iv. medications and other biological treatments 2001.pdf · iv. medications and other biological...

IV. Medications and Other Biological Treatments

Sid ney H Ken nedy, MD, FRCPC1, Ray mond W Lam, MD, FRCPC2, Ni cole L Co hen, MA3, Arun V Rav in dran, MD, PhD, FRCPC4, and the CAN MAT De pres sion Work Group5

Back ground: The Ca na dian Psy chi at ric As so cia tion and the Ca na dian Net work for Mood and Anxi ety Treat ments part nered topro duce clini cal guide lines for psy chia trists for the treat ment of de pres sive dis or ders.

Meth ods: A stan dard guide lines de vel op ment pro cess was fol lowed. Rele vant lit era ture was iden ti fied us ing a com put er izedMed line search sup ple mented by re view of bib li og ra phies. Op era tional cri te ria were used to rate the qual ity of sci en tific evi -dence, and the line of treat ment rec om men da tions in cluded con sen sus clini cal opin ion. This sec tion, “Medi ca tions and OtherBio logi cal Treat ments,” is 1 of 7 ar ti cles that were drafted and re viewed by cli ni cians. Re vised drafts un der went na tional and in -ter na tional ex pert peer re view.

Re sults: Evidence- based rec om men da tions are pre sented for 1) choos ing an an ti de pres sant, based on ef fi cacy, tol er abil ity, andsafety; 2) the op ti mal use of an ti de pres sants, in clud ing aug men ta tion, com bi na tion, and switch ing strate gies; 3) main te nancetreat ment; and 4) elec tro con vul sive ther apy (ECT), light ther apy, and ad di tional so matic treat ments. Evi dence from metaana ly -ses is pre sented first, fol lowed by con clu sions from ran dom ized con trolled tri als (RCTs) and, if ap pro pri ate, open- label data.

Con clu sions: There is sig nifi cant evi dence to sup port the role of se lec tive se ro tonin re up take in hibi tors (SSRIs), novel agents,and clas sic agents in the treat ment of ma jor de pres sive dis or der (MDD). There is also evi dence to sup port the use of so matictreat ments, in clud ing ECT and light ther apy, for some pa tients with MDD. There is lim ited evi dence for the use of spe cific medi -ca tions to treat sub types of MDD. There is emerg ing evi dence to sup port aug men ta tion and com bi na tion strate gies for pa tientspre vi ously non re spon sive to medi ca tion.

IN TRO DUC TION

In an ideal world, spe cific pa tient pro files would de ter minespe cific an ti de pres sant treat ments. To date, the evi dence

from ran dom ized con trolled tri als (RCTs) for such tar getedphar ma co ther apy is lim ited. Many vari ables are likely to in -flu ence the out come of treat ment, in clud ing the popu la tionun der in ves ti ga tion (for ex am ple, sub types within ma jor de -pres sion, in pa tient or out pa tient status, se ver ity, age, sex, cul -ture, or comor bid psy chi at ric and medi cal ill ness), thedu ra tion of treat ment (acute or main te nance), and clini caltrial con di tions (for ex am ple, placebo- controlled, active- drug com pari son, or both).

The in ter pre ta tion of data may also be in flu enced by how out -come is de fined. The gen er ally ac cepted meas ure of re sponseis a 50% re duc tion in symp toms (usu ally on a 17- item Ham il -ton De pres sion Rat ing Scale [HDRS]); this is the ac cepted

stan dard for ap proval of an ti de pres sants by regu la tory bod -ies. Re mis sion, on the other hand, re quires a nearly com pletelevel of symp tom re duc tion; this has been ap plied as an out -come cri te rion in re cent tri als (1,2).

Most evidence- based rec om men da tions are de rived fromrela tively brief acute phar ma co ther apy tri als in volv ing care -fully se lected adult pa tients with vari ous phe no types of ma jor de pres sive dis or der (MDD). In most in stances, these aremono ther apy tri als, with sig nifi cantly fewer stud ies of aug -men ta tion and com bi na tion strate gies. De spite these limi ta -tions, there has been an in creas ing ten dency to pre scribepo lyphar macy for pa tients with de pres sion (3). More re -cently, evi dence for main te nance of ef fi cacy up to at least 1year has been a regu la tory re quire ment for newer an ti de pres -sant agents.

The goal of this sec tion is to pro vide the cli ni cian withevidence- based rec om men da tions for (1) choos ing an an ti de -pres sant based on ef fi cacy, tol er abil ity, and safety; (2) op ti -mal use of an ti de pres sants, in clud ing aug men ta tion,com bi na tion, and switch ing strate gies; (3) main te nance treat -ment; and (4) other bio logi cal treat ments, in clud ing elec tro -con vul sive ther apy (ECT), light ther apy, and other so matictreat ments. Evi dence from metaana ly ses is pre sented first,fol lowed by con clu sions from RCTs and, if ap pro pri ate,open- label data.

CHOOS ING AN AN TI DE PRES SANT

Cli ni cians have 3 ma jor classes of an ti de pres sants to choosefrom. Clas sic agents, mainly tri cyc lic and other het ero cyc lican ti de pres sants (TCAs) in clud ing amitrip tyline, amoxap ine,

The Ca na dian Journal of Psy chia try 38S

Clini cal Guide lines for the Treat ment of De pres sive Dis or ders

1Professor and Cameron Parker Holcombe Wilson Chair in DepressionStudies, Department of Psychiatry, University of Toronto, Toronto, Ontario.2Pro fes sor and Head, Di vi sion of Mood Dis or ders, De part ment of Psy chia -try, Uni ver sity of Brit ish Co lum bia, Van cou ver, Brit ish Co lum bia.3Re search As so ci ate, De part ment of Psy chia try, Uni ver sity of To ronto, To -ronto, On tario.4Pro fes sor, De part ments of Psy chia try and Cel lu lar and Mo lecu lar Medi -cine, Uni ver sity of Ot tawa, Ot tawa, On tario.5 The mem bers of the CAN MAT De pres sion Work Group in clude Sidney HKen nedy (co- chair), Ray mond W Lam (co- chair), Mur ray W Enns, Stan ley P Kutcher, Sa gar V Parikh, Arun V Rav in dran, Robin T Ree sal, Zindel V Segal, Lilian Thorpe, Pi erre Vin cent, and Di ane K Whit ney.

CPA CPA

Copyright 2001 Canadian Psychiatric Association. This guideline is copyrighted and may not be reproduced by any means without prior permission from the Canadian Psychiatric Association. You may print one copy for your own use.

clo mi pramine, de si pramine, doxepin, imi pramine, ma pro ti -line, nor trip tyline, pro trip tyline, and trimi pramine, as well asmonoamine oxi dase in hibi tors (MAOIs)—phe nelz ine andtra nyl cy promine—con tinue to be fa voured in treatment- refractory cases. Se lec tive se ro tonin re up take in hibi tors(SSRIs)—cita lo pram, fluoxet ine, fluvox am ine, par oxet ine,and ser tra line—have been a lead ing class through out the1990s and are con sid ered first- line agents to day. Novelagents (non- TCA agents with nora dren er gic or mixed ac -tions), in clud ing bupropion, mir tazap ine, mo clobe mide, ne -fa zo done, re boxet ine, tra zo done, and ven la fax ine rep re sentan al ter na tive class of agents. Re boxet ine is not rou tinelyavail able in Can ada as of June 2001. De tails of dos ing andme tabo lism for SSRIs and novel agents are pre sented in Ta -ble 4.1.

Al though this clas si fi ca tion can le giti mately be chal lenged,par ticu larly in terms of se ro tonin se lec tiv ity among theSSRIs, it is a frame work from which to de velop guide lines for the use of an ti de pres sants. Cor ner stones in the decision- making pro cess are ef fec tive ness (usu ally in ferred from RCTef fi cacy data), tol er abil ity (side- effect pro files), and safety.

1. How do SSRIs and novel agents compare with classicagents and with each other in the treatment of MDD?

Sev eral metaana ly ses have dem on strated com pa ra ble ef fi -cacy be tween SSRIs and TCAs (4–9). One metaana ly sis,while find ing no dif fer ences be tween in di vid ual SSRIs andTCAs, noted a slight ad van tage to TCAs in a sub group of pa -tients with high HDRS scores (10). Fol low ing Dan ish re portsthat clo mi pramine was more ef fec tive than cita lo pram,

par oxet ine, and mo clobe mide in the treat ment of hos pi tal ized pa tients (11–13), a metaana ly sis of stud ies of in pa tients withde pres sion found that amitrip tyline was sig nifi cantly moreef fec tive than com para tor SSRIs, al though a sen si tiv ityanaly sis in volv ing larger stud ies (100 sub jects) re duced anyTCA ad van tage to a non sig nifi cant trend (14). These re sultswere es sen tially con firmed in an up dated metaana ly sis (4). Itshould be noted, how ever, that an other metaana ly sis foundthat par oxet ine is equiva lent to TCAs in hos pi tal ized pa tientswith de pres sion (15).

For each of the novel agents, there are also metaana ly sesand/or rep li cated RCTs dem on strat ing ef fi cacy com pa ra bleto TCAs al though they have not been sub ject to the same level of metaana lytic rig our (9,16–18). Re cent stud ies have fo -cused on mode of ac tion of novel an ti de pres sants and rela tiveef fi cacy. A metaana ly sis of 105 stud ies did not find any re la -tion ship be tween pos tu lated mecha nism of ac tion of an ti de -pres sants (se ro tonin or nore pi neph rine re up take in hi bi tion,5-HT2 an tago nism, or vari ous com bi na tions of these ef fects)and rela tive ef fi cacy (19). Metaana ly ses in volv ing ven la fax -ine stud ies found greater re sponse and re mis sion rates forven la fax ine, com pared with SSRIs (20,21). The re mis sionrates for SSRIs in these com pari son stud ies are how ever, gen -er ally much lower than those re ported in other SSRI tri als(22,23). Hence, this is sue re mains con tro ver sial.

An other rele vant is sue in these com pari sons is dos ing. Thereis evi dence that an ti de pres sants such as ven la fax ine may in -volve dif fer ent modes of ac tion at dif fer ent dos ages (24). Ametaana ly sis failed, how ever, to dem on strate a dos age–re -sponse re la tion when in di vid ual drug dos ages were con verted to imi pramine equiva lents, and no ad van tage was found fordos ages above the equiva lent of imi pramine 100 to 200 mg

June 2001 IV. Medications and Other Biological Treatments 39S

The Ca na dian Jour nal of Psy chia try

Table 4.1 Pharmacokinetics and dosing of selective serotonin reuptake inhibitors (SSRIs) and novel antidepressantsa

MedicationUsual Effective Dosing (mg daily)

Range(mg daily) Biotransformation Pathways Half-life Protein Binding

SSRI

Cita lo pram 20–40 10–60 Demethylation in 2 steps involvesCYP2C19, 2D6 and 3A4

37 hrs 80%

Fluoxet ine 20–40 10–80 Demethylation involves CYP2D6 4–6 days 95%

Fluvox am ine 100–200 50–300 Demethylation and deaminationinvolves CYP2D6 and 1A2

17–22 hrs 80%

Par oxet ine 20–40 10–60 Oxidation and Demethylation involvesCYP2D6

24 hrs 95%

Ser tra line 50–100 50–200 Demethylation involves CYP3A4 25–26 hrs 98%

Novel

Bupropion SR 150–300 150–300 (divideddose)

Hydroxylation involves CYP2B6 21 hrs 84% (parent)

Mir tazap ine 15–45 15–45 (single dose) Demethylation and hydroxylationinvolve CYP2D6, 1A2 and 3A4

20–40 hrs (parent) 85% (parent)

Ne fa zo done 300–500 300–500 (divideddose)

Dealkylation and hydroxylationinvolve CYP3A4 and 2D6

1.5–4 hrs (parent) 2–4 hrs(HO-Hef 18–33 hrs (TAD)4-9 (mCPP)

99% (parent)

Re boxet ineb 4–8 4–10 (divided dose) Dealkylation and hydroxylationinvolve CYP3A4

13 hrs (parent) 97% (parent)

α-1-acid glycoprotein >albumin

Ven la fax ine IR/XR 75–225 37.5–375 O-desmethylation involves CYP2D6and others

5–7 hrs (parent) 11–13 hrs(ODV)

27% (parent) 30% (ODV)

aAdapted from Ca na dian Phar ma ceu ti cal As so cia tion. Com pen dium of Phar ma ceu ti cals and Spe cial ties. To ronto, Ca na dian Phar ma ceu ti cal As so cia tion, 2001; and Kent JM.SNaRIs, NaS SAs, and NaRIs: new agents for the treat ment of de pres sion. Lan cet 2000; 355:911–8. bNot cur rently avail able in Can ada.

daily (25). Meth ods used to es tab lish dos ing equiva lenceacross an ti de pres sants may be ques tion able and may ac countfor these nega tive find ings.

In the ab sence of clear and sig nifi cant dif fer ences in ef fi cacy,the choice of medi ca tion must be in di vidu al ized for a par ticu -lar pa tient. Sub se quent ques tions will ad dress both pa tientand treat ment fac tors, in clud ing sub type and se ver ity of de -pres sion, tol er abil ity and safety, and se quenced ap proaches to treat ment. Re fer to Sec tion V for com bined psy cho ther -apy–phar ma co ther apy rec om men da tions; Sec tion VI for spe -cific in flu ences of age, sex, and cul ture on treat ment; andSec tion VII for the treat ment of de pres sion as so ci ated withcomor bid physi cal and psy chi at ric con di tions.

Recommendations for Treatment of Major Depressive Disorder

(see Table 4.6)

First-linetreatments

• Se lec tive se ro tonin re up take in hibi tors (SSRIs) andnovel agents (Level 1 evi dence).

• Ven la fax ine may have higher re mis sion rates thanSSRIs (Level 1 evi dence).

Second-linetreatments

• Amitrip tyline and clo mi pramine have greater ef fi cacythan SSRIs in hos pi tal ized pa tients with de pres sion(Level 2 evi dence). Safety and tol er abil ity is sues how -ever, need to be con sid ered.

Third-linetreatments

• Other tri cyc lic an ti de pres sants (TCAs) andmonoamine oxi dase in hibi tors (MAOIs), be cause ofsafety and tol er abil ity is sues (Level 2 evi dence).

2. How does the subtype of depressive disorderinfluence selection of treatment?

MDD with Atypi cal Fea tures

Based on RCTs, phe nelz ine was su pe rior to imi pramine,which in turn was su pe rior to pla cebo (26,27). Fluoxet ine andimi pramine were equally ef fec tive, and both were su pe rior topla cebo, al though fluoxet ine was bet ter tol er ated (28). Ser tra -line and mo clobe mide were also ef fec tive for atypi cal de pres -sion, al though sertraline- treated pa tients had a sig nifi cantlygreater de gree of im prove ment on sev eral psy cho so cial andef fi cacy pa rame ters (29).

Recommendations for Treatment of Major DepressiveDisorder with Atypical Features

(see Table 4.6)

First-line treatments • Fluoxet ine, mo clobe mide, ser tra line(Level 2 evi dence).

Second-line treatments • Phe nelz ine (Level 2 evi dence).

Third-line treatments • Imi pramine (Level 2 evi dence).

MDD with Mel an cholic Fea tures

Most pa tients who meet ad di tional cri te ria for mel an cho liaalso have high lev els of se ver ity, but not all pa tients with “se -vere de pres sion” (for ex am ple, HDRS score > 24) nec es sar ily

have mel an cholic fea tures. There is also an over lap be tweenhos pi tal ized (in pa tient) de pres sion and mel an cho lia (30).

In metaana ly ses, par oxet ine (15), mo clobe mide (16), andven la fax ine (31) were sig nifi cantly more ef fec tive than pla -cebo in treat ing pa tients with mel an cho lia; they were simi larin ef fi cacy to TCA com para tors. Both cita lo pram (32) andfluoxet ine (33) were su pe rior to pla cebo in out pa tients withmel an cho lia. Dan ish stud ies, how ever, found that for hos pi -tal ized pa tients, most with mel an cholic fea tures, re mis sionrates were sig nifi cantly higher with clo mi pramine com paredto par oxet ine, cita lo pram, and mo clobe mide (11–13). Ven la -fax ine also had higher re mis sion rates, com pared withfluoxet ine in some (34), but not all (35), stud ies of pa tientswith mel an cho lia. Nor trip tyline was also sig nifi cantly moreef fec tive than fluoxet ine in a sam ple of mostly older pa tientswith mel an cho lia and con cur rent car dio vas cu lar dis ease(70%) (36). Fur ther re view sup ports the su pe ri or ity of TCAsover SSRIs in mel an cholic de pres sion (37), al though tox ic ityand com pli ance is sues lower the rec om men da tion for TCAs.

Recommendations for Treatment of Major DepressiveDisorder with Melancholic Features

(see Table 4.6)

First-line treatments • Par oxet ine, ven la fax ine (Level 1 evi -dence).

Second-line treatments • Tri cyc lic an ti de pres sants (TCAs), mo clobe mide (Level 1 evi dence).

Third-line treatments • Cita lo pram, fluoxet ine (Level 2 evi dence).

MDD with Psy chotic Fea tures (De lu sional De pres sion)

Few RCTs were spe cifi cally de signed to evalu ate ef fi -cacy in psy chotic or de lu sional de pres sion, and there areno placebo- controlled stud ies. Most of the stud ies com -pare mixed psy chotic and nonpsy chotic de pres sion co -hor t s who re ce ived na tu ra l i s t i c t rea t ment . Thecom bi na tion of amitrip tyline plus per phe nazine was sig -nifi cantly su pe rior to ei ther alone (38). A metaana ly sisfound that ECT was sig nifi cantly su pe rior to TCAs alone (39). How ever, only a trend was found for greater ef fi -cacy of ECT over com bi na tion (TCA plus an tipsy chotic) ther apy, with bi lat eral ECT dis tinctly more ef fec tivethan uni lat eral. There was also a trend to ward com bi na -tion TCA/an tipsy chotic treat ment be ing more ef fec tiveover ei ther medi ca tion alone. Re sults of this metaana ly -sis are lim ited by the in clu sion of open- label tri als.Fluvox am ine alone or in com bi na tion with pin dolol maybe ef fec tive for psy chotic de pres sion (40). The samegroup re ported that ser tra line was sig nifi cantly more ef -fec tive than par oxet ine (41), al though con clu sions werelim ited by ex tremely high drop- out rates in the par oxet -ine group. Most stud ies used the older, typi cal an tipsy -chotic medi ca tions. Emerg ing evi dence from open- label stud ies in psy chotic de pres sion sup ports the use ofnewer atypi cal an tipsy chotic medi ca tions (for ex am ple,ol an zap ine [42]) in com bi na tion with an ti de pres sants.Thus, even though there are few data for the atypi cal

40S CLINI CAL GUIDE LINES FOR THE TREAT MENT OF DE PRES SIVE DIS OR DERS Vol 46, Suppl 1

The Ca na dian Journal of Psy chia try

an tipsy chot ics (ol an zap ine, queti apine, and risperi done),they should be con sid ered in com bi na tion treat ments for psy -chotic de pres sion.

Recommendations for Treatment of Major Depressive Disorder with Psychotic Features

(see Table 4.6)

First-line treatments • Elec tro con vul sive ther apy (ECT) or anan tipsy chotic plus an ti de pres sant (Level 1 evi dence).

Second-line treatments • Ol an zap ine plus an an ti de pres sant(Level 3 evi dence).

Not recommended • Mono ther apy with se lec tive se ro toninre up take in hibi tors (SSRIs) (de spiteLevel 2 evi dence that there is lim itedclini cal sup port).

MDD with Sea sonal Pat tern

Clini cal guide lines (43) have now been pub lished for sea -sonal af fec tive dis or der (SAD), equiva lent to sea sonal pat tern speci fier in DSM- IV. Metaana ly ses and rep li cated RCTshave shown that light ther apy is ef fec tive in treat ing SAD (see ques tion 20). There are fewer medi ca tion stud ies. RCTs show that fluoxet ine and mo clobe mide are ef fec tive (44–46),while bupropion (47), cita lo pram (48), and tra nyl cy promine(49) were bene fi cial in open tri als. Com para tive stud ies be -tween light ther apy and an ti de pres sants are not yet com -pleted, so choice of treat ment de pends on pa tient pref er enceand other clini cal fac tors, in clud ing com para tive side ef fects,se ver ity of de pres sion, pres ence of atypi cal fea tures, and mo -ti va tion for treat ment (43).

Recommendations for Treatment of Major Depressive Disorder with Seasonal Pattern

(see Table 4.6)

First-line treatments • Bright- light ther apy (Level 1 evi dence).

Second-line treatments • Fluoxet ine, mo clobe mide (Level 2 evi dence).

Third-line treatments • Bupropion, cita lo pram, tra nyl cy promine(Level 3 evi dence).

“Anx ious” De pres sion

Al though “anx ious” de pres sion is not a sub type of MDDwithin DSM- IV (50), 60% to 90% of in di vidu als with a pri -mary di ag no sis of de pres sion also ex pe ri ence symp toms ofanxi ety (51). In di vidu als with de pres sion who suf fer fromsig nifi cant anxi ety have more se vere de pres sive symp to -matol ogy, greater func tional and psy cho so cial im pair ment,and poorer prog no sis fol low ing treat ment, com pared with in -di vidu als with de pres sion and low lev els of anxi ety (51,52).

In 3 metaana ly ses, mo clobe mide (53), mir tazap ine (54), and ven la fax ine (52) were as ef fec tive as ac tive com para tors(imi pramine, amitrip tyline, and tra zo done) and su pe rior topla cebo. Par oxet ine (55) and ser tra line (56) were as ef fec -tive as clo mi pramine, while fluvox am ine and lo raze pam(57) were com pa ra ble in ef fi cacy. Other stud ies (58,59) also

sup port the ef fi cacy of par oxet ine in re duc ing symp toms ofanxi ety as so ci ated with de pres sion. Clo naze pam aug men ta -tion of fluoxet ine was su pe rior to fluoxet ine alone in the first 3 weeks of a double- blind trial (60).

Recommendations for Treatment of “Anxious” Depression(see Table 4.6)

First-line treatments • Mir tazap ine, mo clobe mide, par oxet ine, ser -tra line, ven la fax ine (Level 1 evi dence).

Second-linetreatments

• Amitrip tyline, fluvox am ine, imi pramine, tra -zo done (Level 1 evi dence).

Not recommended • Lo raze pam or other ben zo di azepines are notrec om mended as mono ther apy, due to con -cerns about de pend ency, but may be used asshort- term ad junc tive thera pies (Level 2 evi -dence).

Chronic Depression/Dysthymic Disorder

From a clini cal per spec tive, pa tients with pure dysthymia andthose who meet di ag nos tic cri te ria for dou ble de pres sion (61)are of ten in dis tin guish able. The du ra tion of symp toms may beas long as 30 years prior to treat ment. Al though the number ofRCTs in dysthymia is lim ited, there is gen eral agree ment thatac tive phar ma col ogi cal treat ments are more ef fec tive than pla -cebo in the short- term treat ment of pa tients with dysthymia(62,63). A metaana ly sis con firmed the equiva lent ef fi cacy ofTCAs, SSRIs (fluoxet ine and ser tra line), MAOIs, and otheragents, in clud ing mo clobe mide and ri tan serin (64). TCAswere as so ci ated with more ad verse events and higher drop- outrates. There were no sig nifi cant dif fer ences in rates of re sponse be tween pa tients with pure dysthymia and those with dou blede pres sion, sug gest ing that such dif fer en tia tion may not haveany treat ment im pli ca tions.

In in di vid ual RCTs, ser tra line is the most in ves ti gated SSRI(65,66). Par oxet ine was su pe rior to pla cebo in pa tients aged60 years or over who were treated in pri mary care set tings(67). Over all, data from RCTs sug gest no dif fer ence in doserange, fre quency of ad verse ef fects, or pla cebo re sponse be -tween dysthymia and ma jor de pres sion sub jects. In open- label stud ies, ven la fax ine (68) and mir tazap ine (69) were ef -fec tive.

Chronic de pres sion is one con di tion in which com bined phar -ma co ther apy and psy cho ther apy may pro duce su pe rior re -sults (70) (see Sec tion V).

Recommendations for Treatment of ChronicDepression/Dysthymic Disorder

(see Table 4.6)

First-line treatments • Fluoxet ine, fluvox am ine, mo clobe mide, ne faza done, par oxet ine, ser tra line (Level 2evi dence).

Second-line treatments • De si pramine, imi pramine (Level 2 evi dence).

Third-line treatments • Mir tazap ine, ven la fax ine (Level 3 evi dence).



Subthreshold Depressions

Pro posed cri te ria for sev eral ad di tional subthresh old de pres -sive dis or ders, in clud ing mi nor de pres sive dis or der and re -cur rent brief de pres sion, are listed in DSM- IV Ap pen dix B(50). There are cur rently lim ited data on op ti mal treat mentsfor these dis or ders. For a re view of pre men strual dys pho ricdis or der (PMDD) see Sec tion VI.

Mi nor De pres sive Dis or der

Mi nor de pres sive dis or der de scribes a sub type of de pres sivedis or der with the same symp toms and du ra tion as ma jor de -pres sion, but the symp toms are fewer in number. It re mains an elu sive cate gory that is poorly char ac ter ized in both clini calprac tice and in re search stud ies. Re marka bly high re sponserates to par oxet ine, and slightly lower rates to ma pro ti line,were re ported (71). Par oxet ine and problem- solving treat -ment were both su pe rior to pla cebo in a pri mary care set ting,in pa tients aged 60 years or older (67). In an open- label trial,fluvox am ine was as so ci ated with a sig nifi cant de crease in de -pres sive symp to matol ogy and im prove ment in psy cho so cialim pair ment (72).

Recommendations for Treatment of Minor Depressive Disorder

(see Table 4.6)

First-line treatments • Par oxet ine (Level 2 evi dence, based onin di vidu als aged 60 years and older).

Second-line treatments • Fluvox am ine (Level 3 evi dence).• Ma pro ti line (Level 2 evi dence).

Re cur rent Brief De pres sion

In re cur rent brief de pres sion, the number and se ver ity ofsymp toms are com pa ra ble with MDD and last at least 2 daysbut less than 2 weeks. Based on small RCTs, fluoxet ine andpar oxet ine were not su pe rior to pla cebo in ei ther de creas ingrate of re cur rence or suicide- attempt rate (73,74). Al thoughcase re ports sup port the use of tra nyl cy promine (75) and mir -tazap ine (76), con trolled stud ies are re quired.

Recommendations for Treatment of Recurrent Brief Depression

(see Table 4.6)

Not recommended • No avail able evi dence for an ti de pres -sant treat ment.

3. What are the differences in side effects acrossantidepressants?

There are no multiple- drug, placebo- controlled stud ies com -par ing side ef fects. There fore, com para tive rates usu ally rep -re sent a com pos ite of clini cal tri als in volv ing dif fer entmeth ods of side- effect re port ing that may not ac cu rately

cap ture the side- effect pro files of drugs in long- term use instan dard clini cal prac tice. An ti cho liner gic and car dio vas cu -lar side ef fects are most preva lent with TCAs, es pe cially inthe eld erly (see Sec tion VI), while gas tro in tes ti nal and otherside ef fects are more prob lem atic with SSRIs and other novelagents. Re cent data sug gest that sex ual dys func tion andweight change have been un der re ported dur ing an ti de pres -sant treat ment. A de tailed fre quency com pari son of side ef -fects as so ci ated with older and newer an ti de pres sants ispre sented in Ta bles 4.2a, 4.2b, and 4.2c (adapted from [77]).

Sex ual Dys func tion

Un til re cently, in for ma tion about sex ual side ef fects has notbeen sys tem ati cally evalu ated in clini cal tri als. There are nometaana lytic stud ies in this area. There is, how ever, evi denceto sug gest that drug- related sex ual dys func tion is more fre -quent in men (78,79). TCAs such as imi pramine re duced sex -ual in ter est in men more than in women (80), whileen joy ment and abil ity to achieve or gasm were more se verelyim paired with phe nelz ine than imi pramine (81). Con trolledstud ies show that 30% to 50% of pa tients on SSRIs ex pe ri -ence im pair ment of de sire and or gasm (79,82,83), whilebupropion (84–86), ne fa zo done (87), and mo clobe mide (88)cause sig nifi cantly less dys func tion. Large un con trolled stud -ies also sug gest that bupropion (89) and mo clobe mide causeless sex ual dys func tion than do SSRIs, while ven la fax ine ap -pears to be in ter me di ate in its ef fects on women (79). Mir -tazap ine also ap pears to be rela tively free of sex ual sideef fects (90). Some of the sex ual side ef fects of SSRIs havebeen used thera peu ti cally. For ex am ple, SSRIs have beenshown to de lay ejacu la tion time in men with pre ma ture ejacu -la tion (91).

Conclusions About Sexual Dysfunction DuringAntidepressant Treatment

(see Table 4.6)

• Low rates of sex ual dys func tion (10%) are re ported with bupropion,mo clobe mide, and ne fa zo done (Level 2 evi dence).

• In ter me di ate rates of sex ual dys func tion (10% to 30%) are re portedwith ven la fax ine (Level 3 evi dence).

• Mod er ate rates of sex ual dys func tion (ap proxi mately 30% to 50%)are re ported with se lec tive se ro tonin re up take in hibi tors (SSRIs)(Level 1 evi dence).

Weight Gain

Weight change has not been a pri mary out come meas ure ofmost an ti de pres sant stud ies. Weight gain dur ing treat mentwith TCAs (par ticu larly amitrip tyline, imi pramine, andtrimi pramine) is well rec og nized (92–94). Up to 15% ofimipramine- treated in di vidu als will ex pe ri ence weight gain(> 5 kg) af ter 16 weeks. Stud ies of 6- month du ra tion sug gestthat TCA- induced weight gain ap proxi mates 1 kg per month.Acute- treatment tri als em ploy ing SSRIs are fre quently as so -ci ated with mod est re duc tions in weight (95). There is, how -ever, no evi dence from main te nance stud ies that weight loss



is sus tained with SSRIs. In fact, there is con flict ing evi dencere gard ing weight gain dur ing main te nance treat ment (95,96).Ne fa zo done has been shown through acute and main te nancestud ies to be weight- neutral (97). Bupropion SR ap pears to be weight- neutral or weight loss- inducing in a dosage- dependent fash ion, with up to 25% of pa tients los ing 2.5 kg ormore in short term tri als (98). Short- term tri als also sug gestthat ven la fax ine is weight- neutral; how ever, there is an ab -sence of con trolled, long- term weight data (97).

Moclobemide- treated pa tients ap pear to be at low risk ofweight gain in acute tri als and its weight- gain po ten tial islower than that of con ven tional MAOIs (99–101). Mir tazap -ine ex hib its sig nifi cantly greater weight gain than does pla -cebo in short- term tri als, which may be in part due to itshis ta min er gic af fin ity (102).

Conclusions about Weight Gain During Antidepressant Treatment

(see Table 4.6)

• Se lec tive se ro tonin re up take in hibi tors (SSRIs), bupropion, mo -clobe mide, ne fa zo done, and ven la fax ine are weight- neutral dur ingthe acute phase of treat ment (Level 2 evi dence).

• Mir tazap ine re sults in weight gain of greater than 7% of bodyweight in up to 14% of pa tients (Level 2 evi dence).

• Data on weight change with main te nance treat ment are in con clu -sive. Low rates of weight gain (10%) are re ported with bupropion,mo clobe mide, and ne fa zo done (Level 2 evi dence).

4. What are the differences in discontinuation rates?

Based on a sum mary of metaana ly ses of dis con tinua tion rates (8,14,103–105), the range for TCA dis con tinua tion is 19% to31%, and the range for SSRI dis con tinua tion is 15% to 25%.The range for dif fer ences in each of the analy ses is 3% to10%.

Conclusions About Discontinuation Rates for Antidepressants

(see Table 4.6)

• There is a mod est but clini cally sig nifi cant dif fer ence in fa vour ofse lec tive se ro tonin re up take in hibi tors (SSRIs) over tri cyc lic an ti de -pres sants (TCAs) (Level 1 evi dence).

• There is in suf fi cient evi dence to re port on novel an ti de pres santscom pared with other agents.

5. What are the differences in potential for drug–druginteractions?

All an ti de pres sants are me tabo lized by 1 or more of the cy to -chrome P450 (CYP) he patic isoen zymes. Phar ma coki neticdrug in ter ac tions may oc cur when an an ti de pres sant or othercon comi tant medi ca tion also acts as an in hibi tor or in ducer of1 or more of these isoen zymes (see Ta bles 4.3 and 4.4).Among the an ti de pres sants, fluvox am ine is a po tent CYP1A2 in hibi tor, fluoxet ine and fluvox am ine are po tent CYP2C9 in -hibi tors, and fluvox am ine also in hib its CYP2C19. Fluoxet ine

and par oxet ine most potently in hibit CYP2D6, while ne fa zo -done is the most po tent in hibi tor of CYP3A4. Pro tein bind ingis also an im por tant vari able to be taken into ac count; it islower for cita lo pram (80%) and ven la fax ine (30%) than forother SSRI and novel an ti de pres sants.

OP TI MAL USE OF AN TI DE PRES SANTS

6. What is a rational sequential approach to managing a patient with depression beyond the first antidepressanttrial?

The gen eral prin ci ples of op ti mal as sess ment (in clud ing sui -cide risk) and man age ment of de pres sion are out lined in Sec -tion II. All an ti de pres sants have a mini mal thera peu tic dose,but some pa tients re quire higher dos ages for op ti mal ef fect.The SSRIs have a rela tively flat dosage- response curve,based on fixed- dosage group com pari sons. Nev er the less, in -di vid ual pa tients may bene fit from higher- than- usual dos ingstrate gies. Medi ca tions like ven la fax ine may have a posi tivedosage- response curve, with higher re sponse rates at thehigher dos age ranges (106).

Fig ure 4.1 (p 48S) il lus trates an evidence- based al go rithmsum ma riz ing treat ment strate gies. Stud ies on aug men ta tion,com bi na tion, and switch ing strate gies are re viewed in thecon text of no re sponse (gen er ally < 20% re duc tion in symp -tom rat ing scales), mini mal re sponse (20% to 50% re duc -tion), par tial re sponse ( > 50% re duc tion, but stillsymp to matic), and full re sponse (symp tom rat ings within thenor mal range).

7. How long do you wait for a clinical response?

There should be some evi dence of at least mini mal or par tialre sponse af ter 4 weeks at a thera peu tic dos age. The prob abil -ity of show ing a clini cal re sponse af ter 6 to 8 weeks of an ti de -pres sant treat ment is less than 20% if there has been mini malor no re sponse af ter 3 to 4 weeks (107–109). The cor ol lary isevi dence that even a mini mal re sponse (for ex am ple, 20% re -duc tion in symp tom rat ings) af ter 2 weeks is a sig nifi cant pre -dic tor of sub se quent re sponse af ter 6 to 8 weeks (110).There fore, some al tera tion of treat ment is in di cated (for ex -am ple, in crease of dose) if there is no re sponse af ter 3 to 4weeks of an ti de pres sant ini tia tion. For mini mal or par tial re -sponses, the cli ni cian can con tinue to wait an other week or soto de ter mine whether there is con tin ued im prove ment, be forede cid ing on an al tera tion of strat egy.

8. What do you do when a patient does not respond?

When a pa tient has shown only par tial or no re sponse to anop ti mized (usu ally an in creased) dose of an an ti de pres sant,the cli ni cian should re evalu ate di ag nos tic is sues (for ex am -ple, bi po lar ity, de pres sive sub type, comor bid ity, or sub stance



abuse) and treat ment is sues (for ex am ple, ad her ence, and side ef fects), in clud ing sui cide re as sess ment. Psy cho ther apystrate gies can be con sid ered at this time (see Sec tion V).

For phar ma cologic strate gies, the cli ni cian has the choice ofswitch ing, aug ment ing, or com bin ing an ti de pres sant medi ca -tions. Aug men ta tion strate gies in volve add ing a medi ca tion

that by it self is not an an ti de pres sant, while com bi na tionstrate gies in volve add ing a sec ond an ti de pres sant. Given thepau city of com para tive data to in form the cli ni cian whether to switch or aug ment/com bine, the psy chia trist and pa tient must con sider sev eral fac tors, in clud ing the side- effect bur den of apar ticu lar medi ca tion, whether there is a par tial re sponse to



Table 4.2a Frequency of side effects to cyclic antidepressants at therapeutic dosagesi

Reaction Amitrip tyline Clo mi pramine De si pramine Doxepin Imi pramine Nor trip tyline Pro trip tyline Trimi pramine Amoxap ine Ma pro ti line

CNS effects

Drowsiness,sedation

X n n X m n * X m m

Insomnia n m n n m * m nh m *

Excitement,hypomaniaa

* * n * m n m * n n

Disorientation/confusion

m n — * n m — m n n

Headache n n * * m * — n n *

Asthenia, fatigue m n n n m m m n n n

Anticholinergiceffects

Dry mouth X X m X X m m m X X

Blurred vision m m n m m n m n n m

Constipation m m n m m m m m X m

Sweating m m n n m * m n n n

Delayedmicturitionb

n n — * m * * * m n

Extrapyramidaleffects

Unspecified ne *e * ne * — — * ne n

Tremor m m n n m m n m n m

Cardiovasculareffects

Orthostatichypotension/dizziness

m m n m X n m m m n

Tachycardia,palpitations

m m m n m n n n m n

ECG changesc mf mf nf nf mf nf mf mf *f *f

Cardiacarrhythmia

n n n n n n n n * *

GI distress n m n * m * — * n n

Dermatitis, rash n n n * n * * * m m

Weight gain (over 6 kg)

X m n m m n * m * m

Sexualdisturbances

n X n n X * * * n *

Seizuresd * *g * * * * * * *g ng

* < 2%, n ≥ 2%, m ≥ 10%, X ≥ 30%, — None reported in literature pe rused. aMore likely in pa tients with bi po lar illness; bPrimarily in the eld erly; cECG abnormalities usually without cardiac in jury; dIn pa tients with out epilepsy; eTar dive dyskinesiareported (rarely); fConduction delays: increased PR, QRS or QTc in ter val; gHigher incidence if dose above 250 mg daily clomipramine, 225 mg daily maprotiline or 300 mg dailyamoxap ine; hNo effect on REM sleep; iAdapted from Bezchlibnyk- Butler and Jef fries (77).

the in dex an ti de pres sant, po ten tial side ef fects of new treat -ments, and pre vi ous medi ca tion his tory.

9. How effective are switching strategies?

An ti de pres sants can be switched ei ther within the same medi -ca tion class (or neu ro chemi cal ac tion) or to a dif fer ent class(or neu ro chemi cal ac tion). Switch ing within the class is apoor choice with the TCAs, but can be an ef fec tive strat egyfor SSRIs. The re sponse rate when switch ing to an other SSRIis gen er ally bet ter when the first SSRI is poorly tol er ated(66%) than when the pa tient is re frac tory (48%) (111). Ex pert

cli ni cians would gen er ally switch pa tients who have not re -sponded to an op ti mized and tol er ated dose of the first drug toa medi ca tion in a dif fer ent class—for ex am ple, from an SSRIto a se ro tonin nore pi neph rine re up take in hibi tor (SNRI), a se -ro tonin an tago nist and re up take in hibi tor (SARI), a nora -drena l ine and do pa mine modu la tor (NDM) or anora drena line re up take in hibi tor (NRI).

When con sid er ing a switch to an other an ti de pres sant,drug–drug in ter ac tions must be con sid ered in choos ing astart ing dos age. For ex am ple, a switch from fluoxet ine, a long half- life SSRI that sig nifi cantly in hib its CYP2D6, to a TCAthat re quires this isoen zyme for me tabo lism (for ex am ple,



Table 4.2b Frequency of side effects to selective serotonin reuptake inhibitors (SSRIs) and novel antidepressants at therapeutic dosagess

Reaction SSRIs Novel agents

Cita lo pram Fluoxet ine Fluvox am ine Par oxet ine Ser tra line Ne fa zo done Tra zo done Bupropion Ven la fax ine Mir tazap ine

CNS effects

Drowsiness,sedation

m m m m m m X n m Xr

In som nia m mf m m m n n m mf n

Ex cite ment,hy po ma niaa

n n m n m n —k mk mk n

Dis ori en ta tion/con fu sion

* m n * * m * n n n

Head ache m m m m m m n m m n

As the nia,fa tigue

m m m m n m m n m m

An ti cho liner gicef fects

Dry mouth m m m m m m m m m X

Blurred vi sion n n n n n m nl m n m

Con sti pa tion n n m m n m n m m m

Sweat ing m n m m n n — m m n

De layedmic tu ri tionb

n n n n * * * n * n

Ex tra py ra mi dalef fects

Un speci fied * ng n n n * ng n n *

Tremor m m m m m * n m n n

Car dio vas cu laref fects

Or thostatichy poten sion/diz zi ness

n m n m m m mm no mo n

Tachy car dia,pal pi ta tions

ne *e *e ne ne *e n n n n

ECG changesc * * * * * * n * * *

Car diacar rhyth mia

* *h * * * * nn * * *

GI dis tress m m X m X m m m X n

Der ma ti tis,rash

* n n * n * * n n *

Weight gain(over 6 kg)

*i *i * ni *i — n *i *i X

Sex ualdis tur bances

n X j X Xj Xj * *j *p,j X j *

Seizuresd * * * * * * * *q * *

*< 2%, n ≥ 2%, m ≥ 10%, X ≥ 30%, — None re ported in lit era ture pe rused.aMore likely in pa tients with bi po lar illness; b Primarily in the eld erly; cECG abnormalities usually without cardiac in jury; dIn pa tients with out epilepsy; risk increased withelevated plasma lev els; e De creased heart rate re ported; fEspecially if given in the even ing; gTar dive dyskinesia reported (rarely); hSlow ing of sinus node and atrial dys rhyth mia;iWeight loss reported ini tially; jPriapism re ported; kLess likely to precipitate ma nia; lFound to lower intraocular pres sure; mLess frequent if drugs given after meals; nPa tients withpreexisting cardiac disease have a 10% incidence of premature ventricular con trac tions; oHypertension re ported; pIm proved sexual func tion ing; qHigher incidence if doses usedabove 450 mg daily of bupropion or in patients with bu li mia; rSe da tion decreased at higher doses (above 15 mg); sAdapted from Bezchlibnyk- Butler and Jef fries (77).

de si pramine) in di cates that lower- than- usual start ing dos ages of the TCA should be used. This is also a situa tion whereplasma TCA moni tor ing would be ap pro pri ate.

Gen er ally, there is no need to stop one an ti de pres sant for atime be fore start ing an other. With most drugs, the first an ti de -pres sant can be ta pered while start ing an other, but pa tientsmay ex pe ri ence ad di tive side ef fects in the over lap pe riod.For a pa tient who has not tol er ated the first an ti de pres sant, itmay be wise to pro vide a wash out pe riod, al low ing the sideef fects to dis si pate be fore start ing a sec ond an ti de pres sant.Clas sic MAOIs re quire a 2- week medi ca tion wash out prior to start ing an other an ti de pres sant, while a 3- day wash out formo clobe mide is rec om mended (see Ta ble 4.5).

10. How effective are augmentation strategies?

Aug men ta tion strate gies are among the best vali dated phar -ma cologic treat ments for re frac tory de pres sive dis or ders.There are, how ever, sig nifi cant limi ta tions to our knowl edge,

be cause most of the stud ies in -volve small sam ple sizes andre port aug men ta tion of TCAsmore of ten than SSRIs or otheragents. There are also fewplacebo- controlled tri als ofaug men ta tion strate gies, andthere are even fewer di rectcom pari sons of dif fer ent aug -men ta tion strate gies.

Lith ium aug men ta tion has themost evi dence sup port ing itsuse. Two metaana ly ses haveshown that lith ium aug men ta -tion is ef fec tive in about 60%of re f rac tory pa t ients(112,113). Lith ium should begiven at dos ages of greaterthan 750 mg daily, or at a dos -age that achieves se rum lev elsof at least 0.5 meq/L. A sug -gested dos age sched ule is 600mg daily for 1 week, in creas -ing to 900 mg daily for 1 week,and then ti trat ing to ade quatese rum lev els for an other week.If there is no re sponse af ter 3 to4 weeks, then al ter nate strate -gies can be used. Lith ium aug -men ta tion is as so ci ated withthe usual side ef fects of lith ium use.

Triio do thy ronine (T3, lio thy -ronine) has also been shown tobe ef fec t ive in placebo- controlled RCTs. One RCTfound com pa ra ble ef fi cacy be -tween T3 and lith ium, both of

which were su pe rior to pla cebo (114). T3 also ap pears to bemore ef fec tive than thy rox ine (T4) (115). A metaana ly sis,how ever, showed equivo cal re sults, due to the in flu ence of asin gle larger nega tive study (116). T3 is usu ally started at adose of 25 mcg daily and in creased to 50 mcg af ter 1 week, ifnec es sary. With no re sponse at the higher dose for 2 weeks,an other strat egy should be used. T3 is gen er ally well tol er ated.

Both lith ium and T3 stud ies have pri mar ily in volved pa tientswho were re frac tory to mono ther apy with TCAs or MAOIs.Only 2 stud ies ex am ined lith ium aug men ta tion in SSRI non -re spond ers (with cita lo pram [117] and fluoxet ine [118]).There is only lim ited evi dence to date to sup port lith ium or T3

aug men ta tion of the novel- action an ti de pres sants.

Other strate gies have fo cused on SSRI non re spond ers. Bus pi -rone, a par tial post synap tic 5-HT1A ago nist with cate cho la -mine ef fects, was bene fi cial in a number of open- labelstud ies, but a placebo- controlled RCT was nega tive, likely



Table 4.2c. Frequency of side effects to monamine oxidase inhibitors and reversible inhibitor ofMAO-A antidepressants at therapeutic dosages k

Reaction Iso car boxa zid j Phe nelz ine Tra nyl cy promine Mo clobe mide

CNS effects

Drowsi ness, se da tion n m m n

In som nia ne me me me

Ex cite ment, hy po ma niaa n m m m

Dis ori en ta tion/con fu sion n n n n

Head ache m n — m

As the nia, fa tigue n * * *

Anticholinergic effects

Dry mouth m X m m

Blurred vi sion n m n m

Con sti pa tion n m n n

Sweat ing * n — n

De layed mic tu ri tionb n n n *

Extrapyramidal effects

Un speci fied n m * *

Tremor m m n n

Cardiovascular effects

Or thostatic hy poten sion/diz zi ness

m m m m

Tachy car dia, pal pi ta tions — mf mf n

ECG changesc n *g *g n

Car diac ar rhyth mia n * * n

GI dis tress m m n m

Der ma ti tis, rash n * n n

Weight gain (over 6 kg) n m n *

Sex ual dis tur bances n Xh nh *

Sei zuresd — * —i *

* < 2%, n ≥ 2%, m ≥ 10%, X ≥ 30%, — None re ported in lit era ture pe rused.aMore likely in pa tients with bi po lar ill ness; bPri mar ily in the elderly; cECG ab nor mali ties usually with out car diac in jury; dIn pa tientswith out epi lepsy; eEs pe cially if given in the even ing; f De creased heart rate re ported; gShort ened QTc in ter val; hPria pism re ported;iMay have an ti con vul sant ac tiv ity; jNot cur rently avail able in Can ada; kAdapted from Bezchlibnyk- Butler and Jef fries (77).



Table 4.4 Clinically significant substrates for antidepressant-inhibited isoenzymes

The fol low ing ta ble pro vides in for ma tion on po ten tial drug in ter ac tions be tween fre quently pre scribed an ti de pres sants and other medi ca tions.Less con cern when > 20% re nal drug clear ance. Less con cern when > 1 isoen zyme in volved in meta bolic deg ra da tion. More con cern when con comi tant drug has sig nifi cantdose-re lated side ef fects or nar row thera peu tic-to-toxic range. Metabolites may inhibit different isoenzymes or be substrates for different isoenzymes than the parent compound.

Antidepressant Isoenzyme inhibited Drugs whose blood levels may be elevated if used concomitantly with the specified antidepressant

Bupropion 2B6 Cyclophosphamide Ifosphamide Orphenadrine

Citalopram 1A2, 2D6, 2C19 (weak) None known or predicted to be clinically important.

Fluoxetine

(Norfluoxetine)

2D62C9

(2D6)(3A3/4)

Alprazolamb

Amitriptylineb

Astemizolea

Carbamazepined

Cisapridea

Desipramineb,d

Diazepamc

Doxepinb

Flecainidea

Haloperidolc

Imipramineb

Labetalolc

Metoprololc

Mexiletinea

Midazolamb

Nortriptylineb,d

Perphenazinec

Phenytoind

Pindololc

Propafenonea

Propranololc

Risperidonec

Terfenadinea

Thioridazinec

Timololc

Trazadoneb

Triazolamb

S-warfarind (active)

Fluvoxamine 1A23A3/4

Alprazolamb

Amitriptylinec

Carbamazepined

Cisapridea

Clomipraminec

Clozapinec

Desipraminec Haloperidolc

Imipraminec

Midazolamb

Theophyllined

Triazolamb

S-warfarind (active)

Mirtazapine 1A2 None known or predicted to be clinically important

Moclobemide None None known or predicted to be clinically important

Nefazodone 3A3/4 Alprazolamb

Astemizolea

Carbamazepined

Cisapridea

Midazolamb

Paroxetine 2D6 Amitriptylineb

DesipraminebDoxepinb

FlecainideaHaloperidolc

Imipramineb

Labetalolc

Metoprololc

Mexiletinea

Nortriptylinea

Perphenazinec

Pindololc

Propafenonea

Propranololc

Thioridazinec

Timololc

Trazodoneb

Sertraline 1A2, 2C9, 2C19, 2D6,3A4

Venlafaxine 2D6 (weak) None known or predicted to be clinically important

aAvoid; bRe duce dose of sub strate; cMoni tor for side ef fects; dMeas ure drug plasma lev els. No no ta tion—no in for ma tion to in di cate in ter ac tion is of clini cal sig nifi cance. Adapted from Mi cha lets, EL. Up date: clini cally sig nifi cant cy to chrome P450 drug in ter ac tions.Phar ma co therapy 1998;18:84–112.

Table 4.3 Degree of CYP inhibition by SSRIs and novel antidepressants at usually effective dosages

Drug CYP1A2 CYP 2C9 CYP 2C19 CYP 2D6 CYP 3A4

Citalopram Mild None Mild Mild None

Fluoxetine Mild Mild Mild Sig nifi cant Mild

Fluvoxamine Sig nifi cant Mild Mod er ate Mild Mod er ate

Paroxetine Mild Mild Mild Sig nifi cant Mild

Sertraline Mild Mild Mild Mild Mild

Bupropion None None None Mod er ate None

Moclobemide Mild None Mild Mild None

Nefazodone Mild None None Mild Sig nifi cant

Reboxetine None None None Mild Mild

Venlafaxine None None None Mild None

Mirtazapine None None None Mild None

Modi fied from: Preskorn, SH. An ti de pres sant op tions in pri mary care. Clini cal cor ner stone - De pres sion 1999;1:1–16 and Kent, JM. SNaRIs, NaS SAs, and NaRIs: new agents forthe treat ment of de pres sion. Lan cet 2000;355:911–8.

Mild 20% – 50%, Mod er ate > 50% – 150%, Sig nifi cant > 150%



Fig ure 4.1 Evidence- based al go rithm for treat ment of re frac tory ma jor de pres sion. Note that treatment may not be se quen tial—any step of thealgorithm can be bypassed depending on the clinical state of the patient.

LithiumT3 to another

first-line medication

1st line medication

Response?

Maintenance

Full

None

Partial

Response?

Maintenance

Full

None

Partial

SwitchAugment

to a drug with a

differentmechanism

Switch

to a drug with a

differentmechanism

Switch

with a drug with a

different mechanism

Combine

LithiumT3

BuspironeAtypical

antipsychoticPsycho-stimulant

Augment

3

45

8

9

7 10

Level 1 Evidence

Level 2 Evidence

Level 3 Evidence

Level 1 Evidence

Level 2 Evidence

Level 3 Evidence

1

2

6



Notes to Figure 4.1 Evidence-based algorithm for treatment of refractory major depression

1. First- line medi ca tion is cho sen based on clini cal fac torssuch as pre vi ous re sponse, de pres sive sub type, comor bid ity,side ef fects, and po ten tial for drug–drug in ter ac tions. Dosesshould be in creased every 2 to 4 weeks as nec es sary toachieve op ti mal re sponse.

2. At each de ci sion point, when there is par tial or nore sponse, the cli ni cian should

• Re evalu ate di ag nos tic is sues (for ex am ple, sub type,comor bid ity, bi po lar ity, sub stance abuse)

• Re as sess sui cide risk

• Con sider add ing psy cho ther apy (see Sec tion V)

• Con sider elec tro con vul sive ther apy (ECT) (es pe cially ifhigh sui cide risk, se vere or chronic ill ness, psy chotic fea -tures, physi cal de te rio ra tion, or preg nancy).

3. Defi ni tions of re sponse have usu ally used scores fromstan dard ized rat ing scales like the Ham il ton De pres sionRat ing Scale (HDRS). Gen er ally, mini mal im prove ment isde fined as less than 20% re duc tion in HDRS scores com pared with base line; par tial re mission as 20% to 50% re duc tion inHDRS score, or greater than 50% re duc tion in HDRS butre sid ual score still out side the nor mal range; remission as ascore within the nor mal range. Busy cli ni cians may be morelikely to use global rat ing scales like the Clini cal GlobalIm pres sion Im prove ment Scale—very much im proved(remission), much im proved or mini mally im proved (par tialre mission), not im proved or worse (no re sponse).

4. Switch ing to an other an ti de pres sant with a dif fer entneu ro chemi cal ac tion is generally rec om mended if there is no re sponse af ter op ti miz ing the first an ti de pres sant. Switch ingfrom one selective serotonin reuptake inhibitor (SSRI) toan other can be con sid ered, but re sponse rates are higher inpa tients who are SSRI- intolerant rather than nonresponsive.

5. Many ex perts hold off aug men ta tion un til af ter thesec ond an ti de pres sant; how ever, there is some opin ion thataug men ta tion can be tried if there is a par tial re sponse withthe first medi ca tion. In this situa tion, only vali datedaug men ta tion strate gies with Level 1 evi dence (lith ium, T3)should be con sid ered. Note that lith ium and T3 aug men ta tionhas only been stud ied with tricyclic antidepressant (TCAs)and SSRIs.Fac tors to con sider in de cid ing be tween aug men ta tion andswitch ing af ter the first an ti de pres sant:

• Aug men ta tion strate gies us ing lith ium and T3 are the bestvali dated treat ments.

• Side ef fects of aug men ta tion, es pe cially with lith ium, aregen er ally greater than with an ti de pres sant mono ther apy.

• Bene fits of aug men ta tion in clude build ing on a par tial re -sponse, rapid on set of ef fect, al low ing a longer time onthe ini tial an ti de pres sant, main tain ing thera peu tic op ti -mism with pa tients.

• Bene fits of switch ing to an other mono ther apy in clude:sim pler treat ment, fewer side ef fects, no con cerns aboutdrug–drug in ter ac tions, bet ter com pli ance.

6. Main te nance medi ca tions should be con tin ued at thesame dose for at least 6 months. Longer-term main te nance (at least 2 years) should be used for fre quent epi sodes (2 or more in 5 years), re cur rent epi sodes (3 or more, life time), chronicepi sodes, se vere epi sodes (for example, with psy choticsymp toms or marked sui ci dal idea tion), difficult- to- treatepi sodes, and in older-age pa tients.

7. Af ter an un suc cess ful aug men ta tion, con sider com bin ingan other an ti de pres sant for par tial or no re sponse.

8. Af ter 2 second- generation an ti de pres sants with dif fer entneu ro chemi cal ac tions have been tried, con sider a TCA(nor trip tyline or de si pramine) with thera peu tic drugmoni tor ing (se rum lev els), or a monoamine oxidase inhibitor(MAOI).

9. There is now con sid er able theo reti cal ra tion ale tocom bine an ti de pres sants with dif fer ent neu ro chemi cal ac tions for mono ther apy non re spond ers. How ever, there is as yetonly lim ited evi dence to sup port com bi na tion an ti de pres santtreat ment. Only open- label case se ries (Level 3 evi dence) areavail able. This lim ited evi dence sug gests that SSRI plusde si pramine, SSRI plus mo clobe mide, SSRI plus bupropion,bupropion plus ven la fax ine, TCA plus ven la fax ine,mir tazap ine com bi na tions, and the older TCA plus MAOIcom bi na tion may be bene fi cial for re frac tory de pres sion.Limi ta tions of com bi na tion an ti de pres sants in clude in creasedside ef fects, po ten tial drug –drug in ter ac tions, higher cost,and the pos si bil ity of re sponse to mono ther apy with the newan ti de pres sant.

10. Adding augmentation to combination antidepressants canbe considered for the most refractory patients.

be cause of a high pla cebo re sponse rate (54%) (119). Therehas been much in ter est in pin dolol, a beta- blocking drug that,in low doses, acts as a spe cific an tago nist of the 5-HT1A pre -synap tic auto re cep tor. A large RCT of pin dolol to aug mentSSRIs and clo mi pramine in re frac tory de pres sion was nega -tive, how ever, with pin dolol and pla cebo hav ing iden ti cal re -sponse rates of ap proxi mately 13% (120). Fi nally, a smallRCT in pa tients with nonpsy chotic de pres sion re frac tory tofluoxet ine showed that ad di tion of ol an zap ine was more ef -fec tive than ei ther drug alone (121).

11. How effective are combination antidepressantstrategies?

While the use of com bi na tion an ti de pres sant ther apy is com -mon in clini cal prac tice (122), there are few RCTs to sup portthis strat egy in treat ing re frac tory de pres sion. Com binedMAOI and TCA ther apy was not su pe rior to treat ment withei ther agent alone (123), nor was the com bi na tion of de si -pramine and fluoxet ine su pe rior to ei ther high- dose fluoxet -ine alone or lith ium aug men ta tion of fluoxet ine (124). Openstud ies of com bi na tion an ti de pres sant medi ca tions have beende scribed for an SSRI plus mo clobe mide, an SSRI plusbupropion, bupropion plus ven la fax ine, and a TCA plus ven -la fax ine, as well as for mir tazap ine com bi na tions. The ra tion -ale for com bin ing medi ca tions, how ever, is more sound nowthat there are an ti de pres sants with very dif fer ent neu ro chemi -cal ac tions. Pre limi nary stud ies sug gest that com bin ingfluoxet ine with de si pramine is ef fec tive for re frac tory de pres -sion but must be used with cau tion (125). Other open- labelstud ies have sug gested that vari ous com bi na tion an ti de pres -sant treat ments are bene fi cial when there has not been re -sponse to a sin gle an ti de pres sant, but no con trolled stud ies are yet avail able. More re search on com bi na tion an ti de pres santtreat ment is ur gently needed.

12. What are the relative benefits of switching versusaugmentation/combination?

Ad van tages of switch ing to an other an ti de pres sant, com -pared with aug men ta tion/com bi na tion strate gies, in clude thesim plic ity of mono ther apy and the lack of po ten tialdrug–drug in ter ac tions. Mono ther apy may also, but not nec -es sar ily, have fewer side ef fects and bet ter ad her ence. Po ten -tial ad van tages of aug men ta tion, com pared with switch ing,in clude main te nance of thera peu tic op ti mism (not “giv ingup” on a drug), avoid ance of po ten tial dis con tinua tion symp -toms, and the pos si bil ity of rapid re sponse for some aug men -ta tion agents. Add ing an other agent also “buys more time” onthe ini tial medi ca tion and may build on par tial re sponses.

The bene fits of com bin ing an ti de pres sants in clude the pos si -ble bene fits of us ing mul ti ple neu ro chemi cal ac tions, build -ing on a par tial re sponse, and us ing lower doses of each agent, com pared with mono ther apy. The draw backs of com bi na tionan ti de pres sant use in clude the pos si bil ity that the pa tient

might sim ply re spond to mono ther apy with the sec ond agent,the risk of ad di tive side ef fects and poorer com pli ance, andthe cost of us ing mul ti ple medi ca tions.

Recommendations for Managing Nonresponse to an Antidepressant

(see Table 4.6)

Once an antidepressant is selected, an initial improvement (at least 20%reduction in depression scores) should be seen within 3 to 4 weeks.Otherwise, the following interventions are indicated:

First-line treatments • Op ti mize the an ti de pres sant by in creas ingthe dose as tol er ated (Level 2 evi dence).

Second-line treatments • Switch to an an ti de pres sant with a dif fer -ent neu ro chemi cal ac tion (Level 2 evi -dence).

• Aug ment with lith ium or triio do thy ronine(T3) (Level 1 evi dence).

Third-line treatments • Switch to an an ti de pres sant with a simi larneu ro chemi cal ac tion (Level 2 evi dence).

• Aug ment with bus pi rone or an atypi calan tipsy chotic such as ol an zap ine (Level 2evi dence).

• Com bine with an other an ti de pres sant(Level 3 evi dence).

Not recommended • Aug ment with pin dolol (Level 2 evi -dence).

MAIN TE NANCE THERA PIES

13. How long do you keep patients on an antidepressantonce they are better?

Pre ven tion of re lapse and re cur rence is im por tant be cause de -pres sion is re garded as a chronic and/or re cur rent ill ness. Pre -vi ous at ten tion fo cused on phases of treat ment, with an acutephase last ing 8 to 12 weeks, a con tinua tion phase of treat mentlast ing 4 to 6 months, and a main te nance phase last ing 6months to life time (126). In part, this was based on the ideathat the av er age du ra tion for an un treated de pres sive epi sodeis ap proxi mately 6 months. Re turn of symp toms dur ing thispe riod was seen as re lapse of the same epi sode, whereas re -turn of symp toms later in di cated a new epi sode, or re cur -rence. The ob jec tives of treat ment dif fer among phases: forthe acute phase, sup pres sion of symp toms; for the con tinua -tion phase, pre ven tion of re lapse; and for the main te nancephase, pre ven tion of re cur rence.

In clini cal prac tice, how ever, it is of ten dif fi cult or im pos si ble to de ter mine whether a re turn of symp toms is a re lapse or a re -cur rence. A re cent metaana ly sis of con tinua tion and main te -nance stud ies in di cates that the pe riod of con tinua tion ormain te nance ther apy does not seem to in flu ence the re lapserate once medi ca tions are dis con tin ued (127). A sin gle con -cept of main te nance treat ment—how long a pa tient shouldstay on an an ti de pres sant once the pa tient is bet ter—is sim -pler and more rele vant to the clini cal situa tion.



Rec om men da tions for the du ra tion of an ti de pres sant ther apyhave gen er ally been ex tended dur ing the past dec ade, and thisim por tant as pect of de pres sion man age ment con tin ues to bein flu enced by new data. For pa tients with a sin gle epi sode ofde pres sion, the risk of re lapse is higher in the first 6 monthsfrom the time of full re mis sion of symp toms. Af ter that time,the dif fer ence in re lapse rates be tween an ac tive an ti de pres -sant and a pla cebo may not be sig nifi cant (128). There fore, all pa tients should stay on an ti de pres sants for a mini mum of 9months (8 to 12 weeks or more of ac tive treat ment lead ing tofull re mis sion, fol lowed by 6 months of main te nance treat -ment). Pa tients with risk fac tors need longer main te nancetreat ment (see be low). The dos age of the an ti de pres sant in the main te nance phase should be the same as in the acute phase.

There are few data on the du ra tion of main te nance treat mentfor aug men ta tion or com bi na tion strate gies. Af ter suc cess fullith ium aug men ta tion, pa tients should be kept on lith ium andan ti de pres sants for 6 months or longer (129).

14. Who should be maintained longer on anantidepressant?

Un for tu nately, there are no con sis tent, evidence- based bio -logi cal mark ers that pre dict re cur rence. Dif fer ent demo -graphic and clini cal pa rame ters have been iden ti fied as riskfac tors for re lapse and re cur rence. Pa tients with the fol low ing risk fac tors should be main tained on medi ca tions for a mini -mum of 2 years: older age, chronic epi sodes, se vere or life- threatening epi sodes, psy chotic epi sodes, difficult- to- treatepi sodes, re cur rent epi sodes (3 or more, life time), and fre -quent epi sodes (2 or more epi sodes in 5 years) (127). At theend of the rec om mended pe riod, there should be a col labo ra -tive re as sess ment by the pa tient and phy si cian about con tinu -ing main te nance medi ca tion. The de ci sion needs to bal ancethe bene fits (that is, pre vent ing re cur rence) and the risks (forex am ple, side ef fects, cost, and in con ven ience) of stay ing onmedi ca tions against the op tion of dis con tinu ing medi ca tionand moni tor ing for re cur rence. If the an ti de pres sant is dis con -tin ued, it should be gradu ally ta pered to avoid dis con tinua -tion symp toms (130,131).

Recommendations for Maintenance Pharmacotherapy(see Table 4.6)

• All pa tients should be main tained on an ti de pres sants for at least 6 months af ter clini cal re mis sion (Level 1 evi dence).

• Pa tients with the fol low ing risk fac tors should be main tained on an ti de pres sants for at least 2 years: older age, psy chotic fea tures,chronic epi sodes, re cur rent epi sodes (3 or more life time), fre quentepi sodes (2 or more in 5 years), difficult- to- treat epi sodes, and se -vere epi sodes (Level 2 evi dence).

• The an ti de pres sant dos age in the main te nance phase should be thesame as in the acute phase (Level 2 evi dence).

• An ti de pres sants should be ta pered slowly to avoid dis con tinua tionsymp toms (Level 3 evi dence).

OTHER BIO LOGI CAL TREAT MENTS FOR DE PRES SIVE DIS OR DERS

Electroconvulsive Therapy

15. How effective is electroconvulsive therapy (ECT) for depressive disorders?

ECT is gen er ally ac cepted to be the most rapid and ef fec tivetreat ment avail able for se vere ma jor de pres sive epi sodes(MDEs). ECT is as so ci ated with a 60% to 80% re mis sion rate, with maxi mum re sponse typi cally at tained af ter 2 to 4 weeks.There are few data, how ever, com par ing ECT with SSRIs and newer an ti de pres sant medi ca tions (132).

ECT is pre domi nantly used to treat re frac tory de pres sion butshould be con sid ered as a first- line treat ment in acutely sui ci -dal or se verely medi cally com pro mised pa tients, in clud ingpreg nant women. Some con trolled stud ies sug gest that pa -tients who are re frac tory to an ti de pres sants have a lower re -sponse rate to ECT (133,134) or higher re lapse rates within 6months (135,136), but natu ral is tic stud ies did not show a re la -tion be tween an ti de pres sant re frac to ri ness and clini cal out -come (137).

16. What are the recommended treatment parametersfor ECT?

Dur ing the acute- treatment phase, ECT should be car ried out2 to 3 times weekly, for up to 6 weeks. The less fre quentsched ule may mini mize some of the im me di ate cog ni tive side ef fects of ECT, but the on set of re sponse is slower (138). Re -cent re search has shown that it is im por tant, es pe cially withuni lat eral elec trode place ment, to op ti mize the elec tri cal doserela tive to the sei zure thresh old of each pa tient at the firsttreat ment. In di vid ual as sess ment of sei zure thresh old is criti -cal be cause there is at least a 12- fold range in sei zure thresh -old among pa tients with de pres sion who are re ceiv ing ECT(139).

Com pared with bi lat eral elec trode place ment, uni lat eralplace ment at a su pra thresh old dose pro duces fewer im me di -ate and per sis tent cog ni tive side ef fects, with out com pro mis -ing treat ment ef fi cacy. An elec tri cal dose of 3 to 6 times thesei zure thresh old is re quired for uni lat eral elec trode place -ment to have the same ef fi cacy as bi lat eral ECT (140). It isim por tant that ECT equip ment be ade quately pow ered to de -liver these su pra thresh old doses (141). Du ra tion of the sei -zure is a less im por tant marker of ef fi cacy thanelec tro en cepha lo gram (EEG) evi dence of high sei zure in ten -sity (142).

Al though ECT is usu ally con ducted on in pa tients, out pa tientECT prac tice is grow ing, largely be cause of its use for main -te nance treat ment. Am bu la tory ECT may be ad min is tered onan out pa tient ba sis for a care fully se lected popu la tion of pa -tients in a fa cil ity that is prop erly equipped to do so, ac cord ing to prac tice guide lines (143).



17. What are the adverse effects associated with ECT?

ECT is a safe pro ce dure, and mor tal ity and mor bid ity rates for ECT are ex tremely low in the mod ern era. Apart from raisedin trac ra nial pres sure, there are no ab so lute con tra in di ca tionsto ECT. Care ful pre an es thetic ex ami na tion is es sen tial, par -ticu larly for pa tients at higher risk, in clud ing those with myo -car dial ische mia, car diac ar rhyth mias or pace mak ers, orab domi nal an eu rysm. Mor tal ity as so ci ated with ECT is es ti -mated to be 0.2 deaths per 100 000 treat ments and is mostlikely re lated to car dio vas cu lar com pli ca tions (144). Theadverse- event rate af ter ECT is around 0.4%. Ad verse eventscan in clude mus cu lo skele tal in ju ries, den tal dam age, oral lac -era tions, and per sis tent my al gia (132). Nau sea, head aches,and mus cle aches post- ECT are com mon but can be treatedwith an tie met ics and an al ge sics, if nec es sary.

The cog ni tive side ef fects as so ci ated with ECT are well docu -mented. Ob jec tive mem ory test ing shows a tran sient ret ro -grade am ne sia that less ens with time, so that no cog ni tivedefi cits are ap par ent at 6 months post- ECT, al though theremay be per sis tent spotty mem ory gaps for events oc cur ringaround the time of the ECT. Sub jec tive mem ory com plaintsin clude im me di ate, and some per sis tent (at least 2 months),defi cits in some auto bio graphi cal mem ory, mostly with im -per sonal memo ries (for ex am ple, pub lic events) rather thanwith per sonal memo ries (145).

Greater cog ni tive side ef fects are gen er ally as so ci ated withuse of bi lat eral elec trode place ment; elec tri cal dos ages at

higher su pra thresh old lev els; treat ment 3 times weekly, com -pared with twice weekly; and per sis tent de pres sive moodstate. Con tinua tion and main te nance ECT treat ment is lesslikely to cause cog ni tive side ef fects, pos si bly due to thelonger time in ter val be tween treat ments (146). There is nocredi ble evi dence that ECT causes any form of struc turalbrain dam age (147). Pro spec tive neu roi mag ing stud ies withCT and MRI show no brain struc tural changes af ter ECT.

18. Should ECT be combined with antidepressantmedication?

There is lit tle evi dence that com bin ing an ti de pres sants withECT im proves the re sponse, es pe cially with pro spec tivestud ies (148,149). Pa tients un der go ing ECT have usu allyfailed an ti de pres sant treat ment, so there is lit tle ra tion ale tocon tinue the same medi ca tion. Start ing a new medi ca tioncauses clini cal con fu sion when pa tients re spond and in de ter -min ing ad verse ef fects. There is no evi dence that start ing amedi ca tion with ECT has any ef fect on re lapse rates, com -pared with start ing a medi ca tion fol low ing the last ECT ses -sion.

There has been a sug ges tion that lith ium is as so ci ated with de -lir ium or pro longed con fu sion af ter ECT, but this re quiresfur ther in ves ti ga tion. The use of seda tives, such as ben zo di -azepines, and an ti con vul sants, such as car ba mazepine andval proic acid, may in hibit sei zures or shorten sei zure



Table 4.5 Washout recommendations for switching antidepressants

Switch ing from Switch ing to

SSRI Novel (mixed action) TCA RIMA MAOI

SSRICita lo pramFluoxet ineFluvox am inePar oxet ineSer tra line

No washoutBe aware of additiveserotonergic effects for up to 1 week after stoppingSSRI(5 weeks for fluoxetine)

No washoutBe aware of additiveserotonergic effects for up to 1 week after stoppingSSRI (5 weeks for fluoxetine)

No washoutBe aware that SSRIs canincrease serum TCAlevels for up to 1 weekafter stopping SSRI (5 weeks for fluoxetine)

1 week(5 weeks for fluoxetine)Some data indicate that an SSRI and RIMA can besafely combined, but thisrequires specialist input.

1 week(5 weeks for fluoxetine)

Novel (mixed action)antidepressants

Bupropion SRMir tazap ineNe fa zo doneTra zo doneVen la fax ineVen la fax ine XR

No washout No washoutVenlafaxine should bestarted at a lower dose toavoid additivenoradrenergic effectsBupropion should bestarted at a lower dose toavoid additivenorardrenergic effects

No WashoutTCA should be started ata lower dose to avoidadditive noradrenergiceffects

3–5 daysNo Washout withbupropion

1 week

TCAAmitrip tyline De si pramineImi pramineNor trip tyline and oth ers

No washoutBe aware that SSRI canincrease serum levels ofTCAs for up to 1 weekafter stopping TCA.Be aware of additiveserotonergic effects whenswitching fromclomipramine to SSRIs.

No washoutVenlafaxine should bestarted at a lower dose toavoid additivenoradrenergic effectsBupropion should bestarted at a lower dose toavoid additivenoradrenergic effects

No washout No washout 1 week

RIMAMo clobe mide 3 days 3 days 3 days Not applicable 3 days

MAOIPhe nelz ineTra nyl cy promine

2 weeks 2 weeks 2 weeks 2 weeks 2 weeks

SSRI = se lec tive se ro tonin re up take in hibi tor; TCA = tri cyc lic an ti de pres sant; RIMA = re versi ble in hibi tor of MAO-A; MAOI = monoamine oxi dase in hibi tor.

du ra tion; there fore, their use with ECT should be mini mized(143,146).

19. How can you prevent relapse after ECT? Howeffective is maintenance ECT?

The re lapse rate with out main te nance treat ment fol low ingECT is high: be tween 50% and 95% of pa tients re lapse within 6 months. Pre dic tors of re lapse in clude pre- ECT medi ca tionre sis tance and greater se ver ity of de pres sion. Main te nancemedi ca tion treat ment for at least 1 to 2 years is nearly al waysin di cated to pre vent re lapse. Lith ium and an ti de pres sants canre duce the re lapse rate. Medi ca tions that were used op ti mallyand failed prior to ECT should not be used for post- ECTmain te nance.

ECT it self may also be used for main te nance treat ment, withtreat ment sched ules rang ing from once weekly to oncemonthly (132,135). There is, how ever, a lack of con trolledand well- defined out come stud ies; evi dence is lim ited to casese ries and re ports. Sug gested in di ca tions for use of main te -nance ECT are as fol lows: a his tory of re spon sive ness to ECTin re cur rent epi sodes; re frac to ri ness to, or in tol er ance of, pro -phy lac tic phar ma co ther apy; highly re cur rent and eas ily re -laps ing symp toms; psy chotic symp toms; and pa tientpref er ence (146).

Recommendations for Electroconvulsive Therapy(see Table 4.6)

• Elec tro con vul sive ther apy (ECT) is an ef fec tive treat ment for ma jorde pres sive dis or der (MDD) (Level 1 evi dence).

• In di ca tions for ECT in clude acute sui ci dal risk, se vere physi cal de -te rio ra tion, psy chotic fea tures, re frac to ri ness to medi ca tions, and pa -tient pref er ence.

• Uni lat eral elec trode place ment re quires su pra thresh old doses ofECT (Level 2 evi dence).

• Side ef fects of ECT are gen er ally mild, with evi dence for a tran -sient, short- term mem ory dis tur bance (Level 2 evi dence).

Light Therapy

20. How effective is light therapy in treating depressivedisorders?

The ef fi cacy of light ther apy for sea sonal mood dis or ders hasbeen es tab lished by sev eral large- sample RCTs com par inglight ther apy to plau si ble pla ce bos (150,151), metaana ly ses(152), and pub lished clini cal guide lines (43). The re sponserate for light ther apy is 60% to 90%, with re sponse oc cur ringwithin 2 to 3 weeks.

The ef fi cacy of light ther apy for non sea sonal de pres sive dis -or ders is more con tro ver sial. The con trolled stud ies havesmall sam ples, short treat ment du ra tions (1 to 4 weeks), andequivo cal re sults; thus, there is cur rently in suf fi cient evi -dence to rec om mend light ther apy for non sea sonal de pres -sion. Its use as an ad junc tive treat ment with medi ca tions may, how ever, be con sid ered for some pa tients (153).

21. What are the recommended treatment parametersfor light therapy?

The fluo res cent light box is the gold stan dard light de vice, be -cause there are not yet ade quate ef fi cacy data for head- mounted de vices and dawn simu la tors (43). The rec om -mended treat ment regi men is 10 000 lux in ten sity for 30 min -utes daily in the early morn ing, as soon as pos si ble af terawak en ing. The light box should have an ul tra vio let fil ter toscreen out harm ful ul tra vio let rays. Re sponse to light ther apygen er ally oc curs within 2 to 4 days, and meas ur able im prove -ment is of ten seen in 1 week, but some re sponses may oc curover 2 to 4 weeks.

Side ef fects of light ther apy, in clud ing eye strain, vis ual dis -tur bances, head ache, agi ta tion or feel ing “wired,” nau sea,sweat ing, and se da tion, are gen er ally mild and sub side withtime or with a re duced dose of light. Light ther apy can pre -cipi tate hy po ma nia or ma nia in sus cep ti ble pa tients. There isno evi dence of ocu lar dam age with proper use of light ther -apy. Nev er the less, oph thal mo logi cal as sess ment and moni -tor ing are rec om mended if pa tients have ocu lar risk fac torsfor po ten tial light tox ic ity, in clud ing pre ex ist ing reti nal dis -ease; sys temic ill nesses that af fect the ret ina (for ex am ple,dia be tes); use of pho to sen si tiz ing medi ca tions such as phe -nothi azine an tipsy chot ics, lith ium, mela tonin, or St John’swort; and the eld erly.

Most pa tients (but not all) ex pe ri ence rapid re cur rence ofsymp toms af ter dis con tinua tion of light ther apy. There fore,pa tients should con tinue us ing light ther apy through out theirpe riod of risk for win ter de pres sion and, usu ally, dis con tinuetreat ment dur ing the as ymp to matic sum mer months. Thereare few data on long- term or main te nance use of light ther apyfor SAD or non sea sonal de pres sion.

Recommendations for Light Therapy(see Table 4.6)

• Light ther apy is an ef fec tive treat ment for re cur rent ma jor de pres -sive dis or der (MDD) with a sea sonal pat tern, of mild- to- moderatese ver ity (Level 1 evi dence).

• An ade quate trial of light ther apy should be 2 to 4 weeks of 10 000lux fluo res cent light for 30 min utes daily, in the early morn ing(Level 2 evi dence).

• Pa tients usu ally need to con tinue daily light ther apy through out thewin ter and can dis con tinue treat ment in the sum mer (Level 3 evi -dence).

• There is in suf fi cient evi dence re gard ing the ef fi cacy of light ther apy for long- term or main te nance use.

22. How effective are other biological treatments fordepressive disorders?

St John’s Wort

Sev eral metaana ly ses have sum ma rized RCTs show ing thatex tracts of Hypericum per fo ra tum, also known as the plant StJohn’s wort, are more ef fec tive than pla cebo and equiva lent



to low- dose an ti de pres sants for the short- term treat ment ofmild- to- moderate de pres sion (154). The dos ages ofHypericum per fo ra tum used in the RCTs ranged from 300 to900 mg daily. Side ef fects of Hypericum per fo ra tum ap pearto be milder and less fre quent than those ex pe ri enced withstan dard an ti de pres sants. These metaana ly ses, how ever, alsohigh lighted se ri ous meth odo logi cal limi ta tions in all the pre -vi ous RCTs (154,155). A large placebo- controlled RCT thatad dressed these limi ta tions found that St John’s wort was notef fec tive for treat ment of mod er ately se vere MDD (121).

Some cau tions for the use of Hypericum per fo ra tum shouldbe noted. There is no regu la tion of dos age or qual ity ofHypericum per fo ra tum prepa ra tions in Can ada or the US.The mecha nism of ac tion is un known, but se ro to ner gic anddo pa min er gic ef fects have been de scribed. There is no evi -dence that it is an MAOI. Hypericum per fo ra tum is as so ci ated with ad verse ef fects, in clud ing pho to sen si tiv ity, and there are stud ies show ing that it has im por tant phar ma coki netic in ter -ac tions with dis pa rate agents, in clud ing im mu noregu la torycom pounds, an ti co agu lants, and an ti in fec tive agents. Drug- induced hy po ma nia and the se ro tonin syn drome have beenre ported. There fore, com bin ing Hypericum per fo ra tum withother medi ca tions, in clud ing an ti de pres sants, should be doneonly with cau tion.

Sleep Dep ri va tion

Many stud ies have shown that to tal sleep dep ri va tion, wherepa tients are kept awake for up to 40 hours, can tem po rar ilyim prove de pres sive symp toms, some times with dra matic re -sults (156–158). The an ti de pres sant ef fect is usu ally tran -sient in that most pa tients re lapse af ter even a brief re cov erysleep; how ever, up to 15% of pa tients in clini cal stud ies canhave sus tained re sponses fol low ing to tal sleep dep ri va tion.Ef forts to pre vent the rapid re lapse af ter to tal sleep dep ri va -tion have in cluded use of modi fied par tial sleep- deprivationpro to cols, es pe cially with sleep dep ri va tion in the sec ond half of the night (for ex am ple, sleep ing from 10:00 pm to 2:00am). Even par tial sleep dep ri va tion, how ever, is dif fi cult forpa tients to con tinue for more than a week. Other strate gies tosus tain the an ti de pres sant ef fect in clude com bin ing sleepdep ri va tion with an ti de pres sant medi ca tions, lith ium, pin -dolol, and bright- light treat ment. Placebo- controlled stud iesare ob vi ously dif fi cult to con duct for such a treat ment, butsev eral con trolled stud ies of sleep dep ri va tion sup port thesecom bined treat ments.

Given the rather sim ple pro ce dure, sleep dep ri va tion can becon sid ered an ad junc tive treat ment to medi ca tions, par ticu -larly when a rapid re sponse is re quired (for ex am ple, inacutely sui ci dal in pa tients). One pro to col sup ported in the lit -era ture in cludes 3 to tal sleep- deprivation pe ri ods within 1week (159). Dur ing each 48- hour pe riod, the pa tient is awakefrom 7 am on day 1 to 7 pm on day 2 (36 hours of sleep dep ri -va tion). This is fol lowed by re cov ery sleep from 7 pm to 7 amon day 3 (12 hours sleep). The next sleep dep ri va tion pe riodthen be gins anew.

Ex er cise

A metaana ly sis con cluded that the few stud ies avail able havemeth odo logi cal short com ings and that con clu sions about ef -fec tive ness could not be made (160). One RCT, how ever, forelders with de pres sion, found that ex er cise was as ef fec tive as ser tra line and the com bi na tion af ter 16 weeks but that the ser -tra line sam ple had a faster on set of re sponse (161). In a 6- month follow- up to this study, the ex er cise sam ple had alower re lapse rate than did the medi ca tion sam ple (162). In -clud ing ex er cise in a treat ment pro gram, when ap pro pri ate,may have sev eral ad van tages in terms of cost and side- effectpro file, and it may there fore be po ten tially use ful as a pre ven -ta tive strat egy.

Sur gi cal Treat ments

Mod ern psy cho sur gery, also called lim bic sur gery, con sistsof stereo tac tic neu ro sur gery that places small, se lec tive le -sions in the brain. Mod ern lim bic sur gery pro grams havestrict cri te ria and guide lines to iden tify ap pro pri ate pa tientsthat have been re frac tory to medi ca tions and ECT and to fol -low re sults of treat ment. The pro ce dures cur rently used arean te rior cin gu lotomy or cap su lotomy (163) and sub cau datetrac totomy (164). There are no con trolled stud ies of lim bicsur gery, but large open se ries in volv ing the most treatment- refractory pa tients have shown that 34% to 62% of pa tientshave good or very good im prove ment with these lim bic sur -ger ies. Mor tal ity and mor bid ity rates are very low. No sig nifi -cant be hav ioural or in tel lec tual defi cits, and no changes inper son al ity, have been re ported for the pro ce dures. Psy cho -sur gery can be con sid ered for pa tients with the most in trac ta -ble ill nesses who are sui ci dal or chroni cally se verelyim paired due to de pres sion.

Vagus Nerve Stimu la tion

Vagus nerve stimu la tion (VNS) is an ap proved treat ment fordrug- refractory epi lepsy that is be ing in ves ti gated for re frac -tory MDD. An elec tri cal wire is sur gi cally im planted in thevagus nerve in the neck and con nected to a stimu la tor lo cated



Table 4.6 Criteria for levels of evidence andlines of treatment recommendations

Level of evidence Criteria

1 Metaanalysis or replicated randomized controlled trial (RCT) that includes a placebo condition

2 At least 1 RCT with placebo or activecomparison condition

3 Uncontrolled trial with 10 or more subjects

4 Anecdotal case reports

Line of Treatment Criteria

First-line Level 1 or Level 2 evidence plus clinical support

Second-line Level 3 evidence or highera plus clinical support

Third-line Level 4 evidence or highera plus clinical support

Not recommended Level 1 or Level 2 evidence for lack of efficacy

aTreat ments with higher lev els of evi dence may be listed as lower lines of treat mentdue to clini cal is sues such as side ef fect or safety pro file.

in the chest wall (165). In an open trial, in ter mit tent stimu la -tion of the vagus nerve re sulted in a 40% re sponse rate intreatment- resistant pa tients (166). Con trolled stud ies are now in prog ress.

Tran scra nial Mag netic Stimu la tion

Tran scra nial mag netic stimu la tion (TMS) in volves the stimu -la tion of cor ti cal neu rons by mag netic in duc tion, us ing abrief, high- intensity mag netic field. The elec tri cal cur rent israp idly turned on and off to pro duce a time- varying mag neticfield last ing for about 100 to 200 mi cro sec onds. For treat -ment, re peti tive stimu la tion with fre quen cies in the range of 1to 20 Hz is con ducted over the course of about 30 min utes.The use of a hand- held wand over the scalp re sults in neu ronal stimu la tion and de po lari za tion to a depth of about 2 cm be lowthe brain’s sur face. Al though pri mar ily used as a di ag nos tictool for neu ro logi cal ex ami na tion, pre limi nary stud ies sug -gest that TMS can ex ert short- term an ti de pres sant ef fects(167).

TMS is an at trac tive po ten tial treat ment be cause it can be con -ducted in out pa tient set tings, does not re quire an es the sia orse da tion, has no ef fects on cog ni tion, and has mini mal side ef -fects, such as mild head ache and dis com fort at the site ofstimu la tion. Re cent sham- controlled, short- duration stud ieshave shown ef fects com pa ra ble with an ti de pres sant medi ca -tions (165). Po ten tial clini cal uses in clude the treat ment ofmedication- refractory pa tients (per haps in place of ECT) or,be cause of its rapid on set of ef fect, to has ten clini cal re sponsein con junc tion with an ti de pres sants. There is still, how ever,no clini cal con sen sus about the op ti mal pa rame ters of treat -ment, in clud ing fre quency and anat omic lo ca tion of stimu la -tion. Most of the posi tive re sults have not yet been rep li cated,the follow- up pe ri ods have been short, and there are no long- term or main te nance stud ies.

Recommendations for Other Biological Treatments (see Table 4.6)

• St. John’s wort may be ef fec tive in treat ing de pres sive dis or ders ofmild se ver ity (Level 1 evi dence) but is not rec om mended for morese vere con di tions (Level 2 evi dence). While side ef fects are mini -mal, there is no regu la tion of Hypericum per fo ra tum in Can ada, andthe drug–drug in ter ac tions are not known.

• Sleep dep ri va tion is an ef fec tive tran sient treat ment for de pres sivedis or ders (Level 2 evi dence). Re sponse may be main tained us ingmedi ca tions (for ex am ple, an ti de pres sants, lith ium, or pin dolol) orbright light (Level 2 evi dence). Sleep dep ri va tion may be most use -ful as an ad junc tive treat ment in hos pi tal ized pa tients.

• Ex er cise alone may be ef fec tive in pa tients with mild de pres sion(Level 2 evi dence) and can be used as an ad junc tive treat ment tofirst- line treat ments (Level 3 evi dence).

• Lim bic sur gery (psy cho sur gery) has lim ited evi dence for ef fi cacybut may be con sid ered for the most re frac tory and chronic cases(Level 3 evi dence).

• Tran scra nial mag netic stimu la tion (TMS) and vagus nerve stimu la -tion (VNS) are prom is ing new bio logi cal treat ments, but there is too lit tle evi dence to war rant rec om men da tions for gen eral clini cal use(Level 2 and 3 evi dence).

AC KNOW LEDGE MENTS

The authors thank the fol low ing for their con tri bu tions: Dr DavidBak ish, Dr Mur ray W Enns, Dr Stan ley P Kutcher, Dr David Gard -ner, Dr Ron ald A Remick, and Dr Roger S McIn tyre con trib uted toear lier drafts of this pa per; Dr Alan J Ge len berg pro vided ex ter nalpeer re view prior to pub li ca tion.

REF ER ENCES

1. Frank E, Prien RF, Jar rett RB, Kel ler MB, Kup fer DJ, La vori PW, and oth ers. Con -cep tu ali za tion and ra tion ale for con sen sus defi ni tions of terms in ma jor de pres sivedis or der. Re mis sion, re cov ery, re lapse, and re cur rence. Arch Gen Psy chia try1991;48:851–5.

2. Rush AJ, Trivedi MH. Treat ing de pres sion to re mis sion. Psy chi atr An nals 1995;25:704–9.

3. Frye MA, Ket ter TA, Leverich GS, Hug gins T, Lantz C, Deni coff KD, and oth ers.The in creas ing use of po lyphar ma co ther apy for re frac tory mood dis or ders: 22years of study. J Clin Psy chia try 2000;61:9–15.

4. An der son IM. Se lec tive se ro tonin re up take in hibi tors ver sus tri cyc lic an ti de pres -sants: a metaanalysis of ef fi cacy and tol er abil ity. J Af fect Dis ord 2000;58:19–36.

5. Joffe R, Sok olov S, Strei ner D. An ti de pres sant treat ment of de pres sion: a metaana -ly sis. Can J Psy chia try 1996;41:613–6.

6. Mol ler HJ, Fuger J, Kas per S. Ef fi cacy of new gen era tion an ti de pres sants: metaanalysis of imipramine- controlled stud ies. Phar ma cop sy chia try 1994;27:215–23.

7. Song F, Free man tle N, Shel don TA, House A, Wat son P, Long A, and oth ers. Se lec -tive se ro tonin re up take in hibi tors: metaanalysis of ef fi cacy and ac cept abil ity. BMJ 1993;306:683–7.

8. Stef fens DC, Krishnan KR, Helms MJ. Are SSRIs bet ter than TCAs? Com pari sonof SSRIs and TCAs: a metaanalysis. De press Anxi ety 1997;6:10–8.

9. Work man EA, Short DD. Atypi cal an ti de pres sants ver sus imi pramine in the treat -ment of ma jor de pres sion: a metaanalysis. J Clin Psy chia try 1993;54:5–12.

10. An der son IM, To men son BM. The ef fi cacy of se lec tive se ro tonin reuptake in hibi -tors in de pres sion: A metaanalysis of stud ies against tri cyc lic an ti de pres sants.Jour nal of Psy cho phar ma col ogy 1994;8:238–49.

11. Dan ish Uni ver sity An ti de pres sant Group. Cita lo pram: clini cal ef fect pro file incom pari son with clo mi pramine. A con trolled mul ti cen ter study. Psy cho phar ma -col ogy (Berl) 1986;1:131–8.

12. Dan ish Uni ver sity An ti de pres sant Group. Par oxet ine: a se lec tive re up take in hibi tor show ing bet ter tol er ance but weaker an ti de pres sant ef fect than clo mi pramine in acon trolled mul ti cen ter study. J Af fect Dis ord 1990;18:289–99.

13. Dan ish Uni ver sity An ti de pres sant Group. Mo clobe mide: a re versi ble MAO- A- inhibitor show ing weaker an ti de pres sant ef fect than clo mi pramine in a con trolledmul ti cen ter study. J Af fect Dis ord 1993;28:105–16.

14. An der son IM. SSRIS ver sus tri cyc lic an ti de pres sants in de pressed in pa tients: ameta- analysis of ef fi cacy and tol er abil ity. De press Anxi ety 1998;7 (Suppl 1):11–7.

15. Tig nol J, Stoker MJ, Dun bar GC. Par oxet ine in the treat ment of mel an cho lia and se -vere de pres sion. Int Clin Psy cho phar ma col 1992;7:91–4.

16. Angst J, Stabl M. Ef fi cacy of mo clobe mide in dif fer ent pa tient groups: a metaanalysis of stud ies. Psy cho phar ma col ogy (Berl) 1992;106 (Suppl):109S–113S.

17. Kas per S. Clini cal ef fi cacy of mir tazap ine: a re view of meta- analyses of pooleddata. Int Clin Psy cho phar ma col 1995;10 (Suppl 4):25–35.

18. Pat ten SB. The com para tive ef fi cacy of tra zo done and imi pramine in the treat mentof de pres sion. CMAJ 1992;146:1177–82.

19. Free man tle N, An der son IM, Young P. Pre dic tive value of phar ma col ogi cal ac tiv -ity for the rela tive ef fi cacy of an ti de pres sant drugs. Meta- regression analy sis. Br JPsy chia try 2000;177:292–302.

20. Einar son TR, Arikian SR, Cas ci ano J, Doyle JJ. Com pari son of extended- releaseven la fax ine, se lec tive se ro tonin re up take in hibi tors, and tri cyc lic an ti de pres santsin the treat ment of de pres sion: a metaanalysis of ran dom ized con trolled tri als. Clin Ther 1999;21:296–308.

21. Thase ME, Ent suah AR, Ru dolph RL. Re mis sion rates dur ing treat ment with ven la -fax ine or se lec tive se ro tonin re up take in hibi tors. Br J Psy chia try 2001;178:234–41.

22. Pat ris M, Bouchard JM, Bougerol T, Char bon nier JF, Cheva lier JF, Clerc G, andoth ers. Cita lo pram ver sus fluoxet ine: a double- blind, con trolled, mul ti cen tre,phase III trial in pa tients with uni po lar ma jor de pres sion treated in gen eral prac tice.Int Clin Psy cho phar ma col 1996;11:129–36.

23. Bea sley CM Jr, Nils son ME, Koke SC, Gon zales JS. Ef fi cacy, ad verse events, andtreat ment dis con tinua tions in fluoxet ine clini cal stud ies of ma jor de pres sion: ameta- analysis of the 20- mg/day dose. J Clin Psy chia try 2000;61:722–8.

24. Har vey AT, Ru dolph RL, Preskorn SH. Evi dence of the dual mecha nisms of ac tionof ven la fax ine. Arch Gen Psy chia try 2000;57:503–9.

25. Bol lini P, Pam pal lona S, Tibaldi G, Ku pel nick B, Mu nizza C. Ef fec tive ness of an ti -de pres sants. Meta- analysis of dose- effect re la tion ships in ran dom ised clini cal tri -als. Br J Psy chia try 1999;174:297–303.

26. Quit kin FM, Stewart JW, McGrath PJ, Lie bow itz MR, Har ri son WM, Tri camo E,and oth ers. Phe nelz ine ver sus imi pramine in the treat ment of prob able atypi cal de -

55S The Ca na dian Jour nal of Psy chia try


pres sion: de fin ing syn drome bounda ries of se lec tive MAOI re spond ers. Am J Psy -chia try 1988;145:306–11.

27. Quit kin FM, McGrath PJ, Stewart JW, Har ri son W, Tri camo E, Wa ger SG, and oth -ers. Atypi cal de pres sion, panic at tacks, and re sponse to imi pramine and phe nelz -ine. A rep li ca tion. Arch Gen Psy chia try 1990;47:935–41.

28. McGrath PJ, Stewart JW, Janal MN, Petkova E, Quit kin FM, Klein DF. A placebo- controlled study of fluoxet ine ver sus imi pramine in the acute treat ment of atypi calde pres sion. Am J Psy chia try 2000;157:344–50.

29. So gaard J, Lane R, La ti mer P, Behnke K, Chris tian sen PE, Niel sen B, and oth ers. A12- week study com par ing mo clobe mide and ser tra line in the treat ment of out pa -tients with atypi cal de pres sion. J Psy cho phar ma col 1999;13:406–14.

30. Kel ler MB, McCullough JP, Klein DN, Ar now B, Dun ner DL, Ge len berg AJ, andoth ers. A com pari son of ne fa zo done, the cog ni tive behavioral- analysis sys tem ofpsy cho ther apy, and their com bi na tion for the treat ment of chronic de pres sion. NEngl J Med 2000;342:1462–70.

31. Ent suah AR, Ru dolph RL, Chi tra R. Ef fec tive ness of ven la fax ine treat ment in abroad spec trum of de pressed pa tients: a metaanalysis. Psy cho phar ma col Bull1995;31:759–66.

32. Men dels J, Kiev A, Fabre LF. Double- blind com pari son of cita lo pram and pla ceboin de pressed out pa tients with mel an cho lia. De press Anxi ety 1999;9:54–60.

33. Heili gen stein JH, Tol lef son GD, Far ies DE. Re sponse pat terns of de pressed out pa -tients with and with out mel an cho lia: a double- blind, placebo- controlled trial offluoxet ine ver sus pla cebo. J Af fect Dis ord 1994;30:163–73.

34. Clerc GE, Ruimy P, Verdeau- Palles J. A double- blind com pari son of ven la fax ineand fluoxet ine in pa tients hos pi tal ized for ma jor de pres sion and mel an cho lia. Theven la fax ine French in pa tient study group. Int Clin Psy cho phar ma col1994;9:139–43.

35. Tzanakaki M, Guazzelli M, Ni ma toudis I, Zis sis NP, Smer aldi E, Rizzo F. In -creased re mis sion rates with ven la fax ine com pared with fluoxet ine in hos pi tal izedpa tients with ma jor de pres sion and mel an cho lia. Int Clin Psy cho phar ma col2000;15:29–34.

36. Roose SP, Glass man AH, At tia E, Wood ring S. Com para tive ef fi cacy of se lec tivese ro tonin re up take in hibi tors and tri cyc lics in the treat ment of mel an cho lia. Am JPsy chia try 1994;151:1735–9.

37. Perry PJ. Phar ma co ther apy for ma jor de pres sion with mel an cholic fea tures: rela -tive ef fi cacy of tri cyc lic ver sus se lec tive se ro tonin re up take in hibi tor an ti de pres -sants. J Af fect Dis ord 1996;39:1–6.

38. Spi ker DG, Weiss JC, De aly RS, Grif fin SJ, Hanin I, Neil JF, and oth ers. The phar -ma col ogi cal treat ment of de lu sional de pres sion. Am J Psy chia try1985;142:430–6.

39. Parker G, Roy K, Hadzi- Pavlovic D, Pe dic F. Psy chotic (de lu sional) de pres sion: ameta- analysis of physi cal treat ments. J Af fect Dis ord 1992;24:17–24.

40. Zanardi R, Franchini L, Gas per ini M, Lucca A, Smer aldi E, Perez J. Faster on set ofac tion of fluvox am ine in com bi na tion with pin dolol in the treat ment of de lu sionalde pres sion: a con trolled study. J Clin Psy cho phar ma col 1998;18:441–6.

41. Zanardi R, Franchini L, Gas per ini M, Perez J, Smer aldi E. Double- blind con trolledtrial of ser tra line ver sus par oxet ine in the treat ment of de lu sional de pres sion. Am JPsy chia try 1996;153:1631–3.

42. Rothschild AJ, Bates KS, Boe hrin ger KL, Syed A. Ol an zap ine re sponse in psy -chotic de pres sion. J Clin Psy chia try 1999;60:116–8.

43. Lam RW, Le vitt AJ. Ca na dian con sen sus guide lines for the treat ment of sea sonalaf fec tive dis or der. Van cou ver (BC): Clini cal & Aca demic Pub lish ing; 1999.

44. Lam RW, Gor man CP, Mi cha lon M, Stei ner M, Le vitt AJ, Cor ral MR, and oth ers.Mul ti cen ter, placebo- controlled study of fluoxet ine in sea sonal af fec tive dis or der.Am J Psy chia try 1995;152:1765–70.

45. Par tonen T, Lonnqvist J. Mo clobe mide and fluoxet ine in treat ment of sea sonal af -fec tive dis or der. J Af fect Dis ord 1996;41:93–9.

46. Ruhr mann S, Kas per S, Hawel lek B, Mar ti nez B, Hoflich G, Nick el sen T, and oth -ers. Ef fects of fluoxet ine ver sus bright light in the treat ment of sea sonal af fec tivedis or der. Psy cho logi cal Medi cine 1998;28:923–33.

47. Dil saver SC, Qa mar AB, Del Med ico VJ. The ef fi cacy of bupropion in win ter de -pres sion: re sults of an open trial. J Clin Psy chia try 1992;53:252–5.

48. Thorell LH, Kjell man B, Arned M, Lindwall- Sundel K, Wal in der J, Wet ter berg L.Light treat ment of sea sonal af fec tive dis or der in com bi na tion with cita lo pram orpla cebo with 1- year follow- up. Int Clin Psy cho phar ma col 1999;14 (Suppl2):7S–11S.

49. Dil saver SC, Jaeckle RS. Win ter de pres sion re sponds to an open trial of tra nyl cy -promine. J Clin Psy chia try 1990;51:326–9.

50. Ameri can Psy chi at ric As so cia tion. Di ag nos tic and sta tis ti cal man ual of men tal dis -or ders. 4th ed. Wash ing ton (DC): Ameri can Psy chi at ric As so cia tion; 1994.

51. Rav in dran AV, Judge R, Hunter BN, Bray J, Mor ton NH. A double- blind, mul ti -cen ter study in pri mary care com par ing par oxet ine and clo mi pramine in pa tientswith de pres sion and as so ci ated anxi ety. Par oxet ine Study Group. J Clin Psy chia try1997;58:112–8.

52. Ru dolph RL, Ent suah R, Chi tra R. A metaanalysis of the ef fects of ven la fax ine onanxi ety as so ci ated with de pres sion. J Clin Psy cho phar ma col 1998;18:136–44.

53. Delini- Stula A, Mik kel sen H, Angst J. Thera peu tic ef fi cacy of an ti de pres sants inagi tated anx ious de pres sion: a metaanalysis of mo clobe mide stud ies. J Af fect Dis -ord 1995;35:21–30.

54. Faw cett J, Barkin RL. A metaanalysis of 8 ran dom ized, double- blind, con trolledclini cal tri als of mir tazap ine for the treat ment of pa tients with ma jor de pres sion andsymp toms of anxi ety. J Clin Psy chia try 1998;59:123–7.

55. Rav in dran AV, Judge R, Hunter BN, Bray J, Mor ton NH. A double- blind, mul ti -cen ter study in pri mary care com par ing par oxet ine and clo mi pramine in pa tientswith de pres sion and as so ci ated anxi ety. Par oxet ine study group. J Clin Psy chia try1997;58:112–8.

56. Moon CA, Jago W, Wood K, Doo gan DP. A double- blind com pari son of ser tra lineand clo mi pramine in the treat ment of ma jor de pres sive dis or der and as so ci atedanxi ety in gen eral prac tice. J Psy cho phar ma col 1994;8:171–6.

57. Laws D, Ash ford JJ, An stee JA. A mul ti cen tre double- blind com para tive trial offluvox am ine ver sus lo raze pam in mixed anxi ety and de pres sion treated in gen eralprac tice. Acta Psy chi atr Scand 1990;81:185–9.

58. Dun ner DL, Dun bar GC. Man ag ing the pa tient with de pres sion and anxi ety. EurPsy chia try 1993;8 (Suppl):9S–12S.

59. Dun bar GC, Cohn JB, Fabre LF, Feigh ner JP, Fieve RR, Men dels J, Shri vas tavaRK. A com pari son of par oxet ine, imi pramine and pla cebo in de pressed out- patients. Br J Psy chia try 1991;159:394–8.

60. Smith WT, Lond borg PD, Glaudin V, Painter JR. Short- term aug men ta tion offluoxet ine with clo naze pam in the treat ment of de pres sion: a double- blind study.Am J Psy chia try 1998;155:1339–45.

61. Kel ler MB, Hirsch feld RM, Hanks D. Dou ble de pres sion: a dis tinc tive sub type ofuni po lar de pres sion. J Af fect Dis ord 1997;45:65–73.

62. Rav in dran AV, Merali Z, An is man H. Dysthymia: A bio logi cal per spec tive. In:Licinio J, Bo lis CL, Gold P, edi tors. Dysthymia: from clini cal neu ro sci ence to treat -ment. Ge neva: World Health Or gani za tion; 1997. p 21– 4 4.

63. The WPA dysthymia work ing group. Dysthymia and clini cal prac tice. Br J Psy -chia try 1995;166:174–83.

64. de Lima MS, Ho toph M, Wes sely S. The ef fi cacy of drug treat ments for dysthymia:a sys tem atic re view and metaanalysis. Psy chol Med 1999;29:1273–89.

65. Rav in dran AV, Guelfi JD, Lane RM, Cas sano GB. Treat ment of dysthymia withser tra line: a double- blind, placebo- controlled trial in dysthymic pa tients with outma jor de pres sion. J Clin Psy chia try 2000;61:821–7.

66. Thase ME, Fava M, Hal breich U, Koc sis JH, Ko ran L, David son J, and oth ers. Aplacebo- controlled, ran dom ized clini cal trial com par ing ser tra line and imi praminefor the treat ment of dysthymia. Arch Gen Psy chia try 1996;53:777–84.

67. Teasdale JD, Se gal ZV, Wil liams JM, Ridge way VA, Soulsby JM, Lau MA. Pre -ven tion of re lapse/re cur rence in ma jor de pres sion by mindfulness- based cog ni tivether apy. J Con sult Clin Psy chol 2000;68:615–23.

68. Heller stein DJ, Batchelder ST, Lit tle SA, Fe dak MJ, Kre di tor D, Ro sen thal J. Ven -la fax ine in the treat ment of dysthymia: an open- label study. J Clin Psy chia try1999;60:845–9.

69. Dun ner DL, Hen drick son HE, Bea C, Budech CB, O’Con nor E. Dysthymic dis or -der: treat ment with mir tazap ine. De press Anxi ety 1999;10:68–72.

70. Kel ler MB, McCullough JP, Klein DN, Ar now B, Dun ner DL, Ge len berg AJ, andoth ers. A com pari son of ne fa zo done, the cog ni tive behavioral- analysis sys tem ofpsy cho ther apy, and their com bi na tion for the treat ment of chronic de pres sion. NEngl J Med 2000;342:1462–70.

71. Szegedi A, Wet zel H, Angers bach D, Philipp M, Benkert O. Re sponse to treat mentin mi nor and ma jor de pres sion: re sults of a double- blind com para tive study withpar oxet ine and ma pro ti line. J Af fect Dis ord 1997;45:167–78.

72. Ra pa port MH, Judd LL. Mi nor de pres sive dis or der and sub syn dro mal de pres sivesymp toms: func tional im pair ment and re sponse to treat ment. J Af fect Dis ord1998;48:227–32.

73. Mont gom ery DB, Rob erts A, Green M, Bul lock T, Bald win D, Mont gom ery SA.Lack of ef fi cacy of fluoxet ine in re cur rent brief de pres sion and sui ci dal at tempts.Eur Arch Psy chia try Clin Neu ro sci 1994;244:211–5.

74. Mont gom ery SA, Dun bar G. Par oxet ine is bet ter than pla cebo in re lapse pre ven tion and the pro phy laxis of re cur rent de pres sion. Int Clin Psy cho phar ma col1993;8:189–95.

75. Joffe RT. Tra nyl cy promine in re cur rent brief de pres sion: two case re ports. Int ClinPsy cho phar ma col 1996;11:287–8.

76. Sta menko vic M, Peza was L, de Zwaan M, As chauer HN, Kas per S. Mir tazap ine inre cur rent brief de pres sion. Int Clin Psy cho phar ma col 1998;13:39–40.

77. Bezchlibnyk- Butler KZ, Jef fries JJ. Clini cal Hand book of Psy cho tropic Drugs.10th ed. To ronto: Ho grefe & Hu ber Pub lish ers; 2000.

78. Montejo- Gonzalez AL, Llorca G, Izquierdo JA, Le desma A, Bou sono M, Cal cedoA, and oth ers. SSRI- induced sex ual dys func tion: fluoxet ine, par oxet ine, ser tra line,and fluvox am ine in a pro spec tive, mul ti cen ter, and de scrip tive clini cal study of 344 pa tients. J Sex Mari tal Ther 1997;23:176–94.

79. Ken nedy SH, Eis feld BS, Dick ens SE, Bac chio chi JR, Bagby RM. Anti-depressant- induced sex ual dys func tion dur ing treat ment with mo clobe mide, par -oxet ine, ser tra line, and ven la fax ine. J Clin Psy chia try 2000;61:276–81.

80. Karp JF, Frank E, Ritenour A, McEachran A, Kup fer DJ. Imi pramine and sex ualdys func tion dur ing the long- term treat ment of re cur rent de pres sion. Neu rop sy cho -phar ma col ogy 1994;11:21–7.

81. Har ri son WM, Rab kin JG, Ehr hardt AA, Stewart JW, McGrath PJ, Ross D, and oth -ers. Ef fects of an ti de pres sant medi ca tion on sex ual func tion: a con trolled study. JClin Psy cho phar ma col 1986;6:144–9.



82. Ek se lius L, von Knor ring L, Eber hard G. A double- blind mul ti cen ter trial com par -ing ser tra line and cita lo pram in pa tients with ma jor de pres sion treated in gen eralprac tice. Int Clin Psy cho phar ma col 1997;12:323–31.

83. Ro sen RC, Lane RM, Menza M. Ef fects of SSRIs on sex ual func tion: a criti cal re -view. J Clin Psy cho phar ma col 1999;19:67–85.

84. Cole man CC, Cun ning ham LA, Fos ter VJ, Batey SR, Do na hue RM, Houser TL,and oth ers. Sex ual dys func tion as so ci ated with the treat ment of de pres sion: aplacebo- controlled com pari son of bupropion sus tained re lease and ser tra line treat -ment. Ann Clin Psy chia try 1999;11:205–15.

85. Ka voussi RJ, Segraves RT, Hughes AR, Ascher JA, John ston JA. Double- blindcom pari son of bupropion sus tained re lease and ser tra line in de pressed out pa tients.J Clin Psy chia try 1997;58:532–7.

86. Croft H, Set tle E Jr, Houser T, Batey SR, Do na hue RM, Ascher JA. A placebo- controlled com pari son of the an ti de pres sant ef fi cacy and ef fects on sex ual func -tion ing of sustained- release bupropion and ser tra line. Clin Ther 1999;21:643–58.

87. Fei ger A, Kiev A, Shri vas tava RK, Wis se link PG, Wil cox CS. Ne fa zo done ver susser tra line in out pa tients with ma jor de pres sion: fo cus on ef fi cacy, tol er abil ity, andef fects on sex ual func tion and sat is fac tion. J Clin Psy chia try 1996;57 (Suppl2):53–62.

88. Philipp M, Koh nen R, Benkert O. A com pari son study of mo clobe mide anddoxepin in ma jor de pres sion with spe cial ref er ence to ef fects on sex ual dys func -tion. Int Clin Psy cho phar ma col 1993;7:149–53.

89. Mod ell JG, Katholi CR, Mod ell JD, De Palma RL. Com para tive sex ual side ef fectsof bupropion, fluoxet ine, par oxet ine, and ser tra line. Clin Phar ma col Ther1997;61:476–87.

90. Boyar sky BK, Haque W, Rouleau MR, Hirsch feld RM. Sex ual func tion ing in de -pressed out pa tients tak ing mir tazap ine. De press Anxi ety 1999;9:175–9.

91. Wald inger MD, Hen gev eld MW, Zwin der man AH, Olivier B. Ef fect of SSRI an ti -de pres sants on ejacu la tion: a double- blind, ran dom ized, placebo- controlled studywith fluoxet ine, fluvox am ine, par oxet ine, and ser tra line. J Clin Psy cho phar ma col1998;18:274–81.

92. Ans seau M, von Frenck ell R, Mer tens C, de Wilde J, Botte L, De voitille JM, andoth ers. Con trolled com pari son of two doses of mil na ci pran (F 2207) and amitrip -tyline in ma jor de pres sive in pa tients. Psy cho phar ma col ogy (Berl) 1989;98:163–8.

93. Fern strom MH, Krow in ski RL, Kup fer DJ. Chronic imi pramine treat ment andweight gain. Psy chia try Res 1986;17:269–73.

94. Har ris B, Young J, Hughes B. Com para tive ef fects of 7 an ti de pres sant re gimes onap pe tite, weight and car bo hy drate pref er ence. Br J Psy chia try 1986;148:590–2.

95. Mi chel son D, Am ster dam JD, Quit kin FM, Re im herr FW, Ro sen baum JF, Za jeckaJ, and oth ers. Changes in weight dur ing a 1- year trial of fluoxet ine. Am J Psy chia try 1999;156:1170–6.

96. Fava M, Judge R, Hoog SL, Nils son ME, Koke SC. Fluoxet ine ver sus ser tra line and par oxet ine in ma jor de pres sive dis or der: changes in weight with long- term treat -ment. J Clin Psy chia try 2000;61:863–7.

97. Suss man N, Gins berg DL. Weight ef fects of ne fa zo done, bupropion, mir tazap ine,and ven la fax ine: A re view of avail able evi dence. Pri mary Psy chia try2000;7:33–48.

98. Feigh ner JP, Gard ner EA, John ston JA, Batey SR, Khay ral lah MA, Ascher JA, andoth ers. Double- blind com pari son of bupropion and fluoxet ine in de pressed out pa -tients. J Clin Psy chia try 1991;52:329–35.

99. Bak ish D, Wiens A, El lis J, Alda M, Lapi erre Y. A double- blind placebo- controlled com pari son of mo clobe mide and amitrip tyline in the treat ment of de pres sion. Can J Psy chia try 1992;37 (Suppl 1):12S–17S.

100. Moll E, Neu mann N, Schmid- Burgk W, Stabl M, Am rein R. Safety and ef fi cacydur ing long- term treat ment with mo clobe mide. Clin Neu ro phar ma col 1994;17(Suppl 1):74S–87S.

101. Vaz- Serra A, Figueira ML, Firmino H, Al bu quer que AJ, Jara JM, Pes tana LC.Mul ti cen ter double- blind study of mo clobe mide and ma pro ti line. Clin Neu ro phar -ma col 1994;17 (Suppl 1):38S–49S.

102. Good nick PJ, Kre mer C, Win gard P. Weight change dur ing mir tazap ine ther apy.Pri mary Psy chia try 1999;6:103–8.

103. An der son IM, To men son BM. Treat ment dis con tinua tion with se lec tive se ro toninre up take in hibi tors com pared with tri cyc lic an ti de pres sants: a metaanalysis. BMJ1995;310:1433–8.

104. Ho topf M, Hardy R, Lewis G. Dis con tinua tion rates of SSRIs and tri cyc lic an ti de -pres sants: a metaanalysis and in ves ti ga tion of het ero ge ne ity. Br J Psy chia try1997;170:120–7.

105. Mont gom ery SA, Henry J, McDon ald G, Di nan T, Lader M, Hind march I, and oth -ers. Se lec tive se ro tonin re up take in hibi tors: metaanalysis of dis con tinua tion rates.Int Clin Psy cho phar ma col 1994;9:47–53.

106. Mendle wicz J. Phar ma cologic pro file and ef fi cacy of ven la fax ine. Int Clin Psy cho -phar ma col 1995;10 (Suppl 2):5–13.

107. Nier en berg AA, McLean NE, Alp ert JE, Wor thing ton JJ, Ro sen baum JF, Fava M.Early non re sponse to fluoxet ine as a pre dic tor of poor 8- week out come. Am J Psy -chia try 1995;152:1500–3.

108. Quit kin FM, McGrath PJ, Stewart JW, Tay lor BP, Klein DF. Can the ef fects of an ti -de pres sants be ob served in the first two weeks of treat ment? Neu rop sy cho phar ma -col ogy 1996;15:390–4.

109. Nier en berg AA, Fara baugh AH, Alp ert JE, Gor don J, Wor thing ton JJ, Ro sen baumJF, and oth ers. Tim ing of on set of an ti de pres sant re sponse with fluoxet ine treat -ment. Am J Psy chia try 2000;157:1423–8.

110. Stas sen HH, Angst J, Delini- Stula A. Fluoxet ine ver sus mo clobe mide: cross- comparison be tween the time courses of im prove ment. Phar ma cop sy chia try 1999;32:56–60.

111. Thase ME, Rush AJ. When at first you don’t suc ceed: se quen tial strate gies for an ti -de pres sant non re spond ers. J Clin Psy chia try 1997;58 (Suppl 13):23–9.

112. Aus tin MP, Souza FG, Good win GM. Lith ium aug men ta tion in antidepressant- resistant pa tients. A quan ti ta tive analy sis. Br J Psy chia try 1991;159:510–4.

113. Bauer M, Dopf mer S. Lith ium aug men ta tion in treatment- resistant de pres sion:metaanalysis of placebo- controlled stud ies. J Clin Psy cho phar ma col 1999;19:427–34.

114. Joffe RT, Singer W, Le vitt AJ, Mac Don ald C. A placebo- controlled com pari son oflith ium and triio do thy ronine aug men ta tion of tri cyc lic an ti de pres sants in uni po larre frac tory de pres sion. Arch Gen Psy chia try 1993;50:387–93.

115. Joffe RT, Singer W. A com pari son of triio do thy ronine and thy rox ine in the po ten -tia tion of tri cyc lic an ti de pres sants. Psy chia try Res 1990;32:241–51.

116. Ar on son R, Off man HJ, Joffe RT, Nay lor CD. Triio do thy ronine aug men ta tion inthe treat ment of re frac tory de pres sion. A metaanalysis. Arch Gen Psy chia try1996;53:842–8.

117. Bau mann P, Nil R, Souche A, Mon taldi S, Baet tig D, Lam bert S, and oth ers. Adouble- blind, placebo- controlled study of cita lo pram with and with out lith ium inthe treat ment of therapy- resistant de pres sive pa tients: a clini cal, phar ma coki netic,and phar ma co ge netic in ves ti ga tion. J Clin Psy cho phar ma col 1996;16:307–14.

118. Ka tona CL, Rob ert son MM, Abou- Saleh MT, Nairac BL, Ed wards DR, Lock T,and oth ers. Placebo- controlled trial of lith ium aug men ta tion of fluoxet ine and lo fe -pramine. Int Clin Psy cho phar ma col 1993;8:323.

119. Lan den M, Bjor ling G, Agren H, Fah len T. A ran dom ized, double- blind, placebo- controlled trial of bus pi rone in com bi na tion with an SSRI in pa tients withtreatment- refractory de pres sion. J Clin Psy chia try 1998;59:664–8.

120. Perez V, Soler J, Puig de mont D, Al varez E, Ar ti gas F. A double- blind, ran dom ized, placebo- controlled trial of pin dolol aug men ta tion in de pres sive pa tients re sis tant to se ro tonin re up take in hibi tors. Grup de re cerca en tras torns afec tius. Arch Gen Psy -chia try 1999;56:375–9.

121. Shel ton RC, Tol lef son GD, To hen M, Stahl S, Gan non KS, Ja cobs TG, and oth ers.A novel aug men ta tion strat egy for treat ing re sis tant ma jor de pres sion. Am J Psy -chia try 2001;158:131–4.

122. Frye MA, Ket ter TA, Leverich GS, Hug gins T, Lantz C, Deni coff KD, and oth ers.The in creas ing use of po lyphar ma co ther apy for re frac tory mood dis or ders: 22years of study. J Clin Psy chia try 2000;61:9–15.

123. White K, Simp son G. The com bined use of MAOIs and tri cyc lics. J Clin Psy chia try 1984;45:67–9.

124. Fava M, Ro sen baum JF, McGrath PJ, Stewart JW, Am ster dam JD, Quit kin FM.Lith ium and tri cyc lic aug men ta tion of fluoxet ine treat ment for re sis tant ma jor de -pres sion: a double- blind, con trolled study. Am J Psy chia try 1994;151:1372–4.

125. Nel son JC, Mazure CM, Bow ers MB, Jr., Jat low PI. A pre limi nary, open study ofthe com bi na tion of fluoxet ine and de si pramine for rapid treat ment of ma jor de pres -sion. Arch Gen Psy chia try 1991;48:303–7.

126. Frank E, Prien RF, Jar rett RB, Kel ler MB, Kup fer DJ, La vori PW, and oth ers. Con -cep tu ali za tion and ra tion ale for con sen sus defi ni tions of terms in ma jor de pres sivedis or der. Re mis sion, re cov ery, re lapse, and re cur rence. Arch Gen Psy chia try1991;48:851–5.

127. Viguera AC, Bald es sarini RJ, Fried berg J. Dis con tinu ing an ti de pres sant treat mentin ma jor de pres sion. Harv Rev Psy chia try 1998;5:293–306.

128. Re im herr FW, Am ster dam JD, Quit kin FM, Ro sen baum JF, Fava M, Za jecka J, and oth ers. Op ti mal length of con tinua tion ther apy in de pres sion: a pro spec tive as sess -ment dur ing long- term fluoxet ine treat ment. Am J Psy chia try 1998;155:1247–53.

129. Bauer M, Bschor T, Kunz D, Berg ho fer A, Strohle A, Muller- Oerlinghausen B.Double- blind, placebo- controlled trial of the use of lith ium to aug ment an ti de pres -sant medi ca tion in con tinua tion treat ment of uni po lar ma jor de pres sion. Am J Psy -chia try 2000;157:1429–35.

130. Za jecka J, Tracy KA, Mitchell S. Dis con tinua tion symp toms af ter treat ment withse ro tonin re up take in hibi tors: a lit era ture re view. J Clin Psy chia try 1997;58:291–7.

131. Had dad P. The SSRI dis con tinua tion syn drome. J Psy cho phar ma col1998;12:305–13.

132. Wijer atne C, Hal li day GS, Lyn don RW. The pres ent status of elec tro con vul sivether apy: a sys tem atic re view. Med J Aust 1999;171:250–4.

133. Pru dic J, Sackeim HA, De vanand DP. Medi ca tion re sis tance and clini cal re sponseto elec tro con vul sive ther apy. Psy chia try Res 1990;31:287–96.

134. Pru dic J, Haskett RF, Mul sant B, Ma lone KM, Petti nati HM, Ste phens S, and oth -ers. Re sis tance to an ti de pres sant medi ca tions and short- term clini cal re sponse toECT. Am J Psy chia try 1996;153:985–92.

135. Bour gon LN, Kell ner CH. Re lapse of de pres sion af ter ECT: a re view. J ECT2000;16:19–31.

136. Sackeim HA, Pru dic J, De vanand DP, Decina P, Kerr B, Mal itz S. The im pact ofmedi ca tion re sis tance and con tinua tion phar ma co ther apy on re lapse fol low ing re -sponse to elec tro con vul sive ther apy in ma jor de pres sion. J Clin Psy cho phar ma col1990;10:96–104.



137. Lam RW, Bartley S, Yatham LN, Tam EM, Zis AP. Clini cal pre dic tors of short- term out come in elec tro con vul sive ther apy. Can J Psy chia try 1999;44:158–63.

138. Shapira B, Tubi N, Drex ler H, Lid sky D, Calev A, Lerer B. Cost and bene fit in thechoice of ECT sched ule. Twice ver sus 3 times weekly ECT. Br J Psy chia try1998;172:44–48.

139. Sackeim H, Decina P, Pro hov nik I, Mal itz S. Sei zure thresh old in elec tro con vul -sive ther apy. Ef fects of sex, age, elec trode place ment, and number of treat ments.Arch Gen Psy chia try 1987;44:355–60.

140. Sackeim HA. Re peti tive tran scra nial mag netic stimu la tion: what are the nextsteps? Biol Psy chia try 2000;48:959–61.

141. Abrams R. Elec tro con vul sive ther apy re quires higher dos age lev els: Food andDrug Ad min istra tion ac tion is re quired. Arch Gen Psy chia try 2000;57:445–6.

142. Krys tal AD, Wei ner RD, Cof fey CE. The ic tal EEG as a marker of ade quate stimu -lus in ten sity with uni lat eral ECT. J Neu rop sy chia try Clin Neu ro sci1995;7:295–303.

143. Fink M, Abrams R, Bail ine S, Jaffe R. Am bu la tory elec tro con vul sive ther apy: re -port of a task force of the as so cia tion for con vul sive ther apy. As so cia tion for con -vul sive ther apy. Con vuls Ther 1996;12:42–55.

144. Kra mer BA. Use of ECT in Cali for nia, re vis ited: 1984–1994. J ECT1999;15:245–51.

145. Li sanby SH, Mad dox JH, Pru dic J, De vanand DP, Sackeim HA. The ef fects of elec -tro con vul sive ther apy on mem ory of auto bio graphi cal and pub lic events. Arch GenPsy chia try 2000;57:581–90.

146. Rab heru K, Per sad E. A re view of con tinua tion and main te nance elec tro con vul sivether apy. Can J Psy chia try 1997;42:476–84.

147. De vanand DP, Dwork AJ, Hutchin son ER, Bol wig TG, Sackeim HA. Does ECT al -ter brain struc ture? Am J Psy chia try 1994;151:957–70.

148. Ma yur PM, Gan gad har BN, Sub bakrishna DK, Janaki ra maiah N. Dis con tinua tionof an ti de pres sant drugs dur ing elec tro con vul sive ther apy: a con trolled study. J Af -fect Dis ord 2000;58:37–41.

149. Lau rit zen L, Od gaard K, Clemmesen L, Lunde M, Ohr strom J, Black C, and oth ers.Re lapse pre ven tion by means of par oxet ine in ECT- treated pa tients with ma jor de -pres sion: a com pari son with imi pramine and pla cebo in medium- term con tinua tionther apy. Acta Psy chi atr Scand 1996;94:241–51.

150. Ter man M, Ter man JS, Ross DC. A con trolled trial of timed bright light and nega -tive air ioni za tion for treat ment of win ter de pres sion. Arch Gen Psy chia try1998;55:875–82.

151. East man CI, Young MA, Fogg LF, Liu L, Me aden PM. Bright light treat ment ofwin ter de pres sion: a placebo- controlled trial. Arch Gen Psy chia try 1998;55:883–9.

152. Lee TM, Chan CC. Dose- response re la tion ship of pho to ther apy for sea sonal af fec -tive dis or der: a metaanalysis. Acta Psy chi atr Scand 1999;99:315–23.

153. Kripke DF. Light treat ment for non sea sonal de pres sion: speed, ef fi cacy, and com -bined treat ment. J Af fect Dis ord 1998;49:109–117.

154. Linde K, Mul row CD. St John’s wort for de pres sion. Co chrane Da ta base Syst Rev2000;CD000448.

155. Gas ter B, Hol royd J. St John’s wort for de pres sion: a sys tem atic re view. Arch In -tern Med 2000;160:152–6.

156. Lei ben luft E, Wehr TA. Is sleep dep ri va tion use ful in the treat ment of de pres sion?Am J Psy chia try 1992;149:159–68.

157. Kuhs H, Tolle R. Sleep dep ri va tion ther apy. Biol Psy chia try 1991;29:1129–48.158. Wirz- Justice A, van den Hoof dak ker RH. Sleep dep ri va tion in de pres sion: what do

we know, where do we go? Biol Psy chia try 1999;46:445–53.159. Smer aldi E, Benedetti F, Barb ini B, Cam pori E, Co lombo C. Sus tained an ti de pres -

sant ef fect of sleep dep ri va tion com bined with pin dolol in bi po lar de pres sion. Aplacebo- controlled trial. Neu rop sy cho phar ma col ogy 1999;20:380–5.

160. Law lor DA, Hop ker SW. The ef fec tive ness of ex er cise as an in ter ven tion in theman age ment of de pres sion: sys tem atic re view and meta- regression analy sis of ran -dom ised con trolled tri als. BMJ 2001;322:763.

161. Blu men thal JA, Ba byak MA, Moore KA, Craig head WE, Her man S, Kha tri P, andoth ers. Ef fects of ex er cise train ing on older pa tients with ma jor de pres sion. ArchIn tern Med 1999;159:2349–56.

162. Ba byak M, Blu men thal JA, Her man S, Kha tri P, Do rais wamy M, Moore K, andoth ers. Ex er cise treat ment for ma jor de pres sion: main te nance of thera peu tic bene fit at 10 months. Psy cho som Med 2000;62:633–8.

163. Cos grove GR, Rauch SL. Psy cho sur gery. Neu ro surg Clin N Am 1995;6:167–76.164. Bridges PK, Bartlett JR, Hale AS, Poyn ton AM, Mal izia AL, Hodgkiss AD. Psy -

cho sur gery: stereo tac tic sub cau date trac tomy. An in dis pen sa ble treat ment. Br JPsy chia try 1994;165:599–611.

165. George MS, Sackeim HA, Rush AJ, Ma rangell LB, Na has Z, Hu sain MM, and oth -ers. Vagus nerve stimu la tion: a new tool for brain re search and ther apy. Biol Psy -chia try 2000;47:287–95.

166. Rush AJ, George MS, Sackeim HA, Ma rangell LB, Hu sain MM, Gil ler C, and oth -ers. Vagus nerve stimu la tion (VNS) for treatment- resistant de pres sions: a mul ti -cen ter study. Biol Psy chia try 2000;47:276–86.

167. George MS, Li sanby SH, Sackeim HA. Tran scra nial mag netic stimu la tion: ap pli -ca tions in neu rop sy chia try. Arch Gen Psy chia try 1999;56:300–11.



iv. medications and other biological treatments 2001.pdf · iv. medications and other biological...

Documents