Use of antiviral prophylaxis in influenza outbreaks in long term care facilities

De part ment of Mi cro bi ol ogy, Mount Si nai Hos pi tal, Uni ver sity of To ronto, To ronto, On tario; De part ments of Phar ma col ogy and Thera peu tics, Medi cine and Medi cal Mi cro bi ol ogy, Uni ver sity of Mani toba, Win ni peg, Mani toba; Perth Dis trict Health Unit, Stratford, On tario; York Re gion Health Serv ices De part ment, Newmarket, On tario Cor re spon dence and re prints: Dr Al li son McGeer, Room 1460, De part ment of Mi cro bi ol ogy, Mount Si nai Hos pi tal, 600 Uni ver sity Ave nue, To ronto, On tario M5G 1X5. Tele phone 4165863118, fax 416-586-3140, email amcgeer@mtsi nai.on.ca A McGeer, DS Sitar, SE Tam blyn, F Kolbe, P Orr, FY Aoki. Use of antiviral pro phy laxis in influenza outbreaks in long term care fa cili ties. Can J In fect Dis 2000;11(4):187192.

D e spite the fact that more than 90% of resi dents of long term care fa cili ties in Can ada are vac ci nated against influ enza an nu ally, al most half of such fa cili ties re port de tecting at least one in flu enza out break each year (1,2).Al though there are no ran dom ized, con trolled tri als as sess ing the ef fective ness of an tivi ral pro phy laxis in the con trol of out breaks, aman tadine has been shown to be ef fec tive in pre vent ing influ enza A in ex posed per sons (3,4), and nu mer ous re ports docu ment its suc cess in ter mi nat ing in flu enza A spread in the long term care set ting (1,2,(5)(6)(7)(8)(9). Thus, both Ameri can and Cana dian ex pert ad vi sory com mit tees rec om mend an tivi ral prophy laxis for resi dents for the con trol of in flu enza A out breaks (10,11), and such pro phy laxis has be come a stan dard part of out break man age ment in Ca na dian long term care fa cili ties (1,2).There are, how ever, nu mer ous ar eas of dis agree ment about how best to man age mass pro phy laxis, and the ad vent of neu ramini dase in hibi tors of fers new chal lenges in se lect ing the best op tions for pre ven tion of in flu enza in this set ting.In the pres ent pa per, we dis cuss is sues sur round ing the ini tiation and dis con tinua tion of pro phy laxis, the use of amantadine and the po ten tial place of neu ramini dase in hibi tors in out break con trol.

DE CI SIONS ABOUT THE INI TIA TION OF PRO PHY LAXIS
Aman tadine for an tivi ral pro phy laxis of resi dents is use ful to pre vent mor bid ity and mor tal ity when in flu enza A is be ing trans mit ted in a long term care fa cil ity.For op ti mal use of prophy laxis, it is im por tant that clus ters of acute res pi ra tory infec tion are de tected early, that fa cili ties have the abil ity to ob tain na so pha ryn geal swabs and have rapid an ti gen test ing for in flu enza per formed seven days/week, and that con sent for the use of pro phy laxis be ob tained ei ther on ad mis sion to the long term care fa cil ity or an nu ally be fore in flu enza sea son.
Di ag nos ing in flu enza us ing clini cal in quiry and ex ami nation is dif fi cult.Over all, the symp tom com plex with the best pre dic tive value (ill ness as so ci ated with the abrupt on set of fever higher than 38.5°C and dry cough) has only a 35% posi tive pre dic tive value for the di ag no sis of in flu enza among un vac cinated, eld erly adults liv ing in de pend ently (12).When in fluenza is pres ent in the com mu nity, a simi lar con stel la tion of signs and symp toms in healthy younger adults is about 60% pre dic tive of in flu enza (13).How ever, in vac ci nated resi dents of long term care fa cili ties, whose ill ness may be modi fied by vac cine, who may not mount a feb rile re sponse and who are of ten un able to de scribe symp toms clearly, the pre dic tive value of such signs and symp toms is much poorer (14).Therefore, in flu enza out breaks in nurs ing homes can only be re liably di ag nosed by labo ra tory test ing in the set ting of clus ters of acute res pi ra tory ill ness.
Go mo lin et al (15) have sug gested that a clus ter of in fection should be con sid ered to be three resi dents on one unit who de velop acute res pi ra tory ill ness within 72 h of each other.In this cir cum stance, case find ing should be en hanced, and na so pha ryn geal swabs should be ob tained from the ini tial cases as well as from the next three to five new cases.The iden ti fi ca tion of two resi dents with laboratory-confirmed in -flu enza con firms that in flu enza is be ing trans mit ted, and prophy laxis should be started for all as ymp to matic resi dents.With a sin gle labo ra tory con fir ma tion from the clus ter, judg ment should be used, and a de ci sion about whether to start pro phylaxis should be made jointly by the fa cil ity and pub lic health.
Pro phy laxis should be of fered to resi dents who are as ympto matic at the time of the de ci sion to ini ti ate mass pro phylaxis.Treat ment may be con sid ered for those who have had symp toms for less than 48 h (16).It is im por tant to re mem ber that in flu enza is a self-limited dis ease even in eld erly nurs ing home resi dents, and hos pi tali za tion and death are most of ten due to com pli ca tions rather than to the in flu enza it self.Residents who have had symp toms for more than 48 h will not bene fit from an tivi ral treat ment (17).Mini miz ing the use of an tivi rals for treat ment is par ticu larly im por tant for amantadine, be cause in flu enza A strains de velop re sis tance to aman tadine very eas ily when ex posed to it (3).Amantadineresistant vi ruses are as viru lent and trans mis si ble as sus cep tible vi ruses, and fail ure of aman tadine to con trol out breaks due to the emer gence of re sis tance has been iden ti fied (17)(18)(19)(20)(21). Lim it ing aman tadine treat ment to three to five days and discon tinu ing pro phy laxis in resi dents who de velop symp toms may help to ob vi ate the emer gence of re sis tance (22); the use of a neu ramini dase in hibi tor in stead of aman tadine for treatment may also be help ful (see be low).
In smaller fa cili ties, there is al most al ways sub stan tial mix ing of both resi dents and staff on dif fer ent units, so that it is gen er ally es sen tial to of fer pro phy laxis to all as ymp to matic resi dents in the fa cil ity.In larger fa cili ties, it may be pos si ble to limit pro phy laxis to one or more geo graphi cally sepa rate units.It is im por tant to re al ize that, when an out break is recog nized, a sub stan tial number of ex posed resi dents and staff may be in the in cu ba tion phase but not yet be ill.The abil ity to limit pro phy laxis to one unit suc cess fully de pends on both the de gree of con tact be tween staff and resi dents on the af fected unit and other units in the three days be fore the de tec tion of the out break (ie, are resi dents and staff on other units in cubat ing in flu enza?), and the ex tent to which such con tact can be pre vented over the next few days.Clearly, the vac ci na tion rate among staff is im por tant, be cause ex posed vac ci nated staff are less likely to be come ill.
Vac cine ef fi cacy in healthy adults younger than 65 years of age is 80% or greater (23,24), so that cur rent rec om men dations spec ify that only un vac ci nated staff re quire che mo prophy laxis (10,11).The ma jor ity (65%) of fa cili ties across Canada now of fer pro phy laxis rou tinely to un vac ci nated staff (1).This pro tects staff and their fami lies from ill ness, and re duces ab sen tee ism dur ing the out break, a time when staff ing may be dif fi cult.In ad di tion, be cause staff are in fec tious at or be fore the on set of symp toms (25) and on set may oc cur in the mid dle of a shift, pro phy laxis likely adds a de gree of pro tec tion for resi dents and re duces the risk of propa ga tion of the out break.This ar gu ment has led a number of long term care fa cili ties and pub lic health units across Can ada to re quire un vac ci nated staff to take pro phy laxis if they wish to con tinue to work during out breaks.The On tario La bour Re la tions Board and the On tario Nurses As so cia tion have sup ported such poli cies.

DE CI SIONS ABOUT DIS CON TINU ING PRO PHY LAXIS
In more than 75% of out breaks, the ini tia tion of mass an tivi ral pro phy laxis is as so ci ated with ter mi na tion of the outbreak (1,2,(4)(5)(6)(7)(8)(9)21).Be cause the ef fi cacy of an tivi rals in prevent ing in fec tion is not ab so lute, par ticu larly in those resi dents and staff who are in cu bat ing the in fec tion when pro phy laxis is ini ti ated, a few cases may oc cur in the first 72 h af ter the initia tion of pro phy laxis.Pro phy laxis should be con tin ued un til the out break is over: that is, un til one com plete in cu ba tion period passes fol low ing the in fec tious pe riod (or pe riod of commu ni ca bil ity) in the last case in the fa cil ity.In gen eral, the last in fec tious case oc curs in a resi dent, and sig nifi cant vi ral shedding oc curs for three to five days af ter the on set of symp toms (26,27).The in cu ba tion pe riod of in flu enza is one to three days.Thus, pro phy laxis should be con tin ued un til eight days af ter the on set of symp toms in the last case.
In about 20% of out breaks, new cases may con tinue to occur more than 72 h af ter pro phy laxis is started (1,2,9,21), and fur ther in ves ti ga tion is nec es sary.There are a number of reasons why an tivi ral pro phy laxis may fail to stop the out break.First, if aman tadine is be ing used, the vi rus may be re sis tant (ei ther at the start of the out break or be cause resistance has de vel oped dur ing the out break).Sec ond, an other res pi ra tory vi rus may be co-circulating and caus ing ill ness clini cally that is in dis tin guish able from in flu enza.In one study, at least one case of ill ness due to an other vi rus was iden ti fied in five of six long term care fa cil ity out breaks of in flu enza (28).Third, new cases may be oc cur ring be cause non im mune, un pro tected resi dents or staff con tinue to propa gate the out break.Which of these pos si bili ties is oc cur ring can only be de ter mined by di ag nos tic test ing of na so pha ryn geal swabs from new cases.Be cause rapid an ti gen test ing by ELISA is cur rently avail able only for in flu enza and res pi ra tory syn cytial vi rus, test ing by di rect fluo res cent an ti body, which can de tect in flu enza, res pira tory syn cytial vi rus, parain flu enza and ad eno vi rus, of fers ad van tages in this situa tion.
Aman tadine-re sis tance test ing is not yet avail able in a suffi ciently timely man ner for use in out break man age ment.Aman tadine re sis tance should be sus pected when laboratory-confirmed cases of in flu enza A con tinue to oc cur in resi dents or staff re ceiv ing ade quate pro phy laxis, par ticu larly if the number of new cases starts to in crease again.If re sis tance is suspected, aman tadine should be dis con tin ued and pro phy laxis with a neu ramini dase in hibi tor sub sti tuted (see be low).If ill ness is due to a dif fer ent vi rus, a clini cal de ci sion must be made as to when the last case of in flu enza oc curred.Aman tadine may be dis con tin ued eight days af ter the on set of this case.

DOS ING REGI MENS FOR AMAN TADINE
Se rum lev els of aman tadine are af fected by varia tion in both its ap par ent vol ume of dis tri bu tion and the rate of its renal elimi na tion.The ap par ent vol ume of dis tri bu tion of amantadine is most di rectly re lated to body weight but is in versely re lated to dose.Re nal elimi na tion is di rectly re lated to cre atinine clear ance.In ad di tion, aman tadine re nal clear ance is one-third less in fe males than males of the same weight, proba bly due to a sex dif fer ence in re nal tu bu lar se cre tion rate (29,30).The net ef fect of these in ter de pend ent fac tors in a given pa tient con trib utes to the dif fi culty of de sign ing ef fective and well tol er ated aman tadine dos ing sched ules for frail eld erly resi dents of in sti tu tions.
When doses rec om mended for pro phy laxis in younger adults are used in resi dents of nurs ing homes, a sig nifi cant increase in the rate of dose-related side ef fects of aman tadine, in clud ing diz zi ness, ir ri ta bil ity, con fu sion, and the po ten tiation of ad verse events due to drugs with an ti cho liner gic side ef fects, have been re ported (3,31).These side ef fects may result in falls, frac tured hips and deaths in this popu la tion (18).
The Ca na dian Na tional Ad vi sory Com mit tee on Im mu ni zation (NACI) (10) has pub lished rec om men da tions on in di vidual ized aman tadine dos ing, tak ing into ac count age and es timated cre ati nine clear ance (Ta ble 1).The ma jor ity (79%) of long term care fa cili ties in Can ada use these rec om men da tions and re port that, if this dos ing regi men is used, fewer than 2% of resi dents started on aman tadine need to have their medi cation dis con tin ued due to side ef fects (1,9).Ob vi ously, cal culated cre ati nine clear ances are only es ti mates of true cre atinine clear ances, and a re cent study of aman tadine lev els in resi dents (32) found that se rum lev els may be be low those predicted to be ef fec tive when this dos ing regi men is used.Nonethe less, cu mu la tive ex pe ri ence in Ca na dian nurs ing homes sug gests that this regi men is safe and ef fec tive in con trol ling in flu enza A out breaks (1,2,9,21).
The NACI in di vidu al ized dos ing rec om men da tions have three draw backs.First, the ini tial dose of aman tadine is 100 mg for most resi dents, but 50 mg for one group.Be cause the vol ume of dis tri bu tion is in de pend ent of cre ati nine clearance, the load ing dose should be based only on weight, not cre ati nine clear ance.Given that the ini tial dose se lected by the NACI guide lines has been found to be as so ci ated with a low risk of side ef fects, it is safe and rea son able to give each resident an ini tial dose of 100 mg.If an in flu enza out break oc curs and in di vidu al ized doses have not been cal cu lated in ad vance for each resi dent (as is de sir able), this means that an ini tial dose may be given to each resi dent be fore in di vidu al ized dosing regi mens are cal cu lated.Sec ond, the in ter mit tent dos ing sched ule, with in ter vals of two to seven days be tween doses, re sults in sub stan tial peaks in drug con cen tra tions af ter subse quent 100 mg doses, which may put resi dents at in creased risk of side ef fects (33).Fi nally, some fa cili ties have felt that it is con fus ing to have dos ing sched ules for which dif fer ent residents re ceive medi ca tion in dif fer ent amounts and on dif fer ent days.
For these rea sons, the authors have de vel oped a sec ond dos ing sched ule in which all resi dents re ceive an ini tial dose of 100 mg of aman tadine, fol lowed by a daily dose of amantadine so lu tion, ad justed for es ti mated cre ati nine clear ance (Ta ble 2).As with the dos ing regi men rec om mended by NACI, this regi men takes cre ati nine clear ance into ac count, but does not ad just for other phar ma coki netic ef fects of the resi dent's weight and sex (eg, on vol ume of dis tri bu tion).It is also simpli fied to ac count for the fact that aman tadine so lu tion is likely to be dis pensed in medi ca tion cups marked in 2.5 mL incre ments.Phar ma coki netic cal cu la tions sug gest that this dosing regi men should be as ef fec tive as the stan dard NACI guide lines, with out an in crease in side ef fects.It has the disad van tage that, over all, more doses of medi ca tion need to be ad min is tered.Fa cili ties should take the ad van tages and disad van tages of the two dif fer ent sched ules into con sid era tion when se lect ing a regi men for their resi dents.
For pro phy laxis, ini tial stud ies in healthy adults un der the age of 65 years used the cur rently rec om mended dose of 100 mg twice daily.At this dose, an noy ing neu ro logi cal side ef fects (eg, in som nia, dry mouth) are re ported by as many as 30% of sub jects.In the larg est ran dom ized, con trolled trial of aman tadine pro phy laxis, 20% of sub jects dis con tin ued the drug be cause of side ef fects (34).A dose of 100 mg daily has been shown to be ef fec tive in pro phy laxis in one trial (35).Because this dose is as so ci ated with a sig nifi cantly re duced rate of side ef fects, it may be pref er able for staff pro phy laxis.

ROLE OF NEU RAMINI DASE IN HIBI TORS
By late 1999, two neu ramini dase in hibi tors with ac tiv ity against in flu enza, zanami var and oselta mivir had been licensed in Can ada for the treat ment of in flu enza in adults.There is good evi dence from ran dom ized, con trolled tri als that these medi ca tions are also ef fec tive in pro phy laxis (36,37).
Oselta mivir has been shown to be 80% ef fec tive in pre vent ing in flu enza in nurs ing home resi dents ex posed to in flu enza (38), and zanamivir has been shown to be ef fec tive in con trol of in flu enza in at least two out breaks (39,40).Through the influ enza sea son of 1999/2000, nu mer ous pub lic health units and long term care fa cili ties used these medi ca tions off la bel in the man age ment of in flu enza out breaks in in sti tu tions.
Zanamivir is a pow der that is taken via an in haler.The recom mended treat ment dose is 10 mg (two puffs) bid; the prophylactic dose is 10 mg (two puffs) daily.About 20% of long term care fa cil ity resi dents have some dif fi culty co or di nat ing the in ha la tions (20).Only about 3% of a dose is ab sorbed.In ran dom ized, con trolled tri als to date, no side ef fects have been iden ti fied, but there con tin ues to be con cern about the risk of bron cho spasm in pa tients with asthma.In clini cal tri als, zanamivir ap pears well tol er ated in pa tients with mild to moder ate asthma (13).How ever, one pa tient with se vere chronic ob struc tive lung dis ease who took re peated doses of zanamivir noted wheez ing af ter each dose and re quired hos pi taliza tion for res pi ra tory dis tress on the third day of ther apy (41).The United States Food and Drug Ad min istra tion has re ported that other pa tients with asthma or un der ly ing chronic obstruc tive lung dis ease have also ex pe ri enced de te rio ra tion after zanamivir in ha la tion.
Oselta mivir is sup plied as a 75 mg cap sule, with the adult treat ment dose be ing 75 mg bid and the pro phylatic dose being 75 mg daily.A sus pen sion form of this medi ca tion is expected to be come avail able within the next two years.Oseltamivir is ex creted re nally.It is rec om mended that the treat ment dose be halved in per sons with cre ati nine clear ances less than 30 mL/min; no ad just ment is re quired for the pro phy lac tic dose for those with a lesser de gree of re nal dys func tion.No inter ac tions be tween oselta mivir and other drugs have been iden ti fied.The most com mon side ef fects are nau sea and vomit ing (13,37,38).They are re ported to oc cur most promi nently after the first dose, and can be re duced by tak ing the first dose with food.These symp toms are also more com mon in fe males and younger adults (ex cess rate over pla cebo 5% to 9%) than in nurs ing home resi dents (ex cess rate over pla cebo 2.5%).
An tivi ral re sis tance can be in duced in the labo ra tory to both neu ramini dase in hibi tors; how ever, this resistance is much more dif fi cult to in duce than re sis tance to aman tadine (13).In ad di tion, the re sis tant vi ruses iden ti fied to date have been less in fec tious than their sus cep ti ble coun ter parts.Resis tance to zanamivir has been iden ti fied in only one clini cal iso late and to oselta mivir in fewer than 10 clini cal iso lates (13).Be cause of their ac tiv ity against in flu enza A and B, their im proved side ef fect pro file, the re duced risk of medi ca tion errors when a sin gle dose is used, and the re duced se lec tion of re sis tance, it seems likely that neu ramini dase in hibi tors will be come the drugs of choice for mass an tivi ral pro phy laxis in long term care fa cili ties.How ever, more data are re quired to es tab lish their ef fi cacy, and, at the mo ment, they are con sidera bly more ex pen sive than aman tadine.In sti tu tions, as well as those re spon si ble for the pay ment of an tivi ral pro phy laxis in nurs ing home out breaks, need to look care fully at the overall costs of aman tadine (in clud ing the cost of an nual resi dent as sess ment and in di vidu al ized dose cal cu la tions), and the poten tial risks and bene fits of each drug be fore de cid ing which should be rec om mended and re im bursed in fu ture sea sons.Simi larly, fa cili ties and pub lic health de part ments that are con sid er ing of fer ing or re quir ing staff pro phy laxis should take into ac count not only the drug cost but also the rate of perceived and ac tual side ef fects, and its impact on staff ing during an out break.There are, how ever, sev eral situa tions for which neu ramini dase in hibi tors are al ready in di cated (Ta ble 3).In the set ting of clini cal aman tadine fail ures dur ing in flu enza A outbreaks, con tinu ing in flu enza A causes se ri ous dis ease (18-21) and neu ramini dase in hibi tors are ef fec tive in its pre ven tion (21,(38)(39)(40)42,43).In flu enza B out breaks are as so ci ated with sub stan tial mor bid ity and mor tal ity in long term care fa cili ties (2), and pro phy laxis will bene fit resi dents in at least some outbreaks.Data on the im pact of pre ven tion of in flu enza B outbreaks in long term care fa cili ties will be dif fi cult to ob tain but are ur gently needed.
Aman tadine has been as so ci ated with an in creased risk of sei zures in those with sei zure dis or ders (31) and with po tentia tion of an ti cho liner gic side ef fects in pa tients on an ti choliner gic medi ca tions.In such pa tients, it is dif fi cult to jus tify the risks of aman tadine side ef fects when other, equally ef fective medi ca tions are avail able, and the cost of us ing neuramini dase in hibi tors may be off set by the re duced need for added care and in ves ti ga tion when ad verse events oc cur.This is par ticu larly true in set tings where a ma jor ity of resi dents have con tra in di ca tions or rela tive con tra in di ca tions.In such set tings, us ing aman tadine for some resi dents and neu raminidase in hibi tors for oth ers sig nifi cantly com pli cates man agement in a set ting of great stress, and us ing a neu ramini dase in hibi tor for all resi dents may be the pre ferred strat egy.
Fi nally, the value of aman tadine pro phy laxis in out breaks may be com pro mised by emerg ing re sis tance if aman tadine is used con comi tantly to treat resi dents with in flu enza (3).Both aman tadine and neu ramini dase in hibi tors have been shown to re duce the du ra tion and se ver ity of ill ness in acutely ill adults if treat ment can be started within 48 h of the on set of symp toms (3,12).Al though data on the bene fits of treat ment in the frail eld erly are few, there is no rea son to be lieve that treat ment ef fi cacy will be dif fer ent in this age group, and treatment of acute in flu enza in the frail in sti tu tion al ized eld erly may be rea son able, as long as it is started within the first 48 h of symp toms.Treat ment with aman tadine will in crease the risk of se lec tion for aman tadine re sis tance and fail ure to con trol the out break.Thus, in an out break in which aman tadine is be ing used for mass pro phy laxis of resi dents, neu ramini dase in hibitors should be con sid ered for the treat ment of ill ness.
An tivi ral pro phy laxis and out breaks TA BLE 3 Situa tions dur ing in sti tu tional out breaks of in flu enza in which neu ramini dase in hibi tors may be in di cated

Situa tion Ra tion ale
Out breaks of in flu enza A in which cases of con firmed in flu enza con tinue to oc cur, with an on set of more than 96 h af ter the ini tia tion of aman tadine pro phy laxis These cases may be due to the emer gence of aman tadine re sis tance.In this set ting, aman tadine is no longer ef fec tive Out breaks of in flu enza B Aman tadine is not ef fec tive against in flu enza B; neu ramini dase in hibi tors are Resi dents who are at risk of se ri ous side ef fects from aman tadine* In cludes those who: • re quire mul ti ple ma jor tran quil iz ers or other medi ca tions with an ti cho liner gic side ef fects • have sei zure dis or ders Treat ment of symp to matic resi dents in out breaks of in flu enza A in which aman tadine is be ing used for pro phy laxis Aman tadine re sis tance is more likely to arise in this set ting due to the ex po sure of the vi rus to aman tadine in treated pa tients.Re sis tance arises much less com monly to neu ramini dase in hibi tors.In ad di tion, data sup port ing the use of neu ramini dase in hibi tors for treat ment are bet ter than those for aman tadine *Where pa tients with rela tive con tra in di ca tions to aman tadine make up a ma jor ity of resi dents (eg, geri at ric psy chi at ric units), us ing a neu ramini dase in hibi tor for all resi dents may sim plify out break man age ment

An tivi ral pro phy laxis and out breaks TA BLE 1 Na tional Ad vi sory Com mit tee on Im mu ni za tion rec om men da tions for dos age of aman tadine in per sons over the age of 65 years Cre ati nine clear ance* Dose
J In fect Dis Vol 11 No 4 July/August 2000 189

McGeer et al TA BLE 2 Pro posed once daily dos ing sched ule for aman tadine so lu tion (10 mg/mL) in per sons over the age of 65 years*
Dos ing sched ule de vel oped based on Na tional Ad vi sory Com mit tee on Im mu ni za tion guide lines, with daily dos ing in cre ments set at 2.5 mL to per mit the use of medi cine cups marked at each 2.5 mL.† See Ta ble 1 for method to es ti mate cre ati nine clear ance.‡ No daily dose; if out break continues, re peat 100 mg dose every seven days dur ing the out break 190 Can J In fect Dis Vol 11 No 4 July/August 2000 *