answer the Questions from articals

articlecritique1_3.pdf

Home >Biology homework help >answer the Questions from articals

ORIGINAL INVESTIGATION

Use of Alendronate and Risk of Incident Atrial Fibrillation in Women Susan R. Heckbert, MD, PhD; Guo Li, MS; Steven R. Cummings, MD; Nicholas L. Smith, PhD; Bruce M. Psaty, MD, PhD

Background: A recent publication from the HORIZON (Health Outcomes and Reduced Incidence With Zoled- ronic Acid Once Yearly) trial in women with postmeno- pausal osteoporosis reported a higher risk of serious atrial fibrillation (AF) in zoledronic acid recipients than in pla- cebo recipients. This adverse effect was unexpected and had not been recognized previously.

Methods: We studied alendronate sodium ever use in relation to the risk of incident AF in women in a clinical practice setting. This population-based case-control study was conducted at Group Health, an integrated health care delivery system in Washington State. We identified 719 women with confirmed incident AF between October 1, 2001, and December 31, 2004, and 966 female control subjects without AF, selected at ran- dom from the Group Health enrollment and frequency

matched on age, presence or absence of treated hyper- tension, and calendar year.

Results: More AF case patients than controls had ever used alendronate (6.5% [n = 47] vs 4.1% [n = 40]; P = .03). Compared with never use of any bisphosphonate, ever use of alendronate was associated with a higher risk of incident AF (odds ratio, 1.86; 95% confidence interval, 1.09-3.15) after adjustment for the matching variables, a diagnosis of osteoporosis, and a history of cardiovas- cular disease. Based on the population-attributable frac- tion, we estimated that 3% of incident AF in this popu- lation might be explained by alendronate use.

Conclusion: Ever use of alendronate was associated with an increased risk of incident AF in clinical practice.

Arch Intern Med. 2008;168(8):826-831

A R E C E N T P U B L I C A T I O N 1

f r o m t h e H O R I Z O N (Health Outcomes and Re- duced Incidence With Zoledronic Acid Once

Yearly) trial of once-yearly zoledronic acid for postmenopausal osteoporosis re- ported an unexpected adverse effect: a higher risk of serious atrial fibrillation

(AF) in zoledronic acid recipients than in placebo recipients (50 [1.3%] vs 20 [0.5%] patients; P � .001). A similar finding was reported from the Fracture Intervention Trial of alendronate sodium for osteopo- rosis; the risk of serious AF events was higher in alendronate recipients than in placebo recipients (47 [1.5%] vs 31 [1.0%] patients; P = .07), but this difference did not reach statistical significance.2 The ad- verse effect of AF was not previously re- ported in users of bisphosphonates. We

used data from the population-based Group Health Atrial Fibrillation Study to examine whether alendronate use was as- sociated with risk of incident AF in women in a clinical practice setting.

METHODS

STUDY SETTING

This ongoing case-control study of risk fac- tors for incident AF3 is conducted at Group Health (GH), a large integrated health care de- livery system in Washington State. The GH hu- man subjects review committee approved the study procedures. Waiver of consent was granted for patients with language or cogni- tive difficulty and for patients who had died. All other participants provided verbal con- sent by telephone or written consent.

STUDY PARTICIPANTS

We identified GH enrollees assigned an Inter- national Classification of Diseases, Ninth Revi- sion, code for AF (427.31 [AF] or 427.32 [atrial flutter]) during any inpatient or outpatient visit

See also pages 793 and 820

Author Affiliations: Cardiovascular Health Research Unit (Drs Heckbert, Smith, and Psaty and Mr Li) and Departments of Epidemiology (Drs Heckbert, Smith, and Psaty) and Medicine (Mr Li and Dr Psaty), University of Washington, Seattle; Center for Health Studies, Group Health, Seattle (Drs Heckbert and Psaty); San Francisco Coordinating Center, California Pacific Medical Center Research Institute and University of California, San Francisco (Dr Cummings); and Seattle Epidemiologic Research and Information Center, Veterans Administration Puget Sound Health Care System, Seattle (Dr Smith).

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 826

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014

between October 1, 2001, and December 31, 2004, who had never before been assigned an International Classification of Dis- eases, Ninth Revision, code for AF during their enrollment at GH. Control subjects were shared between this study and com- panion case-control studies of myocardial infarction (MI), stroke, and venous thrombosis at GH.4-6 Controls were chosen at ran- dom from GH enrollment lists and were frequency matched to the largest of the case groups, MI cases, by age (by decade), presence or absence of treated hypertension, and calendar year. We defined an index date for all the participants as the date the AF came to clinical attention for case patients or a random date within the range of the case index dates for controls. Eligible participants were women aged 30 to 84 years with at least 4 health care visits before the index date.

DATA COLLECTION

The GH ambulatory medical record includes notes from pri- mary care and specialty physician visits, emergency depart- ment visit notes, discharge summaries, information from tele- phone contacts, electrocardiograms, and laboratory and diagnostic test reports. Trained abstractors reviewed the medi- cal records, covering a median of 20 years of clinical care be- fore the index date, in an identical manner for case patients and controls. Abstractors recorded the presence of diabetes melli- tus, hypertension, osteoporosis, congestive heart failure, val- vular heart disease, MI, angina, stroke, peripheral vascular dis- ease, revascularization procedures, cholesterol levels, height, and the most recently measured blood pressure and weight be- fore the index date. In telephone interviews, participants were asked about race (to provide demographic information about the study population), smoking status, and alcohol consump- tion before the index date. For participants who did not com- plete a telephone interview (48.3%), these data were obtained from medical record review. Information on bisphosphonate, statin, �-blocker, angiotensin-converting enzyme inhibitor, and postmenopausal estrogen use during enrollment came from the GH automated pharmacy database.

An AF diagnosis was verified if the medical record docu- mented a diagnosis of AF by electrocardiogram and clinical rec- ognition of the diagnosis by a physician, with no previous evi- dence of AF in the medical record. Postoperative AF was included only if it persisted to the time of hospital discharge, and AF as part of a terminal hospitalized illness was excluded.

AF CLASSIFICATION

For AF case patients, we defined categories to describe the du- ration and persistence of AF (AF classification) during the first 6 months after clinical recognition using evidence available in the medical record.7 The AF case patients were classified into 3 mutually exclusive groups based on the pattern of AF: transi- tory, persistent/intermittent, or sustained AF. This classifica- tion scheme closely followed that of the American College of Car- diology/American Heart Association/European Society of Cardiology (ACC/AHA/ESC) guidelines,8 but some modifica- tions were required because we were evaluating AF patterns across 6 months rather than at a single point in time. Transitory AF was defined as a single episode of AF lasting 7 days or less, without recognized recurrence of AF during the next 6 months. The AF was classified as persistent/intermittent if the initial AF episode lasted more than 7 days or if AF recurred but sinus rhythm was also present during the next 6 months (similar to the ACC/AHA/ ESC categories of persistent and paroxysmal). The AF was clas- sified as sustained if the patient was continuously in AF during the 6 months after AF onset (similar to the ACC/AHA/ESC cat- egory of permanent). Case patients were further classified as to

whether the AF diagnosis was made in an outpatient setting or in an acute care setting, which included an urgent care clinic, emergency department, or hospital admission.

MEDICATION USE

Since 1977, the pharmacy database has included a record for each prescription dispensed by a GH pharmacy. Of patients in this age group, 95.5% of GH members report filling all or al- most all prescriptions through GH pharmacies.9 Each phar- macy record includes the drug type and dose, quantity dis- pensed, and intended days’ supply of the prescription.

Ever use of alendronate was defined as the receipt of at least 2 alendronate prescriptions. To determine current use of any par- ticular medication at the index date, we searched the pharmacy data for the prescription immediately preceding the index date. If the patient received enough pills to last until the index date, assuming 80% compliance with prescribing instructions, then the patient was considered to be a current user of the drug. The total amount of alendronate used during GH enrollment up to the index date was determined by summing the milligrams of drug dispensed at each fill (pill strength multiplied by number of pills). Time since first use of alendronate was calculated as the interval, in years, between first use and the index date.

DATA ANALYSES

Diabetes mellitus was defined by a physician diagnosis in the medical record and use of insulin or an oral hypoglycemic agent before the index date. Osteoporosis was defined by a physi- cian diagnosis before the index date. Cardiovascular disease was defined as a history of MI, coronary revascularization, angina, stroke, transient ischemic attack, carotid endarterectomy, clau- dication, peripheral vascular surgery, or congestive heart fail- ure before the index date.

Multivariate logistic regression with robust variance estima- tion was used to examine the adjusted risk of new-onset AF as- sociated with ever use of alendronate, the primary study ques- tion. Preplanned secondary analyses were conducted in subgroups defined by age, hypertension, any cardiovascular disease, dia- betes mellitus, and current statin use. Polytomous logistic re- gression was used in secondary analyses to examine the rela- tionship of alendronate use to the 3 AF classes (transitory, persistent/intermittent, and sustained AF) and to the setting in which AF was clinically recognized (outpatient vs acute care). Models were adjusted for the matching variables of age (as a con- tinuous variable, with age measured finely in terms of days), treated hypertension, and calendar year and other potentially con- founding characteristics. The population-attributable fraction was calculated to estimate the AF incidence in this population that might be explained by alendronate use.10 All the analyses were conducted using a software program (Stata 8.2; Stata Corp, Col- lege Station, Texas).

RESULTS

CHARACTERISTICS OF CASE PATIENTS AND CONTROLS

We identified all female AF case patients (n = 727) and con- trols (n = 1057) with index dates during the study period. We excluded from the control group those with a history of AF (n = 77) and those with a pacemaker (n = 3) before the index date to match the exclusion criteria for the case patients. We also excluded 8 AF case patients and 11 con-

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 827

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014

trols who had ever used a bisphosphonate other than alen- dronate during GH enrollment, leaving 719 AF case pa- tients and 966 controls for analysis. The median age was 75 years for case patients and 71 years for controls, and the median duration of GH enrollment was 20 years (Table 1). As expected, AF case patients had a higher prevalence before the index date of diabetes mellitus, an- gina, MI, valvular heart disease, and congestive heart fail- ure than controls. The prevalence of osteoporosis was simi- lar in case patients and controls (10.4% vs 9.6%).

ALENDRONATE USE

Alendronate is the preferred drug on the GH formulary for the treatment of osteoporosis; other bisphospho- nates are available with previous authorization. More AF

case patients than controls had ever used alendronate (6.5% [n = 47] vs 4.1% [n = 40]; P = .03) (Table 2). The proportion with current use at the index date was simi- lar in AF case patients and controls, whereas the pro- portion with past use was higher in AF case patients than in controls (2.8% vs 1.0%; P = .01).

CHARACTERISTICS ASSOCIATED WITH ALENDRONATE USE IN CASE PATIENTS

AND CONTROLS

In case patients and controls, ever users of alendronate differed from never users in many respects (Table 3). Alendronate use was associated with older age, substan- tially lower body mass index, and higher high-density li- poprotein cholesterol levels, and fewer users had diabe- tes mellitus or any cardiovascular disease. A diagnosis of osteoporosis had been made in most alendronate us- ers and in few never users.

RELATIONSHIP OF ALENDRONATE USE TO RISK OF INCIDENT AF

Ever use of alendronate was associated with an in- creased risk of incident AF after adjustment for the match- ing variables, a diagnosis of osteoporosis, and any car- diovascular disease (odds ratio [OR], 1.86; 95% confidence interval [CI], 1.09-3.15) (Table 4). Further adjust- ment for all other variables significant in Tables 1 and 3, including race, number of physician visits in the year be- fore the index date, height, body mass index, diabetes mellitus, valvular heart disease, high-density lipopro- tein cholesterol level, and estrogen use, did not materi- ally change the estimated OR (OR, 1.83; 95% CI, 1.04- 3.24). The ORs were slightly higher for past users than for current users, but this difference did not reach sig- nificance (test for difference in ORs, P = .08) (Table 4). In sensitivity analyses with an assumption of 100% com- pliance with prescribing instructions rather than 80%, results were similar to those given in Table 4. There was

Table 1. Characteristics of AF Case Patients and Control Subjects

Characteristica

AF Case Patients (n = 719)

Control Subjects (n = 966)

Age, median (IQR), yb 75 (66-80) 71 (60-77) Treated hypertension, %b 59.8 59.6 Black race, % 1.5c 4.0 Physician visits in the past year,

median (IQR), No. 6 (3-10)c 5 (2-8)

Years enrolled at GH, median (IQR) 20 (11-31) 20 (10-30) Height, median (IQR), cm 163 (160-168)c 163 (157-168) BMI, median (IQR) 28 (24-34) 28 (24-33) Total cholesterol, median (IQR),

mg/dL 226 (198-258) 224 (196-253)

HDL cholesterol, median (IQR), mg/dL

58 (48-70) 59 (49-73)

Systolic BP, median (IQR), mm Hg 136 (122-150) 134 (122-146) Current smoking, % 8.3 8.6 �10 Drinks weekly or alcoholism, % 8.1 8.4 Diabetes mellitus, % 14.1c 10.7 Angina, % 13.5c 9.2 Myocardial infarction, % 7.5c 4.2 Valvular heart disease, % 4.5c 2.3 Congestive heart failure, % 15.3c 3.3 Ischemic stroke, % 6.1 4.1 Any cardiovascular disease, % 29.9c 19.1 Diagnosis of osteoporosis, % 10.4 9.6 Current drug use at the index date, %

Statin 14.9 13.8 Postmenopausal estrogen 20.9c 25.6 �-Blocker 28.5 27.0 ACE inhibitor or ARB 29.2 25.6 Diuretic 36.9 34.1

Abbreviations: ACE, angiotensin-converting enzyme; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); BP, blood pressure; GH, Group Health; HDL, high-density lipoprotein; IQR, interquartile range.

SI conversion factor: To convert total and HDL cholesterol to millimoles per liter, multiply by 0.0259.

a Race was missing in 1 case patient and 1 control subject, years enrolled in GH in 4 case patients and 3 control subjects, height and BMI in 10 case patients and 16 control subjects, current smoking in 0 case patients and 6 control subjects, alcohol use in 30 case patients and 58 control subjects, total cholesterol in 43 case patients and 48 control subjects, and HDL cholesterol in 56 case patients and 54 control subjects.

b Age and treated hypertension were matching variables. c P � .05, AF case patients vs control subjects.

Table 2. Alendronate Use by AF Case Patients and Control Subjects

Alendronate Use

AF Case Patients (n = 719)

Control Subjects (n = 966)

Ever use, No. (%) 47 (6.5) a 40 (4.1) Total No. of fills, median (IQR) 9 (4-20) 9 (5-17) Cumulative amount dispensed,

median (IQR), g 3.9 (1.4-8.3) 3.8 (2.1-7.8)

Ever users who were current users, % 57.4 75.0 Current use, No. (%) 27 (3.8) 30 (3.1)

Total No. of fills, median (IQR) 10 (6-23) 12 (7-19) Weekly dose, median (IQR), mg 65 (65-70) 65 (65-70)

Past use, No. (%) 20 (2.8) a 10 (1.0) Total No. of fills, median (IQR) 7 (3-16) 5 (4-8) Time since last prescription,

median (IQR), y 1.3 (0.6-2.5) 0.8 (0.3-1.7)

Abbreviations: AF, atrial fibrillation; IQR, interquartile range. a P � .05, AF case patients vs control subjects.

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 828

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014

no evidence of a difference in AF risk according to the total cumulative grams of alendronate dispensed (� 3.9 g [the median] vs � 3.9 g: OR, 1.84 [95% CI, 0.94-3.60] and 1.87 [95% CI, 0.94-3.73], respectively) or accord- ing to the interval since first prescription of alendronate (� 2.1 years [the median] vs � 2.1 years: OR, 1.57 [95% CI, 0.81-3.07] and 2.21 [95% CI, 1.12-4.37], respec- tively). Based on the population-attributable fraction, we estimated that 3.0% (95% CI, 0.4%-5.6%) of incident AF in this population might be explained by alendronate use.

Of the 719 AF case patients, 299 (41.6%) had transi- tory AF, 328 (45.6%) had persistent/intermittent AF, and 83 (11.5%) had sustained AF during the first 6 months af- ter presentation; AF classification could not be deter- mined in 9 case patients (1.3%). The risk of sustained AF associated with alendronate ever use was higher than the risk of transitory or intermittent AF (sustained AF: OR, 5.75 [95% CI, 2.50-13.25]; transitory AF: OR, 1.93 [95% CI, 0.95-3.92]; and intermittent AF: OR, 1.25 [95% CI, 0.64- 2.44]; test of difference in ORs across strata, P = .005). We did not detect a difference in the risk of AF in an outpa- tient setting (33.3% of case patients: OR, 1.57; 95% CI, 0.80- 3.11) vs an acute care setting (66.7% of case patients: OR, 2.14; 95% CI, 1.18-3.89) (test of difference in ORs, P = .39).

Risk of AF associated with alendronate use did not dif- fer in subgroups defined by age, presence of treated hy- pertension, or history of cardiovascular disease (Table 5).

Risk of AF associated with alendronate use was higher in patients with vs without diabetes mellitus (P = .03) and in those currently taking vs not taking statins (P = .02).

COMMENT

In this case-control study, the risk of incident AF was higher in women who were ever users of alendronate than in never users of bisphosphonates. We did not find evi-

Table 3. Characteristics of AF Case Patients and Control Subjects Who Were Ever Users of Alendronate and Never Users of Any Bisphosphonate

Characteristic

AF Case Patients Control Subjects

Ever Users (n = 47)

Never Users (n = 672)

Ever Users (n = 40)

Never Users (n = 926)

Age, median (IQR), y 79 (72-82) a 74 (66-80) 74 (70-79) a 71 (60-77) Treated hypertension, % 55.3 60.1 47.5 60.2 Black race, % 0 1.6 2.5 3.9 Years enrolled at GH, median (IQR) 22 (10-36) 20 (11-31) 22 (12-37) 20 (10-30) Physician visits in the past year, median (IQR), No. 8 (4-12) a 6 (3-10) 8 (3-12) a 5 (2-8) Height, median (IQR), cm 165 (160-170) 163 (160-168) 160 (156-165) 163 (157-168) BMI, median (IQR) 25 (23-29) a 28 (24-34) 25 (22-29) a 28 (24-33) Total cholesterol, median (IQR), mg/dL 220 (196-246) 227 (198-258) 221 (208-250) 224 (196-253) HDL cholesterol, median (IQR), mg/dL 63 (56-85) a 58 (48-70) 68 (49-78) a 59 (49-73) Systolic blood pressure, median (IQR), mm Hg 136 (130-148) 136 (122-150) 135 (120-143) 134 (122-146) Current smoking, % 4.3 8.6 5.1 8.8 � 10 Drinks weekly or alcoholism, % 14.0 7.7 2.6 8.6 Diabetes mellitus, % 8.5 14.4 0 a 11.1 Angina, % 4.3 14.1 2.5 9.5 Myocardial infarction, % 2.1 7.9 2.5 4.3 Valvular heart disease, % 6.4 4.3 7.5 a 2.1 Congestive heart failure, % 8.5 15.8 0 3.5 Ischemic stroke, % 8.5 6.0 0 4.3 Any cardiovascular disease, % 19.2 30.7 5.0 a 19.7 Diagnosis of osteoporosis, % 83.0 a 5.4 72.5 a 6.9 Current drug use at the index date, %

Statin 19.2 14.6 5.0 14.2 Postmenopausal estrogen 12.8 21.4 20.0 25.8 �-Blocker 25.5 28.7 32.5 26.8 ACE inhibitor or ARB 23.4 29.6 25.0 25.6 Diuretic 31.9 37.2 30.0 34.2

Abbreviations: See Table 1. SI conversion factor: To convert total and HDL cholesterol to millimoles per liter, multiply by 0.0259. a P � .05, ever users vs never users, separately in case patients and control subjects.

Table 4. Association of Incident AF With Ever Use of Alendronate Compared With Never Use of Any Bisphosphonate

Use Status AF Case

Patients, No. Control

Subjects, No. Odds Ratio (95% CI) a

Never users 672 926 1 [Reference] Ever users 47 40 1.86 (1.09-3.15)

Current users 27 30 1.42 (0.78-2.59) Past users 20 10 3.27 (1.43-7.47) b

Abbreviations: AF, atrial fibrillation; CI, confidence interval. a Adjusted for the matching variables of age, treated hypertension, and

calendar year and for the diagnosis of osteoporosis and any cardiovascular disease.

b Comparing a model that included current, past, and never use vs a model that included ever and never use (P = .08).

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 829

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014

dence of a dose-response relationship according to the number of grams of alendronate dispensed or the inter- val since first prescription. Alendronate was more strongly associated with AF that was sustained during the first 6 months after clinical recognition than with AF that was transitory or persistent/intermittent, and the risk of AF associated with alendronate use was higher in patients with diabetes mellitus and in patients taking statins than in those without these characteristics. Because alendro- nate use was relatively rare in this population, the pro- portion of incident AF that might be explained by alen- dronate use was small.

The strengths of this study include the large number of AF case patients, the population-based design, the vali- dation of incident AF, the inclusion of case patients pre- senting in outpatient and acute care settings, the compa- rable ascertainment of potential confounding factors, the use of the GH pharmacy database to assess alendronate use in an unbiased manner, and the availability of de- tailed information on dose, duration, and timing of alen- dronate use. Restriction, stratification, and adjustment were used to minimize the possibility of confounding. All the participants were enrollees of a health maintenance orga- nization and, thus, had similar access to health care.

Nevertheless, this study was observational, and pa- tients were not assigned alendronate use at random. There may have been unknown or unmeasured confounding fac- tors for which adjustment was not possible. Measure- ment error in the assessment or estimation of covariates and their severity may have resulted in incomplete adjust- ment and residual confounding. Ascertainment of AF may not have been complete because we were able to identify only AF case patients who came to clinical attention. Few patients used bisphosphonates other than alendronate, so this study could not address the possibility of a similar as- sociation with other bisphosphonates. We did not exam- ine the risk of AF associated with bisphosphonate use in men. Finally, the number of alendronate users in some sub-

groups was small. In particular, the findings regarding in- creased risk in patients with diabetes mellitus and those taking statins require confirmation, and power was lim- ited to draw conclusions about the risk of AF detected in acute care vs outpatient settings and for risk associated with current vs past use of alendronate.

Atrial fibrillation was not recognized as a possible ad- verse effect of bisphosphonates11 until the recent stud- ies from the HORIZON trial of zoledronic acid1 and the Fracture Intervention Trial of alendronate.2 In those stud- ies, there was a higher rate of serious adverse events of AF, but not all AF, in bisphosphonate recipients than in placebo recipients. In contrast, in the present study, which was designed specifically to study AF, alendronate use was associated with an increased risk of AF overall and specifically for AF diagnosed in an acute care setting. In the clinical trials, the finding of an association with se- rious adverse events of AF, but not all AF events, might be owing to more accurate diagnosis of the arrhythmia when patients are hospitalized, prompting centralized re- view of the medical records, whereas nonserious ad- verse events can be based on self-report or less rigorous clinical data. The present study rigorously confirmed all cases regardless of the health care setting.

In the HORIZON trial, most AF events occurred more than 30 days after the zoledronic acid infusion, and in the Fracture Intervention Trial, the AF risk seemed to be elevated throughout the average 4 years of follow- up, suggesting that AF was not an acute adverse effect. Consistent with that finding, in the present analysis, AF risk was not higher in current users than in past users of alendronate.

The potent nitrogen-containing bisphosphonates, such as zoledronic acid and alendronate, are known to inhibit protein prenylation, thus disrupting the function of key regulatory proteins.12 Bisphosphonates accumulate in bone, where they persist for years and are detectable in body flu- ids for at least several months after administration.12 An

Table 5. Association of Alendronate Ever Use vs Never Use of Bisphosphonates With AF Risk in Subgroups Defined by Patient Characteristics

Characteristic AF Case Patients,

Use/Never Use, No. Control Subjects,

Use/Never Use, No. Odds Ratio (95% CI) a

Interaction P Value

Age, y � 70 9/234 10/434 3.01 (0.93-9.76)

.92 � 70 38/438 30/492 1.69 (0.93-3.07)

Hypertension No 21/268 21/369 1.60 (0.73-3.49)

.41 Yes 26/404 19/557 2.08 (1.00-4.35)

Any cardiovascular disease No 38/466 38/744 1.62 (0.90-2.89)

.25 Yes 9/206 2/182 4.68 (0.77-28.57)

Diabetes mellitus b

No 43/575 40/823 1.59 (0.93-2.71) .03

Yes 4/97 0/103 � Current statin use

No 38/574 38/795 1.41 (0.81-2.45) .02

Yes 9/98 2/131 12.94 (1.65-101.71)

Abbreviations: AF, atrial fibrillation; CI, confidence interval. a Adjusted for the matching variables of age, treated hypertension, and calendar year and for the diagnosis of osteoporosis and any cardiovascular disease. b There were no control subjects with diabetes mellitus who were ever users of alendronate; thus, the odds ratio for the diabetes mellitus, yes, group is infinite.

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 830

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014

acute syndrome of fever and elevation of tumor necrosis factor � and interleukin 6 levels for 1 to 2 days com- monly follow intravenous administration of bisphospho- nates13; less is known about long-term effects on inflam- matory markers. Several authors14-16 have hypothesized that inflammation is related to atrial remodeling and fibrosis and may be involved in the pathogenesis of AF. Bisphos- phonates can cause small decreases in serum calcium and phosphate levels.17-19 The atrium is sensitive to fluxes in calcium concentration,20 but whether bisphosphonates have an effect sufficient to affect atrial conduction is uncer- tain. More information is needed about whether bisphos- phonates could have effects on atrial tissue in the long term through these or other mechanisms that favor the initia- tion or persistence of AF.

Bisphosphonates differ in their affinity for binding with bone. Zoledronic acid binds most strongly, followed, in order, by alendronate, ibandronate sodium, and rised- ronate sodium.12 A recent analysis21 of combined data from several clinical trials of risedronate found no increase in AF adverse events, in contrast to the findings to date for zoledronic acid and alendronate. Further studies are needed to determine whether the association with AF var- ies by the bone-binding affinity of bisphosphonates.

In conclusion, all drugs have benefits and adverse ef- fects. Bisphosphonates reduce the risk of fracture; in par- ticular, alendronate has been shown to reduce the risk of vertebral, hip, and nonspine fractures in women with osteoporosis.1,22,23 When new information becomes avail- able about a previously unrecognized benefit or adverse effect, physicians and patients must reweigh the current knowledge about benefits and risks in making treat- ment decisions for each patient. The benefits of fracture prevention in patients at high risk for fracture will gen- erally outweigh the possible risk of AF. However, it is important to carefully weigh the benefits against the pos- sible risk of AF in women who have only modestly in- creased fracture risk and in women who have risk fac- tors for AF, such as diabetes mellitus, coronary disease, or heart failure.

Accepted for Publication: August 19, 2007. Correspondence: Susan R. Heckbert, MD, PhD, Cardio- vascular Health Research Unit, University of Washing- ton, 1730 Minor Ave, Ste 1360, Seattle, WA 98101- 1448 ([email protected]). Author Contributions: Dr Heckbert had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analy- sis. Study concept and design: Heckbert and Psaty. Acqui- sition of data: Heckbert, Smith, and Psaty. Analysis and interpretation of data: Heckbert, Li, Cummings, and Smith. Drafting of the manuscript: Heckbert and Li. Critical re- vision of the manuscript for important intellectual content: Cummings, Smith, and Psaty. Statistical analysis: Heckbert and Li. Obtained funding: Heckbert, Smith, and Psaty. Ad- ministrative, technical, and material support: Heckbert and Smith. Study supervision: Heckbert and Psaty. Financial Disclosure: Dr Cummings has received re- search support from Amgen, Novartis, Lilly, Pfizer, and

Zelos and consulting fees and honoraria from Amgen, Novartis, Lilly, Zelos, Merck, and P&G-Aventis. Funding/Support: This study was supported by grants HL068986, HL043201, HL073410, and HL068639 from the National Heart, Lung, and Blood Institute. Role of the Sponsor: The funding agency had no role in the design and conduct of the study; collection, man- agement, analysis, and interpretation of the data; and preparation, review, and approval of the manuscript.

REFERENCES

1. Black DM, Delmas PD, Eastell R, et al. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med. 2007;356(18):1809-1822.

2. Cummings SR, Schwartz AV, Black DM. Alendronate and atrial fibrillation [letter]. N Engl J Med. 2007;356(18):1895-1896.

3. Dublin S, French B, Glazer NL, et al. Risk of new-onset atrial fibrillation in rela- tion to body mass index. Arch Intern Med. 2006;166(21):2322-2328.

4. Psaty BM, Heckbert SR, Atkins D, et al. The risk of myocardial infarction asso- ciated with the combined use of estrogens and progestins in post-menopausal women. Arch Intern Med. 1994;154(12):1333-1339.

5. Psaty BM, Heckbert SR, Koepsell TD, et al. The risk of myocardial infarction as- sociated with antihypertensive drug therapies. JAMA. 1995;274(8):620-625.

6. Smith NL, Heckbert SR, Lemaitre RN, et al. Esterified estrogens and conjugated equine estrogens and the risk of venous thrombosis. JAMA. 2004;292(13): 1581-1587.

7. Glazer NL, Dublin S, Smith NL, et al. Newly detected atrial fibrillation and com- pliance with antithrombotic guidelines. Arch Intern Med. 2007;167(3): 246-252.

8. Fuster V, Rydén LE, Asinger RW, et al. ACC/AHA/ESC guidelines for the manage- ment of patients with atrial fibrillation: executive summary: a report of the Ameri- can College of Cardiology/American Heart Association Task Force on Practice Guide- lines and the European Society of Cardiology Committee for Practice Guidelines and Policy Conferences (Committee to Develop Guidelines for the Management of Patients With Atrial Fibrillation). Circulation. 2001;104(17):2118-2150.

9. Saunders KW, Davis RL, Stergachis A. Group Health Cooperative. In: Strom BL, ed. Pharmacoepidemiology. 4th ed. Chichester, England: John Wiley & Sons Ltd; 2005:234.

10. Cole P, MacMahon B. Attributable risk percent in case-control studies. Br J Prev Soc Med. 1971;25(4):242-244.

11. van der Hooft CS, Heeringa J, van Herpen G, Kors JA, Kingma JH, Stricker BH. Drug-induced atrial fibrillation. J Am Coll Cardiol. 2004;44(11):2117-2124.

12. Russell RG. Determinants of structure-function relationships among bisphosphonates. Bone. 2007;40(5)(suppl 2):S21-S25.

13. Dicuonzo G, Vincenzi B, Santini D, et al. Fever after zoledronic acid administra- tion is due to increase in TNF-� and IL-6. J Interferon Cytokine Res. 2003; 23(11):649-654.

14. Boos CJ, Anderson RA, Lip GY. Is atrial fibrillation an inflammatory disorder? Eur Heart J. 2006;27(2):136-149.

15. Page RL, Roden DM. Drug therapy for atrial fibrillation: where do we go from here? Nat Rev Drug Discov. 2005;4(11):899-910.

16. Nattel S, Shiroshita-Takeshita A, Cardin S, Pelletier P. Mechanisms of atrial re- modeling and clinical relevance. Curr Opin Cardiol. 2005;20(1):21-25.

17. Vasikaran SD. Bisphosphonates: an overview with special reference to alendronate. Ann Clin Biochem. 2001;38(pt 6):608-623.

18. Poole KE, Loveridge N, Rose CM, Warburton EA, Reeve J. A single infusion of zoledronate prevents bone loss after stroke. Stroke. 2007;38(5):1519-1525.

19. Reid IR, Brown JP, Burckhardt P, et al. Intravenous zoledronic acid in postmeno- pausal women with low bone mineral density. N Engl J Med. 2002;346(9): 653-661.

20. Van Wagoner DR, Nerbonne JM. Molecular basis of electrical remodeling in atrial fibrillation. J Mol Cell Cardiol. 2000;32(6):1101-1117.

21. Karam R, Camm J, McClung M. Yearly zoledronic acid in postmenopausal os- teoporosis [letter]. N Engl J Med. 2007;357(7):712-713.

22. Black DM, Cummings SR, Karpf DB, et al; Fracture Intervention Trial Research Group. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Lancet. 1996;348(9041):1535-1541.

23. Cummings SR, Black DM, Thompson DE, et al. Effect of alendronate on risk of fracture in women with low bone density but without vertebral fractures: results from the Fracture Intervention Trial. JAMA. 1998;280(24):2077-2082.

(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 8), APR 28, 2008 WWW.ARCHINTERNMED.COM 831

Downloaded From: http://archinte.jamanetwork.com/ by a University of Alabama-Birmingham User on 09/02/2014