Effectiveness of Cardiovascular Evaluations and Interventions on Fall Risk: A Scoping Review

Objective cardiovascular abnormalities are consistently associated with fall risk in older people. However, little research has been done to assess the effect of cardiovascular interventions on fall risk. The aim of this scoping review is to explore the current literature on the effectiveness of cardiovascular evaluations and interventions in reducing fall risk in older people. Design scoping review. Data sources Medline, Cochrane Library, and WHO ICTRP Search Portal were systematically searched. Selection criteria randomized controlled trials (RCTs) and intervention studies of community-dwelling adults aged ≥50 years or with a mean age of >60 years that assessed the effect of a cardiovascular assessment and interventions in reducing fall risk. Key search concepts were “falls” and “aged”, and terms for different cardiovascular evaluations and interventions were included. The Cochrane Checklist for risk of bias and the ROBINS-I tool were used to assess the quality of the studies. Results seven studies were included. The majority showed a reduction in falls after cardiovascular evaluation and intervention. Two out of four studies that focused on carotid sinus hypersensitivity (CSH) as a modifiable cardiovascular risk factor for falls, showed a significant reduction in falls after pacemaker implantation. Two studies that looked at sinus node dysfunction (SND) both showed a significant reduction in falls after pacemaker implantation. One study showed that 33% of the patients experienced a fall after cardiovascular evaluation and intervention, whereas all patients fell before assessment. Conclusions the majority of the included studies showed a reduction in falls after the intervention. However, the number of published papers regarding the effect of cardiovascular assessment and interventions on falls is small. A standardized assessment of cardiovascular risk factors may be essential in preventing falls in older adults and could consequently reduce injuries, loss of quality of life, deaths, and fall-related expenditures. Electronic Supplementary Material Supplementary material is available for this article at 10.1007/s12603-019-1165-2 and is accessible for authorized users.

Although the association between cardiovascular abnormalities and fall risk is clear, little research has been done to assess the effect of cardiovascular interventions on fall risk (25). The aim of this scoping review is to explore the current literature on the effectiveness of single and comprehensive cardiovascular evaluations and interventions in reducing fall risk in older people.

Methods
To answer the research question we conducted a scoping review according to the framework outlined by Levac (26), adapted from Arksey and O'Malley (27). The framework consists of five stages in which the first stage refers to identifying the research question (see above). The other consecutive stages subsequently address the identification of relevant studies, study selection, data charting, and summarizing and reporting the results. . Search terms for cardiovascular evaluations included blood pressure measurements, tilt-table test, carotid sinus massage, electrocardiogram (ECG), echocardiography, Holter monitoring and loop recorder. Search terms for cardiovascular interventions included pacemaker, cardiac valve replacement, coronary angioplasty and catheter ablation. The complete search strategy is provided as a supplement in Appendix I.

Stage 3: Study selection
We included all randomized controlled trials and intervention studies of community-dwelling adults aged ≥50 years or with a mean age of >60 years that looked at falls as an outcome measure, and that assessed the effect of a cardiovascular assessment and intervention. Interventions could comprise either multifactorial or single cardiovascular interventions. Studies with hospitalized or other noncommunity-dwelling participants were also included. Articles were excluded if they were reviews or conference abstracts, if they were not written in English, if the intervention was not clearly defined, or if they applied to a specific subgroup (e.g. patients with Parkinson's disease).
The articles found were selected for inclusion independently by two reviewers (SJ and SL) by screening titles and abstracts. Discrepancies in article inclusion were solved by consulting a third reviewer (NV). The two reviewers (SJ and SL) independently assessed all full-text articles for eligibility.

Stage 4: Charting the data
To address the research question, a data charting form was developed with relevant variables, including: author, year, design, sample size, (mean) age of participants, setting, intervention, control, duration of follow-up, outcome of falls, results, and conclusions. The two reviewers (SJ and SL) independently extracted data from all included studies using this charting form. Discrepancies between results were resolved through discussion. If discrepancies remained unsolved, the third reviewer (NV) was consulted.

Stage 5: Collating, summarizing, and reporting the results
The data found was collated and summarized in a descriptive table (Table 1). To assess quality of the included studies the Cochrane Checklist for risk of bias was used for RCTs (28), and the ROBINS-I tool for included intervention studies (29). Details of the quality assessment are described in Appendix II.

Search Result and Study Characteristics
Our initial search retrieved a total of 3131 studies; of 28 papers the full text was reviewed. Seven articles were included in this review. Figure 1 shows the search and selection process. Table 1 summarizes the findings relevant to the research question. Four studies selected fallers with carotid sinus hypersensitivity (CSH) (30)(31)(32)(33), two studies selected fallers with sinus node dysfunction (SND) (34,35), and one study selected geriatric patients from a specialized falls clinic. Three studies were randomized controlled trials (31)(32)(33), four were non-randomized intervention studies of which one was a pilot study (30,(34)(35)(36). The number of participants varied between 15 and 159 and mean age varied from 71.9 to 79.0 years. Two studies were conducted at a cardiology department (34,35), two in a syncope unit (32,33), one at a geriatric department (30), one at a specialized falls clinic (36), and one at an emergency department (31). All studies that were carried out in hospital included patients that had presented with falls from the community. Fall data were collected prospectively in four studies (31)(32)(33)36), and retrospectively in two studies (30,34). One study collected fall data retrospectively before intervention and prospectively after intervention (35).

Carotid Sinus Hypersensitivity (CSH)
Effectiveness of cardiovascular evaluations and interventions on fall risk in patients with CSH was assessed in four studies, of which three were RCTs and one was an intervention study (30)(31)(32)(33). Two studies were conducted at a syncope unit, one at a geriatric department, and one at an emergency department/ cardiovascular investigation unit. All studies defined CSH as at least asystole induced by carotid sinus massage (CSM). Additionally performed cardiovascular evaluations to Table 1 Overview of studies published on cardiovascular evaluation/interventions and falls  (31). Second, the non-randomized study by Crilley et al. showed that 81% of patients with CSH had falls in the year before implantation, compared with 30% after implantation (30). The other two studies showed no effect of pacemaker implantation. The multicenter RCT of Ryan et al. showed a risk ratio (RR) of 0.79 (95% CI 0.41-1.50) of falling with pacemaker implantation compared with loop recorder implantation (33). The cross-over RCT of Parry et al. showed a RR of 0.82 (95% CI 0.62-1.10) of falling with pacemaker turned on compared with pacemaker turned off (32). All four studies had a low or moderate risk of bias according to the Cochrane Checklist and ROBINS-I tool ( Table 2 and Table 3).

Sinus Node Dysfunction (SND)
Two studies studied the effect of pacemaker implantation on fall risk in patients with SND (34,35). Both studies were non-randomized intervention studies and selected cardiology referrals with SND based on either symptoms compatible with SND, 12-lead ECG and/or Holter ECG assessment. All included patients had a pacemaker implanted, and fall rates were assessed after at least 12 months of follow-up. The two studies reported a significant decrease in fall rates, total number of falls, and fall-related injuries in patients with a pacemaker implanted. The study of Brenner et al. had a low risk of bias according to the ROBINS-I tool. The study of Krasniqi et al. had a serious risk of bias according to the ROBINS-I tool due to bias of selection of participants and bias due to confounding (Table 3).    (36). Almost half of older fallers had a treatable cardiovascular abnormality contributing to fall risk of the individual patient, and a cardiovascular evaluation with subsequent intervention could identify these abnormalities. Furthermore, 33% of the patients experienced a fall after cardiovascular evaluation and intervention, whereas all patients fell before assessment. The study had a low risk of bias according to the ROBINS-I tool (Table 3).

Discussion
This scoping review shows that the majority of studies that investigated the effect of a cardiovascular assessment with subsequent intervention in older fallers showed a reduction in falls after the intervention. Main cardiovascular risk factors for falls focused on in the included studies were CSH and SND. These cardiovascular abnormalities are associated with syncope and falls in older people (37,38). However, this review of current literature shows that evidence about the effectiveness of cardiovascular assessment and subsequent interventions on fall risk is scarce, and that most studies have only focused on single cardiovascular fall risk increasing factors.
Four studies focused on CSH as a modifiable cardiovascular risk factor for falls, and half of them showed a significant reduction in falls after pacemaker implantation. Kenny et al.
showed that falls and injurious events were less common in subjects with CSH after pacemaker implantation. Crilley et al. also showed that patients with CSH had less falls after pacemaker implantation. However, the study of Crilley et al. collected fall-data retrospectively which is a potential risk of bias. The other two studies that assessed the effect of pacemaker implantation in patients with CSH showed no effect of pacing on fall risk, although the study of Parry et al. might be underpowered due to a high drop-out rate.
The two studies that looked at SND both showed a significant reduction in falls after pacemaker implantation. However, although the outcome of these studies was falls, the studies primarily included patients with established SND, and information on falls before pacemaker implantation was recorded retrospectively. Furthermore, these studies lacked a control group, so a causal relationship cannot be established.
The pilot study of Jansen et al. is the only study that has broadly assessed modifiable cardiovascular fall risk factors. They showed that adding a standardized cardiovascular assessment and intervention including structured history taking to the multifactorial falls evaluation led to the identification of cardiovascular abnormalities contributing to fall risk in 47% of older fallers. Furthermore, Jansen et al. found that 33% of the patients experienced a fall after cardiovascular evaluation and intervention, whereas all patients fell before assessment (36). However, being a pilot study, the study primarily assessed feasibility and not effects on fall risk. Thus it comprised a small group of participants and lacked a control group.
Besides the studies in our review, we found a study that assessed the effect of a cardiovascular intervention on fall risk, but they lacked the appropriate outcome to meet the inclusion criteria of this review. Tarro Genta et al. evaluated a cardiac rehabilitation program in patients that underwent transcatheter aortic valve implantation (TAVI) compared with patients that had surgical aortic valve replacement (sAVR). Fall risk was measured with the Morse Fall Scale (MFS), and the results showed that fall risk on admission and at discharge was worse in patients who underwent TAVI, compared with patients that underwent sAVR (39). However, this study was not an RCT and patients that underwent TAVI had more risk factors for falling at baseline (e.g. older, higher proportion of coronary heart disease, more comorbidity, lower left ventricular systolic function). Moreover, TAVI-patients were more disabled according to the Barthel-index. Thus, a causal relationship between TAVI, sAVR and fall risk could not be established with this study.
As mentioned above, up to now little research has been performed on the effectiveness of cardiovascular evaluation and interventions on fall risk, despite the evidence of cardiovascular risk factors for falling (11). Several studies that looked at multifactorial falls interventions have included cardiovascular assessments and interventions. Gobierno Hernandez et al. included referral to a cardiologist as part of their multifactorial intervention and found no significant reduction of falls (40).
Tan et al. included a comprehensive cardiovascular assessment and interventions as part of their multifactorial intervention and also did not find a difference in fall rate compared with controls (41). Rubenstein et al. included ECG and 24-hour Holter monitoring in their intervention group and found 9% less falls in the intervention group, but this was not statistically significant (42). Lightbody et al. evaluated a multifactorial fall prevention program including a cardiovascular assessment with ECG and blood pressure measurement and reported fewer falls in the intervention group compared with controls, although not significant (43). Due to the multifactorial nature of these trials, the contribution of the cardiovascular assessment to the reduction of falls cannot be established. Moreover, the cardiovascular components of these study-assessments were often very limited. However, Tinetti et al. showed that participants in a multifactorial intervention group had significantly fewer falls compared with controls, and that improvement in postural blood pressure change was partly responsible for this reduction in falls (44).
As syncope and symptoms of falls overlap, syncope in older persons is often mistaken for falls (45). Recent research on falls and syncope in older persons has shown that a multidisciplinary comprehensive assessment at a falls and syncope clinic consisting of a 12-lead ECG, blood pressure measurements (supine and active standing), echocardiogram, 24-hour Holter monitoring, and tilt testing with carotid sinus massage (on indication) could identify possible causes for falls and syncope in 94% of the patients (20). Eighty-three percent of the patients were diagnosed with hypotension, 44% of the patients had a cardiac cause for syncope, and 21% had reflex syncope. Only in 6% of the patients the cause of syncope remained unexplained. Remarkably, 50% of the syncope patients in this study presented with falls only, thus underlining the need of a cardiovascular assessment in older fallers. Moreover, Zwart et al. found that atrial fibrillation (AF) prevalence is underestimated in the geriatric population, and that an additional 50% of AF cases were found by using 24-hour Holter ECG in comparison to a 12-lead ECG and medical history taking (22). The results of De Ruiter et al., Zwart et al., and the pilot study of Jansen et al., showed that cardiovascular causes for falls (and syncope) can be easily overlooked. A comprehensive standardized cardiovascular assessment will likely contribute to recognition of these cardiovascular causes, thereby presenting modifiable factors to reduce fall risk in older persons. However, it is important to remark that these additional tests should always be preceded by detailed history taking and physical examination as this remains the cornerstone in identifying cardiovascular risk factors for syncope or falls (46)(47)(48)(49).
Limitations of our review are the limited number of studies. Also, the differences in study populations make it difficult to compare studies and draw conclusions. A meta-analysis could not be performed because of the large heterogeneity of studies.

Conclusion
There is clear evidence that cardiovascular abnormalities are associated with increased fall risk in older persons. However, only few studies have investigated whether cardiovascular evaluations and interventions reduce falls, even though an increasing body of evidence shows that cardiovascular causes of falls are often underestimated. Furthermore, most studies have included only single cardiovascular interventions. This underlines the need for a well-designed randomized controlled trial to evaluate the efficacy of a broad cardiovascular evaluation and subsequent intervention in reducing falls. Additional testing should always be preceded by detailed and structured history taking and physical examination to identify cardiovascular risk factors for syncope or falls. A standardized assessment of cardiovascular risk factors with subsequent additional testing and accompanying interventions may be essential in preventing falls in older adults, and could consequently reduce injuries, loss of quality of life, deaths, and fall-related expenditures.