Effects of Transcranial Magnetic Stimulation on the Hypothalamic-Pituitary Axis in Depression: Results of a Pilot Study

Some studies have reported that repetitive transcranial magnetic stimulation (rTMS) applied to the dorsolateral prefrontal cortex (DLPFC) is able to induce changes in the hypothalamic-pituitary (HP) axis in subjects with major depression and posttraumatic stress disorder.^1–6 The causes of these neuroendocrine effects, which have been reported for serum cortisol, prolactin, and thyroid hormones, are unknown. On the one hand, as studies have reported changes in the HP axis associated with the alleviation of depressive symptoms,² we can hypothesize that alterations in the HP axis are related to a physiological response that accompanies the therapeutic effect and could thus be considered a biological marker of clinical outcome. On the other hand, rTMS may modulate brain networks implicated in the HP axis, thanks to brain connectivity. In this case, the effects of rTMS on the HP axis could be interpreted as undesirable if they make no contribution to improving the subject’s clinical condition. This is what we previously reported in a subject treated by rTMS for major depressive disorder (MDD) who presented with an increase in plasma thyroid-stimulating hormone (TSH) levels above the normal range without clinical improvement.⁵

In order to clarify the issue, we monitored neuroendocrine levels in subjects with MDD included in a pilot study.^5,7 Our objectives were as listed below:

1. To screen for abnormal levels of HP axis hormones induced by rTMS in these subjects;

2. To explore correlations between abnormal changes in HP axis hormone levels or subclinical endocrine changes and the therapeutic efficacy of rTMS on mood.

Methods

Participants

Subjects older than 18 years and suffering from treatment-resistant depression were enrolled in the pilot study.⁷ To qualify for enrollment, they had to meet DSM-IV criteria for MDD,⁸ have a total score of ≥22 on the 21-item Hamilton Depression Rating Scale (HAM-D), and have failed to respond to a minimum of two courses of antidepressant drugs. Subjects were excluded if they had any diagnosis of bipolar I or II disorder, a history of current substance abuse (except nicotine), and a history of seizures or other neurological conditions. Additional exclusion criteria were clinically significant comorbid disease such as liver, kidney, or heart failure and a pacemaker. Female subjects were excluded if they were pregnant or lactating.

Psychotropic medications were broken off 7 days before the start of the stimulation sessions. The concomitant use of psychoactive drugs was restricted, with the exception of cyamemazine and hydroxyzine as rescue medication for severe anxiety or insomnia.

Stimulation Parameters

The subjects received once-daily rTMS sessions Monday to Friday for two weeks using MagPro×100 (MagVenture A/S, Farum, Denmark) with 75-mm figure-of-8 coils. The frequency of stimulation was 1 Hz, with an intensity at 120% of the resting motor threshold. One rTMS session consisted of six trains of 60 pulses, separated by five intertrains of 30 seconds (360 pulses per day; total duration of one rTMS session: 8 minutes, 30 seconds). Brain stimulations were applied over the right DLPFC (Brodmann’s area 9 or 46).⁷ These stimulation parameters have been found to be effective to treat MDD in previous studies.⁹

Outcome Measures

The HAM-D and Montgomery-Åsberg Depression Rating Scale (MADRS) were used to assess depressive symptoms. A therapeutic response was defined as a HAM-D reduction or MADRS reduction ≥50% relative to the pretreatment baseline.

The first and second blood samples were collected between 9:00 a.m. and 10:00 a.m. We took these blood samples within 15 minutes before and after the first and the 10th rTMS sessions, respectively. The third blood sample was taken 4 weeks later at 9:00 a.m. (Figure 1). Blood samples were centrifuged, and the serum was frozen at –20° for future analysis. Thyroid-stimulating hormone (TSH), free triiodothyronine (fT3), and free thyroxine (fT4) were quantified on a Dimension Vista analyzer with dedicated reagents (Siemens, Saint Denis, France). Cortisol, prolactin, growth hormone (GH), and estrogen were quantified on an Immulite 2000 analyzer with dedicated reagents (Siemens). Testosterone was quantified using a radioimmunoassay.

Data Analysis

To assess the effects of rTMS on the HP axis, we first identified post-rTMS changes in HP axis values up to abnormal values. We then compared HP axis hormone levels after rTMS with baseline values using univariate nonparametric methods based on ranked hormone values (Wilcoxon signed-rank test). This comparison of HP changes was done for all 15 subjects and in a subgroup analysis according to the therapeutic results; p≤0.05 were considered significant.

Ethics

The addition of neuroendocrine monitoring to the pilot study was approved by a human research ethics committee (Committee Protection Persons number II from East of France; registration number: 2009-A00772–55). After complete description of the neuroendocrine monitoring to the subjects of the pilot study, written informed consent was obtained.

Results

Eleven women and four men were enrolled in the pilot study. Mean age was 55 years (SD=9.5; range: 41–71). Average number of hospitalizations for depression was three (SD=2.8), and mean duration of the current MDD was 20 months (SD=2.5). All but one subject was right handed.

Among the 15 subjects included in the pilot study, whatever the therapeutic result, analysis of the blood samples after rTMS revealed no significant change after rTMS in HPA hormones levels above the normal range either in short or medium term. With regard to subclinical endocrine changes induced by rTMS, no significant changes in TSH, fT3, fT4, cortisol, prolactin, estrogen, and testosterone were observed at the end of the rTMS sessions or at 4 weeks thereafter. However, we did find a trend toward a decrease in GH levels just after the 10th session (p=0.0652), but not four weeks later (p=0.3495). By analyzing levels of HP axis hormones according to the therapeutic response assessed using HAM-D scores, we found a significant decrease in GH levels at the end of the stimulation session in nonresponders (p=0.0098) (Table 1). This result tended to parallel the result for GH and MADRS scores (p=0.0547). Using effect size (Cohen's d) to measure the strength of the rTMS effect on GH levels on the basis of HRDS results, we found a large size effect (d=0.86) by comparing pre- and post-rTMS changes in GH levels between nonresponders and responders.

Discussion

Neuroendocrine abnormalities in depression have been reported for over 60 years and include changes in cortisol, thyroid hormones, growth hormone, and prolactin.^10,11 These were identified by elevated basal levels of hormones and metabolites in blood and urine, but also by dynamic functions test of the neuroendocrine system, such as the overnight dexamethasone suppression test (DST). Indeed, some studies found abnormal DST results (nonsuppression of cortisol secretion) in psychotic depression, and such results were thus claimed to be a specific episode-related biological marker for melancholia.^10,12,13 Taken as a whole, these neuroendocrine abnormalities tend to normalize with recovery from depression.¹² Changes in neuroendocrine functions were also found during treatment by electroconvulsive therapy (ECT), and to date this is one of the most robust hypotheses to explain the antidepressant effects of this brain stimulation technique. According to several studies, hypothalamic stimulation occurring during ECT led to a transient release of adrenocorticotropic hormone, cortisol, and prolactin into the blood. However, the magnitude of the changes would not have led to the antidepressant effects of ECT.¹² There is controversy regarding the release of the other HP axis hormones, such as GH and thyroid hormones.^14–18

Compared with ECT data, there are few data on the effect of rTMS on the HP axis, probably because of the relatively recent use of this NIBS technique as a treatment for depression. These data suggest that rTMS is also associated with changes in HP axis hormones, mainly among the hypothalamic pituitary thyroid axis, compared with ECT.^2,4,19,20

Our biological screening provides additional information on HP axis changes following TMS in patients with depression. First, we found no abnormal values immediately, after 10 sessions, or 4 weeks later. This led us to believe that the significant increase in plasma TSH above the normal range found using the same rTMS parameters and previously reported by our staff was probably an “accidental” event.⁵ Secondly, unlike previous studies that showed that rTMS did affect the HP axis, we found no changes in TSH, T4, or prolactin levels.^2–4 The small size of our sample could explain why we could not replicate the previously reported data, even though it was not smaller than in some studies that found a hormone-modulating effect.^2–4 The diverse stimulation parameters (frequency, left-right DLPFC) as well as the various methods used to target the DLPFC could also explain these differences in the results, even though some studies using stimulation parameters similar to ours (1 Hz, right DLPFC) have reported the effects of rTMS on serum prolactin and thyroid hormones.^2,3 The concomitant use of psychotropic medications with rTMS, which varies from one study to another, should also be taken into account to explain this different result, as changes in HP axis can be attributed to brain stimulation only in a medication-free context. Thirdly, we found a significant decrease in GH levels in nonresponders. This result has never been observed, since studies assessing the effects of rTMS on the HP axis did not measure GH levels.^1–4,6 As growth hormone deficiency is known to be associated with depressed mood, and as GH treatment has been associated with significant improvements in depression scales in subjects with GH deficiency, we may hypothesize that a link could exist between the effect of rTMS on GH levels and the therapeutic result. According to our results, the absence of a decrease in GH levels after rTMS among responders might be perceived as a relative increase in the levels of this hormone.²¹ Due to the small numbers of subjects involved in this pilot study, the results concerning GH hormone levels need to be replicated to confirm that rTMS is able to act on depression by way of the HP axis. Another limitation due to the small sample size (four men and 11 women) was that we were, unfortunately, unable to perform statistical analyses using changes in estradiol and testosterone levels according to the therapeutic results, as we did with the other HP axis hormones. Finally, caution must be exercised when linking changes in HP axis hormones with changes of mood when this device is used in subjects with depression. Indeed, changes in serum thyroid levels have been reported in healthy subjects who have undergone rTMS.^19,20

In conclusion, changes in HP axis hormones in subjects with major depression treated by rTMS are unclear. Further studies using larger samples and higher levels of proof (i.e., using sham-controlled conditions) are needed to assess the effects of rTMS on HP axis hormones. Moreover, these studies must include screening for GH in addition to the other HP axis hormones to clarify the mechanisms underlying the interaction between the HP axis and depression treated by rTMS. Better understanding of the effects of TMS on HP axis hormones may also shed light on the relationship between neuronendocrine responses and the therapeutic effect of ECT.