A population-based analysis of statin utilization in British Columbia

A Population-Based Analysis of Statin Utilization in British Columbia C o l e t t e B. R a y m o n d , P h a r m D , M S c l ; S t e v e n G. M o r g a n , PhD2; A l a n Katz, M B C h B , M S c , CCFp3; a n d A n i t a L. Kozyrskyj, P h D 4

7University of Manitoba, Faculty of Pharmacy, Winnipeg, Manitoba, Canada; 2University of British Columbia, Centre for Healtk Services and Policy Research, Vancouver, Britisk Columbia, Canada; 3University of Manitoba, Faculty of Medicine, Winnipeg, Manitoba, Canada; and 4Universityof Manitoba, Faculties of Medicine and Pkarmacy, Winnipeg, Manitoba, Canada ABSTRACT Objective: The purpose of this research was to measure prevalent and incident statin use in the population of British Columbia from 1996 to 2004 across specific patient characteristics (sociodemographic and clinical). Methods: Statin utilization and demographic data were assessed with the use of prescription drug claims. Medical and hospital claims for statin users were examined for evidence of ischemic heart disease (IHD), diabetes mellitus (DM), atherosclerosis, cerebrovascular disease, peripheral vascular disease (PVD), and disorders of lipid metabolism during the 3 years before the first statin prescription. Results: Statin prevalence increased from 1996 to 2004 (1.28 %-6.59%). The greatest use was among those aged 65 to 84 years. Prevalent use of atorvastatin, simvastatin, and rosuvastatin increased over time. There were 211,964 new statin users between 1999 and 2004. Quarterly incident statin use increased over time from 1999 to 2004 (0.29%-0.49%). A socioeconomic gradient, whereby use was greater in those with low socioeconomic status, was observed for incident statin use. Incident atorvastatin use increased over time; simvastatin, cerivastatin, and rosuvastatin peaked and then declined; and new use of other statins decreased. Among 211,964 incident statin users, 74,542 (35.17%) had evidence of IHD only; 43,257 (20.41%) had DM but no IHD; 9781 (4.61%) had no DM or IHD but had atherosclerosis, cerebrovascular disease, or PVD; 47,634 (22.47%) had disorders of lipid metabolism only; and 36,750 (17.34%) had none of the medical conditions evaluated. Conclusions: Prevalent use (1996-2004) and incident use (1999-2004) of statins in an entire population have increased dramatically. Although many statin users (60.19%) had evidence of medical conditions

that indicate appropriate statin use, 39.91% of users were at low risk for cardiovascular disease, and therefore the benefit of statins in this group remains small. (Clin Ther. 2007;29:2107-2119) Copyright © 2007 Excerpta Medica, Inc. Key words: drug utilization, claims data, pharmacoepidemiology, cholesterol treatment.

INTRODUCTION The cholesterol-lowering medications known as the 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (ie, statins) maintain the largest share of pharmaceutical sales worldwide. In 2005, total sales of cholesterol-lowering drugs exceeded $32 billion, an increase of 7% over the previous year. 1 In Canada between 1998 and 2002, per capita spending on cholesterol-lowering agents doubled.2 For patients with established vascular disease or diabetes mellitus (DM), large randomized controlled trials (RCTs) have demonstrated that statins prevent recurrence of cardiovascular events and decrease mortality) However, for patients who do not have established vascular disease or DM, the data describing the benefit of statins remain inconsistent. 4,s Thus, the routine use of statins for primary prevention in patients who have not had a major cardiac event remains controversial, s-7

Data from this study were presented in abstract form at the International Society for Pharmacoepidemiology Conference, August 24-27, 2006, Lisbon, Portugal.

Accepted for publicationJune 20, 2007. doi:l 0.10 ~6/j.clinthera.2007.09.022 0149-2918/$32.00 Printed in the USA. Reproduction in whole or part is not permitted. Copyright © 2007 Excerpta Medica, Inc.

Despite the known benefits of statins, a treatmentrisk paradox seems to have emerged. Statins appear to be underused in populations that could benefit greatly from them, particularly those with cardiovascular disease or the elderly (aged >65 years). 8-3s At the same time, statins may be overused in patients with low cardiovascular risk (no previous cardiovascular event, DM, ischemic heart disease [IHD], atherosclerosis, cerebrovascular disease, or peripheral vascular disease [PVD]). 13,31,34,35 Such utilization dynamics would have obvious implications for public health and for the cost-effectiveness of both private and public investment in this form of pharmaceutical care. It is unclear which factors are barriers to potentially appropriate statin use among patients with multiple cardiovascular risks and which factors facilitate potentially excessive use among those with low cardiovascular risk. Most studies that describe statin use in the population are limited by a lack of relevant medical information about patients to identify the indications for statin use. 13,18,23,24,27-33,36-39 F e w studies have evaluated the medical history of statin users in an attempt to quantify statin use by indication 12,34,4° or have specifically investigated the incident use of statins among populations. 9,29,3°,4°,41 The objective of this study was to describe both prevalent and incident statin use in the entire population of British Columbia from 1996 to 2004 across specific sociodemographic and clinical characteristics with the use of an administrative health database.

MATERIALS AND METHODS Study Design This study used a population-based, person-specific data set that contains prescription dispensation records and demographic information for -4.1 million unique residents living in British Columbia from 1996 to 2004 (calendar years). Data describing prescription drug utilization were extracted from the British Columbia PharmaNet drug information system. Provincial law requires that information from every prescription dispensed at retail pharmacies to residents living in the community or in long-term care facilities must be entered into this database. Information extracted for this study included dispensation date, anonymous patient identifier, drug quantity, and drug identifier. Individuals were considered to be statin users if they filled prescriptions for any of the following

statins: atorvastatin, cerivastatin, lovastatin, pravastatin, simvastatin, or rosuvastatin. Anonymized patient identifiers were linked to population registries to obtain sociodemographic information and relevant population denominators. The study cohort comprised every resident eligible for the provincially administered, universal public health insurance (British Columbia Medical Services Plan). This excludes First Nations, veterans, and Royal Canadian Mounted Police (total excluded: - 4 % of the population). Sociodemographic information included age, sex, and a measure of socioeconomic status. Each individual was assigned a measure of socioeconomic status equal to the median census-reported income for all households within the individual's neighborhood. Neighborhoods were defined by >7000 Census Dissemination Areas in the province, each including 400 to 700 persons. The population was divided into deciles ranging from those living in the lowest-income neighborhoods to those living in the highest-income neighborhoods. For each incident statin user, the Medical Services Plan physician-claims database and the hospitaldischarge abstract database were searched for each of the 3 years before study cohort entry to locate any medical conditions that would indicate a reason for the statin prescription according to the International Classification of Diseases, Ninth Revision (ICD-9) codes. 42 Hospital-discharge abstracts for hospital services include information on up to 25 ICD-9 diagnostic codes and up to 20 Canadian Classification of Procedures (CCP) codes. 43 Medical conditions were grouped according to the Expanded Diagnostic Clusters of the Johns Hopkins Adjusted Clinical Group case mix system. 44 Using the physician-claims and hospitaldischarge abstracts databases, we searched the medical histories for the following: IHD (ICD-9 410-414; CCP 480-483), DM (ICD-9 250), cerebrovascular disease (ICD-9 430-438), PVD (ICD-9 440.2, 440.3; CCP 501-503, 512), atherosclerosis (ICD-9 429.2, 440.0, 440.8, 440.9), and disorders of lipid metabolism (ICD-9 272). To expand the definitions of DM and IHD, the PharmaNet records were searched for any prescription for a blood glucose-lowering agent

(Anatomical Therapeutic Chemical Classification [ATC] A10) 45 or for a nitrate (ATC C01D) 45 during the 1 year before the first statin prescription. Incident statin users were divided into 5 mutually exclusive

medical-condition categories based on physician and prescription drug claims: (1) IHD; (2) DM with no IHD; (3) risk equivalent (PVD, cerebrovascular disease, or atherosclerosis) but no IHD or DM; (4) disorders of lipid metabolism but no IHD, DM, or riskequivalent medical conditions; and (5) no medical or prescription claims for any of the medical conditions sought (none of 1-4). Incident statin users were further grouped according to medical conditions indicative of high cardiovascular risk (IHD, DM, PVD, cerebrovascular disease, or atherosclerosis) or low cardiovascular risk (disorders of lipid metabolism or no claims for any medical conditions sought).

Analysis For 1996 through 2004, we calculated the quarterly incidence of statin use per 100 inhabitants, stratified by age, sex, socioeconomic status, drug prescribed, and medical history. Incident statin users were individuals who resided within the province for at least 275 days of each of 3 years before filling their first-ever prescription for a statin. The 3-year residency requirement ensured that incident users were residents of British Columbia who had not previously filled a statin prescription for at least 3 years. Denominators for the incidence calculation included all individuals who satisfied the same residency requirement (at least 275 days of each of 3 years before the year of interest). Prevalence was measured as the number of individuals who filled at least 1 prescription for a statin in the given quarter. Individuals filling prescriptions for >1 specific statin within a given quarter were counted as 1 user of each particular statin. Denominators for prevalence calculations included all individuals in the study cohort during the period of interest, stratified by age and socioeconomic status as necessary. Population denominators were determined for each year of the study. Linear regression was performed to determine whether the series of values for quarterly prevalence and incidence demonstrated a linear trend over the study period. For statin prevalence, regression equations were developed for each age group, socioeconomic stratum, and drug. For statin incidence, regression equations were developed for each age group, socioeconomic stratum, sex, drug, dose, and medical-condition category. Regression lines were considered statistically significant (P < 0.05) when the slope of the entire line was significantly different from 0. All analyses were

conducted in SAS version 9.0 (SAS Institute Inc., Cary, North Carolina). The University of British Columbia Behavioural Research Ethics Board approved this study.46 RES U LTS Prevalence Between 1996 and 2004, the number of adults using statins in British Columbia increased from 1.28% to 6.59% (P < 0.05). Statin use in all age groups increased over time (P < 0.05), although there was no clear age gradient in statin use (Figure 1). Prevalent statin use was highest and increased the most among those aged 65 to <85 years (4.25%- 20.92%), followed by those aged 45 to <65 years (1.82%7.45%). Prevalent statin use in the elderly population (>85 years) increased from 0.45% to 8.11%, nearly a 20-fold increase. The age group with the lowest prevalence of statin use at all time points was aged 20 to <45 years (0.10%-0.56%). Prevalent statin use increased across all socioeconomic strata over time (P < 0.05), although there was no clear socioeconomic gradient (Figure 2). The prevalence of statin use in the group with the lowest socioeconomic status (SES 1) increased from 1.48% to 7.20% over the study period, whereas the prevalence of statin use in the group with the highest socioeconomic status (SES 10) increased from 1.46% to 6.69%. Analysis of prevalent statin use by drug type revealed a changing practice pattern (Figure 3). Prevalent use of atorvastatin increased from 0.0008% at the beginning of 1997, when the product was launched, to 3.05% by the end of 2004 (P < 0.05). The prevalent use of simvastatin also increased throughout the study period (0.27%-1.15%; P < 0.05). The use of pravastatin increased slightly (0.21%-0.36%), whereas the use of fluvastatin decreased from 0.11% to 0.06% and the use of lovastatin decreased from 0.36% to 0.17% (P < 0.05 for fluvastatin and lovastatin). The quick uptake of new agents into routine use is exemplified by cerivastatin and rosuvastatin. Cerivastatin use increased from 0.002% in the first quarter of 1998, when it was launched, and quickly reached 0.33% in the second quarter of 2001 (P < 0.05), followed by a rapid decline at the end of 2001 after market withdrawal because of reports of hepatotoxicity. Similarly, after the launch of rosuvastatin at the beginning of 2003, its prevalent use increased from 0.01%

-

Age 2 0 - < 4 5 y

-

- - Age 4 5 - < 6 5 y . . . . . . Age 6 5 - < 8 5 y .......... Age ->85 -

25-

-

Overall

•

/ i"

20-

/ I.F-~.

k/l If=

o .~

,/"

I*

15-

i "~.

~-- 0 0 l~. Ca 0

/

/

10>o_

5/ .........................

..-o

I

I

I

I

I

I

I

I

1996

1997

1998

1999

2000

2001

2002

2003

2004

QI

QI

QI

QI

QI

QI

QI

QI

QI

Figure 1. Prevalence o f s t a t i n use in British C o l u m b i a , 1 9 9 6 to 2 0 0 4 , by age group. Q = quarter.

- -

SES 1 ( l o w e s t )

- - - - SES 2 . . . . . . SES4 .......... SES 6 . . . . . . SES 8 - - --~ SES 10 (highest) 7-

...... /~

-~g

/'.-.,,...'

s

.c: .~ ~ - 3O .

4-J ~q 0

4

Q-

o

3

Q.

I

I

I

I

I

I

I

I

I

1996

1997

1998

1999

2000

2001

2002

2003

2004

Q1

Q1

Q1

Q1

Q1

Q1

Q1

Q1

Q1

Figure 2. Prevalence o f s t a t i n use in British C o l u m b i a , 1 9 9 6 to 2 0 0 4 , by socioeconomic status (SES). Q = quarter.

- -

Atorvastatin

-- --

Cerivastatin

......

Fluvastadn

.......... Lovastatin --

---- Pravastatin

-----

Rosuvastatin

......

Simvastatin

3.53.0-

2.50 .c_ .,~ n:l -~0

2.0-

0 O_

1.51.0(3-

0.5:~ ......

0

I 1996 QI

"""22~'=-~ .............

~- .............................................

. ~ :..~._. . . . . . . . . . . . . . .

I 1997

I 1998

I 1999

I 2000

I 2001

I 2002

I 2003

I 2004

Q1

Q1

Q1

Q1

Q1

Q1

Q1

Q1

Figure 3. Prevalence ofstatin use in British Columbia, 1996 to 2004, by drug. Q = quarter.

to 0.46% by the end of the study period (P < 0.05), indicating rapid adoption of rosuvastatin.

Incidence Between 1999 and 2004, 211,964 adults who met our inclusion criteria filled a first statin prescription. The mean age of incident statin users was 61.2 years, and the cohort was 116,990/211,964 (55.19%) men. The largest group of incident statin users was aged 45 to <65 years (105,565/211,964 [49.80%]), followed by 65 to <85 years (84,369/211,964 [39.80%]), 20 to <45 years (18,224/211,964 [8.60%]), and finally age ___85years (3,806/211,964 [1.80%]). The quarterly incidence of overall statin initiation among British Columbia adults increased from 0.29% at the beginning of 1999 to 0.49% by the end of 2004 (P < 0.05; Figure 4). Quarterly statin initiation in those aged 45 to <65 years increased from 0.40% to 0.64% (P < 0.05). Quarterly statin initiation in the age group 65 to <85 years increased from 0.66% to 0.88% between 1999 and 2004. Incident statin initiation in the elderly population (aged ___85 years) increased from 0.09% to 0.35% over the study period,

an increase of nearly 4 times (P < 0.05). The age group with the lowest incident statin initiation at all time points was 20 to <45 years (0.05%-0.12%) (P < 0.05). Men had a higher incidence of statin initiation than women at all time points (0.29%-0.47% for men and 0.20%-0.33% for women [data not shown; P < 0.05]). When stratified by socioeconomic status, all groups demonstrated an increase in incident statin initiation over time (P < 0.05); however, a clear socioeconomic gradient was observed (Figure 5). Those with SES 1 demonstrated greater incident statin utilization than those with the highest SES 10. The quarterly incidence in SES 1 increased from 0.31% to 0.53% (P < 0.05); the quarterly incidence in SES 10 increased from 0.27% to 0.46% (P < 0.05). When evaluated by particular drug, the most commonly initiated statin was atorvastatin (128,538/ 211,964 [60.64%]), followed by simvastatin (40,027/ 211,964 [18.88%]), pravastatin (13,091/211,964 [6.18%]), cerivastatin (12,312/211,964 [5.81%]), rosuvastatin (11,422/211,964 [5.39%]), lovastatin (3500/ 211,964 [1.65%]), and fluvastatin (3074/211,964

Age 20-<45 y Age 45-<65 y . . . . . . Age 65-<85 y .......... Age ->85 . . . . . . . Overall -

-

1.00.9--

/

0.8C

0.7-i 13_ L~ 0 C tm 0

i

\

-

,/

,

i

,\ i

~..I

I

\,

"v"

,k

/

'k,

/

~ .....

\'1" I,

\d"

I

0.6-

/

/~v----\

/ ~

0.5-

/

g

2

/

\

'

-

,/ 'k, ,

/i I \ .%~../I • " "

\,

-

-

\./

.......... ~ /

0.4-

..... -~

I/"

,,,,

\/

\-

\ v. /

/r"----~""-~..~

''-

/

\\..~. / /

I-

>.

J -'t"-"-~'~/

0.3-

.........."

%,. """--...--""

................. y -

.--" ..................

........................................... ., .......... ..°,'"""

0.1 0

,.-"

. . , . ............................

0.2-

U

y..

i

I

I

I

I

I

1999

2000

2001

2002

2003

2004

QI

Q1

Q1

Q1

Q1

Q1

Figure 4. Incident statin use in British Columbia, 1999 to 2004, by age group. Q = quarter.

SES I ( l o w e s t ) --

--

......

SES 2 SES 4

.......... SES 6 , --

SES 8 --

SES 1 0 ( h i g h e s t )

0.55 0.50 t-

A

~ g o.45 ~.m =

\

/,'

0.40

ga_ 'U o ° IZ

0.35

'r'-

n

o_ 0.30

O'

/,

0.25 0.20

I

I

I

I

I

I

1999

2000

2001

2002

2003

2004

Q1

Q1

Q1

Q1

Q1

Q1

Figure 5. Incidentstatin usein British Columbia, 1999 to 2004, by socioeconomic status(SES). Q = qua~er.

[1.45%]). Atorvastatin demonstrated the greatest increase in statin initiation over time (0.14%-0.32%; P < 0.05) (Figure 6). The quarterly incidence of simvastatin prescribing increased from 0.05% to 0.14% at the end of 2002, but then declined through the end of the study period to 0.07% (P < 0.05). The quarterly incidence of cerivastatin use increased from 0.03% at the beginning of 1999, peaked at 0.08% in mid-2001 (P < 0.05), but then declined to 0 after being withdrawn from the market. The quarterly incidence of rosuvastatin use increased from 0.006% to 0.11% in mid-2004 (P < 0.05), indicating rapid adoption of rosuvastatin. The distribution of medical conditions in the 211,964 incident statin users over the entire study period is described in the Table. Among incident statin users, 74,542/211,964 (35.17%) had evidence of IHD only; 43,257/211,964 (20.41%) had DM but no IHD; 9781/211,964 (4.61%) had no DM or IHD, but had atherosclerosis, cerebrovascular disease, or PVD; 47,634/211,964 (22.47%) had disorders of lipid metabolism only; and 36,750/211,964 (17.34%) had none of the medical conditions evaluated. Incident statin users with medical conditions indicative of high

cardiovascular risk (IHD, DM, PVD, cerebrovascular disease, or atherosclerosis) in medical and prescription drug records accounted for 127,580/211,964 (60.19%) of all incident statin users over the study period. The greatest use occurred among those with a recorded diagnosis of IHD (Figure 7). For this group, the quarterly incidence increased from 0.12% to 0.15%. The quarterly incidence for those with DM but no IHD showed the greatest change over the study period; the incidence in this group increased from 0.04% to 0.12% (P < 0.05). Over the study period, the incident use in those with PVD, cerebrovascular disease, or atherosclerosis (risk equivalent) increased from 0.01% to 0.02% (P < 0.05). For patients with disorders of lipid metabolism only, quarterly incidence increased from 0.07% to 0.12% over the study period (P < 0.05). For patients with no medical or prescription claims for any of the medical conditions sought, the quarterly statin incidence increased from 0.05% to 0.08%. When the 3 medical-condition groups that indicate high cardiovascular risk (IHD, DM, or risk-equivalent conditions in medical and prescription drug records) were combined, quarterly incidence

Atorvastatin Cerivastatin Fluvastatin .......... Lovastatin Pravastatin - - - - - - Rosuvastatin Simvastatin -

-

-

-

.

.

-

.

-

.

-

-

-

-

.

.

-

. . . . . . . . . . .

0.35

-

/

0.30 c

'-~

0.25 -

"6 -5

0.20-

o "~ o

0.15-

'= ~ "E

O.IO-

r-

,r-'-

/

J

./

,.

~

f

J .. ................

~'.~

rg

C7

o.o5 O

..... _.217~:~......... 2 : : : 2 - Z £ : . £ . . : 2 £ Z T Z ~ I I

-

r

..........

. . . .

7

.....

".

~'"'~

...

.

/

.

.

.

.

.

.

~

.......

...............

T

1999

2000

2001

2002

2003

2004

Q1

Q1

Q1

Q1

Q1

Q1

Figure 6. Incidentstatin usein British Columbia, 1999 to 2004, by drug. Q = qua~er.

%'%"

__-

Table. Medical conditions and treatments in incident statin users (N = 211,964) in British Columbia, 1999 to 2004. Incident Statin Users with Condition, No. (%)

Medical Condition Ischemic heart disease

74,542 (35.17)

Diabetes mellitus

63,644 (30.02)

Peripheral vascular disease

7427 (3.50)

Cerebrovascular disease

15,608 (7.36) 8285 (3.91)

Atherosclerosis Disorders of lipid metabolism

88,286 (41.65)

Diabetes drugs*

41,424 (19.54)

Nitroglycerin*

32,551 (15.36)

Risk equivalent (peripheral vascular disease, cerebrovascular disease, or atherosclerosis)

26,616 (12.56)

*Any prescription during the year before the incident statin prescription.

increased from 0.17% to 0.29% (P < 0.05) (data not shown). In the groups with low cardiovascular risk (disorders of lipid metabolism only or no medical or prescription claims for any medical conditions sought), the incidence of statin prescribing increased from 0.12% to 0.19% per quarter (P < 0.05) (data not shown). A greater proportion of younger incident statin users had low cardiovascular risk compared with older incident statin users. The proportion of incident statin users aged 20 to 44 years with no IHD, DM, PVD, cerebrovascular disease, or atherosclerosis was 10,525/18,224 (57.75 %). This proportion decreased to 48,358/105,565 (45.81%), 24,878/84,369 (29.49%), and 623/3806 (16.36%) in the age groups 45 to 64, 65 to 84, and >_85 years, respectively. Incident statin users aged >_85 years with lower cardiovascular risk (disorders of lipid metabolism only or no medical or prescription claims for any medical conditions) comprised 623/211,964 (0.29%) of all incident statin users. Incident statin users aged 65 to 84 years with lower cardiovascular risk, however, comprised 24,878/ 211,964 (11.74 %) of all incident statin users.

-

--

IHD

-

--

......

DM no IHD RE

.......... H y p e r l i p i d i m i a , n o I H D , D M , o r RE ........ N o c l a i m s For m e d i c a l c o n d i t i o n s s o u g h t

0.18 0.16 0.14

t-r~

O

~q .~

0.12-

O

o_ 0.10~ oi ~ .

..

e..

g

_

tO

v--

--

~

..,'~'"~........'"".,

,,- ..............

.

,,..'"'"".

°....~ ,,

Z °.'~."........ -....

,,,,"'/"

0.080.06/

0.04(:7

/

0.02 0 1999 Q1

I

I

I

I

I

2000 Q1

2001 Q1

2002 Q1

2003 Q1

2004 Q1

Figure 7. Incident statin use in British Columbia, 1999 to 2004, by medical-condition category. Q = quarter; IHD = ischemic heart disease; DM = diabetes mellitus; RE = risk equivalent.

A greater proportion of female incident users than male users had low cardiovascular risk: 40,567/94,415 (42.97%) of female and 43,660/116,990 (37.31%) of male incident statin users had no IHD, DM, PVD, cerebrovascular disease, or atherosclerosis in medical and prescription drug records. When further analyzed by the total study population, women and men with low cardiovascular risk represented 40,567/211,964 (19.14%) and 43,660/211,964 (20.60%) of the total group of incident statin users, respectively. DISCUSSION This comprehensive population-based analysis of prevalent and incident statin utilization in an entire province demonstrated a 5-fold increase in prevalent use from 1996 to 2004 and a 1.7-fold increase in incident use from 1999 to 2004. Statin use was greater in men than in women and in older age groups. A socioeconomic gradient, with greater use in lower socioeconomic strata, was observed for incident statin use. However, statin use in the population remains suboptimal. Although the majority (60.19%) of incident statin users were at high cardiovascular risk (as defined by medical and prescription drug records), 39.81% of statin users had no medical conditions indicating high cardiovascular risk. The findings of increased incident and prevalent statin use are consistent with other studies that have evaluated statin utilization o v e r t im e . 18,23,27,2s,3°,31,33,37~° Like other studies of statin utilization across several age g r o u p s , 13,20,30,32,34,37 this study observed increased statin use with age, up to age 65 to 84 years, and then relatively low use in those aged _>85 years. Because limited RCT data exist for statin efficacy in patients aged >85 years, 47,48 no large statin trials have included patients aged >82 years, 4,49 and no statin trial has shown a mortality benefit of statin therapy in patients aged >70 years, 4 it is likely appropriate that incident statin use in the oldest age group was low. We observed greater incident statin use in men as compared with women, consistent with other studies that have evaluated statin utilization by s e x . 13'20'30'32'34'37 Because cardiovascular disease occurs more frequently in men than women, and because the benefits of statin therapy have been demonstrated in RCTs more clearly in men than in women, s° it is likely appropriate that there was greater overall use of statins in men. No obvious socioeconomic gradient in prevalent statin users was observed in this study, consistent with

other studies that have shown no clear socioeconomic gradient in statin use for general or specific (eg, after myocardial infarction) populations. 11,15,21,33,3s However, this study did observe a socioeconomic gradient in incident statin use, whereby those with higher socioeconomic status had lower use. This finding is likely to be appropriate given the higher burden of cardiovascular risk factors in those with low socioeconomic status, sl,s2 Two other Canadian studies have evaluated the impact of socioeconomic status on statin utilization and reported conflicting results. 21,3s The variable nature of provincial drug formularies in Canada s3 makes it difficult to determine whether a clear socioeconomic gradient in statin prescribing exists in Canada. Perhaps the socioeconomic differences in incident and prevalent statin use can be explained by differences in persistence with therapy over long periods (>1 year) or the relative out-of-pocket expenses for long-term therapies such as statins. As in other studies, this analysis observed a general increase in prescriptions for atorvastatin, rosuvastatin, and simvastatin over time, with a corresponding decrease in the use of other statins. 22-24,27,3°,39,s4,ss Prevalent use and incident use were most influenced by the introduction of atorvastatin, which gradually took over statin market share since its release in 1997. Several large RCTs have described the ability of atorvastatin to prevent major cardiac events, s,s6 Atorvastatin gained market share before published RCT evidence appeared describing the impact of atorvastatin on mortality, only first published in 2003. s7 Similarly, despite an absence of data demonstrating the potential benefit of rosuvastatin on mortality or even cardiovascular morbidity at the time of this study, s8 this drug enjoyed rapid uptake among incident statin users, further suggesting a utilization trend that was not based on RCT evidence. The majority of incident statin users (60.19%) had medical conditions indicative of high cardiovascular risk, including IHD, DM, PVD, cerebrovascular disease, and atherosclerosis; therefore, this represents appropriate and evidence-based use of statins. Although the proportion of incident statin users with high cardiovascular risk did not change significantly over the study period, there were still 39.81% of statin users without high cardiovascular risk (based on medical and prescription drug records). The analysis of medical history for incident statin users could not be exhaustive because the analysis was based upon ad-

ministrative data and not a verbal history, physical examination of a patient, or a chart review. Although the analysis searched for a 3-year history of cardiovascular disease or risk, the reliance on ICD coding may have underestimated cardiovascular risk. Despite evaluating medical and prescription claims for many ICD codes, and claims for medications, it is possible that the study underestimated the proportion of incident statin users with high cardiovascular risk. It is possible that a proportion of these patients with seemingly low cardiovascular risk had cardiovascular risk factors (eg, obesity, smoking, family history, or hypertension) that were not well recorded in the medical or hospital administrative databases but that would put them at high cardiovascular risk. It is also possible that there were statin users with high cholesterol that we did not identify because of the relatively short evaluation period of 3 years (given the nature of administrative data); these individuals would therefore be classified as having no medical or prescription claims for any medical conditions sought. Yet, despite this form of misclassification, there is likely still a considerable proportion of statin users with single or minimal cardiovascular risk factors who are being treated with statins. For these individuals, RCT evidence to support the use of statins is a contentious issue, particularly from a pharmacoeconomic perspective. Moreover, despite a paucity of RCT evidence to demonstrate that statins prevent major cardiac events in women or the elderly (>65 years) with low cardiovascular risk, these populations accounted for a large proportion of the incident statin users in the study cohort. For example, 19% of incident statin users over the study period were women with no medical conditions indicative of high cardiovascular risk in medical and prescription drug records. It is understood that the incidence of major cardiac events in women is lower and lags -10 years behind that of men. s9 However, no RCT has demonstrated any beneficial impact of statins on major cardiac events for women or elderly patients taking statins for primary prevention. Major statin trials for primary prevention included <20% women, and primary outcomes from these trials were not significant for women, s7,6°,61 Furthermore, a meta-analysis of primary-prevention trials of statins in women has revealed no statistically significant effect of statins for any cardiovascular end points, s° Similarly, the Prospective Study of Pravastatin in the Elderly at Risk 4 showed no benefit of statin therapy among patients

aged 70 to 82 years in the primary-prevention cohort (n = 3239) and no mortality benefit of statins for primary- or secondary-prevention patients (those who had already experienced a cardiac event). Despite lack of RCT evidence of efficacy, use of statins in these populations in British Columbia was considerable. The results of this study can be placed in context through comparison with other studies that evaluated the medical history of statin users. 12,34,40 Savoie and Kazanjian 34 observed that only a minority (14%) of prevalent statin users in British Columbia had high cardiovascular risk (as indicated by physician claims) in 1995. Similarly, Dubois et a112 observed that only 33% of statin users in a large managed care organization had coronary artery or atherosclerotic disease between 1997 and 1999, and another 14% had DM. Paterson et al4° observed that among incident statin users aged >65 years in British Columbia and Ontario, 8%, 25%, and 14% had acute coronary syndromes, chronic IHD, and DM, respectively, from 1998 to 2001. The present study observed that 35.16% of incident statin users from 1999 to 2004 had medical history claims indicative of acute or chronic IHD up to 3 years before their incident prescription, and that another 20.41% had DM but no IHD. Differences in methods and years of analysis may account for the differences observed between these studies of British Columbia statin users. It is important to consider the findings of this study in terms of appropriate, evidence-based statin use. That is, what proportion of the population of British Columbia should be taking statins? Some research suggests that statins are underprescribed to certain segments of the population, especially those with a clear indication for these drugs. 34,3s Estimates of cardiovascular risk in Canada suggest that -13% of Canadians have a high cardiovascular risk (history of cardiovascular disease or >15% chance of developing cardiovascular disease within 10 years). 62,63 The resuits of the present study indicated that by the end of 2004, 6.6% of British Columbian adults were taking statins (prevalent users), suggesting underuse of statins in populations with high cardiovascular risk. The results of this study also revealed that patients with low cardiovascular risk, or minimal RCT evidence to support the use of statins, such as those aged ___85 years without previous cardiovascular disease, may be overtreated with statins.

An important factor to consider when evaluating the possible overuse of statins in British Columbia is the Canadian Lipid Guidelines. 3 The 2003 version of these guidelines has been criticized for recommending statin therapy for patients who are at low cardiovascular risk, a population for which no RCT data demonstrate a benefit of statin therapy. 6,63 In the context of population-based effectiveness and cost-effectiveness, some have suggested that statins be reserved for individuals with high cardiovascular risk to gain maximal benefit from these drugs. 6,62,63 CONCLUSIONS This population-based analysis of statin utilization revealed that both prevalent use and incident use in the British Columbia population have increased over time. The greatest use occurred in the age group 65 to 84 years, and in men. Although the majority of new statin prescriptions are for people with high cardiovascular risk, 39.81% of new statin prescriptions may be for those with low cardiovascular risk. For the latter population, RCT evidence available at the time did not show a significant benefit to statin therapy, although such treatment may be consistent with current Canadian Lipid Guidelines. ACKNOWLEDGMENTS This research was made possible by funding from a Western Regional Training Centre Award (funded by the Canadian Health Services Research Foundation, Alberta Heritage Foundation for Medical Research, and Canadian Institutes of Health Research) and the Canadian Institutes of Health Research. The British Columbia Ministry of Health approved access and use of the provincial data provided by the Centre for Health Services and Policy Research in conformance with provincial legislation and policy relating to privacy protection. REFERENCES 1. IMS. Leading therapy classes by global pharmaceutical sales, 2005. Available at: http://www.imshealth.com/ims/ portal/front/articleC/0,2777,6599 77478579 7747968 3,00.html. Accessed August 200@ 2. Morgan S. Drug spending in Canada: Recent trends and causes. Med Care. 2004;42:635-642. 3. McPherson R, Frohlich J, Fodor G, Genest J, for the Canadian Cardiovascular Society. Canadian Cardiovascular Society position statement-recommendations for the diagnosis and treatment ofdyslipidemia and prevention of car-

4.

5.

6.

7.

8.

9.

I 0.

11.

12.

diovascular disease [published correction appears in CanJ Cardiol. 2006;22:1077]. CanJ Cardiol. 2006;22:913-927. Shepherd J, Blauw GJ, Murphy MB, et al, for the PROSPER Study Group (PROspective Study of Pravastatin in the Elderly at Risk). Pravastatin in elderly individuals at risk of vascular disease (PROSPER): A randomised controlled trial. Lancet. 2002;360:1623-1630. Cheung BM, Lauder IJ, Lau CP, Kumana CR. Meta-analysis of large randomized controlled trials to evaluate the impact of statins on cardiovascular outcomes. Br J Olin Pharmacol. 2004;57:640-651, Manuel DG, Tanuseputro P, Mustard CA, et al. The 2003 Canadian recommendations for dyslipidemia management: Revisions are needed [published correction appears in CMAJ. 2005;173:133]. CMAJ. 2005;172:1027-1031. GenestJ, McPherson R, FrohlichJ, Fodor G. The analysis by Manuel and colleagues creates controversy with headlines, not data. CMAJ. 2005;172:1033-1034, discussion 1037. Bennett KE, Williams D, FeelyJ. Under-prescribing of cardiovascular therapies for diabetes in primary care. EurJ Clin Pharmacol. 2003;58:835-841. Bjerrum L, Larsen J, KragstrupJ. Guidelines accompanied by changes in reimbursement rules. Effects on lipid-lowering drug prescribing. ScandJ Prim Health Care. 2001 ;I 9:158162. Degli Esposti L, Di Martino M, Saragoni S, et al. Pharmacoutilization ofstatin therapy after acute myocardial infarction. A real practice analysis based on administrative data. Ital HeartJ. 2004;5:120-I 26. DeWilde S, Carey IM, Bremner SA, et al. Evolution of statin prescribing 1994-2001: A case ofagism but not of sexism? Heart. 2003;89:417-421. Dubois RW, Alexander CM, Wade S, et al. Growth in use of lipid-lowering therapies: Are we targeting the right patients? AmJ Manag Care. 2002;8:862-867.

13. FeelyJ, McGettigan P, KellyA. Growth in use ofstatins after trials is not targeted to most appropriate patients. Clin Pharmacol Ther. 2000;67:438-441. 14. FoodyJM,Ferdinand FD, Galusha D, et al. Patterns of secondary prevention in older patients undergoing coronary artery bypass grafting during hospitalization for acute myocardial infarction. Circulation. 2003;108(Suppl I): 1124-1128. 15. Hartz I, Eggen AE, Grimsgaard S, et al. Whom are we treating with lipid-lowering drugs? Are we following the guidelines? Evidence from a population-based study: The Tromso Study 2001. EurJ Clin Pharmacol. 2004;60:643-649. 16. Kanstrup H, Lassen JF, Heickendorff L, et al. Quality of lipid-lowering therapy in patients with ischaemic heart disease: A register-based study in 3477 patients. J Intern Med. 2004;255:367-372.

17. Kopjar B, Sales AE, Pineros SL, et al. Comparison of characteristics of patients with coronary heart disease receiving lipid-lowering therapy versus those not receiving such therapy. AmJ Cardiol. 2003;91 :I 352-I 354. 18. Kucera Z, Vlcek J, Hejdova M. Theoretical exposure of chronically treated patients to lipid lowering agents. Pharmacoepidemiol Drug Sa~ 2005;14:61-67. 19. Lewis BE, McDonough K. Dyslipidemia treatment among patients with coronary artery disease in a managed care organization. Am J Health S_yst Pharm. 2004;61:10321038. 20. Majeed A, Moser K, Maxwell R. Age, sex and practice variations in the use of statins in general practice in England and Wales. J Public Health

Med. 2000;22:275-279. 21. Pilote L, Beck CA, Karp I, et al, for the Canadian Cardiovascular Outcomes Research Team. Secondary prevention after acute myocardial infarction in four Canadian provinces, 1997-2000. Can J Cardiol. 2004;20:61-67. 22. Rasmussen iN, Gislason GH, Abildstrom SZ, et al. Statin use after acute myocardial infarction: A nationwide study in Denmark [published correction appears in BrJ C/in Pharmacol. 2005;60:343]. Br J C/in Pharmacot. 2005;60:150-t 58. 23. Siegel D, Lopez J, Meier J. Use of cholesterol-lowering medications in the United States from 1991 to 1997. AmJ Med. 2000;108:496-499. 24. Teeling M, Bennett K, FeelyJ. The influence of guidelines on the use of statins: Analysis of prescribing trends 1998-2002. Br J Clin Pharmacol. 2005;59:227-232. 25. Williams D, Bennett K, Feely J. Evidence for an age and gender bias in the secondary prevention of ischaemic heart disease in primary care. Br J Clin Pharmacol. 2003;55: 604-608. 26. Yang CC, Jick SS, Testa MA. Who receives lipid-lowering drugs: The el-

fects of comorbidities and patient characteristics on treatment initiation. Br J Clin Pharmacol. 2003;55: 288-298. 27. Cooke C, Nissen L, Sketris I, Tett SE. Quantifying the use of the statin antilipemic drugs: Comparisons and contrasts between Nova Scotia, Canada, and Queensland, Australia. C/in Ther. 2005;27:497-508. 28. Jackevicius CA, Tu K, Filate WA, et al, for the Canadian Cardiovascular Outcomes Research Team. Trends in cardiovascular drug utilization and drug expenditures in Canada between 1996 and 2001. Can J CardioL 2003;19:1359-1366. 29. Larsen J, Andersen M, Kragstrup J, Gram LF. Changes in the utilisation of lipid-lowering drugs over a 6-year period (1993-1998)in a Danish population. Eur J Clin Pharmacol. 2001 ; 57:343-348. 30. LevyAR, O'Brien BJ, McMullen E, et al. Rapid increase in statins newly dispensed to Ontario seniors between 1994 and 2000. CanJ Cardiol. 2003;19:665-669. 31. Magrini N, Einarson T, Vaccheri A, et al. Use of lipid-lowering drugs from 1990 to 1994: An international comparison among Australia, Finland, Italy (Emilia Romagna Region), Norway and Sweden. EurJ Clin Pharmacol. 1997;53:185-189. 32. Martikainen J, Klaukka T, Reunanen A, et al. Recent trends in the consumption of lipid-lowering drugs in Finland. J Clin Epidemiol. 1996;49: 1453-1457. 33. Packham C, Pearson J, Robinson J, Gray D. Use of statins in general practices, 1996-8: Cross sectional study. BMJ. 2000;320:1583-1584. 34. Savoie I, Kazanjian A. Utilization of lipid-lowering drugs in men and women. A reflection of the research evidence? J Clin Epidemiol. 2002;55: 95-101. 35. Ko DT, Mamdani M, Alter DA. Lipid-lowering therapy with statins in high-risk elderly patients: The treatment-risk paradox.JAMA. 2004; 291:1864-1870.

36. WalleyT, Folino-Gallo P, Schwabe U, van Ganse E, for the EuroMedStat Group. Variations and increase in use of statins across Europe: Data from administrative databases [published correction appears in BMJ. 2004;329:323]. BMJ. 2004;328:385386. 37. Metge C, Kozyrskyj A, Dahl M, et al. Drug prescribing in Manitoba: How appropriate is it? Winnipeg, Manitoba, Canada: Manitoba Centre for Health Policy; 2003. 38. Silwer L, Lundborg CS. Patterns of drug use during a 15 year period: Data from a Swedish county, 19882002. Pharmacoepidemiol Drug Sa~ 2005;14:813-820. 39. Walley 7-, Folino-Gallo P, Stephens P, Van Ganse E. Trends in prescribing and utilization of statins and other lipid lowering drugs across Europe 1997-2003. Br J C/in Pharmacol. 2005;60:543-551. 40. Paterson JM, Carney G, Anderson GM, et al, for the BC-Ontario Pharmacosurveillance for Decisionmaking Collaborative. Case selection for statins was similar in two Canadian provinces: BC and Ontario.

J Clin Epidemiol. 2007;60:73-78. 41. Riahi S, Fonager K, Tort E, et al. Use of lipid-lowering drugs during 1991-98 in Northern Jutland, Denmark. BrJ Clin Pharmacol. 2001;52: 307-311. 42. World Health International Classification of Diseases, 9th Revision, 2003. Available at: ftp://ftp.cdc. gov/pub/Health_Statistics/NCHS/ Publications/ICDg/1998. Accessed August 2, 2007. 43. Canadian Classification of Diagnostic, Therapeutic and Surgical Procedures. Ottawa, ON: Statistics Canada; 1986. Catalog no. 82-562-XPE. 44. ACG Software Documentation and Users Manual. Baltimore, Md: Johns Hopkins University; 2001. 45. Norwegian Institute of Public Health. Guidelines for ATC Classification and DDD Assignment. Oslo, Norway: WHO Collaborating Centre for Drug Statistics Methodology; 2003.

46. Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, Social Sciences and Humanities Research Council of Canada. TriCouncil Policy Statement." Ethical Conduct for Research Involving Humans. 1998 (with 2000, 2002 and 2005 amendments). Available at: http://pre.ethics. gc.ca/ engl ish / policystatement/ policy statement.cfm. Accessed August 13, 2007. 47. Foody JM, Rathore SS, Galusha D, et al. HydroxymethylglutaryI-CoA reductase inhibitors in older persons with acute myocardial infarction: Evidence for an age-statin interaction. J Am Geriatr Soc. 2006; 54:421-430. 48. Andrawes WF, Bussy C, Belmin J. Prevention of cardiovascular events in elderly people. DrugsAging. 2005; 22:859-876. 49. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: A randomised placebocontrolled trial. Lancet. 2002;360: 7-22. 50. Walsh JM, Pignone M. Drug treatment of" hyperlipidemia in women. JAMA. 2004;291:2243-2252. 51. Alter DA, Iron K, Austin PC, Naylor CD, for the SESAMI Study Group. Influence of education and income on atherogenic risk factor profiles among patients hospitalized with acute myocardial infarction. Can J Cardiol. 2004;20:1219-I 228. 52. Marmot MG, Kogevinas M, Elston MA. Social/economic status and disease. Annu Rev Public Health. 1987; 8:111-135. 53. Anis AH, Guh D, Wang X. A dog's breakfast: Prescription drug coverage varies widely across Canada. Med Care. 2001 ;39:315-326. 54. Bazalo GR. Managed care trends in statin usage. Manag Care. 2001;10: 48-50, 53-59. 55. Mamdani MM, TuJV. Did the major clinical trials of statins affect pre-

56.

57.

58.

59.

scribing behaviour? CMAJ. 2001 ;I 64: 1695-1696. Baigent C, Keech A, Kearney PM, et al, for the Cholesterol Treatment Trialists' (CTF) Collaborators. Efficacy and safety of cholesterol-lowering treatment: Prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins [published correction appears in Lancet. 2005;366:1358]. Lancet. 2005;366: 1267-I 278. Sever PS, Dahlof B, Poulter NR, et al, for the ASCOT Investigators. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower-than-average cholesterol concentrations, in the AngloScandinavian Cardiac Outcomes Trial-Lipid Lowering Arm (ASCOTLLA): A multicentre randomised controlled trial. Lancet. 2003;361: 1149-1158. Rosuvastatin: New preparation. Opt for statins with evidence of efficacy on clinical outcome. Prescrire Int. 2004;13:132-134. Lerner DJ, Kannel WB. Patterns of coronary heart disease morbidity

60.

61.

62.

63.

and mortality in the sexes:A 26-year follow-up of the Framingham population. Am Heart J. 1986;111:383390. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute coronary events with ]ovastatin in men and women with averagecholesterol levels: Results of AFCAPS/TexCAPS. Air Force/Texas Coronary Atherosclerosis Prevention Study. JAMA. 1998;279:1615-I 622. Shepherd J, Cobbe SM, Ford I, et al, for the West of Scotland Coronary Prevention Study Group. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. N EnglJ Med. 1995;333: 1301-1307. Manuel DG, Lim J, Tanuseputro P, et al. Revisiting Rose: Strategies for reducing coronary heart disease. BMJ. 2006;332:659-662. Manuel DG, Kwong K, Tanuseputro P, et al. Effectivenessand efficiency of different guidelines on statin treatment for preventing deaths From coronary heart disease: Modelling study. BMJ. 2006;332:1419.

Address correspondence to: Colette B. Raymond, PharmD, MSc, Manitoba Renal Program/Faculty of Pharmacy, University of Manitoba Department of Pharmaceutical Services Health Sciences Centre, Hospital MS189-820, Sherbrook Street, Winnipeg MB R3A 1R9 Canada. E-maih [email protected]