BRITISH JOURNAL OF PSYCHIATRY (2006), 188, 432^438

Agoraphobia in adults: incidence and longitudinal Between 1993 and 1996, in the Baltimore ECA Follow-up Study, 88% of relationship with panic the original Baltimore cohort were traced, and 73% of those known to be alive were interviewed again using the DIS (nn¼1920).1920). O. JOSEPH BIENVENU, CHIADI U. ONYIKE, MURRAY B. STEIN, The survey method is described in detail LI-SHIUN CHEN, JACK SAMUELS, GERALD NESTADT and WILLIAM W. EATON elsewhere (Eaton et aletal, 1997). Briefly, the same survey procedures were used as in 1981. The interview was augmented with a life chart, in order to assist recall over the follow-up period. Mortality in the inter- vening 13 years was substantial as a result Background Theories regarding how The DSM–IV implies that agoraphobia is of the high proportion of elderly respon- spontaneous panic and agoraphobia relate almost always a consequence of sponta- dents originally sampled at this site. neous panic attacks (American Psychiatric Neither baseline phobias nor panic disorder are based mostly on cross-sectional Association, 1994, 2000). Recent editions were significant predictors of loss to and/or clinic data. of the DSM have been influenced by the follow-up or refusal to participate in the work of Klein (1980) and others, whose follow-up study, although phobias were Aims Todetermine how spontaneous patients usually reported that their agora- weakly predictive of mortality between panic and agoraphobia relate longitu- phobic symptoms followed unexpected interview waves (Badawi et aletal, 1999).,1999). dinally,anddinally, and to estimate the incidence rate spells of panic. Although epidemiological A subset of the 1920 participants in the of and other possible risk factors for studies found many people with agora- Baltimore ECA Follow-up Study was first-onset agoraphobia, using a general phobia but no history of panic, typical additionally interviewed by psychiatrists epidemiological methods, employing fully ((nn¼816). Individuals were selected for this population cohort. structured non-clinician interviews, were interview mainly if they showed evidence suspect (Horwath et aletal, 1993). Also, of Axis I psychopathology (especially with MethodMethod A sample of1920 adultsin east although retrospective investigations found the DIS) in 1981 and/or in the Follow-up BaltimorewereBaltimorewereassessedin1981^1982 assessedin1981^1982 and that agoraphobic symptoms preceded panic Study, although about a quarter were the mid-1990s with the Diagnostic Inter- in some patients with both conditions selected at random from the remaining view Schedule (DIS).Psychiatristdiagnoses (reviewed in Fava & Mangelli, 1999), the members of the sample (Samuels et aletal,, were made in a subset of the sample at authors of DSM–IV were concerned about 2002). A total of 816 persons participated possible recall bias (Craske, 1996) given in the psychiatric evaluation, which follow-up (nn¼816).816). the absence of prospective longitudinal included the use of a semi-structured diag- ResultsResults Forty-one new cases of DIS/ studies such as the one reported here. We nostic instrument, the Schedule for Clinical sought to determine first, how spontaneous Assessment in Neuropsychiatry (SCAN; DSM ^ III ^ R agoraphobia were identified panic and agoraphobia relate longitudi- WingWing et aletal, 1990). Participants provided (about 2 per1000 person-years at risk). nally, and second, the incidence rate of informed consent for each interview, and As expected, baseline DIS/DSM^IIIpanic and other possible risk factors for first- the current study was approved by the disorderpredictedfirstincidenceofagora--disorderpredictedfirstincidenceofagora onset agoraphobia, using a longitudinally Johns Hopkins Medicine Institutional assessed general population cohort, Review Board 3. phobia (OR(ORphobia ¼12,95% CI 3.2^45), asdidas did with clinical reappraisal at follow-up to youngeryoungerage,femalegenderandother age, female gender and other assess effects of possible diagnostic phobias.Importantly,baselinephobias.Importantly, baseline agora- misclassification. Diagnostic interviews phobiawithout spontaneous panic attacks The version of the DIS used in the first two also predicted first incidence of panic dis- interview waves (1981 and 1982) included a general description of a phobia (‘such a order (OR(ORorder ¼3.9,95% CI1.8^8.4).Longitu- METHOD strong fear of something or some situation dinalrelationships between panic disorder SampleSample that [you] try to avoid it, even though and psychiatrist-confirmed agoraphobia The Epidemiologic Catchment Area (ECA) [you] know there is no real danger’), fol- were strong (panic before agoraphobia programme was completed in five centres lowed by screening questions about fears OR¼20,95% CI 2.3^180; agoraphobia in the USA from 1981 to 1982 (Regier etet of ‘tunnels or bridges’, ‘public transporta- alal, 1984). Trained non-clinician inter- tion (airplanes, buses, elevators)’, ‘crowds’, before panic OR¼16, 95% CI 3.2^78). viewers administered the National Institute ‘going out of the house alone’ and ‘being of Mental Health Diagnostic Interview alone’. If respondents acknowledged having Conclusions The implied one-way Schedule (DIS; Robins et aletal, 1981) to a any of these unreasonable fears that causalrelationship between spontaneous probabilistic sample of the adult population resulted in avoidance and ‘interfered with panic attacks and agoraphobia in DSM ^ IV in each centre at two points in time, (their lives) or activities a lot’, they were appearsappearsincorrect. incorrect. separated by a 1-year interval. In given DIS/DSM–III diagnoses of agora- Baltimore, the original sample included phobia (Boyd et aletal, 1985). As noted by Declaration of interest None. 3481 participants (Fig. 1). WeissmanWeissman et aletal (1986), the DIS did not

432

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. AGORAPHOBIA:AGOR APHOBIA: INCIDENCE AND RELATION WITH PANIC

(Rubio-Stipec et aletal, 1992). Seventeen respondents with incident DSM–III–R agoraphobia reported no disorder-level symptoms in the first wave but, at follow- up, reported the onset of their earliest related fears as before 1981. It should be noted that the DIS–III–R did not assess age at onset of associated symptoms or interference, so these respondents could have had agoraphobia-related fears without meeting full criteria for agoraphobia until the follow-up period. Consistent with this, we found that more than half of these individuals (9 out of 17) reported ‘sub- Fig. 11Fig. Overview of the Baltimore Epidemiologic Catchment Area (ECA) Study.DIS,Study. DIS, Diagnostic Interview threshold’ DIS/agoraphobia-related fears Schedule; SCAN, Schedules for Clinical Assessment in Neuropsychiatry. in 1981. An additional three people re- ported that both the onset and offset of address the DSM–III concept of agora- Psychiatrists made agoraphobia diag- their fears occurred before 1981. All 20 phobia as ‘fear of being in a place from noses according to DSM–III–R criteria, were considered incident cases (in effect, which escape might be impossible or diffi- using the SCAN. Agoraphobic fears were giving priority to the time of assessment cult in case of incapacitation’ (American grouped together, with the following six data and assuming errors in recall for the Psychiatric Association, 1980). probes: ‘being in a public place when alone, three respondents mentioned above). We The version of the DIS used from 1993 open spaces, empty streets’, ‘crowds, shops, conducted additional analyses in which to 1996 was modified to reflect the changes theatres – places with no easy exit’, these three cases were excluded, and there in diagnostic criteria in DSM–III–R ‘travelling alone – buses, trains, planes’, was no substantive difference in results. (American Psychiatric Association, 1987). ‘going out alone – being alone away from There were also slight differences in the Screening questions for agoraphobia read: home’, ‘collapsing while alone or with no screening questions for DIS/DSM–III–R help near’ and ‘being alone indoors’. The agoraphobia; that is, it was conceivable (a)(a)‘Some‘Some people have such an unreason- SCAN glossary describes agoraphobia as that participants with phobias of riding able fear of being in a crowd, anxiety related to being in such situations, in cars or trains were missed with the DIS/ travelling in buses, cars or trains, or lessened in the presence of a trusted com- DSM–III interview and incorrectly classi- crossing a bridge that they always get panion. The anxiety is related to difficulty fied as incident DIS/DSM–III–R cases. very upset in such a situation or avoid escaping quickly (feeling trapped) or leav- However, we found that incident cases of it altogether. Did you ever go through DIS/DSM–III–R agoraphobia respondents a period when being in such a situation ing without embarrassment, or fear of always frightened you badly?’ being incapacitated, with no help near. who endorsed items containing these fears Thus, psychiatrists administering the SCAN were, if anything, less likely to report onset (b)(b)‘Did‘Did you ever have such an unreason- assessed the nature of the fears, not just the before 1981, compared with those who en- ably strong fear of being alone away situations in which they occurred. dorsed items containing the same situations from home that you always got very in both interviews (further information upset in this situation, or you avoided available from the author on request). it altogether?’ Defining incident DIS/DSM^III^R agoraphobia (c)(c)‘Have‘Have you ever had an unreasonable Statistical analysis fear of being alone in your own home?’ In the follow-up sample, 221 cases of life- time DIS/DSM–III agoraphobia were iden- A weighted cumulative incidence propor- If positive, these screening questions tified at wave 1 (1981); in these cases the tion was calculated with an elaboration of were followed with questions pertaining individuals were considered not at risk for the Woodbury down-weighting procedure to onset, recency, associated anxiety symp- incidence of the disorder. Of the remaining (Kessler(Kessler et aletal, 1985). This procedure was toms and level of interference (including participants, 1557 were considered at risk developed for the ECA programme to limits on being able to travel or leave for development of the disorder during the ensure that parameter estimates were home). In order to meet DIS/DSM–III–R 13 years of follow-up, and 140 had missing accurate. Weighting, in this case, corrects criteria for agoraphobia, respondents had data. Incident cases for this study were the parameter estimate by the probability to endorse at least one of the screening those who were at risk and met criteria of selection into the sample; also, it adjusts questions, as well as at least one associated for lifetime DSM–III–R agoraphobia at for non-response to replicate the target anxiety symptom (specifically dizziness, wave 3 (1993–1996). Participants with population of survivors, according to age, palpitations, nausea/vomiting, feeling of incident DIS/DSM–III agoraphobia at wave ethnicity and gender categories. Putative loss of control of bodily functions, or 2 (1982,2(1982, nn¼3) were considered new cases baseline risk factors, including demo- derealisation) and interference. Again, the only if they met DSM–III–R criteria at wave graphic factors and lifetime psychiatric nature of the fear was not directly assessed, 3 as well (nn¼1). Given the multiwave diagnoses (as recorded in 1981), were only the typical situations in which such nature of the study, there was a potential examined in relation to subsequent fear was likely to occur. for discrepancies in timing of onsets development of agoraphobia. Baseline

433433

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. BIENVENU ET AL

agoraphobia (with and without a history of Ta b l e 1 Odds of first-onset agoraphobia over aa13-year 13-year period, given baseline (1981) demographic spontaneous panic attacks) and other Axis I characteristics conditions were also examined in relation to subsequent development of panic disor- New-onset agoraphobia11 UnivariateAdjusted 22 der. Univariate and adjusted odds ratios were generated in logistic regression Present Absent OR (95% CI)CI)(95% OROR (95% CI)CI)(95% analyses.analyses. nn nn

Age, years 18^29 21 5085081.0 1.01.0 RESULTSRESULTS 30^44 1414 4554550.7 (0.4^1.5) 0.7 (0.4^1.4) 45^6445^64 63856 385 0.40.4 (0.2^0.9)*(0.2^0.9)*0.3 (0.1^0.8)* Incidence of agoraphobia 65^86 01680 168 There were 41 new cases of DIS/DSM–III–R Gender agoraphobia identified among the 1557 FemaleFemale 3434900 3.3 (1.5^7.5)* 3.63.6 (1.6^8.2)*(1.6^8.2)* participants at risk. Of these, 14 (34%) re- MaleMale 76167 6161.0 1.01.0 ported fears of being in a crowd or standing EthnicityEthnicity in line; 10 (24%) reported fears of riding on WhiteWhite 20 9669661.0 1.01.0 trains or buses, planes or in a car; 14 (34%) African American 21497 2.02.0(1.1^3.8)* 1.6 (0.8^2.9) reported fears of crossing a bridge; 11 OtherOther 050 533 (27%) reported fears of being alone away Education33 from home and 11 (27%) reported fears of being alone at home. The unweighted Up to grade 8 21262 1260.7 (0.2^3.6)(0.2^3.6)1.5 (0.3^7.8) 13-year cumulative incidence proportion Grades 9^11 14378 1.7 (0.7^4.2)(0.7^4.2)1.9 (0.8^4.7) of DIS/DSM–III–R agoraphobia is esti- Grade 12 16 5045041.5 (0.6^3.5) 1.4 (0.6^3.3) mated at 2.6 per 100 persons at risk. The More than grade 12 83778 3771.0 1.01.0 weighted 13-year cumulative incidence pro- Marital status33 portion is estimated at 2.4 per 100 persons Married 13701 70 1 1.0 1.01.0 at risk. These estimates correspond to an WidowedWidowed 11431 1430.4 (0.0^2.9) 0.6 (0.1^5.0) annual incidence rate of approximately 2 Separated 61286 1282.5 (0.9^6.8) 2.02.0 (0.7^5.3)(0.7^5.3) per 1000. The mean age at onset of Divorced 31433 1431.1 (0.3^4.0) 1.0 (0.3^3.5) agoraphobia-related fears for these incident Never married 18400 2.4 (1.2^5.0)*(1.2^5.0)*1.9 (0.9^4.3) cases was 21.2 years (s.d.¼16.8).16.8). Annual household income33 Up to $4999 81918 1914.6 (0.6^38) 3.8 (0.4^31) $5000^9999 52235 223 2.52.5 (0.3^22)(0.3^22)2.4 (0.3^21) $10 000^14 999 72527 252 3.13.1 (0.4^25)(0.4^25) 2.62.6 (0.3^22)(0.3^22) Baseline demographic $15 000^19 999 41994 1992.2 (0.2^20)(0.2^20) 2.12.1 (0.2^20)(0.2^20) characteristics and incident agoraphobia $20 000^34 999 94179 4172.4 (0.3^19)(0.3^19)2.3 (0.3^18) 44$35 000 111 111111.0 1.01.0 The 41 respondents with incident DIS/ DSM–III–R agoraphobia were compared **PP440.05.0.05. 1. Diagnosed with the Diagnostic Interview Schedule modified to reflect DSM^III^R criteria. with the 1516 respondents who were at risk 2. Adjusted for age group and gender. but did not develop agoraphobia. Younger 3. Education, marital status and household income data were missing for nn¼132,132, nn¼1and1andnn¼130 respecrespectively. tively. age, female gender, African American eth- nicity and having never married (at base- line, 1981) were statistically significant Baseline Axis I disorders OROR¼12, as opposed to odds ratios of about predictors of incident agoraphobia in uni- and incident agoraphobia 3–43–4forfor the other significant Axis I predic- variate models (Table 1). The odds ratio tors). The relationships between baseline for incident agoraphobia was more than In univariate analyses, participants with major depression or spontaneous panic twice as high in young adults (18–29 years lifetime DIS/DSM–III major depression, attacks and incident agoraphobia were no old at baseline) compared with middle-aged spontaneous panic attacks (not cued by longer statistically significant when age adults (45–64 years old; there was no new phobic stimuli), panic disorder and other group and gender were taken into account. case among those over 65 years old), and phobias at baseline (i.e. in 1981) were Although baseline alcohol use disorders did more than three times as high in women. significantly more likely to develop agora- significantly predict onset of agoraphobia When age group and gender were taken phobia over the follow-up period, com- when these demographic factors (particu- into account, ethnicity and marital status pared with those without these conditions larly gender) were taken into account were not significantly related to incidence at baseline (Table 2). By far the strongest (women were less likely to have baseline al- (e.g. as might be expected, younger partici- predictor of incident agoraphobia was cohol use disorders), this relationship was pants were less likely to be married). baseline panic disorder (unadjusted not significant when baseline comorbid

434

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. AGORAPHOBIA:AGOR APHOBIA: INCIDENCE AND RELATION WITH PANIC

Ta b l e 22Tab Odds of first-onset agoraphobia over a13-yeara 13-year period, given baseline (1981) lifetime Axis I disorders cases, two people were assessed as unreli- able informants, one young woman said Lifetime Axis I First-onset agoraphobia22 UnivariateAdjusted 33 Adjusted44 that her physician had recommended she avoid stressful situations following possible disorder at baseline11 Present Absent OR (95% CI)CI)(95% OROR (95% CI)CI)(95% OROR (95% CI)CI)(95% transient ischaemic attacks, and three nn nn appeared to give discrepant answers to similar questions in the two interviews Alcohol use disorder (i.e. the non-clinician-administered DIS Present 71727 1721.6 (0.7^3.6) 3.1 (1.2^7.8)*(1.2^7.8)* 2.42.4 (0.9^6.2)(0.9^6.2) and the psychiatrist-administered SCAN). Absent 341316 In addition, 26 participants at risk of Drug use disorder DIS/DSM–III–R agoraphobia but without Present 2812 810.9 (0.2^3.8) 0.8 (0.2^3.3)(0.2^3.3)0.6 (0.1^2.7) the DIS diagnosis were diagnosed as having Absent 381401 agoraphobia by psychiatrists. Twelve Major depression of theseoftheseparticipants endorsed DIS agoraphobia-related fears but did not Present 4504 503.1 (1.1^9.1)* 2.4 (0.8^7.2)(0.8^7.2)1.5 (0.4^5.3) endorse sufficient additional criteria for Absent 371439 the diagnosis, and 10 appeared to provide Dysthymia discrepant answers to similar questions in Present 2292 29 2.62.6 (0.6^11.2)(0.6^11.2)1.9 (0.4^8.5) 0.3 (0.0^3.6) the two interviews. In the remaining 4 cases Absent 391460 it was possible that psychiatrists adminis- Panic disorder tering the SCAN had elicited symptoms Present 3103 1012 (3.2^45)*(3.2^45)* 1111(2.8^46)* 8.3 (1.9^35)* not specifically addressed with the DIS: fear Absent 371478 of shops and theatres, and fear of collapsing Spontaneous panic attack with no help nearby. Present 3303 303.9 (1.2^13)*(1.2^13)*2.9 (0.8^10) 2.1 (0.6^7.8) 55 For the following analyses, we defined Absent 371457 psychiatrist-confirmed incident agorapho- bia as present if the participant met criteria Other phobia for incident DIS/DSM–III–R agoraphobia Present 13 227227 2.62.6 (1.3^5.1)*(1.3^5.1)* 2.32.3 (1.2^4.5)*(1.2^4.5)* 2.22.2(1.1^4.4)* and was diagnosed by a psychiatrist as Absent 281262 having lifetime agoraphobia at follow-up **PP440.05. ((nn¼9). We excluded discrepant cases (with 1. Diagnosed with the unmodified Diagnostic Interview Schedule (DSM^III). unconfirmed incident DIS/DSM–III–R 2. Diagnosed with the Diagnostic Interview Schedule modified to reflect DSM^III^R criteria. 3. Adjusted for baseline age group and gender. agoraphobia, nn¼18; or psychiatrist- 4. Adjusted for baseline age group, gender, alcohol use disorder, panic disorder and other phobia. diagnosed agoraphobia without the 5. Adjusted for baseline age group, gender, alcohol use disorder and other phobia. incident DIS diagnosis, nn¼26), so the comparison group was also slightly smaller conditions were also taken into account Limiting to psychiatrist-confirmed than in the previous sections (nn¼1490). The1490).The (i.e. panic disorder and other phobias). agoraphobia pattern of predictors for psychiatrist- Twenty-seven of the participants with confirmed incident agoraphobia was simi- incident DIS/DSM–III–R agoraphobia were lar to that reported above; however, given Baseline agoraphobia and other seen by psychiatrists. Masked to DIS the restricted number of cases, statistical Axis I disorders, and incident panic diagnoses, psychiatrists rated 15 of these significance was not usually present. Never- disorderdisorder individuals as having substantial, unambig- theless, the expected relationship between As reported previously by Eaton et aletal uous lifetime agoraphobic phenomena, baseline DIS/DSM–III panic disorder and (1998), there were 35 new cases of panic although the psychiatrists thought 6 of psychiatrist-confirmed incident agora- disorder during the follow-up period out these 15 cases were below the rigorous phobia was highly significant (OR¼20,20, of 1731 at risk; demographic risk factors SCAN diagnostic threshold (in all 6 of these 95% CI 2.3–180). Given the relatively included female gender, younger age and cases the individual had at least two small number of cases here, we did not White ethnicity. Significant baseline agoraphobia-type fears). Of the remaining think it wise to include covariates in the (1981) Axis I predictors of incident panic 12 cases, it appeared that 6 had been mis- logistic regression models. We also con- disorder included lifetime DIS/DSM–III classified with the DIS (i.e. using the psychi- ducted additional analyses of predictors drug use disorder, major depression, dys- atric interview as the standard); that is, of ‘incident’ psychiatrist-diagnosed thymia, agoraphobia, and agoraphobia their avoidances were understandable in DSM–III–Ragoraphobia; cases were classi- without a history of spontaneous panic at- terms of other kinds of fears. In the latter fied as incident if they met lifetime criteria tacks (last OR¼3.9 95% CI 1.8–8.4) (Table cases one person had a fear of flying, one at follow-up but did not meet criteria for 3). However, there was a substantial degree had a fear of heights, and four had social DIS/DSM–III agoraphobia in 1981 (nn¼35/35/ of comorbidity among baseline disorders. and other fears (one was practically house- 667; unconfirmed DIS cases were ex- When baseline comorbidity was taken into bound with fear of people; the rest had cluded). As with incident DIS/DSM–III–R account, agoraphobia was the only signifi- additional fears of enclosed spaces, being agoraphobia, significant predictors in- cant predictor of incident panic disorder. attacked or heights/bridges). In the final 6 cluded baseline young adult age, female

435435

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. BIENVENU ET AL

Ta b l e 3 Odds of first-onsetpanicfirst-onset panic disorder over a13-year period, given baseline (1981) lifetime Axis I disorders the only baseline Axis I disorder studied that significantly predicted incident panic disorder when baseline comorbidity was Lifetime Axis I First-onset panic disorder22 UnivariateUnivariateAdjusted 33 Adjusted44 taken into account. disorder at baseline11 Present Absent OR (95% CI)CI)(95% OROR (95% CI)CI)(95% OROR (95% CI)CI)(95% Our results are consistent with previous nn nn retrospective reports in which patients with both agoraphobia and panic sometimes Alcohol use disorder reported agoraphobia preceding panic Present 51955 1951.2 (0.5^3.3)(0.5^3.3) 2.5 (0.9^7.1)(0.9^7.1)1.7 (0.6^5.1)(0.6^5.1) (e.g. Fava(e.g.Fava et aletal, 1988; Argyle & Roth, Absent 301471 1989; Lelliott et aletal, 1989). Given that the Drug use disorder causal relationship between spontaneous Present 7967 964.1 (1.7^9.6)* 3.9 (1.6^10)* 2.6 (1.0^7.0)(1.0^7.0) panic attacks and agoraphobia does not ap- Absent 28 15641564 pear as straightforward as Klein’s theory Major depression suggests, we suggest that DSM–V should Present 7667 666.1 (2.6^14)*(2.6^14)* 3.73.7 (1.5^9.2)*(1.5^9.2)* 2.2 (0.8^5.9)(0.8^5.9) de-emphasise the implied one-way causal Absent 28 16021602 relationship from spontaneous panic to agoraphobia and make agoraphobia itself DysthymiaDysthymia a stand-alone diagnosis again, as in Present 4334 336.4 (2.1^19)*(2.1^19)* 3.83.8 (1.2^12)*(1.2^12)* 2.4 (0.7^8.6)(0.7^8.6) ICD–10ICD–10(World(World Health Organization, Absent 31 16351635 1993). That is, although panic does appear Agoraphobia to be a potent risk factor for agoraphobia, Present 11169 4.1 3.73.7 2.9 (2.0^8.5)*(2.0^8.5)* (1.7^7.9)* (1.3^6.5)* agoraphobia also appears to be a risk factor Absent 241516 for panic disorder. 55 Agoraphobia without panic Other recent epidemiological research Present 10161 3.9 (1.8^8.4)*(1.8^8.4)* 3.53.5 (1.6^7.8)*(1.6^7.8)* 2.7 (1.2^6.1)*(1.2^6.1)*,6 is relevant to Marks’s argument that agora- Absent 241516 phobia should be a stand-alone diagnosis, Other phobia like other phobias, often but not always Present 42314 2312.5 (1.3^5.0)(1.3^5.0) 0.70.7 (0.2^2.2)(0.2^2.2) 1.11.1 (0.4^3.4) associated with panic (Marks, 1987aa).). Absent 311436 1436 Three cross-sectional community studies that employed rigorous clinical methods **PP440.05. 1. Diagnosed with the unmodified Diagnostic Interview Schedule (DSM^III). found that agoraphobia was sometimes 2. Diagnosed with the Diagnostic Interview Schedule modified to reflect DSM^III^R criteria. present in people with no history of panic: 3. Adjusted for baseline age group, gender and ethnicity. 4. Adjusted for baseline age group, gender, ethnicity, drug use disorder, major depression, dysthymia and agoraphobia. WittchenWittchen et aletal (1998) found that many ado- 5. Agoraphobia without a history of spontaneous panic attacks. lescents and young adults had clinician- 6. Adjusted for baseline age group, gender, ethnicity, drug use disorder, major depression and dysthymia. confirmed agoraphobia but no history of gender, spontaneous panic attacks and DISCUSSION panic; Hayward et aletal (2003) found that rig- panic disorder (further information avail- orously defined agoraphobic fear and able from the author on request). Panic and agoraphobia avoidance, although relatively rare in a The good concordance between DIS/ A primary goal of this study was to deter- sample of high-school students, was usually DSM–III–R and psychiatrist-diagnosed mine how spontaneous panic attacks/panic not associated with a history of panic at- (SCAN) incident panic disorder has been disorder and agoraphobia relate longitudin- tacks; and Faravelli et aletal (2004), who – like reported elsewhere (Eaton et aletal, 1998).,1998). ally. Consistent with Klein’s cogent many in the USA – were sceptical that For the purposes of the current analyses, argument regarding the unconditioned agoraphobia without a history of panic we defined psychiatrist-confirmed baseline stimulus property of spontaneous panic existed, found several such cases in their agoraphobia as present if DIS/DSM–III attacks causing a conditioned avoidance mostly adult sample. Epidemiologists have agoraphobia was present in 1981 and a response (Klein, 1980), several previous long noted that one reason clinicians rarely psychiatrist made the diagnosis of lifetime retrospective clinical studies (reviewed by encounter patients with agoraphobia but agoraphobia at follow-up (nn¼15, of whom Craske, 1996; Fava & Mangelli, 1999) without panic is that panic itself influences 11 had no history of DIS/DSM–III spon- and the one previous community study of treatment-seeking. Two epidemiological taneous panic attacks at baseline). We incident agoraphobia we know of (Eaton studies have provided empirical evidence excluded participants with unconfirmed & Keyl, 1990), baseline spontaneous panic that individuals experiencing agoraphobia baseline agoraphobia (nn¼97). Psychiatrist- attacks (especially frequent ones, i.e. DSM– without panic are less likely to seek confirmed baseline agoraphobia, with and III panic disorder) strongly predicted new- psychiatric care (Wittchen et aletal, 1998;,1998; without a history of spontaneous panic onset agoraphobia. This finding was not Andrews & Slade, 2002); however, these attacks, was a strong predictor of incident surprising. However, the finding that base- studies did not find that such individuals DIS/DSM–III–R panic disorder (OR¼16,16, line agoraphobia without spontaneous nevernever seek care, so it is important that clin- 95% CI 3.2–78 and OR¼15, 95% CI panic attacks also predicted first-onset icians recognise agoraphobia in the absence 3.1–75 respectively); again, given relatively panic disorder is clearly inconsistent with of panic. In the USA at least, many younger small numbers, we did not include Klein’s theory. In fact, in the present study, clinicians tend to consider the diagnosis of covariates in these models. baseline agoraphobia without panic was agoraphobia only when their patients

436

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. AGORAPHOBIA:AGOR APHOBIA: INCIDENCE AND RELATION WITH PANIC

report panic attacks – this is understand- as having agoraphobia with the DIS was of situations such as driving on express- able as a result of changes in recent versions smaller than the number of people appar- ways or through tunnels, but their agora- of DSM. Diagnostic criteria clearly influ- ently misclassified as not agoraphobic). phobia symptoms become worse and ence the conceptualisation of psychiatric Our findings regarding demographic pre- generalise further after the onset of panic conditions, and this is not limited to the dictors of agoraphobia (i.e. younger age in those situations: see Fava & Mangelli clinical setting: if researchers conceptualise and female gender) were in agreement with (1999) for a review. Second, we focused agoraphobia only as avoidance of certain those of Eaton & Keyl (1990). on time-of-interview data, ignoring the spe- situations because of fear of panic, the We found that other baseline phobias cific timing of onset of syndromes between suffering of many people with agoraphobia also predicted onset of agoraphobia, interview waves (e.g. there were seven par- is ignored.isignored. although these were a much weaker pre- ticipants with first onset of both panic One reason Marks contended that dictor than baseline panic disorder. The disorder and agoraphobia during the follow- agoraphobia is a syndrome is that agora- difference in strength of this relationship, up period). We chose to do this because phobic fears have repeatedly been found and the fact that baseline depressive dis- longitudinally collected time-of-interview to cluster together within individuals, sepa- orders were not significantly related to data are less susceptible to recall bias, and rate from other types of fears (Marks, onset of agoraphobia (at least when demo- they are a particular strength of this data- 19871987bb; Arrindell;Arrindell et aletal, 2003). Further graphic correlates were taken into account), set. Third, since the psychiatric interviews support for allowing agoraphobia to stand suggest a somewhat specific longitudinal were conducted at follow-up, and only on apart from panic comes from a recent study relationship between panic disorder and a selected subset of our sample, it was not of data from the National Comorbidity agoraphobia in adults, beyond membership possible to adjust incidence rates accurately Survey (Kessler et aletal, 1994), in which Cox in a larger group of ‘internalising’ or using this method. Fourth, the relatively small et aletal (2003) found that agoraphobic fears ‘neurotic’ disorders (Krueger, 1999; Tyrer, number of cases with psychiatrist- clustered together within individuals 1985).1985). confirmed incident DIS/DSM–III–R agora- whether or not they had a history of panic phobia precluded statistical confirmation attacks.attacks. of baseline predictors other than panic HettemaHettema et aletal (2005) have recently pro- Strengths and limitations disorder. Nevertheless, for most cases in vided evidence that genetic factors in com- of the study which the diagnosis of incident DIS/ mon may predispose to both agoraphobiaphobiaagora This study’s strengths include the use of a DSM–III–R agoraphobia was not con- and panic disorder, in a cross-sectionalsectionalcross- longitudinally assessed population-based firmed by psychiatrists, the issue was not multivariate twin study of common anxiety cohort, with psychiatrist interviews to chiefly frank misclassification (e.g. the disorders. A genetically informative bi- determine the possible effects of misclassi- symptoms were better conceptualised as variate longitudinal study of panic and fication. To our knowledge, this is the first another phobia), but rather a combination agoraphobia could potentially provide comprehensive report of longitudinal asso- of stricter thresholds in the psychiatric additional information about aetiological ciations between panic and agoraphobia interviews or inconsistent responses by the mechanisms, in that there may be direct using a prospective design. participant concerned to similar questions causal paths from the experience of panic We note the following caveats. First, in the two interviews (thus, DIS/ to agoraphobia (Klein, 1980) and/or vicevice since the psychiatric interviews were con- DSM–III–Ragoraphobia appears to versaversa (Fava & Mangelli, 1999). ducted at follow-up, it is possible that some resemble more closely what clinicians call Regardless of specific aetiological mechan- individuals classified here as having base- ‘agoraphobia’ than did DIS/DSM–III agora- isms, though, clinicians should keep in line (1981) DIS/DSM–III agoraphobia phobia). We gain further confidence in our mind that agoraphobia without panic without panic and subsequent first-onset reported risk factors since we found similar appears to be at least a marker of risk for DIS/DSM–III–R panic disorder would have results in predictors of psychiatrist- later-onset panic disorder. been found to have a history of baseline diagnosed ‘incident’ agoraphobia. panic with a thorough clinical assessment of lifetime symptoms at baseline. However, ACKNOWLEDGEMENTS Agoraphobia incidence given results of a previous clinical reapprai- The National Institute of Mental Health (grants and other predictors sal of DIS/DSM–III agoraphobia without R01-MH474 47, R01-MH5 0616 and K23 -MH6-MH6454) 454) Eaton & Keyl (1990) estimated the annual panic, the most likely form of misclassifica- supported this study.We thank the many individuals first incidence of DIS/DSM–III agoraphobia tion by far involved diagnosing simple who have taken part over the years as administra- at 22 per 1000 population. We believe our (specific) phobias as DIS/DSM–III agora- tors, trainers, interviewers and participants in the Baltimore Epidemiologic Catchment Area Study. estimate of DSM–III–R agoraphobia inci- phobia (Horwath et aletal, 1993). Also, partici- dence (approximately 2 per 1000 per year) pants who had mild agoraphobia at REFERENCES is likely to be a conservative one for two baseline and developed first-onset panic American Psychiatric Association (1980) Diagnostic reasons. First, the 13-year period of risk disorder during the follow-up period could and Statistical Manual of Mental Disorders (3rd edn) probably affected recall for some agora- have had worsening of agoraphobia symp- (DSM^III).Washington,(DSM ^ III).Washington, DC: APA. phobic symptoms at follow-up. Second, toms after the onset of panic, thus making American Psychiatric Association (1987) Diagnostic many of the sample who might have been them more likely to be diagnosed with life- and Statistical Manual of Mental Disorders (3rd edn,edn,(3rd at risk and developed agoraphobia over time agoraphobia by psychiatrists at revised) (DSM^III^R).Washington, DC: APA. the follow-up period were not diagnosed follow-up. This is consistent with our and American Psychiatric Association (1994) Diagnostic with DIS/DSM–III–R agoraphobia (the others’ clinical experience; for example, and Statistical Manual of Mental Disorders (4th edn) number of people apparently misclassified some patients report always being afraid (DSM^IV).Washington, DC: APA.

437437

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use. BIENVENU ET AL

American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders (4th edn, text revision) (DSM^IV^TR).Washington, DC: APA. CLINICAL IMPLICATIONS Andrews, G. & Slade, T. (2002) Agoraphobia without a history of panic disorder may be part of the panic && Agoraphobia without panic appears to be a risk factor,or at least a marker of risk, disorder syndrome. Journal of Nervous and Mental for the onset of panic. Disease,, 190,,624^630. 624^630.

Argyle,N.&Roth,M.(1989)The phenomenological && Clinicians should be aware that agoraphobia can exist without a history of panic, study of 90 patients with panic disorder. PsychiatricPsychiatric despite what is implied in DSM^IV. Developments,, 33,187^209. Arrindell,W. A., Eisemann, M., Richter, J., et aletal && Young women are at greatest risk for the onset of agoraphobia. (2003)(2003) Phobic anxiety in 11nations.11 nations. Part I: Dimensional constancy of the five-factor model. Behaviour Research LIMITATIONS and Therapy,, 4141, 461^479.,461^479. Badawi, M. A., Eaton,W.W., Myllyluoma, J., et aletal && Non-clinician-administered, fully structured interviews such as the Diagnostic (19(1999) 9 9) Psychopathology and attrition in the Baltimore Interview Schedule sometimes misclassify people with other phobias as having ECA15-year follow-up 1981^1996. Social Psychiatry and Psychiatric Epidemiology,, 34, 91^98.,91^98. agoraphobia.

Boyd, J. H., Robins, L. N., Holzer,C.Holzer, C. E., et aletal (19 8 5) && Psychiatrists only performed (lifetime) diagnostic interviews at follow-up. Making diagnoses from DIS data. In Epidemiologic Field Methods in Psychiatry:ThePsychiatry: The NIMH Epidemiologic && Catchment Area Program (eds( e d s W.W.EatonW. W. E a t o n & L.G.L . G . It was not possible to adjust incidence rates using data obtained from psychiatrists. Kessler), pp. 209^231.Orlando, FL: Academic Press.

Cox, B. J., McWilliams, L. A., Clara, I. P., et aletal (2003)(2003) The structure of feared situations in a nationally representative sample. Journal of Anxiety Disorders,, 1717,, O.JOSEPHBIENVENU,MD,PhD,CHIADIU.ONYIKE,MB,BS,MHS,DepartmentofPsychiatryandBehavioral 89^101.89^101. Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland; MURRAYB. STEIN, MD,MD,MPH, MPH, Craske, M. G. (1996) Is agoraphobic avoidance Departments of Psychiatry and Family and Preventive Medicine,UniversityMedicine, University of California, San Diego,California;Diego, California; secondary to panic attacks? In DSM^IVDSM ^ IV Sourcebook LI-SHIUN CHEN, MD, ScD,Department of Psychiatry,Washington University School of Medicine, St Louis, (edsT. A.Widiger, A. J. Frances, H. A. Pincus, et al), Missouri; JACK SAMUELS, PhD,GERALD NESTADT,MB, BCh, MPH, Department of Psychiatry and Behavioral pp. 448^459.Washington,DC:448^459.Washington, DC: American Psychiatric Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland; WILLIAM W.EATON, PhD, Association.Association. Department of Mental Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, E a t o n , W. W. & Ke y l , P. (1 9 9 0 ) Risk factors for the onset Maryland, USA of Diagnostic Interview Schedule/DSM^Schedule/DSM^III III agoraphobia in a prospective, population-based study. Archives ofofArchives CorrespCorrespondence:ondence: Dr O.J.Bienvenu, 600 North WWolfeolfe Street, Meyer101,Baltimore, MD 21287,USA. General Psychiatry,, 47,,819^824. 819^824. Tel: +1 410 614 9063; fax: +1 410 614 5913; e-mail: jbienven@@jhmi.edu Eaton,W.W., Anthony, J. C., Gallo, J., et aletal (19 9 7) Natural history of Diagnostic Interview Schedule/ (First received 4 March 2005, final revision 6 June 2005, accepted 1July1 July 2005) DSM ^ IV major depression: the Baltimore Epidemiologic Catchment Area follow-up. Archives of General Kessler, L. G., Folsom, R., Royall, R., et aletal (19 8 5) population characteristics. Archives of General Psychiatry,, Psychiatry,, 54, 993^999. Parameter and variance estimation.Inestimation. In Epidemiologic Field 4141,934^941., 934^941. Eaton,W.W., Anthony, J. C., Romanoski, A., et aletal Methods in Psychiatry:Psychiatry:The The NIMH Epidemiologic Robins, L. N., Helzer, J. E., Croughan, J., et al (19 81) (19 9 8) Onset and recovery from panic disorder in the Catchment Area Program (eds( e d s W.W.EatonW. W. E a t o n & L.G.L . G . National Institute of Mental Health Diagnostic Interview Baltimore Epidemiologic Catchment Area follow-up. Kessler), pp. 327^349.Orlando, FL: Academic Press. Schedule: its history, characteristics and validity. ArchivesArchives British Journal of Psychiatry,, 173, 501^507. Kessler, R. C., McGonagle, K. A., Zhao, S., et aletal of General Psychiatry,, 3838,381^389., 381^389. Faravelli, C., Abrardi, L., Bartolozzi, D., et aletal (2004) (19(1994) 94) LifetimeLifetimeand12-monthprevalenceofDSM^III^R and 12-month prevalence of DSM ^ III ^ R Rubio-Stipec, M., Freeman, D. H., Robins, L., et aletal The Sesto Fiorentino study: point and one-year psychiatric disorders in the United States: results from (19 92) Response error and the estimation of lifetime prevalences of psychiatric disorders in an Italian the National Comorbidity Survey. Archives of General prevalence and incidence of alcoholism: experience in a community sample using clinical interviewers. PsychiatryPsychiatry,, 5151,,8^19. 8^19. community survey. International Journal of Methods in Psychotherapy and Psychosomatics,, 7373, 226^234. Klein, D. F. (1980) Anxiety reconceptualized. Psychiatric Research,, 22 217^224. Fava, G. A. & Mangelli, L. (1999) SubclinicalSubclinical Comprehensive Psychiatry,, 2121,,411^427. 411^427. Samuels, J., Eaton,W.W., Bienvenu,O.Bienvenu, O. J., et al (2002)(2002) symptoms of panic disorder: new insights into Prevalence and correlates of personality disorders in a pathophysiology and treatment. Psychotherapy and Krueger,Krueger,R. R. F. (1999) The structure of common mental community sample. British Journal of Psychiatry,, 180180,, Psychosomatics,, 6868, 281^289.,281^289. disorders.disorders. Archives of General Psychiatry,, 5656, 921^926.,921^926. 536^542.536^542. Fava, G. A., Kellner, R. & Zielezny, M. A. (1988) Lelliott,P.,Marks,I.,McNamee,G.,et aletal (19 8 9) Tyrer,P.(1985)Ty r e r, P. (1 9 8 5 ) Neurosis divisible? Lancet,, ii, 685^688. Prodromal symptoms in panic disorder with Onset of panic disorder with agoraphobia: toward an agoraphobia. American Journal of Psychiatry,, 145145,, integrated model. Archives of General Psychiatry,, 4646,, Weissman, M. M., Leaf, P.J., Blazer, D. G., et aletal (19 8 6) 1564^1567.156 4^ 1567. 1000^1004. The relationship between panic disorder and agoraphobia: an epidemiologic perspective. Hayward, C., Killen, J. D. & Taylor,C. B. (2003) The Marks, I. M. (1987aa)) The agoraphobic syndrome (panic Psychopharmacology Bulletin,, 22, 787^791.,787^791. relationship between agoraphobia symptoms and panic disorder with agoraphobia). In Fears, Phobias, and disorder in a non-clinical sample of adolescents. Rituals: Panic, Anxiety, and Their Disorders, pp. 323^361. Wing, J. K., Babor,T., Brugha,Brugha,T., T., et al (1990) SCAN: Psychological Medicine,, 33, 733^738.,733^738. New York: Oxford University Press. Schedules for Clinical Assessment in Neuropsychiatry. Archives of General Psychiatry,, 47, 589^593.,589^593. Hettema, J. M., Prescott, C. A., Myers, J. M., et aletal Marks, I. M. (1987bb)) Phobic and obsessive ^ compulsive (2005)(2005) The structure of genetic and environmental risk phenomena: classification, prevalence, and relationship Wittchen, H.-U., Reed,V. & Kessler, R. C. (1998) TheThe factors for anxiety disorders in men and women. to other problems. In Fears, Phobias, and Rituals: Panic, relationship of agoraphobia and panic disorder in a Archives of General Psychiatry,, 6262,182^189.,182^189. Anxiety, and Their Disorders, pp. 290^296. NewYork:New York: community sample of adolescents and young adults. Oxford University Press. Archives of General Psychiatry,, 5555,,1017^1024. 1017^1024. Horwath, E., Lish, J. D., Johnson, J., et aletal (19 93)93)(19 Agoraphobia without panic: clinical reappraisal of an Regier, D., Myers, J., Kramer, M., et aletal (19 8 4) TheThe WorldWorldHealthOrganization(1993) Health Organization (1993) The ICDICD^10The ^10 epidemiologic finding. American Journal of Psychiatry,, 150150,, NIMH Epidemiologic Catchment Area program: Classification of Mental and Behavioural Disorders: 1496^1501.1496^1501. historical context, major objectives, and study Diagnostic Criteria for Research..Geneva: Geneva: WHO.

438

Downloaded from https://www.cambridge.org/core. 26 Sep 2021 at 20:36:29, subject to the Cambridge Core terms of use.