trends in prescription and determinants of persistence to antihypertensive therapy

Trends in Prescription and Determinants ofPersistence to Antihypertensive TherapyThe PAPEETE Studyy

Francesco Vittorio Costa,1 Luca Degli Esposti,2 Carlo Cerra,3 Chiara Veronesi2 and Stefano Buda2

1 Madre Fortunata Toniolo Hospital, Bologna, Italy

2 CliCon S.r.l. Health, Economics & Outcomes Research, Ravenna, Italy

3 Dipartimento Programmazione, Acquisto e Controllo, ASL di Pavia, Pavia, Italy

Abstract Objectives: To assess trends in prescriptions, determinants and timing of treatment discontinuation and/orchanges in antihypertensive drug therapy in a cohort of hypertensive patients living in Pavia, a city in the

north of Italy.

Methods: The cohort included 61493 patients aged ‡18 years who received their first antihypertensive drug

prescription (monotherapy, fixed or extemporaneous combination) during the period 2003–6. Patients were

classified as ‘persistent’ if 12 months after the beginning of treatment they were still taking a regular therapy

(same drug= ‘same therapy users’, added one or more drugs= ‘add-on therapy users’, different drug= ‘switch-

ers’).Otherwise, theywere classified as ‘non-persistent’ (stopping therapy after the first prescription= ‘occasional

users’; stopping treatment early= ‘stoppers’; taking medicines in an erratic fashion= ‘intermittent users’).

Results: ACE inhibitors were the most frequently prescribed drugs (22.8%), followed by b-adrenoceptorantagonists (b-blockers) [14.3%], diuretics (13.9%), Ca2+ antagonists (11.4%) and angiotensin II type 1 re-

ceptor antagonists (angiotensin receptor blockers [ARBs]) [9.3%]. After 12 months, persistent patients were

only 11.2% (same therapy users 6.7%, switchers 3.2%, add-on therapy users 1.3%). Non-persistent patients

were 88.8% (35.3% occasional users, 20.6% stoppers, 32.8% intermittent users). Patient-related predictors of

persistence were older age, male sex, concomitant treatment with antidiabetic and hypolipidaemic drugs and

previous hospitalizations for cardiovascular events. Highest level of persistence was seen in patients starting

with ARBs (18.8%), followed by ACE inhibitors (11.4%), b-blockers (11.0%), Ca2+ antagonists (10.8%) and

diuretics (3.0%). Among ARBs, considering separately monotherapy and fixed-combination therapy,

highest level of persistence was observed in patients starting with candesartan, irbesartan, valsartan and

telmisartan given in monotherapy, and with valsartan and telmisartan given in fixed-dose combination.

Conclusions: Persistence to antihypertensive treatment at 12 months is only 11.2%, being the lowest with

diuretics (3.0%) and the highest with ARBs (18.8%).

Received for publication 7 August 2009; accepted for publication 3 September 2009.

Keywords: antihypertensive therapy, drug utilization, persistence to treatment, angiotensin receptor

blockers, predictors of non-persistence, administrative databases.

Background

Controlled clinical trials have shown that continuous treat-

ment with antihypertensive drugs can significantly lower blood

pressure in a very high proportion of patients, thus producing a

striking decrease in cardiovascular (CV) events.[1-4] However, in

everyday practice, results of treatment are quite worse and the

proportion of treated patients achieving a normal blood pres-

sure is usually <25%. In European countries, on average only

8% of hypertensive patients have their blood pressure at goal

compared with 23% in Canada and the US.[5]

Recently, the EUROASPIRE III survey showed that blood

pressure control has had no improvement with time, but instead

a 10% worsening over the three surveys.[6]

y Population-Based Analysis of PErsistence with Treatment and Economics of TElmisartan Study.

ORIGINAL RESEARCH ARTICLEHigh Blood Press Cardiovasc Prev 2009; 16 (4): 167-176

1120-9879/09/0004-0167/$49.95/0

ª 2009 Adis Data Information BV. All rights reserved.

The reasons for this therapeutic fiasco are complex and

multifactorial. There certainly is no shortage of pharmacolo-

gical treatments nor is there a lack of detailed guidelines. Main

factors are probably low-dose prescriptions, inadequate up-

titration of doses when needed (the so-called ‘therapeutic in-

ertia’),[7] patients and doctors underestimation of risk[8] and

poor compliance and persistence to treatment.[9]

Poor compliance (taking medications at the prescribed in-

tervals and dosing regimen) and persistence (continuous use of

medications for the specified treatment time period) are quite

common in hypertensive patients[10-18] and are probably the

single most relevant factors affecting the results of treatment.

The consequences of lack of adherence to treatment cannot

be easily understood from clinical trials data because the tight

control imposed in trials may artificially increase compliance.

This study was designed to evaluate persistence to antihyper-

tensive drugs in actual everyday practice. We examined newly

treated patients living in Pavia, a city located in the north-west

part of Italy in which an automated database is utilized to get

information on the use of health services. The database has

been analysed to determine the initially prescribed antihyper-

tensive drug therapy, rate and time of treatment discontinua-

tion, combination with other drugs and switching of the initial

treatment with another one.We also evaluated factors (patient-

and drug-related) affecting persistence to therapy and econo-

mic cost of drugs in persistent and non-persistent patients.

Methods

Data Source

The study subjects were enrolled from individuals registered

with the Local Health Unit (LHU) of Pavia – an area located in

the north of Italy (approximately 500 000 inhabitants) – and

permanently eligible over the study period. The Ethic Com-

mittee of the LHU of Pavia approved this study. The LHU is a

body delegated by theNationalHealth System (NHS) to serve a

specific geographical area, generally a province, providing

healthcare. The LHU, being a point of delivery for the central

health system, has an information network that routinely

measures the volume of expenditure generated by the dispen-

sing of drugs to beneficiaries. In particular, this adminis-

trative/accounting type archive is used conventionally for

recording the amounts pharmacies are entitled to receive from

the LHU by way of refund in respect of drugs reimbursable by

the NHS and dispensed free of charge. This archive has been

structured in such away as to enable a patient-oriented reading:

the prescriptions recorded are attributed in each case to the

patient-recipient. The data available in each prescription in-

clude the patient’s national health number, the prescribing

physician’s number, the Anatomical-Therapeutic-Chemical

(ATC) code of the drug purchased, the number of packs, the

number of units per pack, the dosages, the unit cost per pack

and the prescription date. The identification of the patient given

by the personal health number, cross-checked with the registry

office and hospital database, allows the information to be in-

tegrated with date of birth, sex and any record of previous

hospitalizations for cardio-cerebrovascular diseases.

Study Design

This was a retrospective cohort study, which included only the

records of patients who were receiving antihypertensive drugs

for the first time. Those enrolled for the study were new users,

18 years of age or over, who had received a first prescription for

one of the following classes of antihypertensive drugs between

1 January 2003 and 31December 2006 (enrolment period): diuret-

ics (ATC code C03), b-adrenoceptor antagonists (b-blockers)[ATC code C07A/B], calcium channel-blockers (ATC code

C08), ACE inhibitors (ATC code C09A/B), angiotensin II type

1 receptor antagonists (angiotensin receptor blockers [ARBs])

[ATC code C09C/D] or other antihypertensive drugs (ATC

code C02). We included ARBs being already on the market at

the beginning of 2002.

New users were defined as patients who had not been pre-

scribed any antihypertensivedrugs in the 12months preceding the

enrolment date (date of the first prescriptionwithin the enrolment

period). Patients who died or moved away during the follow-up

period (12 months following the enrolment date) were excluded

from the study. A history of two or more prescriptions for lipid-

lowering drugs (ATC code C10), platelet inhibitors (ATC code

B01) and heart disease drugs (ATC code C01) during the

12 months preceding the enrolment date, and a history of two or

more prescriptions for antidiabetics (ATC code A10) and of one

or more hospitalizations for cardio-cerebrovascular diseases

(ICD9 codes: 401–405, arterial hypertension; 410, acute myo-

cardial infarction; 411–414, coronary disease; 428, heart failure;

430–438, cerebral circulatory dysfunction; 440–442, arterio-

sclerosis of the main arteries and aneurysm; 535, chronic renal

insufficiency) between 1 January 2002 and the enrolment date,

were used to characterize included patients.

Drug Utilization Analysis

The follow-up period for each patient was 12 months start-

ing from the enrolment date. The duration of therapy was

168 Costa et al.

ª 2009 Adis Data Information BV. All rights reserved. High Blood Press Cardiovasc Prev 2009; 16 (4)

calculated following the method of Catalan and LeLorier.[19]

The continuous use of antihypertensive treatment is expressed

in days from the date of the enrolment prescription to the

time of first failure to continue renewals of antihypertensive

drugs with the permissible time period (gap) between the

prescribed end of the given dispensation and the date of the

next dispensation. The permissible gap between the end date

of a given antihypertensive dispensation and the date of

renewal was defined as a time period equal to one-half the

duration of the given antihypertensive dispensation, or 7 days,

whichever was longer. In the presence of continuous therapy

from the enrolment date onwards, the compliant renewers

have the number of days of continuous use equal to 365 days

(duration of follow-up period). Based on this, patients were

defined as ‘persistent’ and ‘not-persistent’ users. There can be

three different types of persistent users: ‘same therapy users’

(patients maintaining the same class of drug), ‘add-on therapy

users’ (patients adding one or more drugs to initial therapy)

and ‘switchers’ (patients changing enrolment class of drug).

Non-persistent users were classified as: ‘occasional users’

(patients with a single drug refill), ‘stoppers’ (patients stopping

treatment early) and ‘intermittent users’ (patients having

an intermittent therapy with intervals of variable duration be-

tween prescriptions).

Statistical Analysis

Continuous variables are presented as mean values– standard

deviations. The statistical significance between means was

calculatedby one-wayANOVA.The associationbetween catego-

rical variables was tested using the Pearson’s w2 test. A multi-

variable logistic regression model was developed to evaluate the

drug class effects on non-persistence to antihypertensive treat-

ment or to initial antihypertensive therapy (the dependent vari-

able) adjusting for potential external influences. The following

independent variables were considered: class of drug on enrol-

ment, patient age, sex, presence/absence of antidiabetic drugs,

presence/absence of lipid-lowering drugs, presence/absence of

platelet inhibitors, presence/absence of heart disease drugs, pre-sence/absence of previous CV hospitalizations and presence/absence of two or more co-morbidities (included CV hospitali-

zations). To measure the strength of the linear relationship

between trends in utilization of ARBs and persistence to

treatment from 2003 to 2006 Pearson’s r correlation coefficient

was performed. For all analyses, p-values of <0.05 were

considered significant. We performed all statistical analyses

with SPSS for Windows version 15.0 (SPSS Inc., Chicago, IL,

USA, 2008).Tab

leI.

Featu

res

ofpatients

by

enro

lmentantihypert

ensiv

ecla

ss

Diu

r.B

BC

CB

AC

EI

AR

BO

ther

AH

TB

B+

diu

r.A

CE

I+diu

r.

AR

B+

diu

r.A

llcom

b.

Tota

l

Patients

[n(%

)]8538

8793

7012

14

010

5712

1438

737

5352

3139

6762

61

493

(13.9

)(1

4.3

)(1

1.4

)(2

2.8

)(9

.3)

(2.3

)(1

.2)

(8.7

)(5

.1)

(11.0

)(1

00.0

)

Mean

age

*(y

ears

–S

D)

61.8

–17.3

54.9

–14.4

57.4

–16.7

59.4

–15.2

59.6

–14.2

54.8

–15.7

58.4

–15.1

58.3

–16.2

59.5

–14.8

60.6

–15.9

58.8

–15.8

Sex

(%m

ale

)32.7

41.4

47.9

50.6

49.7

50.8

35.4

44.7

46.4

53.0

45.8

Antidia

betic

thera

py*

(%)

5.4

3.3

4.1

8.4

9.3

4.7

4.1

4.5

4.0

5.8

5.9

Hypolip

idaem

icth

era

py

*(%

)4.0

6.7

4.2

5.2

6.1

3.8

3.7

4.0

4.2

2.6

4.7

Antipla

tele

tth

era

py

*(%

)10.2

7.1

5.5

6.9

5.9

5.3

2.7

4.6

4.5

4.9

6.5

Card

iac

thera

py

*(%

)5.7

3.8

2.4

2.4

2.1

1.4

0.7

1.4

1.6

2.4

2.9

Pre

vio

us

CV

hospitaliz

ations*

(%)

7.7

11.0

7.8

9.7

7.2

6.1

4.2

5.5

5.9

21.0

9.7

Patients

with

‡2co-m

orb

ilities*

(%)

7.4

6.8

4.6

6.3

5.7

4.5

1.6

3.3

3.3

6.5

5.8

AC

EI=

AC

Ein

hib

itors

;A

HT

=antihypert

ensiv

edru

gs;

AR

B=

angio

tensin

IIty

pe

1re

cepto

ranta

gonis

ts(a

ngio

tensin

recepto

rblo

ckers

);B

B=b-

adre

nocepto

ranta

gonis

ts(b

-blo

ckers

);

CC

B=

calc

ium

channelb

lockers

;co

mb

.=com

bin

ations;C

V=

card

iovascula

r;d

iur.

=diu

retics;

*p

<0.0

01.

Persistence to Antihypertensive Therapy 169


Results

Population Features and Trends in Drugs Prescription

A total of 61 493 newly treated hypertensive patients were

enrolled in the study. As shown in table I, their mean age was

58.8 – 15.8 years, 45.8% were males and 54.2% females.

The most frequently prescribed drugs were ACE inhibitors,

followed by b-blockers, diuretics, calcium antagonists and

ARBs. An extemporaneous combination of two or more drugs

was prescribed to 11.0% of patients, whereas the most com-

monly prescribed fixed-dose combination was ACE inhibitor +hydrochlorothiazide. Patients receiving b-blockers were youngerandmore frequently had previous hospitalizations for CV events,

while diuretics and extemporaneous combinations were more

frequently prescribed to older subjects. On the other hand, ACE

inhibitors and ARBs were more often prescribed to patients re-

ceiving an antidiabetic treatment.

Table II shows the trends of prescriptions with time.

Prescription of diuretics from 2003 to 2006 had a 0.8% re-

duction, b-blockers a 2.9% increase, while the most evident

decrease was that of calcium antagonists (-2.7%). ACE in-

hibitors were slightly reduced (-0.2%), while ARBs prescrip-

tion rose by 1.0%. Also, when used in fixed combination with

low-dose hydrochlorothiazide, ARBs showed a 2.1% increase

in prescriptions, while fixed combinations of ACE inhibitors

had a 2.2% decrease and b-blockers fixed combinations re-

mained almost unchanged.

Persistence to Treatment

Figure 1 shows distribution of persistent and non-persistent

patients 1 year after the beginning of therapy.

Persistent patients were 11.2% (6.7% same therapy users, 1.3%add-on therapy users and 3.2% switchers). Not-persistent pa-

tients were 88.8% of the total. Of these, 35.3% were occasional

Table II. Trends of prescription of the different classes of drugs by year of enrolment

Enrolment year Diur.

(%)

BB

(%)

CCB

(%)

ACE I

(%)

ARB

(%)

Other AHT

(%)

BB + diur.

(%)

ACE I + diur.

(%)

ARB + diur.

(%)

All comb.

(%)

2003 14.4 13.3 12.4 22.4 8.7 2.9 1.4 9.9 3.9 10.7

2004 14.1 13.5 11.7 22.6 9.6 2.5 1.1 9.4 5.2 10.2

2005 14.6 16.1 10.8 22.8 9.1 1.9 1.2 6.8 6.0 10.6

2006 13.6 16.2 9.7 22.2 9.7 1.7 1.2 7.7 6.0 11.9

ACE I = ACE inhibitors; AHT = antihypertensive drugs; ARB = angiotensin II type 1 receptor antagonists (angiotensin receptor blockers); BB = b-adrenoceptor

antagonists (b-blockers); CCB = calcium channel blockers; comb. = combinations; diur. = diuretics.

Newly treated pts with antihypertensive drugsfrom 2003 to 2006 in the local health unit of Pavia

61 493 (100.0%)

Persistent users(pts with continuous drug refills

during the 1-year follow-up period)6882 (11.2%)

Non-persistent users(pts without continuous drug refillsduring the 1-year follow-up period)

54 611 (88.8%)

Same therapy users(pts maintaining initial therapy)

4109 (6.7%)1 (59.7%)2

Switchers(pts switching initial therapy)

1945 (3.2%)1 (28.3%)2

Add-on therapy users(pts adding drugs to initial therapy)

828 (1.3%)1 (12.0%)2

Occasional users(pts with single drug refills)21 732 (35.3%)1 (39.8%)2

Stoppers(pts interrupting therapy early)

12 680 (20.6%)1 (23.2%)2

Intermittent users(pts with intermittent drug refills)

20 199 (32.8%)1 (37.0%)2

Fig. 1. Distribution of persistent and non-persistent patients (pts). 1 indicates percentages calculated on overall enrolled population; 2 indicates percentages

calculated on persistent/non-persistent users.

170 Costa et al.


users, 20.6% were stoppers, while 32.8% were intermittent users.

Persistent patients were older, more frequently males and had

higher CV risk as testified by concomitant drug treatments and

previous hospitalizations for CV events (table III).

Figures 2a and b show persistence level with the different

drug treatments.

The lowest level of persistence was seen with diuretics

(figure 2a).After 1 year, only 3.0%of patients starting therapywith

a diuretic were still on treatment (1.7% still taking the same ther-

apy). The highest level of persistence was seen in patients starting

with ARBs (13.3% maintaining the same therapy and 18.8% still

on treatment). When considering together monotherapy and

fixed-dose combinations (figure 2b), persistencewithARBs rose to

34.0%, ACE inhibitors to 18.5% and b-blockers to 17.9%.

Predictors of Non-Persistence

Figure 3 shows patient-related predictors of non-persistence.

Older patients had a significantly reduced risk of non-persis-

tence (-1.5% for a 1-year increase in age). Females have a 37%higher risk of non-persistence, while patients treated with anti-

diabetic and hypolipidaemic drugs are significantly more

persistent and likewise people having had a previous hospita-

lization for a CV event (+74%).

Table IV shows drug-related predictors of non-persistence to

antihypertensive treatment. Taking ARBs as reference drugs,

the risk of non-persistence was 75% greater with b-blockers,87% greater with calcium antagonists, 85% greater with ACE

inhibitors and 641% greater with diuretics. Starting with a fixed

combination increased risk of non-persistence by 191% with

ACE inhibitors, by 188% with b-blockers and only by 26% with

ARBs. Table V shows drug-related predictors of risk of non-

persistence to initial antihypertensive therapy.

Figure 4 shows the relative risk of non-persistence to initial

therapy with different ARBs in comparison with the progenitor

losartan. In figure 4a, monotherapies and fixed-dose combi-

nations are considered separately, whereas in figure 4b they are

considered as a whole.

Table III. Features of persistent and non-persistent patients

Persistent patients Non-persistent patients

Patients [n (%)] 6882 (11.2) 54 611 (88.8)

Mean age* (years – SD) 61.9 – 12.6 58.4 – 16.1

Sex* (% male) 54.5 44.7

Antidiabetic therapy* (%) 8.2 5.6

Hypolipidaemic therapy* (%) 6.0 4.6

Antiplatelet therapy** (%) 7.3 6.4

Cardiac therapy*** (%) 3.3 2.8

Previous cardiovascular hospitalizations* (%) 17.2 8.7

Patients with ‡2 co-morbidities* (%) 8.5 5.4

* p < 0.001, ** p < 0.01, *** p < 0.05.

02468

101214161820

New

ly tr

eate

d pa

tient

s w

ithan

tihyp

erte

nsiv

e dr

ugs

(%)

New

ly tr

eate

d pa

tient

s w

ithan

tihyp

erte

nsiv

e dr

ugs

(%)

0

5

10

15

20

25

30

35

b

a

ARB + diu

r.

ACE I +

diur.

BB + diu

r.

Other

AHT

ARBACE I

CCBBBDiur

.

Other

AHT

ARBACE I

CCBBBDiur

.

SameAdd-onSwitch

Fig. 2. Persistence after 1 year with monotherapy or fixed combinations with

different drugs considered (a) separately or (b) considered as a single class.

ACE I = ACE inhibitors; AHT = antihypertensive drugs; ARB = angiotensin II

type 1 receptor antagonists (angiotensin receptor blockers); BB = b-adrenoceptor

antagonists (b-blockers); CCB = calcium channel blockers; diur.= diuretics.



Considering separately monotherapy and fixed-combination

therapy (figure 4a), a reduced risk of non-persistence was seen

with candesartan, irbesartan, valsartan and telmisartan given in

monotherapy, and with valsartan and telmisartan given in

fixed-dose combination. Considering monotherapy and fixed-

dose combination as a whole (figure 4b), the only two ARBs

with a non-persistence risk lower than losartan were valsartan

and telmisartan.

Trends in Persistence

The proportion of persistent patients showed a continuous

annual increase from 2003 to 2006 (8.4%, 9.8%, 12.1% and

14.6%, respectively). This trend was found correlated to the

increase in ARBs utilization (r = 0.85) [figure 5].

Economic Cost of Non-Persistence

The total cost of drugs in the population we examined was

h8572 049.23 of which h2431 435.08 (28.4%) were spent by

persistent patients and h6140 614.15 (71.6%) by non-persistent

patients (h509 334.80 by occasional patients, h2602 920.47 by

stoppers and h3028 358.88 by intermittent patients). The per-

capita cost of persistent patients was h353.30/year, whereas thatof non-persistent patients was h112.44/year. This calculation

takes into account the sole cost of drug therapy and excludes

the expected reduction of total economic and social costs due to

reduced events linked to a better blood pressure control in

persistent patients.

Discussion

Our data confirm the results of previous studies showing that

antihypertensive treatment is frequently interrupted just a few

months after its prescription.[18] In our study, after 1 year, only

11.2% of newly treated hypertensive patients were still on regular

drug therapy. It is noteworthy that 35.3% of patients stopped

treatment after the first prescription. This observation is similar

to that of the study by Corrao et al.[17] where 38.4% of patients

stopped treatment after the first prescription. However, in that

study, patients stopping treatment after the first prescriptionwere

excluded from final evaluation and this, along with our more

stringent method for classifying persistence,[19] justifies the lower

levels of persistence found in our study.

In our population, a better persistence was found to be as-

sociated with increasing age and increasing level of CV risk. In

fact, more persistent subjects were those having had a previous

CV event and those receiving, along with antihypertensive

drugs, antidiabetic, hypolipidaemic and antiplatelet drugs,

suggesting associated risk factors or previous events. In other

words, better persistence seems to be associated to a better

awareness of the risk linked to the disease.

Our results also confirm that treatment discontinuation rate

differs between antihypertensive drug classes,[12-18,20-22] the best

persistence being exhibited by patients in whom the drug initially

prescribedwas anARBand theworst by adiuretic. It is remarkable

0.8980.985*

1.181**1.371*

1.224*

1.741*

00.2000.4000.6000.8001.0001.2001.4001.6001.8002.000

RR

No pr

eviou

s CV h

ospit

aliza

tions

No an

tiplat

elet t

hera

py

No an

tidys

lipida

emic

ther

apy

No an

tidiab

etic

ther

apy

Female

Age1

Fig. 3. Patient-related predictors of risk of non-persistence to anti-

hypertensive treatment. CV = cardiovascular; RR = relative risk. 1 indicates

for 1 year age increase; * p < 0.001, ** p < 0.01.

Table IV. Drug-related predictors of risk of non-persistence to antihyper-

tensive treatment

Initial class therapy p-Value Relative risk of

discontinuationa

95% CI

Diuretics <0.001 7.412 6.433, 8.539

ACE I + diuretics <0.001 2.908 2.568, 3.293

BB + diuretics <0.001 2.875 2.144, 3.855

Other AHT <0.001 2.829 2.286, 3.501

CCB <0.001 1.867 1.687, 2.066

ACE I <0.001 1.853 1.701, 2.017

BB <0.001 1.753 1.593, 1.929

ARB + diuretics <0.001 1.256 1.115, 1.415

All combinations =0.001 1.163 1.060, 1.276

ARB 1.000

a Relative risks refer to the discontinuation of AHT treatment. Relative risks

were adjusted for: patient age, gender, presence of antidiabetic drugs,

lipid-lowering drugs, platelet inhibitors, heart disease drugs, previous

cardiovascular hospitalizations, two or more co-morbidities.



antagonists (b-blockers); CCB = calcium channel blockers.

172 Costa et al.


that a fixed-dose combination of ARBs and a low-dose thiazide

diuretic, is associated with the second best persistence level.

A variety of factors influence persistence with anti-

hypertensivemedication, but one of themost significantmay be

adverse effects and the placebo-like tolerability of ARBs jus-

tifies their higher levels of persistence.

In fact, starting with drugs different from ARBs always

implies an excess risk of non-persistence. The relative risk of

treatment discontinuation compared with ARBs was +75%with b-blockers, +85% withACE inhibitors, +87% with calcium

channel blockers and a striking +641% with diuretics.

On the other hand, the relative risk of non-persistence was

only slightly increased by starting treatment with a fixed com-

bination of ARBs and a diuretic (+26%), but was significantly

greater with fixed combinations of b-blockers (+188%) and

ACE inhibitors (+191%).

These observations have important clinical implications

because treatment discontinuation is the major determinant[23]

of the poor blood pressure control observed in the hypertensive

populationworldwide,[5,6,24,25] and poor blood pressure control

is responsible for a high incidence of CV events. A recent paper

showed that stroke incidence is 22% lower in hypertensive pa-

tients having a good adherence to treatment.[25] Another study

demonstrated that non-persistence with antihypertensive

drug therapy was associated with a 15% increase in the risk

of acute myocardial infarction and a 28% increase in the

risk of stroke.[26] A retrospective cohort study[27] of hypertensive

patients found that the 1-year risk of hospitalization was

significantly lower for patients who had a 80–100% adherence

with their antihypertensive regimen compared with those who

were not. Hospitalization, beside clinical considerations, is par-

ticularly relevant for cost analysis. In that study, in fact, although

drug costs were higher for patients who were 80–100% adherent,

overall medical costs were lower thanks to a lower hospitalization

rate.[27] In the US, the American Heart Association estimated that

in 2005 about 71% of hypertensive Americans had suboptimal

blood pressure control, which resulted in 39702 CV events and

8374deaths for a totalmedical expenditureof $US964million.[28] In

Europe, it has been estimated that if hypertension management

achieved blood pressure targets, the European healthcare system

could save up to h1.26billion.[29] A recent study calculated that in

2001 nonoptimal blood pressure control had a $US0.37billion

global cost. This represents about 10% of the world’s overall

healthcare expenditure. Over a 10-year period, elevated blood

pressure may cost nearly $US1billion globally in health spending,

if current blood pressure levels persist then indirect costs could be

as high as $US3.6billion annually.[30]

Our study showed that 71.6% of expenses for drugs were

consumed by non-persistent patients in whom drugs had re-

duced or had partial or no positive effect at all. This is an

incredible waste of economic resources.

Our study, however, also showed that persistence levels had

a constant improvement passing from 8.4% in 2003 to 14.6% in

2006. This trend was found correlated to the increase of ARB

users.

A similar phenomenon was recently reported in a study

showing that between 2003 and 2006 there was a blood pressure

control improvement in England and that it was associated

with changes in prescription habits, with a reduction of

prescriptions of diuretics and b-blockers in favour of ACE

inhibitors and ARBs.[31]

This should be seriously taken into account when calculating

the economic impact of treatment with the more expensive (but

also much less frequently abandoned) ARBs on NHS financial

resources.

A recent review suggested that ARBs, thanks to the high

level of adherence and persistence, represent a cost-saving and

cost-effective treatment compared with other conventional

treatment options for patients with hypertension and asso-

ciated conditions.[32]

Another interesting point is that among ARBs some mole-

cules seem to be associatedwith a better persistence profile.When

Table V. Drug-related predictors of risk of non-persistence to initial antihyper-

tensive therapy

Initial class therapy p-Value Relative risk of

discontinuation

initial therapya

95% CI

Diuretics <0.001 9.666 8.158, 11.452

Other AHT <0.001 3.502 2.729, 4.494

All combinations <0.001 3.221 2.848, 3.643

ACE I + diuretics <0.001 3.030 2.647, 3.468

BB + diuretics <0.001 2.903 2.115, 3.986

CCB <0.001 2.144 1.918, 2.398

ACE I <0.001 2.102 1.916, 2.305

BB <0.001 1.797 1.622, 1.991

ARB + diuretics <0.001 1.254 1.105, 1.422

ARB 1.000

a Relative risks refer to the discontinuation of AHT treatment or to the switch

of initial therapy. Relative risks were adjusted for: patient age, gender, pre-

sence of antidiabetic drugs, lipid-lowering drugs, platelet inhibitors, heart

disease drugs, previous cardiovascular hospitalizations, two or more

co-morbidities.



antagonists (b-blockers); CCB = calcium channel blockers.



compared with the progenitor losartan, better persistence was

seen with monotherapy with candesartan, irbesartan, valsartan

and telmisartan, whereas when considering monotherapy and

fixed-dose combination therapy together, only valsartan and

telmisartan performed better than losartan.

Differences in persistence to different drugs are probably

determined by a complex and composite interaction between

many clinical and non-clinical factors.

Besides the combination efficacy-tolerability, which prob-

ably accounts for the highest levels of persistence observed

with ARBs, many other variables can play a role. Physicians’

attitude, patient’s characteristics and clinical conditions, ad-

verse effects, costs, but also marketing influences and NHS

pressures to reduce prescriptions of higher price drugs, can

explain the wide variability between classes of drugs and also

within the ARBs class.

Our study also has some limitations. First, in our database,

there is no information on clinical diagnosis, and patients

may thus have taken antihypertensive drugs for other clinical

conditions, such as heart failure and coronary artery disease.

An Italian study,[33] however, has shown that hypertension is by

far the most common diagnosis for use of antihypertensive

drugs (73%).

0.8330.904 0.911

1.000

0.714**0.596*

0

0.200

0.400

0.600

0.800

1.000

1.200

1.400

1.600

b

a

CANDE EPRO IRBE TELMI VALSA LOSA

RR

0.841

0.708

0.872

1.000

0.697***0.625**

0.384*

0.559*

0.420*

0.649**0.604*

0

0.200

0.400

0.600

0.800

1.000

1.200

1.400

CANDE CANDE +diur.

EPRO IRBE IRBE +diur.

VALSA VALSA +diur.

TELMI TELMI +diur.

LOSA LOSA +diur.

RR

Fig. 4. Relative risk (RR) of non-persistence to initial therapy with different angiotensin II type 1 receptor antagonists (angiotensin receptor blockers) compared with

losartan. In (a), monotherapies and fixed-dose combination are considered separately, whereas in (b), they are considered as a whole. CANDE = candesartan;

diur. = diuretic; EPRO = eprosartan; IRBE = irbesartan; LOSA = losartan; TELMI = telmisartan; VALSA = valsartan; * p £ 0.001, ** p < 0.01, *** p < 0.05.

0

2

4

6

8

10

12

14

16

18

2003 2004 2005 2006

Enrolment year

New

ly tr

eate

d pa

tient

s w

ith a

ntih

yper

tens

ive

drug

s (%

)

r = 0.85

Persistent patientsARB users

Fig. 5. Changes with time of persistent patients and angiotensin II type 1

receptor antagonist (angiotensin receptor blocker [ARB]) users.

174 Costa et al.


The second problem with the present study is that the data

source does not include clinical data (such as blood pressure

levels, drug-related adverse effects) that lead doctors to choose

or to switch an antihypertensive drug.

Conclusions

This study confirms that the vast majority of newly treated

hypertensive patients do not continue treatment in a regular

and constant manner, diuretics being the drugs conditioning

the lowest persistence and ARBs the highest. Since persistence

is the key to a satisfactory blood pressure control, economic

analysis should take into account the highest persistence levels

observed with ARBs and these drugs should be regarded as a

tool to decrease, not to increase, global health expenses. Thus,

the choice of first antihypertensive drug is crucial since persis-

tence to treatment varies widely between and within drug

classes and has a deep impact on blood pressure control and

thus on CV complications of hypertension, which are the main

factors responsible for the disease cost.

Acknowledgements

This study was funded by Boehringer Ingelheim Italia.

For all authors, there is no conflict of interest and there are no persons

who contributed to the work who do not meet the criteria for authorship.

References1. Psaty BM, Lumley T, Furberg CD, et al. Health outcomes associated with

various antihypertensive therapies used as first-line agents: a network meta-

analysis. JAMA 2003; 289: 2534-44

2. Blood Pressure Lowering Treatment Trialists’ Collaboration. Effects of

different blood pressure lowering regimens on major cardiovascular

events: second cycle of prospectively designed overviews. Lancet 2003; 362:

1527-35

3. Chalmers J. Comparison of various blood pressure lowering treatments on the

primary prevention of cardiovascular outcomes in recent randomised clinical

trials. Clin Exp Hypertens 2004; 26: 709-19

4. Turnbull F, Neal B, Ninomiya T, et al. Effects of different regimens to lower

blood pressure on major cardiovascular events in older and younger adults:

meta-analysis of randomised trials. Blood Pressure Lowering Treatment

Trialists’ Collaboration. BMJ 2008; 336 (7653): 1121-3

5. Wolf-Maier K, Cooper RS, Banegas JR, et al. Hypertension prevalence and

blood pressure levels in 6 European Countries, Canada, and the United

States. JAMA 2003; 289 (18): 2363-9

6. KotsevaK,WoodD,DeBackerD, et al., for the EUROASPIREStudyGroup.

Cardiovascular prevention guidelines in daily practice: a comparison of

EUROASPIRE I, II, and III surveys in eight European countries. Lancet

2009; 373: 929-40

7. Eni C, Okonofua EC, Simpson KN, et al. Therapeutic inertia is an impediment

to achieving the Healthy People 2010. Blood Pressure Control Goals Hyper-

tens 2006; 47: 345-51

8. Fujita T, Shimamoto K, Wu Z, et al. What are the major challenges in getting

patients to the optimal BP goal? Difficulties in educating doctors and patients.

Int J Clin Pract Suppl 2006; 150: 20-2

9. Costa FV. Compliance with antihypertensive treatment. Clin Exp Hypertens

1996 Apr-May; 18 (3-4): 463-72

10. Burnier M. Medication adherence and persistence as the cornerstone of

effective antihypertensive therapy. Am J Hypertens 2006; 19: 1190-6

11. Lopatriello S, Berto P, Cramer J, et al. Different aspects of adherence to

antihypertensive treatments [abstract]. Expert Rev Pharmacoeconomics

Outcomes Res 2004; 4 (3): 317-33

12. Hyre AD, Krousel-Wood MA, Muntner P, et al. Prevalence and predictors of

poor antihypertensive medication adherence in an urban health clinic setting.

J Clin Hypertens (Greenwich) 2007 Mar; 9 (3): 179-86

13. Degli Esposti L,Degli Esposti E, Valpiani G, et al. A retrospective, population-

based analysis of persistence with antihypertensive drug therapy in primary

care practice in Italy. Clin Ther 2002 Aug; 24 (8): 1347-57

14. Degli Esposti E, Sturani A, Di Martino M, et al. Long-term persistence

with antihypertensive drugs in new patients. J Hum Hypertens 2002; 16:

439-44

15. Mazzaglia G, Mantovani L, Sturkenboom MC, et al. Patterns of persistence

with antihypertensive medications in newly diagnosed hypertensive patients

in Italy: a retrospective cohort study in primary care. J Hypertens 2005; 23:

2093-100

16. Bourgault C, Senecal M, Brisson M, et al. Persistence and discontinua-

tion patterns of antihypertensive therapy among newly treated patients: a

population-based study. J Hum Hypertens 2005 Aug; 19 (8): 607-13

17. Corrao G, Zambon A, Parodi A, et al. Discontinuation of and changes in drug

therapy for hypertension among newly-treated patients: a population-based

study in Italy. J Hypertens 2008 Apr; 26 (4): 819-24

18. Costa FV, D’Ausilio A, Bianchi C, et al. Adherence to anti-hypertensive

medications: a review and update. High Blood Pressure Cardiovasc Prev 2009

Sep; 16 (3): 101-10

19. Catalan VS, LeLorier J. Predictors of long-term persistence on statins

in a subsidized clinical population. Value Health 2000 Nov-Dec; 3 (6):

417-26

20. Elliott WJ, Plauschinat CA, Skrepnek GH, et al. Persistence, adherence,

and risk of discontinuation associated with commonly prescribed

antihypertensive drug monotherapies. J Am Board Fam Med 2007; 20:

72-80

21. Hasford J, Mimran A, Simons WR. A population-based European cohort

study of persistence in newly diagnosed hypertensive patients. J Hum

Hypertens 2002; 16: 569-75

22. Erkens JA, PannemanMM,Klungel OH, et al. Differences in antihypertensive

drug persistence associated with drug class and gender: a PHARMO study.

Pharmacoepidemiol Drug Saf 2005 Nov; 14 (11): 795-803

23. Haynes RB, McDonald HP, Garg AX. Helping patients follow prescribed

treatment: clinical applications. JAMA 2002; 288: 2880-3

24. Wolf-Maier K, Cooper RS, Kramer H, et al. Hypertension treatment

and control in five European countries, Canada, and the United States.

Hypertension 2004; 43: 10-7

25. Kettani FZ, Dragomir A, Cote R et al. Impact of a better adherence to anti-

hypertensive agents on cerebrovascular disease for primary prevention.

Stroke 2009; 40: 213-20

26. Breekveldt-Postma NS, Penning-van Beest FJA, Siiskonen SJ, et al. The effect

of discontinuation of antihypertensives on the risk of acute myocardial in-

farction and stroke. Curr Med Res Opin 2008; 24: 121-7

27. Sokol MC, McGuigan KA, Verbrugge RR, et al. Impact of medication

adherence on hospitalization risk and healthcare cost. Med Care 2005; 43:

521-30



28. American Heart Association. High blood pressure 2009 [online]. Available

from URL: http://www.americanheart.org/presenter.jhtml?identifier=2139[Accessed 2009 Sep 15]

29. Hansson L, Lloyd A, Anderson P, et al. Excess morbidity and cost of failure to

achieve targets for blood pressure control in Europe. Blood Press 2002; 11 (1):

35-45

30. Gaziano TA, Bitton A, Anand S, et al. The global cost of non optimal blood

pressure. J Hypertens 2009; 27: 1472-7

31. Falaschetti E, Chaudhury M, Mindell J, et al. Continued improvement in hy-

pertension management in England: results from the Health Survey

for England 2006. Hypertension 2009; 53: 480-6

32. Theodoratou D, Maniadakis N, Fragoulakis V, et al. Analysis of published

economic evaluations of angiotensin receptor blockers. Hellenic J Cardiol

2009; 50: 105-18

33. Poluzzi E, Strahinja P, Vargiu A, et al. Initial treatment of hypertension and

adherence to therapy in general practice in Italy. Eur J Clin Pharmacol 2005;

61: 603-9

Correspondence: Dr Luca Degli Esposti, CliCon S.r.l., Via Salara, 36, 48100,

Ravenna, Italy.

E-mail: [email protected]

176 Costa et al.


trends in prescription and determinants of persistence to antihypertensive therapy

Documents