Vagus nerve stimulation for treatment-resistant mood disorders: a long-term naturalistic study

One hundred-fifty-nine subjects were interviewed through telephone in order to assess eligibility criteria for inclusion into the study; of them, 39 were asked to present for a face-to-face clinical interview. Thirty-three patients were excluded because of not having a current Major Depressive Episode (N = 7), not having a chronic (actual episode ≥2 years) or recurrent (history of at least 4 lifetime MDEs) disorder (N = 20), having a medical illness that contraindicated surgery (N = 4), or having a current Substance Use Disorder (N = 2). One additional patient, fulfilling eligibility criteria, first accepted to undergo surgical procedure, and then refused after the screening visit.

A total of five patients met the inclusion criteria and had the VNS generator implanted. The mean ± SD age of the patients was 56.6 ± 7.3 years (range, 48–66 years). Three patients had a Major Depressive Disorder (60.0%) and the remaining two patients had a Bipolar Disorder (one BD type I and one BD type II). Socio-demographic and clinical data are summarized in Table 1. Concerning the duration of the current MDE, three patients fulfilled DSM-IV criteria for chronic (>2 years) major depression (all three had a diagnosis of unipolar major depression). Pre-implant history of the two BD patients was carefully recorded by means of direct interview, family members’ interview (when available) and medical records review. The first patient (Figure 1) had a history of BD type I with onset at age 26; he had more than 10 lifetime MDEs, with 8 admissions to psychiatric wards and 2 suicide attempts during previous episodes. Despite being on mood stabilizer (valproic acid), during the last two years preceding VNS implant he showed 3 MDEs (not considering the current one) without (hypo) manic episodes, but with intervals between depressive episodes lasting less than 3 months. Treatments with adjunctive lamotrigine or quetiapine did not result in stable remission; adjunctive antidepressants in the current episode did not resolve depressive symptomatology. The second patient (Figure 2) had BD type II, with onset at age 36. He also had more than 10 lifetime MDEs, a history of 1 prior suicide attempt, and had, during the last two years prior to VNS implant, 2 long-lasting MDEs (not considering the current one). He spent, during the last two years prior to surgery, a total of 5 months only without a MDE (although without complete remission), despite being treated with valproic acid, quetiapine, lamotrigine, and four different antidepressants (two of them during the current MDE). No hypomanic episodes were recorded in the last two years.

All patients had the VNS generator implanted but for the long-term follow-up, four patients (80.0%) were evaluable at 24 months and three (60.0%) were evaluable at 60 months. Two patients withdrew from the study: one subject (Figure 3), whose depression slightly improved (HDRS-17: 14), withdrew after 18 months because of side effects judged by the patient as intolerable (hoarseness, sore throat and neck pain), and had the VNS generator switched off. The other (Figure 2) withdrew after 12 months because symptoms remained unchanged (HDRS-17: 21); although the clinician in charge of that patient clearly advised him that VNS efficacy sometimes is evident after 12 months of treatment, he was unwilling to attend follow-up assessments and dropped out from the study.

Table 2 shows results of the repeated measures ANOVA on the total scores of the HDRS, MADRS, CGI-S and SF-36 scales (all not statistically significant). Figures 1, 2, 3, 4 and 5 show individual scores for each patient and stimulation parameters, as patients had different outcomes over the 5-year follow-up. According to the HDRS-17, 2 patients were responders (40%) and remitters (40%) after 1 year of VNS treatment (patients 2 and 4). At 2 years, the three patients still on treatment were all responders and remitters. Two of the three patients who were followed-up for 5 years had one depressive recurrence during the 4th year and the third subject showed no recurrences for the whole period.

Table 3 summarizes the rates of adverse events (AEs) after VNS implant and during the follow-up period. With regard to tolerability, patients reported common side effects traditionally observed in previous clinical studies. The only side effect related to the surgical procedure was neck pain located at the level of the surgical incision, which generally disappeared within 2 weeks from implant. Only in one case the pain persisted up to month 18. Stimulation-related adverse effects experienced by patients were hoarseness, discomfort, sore throat, headache, alteration of voice, and described as mild to moderate. In one patient, hoarseness, sore throat and neck pain were described as severe and intolerable, and the patient asked to have the VNS generator switched off at 1 year; after 5 months we tried again to switch on the generator, but she reported again intolerable side effects and 18 months after the implant she asked to have the generator switched off definitely. Pulse-width was not reduced, although this strategy is commonly used to improve tolerability. No (hypo) manic episodes, a rare but yet documented adverse event potentially occurring with VNS, were observed during the 60 months of stimulation.

Our study shows that VNS effectiveness in a clinical practice devoted to the care of individuals with TRD was comparable to the efficacy outcomes reported from prior controlled and uncontrolled clinical trials: the 1-year response and remission rates (based upon the HDRS-17) were 40% and 40% respectively in our case series. Our response rate at 1 year is comparable to the 31.8% response rate (23.2 to 41.8%) after a mean of 20 weeks found in a recent meta-analytical study [38].

To our knowledge, the present study is one of the few prospective ones with a long (5-year for three patients) follow-up of patients who underwent VNS implant for TRD. Findings from our case series indicate a progressive symptom reduction over the first year, although statistical analyses were not significant; moreover, the analysis of individual rating scales scores over time shows that VNS, when effective (not in all patients), is associated with a reduction of the number of depressive recurrences. As already shown by previous reports [21,22,36,39], the present 5-year follow-up study suggests that VNS may prove more useful as a long-term maintenance therapy for chronic depression rather than for acute stabilization of an episode. Preliminary evidence suggests that adjunctive VNS has two principal effects: first, it reduces depressive symptoms in a small but significant proportion of patients who otherwise would be unresponsive to standard treatments, and, second, it reduces exacerbations over the long-term for the vast majority of patients who responded to VNS addition (sustained response) [40]. This long-term effect, however, appears from naturalistic studies or longitudinal pivotal trials without a control or comparison group [22,26,28,36]; long-term observations, being very few in the literature, also small case series like the present one, may be then of interest to clinicians.

Caution has to be used in interpreting our results, as we observed no statistically significant changes in HDRS or MADRS scores (nor in SF-36 scores) in the first 12 months of treatment (see Table 2); moreover, two patients dropped out by 12–18 months. Conclusions about long-term effectiveness (and/or tolerability), being based on the three subjects who remained on treatment, have to be mitigated on the basis of these limitations.

VNS is not a procedure devoid of side effects; one patient from our case series was unable to tolerate hoarseness, sore throat and neck pain, and the VNS generator had to be switched off. Moreover, the majority of patients reported stimulation-related adverse effects, although to a lesser degree. A careful screening of subjects with TRD is then indicated, with particular attention not only to the history of resistant depression (number of previous failed antidepressant trials, compliance, etc.) but also to the history of side effects of previous treatments and to the ability of the patient to tolerate such side effects. Unfortunately, we did not change stimulation settings (e.g. reduced pulse-width) during the treatment phase in order to counterbalance reported adverse effects; we then acknowledge this as a limitation to be kept in mind when interpreting our results concerning side effects of VNS.

One patient refused to attend follow-up assessments after 12 months of treatment, being depressive symptoms unchanged. We acknowledge that a 12-month VNS treatment may be too short to conclude that this procedure is ineffective. Despite several efforts by the psychiatrist in charge of that patient to maintain him in the study, he withdrew the consent and dropped out. Although it is unusual for someone to consent to having a neurosurgical procedure and then refuse to be followed up, he did so and told the psychiatrist that he was willing to try transcranial magnetic stimulation at another University hospital. This case underlines, to our opinion, that a careful screening of subjects with treatment-resistant mood disorders is compulsory before implementing neurosurgical procedures, with a particular attention to motivations to attend complex and expensive procedures such as VNS implantation.

Our study, even though limited by its small sample size and by its observational nature, shows that VNS may be effective in clinical practice in a small but significant minority of patients. The investigation of potential predictors of VNS response is strongly needed and future studies in larger samples should focus on this matter. Further long-term follow-up studies are also strongly needed in order to confirm the putative stabilizing property of VNS.