Antihypertensive drugs and the risk of fall injuries: A systematic review and meta-analysis

Abstract

Objective

A meta-analysis of published studies was performed to determine whether administration of any of five antihypertensive drug classes (thiazide diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers and β-blockers) affected the risk of fall injuries in the elderly (aged ≥60 years).

Methods

Articles reporting the risk of fall injury in elderly people being treated with the five main classes of antihypertensive drugs were retrieved using MEDLINE®, EMBASE, SCOPUS® and the Cochrane Database. Trial eligibility and methodological quality were assessed before data extraction and analysed using odds ratios with 95% confidence intervals.

Results

Sixty-two articles, included in two meta-analyses, were identified. These meta-analyses drew opposite conclusions about the role of antihypertensive drugs in fall injuries in the elderly. However, the present analysis did not reveal a clear association (or the lack of one) between antihypertensive drugs and risk of fall injuries.

Conclusions

There is no clear, statistically significant clinical precedent indicating that the use of any of the antihypertensive drugs considered here increases the risk of fall injuries in the elderly. Nonetheless, in following standard clinical guidelines for hypertension management, physicians need to be aware of the impact of drug therapies on fall injuries.

Introduction

Hypertension is one of the most common medical conditions in elderly people (defined as those aged ≥60 years) and antihypertensive medications used to treat hypertension are among the most widely prescribed drugs for this age group.^1–4 However, these agents can potentially cause fall injuries (as a result of symptoms such as dizziness or syncope caused by orthostatic hypotension) in elderly people initiating therapy for hypertension.^5,6 Such injuries can potentially lead to fractures, particularly of the hip, and are associated with substantial morbidity and mortality.^7–9 At present, there is little evidence to support an immediate increase in the risk of fall injuries during the initiation of antihypertensive therapy in this patient population.^10,11 In fact, a systematic review of the different aetiologies of fall injuries did not identify the use of antihypertensive drugs as a definite risk factor for falls in elderly patients, even though it is well documented that they can cause orthostasis and dizziness. ¹²

Epidemiological studies in the elderly show that serious fall-related injuries such as fractures have functional, cognitive and physical effects similar to those of myocardial infarction and stroke.^7,13,14 The incidences of nonfatal cardiovascular events in hypertensive elderly patients, and of serious fall injury in elderly people at risk of falls, are both 16%.^15,16 Adverse drug events are three times more common in elderly people than in the general population: approximately 35% of ambulatory older adults experience an adverse drug event and 29% require health services as a result. ¹⁷ The objective of the present study was to perform a meta-analysis of published studies that looked at the association between the initiation of monotherapy with any of the five commonly used antihypertensive drug classes (thiazide diuretics, angiotensin-converting enzyme [ACE] inhibitors, angiotensin receptor blockers [ARBs], calcium channel blockers [CCBs] and β-blockers [BB]) in the elderly and the risk of fall injuries.

Materials and methods

Data sources and searches

Clinical literature was identified using MEDLINE®, EMBASE, SCOPUS® and the Cochrane Database (January 1975–June 2012). MeSH terms used were ‘antihypertensive drugs’, ‘thiazide’, ‘ACE inhibitor’, ‘calcium channel blockers, CCB’, ‘beta-blocker, BB’, ‘angiotensin receptor blocker, ARB’, ‘elderly population’ and ‘fall injuries’. Cohort and case–control studies documenting relative risks and confidence intervals (CIs) for the association between fall injuries (fracture outcomes) and exposure to antihypertensive agents were identified. No language restrictions were applied.

Assessment of methodological quality

Searches were limited to literature on humans. Hand searching of key article reference lists was used to locate additional relevant articles. Eligibility assessment and data abstraction were both performed independently (at different times) in an unblended, standardized manner. The inclusion of all participants in the final analysis was evaluated, and for each timepoint adequate follow-up was defined as the availability of data on ≥80% of participants.

Data extraction and analysis

Results from all searches were combined and duplicates were removed. Studies were eligible for inclusion if they fulfilled the following criteria: the study design was described; the number of included patients was stated; mean patient age was ≥60 years; numbers of patients of each sex were stated; the type of diagnostic criteria and/or intervention strategy used was stated; timing of diagnosis was stated. The outcomes described in the collected manuscripts were synthesized and formed the basis for further analysis and description, which was done by following recommendations from The Cochrane Collaboration and the Quality of Reporting of Meta-analyses guidelines. ¹⁸ The odds ratio (OR) was used to analyse all dichotomous outcome measures and each ratio is presented with its 95% confidence interval (CI).

Assessment of heterogeneity

The risk of bias of each relevant article was assessed using the 12 criteria recommended by the Cochrane Back Review Group. ¹⁹ Criteria were scored ‘yes’ (criterion met), ‘no’ (criterion not met) or ‘unsure’ (not enough information to make the decision). Articles that met six of the 12 criteria were considered to have a low risk of bias. Clinical relevance was assessed using the five questions recommended by the Cochrane Back Review Group. ¹⁹ All results reported were based on a relatively small sample size. Potential publication bias was tested using a funnel plot. The overall quality of the evidence for each outcome was assessed using an adapted GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach, as recommended by the Cochrane Back Review Group. ²⁰ The quality of the evidence for a specific outcome was based on the study design, risk of bias, consistency of results, directness (generalizability), precision (sufficient data) and potential bias for the reporting of results across all studies that measured that particular outcome.

Results

Findings of the literature search are summarized in Figure 1. Based on the inclusion criteria described in the previous section, the search yielded 62 studies, which were included in two meta-analyses.^21,22 A meta-analysis published in 2009 ²³ was not included in the final analysis as it did not take into consideration measures of uncertainty, which is imperative for the determination of accurate confidence intervals. As is evident from the search results, there is a significant lack of published studies that specifically look at the risk of fall injuries in elderly patients when they are starting antihypertensive monotherapy.

Figure 1.

Schematic diagram of searches performed in the present systematic review and meta-analysis of the association between antihypertensive medication and the risk of fall injury in the elderly, using MEDLINE®, EMBASE, SCOPUS® and the Cochrane Database for the period 1975–2012.

In 2000 a Cochrane Database Systematic review was published dealing with pharmacotherapeutic options in hypertension. ²⁴ This review summarized 15 randomized controlled trials (RCTs). However, a glaring deficiency in most of the RCTs that were included in the Cochrane review ²⁴ was that they did not report the percentage of dropouts based on adverse drug effects. This points to a serious deficiency in these studies, i.e. lack of appreciation of the great importance of reporting adverse drug effects, which is recommended by the Consolidated Standards of Reporting Trials (CONSORT) statement. ²⁵ This systematic review ²⁴ was not further analysed in the present study.

Leipzig et al. ²¹ performed a meta-analysis of the effects of antihypertensive drugs and fall injuries. Their analysis characterized diuretics, including furosemide and thiazides, as a weak risk factor for fall injuries in the elderly population (pooled OR 1.08; 95% CI 1.02, 1.16) (Table 1). One major limitation of this analysis was its study design. Because it was an observational study, the results could not be adjusted for potential confounders. Another obvious limitation was the failure to document and report the dosage or duration of diuretic usage that would constitute a significant risk factor for fall injuries. However, the updated meta-analysis published in 2009 ²³ comprehensively showed that furosemide and thiazide had no association with fall injuries in the elderly.

Table 1.

Details of meta-analyses included in the present systematic analysis of the association between antihypertensive medication and risk of fall injury in elderly people (aged ≥60 years).

Study	No. on prescription medication/total no. of subjects	Study population	Results, OR or RR (95% CI)	Major limitations
Leipzig et al.21,a	BBs, 629/5551	Long-term care and community-dwelling adults aged ≥60 years	0.93 (0.77, 1.11)	Not all classes of antihypertensive agents tested; no data on dosage of medication or period patients were on the drugs; pooled OR not adjusted for any potential or obvious confounders
	Thiazides, 3054/12475		1.06 (0.97, 1.16)
	ACE inhibitors, 3612		1.20 (0.92, 1.58)
	CCBs, 552/3929		0.94 (0.77, 1.14)
Wiens et al.22,b	BBs, thiazides or ACE inhibitors, 30601/120819 CCBs, 169/315	Nursing home or hospital inpatients and community-dwelling adults >44 years old	BBs, 0.86 (0.70, 0.98)	ARBs not tested; duration of drug administration was long, so there were no data on immediate effects of drug exposure; pooled RR only minimally adjusted for potential or obvious confounders
			Thiazides, 0.86 (0.81, 0.92) ACE inhibitors, 0.81 (0.73, 0.89) CCBs, 1.96 (1.16, 3.30)

a

Data were analysed using the pooled odds ratio (OR).

b

Data were analysed using the pooled relative risk (RR).

CI, confidence interval; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BB, β-blocker; CCB, calcium channel blocker.

Another meta-analysis, which included 54 observational studies (14 cohort and 40 case–control studies), was performed by Wiens et al. in 2006. ²² This analysis used the pooled relative risk [RR] rather than the OR. Of note, 74% of the articles that were included ²² described case–control studies, which would easily allow calculation of ORs. It is imperative that uniformly accepted statistical principles are followed in such analyses, so that results across different platforms can be compared in order to generate useful clinical information. Surprisingly, the analyses by Wiens et al. revealed a lower risk of fractures in elderly people using BBs and diuretics compared with those not using these drugs (pooled RR 0.86 and 0.81; 95% CI 0.81, 0.92 and 0.73, 0.89 for diuretics and BBs, respectively). ²² This analysis encompassed one study each with CCBs ²⁶ and ACE inhibitors. ²⁷ Whereas CCBs showed a significantly increased risk of fracture (pooled RR 1.96; 95% CI 1.16, 3.30), ²⁶ ACE inhibitors surprisingly showed a decreased risk of hip fracture (pooled RR 0.81; 95% CI 0.73. 0.89). ²⁷ A serious limitation of these studies26,27 was the absence of a matched control group (in each case a comparison was made with patients admitted with other comorbidities). In addition, neither study provided information on the actual drugs used (only drug classes were mentioned) and the exact duration of drug treatment (it was only mentioned that treatment duration was ≥1 year), making it impossible to draw a relevant and dependable clinical conclusion. The other obvious limitation of the meta-analysis by Wiens et al. was its predominant inclusion of female elderly patients (>74%). ²² In the absence of clinical evidence that there is no sex-specific correlation between usage of antihypertensive drugs and fall injuries in the elderly, this reduces the overall relevance of the analysis (Table 1).

Our analysis revealed a few published studies investigating the immediate risk of fall injuries in the elderly during the initiation of monotherapy with antihypertensive drugs. One of them ¹⁰ was a case–control study conducted in the UK, which involved 9682 elderly people with fall injuries and 52 100 matched controls. An increased risk of fall injuries within the first 3 weeks of thiazide initiation was observed (OR 4.28; 95% CI 1.19, 15.42); however, no correlation was observed with sustained usage of ACE inhibitors or CCBs. However, the results given in this article ¹⁰ do not corroborate those of a subsequent study published in 2011. ²⁸ This cohort study, involving 376 061 elderly people with hypertension, found that patients using thiazides had the lowest fracture rate, whereas the use of loop diuretics resulted in a high rate of fractures resulting from fall injuries in the study population. ²⁸ Of note, patients prescribed CCBs constituted the control group in the study by Solomon et al., ²⁸ whereas Gribbin et al. ¹⁰ used matched controls; this might explain the apparent discrepancy between the findings of both studies.

In a case-crossover study performed in 349 patients Japan, it was observed that monotherapy with ARBs, CCBs, α-adrenergic blockers or ACE inhibitors for 3 days was a risk factor for fall injuries (OR 8.42; 95% CI 3.12, 22.72). ¹¹ It should be noted that this study enrolled patients who were ≥20 years of age, hence the study findings are not entirely relevant to the goals of the present study. Moreover, the data analysis was based on prescriptions rather than actual intake levels by the enrolled patients, which in itself is a serious confounding factor in such analyses. ¹¹

Publication bias within the studies covered in the two meta-analyses^21,22 was assessed both visually, using a funnel plot (Figure 2), and with the 12 criteria recommended by the Cochrane Back Review Group ¹⁹ (Figure 3). The funnel plot (Figure 2) suggested a potential lack of studies with an RR of >1, which is corroborated by Figure 3. Moreover, none of the studies included in the two meta-analyses^21,22 reported (or appeared to use) an intention-to-treat analysis (Figure 3). The number of criteria met varied between 2/12 ²⁰ and 3/12, ²² with a mean of 2.5/12. All studies included in the two meta-analyses^21,22 were considered to have a high risk of bias. Neither of the included studies had an adequate description of withdrawals and dropouts.

Figure 2.

Funnel plot of studies on antihypertensive drug usage use and risk of fall injuries in the elderly. Publication bias suggested a deficit of studies with a relative risk >1 and, hence, potential bias. Studies were retrieved using MEDLINE®, EMBASE, SCOPUS® and the Cochrane Database for the period 1975–2012.

Figure 3.

Overall quality of the evidence for each outcome (in a systematic review and meta-analysis of the association between antihypertensive medication and the risk of fall injury in the elderly) was measured using an adapted GRADE approach, as recommended by the Cochrane Back Review Group. ¹⁹ Quality of the evidence for a specific outcome was based on study design, risk of bias, consistency of results, directness (generalizability), precision (sufficiency of data) and potential bias for the reporting of results across all studies that measured that particular outcome. Studies were retrieved using MEDLINE®, EMBASE, SCOPUS® and the Cochrane Database for the period 1975–2012.

Discussion

It has been known for several decades that sudden hypotension caused by initiation of antihypertensive monotherapy can result in fall injuries, and is especially severe in the elderly population for two reasons: they are the people most affected by hypertension and they are more prone to serious consequences from a fall, compared with younger persons. ⁸ In fact, fall injuries contribute to 67% of deaths resulting from accidental injuries. ²⁹ Factoring into this the finding that 33–50% of elderly people suffer at least one fall per year, ³⁰ half of which result in injury, ³¹ it becomes imperative to analyse properly any risk of fall injuries associated with the initiation of antihypertensive monotherapy.

However, the currently available literature reveals divided opinion on the risk of falls associated with initiation of antihypertensive monotherapy. Whereas a few studies have, interestingly, pointed to a significant lowering of the risk of fall injuries with the introduction of antihypertensive agents,^22,26,28 others have shown that the highest risk of such injuries occurs in the period immediately after initiation of antihypertensive monotherapy.^10,11 The major reasons for such conflicting reports on are the lack of randomized controlled trials, the lack of inclusion of a wide spectrum of antihypertensive drugs in different studies, the lack of adjustment for confounders, drug dosage and duration of usage, and inadequate statistical analyses of observational study results.

However, if considered from a global perspective, the disparate findings make sense. Prolonged, sustained and regulated use of antihypertensive drugs will definitely reduce the chances of fluctuation in blood pressure resulting in hypotension and consequent fall injuries. At the same time, immediately after initiation of an antihypertensive therapeutic regimen is the most obvious period when homeostatic regulation of blood pressure will be wayward and can result in hypotension. It would perhaps be clinically ideal to initiate antihypertensive monotherapy under close monitoring ,so that hypotension and consequent fall injuries can be prevented (and rapid treatment can be provided) if hypotension occurs. A limitation of the present study is that, even though our analysis focuses on the association of antihypertensive drug use with fall injuries/fractures, the included studies involved drug exposures of long duration (≥1year), where the underlying mechanism is thought to be secondary to the effects on bone metabolism. Hence, the clinical management of hypertension needs to factor in the impact of antihypertensive agents on fall injuries. Another potential limitation is that we used the Cochrane QUOROM guidelines rather than the PRISMA guidelines ³² for the present analysis. Our analysis showed a good match with the checklist; the major differences were that we did not apply measures of consistency for each of the meta-analyses that were included. Moreover, we defined the risk of bias across studies rather than within studies, both of which are recommended by the PRISMA guidelines. ³² The results of our analysis indicate the need for multicentre randomized clinical trials to analyse comprehensively the role of antihypertensive drugs in fall injuries, including hip fracture, in the elderly population.