Original article 87 The relationship between early changes in the HAMD-17 anxiety/somatization factor items and treatment outcome among depressed outpatients Amy Farabaugh, David Mischoulon, Maurizio Fava, Shirley L. Wu, Alessandra Mascarini, Eliana Tossani and Jonathan E. Alpert The 17-item Hamilton Rating Scale for Depression (HAMD-17) Anxiety/Somatization factor includes six items: Anxiety (psychic), Anxiety (somatic), Somatic Symptoms (gastrointestinal), Somatic Symptoms (general), Hypo- chondriasis and Insight. This study examines the relationship between early changes (defined as those observed between baseline and week 1) in these HAMD-17 Anxiety/Somatization Factor items and treatment outcome among major depressive disorder (MDD) patients who participated in a study comparing the antidepressant efficacy of a standardized extract of hypericum with both placebo and fluoxetine. Following a 1-week, single-blind washout, patients with MDD diagnosed by the Structured Clinical Interview for DSM-IV (SCID) were randomized to 12 weeks of double-blind treatment with hypericum extract (900 mg/day), fluoxetine (20 mg/day) or placebo. The relationship between early changes in HAMD-17 anxiety/ somatization factor items and treatment outcome was assessed separately for patients who received study treatment (hypericum or fluoxetine) versus placebo with a logistic regression method. One hundred and thirty-five patients (female 57%, mean age = 37.3 ± 11.0 years; mean baseline HAMD-17 = 19.7 ± 3.2 years) were randomized to double-blind treatment and were included in the intent-to- treat (ITT) analyses. After adjusting for baseline HAMD-17 scores and for multiple comparisons with the Bonferroni correction, patients who remitted (HAMD-17 score < 8) after study treatment had significantly greater early improvement in Somatic Symptoms (General) scores than non-remitters. No other significant differences in early changes were noted for the remaining items between remitters versus non-remitters who received active treatment. For patients treated with placebo, early change was not predictive of remission for any of the items after Bonferroni correction. In conclusion, the presence of early improvement on the HAMD-17 item concerning fatigue and general somatic symptoms is significantly predictive of achieving remission at endpoint with active study treatment but not with placebo. Int Clin Psychopharmacol 20:87–91  c 2005 Lippincott Williams & Wilkins. International Clinical Psychopharmacology 2005, 20:87–91 Keywords: Depression, Hamilton Rating Scale for Depression (HAMD-17), treatment outcome Depression Clinical and Research Program, Massachusetts General Hospital, Massachusetts, USA. Correspondence and requests for reprints to Amy Farabaugh, Depression Clinical and Research Program, Massachusetts General Hospital, 15 Parkman Street, MA WAC 812, USA. E-mail: afarabaugh@partners.org Received 11 August 2004 Accepted 24 November 2004 Introduction An appreciable percentage of patients receiving anti- depressant treatment for major depressive disorder (MDD) will have a partial or non-response and require more aggressive next-step treatments (Fava and David- son, 1996). Unfortunately, it is often necessary to wait 2–3 months before ascertaining whether a depressive episode will remit on a given treatment (Quitkin et al., 2003). Studies have examined the time to onset of response comparing active treatments to placebo, as well as comparing active treatments to one another. Furthermore, studies have often considered a delayed and persistent response as a true drug response, and an early or nonpersistent response as a placebo response pattern (Stewart et al., 1998). Quitkin et al. (1984) defined significant symptomatic improvement by the end of the first or second week as early onset or a placebo pattern of response. Stassen et al. (1993, 1994) and Nierenberg et al. (1996) have suggested that early improvement is a predictor of later stable response. Hence, it is important to identify factors that may change early in treatment and predict eventual outcome. Some of these factors may be neurobiological (Cook and Leuchter, 2001), while others may be related to clinical features. For example, Bech et al. (2001) completed a meta-analysis on seven clinical trials evaluating the effectiveness of mirtazapine compared to placebo or amitriptyline to determine whether early changes in the Hamilton Rating Scale for Depression (HAMD) item of depressed mood as well as the HAMD depression factor would predict subsequent remission. They found that early changes on 0268-1315  c 2005 Lippincott Williams & Wilkins Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.