Toshihiko Fukuchi M.D.*,
Kousuke Kanemoto M.D.**

*Fukuchi Clinic
**Aichi Medical University, Dept. Neuropsychiatry

Abstract

We attempted to compare the antidepressant efficacy of milnacipran and fluvoxamine in 202 outpatients with major depression, using the 17-item Hamilton Depression Rating Scale (HDRS). Special attention was paid to the difference of responsiveness as a function of the severity of depression and individual HDRS factors. As a result, while no significant difference between the treatment groups was found overall, a positive response (50% or more decrease in total score from the baseline) was recorded significantly more often with milnacipran than fluvoxamine recipients whose baseline HDRS total score was greater than 19 points. Further, there was a significant difference of response for the 'agitation' and 'insomnia' factors in favor of milnacipran. In both treatment groups, the incidence of adverse events, characteristic of tricyclic antidepressants such as dry mouth, constipation, somnolence, and postural hypotension, was low. While complaints concerning the upper intestinal tract such as epigastric distress were predominant in the fluvoxamine group, urological complications and palpitations were reported only in the milnacipran group. In conclusion, we suggest that milnacipran is preferred to SSRIs for the treatment of depressed patients with agitation as well as severely depressed patients.

Introduction

The introduction of selective serotonin reuptake inhibitors (SSRIs) has dramatically added to the therapeutic strategies possible for patients with depression. Since they offer a drastic alleviation of various upsetting side-effects commonly produced by non-selective interactions with cholinergic-muscarinic, α1-adrenergic, and histamine H1 receptors when used with tricyclic antidepressants (TCAs), a rapid change from TCAs to SSRIs has occurred in many countries. However, especially in severe cases, it has become increasingly apparent that SSRIs may not be as effective as traditional TCAs (1, 10, 11). A so-called fourth generation of antidepressants, such as milnacipran, venlafaxine, and duloxetine, have been developed to selectively inhibit the reuptake of noradrenaline as well as serotonin (24), thus, if the superior effectiveness of TCAs over SSRIs is actually based on their differential influence over noradrenalinergic neurons, theoretically, a therapeutic efficacy similar to TCAs would be expected in selective serotonin and noradrenaline reuptake inhibitors (SNRIs). In Japan, due to the delay in introducing SSRIs, a unique situation has emerged in which SSRIs and an SNRI (milnacipran) have begun to be marketed nearly simultaneously. This event has provided a special opportunity to directly compare the clinical efficacy and effectiveness profiles of both SSRIs and SNRIs. Thus far, most analyses and meta-analyses focusing on milnacipran have come from trials performed by study groups in France (2, 6, 12, 15-17). The present study is an attempt to determine, first, if there are any differences of effectiveness between milnacipran and fluvoxamine as a function of the severity of depression and, second, if the efficacy profiles differ as a function of individual clinical symptoms of depression.

Subjects and methods

All of the study subjects in the milnacipran group were outpatients (n=112) who fulfilled the DSM IV criteria for major depression (American Psychiatric Association, 1994) and had been treated with milnacipran by the first author from April 1999 to February 2001. Exclusion criteria were; serious or uncontrolled medical illness including glaucoma and diabetes mellitus, major risk of suicide, previous history of epileptic seizures and allergic reactions to tri- or tetracyclic antidepressants, pregnancy and lack of contraception in non-menopausal women, organic brain disease, and schizophrenia. As controls, we chose 112 patients out of 241 who fulfilled the above-mentioned inclusion and exclusion criteria, except for being given fluvoxamine instead of milnacipran and who had been treated with fluvoxamine by the first author from April 1999 to February 2001. These were matched in age, sex, and depression severity, and were chosen randomly using a computer. Both groups were studied on an open naturalistic basis and all subjects, including the controls, were evaluated by the first author. The initial choice of therapy (fluvoxamine or milnacipran) was done randomly in principle, because there was no definite clinical evidence regarding differences between the two at that time. Drug dosages were initiated in a range of 30-100 mg and titrated maximally to 150 mg in the milnacipran group. Fluvoxamine was initiated at 50-150 mg and titrated maximally to 250 mg. Thirty-two patients were later excluded because of lost contact immediately after the first visit. Thus, a total of 192 patients were studied, which included 102 patients in the milnacipran group (MIL group) and 90 in the fluvoxamine group (FLU group).
In Table 1, basic patient demographic data are presented. No statistically significant relationship was found between the basic demographic data and drug efficacy as a function of responder and remission ratios by examination with a multinomial logistic regression model. Besides milnacipran and fluvoxamine, fourteen other psychotropic drugs were co-medicated. There were no statistically significant differences or trends between the groups with regard to the other medications. Each patient was fully informed and provided subsequent written consent.
The main parameters for efficacy was the change in total and individual scores on a 17-item HDRS test (Hamilton, 1960) (26) between baseline and end-point, and the patients were examined and assessed every 1 or 2 weeks during the observation period. Treatment was withdrawn prematurely when intolerable side effects or intercurrent serious medical illness emerged newly, psychiatric symptoms worsened, and/or the patient or a family member decided to discontinue the therapy. In such cases, the final assessment was done at that point. Otherwise, the observations continued until the HDRS score remained at a plateau for more than 3 successive weeks after initial improvement or decreased to less than 7 points. Those patients whose total HDRS score decreased by 50% from the baseline were considered as 'responders', while 'remission' was defined as a state in which the total HDRS score was less than 7 points. Further, we defined patients with a baseline HDRS score of more than 19 points as severe. Adverse events were obtained by open questioning, and recorded at the initial interview (baseline) and each subsequent visit.
Statistical analyses were performed using Fisher's exact test and Mann-Whitney's U test. All tests of hypotheses were two-sided. Results were considered statistically significant at a risk level of 0.05.

Results

1) Drop-outs. (Table 2) summarizes the reasons for treatment withdrawal. There were no statistically significant differences between the two groups.
2) Efficacy of treatment (Table 3). There was no significant difference between groups in change of HDRS total score between baseline and end-point. A response (more than 50% decrease in total score from baseline) was recorded for 52.9% of the patients treated with milnacipran and 46.7% of those treated with fluvoxamine. The onset of effect was slightly faster with milnacipran, as the estimated time to 50% response was 3.64 weeks in milnacipran-treated patients and 4.62 weeks in fluvoxamine-treated patients. No statistically significant between-treatment differences were found for responder ratio or onset of effect. However, when analysis was limited to patients with severe depression, there was a statistically significant difference in favor of milnacipran with respect to the change in total HDRS score (p=0.0276). In the same limited sample, 68.9% of those treated with milnacipran and 46.2% of those treated with fluvoxamine were responders, which, as with the change of total score, was statistically significant (p=0.04648). With respect to individual HDRS factors, there were several significant differences between the groups in responder ratio for 'insomnia, initial' (p=0.02093), 'insomnia, middle' (p=0.03201), 'insomnia, delayed' (p=0.01492), 'agitation' (p=0.01839), and 'gastrointestinal, somatic' factors (p=0.03724) (Fisher's exact test). Except for the 'gastrointestinal, somatic' factor, these differences remained statistically significant when analysis was limited to patients with severe depression ['insomnia, initial' (p=0.00733), 'insomnia, middle' (p=0.01311), 'insomnia, delayed' (p=0.00914), 'agitation' (p=0.02613) (Fisher's exact test)].
3) Adverse events. Among 38 adverse events spontaneously reported, representative complaints were selected and are listed In Table 4. Incidences of adverse effects commonly noted during treatment with TCAs, such as constipation, dry mouth, vertigo, and somnolence, were remarkably low in both groups. No significant difference was observed between the treatment groups for either individual or total incidence of adverse events. However, there was a trend (p=0.0608, Fisher's exact test) showing that upper digestive tract symptoms such as nausea, pain and discomfort in epigastric regions, and loss of appetite were relatively frequent in the FLU group (10.7% vs. 22%), while urinary symptoms such as discomfort or difficulty in urination as well as anuria were more prevalent in the MIL group (2.7% vs. 0.0%).

Discussion

There is a discrepancy of opinions about the relative efficacy of TCAs and SSRIs. For the latter, some authors have postulated a better tolerability instead of a weaker antidepressant effect in comparison with TCAs (3,10 ,11,20), while others have found an equal efficacy between the two (7,18,23). However, a meta-analysis of cases whose HDRS scores were relatively high has demonstrated the superiority of TCAs over SSRIs (1).
Recently, evidence has emerged that SNRIs may be as effective as TCAs, even in severely depressed patients. In a series reported by Cole et al.(9), one third of the severely depressed patients, who had not been responsive to either conventional tri- and tetracyclics or to electroconvulsive therapy, remitted with venlafaxine treatment. In the meta-analyses of Puech et al. (19) and Kasper et al. (15), in which most patients were hospitalized at the beginning of treatment, milnacipran was compared with imipramine and clomipramine. Their results showed an equal efficacy between milnacipran and the two TCAs when measured using HDRS scores and/or MADRS (Montgomery-Asberg Depression Rating Scale) scores. Overall, 64% of the milnacipran and 67% of the TCA recipients achieved a greater than 50% decrease in HDRS score. Further, results of other trials, which compared the efficacy of 100 mg/day milnacipran with that of amitriptyline, showed no significant between-treatment differences after 4 weeks (22).
contrast to the relatively abundant reports comparing the efficacy of TCAs vs. SSRIs and vs. SNRIs, direct comparisons between SSRIs and SNRIs are scarce. With regard to the comparison in efficacy between milnacipran and SSRIs, Spencer pointed out that only two published reports are available (2,14). Following a meta-analysis of trials that included unpublished data, Lopez-Ibor et al. (17) suggested that milnacipran at 50 mg twice a day produced significantly greater improvement in HDRS and MADRS scores than SSRIs, and that it was particularly marked when more rigorous criteria, such as the percentage of patients responding or in remission, were used. In the current study, though there were no statistically significant differences or trend of difference between the two groups overall, a statistically significant difference in favor of milnacipran over fluvoxamine was found in moderately or severely depressed patients with an HDRS score greater than 19 points. The advantage of SNRIs over SSRIs in severely depressed patients has already been pointed out by previous authors such as Clerc et al.(5), who postulated that reductions in HDRS scores were greater in venlafaxine than fluoxetine recipients among those with severe depression. Furthermore, a very recent study by Clerc (7) provides additional evidence in support of the advantage of milnacipran over fluvoxamine in 113 patients with moderate to severe major depression. However, no previous study has performed two-step analyses, that is, first with overall patients and second with only more severely depressed patients from among the same sample. Our results, using such analyses, demonstrated that the severity of depression augmented the superiority of milnacipran over the SSRI.
In the current analysis of HDRS factors, significant differences between the treatment groups in favor of milnacipran were found for the three 'insomnia' factors as well as the 'agitation' factor. In a comparison between depressive mania versus agitated depression, Swann et al. (24) failed to confirm any elevation of norepinephrine and its metabolite either in the cerebrospinal fluid or urine of patients with agitated depression, which was in sharp contrast with their results for manic agitation. This suggests that agitation in severe depression cannot be explained by an elevation of norepinephrine, as in the case of depressive mania. Further, Filteau et al. (13) reported that SSRI responders were more anxious-agitated than norepinephrine reuptake inhibitor responders and interpreted this as proof of the preferential efficacy of SSRIs in agitated depression. Although their conclusion apparently contradicts our results, a preferential efficacy of SSRIs does not exclude the possibility that selective reuptake inhibition of serotonin plus norepinephrine could be more effective than that of serotonin alone. However, because previous analyses of HDRS factors have provided inconsistent results (4,14), the questions remain open for future investigation.
In agreement with Ansseau et al. (2), the tolerability profile of milnacipran was not significantly different from that of fluvoxamine in the present study. With both drugs, the incidences of conventional adverse events such as anticholinergic and antihistamine effects were remarkably low. On the other hand, there was a characteristic trend for each drug. As Kasper et al. (15) also noted, urological manifestations occurred at a higher rate in the milnacipran recipients in the current series, whereas upper digestive tract symptoms such as epigastric distress were predominate in the SSRI-treated patients. It should also be mentioned that cardiovascular complaints such as palpitations were reported only in the milnacipran-treated patients (14).
The uniqueness of the current study results lies partially in the method of patient sampling, as both control and study subjects consisted exclusively of outpatients. However, though it is apparent that further amplification is needed before making any firm conclusions, because the patient sampling was done on an open naturalistic basis, this limiting of study subjects to an outpatient population enabled us to perform two-step analyses, In one, we included patients with mild depression and in the other we excluded them, thus revealing the different pharmacological profiles of milnacipran and fluvoxamine as a function of depressive severity. Considering that the overwhelming majority of patients treated by SSRIs or SNRIs are outpatients, the significance of our data is apparent. Because the duration of illness prior to the initial rating would inevitably be prolonged if the study subjects included inpatients, the shortness of duration in the current study can be explained by the lack of inpatients in the study group. We are not able to provide a well-founded explanation for the male preponderance seen in the current study, however, sex-ratio could also vary as a function of inpatient/outpatient ratio. The present results strongly support and supplement the recent data of Clerc (7), indicating that SNRIs are preferred to SSRIs for moderately or severely depressed outpatients with HDRS scores greater than 19 points, and especially for those with marked agitation and insomnia.

Acknowledgements

The authors wish to thank Dr. Susumu Nagai, Miss Mikako Kitazawa, Mr. Takuji Sugimoto, and Miss Akiko Hayashi for their help in statistical analyses of the data and transcribing the raw data into solid formats, along with Mr. Kazutoshi Sugiyama for his advice regarding pharmaceutical issues.

Reference

1) Anderson IM (1998) SSRI versus TCAs in depressed inpatients: a meta-analysis of efficacy and tolerability. Depress. Anxiety 7(Suppl.1):11-17.
2) Ansseau M, Papart P, Troisfontaines B (1994) Controlled comparison of milnacipran and fluoxetine in major depression. Psychopahrmacology 131-137
3) Bowden CL, Schatzberg AF, Rosebaum A et al. (1993) Fluoxetine and desipramine in major depressive disorder. J Clin Psychopharmacol 13:305-310
4) Burnett FE (1998) The clinical efficacy of venlafaxine in the treatment of depression. Rev Contemp Pharmacother 9:303-320
5) Clerc GE, Ruimy P, Verdeau-Pailles J (On behalf of the Venlafaxine French Inpatient Study Group) (1994) A double-blind comparison of venlafaxine and fluoxetine in patients hospitalized for major depression and melancholia. Int Clin Psychopharmacol 9:139-143
6) Clerc GE, Pagot JM, Bouchard JM, Oules J, Guilbert M, Assicot M et al (1990) Therapeutic contribution of milnacipran and of clomipramine during a 3-month treatment: results of a comparative study. Psychiatrie Psychobiologie 5:137-143
7) Clerc GE (2001) Antidepressant efficacy and tolerability of milnacipran, a dual serotonin and noradrenaline reuptake inhibitor: a comparison with fluvoxamine. Int Clin Psychopharmacol, 16: 145-51
8) Cohn CK, Shrivastava R, Mendels J et al. (1990) Double-blind, multicenter comparison of sertraline and amitriptyline in elderly depressed patients. J Clin Psychiatry 51(Suppl B):28-33
9) Cole J, Schatzberg A, Sniffin C et al. (1981) Trazodone in treatment of resistant depression : an open study. J Clin Psychopharmacol 1(Suppl): 45-54
10) Danish University Antidepressant Group (1990) Paroxetine: a selective serotonin reuptake inhibitor showing bettwe tolerance, but weaker antidepressant effect than clomipramine in a controlled multicenter study. J Affect Disord 18:289-299
11) Danish University Antidepressant Group (1986) Citalopram: clinical effect profile in comparison with clomipramine. A controlled multicentre study. Psychopharmacology 90:131-138
12) Dardennes R, Berdeaux G, Lafuma A, Fagnani F (1999) Comparison of the cost-effectiveness of milnacipran (a SNRI) with TCAs and SSRIs: a modeling approach. Eur Psychiatry 14:152-162
13) Filteau MJ, Baruch P, Lapierre YD, Bakish D, Blanchard A (1995) SSRIs in anxious-agitated depression: a post-hoc analysis of 279 patients. Int Clin Psychopharmacol 10:51-54
14) Guelfi JD, Ansseau M, Corruble E, Samuelian JC, Tonelli I, Tournoux A, Pletan Y (1998) A double-blind comparison of the efficacy and safety of milnacipran and fluoxetine in depressed patients. Int Clin Psychopharmacol 13:121-128
15) Kasper S, Pletan Y, Solles A, et al. (1996) Comparative studies with milnacipran and TCAs in the treatment of patients with major depression: a summary of clinical trial results. Int Clin Psychopharmacol 11(Suppl 4):35-39
16) Lecrubier Y, Pletan Y, Solles A, Tournoux A, Magne V (1996) Clinical efficacy of milnacipran. Placebo-controlled trials. Int Clin Psychopharmacol 11(Suppl 4):29-33
17) Lopez-Ibor J, Guelfi JD, Pletan Y, Tornoux A, Prost JF (1996) Milnacipran and selective serotonin reuptake inhibitors in major depression. Int Clin Psychopharmacol 11(Suppl 4):41-46
18) Ottevanger EA (1994) The efficacy of fluoxemine in patients with severe depression. Prog Neuropsychopharmacol Biol Psychiatry 18:731-740
19) Puech A, Montgomery SA, Prost JF et al (1997) Milnacipran, a new serotonin and noradrenaline reuptake inhibitor: an overview of its antidepressant activity and clinical tolerability. Int Clin Psychopharmacol. 12:99-108
20) Roose SP, Glassman AH, Attia E et al. (1994) Comparative efficacy of selective serotonin reuptake inhibitors and tricyclic in the treatment of melancholia. Am J Psychiatry 151:1735-1739
21) Rouillon F, Warner B, Pezous N, Bisserbe JC et al. (2000) Milnacipran efficacy in the prevention of recurrent depression: a 12-month placebo-controlled study. Int Clin Psychopharmacol 15:133-140
22) Spenceer CM (1998) Milnacipran. A review of its use in depression. Drugs 56:405-427
23) Stuppeak CH, Gerestsegger C, Whitworth AB et al. (1994) A multicenter double-blind trial of paroxetine versus amitriptyline in depressed inpatients. J Clin Psychopharmacol 14:241-246
24) Swann AC, Stokes PE, Secunda SK, et al. (1994) Depressive mania versus agitated depression: biogenic amine and hypothalamic-pituitary-adrenocortical function. Biol Psychiatry 35:803-13
25) Wong DT (1993) LY 248686, a new inhibitor of serotonin and norepinephrine uptake. Neuropsychopharmacology 8:23-33
26) Zitman FG, Mennen MFG, Griez E, Hoover C (1990) The different versions of the Hamilton Depression Rating Scale. In Beck P and Coppen A eds. The Hamilton Scales. Springer, Berlin, pp28-34