Reexamining the Effect of Antihypertensive Medications on Falls in Old AgeNovelty and Significance
Conflicting data on the relationship between antihypertensive medications and falls in elderly people may lead to inappropriate undertreatment of hypertension in an effort to prevent falls. We aimed to clarify the relationships between the chronic use of different classes of antihypertensive medications and different types of falls, to determine the effect of medication dose, and to assess whether the risk of falls is associated with differences in cerebral blood flow. We assessed demographics, clinical characteristics, and chronic antihypertensive medication use in 598 community-dwelling people with hypertension, aged 70 to 97 years, then followed them prospectively for self-reported falls using monthly calendar postcards and telephone interviews. Antihypertensive medication use was not associated with an increased risk of falls. Participants reporting use of angiotensin-converting enzyme inhibitors had a significantly decreased 1-year risk of injurious falls (odds ratio, 0.62; 95% confidence interval, 0.39–0.96), whereas those using calcium channel blockers had a decreased risk of all falls (odds ratio, 0.62; 95% confidence interval, 0.42–0.91) and indoor falls (odds ratio, 0.57; 95% confidence interval, 0.36–0.91), compared with participants not taking these drugs. Larger doses of these classes were associated with a lower fall risk. Participants taking calcium channel blockers had higher cerebral blood flow than those not taking these medications. In relatively healthy community-dwelling elderly people, high doses of antihypertensive agents are not associated with an increased risk of falls.
- angiotensin-converting enzyme inhibitors
- antihypertensive agents
- blood flow velocity
- calcium channel blockers
Several studies have suggested that there is an increased risk of falls in elderly people taking antihypertensive medications.1,2 A recent observational study of 4961 hypertensive community-dwelling participants in the Medicare Current Beneficiary Survey showed an increased risk of injurious falls in those who received moderately intense antihypertensive therapy compared with those on no treatment, but no particular drug class was implicated.3 The acute administration of antihypertensive medications, such as diuretics, can produce significant hypotension in elderly people,4 which may threaten cerebral perfusion and produce falls and syncope. However, this may be a short-term effect that wanes after prolonged drug administration.5
In contrast, other studies have suggested that chronic blood pressure lowering with regimens that include angiotensin system–blocking medications may preserve cerebral hemodynamics,6 increase cerebral perfusion,7–10 and reduce the progression of cerebral white matter hyperintensities,11 which are associated with abnormal gait and falls.12,13 Consistent with these reports, a recent prospective cohort study by Wong et al14 showed a lower risk of falls among elderly community-dwelling participants taking angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs).
The conflicting data on the relationship between antihypertensive medications and falls may be because of the duration of use, an effect on cerebral perfusion, or a dose effect. Lower doses of certain antihypertensive medications may be protective, whereas higher doses may be associated with falls.3,15 We examined 1-year fall risk in a population-based cohort of hypertensive elderly people to help clarify the relationships between the chronic use of different classes of antihypertensive medications and falls. We also examined the effect of medication dose and whether antihypertensive agents are associated with an increase in cerebral blood flow.
Study Design and Sample
We conducted a 1-year prospective observational study of falls after the collection of baseline data in 598 people aged 70 to 97 years with hypertension who were enrolled in the Maintenance of Balance, Independent Living, Intellect, and Zest in the Elderly Study in Boston, Massachusetts (MOBILIZE Boston Study). The study design and methods have been described previously.16 At enrollment, participants met the following inclusion criteria: age ≥70 years, ability to speak and understand English, ability to walk across a room, visual ability to read written material, mini-mental state examination score ≥18, and expected residence in the area for at least 2 years. All participants in this study also met criteria for hypertension, which included an average baseline blood pressure >140 systolic or >90 diastolic on 2 seated measures or the use of antihypertensive medications. All participants underwent a complete baseline home interview and clinical laboratory assessment as previously described,16 which included data on comorbidities, medications, functional and cognitive status, physical performance, blood pressure, height, and weight. All data were collected by trained research assistants.
A subset of 313 participants with accessible temporal acoustic windows underwent transcranial Doppler ultrasound measures of cerebral blood flow velocity (BFV). All participants submitted completed falls study calendars monthly during the follow-up period; however, 14 participants (2.3%) did not know, refused to answer, or could not be reached for additional follow-up questions on the time, location (indoor versus outdoor), and injury status of falls. The study was approved by the Institutional Review Board of Hebrew SeniorLife, and all participants provided written informed consent.
Use of antihypertensive medications was assessed at baseline during an in-person interview by asking the participant to report all medications that were taken for at least a 2-week period. Dose information was recorded from medication bottles. All drug data were coded according to the Iowa Drug Information System, and antihypertensives were classified into the following 6 categories: α-blockers, ACEIs, ARBs, β-blockers, diuretics, and calcium channel blockers. In addition, we grouped participants into those taking any antihypertensive medications and those on angiotensin system-blocking medications (ACE or ARBs). We also assessed the use of statins and psychotropic medications.
Daily medication dose was calculated by multiplying the prescribed strength by the number of times taken per day and the number of days taken per week, divided by 7. We then computed the median dose among all users and dichotomized the dose into standard or low (<median) and high (>median) levels. For example, high-dose amlodipine was >7.5 mg daily, whereas standard or low dose was <5.0 mg. Details are provided in Table S1 in the online-only Data Supplement. A geriatrician (L.A.L.) and pharmacist (I.I.) reviewed these dose categories for each antihypertensive medication to ensure face validity.
Cerebral Blood Flow
During the baseline clinic visit, heart rate was measured by a continuous ECG and beat-to-beat arterial pressure was measured by photoplethysmography (Finapres, Ohmeda Monitoring Systems, Englewood, CO).17 End-tidal CO2 was measured using a vacuumed CO2 analyzer (Ventura, CA) attached to a nasal cannula. Transcranial Doppler ultrasonography (MultiDop X4, DWL-Transcranial Doppler Systems Inc, Sterling, VA) was used to measure middle cerebral artery (MCA) mean BFV while sitting and during changes in end-tidal carbon dioxide (CO2 vasoreactivity).18 The MCA signal was identified according to standard criteria19 and recorded at a depth of 50 to 60 mm. The envelope of the velocity waveform, derived from a fast-Fourier analysis of the Doppler frequency signal, was digitized at 500 Hz, displayed simultaneously with the arterial blood pressure, ECG, and end-tidal CO2 signals, and stored for later off-line analysis. Previous studies, using a variety of techniques (133Xe, SPECT [single-photon emission computerized tomography], magnetic resonance imaging) and stimuli, have validated transcranial Doppler measures of relative changes in cerebral flow velocity as representative of changes in cerebral blood flow.20,21
To assess CO2 vasoreactivity, BFV was measured continuously in the MCA while participants inspired a gas mixture of 8% CO2, 21% O2, and balance nitrogen for 2 minutes and then mildly hyperventilated to an end-tidal CO2 of ≈25 mm Hg for 2 minutes. MCA BFV was plotted against end-tidal CO2 while breathing room air, 8% CO2, or hyperventilating. Cerebral vasoreactivity was measured as the slope of this relationship and expressed as change in BFV per mm Hg change in end-tidal CO2.
A fall was defined as unintentionally coming to rest on the floor, ground, or other lower surface. Falls were ascertained by self-report using monthly falls calendar postcards and telephone interviews for participants who did not return the calendar. Any participant reporting a fall received a follow-up telephone call to collect more information about the location and circumstances of the fall. An indoor fall was defined as one that occurred inside the participant’s home, inside someone else’s home, inside another building, or inside another location. Outdoor falls were those reported to have occurred anywhere outside of these locations. Injurious falls were those reported to have resulted in any physical injury or required treatment in an emergency department or hospital. The primary outcome variable was ≥1 falls within a 1-year follow-up period.
Participant characteristics and fall risk factors were summarized by means and SDs or percentages, and compared between fallers and nonfallers, and treated and untreated participants, using t tests for continuous variables and χ2 tests for categorical variables. We determined the proportion of individuals reporting any fall, and ≥1 indoor, outdoor, and injurious falls during a 1-year follow-up period. We also determined the prevalence of the use of 6 classes of antihypertensive medications, including α-blockers, ACEIs, ARBs, β-blockers, diuretics, and calcium channel blockers. Because only 4 participants were taking vasodilating β-blockers, we were not able to examine this subclass.
Multivariate logistic regression was used to assess the association between antihypertensive medication use and each category of falls, adjusting for relevant covariates and factors found in bivariate analyses to be significantly associated with falls. Covariates included baseline blood pressure, a sum of comorbidities,22 education, race, body mass index, psychotropic medication use (sedatives, hypnotics, antidepressants, and antipsychotics), and time spent engaging in outdoor activities according to the Physical Activity Scale for the Elderly.23 In post hoc analyses, we also stratified our sample by participants with and without falls in the year before the baseline assessment and examined the relationship between antihypertensive medications and subsequent falls in each group.
In subgroup analyses, we studied the effects of each of the 6 classes of antihypertensive medication on 4 different outcome variables: total falls, outdoor falls, indoor falls, and injurious falls. For outdoor and indoor falls, we also adjusted the analyses for estimates of time spent engaging in outdoor and indoor activities using self-reported data from the Physical Activity Scale for the Elderly instrument.
Finally, we examined the relationship between the dose of each antihypertensive medication (high dose, low or standard dose, and none), grouped by medication class, and each of the outcomes of falls using similar multivariate logistic regression models, adjusted for the same covariates as in the 2-group models. We used ANOVA-based F statistics for omnibus tests of differences in risk between dose groups, followed by Tukey post hoc tests for comparisons among groups.
Finally, to explore the degree to which cerebral blood flow could account for associations between antihypertensive medications and falls, we compared odds ratios (ORs) generated in the presence and absence of measures of cerebral blood flow. These analyses were restricted to the subsample of participants with transcranial Doppler ultrasound measures and complete data (n=301). The percent change in the OR relating a given medication to falls when cerebral blood flow was added to the model was used to quantify a possible mediating effect.
Data were analyzed using SAS software version 9.3 (SAS Institute, Cary, NC). Means are presented with their SD in parentheses. Differences were considered statistically significant at P<0.05 (2-tailed).
The mean age of the 598 hypertensive participants was 78.4 (5.4) years; 372 (62.2%) were women, 449 (75.1%) were whites, and 259 (43.4%) had a bachelor’s degree or higher. Two hundred sixty-two participants (89.5%) were regularly taking antihypertensive medications at the start of the study. The proportion of participants reporting the use of antihypertensive medication subclasses were ARBs (12.8%), ACEIs (34.7%), α-blockers (7.1%), β-blockers (51.9%), calcium channel blockers (29.3%), and diuretics (47.0%).
Among the 598 participants, 541 falls were recorded during the 1-year follow-up; 331 (55.3%) participants reported no falls and 267 (44.7%) reported ≥1 falls. Eighty-two (13.7%) participants reported falling outdoors, 127 (21.2%) reported falling indoors, 44 (7.4%) reported falling both indoors and outdoors, and 164 (27.4%) reported falls resulting in injury. Fourteen (2.3%) participants reported falling but were unavailable for follow-up questioning on details of the fall event. The total number of falls per person during the follow-up period ranged from 0 to 17.
Table 1 presents the demographic and clinical characteristics of hypertensive participants who did and did not fall during the 1-year follow-up period. Fallers reported higher levels of education, greater use of psychotropic medications, more depressive symptoms, and a higher rate of previous falls. They were more likely to be white, less likely to be cognitively impaired, and had slightly lower systolic blood pressures compared with nonfallers. Although fallers had a slightly lower prevalence of diabetes mellitus, their total number of comorbidities, including congestive heart failure, was higher.
Table 2 shows the cerebral hemodynamic characteristics of participants who underwent the transcranial Doppler assessment. Fallers had significantly lower cerebral BFV and CO2 vasoreactivity than nonfallers.
We also compared characteristics of participants taking antihypertensive medications to those not taking these medications (Tables S2 and S3). The group taking antihypertensive medications had significantly lower blood pressures, more comorbidities, a higher prevalence of diabetes mellitus, and a lower prevalence of orthostatic hypotension but no difference in previous fall rates.
Relationship Between Antihypertensive Use and Falls
As shown in Table 1, nonfallers were more likely to be taking calcium channel blockers than fallers. There were no differences in the use of other antihypertensive medications between fallers and nonfallers. Multivariate logistic regression analyses revealed that the association between overall antihypertensive medication use and all falls were not statistically significant after adjusting for significant covariates (OR, 0.76; 95% confidence interval [CI], 0.42–1.35). Table 3 shows the results of adjusted multivariate logistic regression analyses for the relationships between antihypertensive medication classes and different types of falls during a 1-year follow-up period. No single class of antihypertensive medication was associated with an increased risk of falls. In fact, calcium channel blocker use was associated with a reduced risk of all falls (OR, 0.62; 95% CI, 0.42–0.91) and indoor falls (OR, 0.57; 95% CI, 0.36–0.91), and ACEI use was associated with a reduced risk of injurious falls (OR, 0.62; 95% CI, 0.39–0.96). When the participants were stratified according to a history of falls, those with previous falls also had a lower risk of all, outdoor, and injurious falls during the 1-year follow-up period if they were taking calcium channel blockers or ACEIs (Table S4).
Effect of Medication Dose
As shown in Figure, higher doses of calcium channel blockers were associated with a lower risk of all falls (OR, 0.44; 95% CI, 0.24–0.82; P=0.03) and indoor falls (OR, 0.31; 95% CI, 0.14–0.69; P=0.01) compared with those not taking calcium channel blocking medications. Higher doses of ACEIs were associated with a significantly lower risk of outdoor falls (OR, 0.40; 95% CI, 0.18–0.92; P=0.04), and all doses were associated with marginally significant reductions in injurious falls (OR, 0.58; 95% CI, 0.34–0.99; P=0.05 for low or standard doses and OR, 0.53; 95% CI, 0.27–1.04; P=0.07 for high doses) compared with those not taking ACEIs.
Effect of Medications on Cerebral Blood Flow
Among the 6 classes of antihypertensive medications examined, only calcium channel blocker use was associated with a statistically significant increase in cerebral BFV. Participants taking calcium channel blockers had a baseline cerebral BFV of 43.3±12.5 cm/s compared with 40.0±9.8 cm/s in those not taking these medications (P=0.02). There was also a modest dose effect of calcium channel blockers on cerebral BFV. Blood flow velocity in participants taking no calcium blocking medications was 40.0±9.8 cm/s, those taking standard/low doses was 43.7±13.5 cm/s, and those taking high doses was 42.3±10.0 cm/s (P=0.03). Post-hoc pairwise comparisons showed that the standard/low-dose group had significantly higher cerebral blood flow values than the no calcium blocker group (P=0.01) and the remaining pairwise comparisons failed to meet statistical significance.
There was a nonsignificant trend toward increased cerebral BFV in those taking ACEIs (42.0+9.9 cm/s) compared with those not taking these drugs (40.1+11.0 cm/s; P=0.14). There was a monotonic trend toward higher doses of ACEIs being associated with higher average cerebral blood flow values, but this relationship also failed to meet statistical significance (none 40.1±11.0 cm/s, standard/low 41.8±9.0 cm/s, and high 42.5±11.2 cm/s; P=0.15).
None of the antihypertensive medications was associated with CO2 vasoreactivity.
Relationship Between Cerebral Blood Flow and Falls
Increased cerebral BFV was also associated with a reduced risk of all falls (OR per 10-cm/s increase in BFV, 0.75; 95% CI, 0.59–0.94; P=0.01), indoor falls (OR per 10-cm/s increase in BFV, 0.76; 95% CI, 0.57–1.00; P=0.05), and injurious falls (OR per 10-cm/s increase in BFV, 0.74; 95% CI, 0.56–0.98; P=0.04). When cerebral BFV was added to the regression model relating calcium channel blockers to falls, the relationship between calcium channel blockers and total and indoor falls was attenuated by 19% and 29%, respectively (Table 4).
The results of this study suggest that the chronic use of antihypertensive medications by community-dwelling elderly people with hypertension may not be as strong a risk factor for falls as previously thought. Neither standard nor high doses of antihypertensive agents were associated with falls. Moreover, ACEIs were associated with a reduced risk of injurious falls and calcium channel blockers were associated with a reduced risk of all falls and indoor falls.
Our findings are supported by the previous studies of Wong et al,14 which also showed a lower risk of falls among community-dwelling older adults taking renin–angiotensin system–blocking medications, and Callisaya et al,15 who showed a minimal effect of antihypertensive drugs on falls in therapeutic doses. In the latter study, only a high dose of these medications that was 3 times the daily defined dose, was significantly associated with falls. Although our participants were not on doses this high, we found that elderly people taking relatively large doses of calcium channel blockers and ACEIs had a lower risk of falls compared with those not taking these medications.
Our results differ from those of Tinetti et al,3 who found a statistically significant increased risk of serious fall injuries only with moderately intense doses of antihypertensive medications in the Medicare Current Beneficiary Survey population and an increased risk with high-intensity doses in participants with a history of falls. It is hard to know how the doses used in the study of Tinetti et al3 compared with ours. Furthermore, our population was slightly younger (78 versus 80 years), healthier (less stroke, diabetes mellitus, and heart failure), and better educated (82%–90% completed more than a high-school education compared with 30%–34%). When we stratified our participants according to those with and without previous falls, even the higher risk participants with previous falls had a lower risk of subsequent falls if they were taking calcium channel blockers or ACEIs. In the study of Tinetti et al,3 no particular class of antihypertensive medication was associated with injurious falls.
One potential mechanism by which the calcium channel blockers could participate in the prevention of falls is through an increase in cerebral blood flow and prevention of ischemic brain injury. Our analysis showing attenuation of the association between calcium channel blockers and falls when controlled for cerebral blood flow is consistent with this mechanism but does not imply causality and requires validation in future clinical trials. This line of reasoning may seem counterintuitive because antihypertensives are thought to decrease cerebral perfusion by reducing blood pressure. It is supported, however, by a small body of evidence showing that the long-term control of high blood pressure may improve systemic and cerebral hemodynamics.7,8,10 We have previously shown that the use of the calcium channel blocker nicardipine in elderly hypertensive people with a history of coronary artery disease can reduce postprandial hypotension.24 The chronic use of several antihypertensive medications has also been shown to reduce orthostatic hypotension.25 Nimodipine has been demonstrated to improve regional cerebral blood flow in rats6 and cerebral oxygenation and blood volume in healthy humans.26 Other studies have shown that the ARB, telmisartan, can increase regional cerebral blood flow in hypertensive patients with chronic-stage stroke27 or Alzheimer disease,28 and candesartan can increase cerebral blood flow in healthy hypertensive participants with low baseline cerebral blood flow.7 We previously showed increases in cerebral blood flow in hypertensive participants whose blood pressure was controlled with a lisinopril-based regimen for 6 months.8 Our present results with ACEIs are limited by relatively small sample sizes, but they suggest that this class of medication may be associated with a lower incidence of injurious falls because of other mechanisms.
One unique aspect of this study is that it examined different types of falls that may have different pathophysiologic mechanisms. For example, indoor falls are more likely to occur in frail individuals with multiple chronic conditions, whereas outdoor falls tend to occur in healthier, more ambulatory individuals with greater exposures to environmental hazards.29 Calcium channel blockers were associated with fewer indoors and all types of falls, whereas ACEIs were associated with fewer injurious falls. The reasons for these differences are not known, but it is possible that the ACEIs improved skeletal muscle function,30 thereby enabling people taking them to better protect themselves during a fall through compensatory motor responses.
There are several limitations to our study. Unfortunately, we could not conduct a randomized trial comparing different classes of antihypertensive agents on rates of falls. We did not have information on adherence or changes in medications during the follow-up period, which precluded us from determining whether the risk of falling was affected by changes in the frequency or dose of medications. In addition, the number of participants taking some classes of medications (eg, α-blockers and ARBs) was relatively small, limiting our power to detect statistically significant effects on falls. We considered the possibility that participants taking antihypertensive medications were healthier and therefore less likely to fall. However, those taking these medications had more comorbidities and diabetes mellitus, which would have biased the results toward more falls in this group.
Not all of our participants had an adequate temporal acoustic window for transcranial Doppler measurements of cerebral BFV. Previous studies using the MOBILIZE Boston Study cohort have shown that elderly participants with suitable bone windows tend to be healthier than those without.31 It is possible that antihypertensive medications have different effects on cerebral hemodynamics in people with and without transcranial Doppler windows. Unfortunately, our cerebral blood flow measurements were limited to only the MCA territory of the brain. However, this territory includes the prefrontal motor cortex that largely controls mobility. Unfortunately, we were unable to assess cerebral blood flow at the time of a fall. Despite these limitations, this study has the unique advantages of using a population-based sample, rigorous longitudinal falls follow-up, novel subcategories of falls, and well-established noninvasive measures of cerebral hemodynamics.
Given the known benefits of treating hypertension in elderly people,32 the withholding of antihypertensive medications to prevent falls may not be a justifiable medical practice. Although the acute administration of any hypotensive medication can precipitate a fall or syncope, the careful uptitration and long-term administration of the lowest effective doses of these medications do not seem to increase the risk of falls.
Sources of Funding
This study was supported by grants P01 AG04390, P30 AG031679, R37 AG25037, and R01 AG041785 from the National Institute on Aging, and R01 NS085002 from the National Institute of Neurological Disorders and Stroke, Bethesda, MD. Drs Tchalla, Dantoine, and Lipsitz were supported in part by the Limoges University Foundation, University Hospital Center of Limoges (CHU de Limoges), and Regional Council of Limousin, France. Dr Lipsitz holds the Irving and Edyth S. Usen and Family Chair in Geriatric Medicine at Hebrew SeniorLife.
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.115.05513/-/DC1.
- Received March 23, 2015.
- Revision received March 30, 2015.
- Accepted April 14, 2015.
- © 2015 American Heart Association, Inc.
- Shannon RP,
- Wei JY,
- Rosa RM,
- Epstein FH,
- Rowe JW.
- Gribbin J,
- Hubbard R,
- Gladman JR,
- Smith C,
- Lewis S.
- Zhang R,
- Witkowski S,
- Fu Q,
- Claassen JA,
- Levine BD.
- Lipsitz LA,
- Gagnon M,
- Vyas M,
- Iloputaife I,
- Kiely DK,
- Sorond F,
- Serrador J,
- Cheng DM,
- Babikian V,
- Cupples LA.
- Tryambake D,
- He J,
- Firbank MJ,
- O’Brien JT,
- Blamire AM,
- Ford GA.
- Verhaaren BF,
- Vernooij MW,
- de Boer R,
- Hofman A,
- Niessen WJ,
- van der Lugt A,
- Ikram MA.
- Srikanth V,
- Beare R,
- Blizzard L,
- Phan T,
- Stapleton J,
- Chen J,
- Callisaya M,
- Martin K,
- Reutens D.
- Zheng JJ,
- Delbaere K,
- Close JC,
- Sachdev PS,
- Lord SR.
- Leveille SG,
- Kiel DP,
- Jones RN,
- Roman A,
- Hannan MT,
- Sorond FA,
- Kang HG,
- Samelson EJ,
- Gagnon M,
- Freeman M,
- Lipsitz LA.
- Sorond FA,
- Khavari R,
- Serrador JM,
- Lipsitz LA.
- Serrador JM,
- Picot PA,
- Rutt BK,
- Shoemaker JK,
- Bondar RL.
- Masuo K,
- Mikami H,
- Ogihara T,
- Tuck ML.
Novelty and Significance
What Is New?
This article provides new information on the possible roles of medication class, dose, and cerebral blood flow on the relationship between antihypertensive medications and different types of falls in elderly people.
What Is Relevant?
Both the treatment of hypertension and prevention of falls are important health promotion goals, yet they may be in conflict if antihypertensive medications increase the risk of falls, as some previous studies suggest.
In community-dwelling elderly people, standard and high doses of antihypertensive medications are not associated with an increased risk of falls. Treatment with calcium channel blockers or angiotensin-converting enzyme inhibitors is associated with a lower risk of falls.