Comparison of health care costs and co-morbidities between men diagnosed with benign prostatic hyperplasia and cardiovascular disease (CVD) and men with CVD alone in a US commercial

ORIGINAL ARTICLE

Comparison of health care costs and co-morbidities between men diagnosed with benign prostatic hyperplasia and cardiovascular disease (CVD) and men with CVD alone in a US commercial populationManan Shah a, Melissa butler a, thomas bramley a, tammy G. Curtice b and Shari Fine c

a Xcenda (formerly Applied Health Outcomes), Salt Lake City, UT, USAb Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, CT, USAc Christ Hospital, Jersey City, NJ, USA

Address for correspondence: Thomas Bramley, RPh, PhD, Director, Xcenda (formerly Applied Health Outcomes), 1528 Preston Street, Salt Lake City, UT 84108, USA. Tel.: +1 801 983 6784; Fax: +1 801 983 7337; email: [email protected]

Key words: Benign prostatic hyperplasia – Cardiovascular disease – Costs

0300-7995

doi:10.1185/030079906X167345

All rights reserved: reproduction in whole or part not permitted

CurrEnt MEDiCAl rESEArCH AnD OPiniOn®

VoL. 23, No. 2, 2007, 417–426

© 2007 librAPHArM liMitED

Paper 3599 417

Objective: The purpose of this study was to compare costs and treatment patterns between men with concomitant benign prostatic hyperplasia (BPH) and CVD to men with CVD (but not BPH).

Study design: A retrospective, matched cohort study was utilized to assess costs and treatment between two study populations.

Methods: The data source was administrative claims from managed care organizations between January 1, 1997 and December 31, 2004. A control group of men with CVD only was created matching by age, index CVD diagnosis date, and CVD diagnoses. Diagnosis and procedure codes identified men with BPH and CVD. Differences in medical costs, co‑morbidities, and drug treatments were assessed.

Results: Approximately 39% of men identified with BPH also had some form of CVD at the time of BPH diagnosis. Men with BPH and CVD were more likely to have additional co‑morbidities, more

frequently received medications for CVD and non‑CVD disorders, had 44% higher total medical costs than men with CVD only ( p < 0.001), and had 42% higher CVD‑related costs ( p < 0.001) than men with CVD only.

Limitations: The population studied in this analysis was primarily working individuals with health benefits provided by managed care plans; therefore, the results may not generalize to other populations.

Conclusions: This study demonstrates in a commercial payer population that men with concomitant BPH and CVD have more co‑morbidities, receive pharmacologic agents more frequently, and have higher health care resource utilization than men with CVD only. Due to the high prevalence of co‑morbid BPH and CVD, screening for BPH in men presenting with CVD may assist with earlier disease identification and cost management over time.

A B S T R A C T

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.

418 Costs and characteristics of men with co-morbid BPH and CVD © 2007 librAPHArM ltD – Curr Med res Opin 2007; 23(2)

Introduction

Benign prostatic hyperplasia (BPH) is one of the most common disorders in aging men. Approximately 50% of all men have BPH identifiable histology by the age of 60 years, and, by the age of 85 years, BPH histology is identified in almost 90% of men1. Similar to BPH, the incidence and prevalence of cardiovascular disease (CVD) increases with age. In 2002, the prevalence of CVD in the US was 34.4% in men; over 32 million men had some form of CVD2. Concomitant HTN with BPH has a reported prevalence of 25%3,4.

Clinical BPH is differentiated from histological BPH by the progressive development of lower urinary tract symptoms (LUTS)5. The hallmark symptoms of BPH include a weak urinary stream, hesitancy, intermittency and sensations of incomplete emptying, as well as frequency, urgency, urge incontinence, and nocturia5. Twenty‑six percent of men in their 50s experience moderate to severe LUTS, increasing to 46% of men aged 80 years or older6. Moderate to severe symptoms are considered bothersome, and men tend to seek care when their symptoms become bothersome7.

There are a number of treatment options for BPH, which include watchful waiting, medical therapy, balloon dilatation, and various surgical procedures8. Historically, surgery was the most common treatment for BPH; however, the number of surgeries conducted for BPH has declined, while the number of patients receiving medications to treat BPH has increased4,5. In 2000, medications were prescribed in 7.3–12% of BPH or LUTS visits5. Between 1996 and 1998, the average annual expenditure for BPH prescription drugs was $USD194 million5. BPH expenditures have decreased over time, mainly due to the decline in hospitalizations5. Despite the decline, men between the ages of 45 and 64 years, with BPH, have higher total medical expenditures ($USD5729) than men without BPH ($USD4193)9. Similar results were found in a study looking at direct annual medical costs in men between the ages of 18 and 64 years. The difference in annual costs was $USD2577 in men without a BPH claim ($USD3138) compared to men with a BPH claim ($USD5715)5. Exclusive of pharmaceuticals, costs associated with BPH for hospitalizations, outpatient department visits, office visits, and emergency room (ER) visits were $USD1.1 billion in 20005.

The most prevalent CVD is hypertension (HTN); 31.5% of men have HTN2. In 2005, the estimated direct and indirect costs of CVD are estimated at $USD393.5 billion2. This figure includes $USD241.9 billion in direct medical costs with hospitalizations accounting for almost half of the direct medical costs at $USD109.8 billion2. Medications and devices are the

next most costly expenditures in CVD, with $USD45 billion estimated to be spent on these in 20052. The number of visits to physician offices, the emergency room (ER), and outpatient departments in 2002 were 80 million, 4.6 million, and 6 million, respectively2. Approximately $USD151.6 billion is associated with lost productivity from morbidity and mortality associated with CVD2.

BPH symptoms appear to increase when men have measured HTN10. The reverse has also been noted; persons experiencing BPH symptoms are more likely to have a diagnosis of HTN10. Age‑related increases in sympathetic tone could contribute to the patho‑physiology of both diseases3. Another study using retrospective chart review examined the relationship between BPH and coronary artery disease (CAD)11. The prevalence of CAD in men with no evidence of BPH was 9%; however, men with BPH had a significantly higher prevalence of CAD (29%). Both groups had equal risk factors for CVD, such as smoking, diabetes, and HTN.

Men with both BPH and HTN tend to have worse BPH symptoms10. As such, it can be hypothesized that because symptoms may be worse, health care utilization and expenditures would be greater in this population compared to men with only BPH or only HTN. In addition to differences in costs, co‑morbidities and medication use in this group may also differ. Furthermore, assessment of such differences may highlight opportunities for earlier disease identification and cost management strategies for men with BPH and CVD. To this end, the purpose of this analysis is to compare the costs and treatment patterns between men with concomitant BPH and CVD and men with CVD only, in a commercial population.

MethodsStudy design

A retrospective, matched cohort study was utilized to assess the incremental impact of BPH in terms of medical costs, co‑morbidities, and drug treatments by comparing a cohort of men with BPH and CVD (cases) to another cohort of men with CVD only (controls).

Data source

A large, integrated database comprised of both medical and pharmacy claims (PHARMetrics, Watertown, MA), including inpatient and outpatient diagnoses as well as procedure codes, was selected as the source of data for this study. This database includes data from 73 managed care organizations (MCOs) and encompasses 57 million lives from commercial payers in the United States. The data received are de‑identified.

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.

© 2007 librAPHArM ltD – Curr Med res Opin 2007; 23(2) Costs and characteristics of men with co-morbid BPH and CVD Shah et al. 419

Study timeframe

The study period for the analysis was January 1, 1997 through June 30, 2004 (Figure 1). The enrollment period was from July 1, 1997 through June 30, 2003. The index BPH diagnosis date used to identify patients with BPH was defined as the date of the first diagnosis for BPH within the enrollment period. BPH was defined as International Classification of Diseases, Ninth Revision, Clinical Modifications (ICD‑9‑CM), code 600.xx. The 6 months prior to the index BPH diagnosis date was defined as the pre‑period. The follow‑up period was defined as the 12 months after the index BPH diagnosis date. The patient‑specific period was the entire time the patients were followed, which was the pre‑period plus the follow‑up period totaling 18 months. The age of the patient was based on the date of index BPH diagnosis.

identification of patients with bPH and CVD (cases)

All patients were male, ≥ 35 years of age, had at least one BPH diagnosis (ICD‑9‑CM code 600.xx) during the enrollment period, and had continuous enrollment during the patient‑specific period. Only newly diagnosed patients with BPH were included as defined by a 6‑month clean period prior to the index diagnosis. Men with BPH were defined as having co‑morbid CVD if they had a CVD‑related diagnosis and/or procedure code during the pre‑period (Table 1). Men were excluded if they had a history of prostate cancer, bladder cancer, or prostate surgery during the pre‑period.

identification of patients without bPH but with CVD (controls)

Controls were men without a diagnosis of BPH during the study period, but with evidence of CVD. The same diagnosis and procedure codes used to identify a

CVD diagnosis in cases were used to identify a CVD diagnosis in controls. Controls were included only if they matched with cases on age (± 2 years), index CVD diagnosis date (± 3 months), and type of CVD diagnoses (HTN, stroke, angina, myocardial infarction, and heart failure). Up to four controls were selected for each case. All controls were given the same index BPH diagnosis date as their matched case to allow for comparisons over the same patient‑specific period.

Variables of interestCosts

Direct medical costs were captured in this analysis and represent reported costs from commercial payers in the US. Both total costs of care and CVD‑specific costs were measured. The cost differences between the control group and the cases are those costs attributable to BPH. All medical and pharmacy claims occurring during the patient‑specific period were used to calculate the total cost of care. Costs were categorized into inpatient, outpatient, ER, and pharmacy costs. The database included episode treatment group codes (ETGs). The ETG methodology was used to identify CVD‑related costs12,13. ETGs identify and group healthcare services utilized by a patient and, thus, the claims produced from those services to an underlying cause of illness. They allow for case mix adjustment and the establishment of clinical homogeneity between individuals. ETGs differ from diagnostic‑related groups in that both inpatient and outpatient episodes of care are applied when forming them. Since data were used from various years, the Medical Consumer Price Index (M‑CPI) for medical care was used to adjust all costs to 2004 dollars14.

Co-morbidities

The top ten co‑morbidities were examined for men with BPH and CVD (cases) using ICD‑9 codes. The

Figure 1. Study period

Study period

Enrollment period

January 1, 1997 July 1, 1997 June 30, 2003 June 30, 2004

Index BPHdiagnosis date

6-month pre-period 12-month follow-up period

Index CVD diagnosis date

Patient-speci�c period

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


Table 1. Three-digit cardiovascular ICD-9-codes and cardiovascular CPT code

ICD-9-CM codes/CPT codes

Category description ICD-9-CM codes/CPT codes

Category description

391 Rheumatic fever with heart involvement 441 Aortic aneurysm and dissection 393 Chronic rheumatic pericarditis 442 Other aneurysm 394 Diseases of mitral valve 443 Other peripheral vascular disease 395 Diseases of aortic valve 444 Arterial embolism and thrombosis 396 Diseases of mitral and aortic valves 446 Polyarteritis nodosa and allied conditions 397 Diseases of other endocardial structures 447 Other disorders of arteries and arterioles 398 Other rheumatic heart disease 448 Disease of capillaries 401 Essential hypertension 451 Phlebitis and thrombophlebitis 402 Hypertensive heart disease 452 Portal vein thrombosis 403 Hypertensive renal disease 453 Other venous embolism and thrombosis 404 Hypertensive heart and renal disease 454 Varicose veins of lower extremities 405 Secondary hypertension 455 Hemorrhoids 410 Acute myocardial infarction 456 Varicose veins of other sites 411 Other acute and subacute forms of

ischemic heart disease 458 Hypotension

412 Old myocardial infarction 459 Other disorders of circulatory system 413 Angina pectoris 35001–35162 Direct repair of aneurysm of excision

(partial or total) and graft insertion for aneurysm, pseudoaneurysm, ruptured aneurysm, and associated occlusive disease

414 Other forms of chronic ischemic heart disease

35180–35190 Repair arteriovenous fistula

415 Acute pulmonary heart disease 35201–35286 Repair blood vessel other than for fistula, with or without patch angioplasty

416 Chronic pulmonary heart disease 35301–35381 Thromboendarterectomy 417 Other diseases of pulmonary circulation 35400 Angioscopy 420 Acute pericarditis 35450–35460 Transluminal angioplasty (open) 421 Acute and subacute endocarditis 35470–35473 Transluminal angioplasty (percutaneous) 422 Acute myocarditis 35474–35476 Transluminal angioplasty (transluminal

balloon angioplasty) 423 Other diseases of pericardium 35480–35485 Transluminal atherectomy (open) 424 Other diseases of endocardium 35490–35495 Transluminal atherectomy (percutaneous) 425 Cardiomyopathy 35500–35572 Bypass graft (vein) 426 Conduction disorders 35582–35587 Bypass graft (in-situ vein) 427 Cardiac dysrhythmias 35600–35671 Bypass graft (other than vein) 428 Heart failure 35681–35683 Composite grafts 429 Ill-defined descriptions and

complications of heart disease 35685–35686 Adjuvant techniques

430 Subarachnoid hemorrhage 35691–35695 Arterial transportation 431 Intracerebral hemorrhage 35700–35907 Exploration/revision 432 Other and unspecified intracranial

hemorrhage 36000–36015 Intravenous

433 Occlusion and stenosis of precerebral arteries

36100–36299 Intra-arterial/intra-aortic

434 Occlusion of cerebral arteries 36400–36550 Venous 435 Transient cerebral ischemia 36600–36660 Arterial 436 Acute, but ill-defined, cerebrovascular

disease 36680 Intra-osseous

437 Other and ill-defined cerebrovascular disease

36800–36870 Hemodialysis access, intervascular cannulation for extracorporeal circulation, or shunt insertion

438 Late effects of cerebrovascular disease 37195–37209 Transcatheter procedures 440 Atherosclerosis 37250–37251 Intravascular ultrasound services

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


co‑morbidities found most commonly in cases were assessed in the control patients as well. Additionally, co‑morbidities of interest in a BPH population were examined, if not already found among the top ten prevalent co‑morbidities. These included cardiac dys‑rhythmias (458), hypotension (427), angina pectoris (413), renal insufficiency (539.9), chronic renal disease (582), decreased libido (799.81), and erectile dysfunc‑tion (ED) (607.84, 302.72).

The Dartmouth–Manitoba and Deyo modified Charlson index was created for each patient. This index contains 19 categories of co‑morbidities, primarily defined by ICD‑9‑CM diagnosis codes. Higher scores represent a higher burden of co‑morbidity. Charlson index scores for this study were derived by evaluating the presence of one of the indicated diagnoses in the inpatient or outpatient setting in the 6‑month pre‑period. The 6‑month time period was selected because all patients were known to have at least 6 months of continuous eligibility as defined by the inclusion criteria. In addition, a count variable was created which calculated the number of unique disease states during the pre‑period that were not used to calculate the Charlson index.

Therapeutic drug classes

Pharmacy claims were analyzed by therapeutic drug classes to ascertain the most frequent classes received by men with BPH and CVD. In addition, specific thera‑peutic drug classes of interest were evaluated, including antihypertensives (angiotensin‑converting enzyme [ACE] inhibitors, diuretics, calcium channel blockers, beta‑blockers, alpha‑blockers, angiotensin II receptor blockers [ARBS]) and ED drugs. A count variable for the number of unique prescription drug categories received during the pre‑period was also created.

Statistical analyses

Chi‑square tests of proportionality were used to compare categorical variables, while continuous variables were assessed using t‑tests on untransformed data. All categorical and continuous variables had samples sizes greater than 30. Except for costs, data were normally distributed. The a priori level of significance was α =0.05. Multivariate, OLS regression models were used to predict total and CVD‑related spending during the 1‑year follow‑up period. The models controlled for age, co‑morbidities, and total costs during the pre‑period to isolate the effects of the cohort (BPH and CVD vs. CVD alone). Both the dependent and independent cost variables were log‑transformed to normalize the distribution. All statistical analyses were done using SAS software, version 8.0 (SAS Corp, Cary, NC).

ResultsPopulation demographics and characteristics

Initially, 704 658 men were identified as having BPH by ICD‑9 code. Approximately 69% were eliminated due to one or more exclusion criteria leaving a total of 217 933 patients with BPH available for CVD assessment (Table 2). Of the remaining men, approximately 39.2% (n = 85 564) had established CVD at the time of BPH diagnosis and were used in this study.

A maximum of four control patients (CVD only) were identified for each of the 85 564 cases (BPH plus established CVD). The final sample population consisted of 82 828 cases and 271 440 controls, as 2736 cases (3.2%) did not adequately match with any of the potential controls. Table 3 provides information on the matching variables. All are statistically significant different due to the sample size.

Co-morbidities – bPH and CVD versus CVD only

Hyperlipidemia and diabetes were observed more frequently in men with BPH and CVD than in men with CVD only (both p < 0.001). In fact, all of the top co‑morbidities were observed more frequently in men with BPH and CVD than in men with CVD only (all p < 0.001; Table 4). For the additional co‑morbidities of interest, ED, premature ejaculation, hypotension, and chronic renal disease occurred more frequently

Table 2. Inclusion and exclusion criteria

Inclusion criterion N (%)

Patients with BPH diagnosis 704 658 (100%)

Exclusion criteria N (%)

Female 4477 (0.6%)

< 35 years of age 8364 (1.2%)

Index BPH Dx date outside of enroll-ment period (07/1997–06/2003)

48 968 (6.9%)

Not continuously eligible during patient-specific period

456 945 (64.8%)

Prostate cancer 60 657 (8.6%)

Bladder cancer 10 401 (14.7%)

Prostate surgery during the pre-period 5690 (0.8%)

Total population available for CVD assessment

217 933 (30.9%)

Patients with BPH and no evidence CVD

78 822 (36.2%)

Patients with BPH and incident CVD 53 547 (24.6%)

Patients with BPH and prevalent CVD 85 564 (39.2%)

*Exclusion criteria not sequential: an excluded patient may have met > 1 exclusion criterion

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


in men with BPH and CVD than in men with CVD only. Since only ED occurred in more than 2% of the population, it is the only one reported in Table 4, along with the top 10 co‑morbidities identified.

therapeutic drug classes – bPH and CVD versus CVD only

Men with BPH and CVD were consistently more likely to receive agents in CVD therapeutic drug categories than men with CVD only (all p < 0.0001; Table 5). The largest difference in use was seen in peripherally acting alpha‑blocking agents with 2% use in men with CVD only and 22% use in men with BPH and CVD. This is likely related to dual indications for BPH and hypertension with some of the peripherally acting alpha‑blocking agents. The remaining CVD treatments are not indicated for BPH, and as such much smaller differences between the two cohorts were observed (~1% for inotropic agents, thiazide diuretics, non‑cardioselective beta blockers). Although these smaller differences are statistically significant, this likely is due

to large sample sizes. Within the non‑CVD therapeutic categories, men with BPH and CVD were significantly more likely to receive each assessed therapeutic category than men with CVD only (all p < 0.0001).

Annual costs – bPH and CVD versus CVD only

Total and CVD‑related annual costs were compared between the cases and controls. The average total annual costs for men with BPH and CVD were $USD7810 as compared to men with CVD only with costs of $USD5335. Men with CVD only had proportionally similar sources of costs as men with BPH and CVD across the categories of cost assessed, albeit men with CVD only were less costly overall (Table 6). Pharmacy costs contributed to total costs similarly in each group, at 18% in men with BPH and CVD, and 19% in men with CVD only.

As with total costs, average CVD‑related annual costs for men with BPH and CVD were $USD2204 and $USD1613 for controls with CVD only. Also

Table 3. Demographic comparison of cases and controls

Case (BPH & CVD)(N = 82 828)

Control (CVD only)(N = 271 440)

p-value

Average age (standard deviation) 59.2 (9.39) 59.3 (8.97) < 0.0001

Hypertension* (%) 67.9% 69.8% < 0.0001

Myocardial infarction* (%) 9.1% 9.7% < 0.0001

Heart failure* (%) 6.4% 6.4% < 0.0001

Angina* (%) 28.0% 28.8% < 0.0001

Stroke* (%) 3.9% 3.9% < 0.0001

*Patient-specific period

Table 4. Common co-morbidities and co-morbidities of interest

BPH & CVD CVD only ICD-9-CM code Medical condition

No. of patients (%)(N = 82 828)

No. of patients (%)(N = 271 440)

p-value

272.xx Hyperlipidemia 49 574 (60%) 127 166 (47%) < 0.001

786.xx Respiratory/chest symptoms 32 129 (39%) 87 361 (32%) < 0.001

250.xx Diabetes mellitus 18 677 (23%) 53 511 (20%) < 0.001

719.xx Joint disorders 17 775 (21%) 46 124 (17%) < 0.001

789.xx Abdomen/pelvis symptoms 17 355 (21%) 37 123 (14%) < 0.001

724.xx Back disorders 16 611 (20%) 39 837 (15%) < 0.001

599.xx Urethra/urinary disorders 15 635 (19%) 15 064 (6%) < 0.001

788.xx Urinary system symptoms 15 321 (18%) 11 214 (4%) < 0.001

427.xx Cardiac dysrhythmias 12 660 (15%) 28 103 (10%) < 0.001

367.xx Eye disorders 11 924 (14%) 31 540 (12%) < 0.001

607.84 302.72 Erectile dysfunction (ED) 7865 (9%) 10 496 (4%) < 0.001

*Patient-specific period

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


similar to total costs, CVD‑related costs by source were proportionally nearly equivalent between the two cohorts (Table 6). Inpatient hospitalizations accounted for 42–45% of the CVD‑related costs and 28–30% of the total costs.

After controlling for age, co‑morbidities (diabetes and hyperlipidemia), year enrolled, and pre‑period measures such as the Charlson index, costs, and number of therapeutic classes prescribed, log‑linear regression modeling indicated overall total medical costs were 44% higher in cases (BPH and CVD) as compared to controls (CVD only; p < 0.0001) (Table 7) and CVD‑related costs were 42% higher in cases with BPH and CVD as compared to controls with CVD only ( p < 0.0001) (Table 8). In both models, several variables were associated with higher costs including age, Charlson index, and pre‑period total costs.

Discussion

Men with newly diagnosed BPH frequently have CVD complications. Approximately 64% of men in this

Table 5. Common therapeutic drug categories

Table 6. Mean annual costs by source

Therapeutic category* BPH & CVD No. of patients (%)

(N = 82 828)

CVD only No. of patients (%)

(N = 271 440)

p-value

CVD therapeutic categories

Angiotensin converting enzyme inhibitors 20 693 (25%) 62 133 (23%) < 0.001

Antiadrenergic agents, peripherally acting 18 369 (22%) 5724 (2%) < 0.001

Calcium channel blocking agents 16 013 (19%) 41 123 (15%) < 0.001

Inotropic agents 3148 (4%) 7338 (3%) < 0.001

Antihypertensive combinations 11 176 (13%) 32 771 (12%) < 0.001

Angiotensin II inhibitors 5414 (7%) 14 083 (5%) < 0.001

Loop diuretics 5819 (7%) 14 046 (5%) < 0.001

Thiazide diuretics 6782 (8%) 19 245 (7%) < 0.001

Cardioselective beta blockers 16 738 (20%) 47 433 (17%) < 0.001

Non-cardioselective beta blockers 3053 (4%) 8031 (3%) < 0.001

Other therapeutic categories

HMG-CoA reductase inhibitors 27 444 (33%) 72 952 (27%) < 0.001

Narcotic analgesic combinations 24 500 (30%) 66 059 (24%) < 0.001

Nonsteroidal anti-inflammatory agents 19 912 (24%) 54 421 (20%) < 0.001

Quinolones 17 904 (22%) 27 870 (10%) < 0.001

Macrolides 14 113 (17%) 38 618 (14%) < 0.001

Upper respiratory combinations 13 654 (16%) 37 172 (14%) < 0.001

Proton pump inhibitors 13 040 (16%) 30 984 (11%) < 0.001

Aminopenicillins 13 040 (16%) 31 305 (12%) < 0.001

Topical steroids 11 739 (14%) 28 348 (10%) < 0.001

Impotence agents 6242 (8%) 11 146 (4%) < 0.001

5 PDE inhibitors† 5478 (7%) 10 418 (5%) < 0.001

*Patient-specific period †Sildenafil, tadalafil, vardenafil

Source BPH & CVD Mean $USD (%)

(N = 82 828)

CVD only Mean $USD (%)(N = 271 440)

Total costs

ER $119 (2%) $89 (2%)

Inpatient $2225 (28%) $1578(30%)

Outpatient $1178 (15%) $701 (13%)

Office visits $1132 (14%) $693 (13%)

Other $1761 (22%) $1275 (24%)

Pharmacy $1396 (18%) $1000 (19%)

Overall $7810 (100%) $5335 (100%)

CVD-related costs

ER $37 (2%) $25 (2%)

Inpatient $934 (42%) $720 (45%)

Outpatient $242 (11%) $169 (10%)

Office visits $238 (11%) $155 (10%)

Other $463 (21%) $359 (22%)

Pharmacy $290 (13%) $184 (11%)

Overall $2204 (100%) $1613 (100%)

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


analysis had both disorders, with 61% presenting with CVD at the time of BPH diagnosis and the remaining 39% of this population diagnosed in the 12 months following BPH diagnosis. The most common diagnosis in men with CVD at the time of BPH diagnosis was hypertension. The overall rate of concomitant BPH and

hypertension in this population (21%) was very similar to a previously reported rate (25%)3.

Hypertensive patients are more likely to have renal artery stenosis, which leads to decreased urine output possibly creating a relationship with BPH15. Moreover, it is theorized that androgen dependence with BPH may

Table 7. Regression results – total healthcare costs

Table 8. Regression results – total CV-related costs

Parameter estimate T-statistic p-value

Cohort* 0.44082 53.63 < 0.0001

Age 0.02125 60.39 < 0.0001

Diabetes mellitus 0.51919 66.00 < 0.0001

Hyperlipidemia 0.58330 91.25 < 0.0001

Year enrolled†

1997 –0.15891 –8.40 < 0.0001

1998 –0.08268 –6.93 < 0.0001

1999 –0.03559 –3.57 0.0004

2001 0.52773 57.71 < 0.0001

2002 0.50385 47.66 < 0.0001

2003 0.51635 14.87 < 0.0001

Charlson index‡ 0.06630 18.16 < 0.0001

No. of unique diagnosis codes‡ 0.02825 23.78 < 0.0001

No. of unique Rx classes‡ 0.13927 106.87 < 0.0001

Total costs‡ 0.00000292 14.64 < 0.0001

*Reference group: CV disease and no BPH †Reference group: 2000 ‡Based on 6-month data prior to BPH diagnosis date

Parameter estimate T-statistic p-value

Cohort* 0.42056 33.23 < 0.0001

Age 0.07405 136.68 < 0.0001

Diabetes mellitus 0.56363 46.53 < 0.0001

Hyperlipidemia 0.95033 96.55 < 0.0001

Year enrolled†

1997 –0.35763 –12.27 < 0.0001

1998 –0.30598 –16.66 < 0.0001

1999 0.02597 1.69 0.0912

2001 0.33848 24.04 < 0.0001

2002 0.35168 21.60 < 0.0001

2003 0.25842 4.83 < 0.0001

Charlson index‡ 0.03045 5.42 < 0.0001

No. of unique diagnosis codes‡ –0.10644 –58.19 < 0.0001

No. of unique Rx classes‡ 0.28442 141.73 < 0.0001

Total costs‡ 0.00001047 34.04 < 0.0001

*Reference group: CV disease and no BPH †Reference group: 2000 ‡Based on 6-month data prior to BPH diagnosis date

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


be associated with CAD, in particular atherosclerosis11. Another common feature these two disorders share is a sympathetic nervous system etiology which further establishes a potential association between these disease states16.

As expected, in the matched cohort portion of the analysis, men with BPH and CVD had higher rates of urologic disorders than men with CVD only. Men with BPH and CVD also had higher rates of hyperlipidemia and diabetes. The higher levels of hyperlipidemia may be explained by the androgen theory11. Greater use of statin medications in the men with BPH and CVD is related to higher rates of hyperlipidemia in these men. Higher rates of hypotension and cardiac dysarythmias were also noted in men with BPH and CVD. This difference may be the result of using peripherally acting alpha‑blockers at higher rates (22% vs. 2%), since alpha‑blockers, especially the non‑selective, have been found to have hypotensive adverse events and have also been associated with cardiovascular side effects in patients with hypertension17,18. There was a higher use of quinolones in the BPH and CVD group, likely due to urinary tract infection complications, common in patients with BPH. We also noted greater use of all classes of antihypertensive medications in the cases, even though the controls were slightly more likely to have received a diagnosis of hypertension than cases.

The total costs of care between the cases and controls differed, with the men with BPH and CVD having higher costs ($USD7810 vs. $USD5335). Because this analysis matched on age and CVD complications, this controlled for many confounding elements. In addition, log‑linear regression further controlled for other patient characteristics that may have influenced these results further eliminating plausible alternative explanations for the higher costs in cases versus controls.

Greater total and CVD‑related health care resource utilization among men with BPH and CVD as compared to men with CVD only, suggests that screening men presenting with CVD may provide an opportunity for earlier disease identification and cost management strategies. One example may include evidenced‑based prescribing guidelines as clinicians must be diligent in medications prescribed for men with both CVD and BPH to ensure successful and efficient treatment of these conditions. Historically, the use of non‑selective alpha‑blockers (doxazosin and terazosin) in men with both BPH and HTN was often encouraged due to the availability of generics coupled with the ability to treat two diseases at once. The antihypertensive and lipid‑lowering treatment to prevent heart attack trial (ALLHAT) study, however, highlighted several problems with attempting to treat HTN with a non‑selective alpha‑blocker18. In the ALLHAT study, the doxazosin arm in the blood pressure component of the trial was discontinued due to

overwhelming evidence of negative clinical outcomes. Participants assigned to the doxazosin treatment arm had a statistically significant 25% higher incidence of major CVD events than participants assigned to the thiazide diuretic (chlorthalidone) arm. In addition, there were trends towards higher rates of stroke and for combined coronary heart disease endpoints, particularly coronary revascularization and angina, for the doxazosin arm as compared to the diuretic arm. Based on these findings, the American Urological Association (AUA) guidelines suggest that men with BPH and concomitant hypertension and cardiac risk factors are sub‑optimally treated if prescribed only a non‑selective alpha‑blocker to treat both LUTS and hypertension. The guidelines recommend making separate treatment decisions for each disorder19.

When comparing the prevalence of men with symptoms to the prevalence of men that have been treated, both BPH and CVD appear to be under‑diagnosed20,21. Therefore, the co‑existence of these diseases, especially in men over 60 years of age, may be even higher than originally estimated, making it more important that the primary care physicians and specialists (i.e., urologists and cardiologists) are considering both BPH and CVD when making treatment decisions3. As demonstrated in this study, of the 60.7% (n = 132 369) of men who presented without established CVD, 40.5% (n = 53 547) developed CVD in the 12 months following BPH diagnosis.

limitations

The population studied in this analysis was primarily working men in the US with health benefits provided by managed care plans; therefore, the results may not generalize to other populations. Additionally, as with all retrospective claims database analyses, coding errors may have occurred that could lead to misclassification of patients. Approximately 0.6% of the original sample of patients with a BPH diagnosis was female, suggesting coding errors for either gender or BPH. An unmeasured variable is disease severity with regards to both CVD and BPH, which mitigates our ability to specifically pinpoint why there are differences in CVD‑related resource utilization between men with CVD only and men with BPH and CVD. Additional research is warranted to better understand the dynamic interplay of these two conditions so that appropriate interventions can be promoted to improve management of both conditions in men.

Conclusion

This study demonstrates that, in a commercial payer population, men with concomitant BPH and CVD have

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.


more co‑morbidities (i.e., hyperlipidemia and diabetes), receive pharmacologic agents more frequently, and have higher health care resource utilization than men with CVD only. Greater resource utilization is noted not only with total medical costs, but also with CVD‑related medical costs. Due to the increased total and CVD‑related costs associated with co‑morbid BPH and CVD, screening for BPH in men presenting with CVD may assist with earlier disease identification and yield cost management opportunities.

Acknowledgment

Declaration of interest: Manan Shah, Melissa Butler, and Thomas Bramley received funding from Boehringer Ingelheim in support of this research. Tammy Curtice is an employee of Boehringer Ingelheim. Shari Fine has received an honorarium from Boehringer Ingelheim for consulting work.

References 1. Berry SJ, Coffey DS, Walsh PC, Ewing LL. The development

of human benign prostatic hyperplasia with age. J Urol 1984;132:474‑9

2. American Heart Association. Heart disease and stroke statistics – 2005 update. Dallas (TX): American Heart Association; 2005

3. Boyle P, Napalkov P. The epidemiology of benign prostatic hyperplasia and observations on concomitant hypertension. Scand J Urol Nephrol Suppl 1995;168:7‑12

4. Man in’t Veld AJ. Symptomatic BPH and hypertension: does comorbidity affect quality of life? Eur Urol 1998;34(Suppl 2):29‑36

5. Wei JT, Calhoun E, Jacobsen SJ. Urologic diseases in America project: benign prostatic hyperplasia. J Urol 2005;173:1256‑61

6. Barry MJ, Fowler Jr FJ, O’Leary MP, et al. The American Urological Association symptom index for benign prostatic hyperplasia [The Measurement Committee of the American Urological Association]. J Urol 1992;148:1549‑57

7. Chute CG, Panser LA, Girman CJ, et al. The prevalence of prostatism: a population‑based survey of urinary symptoms. J Urol 1993;150:85‑9

8. Sarma AV, Jacobson DJ, McGree ME et al. A population based study of incidence and treatment of benign prostatic hyperplasia among residents of Olmsted County, Minnesota: 1987 to 1997. J Urol 2005;173:2048‑53

9. Saigal CS, Joyce G. Economic costs of benign prostatic hyperplasia in the private sector. J Urol 2005;173:1309‑13

10. Michel MC, Heemann U, Schumacher H, et al. Association of hypertension with symptoms of benign prostatic hyperplasia. J Urol 2004;172:1390‑3

11. Weisman KM, Larijani GE, Goldstein MR, et al. Relationship between benign prostatic hyperplasia and history of coronary artery disease in elderly men. Pharmacotherapy 2000;20: 383‑6

12. Dang DK, Pont JM, Portnoy MA. Episode treatment groups: an illness classification and episode building system – part I. Med Interface 1996;9:118‑22

13. Dang DK, Pont JM, Portnoy MA. Episode treatment groups: an illness classification and episode building system – part II. Med Interface 1996;9:122‑8

14. US Department of Labor, Bureau of Labor Statistics. Available at www.bls.gov [accessed January 2006]

15. Sugaya K, Kadekawa K, Ikehara A, et al. Influence of hypertension on lower urinary tract symptoms in benign prostatic hyperplasia. Int J Urol 2003;10:569‑74

16. Pool JL. Role of the sympathetic nervous system in hypertension and benign prostatic hyperplasia. Br J Clin Pract Suppl 1994;74:13‑7

17. de Mey C. Cardiovascular effects of alpha‑blockers for the treat‑ment of symptomatic BPH. Eur Urol 1998;34(Suppl 2):18‑28

18. ALLHAT Collaborative Research Group. Major cardiovascular events in hypertensive patients randomized to doxazosin vs chlorthalidone: the antihypertensive and lipid‑lowering treatment to prevent heart attack trial (ALLHAT). J Am Med Assoc 2000;283:1967‑75

19. AUA Practice Guidelines Committee. AUA guideline on management of benign prostatic hyperplasia (2003); chapter 1: diagnosis and treatment recommendations. J Urol 2003;170: 530‑47

20. Trueman P, Hood SC, Nayak USL, et al. Prevalence of lower urinary tract symptoms and self‑reported diagnosed ‘benign prostatic hyperplasia’ and their effects on quality of life in a community‑based survey of men in the UK. Br J Urol Int 1999;83:410‑5

21. Lloyd‑Jones DM, Evans JC, Larson MG, et al. Treatment and control of hypertension in the community: a prospective analysis. Hypertension 2002;40:640‑6

CrossRef links are available in the online published version of this paper:http://www.cmrojournal.com

Paper CMRO‑3599_3, Accepted for publication: 11 December 2006Published Online: 23 January 2007doi:10.1185/030079906X167345

Cur

r M

ed R

es O

pin

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f A

uckl

and

on 1

1/12

/14

For

pers

onal

use

onl

y.

Documents

Comparison of health care costs and co-morbidities between men diagnosed with benign prostatic hyperplasia and cardiovascular disease (CVD) and men with CVD alone in a US commercial