham in 40-90 years old subjects with TRD	No sham vs active differences
Jorge et al[32] 2008 (Trial 1)	62.9 (7.2)	10 Hz; 110%MT; MRI-based target localisation of left DLPFC; 1200 pulses/session; 10 sessions; no. of pulses = 12000	30 (sham = 15)	Double blind RCT with look-alike sham coil in subjects > 50 yr age with vascular depression and failed at least one antidepressant; all patients medication-free for at least 4 d before TMS	Age inversely correlated with response; frontal volume positively correlated with response. Active treatment: 33.3% responders; sham: 6.7% responders. No drop-outs due to adverse events
Jorge et al[32] 2008 (Trial 2)	64.3 (9.4)	10 Hz; 110%MT; MRI-based target localisation of left DLPFC; 1200 pulses/session; 15 sessions; no. of pulses = 18000	62 (sham = 29)	Double blind RCT with look-alike sham coil in subjects > 50 yr age with vascular depression and failed at least one antidepressant; all patients medication-free for at least 4 d before TMS	Active treatment: 39.4% responders; sham: 6.9% responders. Older subjects had better response for higher dose. No drop-outs due to adverse events

DLPFC: Dorsolateral Prefrontal Cortex; MT: Motor Threshold; RCT: Randomised Controlled Trial; TMS: Transcranial magnetic stimulation.

Table 3 Uncontrolled studies exploring the effect of age on the antidepressant effect of transcranial magnetic stimulation by recruiting older subjects (mean age of sample > 60) or undertaking analyses in subgroups with mean age > 60

Trial	Age range	TMS parameters	Sample size	Methodology	Results
Figiel et al[37] 1998	60.0 (22-89)	10 Hz; 110%MT; left DLPFC; 500 pulses/session; 5 sessions; no. of pulses = 2500	50	Uncontrolled trial in patients > 18 yr attending a Mood Disorder Clinic (most referred for ECT)	Age associated with treatment response; < 65 (n = 28) responded better (56%) than > 65 group (n = 22; 23% response). Overall 42% responded after the 5 sessions
Nahas et al[38] 2004	61.2 (7.3)	5Hz; 103%-141%MT (distance adjusted); left DLPFC; 1600 pulses/session; 15 sessions; no. of pulses = 18000	18	Uncontrolled trial in patients 55-75 yr; not selected for TRD	No correlation between age and response; 27% responded; 22% remitted; No drop-outs due to adverse events
Fabre et al[35] 2004	67.9 (6.7)	10 Hz; 100%MT; left DLPFC; 1600 pulses/session; 10 sessions; no. of pulses = 16000	11	Uncontrolled trial in patients age > 55 with vascular depression (first episode) and TRD but kept antidepressant free for 1 wk	5 out of 11 patients had clinically meaningful improvement in HDRS scores; response inversely related to frontal volume. No drop-outs due to adverse events
Abraham et al[39] 2007	66.8 (6.4)	10 Hz; 100%MT; left DLPFC; 1600 pulses/session; 10 sessions; no. of pulses = 16000	20	Uncontrolled trial in patients > 60 yr attending a specialist clinic - most referred for ECT	30% responded at the end of treatment; 1 dropout due to discomfort
Milev et al[78] 2009	69.0 (6.7)	Variable parameters: LF (1 Hz, 1200 pulses/session, n = 14), HF (10 Hz, 1600 pulses/session n = 31); both LF and HF (n = 4); 80%-110%MT; right or left DLPFC	49	Uncontrolled trial that includes patients with TRD referred to 2 specialist mood disorder clinics; all except 3 medicated	24.7% mean reduction in HDRS scores; 18% responders; 1 dropout due to discomfort out of 49. (Note: This sample includes n = 20 from Abraham 2007)
Pallanti et al[33] 2012	51.8 (14.1)	1 Hz; 110%MT; right DLPFC; 420 pulses/session; 15 sessions.; no. of pulses = 6300	102	Uncontrolled trial in consecutively enrolled nonpsychotic subjects in a TMS clinic with TRD	Age inversely related to response especially in patients > 60 yr; overall 56.9% responded. 62.1% of < 60 (n = 66) and 47.2% of > 60 (n = 36) responded at 3 wk. 18.6% drop-outs due to intolerance
Hizli Sayar et al[36] 2013	66.6 (5.8)	25 Hz; 100%MT; left DLPFC; 1000 pulses/session; 18 sessions; no. of pulses = 18000	70	Uncontrolled trial in patients > 60 yr with TRD	58.5% responded; 29.2% remitted; No drop-outs due to adverse events
Ciobanu et al[34] 2013	58.7 (14.0)	LF (1 Hz, 1200 pulses/session, n = 80), HF (10 Hz, 2000 pulses/session n = 13); 90%MT; right or left DLPFC; 15 sessions; no. of pulses = 18000	93	Uncontrolled trial in > 18 yr old subjects with TRD	Age not related to response; No difference between < 65 (n = 63; 53.3% responded) and > 65 age (n = 30; 46.7% responded) groups immediately and at 3 mo

DLPFC: Dorsolateral Prefrontal Cortex; MT: Motor Threshold; RCT: Randomised Controlled Trial; ECT: Electroconvulsive therapy; TRD: Treatment resistant depression; TMS: Transcranial magnetic stimulation.

Table 4 Summary of meta-analyses comparing transcranial magnetic stimulation and electroconvulsive therapy in depression

Meta-analysis	No. of studies/sample size	Mean age range	Summary of effect
Berlim et al[47] 2013	7/294	31.0-63.6	NNT = 6 favouring ECT for short-term response; at baseline, ECT samples had shorter illness duration and higher HAMD scores than rTMS samples. Age has no moderating effect on the differences
Ren et al[46] 2014	9/425	31.0-63.6	NNT = 7 for response; 6 for remission favouring ECT; No significant group difference when continuous change in HAMD scores is considered as outcome; In the absence of psychosis, rTMS as efficacious as ECT; Cognitive domains are better preserved after rTMS than ECT
Micallef-Trigona[49] 2014	9/384	34.0-63.6	Hedges’g = 1.28 for rTMS and 2.15 for ECT. rTMS produces a mean reduction of 9.3 points; ECT produces a mean reduction of 15.42 points on the HDRS
Xie et al[48] 2013	9/395	31.0-63.6	OR = 0.55 for response and 0.49 for remission in favour of ECT; rTMS is better tolerated than ECT (OR = 0.70); rTMS > 1200 stimuli/d is as efficacious as ECT

NNT: Number needed to treat; TMS: Transcranial magnetic stimulation; ECT: Electroconvulsive therapy; HAMD: Hamilton Depression Rating Scale.