Pe sis ence wi h s a in he apy in Hunga y
Zol an Kiss1, Laszlo Nagy1, Is an Reibe 2, Gyö gy Pa agh3, Ma k Pe e Molna 4, Gyö gy Rokszin5,
Zsol Abonyi-To h5, Laszlo Ma k6
Abs ac
IInn oodduucc iioonn:: Pe sis ence wi h lipid-lowe ing d ug he apy by ca dio ascula
pa ien s in Hunga y has no been s udied p e iously. This s udy was designed
o de e mine he a e wi h which Hunga ian pa ien s wi h hype lipidemia pe -
sis in aking lipid-lowe ing agen s, and o compa e his wi h a es epo ed
om o he coun ies.
MMaa ee iiaall aanndd mmee hhooddss:: This was a e ospec i e s udy ha u ilized da a om
he Ins i u ional Da abase o he Na ional Heal h Insu ance Fund o analyze pe -
sis ence a es wi h s a ins and eze imibe. The s udy included da a o pa ien s
who s a ed lipid-lowe ing he apy be ween Janua y 1, 2007, and Ma ch 31, 2009.
Va iables included ype o lipid-lowe ing he apy, yea o he apy s a , and
pa ien age. Main ou come measu es we e medians o pe sis ence in mon hs,
pe cen ages o pa ien s pe sis ing in he apy o 6 and 12 mon hs, and Kaplan-
Meie pe sis ence plo s.
RReessuull ss:: The pe cen age o pa ien s who pe sis ed wi h o e all s a in he apy
was 46% a e 1 mon h, 40.3% a e 2 mon hs, 27% a e 6 mon hs, and 20.1%
a e 12 mon hs. Pe sis ence was sligh ly g ea e o s a in he apy s a ed du -
ing 2008 han du ing 2007. Olde pa ien s we e mo e pe sis en wi h he apy
han younge pa ien s. Pe sis ence wi h he combina ion o eze imibe-s a in
he apy was g ea e han wi h s a in o eze imibe mono he apy.
CCoonncclluussiioonnss:: Pe sis ence wi h s a in he apy by pa ien s in Hunga y was low
compa ed wi h o he coun ies. Low pe sis ence may ha e nega ed po en ial
clinical bene i s o long- e m s a in he apy.
KKeeyy wwoo ddss:: pe sis ence, adhe ence, s a ins, eze imibe, Hunga y.
In oduc ion
Hunga ians ha e a highe isk o p ema u e dea h due o ca dio ascu-
la disease (CVD) compa ed wi h esiden s o wes e n Eu opean coun ies
[1, 2] and he lowes li e expec ancy a bi h (72.4 yea s) among a ea
na ions in he O ganiza ion o Economic Coope a ion and De elopmen
(OECD) [3]. The impo ance o lipid-lowe ing he apy is well es ablished
[4, 5]. Rega ding he Hunga ian lipid goal achie emen a es in ecen yea s
he e a e signi ican imp o emen s, bu he esul is s ill a om expec-
CCoo eessppoonnddiinngg aauu hhoo ::
Laszlo Ma k MD, PhD
2nd Depa men o In e nal
Medicine – Ca diology
Pandy Kalman Bekes
Coun y Hospi al
Semmelweis u. 1
P.O. Box 46
5701 Gyula, Hunga y
Phone: +36-209288053
Fax: +36-66526543
E-mail: ma [email protected]
Clinical esea ch
1MULTI GAP S udy G oup, Budapes , Hunga y
2S . Geo ge Feje Coun y Hospi al, 4 h Depa men o Medicine, Szekes ehe a ,
Hunga y
3Medical and Heal h Science Cen e, Uni e si y o Deb ecen, 1s Depa men o Medicine,
Deb ecen, Hunga y
4Co inus Uni e si y, Budapes , Hunga y
5Rx Ta ge S a is ical Agency, Budapes , Hunga y
6Pandy Kalman Bekes Coun y Hospi al, 2nd Depa men o Medicine – Ca diology,
Gyula, Hunga y
SSuubbmmii eedd:: 7 Janua y 2013
AAcccceepp eedd:: 13 Ap il 2013
A ch Med Sci 2013; 9, 3: 409-417
DOI: 10.5114/aoms.2013.35327
Copy igh © 2013 Te media & Banach
410 A ch Med Sci 3, June / 2013
Zol an Kiss, Laszlo Nagy, Is an Reibe , Gyö gy Pa agh, Ma k Pe e Molna , Gyö gy Rokszin, Zsol Abonyi-To h, Laszlo Ma k
a ions [6-8]. In a 2007 summa y o 14 key clinical
ials o lipid lowe ing conduc ed since 2004 and
in ol ing 90 056 indi iduals [9]. Ray e al. de e -
mined ha s a in he apy was sa e and signi ican ly
educed he ca dio ascula endpoin s. Fu he mo e,
a ecen me a-analysis o da a om 170 000 pa -
icipan s in 26 andomized ials o s a in he apy
demons a ed ha mo e in ensi e lowe ing o low-
densi y lipop o ein choles e ol (LDL-C) sa ely p o-
duced signi ican u he educ ions (15%; 95% con-
idence in e al (CI) 11-18; p< 0.0001) in majo
ascula e en s, wi h no e idence o a h eshold
wi hin he choles e ol ange s udies below which
u he educ ions did no esul in inc emen al
educ ions in CVD isk [10]. E en among hose a -
aining 1.8 mmol/l o lowe wi h a s anda d s a in
egimen, u he educ ion yielded bene i ( ela i e
isk (RR) 0.63, 99% CI 0.41-0.95; p = 0.004).
In e na ional clinical ials ha e shown ha he
op imal ca dio ascula ad an ages o s a in he a-
py a e achie ed only wi h he pa ien ’s long- e m
pe sis ence (i.e., con inuing he ea men o he
p esc ibed du a ion) and adhe ence (i.e., ex en o
con o mi y wi h he p esc ibed iming, dosage, and
equency o a egimen) [11, 12]. The ea lies poin
in some ials a which pa ien s who pe sis ed wi h
ac i e he apy began o expe ience ca dio ascula
ad an ages o s a in compa ed wi h placebo he -
apy ( he poin o isual di e gence – PVD) [13] was
a 6 [14] o 12 mon hs o he apy [15]. In a s udy
ha assessed pa ien adhe ence jus a e hospi-
aliza ion, Ho e al. ound ha pa ien s who dis-
con inued s a in he apy 1 mon h a e ini ia ion
ollowing an acu e myoca dial in a c ion had sig-
ni ican ly lowe 1-yea su i al compa ed wi h hose
who con inued wi h s a in he apy [16].
Lack o pe sis ence wi h medica ions in gene al,
and wi h s a in he apy in pa icula , emains a wide-
sp ead p oblem wo ldwide [17-19]. Se e al decades
o esea ch indica e ha up o 20% o all pa ien s
do no ill a new p esc ip ion, ega dless o diagno-
sis, and app oxima ely one hal o hose who do ill
a new p esc ip ion discon inue he apy in he i s
6 mon hs [20]. Up o 69% o all medica ion- ela ed
hospi al admissions a e due o subop imal medica-
ion adhe ence [19]. Lack o pe sis ence wi h s a in
he apy in pa icula has been associa ed wi h in -
c eased ca dio ascula e en s (odds a io (OR) 1.40,
95% CI 1.35-1.45 in p ima y p e en ion; OR 1.59,
95% CI 1.51-1.68 in seconda y p e en ion) and in -
c eased all-cause mo ali y (haza d a io (HR) 1.25,
95% CI 1.09-1.42; p = 0.001) [17, 18, 21-23]. In addi ion,
s a in he apy o insu icien pe sis ence is associ-
a ed wi h highe hospi aliza ion a es and o al
di ec heal hca e cos s compa ed wi h good adhe -
ence o he apy o e he i s 2 yea s o use [24].
Po en ial ba ie s o illing and adhe ing o p e-
sc ip ion ea men s include he p esence o psy-
chological diso de s (e.g. dep ession) o cogni i e
impai men ; ad e se e ec s, cos s (and copaymen
s a us), and complexi y o medica ion egimens;
pa ien s' lack o insigh in o hei illnesses and/o
belie in he bene i s o hei ea men s; and poo
p o ide -pa ien ela ionships, such as inadequa e
ollow-up o ea men planning and/o missed
appoin men s; and logis ical issues such as ans-
po a ion p oblems/incon enience/long pha macy
wai imes, and/o ha ing a su icien supply o
medica ions a home [19, 25].
Cu en ly, only a ew s udies add ess pe sis ence
wi h s a in he apy, he deg ee o which his he -
apy con ibu es o educ ion o ca dio ascula
e en s, o e ec s o he Na ional Heal h Insu ance
Fund paymen s o s a in he apy on imp o ing he
heal h o he gene al popula ion in Hunga y [26].
We unde ook he cu en s udy o de e mine
pe sis ence wi h s a in he apy o Hunga ian pa -
ien s wi h hype choles e olemia, and o compa e
his wi h pe sis ence o he apy in o he coun ies.
The s udy examined gene al pe sis ence, as well as
pe sis ence wi h ini ial he apy and pe sis ence wi h
he apies con aining eze imibe in ce ain g oups.
The s udy also e alua ed he in luence o s a in, age,
and geog aphic loca ion on pe sis ence.
Ma e ial and me hods
This was a e ospec i e s udy o da a main ained
in he d ug-dispensing da abase o he Hunga ian
Na ional Heal h Insu ance Fund, which con ains p e-
sc ip ion da a o all pa ien s ecei ing heal h insu -
ance subsidies. The s udy included da a o all
pa ien s o any age and ei he sex who s a ed lipid-
lowe ing he apy wi h a s a in, eze imibe, o a combi -
na ion o a s a in and eze imibe be ween Janua y 1,
2007, and Ma ch 31, 2009. To help ensu e ha only
pa ien s who began he apy du ing his pe iod we e
included, da a o any pa ien who ecei ed such
he apy in 2006 we e excluded. Da a we e also
excluded o pa ien s who ecei ed a combina ion
o a s a in and any lipid-lowe ing d ug o he han
eze imibe on he ini ial day o he apy, and o pa -
ien s who died du ing he pe iod o he s udy. The
ollow-up pe iod was be ween 4 and 31 mon hs
depending on he beginning o he he apy.
Da a analyses we e conduc ed acco ding o he
yea du ing which he apy was s a ed, he ac i e
ing edien and s eng h o he p oduc (s) ini ially
dispensed, and pa ien age and esidence. Diag-
nos ic da a we e limi ed o he In e na ional Clas-
si ica ion o Diseases (ICD) code w i en on p e-
sc ip ions, limi ing he abili y o de e mine whe he
he apy was aimed a p ima y o seconda y p e-
en ion. On he basis o na ional he apeu ic guide-
lines, howe e , i could be assumed ha a o as-
a in 80 mg and eze imibe we e p esc ibed only o
seconda y p e en ion.
A ch Med Sci 3, June / 2013 411
Pe sis ence wi h s a in he apy in Hunga y
In de e mining pe sis ence, he i s dispensing
o a p esc ip ion was conside ed he s a o s udy
he apy. We conside ed he apy o ha e been dis-
con inued i a pa ien did no ob ain a eplenishmen
o d ug o a leas 60 days ( he “g ace pe iod”), ac -
co ding o he ecommenda ion o he In e na ion-
al Socie y o Pha macoeconomics and Ou comes
Resea ch (ISPOR) [19]. An addi ional analysis was con-
duc ed o all pa ien s ea ed wi h a s a in using
a g ace pe iod o 180 days. Da a we e examined and
esul s epo ed in 1-mon h uni s; a mon h is he
sho es ime span o which he Na ional Heal h
Insu ance Fund p o ides da a. We a emp ed o
de e mine he end o he he apy as accu a ely as
possible by es ima ing he pa ien ’s ac ual use o an
agen , assuming ha , om he i s p esc ip ion,
a pa ien would ha e aken one able daily. S a ing
wi h he second p esc ip ion, we calcula ed how
many days o he apy would ha e been possible wi h
he numbe o able s pu chased o ha da e.
S udy analyses we e pe o med only o hose
pa ien s whose consump ion o able s was calcu-
la ed o be 0.75 o 1.25 imes he numbe p esc ibed
(calcula ed om he numbe o able s dispensed
and he ime pe iods o e which he able s we e
dispensed). By doing so, small luc ua ions could
be excluded, which migh be caused by in e up-
ion o he apy o pa ien s ockpiling o d ugs. In
calcula ing gene al o o al s a in pe sis ence, all
s a in p oduc s we e conside ed oge he , dis e-
ga ding swi ches o a di e en d ug o a di e en
dosage o he same d ug.
Fo ea men g oups o he han he o al s a in
g oup, we conside ed ha ini ial he apy was e -
mina ed when a pa ien had a 60-day pe iod du -
ing which he ini ial he apy was no longe aken
exac ly as p esc ibed, e en i he pa ien hen ook
a d ug ha was simila o he ini ial d ug (e.g., a o -
as a in o sim as a in-eze imibe ixed combina-
ion ins ead o sim as a in), used he ini ial d ug
subsequen ly, o ook a p oduc wi h he same
ac i e ing edien as he ini ial d ug bu wi h a di -
e en dosage. Fo hese speci ic g oups, we he e-
o e de e mined du a ion o pe sis ence wi h he
ini ially p esc ibed d ug egimen, e en hough he
p esc ip ion migh ha e changed. The o al s a in
g oup consis ed o all pa ien s who con inued wi h
any s a in a any dose, while g oups o use s o indi-
idual s a ins consis ed o only hose who emained
on he ini ially p esc ibed s a in and dosage.
The Na ional Heal h Insu ance Fund da abase
enabled us o s udy eze imibe he apy, whe he i
consis ed o eze imibe mono he apy, eze imibe-
s a in ixed combina ion he apy, o he apy wi h
eze imibe plus a s a in as sepa a e p oduc s. We
calcula ed he leng h o he apy om he i s pu -
chase o eze imibe o he comple ion o he eze-
imibe he apy, ega dless o when he eze imibe
he apy was s a ed. The s a o eze imibe he a-
py was no necessa ily simul aneous wi h he s a
o s a in he apy.
SS aa iiss iiccaall aannaallyyssiiss
Su i al analysis me hods we e used o assess
pe sis ence; censo ed da a necessa y o analyses
we e a ailable. Su i al cu es we e cons uc ed
acco ding o Kaplan-Meie analysis (unadjus ed); he
a io o pa ien s s ill ecei ing he apy om mon h o
mon h was calcula ed. Obse a ions we e censo ed
when a pe son died o became ins i u ionalized, o
when he end o he s udy pe iod was eached.
We de e mined he medians o pe sis ence and hei
95% CIs. The medians o pe sis ence in di e en
g oups we e compa ed wi h he Pe o-Wilcoxon es .
Resul s
A o al o 459 034 pa ien s s a ed on any s a -
in he apy: 188 245 in 2007 and 209 716 in 2008
(Table I). The apy was ini ia ed wi h a o as a in in
277 378 pa ien s, sim as a in in 120 921, and osu-
as a in in 19 687. The apy was ini ia ed wi h eze-
imibe in 8893 pa ien s: eze imibe mono he apy in
2112, eze imibe plus a s a in in 2044, and a ixed-
combina ion eze imibe-s a in p oduc in 5145. Age
g oupings o pa ien s anged om < 20 (n = 2568)
o ≥70 yea s (n = 101 301).
In he o al g oup o pa ien s who ecei ed s a in
he apy, 54% discon inued du ing he i s mon h,
based on a g ace pe iod o 60 days. Du ing he sec-
ond mon h, 40.3% o pa ien s we e s ill ecei ing
s a in-based lipid-lowe ing ea men (Figu e 1). This
pe cen age dec eased o 27% a 6 mon hs and o
20.1% a he end o he i s yea (Table I). Because
o he high deg ee o e osion (54%) du ing he i s
mon h, an analysis was conduc ed based on a g ace
pe iod o 180 days and yielded 27% pe sis ence
a e 12 mon hs. F om he o al 459 034 pa ien s in
5590 cases he e we e enough da a abou a o me
myoca dial in a c ion o pe cu aneous co ona y
in e en ion in he da abase. In hese cases he pe -
sis ence a 1 mon h was 76%, a 6 mon hs 61%,
and a 12 mon hs 50%. These a e a o able da a,
bu no absolu e. The eal seconda y p e en ion
popula ion migh be highe .
Du ing he i s yea o he apy, pe sis ence o
all s a in he apy s a ed in 2008 was sligh ly
g ea e han o he apy s a ed in 2007. Du ing no
mon h was he di e ence be ween yea s as much
as 4%. The e o e, al hough he di e ence be ween
he median pe sis ence o he apies s a ed in 2007
and 2008 was s a is ically signi ican (p < 0.001), i
was no deemed clinically ele an .
The median pe sis ence o a o as a in and sim-
as a in was 1 mon h, and o osu as a in i was
2 mon hs (Table I); di e ences we e s a is ically
412 A ch Med Sci 3, June / 2013
Zol an Kiss, Laszlo Nagy, Is an Reibe , Gyö gy Pa agh, Ma k Pe e Molna , Gyö gy Rokszin, Zsol Abonyi-To h, Laszlo Ma k
signi ican (p < 0.001). Pe sis ence cu es we e also
de e mined sepa a ely o he indi idual s a ins. The
g ea es di e ence be ween pe sis ence cu es was
only 6% (obse ed in he i s mon h), indica ing
ha he di e ences we e no o clinical signi icance.
Cu es o pe sis ence wi h di e en doses o sim as -
a in a e shown in Figu e 2. Pe sis ence was g ea es
wi h he 20-mg dose, lowe wi h he 40-mg dose, and
lowes wi h he 10-mg dose. The di e ence be ween
he g oups was s a is ically signi ican (p < 0.001)
bu no clinically signi ican , wi h he g ea es di -
e ence being only 6.3%, seen in he hi d mon h.
Cu es o pe sis ence wi h di e en doses o
a o as a in a e shown in Figu e 3. A e he i s
mon h, pe sis ence dec eased p og essi ely in he
ollowing dose o de : 20 mg, 10 mg, 40 mg, and
GG oouuppNNuummbbee oo PPee ssiiss eennccee mmeeddiiaann iinn mmoonn hhss PPee cceenn aaggee oo ppaa iieenn ss
ppaa iieenn ss((9955%% ccoonn iiddeennccee iinn ee aall))ppee ssiiss iinngg iinn hhee aappyy
aa 66 aanndd 1122 mmoonn hhss
AAllll ((oo oo aall)) ss aa iinnss**
All pa ien s ecei ing s a in he apy 459034 1 (1.1) 27%, 20.1%
2007 s a o s a in he apy 188245 1 (1.1) 25.4%, 20.3%
2008 s a o s a in he apy 209716 1 (1.1) 27.4%, 19.9%
IInnddii iidduuaall ss aa iinn hhee aappiieess****
A o as a in, all 277378 1 (1.1) 27.3%, 20.9%
A o as a in 10 mg 42037 1 (1.1) 22.7%, 16.1%
A o as a in 20 mg 139850 1 (1.1) 24.5%, 17.5%
A o as a in 40 mg 94678 1 (1.1) 20.2%, 14.5%
A o as a in 80 mg 6409 1 (1.1) 13.2%, 8.2%
Rosu as a in, all 19687 2 (1.2) 28.1%, 21.3%
Sim as a in, all 120921 1 (1,1) 26%, 20.5%
Sim as a in 10 mg 12217 1 (1,1) 15.9%, 10%
Sim as a in 20 mg 68778 1 (1.1) 21.6%, 15.3%
Sim as a in 40 mg 41919 1 (1.1) 18.3%, 12.6%
AAggee gg oouuppss [[yyeeaa ss]]******
< 20 2568 1 (1.1) 1%, no da a
≥20 o < 30 9084 1 (1.1) 6.9%, 3.7%
≥30 o < 40 30229 1 (1.1) 12.1%, 7.9%
≥40 o < 50 65137 1 (1.1) 17.6%, 11.8%
≥50 o < 60 139142 1 (1.1) 22.3%, 15.9%
≥60 o < 70 117096 1 (1.1) 28.6%, 21.5%
≥70 101301 1 (1.1) 30.8%, 23.5%
EEzzee iimmiibbee********
All eze imibe g oups n= 8893 2 (2.2) 36.6%, 26.7%
Eze imibemono he apy n= 2112 1 (1.1) 24.1%, 16.6%
Eze imibe add-on o a s a in n= 2044 3 (3.4) 39.4%, 25.5%
Fixed eze imibe-sim as a in combina ion n= 5145 3 (3.3) 38.5%, 27.5%
*This able con ains pa ien s wi h s a in he apy only he e o e pa ien s wi h Eze imibemono he apy a e no lis ed he e. 61,073 pa ien s s a ed
hei he apy in 2009 Q1 pe iod and hei da a was no compa ed o he pa ien s s a ed in 2007 o 2008 because o he much sho e ollow up
ime. **Some ini ial he apies a e no lis ed he e, eg. lu as a in, p a as a in because hei impo ance is qui e low in Hunga y. In he ac i e sub-
s ance le el (eg. sim as a in) he numbe o pa ien s migh be less han in he s eng h le el (eg. sim as a in 10 mg, 20 mg, 40 mg). I is possible
ha a pa ien daily dosage is 1 pill du ing he ini ial sim as a in 10 mg he apy so he/she is isible in he able. He/she may swi ch o sim as a in
20 mg and du ing he he apy may ha e less o mo e han one pill dosage. In his case he/she is excluded om he examina ion in he ac i e sub-
s ance le el. ***The Na ional Heal h Insu ance Found doesn' p o ide da a o g oups whe e he numbe o pa ien s is less han 10. In he able o
age g oups 15 pa ien s ha e been dele ed o ha eason because he o iginal da a was mo e de ailed (age + agen ). ****Eze imibe add-on and
ixed Eze imibe he apy is possible o he same pa ien i he/she swi ches om add-on he apy o ixed combina ion he apy.
TTaabbllee II..G oupings o pa ien s by he apy, age, and geog aphic loca ion, wi h numbe s, median mon hs o pe sis -
ence, and pe sis ence pe cen ages a 6 and 12 mon hs
A ch Med Sci 3, June / 2013 413
Pe sis ence wi h s a in he apy in Hunga y
inally 80 mg, which had a pe sis ence conside ably
lowe han ha o he lowe doses.
The pe sis ence o ini ial he apies inc eased as
he age o pa ien s inc eased (Table I, Figu e 4). A
6 mon hs, pe sis ence was 1% o he < 20-yea age
g oup and 30.8% o he ≥70-yea age g oup, an
app oxima e 30% di e ence be ween he younges
and oldes age g oups. The median pe sis ence o
all age g oups was 1 mon h; he di e ences we e
s a is ically signi ican (p < 0.001).
Pe sis ence da a o di e en eze imibe egimens
a e shown in Table I and Figu e 5. The median pe -
sis ence wi h eze imibe mono he apy was 1 mon h
and ha o eze imibe added o s a in o o ixed
eze imibe-s a in combina ion he apy was 3 mon hs.
The e was no s a is ically signi ican di e ence
be ween he add-on and ixed combina ion egimens
(p = 0.94). Pe sis ence wi h all eze imibe he apies
(median 2 mon hs) was g ea e han o sim as a in
and a o as a in he apies (Table I and Figu e 6).
Discussion
This is he i s s udy, o ou knowledge, o de -
ailed pe sis ence pa e ns wi h s a in he apy
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 11..Kaplan-Meie plo o o al s a in pe sis ence
TToo aall ss aa iinn ppee ssiiss eennccee [[%%]]
012345678910 11 12 13 14 15 16 17 18 19
TTiimmee [[mmoonn hh]]
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 22..Kaplan-Meie plo o pe sis ence o di e -
en sim as a in doses
PPee ssiiss eennccee [[%%]]
012345678910 11 12 13 14 15 16 17 18 19 20
TTiimmee [[mmoonn hh]]
Sim as a in 10 mg Sim as a in 20 mg Sim as a in 40 mg
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 33.. Kaplan-Meie plo o pe sis ence o di e -
en a o as a in doses
PPee ssiiss eennccee [[%%]]
012345678910 11 12 13 14 15 16 17 18 19 20
TTiimmee [[mmoonn hh]]
A o as a in 10 mg A o as a in 20 mg
A o as a in 40 mg A o as a in 80 mg
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 44..Kaplan-Meie plo o pe sis ence o di -
e en age coho s
PPee ssiiss eennccee [[%%]]
012345678910 11 12 13 14 15 16 17 18 19 20
TTiimmee [[mmoonn hh]]
Age < 20 20 ≤Age < 30 30 ≤Age < 40
40 ≤Age < 50 50 ≤Age < 60
60 ≤Age < 70 70 ≤Age
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 55..Kaplan-Meie plo o pe sis ence o eze-
imibe as mono he apy, add-on o a s a in, and eze-
imibe-s a in ixed combina ion
PPee ssiiss eennccee [[%%]]
012345678910 11 12 13 14 15 16 17 18 19 20
TTiimmee [[mmoonn hh]]
Mono EZT Add-on EZT Inegy
100
90
80
70
60
50
40
30
20
10
0
FFiigguu ee 66..Kaplan-Meie plo o o al eze imibe and
o al s a in pe sis ence
PPee ssiiss eennccee [[%%]]
0123456789 11 13 15 17 19 21 23 25 27
TTiimmee [[mmoonn hh]]
Eze imib S a in
414 A ch Med Sci 3, June / 2013
Zol an Kiss, Laszlo Nagy, Is an Reibe , Gyö gy Pa agh, Ma k Pe e Molna , Gyö gy Rokszin, Zsol Abonyi-To h, Laszlo Ma k
among Hunga ian pa ien s. We de e mined ha ,
when pe sis ence was based on a 60-day g ace pe i-
od, he pe cen age o pa ien s who pe sis ed wi h
he apy was only 46% a e 1 mon h, 40.3% a e
2 mon hs, 27% a e 6 mon hs, and 20.1% a e
12 mon hs. Almos 80% o pa ien s had discon inued
he apy du ing he i s yea . When pe sis ence was
based on a mo e lenien g ace pe iod o 180 days,
pe sis ence a e 12 mon hs was inc eased only om
20.1% o 27%. These esul s indica e ha , because
o ea ly discon inua ion, a conside able p opo ion
o Hunga ian pa ien s do no expe ience ad an ages
o s a in he apy. Se e al s udies ha e shown ha
he du a ion o s a in he apy can be a majo ac o
in de e mining he apeu ic bene i ; he apy ha is
no su icien ly pe sis en may p o ide no signi i-
can ca diop o ec ion [14, 15].
Con a y o ou expec a ion, pe sis ence wi h
s a in he apy in Hunga y was no as g ea as in
de eloped coun ies. Compa ed wi h he 1-mon h
median pe sis ence we ound, Kamal-Bahl e al.,
using less s ic c i e ia han he cu en s udy o
de e mining discon inua ion, ound a median pe -
sis ence o 27.5 mon hs in he Uni ed S a es (US)
[27]. In ano he US s udy, Foody e al. used discon-
inua ion c i e ia simila o his s udy, and ound
12-mon h pe sis ence o a o as a in in p ima y
p e en ion o be abou 40%, nea ly wice as high
as ou inding o 20.9%; o sim as a in, hey ound
12-mon h pe sis ence o be abou 30%, compa ed
o 20.5% o ou pa ien s [28]. La sen e al., using
a 30-day g ace pe iod, ound ha only 11.2% o Dan-
ish pa ien s d opped ou o he apy a e 1 mon h
[29], compa ed wi h 54% o Hunga ian pa ien s.
The Man el-Teeuwisse e al. s udy in he Ne he -
lands showed signi ican ly be e 1-yea pe sis ence
han in ou s udy: 61.5% s. 20.1% [30]. Pe sis ence
a es highe han hose in ou s udy we e also
obse ed by Pe eaul e al. in Quebec [31], Helin-
Salmi aa a e al. in Finland [32], Deamb osis e al.
in I aly [33], and Chodick e al. in Is ael [34].
Howe e , ou s udy is no he only one o ha e
ound a disappoin ingly apid d op in d ug usage
a e ini ia ion o he apy o p ima y o seconda y
p e en ion o ch onic condi ions. In se e al la ge,
e ospec i e coho s udies o s a in use in he US
and Canada, he g ea es decline in pe sis ence
occu ed du ing he i s 3-6 mon hs o ea men ,
wi h 25% o almos 50% o pa ien s discon inuing
s a ins wi hin 6 mon hs o s a ing he apy [35]. In
a No h Ame ican coho s udy o elde ly pa ien s
(65 yea s o olde ), almos 15% o he coho dis-
con inued lipid-lowe ing he apy a e hei ini ial
p esc ip ion ill, and mo e han one- hi d o pa ien s
discon inued ea men wi hin 1 yea [36]. The s udy
by Chodick e al. in Is ael ound ha nea ly one-
hi d o he p ima y-p e en ion coho and one-
qua e o he seconda y-p e en ion coho dis-
con inued s a in ea men a e only one p e-
sc ip ion ill [34]. These obse a ions a e consis en
wi h he heal h belie model, acco ding o which
he likelihood o adhe ence o a medica ion egi-
men is de e mined by he pe cei ed h ea o dis-
ease (suscep ibili y and se e i y) and pe cei ed ben-
e i s o he apy [37]. Gi en ha hype lipidemia is
an asymp oma ic disease un il a ca dio ascula
e en occu s, he pe cei ed h ea o disease and
bene i s o lipid-lowe ing he apy may no be an-
gible o many pa ien s, comp omising adhe ence
and pe sis ence.
Ou s udy included ai ly adhe en pa ien s, by
es ic ing da a analyses o pa ien s who we e ak-
ing 0.75-1.25 imes he numbe o able s p esc ibed
o hem. We could no de e mine whe he pa ien s
who ook wice as long as hey should ha e o
ob ain a enewal p esc ip ion we e cu ing hei
able s in hal and we e egula ly aking hal hei
p esc ibed dose daily o we e aking a ull able
e e y o he day. Resul s we e a ec ed o a non-sig-
ni ican ex en (by 8% o 17%) by elimina ing om
he analysis he da a om hose pa ien s who
e illed p esc ip ions egula ly bu less equen ly
(< 0.75) o mo e equen ly (> 1.25) han called o by
hei p esc ip ion.
Medica ion adhe ence and pe sis ence a e in e -
ela ed, and bo h impac clinical ou comes o ea -
men [12]. Wisniowska and Skow on in Poland ound
pe sis ence wi h s a in he apy o be 160 days o
highly adhe en (≥80%) pa ien s and 90 days o
poo ly adhe en (< 80%) pa ien s [38]. The di e -
ence in ou s udy be ween 1-yea pe sis ence based
on 60- and 180-day g ace pe iods indica es ha
ewe han 7% o pa ien s who we e non-adhe en
o a 60-day g ace pe iod esumed he apy du ing
he subsequen 120 days.
A ecen sys ema ic e iew o 19 s udies exam-
ining he ela ionship be ween adhe ence o pe -
sis ence o s a in he apy and clinical ou comes
indica ed ha high le els o adhe ence and longe
du a ions o pe sis ence wi h s a ins a e associa -
ed wi h p og essi ely inc easing clinical bene i s in
p ima y and seconda y p e en ion popula ions,
including signi ican educ ions in all-cause mo -
ali y (OR 0.49-0.66), a al and non a al co ona y
hea disease e en s (OR 0.74-0.83), and hospi al-
iza ions (OR 0.19-0.70) [15]. A ecen compa a i e
cos -e ec i eness analysis de e mined ha bo h
eminde s/educa ional ma e ials and pha ma-
cis /nu se managemen p og ams we e cos -e ec-
i e in e en ions o imp o e adhe ence wi h CVD
medica ions associa ed wi h inc emen al cos -e ec-
i eness a ios (ICER) o $4984 and $6358 pe qual-
i y-adjus ed li e-yea (QALY) gained, espec i ely
[39]. Acco ding o a Wo ld Heal h O ganiza ion
epo on medica ion adhe ence o long- e m he -
apies, “Inc easing he e ec i eness o adhe ence
A ch Med Sci 3, June / 2013 415
Pe sis ence wi h s a in he apy in Hunga y
in e en ions may ha e a a g ea e impac on he
heal h o he popula ion han any imp o emen in
speci ic medical ea men s” [40].
The inc eased pe sis ence obse ed in ou s udy
when compa ing he apy begun in 2008 wi h ha
begun in 2007 was no ma kedly imp o ed, bu
may signal a end owa d con inuing imp o emen
o e ime.
We ound ha discon inua ion o he apy de -
c eased wi h inc easing age; olde pa ien s ended
o be mo e adhe en o physicians’ ins uc ions. In
he six h mon h o he apy, he e was app oxi-
ma ely a 30% di e ence in median pe sis ence
be ween he younges and oldes age g oups. Al -
hough his di e ence was s a is ically signi ican
(p < 0.001), he median pe sis ence o all g oups
was only 1 mon h. One mon h was also he medi-
an pe sis ence o he o al s a in g oup and he
g oups o use s o indi idual s a ins (o he han
osu as a in, which had a median pe sis ence o
2 mon hs). Thus, he pe cen o pa ien s d opping
ou o he apy ea ly is conside able o all age and
he apy g oups, ega dless o any s a in swi ching
o dose changing ha may ha e aken place. The
causes o he discon inua ion we e no s udied. On
eason migh be ha he s a in induced muscle
complain s, he equency o which in obse a ion
s udies could be 5-10% [41]. In a p ospec i e coho
s udy in es iga ing he e ec o physical ac i i y
and s a ins conduc ed on mo e han 10 000 pa ien s
du ing a 10-yea ollow-up he mo ali y isk was
si gni ican ly be e in people aking s a ins han
in hose no aking s a ins (18.5% s. 27.7%)
(p < 0.0001). In pa ien s who ook s a ins, mo al-
i y isk dec eased as i ness inc eased (a simila
end be ween i ness and mo ali y was obse ed
in pa ien s wi hou s a ins) [42].
O he s udies ha e yielded con lic ing esul s
ega ding associa ion o pe sis ence wi h age. Se -
e al ound, as ha e we, ha pe sis ence imp o es
wi h inc easing pa ien age (up o a ce ain ‘ h esh-
old’ in some cases [e.g., 55-64 yea s o age], beyond
which pe sis ence declines) [30, 43, 44], while o h-
e s ound olde age o be a p edic o o poo long-
e m pe sis ence [28, 36].
The median pe sis ence wi h combined eze im-
ibe and s a in he apy – whe he eze imibe was
added on o a s a in o was p esc ibed as an eze-
imibe-s a in ixed combina ion – was 3 mon hs,
which was longe han he 1-mon h pe sis ence
wi h o al s a in he apy. G ea e pe sis ence wi h
combina ion eze imibe and s a in he apy han wi h
o al s a in he apy was obse ed h ough a ull
yea : 25.5% wi h eze imibe add-on he apy and
27.5% wi h ixed-combina ion he apy compa ed
o 20.1% a 12 mon hs o o al s a in pe sis ence.
Based on he Hunga ian subsidy sys em, we
we e able o conside ea men s con aining eze-
imibe p esc ibed o pa ien s wi h CVD as sec-
onda y p e en ion ea men s. We could no de e -
mine whe he s a in mono he apy was used o p i-
ma y o seconda y p e en ion, o he han o
a o as a in 80-mg he apy, which, in Hunga y, is
subsidized only o seconda y p e en ion. Se e al
g oups o in es iga o s ha e epo ed ha pa ien s
ecei ing seconda y-p e en ion he apy we e mo e
adhe en o hei d ug egimen han hose ecei -
ing p ima y-p e en ion he apy, possibly due o an
inc eased app ecia ion o he impo ance o man-
aging hei isk ac o s) [33, 43, 44].
Kamal-Bahl e al. s udied a US popula ion wi h
p i a e heal h insu ance and ound ha , using
a 180-day g ace pe iod, pe sis ence a e 1 yea was
g ea e o s a in use s (71.1%) han o eze imibe
use s (67%) [27]. These in es iga o s did no spec-
i y whe he he apy was used o p ima y o sec-
onda y p e en ion.
The e is a limi a ion o his in es iga ion in ha
we s udied only hose p esc ip ions ha we e sub-
sidized by he Na ional Heal h Insu ance Fund.
Undoub edly he e we e pa ien s who we e pe -
sis en wi h s a in he apy bu we e unknown o
us because hey we e able o con inue aking a d ug
o longe han he 60-day g ace pe iod wi hou
applying o a und subsidy. These pa ien s may
ha e ecei ed he d ug while in he hospi al, been
discha ged om he hospi al wi h a supply o he
d ug, used d ugs le o e om a ela i e’s o
iend’s supply, o ecei ed samples o d ugs.
We did no assess p opo ions o pa ien s mak-
ing and keeping appoin men s o ollow-up lipid
es s and physician isi s, ac o s which can impac
adhe ence o he apy. In a la ge coho s udy, Ben-
ne e al. demons a ed ha pa ien s who ecei ed
ollow-up physician isi s and lipid es s we e 45%
mo e likely o be adhe en (95% CI 1.34-1.55) o lipid-
lowe ing he apy [45].
In conclusion, we can s a e ha he low a e o
pe sis ence wi h s a in he apy ha we obse ed
in Hunga y is likely associa ed wi h occu ence o
ca dio ascula e en s o dea hs ha o he wise
migh ha e been p e en ed by app op ia ely long-
e m con inua ion o s a in he apy. These da a a e
consis en wi h a likely inancial d ain on na ional
esou ces used o subsidize subop imal s a in he -
apy, wi h insu icien yields o heal h imp o emen
in he Hunga ian popula ion. Pe sis ence wi h lipid-
lowe ing he apy among Hunga ian pa ien s needs
o be inc eased. The key o imp o emen is he
pa ien -physician ela ion. E e y ime a pa ien
mee s a heal h ca e p o essional he a en ion o
he impo ance o aking he pills p ope ly has o
be d awn up. The possible side e ec s mus be dis-
cussed in de ail. The pa ien s’ mo i a ion could be
inc eased by explaining he ole o choles e ol in
he p e en ion o a he oscle osis. B oad p o essional
416 A ch Med Sci 3, June / 2013
Zol an Kiss, Laszlo Nagy, Is an Reibe , Gyö gy Pa agh, Ma k Pe e Molna , Gyö gy Rokszin, Zsol Abonyi-To h, Laszlo Ma k
coope a i e e o s, including hose o pha macis s,
in o de o op imize he bene icial e ec s o lipid he -
apies, o imp o e he pa ien s’ pe sis ence, could u -
he educe ca dio ascula mo bidi y and mo ali y.
Acknowledgmen s
Da a om he Hunga ian Ins i u ional Da abase
o he Na ional Heal h Insu ance Fund we e used
in he s udy.
Re e ences
1. Józan P. Changes in he socio-economic sys em and
poli ical egime and he ou h s age o he epidemio -
logical ansi ion [Hunga ian]. O He il 2012; 153: 662-77.
2. Helis E, Augus incic L, S eine S, e al. Time ends in
ca dio ascula and all-cause mo ali y in he 'old' and
'new' Eu opean Union coun ies. Eu J Ca dio asc P e
Rehabil 2011; 18: 347-59.
3. Szucs TD. Healing unheal hy Hunga ian hea s – a wha
cos ? Wi h wha bene i ? Ca dio asc D ugs The 2006;
20: 239-43.
4. Me cando AD, Lai HM, A onow WS, e al. Reduc ion in
a he oscle o ic e en s: a e ospec i e s udy in an ou -
pa ien ca diology p ac ice. A ch Med Sci 2012; 8: 57-62.
5. Lai HM, A onow WS, Me cando AD, e al. The impac o
s a in he apy on long- e m ca dio ascula ou comes in
an ou pa ien ca diology p ac ice. A ch Med Sci 2012; 8:
53-56.
6. Ma k L, Pa agh GY, Ka adi I, e al. Changes in a ainmen
o lipid goals by gene al p ac i ione s and specialis s in
pa ien s a high ca dio ascula isk in Hunga y du ing
2004-2008. A ch Med Sci 2010; 6: 695-700.
7. Ma k L, Pa agh GY, Ka adi I, e al. An a emp o make
lipid-lowe ing he apy mo e e ec i e in Hunga y. The
esul s o MULTI GAP 2010 and he Plus P og am. A ch
Med Sci 2011; 7: 760-6.
8. Ma k L, Pa agh GY, Ka adi I, e al. How can we u he im -
p o e he LDL-choles e ol a ge le el achie emen a e
based on he Hunga ian MULTI GAP 2011 s udy esul s
and conside ing he new Eu opean dyslipidaemia
guidelines? A ch Med Sci 2012; 8: 608-13.
9. Ray KK, Cannon CP, B aunwald E. Recen ials o lipid
lowe ing. In J Clin P ac 2007; 61: 1145-59.
10. Choles e ol T ea men T ialis s’ (CTT) Collabo a ion. E icacy
and sa e y o mo e in ensi e lowe ing o LDL choles e ol:
a me a-analysis o da a om 170 000 pa icipan s in 26
andomised ials. Lance 2010; 376: 1670-81.
11. Pe e son AM, Nau DP, C ame JA, e al. A checklis o
medica ion compliance and pe sis ence s udies using
e ospec i e da abases. Value Heal h 2007; 10: 3-12.
12. C ame JA, Roy A, Bu ell A, e al. Medica ion compliance
and pe sis ence: e minology and de ini ions. Value Heal h
2008; 11: 44-7.
13. Ca dinal H, Mon a ed AA, Do ais M, e al. The concep o
he 'pe cen was ed pa ien s' in p e en i e heal h s a -
egies. Pha macoepidemiol D ug Sa 2006; 15: 57-61.
14. Downs JR, Clea ield M, Weis S, e al. P ima y p e en ion
o acu e co ona y e en s wi h lo as a in in men and
women wi h a e age choles e ol le els: esul s o AFCAPS/
TexCAPS. Ai Fo ce/Texas Co ona y A he o scle o sis P e -
en ion S udy. JAMA 1998; 279: 1615-22.
15. Simpson RJ J , Mendys P. The e ec s o adhe ence and
pe sis ence on clinical ou comes in pa ien s ea ed wi h
s a ins: a sys ema ic e iew. J Clin Lipidol 2010; 4: 462-71.
16. Ho PM, Spe us JA, Masoudi FA, e al. Impac o me -
dica ion he apy discon inua ion on mo ali y a e myo -
ca dial in a c ion. A ch In e n Med 2006; 166: 1842-7.
17. Mo ide Y, Hegele RA, Lange A, e al. Clinical and public
heal h assessmen o bene i s and isks o s a ins in
p ima y p e en ion o co ona y e en s: esol ed and
un esol ed issues. Can J Ca diol 2008; 24: 293-300.
18. Pe eaul S, Blais L, Lama e D, e al. Pe sis ence and
de e minan s o s a in he apy among middle-aged
pa ien s o p ima y and seconda y p e en ion. B J Clin
Pha macol 2005; 59: 564-73.
19. Os e be g L, Blaschke T. Adhe ence o medica ion. N Engl
J Med 2005; 353: 487-97.
20. McHo ney CA. The adhe ence es ima o : a b ie , p oximal
sc eene o pa ien p opensi y o adhe e o p esc ip ion
medica ions o ch onic disease. Cu Med Res Opin 2009;
25: 215-38.
21. Poluzzi E, Piccinni C, Ca a P, e al. Ca dio ascula e en s
in s a in ecipien s: impac o adhe ence o ea men in
a 3-yea eco d linkage s udy. Eu J Clin Pha macol 2011;
67: 407-14.
22. Rasmussen JN, Chong A, Al e DA. Rela ionship be ween
adhe ence o e idence-based pha maco he apy and long-
e m mo ali y a e acu e myoca dial in a c ion. JAMA
2007; 297: 177-86.
23. Lindg en P, E iksson J, Bux on M, e al.; he ASCOT ial
in es iga o s. The economic consequences o non-
adhe ence o lipid-lowe ing he apy: esul s om he
Anglo-Scandina ian-Ca diac Ou comes T ial. In J Clin P ac
2010; 64: 1228-34.
24. Aube RE, Yao J, Xia F, Ga a aglia SB. Is he e a e -
la ionship be ween ea ly s a in compliance and a e -
duc ion in heal hca e u iliza ion? Am J Manag Ca e 2010;
16: 459-66.
25. Gadka i AS, McHo ney CA. Medica ion non ul illmen a es
and easons: na a i e sys ema ic e iew. Cu Med Res
Opin 2010; 26: 683-705.
26. Je mendy G, Wi mann I, Nagy L, Kiss Z, e al. Pe sis ence
o ini ial o al an idiabe ic ea men in pa ien s wi h ype
2 diabe es melli us. Med Sci Moni 2012; 18: CR72-7.
27. Kamal-Bahl SJ, Bu ke T, Wa son D, e al. Discon inua ion
o lipid modi ying d ugs among comme cially insu ed
Uni ed S a es pa ien s in ecen clinical p ac ice. Am
J Ca diol 2007; 99: 530-4.
28. Foody JM, Joyce AT, Rudolph AE, e al. Pe sis ence o
a o as a in and sim as a in among pa ien s wi h and
wi hou p io ca dio ascula diseases: a US managed ca e
s udy. Cu Med Res Opin 2008; 24: 1987-2000.
29. La sen J, Ande sen M, K ags up J, e al. High pe sis ence
o s a in use in a Danish popula ion: compliance s udy
1993-1998. B J Clin Pha macol 2002; 53: 375-8.
30. Man el-Teeuwisse AK, Goe sch WG, Klungel OH, e al.
Long e m pe sis ence wi h s a in ea men in daily
medical p ac ice. Hea 2004; 90: 1065-6.
31. Pe eaul S, Blais L, D agomi A, e al. Pe sis ence and
de e minan s o s a in he apy among middle-aged
pa ien s ee o ca dio ascula disease. Eu J Clin
Pha macol 2005; 61: 667-74.
32. Helin-Salmi aa a A, La ikainen P, Ruokoniemi P, e al.
Pe sis ence wi h s a in he apy in diabe ic and non-
diabe ic pe sons: a na ion-wide egis e s udy in 1995-
2005 in Finland. Diabe es Res Clin P ac 2009; 84: e9-11
33. Deamb osis P, Sa amin C, Te azzani G, e al. E alua ion
o he p esc ip ion and u iliza ion pa e ns o s a ins in
an I alian local heal h uni du ing he pe iod 1994-2003.
Eu J Clin Pha macol 2007; 63: 197-203.
34. Chodick G, Shale V, Ge be Y, e al. Long- e m pe sis ence
wi h s a in ea men in a no - o -p o i heal h main -
A ch Med Sci 3, June / 2013 417
Pe sis ence wi h s a in he apy in Hunga y
enance o ganiza ion: a popula ion-based e ospec i e
coho s udy in Is ael. Clin The 2008; 30: 2167-79.
35. Jacke icius CA, Mamdani M, Tu JV. Adhe ence wi h s a in
he apy in elde ly pa ien s wi h and wi hou acu e
co ona y synd omes. JAMA 2002; 288: 462-7.
36. Benne JS, Glynn RJ, Mogun H, e al. Long- e m pe -
sis ence in use o s a in he apy in elde ly pa ien s. JAMA
2002; 288: 455-61.
37. Rosens ock IM, S eche VJ, Becke MH. Social lea ning
heo y and he Heal h Belie Model. Heal h Educ Q 1988;
15: 175-83.
38. Wisniowska B, Skow on A. E alua ion o pa ien s' ad -
he ence o s a ins in Poland. Cu Med Res Opin 2011; 27:
99-105.
39. Chapman RH, Kowal SL, Che y SB, e al. The modeled
li e ime cos -e ec i eness o published adhe ence-
imp o ing in e en ions o an ihype ensi e and lipid-
lowe ing medica ions. Value Heal h 2010; 13: 685-94.
40. Wo ld Heal h O ganiza ion. Adhe ence o long- e m
he apies. E idence o ac ion. Wo ld Heal h O ganiza ion
2003. h p://www.who.in /chp/knowledge/publica ions/
adhe ence_ epo /en/. Accessed Janua y 7, 2011.
41. B ucke E, Hayem G, Dejage S, e al. Mild o mode a e
muscula symp oms wi h high-dosage s a in he apy in
hype lipidemic pa ien s: he PRIMO s udy. Ca dio asc
D ugs The 2005; 19: 403-14.
42. Kokkinos PF, Faselis C, Mye s J, e al. In e ac i e e ec s
o i ness and s a in ea men on mo ali y isk in
e e ans wi h dyslipidaemia: a coho s udy. Lance 2013;
381: 394-9.
43. Benne JS, Chapman RH, Pe illa AA, Tank SS, Rosenbe g N,
Schwa z JS. Associa ion be ween p esc ip ion bu den and
medica ion adhe ence in pa ien s ini ia ing an ihype -
ensi e and lipid-lowe ing he apy. Am J Heal h Sys
Pha m 2009; 66: 1471-7.
44. Chapman RH, Benne JS, Pe illa AA, e al. P edic o s o
adhe ence wi h an ihype ensi e and lipid-lowe ing
he apy. A ch In e n Med 2005; 165: 1147-52.
45. Benne JS, Tie ce JC, Ballan yne CM, e al. Follow-up lipid
es s and physician isi s a e associa ed wi h imp o ed
adhe ence o s a in he apy. Pha macoeconomics 2004;
22 (suppl. 3): 13-23.
