Osteoporos Int (2008) 19:349–356 DOI 10.1007/s00198-007-0471-4

ORIGINAL ARTICLE

Adherence to alendronate in male veterans K. E. Hansen & E. D. Swenson & B. Baltz & A. A. Schuna & A. N. Jones & M. E. Elliott

Received: 15 May 2007 / Accepted: 7 August 2007 / Published online: 26 September 2007 # International Osteoporosis Foundation and National Osteoporosis Foundation 2007

Abstract Summary In one Veterans Affairs' medical center, alendronate non-adherence was more likely in male veterans who smoke or report side effects, and less likely in men undergoing bone densitometry during therapy. Providers urgently need programs to increase adherence to osteoporosis medications. Initial programs should target patients with risk factors for non-adherence. Introduction Adherence to osteoporosis therapy in men is unknown. We hypothesized that ca. 50% of men at one center would be adherent to alendronate and one or more patient-specific factors would associate with adherence. Methods We conducted a retrospective chart review study of male veterans to determine the rates and predictors of

K. E. Hansen : E. D. Swenson : A. N. Jones University of Wisconsin School of Medicine and Public Health, Madison, WI, USA

alendronate adherence over two years. We excluded women, men who received primary care elsewhere and those who took alendronate for indications other than low bone mass. We defined adherence as a medication possession ratio ≥80% in the first 24 months of therapy. Results Adherence in the first 12 and 24 months of therapy was 59% and 54%, respectively. In multivariate analyses, non-adherence was more likely in men using tobacco (OR 2.08, 95% CI 1.13, 3.84, p=0.02) and reporting side effects (OR 2.06, 95% CI 1.14, 3.73, p=0.02) and less likely in men undergoing bone density during therapy (OR 0.49, 95% CI 0.26, 0.90, p=0.02). Conclusions Alendronate non-adherence is more likely in male veterans who smoke or report side effects, and less likely in men having bone densitometry during therapy. Providers urgently need programs to increase adherence to osteoporosis medications. Initial programs should target patients with risk factors for non-adherence.

K. E. Hansen : A. N. Jones University of Wisconsin Osteoporosis Clinical Center & Research Program, Madison, WI, USA

Keywords Adherence . Alendronate . Medication possession ratio . Men . Osteoporosis . Treatment

B. Baltz : M. E. Elliott University of Wisconsin School of Pharmacy, Madison, WI, USA

Introduction

A. A. Schuna : M. E. Elliott William S. Middleton Veterans Affairs Medical Center, Madison, WI, USA K. E. Hansen (*) Department of Medicine, UW School of Medicine and Public Health, Mailbox 3244, Clinical Science Center, 600 Highland Avenue, Madison, WI 53792, USA e-mail: [email protected]

The rate of osteoporotic fracture is increasing worldwide as the population ages, with substantial human, economic and social costs [1, 2]. Postmenopausal women are most affected, but men are also at risk. Men experience 25– 30% of hip fractures and by age 50 have a 14–25% lifetime risk of osteoporotic fracture [2–4]. Men may develop osteoporosis secondary to excess alcohol, smoking, glucocorticoid therapy, hypogonadism, immobility, low body weight, increasing age and vitamin D deficiency with secondary hyperparathyroidism [5–7]. Clinical trials docu-

350

ment the anti-fracture efficacy of osteoporosis therapy in men [8–12]. Significant advances in osteoporosis therapies should substantially reduce rates of fragility fracture in older individuals. However, the efficacy of osteoporosis medication depends on medication adherence, commonly evaluated using the medication possession ratio. The medication possession ratio is calculated as the number of days a medication was supplied between the dates of the first and last prescription, divided by the number of days between the dates of the first and last prescription [13]. Fracture risk reduction is minimal below a bisphosphonate medication possession ratio of 50%, and steadily increases as this ratio rises to ca. 80% and beyond [14–17]. Unfortunately, the medication possession ratio of postmenopausal women ranges from 34% to 70% [16, 18–21]. Although men often suffer from osteoporosis, their adherence rate is unknown, as they are rarely included in osteoporosis adherence studies [20]. Factors associated with non-adherence to osteoporosis medications in women include age [18, 20, 22–24], daily versus weekly use [25], side effects [25, 26], number of concurrent medications [20, 27], history of fracture [18, 20, 22, 25], comorbid conditions [20, 23] and bone density measurement before [20, 23, 25, 27] and during therapy [20]. Many studies use claims or pharmacy databases to determine adherence. These databases may not include patient-specific variables, found only in medical records, that might affect adherence [18–21, 28]. Additionally, many studies measure adherence for less than 12 months, whereas optimal osteoporosis therapy requires several years of therapy [29, 30]. Patient-specific risk factors for non-adherence may help health care providers target certain patients to improve their adherence and ultimately reduce fracture risk. Accordingly, in a retrospective study of male veterans receiving care at a Veterans Affairs Medical Center (VAMC), we determined adherence rates to alendronate over two years. We hypothesized that 50% of veterans would be non-adherent, and that one or more patient-specific factors would significantly associate with non-adherence.

Methods Identification of subjects and eligibility criteria Using VA Pharmacy Service computerized records, we identified male veterans at least 50 years of age who received their first prescription for alendronate between July 2000 and May 2004. Subjects eligible for study inclusion received primary care through the Madison, Wisconsin VAMC. Exclusion criteria included alendronate therapy for indications other than low bone mass (e.g.,

Osteoporos Int (2008) 19:349–356

Paget's disease of bone, steroid-induced osteoporosis), receipt of other bisphosphonates before the first alendronate prescription, receipt of primary care elsewhere and death or nursing home admission within the two-year interval. The University of Wisconsin Human Subjects and William S. Middleton Veterans Affairs Research and Development Committees approved the study and authorized a waiver of written consent. Chart review and data collection We reviewed subjects’ medical records, including preventive care and primary care provider notes, hospital discharge summaries and problem lists. We accessed records using the Computerized Patient Record System (CPRS), a VA-wide healthcare documentation system. We recorded patient characteristics potentially associated with adherence, using a standardized chart review form (Table 1). To avoid a potential identifier, we recorded the age of men over 89 years old as “age > 89 years.” For each veteran, the number of medications being used when alendronate was first prescribed was recorded from CPRS and pharmacy records. These included prescribed oral, topical and injected medications as well as calcium,

Table 1 Demographic and other characteristics of study participants Characteristic

n (%) or mean (standard deviation)

Age, years Race Caucasian Other Unknown Weight, pounds Current tobacco use Current alcohol use Clinical fracture Prior to alendronate During alendronate Bone densitometry Before alendronate During alendronate T-score Lumbar spine Total hip 33% radius Number requiring alendronate co-pay Concomitant medications Prescription for Calcium Vitamin D Multivitamin

71 (10) 178 (90%) 5 (3%) 15 (7%) 178 (34) 66 (33%) 104 (53%) 114 (58%) 17 (9%) 188 (95%) 128 (65%) −1.1 (1.7) −1.9 (0.9) −1.4 (1.2) 111 (56%) 9 (5) 164 (83%) 43 (22%) 141 (71%)

Osteoporos Int (2008) 19:349–356

120

100

80

60

40

20

90

-1 00 90

-< 80

-<

80

70 70

-< 60

-<

60

50 50

-< 40

-<

40

30 30

-< 20

-<

20

0 <1 0

We defined adherence as a medication possession ratio (fill rate) ≥ 80% for a given time period during which the veteran was eligible to receive alendronate, based on work indicating a higher fracture risk for adherence below 80% [14, 16]. “Non-adherence” was defined as having a medication possession ratio <80% for a given study interval. Based on our clinical experience at the VAMC, we estimated that 50% of men would be adherent to alendronate therapy. We calculated that a sample size of ca. 200 men would provide a 95% confidence interval (CI) of 43–57% adherence, using the large sample approximation for binomial variables [31]. A sample size of 200 also allowed us to compare patient variables in men adherent and non-adherent to alendronate. We compared the characteristics of adherent and nonadherent men using the Wilcoxon rank sum and chi square tests. We calculated 95% confidence intervals for proportions using the large sample approximation for binomial variables [31]. We used univariate and multivariate tests to determine the odds ratios for non-adherence as a function of each patient variable. In multivariate analysis, we controlled for the presence of side effects and current use of tobacco. Analyses were completed using SAS (version 9.1, SAS Institute, Cary, NC) and R (version 2.4.0, The R Project for Statistical Computing, http://www.r-project.org).

The pharmacy identified 1,283 veterans receiving alendronate. From this group, we excluded 530 veterans based on gender, transplant status and use of another bisphosphonate prior to alendronate. We excluded an additional 392 veterans due to receipt of primary care elsewhere, death and nursing home admission within the two-year study period. Receipt of alendronate for indications other than low bone mass and onset of alendronate outside the dates of July 2000 to May 2004 prompted exclusion of an additional 163 veterans. The remaining 198 veterans met all eligibility criteria allowing inclusion into the study. Subjects included in the study had a mean (standard deviation) age of 71 (10) years and most (90%) were Caucasian (Table 1). Fifty-eight percent had a clinical fracture before beginning alendronate therapy. The mean Tscores in the lumbar spine, total hip and 33% radius at baseline were −1.1 (1.7), −1.9 (0.9) and −1.4 (1.2), respectively. Thirty-three percent of subjects smoked, while 53% used alcohol. Adherence to alendronate decreased over the two-year study interval. In year one, 59% of men (CI 52–66%) were adherent. Over the two-year study interval, 54% (CI 47– 60%) were adherent. By the end of year two, 28% (n=56) had completely stopped alendronate, 14% (n=28) at their own instigation, 14% (n=27) with the approval of their health care provider, and for one veteran, this information was not documented. Figures 1 and 2 depict the medication possession ratios in deciles for all subjects for year one and for both years combined, respectively. For each time interval, the majority of men were adherent. In nonadherent men, Figs. 1 and 2 indicate that mean possession ratios were broadly distributed from 0% to 80%.

10

Statistical analysis

Results

Number of Men

vitamin D and multivitamins. We noted for each veteran whether a co-payment was required for alendronate. Adherence was calculated using a modification of the medication possession ratio, which we defined as the number of actual alendronate fills divided by the number of potential fills over the study interval [13]. We reviewed progress notes to record side effects and/or discontinuation of alendronate. When a health care provider stopped alendronate, the reason and the months of possible fills to that date were recorded. If the patient stopped therapy with the provider’s agreement, we recorded the number of potential fills to that date. If the veteran stopped alendronate without endorsement from his health care provider, the number of potential fills extended through the entire two-year study interval. We recorded bone mass measurements (g/cm2) and corresponding T-scores obtained at the L1–L4 spine, total hip and 33% radius before and during the first two years of therapy. At the time of each bone density test, height and weight were determined using a wall-mounted stadiometer and balance beam scale. Throughout the study, a single radiology technician made all measurements of bone mass, height and weight. We used body weight at time of first alendronate prescription and body mass index (obese or not obese) at the time of baseline bone density as potential variables influencing adherence.

351

Mean Possession Ratio (%)

Fig. 1 Veterans’ one-year medication possession ratios for alendronate. Medication possession ratios (prescriptions filled/prescriptions available) were calculated as described in the text. Bars depict the number of men with the indicated medication possession ratio

352

Osteoporos Int (2008) 19:349–356 120

Table 2 Characteristics associated with adherence in year one1,2 Non-adherent men, n=81 (41%)

100

Number of Men

80

Adherent men, n=117 (59%)

p value3

72 (9) 106 (99%) 177 (32) 29 (25%) 57 (50%)

0.03 0.08 0.77 0.002 0.68

63 (54%) 10 (9%)

0.20 0.98

111 (95%) 82 (70%)

0.95 0.05

−1.1 (1.8) −1.9 (0.9) −1.5 (1.3)

0.43 0.66 0.55

0.046 (0.051) 0.007 (0.024) −0.007 (0.081)

0.20

44 (54%) 9 (6)

67 (57%) 9 (5)

0.68 0.57

68 18 55 34

99 25 86 40

0.42 0.89 0.39 0.27

60

40

20

90

-1 00 90

-<

80 80

-<

70 70

-< 60

-<

60

50 50

-<

40 40

-< 30

-<

30

20 20

-< 10

<1 0

0

Mean Possession Ratio (%)

Fig. 2 Veterans’ two-year medication possession ratios for alendronate. Medication possession ratios (prescriptions filled/prescriptions available) were calculated as described in the text. Bars depict the number of men with the indicated medication possession ratio

Variables potentially associated with adherence were examined. In year one, three variables were significantly associated with non-adherence (Table 2). Non-adherent men were younger (69 (9) versus 72 (9) years, p=0.03) and were more likely to smoke (46% versus 24%, p= 0.002). Additionally, non-adherent men were less likely to undergo measurement of bone mass during alendronate therapy (57% versus 70%, p=0.05). During the two-year interval following onset of alendronate therapy, four variables were associated with non-adherence (Table 3). Non-adherent men were more likely to smoke (42% versus 25%, p=0.01) and describe side effects of alendronate (47% versus 29%, p=0.01). Non-adherent men were less likely to undergo bone mass measurement during therapy (57% versus 72%, p=0.03). Non-adherent men who underwent bone mass measurement experienced a smaller increment in lumbar spine bone mass (0.028 (0.040) versus 0.052 (0.051) g/cm2, p=0.003) compared to adherent men. In order to determine if variables noted above were independently associated with non-adherence, we carried out univariate and multivariate analyses (Table 4). In multivariate analyses, we determined that non-adherent men were significantly more likely to use tobacco (OR 2.08, 95% CI 1.13, 3.84, p=0.02) or to report side effects (OR 2.06, 95% CI 1.14, 3.73, p=0.02) but less likely to undergo measurement of bone mass during therapy (OR 0.49, 95% CI 0.26, 0.90, p=0.02). In univariate analysis, more than nine current medications conferred a greater odds ratio for non-adherence, but this relationship did not persist in multivariate analysis. Discussion Among 198 male veterans who started alendronate for treatment of low bone mass, we hypothesized a 50%

Demographic characteristics and habits Age, years 69 (9) Caucasian race 72 (95%) Weight, pounds 179 (38) Current tobacco use 37 (46%) Current acohol use 40 (53%) Clinical fracture Prior to alendronate 51 (63%) During alendronate 7 (9%) Bone density characteristics Bone densitometry measurement Before alendronate 77 (95%) During alendronate 46 (57%) Baseline T-score Lumbar spine −1.0 (1.5) Mean total hip −2.0 (1.0) 33% radius −1.4 (1.2) Change in bone mass, g/cm2 Lumbar spine 0.036 (0.041) Mean total hip

0.002 (0.023)

33% radius

−0.005 (0.032)

Treatment characteristics Required co-pay Concomitant medications Prescription for Calcium Vitamin D Multivitamin Side effects

(80%) (22%) (68%) (42%)

(85%) (21%) (74%) (34%)

0.22 0.60

Adherence was defined as filling ≥80% of alendronate fills over the study interval 2 Data are shown as the mean (standard deviation) or the raw number (percent) 3 Analyses were conducted using the Chi-square or Wilcoxon rank sum test 1

adherence rate. We found adherence rates in the first 12 and 24 months of therapy of 59% (CI 52–66%) and 54% (CI 47– 60%), respectively. We also hypothesized that one or more patient-specific characteristics would significantly relate to adherence. In univariate and multivariate analyses, nonadherent men were more likely to use tobacco and report side effects but less likely to undergo measurement of bone mass during therapy, confirming our secondary hypothesis. Ours is the first study to investigate rates of alendronate adherence exclusively in men. Male adherence to alendronate proved similar to that reported in postmenopausal

Osteoporos Int (2008) 19:349–356

353

Table 3 Characteristics associated with adherence over two years1 Non-adherent men, n=92 (46%)

Adherent men n=106 (54%)

p value2

72 (9) 96/97 (99%) 175 (28) 27 (25%) 50/104 (48%)

0.27 0.13 0.47 0.01 0.41

57 (54%) 10 (9%)

0.25 0.65

100 (94%) 76 (72%)

0.67 0.03

−1.2 (1.8) −1.9 (0.9) −1.6 (1.3)

0.40 0.36 0.12

0.052 (0.051) 0.008 (0.022) −0.006 (0.085)

0.003

48 (52%) 10 (5)

63 (59%) 9 (5)

0.30 0.34

77 21 65 43 18

87 22 76 31 13

0.76 0.72 0.87 0.01 0.20

Demographic characteristics and habits Age, years 70 (10) Caucasian race 82/86 (95%) Weight, pounds 181 (40) Tobacco use 39 (42%) Alcohol use 47/87 (54%) Clinical fracture Prior to alendronate 57 (62%) During alendronate 7 (8%) Bone density characteristics Bone densitometry measurement Before alendronate 88 (96%) During alendronate 52 (57%) Baseline T-score Lumbar spine −1.0 (1.6) Mean total hip −2.0 (0.9) 33% Radius −1.3 (1.2) Change in bone mass, g/cm2 Lumbar spine 0.028 (0.040) Mean total hip

0.001 (0.025)

33% radius

−0.006 (0.030)

Treatment characteristics Required co-pay Concomitant medications Prescription for Calcium Vitamin D Multivitamin Side effects Osteoporosis clinic referral

(84%) (23%) (71%) (47%) (20%)

(82%) (21%) (72%) (29%) (12%)

0.06 0.28

Adherence was defined as filling ≥80% of alendronate fills over the study interval 2 Analyses were conducted using the Chi-square or Wilcoxon rank sum test 1

women, for whom bisphosphonate adherence is 34% to 70% in year one [18–21, 28] and 27% to 43% in year two of therapy [16, 18]. In one study, in which 4% of 40,000 study subjects were men, female gender associated with increased adherence (relative risk 1.16, CI 1.08–1.25) [20], but the adherence rate in men was not separately reported. To our knowledge, ours is the first study to link smoking with non-adherence to osteoporosis medications. In studies with other medications, smoking associated with nonadherence [32–34]. The link may result from psychosocial

factors or patients’ health beliefs that undermine medication adherence and hamper other health pursuits including smoking cessation [32–34]. Additionally, a greater number of comorbidities and/or total number of medications associated with the health consequences of smoking may negatively affect adherence [32–34]. We recommend smoking cessation to all men with osteoporosis. We base this on studies indicating that approximately five years after men stop smoking, their risk of fragility fracture decreases to that of non-smokers [35]. However, there is no evidence that smoking cessation itself improves non-adherence. Further research is needed to determine why smoking associates with medication nonadherence. Previously reported patient characteristics associated with bisphosphonate non-adherence include side effects, daily versus weekly dosing, greater number of co-morbid conditions and medications, younger age, high cost of medication, lack of fracture and lack of bone density measurement [18, 22–27, 36]. These patient characteristics are not necessarily unique to osteoporosis adherence studies, as many of these factors associated with adherence to antihypertensive and lipid-lowering medications [37, 38]. Unlike other researchers, we found no association between non-adherence and the cost of medication, number of comorbid diseases or history of fracture [20, 25, 39]. This may be due to several factors. First, the out-of-pocket cost of alendronate is less than in other studies, as our veterans have a low monthly co-pay ($3 to $8) for alendronate. Second, although we did not record the number of comorbidities, the number of current medications may serve as a surrogate marker of this parameter. In univariate analysis, more than nine current medications conferred a greater odds ratio for non-adherence, but this relationship did not persist in multivariate analysis. Third, the lack of association between clinical fracture and adherence may relate to our small sample size, the lack of fracture as a universal motivator to take medication, or the patient’s lack of belief or understanding that alendronate decreases the risk of new fracture. It is not surprising that patient-reported side effects associate with non- adherence to bisphosphonate therapy. Patients may appropriately or inappropriately discontinue medication due to real or perceived side effects. The clinician may not recognize that the patient has stopped a medication, nor be aware of the side effects that led to this decision. Such a scenario may be most likely in patients without fracture or those with urgent or multiple medical issues. Asking the patient if he is taking the medication and whether he is having side effects is a reasonable approach. Notably, patients who experience gastrointestinal side effects with one bisphosphonate may later tolerate the same or another oral bisphosphonate [40].

354

Osteoporos Int (2008) 19:349–356

Table 4 Odds ratios for non-adherence over two years1 Variable

Univariate odds ratio (95% CI)

Demographic characteristics and habits Age > 70 years 0.69 (0.38, 1.26) Caucasian race 0.22 (0, 2.23) 1.40 (0.77, 2.56) Weight > 178 pounds4 2.14 (1.13, 4.12) Tobacco use2 Alcohol use 1.27 (0.69, 2.34) Clinical fracture Prior to alendronate 1.40 (0.76, 2.57) During alendronate 0.79 (0.24, 2.42) Bone density characteristics Bone mass measurement Before alendronate 1.32 (0.30, 6.56) During alendronate 0.51 (0.27, 0.97) Osteoporosis by bone densitometry Any site 0.72 (0.39, 1.32) Lumbar spine 0.58 (0.26, 1.26) Mean total hip 1.41 (0.72, 2.78) 33% radius 0.95 (0.42, 2.15) Treatment characteristics Required co-pay 0.68 (0.37, 1.26) Over nine current medications 1.86 (0.99, 3.50) Prescription for Calcium 1.12 (0.50, 2.55) Vitamin D 1.13 (0.54, 2.35) Multivitamin 0.95 (0.49, 1.85) 2.11 (1.14, 3.98) Side effects3

p value

Multivariate odds ratio (95% CI)2,3

p value

0.20 0.19 0.25 0.01 0.47

0.98 0.19 1.49 2.08 1.24

1.86) 1.76) 2.67) 3.84) 2.24)

0.96 0.14 0.18 0.02 0.48

0.25 0.80

1.32 (0.73, 2.37) 0.78 (0.28, 2.21)

0.36 0.64

0.75 0.04

1.14 (0.30, 4.29) 0.49 (0.26, 0.90)

0.85 0.02

0.31 0.16 0.34 1.0

0.75 0.55 1.47 1.10

1.36) 1.17) 2.81) 2.39)

0.35 0.12 0.24 0.81

0.20 0.04

0.92 (0.51, 1.68) 1.75 (0.93, 3.29)

0.79 0.08

0.85 0.73 0.88 0.01

1.13 (0.52, 2.43) 1.10 (0.54, 2.23) 1.0 (0.53, 1.89) 2.06 (1.14, 3.73)

0.76 0.79 0.99 0.02

(0.52, (0.02, (0.83, (1.13, (0.69,

(0.42, (0.26, (0.77, (0.51,

1 We compared male veterans with this characteristic to men without this characteristic. We calculated men’s odds ratios for non-adherence in univariate and multivariate analyses, the latter controlling for current tobacco use and presence of side effects. In these analyses ”CI” denotes the confidence interval 2 The multivariate odds ratio was adjusted for the presence of side effects 3 The multivariate odds ratio was adjusted for current tobacco use 4 Data were also analyzed using a body mass index ≥25 to indicate obesity, with similar results

Like Solomon et al. [20], we found that non-adherent men were less likely to undergo measurement of bone mass during alendronate therapy. However, we cannot conclude that measurement of bone mass during therapy increases adherence to alendronate therapy. Men who are motivated to take osteoporosis medication are probably more likely to return for follow up visits and comply with (or request) additional tests, including bone densitometry. Additionally, clinicians may opt to defer repeat bone mineral density studies on patients whom they know to be poorly compliant with osteoporosis therapy. Further studies are needed to clarify the influence of repeated bone density testing on medication adherence. Likewise, it is probable that good adherence causes increased lumbar spine bone mass during therapy, rather than the other way around. Our study and others emphasize the need to improve adherence to osteoporosis therapy. Improved adherence may result from increased patient-provider communication or nurse monitoring of therapy [26, 41, 42]. A nurse clinic visit at 12, 24 and 36 weeks improved adherence by 57%

when compared to a single clinic visit one year following onset of raloxifene therapy [42]. Phone calls, letters to patients, educational programs, or rewards for excellent adherence (e.g., medication discounts, clinic visit travel reimbursement, or cafeteria vouchers) may increase adherence and improve the care of patients with chronic illnesses [26, 41–45]. The VA CPRS could serve as a tool to enhance adherence. For patients starting alendronate, the CPRS could be used to automatically send letters at predefined intervals, encouraging adherence and requesting the patient to call the VA regarding medication problems. During bone densitometry measurements the radiology technician could confirm adherence and tolerability of alendronate with simple yes/no questions, recording responses on the report sent to the provider. Programmers could design a template of orders ensuring that a bone density test is ordered one year following initiation of osteoporosis medication. Our study has limitations. Patient variables associated with adherence in male veterans at one center may not apply to

Osteoporos Int (2008) 19:349–356

men treated elsewhere. Information acquired from chart reviews may be incomplete. For example, patients or providers may under-report side effects when compared to prospectively collected rates of side effects within a clinical trial. Two different investigators reviewed each chart to verify data accuracy prior to analysis, yet this data collection may still be subjective. We based adherence on number of fills, rather than number of tablets ingested or proper intake of alendronate. However, the medication possession ratio is the most commonly used research method to judge adherence [13, 26, 45]. Finally, our relatively small sample size may limit detection of all factors associated with non-adherence among male veterans. Despite these limitations, our study has strengths. We examined the charts of individual subjects to document patient-specific factors including tobacco use and side effects, information that is often unavailable in claims or pharmacy databases. The study focused exclusively on men, a population rarely included in osteoporosis adherence studies. In multivariate analysis, we found that smoking, side effects and lack of bone mass measurement during therapy were independently associated with non-adherence to alendronate treatment. Side effects and lack of bone mass measurement have previously been associated with nonadherence to antiresorptive therapy in women. Within the current report, similar relationships have now been observed in men, and an association between smoking and non-adherence was also found. In summary, men are at high risk of osteoporotic fracture. Medications, if taken regularly, can reduce the risk of fracture. Despite this, we found only half of the veterans at one VAMC remained adherent to alendronate therapy in the two years following initial prescription. Current tobacco use, occurrence of side effects and lack of bone mass measurement significantly associated with non-adherence. Recognition of these and other factors associated with non-adherence may help clinicians identify and counsel high-risk patients. The nonadherence rates observed in this study and others demand a formal program to improve patient adherence to osteoporosis therapy. Phone calls, letters or email reminders, educational programs or rewards for excellent adherence may prove to increase adherence to osteoporosis therapies and thereby decrease rates of osteoporotic fracture. Prospective studies should investigate strategies to increase adherence, improve skeletal health and lessen the personal and economic costs of fractures in men. References 1. Woolf AD, Pfleger B (2003) Burden of major musculoskeletal conditions. Bull World Health Organ 81:646–656 2. (2001) Osteoporosis prevention, diagnosis, and therapy. JAMA 285:785–795

355 3. (2003) Physician’s guide to prevention and treatment of osteoporosis. National Osteoporosis Foundation: Washington, DC 4. Binkley N, Krueger D (2002) Osteoporosis in men. WMJ 101:28–32 5. Compston J (2001) Secondary causes of osteoporosis in men. Calcif Tissue Int 69:193–195 6. Kelepouris N, Harper KD, Gannon F et al (1995) Severe osteoporosis in men. Ann Intern Med 123:452–460 7. Riggs BL, Khosla S, Melton LJ 3rd (1998) A unitary model for involutional osteoporosis: estrogen deficiency causes both type I and type II osteoporosis in postmenopausal women and contributes to bone loss in aging men. J Bone Miner Res 13:763–773 8. Orwoll E, Ettinger M, Weiss S et al (2000) Alendronate for the treatment of osteoporosis in men. N Engl J Med 343:604–610 9. Adachi JD, Saag KG, Delmas PD et al (2001) Two-year effects of alendronate on bone mineral density and vertebral fracture in patients receiving glucocorticoids: a randomized, double-blind, placebo-controlled extension trial. Arthritis Rheum 44:202–211 10. Wallach S, Cohen S, Reid DM et al (2000) Effects of risedronate treatment on bone density and vertebral fracture in patients on corticosteroid therapy. Calcif Tissue Int 67:277–285 11. Finkelstein JS, Hayes A, Hunzelman JL et al (2003) The effects of parathyroid hormone, alendronate, or both in men with osteoporosis. N Engl J Med 349:1216–1226 12. Orwoll ES, Scheele WH, Paul S et al (2003) The effect of teriparatide [human parathyroid hormone (1–34)] therapy on bone density in men with osteoporosis. J Bone Miner Res 18:9–17 13. Badamgarav E, Fitzpatrick LA (2006) A new look at osteoporosis outcomes: the influence of treatment, compliance, persistence, and adherence. Mayo Clin Proc 81:1009–1012 14. Caro JJ, Ishak KJ, Huybrechts KF et al (2004) The impact of compliance with osteoporosis therapy on fracture rates in actual practice. Osteoporos Int 15:1003–1008 15. McCombs JS, Thiebaud P, McLaughlin-Miley C et al (2004) Compliance with drug therapies for the treatment and prevention of osteoporosis. Maturitas 48:271–287 16. Siris ES, Harris ST, Rosen CJ et al (2006) Adherence to bisphosphonate therapy and fracture rates in osteoporotic women: relationship to vertebral and nonvertebral fractures from 2 US claims databases. Mayo Clin Proc 81:1013–1022 17. Huybrechts KF, Ishak KJ, Caro JJ (2006) Assessment of compliance with osteoporosis treatment and its consequences in a managed care population. Bone 38:922–928 18. Weycker D, Macarios D, Edelsberg J et al (2006) Compliance with drug therapy for postmenopausal osteoporosis. Osteoporos Int 17:1645–1652 19. Recker RR, Gallagher R, MacCosbe PE (2005) Effect of dosing frequency on bisphosphonate medication adherence in a large longitudinal cohort of women. Mayo Clin Proc 80:856–861 20. Solomon DH, Avorn J, Katz JN et al (2005) Compliance with osteoporosis medications. Arch Intern Med 165:2414–2419 21. Yood RA, Emani S, Reed JI et al (2003) Compliance with pharmacologic therapy for osteoporosis. Osteoporos Int 14:965–968 22. Papaioannou A, Ioannidis G, Adachi JD et al (2003) Adherence to bisphosphonates and hormone replacement therapy in a tertiary care setting of patients in the CANDOO database. Osteoporos Int 14:808–813 23. Curtis, JR, Westfall AO, Allison JJ et al (2006) Channeling and adherence with alendronate and risedronate among chronic glucocorticoid users. Osteoporos Int 17:1268–1274 24. Cooper A, Drake J, Brankin E (2006) Treatment persistence with once-monthly ibandronate and patient support vs. once-weekly alendronate: results from the PERSIST study. Int J Clin Pract 60:896–905 25. Rossini M, Bianchi G, Di Munno O et al (2006) Determinants of adherence to osteoporosis treatment in clinical practice. Osteoporos Int 17:914–921

356 26. Gold DT, Alexander IM, Ettinger MP (2006) How can osteoporosis patients benefit more from their therapy? Adherence issues with bisphosphonate therapy. Ann Pharmacother 40:1143–1150 27. Lo JC, Pressman AR, Omar MA et al (2006) Persistence with weekly alendronate therapy among postmenopausal women. Osteoporos Int 17:922–928 28. Downey TW, Foltz SH, Boccuzzi SJ et al (2006) Adherence and persistence associated with the pharmacologic treatment of osteoporosis in a managed care setting. South Med J 99:570–575 29. Ensrud KE, Barrett-Connor EL, Schwartz A et al (2004) Randomized trial of effect of alendronate continuation versus discontinuation in women with low BMD: results from the Fracture Intervention Trial long-term extension. J Bone Miner Res 19:1259–1269 30. Black DM, Schwartz AV, Ensrud KE et al (2006) Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA 296:2927–2938 31. Fisher L, Belle GV (1993) Biostatistics: a methodology for the health sciences. Wiley, New York 32. Vik SA, Hogan DB, Patten SB et al (2006) Medication nonadherence and subsequent risk of hospitalisation and mortality among older adults. Drugs Aging 23:345–356 33. Barr RG, Somers SC, Speizer FE et al (2002) Patient factors and medication guideline adherence among older women with asthma. Arch Intern Med 162:1761–1768 34. Lavigne M, Rocher I, Steensma C et al (2006) The impact of smoking on adherence to treatment for latent tuberculosis infection. BMC Public Health 6:66

Osteoporos Int (2008) 19:349–356 35. Law MR, Hackshaw AK (1997) A meta-analysis of cigarette smoking, bone mineral density and risk of hip fracture: recognition of a major effect. BMJ 315:841–846 36. Solomon DH, Avorn J (2003) Pharmacoepidemiology and rheumatic diseases: 2001–2002. Curr Opin Rheumatol 15:122–126 37. Morris AB, Li J, Kroenke K et al (2006) Factors associated with drug adherence and blood pressure control in patients with hypertension. Pharmacotherapy 26:483–492 38. Benner JS, Glynn RJ, Mogun H et al (2002) Long-term persistence in use of statin therapy in elderly patients. JAMA 288:455–461 39. Carr AJ, Thompson PW, Cooper C (2006) Factors associated with adherence and persistence to bisphosphonate therapy in osteoporosis: a cross-sectional survey. Osteoporos Int 17:1638–1644 40. Miller PD, Woodson G, Licata AA et al (2000) Rechallenge of patients who had discontinued alendronate therapy because of upper gastrointestinal symptoms. Clin Ther 22:1433–1442 41. Compston JE, Seeman E (2006) Compliance with osteoporosis therapy is the weakest link. Lancet 368:973–974 42. Clowes JA, Peel NF, Eastell R (2004) The impact of monitoring on adherence and persistence with antiresorptive treatment for postmenopausal osteoporosis: a randomized controlled trial. J Clin Endocrinol Metab 89:1117–1123 43. Wu JY, Leung WY, Chang S et al (2006) Effectiveness of telephone counselling by a pharmacist in reducing mortality in patients receiving polypharmacy: randomised controlled trial. BMJ 333:522 44. Raynor DK, Booth TG, Blenkinsopp A (1993) Effects of computer generated reminder charts on patients’ compliance with drug regimens. BMJ 306:1158–1161 45. Osterberg L, Blaschke T (2005) Adherence to medication. N Engl J Med 353:487–497