The Author Published by Oxford University Press on behalf of The Gerontological Society of America.

Transcription

1 Journal of Gerontology: MEDICAL SCIENCES The Author Published by Oxford University Press on behalf of The Gerontological Society of America. Cite journal as: J Gerontol A Biol Sci Med Sci July;65(7): All rights reserved. For permissions, please journals.permissions@oxfordjournals.org. doi: /gerona/glq013 Advance Access published on February 16, 2010 Which Types of Activities Are Associated With Risk of Recurrent Falling in Older Persons? G. M. E. E. (Geeske) Peeters, 1 Lisanne M. Verweij, 2 Natasja M. van Schoor, 1 Mirjam Pijnappels, 3 Saskia M. F. Pluijm, 4 Marjolein Visser, 1,5 and Paul Lips 1,6 1 Department of Epidemiology and Biostatistics and 2 Department of Public and Occupational Health, EMGO Institute for Health and Care Research, VU University Medical Center, Amsterdam, The Netherlands. 3 Research Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, The Netherlands. 4 Department of Public Health, Erasmus MC, Rotterdam, The Netherlands. 5 Department of Health Sciences, Faculty of Earth and Life Sciences, VU University Amsterdam, The Netherlands. 6 Department of Internal Medicine, Section of Endocrinology, VU University Medical Center, Amsterdam, The Netherlands. Address correspondence to P. Lips, MD, PhD, Department of Internal Medicine, Section Endocrinology, VU University Medical Center, Postbus 7057, 1007 MB Amsterdam, The Netherlands. p.lips@vumc.nl Background. This study explored the associations between various types of activities, their underlying physical components, and recurrent falling in community-dwelling older persons. Methods. This study included 1,329 community-dwelling persons ( 65 years) of the Longitudinal Aging Study Amsterdam (LASA). The time spent in walking, cycling, light and heavy household activities, and two sports was measured using the LASA Physical Activity Questionnaire (LAPAQ). Physical activity components included strength, intensity, mechanical strain, and turning. Time to second fall in a 6-month period was measured during 3 years with fall calendars. Cox proportional hazards models were adjusted for confounders and stratified for physical performance and sex in case of significant (p <.10) interaction. Results. During 3 years, 325 (24.5%) persons became recurrent fallers. In women, doing light (hazard ratios [HRs] = 0.40, 95% confidence intervals [CIs] = ) or heavy household activities (HR = 0.63, CI = ) was associated with a decreased risk of recurrent falling. In persons with good physical performance, doing sports (HR = 1.56, CI = ), high intensity (HR > 1.75, CI = ), and high mechanical strain (HR = 1.70, CI = ) activities was associated with an increased risk of recurrent falling. Conclusions. The results suggest that the relationship between physical activity and recurrent falling differs per type of activity and is modified by physical performance. Doing household activities was associated with a decreased risk of recurrent falling in women. In physically fit older persons, doing sports or activities with high intensity or mechanical strain demands was associated with an increased risk of recurrent falling. Key Words: Physical activity Physical performance Accidental falls Aged. Received June 8, 2009; Accepted January 3, 2010 Decision Editor: Luigi Ferrucci, MD, PhD Falling is a major source of injuries and disabilities in older persons. Annually, about 30% of the older persons falls at least once and 15% falls at least twice (1,2). The consequences of falling vary from none to major injuries, such as fractures and nursing home admittance (3,4). Both high and low levels of physical activity have been associated with an increased fall risk (5 7). The relationship between physical activity and falling is complex and influenced by many factors, such as physical capacity, type of activity, and environmental circumstances. The results of one study showed that doing household activities, but not leisure activities, was associated with an increased fall risk (6). A study on the circumstances of falling found that falls frequently occurred during walking, carrying objects, changing position, and reaching (8). These findings imply that the relationship between physical activity and falling may be different for various types of activities. However, the first study included only men and did not examine associations with other common activities, such as walking and cycling. The second study included only women and focused on actions rather than activities. Activities can be split into components, which can be translated into requirements of certain aptitudes, such as strength, mechanical strain, intensity, and turning actions (9). Potential differences in the associations between various types of activities and falling may be explained by their underlying activity components. This hypothesis is supported by studies that reported that engagement in rapid and forceful activities increases the risk of injurious falls (5,10,11). However, the association between 743

2 744 Peeters et al. underlying components of activities and falling has not yet been studied. This study aimed to examine which types of physical activities and which underlying activity components are associated with recurrent falling in older persons. We focused on recurrent falling because the consequences of falling appear to be more severe in recurrent fallers than in once fallers (4). Moreover, a single fall may be coincidental and caused mainly by environmental factors, whereas recurrent falls are mainly caused by physical, cognitive, and behavioral factors within the person (12). Methods Participants This study was performed within the Longitudinal Aging Study Amsterdam (LASA), an ongoing interdisciplinary cohort study on predictors and consequences of changes in physical, cognitive, emotional, and social functioning in older persons (13). A random sample of older persons, stratified for age, sex, and expected 5 years mortality rate, was drawn from the population registries of 11 municipalities in The Netherlands. The sample is representative for the older Dutch population with respect to geographic region and degree of urbanization (14). The sampling and data collection procedures have been described in detail elsewhere (13,15). The medical ethics committee approved the study, and all participants signed informed consent. The sample for this study consisted of 1,509 participants who were aged 65 years or older and who took part in the second cycle ( ) of LASA. Participants were not preselected based on fall risk or fall history. Fall follow-up was available from 1,427 participants, of whom 45 and 53 participants had missing values on physical activity or any of the confounders, respectively. Complete data were available of 1,329 participants. Time to Recurrent Falling Falls were assessed during 3 years following the interview in using fall calendars (1). A fall was defined as an unintentional change in position resulting in coming to rest at a lower level or on the ground (16). Participants were asked to tick per week whether or not they had fallen. Once per 3 months, the calendar page was mailed to the institute. If the calendar procedure was too complicated because of cognitive or physical limitations, if no page was received (even after a reminder), or if the page was completed incorrectly, the participants were contacted per telephone. Proxies were contacted if participants were unable to respond. Recurrent falling was defined as falling at least two times in 6 months (ie, 26 weeks) during the 3-year fall follow-up (17). Time in weeks from baseline (date of interview) to the date of the second fall in a 6-month period was determined as the time to recurrent falling. For example, a person who fell 3, 30, and 33 weeks after the baseline interview was classified as recurrent faller based on the second and third fall, with a time to recurrent falling of 33 weeks. Types of Physical Activities The physical activities were measured using the validated LASA Physical Activity Questionnaire (LAPAQ) (18), an interviewer-mediated questionnaire, which estimates the frequency and duration of participation in six activities in the previous 2 weeks. The activities were walking outside, cycling, light and heavy household work, and a maximum of two sports. Per activity, the frequency and duration were multiplied and subsequently divided by 14 to express the activity in minutes per day. The time spent in walking was dichotomized using the median value into less than 15 versus greater than or equal to 15 minutes per day. The time spent doing light household activities was dichotomized using the median value into less than 60 versus greater than or equal to 60 minutes per day. Heavy household activities, cycling, and sports were dichotomized based on participation (0 vs 1 min/d) because more than half of the participants did not participate in these activities. Participants with missing values on the LAPAQ who were bedridden or wheelchair dependent were assigned to the least active categories (n = 19). Physical Activity Components Four components of physical activity were distinguished: strength, turning, mechanical strain, and intensity (9). Per component, a score was assigned to each activity of the LAPAQ. Strength was defined as the amount of required lower body muscle strength to perform an activity. Scores were based on expert opinion (occupational therapists, movement scientist, and rehabilitation physician) and then validated using the average strength measured in participants performing the activities (range 1 4). Turning was defined as the amount of transversal rotations of the lower extremities in an activity. All activities were scored by experts on rate of transversal rotations of the knee and hip joint (range 1 3). The scores for mechanical strain and intensity were based on scores and definitions available in the literature. Mechanical strain was defined as loading of the bone and was derived from ground reaction forces expressed in multiples of body weight (range 1 4) (19). Intensity was expressed in metabolic equivalents (MET) as described by Ainsworth (20) (range ). An MET score of 2.5 was assigned to activities such as light housekeeping and billiards, and an MET score of 7.0 was assigned to activities such as tennis and running. The scores were processed in three different ways, as explained in Appendix 1. The average score indicates the average intensity (or strength, turning, and/or mechanical strain) required for the person s activity pattern. The sum score indicates the total intensity

3 TYPE OF ACTIVITIES AND RECURRENT FALLING 745 (or strength, turning, and/or mechanical strain) needed for the person s activity pattern. The maximum score indicates the upper limit of intensity (or strength, turning, and/or mechanical strain) required in the person s activity pattern. Each of these scores emphasizes a distinct aspect of the activity pattern and may be associated differently with recurrent falling. The distributions of the average scores, sum scores, and maximum scores were skewed and therefore categorized into tertiles: low, medium, and high. The cutpoints for the tertiles were based on the current sample and are given in Appendix 1. Effect Modifiers Potential effect modifiers were sex and physical performance because women and men and persons with good and poor physical performance may differ in activity patterns and physical capacities. In this study, physical performance was measured using three standardized tests measuring the ability to walk, rise from a chair, and maintain balance (21). The chair stands test measures the time needed to stand up from a chair and sit down for five times. The walk test measures the time needed to walk 3 m along a line, turn 180, and walk back. During the tandem stand, the participant stands with one foot behind the other (heel against toe) up to 10 seconds. The scores of the chair stands and walk test range from 1 (slowest) to 4 (fastest). The score of 0 was assigned when the participant was unable to complete a test. For the tandem stand, 0 points were scored when the participant was able to hold for less than 3 seconds, 2 points for 3 9 seconds, and 4 points for 10 seconds. The three scores were summed (range 0 12), and a score of 12 represents optimal physical performance (17,21). To examine effect modification, the score was dichotomized using the median score of 7 as a cutoff value (poor physical performance: 0 7 vs good physical performance: 8 12). This cutoff value was chosen because of practical considerations: If an interaction with physical performance was significant, further analysis would have to be stratified for physical performance. The median value was chosen to maintain sufficient power per strata. Potential Confounders Potential confounders were age, sex, level of education, body mass index (BMI), chronic diseases, dizziness, vision, alcohol use, psychotropic medication, polypharmacy, cognitive functioning, depressive symptoms, and fear of falling. The relationship between the activity components and time to recurrent falling may be additionally confounded by total physical activity. Level of education was assessed as the highest level of education completed (range 5 years, primary education, to 18 years, university). BMI was calculated as weight (kilograms) per height (square meters). The number of chronic diseases was assessed using a questionnaire, including chronic nonspecific lung diseases, cardiovascular disease, stroke, diabetes mellitus, malignant neoplasms, and joint disorders (range 0 7). Dizziness was assessed by asking, Are you dizzy regularly? (yes or no). Vision was assessed by asking, Can you recognize a face at 4 meters? (yes or no). Self-reported number of glasses of alcohol per week was categorized according to the Garretsen index (range, 0 = does not drink to 4 = very excessive drinker) (22). Medication use was assessed by copying the names of the medications directly from the containers. Use of psychotropic medication (ie, antipsychotics, antidepressants, anxiolytics, or hypnotics) was dichotomized as nonusers versus users. Polypharmacy was defined as taking four or more medications (yes or no). Cognitive functioning was measured using the Mini-Mental State Examination (range 0 30) (23). Depressive symptoms were assessed using the Center for Epidemiological Studies-Depression scale (range 0 60). Fear of falling was measured using a modified version of the Falls Efficacy Scale (24). The participants reported their concern (0 = not concerned to 3 = very concerned) for falling while carrying out 10 activities of daily living (range 0 30). Total physical activity was measured as the minutes per day participated in each of the activities (walking, cycling, light or heavy household activities, and first and second sports) and summed (minutes per day) (20,25). Statistics Differences in baseline characteristics between recurrent fallers and nonrecurrent fallers were tested with the t test for normally distributed variables, Mann Whitney U test for skewed variables, and the chi-squared test for categorical variables. To examine the associations between types and components of physical activity and recurrent falling, a hazard ratio (HR) and 95% confidence interval (CI) were calculated using Cox proportional hazard models. First, we examined interactions of the various activities with sex and physical performance. If an interaction was significant (p <.10) (26), further analyses were stratified. Second, we examined the unadjusted associations between various activities and recurrent falling. Third, we examined these associations after adjustment for confounders that changed the regression coefficient with more than 10%. The least active group was used as a reference group in all analyses. We performed the same procedure to examine the associations between activity components and recurrent falling. p Values were based on two-sided tests and were considered statistically significant at p <.05. All analyses were conducted using SPSS software (version ). Results In this study, 1,329 participants were included, of whom 325 (24.5%) appeared to be recurrent fallers. A total of 180 persons were excluded from the analyses because of incomplete fall follow-up (n = 82) or missing values on physical activity (n = 45) or any of the potential effect modifiers or

4 746 Peeters et al. Table 1. Baseline Characteristics Recurrent Fallers (n = 325) Nonrecurrent Fallers (n = 1,004) p Value Type of activity Walking (% 15 min/d) Cycling (% yes) Light household (% 60 min/d) Heavy household (% yes) <.001 Sports (% yes) Confounders and effect modifiers Age (years) 76.8 (6.8) 74.8 (6.3) <.001 Sex (% women) Body mass index (kg/m 2 ) 26.5 (4.0) 27.1 (4.3).03 Chronic diseases (range 0 7) 1.0 [ ] 1.0 [ ].03 Dizziness (% yes) <.001 Vision (% poor) Alcohol (range 0 4) 1.0 [ ] 1.0 [ ].46 Psychotropic <.001 medication (% yes) Polypharmacy (% 4 medications) Years of education 9.0 [ ] 9.0 [ ].02 (range 5 18 years) Depressive symptoms (range 0 60) 8.0 [ ] 5.0 [ ] <.001 MMSE (range 0 30) 28.0 [ ] 28.0 [ ].29 Fear of falling (range 0 30) 1.0 [ ] 0.0 [ ] <.001 Physical performance 7.0 [ ] 8.0 [ ] <.001 (range 0 12) Total physical activity (min/d) 115 [66 179] 136 [77 205].001 Notes: MMSE = Mini-Mental State Examination. Mean (SD). Median [interquartile range]. confounders (n = 53). Nonresponse analysis was based on the data available from these excluded participants; therefore, the sample size varies per outcome measure that was analyzed. The 180 excluded persons were older, had lower levels of education, had poorer physical and mental health conditions, and were more often recurrent fallers (p.001 for all characteristics). Table 1 presents the baseline characteristics for recurrent fallers and nonrecurrent fallers. Recurrent fallers were older (p <.001), had lower BMI (p =.03), higher levels of education (p =.02), lower total physical activity (p =.001), and participated less often in heavy household activities (p <.001) than nonrecurrent fallers. Furthermore, recurrent fallers had more chronic diseases, dizziness, polypharmacy, depressive symptoms, fear of falling, and used more psychotropic medication (p.03). Sex modified the associations between light and heavy household activities and recurrent falling (p.07). Physical performance modified the associations between sports and recurrent falling and in women between light household activities and recurrent falling (p.04). In women with a good physical performance, those who did light household activities for at least 60 minutes per day had a 60% lower risk of recurrent falling than those who did these activities for less than 60 minutes per day (adjusted HR = 0.40, CI = Table 2. Associations Between the Types of Physical Activity and Time to Recurrent Falling n (% events) Unadjusted Adjusted HR CI HR CI Walking outside 0 15 min/d 624 (26.8) min/d 705 (22.4) Cycling No 722 (26.0) Yes 607 (22.6) Light household activities Women, poor performance 0 60 min/d 71 (35.2) min/d 330 (27.0) Women, good performance 0 60 min/d 27 (40.7) min/d 249 (17.3) Men 0 60 min/d 364 (24.2) min/d 288 (24.0) Heavy household activities Women No 284 (33.1) Yes 389 (18.5) Men No 283 (26.9) Yes 360 (21.7) Sports Poor performance No 483 (30.4) Yes 208 (28.8) Good performance No 337 (14.5) Yes 301 (22.9) Note: HRs and CIs are presented unadjusted and adjusted for age, education, Mini-Mental State Examination, body mass index, chronic diseases, depression, fear of falling, dizziness, alcohol use, and psychotropic medication. The least active groups are the reference categories. CIs = 95% confidence intervals; HRs = hazard ratios ; Table 2). Furthermore, the fall risk in women who did heavy household activities was 37% lower than those who did not do heavy household activities (adjusted HR = 0.63, CI = ). In men, no associations were found between light or heavy household activities and recurrent falling. In participants with a poor physical performance, no associations were found between sports and recurrent falling. Contrastingly, in participants with a good performance, those who participated in sports had a 56% higher risk of recurrent falling than those who did not (adjusted HR = 1.56, CI = ). Participants who walked outside for at least 15 minutes per day had a 22% lower risk of recurrent falling than those who walked outside less than 15 minutes per day (HR = 0.78, CI = ). However, after adjustment for confounding, this relationship was no longer significant. No crude or adjusted associations were found between cycling and recurrent falling. Table 3 shows that recurrent fallers had lower sum scores for each of the four components compared with nonrecurrent fallers (p <.05). Recurrent fallers also had lower average

5 TYPE OF ACTIVITIES AND RECURRENT FALLING 747 Table 3. Average, Sum and Maximum Scores for Each of the Physical Activity Components Recurrent Fallers (n = 325) Nonrecurrent Fallers (n = 1,004) p Value Average scores Strength 2.5 [ ] 2.7 [ ] <.001 * Mechanical strain 1.5 [ ] 1.5 [ ].42 Intensity 3.5 [ ] 3.6 [ ].18 Turning 1.5 [ ] 1.5 [ ].77 Sum scores Strength 8.0 [ ] 8.0 [ ].01 * Mechanical strain 5.0 [ ] 5.0 [ ].03 * Intensity 12.0 [ ] 12.6 [ ].05 * Turning 5.0 [ ] 5.0 [ ].03 * Maximum scores Strength 3.0 [ ] 4.0 [ ] <.001 * Mechanical strain 2.0 [ ] 2.0 [ ].55 Intensity 4.5 [ ] 4.5 [ ].16 Turning 2.0 [ ] 2.0 [ ].67 Notes: Presented are the medians and interquartile ranges for the physical activity component scores. p Values are based on the Mann Whitney U test. The Mann Whitney U test ranks all observations as if they were from a single sample. Per group, the sum of the ranks is calculated and the sums are compared. This procedure and the large sample size allows for statistically significant differences despite similar median values and interquartile ranges. * Component scores are significantly (p <.05) lower in recurrent fallers compared with nonrecurrent fallers. and maximum strength scores (p <.001). The associations between any of the component scores and time to recurrent falling were not modified by sex (p >.10). However, physical performance interacted with the sum scores of mechanical strain, intensity, and turning and with the average intensity scores and maximum strength scores in the association with recurrent falling (p.08). To enhance the comparability of the results between the various component scores, unstratified associations are presented for all component scores and additional stratified analyses are presented if relevant (Table 4). Although most of the unadjusted associations between the component scores and recurrent falling were statistically significant (data not shown), most associations became nonsignificant after adjustment for the confounders. Participants with high sum and maximum scores for intensity had a higher risk of recurrent falling than those with low sum or maximum intensity scores (sum scores: HR = 1.35, CI = and maximum scores: HR = 1.74, CI = ). These associations were mainly explained by higher risks of recurrent falling in persons with a good physical performance who scored high on average and sum scores for intensity (average scores: HR = 1.75, CI = and sum scores: HR = 1.95, CI = ). In addition, in participants with a good performance, those who scored high on sum scores for mechanical strain had an increased risk of recurrent falling compared with low scoring participants (HR = 1.70, CI = ). To test whether the sports activities with high underlying requirements were associated with an increased fall risk rather than those with lower requirements, post hoc analysis were done in which we split the sports activities according to their underlying component scores. Persons with a good physical performance who participated in sports with high scores for intensity (>4), strength (>2), or turning (>1) had a higher risk of recurrent falling than those who did not participate in sports (HR = , CI = ). In addition, persons who participated in sports with low requirements did not differ in recurrent fall risk compared with those who did not participate in sports (data not shown). Discussion This study explored whether and how various types of activities and their underlying components were associated with recurrent falling in older persons. The results showed that a lower risk of recurrent falling was found in women who performed heavy household activities and women with a good physical performance who performed light household activities compared with women who were not active in housekeeping. Among persons with good physical performance, the risk of recurrent falling was higher in those who participated in sports. Especially in persons with a good physical performance, high-intensity activity patterns tended to increase the risk of recurrent falling. In the literature, high levels of physical activity have been associated with both an increased and a decreased fall risk (5 7,27). In addition, some studies reported that the association was modified by measures of physical functioning (6,27), whereas in our previous study, no interaction was found (28). Apart from the many methodological differences (eg, sample size, definition of falling, study design), the current findings suggest that these discrepancies may be explained by the contrasting directions of the associations between the various types of activities and (recurrent) falling. Household activities were associated with a decreased fall risk, whereas sports activities were associated with an increased fall risk. Furthermore, physical performance modified the associations of recurrent falling with some but not all types of activities. In our study, no associations were found between light or heavy household activities and fall risk in men. In the MrOs study, an increased fall risk was found in men of the most active quartile relative to the least active quartile of housekeeping (6). Three differences between the studies may explain these contrasting findings. First, the MrOs study had a larger sample size (5,997 vs 652 men in our study) and thus more power to detect an association. However, the risk ratio in their study (RR = 1.17, CI = ) was similarly small as the hazard ratio in our study (HR = 1.12, CI = ). Second, different measures of physical activity were used and they categorized the level of activity in quartiles and only reported a significant difference between the highest and lowest categories. Third, the outcome measure differed: We used time to recurrent falling, whereas they used the probability of falling within 4 months, including occasional

6 748 Peeters et al. Table 4. The Associations Between Physical Activity Components and Time to Recurrent Falling Strength Mechanical Strain Intensity Turning HR CI HR CI HR CI HR CI Average scores Total group Low Medium High Poor performance Low 1.00 Medium High Good performance Low 1.00 Medium High Sum scores Total group Low Medium High Poor performance Low Medium High Good performance Low Medium High Maximum scores Total group Low Medium High Poor performance Low 1.00 Medium High Good performance Low 1.00 Medium High Note: HR and CI are presented after adjustment for age, body mass index, physical activity, dizziness, depression, and fear of falling. Lowest categories are the reference categories. The analyses were stratified for physical performance in case of significant interaction (p <.1). CI = 95% confidence intervals; HR = hazard ratios. falls caused by external risk factors. Maybe in housekeeping activities, the external risk factors (such as standing on an unstable stepladder or walking on a wet floor) may cause men to fall once, but the active lifestyle may keep them fit enough to prevent recurrent falling, which are more likely to be caused by intrinsic risk factors (29). Among participants with a poor physical performance, no associations were found between sports activities and light household activities and recurrent falling. This may suggest that these persons were aware of their physical limitations and adjusted their activity pattern (avoid hazardous activities) or the performance of activities (eg, slower, more cautious) accordingly. However, lack of significant associations may also be due to selection bias. Nonresponders had poorer physical performance, were less active, and were more often recurrent fallers, which may have resulted in an underestimation of the actual relationships. Further research in this group is needed to verify these explanations. Physically good functioning persons who participated in sports had an increased risk of recurrent falling. Also, in these persons, participation in activities with high intensity and mechanical strain requirements was associated with an increased risk of recurrent falling. We performed post hoc analyses to test whether sports activities with high underlying requirements explained the increased risk of recurrent falling associated with sports activities in this group. The findings support the idea that persons with a good physical performance may have an increased risk of recurrent falling when they participate in sports with high demands on strength, intensity, or turning requirements. This suggests that activities

7 TYPE OF ACTIVITIES AND RECURRENT FALLING 749 with high requirements may improve physical fitness but can also lead to more falls due to increased exposure to hazardous situations. Although it has been shown that effective exercise for fall prevention should challenge balance (30), it is therefore recommended that hazardous activities should be avoided and sports activities should be performed in a safe environment and on a regular base with moderate intensity. This is in line with the Dutch recommendations for physical activity, which state that older persons should spent 30 minutes per day, at least 5 days per week on moderately intensive activities (MET scores of 3 4) (31). These recommendations mainly focus on maintaining physical fitness and health. Examples of activities with medium-intensity demands are cycling, gymnastics, ironing, and sweeping. The component scores were calculated in three different ways emphasizing distinct aspects of the activity pattern. Each of the approaches has its limitations, which will be discussed using Appendix 1. Although the average score is a good representation of ones varied activity pattern, its limitation is that a person who engages in many activities with varying component scores is underscored compared with a person engaging in few activities with higher scores. Although both Persons A and B are assigned to the medium category, the intensity of the activity pattern of B appears higher. Summing the component scores compensated for this limitation. Using the sum scores, Person B was assigned to the low-intensity category, whereas Person A was assigned to the medium-intensity category. However, using the sum scores, a person who engages in many lower demanding activities may be assigned to a higher category than a person with few but higher demanding activities (eg, B vs C). This limitation was overtaken by calculating the maximum component scores, which reflect the maximum demands that are required in the person s activity pattern. Although the various calculations emphasize different aspects of the physical activity pattern, there is no clear preference for one of the calculations based on the current results. Note that none of these measures reflect ones capacity. Strengths of this study are its long-term fall follow-up and large sample size, which increase the power to detect recurrent fallers. The sample is representative for the communitydwelling older population in The Netherlands (14). Physical activity was based on self-report, but the questionnaire has been validated for older persons (18) and includes both leisure and household activities. Although accelerometry is a more objective method to measure physical activity, it does not distinguish between all the types of activities included in the current questionnaire. Yet, three limitations need to be discussed. First, the component scores have been developed and validated in the same data set as used in this study (9). Future use of these scores in different data sets is necessary to evaluate its external validity. Second, many tests were done and the numbers of participants and events in some groups were relatively small due to stratification for both sex and physical performance. Therefore, some of the associations found may be due to Type I error. Third, 180 participants were excluded from the analyses due to missing values on any of the variables. Nonresponse analyses show that these persons were older, had poorer health conditions, tended to be less active, and were more often recurrent fallers. Consequently, the current results may have underestimated the actual associations between types of physical activities and recurrent falling, mainly among persons with a poor physical performance. In conclusion, the relationship between physical activity and falling differs for various types of activities. Doing household activities was associated with a decreased risk of recurrent falling in women. In physically fit older persons, doing sports, especially sports with high intensity, strength, and turning demands, was associated with an increased risk of recurrent falling. High-intensity activity patterns seemed to be associated with an increased risk of recurrent falling, mainly in persons with a good physical performance. Funding This work was supported by the Dutch Ministry of Health, Welfare and Sports, and The Netherlands Organisation for Scientific Research (NWO grant no to M.P.). References 1. Tromp AM, Smit JH, Deeg DJ, Bouter LM, Lips P. Predictors for falls and fractures in the Longitudinal Aging Study Amsterdam. J Bone Miner Res. 1998;13(12): Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med. 1988;319(26): Stel VS, Smit JH, Pluijm SM, Lips P. Consequences of falling in older men and women and risk factors for health service use and functional decline. Age Ageing. 2004;33(1): Dunn JE, Rudberg MA, Furner SE, Cassel CK. Mortality, disability, and falls in older persons: the role of underlying disease and disability. Am J Public Health. 1992;82(3): O Loughlin JL, Robitaille Y, Boivin JF, Suissa S. Incidence of and risk factors for falls and injurious falls among the community-dwelling elderly. Am J Epidemiol. 1993;137(3): Chan BK, Marshall LM, Winters KM, Faulkner KA, Schwartz AV, Orwoll ES. Incident fall risk and physical activity and physical performance among older men: the Osteoporotic Fractures in Men Study. Am J Epidemiol. 2007;165(6): Heesch KC, Byles JE, Brown WJ. Prospective association between physical activity and falls in community-dwelling older women. J Epidemiol Community Health. 2008;62(5): Nachreiner NM, Findorff MJ, Wyman JF, McCarthy TC. Circumstances and consequences of falls in community-dwelling older women. J Womens Health (Larchmt). 2007;16(10): Verweij LM, van Schoor NM, Deeg DJ, Dekker J, Visser M. Physical activity and incident clinical knee osteoarthritis in older adults. Arthritis Rheum. 2009;61(2): Koski K, Luukinen H, Laippala P, Kivela SL. Risk factors for major injurious falls among the home-dwelling elderly by functional abilities. A prospective population-based study. Gerontology. 1998;44(4): Speechley M, Tinetti M. Falls and injuries in frail and vigorous community elderly persons. J Am Geriatr Soc. 1991;39(1): Lord SR, Ward JA, Williams P, Anstey KJ. Physiological factors associated with falls in older community-dwelling women. J Am Geriatr Soc. 1994;42(10):

8 750 Peeters et al. 13. Deeg DJ, van Tilburg T, Smit JH, de Leeuw ED. Attrition in the Longitudinal Aging Study Amsterdam. The effect of differential inclusion in side studies. J Clin Epidemiol. 2002;55(4): Deeg D, Knipscheer C, van Tilburg W. Autonomy and well-being in the aging population: concepts and design of the Longitudinal Aging Study Amsterdam. Bunnik, The Netherlands: Netherlands Institute of Gerontology; Smith J, de Vries M. Procedures and results of the field work. In: DJH Deeg and M Westendorp-de Serriere (eds). Autonomy and wellbeing in the aging population I: report from the Longitudinal Aging Study Amsterdam Amsterdam, The Netherlands: VU University Press; Kellogg Interna tional Work Group.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(suppl 4): Stel VS, Smit JH, Pluijm SM, Lips P. Balance and mobility performance as treatable risk factors for recurrent falling in older persons. J Clin Epidemiol. 2003;56(7): Stel VS, Smit JH, Pluijm SM, Visser M, Deeg DJ, Lips P. Comparison of the LASA Physical Activity Questionnaire with a 7-day diary and pedometer. J Clin Epidemiol. 2004;57(3): Kemper HC, Twisk JW, van Mechelen W, Post GB, Roos JC, Lips P. A fifteen-year longitudinal study in young adults on the relation of physical activity and fitness with the development of the bone mass: the Amsterdam Growth And Health Longitudinal Study. Bone. 2000;27(6): Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc. 1993;25(1): Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol. 1994;49(2):M85 M Garretsen HF. Probleemdrinken, prevalentiebepaling, beinvloedende factoren en preventiemogelijkheden, Theoretische overwegingen en onderzoek in Rotterdam. Lisse, The Netherlands; Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3): Tinetti ME, Richman D, Powell L. Falls efficacy as a measure of fear of falling. J Gerontol. 1990;45(6): Caspersen CJ, Bloemberg BP, Saris WH, Merritt RK, Kromhout D. The prevalence of selected physical activities and their relation with coronary heart disease risk factors in elderly men: the Zutphen Study, Am J Epidemiol. 1991;133(11): Collett D. Modeling survival data in medical research. London, UK: Chapman & Hall; Graafmans WC, Lips P, Wijlhuizen GJ, Pluijm SM, Bouter LM. Daily physical activity and the use of a walking aid in relation to falls in elderly people in a residential care setting. Z Gerontol Geriatr. 2003;36(1): Peeters GM, van Schoor NM, Pluijm SM, Deeg DJ, Lips P. Is there a U-shaped association between physical activity and falling in older persons. Osteoporos Int Campbell AJ, Reinken J, Allan BC, Martinez GS. Falls in old age: a study of frequency and related clinical factors. Age Ageing. 1981;10(4): Sherrington C, Whitney JC, Lord SR, Herbert RD, Cumming RG, Close JC. Effective exercise for the prevention of falls: a systematic review and meta-analysis. J Am Geriatr Soc. 2008;56(12): Kemper H, Ooijendonk W, Stiggelbout M. Consensus over de Nederlandse norm voor gezond bewegen. Tijdschr Soc Gezondheidsz. 2000;78: Appendix 1: Calculation of average, sum, and maximum scores for the component intensity Person activity pattern (scores for intensity) A walking outside (3.5), bowling (3.5). B walking outside (3.5), light (2.5) and heavy (4.5) household activities, bowling (3.5). C heavy household activities (4.5), swimming (5.0). Person A B C Intensity scores per activity Average scores 3.5 (medium) 3.5 (medium) 4.8 (high) Sum scores 7.0 (low) 14.0 (medium) 9.5 (low) Maximum scores 3.5 (low) 4.5 (medium) 5.0 (high) Average scores represent the average level of a component in the person s activity pattern. Sum scores represent the total level of a component in the person s activity pattern. Maximum scores represent the maximum required level of a component in the person s activity pattern. The cutoff values used for the categorization in tertiles of the component scores were the following: - intensity: average scores 3.5/3.7, sum scores 10.0/15.0, maximum scores 4.3/4.8; - strength: average scores 2.3/2.7, sum scores 6.5/10.5, maximum scores 2.5/3.5; - mechanical strain: average scores 1.4/1.6, sum scores 4.5/6.5, maximum scores 1.5/2.5; and - turning: average scores 1.4/1.6, sum scores 4.5/6.5, maximum scores 1.5/2.5.