Abstract
Introduction
angiotensin II, a major effector protein of the renin angiotensin system (RAS), induces bone loss under certain conditions. Drugs that block the RAS may therefore reduce bone loss and fracture incidence. The fracture incidence in older hypertensive men with long-term use of angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) were compared with the incidence in users of calcium channel blockers (CCBs) and non-users.
Methods
a total of 5,994 US men aged 65 years or older who had bone mineral density measured at baseline in the Osteoporotic Fractures in Men Study (MrOS) were followed for fracture incidence for an average of 6.8 years. Men with follow-up dual-energy X-ray absorptiometry bone mineral density data and who reported hypertension at any visit, or use of antihypertensive medications at any visit among those with non-missing mediation data were included in the study (N = 2,573).
Results
six hundred and nineteen men had taken ACE inhibitors, while 182 took ARBs for at least 4 years. Using Cox regression for the incidence of non-vertebral fractures, we found that long-term users of ACE inhibitors and ARBs each had a significantly lower fracture incidence than non-users. The hazard ratio of non-vertebral fractures was three times lower in ARB users than ACE inhibitor users (Hazard ratio (95% confidence interval): 0.194 (0.079–0.474) versus 0.620 (0.453–0.850), P = 0.0168). There was a trend of greater fracture risk reduction with longer duration of ARB use, but not for ACE inhibitor use.
Conclusions
in older hypertensive men, ARBs use was associated with lower incidence of non-vertebral fracture than ACE inhibitors or CCBs.
Keywords: older people, angiotensin-converting enzyme inhibitor, angiotensin receptor blocker, fracture, hypertension, male, aged
Introduction
Hypertension and osteoporosis commonly coexist in the elderly. One common pathophysiological process is the renin angiotensin system (RAS), which reportedly can lead to bone loss by osteoclast activation [1]. RAS blockers may therefore prevent bone loss and fracture in older hypertensive people [2]. However, in the Osteoporotic Fractures in Men Study (MrOS) prospective cohort study of older American men, long-term use of an angiotensin-converting enzyme (ACE) inhibitor, but not angiotensin receptor blocker (ARB), was associated with a small increase in the rate of bone loss over a period of 4 years [3]. In a retrospective analysis of a large health insurance database in the USA, the sole use of ARB for newly diagnosed hypertensive patients was associated with a significantly lower fracture risk in the first year, while an ACE inhibitor had no significant protective effect [4]. These results suggest that ARBs may be superior to ACE inhibitors for fracture prevention in older hypertensive people.
ACE inhibitors have only a limited ability to block RAS because of the feedback increase in renin. However, ARBs block the angiotensin II receptor type 1 (AT1) receptor directly and do not have this limitation. Nevertheless, they have exhibited inconsistent effects on bone loss in different animal models of osteoporosis [1, 5, 6]. The results of these animal studies cannot be directly extrapolated to older people with hypertension, as their hormonal milieu may be different and may also be altered by concomitant drug use.
To ascertain whether ARBs are superior to ACE inhibitors in maintaining bone health in older people, we used the 6-year prospective data of the MrOS study to compare the fracture incidence among older hypertensive men receiving ACE inhibitors or ARBs. We hypothesised that ARB use, when compared with ACE inhibitor use, would be associated with a lower fracture incidence.
Materials and methods
Participants
From March 2000 to April 2002, men aged ≥65 years participated in the baseline examination for the prospective cohort study of risk factors in the MrOS [7]. MrOS was a multicentre study; the design, recruitment methods and measurements have been described previously [8]. Briefly, 5,994 men were recruited from six regions of the USA: Birmingham, AL; Minneapolis, MN; the Monongahela Valley near Pittsburgh, PA; Palo Alto, CA; Portland, OR; and San Diego, CA. All participants were invited to participate in follow-up studies. Among the 5,994 MrOS men enroled at baseline, 5,229 returned at the 4-year follow-up visit. Among them, only 3,633 subjects with hypertension were included. After excluding use of osteoporosis medication, androgen or anti-androgen and missing medication or bone mineral density (BMD) follow-up data, 2,573 men was used in these analyses (Figure 1). Written informed consent was obtained from all participants.
Figure 1.
Flow chart.
Measurements
Demographic information, smoking habits, personal medical history, including any history of hypertension, diabetes mellitus or cardiac failure, as well as medication use, were obtained by self-administered questionnaire or face-to-face interviews performed by certificated research assistants [9].
At each visit, participants were asked to bring all medications they had taken in the past 30 days to the clinical centre. Medications were categorised by using the original Established Populations for Epidemiologic Studies of the Elderly coding system [10]. Only prescription medications were included at baseline. Use of calcium and vitamin D supplements was obtained from a modified food frequency questionnaire developed specifically for MrOS by Block Dietary Data Systems. At the 2-, 4-, and 6-year follow-up visits, both prescribed and over-the-counter medications were recorded. Hypertensive subjects were defined as those who reported a history of hypertension or the use of antihypertensive drugs during any visit. Long-term users were defined as those on drugs for three visits or more while short-term users were defined as those on drugs for only one visit.
Body weight was measured using a balance beam or electronic scale with the participants wearing indoor clothing without shoes. The scale at each clinical centre was calibrated monthly.
Morning fasting blood was obtained and frozen. The estimated glomerular filtration rate (eGFR) was calculated from the following Modification of Diet in Renal Disease (MDRD) study equation [11]: eGFR (mL/min/1.73 m2) = 175 × SCr (mg/dL)(−1.154) × Age(−0.203) × 1.212 (if race = African American), in which an eGFR <60 mL/min/1.73 m2 is indicative of moderate (stage 3) chronic kidney disease. We used this cut-off value of eGFR to define chronic kidney disease.
BMD of the total hip, femoral neck and trochanter was measured at each visit using dual-energy X-ray absorptiometry (DXA) with a Hologic QDR 4500 bone densitometer (Hologic, Waltham, MA, USA). BMD data at baseline and year 6 were used. All measurements were performed on the right hip unless the participant had a right hip replacement or other metal internal fixation materials. The inter-scanner coefficient of variation was 0.9%.
Statistical analysis
The users of the ACE inhibitors, ARBs or calcium channel blockers (CCBs) were compared with non-users of any of these drugs, individually in terms of clinical characteristics, using ANOVA for continuous variables or the chi-square test for categorical variables. ANCOVA was adopted for the baseline BMD, adjusting for age and weight. Fracture was analysed using the Cox regression model.
The duration of the use of the ACE inhibitors, ARB and CCB were analysed for non-vertebral and hip or forearm fractures separately, and each of them was further analysed in three models. In the first model, known confounders including age, tricyclic antidepressant use, previous fracture, inability to complete a narrow walk, falls in the previous year and depressed mood [12] were adjusted in a Cox regression model. In the second model, total hip BMD was added. In the final model, propensity scores of the ACE inhibitors ARB and CCB were further added independently. Propensity scores indicate the likelihood of a participant's use of either of the ACE inhibitors, ARB or CCB. Significant factors for the use of ACE inhibitors, ARB and CCB were determined by stepwise logistic regression. A generalised logits model was applied, when the proportional odds assumption was not satisfied [13].
All statistical analyses were performed using the statistical package SAS, version 9.2 (SAS Institute, Inc., Cary, NC, USA). An alpha level of 5% was used as the level of significance.
Power analysis
For binary factors, by using Cox regression with 10% fracture rate, 0.05 alpha level and 80% power, 2,573 subjects can detect effect size as small as 0.35 (log hazard ratio, equivalent to Hazard ratio = 1.4 or 0.7).
Results
There were 2,573 men with hypertension among all the MrOS participants. The clinical characteristics, concomitant drug use, BMD, BMD changes and the incidence of non-vertebral fractures in participants who took an ACE inhibitor, ARB or CCB along with non-users are shown in Table 1. The non-users primarily used alpha or beta blockers, but 50% of them had not taken any antihypertensive medication for three visits or more. When compared with non-users, long-term users of ACE inhibitors or ARBs were heavier, more likely to have diabetes, heart failure and/or chronic kidney disease. The clinical characteristics and concomitant drug use in users of the ACE inhibitors, ARBs and CCBs were not significantly different from each other.
Table 1.
Fracture rates between users of ACE inhibitor, ARB, CCB and non-users (subject with hypertension)
| Mean (SD), N (%) or N/rate (per 100,000 person-years) | Others | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ACE inhibitor use | ARB use | CCB use | ||||||||
| ≥3 visits | 2 visits | 1 visit | ≥3 visits | 2 visits | 1 visit | ≥3 visits | 2 visits | 1 visit | ||
| Number | 619 | 276 | 342 | 182 | 147 | 171 | 352 | 188 | 234 | 798 |
| Age (year) | 72.8 (5.2) | 73.2 (5.3)* | 72.8 (5.7) | 72.3 (5.4) | 72.7 (5.2) | 72.8 (5.0) | 73.1 (5.4)* | 73.3 (5.2)* | 73.7 (5.6)* | 72.3 (5.3) |
| Body weight (kg) | 86.5 (13.7)* | 84.8 (12.3) | 84.5 (12.7) | 86.3 (12.9)* | 87.4 (13.9)* | 85.2 (12.5) | 85.9 (12.7)* | 84.7 (13.3) | 83.9 (13.8) | 83.7 (12.8) |
| Diabetes mellitus, N (%) | 136 (22%)* | 41 (15%)* | 32 (9%)* | 29 (16%)* | 23 (16%)* | 15 (9%)* | 52 (15%)* | 25 (13%)* | 24 (10%)* | 22 (3%) |
| Cardiac failure, N (%) | 47 (8%)* | 13 (5%)* | 14 (4%)* | 15 (8%)* | 8 (5%)* | 5 (3%) | 15 (4%)* | 13 (7%)* | 10 (4%)* | 12 (2%) |
| Hypertension, N (%) | 494 (80%)* | 159 (58%)* | 157 (46%)* | 152 (84%)* | 95 (65%)* | 92 (54%)* | 300 (85%)* | 142 (76%)* | 150 (64%)* | 201 (25%) |
| Chronic kidney diseasea, N (%) | 103 (18%)* | 36 (14%) | 50 (16%) | 37 (22%)* | 20 (15%) | 28 (17%) | 49 (15%) | 39 (22%)* | 66 (31%)* | 88 (12%) |
| Fracture at or after age 50 | 87 (14%) | 51 (18%) | 47 (14%) | 24 (13%) | 15 (10%) | 19 (11%) | 61 (17%) | 24 (13%) | 28 (12%) | 125 (16%) |
| Inability to complete narrow walk trial | 40 (6%) | 18 (7%) | 16 (5%) | 8 (4%) | 10 (7%) | 8 (5%) | 17 (5%) | 14 (7%) | 17 (7%) | 37 (5%) |
| Falls in past year | 123 (20%) | 63 (23%) | 63 (18%) | 31 (17%) | 27 (18%) | 31 (18%) | 75 (21%) | 46 (24%) | 45 (19%) | 148 (19%) |
| Drug useb, N (%) | ||||||||||
| Thiazide | 124 (20%)* | 38 (14%)* | 42 (12%)* | 48 (26%)* | 30 (20%)* | 25 (15%)* | 68 (19%)* | 36 (19%)* | 46 (20%)* | 55 (7%) |
| Loop diuretic | 53 (9%)* | 8 (3%)* | 11 (3%)* | 14 (8%)* | 6 (4%)* | 9 (5%)* | 22 (6%)* | 14 (7%)* | 9 (4%)* | 8 (1%) |
| Nitrate | 27 (4%)* | 8 (3%) | 8 (2%) | 6 (3%) | 4 (3%) | 7 (4%)* | 18 (5%)* | 6 (3%) | 8 (3%) | 12 (2%) |
| Statin | 313 (51%)* | 116 (42%)* | 104 (30%) | 89 (49%)* | 65 (44%)* | 59 (35%) | 168 (48%)* | 72 (38%)* | 93 (40%)* | 220 (28%) |
| Beta blocker | 187 (30%)* | 75 (27%) | 80 (23%) | 54 (30%)* | 41 (28%) | 40 (23%) | 96 (27%) | 49 (26%) | 79 (34%)* | 171 (21%) |
| Alpha blocker | 105 (17%) | 41 (15%)* | 52 (15%)* | 33 (18%) | 22 (15%) | 22 (13%)* | 75 (21%) | 30 (16%) | 27 (12%)* | 161 (20%) |
| Tricyclic antidepressant | 5 (1%) | 2 (1%) | 0 (0%) | 0 (0%) | 2 (1%) | 0 (0%) | 2 (1%) | 1 (1%) | 1 (0%) | 6 (1%) |
| Vitamin D | 253 (41%) | 130 (47%) | 136 (40%) | 93 (51%) | 53 (36%)* | 75 (44%) | 145 (41%) | 81 (43%) | 90 (38%) | 364 (46%) |
| Calcium | 116 (19%) | 48 (17%) | 55 (16%)* | 34 (19%) | 21 (14%)* | 35 (20%) | 64 (18%) | 24 (13%)* | 45 (19%) | 177 (22%) |
| ACE Inhibitor | 619 (100%)* | 0 (0%) | 0 (0%) | 4 (2%)* | 8 (5%)* | 24 (14%)* | 128 (36%)* | 53 (28%)* | 55 (24%)* | 0 (0%) |
| ARB | 4 (1%)* | 3 (1%)* | 33 (10%)* | 182 (100%)* | 0 (0%) | 0 (0%) | 35 (10%)* | 19 (10%)* | 25 (11%)* | 0 (0%) |
| CCB | 128 (21%)* | 29 (11%)* | 43 (13%)* | 35 (19%)* | 28 (19%)* | 34 (20%)* | 352 (100%)* | 0 (0%) | 0 (0%) | 0 (0%) |
| Baseline BMD (g/cm2) | ||||||||||
| Total hip | 0.99 (0.14)*,** | 0.95 (0.13)** | 0.98 (0.12) | 0.99 (0.15) | 1.01 (0.13)*,** | 0.99 (0.14) | 0.99 (0.13)* | 0.98 (0.13) | 0.97 (0.13) | 0.97 (0.12) |
| Femoral neck | 0.82 (0.13)*,** | 0.78 (0.12) | 0.81 (0.12)*,** | 0.81 (0.13)* | 0.83 (0.13)*,** | 0.81 (0.12)*,** | 0.81 (0.13)*,** | 0.81 (0.12) | 0.79 (0.12) | 0.79 (0.12) |
| Trochanteric | 0.79 (0.13)*,** | 0.76 (0.12)*,** | 0.79 (0.12) | 0.79 (0.14) | 0.81 (0.11)*,** | 0.79 (0.13) | 0.79 (0.12) | 0.78 (0.12) | 0.77 (0.12) | 0.77 (0.11) |
| Total hip BMD % change per year | −0.48 (0.74)*,** | −0.53 (0.84)*,** | −0.39 (0.72) | −0.38 (0.68) | −0.31 (0.61) | −0.45 (0.75) | −0.48 (0.69)*,** | −0.54 (0.83)*,** | −0.42 (0.75) | −0.35 (0.65) |
| Fracture, N/rate (per 100, 000 person- year) | ||||||||||
| Non-spine | 55/977.21* | 34/1,362.12 | 30/950.02* | 5/288.65* | 7/514.11* | 16/1,085.18 | 25/785.18* | 21/1,261.23 | 39/1,772.28 | 133/1,756.07 |
| Hip fracture | 12/201.03 | 8/297.32 | 4/120.80 | 3/169.63 | 0/0.00 | 3/182.48 | 6/177.45 | 5/276.71 | 10/441.75* | 10/128.02 |
| Hip or wrist fracture | 19/321.12 | 12/449.27 | 7/212.41 | 3/169.63 | 2/139.75 | 3/183.56 | 9/269.62 | 7/391.20 | 15/665.11* | 23/295.75 |
*P < 0.05, when compared with non-users, using ANOVA for continuous variables, chi-square test for categorical variables and Cox regression for fractures; **P < 0.05, when compared with non-users, using ANCOVA, adjusting for age and weight.
aDefined by eGFR, 182 missing data.
bMedication use ≥ 3 visits.
Long-term users of an ACE inhibitor, ARB or CCB had slightly higher BMD in the femoral neck at baseline when compared with non-users. The use of either an ACE inhibitor or CCB for two or more visits was associated with greater hip BMD loss. However, no ARB use of any duration was associated with increased bone loss when compared with non-users. The unadjusted incidence rates of non-vertebral fractures of the long-term use of an ACE inhibitor, ARB or CCB were significantly lower than those of non-users.
The long-term use of an ACE inhibitor or ARB was associated with a lower risk of non-vertebral fracture when compared with non-users after adjustment for BMD, clinical risk factors and propensity score (Table 2). There was a significant trend for a greater rate reduction of non-vertebral fractures with a longer period of use among ARB users, but not ACE inhibitor users. Long-term ARB use had a significantly lower (three times lower) risk of non-vertebral fracture when compared with long-term users of an ACE inhibitor or CCB. The use of an ACE inhibitor, ARB or CCB for two visits or more was not significantly associated with hip or forearm fracture incidence. Short-term use of CCB was associated with a greater risk of hip or forearm fracture compared with either non-users or short-term ACE inhibitor users.
Table 2.
The adjusted hazard ratio of fracture in users of ACE inhibitor and ARB (subject with hypertension)
| Site | Hazard ratio of fracture (95% confidence interval) | P-value (Trend across ≥3, 2 and 1 visit) | |||
|---|---|---|---|---|---|
| ACE inhibitor use | |||||
| ≥3 visits (N = 619) | 2 visits (N = 276) | 1 visit (N = 342) | Non-user (N = 1,328) | ||
| Non-spine fracture | |||||
| Multivariable adjusteda | 0.698 (0.515–0.947) | 0.881 (0.610–1.273) | 0.653 (0.444–0.961) | Reference | 0.7027 |
| Multivariablea + hip BMD adjusted | 0.723 (0.534–0.980) | 0.832 (0.575–1.202) | 0.668 (0.454–0.984) | Reference | 0.8126 |
| Multivariablea + hip BMD + propensity score of ACE inhibitor adjustedb | 0.620 (0.453–0.850) | 0.718 (0.494–1.044) | 0.625 (0.423–0.923) | Reference | 0.3778 |
| Hip or wrist fracture | |||||
| Multivariable adjusteda | 1.064 (0.613–1.846) | 1.322 (0.690–2.532) | 0.668 (0.299–1.494) | Reference | 0.4251 |
| Multivariablea + hip BMD adjusted | 1.120 (0.646–1.944) | 1.189 (0.619–2.285) | 0.685 (0.306–1.533) | Reference | 0.3457 |
| Multivariablea + hip BMD + propensity score of ACE inhibitor adjustedb | 0.978 (0.550–1.736) | 1.181 (0.607–2.297) | 0.670 (0.297–1.511) | Reference | 0.7554 |
| ARB use | |||||
| ≥3 visits (N = 182) | 2 visits (N = 147) | 1 visit (N = 171) | Non-user (N = 2,065) | ||
| Non-spine fracture | |||||
| Multivariable adjusteda | 0.220 (0.091–0.534)1 | 0.389 (0.183–0.826)4 | 0.795 (0.479–1.318) | Reference | 0.0086 |
| Multivariablea + hip BMD adjusted | 0.224 (0.092–0.545)2 | 0.421 (0.198–0.893)5 | 0.816 (0.492–1.353) | Reference | 0.0083 |
| Multivariablea + hip BMD + propensity score of ARB adjustedc | 0.194 (0.079–0.474)3 | 0.385 (0.180–0.822) | 0.810 (0.484–1.354) | Reference | 0.0094 |
| Hip or wrist fracture | |||||
| Multivariable adjusteda | 0.532 (0.166–1.698) | 0.440 (0.108–1.803) | 0.536 (0.168–1.707) | Reference | 0.8576 |
| Multivariablea + hip BMD adjusted | 0.539 (0.168–1.724) | 0.503 (0.123–2.062) | 0.548 (0.172–1.749) | Reference | 0.8536 |
| Multivariablea + hip BMD + propensity score of ARB adjustedc | 0.543 (0.166–1.778) | 0.492 (0.119–2.042) | 0.567 (0.175–1.839) | Reference | 0.7621 |
| CCB use | |||||
| ≥3 visits (N = 352) | 2 visits (N = 188) | 1 visit (N = 234) | Non-user (N = 1,791) | ||
| Non-spine fracture | |||||
| Multivariable adjusteda | 0.589 (0.388–0.893)6 | 0.991 (0.631–1.554) | 1.378 (0.976–1.946) | Reference | 0.0017 |
| Multivariablea + hip BMD adjusted | 0.596 (0.393–0.904)7 | 1.002 (0.639–1.572) | 1.336 (0.945–1.889) | Reference | 0.0028 |
| Multivariablea + hip BMD + propensity score of Calcium channel blocker adjustedd | 0.660 (0.426–1.022)8 | 1.157 (0.725–1.846) | 1.421 (0.984–2.052) | Reference | 0.0058 |
| Hip or wrist fracture | |||||
| Multivariable adjusteda | 0.966 (0.471–1.983) | 1.421 (0.640–3.157) | 2.371 (1.312–4.286)9 | Reference | 0.1426 |
| Multivariablea + hip BMD adjusted | 0.980 (0.477–2.012) | 1.518 (0.683–3.372) | 2.179 (1.196–3.969)10 | Reference | 0.1944 |
| Multivariablea + hip BMD + propensity score of Calcium channel blocker adjustedd | 1.146 (0.550–2.385) | 1.445 (0.603–3.465) | 2.422 (1.275–4.604) 11 | Reference | 0.1655 |
aCovariates included age, tricyclic antidepressant use, thiazide use, previous fracture, inability to complete a narrow walk trial, falls in previous year and depressed mood.
bPropensity score of ACE inhibitor was calculated by diabetes mellitus, cardiac failure, hypertension, duration of use of loop diuretic, statin, beta blocker and ARB.
cPropensity score of ARB was calculated by diabetes mellitus, cardiac failure, hypertension, duration of use of loop diuretic, statin, beta blocker, calcium antagonist and ACE inhibitor.
dPropensity score of calcium channel blocker was calculated by hypertension, chronic kidney disease, smoking status, duration of use of loop diuretic, nitrate, beta blocker, ACE inhibitor and ARB.
1P-value = 0.0171,
2P-value = 0.0140 and
3P-value = 0.0168 comparing use of ARB with ACE inhibitor for ≥3 visits.
4P-value = 0.0350
5P-value = 0.0409 comparing use of ARB with ACE inhibitor for 2 visits.
6P-value = 0.0272.
7P-value = 0.0229 and
8P-value = 0.0244 comparing use of ARB with calcium channel blocker for ≥3 visits.
9P-value = 0.0167,
10P-value = 0.0206 and
11P-value = 0.0047 comparing use of ACE inhibitor with calcium channel blocker for 1 visit.
Discussion
This prospective cohort study of older hypertensive men in the MrOS study showed that long-term use of an ARB or ACE inhibitor was associated with a significantly lower incidence of non-vertebral fractures than non-users. Fracture risk reduction was significantly lower with ARB use than with ACE inhibitor or CCB use. These results suggest that only long-term use of an ARB may be associated with lower fracture incidence in older men.
Although cross-sectional studies have suggested that antihypertensive drug use is associated with a lower prevalence of fracture [14], there are limited prospective data on the influence of antihypertensive drugs on fracture incidence. However, the interpretation of prospective cohort studies on the association between drug use and clinical outcomes is confounded by indication bias. For example, the physician may be more inclined to prescribe ACE inhibitors in patients with heart failure or diabetes mellitus. This is one reason our current report focused on the comparison of fracture incidence between ARB and ACE inhibitors. ARBs and ACE inhibitors have very similar profiles in terms of cardiovascular effects and share the same indications. Thus, their comparisons should be less likely biased by indication.
As an additional measure to reduce bias by indication, we analysed our data by the propensity score method, which is designed to reduce bias when comparing exposure groups in observational studies.
Our ARB, ACE inhibitor and CCB users were comparable in clinical characteristics and concomitant drug use. Half of the non-users of these drugs had not taken any antihypertensive drugs for three visits or more, suggesting that their hypertension was milder than the other drug users. This may explain why the non-users had lower body weight, less cardiovascular burden and were less likely to take statins or thiazide diuretics. CCB was used as the major cardiovascular comparator, because it has not been reported to alter bone health.
Both ARB and ACE inhibitor use in the long term was associated with a significant reduction in fracture risk. However, the ARB protective effect on all non-vertebral fractures was more convincing in that it increased with duration of use and was significantly greater than that of the CCB effect. More importantly, the fracture risk reduction in the longer term use of an ARB was significantly greater than that of ACE inhibitor. These results are consistent with a retrospective study which found that the sole use of ARB for newly diagnosed hypertensive patients was associated with a significantly lower risk of fractures, while an ACE inhibitor showed no significant effect [4]. That study had the advantage of a large sample size and no concomitant use of other antihypertensive drugs. However, the data captured the clinically documented fractures only within the period of 1 year. Our study, however, used more reliable fracture data for a period of nearly 7 years, with extensive data on each participant's clinical characteristics. Another retrospective cohort study with propensity score matching in Canada also reported that ARB users had lower risk of non-hip major osteoporotic fracture, compared with ACE inhibitor users [15].
Angiotensin II, the major RAS effector protein, activates osteoclasts and induces bone loss in animal models [16]. ACE inhibitors are not effective in lowering tissue angiotensin II concentrations because of the feedback increase in renin that occurs [17]. All of the RAS components are present in bone tissue. ARB specifically blocks the angiotensin II AT1 receptor, which is primarily responsible for the deleterious cardiovascular effects of angiotensin II. However, its role in bone metabolism is uncertain. In certain animal models, activation of angiotensin II AT2, rather than AT1 receptors, led to osteoclastic activation [1, 18]. Moreover, losartan did not significantly alter bone loss or architecture in male rats after orchidectomy [6], although it did reduce bone loss in female rats 3 months after ovariectomy. It is interesting to note that in the latter study, losartan slightly reduced bone density when given to female rats without ovariectomy [19]. These results suggest that losartan may prevent bone loss only when the RAS in bone is activated.
The protective effect of ARB against fractures may not be directly related to bone loss. First, our previous analysis did not suggest a significant ARB effect against BMD loss over a 4-year period [3]. Second, although most of the fractures were fragility fractures, there was no significant reduction in risk for hip or forearm fractures, which are some of the most common osteoporotic fractures. Third, the Solomon study [4] found a reduction in the fracture rate even within the first year of use. It therefore seems probable that ARBs had a beneficial effect on the fracture rate by preventing falls and/or reducing the severity of the fall. The Lipsitz study found a slightly lower indoor fall risk for ARB user (Odds ratio = 0.64, 95% confidence interval 0.32–1.29) but not statistical significant due to small sample size [20]. However, a beneficial effect on bone architecture cannot be ruled out.
Antihypertensive drug use has been associated with falls, presumably from postural hypotension. Taking more than two antihypertensive drugs, especially around the time of drug initiation or at the time of dose increase, significantly increased the risk of falls [21]. A study showed that ARBs are able to partially restore sympathetic dysfunction in hypertensive people [22].
Other possible mechanisms by which ARBs may reduce fall risk and/or severity include potential effects on muscle and brain function. ACE inhibitors have been suggested to protect muscle function in older people, although more recent prospective studies did not confirm this association [23, 24]. There is some experimental evidence that ARBs may be beneficial to muscle function in a deconditioned animal model [25, 26]. However, the effect of ARB has not been reported in human muscle.
It is interesting that the fracture risk of long-term ACE inhibitor use was similar to that of long-term CCB use and that the short-term users of CCB had an increased risk of osteoporotic fractures. One may speculate that people who were unable to tolerate CCB may have been more frail than those who were able to tolerate it.
The strengths of this study include the detailed data on multiple comorbidities and the fracture risk factors documented by multivariate analysis. The limitations include the lack of data on vertebral deformities, which are often osteoporosis-related, the relatively small number of ARB users, the lack of drug dosing data and the exclusion of women. Moreover, this study was observational, so these findings need to be critically examined in clinical trials.
In summary, we conclude that long-term ARB use may be superior to ACE inhibitor use for fracture prevention in older hypertensive men. Further work is required to elucidate the mechanisms underlying this apparent advantage.
Key points.
In older hypertensive men, ARBs use was associated with lower incidence of non-vertebral fracture than ACE inhibitors or CCBs.
The hazard ratio of non-vertebral fractures was significantly lower in ARB users than ACE inhibitor users (Hazard ratio (95% confidence interval): 0.194 (0.079−0.474) versus 0.620 (0.453−0.850)).
There was a trend of greater fracture risk reduction with longer duration of ARB use, but not for ACE inhibitor use.
Authors' contributions
Timothy Kwok designed the study, wrote and revised the manuscript. Jason Leung conducted statistical analysis and revised the manuscript. Elizabeth Barrett-Connor participated in revision of manuscript. All authors read and approved the final manuscript.
Conflicts of interest
Timothy Kwok, Jason Leung, and Elizabeth Barrett-Connor declare that they have no conflict of interest.
Funding
The Osteoporotic Fractures in Men (MrOS) Study is supported by National Institutes of Health funding. The following institutes provided support: the National Institute on Aging (NIA), the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), the National Center for Advancing Translational Sciences (NCATS), and NIH Roadmap for Medical Research under the following grant numbers: U01 AG027810, U01 AG042124, U01 AG042139, U01 AG042140, U01 AG042143, U01 AG042145, U01 AG042168, U01 AR066160 and UL1 TR000128.
References
- 1. Asaba Y, Ito M, Fumoto T et al. Activation of renin-angiotensin system induces osteoporosis independently of hypertension. J Bone Miner Res 2009; 24: 241–50. [DOI] [PubMed] [Google Scholar]
- 2. Gebru Y, Diao T-Y, Pan H, Mukwaya E, Zhang Y. Potential of RAS inhibition to improve metabolic bone disorders. Biomed Res Int 2013; 10.1155/2013/932691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kwok T, Leung J, Zhang YF, Bauer D et al. and for the Osteoporotic Fractures in Men (MrOS) Research Group Does use of ACE inhibitors or angiotensin receptor blockers affect bone loss in older men. Osteoporos Int 2012; 23: 2159–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Solomon DH, Mogun H, Garneau K, Fischer MA. Risk of fractures in older adults using antihypertensive medications. J Bone Miner Res 2011; 26: 1561–7. [DOI] [PubMed] [Google Scholar]
- 5. Shimizu H, Nakagami H, Osako MK et al. Prevention of osteoporosis by angiotensin-converting enzyme inhibitor in spontaneous hypertensive rats. Hypertens Res 2009; 32: 786–90. [PubMed: 19590507]. [DOI] [PubMed] [Google Scholar]
- 6. Zhang Y-F, Qin L, Kwok TC, Yeung BH, Li G-D, Liu F. Effect of angiotensin II type I receptor blocker losartan on bone deterioration in orchiectomized male hypertensive and normotensive rats. Chin Med J 2013; 126: 2661–5. [PubMed] [Google Scholar]
- 7. Orwoll E, Blank JB, Barrett-Connor E et al. Design and baseline characteristics of the osteoporotic fractures in men (MrOS) study—a large observational study of the determinants of fracture in older men. Contemp Clin Trials 2005; 26: 569–85. [PubMed: 16084776]. [DOI] [PubMed] [Google Scholar]
- 8. Blank JB, Cawthon PM, Carrion-Petersen ML et al. Overview of recruitment for the osteoporotic fractures in men study (MrOS). Contemp Clin Trials 2005; 26: 557–68. [PubMed: 16085466]. [DOI] [PubMed] [Google Scholar]
- 9. Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activity Scale for the Elderly (PASE): development and evaluation. J Clin Epidemiol 1993; 46: 153–62. [PubMed: 8437031]. [DOI] [PubMed] [Google Scholar]
- 10. Pahor M, Chrischilles EA, Guralnik JM, Brown SL, Wallace RB, Carbonin P. Drug data coding and analysis in epidemiologic studies. Eur J Epidemiol 1994; 10: 405–11. [PubMed: 7843344]. [DOI] [PubMed] [Google Scholar]
- 11. National Kidney Foundation K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002; 39(2 Suppl 1):S1–266. [PubMed: 11904577]. [PubMed] [Google Scholar]
- 12. Lewis CE, Ewing SK, Taylor BC et al. Predictors of non-spine fracture in elderly men: the MrOS study. Osteoporotic Fractures in Men (MrOS) Study Research Group. J Bone Miner Res 2007; 22: 211–9. [DOI] [PubMed] [Google Scholar]
- 13. D‘Agostino RB., Jr Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998; 17: 2265–81. [PubMed: 9802183]. [DOI] [PubMed] [Google Scholar]
- 14. Rejnmark L, Vestergaard P, Mosekilde L. Treatment with beta-blockers, ACE inhibitors, and calcium-channel blockers is associated with a reduced fracture risk: a nationwide case-control study. J Hypertens 2006; 24: 581–9. [DOI] [PubMed] [Google Scholar]
- 15. Butt DA, Mamdani M, Gomes T, Lix L, Lu H, Tu K. and Hypertension Outcome Surveillance Team Risk of osteoporotic fractures with angiotensin ii receptor blockers versus angiotensin‐converting enzyme inhibitors in hypertensive community‐dwelling elderly. J Bone Miner Res 2014; 29: 2483–8. [DOI] [PubMed] [Google Scholar]
- 16. Shimizu H, Nakagami H, Osako MK et al. Angiotensin II accelerates osteoporosis by activating osteoclasts. FASEB J 2008; 22: 2465–75. [DOI] [PubMed] [Google Scholar]
- 17. Azizi M, Chatellier G, Guyene TT, Murieta-Geoffroy D, Ménard J. Additive effects of combined angiotensin-converting enzyme inhibition and angiotensin II antagonism on blood pressure and renin release in sodium-depleted normotensives. Circulation 1995; 92: 825–34. [DOI] [PubMed] [Google Scholar]
- 18. Izu Y, Mizoguchi F, Kawamata A et al. Angiotensin II type 2 receptor blockade increases bone mass. J Biol Chem 2008; 284: 4857–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Donmez BO, Ozdemir S, Sarikanat M et al. Effect of angiotensin II type 1 receptor blocker on osteoporotic rat femurs. Pharmacol Rep 2012; 64: 878–88. [DOI] [PubMed] [Google Scholar]
- 20. Lipsitz LA, Habtemariam D, Gagnon M et al. Reexamining the effect of antihypertensive medications on falls in old age. Hypertension 2015; 66: 183–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Slomski A. Falls from taking multiple medications may be a risk for both young and old. JAMA 2012; 307: 1127–8. [DOI] [PubMed] [Google Scholar]
- 22. Sato R, Mizuno M, Miura T et al. Angiotensin receptor blockers regulate the synchronization of circadian rhythms in heart rate and blood pressure. J Hypertens 2013; 31: 1233–8. [DOI] [PubMed] [Google Scholar]
- 23. Gray SL, Boudreau RM, Newman AB. and for the Health et al. Aging and Body Composition Study Angiotensin-converting enzyme inhibitor and statin use and incident mobility limitation in community-dwelling older adults: the Health, Aging and Body Composition study. J Am Geriatr Soc 2011; 59: 2226–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Schellenbaum GD, Smith NL, Heckbert SR et al. Weight loss, muscle strength, and angiotensin-converting enzyme inhibitors in older adults with congestive heart failure or hypertension. J Am Geriatr Soc 2005; 53: 1996–2000. [DOI] [PubMed] [Google Scholar]
- 25. Burks TN, Andres-Mateos E, Marx R et al. Losartan restores skeletal muscle remodeling and protects against disuse atrophy in sarcopenia. Sci Trans Med 2011; 3: 82ra37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Lee EM, Kim DY, Kim AY et al. Chronic effects of losartan on the muscles and the serologic profiles of mdx mice. Life Sci 2015; 143: 35–42. [DOI] [PubMed] [Google Scholar]