- Email: [email protected]
The Association of Alpha-Blockers and 5-Alpha Reductase Inhibitors in Benign Prostatic Hyperplasia With Fractures
CLINICAL INVESTIGATION
The Association of Alpha-Blockers and 5-Alpha Reductase Inhibitors in Benign Prostatic Hyperplasia With Fractures Sian Yik Lim, MD, Pavis Laengvejkal, MD, Ragesh Panikkath, MD and Kenneth Nugent, MD
Abstract: Background: Alpha-blockers and 5-alpha reductase inhibitors are common drugs used to treat benign prostatic hyperplasia (BPH), a prevalent problem in older men associated with significant morbidity and cost. Data regarding how these medications affect skeletal health and fracture risk remain scarce. Methods: Studies were identified by searching PubMed, EMBASE, the Cochrane library and Thomson Reuters Web of Knowledge. Studies involving BPH patients that reported odds ratio (OR) estimates with 95% confidence intervals (CIs) for the association between fractures and exposure to 5-alpha reductase inhibitors or alpha-blockers were included. Pooled ORs were calculated using the random-effects model. Results: Three studies addressed fracture risk in patients exposed to 5-alpha reductase inhibitors (21,366 fracture cases). Four studies addressed fracture risk in patients exposed to alpha-blockers (22,051 fracture cases). The pooled OR for fractures with 5-alpha reductase inhibitor use was 0.9 (95% CI 5 0.7–1.1). For hip/femur fractures with 5-alpha reductase inhibitor use, the pooled OR was 0.8 (95% CI 5 0.7– 1.0). The pooled OR for fractures with alpha-blockers was 1.1 (95% CI 5 0.9–1.3). There was significant statistical heterogeneity among studies for alpha-blockers. Conclusions: In patients with BPH, exposure to 5-alpha reductase inhibitors was not associated with change in fracture risk. The 5-alpha reductase inhibitors may have a small protective effect against hip/femur fractures although this was not statistically significant. Although alpha-blockers were not associated with change in fracture risk, caution is required when interpreting the results as significant heterogeneity was present. Key Indexing Terms: Alpha-blockers; 5-Alpha reductase inhibitors; Fracture; Osteoporosis. [Am J Med Sci 2014;347(6):463–471.]
O
steoporotic fractures are a major public health concern because they are associated with significant mortality, morbidity and health care costs. In 2005, there were more than 2 million osteoporotic fractures in the United States, and this number is expected to increase to more than 3 million by 2025.1,2 Annual medical costs from osteoporotic fractures are estimated at 17 billion dollars.3 Benign prostatic hyperplasia and lower urinary tract symptoms are common health problems encountered in aging men. Benign prostatic hyperplasia (BPH) is the 4th most commonly diagnosed disorder in men aged 50 years or older in the United States.4 More than 50% of men in their 60s and up to as many as 90% of octogenarians present with lower urinary tract symptoms.5 Currently, more men are being treated medically with 5-alpha reductase inhibitors and alpha-blockers as medical management is the first line of treatment.6 From the Department of Internal Medicine, Texas Tech University Health Sciences Center, Lubbock, Texas. Submitted April 1, 2013; accepted in revised form June 25, 2013. The authors have no financial or other conflicts of interest to disclose. Correspondence: Sian Yik Lim, MD, Department of Internal Medicine, Texas Tech University Health Sciences Center, 3601 4th Street, Lubbock, TX 79430 (E-mail: [email protected]).
The American Journal of the Medical Sciences
Androgens play an important role in male bone homeostasis.7 Androgen receptors are present on osteoblasts, and androgens have been shown to stimulate the proliferation of osteoblasts in vivo.8 5-alpha reductase is present in many tissues, including the prostate and bone, and converts testosterone to dihydrotestosterone (DHT, 10 times more potent than testosterone). Five-alpha reductase inhibitors used to treat BPH cause a decrease in serum DHT levels while increasing serum testosterone levels and estradiol levels.9 Five-alpha reductase inhibitors cause decreased DHT levels in the prostate, shrinking the prostate and reducing prostatic growth. The use of 5-alpha reductase inhibitors in the management of benign prostatic hyperplasia is increasing.10 Despite this, data on how hormonal changes induced by 5-alpha reductase inhibitors affect bone health in general and fracture risk specifically remain scarce. Their effect on fracture risk is an important question that warrants more investigation because fractures are associated with significant mortality and morbidity. Rodents do not have spontaneous fractures,11 and hence, animal studies cannot answer the question whether these medications affect fracture risk. This fact underscores the importance of carefully analyzing studies that assess fracture risk with the use of 5-alpha reductase inhibitors in humans. Alpha-blockers, when used to treat hypertension, were associated with increased risk of fracture.12 However, when used in the setting benign prostatic hyperplasia, there have been conflicting data, with some studies suggesting harm13 and others suggesting a protective effect.14 The use of alpha-blockers in benign prostatic hyperplasia is associated with increased incidence of hypotension-related side effects, including dizziness, hypotension and syncope.15 Syncope and dizziness are related to falls and osteoporotic fractures.16,17 In this study, we sought to determine whether 5-alpha reductase inhibitors or alpha-blockers used in the treatment of benign prostatic hyperplasia affect fracture risk. In this article, we report a meta-analysis of three 5-alpha reductase inhibitor studies12,14,18 and 4 alpha-blocker studies12–14,18 to investigate whether these drugs affect fracture risk.
METHODS Data Collection Systematic searches of PubMed, EMBASE, Cochrane Central Register of Controlled Trials and Thomson Reuters Web of Knowledge were performed. We searched these databases from the earliest available online year of indexing up to September 2012. We used a combined text word and Medical Subject Headings search strategy. The following medical subject headings were used: fracture, adrenergic alpha antagonist, alpha-blocker, 5-alpha reductase inhibitor and benign prostatic hyperplasia. Individual drug names were also searched for alphablockers (alfuzosin, prazosin, doxazosin, tamsulosin, terazosin and silodosin) and 5-alpha reductase inhibitors (finasteride, dutasteride). References of relevant articles were also screened
Volume 347, Number 6, June 2014
463
Lim et al
for additional studies. We also manually searched abstracts from scientific meetings (American College of Rheumatology—2009– 2012 and Endocrine Society—2010–2012). Our search included English language and non-English language articles. Study Selection Two independent reviewers (S.Y.L., P.L.) independently screened all titles and abstracts of identified articles. Full-text articles of potentially relevant studies were reviewed to determine article eligibility. All disagreements were resolved by consensus. A third reviewer (K.N.) mediated in the event of disagreement after discussion. We defined a study as an analysis of exposure to either alpha-blockers or 5-alpha reductase inhibitors. Therefore, 1 publication could contain several studies. Studies were included if they met the following inclusion criteria: (1) controlled clinical studies; (2) alpha reductase inhibitor/alpha-blocker users as an exposure group and nonusers as a control group; (3) reported fracture outcome risks (relative risks [RRs], odds ratio [OR], hazard ratios [HRs] and proportion ratio) and its 95% confidence interval (CI) associated with 5-alpha reductase inhibitors/alpha-blockers; (4) adequate data were provided to extract the risk estimates and (5) alphablockers and 5-alpha reductase inhibitors were used to treat benign prostatic hyperplasia. Data Extraction and Quality Assessment The following information was extracted from each article: author identification, year of publication, type of study design, sample size, definition of exposure to medication, adjusted effect estimates (ORs, RRs or HRs) between outcome measures and the use of 5-alpha reductase inhibitors/alphablockers with standard error. The quality of eligible studies was assessed and classified using a 10-point scale used to evaluate observational studies.19 Studies were evaluated based on response rate, control selection, incident cases, whether the duration of exposure was measured, whether there was adequate matching or adjustment. Each item was scored from 0 to 2. A score of 6 or higher is considered “high quality.” The quality of the evidence across studies for main outcomes was classified as “high,” “moderate,” “low” or “very low” according to the approach of the GRADE working group.20 Summary of findings table was developed with the use of GradePro software (GradePro Version 3.2; downloaded from http://ims.cochrane.org/gradepro). Data Analysis The outcome measure of fracture was calculated in each study. All risk estimates were pooled using generic inverse variance weighting and a random effects (DerSimoneon and Laird) model. The Q test and I2 statistic were used to assess heterogeneity among studies. All pooled estimates were reported with associated 95% CIs. Subgroup analyses were performed to investigate the possible sources of heterogeneity. For 5-alpha reductase inhibitors, subgroup analyses of patients who had hip/femur fractures only were performed. For alpha-blockers, factors assessed were age of patient population included (studies including patients’ age 60 years or older and studies including younger patients), type of exposure (current exposure only or any exposure included) and type of fracture included in study (hip/femur fracture or any fracture). Publication bias was assessed by 2 formal tests: the Begg adjusted rank correlation test and the Egger regression asymmetry test. The implications were assessed by the fail safe N and the trim and fill methods. Statistical analysis was performed using Comprehensive Meta-analysis Version 2.21
464
RESULTS Figure 1 schematically shows the study selection process. Four hundred thirty-six articles were identified through database searches; 11 articles were fully evaluated. We found 7 studies (4 articles) that met our inclusion criteria. Four studies evaluated the fracture risk in alpha-blockers (1 case crossover study and 3 case-control studies) and 3 studies (case-control studies) evaluated fracture risk in 5-alpha reductase inhibitors. The summary of findings table can be found in Figure 6. Three studies for 5-alpha reductase inhibitors included in our meta-analysis involved 21,366 cases of fracture (40,803 controls). Four studies for alpha-blockers in our meta-analysis involved 22,051 cases of fracture (43,543 controls). The main characteristics of the studies included in the meta-analysis are summarized in Tables 1 and 2. Fractures With 5-Alpha Reductase Inhibitors Three studies evaluated the association between 5-alpha reductase inhibitors and the risk of fractures (Figure 2). Fivealpha reductase inhibitors use was not significantly associated with increased risk of fractures as compared with nonusers; the pooled OR was 0.9 (95% CI 5 0.7–1.1). Only moderate amount of heterogeneity was present among the studies (P for heterogeneity 5 0.159, I2 5 45.6%). Hip/Femur Fractures and 5-Alpha Reductase Inhibitors When the risk of hip/femur fracture with the use of 5-alpha reductase inhibitors was evaluated, although not statistically significant, there may be a protective effect for patients exposed to 5-alpha reductase inhibitors (Figure 3). The pooled OR was 0.8 (95% CI 5 0.7–1.0). There was less heterogeneity present as compared with analysis for risk of fracture with exposure to 5-alpha reductase inhibitors (P for heterogeneity 5 0.251, I2 5 27.7%). When compared with nonhip/femur fractures, there was a significant difference between the 2 groups (P 5 0.046 for Q test based on analysis of variance).
FIGURE 1. literature.
Schematic summary of a systematic review of the
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
TABLE 1. Summary of studies included for 5-alpha reductase inhibitors Study (reference Patient Fracture No. of No. of Region number) Study type setting evaluated cases controls United Kingdom Souverein et al12
Case-control Men aged 40 years or older
Hip/femur fracture
4,571
4,571
United States
Jacobsen et al14
Hip fracture
7,076
7,076
Denmark
Vestergaard et al16
Case-control Men aged 45 years or older Case-control Men aged 60 years or older
Any fracture
9,719
29,156
Adjustment Adjusted for NSAIDS, hyponotics/ anxiolytics, antipsychotics, antiparkinsonian drugs, glucocorticoids, bronchodilators, heart failure, diabetes mellitus, osteoporosis, prostate cancer and other LUTS Adjusted for age medical center, enrollment, race/ethnicity and comorbidity, alpha-blocker use Adjusted for previous fracture, ever use of corticosteroids, ever use of drugs against epilepsy, ever use of strong analgesics, ever use of weak analgesics, Charlson index, number of bed days in 1999, number of contacts to GP or specialist in 1999, alcoholism, income, working or not, living with someone or living alone, use of beta-blockers, use of calcium channel blockers, alpha-blockers as antihypertensives, loop diuretics, thiazide diuretics, potassium sparing and other types of diuretics, prostate cancer, orchiectomy and use of LHRH agonists
Study quality 9
8
6
GP, general specialist; LHRH, luteinizing hormone–releasing hormone; LUTS, lower urinary tract symptoms; NSAIDs, nonsteroidal antiinflammatory drugs.
Fractures With Alpha-Blockers Four studies evaluated the association between alphablockers and the risk of fractures (Figure 4). Alpha-blockers were not significantly associated with change in risk of fractures as compared with nonusers; the pooled OR was 1.1 (95% CI 5 0.9–1.3). Significant heterogeneity was present among studies (P for heterogeneity 5 0.0001, I2 5 86%). Subgroup Analyses for Alpha-Blockers To find possible sources of heterogeneity, subgroup analyses were performed to analyze the influence of age of patient population included, the definition of exposure to alphablockers (current exposure only versus any exposure) and the type of fracture included in study (Figure 5). There were 2 studies that included older men (60 years or older) only. The pooled OR showed no evidence of change in fracture risk in this group (OR 5 1.1; 95% CI 5 0.8–1.7). Significant heterogeneity was present (P 5 0.00002, I2 5 94.6%). For the 2 studies that included younger men (40– 45 years or older) and older men, no change in fracture risk was demonstrated as well (OR 5 1.1; 95% CI 5 1.0–1.2). No significant heterogeneity was present (P 5 0.876; I2 5 0%). There was, however, no evidence of difference between the 2 subgroups (P 5 0.81 for Q test based on analysis of variance). The pooled OR for studies that included only current exposure as cases showed an increase in fracture risk in this group (OR 5 1.38; 95% CI 5 1.17–1.63). No significant heterogeneity was present (P 5 0.484, I2 5 0%). For the 2 studies Ó 2013 Lippincott Williams & Wilkins
that included any exposure as cases, there was no evidence for a change in fracture risk (OR 5 1.00; 95% CI 5 0.87–1.15). Significant heterogeneity was present (P 5 0.005; I2 5 87.4%), and the 2 subgroups were significantly different (P 5 0.0004 for Q test based on analysis of variance). The pooled OR for studies that included only hip/femur fracture as cases showed no significant change in fracture risk in this group (OR 5 1.0; 95% CI 5 0.8–1.2). Significant heterogeneity was present (P 5 0.025, I2 5 72.8%). For studies/ subgroup that included other fractures (excluding hip/femur fracture), there was no evidence for a change in fracture risk (OR 5 0.9; 95% CI 5 0.8–1.1). The 2 subgroups were not significantly different (P 5 0.739 for Q test based on analysis of variance). Publication Bias For 5-alpha reductase inhibitors and the risk of fracture, both the Begg test (P . 0.05) and the Egger test (P . 0.1) found no evidence to indicate publication bias. The Orwin’s Fail safe N indicated that there would be 29 studies needed to reduce the OR to a trivial one. The trim and fill test did not show any adjustments. For alpha-blockers and the risk of fracture, both the Begg test (P . 0.05) and the Egger test (P . 0.1) found no evidence to indicate publication bias. The Orwin’s Fail safe N indicated that there would be 8 studies needed to reduce the OR to a trivial one. The trim and fill test did not show any adjustments.
465
Lim et al
TABLE 2. Summary of studies included for alpha-blockers Study (reference Patient Fracture Region number) Study type setting evaluated
No. of No. of cases controls
United Kingdom Souverein et al12
Case-control
Men aged 40 years or older
Hip/femur fracture
4,571
4,571
United States
Jacobsen et al14
Case-control
Hip fracture
7,076
7,076
Denmark
Vestergaard et al16
Case-control
Men aged 45 years or older Men aged 60 years or older
Any fracture
9,719
29,156
South Korea
Lee et al13
Case crossover Men aged 65 years or older
Any fracture
685
2740
Adjustment
Study quality
Adjusted for NSAIDS, hyponotics/anxiolytics, antipsychotics, antiparkinsonian drugs, glucocorticoids, bronchodilators, heart failure, diabetes mellitus, osteoporosis, prostate cancer and other LUTS Adjusted for age medical center, enrollment, race/ethnicity and comorbidity, alpha-blocker use Adjusted for previous fracture, ever use of corticosteroids, ever use of drugs against epilepsy, ever use of strong analgesics, ever use of weak analgesics, Charlson index, number of bed days in 1999, number of contacts to GP or specialist in 1999, alcoholism, income, working or not, living with someone or living alone, use of beta-blockers, use of calcium channel blockers, alpha-blockers as antihypertensives, loop diuretics, thiazide diuretics, potassium sparing and other types of diuretics, prostate cancer, orchiectomy and use of LHRH agonists No adjustment
9
7
6
8
GP, general specialist; LHRH, luteinizing hormone–releasing hormone; LUTS, lower urinary tract symptoms; NSAIDs, nonsteroidal antiinflammatory drugs.
Sensitivity Analysis To assess whether a single study had substantial influence on main results, we excluded each study and evaluated its effect on the summary estimates and heterogeneity
of the main analysis. Regarding analysis for 5-alpha reductase inhibitors for fractures and hip/femur fracture, exclusion of the study by Vestergaard et al18 led to the possibility of decreased risk of fractures (OR 5 0.7; 95% CI 5 0.6–0.9) and hip/femur
FIGURE 2. Odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving 5-alpha reductase inhibitors (5-ARI).
466
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
FIGURE 3. (5-ARI).
Odds ratios and 95% confidence intervals (CIs) of hip/femur fracture for patients receiving 5-alpha reductase inhibitors
fracture (OR 5 0.7; 95% CI 5 0.6–0.9). For alpha-blockers, we did not find any major changes in direction or magnitude of summary estimates and P values of heterogeneity.
DISCUSSION
In our meta-analysis, we did not find evidence to support the hypothesis that 5-alpha reductase inhibitors were associated with a change in overall fracture risk. Five-alpha reductase inhibitors may be associated with decreased hip/femur fracture risk, but this did not reach statistical significance. Although there was insufficient evidence to support the hypothesis that alpha-blockers are associated with a change in fracture risk, there was significant heterogeneity among the studies. Subgroup analysis showed that studies that defined exposure as current (ie, patient were taking alpha-blockers when they had the fracture) showed an increased risk for fracture. This subgroup was significantly different from the studies that defined exposure as any exposure (previous exposure and current exposure). A review of the literature showed that bone mineral density is unchanged in patients on finasteride (follow-up time of 4 years and 1 year, respectively)22,23 and in men with 5-alpha reductase insufficiency.24 In a short-term study (3 months), bone turnover markers are not affected by finasteride.25 Therapy with transdermal DHT (3 months) did not affect osteocalcin, a bone turnover marker, when used to treat osteoporotic men.26 The results of our meta-analysis are consistent with these stud-
ies that 5-alpha reductase inhibitors have little effect on bone turnover and, therefore, do not affect fracture risk. Testosterone has an important role in the development of benign prostatic hyperplasia. The principal prostatic androgen is DHT, which causes hyperproliferation of both stromal and epithelial cells in the prostate. Furthermore, complex interactions between cytokines and growth factors from various sources, such as inflammatory cell infiltrates, contribute to the development of benign prostatic hyperplasia.27 Testosterone also has an important role in maintenance of bone health in men. Hypogonadal men are at increased risk for the development of fractures, and supplementation with testosterone decreases this risk.28 The effect of testosterone on bone health is mediated through both androgen and estrogen receptors. Current evidence suggests that estradiol, through the estrogen receptor and aromatization of androgens into estrogen, may have a greater role in maintenance of skeletal health in men than testosterone acting directly through the androgen receptor.8,28 A possible explanation for the unchanged fracture risk with the use of 5-alpha reductase inhibitors is that decreased serum levels of DHT caused by 5-alpha reductase inhibitors may be compensated by increased testosterone levels and shunting of testosterone to estrogen in bone tissue. However, more studies are required to confirm this hypothesis. Also, it is has been found that human osteoblasts predominantly express 5-alpha reductase type 1, although finasteride inhibits 5-alpha reductase type 2.29 Therefore, there may be local DHT
FIGURE 4. Odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving alpha-blockers (AB). Ó 2013 Lippincott Williams & Wilkins
467
Lim et al
FIGURE 5. Subgroup analysis: odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving alpha-blockers (AB). (A) Comparing studies, including older patients only (aged 60 years or older) compared with studies, including younger patients (40–45 years or older). (B) Comparing studies defining exposure as current compared with studies defining exposure as any exposure (current or previous). (C) Comparing studies, including hip/femur fracture compared with studies, including other fracture (excluding hip/femur fracture).
468
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
FIGURE 6. Summary of findings table: use of 5-alpha reductase inhibitors and alpha-blockers in benign prostatic hyperplasia and fracture risk.
production in the skeletal system when finasteride is used to treat benign prostatic hyperplasia. Our meta-analysis reinforces the fact that 5-alpha reductase inhibitors have a relatively benign adverse effect profile30 and do not seem to increase fracture risk. However, caution is still needed with 5-alpha reductase inhibitors, especially in patients who are at high risk for osteoporosis, as most studies that evaluated their effects were of relatively short duration and involved small sample sizes.23,25,26 Also, our meta-analysis had a small number of studies. Subgroup analysis for hip/femur fractures and sensitivity analysis, excluding the study by Vestergaard et al,18 showed that 5-alpha reductase inhibitors may be associated with decreased risk of fracture. Whether there is a true protective effect of 5-alpha reductase inhibitors requires more research and validation because this has important implications. All studies we analyzed did not include data on the effects of treatment on lower urinary tract symptoms. Inclusion of these data in future studies may be helpful as it would determine whether treatment leading to a decrease in nocturia or lower urinary tract symptoms, both known risk factors for fractures,31,32 could decrease fracture risk. In subgroup analysis, decreased heterogeneity was observed when only hip/femur fractures were analyzed, and this was statistically different from the subgroup comprised of other fractures (hip/femur fractures excluded). There may be differences between hip/femur fractures and other fractures. For example, vertebral fractures as compared with hip fractures are more difficult to assess in observational studies because the definition of a significant vertebral fracture and the clinical presentation attributable to vertebral deformity vary.33 Information about hip fractures is more readily available and easier to assess as virtually all hip fractures are treated during a hospital admission, making these fractures relatively easy to detect using Ó 2013 Lippincott Williams & Wilkins
discharge registries.34 Forearm fractures as compared with hip fractures have different risk profiles35; therefore, besides the effect of finasteride on bone strength, there may be other factors involved in the development of fracture at a particular site. The fact that alpha-blockers were not associated with any change in fracture risk needs to be interpreted with caution as significant heterogeneity was present in this meta-analysis. Studies that defined exposure to alpha-blockers as current did demonstrate increased fracture risk as compared with studies that defined exposure as any exposure. This may represent selection bias in which studies that defined exposure as any exposure may have included patients who are not currently on alpha-blockers. Therefore, the risk of fracture may have been underestimated in those studies. Caution is still warranted when prescribing alpha-blockers. Most studies analyzed in our meta-analysis did not provide sufficient data about the blood pressure effects of these medications when used to treat BPH. It is also difficult to infer cause and effect and to determine the timing of initiation of alpha-blockers and the onset of fracture from case-control studies. However, other studies suggest possible mechanisms for increased risk of fracture when alpha-blockers are used. The initiation of antihypertensives in older patients increases fracture risk, partially from associated orthostatic hypotension.36 Alpha-blockers are also associated with an increased risk of vascular events that could precipitate osteoporotic fractures related to falls.15,37 Bone protective mechanisms associated with alpha-blockers could compensate for the increased risk of vascular events. These include the induction of osteoblastic stem cell differentiation by alpha-blockers.38 Blood pressure–lowering effects of alpha-blockers may also decrease urinary calcium excretion.39 Finally, improvement of nocturia and lower urinary tract symptoms could decrease the risk of fracture.
469
Lim et al
There were several limitations in our meta-analysis. First, the number of studies in our meta-analysis was small; therefore, statistical analysis to test for publication bias may be underpowered. To compensate this, a manual search through references in articles reviewed and abstracts from scientific meetings (American College of Rheumatology—2009–2012 and Endocrine Society—2010–2012) was performed to identify unpublished studies. Second, significant heterogeneity was found in the main analyses for alpha-blockers. This could be because of the differences in study design, the definition of exposure, the dose and duration of medications used, and other factors. A random-effects model was used to partly account for differences among studies. Third, the occurrence of osteoporotic fractures may be influenced by common comorbidities in older men, such as hypogonadism, and current medications. Therefore, we tried to include adjusted ORs in our meta-analysis. However, the studies analyzed did not always adjust for the same factors. Also, residual confounding (some confounding factors were not considered/adjusted for or insufficient data collected for certain confounding factors to be adjusted) may limit our analysis. A small number of studies also limited the ability to perform subgroup analyses to evaluate for heterogeneity and to evaluate the effect of different alpha-blockers and 5-alpha reductase inhibitors. Whether there was a dose-response relationship for both drugs could not be determined because studies did not provide information on uniform dosages for evaluation. Limited data on the fracture risk of different age groups and the degree of lower urinary tract symptoms also precluded important analyses of possible causes that may explain the effects of these medications on fracture risk. There have been no studies evaluating combined therapy and fracture risk. Finally, there remains a lot to be learned about the pathophysiology of benign prostatic hyperplasia, the systemic effects of these medications and the complex interactions of calcium, vitamin D, BPH and fractures.
CONCLUSIONS In summary, our meta-analysis provided no evidence that 5-alpha reductase inhibitors and alpha-blockers were associated with any change in fracture risk. It is a reassuring that 5-alpha reductase inhibitors do not increase fracture risk and may even have protective effects. However, more studies are needed to confirm this. Although alpha-blockers were not associated with change in fracture risk, caution is required because significant heterogeneity was present in between the studies. More studies are needed to evaluate these drugs for fracture risk, especially during initiation of alpha-blocker therapy. REFERENCES 1. Lim LS, Hoeksema LJ, Sherin K, et al. Screening for osteoporosis in the adult U.S. population: ACPM position statement on preventive practice. Am J Prev Med 2009;36:366–75. 2. Looker AC, Melton LJ 3rd, Harris TB, et al. Prevalence and trends in low femur bone density among older US adults: NHANES 2005-2006 compared with NHANES III. J Bone Miner Res 2010;25:64–71. 3. Burge R, Dawson-Hughes B, Solomon DH, et al. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. J Bone Miner Res 2007;22:465–75. 4. Issa MM, Fenter TC, Black L, et al. An assessment of the diagnosed prevalence of diseases in men 50 years of age or older. Am J Manag Care 2006;12(suppl 4):S83–9. 5. 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.
470
6. 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. 7. Kasperk CH, Wergedal JE, Farley JR, et al. Androgens directly stimulate proliferation of bone cells in vitro. Endocrinology 1989;124:1576–8. 8. Clarke BL, Khosla S. Androgens and bone. Steroids 2009;74:296–305. 9. Castro-Magana M, Angulo M, Fuentes B, et al. Effect of finasteride on human testicular steroidogenesis. J Androl 1996;17:516–21. 10. Kruep EJ, Goodwin BB, Chaudhari S. Evaluation of recent trends in treatment patterns among men with benign prostatic hyperplasia. Am J Mens Health 2013;7:214–9. 11. Vanderschueren D, Vandenput L, Boonen S, et al. Androgens and bone. Endocr Rev 2004;25:389–425. 12. Souverein PC, Van Staa TP, Egberts AC, et al. Use of alpha-blockers and the risk of hip/femur fractures. J Intern Med 2003;254:548–54. 13. Lee J, Choi NK, Jung SY, et al. The risk of fracture with taking alpha blockers for treating benign prostatic hyperplasia [in Korean]. J Prev Med Public Health 2009;42:165–70. 14. Jacobsen SJ, Cheetham TC, Haque R, et al. Association between 5-alpha reductase inhibition and risk of hip fracture. JAMA 2008;300:1660–4. 15. Nickel JC, Sander S, Moon TD. A meta-analysis of the vascular-related safety profile and efficacy of alpha-adrenergic blockers for symptoms related to benign prostatic hyperplasia. Int J Clin Pract 2008;62:1547–59. 16. Radtke A, Lempert T, von Brevern M, et al. Prevalence and complications of orthostatic dizziness in the general population. Clin Auton Res 2011;21:161–8. 17. Cummings SR, Nevitt MC. Nonskeletal determinants of fractures: the potential importance of the mechanics of falls. Study of Osteoporotic Fractures Research Group. Osteoporos Int 1994;4(suppl 1):67–70. 18. Vestergaard P, Rejnmark L, Mosekilde L. Risk of fractures associated with treatment for benign prostate hyperplasia in men. Osteoporos Int 2011;22:731–7. 19. Wiens M, Etminan M, Gill SS, et al. Effects of antihypertensive drug treatments on fracture outcomes: a meta-analysis of observational studies. J Intern Med 2006;260:350–62. 20. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336:924–6. 21. Borenstein M, Hedges L, Higgins J, et al. Comprehensive Meta-Analysis. Englewood, NJ: Biostat; 2005. 22. Matsumoto AM, Tenover L, McClung M, et al. The long-term effect of specific type II 5alpha-reductase inhibition with finasteride on bone mineral density in men: results of a 4-year placebo controlled trial. J Urol 2002;167:2105–8. 23. Matzkin H, Chen J, Weisman Y, et al. Prolonged treatment with finasteride (a 5 alpha-reductase inhibitor) does not affect bone density and metabolism. Clin Endocrinol (Oxf) 1992;37:432–6. 24. Sobel V, Schwartz B, Zhu YS, et al. Bone mineral density in the complete androgen insensitivity and 5alpha-reductase-2 deficiency syndromes. J Clin Endocrinol Metab 2006;91:3017–23. 25. Tollin SR, Rosen HN, Zurowski K, et al. Finasteride therapy does not alter bone turnover in men with benign prostatic hyperplasia—a Clinical Research Center study. J Clin Endocrinol Metab 1996;81:1031–4. 26. Ly LP, Jimenez M, Zhuang TN, et al. A double-blind, placebo-controlled, randomized clinical trial of transdermal dihydrotestosterone gel on muscular strength, mobility, and quality of life in older men with partial androgen deficiency. J Clin Endocrinol Metab 2001;86:4078–88.
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
27. Wein AJ, Kavoussi LR, Partin AW, et al. Campbell-Walsh urology. Philadelphia, PA: Elsevier Saunders; 2012.
34. Marcus R, Feldman D, Kelsey JL. Osteoporosis. San Diego, London: Academic Press; 2001.
28. Sinnesael M, Boonen S, Claessens F, et al. Testosterone and the male skeleton: a dual mode of action. J Osteoporos 2011;2011:240328.
35. Graafmans WC, Ooms ME, Bezemer PD, et al. Different risk profiles for hip fractures and distal forearm fractures: a prospective study. Osteoporos Int 1996;6:427–31.
29. Issa S, Schnabel D, Feix M, et al. Human osteoblast-like cells express predominantly steroid 5alpha-reductase type 1. J Clin Endocrinol Metab 2002;87:5401–7. 30. Leedahl DD, Vo PH, Maxson PM, et al. Benign prostatic hyperplasia: implications for pharmacologic treatment and perioperative care. J Pharm Pract 2012; Epub ahead of print. 31. Parsons JK, Mougey J, Lambert L, et al. Lower urinary tract symptoms increase the risk of falls in older men. BJU Int. 2009;104:63–8.
36. Butt DA, Mamdani M, Austin PC, et al. The risk of falls on initiation of antihypertensive drugs in the elderly. Osteoporos Int 2013;24:2649– 57. 37. Morrison A, Fan T, Sen SS, et al. Epidemiology of falls and osteoporotic fractures: a systematic review. Clinicoecon Outcomes Res 2013; 5:9–18.
32. Temml C, Ponholzer A, Gutjahr G, et al. Nocturia is an age-independent risk factor for hip-fractures in men. Neurourol Urodyn 2009;28:949–52.
38. Choi YJ, Lee JY, Lee SJ, et al. Alpha-adrenergic blocker mediated osteoblastic stem cell differentiation. Biochem Biophys Res Commun 2011;416:232–8.
33. Cooper C, Cole ZA, Holroyd CR, et al. Secular trends in the incidence of hip and other osteoporotic fractures. Osteoporos Int 2011;22:1277–88.
39. Tsuda K, Nishio I, Masuyama Y. Bone mineral density in women with essential hypertension. Am J Hypertens 2001;14:704–7.
Ó 2013 Lippincott Williams & Wilkins
471
The Association of Alpha-Blockers and 5-Alpha Reductase Inhibitors in Benign Prostatic Hyperplasia With Fractures Sian Yik Lim, MD, Pavis Laengvejkal, MD, Ragesh Panikkath, MD and Kenneth Nugent, MD
Abstract: Background: Alpha-blockers and 5-alpha reductase inhibitors are common drugs used to treat benign prostatic hyperplasia (BPH), a prevalent problem in older men associated with significant morbidity and cost. Data regarding how these medications affect skeletal health and fracture risk remain scarce. Methods: Studies were identified by searching PubMed, EMBASE, the Cochrane library and Thomson Reuters Web of Knowledge. Studies involving BPH patients that reported odds ratio (OR) estimates with 95% confidence intervals (CIs) for the association between fractures and exposure to 5-alpha reductase inhibitors or alpha-blockers were included. Pooled ORs were calculated using the random-effects model. Results: Three studies addressed fracture risk in patients exposed to 5-alpha reductase inhibitors (21,366 fracture cases). Four studies addressed fracture risk in patients exposed to alpha-blockers (22,051 fracture cases). The pooled OR for fractures with 5-alpha reductase inhibitor use was 0.9 (95% CI 5 0.7–1.1). For hip/femur fractures with 5-alpha reductase inhibitor use, the pooled OR was 0.8 (95% CI 5 0.7– 1.0). The pooled OR for fractures with alpha-blockers was 1.1 (95% CI 5 0.9–1.3). There was significant statistical heterogeneity among studies for alpha-blockers. Conclusions: In patients with BPH, exposure to 5-alpha reductase inhibitors was not associated with change in fracture risk. The 5-alpha reductase inhibitors may have a small protective effect against hip/femur fractures although this was not statistically significant. Although alpha-blockers were not associated with change in fracture risk, caution is required when interpreting the results as significant heterogeneity was present. Key Indexing Terms: Alpha-blockers; 5-Alpha reductase inhibitors; Fracture; Osteoporosis. [Am J Med Sci 2014;347(6):463–471.]
O
steoporotic fractures are a major public health concern because they are associated with significant mortality, morbidity and health care costs. In 2005, there were more than 2 million osteoporotic fractures in the United States, and this number is expected to increase to more than 3 million by 2025.1,2 Annual medical costs from osteoporotic fractures are estimated at 17 billion dollars.3 Benign prostatic hyperplasia and lower urinary tract symptoms are common health problems encountered in aging men. Benign prostatic hyperplasia (BPH) is the 4th most commonly diagnosed disorder in men aged 50 years or older in the United States.4 More than 50% of men in their 60s and up to as many as 90% of octogenarians present with lower urinary tract symptoms.5 Currently, more men are being treated medically with 5-alpha reductase inhibitors and alpha-blockers as medical management is the first line of treatment.6 From the Department of Internal Medicine, Texas Tech University Health Sciences Center, Lubbock, Texas. Submitted April 1, 2013; accepted in revised form June 25, 2013. The authors have no financial or other conflicts of interest to disclose. Correspondence: Sian Yik Lim, MD, Department of Internal Medicine, Texas Tech University Health Sciences Center, 3601 4th Street, Lubbock, TX 79430 (E-mail: [email protected]).
The American Journal of the Medical Sciences
Androgens play an important role in male bone homeostasis.7 Androgen receptors are present on osteoblasts, and androgens have been shown to stimulate the proliferation of osteoblasts in vivo.8 5-alpha reductase is present in many tissues, including the prostate and bone, and converts testosterone to dihydrotestosterone (DHT, 10 times more potent than testosterone). Five-alpha reductase inhibitors used to treat BPH cause a decrease in serum DHT levels while increasing serum testosterone levels and estradiol levels.9 Five-alpha reductase inhibitors cause decreased DHT levels in the prostate, shrinking the prostate and reducing prostatic growth. The use of 5-alpha reductase inhibitors in the management of benign prostatic hyperplasia is increasing.10 Despite this, data on how hormonal changes induced by 5-alpha reductase inhibitors affect bone health in general and fracture risk specifically remain scarce. Their effect on fracture risk is an important question that warrants more investigation because fractures are associated with significant mortality and morbidity. Rodents do not have spontaneous fractures,11 and hence, animal studies cannot answer the question whether these medications affect fracture risk. This fact underscores the importance of carefully analyzing studies that assess fracture risk with the use of 5-alpha reductase inhibitors in humans. Alpha-blockers, when used to treat hypertension, were associated with increased risk of fracture.12 However, when used in the setting benign prostatic hyperplasia, there have been conflicting data, with some studies suggesting harm13 and others suggesting a protective effect.14 The use of alpha-blockers in benign prostatic hyperplasia is associated with increased incidence of hypotension-related side effects, including dizziness, hypotension and syncope.15 Syncope and dizziness are related to falls and osteoporotic fractures.16,17 In this study, we sought to determine whether 5-alpha reductase inhibitors or alpha-blockers used in the treatment of benign prostatic hyperplasia affect fracture risk. In this article, we report a meta-analysis of three 5-alpha reductase inhibitor studies12,14,18 and 4 alpha-blocker studies12–14,18 to investigate whether these drugs affect fracture risk.
METHODS Data Collection Systematic searches of PubMed, EMBASE, Cochrane Central Register of Controlled Trials and Thomson Reuters Web of Knowledge were performed. We searched these databases from the earliest available online year of indexing up to September 2012. We used a combined text word and Medical Subject Headings search strategy. The following medical subject headings were used: fracture, adrenergic alpha antagonist, alpha-blocker, 5-alpha reductase inhibitor and benign prostatic hyperplasia. Individual drug names were also searched for alphablockers (alfuzosin, prazosin, doxazosin, tamsulosin, terazosin and silodosin) and 5-alpha reductase inhibitors (finasteride, dutasteride). References of relevant articles were also screened
Volume 347, Number 6, June 2014
463
Lim et al
for additional studies. We also manually searched abstracts from scientific meetings (American College of Rheumatology—2009– 2012 and Endocrine Society—2010–2012). Our search included English language and non-English language articles. Study Selection Two independent reviewers (S.Y.L., P.L.) independently screened all titles and abstracts of identified articles. Full-text articles of potentially relevant studies were reviewed to determine article eligibility. All disagreements were resolved by consensus. A third reviewer (K.N.) mediated in the event of disagreement after discussion. We defined a study as an analysis of exposure to either alpha-blockers or 5-alpha reductase inhibitors. Therefore, 1 publication could contain several studies. Studies were included if they met the following inclusion criteria: (1) controlled clinical studies; (2) alpha reductase inhibitor/alpha-blocker users as an exposure group and nonusers as a control group; (3) reported fracture outcome risks (relative risks [RRs], odds ratio [OR], hazard ratios [HRs] and proportion ratio) and its 95% confidence interval (CI) associated with 5-alpha reductase inhibitors/alpha-blockers; (4) adequate data were provided to extract the risk estimates and (5) alphablockers and 5-alpha reductase inhibitors were used to treat benign prostatic hyperplasia. Data Extraction and Quality Assessment The following information was extracted from each article: author identification, year of publication, type of study design, sample size, definition of exposure to medication, adjusted effect estimates (ORs, RRs or HRs) between outcome measures and the use of 5-alpha reductase inhibitors/alphablockers with standard error. The quality of eligible studies was assessed and classified using a 10-point scale used to evaluate observational studies.19 Studies were evaluated based on response rate, control selection, incident cases, whether the duration of exposure was measured, whether there was adequate matching or adjustment. Each item was scored from 0 to 2. A score of 6 or higher is considered “high quality.” The quality of the evidence across studies for main outcomes was classified as “high,” “moderate,” “low” or “very low” according to the approach of the GRADE working group.20 Summary of findings table was developed with the use of GradePro software (GradePro Version 3.2; downloaded from http://ims.cochrane.org/gradepro). Data Analysis The outcome measure of fracture was calculated in each study. All risk estimates were pooled using generic inverse variance weighting and a random effects (DerSimoneon and Laird) model. The Q test and I2 statistic were used to assess heterogeneity among studies. All pooled estimates were reported with associated 95% CIs. Subgroup analyses were performed to investigate the possible sources of heterogeneity. For 5-alpha reductase inhibitors, subgroup analyses of patients who had hip/femur fractures only were performed. For alpha-blockers, factors assessed were age of patient population included (studies including patients’ age 60 years or older and studies including younger patients), type of exposure (current exposure only or any exposure included) and type of fracture included in study (hip/femur fracture or any fracture). Publication bias was assessed by 2 formal tests: the Begg adjusted rank correlation test and the Egger regression asymmetry test. The implications were assessed by the fail safe N and the trim and fill methods. Statistical analysis was performed using Comprehensive Meta-analysis Version 2.21
464
RESULTS Figure 1 schematically shows the study selection process. Four hundred thirty-six articles were identified through database searches; 11 articles were fully evaluated. We found 7 studies (4 articles) that met our inclusion criteria. Four studies evaluated the fracture risk in alpha-blockers (1 case crossover study and 3 case-control studies) and 3 studies (case-control studies) evaluated fracture risk in 5-alpha reductase inhibitors. The summary of findings table can be found in Figure 6. Three studies for 5-alpha reductase inhibitors included in our meta-analysis involved 21,366 cases of fracture (40,803 controls). Four studies for alpha-blockers in our meta-analysis involved 22,051 cases of fracture (43,543 controls). The main characteristics of the studies included in the meta-analysis are summarized in Tables 1 and 2. Fractures With 5-Alpha Reductase Inhibitors Three studies evaluated the association between 5-alpha reductase inhibitors and the risk of fractures (Figure 2). Fivealpha reductase inhibitors use was not significantly associated with increased risk of fractures as compared with nonusers; the pooled OR was 0.9 (95% CI 5 0.7–1.1). Only moderate amount of heterogeneity was present among the studies (P for heterogeneity 5 0.159, I2 5 45.6%). Hip/Femur Fractures and 5-Alpha Reductase Inhibitors When the risk of hip/femur fracture with the use of 5-alpha reductase inhibitors was evaluated, although not statistically significant, there may be a protective effect for patients exposed to 5-alpha reductase inhibitors (Figure 3). The pooled OR was 0.8 (95% CI 5 0.7–1.0). There was less heterogeneity present as compared with analysis for risk of fracture with exposure to 5-alpha reductase inhibitors (P for heterogeneity 5 0.251, I2 5 27.7%). When compared with nonhip/femur fractures, there was a significant difference between the 2 groups (P 5 0.046 for Q test based on analysis of variance).
FIGURE 1. literature.
Schematic summary of a systematic review of the
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
TABLE 1. Summary of studies included for 5-alpha reductase inhibitors Study (reference Patient Fracture No. of No. of Region number) Study type setting evaluated cases controls United Kingdom Souverein et al12
Case-control Men aged 40 years or older
Hip/femur fracture
4,571
4,571
United States
Jacobsen et al14
Hip fracture
7,076
7,076
Denmark
Vestergaard et al16
Case-control Men aged 45 years or older Case-control Men aged 60 years or older
Any fracture
9,719
29,156
Adjustment Adjusted for NSAIDS, hyponotics/ anxiolytics, antipsychotics, antiparkinsonian drugs, glucocorticoids, bronchodilators, heart failure, diabetes mellitus, osteoporosis, prostate cancer and other LUTS Adjusted for age medical center, enrollment, race/ethnicity and comorbidity, alpha-blocker use Adjusted for previous fracture, ever use of corticosteroids, ever use of drugs against epilepsy, ever use of strong analgesics, ever use of weak analgesics, Charlson index, number of bed days in 1999, number of contacts to GP or specialist in 1999, alcoholism, income, working or not, living with someone or living alone, use of beta-blockers, use of calcium channel blockers, alpha-blockers as antihypertensives, loop diuretics, thiazide diuretics, potassium sparing and other types of diuretics, prostate cancer, orchiectomy and use of LHRH agonists
Study quality 9
8
6
GP, general specialist; LHRH, luteinizing hormone–releasing hormone; LUTS, lower urinary tract symptoms; NSAIDs, nonsteroidal antiinflammatory drugs.
Fractures With Alpha-Blockers Four studies evaluated the association between alphablockers and the risk of fractures (Figure 4). Alpha-blockers were not significantly associated with change in risk of fractures as compared with nonusers; the pooled OR was 1.1 (95% CI 5 0.9–1.3). Significant heterogeneity was present among studies (P for heterogeneity 5 0.0001, I2 5 86%). Subgroup Analyses for Alpha-Blockers To find possible sources of heterogeneity, subgroup analyses were performed to analyze the influence of age of patient population included, the definition of exposure to alphablockers (current exposure only versus any exposure) and the type of fracture included in study (Figure 5). There were 2 studies that included older men (60 years or older) only. The pooled OR showed no evidence of change in fracture risk in this group (OR 5 1.1; 95% CI 5 0.8–1.7). Significant heterogeneity was present (P 5 0.00002, I2 5 94.6%). For the 2 studies that included younger men (40– 45 years or older) and older men, no change in fracture risk was demonstrated as well (OR 5 1.1; 95% CI 5 1.0–1.2). No significant heterogeneity was present (P 5 0.876; I2 5 0%). There was, however, no evidence of difference between the 2 subgroups (P 5 0.81 for Q test based on analysis of variance). The pooled OR for studies that included only current exposure as cases showed an increase in fracture risk in this group (OR 5 1.38; 95% CI 5 1.17–1.63). No significant heterogeneity was present (P 5 0.484, I2 5 0%). For the 2 studies Ó 2013 Lippincott Williams & Wilkins
that included any exposure as cases, there was no evidence for a change in fracture risk (OR 5 1.00; 95% CI 5 0.87–1.15). Significant heterogeneity was present (P 5 0.005; I2 5 87.4%), and the 2 subgroups were significantly different (P 5 0.0004 for Q test based on analysis of variance). The pooled OR for studies that included only hip/femur fracture as cases showed no significant change in fracture risk in this group (OR 5 1.0; 95% CI 5 0.8–1.2). Significant heterogeneity was present (P 5 0.025, I2 5 72.8%). For studies/ subgroup that included other fractures (excluding hip/femur fracture), there was no evidence for a change in fracture risk (OR 5 0.9; 95% CI 5 0.8–1.1). The 2 subgroups were not significantly different (P 5 0.739 for Q test based on analysis of variance). Publication Bias For 5-alpha reductase inhibitors and the risk of fracture, both the Begg test (P . 0.05) and the Egger test (P . 0.1) found no evidence to indicate publication bias. The Orwin’s Fail safe N indicated that there would be 29 studies needed to reduce the OR to a trivial one. The trim and fill test did not show any adjustments. For alpha-blockers and the risk of fracture, both the Begg test (P . 0.05) and the Egger test (P . 0.1) found no evidence to indicate publication bias. The Orwin’s Fail safe N indicated that there would be 8 studies needed to reduce the OR to a trivial one. The trim and fill test did not show any adjustments.
465
Lim et al
TABLE 2. Summary of studies included for alpha-blockers Study (reference Patient Fracture Region number) Study type setting evaluated
No. of No. of cases controls
United Kingdom Souverein et al12
Case-control
Men aged 40 years or older
Hip/femur fracture
4,571
4,571
United States
Jacobsen et al14
Case-control
Hip fracture
7,076
7,076
Denmark
Vestergaard et al16
Case-control
Men aged 45 years or older Men aged 60 years or older
Any fracture
9,719
29,156
South Korea
Lee et al13
Case crossover Men aged 65 years or older
Any fracture
685
2740
Adjustment
Study quality
Adjusted for NSAIDS, hyponotics/anxiolytics, antipsychotics, antiparkinsonian drugs, glucocorticoids, bronchodilators, heart failure, diabetes mellitus, osteoporosis, prostate cancer and other LUTS Adjusted for age medical center, enrollment, race/ethnicity and comorbidity, alpha-blocker use Adjusted for previous fracture, ever use of corticosteroids, ever use of drugs against epilepsy, ever use of strong analgesics, ever use of weak analgesics, Charlson index, number of bed days in 1999, number of contacts to GP or specialist in 1999, alcoholism, income, working or not, living with someone or living alone, use of beta-blockers, use of calcium channel blockers, alpha-blockers as antihypertensives, loop diuretics, thiazide diuretics, potassium sparing and other types of diuretics, prostate cancer, orchiectomy and use of LHRH agonists No adjustment
9
7
6
8
GP, general specialist; LHRH, luteinizing hormone–releasing hormone; LUTS, lower urinary tract symptoms; NSAIDs, nonsteroidal antiinflammatory drugs.
Sensitivity Analysis To assess whether a single study had substantial influence on main results, we excluded each study and evaluated its effect on the summary estimates and heterogeneity
of the main analysis. Regarding analysis for 5-alpha reductase inhibitors for fractures and hip/femur fracture, exclusion of the study by Vestergaard et al18 led to the possibility of decreased risk of fractures (OR 5 0.7; 95% CI 5 0.6–0.9) and hip/femur
FIGURE 2. Odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving 5-alpha reductase inhibitors (5-ARI).
466
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
FIGURE 3. (5-ARI).
Odds ratios and 95% confidence intervals (CIs) of hip/femur fracture for patients receiving 5-alpha reductase inhibitors
fracture (OR 5 0.7; 95% CI 5 0.6–0.9). For alpha-blockers, we did not find any major changes in direction or magnitude of summary estimates and P values of heterogeneity.
DISCUSSION
In our meta-analysis, we did not find evidence to support the hypothesis that 5-alpha reductase inhibitors were associated with a change in overall fracture risk. Five-alpha reductase inhibitors may be associated with decreased hip/femur fracture risk, but this did not reach statistical significance. Although there was insufficient evidence to support the hypothesis that alpha-blockers are associated with a change in fracture risk, there was significant heterogeneity among the studies. Subgroup analysis showed that studies that defined exposure as current (ie, patient were taking alpha-blockers when they had the fracture) showed an increased risk for fracture. This subgroup was significantly different from the studies that defined exposure as any exposure (previous exposure and current exposure). A review of the literature showed that bone mineral density is unchanged in patients on finasteride (follow-up time of 4 years and 1 year, respectively)22,23 and in men with 5-alpha reductase insufficiency.24 In a short-term study (3 months), bone turnover markers are not affected by finasteride.25 Therapy with transdermal DHT (3 months) did not affect osteocalcin, a bone turnover marker, when used to treat osteoporotic men.26 The results of our meta-analysis are consistent with these stud-
ies that 5-alpha reductase inhibitors have little effect on bone turnover and, therefore, do not affect fracture risk. Testosterone has an important role in the development of benign prostatic hyperplasia. The principal prostatic androgen is DHT, which causes hyperproliferation of both stromal and epithelial cells in the prostate. Furthermore, complex interactions between cytokines and growth factors from various sources, such as inflammatory cell infiltrates, contribute to the development of benign prostatic hyperplasia.27 Testosterone also has an important role in maintenance of bone health in men. Hypogonadal men are at increased risk for the development of fractures, and supplementation with testosterone decreases this risk.28 The effect of testosterone on bone health is mediated through both androgen and estrogen receptors. Current evidence suggests that estradiol, through the estrogen receptor and aromatization of androgens into estrogen, may have a greater role in maintenance of skeletal health in men than testosterone acting directly through the androgen receptor.8,28 A possible explanation for the unchanged fracture risk with the use of 5-alpha reductase inhibitors is that decreased serum levels of DHT caused by 5-alpha reductase inhibitors may be compensated by increased testosterone levels and shunting of testosterone to estrogen in bone tissue. However, more studies are required to confirm this hypothesis. Also, it is has been found that human osteoblasts predominantly express 5-alpha reductase type 1, although finasteride inhibits 5-alpha reductase type 2.29 Therefore, there may be local DHT
FIGURE 4. Odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving alpha-blockers (AB). Ó 2013 Lippincott Williams & Wilkins
467
Lim et al
FIGURE 5. Subgroup analysis: odds ratios and 95% confidence intervals (CIs) of fracture for patients receiving alpha-blockers (AB). (A) Comparing studies, including older patients only (aged 60 years or older) compared with studies, including younger patients (40–45 years or older). (B) Comparing studies defining exposure as current compared with studies defining exposure as any exposure (current or previous). (C) Comparing studies, including hip/femur fracture compared with studies, including other fracture (excluding hip/femur fracture).
468
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
FIGURE 6. Summary of findings table: use of 5-alpha reductase inhibitors and alpha-blockers in benign prostatic hyperplasia and fracture risk.
production in the skeletal system when finasteride is used to treat benign prostatic hyperplasia. Our meta-analysis reinforces the fact that 5-alpha reductase inhibitors have a relatively benign adverse effect profile30 and do not seem to increase fracture risk. However, caution is still needed with 5-alpha reductase inhibitors, especially in patients who are at high risk for osteoporosis, as most studies that evaluated their effects were of relatively short duration and involved small sample sizes.23,25,26 Also, our meta-analysis had a small number of studies. Subgroup analysis for hip/femur fractures and sensitivity analysis, excluding the study by Vestergaard et al,18 showed that 5-alpha reductase inhibitors may be associated with decreased risk of fracture. Whether there is a true protective effect of 5-alpha reductase inhibitors requires more research and validation because this has important implications. All studies we analyzed did not include data on the effects of treatment on lower urinary tract symptoms. Inclusion of these data in future studies may be helpful as it would determine whether treatment leading to a decrease in nocturia or lower urinary tract symptoms, both known risk factors for fractures,31,32 could decrease fracture risk. In subgroup analysis, decreased heterogeneity was observed when only hip/femur fractures were analyzed, and this was statistically different from the subgroup comprised of other fractures (hip/femur fractures excluded). There may be differences between hip/femur fractures and other fractures. For example, vertebral fractures as compared with hip fractures are more difficult to assess in observational studies because the definition of a significant vertebral fracture and the clinical presentation attributable to vertebral deformity vary.33 Information about hip fractures is more readily available and easier to assess as virtually all hip fractures are treated during a hospital admission, making these fractures relatively easy to detect using Ó 2013 Lippincott Williams & Wilkins
discharge registries.34 Forearm fractures as compared with hip fractures have different risk profiles35; therefore, besides the effect of finasteride on bone strength, there may be other factors involved in the development of fracture at a particular site. The fact that alpha-blockers were not associated with any change in fracture risk needs to be interpreted with caution as significant heterogeneity was present in this meta-analysis. Studies that defined exposure to alpha-blockers as current did demonstrate increased fracture risk as compared with studies that defined exposure as any exposure. This may represent selection bias in which studies that defined exposure as any exposure may have included patients who are not currently on alpha-blockers. Therefore, the risk of fracture may have been underestimated in those studies. Caution is still warranted when prescribing alpha-blockers. Most studies analyzed in our meta-analysis did not provide sufficient data about the blood pressure effects of these medications when used to treat BPH. It is also difficult to infer cause and effect and to determine the timing of initiation of alpha-blockers and the onset of fracture from case-control studies. However, other studies suggest possible mechanisms for increased risk of fracture when alpha-blockers are used. The initiation of antihypertensives in older patients increases fracture risk, partially from associated orthostatic hypotension.36 Alpha-blockers are also associated with an increased risk of vascular events that could precipitate osteoporotic fractures related to falls.15,37 Bone protective mechanisms associated with alpha-blockers could compensate for the increased risk of vascular events. These include the induction of osteoblastic stem cell differentiation by alpha-blockers.38 Blood pressure–lowering effects of alpha-blockers may also decrease urinary calcium excretion.39 Finally, improvement of nocturia and lower urinary tract symptoms could decrease the risk of fracture.
469
Lim et al
There were several limitations in our meta-analysis. First, the number of studies in our meta-analysis was small; therefore, statistical analysis to test for publication bias may be underpowered. To compensate this, a manual search through references in articles reviewed and abstracts from scientific meetings (American College of Rheumatology—2009–2012 and Endocrine Society—2010–2012) was performed to identify unpublished studies. Second, significant heterogeneity was found in the main analyses for alpha-blockers. This could be because of the differences in study design, the definition of exposure, the dose and duration of medications used, and other factors. A random-effects model was used to partly account for differences among studies. Third, the occurrence of osteoporotic fractures may be influenced by common comorbidities in older men, such as hypogonadism, and current medications. Therefore, we tried to include adjusted ORs in our meta-analysis. However, the studies analyzed did not always adjust for the same factors. Also, residual confounding (some confounding factors were not considered/adjusted for or insufficient data collected for certain confounding factors to be adjusted) may limit our analysis. A small number of studies also limited the ability to perform subgroup analyses to evaluate for heterogeneity and to evaluate the effect of different alpha-blockers and 5-alpha reductase inhibitors. Whether there was a dose-response relationship for both drugs could not be determined because studies did not provide information on uniform dosages for evaluation. Limited data on the fracture risk of different age groups and the degree of lower urinary tract symptoms also precluded important analyses of possible causes that may explain the effects of these medications on fracture risk. There have been no studies evaluating combined therapy and fracture risk. Finally, there remains a lot to be learned about the pathophysiology of benign prostatic hyperplasia, the systemic effects of these medications and the complex interactions of calcium, vitamin D, BPH and fractures.
CONCLUSIONS In summary, our meta-analysis provided no evidence that 5-alpha reductase inhibitors and alpha-blockers were associated with any change in fracture risk. It is a reassuring that 5-alpha reductase inhibitors do not increase fracture risk and may even have protective effects. However, more studies are needed to confirm this. Although alpha-blockers were not associated with change in fracture risk, caution is required because significant heterogeneity was present in between the studies. More studies are needed to evaluate these drugs for fracture risk, especially during initiation of alpha-blocker therapy. REFERENCES 1. Lim LS, Hoeksema LJ, Sherin K, et al. Screening for osteoporosis in the adult U.S. population: ACPM position statement on preventive practice. Am J Prev Med 2009;36:366–75. 2. Looker AC, Melton LJ 3rd, Harris TB, et al. Prevalence and trends in low femur bone density among older US adults: NHANES 2005-2006 compared with NHANES III. J Bone Miner Res 2010;25:64–71. 3. Burge R, Dawson-Hughes B, Solomon DH, et al. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. J Bone Miner Res 2007;22:465–75. 4. Issa MM, Fenter TC, Black L, et al. An assessment of the diagnosed prevalence of diseases in men 50 years of age or older. Am J Manag Care 2006;12(suppl 4):S83–9. 5. 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.
470
6. 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. 7. Kasperk CH, Wergedal JE, Farley JR, et al. Androgens directly stimulate proliferation of bone cells in vitro. Endocrinology 1989;124:1576–8. 8. Clarke BL, Khosla S. Androgens and bone. Steroids 2009;74:296–305. 9. Castro-Magana M, Angulo M, Fuentes B, et al. Effect of finasteride on human testicular steroidogenesis. J Androl 1996;17:516–21. 10. Kruep EJ, Goodwin BB, Chaudhari S. Evaluation of recent trends in treatment patterns among men with benign prostatic hyperplasia. Am J Mens Health 2013;7:214–9. 11. Vanderschueren D, Vandenput L, Boonen S, et al. Androgens and bone. Endocr Rev 2004;25:389–425. 12. Souverein PC, Van Staa TP, Egberts AC, et al. Use of alpha-blockers and the risk of hip/femur fractures. J Intern Med 2003;254:548–54. 13. Lee J, Choi NK, Jung SY, et al. The risk of fracture with taking alpha blockers for treating benign prostatic hyperplasia [in Korean]. J Prev Med Public Health 2009;42:165–70. 14. Jacobsen SJ, Cheetham TC, Haque R, et al. Association between 5-alpha reductase inhibition and risk of hip fracture. JAMA 2008;300:1660–4. 15. Nickel JC, Sander S, Moon TD. A meta-analysis of the vascular-related safety profile and efficacy of alpha-adrenergic blockers for symptoms related to benign prostatic hyperplasia. Int J Clin Pract 2008;62:1547–59. 16. Radtke A, Lempert T, von Brevern M, et al. Prevalence and complications of orthostatic dizziness in the general population. Clin Auton Res 2011;21:161–8. 17. Cummings SR, Nevitt MC. Nonskeletal determinants of fractures: the potential importance of the mechanics of falls. Study of Osteoporotic Fractures Research Group. Osteoporos Int 1994;4(suppl 1):67–70. 18. Vestergaard P, Rejnmark L, Mosekilde L. Risk of fractures associated with treatment for benign prostate hyperplasia in men. Osteoporos Int 2011;22:731–7. 19. Wiens M, Etminan M, Gill SS, et al. Effects of antihypertensive drug treatments on fracture outcomes: a meta-analysis of observational studies. J Intern Med 2006;260:350–62. 20. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336:924–6. 21. Borenstein M, Hedges L, Higgins J, et al. Comprehensive Meta-Analysis. Englewood, NJ: Biostat; 2005. 22. Matsumoto AM, Tenover L, McClung M, et al. The long-term effect of specific type II 5alpha-reductase inhibition with finasteride on bone mineral density in men: results of a 4-year placebo controlled trial. J Urol 2002;167:2105–8. 23. Matzkin H, Chen J, Weisman Y, et al. Prolonged treatment with finasteride (a 5 alpha-reductase inhibitor) does not affect bone density and metabolism. Clin Endocrinol (Oxf) 1992;37:432–6. 24. Sobel V, Schwartz B, Zhu YS, et al. Bone mineral density in the complete androgen insensitivity and 5alpha-reductase-2 deficiency syndromes. J Clin Endocrinol Metab 2006;91:3017–23. 25. Tollin SR, Rosen HN, Zurowski K, et al. Finasteride therapy does not alter bone turnover in men with benign prostatic hyperplasia—a Clinical Research Center study. J Clin Endocrinol Metab 1996;81:1031–4. 26. Ly LP, Jimenez M, Zhuang TN, et al. A double-blind, placebo-controlled, randomized clinical trial of transdermal dihydrotestosterone gel on muscular strength, mobility, and quality of life in older men with partial androgen deficiency. J Clin Endocrinol Metab 2001;86:4078–88.
Volume 347, Number 6, June 2014
Medical Management of BPH and Fractures
27. Wein AJ, Kavoussi LR, Partin AW, et al. Campbell-Walsh urology. Philadelphia, PA: Elsevier Saunders; 2012.
34. Marcus R, Feldman D, Kelsey JL. Osteoporosis. San Diego, London: Academic Press; 2001.
28. Sinnesael M, Boonen S, Claessens F, et al. Testosterone and the male skeleton: a dual mode of action. J Osteoporos 2011;2011:240328.
35. Graafmans WC, Ooms ME, Bezemer PD, et al. Different risk profiles for hip fractures and distal forearm fractures: a prospective study. Osteoporos Int 1996;6:427–31.
29. Issa S, Schnabel D, Feix M, et al. Human osteoblast-like cells express predominantly steroid 5alpha-reductase type 1. J Clin Endocrinol Metab 2002;87:5401–7. 30. Leedahl DD, Vo PH, Maxson PM, et al. Benign prostatic hyperplasia: implications for pharmacologic treatment and perioperative care. J Pharm Pract 2012; Epub ahead of print. 31. Parsons JK, Mougey J, Lambert L, et al. Lower urinary tract symptoms increase the risk of falls in older men. BJU Int. 2009;104:63–8.
36. Butt DA, Mamdani M, Austin PC, et al. The risk of falls on initiation of antihypertensive drugs in the elderly. Osteoporos Int 2013;24:2649– 57. 37. Morrison A, Fan T, Sen SS, et al. Epidemiology of falls and osteoporotic fractures: a systematic review. Clinicoecon Outcomes Res 2013; 5:9–18.
32. Temml C, Ponholzer A, Gutjahr G, et al. Nocturia is an age-independent risk factor for hip-fractures in men. Neurourol Urodyn 2009;28:949–52.
38. Choi YJ, Lee JY, Lee SJ, et al. Alpha-adrenergic blocker mediated osteoblastic stem cell differentiation. Biochem Biophys Res Commun 2011;416:232–8.
33. Cooper C, Cole ZA, Holroyd CR, et al. Secular trends in the incidence of hip and other osteoporotic fractures. Osteoporos Int 2011;22:1277–88.
39. Tsuda K, Nishio I, Masuyama Y. Bone mineral density in women with essential hypertension. Am J Hypertens 2001;14:704–7.
Ó 2013 Lippincott Williams & Wilkins
471