ORIGINAL ARTICLE Partial Validation of the Atypical Features Subtype of Major Depressive Disorder

Michael A. Posternak, MD; Mark Zimmerman, MD

Background: Symptoms of the atypical features sub- semistructured interviews. Using the available litera- type of major depressive disorder include mood reactiv- ture, we made a series of a priori hypotheses regarding ity, , hyperphagia, leaden paralysis, and re- how depressed patients with atypical features (n=130) jection sensitivity. This subtype was introduced into the would differ from those without atypical features (n=449). mood disorders section of the DSM-IV following a series In addition, we tested the strength of the associations be- of trials showing that such patients re- tween each of the 5 atypical symptoms. sponded preferentially to monoamine oxidase inhibi- tors. Studies aimed at validating the atypical features sub- Results: Although many of the predicted hypotheses were type have yielded inconsistent results. Our study sought substantiated, an equal number were not. Correlation to reevaluate the validity of atypical by ex- analyses revealed modest associations between several of amining the demographic and clinical features of a large the atypical symptoms, but mood reactivity was not as- cohort of depressed patients who met DSM-IV criteria for sociated with any of the other symptom criteria. atypical features. Conclusion: Our results provide partial support for the Methods: We evaluated 579 psychiatric outpatients with validity of the atypical features subtype of major depres- a current diagnosis of major depressive disorder for the sive disorder. presence of atypical features. Detailed demographic and clinical information was obtained for each patient through Arch Gen . 2002;59:70-76

TYPICAL depression stands included patients with ; the alone among the DSM-IV1 other, patients with hysteroid . subtypes The latter was characteristic of as being born out of the histrionic patients, whose mood was de- modern psychopharmaco- scribed as “shallow” and excessively sensi- logicA revolution. West and Dally2 first used tive to both admiration (mood reactive) and the qualifier atypical to characterize a co- rejection (rejection sensitive). When de- hort of depressed patients who appeared pressed, such patients purportedly dis- phobic, “overreactive,” and “hysterical” played a propensity to oversleep and over- and exhibited prominent fatigue, re- eat. Although the validity of hysteroid versed diurnal variation, initial insom- dysphoria was questioned by some,5,6 it was nia, and an absence of decreased appe- ultimately incorporated into the DSM-IV as tite. Sargant3 added that such patients also a depressive subtype by virtue of the pref- tended to complain of severe lethargy, hy- erential response such patients showed for persomnia, and irritability. Both sets of in- MAOIs rather than tricyclic antidepres- vestigators believed that these patients had sants. In its current form, the atypical sub- good premorbid personalities but were type includes 5 features: mood reactivity now experiencing a chronic form of de- plus at least 2 of the following 4 symptoms pression. They also noted that such pa- (hereafter referred to as atypical B symp- tients responded particularly well to mono- toms): hypersomnia, hyperphagia, severe amine oxidase inhibitors (MAOIs) and less lethargy (often described as a feeling of From the Department of well to tricyclic and elec- “leaden paralysis”), and a pathological sen- Psychiatry and Human troconvulsive therapy. sitivity to rejection and criticism. 4 Behavior, Brown University Quitkin et al subsequently argued that The decision to include the atypical School of Medicine, Rhode these early investigators were actually de- features subtype in the DSM-IV was con- Island Hospital, Providence. scribing 2 distinct subtypes. One subtype troversial,7 in part because studies that have

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 70

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 SUBJECTS AND METHODS 0.78; and rejection sensitivity, 0.75, indicating excellent in- terrater reliability. The SCID was supplemented with questions from the SUBJECTS Schedule for Affective Disorders and (SADS)37 to assess the severity of symptoms during the week prior to All subjects in our study were recruited from the Rhode the evaluation. From this, we were able to obtain extracted Island Hospital (Providence) Department of Psychiatry’s 21-item Hamilton Depression Rating Scale (HAM-D) scores outpatient practice. During their initial telephone screen, following the algorithm developed by Endicott et al.38 Cur- all patients were invited to participate in an in-depth di- rent social-functioning ratings were obtained using the SADS agnostic evaluation prior to meeting with their treating cli- item that rates the highest level of social relations during nician (psychiatrist, psychologist, or social worker). Only the last 5 years. Baseline Clinical Global Impression– non-English-speaking patients and those with evidence of Severity (CGI-S)39 ratings and current Global Assessment of cognitive impairment were excluded from the study. To date, Functioning (GAF)40 scores were also obtained for each pa- 1130 patients have been evaluated. Of these, 579 (51.2%) tient by the diagnostic interviewer. as- were diagnosed as having a current major depressive dis- sessments were incorporated into the protocol for only the order and form the cohort of interest for the study. The last 530 patients; the first 600 patients did not undergo an Rhode Island Hospital institutional review board ap- Axis II diagnostic evaluation. Thus, PD diagnoses were avail- proved the research protocol, and all patients provided in- able for only 262 (45.3%) of the 579 patients with major formed written consent. depression. Personality disorders were assessed using the Structured Interview for DSM-IV Personality.41 METHODS All diagnoses were made according to DSM-IV crite- ria. Our analyses comparing depressed patients with and Consenting patients were interviewed at baseline using the without atypical features are based on current rather than Structured Clinical Interview for DSM-IV (SCID).35 Diag- lifetime diagnoses of affective disorders. In comparing co- nostic raters were PhD psychologists or college graduate morbidity rates, however, we used lifetime diagnoses. Psy- research assistants who had undergone extensive train- chomotor retardation was rated by interviewer observa- ing, as described elsewhere.36 tion rather than patient report. Prior suicidal behavior was Mood reactivity was assessed according to the SCID by analyzed based on the most serious lifetime attempt. The asking patients whether they felt better, even temporarily, determination of seriousness was made after assessing the when something good happened. Hyperphagia was defined method and purpose of the attempt, the likelihood of res- as increased appetite nearly every day for at least 2 weeks or cue, and the seriousness of injury. an increase in body weight of 5% or more. Hypersomnia was defined as sleeping significantly more than usual. Leaden pa- STATISTICAL ANALYSES ralysis was considered to be present when a patient acknowl- edged often having a heavy, leaden feeling in the arms or legs. We performed ␹2 and t tests to analyze all categorical and Rejection sensitivity was defined as a long-standing pattern continuous variables, respectively. Correlation coeffi- of interpersonal sensitivity (not limited to episodes of mood cients were obtained using the Spearman ␳ because all vari- disturbance) that caused significant social or occupational im- ables tested were dichotomous. For all hypothesized com- pairment. All 5 symptoms were rated as being either pre- parisons, statistical significance was set at PϽ.05, and all sent, absent, or subthreshold, and only the first rating counted tests were 2-tailed. In addition, 34 tests were performed with- as being present. Reliability ratings for each of the atypical out an a priori hypothesis. For these tests, the Bonferroni symptoms were obtained from 24 joint interviews. ␬ Values correction was used to adjust for chance positive findings. for the atypical symptoms were as follows: mood reactivity, Statistical significance in these analyses was set at PϽ.05 0.83; hyperphagia, 1.0; hypersomnia, 0.90; leaden paralysis, divided by 34, or PϽ.0015.

examined the demographic and clinical features of pa- In our study, we sought to assess and reevaluate the tients with atypical depression have frequently yielded con- validity of the atypical subtype by comparing the demo- tradictory results. For example, the atypical features sub- graphic and clinical features of depressed patients with type has been associated with female sex by some and without atypical features. Our analyses are based on investigators8-11 but not others,12-14 younger age by some9,12,15 data collected from the Rhode Island Methods to Im- but not others,8,13,14,16 bipolarity by some17,18 but not oth- prove Diagnostic Assessment and Services (MIDAS) ers,13 greater anxiety by some9,12,19-21 but not oth- project,27 which has overcome many of the limitations ers,11,14,16,22 less severity by some11,22 but not others,8,13,23,24 of previous studies by (1) evaluating a large cohort of un- a longer duration of illness by some8,22,25 but not others,13 selected psychiatric outpatients; (2) abiding by the a younger age of onset by some12,17,22,26 but not others,7,23 current DSM-IV formulation of atypical depression; (3) higher rates of recurrence by some12,13 but not others,22 and using semistructured interviews conducted by research- more suicidality by some,12 but less suicidality by oth- ers with extensive training to obtain diagnoses on most ers.14-22 Integrating these results has been difficult because major Axis I and Axis II disorders; and (4) obtaining ex- investigators have used disparate criteria to define the atypi- tensive information regarding baseline demographic and cal subtype and have often studied distinct and highly se- clinical features. Our study uses the results from the first lected populations of subjects. 579 patients who were evaluated in the MIDAS project

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 71

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 social , , bulimia, and body dys- Table 1. Comparison of Demographic and Clinical Features morphic disorder; (8) amphetamine abuse or depen- of Depressed Patients With and Without Atypical Features* dence; and (9) histrionic and avoidant personality dis- orders (PDs). In addition, we predicted that reversed Atypical Nonatypical (n=3) (n = 449) P diurnal variation, an absence of psychomotor retarda- tion, an absence of guilt, irritability, and psychic anxi- Female, No. (%) 101 (77.7) 286 (63.7) .003† ety would be more common in this cohort. We also pre- Age, mean ± SD, y 37.8 ± 11.0 39.3 ± 12.2 .21‡ Race: white, No. (%) 112 (86.2) 391 (87.1) .78 dicted that patients with atypical depression would have Diagnosis, No. (%) .90‡ more lifetime nonserious suicide attempts and fewer se- Unipolar depression 120 (92.3) 416 (92.7) rious attempts compared with patients who had non- Bipolar I, depressed 4 (3.1) 11 (2.4) atypical depression. Bipolar II, depressed 6 (4.6) 22 (4.9) To test the internal consistency of the criteria set, we Marital status, No. (%) .16 hypothesized that patients who maintained mood reac- Single 34 (26.2) 119 (26.5) Married/living together 58 (44.6) 219 (48.8) tivity would more frequently meet the atypical B criteria Divorced/separated 38 (29.2) 101 (22.5) threshold than those with an unreactive mood. We also Widowed 0 (0.0) 10 (2.2) expected that each of the 5 atypical symptoms would be Level of education, No. (%) .07 significantly correlated with each other, and that hyper- Less than high school diploma 23 (17.7) 49 (10.9) somnia and hyperphagia would be more strongly corre- High school graduate 36 (27.7) 113 (25.2) lated with the other atypical symptoms than would the non- At least some college 71 (54.6) 287 (63.9) atypical symptoms of and decreased appetite. Current social functioning, No. (%) .08 Poor 25 (19.2) 58 (12.9) Last, because some investigators have argued that Fair 30 (23.1) 86 (19.2) mood reactivity plus 1 of 4 atypical B symptoms, or prob- Good 75 (57.7) 305 (67.9) able atypical depression, may better capture the unique Prior psychiatric hospitalization 33 (25.4) 118 (26.3) .87 features of patients who respond preferentially to Severity, mean ± SD MAOIs,7,9,28,29,33,34 we sought to determine whether such a Ham-D score 23.4 ±7.0 21.8 ± 6.6 .02§ modification would be supported by our data. If patients CGI-S score 3.23 ± 0.64 3.08 ± 0.73 .04§ Age of onset, mean ± SD, y 23.1 ± 11.9 26.9 ± 14.3 .003† with probable atypical depression resemble those who meet Duration, wk threshold criteria, we would expect the demographic and Mean ± SD 349 ± 726 205 ± 414 .03† clinical features of these 2 cohorts to be similar. Median 104 40 .004† First episode, No. (%) 54 (41.5) 209 (46.5) .31 No. of episodes, mean ± SD 6.6 ± 17.6 8.0 ± 21.5 .51 RESULTS Suicidal ideation, No. (%) 67 (51.5) 218 (48.6) .55 History of nonserious suicide attempts 11 (8.5) 49 (10.9) .42‡ PREVALENCE OF ATYPICAL FEATURES History of serious suicide attempts 26 (20.0) 74 (16.5) .35‡ IN MAJOR DEPRESSION Current GAF score, mean ± SD 47.3 ± 8.9 49.8 ± 9.4 .006࿣ Of the 579 patients with a current major depressive dis- *N = 579. Ham-D indicates Hamilton Depression Rating Scale; order, 130 (22.5%) met criteria for the atypical subtype. CGI-S, Clinical Global Impression–Severity; and GAF, Global Assessment of Functioning. Rates of atypical symptoms in the 579 patients were as †Statistically significant in the hypothesized direction. follows: mood reactivity, 71.7%; hypersomnia, 16.8%; hy- ‡Contrary to hypothesis, not found to be statistically significant. perphagia, 21.8%; leaden paralysis, 28.0%; and rejec- §Statistically significant in the direction opposite to what was tion sensitivity, 40.9%. Symptom rates in the 130 pa- hypothesized. ࿣Not statistically significant after adjusting for the Bonferroni correction. tients with atypical depression were 100% for mood reactivity, 36.2% for hypersomnia, 53.1% for hyperpha- gia, 60.8% for leaden paralysis, and 75.4% for rejection and who received a current diagnosis of major depres- sensitivity. sive disorder. Because of the inconsistencies in the results of pre- DEMOGRAPHIC AND CLINICAL FEATURES vious studies, there is no straightforward way to gener- ASSOCIATED WITH ATYPICAL DEPRESSION ate hypotheses regarding how patients with atypical fea- tures would be expected to differ from those without The demographic and clinical features of the 130 pa- atypical features. In reviewing the literature, therefore, tients with atypical depression were compared with we have given more weight to the Columbia University those of the 449 patients with nonatypical depression (New York, NY) group’s formulation of the disorder be- (Table 1). As predicted, atypical depression was asso- cause this group developed the current set of diagnostic ciated with female sex, a younger age at onset, and a longer criteria and demonstrated the preferential MAOI re- episode duration. Contrary to expectation, patients with sponse pattern.28-32 We tested the hypotheses that de- atypical features were not younger, the diagnosis was not pressed patients with atypical features would be more more prevalent in patients with , and likely than nonatypical depressed patients to exhibit the atypical depression was associated with a greater rather following characteristics: (1) female sex; (2) younger age; than lesser severity of illness, both on HAM-D scores and (3) a longer episode duration; (4) a younger age at on- CGI-S ratings. set; (5) higher rates of bipolarity; (6) milder illness; (7) Because this latter finding is contrary to the estab- greater comorbidity with panic attacks, , lished opinion regarding atypical depression, we fur-

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 72

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 Table 2. Lifetime Comorbidity Rates in Depressed Patients Table 3. DSM-IV Personality Disorder Dimensional Scores With and Without Atypical Features* in Depressed Patients With and Without Atypical Features*

Atypical Nonatypical Atypical Nonatypical (n = 130) (n = 449) P (n = 54) (n = 208) P Panic disorder without agoraphobia 8 (6.2) 27 (6.0) .95† Paranoid 1.2 ± 1.7 0.86 ± 1.2 .12 Panic disorder with agoraphobia 41 (31.5) 87 (19.4) .003‡ Schizotypal 1.3 ± 1.8 0.95 ± 1.5 .15 24 (18.5) 63 (14.0) .21 Schizoid .98 ± 1.2 0.73 ± 1.1 .14 Social phobia 71 (54.6) 145 (32.3) Ͻ.001‡ Borderline 2.8 ± 2.4 1.7 ± 2.0 .003† Obsessive-compulsive disorder 18 (13.8) 57 (12.7) .73 Narcissistic 1.3 ± 1.4 0.78 ± 1.2 .003† Posttraumatic disorder 40 (30.8) 116 (25.8) .26 Antisocial .46 ± .95 0.39 ± 0.93 .63 Generalized 26 (20.0) 69 (15.4) .21 Histrionic 1.0 ± 1.2 0.65 ± 1.1 .04‡ Hypochondriasis 7 (5.4) 7 (1.6) .01‡ OCPD 1.9 ± 1.5 1.6 ± 1.4 .10 Bulimia 8 (6.2) 12 (2.7) .06† Dependent 1.3 ± 1.3 .84 ± 1.3 .03† 9 (6.9) 8 (1.8) .002‡ Avoidant 3.0 ± 2.3 1.2 ± 1.9 .001‡ Drug or alcohol abuse/dependence 54 (41.5) 194 (43.2) .74 Amphetamine abuse/dependence 5 (3.8) 19 (4.2) .85† *N = 262. Data are presented as mean ± SD unless otherwise indicated. OCPD indicates obsessive-compulsive personality disorder. *N = 579. Patients could be diagnosed as having Ͼ1 comorbid disorder. †Not statistically significant after adjusting for the Bonferroni correction. Data are presented as number (percentage) unless otherwise indicated. ‡Statistically significant in the hypothesized direction. †Contrary to hypothesis, not found to be statistically significant. ‡Statistically significant in the hypothesized direction. phobia but were not more likely to have panic disorder without agoraphobia (Table 2). No differences were found ther explored the relationship between severity and atypi- in rates of other anxiety disorders or substance use disor- cality. With a worsening of mood, as rated on the 6-point ders, including amphetamine abuse or dependence. Life- ␹2 mood severity item from the SADS, rates of hypersom- time diagnoses of hypochondriasis (5.4% vs 1.6%; 1=6.2; nia (r=0.11; P=.01), leaden paralysis (r=0.21; PϽ.001), P=.01) and body dysmorphic disorder (6.9% vs 1.8%; ␹2 and rejection sensitivity (r=0.13; P=.009) all increased. 1=9.4; P=.002) were more common in patients with The percentage of patients who met criteria for atypical atypical depression; bulimia was not found to be more com- ␹2 features also increased with greater mood severity but not mon (6.2% vs 2.7%; 1=3.7; P=.06). to a statistically significant degree (r=0.07; P=.08), largely Of the 579 depressed patients in our sample, 262 because mood reactivity was inversely correlated with underwent an Axis II evaluation. A dimensional analy- mood severity (r=−0.18; PϽ.001). sis, which compares the number of DSM-IV criteria met for each PD, confirmed that histrionic and avoidant traits PRESENCE OF SYMPTOMS HYPOTHESIZED TO were associated with atypical depression (F261=4.2; P=.04, Ͻ BE ASSOCIATED WITH ATYPICAL DEPRESSION and F261=35.2; P .001, respectively). In fact, patients with atypical depression had higher mean dimensional rat- In comparing patients with atypical depression with those ing scores for each of the 10 PDs than patients with non- who had nonatypical depression, the former cohort was atypical depression (Table 3). ␹2 more often rated as irritable (72.3% vs 62.1%; 1=4.5; From a categorical standpoint, 100 (38.2%) of the ␹2 P=.03) and anxious (66.9% vs 55.0%; 1=5.9; P=.02). 262 depressed patients met the diagnostic threshold for However, patients with atypical depression were no more at least 1 comorbid PD. Avoidant PD (but not histrionic likely to display reversed diurnal variation (26.2% vs PD) was significantly associated with a greater likeli- ␹2 ␹2 Ͻ 30.5%; 1=0.9; P=.34), less likely to demonstrate an ab- hood of having atypical features ( 1=20.1; P .001). No ␹2 sence of guilt (34.6% vs 45.4%; 1=4.8; P=.03), and more other PDs were associated with the atypical subtype, al- often rated as having (40.0% though some of the sample sizes were small (Table 4). ␹2 vs 28.1%; 1=6.7; P=.009). Rates of current suicidal ideation were nearly iden- CORRELATIONS AMONG ATYPICAL SYMPTOMS tical in both cohorts (51.5% and 48.6%). Of the de- pressed patients with atypical features, 20% had a his- We first evaluated whether the presence of mood reactiv- tory of a serious suicide attempt compared with 16.5% ity was significantly associated with having at least 2 of 4 of nonatypical depressed patients; rates of nonserious sui- atypical B symptoms. Of the 579 patients diagnosed as hav- cide attempts were likewise similar between the 2 groups ing a current major depressive disorder, 415 (71.7%) were (8.5% and 10.9%, respectively). Thus, contrary to ex- mood reactive and 164 (28.3%) were mood nonreactive. pectation, patients with atypical depression were nei- Of the mood-reactive patients, 130 (31.3%) met the atypi- ther more prone to nonserious suicide attempts nor less cal B criteria; 53 (32.3%) of the mood-nonreactive pa- prone to serious ones. tients also met the criteria. Thus, mood reactivity was not associated with a greater likelihood of meeting the atypi- COMORBIDITY IN PATIENTS cal B diagnostic threshold. WITH ATYPICAL DEPRESSION Next, we assessed the strength of association be- tween the 5 atypical symptoms (Table 5). Mood reac- Patients with atypical depression were significantly more tivity was not correlated with any of the atypical B symp- likely to have panic disorder with agoraphobia and social toms. Hyperphagia was significantly associated with

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 73

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 hypersomnia (r=0.09; P=.03) as well as leaden paraly- validity compared with the nonatypical symptoms of de- sis (r=0.10; P=.02), and leaden paralysis was signifi- creased appetite and insomnia. cantly associated with rejection sensitivity (r=0.09; P=.03). No other symptom domains were significantly CATEGORIZING PATIENTS WITH PROBABLE correlated. ATYPICAL DEPRESSION To assess the discriminant validity of the atypical subtype, we compared the strength of associations be- Probable atypical depression is defined as mood reactiv- tween hyperphagia and hypersomnia with the remain- ity plus exactly 1 of 4 atypical B symptoms. Of the 449 ing 3 atypical symptoms and contrasted these with the patients who did not meet the full criteria for atypical de- strength of associations between the nonatypical symp- pression, 141 (31.5%) met the probable criteria for this toms of decreased appetite and insomnia. In making these disorder. To determine whether these patients more closely comparisons, only the previously mentioned associa- resembled those who met the full criteria or those with tions reached statistical significance. Hyperphagia was nonatypical depression, we performed the following analy- positively associated with each of the other 4 atypical sis: we compared the 141 patients with probable atypical symptoms (r=0.04 to 0.10), whereas decreased appetite depression with both the 130 patients who had atypical was negatively correlated with them (r=−0.04 to 0.00). depression and the 308 patients with nonatypical depres- Hypersomnia was more strongly correlated with hyper- sion across each demographic and clinical variable. Pa- phagia (r=0.09) and rejection sensitivity (r=0.02) than tients with probable atypical depression differed signifi- was insomnia (r=−0.02 for both); however, leaden pa- cantly from those with threshold criteria on 7 variables. ralysis was more strongly correlated with insomnia They were less severely ill; had a shorter episode dura- (r=0.06) than with hypersomnia (r=0.01). These re- tion; had higher rates of comorbid social phobia; had lower sults suggest a positive but modest correlation between rates of borderline, narcissistic, and avoidant traits; and the atypical B symptoms, and fairly good discriminant had lower rates of avoidant PD. In contrast, patients with probable atypical depression differed significantly from pa- tients with nonatypical depression on only 1 variable: they Table 4. Comorbidity Rates in Depressed Patients were more likely to be female. Thus, our results suggest With and Without Atypical Features* that patients with probable atypical depression are dis- tinct from those who meet threshold criteria. Atypical Nonatypical (n = 54) (n = 208) P COMMENT Paranoid 5 (9.3) 11 (5.3) .28 Schizotypal 4 (7.4) 6 (2.9) .12 The ideal method for introducing a new diagnosis into the Schizoid 4 (7.4) 6 (2.9) .12 Borderline 10 (18.5) 24 (11.5) .17 official nomenclature is to postulate its features by draw- Narcissistic 1 (1.9) 4 (1.9) .97 ing from clinical experience, establish its boundaries Antisocial 0 (0) 5 (2.4) .25 through empirical research, test its reliability, and con- Histrionic 1 (1.9) 3 (1.4) .83 firm its validity. Rarely have psychiatric diagnoses fol- OCPD 6 (11.1) 21 (10.1) .83 lowed this route, and atypical depression is no exception. Dependent 1 (1.9) 4 (1.9) .97 This condition is somewhat unique, however, because it Ͻ Avoidant 23 (42.6) 31 (14.9) .001† evolved largely from the results of psychopharmacology Cluster A 9 (16.7) 20 (9.6) .14 Cluster B 11 (20.4) 29 (13.9) .24 trials. Previous studies attempting to validate the sub- Cluster C 24 (44.4) 50 (24.0) .003‡ type by more traditional means had methodological short- Any PD 28 (51.9) 72 (34.6) .03‡ comings and consequently have yielded inconsistent re- sults. Because our hypotheses were based on this incon- *N = 262. Data are presented as number (percentage) unless otherwise sistent literature, it would have been unrealistic to expect indicated. OCPD indicates obsessive-compulsive personality disorder; that all of the hypotheses would be confirmed. The de- PD, personality disorder. †Statistically significant in the hypothesized direction. gree to which our results validate the subtype is therefore ‡Not statistically significant after adjusting for the Bonferroni correction. a matter of judgment.

Table 5. Correlations Among Atypical Symptoms and Comparisons With the Nonatypical Symptoms of Decreased Appetite and Decreased Sleep*

Mood Reactivity Hyperphagia Hypersomnia Leaden Paralysis Rejection Sensitivity Decreased Appetite Decreased Sleep Mood reactivity ...... Hyperphagia .04 ...... Hypersomnia −.05 .09† ...... Leaden paralysis −.02 .10† .01 ...... Rejection sensitivity −.08 .06 .02 .09† ...... Decreased appetite −.04 −.43‡ −.04 .00 −.04 ...... Decreased sleep −.01 −.02 −.32‡ .06 −.02 .13‡ . . .

*N = 579. Ellipses indicate not applicable. †PϽ.05. ‡PϽ.01.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 74

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 Our most salient finding regards mood reactivity. ciated with atypical depression is consistent with previ- Mood reactivity has been considered an essential com- ous studies.21,24 One possible explanation for this find- ponent of the atypical features subtype; of the 5 atypical ing, however, is that rejection sensitivity may be closely symptoms, it is the only one required to make the diag- related to avoidant traits. If so, the comorbidity findings nosis. Consistent with other reports,11,14,22,32,42,43 we did could simply be a consequence of the overlapping na- not find any evidence to suggest that mood reactivity is ture of these constructs; in fact, all of the comorbid con- associated with the atypical B symptoms. In some anti- ditions associated with the atypical features subtype ap- depressant trials, the presence of mood reactivity has been pear to have this phobic-hyperconscious element in shown to predict a preferential MAOI response,33,44 common. To determine whether the other atypical symp- whereas other studies have found that it is not predic- toms were independently associated with these disor- tive34 or that mood unreactivity predicts a superior MAOI ders, we reanalyzed our data set using a modified defi- response.19,22,45 nition of atypical depression that required mood reactivity Of the remaining atypical symptoms, correlation plus 2 of 3 atypical B symptoms (excluding rejection analyses revealed significant but modest associations. Al- sensitivity). Of all the validators we assessed in this study, though correlation coefficients of 0.09 to 0.10 account only 1 remained significantly associated with the atypi- for only about 1% of the variance, the association be- cal subtype: being female. This suggests that the comor- tween insomnia and decreased appetite (r=0.13) was only bidity findings may be a consequence of the fact that re- slightly higher. This suggests that this level of correla- jection sensitivity is an element of most, if not all, of these tion may be clinically meaningful. Our analyses lend sup- disorders. port to the discriminant validity of the subtype because Two limitations of our study should be kept in mind. hyperphagia and hypersomnia were generally more closely First, although multiple raters were used to interview pa- correlated with the remaining atypical symptoms than tients, the study was entirely carried out at one site, and were the nonatypical symptoms of decreased appetite and the results therefore warrant replication. Second, the small insomnia. sample sizes for conditions such as panic disorder with- Three of the most commonly cited validators of atypi- out agoraphobia, bipolar disorder, and certain PDs may cal depression were confirmed in our study: a preponder- have provided insufficient power to detect real differ- ance of women, a younger age at onset, and a longer du- ences that might have been present; these analyses should ration of illness. Two other important validators were not be interpreted cautiously. confirmed. Depressed patients with atypical features were We conclude that our results lend partial support not younger, and they were found to be more rather than for the validity of the DSM-IV atypical features subtype. less severely depressed. Although psychopharmacologic We could find no evidence, however, to suggest that mood studies have consistently reported that patients with atypi- reactivity is a valid component of the subtype, and this cal depression have a milder illness, these studies are vul- feature should perhaps be dropped from the diagnostic nerable to sampling bias because subjects with mild de- criteria set. With MAOIs falling into disuse, it seems un- pression are invariably excluded from these trials. The lower likely that more pharmacologic trials will be performed rates of severity found in these studies could also reflect to further address these issues. Thirty years ago, using the fact that the HAM-D rating scale does not account for antidepressant response rates in probands and family reversed neurovegetative symptoms. In our study, a greater members, Pare and Mack48 suggested that certain pa- severity of illness in patients with atypical depression was tients may have a distinct genetic makeup that predis- corroborated by lower baseline GAF scores and poorer so- poses them to respond to particular antidepressants. Since cial-functioning ratings. then, little progress has been made in our ability to pro- Our analysis of comorbid Axis I disorders con- file specific antidepressant responders. It would be un- firmed that panic disorder with agoraphobia, social pho- fortunate if the one instance in which a clear beneficial bia, hypochondriasis, and body dysmorphic disorder were response pattern was known were not more fully inves- all associated with atypical depression, as predicted. Panic tigated to uncover any underlying clues that might be disorder without agoraphobia was not associated with this available. condition. Although most research assessing the predic- tive value of comorbid anxiety for MAOI response rates Accepted for publication June 26, 2001. in patients with atypical features has focused on the pres- Corresponding author and reprints: Michael A. Poster- ence of comorbid panic attacks,19,20,27,29,30,42,44,46,47 our re- nak, MD, Department of Psychiatry and Human Behavior, sults suggest that the phobic element may be more rel- Brown University School of Medicine, Rhode Island Hospi- evant. tal, 235 Plain St, Suite 501, Providence, RI 02905 (e-mail: Despite the statistically significant findings, we would [email protected]). argue that our results do not support a strong associa- tion between histrionic PD and atypical depression be- cause (1) personality traits were also found to be higher REFERENCES in PDs that have not been postulated to be associated with atypical features; (2) the mean±SD number of DSM-IV 1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental histrionic traits in the 54 patients with atypical depres- Disorders, Fourth Edition. Washington, DC: American Psychiatric Association; 1994. sion was only 1.0±1.2; and (3) only 1.9% of the de- 2. West ED, Dally PJ. Effects of iproniazid in depressive . BMJ. 1959;1: pressed patients with atypical features met the criteria 1491-1494. for histrionic PD. Our finding that avoidant PD is asso- 3. Sargant W. Drugs in the treatment of depression. BMJ. 1961;1:225-227.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 75

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 4. Quitkin F, Rifkin A, Klein DF. Monoamine oxidase inhibitors: a review of antide- 28. Quitkin FM, Stewart JW, McGrath PJ, Liebowitz MR, Tricamo E, Klein DF, Rabkin pressant effectiveness. Arch Gen Psychiatry. 1979;36:749-760. JG, Markowitz JS, Wager SG. Phenelzine vs imipramine in the treatment of prob- 5. Beeber AR, Kline MD, Pies RW, Manring JM Jr. Hysteroid dysphoria in de- able atypical depression: defining syndrome boundaries of selective MAOI re- pressed inpatients. J Clin Psychiatry. 1984;45:164-166. sponders. Am J Psychiatry. 1988;145:306-311. 6. Spitzer RL, Williams JBW. Hysteroid dysphoria: an unsuccessful attempt to dem- 29. Quitkin FM, McGrath PJ, Stewart JW, Harrison W, Wager SG, Nunes E, Rabkin onstrate its syndromal validity. Am J Psychiatry. 1982;139:1286-1291. JG, Tricamo E, Markowitz J, Klein DF. Phenelzine and imipramine in mood re- 7. Rabkin JG, Stewart JW, Quitkin FM, McGrath PJ, Harrison WM, Klein DF. Should active depressives: further delineation of the syndrome of atypical depression. atypical depression be included in DSM-IV? In: Widiger TA, Frances AJ, Pincus Arch Gen Psychiatry. 1989;46:787-793. HA, First MB, Ross R, Davis W. DSM-IV Sourcebook. Vol 2. Washington, DC: 30. Liebowitz MR, Quitkin FM, Stewart JW, McGrath PJ, Harrison W, Rabkin JG, Tri- American Psychiatric Association; 1996:239-260. camo E, Markowitz JS, Klein DF. Psychopharmacologic validation of atypical de- 8. Asnis GM, McGinn LK, Sanderson WC. Atypical depression: clinical aspects and pression. J Clin Psychiatry. 1984;45:22-25. noradrenergic function. Am J Psychiatry. 1995;152:31-36. 31. Liebowitz MR, Quitkin FM, Stewart JW, McGrath PJ, Harrison W, Rabkin J, Tri- 9. Lonnqvist J, Sihvo S, Syvalahti E, Kiviruusu O. and fluoxetine in camo E, Markowitz JS, Klein DF. Phenelzine v imipramine in atypical depres- atypical depression: a double-blind trial. J Affect Disord. 1994;32:169-177. sion: a preliminary report. Arch Gen Psychiatry. 1984;41:669-677. 10. Paykel ES. Depression and appetite. J Psychosom Res. 1977;21:401-407. 32. Liebowitz MR, Quitkin FM, Stewart JW, McGrath PJ, Harrison WM, Markowitz 11. Casper RC, Redmond DE Jr, Katz MM, Schaffer CB, Davis JM, Koslow SH. So- JS, Rabkin JG, Tricamo E, Goetz DM, Klein DF. Antidepressant specificity in atypi- matic symptoms in primary affective disorder: presence and relationship to the cal depression. Arch Gen Psychiatry. 1988;45:129-137. classification of depression. Arch Gen Psychiatry. 1985;42:1098-1104. 33. McGrath PJ, Stewart JW, Harrison W, Ocepek-Welikson, Rabkin JG, Nunes EN, 12. Horwath E, Johnson J, Weissman MM, Hornig CD. The validity of major depres- Wager SG, Tricamo E, Quitkin FM, Klein DF. Predictive value of symptoms of sion with atypical features based on a community study. J Affect Disord. 1992; atypical depression for differential drug treatment outcome. J Clin Psychophar- 26:117-126. macol. 1992;12:197-202. 13. Robertson HA, Lam RW, Stewart JN, Yatham LN, Tam EM, Zis AP. Atypical de- 34. Sotsky SM, Simmens SJ. Pharmacotherapy response and diagnostic validity in pressive symptoms and clusters in unipolar and bipolar depression. Acta Psy- atypical depression. J Affect Disord. 1999;54:237-247. chiatr Scand. 1996;94:421-427. 35. First MB, Spitzer RL, Williams JBW, Gibbon M. Structured Clinical Interview for 14. Zisook S, Shuchter SR, Gallagher T, Sledge P. Atypical depression in an outpa- DSM-IV (SCID). Washington, DC: American Psychiatric Association; 1997. tient psychiatric population. Depression. 1993;1:268-274. 36. Zimmerman M, Mattia JI. Psychiatric diagnosis in clinical practice: is comorbid- 15. Pollitt J, Young J. Anxiety state or ? a study based on the ity being missed? Compr Psychiatry. 1999;40:182-191. action of monoamine oxidase inhibitors. Br J Psychiatry. 1971;119:143-149. 37. Endicott J, Spitzer RL. A diagnostic interview: the Schedule for Affective Disor- 16. Kendler KS, Eaves LJ, Walters EE, Neale MC, Heath AC, Kessler RC. The identi- ders and Schizophrenia. Arch Gen Psychiatry. 1978;35:837-844. fication and validation of distinct depressive syndromes in a population-based 38. Endicott J, Cohen J, Nee J, Fleiss J, Sarantakos S. Hamilton Depression Rating sample of female twins. Arch Gen Psychiatry. 1996;53:391-399. Scale: extracted from regular and change versions of the Schedule for Affective 17. Benazzi F. Atypical depression in private practice depressed outpatients: a 203- Disorders and Schizophrenia. Arch Gen Psychiatry. 1981;38:98-103. case study. Compr Psychiatry. 1999;40:80-83. 39. Guy W. ECDEU Assessment Manual for Psychopharmacology. Rockville, Md: Na- 18. Benazzi F. Exploring aspects of DSM-IV interpersonal sensitivity in bipolar II. tional Institute of ; 1976. US Dept of Health, Education, and Wel- J Affect Disord. 2000;60:43-46. fare publication ADM 76-338. 19. Kayser A, Robinson DS, Yingling K, Howard DB, Corcella J, Laux D. The influ- 40. Endicott J, Spitzer RL, Fleiss JL, Cohen J. The Global Assessment Scale: a pro- ence of panic attacks on response to phenelzine and amitriptyline in depressed cedure for measuring overall severity of psychiatric disturbance. Arch Gen Psy- outpatients. J Clin Psychopharmacol. 1988;8:246-253. chiatry. 1976;33:766-771. 20. Robinson DS, Kayser A, Corcella J, Laux D, Yingling K, Howard D. Panic attacks 41. Pfohl B, Blum N, Zimmerman M. Structured Interview for DSM-IV Personality. in outpatients with depression: response to antidepressant treatment. Psycho- Washington, DC: American Psychiatric Press Inc; 1997. pharmacol Bull. 1985;21:562-567. 42. Paykel ES, Rowan PR, Parker RR, Bhat AV. Response to phenelzine and ami- 21. Alpert JE, Uelelacker LA, McLean NE, Neirenberg AA, Pava JA, Worthington JJ, triptyline in subtypes of outpatient depression. Arch Gen Psychiatry. 1982;39: Tedlow JR, Rosenbaum JF, Fava M. Social phobia, avoidant personality disor- 1041-1049. der and atypical depression: co-occurrence and clinical implications. Psychol Med. 43. Paykel ES, Parker RR, Rowan PR, Rao BM, Taylor CN. Nosology of atypical de- 1997;27:627-633. pression. Psychol Med. 1983;13:131-139. 22. Thase ME, Carpenter L, Kupfer DJ, Frank E. Clinical significance of reversed veg- 44. Ravaris CL, Robinson DS, Ives JO, Nies A, Bartlett D. Phenelzine and amitripty- etative subtypes of recurrent major depression. Psychopharmacol Bull. 1991; line in the treatment of depression: a comparison of present and past studies. 27:17-22. Arch Gen Psychiatry. 1980;37:1075-1080. 23. Fava M, Uebelacker LA, Alpert JE, Nierenberg AA, Pava JA, Rosenbaum JF. Major 45. Davidson JRT, Giller EL, Zisook S, Helms MJ. Predictors of response to mono- depressive subtypes and treatment response. Biol Psychiatry. 1997;42:568-576. amine oxidase inhibitors: do they exist? Eur Arch Psychiatry Clin Neurosci. 1991; 24. Derecho CN, Wetzler S, McGinn LK, Sanderson WC, Asnis GM. Atypical depres- 241:181-186. sion among psychiatric inpatients: clinical features and personality traits. J Affect 46. Davidson J, Miller R, Strickland R. and personality disorder in de- Disord. 1995;39:55-59. pression. J Affect Disord. 1985;8:177-182. 25. Weissenburger J, Rush AJ, Giles DE, Stunkard AJ. Weight change in depres- 47. Liebowitz MR, Quitkin FM, Stewart JW, McGrath PJ, Harrison W, Rabkin J, Klein sion. Psychiatry Res. 1986;17:275-283. DF. Psychopharmacological dissection of nonendogenous depression. Psycho- 26. Stewart JW, McGrath PJ, Rabkin JG, Quitkin FM. Atypical depression: a valid pharmacol Bull. 1984;20:390-392. clinical entity? Psychiatr Clin North Am. 1993;16:479-495. 48. Pare CMB, Mack JW. Differentiation of two genetically specific types of de- 27. Zimmerman M, Mattia JI. Psychiatric diagnosis in clinical practice: is comorbid- pression by the response to antidepressant drugs. J Med Genet. 1971;8:306- ity being missed? Compr Psychiatry. 1999;40:182-191. 309.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 59, JAN 2002 WWW.ARCHGENPSYCHIATRY.COM 76

©2002 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021