|Home | About | Journals | Submit | Contact Us | Français|
Driving is a complex activity that always becomes impaired at some point in older adults with degenerative dementia. Over time, disruption of the visual processing circuits of the brain that link the occipital and prefrontal regions, particularly in the right hemisphere, leads to increasing degrees of driving impairment that ultimately preclude safe driving. Neuropsychological tests of visuospatial ability, executive function and attention that tap into the integrity of these brain regions provide the clinician with important information regarding the need for a formal determination of driving competence. Enhancement of cognitive function in these domains through anti-dementia therapy and exercise may partially mitigate risk; however, all drivers with dementia must ultimately retire from driving when dementia becomes moderately severe, and often in earlier stages of the illness. Future efforts to improve screening tests for hazardous driving and to develop interventions to help prolong the time that drivers with mild dementia can continue to drive safely are needed for our increasingly aged and mobile population.
A wide variety of health-related problems in the elderly can have a negative impact on driving ability, including the neurologic changes of aging (e.g., bradykinesia and hearing deficits), cognitive changes (e.g., slowed reaction times and processing speed), eye disease (e.g., cataracts, macular degeneration and glaucoma) and other disorders (e.g., arthritis, diabetes, sleep disorders, Parkinson’s disease and dementia). A number of editorials and review articles have addressed the issues of driving risk and measures to predict at-risk drivers in older individuals and in those with dementia [1–7,101]. However, until recently, little research has been devoted to defining the underpinnings of hazardous driving in these people. Consequently, efforts to develop strategies and treatments to reduce this risk have been limited. This article seeks to fill this gap by focusing specifically on summarizing research that defines the physiological and cognitive changes related to degenerative dementia of the Alzheimer’s type that affect driving ability in older drivers. The hope is that such studies will inform the future efforts to enhance the safety of cognitively impaired older drivers that are discussed at the end of this special report.
Cognitive impairment is not uncommon in the older population of active drivers. In one community-based study of 3238 drivers aged 65 years and older, applying for renewal of their North Carolina driver’s license, who were examined using the Short Blessed Mental Status Examination, 6.2% of those aged between 65 and 69 years were moderately or severely impaired compared with 7.7% of those aged between 70 and 74 years, 11.9% of those aged between 75 and 79 years, and 18.7% of those aged 80 years or above .
Based on overall crash occurrences per year, older drivers have fewer crashes than younger drivers, largely owing to a reduction in driving miles. However, when gauged on a per mile basis, older drivers are a significant risk group, with crash rates approaching those observed in teenage drivers . Moreover, owing to frailty, older drivers are more likely to die in a motor vehicle accident than teenage drivers [9,10]. However, age itself should not be regarded as the problem but rather as a proxy for physical or cognitive impairments that lead to unsafe driving. With age, right-of-way and traffic sign violations increase, and the frequency of intersection crashes also increases [9,11,12]. These errors are generally thought to be related to changes in the driver’s reaction time, visual perception and attention.
Owing to similar but more severe deficits, drivers with dementia are at a high risk of unsafe driving, particularly if examined on a per miles driven basis. Studies have demonstrated that drivers with dementia have an approximately two to five times greater risk of involvement in a crash compared with age-matched controls . The risk of crashing for a driver with Alzheimer’s disease (AD) rises above that of the highest risk group (teenage males) beyond the third year of disease .
Whereas a number of cognitive deficits have been demonstrated to be related to driving impairment among those with dementia, the underlying neurophysiologic basis for such impairment is unknown. Are these factors caused by global or brain area-specific dysfunction? In order to address this issue, a single photon emission computed tomography (SPECT) study was carried out in drivers with AD .
A total of 79 subjects with very mild to moderate dementia (i.e., with a Clinical Dementia Rating [CDR]  of 0.5–2.0) were studied. The predictor variables were semiquantitative measurements of radionuclide activity in specific brain regions of interest, including the superior and inferior frontal, anterior and posterior temporal, parietal and lateral occipital cortex. These regions captured the major sites of AD pathology in the association cortex as well as regions involved in the dorsal and ventral streams of visual processing in the brain. The outcome variable was the caregiver’s global assessment of the patient’s driving ability.
Among the global measures of function, the Mini-Mental State Examination (MMSE)  did not significantly discriminate between the levels of driving ability described by the caregivers; however, a scale of instrumental activities of daily living (IADL) did discriminate between these levels. Based on factor analysis, changes in the temporoparietal regions, particularly in the right posterior temporal and occipital region, could differentiate between those who could drive alone safely and those who could not. Changes in the frontal lobes, as well as in the right occipital cortex could differentiate those who were unable to drive from those who drove but with some difficulty.
These observations do not prove causality but do mirror the progression of AD, which spreads over time from the temporoparietal cortex to more frontal involvement, with advancing disease impacting the two major visual processing streams in the brain (Figure 1). It is likely that in early-stage disease, involvement of the ventral region of visual processing, particularly in the right hemisphere, disrupts driving ability. This is the brain region that processes what we see.
With more advanced disease, it is likely that additional involvement of the dorsal stream that projects to the frontal lobes leads to further driving impairment. This is the brain region that processes where we see [18,19]. Involvement of the frontal projections, or the frontal lobes themselves, produces crucial deficits in spatial surveillance, effector circuits for rapid arousal and saccadic eye and body movements, which may be necessary for reacting at intersections and to avoid accidents .
Another more recent study reinforces the importance of normal function of the occipital and prefrontal cortex in healthy older adults for detecting driving hazards. In this study, 14 cognitively normal older drivers were studied for brain activation using functional MRI (fMRI) while they were shown video clips of road hazards and neutral road scenes. Hazard detection was associated with activation of the bilateral lateral occipital and right prefrontal cortex . In another fMRI study, performed in 12 healthy subjects during a driving simulation task, activations were observed in occipital and parietal regions bilaterally as well as in the sensorimotor cortex and cerebellum, suggesting that simulated driving requires coordinated activity between occipitoparietal and motor brain areas .
Near infrared spectroscopy is a new brain imaging technique that detects regional changes in blood oxygenation in the brain that relate to cortical activity in real time, similar to fMRI. Using this technique to measure brain activity while individuals were engaged in testing using a driving simulator, differences were demonstrated between 12 drivers with AD and 14 healthy elderly controls. During a collision avoidance task on the simulator, AD drivers showed less prominent increases in oxyhemoglobin levels in the prefrontal cortex compared with controls. These differences were correlated with a delay in braking. The most prominent differences were observed in the right lateral prefrontal cortex. This study reinforces the critical importance of frontal and right hemisphere dysfunction in older drivers with AD and suggests that those with a slow braking response to crash situations may have lost important resources in these areas .
A more direct examination of brain function during real-world driving was carried out by Jeong and colleagues. A total of 30 normal volunteers aged 20–56 years engaged in driving for 30 min immediately after injection with (18F)2-deoxy-2-fluoro-d-glucose (FDG). PET imaging that was performed after driving revealed activation in primary and secondary visual cortices, the primary sensorimotor, premotor, parietal association areas, cingulate and parahippocampal gyrus, as well as in the thalamus and cerebellum, indicating that driving requires coordination between the visuoperceptual and visuomotor areas of the brain, which is similar to findings from previous studies that used driving simulators .
To further examine the hypothesis of regional brain influences, clock drawings performed by drivers with dementia were analyzed . Clock drawing is a simple cognitive task that is often performed in the office or at the bedside. Clock drawing copying is regarded as a measure of visuospatial constructional ability, which is largely related to right posterior hemispheric function, while clock drawing to command is also regarded as a measure of executive function, which is frontally mediated. As predicted, early-stage driving impairment was significantly related to scores on the clock drawing copying task, while later-stage driving impairment was more significantly related to clock drawing on command.
This observation is of practical significance, since the clock drawing task has been recommended as a cognitive screening test for driving impairment in older adults [24–26] and is included in the American Medical Association’s recommendations for physicians , along with a Trail Making Test – another nonverbal test of executive and visuoperceptual function examining function in similar brain regions.
In conclusion, driving impairment in adults with AD has been correlated with changes in regional cortical function, which varies according to the severity of disease. Consequently, cognitive tests of visuoperceptual and executive function that assess frontal and right hemisphere resources appear to be more useful screening tools than global measures such as the MMSE . It is important to note that in other dementia disorders, such as frontotemporal dementia, dementia with Lewy bodies and vascular dementia, pathology may affect brain regions other than those in AD described previously, producing different patterns of cognitive deficits that could impair driving. This is an important area for future research.
The effects of aging on motor function and other disease-related factors, such as eye disorders, add to decrements in driving ability. Environmental factors such as the weather, time of day, familiarity with the route and street conditions, plus the driving behavior of other drivers encountered on the road, add importantly to the risk of the person with dementia becoming involved in a motor vehicle accident. A model of these complex interactions is provided in Figure 2. As we have discussed, the degenerative process of AD itself produces a continuum of evolving cognitive deficits according to specific brain region involvement that manifests as varying but increasing degrees of driving impairment over time.
Early intervention via forced or voluntary driving retirement can avoid serious accidents, yet one does not want to limit driving arbitrarily based on diagnosis alone, since autonomy for the elderly is an extremely important goal both socially and economically. Specifically, driving cessation in the elderly has been associated with a decline in out-of-home activity levels , increased depression  and an increased likelihood of entry into long-term care facilities . Therefore, the major question arises as to when should a cognitively impaired older adult stop driving, and how do we recognize when that time has come. To address these questions, a large number of epidemiologic risk factor studies and neuropsychological test studies have been performed using various outcome measures of driving competence, such as motor vehicle accidents, performance on driving simulators and on-road driving tests.
Based on empirical research, a number of potentially valid screening tests for drivers with dementia have emerged. One approach is the use of composite scores from neuropsychological test batteries [30,31]. A meta-analysis of 27 studies of drivers with dementia was conducted to examine the relationship between neuropsychological functioning and driving. Overall, the highest correlations were observed for individual tests of visuospatial skills and on-road or nonroad driving measures . Visuomotor and executive function tests, such as the Trail Making Test, have been correlated with road test performance [30,33]. In addition, the driving scenes task from the Neuropsychological Assessment Battery has been demonstrated to discriminate safe from unsafe drivers on road tests in a group of subjects with very mild dementia (CDR: 0.5) or normal cognition .
Performance on hand-drawn [35,102] and computerized mazes [33,36] has been demonstrated to be a useful predictor of road test performance in normal elders and in those with mild dementia. In one such study , the drawing time for five simple mazes correctly categorized 69% of a mixed sample of elderly control and dementia drivers as safe or not safe on a road test. Accounting for age improved the accuracy to 74%, and the addition of tests of working memory and trail making improved the accuracy of the model to 81%. Unfortunately, higher levels of accuracy using neuropsychological test predictor models may be limited as much by the characteristics of the outcome measure of driving competence as by the characteristics of the predictor variables.
Using more sophisticated computerized measures of visual attention, recent research has further contributed to our understanding of the critical components of cognition affecting driving in adults with dementia. Duchek and colleagues at Washington University, MO, USA, studied normal older drivers and those with very mild to mild dementia (CDR: 0.5–1.0) with computerized cognitive tests and road tests every 6 months over 3 years. Interestingly, the normal older drivers declined in driving ability as well as the impaired drivers, although not as severely. Road test scores were significantly related to a visual search task, accounting for 34% of the variance .
In a preliminary report from a second longitudinal study at Brown University, RI, USA, involving older drivers with early AD using comparable methods, Festa and colleagues reported that a two-part neurocognitive model could predict deterioration in on-road driving performance . In this study, subjects underwent a standardized road test, in addition to completing a computerized test battery with driving simulation at 6-month intervals over 24 months. As expected, road test driving performance progressively deteriorated over 24 months.
At 12 months, decrements in the ability to select a correct response when presented with distracting information predicted decline in on-road driving. By contrast, poorer road test performance at 24 months was predicted by deterioration in the subjects’ ability to accurately maintain lane position while simultaneously executing the visual search task. These results suggest that specific components of visual attention that are important for safe driving are impaired at different rates in adults with AD. Furthermore, a composite computerized task designed to simultaneously assess these specific processes could be useful for predicting evolving driving risk in AD.
Relatively little attention has been paid to the exploration of possible therapeutic modalities, such as medication or physical exercise, to extend the period of time that persons with early dementia can drive safely for. Discovery of effective treatment interventions might lead to short-term improvements in driving competency and decrease motor vehicle accident risk.
It is important to consider the potential negative and positive pharmacologic impact of commonly used medications on driving abilities in dementia. A number of CNS medications probably amplify crash risk in individuals with pre-existing cognitive impairment. In an observational study of Canadian older adult drivers, Rapoport and colleagues reported that prescription of psychotropic medication for individuals with a diagnosis of dementia significantly increased the risk of being involved in a motor vehicle crash, based on traffic records . Therefore, a review of all medications, with the goal of eliminating sedatives and other agents that may impair cognition or slow reaction time, is a prudent intervention for all persons with dementia who continue to drive.
How might the medications indicated for the treatment of mild-to-moderate AD affect driving skills in the early stages of the illness? In a small group of healthy, middle-aged pilots, cholinergic augmentation with donepezil was reported to improve performance in specific flight simulator tasks that place a high demand on divided and visual attention such as emergency detection and landing approach [40,41]. These observations provide a theoretical framework for speculation concerning how boosting levels of brain acetyl-choline in individuals with early AD might impact aspects of driving that require prompt, accurate reactions and the ability to monitor and shift between multiple attentional demands.
In a preliminary report, Daiello and colleagues reported on the effects of cholinesterase inhibitor (ChEI) treatment on performance in computerized tests of simulated driving, visual attention and executive function in drivers with very mild or mild AD (CDR: 0.5–1.0) . This was a two-part pilot study: first, a case–control comparison was made between the baseline test performance of untreated AD patients (ChEI nonusers) and a demographically-matched group of AD patients treated with stable doses of a ChEI (ChEI users) who had completed a previous longitudinal driving study. Second, an observational cohort study was conducted comparing the performance of the untreated AD patients at pretreatment baseline (pre-ChEI) and after 3 months of ChEI therapy (post-ChEI).
In the case–control comparison, ChEI treatment was found to consistently enhance the ability to maintain lane position in the simulated driving task; this effect was more pronounced when driving was performed alone (single task) in the ChEI users/nonusers comparison and under the dual-task condition (driving performed in conjunction with a visual search) for the pre-/post-observational treatment comparison. ChEI treatment improved the accuracy of target detection, overall response time and response time variability in the visual search task in both comparisons. ChEI treatment also reduced the detrimental impact of increasing distracting information under dual-task conditions for the pre-/post-treatment comparison. However, this effect was not observed in the ChEI users/nonusers comparison.
Taken together, the demonstration of similar effects of ChEI treatment across the two analyses provides converging evidence for the effects of cholinergic augmentation on driving-related attentional abilities in AD patients. However, even if future research finds that ChEI treatment enhances driving safety in individuals with AD, these effects are likely to be of limited duration owing to the progressive disruption of cholinergic pathways associated with increasing dementia severity.
Nonpharmacological interventions may also improve driving-associated cognitive abilities. A large amount of evidence is accumulating suggesting that physical exercise enhances and preserves cognitive function in the elderly. One recent study investigated the effects of participation in an exercise program on several abilities associated with driving performance in older adults. A total of 32 subjects were randomly assigned to either an exercise group or a control group. The 12-week exercise program stressed perceptive, cognitive and physical abilities. Significant positive effects were found at the 12-week follow-up. Behavioral speed improvements were observed in reaction time, movement time and response time (both in single- and dual-task conditions) while operating a driving simulator. In addition, visual attention improvements occurred in the speed of processing and divided attention. Psychomotor performance improvements occurred in lower limb mobility .
Since visual attention and executive cognitive processes are served by right hemisphere and frontal brain regions, discussed previously in association with driving impairment, it appears that these early intervention studies are tapping into relevant areas of partially reversible deficit. Unfortunately, long-lasting results are likely to be limited by the underlying progression of AD over time.
Most research and public policy directives dealing with drivers with dementia have focused on defining methods to detect unsafe drivers and removing them from the active driving pool. A number of consensus groups have published practice guidelines for clinicians in this regard [44–49, ]. There is broad consensus that moderately severe dementia precludes safe driving; however, there is still no consensus on how to deal with those with questionable or mild dementia who are minimally or only mildly dependent on others for assistance with their other daily living activities.
Pooled data from two longitudinal studies [50,51] involving 134 drivers with dementia show that 88% of drivers with very mild dementia (CDR: 0.5) and 69% of drivers with mild dementia (CDR: 1.0) are still able to pass a formal road test. In the Brown study, 77% of early driving terminations at 18 months were due either to hazardous driving (55% were related to road test failure and 2% were related to motor vehicle accident) or family decisions based on the progression of dementia symptoms (20%). Moreover, the median time to discontinuance of driving for those with very mild dementia was 2 years, and for those with mild dementia the mean discontinuance was 1 year .
Future efforts should be made to maximize the safety and driving competence of those who are still actively driving with cognitive impairment owing to degenerative dementia. This issue will only become more compelling as the baby boom generation enters a period of advanced aging yet aims to retain as vigorous a lifestyle as possible.
Currently, a number of potentially valid screening tests are available that could be used in the clinic and in motor vehicle registry settings to identify at-risk drivers who are in need of closer evaluation and could potentially benefit from remedial actions in order to improve their safety by restrictions on driving, use of assistive personnel and devices, medical treatment and exercise. However, before adopting a screening test, research on currently available tools is greatly needed in order to define appropriate cut-off scores that provide acceptable levels of sensitivity and specificity . Development of uniform standards for road testing may improve outcomes; for example, in a longitudinal study based at an academic medical center, when evaluated by road tests every 6 months, crash rates for drivers with dementia declined to the levels of healthy control drivers during a period of 3 years . However, the costs of such detailed surveillance may be prohibitive and it is unknown whether community-based road testing programs would reproduce these results.
Although it is generally accepted as a valid standard for driving competence – since it is used by motor vehicle registries to grant licenses – he road test is not without problems. Anxiety related to taking the test  and environmental cues  are among the factors that may directly affect the road test performance of older drivers with dementia. Road test designs that are specifically tailored to detect the types of driving performance errors encountered in this particular population need to be examined. Naturalistic driving studies that are currently ongoing will inform researchers regarding these issues.
It is generally recognized that memory loss alone does not impair the critical cognitive processes needed for safe driving, but it could impact the ability of a driver to navigate without getting lost and may increase safety risk under challenging circumstances . Often, family members try to compensate by having a nonimpaired driver serve as ‘copilot’ ; however, the advisability of this approach has been challenged . Assistive technologies, such as user-friendly GPS devices and accident warning systems, need to be developed in order to maximize independent living for those with cognitive impairment .
Efforts to enhance physical and cognitive function may enable impaired drivers to drive more safely and should be explored. The preliminary results of one study suggesting benefit for critical cognitive functions after treatment with ChEIs have been discussed. Newer anti-dementia therapies in the future will hopefully provide a more effective reduction of symptoms and prevention of decline.
Based upon our knowledge regarding the pathophysiology, cognitive factors and natural history of driving impairment of those with dementia, it is hoped that we can develop a more enlightened approach to this difficult issue by recognizing that a diagnosis of early dementia or AD does not necessarily preclude safe driving and that efforts to maximize the function and safety of these drivers are highly desirable in order to preserve personal autonomy and limit unnecessary social disability.
Financial & competing interests disclosure
Dr Brian R Ott has received grant funds from Pfizer. This work was supported by the National Institute of Aging Grant no. AG16335 awarded to Dr Brian R Ott and by AHRQ K-08 Grant no. HS017735 awarded to Dr Lori A Daiello. Preliminary reports of studies cited in this special report [38,42] were presented by Dr Daiello and Dr Festa at the 12th International Conference on Alzheimer’s Disease, Chicago, 26–31 July 2008. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Dr Brian R Ott, Department of Neurology, Warren Alpert Medical School of Brown University, RI, USA and The Alzheimer’s Disease & Memory, Disorders Center, Rhode Island Hospital, RI, USA, Tel.: +1 401 444 6440, Fax: +1 401 444 6858, Email: gro.napsefil@ttob.
Dr Lori A Daiello, Department of Neurology, Warren Alpert Medical School of Brown University, RI, USA and The Alzheimer’s Disease & Memory, Disorders Center, Rhode Island Hospital, RI, USA, Tel.: +1 401 444 6440, Fax: +1 401 444 6858, Email: gro.napsefil@olleiadl.
Papers of special note have been highlighted as:
• of interest
•• of considerable interest