Association Between Chronic or Acute Use of Antihypertensive Class of Medications and Falls in Older Adults. A Systematic Review and Meta-Analysis

Association Between Chronic or Acute Use of Antihypertensive Class of Medications and Falls in... Abstract BACKGROUND Evaluating effect of acute or chronic use of antihypertensives on risk of falls in older adults. METHODS Data sources: Systematic search of primary research articles in CINAHL, Cochrane, EBM, EMBASE, and MEDLINE databases from January 1 2007 to June 1 2017. Study selection: Research studies of cohort, case-control, case-crossover, cross-sectional, or randomized controlled trial (RCT) design examining association between antihypertensives and falls in people older than 60 years were evaluated. Data synthesis: Twenty-nine studies (N = 1,234,667 participants) were included. Study quality was assessed using the Newcastle–Ottawa Scale (NOS). PRISMA and MOOSE guidelines were used for abstracting data and random-effects inverse-variance meta-analysis was conducted on 26 articles examining chronic antihypertensive use, with odds ratios (ORs) and hazards ratios (HRs) analyzed separately. Time-risk analysis was performed on 5 articles examining acute use of antihypertensives. Outcomes: Pooled ORs and HRs were calculated to determine the association between chronic antihypertensive use and falls. For time-risk analysis, OR was plotted with respect to number of days since antihypertensive commencement, change, or dose increase. RESULTS There was no significant association between risk of falling and chronic antihypertensive medication use (OR = 0.97, 95% confidence interval [CI] 0.93–1.01, I2 = 64.1%, P = 0.000; and HR = 0.96, 95% CI 0.92–1.00, I2 = 0.0%, P = 0.706). The time-risk analysis demonstrated a significantly elevated risk of falling 0–24 hours after antihypertensive initiation, change, or dose increase. When diuretics were used, the risk remained significantly elevated till day 21. CONCLUSIONS There is no significant association between chronic use of antihypertensives and falls in older adults. Risk of falls is highest on day zero for all antihypertensive medications. accidental falls, older adults, antihypertensives (agents), blood pressure, geriatrics, hypertension, meta-analysis Falls are a major risk for older adults and are associated with increased morbidity and mortality. Between 0.85 and 1.5% of total health care expenditure is consumed on fall-related expenses in the United States, Europe, and Australia.1 A number of widely prescribed medications have been shown to be significant contributors to falls and fractures.2 While there are many risk factors for falls in older adults, none are potentially as preventable or reversible as medication use.3 Cardiovascular medications are the most commonly used medicines among older adults4 and have been identified as one of the main risk factors for falls in many studies.5–11 Importantly, the prescription of medications to older adults has increased significantly over the last 2 decades.11–13 There have been several systematic and literature reviews14–18 examining the effect of different medications on falls since the meta-analysis by Leipzig et al.18 using articles publishes between 1966 and 1996. Hartikainen et al.14 performed a systematic review of studies published between 1996 and 2004 but did not pool the data for meta-analysis. Wiens et al.15 examined the effect of antihypertensive medications on fractures, meta-analyzing articles published between 1996 and 2005. Woolcott et al.16 updated Leipzig et al.18 using articles published between April 1996 and August 2007 and assessed the impact of 9 medication classes on falls in older adults. Despite the number of studies, fundamental differences in study design have hampered meta-analysis. Zang et al.17 combined the results of Leipzig et al.18 and Wiens et al.15 More recent studies have suggested that time after the initiation, change, or increase in dose of antihypertensive medication is a significant predictor of falls.5,6,8,10,11,19–24 The aim of our systematic review was to update previous meta-analyses16,18 in light of recent research and to examine the association between falls and antihypertensive medication initiation, change, or dose increase in older adults. METHODS Data sources Original research studies were collected through a systematic search of English language articles in CINAHL, Cochrane, EBM, EMBASE, and MEDLINE databases. Primary research articles published between 1 January 2007 and 1 June 2017 were eligible for inclusion. We combined the MeSH term “Therapeutic uses”, which includes all indexed classes of drugs and individual agents, with the MeSH terms “Accidents, Home”, and “Accidental Falls”. In addition, the MeSH terms “Epidemiology” and “Pharmacoepidemiology” were incorporated into the systematic search to find studies in which drug exposure may have been the secondary objective of the research. The MeSH terms “Cardiovascular Agents” and “Antihypertensive Agents”, “Diuretics”, “Vasodilator Agents”, “Nitrates”, “Adrenergic beta-Antagonist” (BB), “Calcium Channel Blockers” (CCB), “Angiotensin-Converting Enzyme Inhibitors” (ACE-i), “Angiotensin Receptor Antagonists” (ARB), and agents acting on the “Renin–Angiotensin System” were associated with MeSH terms “Falls” or “Falling”, or “Fractures”. All terms were expanded to cover all headings and subheadings and find potentially appropriate and relevant studies. References of selected studies were checked for other potentially eligible studies (Figure 1). Based on Equator reporting guidelines, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-P)25 was implemented for the review and selection process. Figure 1. View largeDownload slide PRISMA flowchart. Figure 1. View largeDownload slide PRISMA flowchart. Study selection, inclusion, and exclusion criteria Studies were included if they were primary research studies providing original statistical data of cohort, case–control, case-crossover, cross-sectional, or randomized controlled trial (RCT) design in which the association between use of antihypertensive medications and falls, injurious falls, and fractures in people older than 60 years were evaluated. Antihypertensive medications included all medication that had a primary therapeutic indication of reducing blood pressure. This was defined as medications within Anatomical Therapeutic Chemical (ATC) codes26 C02 (antihypertensive), C03 (diuretic), C07 (BB), C08 (CCB), C09 (ACE-i, ARB, and RAS). Articles covering individual subgroups of nitrates and vasodilators (ATC codes C01, C04) have been excluded as these medication classes are not primarily indicated for hypertension. An operational definition for chronic use was conservatively set at medication therapy for at least 28 days, to account for pharmacokinetic, and pharmacodynamics of included medications. Articles covering temporal relationships between medication change or dose increase and falls were included for time-risk analysis.5,7,10,21,23 Data extraction Article titles, key words, indexed terms, abstracts, and full texts were initially screened by H.R.K., with inclusion decided in consensus with C.R.S. and M.D.L. In addition, 10% of the excluded articles were randomly checked by C.R.S. to confirm that they did not meet the inclusion criteria. In order to be included in the meta-analysis or time-risk analysis, effect sizes needed to be reported as odds ratio (OR), relative risk (RR), incidence risk ratio (IRR), standard incidence ratio, or hazard ratio (HR). The adjusted effect sizes and 95% confidence intervals (CI) were used for analysis. When adjusted effect sizes were not reported, crude effect sizes were used.20,27–29 If the reported data did not meet these criteria or clarification was required to assess eligibility, the corresponding authors were contacted. If the articles only provided 2 × 2 tables of fall incidents and exposure, the effect sizes were calculated from the available data. Information extracted from the compiled studies contained study type, study setting (hospital, nursing home, or community dwelling), country of study, mean age, demographic data, period of study, sample size, ATC codes, fall definition, types of medication, exposure period, type of effect size, confounders, medication duration, fall ascertainment method, temporal relationship between medication initiation, change or dose increase and fall, fall risk index,30 and study design (Table 1). Table 1. Summary of the research studies included in the meta-analysis and time-risk analysis. Source  Setting  Publication date, Data collection, Follow up period  Sample size  Age, Mean(SD),eYrs.  Type of Fall  Medicationa  ATC Codes b   Time of Medication Ascertainment  Method of Fall Ascertainment  Study design  NOS results  Askari et al. 19   Hospital (ED) c  2013/72(mon)f  2258  77.7(7.8)  Falls  C, D, BB  C01,C03, C07  Medication self-report  Incident report  Cross- sectional  10/10  Baranzini et al. 43   L/N d  2009/42(mon)  293  84.4(8.2)  Fall injuries  Ahy, D  C, C03  Baseline  Incident report  Cohort  8/9  Berry et al. 5   L/N  201/65(mon)  1191  88(8)  Falls  D  C03  Computerized order entry system  Computerized incident report  Case-cross over  7/9  Butt et al.(1) 6   Community  2012/108(mon)/ 450days  301591  81(7.3)  Hip fracture  Ahy, Tz, BB, CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline,  Baseline  Case-cross over  7/9  Butt et al.(2) 7   Community  2013/108(mon)/ 450days  543572  80(7.6)  Falls  Ahy  C  Baseline, extracted from DB  Baseline: DB  Case-cross over  7/9  Callisaya et al. 8   Community  2014/12(mon)/ 12 mon  409  72(6.9)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C, C03, C07, C08, C09  Baseline, face to face interview/ prescription check  Daily fall calendar, Incident report, bimonthly questionnaire  Cohort  9/9  Charlesworth et al. 38   Community  2015/72(mon)/ 4.5 yrs.  4736  71.7(6.4)  Falls  Ahy  C  Baseline, face to face interview/prescription check  Self-report  Cohort (RCT)  9/9  Duh et al. 44   L/N  2008/64(mon)/ More than five yrs.  47530  75.4(3.1)  Injurious falls  Ahy  C  Baseline extracted from database  Injurious claim within 30 days after fall  Cohort  7/9  Eriksson et al. 20   L/N  2009/6(mon)/ 6 mon  186  83.6(6.6)  Falls  D, BB, ACE-i  C03, C07, C09  Baseline  Incident report  Cohort  8/9  Garcia et al. 50   Hospital  2014/12(mon)  122  83.3(4.8)  Hip fracture  D  C03  Baseline  Incident report  Case- Control  4/9  Gribbin et al. 21   Community  2010/48(mon)/ 3 yrs.  9682  77.5(.)  Falls  Tz, BB, CCB ACE-i, ARB  C03, C07 C08, C09  Baseline  Incident report  Case- Control  7/9  Ham et al. 37   Community  2014/32(mon)/ -2 to -3 yrs.  2407  74.4(6.7)  Falls  Ahy, D, BB, CCB, ACE- i,, ARB  C, C03, C07, C08, C09  Baseline prospective  Fall incident report  Cohort (RCT)  9/9  Hasegawa et al. 27   L/N  2009/12(mon)/ 6 mon  1082  82.5(8.5)  Injurious falls  Ahy, ACE-i, CCB  C, C08, C09  Baseline  Incident report  Cohort  9/9  Lipsitz et al. 45   Community  2015/12(mon)/ 1 yr.  598  78.4(5.4)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline: Interview 2 weeks recall  Self-report, Monthly calendar  Cohort  9/9  Montali et al. 46   Hospital(ED)  2015/12(mon)  2377  81.2(8)  Falls  Ahy  C  Interview reports  Baseline, from Database  Cohort  8/9  Pariente et al. 22   Community  2008/120(mon)/ 10 yrs.  3777  79.9(6.8)  Injurious falls  Ahy,  C  Interview  Baseline, Incident report  Case- Control  7/9  Payne et al. 9   Hospital  2013/12(mon)/ 1 yr.  39813  (NA)  Falls or fractures  CV  C  Baseline  Incident report  Case- Control  8/9  Rafiq et al. 47   Community  2014/60(mon)/ 30 mon  135433  75.4(7.6)  Falls  Ahy, ACE-i  C, C09  Baseline(from Database)  Baseline(from Database)  Cohort  8/9  Rhalimi et al. 51   Hospital  2009/24(mon)  260  89(7)  Falls  CCB  C08  Baseline  Incident report  Case-control  6/9  Shimbo et al. 10   Hospital(ED)  2016/66(mon)/ 365 days  90127  (NA)  Serious fall injuries  Ahy, ACE- i, ARB, BB,CCB,D  C, C03, C07,C08, C09  Baseline prescription  Baseline- ED ,inpatient claims  Case-cross over  8/9  Shuto et al. 23   Hospital  2009/30(mon)  349  71.5(14.8)  Falls  Ahy, ARB  C, C09  Baseline  Incident report  Case-cross over  7/9  Stenhagen et al. 24   Community  2013/72(mon)/ 3 yrs.  1763  (NA)  Falls  Ahy, D  C, C03  Baseline  Interview(6 Months recall)  Cohort  9/9  Sterke et al.(1) 28   L/N  2012/24(mon)/ 350 days  248  82(8)  Falls  Ahy, BB,  C, C07  Baseline  Incident report  Cohort  8/9  Sterke et al.(2) 48   L/N  2012/24(mon)/ 350 days  248  82(8)  Injurious falls  Ahy, BB,  C, C07,  Baseline  Incident report  Cohort  7/9  Thorel et al. 49   Community  2014/24(mon)  38407  (NA)  Hip fracture  Ahy, D, BB, CCB, RAS,  C, C03, C07, C08,C09  Baseline  Incident report  Cohort  8/9  Tinetti et al. 11   Community  2014/36(mon)/ 3 yrs.  4961  80.2(6.8)  Serious fall injuries  D, BB, CCB, RAS  C03, C07, C08, C09  Interview-direct observation  Incident report  Cohort  8/9  Wong et al. 29   Community  2013/17(mon)/ 12 mon  531  79.9(4.4)  Falls  D, Tz, BB, CV, ACE-i, ARB, RAS  C, C03, C07, C09,  Baseline  Fall calendar (monthly)  Cohort  9/9  Zia et al.(1) 52   Community  2015/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Zia et al.(2) 53   Community  2016/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Source  Setting  Publication date, Data collection, Follow up period  Sample size  Age, Mean(SD),eYrs.  Type of Fall  Medicationa  ATC Codes b   Time of Medication Ascertainment  Method of Fall Ascertainment  Study design  NOS results  Askari et al. 19   Hospital (ED) c  2013/72(mon)f  2258  77.7(7.8)  Falls  C, D, BB  C01,C03, C07  Medication self-report  Incident report  Cross- sectional  10/10  Baranzini et al. 43   L/N d  2009/42(mon)  293  84.4(8.2)  Fall injuries  Ahy, D  C, C03  Baseline  Incident report  Cohort  8/9  Berry et al. 5   L/N  201/65(mon)  1191  88(8)  Falls  D  C03  Computerized order entry system  Computerized incident report  Case-cross over  7/9  Butt et al.(1) 6   Community  2012/108(mon)/ 450days  301591  81(7.3)  Hip fracture  Ahy, Tz, BB, CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline,  Baseline  Case-cross over  7/9  Butt et al.(2) 7   Community  2013/108(mon)/ 450days  543572  80(7.6)  Falls  Ahy  C  Baseline, extracted from DB  Baseline: DB  Case-cross over  7/9  Callisaya et al. 8   Community  2014/12(mon)/ 12 mon  409  72(6.9)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C, C03, C07, C08, C09  Baseline, face to face interview/ prescription check  Daily fall calendar, Incident report, bimonthly questionnaire  Cohort  9/9  Charlesworth et al. 38   Community  2015/72(mon)/ 4.5 yrs.  4736  71.7(6.4)  Falls  Ahy  C  Baseline, face to face interview/prescription check  Self-report  Cohort (RCT)  9/9  Duh et al. 44   L/N  2008/64(mon)/ More than five yrs.  47530  75.4(3.1)  Injurious falls  Ahy  C  Baseline extracted from database  Injurious claim within 30 days after fall  Cohort  7/9  Eriksson et al. 20   L/N  2009/6(mon)/ 6 mon  186  83.6(6.6)  Falls  D, BB, ACE-i  C03, C07, C09  Baseline  Incident report  Cohort  8/9  Garcia et al. 50   Hospital  2014/12(mon)  122  83.3(4.8)  Hip fracture  D  C03  Baseline  Incident report  Case- Control  4/9  Gribbin et al. 21   Community  2010/48(mon)/ 3 yrs.  9682  77.5(.)  Falls  Tz, BB, CCB ACE-i, ARB  C03, C07 C08, C09  Baseline  Incident report  Case- Control  7/9  Ham et al. 37   Community  2014/32(mon)/ -2 to -3 yrs.  2407  74.4(6.7)  Falls  Ahy, D, BB, CCB, ACE- i,, ARB  C, C03, C07, C08, C09  Baseline prospective  Fall incident report  Cohort (RCT)  9/9  Hasegawa et al. 27   L/N  2009/12(mon)/ 6 mon  1082  82.5(8.5)  Injurious falls  Ahy, ACE-i, CCB  C, C08, C09  Baseline  Incident report  Cohort  9/9  Lipsitz et al. 45   Community  2015/12(mon)/ 1 yr.  598  78.4(5.4)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline: Interview 2 weeks recall  Self-report, Monthly calendar  Cohort  9/9  Montali et al. 46   Hospital(ED)  2015/12(mon)  2377  81.2(8)  Falls  Ahy  C  Interview reports  Baseline, from Database  Cohort  8/9  Pariente et al. 22   Community  2008/120(mon)/ 10 yrs.  3777  79.9(6.8)  Injurious falls  Ahy,  C  Interview  Baseline, Incident report  Case- Control  7/9  Payne et al. 9   Hospital  2013/12(mon)/ 1 yr.  39813  (NA)  Falls or fractures  CV  C  Baseline  Incident report  Case- Control  8/9  Rafiq et al. 47   Community  2014/60(mon)/ 30 mon  135433  75.4(7.6)  Falls  Ahy, ACE-i  C, C09  Baseline(from Database)  Baseline(from Database)  Cohort  8/9  Rhalimi et al. 51   Hospital  2009/24(mon)  260  89(7)  Falls  CCB  C08  Baseline  Incident report  Case-control  6/9  Shimbo et al. 10   Hospital(ED)  2016/66(mon)/ 365 days  90127  (NA)  Serious fall injuries  Ahy, ACE- i, ARB, BB,CCB,D  C, C03, C07,C08, C09  Baseline prescription  Baseline- ED ,inpatient claims  Case-cross over  8/9  Shuto et al. 23   Hospital  2009/30(mon)  349  71.5(14.8)  Falls  Ahy, ARB  C, C09  Baseline  Incident report  Case-cross over  7/9  Stenhagen et al. 24   Community  2013/72(mon)/ 3 yrs.  1763  (NA)  Falls  Ahy, D  C, C03  Baseline  Interview(6 Months recall)  Cohort  9/9  Sterke et al.(1) 28   L/N  2012/24(mon)/ 350 days  248  82(8)  Falls  Ahy, BB,  C, C07  Baseline  Incident report  Cohort  8/9  Sterke et al.(2) 48   L/N  2012/24(mon)/ 350 days  248  82(8)  Injurious falls  Ahy, BB,  C, C07,  Baseline  Incident report  Cohort  7/9  Thorel et al. 49   Community  2014/24(mon)  38407  (NA)  Hip fracture  Ahy, D, BB, CCB, RAS,  C, C03, C07, C08,C09  Baseline  Incident report  Cohort  8/9  Tinetti et al. 11   Community  2014/36(mon)/ 3 yrs.  4961  80.2(6.8)  Serious fall injuries  D, BB, CCB, RAS  C03, C07, C08, C09  Interview-direct observation  Incident report  Cohort  8/9  Wong et al. 29   Community  2013/17(mon)/ 12 mon  531  79.9(4.4)  Falls  D, Tz, BB, CV, ACE-i, ARB, RAS  C, C03, C07, C09,  Baseline  Fall calendar (monthly)  Cohort  9/9  Zia et al.(1) 52   Community  2015/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Zia et al.(2) 53   Community  2016/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Abbreviations: a Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha-blocker; BB, Beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin-angiotensin system (RAS); Tz, thiazide, NOS, Newcastle-Ottawa Scale; b ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin-angiotensin system (ACE-i, RAS, ARB).c ED, emergency department; d L/N: Long-term care facility/Nursing home; T studies included in time-risk effect analysis; M-A Meta-analysis; e (SD), standard deviation; Yrs, Years; f mon, month; g Adjusted OR was provided via personal correspondence. View Large The definition of a “fall” used by individual studies was compared with the definitions provided by WHO,31 the Prevention of Falls Network Europe (ProFaNE)32 and Kellogg’s international working group.33 Assessment of study quality, risk of bias The 9-item scale for cohort, case–control, and case-crossover studies, and the 10-item scale for cross-sectional studies from the Newcastle–Ottawa scale (NOS)34 were used to assess study quality and risk of bias. Two RCTs, (B-PROOF),35 and (SHEP),36 followed cohorts37,38 of subjects to determine risk of falls. In addition to the quality assessment, methods of medication verification, and falls ascertainment were extracted from each article. Data synthesis and analysis Meta-analyses were performed on 6 medication classes, categorized by their ATC codes.26 If an ATC code, name or class of medication were not reported in the study, the effect sizes were included in the non–class-specified antihypertensive group (ATC code C). The meta-analysis was conducted in accordance with the Meta-analysis Of Observational Studies in Epidemiology Group (MOOSE) protocols.39 Articles examining acute (<28 days) use of antihypertensives were selected for time-risk analysis. Pooled OR was obtained from OR. In one instance when relative risk was reported, the author was contacted and adjusted OR was provided via personal correspondence, pooled HR was obtained from HR and IRR. Separate meta-analyses were conducted for OR and HR, following the recommendations of Egger et al.40 (Figures 2 and 3). Figure 2. View largeDownload slide Antihypertensive medication and falls, odds ratio meta-analysis, including prior evidence. Antihypertensive medications and falls, meta-analysis results. Odds ratios and 95% confidence intervals for each individual or pooled study effect-sizes, ATC codes: C, C02, C03, C07, C08, C09. Statistical analysis calculated by random-effect inverse-variance (REIV) model. Abbreviations: Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blockers; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CI, confidence interval; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive, Cardiovascular system; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 2. View largeDownload slide Antihypertensive medication and falls, odds ratio meta-analysis, including prior evidence. Antihypertensive medications and falls, meta-analysis results. Odds ratios and 95% confidence intervals for each individual or pooled study effect-sizes, ATC codes: C, C02, C03, C07, C08, C09. Statistical analysis calculated by random-effect inverse-variance (REIV) model. Abbreviations: Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blockers; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CI, confidence interval; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive, Cardiovascular system; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 3. View largeDownload slide Antihypertensive medication and falls, hazard ratio meta-analysis. Antihypertensive medication and falls, meta-analysis results. Hazard ratios and 95% confidence interval (95% CI) for each individual or pooled study effect-size, ATC codes: C, C02, C03, C07, C08, C09. Abbreviations: Ahy, antihypertensives; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. Figure 3. View largeDownload slide Antihypertensive medication and falls, hazard ratio meta-analysis. Antihypertensive medication and falls, meta-analysis results. Hazard ratios and 95% confidence interval (95% CI) for each individual or pooled study effect-size, ATC codes: C, C02, C03, C07, C08, C09. Abbreviations: Ahy, antihypertensives; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. Owing to heterogeneity between studies in experimental design and sampling, a random-effects model was used with inverse-variance estimate of the weights, as recommended by DerSimonian and Laird.41,42 Subgroup sensitivity analyses with 95% CI based on Pearson’s heterogeneity Chi-square test were performed for type of study (cohort, case–control, case-crossover, cross-sectional), study setting (hospital, nursing home, or community-dwelling), and mean age of participants (≤75 years old or >75 years old) where heterogeneity was greater than 50% and statistically significant (P < 0.05) (Table 2). Data regarding initiation, change, or dose increase of medications were examined separately. All statistical analyses were conducted using STATA/IC version 14.2, Windows 64 bit (StataCorp, TX). Table 2. Association between use of antihypertensive medications and falls, subgroup sensitivity analysis   C (non–class-specified antihypertensives)  C08 (CCB)  C09 (ACE-i, ARB, RAS)    n, effect size (95% CI)  n, effect size (95% CI)  n, effect size (95% CI)  All studies (OR)  12, 0.90 (0.83, 0.98)  6, 0.95 (0.75, 1.15)  12, 0.98 (0.92, 1.05)  All studies (HR)a  7, 0.94 (0.86, 1.02)  4, 1.01 (0.90, 1.13)  7, 0.94 (0.86, 1.03)  Population (OR)  Community  8, 0.84 (0.68, 0.99)  5, 0.93 (0.74, 1.13)  12, 0.98 (0.92, 1.05)  L/N  1, 1.30 (0.60, 2.00)  0  0  Hospital  3, 1.02 (0.80, 1.25)  1, 2.35 (0.56, 4.14)  0  Mean age of study subjects (OR)  ≤75 years  3, 0.93 (0.91, 0.96)  2, 1.36 (0.96, 1.77)  5, 1.16 (0.99, 1.33)  >75 years  5, 0.88 (0.59, 1.17)  3, 0.90 (0.51, 1.29)  6, 0.96 (0.88, 1.04)  Study type (OR)        Cohort  7, 0.99 (0.79, 1.19)  3, 0.85 (0.56, 1.14)  8, 0.96 (0.86, 1.06)  Case–control  4, 0.82 (0.64, 1.00)  3, 1.13 (0.79, 1.48)  4, 1.02 (0.91, 1.13)  Cross-sectional  0  0  0  Case-crossover  1, 1.00 (0.80, 1.20)  0  0    C (non–class-specified antihypertensives)  C08 (CCB)  C09 (ACE-i, ARB, RAS)    n, effect size (95% CI)  n, effect size (95% CI)  n, effect size (95% CI)  All studies (OR)  12, 0.90 (0.83, 0.98)  6, 0.95 (0.75, 1.15)  12, 0.98 (0.92, 1.05)  All studies (HR)a  7, 0.94 (0.86, 1.02)  4, 1.01 (0.90, 1.13)  7, 0.94 (0.86, 1.03)  Population (OR)  Community  8, 0.84 (0.68, 0.99)  5, 0.93 (0.74, 1.13)  12, 0.98 (0.92, 1.05)  L/N  1, 1.30 (0.60, 2.00)  0  0  Hospital  3, 1.02 (0.80, 1.25)  1, 2.35 (0.56, 4.14)  0  Mean age of study subjects (OR)  ≤75 years  3, 0.93 (0.91, 0.96)  2, 1.36 (0.96, 1.77)  5, 1.16 (0.99, 1.33)  >75 years  5, 0.88 (0.59, 1.17)  3, 0.90 (0.51, 1.29)  6, 0.96 (0.88, 1.04)  Study type (OR)        Cohort  7, 0.99 (0.79, 1.19)  3, 0.85 (0.56, 1.14)  8, 0.96 (0.86, 1.06)  Case–control  4, 0.82 (0.64, 1.00)  3, 1.13 (0.79, 1.48)  4, 1.02 (0.91, 1.13)  Cross-sectional  0  0  0  Case-crossover  1, 1.00 (0.80, 1.20)  0  0  Pooled OR and HR. Stratified by population setting, type of study, and mean age of participants in OR and HR. ATC code: The Anatomical Therapeutic Chemical (ATC) classification system. C, non–class-specified antihypertensive; C02, antihypertensives; C03, diuretics; C07, beta-blocking agents; C08, calcium channel blockers; C09, agents acting on RAS (ACE-i, RAS, ARB). Statistical analysis: calculated by random-effects inverse-variance model (REIV). Abbreviations: ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CI, confidence interval; ED, emergency department, ED were considered as community in the sensitivity analysis as the fall has occurred in the community; HR, hazard ratio; L/N, long-term care/nursing home; OR, odds ratio; HR, hazard ratio; n, number of effect-sizes in the meta-analysis; RAS, renin–angiotensin system. aHeterogeneity was not observed in HR plot, sensitivity analysis is to provide an updated comparison view between OR and HR studies. View Large RESULTS From 5,142 studies, 29 (N = 1,234,667 participants) met our inclusion criteria. Detailed characteristics of the 29 studies are presented in (Table 1). Among the articles included were 16 cohort,8,11,20,24,27–29,37,38,43–49 7 case–control,9,21,22,50–53 5 case-crossover,5–7,10,23 and 1 cross-sectional study.19 Of the 16 cohort studies, 8 were prospective studies with 6 months,20,27 1 year,8,28,29,45,48 and 3- to 6-year follow-up periods.24 Among studies providing a standard definition of a “fall”, 5 used the WHO,31 48,23,52,53 used the ProFaNE,32 and 328,37,48 used the Kellogg definition.33 Most of the studies adjusted their effect sizes for age, sex, and comorbidity. Only 2 studies10,29 using OR and 3 studies using HR20,27,28 presented crude effect sizes. Twenty-six articles involving chronic antihypertensive medication use were meta-analyzed. Heterogeneity was observed in OR forest plots. There was no heterogeneity for studies examining antihypertensive medication classes C02, C03, or C07. There was heterogeneity in ATC codes C, C08 and C09. For ATC C (non–class-specific hypertensive), a significantly reduced risk of falling was noted only in studies of community-dwelling older adults (OR = 0.84, 95% CI, 0.68–0.99) and subjects 75 years of age and younger (OR = 0.93, 95% CI, 0.91–0.96), leading to a reduced risk for ATC code C overall (pooled OR = 0.90, 95% CI, 0.83–0.98). For classes C08 (CCB) and C09 (ACE-i, ARB, and RAS) all effect sizes found no significant association between medications and falls. No significant differences between the OR or HR for different populations (i.e., residents in the community, hospital, or nursing home), mean age (≤75 or >75 years), or type of study (cohort, case–control, case-crossover, or cross-sectional) were noted. Although some minor variations in the stratified OR and HR were noticed, none of these were statistically significant (Table 2). Only one study11 in the meta-analysis used propensity score matching54 to control for confounding by indication,55,56 while most of the studies used multivariable modelling and reported adjusted OR or HR. Five studies5,7,10,21,23 investigated commencement of cardiovascular medication as a risk factor for falls and were included in time-risk analysis. Four studies had case-crossover and 1 study21 had case–control design. Studies varied in setting (community, hospital, or long-term residential facility). One of these studies7 presented the data as IRR with the remainder providing OR as effect size. Two articles10,21 were included in both meta-analysis and time-risk analysis (Table 1). Chronic hypertensive use Table 1 includes results for non–class-specified antihypertensive medication, diuretic, BB, CCB, and ACE-i. Figures 2 and 3 show the OR, HR, and 95% CI of each study, along with the random-effects inverse-variance pooled effect size of each class of medication. OR pooled by Leipzig et al.18 in fixed-effects models and by Woolcott et al.16 in Bayesian models were documented as sources of prior evidence without being included in our meta-analysis. Chronic antihypertensive use was not associated with falls (overall OR = 0.97, 95% CI, 0.93–1.01; overall HR = 0.96, 95% CI, 0.92–1.00). Similarly, when antihypertensive classes were examined, there was no significantly increased risk was associated with diuretics (OR = 1.1, 95% CI, 0.99–1.23; HR = 0.99, 95% CI, 0.83–1.14), BB (OR = 0.93, 95% CI, 0.89–0.98; HR = 0.99, 95% CI, 0.89–1.09), CCB (OR = 0.95, 95% CI, 0.75–1.15; HR = 1.01, 95% CI, 0.90–1.13), or ACE-i and ARB (OR = 0.98, 95% CI, 0.92–1.05; HR = 0.94, 95% CI, 0.86–1.03). For non–class-specific antihypertensives (ATC code C), the OR reported by Leipzig et al.18 (OR = 1.16, 95% CI, 0.87–1.55) and Woolcott et al.16 (OR = 1.26, 95% CI, 1.08–1.46) were significantly different (P = 0.010) to our pooled OR (OR = 0.90, 95% CI, 0.83–0.98). Our pooled effect size for BB use (OR = 0.93, 95% CI, 0.89–0.98) differed from the Bayesian estimate calculated by Woolcott et al.16 (OR = 1.14, 95% CI, 0.97–1.33), although the difference was not statistically significant (P = 0.816). Dose and risk Callisaya et al.8 demonstrated that higher daily defined dose (DDD) is independently associated with falls and with more than 3 DDD, OR of falling increased from 1.17 (95% CI 0.77–1.77) to 1.56 (95% CI 1.02–2.38). Tinetti et al.11 reported that the HR of falling increased from 2.17 (95% CI, 0.98–4.8) for moderate doses to 2.31 (95% CI, 1.01–5.29) for high doses of antihypertensives. Shimbo et al.10 investigated polypharmacy and demonstrated that increasing the number of classes of medication from 1 to 2 or more will increase the OR of falling from 1.31 (95% CI, 1.13–1.52) to 1.58 (95% CI, 1.17–2.13). Lipsitz et al.45 reported that higher doses of CCB were associated with a lower risk of falls compared with those not taking CCB (OR = 0.44, 95% CI, 0.24–0.82). In addition, higher doses of ACE-i associated with a significantly lower risk of falls occurring outside home (OR = 0.40, 95% CI, 0.18–0.92). Low or standard ACE-i doses were associated with a significant reduction in injurious falls (OR = 0.58, 95% CI, 0.34–0.99), while high doses were associated with a nonsignificant reduction (OR = 0.53, 95% CI, 0.27–1.04) compared with those not taking ACE-i.45 Time and risk Two studies21,23 found a highly elevated risk of falling in the first 0–24 hours (day 0) after initiation, increase of dose or change of antihypertensive medication (Figure 4) for non–class-specified antihypertensives, diuretics, BB, CCB, and ACE-i/ARB. The risk of falling on diuretics remained significantly elevated until day 21, with a trend toward being significantly elevated until day 28. The risk of falling for all other antihypertensive classes decreased to nonsignificant levels from day 1 onward. From day 28 onward, no antihypertensive medication was associated with an increased risk if falling. Figure 4. View largeDownload slide Time-effect analysis: association of antihypertensive medication and falls. Abbreviations: ATC code, The Anatomical Therapeutic Chemical (ATC) Classification System; C, non–class specified antihypertensive; Cardiovascular system: C02, antihypertensive; C03, diuretic; C07, beta-blocking agent; C08, calcium channel blocker; C09, agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 4. View largeDownload slide Time-effect analysis: association of antihypertensive medication and falls. Abbreviations: ATC code, The Anatomical Therapeutic Chemical (ATC) Classification System; C, non–class specified antihypertensive; Cardiovascular system: C02, antihypertensive; C03, diuretic; C07, beta-blocking agent; C08, calcium channel blocker; C09, agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 5. View largeDownload slide Publication bias assessed via funnel plot, association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 5. View largeDownload slide Publication bias assessed via funnel plot, association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 6. View largeDownload slide Publication bias assessed by funnel plot: association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 6. View largeDownload slide Publication bias assessed by funnel plot: association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. DISCUSSION Our study found that the risk of falling was influenced by duration of antihypertensive treatment, being significantly elevated and highest during the first 24 hours (day 0) following medication initiation, change, or dose increase, with the odds increasing up to 36 times.21 Chronic antihypertensive medication use (≥28 days) was not associated with an increased risk of falling in older adults. When the diuretic class (ATC code 03) was examined, the risk of falling remained significantly elevated until day 21 and nonsignificantly elevated in chronic use. Chronic BB use appeared to be associated with a significantly reduced risk of falling. These findings suggest that a first dose effect may contribute to falls in susceptible older adults, possibly due to substantial drops in blood pressure and orthostatic hypotension. Antihypertensive medications that cause hypovolemia, bradycardia, and arterial or venous dilatation may predispose to this phenomenon. For diuretics, however, it takes longer, possibly up to 28 days, to decline to the low-risk and no-risk zone. This extended duration of risk may be due to a diuresis effect. Our pooled OR for BB use was nonsignificantly lower than the Bayesian estimate calculated by Woolcott et al.16 As we excluded BB use of less than 28 days from our meta-analysis, our pooled OR reflects chronic BB use more accurately than earlier meta-analyses which included both acute and chronic BB use. Also, there appeared to be a statistically significant reduction in risk of falling with chronic use of BB and non–class-specific antihypertensive use (ATC code C). This finding could be due to confounding by selective prescribing, where frail older adults (those more likely to fall) are less likely to receive chronic treatments (such as BB) than older adults who are not frail. The latter group may also be more likely to pursue and adhere to other therapies, such as exercise, that may reduce the risk of falls. It is also possible that BB may protect against falls in some way, although this would be difficult to evaluate outside the RCT setting. Conversely Lipsitz et al.45 found that higher doses of ACE-i were associated with lower risk of injurious falls and CCB with lower risk of all and indoor falls. Lipsitz et al.45 suggests that one potential mechanism that CCBs contribute to prevention of falls is an increase in cerebral blood flow and prevention of ischemic brain injury. For ACE-i, Lipsitz et al. suggests that possible improvement of skeletal muscle function.57 Leipzig et al.18 found that thiazide diuretics had higher OR than loop diuretics but suggested that the risk of falls had been declining over the years 1975 to 1993, possibly due to lower doses of medications being used. The subsequent meta-analysis by Woolcott et al.,19 including studies published between 1996 to 2007, found that diuretic use was associated with increased fall risk in unadjusted meta-analysis but not when adjustment was made for other risk factors for falls. Similarly, our meta-analysis found that chronic diuretic use was not associated with a significantly increased risk of falling. In our study, OR and HR were meta-analyzed separately. HR provides an estimate of falls risk at each point in time during the follow-up period and should not be meta-analyzed with OR, which estimates the risk of falls across the entire follow-up period. As risk of falling with antihypertensive medication is time-dependent, effect sizes should be calculated and reported as HR, IRR, or standard incidence ratio rather than OR or relative risk. The majority of included studies did not provide a gender-specific effect size. Studies should distinguish between chronic medication use and acute changes in medication. Moreover, it is clearly visible in HR/IRR effect sizes that there is lower heterogeneity between studies.58,59 The degree of heterogeneity observed in our OR plots may have been due to differences in study design, degree of adjustment for confounders, follow-up duration, and outcome variable (falls, injurious falls, or fractures). Studies that did not adjust for confounding may have overestimated the association between medications and falls. Leipzig et al.18 suggested that heterogeneity could also be attributed to varying doses of antihypertensive medications or confounding by indication. For example, studies reporting significantly elevated risk of falling with diuretic45 and CCB10 involved hospital patients, where sicker or frailer patients may have been more likely to fall and receive antihypertensive therapy. Future meta-regression could be conducted to identify sources of heterogeneity. Researchers should incorporate time into their study designs (index, incident, initiation, and exposure times) and distinguish between acute and chronic medication use. A case-crossover or RCT design are more appropriate designs,40 although the propensity score method54 could reduce confounding by indication in future observational studies. Strengths and limitations By analyzing chronic antihypertensive medication use separately to acute use, our study was able to reduce study heterogeneity and examine the relationship between duration of therapy and risk of falls. We used the random-effects inverse-variance method to account for study heterogeneity and conducted separate analyses for OR and HR.40 A limitation to our study was the inclusion of published articles only written in English. In conclusion, there is no significant increase in risk of falls for chronic use of antihypertensive medication. The risk of falling is highest 0 to 24 hours after antihypertensive medication initiation, change, or dose increase (day 0) and remains significantly elevated between days 1 to 21 for diuretics. We recommend that health care practitioners exercise extra vigilance and initiate strategies to reduce the risk of falling in older adults prior to and for these periods. KEY POINTS Question Is there any association between acute (<28days) or chronic (≥28days) use of antihypertensive medications and falls in older adults (age 60 years and older)? Findings The risk of falling is highest during acute antihypertensive use. OR was highest 0 to 24 hours after antihypertensive medication initiation, change, or dose increase (day 0). For diuretics, the OR remained significantly elevated until day 21. There was no association between chronic use of antihypertensive medications and falls in older adults. Meaning Strategies to reduce the risk of falling should be implemented prior to commencing antihypertensive medications in older adults. DISCLOSURE The authors declared no conflict of interest. ACKNOWLEDGMENT We would like to thank Lorraine Evison, the Faculty Liaison Librarian at The University of Sydney Medical Sciences Libraries for her support in building the search strategy for this review and Dr Barbara Mintzes, senior lecturer at the University of Sydney for her support and advice on acquiring proper statistical methods, and Associate Professor Dr Amanda Salis NHMRC Senior Research Fellow at the University of Sydney for her help in manuscript writing and editing. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. REFERENCES 1. Heinrich S Rapp K Rissmann U Becker C König HH. Cost of falls in old age: a systematic review. Osteoporos Int  2010; 21: 891– 902. Google Scholar CrossRef Search ADS PubMed  2. Lord SR March LM Cameron ID Cumming RG Schwarz J Zochling J Chen JS Makaroff J Sitoh YY Lau TC Brnabic A Sambrook PN. Differing risk factors for falls in nursing home and intermediate-care residents who can and cannot stand unaided. J Am Geriatr Soc  2003; 51: 1645– 1650. Google Scholar CrossRef Search ADS PubMed  3. Leipzig RM Cumming RG Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis: I. Psychotropic drugs. J Am Geriatr Soc  1999; 47: 30– 39. Google Scholar CrossRef Search ADS PubMed  4. Aronow WS Frishman WH Cheng-Lai A. Cardiovascular drug therapy in the elderly. Cardiol Rev  2007; 15: 195– 215. Google Scholar CrossRef Search ADS PubMed  5. Berry SD Mittleman MA Zhang Y Solomon DH Lipsitz LA Mostofsky E Goldense D Kiel DP. New loop diuretic prescriptions may be an acute risk factor for falls in the nursing home. Pharmacoepidemiol Drug Saf  2012; 21: 560– 563. Google Scholar CrossRef Search ADS PubMed  6. Butt DA Mamdani M Austin PC Tu K Gomes T Glazier RH. The risk of hip fracture after initiating antihypertensive drugs in the elderly. Arch Intern Med  2012; 172: 1739– 1744. Google Scholar CrossRef Search ADS PubMed  7. Butt DA Mamdani M Austin PC Tu K Gomes T Glazier RH. The risk of falls on initiation of antihypertensive drugs in the elderly. Osteoporos Int  2013; 24: 2649– 2657. Google Scholar CrossRef Search ADS PubMed  8. Callisaya ML Sharman JE Close J Lord SR Srikanth VK. Greater daily defined dose of antihypertensive medication increases the risk of falls in older people–a population-based study. J Am Geriatr Soc  2014; 62: 1527– 1533. Google Scholar CrossRef Search ADS PubMed  9. Payne RA Abel GA Simpson CR Maxwell SR. Association between prescribing of cardiovascular and psychotropic medications and hospital admission for falls or fractures. Drugs Aging  2013; 30: 247– 254. Google Scholar CrossRef Search ADS PubMed  10. Shimbo D Barrett Bowling C Levitan EB Deng L Sim JJ Huang L Reynolds K Muntner P. Short-Term Risk of Serious Fall Injuries in Older Adults Initiating and Intensifying Treatment With Antihypertensive Medication. Circ Cardiovasc Qual Outcomes  2016; 9: 222– 229. Google Scholar CrossRef Search ADS PubMed  11. Tinetti ME Han L Lee DS McAvay GJ Peduzzi P Gross CP Zhou B Lin H. Antihypertensive medications and serious fall injuries in a nationally representative sample of older adults. JAMA Intern Med  2014; 174: 588– 595. Google Scholar CrossRef Search ADS PubMed  12. Smith AD. The worldwide challenge of the dementias: a role for B vitamins and homocysteine? Food Nutr Bull  2008; 29: S143– S172. Google Scholar CrossRef Search ADS PubMed  13. Morgan S Smolina K Mooney D Raymond C Bowen ML Gorczynski C Basham KAR. The Canadian Rx atlas , 3rd edition. 2013; <https://open.library.ubc.ca/collections/52383/items/1.0048514>. Accessed 21 March 2017. 14. Hartikainen S Lönnroos E Louhivuori K. Medication as a risk factor for falls: critical systematic review. J Gerontol A Biol Sci Med Sci  2007; 62: 1172– 1181. Google Scholar CrossRef Search ADS PubMed  15. Wiens M Etminan M Gill SS Takkouche B. Effects of antihypertensive drug treatments on fracture outcomes: a meta-analysis of observational studies. J Intern Med  2006; 260: 350– 362. Google Scholar CrossRef Search ADS PubMed  16. Woolcott JC Richardson KJ Wiens MO Patel B Marin J Khan KM Marra CA. Meta-analysis of the impact of 9 medication classes on falls in elderly persons. Arch Intern Med  2009; 169: 1952– 1960. Google Scholar CrossRef Search ADS PubMed  17. Zang G. Antihypertensive drugs and the risk of fall injuries: a systematic review and meta-analysis. J Int Med Res  2013; 41: 1408– 1417. Google Scholar CrossRef Search ADS PubMed  18. Leipzig RM Cumming RG Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis: II. Cardiac and analgesic drugs. J Am Geriatr Soc  1999; 47: 40– 50. Google Scholar CrossRef Search ADS PubMed  19. Askari M Eslami S Scheffer AC Medlock S de Rooij SE van der Velde N Abu-Hanna A. Different risk-increasing drugs in recurrent versus single fallers: are recurrent fallers a distinct population? Drugs Aging  2013; 30: 845– 851. Google Scholar CrossRef Search ADS PubMed  20. Eriksson S Gustafson Y Lundin-Olsson L. Risk factors for falls in people with and without a diagnose of dementia living in residential care facilities: a prospective study. Arch Gerontol Geriatr  2008; 46: 293– 306. Google Scholar CrossRef Search ADS PubMed  21. Gribbin J Hubbard R Gladman J Smith C Lewis S. Risk of falls associated with antihypertensive medication: self-controlled case series. Pharmacoepidemiol Drug Saf  2011; 20: 879– 884. Google Scholar CrossRef Search ADS PubMed  22. Pariente A Dartigues JF Benichou J Letenneur L Moore N Fourrier-Réglat A. Benzodiazepines and injurious falls in community dwelling elders. Drugs Aging  2008; 25: 61– 70. Google Scholar CrossRef Search ADS PubMed  23. Shuto H Imakyure O Matsumoto J Egawa T Jiang Y Hirakawa M Kataoka Y Yanagawa T. Medication use as a risk factor for inpatient falls in an acute care hospital: a case-crossover study. Br J Clin Pharmacol  2010; 69: 535– 542. Google Scholar CrossRef Search ADS PubMed  24. Stenhagen M Ekstrom H Nordell E Elmstahl S. Falls in the general elderly population: A 3-and 6-year follow-up of risk factors using data from the longitudinal population study “good ageing in Skane”. Eur Geriatr Med  2013; 3: S53. Google Scholar CrossRef Search ADS   25. Moher D Shamseer L Clarke M Ghersi D Liberati A Petticrew M Shekelle P Stewart LA; PRISMA-P Group. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev  2015; 4: 1. Google Scholar CrossRef Search ADS PubMed  26. ATC and DDD Index. WHO Collaborating Centre for Drug Statistics Methodology, Oslo ATC Code: Anatomical Therapeutic Chemical Classification System.<https://www.whocc.no/atc_ddd_index/>. Accessed 20 March 2017. 27. Hasegawa J Kuzuya M Iguchi A. Urinary incontinence and behavioral symptoms are independent risk factors for recurrent and injurious falls, respectively, among residents in long-term care facilities. Arch Gerontol Geriatr  2010; 50: 77– 81. Google Scholar CrossRef Search ADS PubMed  28. Sterke CS van Beeck EF van der Velde N Ziere G Petrovic M Looman CW van der Cammen TJ. New insights: dose-response relationship between psychotropic drugs and falls: a study in nursing home residents with dementia. J Clin Pharmacol  2012; 52: 947– 955. Google Scholar CrossRef Search ADS PubMed  29. Wong AK Lord SR Sturnieks DL Delbaere K Trollor JN Close JC. Angiotensin system-blocking medications are associated with fewer falls over 12 months in community-dwelling older people. J Am Geriatr Soc  2013; 61: 776– 781. Google Scholar CrossRef Search ADS PubMed  30. Tinetti ME Williams TF Mayewski R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med  1986; 80: 429– 434. Google Scholar CrossRef Search ADS PubMed  31. World Health Organization, WHO Falls fact sheet. <http://www.who.int/mediacentre/factsheets/fs344/en/>. Accessed 10 September 2017. 32. Lamb SE Jørstad-Stein EC Hauer K Becker C. Development of a common outcome data set for fall injury prevention trials: the Prevention of Falls Network Europe consensus. J Am Geriatr Soc  2005; 53: 1618– 1622. Google Scholar CrossRef Search ADS PubMed  33. Kellogg IWG. The prevention of falls in later life. A report of the Kellogg International Work Group on the Prevention of Falls by the Elderly. Dan Med Bull  1987; 34: 1– 24. 34. Wells GA Shea B O’Connell D Peterson J Welch D Losos D Tugwell D. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. 2014; <http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp>. Accessed 21 March 2017. 35. van Wijngaarden JP Dhonukshe-Rutten RA van Schoor NM van der Velde N Swart KM Enneman AW van Dijk SC Brouwer-Brolsma EM Zillikens MC van Meurs JB Brug J Uitterlinden AG Lips P de Groot LC. Rationale and design of the B-PROOF study, a randomized controlled trial on the effect of supplemental intake of vitamin B12 and folic acid on fracture incidence. BMC Geriatr  2011; 11: 1– 11. Google Scholar CrossRef Search ADS PubMed  36. Hulley SB Furberg CD Gurland B McDonald R Perry HM Schnaper HW Schoenberger JA Smith WM Vogt TM. Systolic Hypertension in the Elderly Program (SHEP): antihypertensive efficacy of chlorthalidone. Am J Cardiol  1985; 56: 913– 920. Google Scholar CrossRef Search ADS PubMed  37. Ham AC Swart KM Enneman AW van Dijk SC Oliai Araghi S van Wijngaarden JP van der Zwaluw NL Brouwer-Brolsma EM Dhonukshe-Rutten RA van Schoor NM van der Cammen TJ Lips P de Groot LC Uitterlinden AG Witkamp RF Stricker BH van der Velde N. Medication-related fall incidents in an older, ambulant population: the B-PROOF study. Drugs Aging  2014; 31: 917– 927. Google Scholar CrossRef Search ADS PubMed  38. Charlesworth CJ Peralta CA Odden MC. Functional Status and Antihypertensive Therapy in Older Adults: A New Perspective on Old Data. Am J Hypertens  2016; 29: 690– 695. Google Scholar CrossRef Search ADS PubMed  39. Stroup DF Berlin JA Morton SC Olkin I Williamson GD Rennie D Moher D Becker BJ Sipe TA Thacker SB. Meta-analysis of observational studies in epidemiology: a proposal for reporting. JAMA  2000; 283: 2008– 2012. Google Scholar CrossRef Search ADS PubMed  40. Egger M Davey-Smith G Altman D. Systematic reviews in health care: meta-analysis in context . London: John Wiley & Sons; 2008. 41. DerSimonian R Kacker R. Random-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials  2007; 28: 105– 114. Google Scholar CrossRef Search ADS PubMed  42. DerSimonian R Laird N. Meta-analysis in clinical trials. Control Clin Trials  1986; 7: 177– 188. Google Scholar CrossRef Search ADS PubMed  43. Baranzini F Diurni M Ceccon F Poloni N Cazzamalli S Costantini C Colli C Greco L Callegari C. Fall-related injuries in a nursing home setting: is polypharmacy a risk factor? BMC Health Serv Res  2009; 9: 228. Google Scholar CrossRef Search ADS PubMed  44. Duh MS Mody SH Lefebvre P Woodman RC Buteau S Piech CT. Anaemia and the risk of injurious falls in a community-dwelling elderly population. Drugs Aging  2008; 25: 325– 334. Google Scholar CrossRef Search ADS PubMed  45. Lipsitz LA Habtemariam D Gagnon M Iloputaife I Sorond F Tchalla AE Dantoine TF Travison TG. Reexamining the Effect of Antihypertensive Medications on Falls in Old Age. Hypertension  2015; 66: 183– 189. Google Scholar CrossRef Search ADS PubMed  46. Montali F Campaniello G Benatti M Rastelli G Pedrazzoni M Cervellin G. Impact of different drug classes on clinical severity of falls in an elderly population: epidemiological survey in a trauma center. Journal of Clinical Gerontology and Geriatrics  2015; 6: 63– 67. Google Scholar CrossRef Search ADS   47. Rafiq M McGovern A Jones S Harris K Tomson C Gallagher H de Lusignan S. Falls in the elderly were predicted opportunistically using a decision tree and systematically using a database-driven screening tool. J Clin Epidemiol  2014; 67: 877– 886. Google Scholar CrossRef Search ADS PubMed  48. Sterke CS Ziere G van Beeck EF Looman CW van der Cammen TJ. Dose-response relationship between selective serotonin re-uptake inhibitors and injurious falls: a study in nursing home residents with dementia. Br J Clin Pharmacol  2012; 73: 812– 820. Google Scholar CrossRef Search ADS PubMed  49. Thorell K Ranstad K Midlöv P Borgquist L Halling A. Is use of fall risk-increasing drugs in an elderly population associated with an increased risk of hip fracture, after adjustment for multimorbidity level: a cohort study. BMC Geriatr  2014; 14: 131. Google Scholar CrossRef Search ADS PubMed  50. Garcia Molina O Olmos Jimenez R Castellote De Varona FJ Mendoza Otero F Arocas Casano V De La Rubia Nieto MA. Polypharmacy related to increased risk of hip fracture in elderly patients. Atencion Farmaceutica  2014; 16: 117– 123. 51. Rhalimi M Helou R Jaecker P. Medication use and increased risk of falls in hospitalized elderly patients: a retrospective, case-control study. Drugs Aging  2009; 26: 847– 852. Google Scholar CrossRef Search ADS PubMed  52. Zia A Kamaruzzaman SB Myint PK Tan MP. The association of antihypertensives with postural blood pressure and falls among seniors residing in the community: a case-control study. Eur J Clin Invest  2015; 45: 1069– 1076. Google Scholar CrossRef Search ADS PubMed  53. Zia A Kamaruzzaman SB Tan MP. The consumption of two or more fall risk-increasing drugs rather than polypharmacy is associated with falls. Geriatr Gerontol Int  2017; 17: 463– 470. Google Scholar CrossRef Search ADS PubMed  54. Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res  2011; 46: 399– 424. Google Scholar CrossRef Search ADS PubMed  55. Johnston SC. Identifying confounding by indication through blinded prospective review. Am J Epidemiol  2001; 154: 276– 284. Google Scholar CrossRef Search ADS PubMed  56. Salas M Hofman A Stricker BH. Confounding by indication: an example of variation in the use of epidemiologic terminology. Am J Epidemiol  1999; 149: 981– 983. Google Scholar CrossRef Search ADS PubMed  57. Onder G Vedova CD Pahor M. Effects of ACE inhibitors on skeletal muscle. Curr Pharm Des  2006; 12: 2057– 2064. Google Scholar CrossRef Search ADS PubMed  58. Higgins JP. Commentary: Heterogeneity in meta-analysis should be expected and appropriately quantified. Int J Epidemiol  2008; 37: 1158– 1160. Google Scholar CrossRef Search ADS PubMed  59. Higgins JP Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med  2002; 21: 1539– 1558. Google Scholar CrossRef Search ADS PubMed  © American Journal of Hypertension, Ltd 2017. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Hypertension Oxford University Press

Association Between Chronic or Acute Use of Antihypertensive Class of Medications and Falls in Older Adults. A Systematic Review and Meta-Analysis

Loading next page...
 
/lp/ou_press/association-between-chronic-or-acute-use-of-antihypertensive-class-of-22gelf6868
Publisher
Oxford University Press
Copyright
© American Journal of Hypertension, Ltd 2017. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0895-7061
eISSN
1941-7225
D.O.I.
10.1093/ajh/hpx189
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND Evaluating effect of acute or chronic use of antihypertensives on risk of falls in older adults. METHODS Data sources: Systematic search of primary research articles in CINAHL, Cochrane, EBM, EMBASE, and MEDLINE databases from January 1 2007 to June 1 2017. Study selection: Research studies of cohort, case-control, case-crossover, cross-sectional, or randomized controlled trial (RCT) design examining association between antihypertensives and falls in people older than 60 years were evaluated. Data synthesis: Twenty-nine studies (N = 1,234,667 participants) were included. Study quality was assessed using the Newcastle–Ottawa Scale (NOS). PRISMA and MOOSE guidelines were used for abstracting data and random-effects inverse-variance meta-analysis was conducted on 26 articles examining chronic antihypertensive use, with odds ratios (ORs) and hazards ratios (HRs) analyzed separately. Time-risk analysis was performed on 5 articles examining acute use of antihypertensives. Outcomes: Pooled ORs and HRs were calculated to determine the association between chronic antihypertensive use and falls. For time-risk analysis, OR was plotted with respect to number of days since antihypertensive commencement, change, or dose increase. RESULTS There was no significant association between risk of falling and chronic antihypertensive medication use (OR = 0.97, 95% confidence interval [CI] 0.93–1.01, I2 = 64.1%, P = 0.000; and HR = 0.96, 95% CI 0.92–1.00, I2 = 0.0%, P = 0.706). The time-risk analysis demonstrated a significantly elevated risk of falling 0–24 hours after antihypertensive initiation, change, or dose increase. When diuretics were used, the risk remained significantly elevated till day 21. CONCLUSIONS There is no significant association between chronic use of antihypertensives and falls in older adults. Risk of falls is highest on day zero for all antihypertensive medications. accidental falls, older adults, antihypertensives (agents), blood pressure, geriatrics, hypertension, meta-analysis Falls are a major risk for older adults and are associated with increased morbidity and mortality. Between 0.85 and 1.5% of total health care expenditure is consumed on fall-related expenses in the United States, Europe, and Australia.1 A number of widely prescribed medications have been shown to be significant contributors to falls and fractures.2 While there are many risk factors for falls in older adults, none are potentially as preventable or reversible as medication use.3 Cardiovascular medications are the most commonly used medicines among older adults4 and have been identified as one of the main risk factors for falls in many studies.5–11 Importantly, the prescription of medications to older adults has increased significantly over the last 2 decades.11–13 There have been several systematic and literature reviews14–18 examining the effect of different medications on falls since the meta-analysis by Leipzig et al.18 using articles publishes between 1966 and 1996. Hartikainen et al.14 performed a systematic review of studies published between 1996 and 2004 but did not pool the data for meta-analysis. Wiens et al.15 examined the effect of antihypertensive medications on fractures, meta-analyzing articles published between 1996 and 2005. Woolcott et al.16 updated Leipzig et al.18 using articles published between April 1996 and August 2007 and assessed the impact of 9 medication classes on falls in older adults. Despite the number of studies, fundamental differences in study design have hampered meta-analysis. Zang et al.17 combined the results of Leipzig et al.18 and Wiens et al.15 More recent studies have suggested that time after the initiation, change, or increase in dose of antihypertensive medication is a significant predictor of falls.5,6,8,10,11,19–24 The aim of our systematic review was to update previous meta-analyses16,18 in light of recent research and to examine the association between falls and antihypertensive medication initiation, change, or dose increase in older adults. METHODS Data sources Original research studies were collected through a systematic search of English language articles in CINAHL, Cochrane, EBM, EMBASE, and MEDLINE databases. Primary research articles published between 1 January 2007 and 1 June 2017 were eligible for inclusion. We combined the MeSH term “Therapeutic uses”, which includes all indexed classes of drugs and individual agents, with the MeSH terms “Accidents, Home”, and “Accidental Falls”. In addition, the MeSH terms “Epidemiology” and “Pharmacoepidemiology” were incorporated into the systematic search to find studies in which drug exposure may have been the secondary objective of the research. The MeSH terms “Cardiovascular Agents” and “Antihypertensive Agents”, “Diuretics”, “Vasodilator Agents”, “Nitrates”, “Adrenergic beta-Antagonist” (BB), “Calcium Channel Blockers” (CCB), “Angiotensin-Converting Enzyme Inhibitors” (ACE-i), “Angiotensin Receptor Antagonists” (ARB), and agents acting on the “Renin–Angiotensin System” were associated with MeSH terms “Falls” or “Falling”, or “Fractures”. All terms were expanded to cover all headings and subheadings and find potentially appropriate and relevant studies. References of selected studies were checked for other potentially eligible studies (Figure 1). Based on Equator reporting guidelines, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-P)25 was implemented for the review and selection process. Figure 1. View largeDownload slide PRISMA flowchart. Figure 1. View largeDownload slide PRISMA flowchart. Study selection, inclusion, and exclusion criteria Studies were included if they were primary research studies providing original statistical data of cohort, case–control, case-crossover, cross-sectional, or randomized controlled trial (RCT) design in which the association between use of antihypertensive medications and falls, injurious falls, and fractures in people older than 60 years were evaluated. Antihypertensive medications included all medication that had a primary therapeutic indication of reducing blood pressure. This was defined as medications within Anatomical Therapeutic Chemical (ATC) codes26 C02 (antihypertensive), C03 (diuretic), C07 (BB), C08 (CCB), C09 (ACE-i, ARB, and RAS). Articles covering individual subgroups of nitrates and vasodilators (ATC codes C01, C04) have been excluded as these medication classes are not primarily indicated for hypertension. An operational definition for chronic use was conservatively set at medication therapy for at least 28 days, to account for pharmacokinetic, and pharmacodynamics of included medications. Articles covering temporal relationships between medication change or dose increase and falls were included for time-risk analysis.5,7,10,21,23 Data extraction Article titles, key words, indexed terms, abstracts, and full texts were initially screened by H.R.K., with inclusion decided in consensus with C.R.S. and M.D.L. In addition, 10% of the excluded articles were randomly checked by C.R.S. to confirm that they did not meet the inclusion criteria. In order to be included in the meta-analysis or time-risk analysis, effect sizes needed to be reported as odds ratio (OR), relative risk (RR), incidence risk ratio (IRR), standard incidence ratio, or hazard ratio (HR). The adjusted effect sizes and 95% confidence intervals (CI) were used for analysis. When adjusted effect sizes were not reported, crude effect sizes were used.20,27–29 If the reported data did not meet these criteria or clarification was required to assess eligibility, the corresponding authors were contacted. If the articles only provided 2 × 2 tables of fall incidents and exposure, the effect sizes were calculated from the available data. Information extracted from the compiled studies contained study type, study setting (hospital, nursing home, or community dwelling), country of study, mean age, demographic data, period of study, sample size, ATC codes, fall definition, types of medication, exposure period, type of effect size, confounders, medication duration, fall ascertainment method, temporal relationship between medication initiation, change or dose increase and fall, fall risk index,30 and study design (Table 1). Table 1. Summary of the research studies included in the meta-analysis and time-risk analysis. Source  Setting  Publication date, Data collection, Follow up period  Sample size  Age, Mean(SD),eYrs.  Type of Fall  Medicationa  ATC Codes b   Time of Medication Ascertainment  Method of Fall Ascertainment  Study design  NOS results  Askari et al. 19   Hospital (ED) c  2013/72(mon)f  2258  77.7(7.8)  Falls  C, D, BB  C01,C03, C07  Medication self-report  Incident report  Cross- sectional  10/10  Baranzini et al. 43   L/N d  2009/42(mon)  293  84.4(8.2)  Fall injuries  Ahy, D  C, C03  Baseline  Incident report  Cohort  8/9  Berry et al. 5   L/N  201/65(mon)  1191  88(8)  Falls  D  C03  Computerized order entry system  Computerized incident report  Case-cross over  7/9  Butt et al.(1) 6   Community  2012/108(mon)/ 450days  301591  81(7.3)  Hip fracture  Ahy, Tz, BB, CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline,  Baseline  Case-cross over  7/9  Butt et al.(2) 7   Community  2013/108(mon)/ 450days  543572  80(7.6)  Falls  Ahy  C  Baseline, extracted from DB  Baseline: DB  Case-cross over  7/9  Callisaya et al. 8   Community  2014/12(mon)/ 12 mon  409  72(6.9)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C, C03, C07, C08, C09  Baseline, face to face interview/ prescription check  Daily fall calendar, Incident report, bimonthly questionnaire  Cohort  9/9  Charlesworth et al. 38   Community  2015/72(mon)/ 4.5 yrs.  4736  71.7(6.4)  Falls  Ahy  C  Baseline, face to face interview/prescription check  Self-report  Cohort (RCT)  9/9  Duh et al. 44   L/N  2008/64(mon)/ More than five yrs.  47530  75.4(3.1)  Injurious falls  Ahy  C  Baseline extracted from database  Injurious claim within 30 days after fall  Cohort  7/9  Eriksson et al. 20   L/N  2009/6(mon)/ 6 mon  186  83.6(6.6)  Falls  D, BB, ACE-i  C03, C07, C09  Baseline  Incident report  Cohort  8/9  Garcia et al. 50   Hospital  2014/12(mon)  122  83.3(4.8)  Hip fracture  D  C03  Baseline  Incident report  Case- Control  4/9  Gribbin et al. 21   Community  2010/48(mon)/ 3 yrs.  9682  77.5(.)  Falls  Tz, BB, CCB ACE-i, ARB  C03, C07 C08, C09  Baseline  Incident report  Case- Control  7/9  Ham et al. 37   Community  2014/32(mon)/ -2 to -3 yrs.  2407  74.4(6.7)  Falls  Ahy, D, BB, CCB, ACE- i,, ARB  C, C03, C07, C08, C09  Baseline prospective  Fall incident report  Cohort (RCT)  9/9  Hasegawa et al. 27   L/N  2009/12(mon)/ 6 mon  1082  82.5(8.5)  Injurious falls  Ahy, ACE-i, CCB  C, C08, C09  Baseline  Incident report  Cohort  9/9  Lipsitz et al. 45   Community  2015/12(mon)/ 1 yr.  598  78.4(5.4)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline: Interview 2 weeks recall  Self-report, Monthly calendar  Cohort  9/9  Montali et al. 46   Hospital(ED)  2015/12(mon)  2377  81.2(8)  Falls  Ahy  C  Interview reports  Baseline, from Database  Cohort  8/9  Pariente et al. 22   Community  2008/120(mon)/ 10 yrs.  3777  79.9(6.8)  Injurious falls  Ahy,  C  Interview  Baseline, Incident report  Case- Control  7/9  Payne et al. 9   Hospital  2013/12(mon)/ 1 yr.  39813  (NA)  Falls or fractures  CV  C  Baseline  Incident report  Case- Control  8/9  Rafiq et al. 47   Community  2014/60(mon)/ 30 mon  135433  75.4(7.6)  Falls  Ahy, ACE-i  C, C09  Baseline(from Database)  Baseline(from Database)  Cohort  8/9  Rhalimi et al. 51   Hospital  2009/24(mon)  260  89(7)  Falls  CCB  C08  Baseline  Incident report  Case-control  6/9  Shimbo et al. 10   Hospital(ED)  2016/66(mon)/ 365 days  90127  (NA)  Serious fall injuries  Ahy, ACE- i, ARB, BB,CCB,D  C, C03, C07,C08, C09  Baseline prescription  Baseline- ED ,inpatient claims  Case-cross over  8/9  Shuto et al. 23   Hospital  2009/30(mon)  349  71.5(14.8)  Falls  Ahy, ARB  C, C09  Baseline  Incident report  Case-cross over  7/9  Stenhagen et al. 24   Community  2013/72(mon)/ 3 yrs.  1763  (NA)  Falls  Ahy, D  C, C03  Baseline  Interview(6 Months recall)  Cohort  9/9  Sterke et al.(1) 28   L/N  2012/24(mon)/ 350 days  248  82(8)  Falls  Ahy, BB,  C, C07  Baseline  Incident report  Cohort  8/9  Sterke et al.(2) 48   L/N  2012/24(mon)/ 350 days  248  82(8)  Injurious falls  Ahy, BB,  C, C07,  Baseline  Incident report  Cohort  7/9  Thorel et al. 49   Community  2014/24(mon)  38407  (NA)  Hip fracture  Ahy, D, BB, CCB, RAS,  C, C03, C07, C08,C09  Baseline  Incident report  Cohort  8/9  Tinetti et al. 11   Community  2014/36(mon)/ 3 yrs.  4961  80.2(6.8)  Serious fall injuries  D, BB, CCB, RAS  C03, C07, C08, C09  Interview-direct observation  Incident report  Cohort  8/9  Wong et al. 29   Community  2013/17(mon)/ 12 mon  531  79.9(4.4)  Falls  D, Tz, BB, CV, ACE-i, ARB, RAS  C, C03, C07, C09,  Baseline  Fall calendar (monthly)  Cohort  9/9  Zia et al.(1) 52   Community  2015/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Zia et al.(2) 53   Community  2016/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Source  Setting  Publication date, Data collection, Follow up period  Sample size  Age, Mean(SD),eYrs.  Type of Fall  Medicationa  ATC Codes b   Time of Medication Ascertainment  Method of Fall Ascertainment  Study design  NOS results  Askari et al. 19   Hospital (ED) c  2013/72(mon)f  2258  77.7(7.8)  Falls  C, D, BB  C01,C03, C07  Medication self-report  Incident report  Cross- sectional  10/10  Baranzini et al. 43   L/N d  2009/42(mon)  293  84.4(8.2)  Fall injuries  Ahy, D  C, C03  Baseline  Incident report  Cohort  8/9  Berry et al. 5   L/N  201/65(mon)  1191  88(8)  Falls  D  C03  Computerized order entry system  Computerized incident report  Case-cross over  7/9  Butt et al.(1) 6   Community  2012/108(mon)/ 450days  301591  81(7.3)  Hip fracture  Ahy, Tz, BB, CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline,  Baseline  Case-cross over  7/9  Butt et al.(2) 7   Community  2013/108(mon)/ 450days  543572  80(7.6)  Falls  Ahy  C  Baseline, extracted from DB  Baseline: DB  Case-cross over  7/9  Callisaya et al. 8   Community  2014/12(mon)/ 12 mon  409  72(6.9)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C, C03, C07, C08, C09  Baseline, face to face interview/ prescription check  Daily fall calendar, Incident report, bimonthly questionnaire  Cohort  9/9  Charlesworth et al. 38   Community  2015/72(mon)/ 4.5 yrs.  4736  71.7(6.4)  Falls  Ahy  C  Baseline, face to face interview/prescription check  Self-report  Cohort (RCT)  9/9  Duh et al. 44   L/N  2008/64(mon)/ More than five yrs.  47530  75.4(3.1)  Injurious falls  Ahy  C  Baseline extracted from database  Injurious claim within 30 days after fall  Cohort  7/9  Eriksson et al. 20   L/N  2009/6(mon)/ 6 mon  186  83.6(6.6)  Falls  D, BB, ACE-i  C03, C07, C09  Baseline  Incident report  Cohort  8/9  Garcia et al. 50   Hospital  2014/12(mon)  122  83.3(4.8)  Hip fracture  D  C03  Baseline  Incident report  Case- Control  4/9  Gribbin et al. 21   Community  2010/48(mon)/ 3 yrs.  9682  77.5(.)  Falls  Tz, BB, CCB ACE-i, ARB  C03, C07 C08, C09  Baseline  Incident report  Case- Control  7/9  Ham et al. 37   Community  2014/32(mon)/ -2 to -3 yrs.  2407  74.4(6.7)  Falls  Ahy, D, BB, CCB, ACE- i,, ARB  C, C03, C07, C08, C09  Baseline prospective  Fall incident report  Cohort (RCT)  9/9  Hasegawa et al. 27   L/N  2009/12(mon)/ 6 mon  1082  82.5(8.5)  Injurious falls  Ahy, ACE-i, CCB  C, C08, C09  Baseline  Incident report  Cohort  9/9  Lipsitz et al. 45   Community  2015/12(mon)/ 1 yr.  598  78.4(5.4)  Falls  Ahy, D, BB,RAS CCB, ACE-i, ARB  C,C03,C07, C08, C09  Baseline: Interview 2 weeks recall  Self-report, Monthly calendar  Cohort  9/9  Montali et al. 46   Hospital(ED)  2015/12(mon)  2377  81.2(8)  Falls  Ahy  C  Interview reports  Baseline, from Database  Cohort  8/9  Pariente et al. 22   Community  2008/120(mon)/ 10 yrs.  3777  79.9(6.8)  Injurious falls  Ahy,  C  Interview  Baseline, Incident report  Case- Control  7/9  Payne et al. 9   Hospital  2013/12(mon)/ 1 yr.  39813  (NA)  Falls or fractures  CV  C  Baseline  Incident report  Case- Control  8/9  Rafiq et al. 47   Community  2014/60(mon)/ 30 mon  135433  75.4(7.6)  Falls  Ahy, ACE-i  C, C09  Baseline(from Database)  Baseline(from Database)  Cohort  8/9  Rhalimi et al. 51   Hospital  2009/24(mon)  260  89(7)  Falls  CCB  C08  Baseline  Incident report  Case-control  6/9  Shimbo et al. 10   Hospital(ED)  2016/66(mon)/ 365 days  90127  (NA)  Serious fall injuries  Ahy, ACE- i, ARB, BB,CCB,D  C, C03, C07,C08, C09  Baseline prescription  Baseline- ED ,inpatient claims  Case-cross over  8/9  Shuto et al. 23   Hospital  2009/30(mon)  349  71.5(14.8)  Falls  Ahy, ARB  C, C09  Baseline  Incident report  Case-cross over  7/9  Stenhagen et al. 24   Community  2013/72(mon)/ 3 yrs.  1763  (NA)  Falls  Ahy, D  C, C03  Baseline  Interview(6 Months recall)  Cohort  9/9  Sterke et al.(1) 28   L/N  2012/24(mon)/ 350 days  248  82(8)  Falls  Ahy, BB,  C, C07  Baseline  Incident report  Cohort  8/9  Sterke et al.(2) 48   L/N  2012/24(mon)/ 350 days  248  82(8)  Injurious falls  Ahy, BB,  C, C07,  Baseline  Incident report  Cohort  7/9  Thorel et al. 49   Community  2014/24(mon)  38407  (NA)  Hip fracture  Ahy, D, BB, CCB, RAS,  C, C03, C07, C08,C09  Baseline  Incident report  Cohort  8/9  Tinetti et al. 11   Community  2014/36(mon)/ 3 yrs.  4961  80.2(6.8)  Serious fall injuries  D, BB, CCB, RAS  C03, C07, C08, C09  Interview-direct observation  Incident report  Cohort  8/9  Wong et al. 29   Community  2013/17(mon)/ 12 mon  531  79.9(4.4)  Falls  D, Tz, BB, CV, ACE-i, ARB, RAS  C, C03, C07, C09,  Baseline  Fall calendar (monthly)  Cohort  9/9  Zia et al.(1) 52   Community  2015/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Zia et al.(2) 53   Community  2016/12(mon)  358  75.2(7.1)  Falls  Ahy, D, BB, αB CCB, ACE-i, RAS  C, C03, C07, C08, C09  Baseline  Baseline-Enquiry of fall occurrence  Case-control  9/9  Abbreviations: a Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha-blocker; BB, Beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin-angiotensin system (RAS); Tz, thiazide, NOS, Newcastle-Ottawa Scale; b ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin-angiotensin system (ACE-i, RAS, ARB).c ED, emergency department; d L/N: Long-term care facility/Nursing home; T studies included in time-risk effect analysis; M-A Meta-analysis; e (SD), standard deviation; Yrs, Years; f mon, month; g Adjusted OR was provided via personal correspondence. View Large The definition of a “fall” used by individual studies was compared with the definitions provided by WHO,31 the Prevention of Falls Network Europe (ProFaNE)32 and Kellogg’s international working group.33 Assessment of study quality, risk of bias The 9-item scale for cohort, case–control, and case-crossover studies, and the 10-item scale for cross-sectional studies from the Newcastle–Ottawa scale (NOS)34 were used to assess study quality and risk of bias. Two RCTs, (B-PROOF),35 and (SHEP),36 followed cohorts37,38 of subjects to determine risk of falls. In addition to the quality assessment, methods of medication verification, and falls ascertainment were extracted from each article. Data synthesis and analysis Meta-analyses were performed on 6 medication classes, categorized by their ATC codes.26 If an ATC code, name or class of medication were not reported in the study, the effect sizes were included in the non–class-specified antihypertensive group (ATC code C). The meta-analysis was conducted in accordance with the Meta-analysis Of Observational Studies in Epidemiology Group (MOOSE) protocols.39 Articles examining acute (<28 days) use of antihypertensives were selected for time-risk analysis. Pooled OR was obtained from OR. In one instance when relative risk was reported, the author was contacted and adjusted OR was provided via personal correspondence, pooled HR was obtained from HR and IRR. Separate meta-analyses were conducted for OR and HR, following the recommendations of Egger et al.40 (Figures 2 and 3). Figure 2. View largeDownload slide Antihypertensive medication and falls, odds ratio meta-analysis, including prior evidence. Antihypertensive medications and falls, meta-analysis results. Odds ratios and 95% confidence intervals for each individual or pooled study effect-sizes, ATC codes: C, C02, C03, C07, C08, C09. Statistical analysis calculated by random-effect inverse-variance (REIV) model. Abbreviations: Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blockers; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CI, confidence interval; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive, Cardiovascular system; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 2. View largeDownload slide Antihypertensive medication and falls, odds ratio meta-analysis, including prior evidence. Antihypertensive medications and falls, meta-analysis results. Odds ratios and 95% confidence intervals for each individual or pooled study effect-sizes, ATC codes: C, C02, C03, C07, C08, C09. Statistical analysis calculated by random-effect inverse-variance (REIV) model. Abbreviations: Ahy, antihypertensive; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blockers; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CI, confidence interval; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. ATC code: The Anatomical Therapeutic Chemical (ATC) Classification System. C: Non-class specified antihypertensive, Cardiovascular system; C02: Antihypertensive; C03: Diuretic; C07: Beta-blocking agent; C08: Calcium channel blocker; C09: Agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 3. View largeDownload slide Antihypertensive medication and falls, hazard ratio meta-analysis. Antihypertensive medication and falls, meta-analysis results. Hazard ratios and 95% confidence interval (95% CI) for each individual or pooled study effect-size, ATC codes: C, C02, C03, C07, C08, C09. Abbreviations: Ahy, antihypertensives; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. Figure 3. View largeDownload slide Antihypertensive medication and falls, hazard ratio meta-analysis. Antihypertensive medication and falls, meta-analysis results. Hazard ratios and 95% confidence interval (95% CI) for each individual or pooled study effect-size, ATC codes: C, C02, C03, C07, C08, C09. Abbreviations: Ahy, antihypertensives; ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; αB, alpha blocker; BB, beta-blocker; CCB, calcium channel blocker; CT, cardiac therapy; CV, cardiovascular; D, diuretic; Ni, nitrate; RAS, renin–angiotensin system (RAS); Tz, thiazide. Owing to heterogeneity between studies in experimental design and sampling, a random-effects model was used with inverse-variance estimate of the weights, as recommended by DerSimonian and Laird.41,42 Subgroup sensitivity analyses with 95% CI based on Pearson’s heterogeneity Chi-square test were performed for type of study (cohort, case–control, case-crossover, cross-sectional), study setting (hospital, nursing home, or community-dwelling), and mean age of participants (≤75 years old or >75 years old) where heterogeneity was greater than 50% and statistically significant (P < 0.05) (Table 2). Data regarding initiation, change, or dose increase of medications were examined separately. All statistical analyses were conducted using STATA/IC version 14.2, Windows 64 bit (StataCorp, TX). Table 2. Association between use of antihypertensive medications and falls, subgroup sensitivity analysis   C (non–class-specified antihypertensives)  C08 (CCB)  C09 (ACE-i, ARB, RAS)    n, effect size (95% CI)  n, effect size (95% CI)  n, effect size (95% CI)  All studies (OR)  12, 0.90 (0.83, 0.98)  6, 0.95 (0.75, 1.15)  12, 0.98 (0.92, 1.05)  All studies (HR)a  7, 0.94 (0.86, 1.02)  4, 1.01 (0.90, 1.13)  7, 0.94 (0.86, 1.03)  Population (OR)  Community  8, 0.84 (0.68, 0.99)  5, 0.93 (0.74, 1.13)  12, 0.98 (0.92, 1.05)  L/N  1, 1.30 (0.60, 2.00)  0  0  Hospital  3, 1.02 (0.80, 1.25)  1, 2.35 (0.56, 4.14)  0  Mean age of study subjects (OR)  ≤75 years  3, 0.93 (0.91, 0.96)  2, 1.36 (0.96, 1.77)  5, 1.16 (0.99, 1.33)  >75 years  5, 0.88 (0.59, 1.17)  3, 0.90 (0.51, 1.29)  6, 0.96 (0.88, 1.04)  Study type (OR)        Cohort  7, 0.99 (0.79, 1.19)  3, 0.85 (0.56, 1.14)  8, 0.96 (0.86, 1.06)  Case–control  4, 0.82 (0.64, 1.00)  3, 1.13 (0.79, 1.48)  4, 1.02 (0.91, 1.13)  Cross-sectional  0  0  0  Case-crossover  1, 1.00 (0.80, 1.20)  0  0    C (non–class-specified antihypertensives)  C08 (CCB)  C09 (ACE-i, ARB, RAS)    n, effect size (95% CI)  n, effect size (95% CI)  n, effect size (95% CI)  All studies (OR)  12, 0.90 (0.83, 0.98)  6, 0.95 (0.75, 1.15)  12, 0.98 (0.92, 1.05)  All studies (HR)a  7, 0.94 (0.86, 1.02)  4, 1.01 (0.90, 1.13)  7, 0.94 (0.86, 1.03)  Population (OR)  Community  8, 0.84 (0.68, 0.99)  5, 0.93 (0.74, 1.13)  12, 0.98 (0.92, 1.05)  L/N  1, 1.30 (0.60, 2.00)  0  0  Hospital  3, 1.02 (0.80, 1.25)  1, 2.35 (0.56, 4.14)  0  Mean age of study subjects (OR)  ≤75 years  3, 0.93 (0.91, 0.96)  2, 1.36 (0.96, 1.77)  5, 1.16 (0.99, 1.33)  >75 years  5, 0.88 (0.59, 1.17)  3, 0.90 (0.51, 1.29)  6, 0.96 (0.88, 1.04)  Study type (OR)        Cohort  7, 0.99 (0.79, 1.19)  3, 0.85 (0.56, 1.14)  8, 0.96 (0.86, 1.06)  Case–control  4, 0.82 (0.64, 1.00)  3, 1.13 (0.79, 1.48)  4, 1.02 (0.91, 1.13)  Cross-sectional  0  0  0  Case-crossover  1, 1.00 (0.80, 1.20)  0  0  Pooled OR and HR. Stratified by population setting, type of study, and mean age of participants in OR and HR. ATC code: The Anatomical Therapeutic Chemical (ATC) classification system. C, non–class-specified antihypertensive; C02, antihypertensives; C03, diuretics; C07, beta-blocking agents; C08, calcium channel blockers; C09, agents acting on RAS (ACE-i, RAS, ARB). Statistical analysis: calculated by random-effects inverse-variance model (REIV). Abbreviations: ACE-i (ACE inhibitor), angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CI, confidence interval; ED, emergency department, ED were considered as community in the sensitivity analysis as the fall has occurred in the community; HR, hazard ratio; L/N, long-term care/nursing home; OR, odds ratio; HR, hazard ratio; n, number of effect-sizes in the meta-analysis; RAS, renin–angiotensin system. aHeterogeneity was not observed in HR plot, sensitivity analysis is to provide an updated comparison view between OR and HR studies. View Large RESULTS From 5,142 studies, 29 (N = 1,234,667 participants) met our inclusion criteria. Detailed characteristics of the 29 studies are presented in (Table 1). Among the articles included were 16 cohort,8,11,20,24,27–29,37,38,43–49 7 case–control,9,21,22,50–53 5 case-crossover,5–7,10,23 and 1 cross-sectional study.19 Of the 16 cohort studies, 8 were prospective studies with 6 months,20,27 1 year,8,28,29,45,48 and 3- to 6-year follow-up periods.24 Among studies providing a standard definition of a “fall”, 5 used the WHO,31 48,23,52,53 used the ProFaNE,32 and 328,37,48 used the Kellogg definition.33 Most of the studies adjusted their effect sizes for age, sex, and comorbidity. Only 2 studies10,29 using OR and 3 studies using HR20,27,28 presented crude effect sizes. Twenty-six articles involving chronic antihypertensive medication use were meta-analyzed. Heterogeneity was observed in OR forest plots. There was no heterogeneity for studies examining antihypertensive medication classes C02, C03, or C07. There was heterogeneity in ATC codes C, C08 and C09. For ATC C (non–class-specific hypertensive), a significantly reduced risk of falling was noted only in studies of community-dwelling older adults (OR = 0.84, 95% CI, 0.68–0.99) and subjects 75 years of age and younger (OR = 0.93, 95% CI, 0.91–0.96), leading to a reduced risk for ATC code C overall (pooled OR = 0.90, 95% CI, 0.83–0.98). For classes C08 (CCB) and C09 (ACE-i, ARB, and RAS) all effect sizes found no significant association between medications and falls. No significant differences between the OR or HR for different populations (i.e., residents in the community, hospital, or nursing home), mean age (≤75 or >75 years), or type of study (cohort, case–control, case-crossover, or cross-sectional) were noted. Although some minor variations in the stratified OR and HR were noticed, none of these were statistically significant (Table 2). Only one study11 in the meta-analysis used propensity score matching54 to control for confounding by indication,55,56 while most of the studies used multivariable modelling and reported adjusted OR or HR. Five studies5,7,10,21,23 investigated commencement of cardiovascular medication as a risk factor for falls and were included in time-risk analysis. Four studies had case-crossover and 1 study21 had case–control design. Studies varied in setting (community, hospital, or long-term residential facility). One of these studies7 presented the data as IRR with the remainder providing OR as effect size. Two articles10,21 were included in both meta-analysis and time-risk analysis (Table 1). Chronic hypertensive use Table 1 includes results for non–class-specified antihypertensive medication, diuretic, BB, CCB, and ACE-i. Figures 2 and 3 show the OR, HR, and 95% CI of each study, along with the random-effects inverse-variance pooled effect size of each class of medication. OR pooled by Leipzig et al.18 in fixed-effects models and by Woolcott et al.16 in Bayesian models were documented as sources of prior evidence without being included in our meta-analysis. Chronic antihypertensive use was not associated with falls (overall OR = 0.97, 95% CI, 0.93–1.01; overall HR = 0.96, 95% CI, 0.92–1.00). Similarly, when antihypertensive classes were examined, there was no significantly increased risk was associated with diuretics (OR = 1.1, 95% CI, 0.99–1.23; HR = 0.99, 95% CI, 0.83–1.14), BB (OR = 0.93, 95% CI, 0.89–0.98; HR = 0.99, 95% CI, 0.89–1.09), CCB (OR = 0.95, 95% CI, 0.75–1.15; HR = 1.01, 95% CI, 0.90–1.13), or ACE-i and ARB (OR = 0.98, 95% CI, 0.92–1.05; HR = 0.94, 95% CI, 0.86–1.03). For non–class-specific antihypertensives (ATC code C), the OR reported by Leipzig et al.18 (OR = 1.16, 95% CI, 0.87–1.55) and Woolcott et al.16 (OR = 1.26, 95% CI, 1.08–1.46) were significantly different (P = 0.010) to our pooled OR (OR = 0.90, 95% CI, 0.83–0.98). Our pooled effect size for BB use (OR = 0.93, 95% CI, 0.89–0.98) differed from the Bayesian estimate calculated by Woolcott et al.16 (OR = 1.14, 95% CI, 0.97–1.33), although the difference was not statistically significant (P = 0.816). Dose and risk Callisaya et al.8 demonstrated that higher daily defined dose (DDD) is independently associated with falls and with more than 3 DDD, OR of falling increased from 1.17 (95% CI 0.77–1.77) to 1.56 (95% CI 1.02–2.38). Tinetti et al.11 reported that the HR of falling increased from 2.17 (95% CI, 0.98–4.8) for moderate doses to 2.31 (95% CI, 1.01–5.29) for high doses of antihypertensives. Shimbo et al.10 investigated polypharmacy and demonstrated that increasing the number of classes of medication from 1 to 2 or more will increase the OR of falling from 1.31 (95% CI, 1.13–1.52) to 1.58 (95% CI, 1.17–2.13). Lipsitz et al.45 reported that higher doses of CCB were associated with a lower risk of falls compared with those not taking CCB (OR = 0.44, 95% CI, 0.24–0.82). In addition, higher doses of ACE-i associated with a significantly lower risk of falls occurring outside home (OR = 0.40, 95% CI, 0.18–0.92). Low or standard ACE-i doses were associated with a significant reduction in injurious falls (OR = 0.58, 95% CI, 0.34–0.99), while high doses were associated with a nonsignificant reduction (OR = 0.53, 95% CI, 0.27–1.04) compared with those not taking ACE-i.45 Time and risk Two studies21,23 found a highly elevated risk of falling in the first 0–24 hours (day 0) after initiation, increase of dose or change of antihypertensive medication (Figure 4) for non–class-specified antihypertensives, diuretics, BB, CCB, and ACE-i/ARB. The risk of falling on diuretics remained significantly elevated until day 21, with a trend toward being significantly elevated until day 28. The risk of falling for all other antihypertensive classes decreased to nonsignificant levels from day 1 onward. From day 28 onward, no antihypertensive medication was associated with an increased risk if falling. Figure 4. View largeDownload slide Time-effect analysis: association of antihypertensive medication and falls. Abbreviations: ATC code, The Anatomical Therapeutic Chemical (ATC) Classification System; C, non–class specified antihypertensive; Cardiovascular system: C02, antihypertensive; C03, diuretic; C07, beta-blocking agent; C08, calcium channel blocker; C09, agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 4. View largeDownload slide Time-effect analysis: association of antihypertensive medication and falls. Abbreviations: ATC code, The Anatomical Therapeutic Chemical (ATC) Classification System; C, non–class specified antihypertensive; Cardiovascular system: C02, antihypertensive; C03, diuretic; C07, beta-blocking agent; C08, calcium channel blocker; C09, agents acting on renin–angiotensin system (ACE-i, RAS, ARB). Figure 5. View largeDownload slide Publication bias assessed via funnel plot, association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 5. View largeDownload slide Publication bias assessed via funnel plot, association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 6. View largeDownload slide Publication bias assessed by funnel plot: association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. Figure 6. View largeDownload slide Publication bias assessed by funnel plot: association between antihypertensive medication and falls (OR studies). Abbreviation: OR, odds ratio. DISCUSSION Our study found that the risk of falling was influenced by duration of antihypertensive treatment, being significantly elevated and highest during the first 24 hours (day 0) following medication initiation, change, or dose increase, with the odds increasing up to 36 times.21 Chronic antihypertensive medication use (≥28 days) was not associated with an increased risk of falling in older adults. When the diuretic class (ATC code 03) was examined, the risk of falling remained significantly elevated until day 21 and nonsignificantly elevated in chronic use. Chronic BB use appeared to be associated with a significantly reduced risk of falling. These findings suggest that a first dose effect may contribute to falls in susceptible older adults, possibly due to substantial drops in blood pressure and orthostatic hypotension. Antihypertensive medications that cause hypovolemia, bradycardia, and arterial or venous dilatation may predispose to this phenomenon. For diuretics, however, it takes longer, possibly up to 28 days, to decline to the low-risk and no-risk zone. This extended duration of risk may be due to a diuresis effect. Our pooled OR for BB use was nonsignificantly lower than the Bayesian estimate calculated by Woolcott et al.16 As we excluded BB use of less than 28 days from our meta-analysis, our pooled OR reflects chronic BB use more accurately than earlier meta-analyses which included both acute and chronic BB use. Also, there appeared to be a statistically significant reduction in risk of falling with chronic use of BB and non–class-specific antihypertensive use (ATC code C). This finding could be due to confounding by selective prescribing, where frail older adults (those more likely to fall) are less likely to receive chronic treatments (such as BB) than older adults who are not frail. The latter group may also be more likely to pursue and adhere to other therapies, such as exercise, that may reduce the risk of falls. It is also possible that BB may protect against falls in some way, although this would be difficult to evaluate outside the RCT setting. Conversely Lipsitz et al.45 found that higher doses of ACE-i were associated with lower risk of injurious falls and CCB with lower risk of all and indoor falls. Lipsitz et al.45 suggests that one potential mechanism that CCBs contribute to prevention of falls is an increase in cerebral blood flow and prevention of ischemic brain injury. For ACE-i, Lipsitz et al. suggests that possible improvement of skeletal muscle function.57 Leipzig et al.18 found that thiazide diuretics had higher OR than loop diuretics but suggested that the risk of falls had been declining over the years 1975 to 1993, possibly due to lower doses of medications being used. The subsequent meta-analysis by Woolcott et al.,19 including studies published between 1996 to 2007, found that diuretic use was associated with increased fall risk in unadjusted meta-analysis but not when adjustment was made for other risk factors for falls. Similarly, our meta-analysis found that chronic diuretic use was not associated with a significantly increased risk of falling. In our study, OR and HR were meta-analyzed separately. HR provides an estimate of falls risk at each point in time during the follow-up period and should not be meta-analyzed with OR, which estimates the risk of falls across the entire follow-up period. As risk of falling with antihypertensive medication is time-dependent, effect sizes should be calculated and reported as HR, IRR, or standard incidence ratio rather than OR or relative risk. The majority of included studies did not provide a gender-specific effect size. Studies should distinguish between chronic medication use and acute changes in medication. Moreover, it is clearly visible in HR/IRR effect sizes that there is lower heterogeneity between studies.58,59 The degree of heterogeneity observed in our OR plots may have been due to differences in study design, degree of adjustment for confounders, follow-up duration, and outcome variable (falls, injurious falls, or fractures). Studies that did not adjust for confounding may have overestimated the association between medications and falls. Leipzig et al.18 suggested that heterogeneity could also be attributed to varying doses of antihypertensive medications or confounding by indication. For example, studies reporting significantly elevated risk of falling with diuretic45 and CCB10 involved hospital patients, where sicker or frailer patients may have been more likely to fall and receive antihypertensive therapy. Future meta-regression could be conducted to identify sources of heterogeneity. Researchers should incorporate time into their study designs (index, incident, initiation, and exposure times) and distinguish between acute and chronic medication use. A case-crossover or RCT design are more appropriate designs,40 although the propensity score method54 could reduce confounding by indication in future observational studies. Strengths and limitations By analyzing chronic antihypertensive medication use separately to acute use, our study was able to reduce study heterogeneity and examine the relationship between duration of therapy and risk of falls. We used the random-effects inverse-variance method to account for study heterogeneity and conducted separate analyses for OR and HR.40 A limitation to our study was the inclusion of published articles only written in English. In conclusion, there is no significant increase in risk of falls for chronic use of antihypertensive medication. The risk of falling is highest 0 to 24 hours after antihypertensive medication initiation, change, or dose increase (day 0) and remains significantly elevated between days 1 to 21 for diuretics. We recommend that health care practitioners exercise extra vigilance and initiate strategies to reduce the risk of falling in older adults prior to and for these periods. KEY POINTS Question Is there any association between acute (<28days) or chronic (≥28days) use of antihypertensive medications and falls in older adults (age 60 years and older)? Findings The risk of falling is highest during acute antihypertensive use. OR was highest 0 to 24 hours after antihypertensive medication initiation, change, or dose increase (day 0). For diuretics, the OR remained significantly elevated until day 21. There was no association between chronic use of antihypertensive medications and falls in older adults. Meaning Strategies to reduce the risk of falling should be implemented prior to commencing antihypertensive medications in older adults. DISCLOSURE The authors declared no conflict of interest. ACKNOWLEDGMENT We would like to thank Lorraine Evison, the Faculty Liaison Librarian at The University of Sydney Medical Sciences Libraries for her support in building the search strategy for this review and Dr Barbara Mintzes, senior lecturer at the University of Sydney for her support and advice on acquiring proper statistical methods, and Associate Professor Dr Amanda Salis NHMRC Senior Research Fellow at the University of Sydney for her help in manuscript writing and editing. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. REFERENCES 1. Heinrich S Rapp K Rissmann U Becker C König HH. Cost of falls in old age: a systematic review. Osteoporos Int  2010; 21: 891– 902. Google Scholar CrossRef Search ADS PubMed  2. Lord SR March LM Cameron ID Cumming RG Schwarz J Zochling J Chen JS Makaroff J Sitoh YY Lau TC Brnabic A Sambrook PN. Differing risk factors for falls in nursing home and intermediate-care residents who can and cannot stand unaided. J Am Geriatr Soc  2003; 51: 1645– 1650. Google Scholar CrossRef Search ADS PubMed  3. Leipzig RM Cumming RG Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis: I. Psychotropic drugs. J Am Geriatr Soc  1999; 47: 30– 39. Google Scholar CrossRef Search ADS PubMed  4. Aronow WS Frishman WH Cheng-Lai A. Cardiovascular drug therapy in the elderly. Cardiol Rev  2007; 15: 195– 215. Google Scholar CrossRef Search ADS PubMed  5. Berry SD Mittleman MA Zhang Y Solomon DH Lipsitz LA Mostofsky E Goldense D Kiel DP. New loop diuretic prescriptions may be an acute risk factor for falls in the nursing home. Pharmacoepidemiol Drug Saf  2012; 21: 560– 563. Google Scholar CrossRef Search ADS PubMed  6. Butt DA Mamdani M Austin PC Tu K Gomes T Glazier RH. The risk of hip fracture after initiating antihypertensive drugs in the elderly. Arch Intern Med  2012; 172: 1739– 1744. Google Scholar CrossRef Search ADS PubMed  7. Butt DA Mamdani M Austin PC Tu K Gomes T Glazier RH. The risk of falls on initiation of antihypertensive drugs in the elderly. Osteoporos Int  2013; 24: 2649– 2657. Google Scholar CrossRef Search ADS PubMed  8. Callisaya ML Sharman JE Close J Lord SR Srikanth VK. Greater daily defined dose of antihypertensive medication increases the risk of falls in older people–a population-based study. J Am Geriatr Soc  2014; 62: 1527– 1533. Google Scholar CrossRef Search ADS PubMed  9. Payne RA Abel GA Simpson CR Maxwell SR. Association between prescribing of cardiovascular and psychotropic medications and hospital admission for falls or fractures. Drugs Aging  2013; 30: 247– 254. Google Scholar CrossRef Search ADS PubMed  10. Shimbo D Barrett Bowling C Levitan EB Deng L Sim JJ Huang L Reynolds K Muntner P. Short-Term Risk of Serious Fall Injuries in Older Adults Initiating and Intensifying Treatment With Antihypertensive Medication. Circ Cardiovasc Qual Outcomes  2016; 9: 222– 229. Google Scholar CrossRef Search ADS PubMed  11. Tinetti ME Han L Lee DS McAvay GJ Peduzzi P Gross CP Zhou B Lin H. Antihypertensive medications and serious fall injuries in a nationally representative sample of older adults. JAMA Intern Med  2014; 174: 588– 595. Google Scholar CrossRef Search ADS PubMed  12. Smith AD. The worldwide challenge of the dementias: a role for B vitamins and homocysteine? Food Nutr Bull  2008; 29: S143– S172. Google Scholar CrossRef Search ADS PubMed  13. Morgan S Smolina K Mooney D Raymond C Bowen ML Gorczynski C Basham KAR. The Canadian Rx atlas , 3rd edition. 2013; <https://open.library.ubc.ca/collections/52383/items/1.0048514>. Accessed 21 March 2017. 14. Hartikainen S Lönnroos E Louhivuori K. Medication as a risk factor for falls: critical systematic review. J Gerontol A Biol Sci Med Sci  2007; 62: 1172– 1181. Google Scholar CrossRef Search ADS PubMed  15. Wiens M Etminan M Gill SS Takkouche B. Effects of antihypertensive drug treatments on fracture outcomes: a meta-analysis of observational studies. J Intern Med  2006; 260: 350– 362. Google Scholar CrossRef Search ADS PubMed  16. Woolcott JC Richardson KJ Wiens MO Patel B Marin J Khan KM Marra CA. Meta-analysis of the impact of 9 medication classes on falls in elderly persons. Arch Intern Med  2009; 169: 1952– 1960. Google Scholar CrossRef Search ADS PubMed  17. Zang G. Antihypertensive drugs and the risk of fall injuries: a systematic review and meta-analysis. J Int Med Res  2013; 41: 1408– 1417. Google Scholar CrossRef Search ADS PubMed  18. Leipzig RM Cumming RG Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis: II. Cardiac and analgesic drugs. J Am Geriatr Soc  1999; 47: 40– 50. Google Scholar CrossRef Search ADS PubMed  19. Askari M Eslami S Scheffer AC Medlock S de Rooij SE van der Velde N Abu-Hanna A. Different risk-increasing drugs in recurrent versus single fallers: are recurrent fallers a distinct population? Drugs Aging  2013; 30: 845– 851. Google Scholar CrossRef Search ADS PubMed  20. Eriksson S Gustafson Y Lundin-Olsson L. Risk factors for falls in people with and without a diagnose of dementia living in residential care facilities: a prospective study. Arch Gerontol Geriatr  2008; 46: 293– 306. Google Scholar CrossRef Search ADS PubMed  21. Gribbin J Hubbard R Gladman J Smith C Lewis S. Risk of falls associated with antihypertensive medication: self-controlled case series. Pharmacoepidemiol Drug Saf  2011; 20: 879– 884. Google Scholar CrossRef Search ADS PubMed  22. Pariente A Dartigues JF Benichou J Letenneur L Moore N Fourrier-Réglat A. Benzodiazepines and injurious falls in community dwelling elders. Drugs Aging  2008; 25: 61– 70. Google Scholar CrossRef Search ADS PubMed  23. Shuto H Imakyure O Matsumoto J Egawa T Jiang Y Hirakawa M Kataoka Y Yanagawa T. Medication use as a risk factor for inpatient falls in an acute care hospital: a case-crossover study. Br J Clin Pharmacol  2010; 69: 535– 542. Google Scholar CrossRef Search ADS PubMed  24. Stenhagen M Ekstrom H Nordell E Elmstahl S. Falls in the general elderly population: A 3-and 6-year follow-up of risk factors using data from the longitudinal population study “good ageing in Skane”. Eur Geriatr Med  2013; 3: S53. Google Scholar CrossRef Search ADS   25. Moher D Shamseer L Clarke M Ghersi D Liberati A Petticrew M Shekelle P Stewart LA; PRISMA-P Group. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev  2015; 4: 1. Google Scholar CrossRef Search ADS PubMed  26. ATC and DDD Index. WHO Collaborating Centre for Drug Statistics Methodology, Oslo ATC Code: Anatomical Therapeutic Chemical Classification System.<https://www.whocc.no/atc_ddd_index/>. Accessed 20 March 2017. 27. Hasegawa J Kuzuya M Iguchi A. Urinary incontinence and behavioral symptoms are independent risk factors for recurrent and injurious falls, respectively, among residents in long-term care facilities. Arch Gerontol Geriatr  2010; 50: 77– 81. Google Scholar CrossRef Search ADS PubMed  28. Sterke CS van Beeck EF van der Velde N Ziere G Petrovic M Looman CW van der Cammen TJ. New insights: dose-response relationship between psychotropic drugs and falls: a study in nursing home residents with dementia. J Clin Pharmacol  2012; 52: 947– 955. Google Scholar CrossRef Search ADS PubMed  29. Wong AK Lord SR Sturnieks DL Delbaere K Trollor JN Close JC. Angiotensin system-blocking medications are associated with fewer falls over 12 months in community-dwelling older people. J Am Geriatr Soc  2013; 61: 776– 781. Google Scholar CrossRef Search ADS PubMed  30. Tinetti ME Williams TF Mayewski R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med  1986; 80: 429– 434. Google Scholar CrossRef Search ADS PubMed  31. World Health Organization, WHO Falls fact sheet. <http://www.who.int/mediacentre/factsheets/fs344/en/>. Accessed 10 September 2017. 32. Lamb SE Jørstad-Stein EC Hauer K Becker C. Development of a common outcome data set for fall injury prevention trials: the Prevention of Falls Network Europe consensus. J Am Geriatr Soc  2005; 53: 1618– 1622. Google Scholar CrossRef Search ADS PubMed  33. Kellogg IWG. The prevention of falls in later life. A report of the Kellogg International Work Group on the Prevention of Falls by the Elderly. Dan Med Bull  1987; 34: 1– 24. 34. Wells GA Shea B O’Connell D Peterson J Welch D Losos D Tugwell D. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. 2014; <http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp>. Accessed 21 March 2017. 35. van Wijngaarden JP Dhonukshe-Rutten RA van Schoor NM van der Velde N Swart KM Enneman AW van Dijk SC Brouwer-Brolsma EM Zillikens MC van Meurs JB Brug J Uitterlinden AG Lips P de Groot LC. Rationale and design of the B-PROOF study, a randomized controlled trial on the effect of supplemental intake of vitamin B12 and folic acid on fracture incidence. BMC Geriatr  2011; 11: 1– 11. Google Scholar CrossRef Search ADS PubMed  36. Hulley SB Furberg CD Gurland B McDonald R Perry HM Schnaper HW Schoenberger JA Smith WM Vogt TM. Systolic Hypertension in the Elderly Program (SHEP): antihypertensive efficacy of chlorthalidone. Am J Cardiol  1985; 56: 913– 920. Google Scholar CrossRef Search ADS PubMed  37. Ham AC Swart KM Enneman AW van Dijk SC Oliai Araghi S van Wijngaarden JP van der Zwaluw NL Brouwer-Brolsma EM Dhonukshe-Rutten RA van Schoor NM van der Cammen TJ Lips P de Groot LC Uitterlinden AG Witkamp RF Stricker BH van der Velde N. Medication-related fall incidents in an older, ambulant population: the B-PROOF study. Drugs Aging  2014; 31: 917– 927. Google Scholar CrossRef Search ADS PubMed  38. Charlesworth CJ Peralta CA Odden MC. Functional Status and Antihypertensive Therapy in Older Adults: A New Perspective on Old Data. Am J Hypertens  2016; 29: 690– 695. Google Scholar CrossRef Search ADS PubMed  39. Stroup DF Berlin JA Morton SC Olkin I Williamson GD Rennie D Moher D Becker BJ Sipe TA Thacker SB. Meta-analysis of observational studies in epidemiology: a proposal for reporting. JAMA  2000; 283: 2008– 2012. Google Scholar CrossRef Search ADS PubMed  40. Egger M Davey-Smith G Altman D. Systematic reviews in health care: meta-analysis in context . London: John Wiley & Sons; 2008. 41. DerSimonian R Kacker R. Random-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials  2007; 28: 105– 114. Google Scholar CrossRef Search ADS PubMed  42. DerSimonian R Laird N. Meta-analysis in clinical trials. Control Clin Trials  1986; 7: 177– 188. Google Scholar CrossRef Search ADS PubMed  43. Baranzini F Diurni M Ceccon F Poloni N Cazzamalli S Costantini C Colli C Greco L Callegari C. Fall-related injuries in a nursing home setting: is polypharmacy a risk factor? BMC Health Serv Res  2009; 9: 228. Google Scholar CrossRef Search ADS PubMed  44. Duh MS Mody SH Lefebvre P Woodman RC Buteau S Piech CT. Anaemia and the risk of injurious falls in a community-dwelling elderly population. Drugs Aging  2008; 25: 325– 334. Google Scholar CrossRef Search ADS PubMed  45. Lipsitz LA Habtemariam D Gagnon M Iloputaife I Sorond F Tchalla AE Dantoine TF Travison TG. Reexamining the Effect of Antihypertensive Medications on Falls in Old Age. Hypertension  2015; 66: 183– 189. Google Scholar CrossRef Search ADS PubMed  46. Montali F Campaniello G Benatti M Rastelli G Pedrazzoni M Cervellin G. Impact of different drug classes on clinical severity of falls in an elderly population: epidemiological survey in a trauma center. Journal of Clinical Gerontology and Geriatrics  2015; 6: 63– 67. Google Scholar CrossRef Search ADS   47. Rafiq M McGovern A Jones S Harris K Tomson C Gallagher H de Lusignan S. Falls in the elderly were predicted opportunistically using a decision tree and systematically using a database-driven screening tool. J Clin Epidemiol  2014; 67: 877– 886. Google Scholar CrossRef Search ADS PubMed  48. Sterke CS Ziere G van Beeck EF Looman CW van der Cammen TJ. Dose-response relationship between selective serotonin re-uptake inhibitors and injurious falls: a study in nursing home residents with dementia. Br J Clin Pharmacol  2012; 73: 812– 820. Google Scholar CrossRef Search ADS PubMed  49. Thorell K Ranstad K Midlöv P Borgquist L Halling A. Is use of fall risk-increasing drugs in an elderly population associated with an increased risk of hip fracture, after adjustment for multimorbidity level: a cohort study. BMC Geriatr  2014; 14: 131. Google Scholar CrossRef Search ADS PubMed  50. Garcia Molina O Olmos Jimenez R Castellote De Varona FJ Mendoza Otero F Arocas Casano V De La Rubia Nieto MA. Polypharmacy related to increased risk of hip fracture in elderly patients. Atencion Farmaceutica  2014; 16: 117– 123. 51. Rhalimi M Helou R Jaecker P. Medication use and increased risk of falls in hospitalized elderly patients: a retrospective, case-control study. Drugs Aging  2009; 26: 847– 852. Google Scholar CrossRef Search ADS PubMed  52. Zia A Kamaruzzaman SB Myint PK Tan MP. The association of antihypertensives with postural blood pressure and falls among seniors residing in the community: a case-control study. Eur J Clin Invest  2015; 45: 1069– 1076. Google Scholar CrossRef Search ADS PubMed  53. Zia A Kamaruzzaman SB Tan MP. The consumption of two or more fall risk-increasing drugs rather than polypharmacy is associated with falls. Geriatr Gerontol Int  2017; 17: 463– 470. Google Scholar CrossRef Search ADS PubMed  54. Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res  2011; 46: 399– 424. Google Scholar CrossRef Search ADS PubMed  55. Johnston SC. Identifying confounding by indication through blinded prospective review. Am J Epidemiol  2001; 154: 276– 284. Google Scholar CrossRef Search ADS PubMed  56. Salas M Hofman A Stricker BH. Confounding by indication: an example of variation in the use of epidemiologic terminology. Am J Epidemiol  1999; 149: 981– 983. Google Scholar CrossRef Search ADS PubMed  57. Onder G Vedova CD Pahor M. Effects of ACE inhibitors on skeletal muscle. Curr Pharm Des  2006; 12: 2057– 2064. Google Scholar CrossRef Search ADS PubMed  58. Higgins JP. Commentary: Heterogeneity in meta-analysis should be expected and appropriately quantified. Int J Epidemiol  2008; 37: 1158– 1160. Google Scholar CrossRef Search ADS PubMed  59. Higgins JP Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med  2002; 21: 1539– 1558. Google Scholar CrossRef Search ADS PubMed  © American Journal of Hypertension, Ltd 2017. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

American Journal of HypertensionOxford University Press

Published: Apr 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial