Amencan Journal of Epidemiology Copyright C 1992 by The Johns Hopkns University School of Hygiene and Public Health All rights reserved

Vol. 136, No. 5 Printed in U.S.A.

Peripheral Neuromuscular Dysfunction and Falls in an Elderly Cohort

Gary S. Sorock1 and David M. Labiner2

In a prospective study of 169 tenants of senior citizen housing in New Jersey in 1986-1987, the relations between tests of peripheral sensory and motor functions in the lower extremities and the rate of first falls were evaluated. The mean age of the cohort was 79.8 years. Fifty-seven persons fell at least once during the follow-up period (mean, 5.6 months). After adjustment for history of stroke, heart failure, emphysema, and use of a walker or cane, rate ratios for first falls were elevated in subjects with reduced toe joint position sense (rate ratio (RR) = 2.2) and sharp-dull discrimination (RR = 2.0), but to a lesser extent for reduced ankle strength (RR = 1.5). Presence of one or more of these three deficits was defined as a peripheral neuromuscular dysfunction and was associated with first falls after adjustment for multiple covariates (RR = 2.4, 95% confidence interval 1.3-4.5). Having two or all three sensory or motor deficits increased the rate of falling 3.9 times (95% confidence interval 2.1-7.0) compared with persons without these deficits. These data suggest that impaired sensory and motor function of the lower extremities plays an important role in falls in the elderly. Am J Epidemiol 1992;136:584-91. accidental falls; aged; extremities; peripheral nerves; risk factors

Upright posture and coordinated gait require multiple sensory inputs and intact motor function. The sensory input is derived from the visual, vestibular, and proprioceptive systems. This afferent information is

Received for publication November 13, 1991, and in final form Aprt 13, 1992. Abbreviations Q, confidence interval; PND, peripheral neuromuscular dysfunction; RR, rate ratio. 1 Division of Epidemiology, Columbia University School of Public Health, New York, NY. 2 Department of Neurology, Neurological InstituteColumbia University, New York, NY. Reprint requests to Dr. Gary S. Sorock, 41 DeJmar Avenue, Framingham, MA 01701. Dr Sorock was supported by grant AR07486 from the National Institute of Arthntis and Musculoskeletal and Skin Diseases Dr Allen Hauser assisted with the neurological aspects of this study. Drs. Jennifer Kelsey, Kenneth O'Dowd, and Eleanor Shimkin made helpful suggestions on the manuscript. This paper was presented at the Twenty-second Annual Meeting of the Society for Epidemiolocpc Research, Birmingham, AL, June 14-16, 1989.

processed in the central nervous system with the resulting efferent motor and reflex responses maintaining posture. Sensory and motor functions may be compromised by disease, normal aging, or both and may account in part for the increased frequency of falling seen at older ages (1,2). Numerous studies have evaluated the relation between sensory and motor functions of the lower extremities and falls in the elderly (3-8). Among community-dwelling older adults, a history of falls during the previous year was associated with reduced position sense in the toes and ankles (3), but not with abnormal vibratory sensation measured at the ankles (3) or knees (4). A history of falls in nursing home residents was associated with reduced strength and speed of ankle dorsiflexion (5). Prospective data from studies of noninstitutionalized elderly have shown that risk of falling increased in persons with weak distal lower extremities (6, 7). Recurrent falls have been associated with 584

Neuropathy and Falls in the Elderly

difficulty in standing up from a chair, but not with reduced ankle position sense (8). This prospective study was designed in part to examine the relations between 10 tests of sensory and motor functions of the lower extremities and the rate of first falls in a community-dwelling elderly cohort. MATERIALS AND METHODS Study population

Tenants of six senior-citizen apartment buildings in New Jersey were recruited. All but one of the six buildings were subsidized housing for the elderly. Only independently living older persons were accepted as new tenants in these buildings. English-speaking, non-wheelchair-bound tenants were eligible for the study. Tenants were told that the purpose of the study was to learn about the causes of falls in the elderly. Overall, 15.4 percent (169/1,096) of the tenants of the six buildings were enrolled in the study. Participation in the six buildings ranged from 5.5 to 36.0 percent. Baseline data

After informed consent was obtained, a baseline interview and physical assessment were performed in the subjects' apartments by a single interviewer (G. S.). Interviews consisted of questions regarding the presence or absence of specific diseases (as told to subjects by a physician), current medication used, mental status questions, history of falls, physical activity, and hospitalizations in the previous year. After questions about health status, a physical assessment of neuromuscular function and gait was conducted. The neuromuscular assessment protocol was developed by the authors based on standard physical diagnostic principles (9, 10). The neuromuscular assessment consisted of nine sensory tests of the lower extremities, one test of muscle strength, and gait assessment. Subjects were seated and were asked to remove their shoes and stockings. Proprioception was assessed by slightly moving the first metatarsophalangeal joint of each great

585

toe up and down. The subject was asked to determine whether his or her toe was moving up or down without seeing it. Three or more incorrect responses out of six movements were defined as reduced proprioception at that joint. The term "reduced" refers to clinically significant reduction in sensory or motor function for this age group. Vibration sense was measured at the lateral malleolus with a tuning fork. If the vibration could not be sensed, vibratory sensation was determined to be reduced. Deep tendon reflexes at both the ankle and knee were judged to be reduced if no reflex movement was seen; no reinforcement techniques were used. The plantar response was abnormal (Babinski sign) if the great toe visibly moved upward (extension) with appropriate plantar stimulation. Sharp-dull discrimination was measured with both the sharp and dull ends of a safety pin. Each lower leg was tested six times; three or more incorrect interpretations out of six trials was defined as a reduction of this modality. Stocking-type sensory loss was measured with the dull end of the pin; if there were any reported changes in sensation from three stimuli moving upward from the foot to the lower leg, stocking changes were said to be present. Touch sensation was examined by having the subject state whether or not he or she felt a cotton swab. Each leg was tested six times; if the subject responded incorrectly three or more times, touch sensation was judged to be reduced. The subject was then asked to stand barefoot with eyes closed, heels together, and toes pointing outward for 10 seconds for the Romberg test. If the subject opened an eye, reached out to hold onto something, or moved a foot during this 10-second period, the Romberg test was judged positive. Ankle strength was assessed with the subject sitting with legs uncrossed, heels on the floor, and toes pointing upward. Both feet were pulled downward while the subject resisted. If a foot was pulled downward easily, ankle strength was reduced. For all of the above tests, reduced sensory or motor function was recorded as unilateral or bilateral. Gait was then assessed by asking subjects

586

Sorock and Labiner

to walk 12 feet (368.4 cm), turn around, and walk back. The number of steps taken was used to assess gait function. Fifteen or more steps to walk 12 feet (87th percentile) or more than six steps to turn around (86th percentile) was considered abnormal gait function. Use of a cane or walker during this measured walk was also recorded. After these physical assessments, the Folstein Mini-Mental State examination (11) was used to assess cognitive function. Next, questions regarding medication usage, including psychoactive medications, were asked. Finally, questions pertaining to history of falls in the previous year, physical activity (blocks walked outside per week on average), and hospitalization history in the previous year were asked. Follow-up data

All baseline interviews were performed between September 1986 and March 1987. The follow-up period lasted from October 1986 to June 1987. Each study subject was called monthly on the telephone and asked if he or she had fallen in the previous month. A fall was defined as unintentional contact with the floor or ground or contact with an object below the level of the hip while standing, without coming to rest on the floor. For example, falling onto a coffee table without landing on the floor would be counted as a fall. Seizure-related falls or rolling out of bed while asleep were excluded. All follow-up data were gathered by a single interviewer (G. S.). Statistical analysis

All rates were calculated as the number of persons who fell for the first time during the follow-up period divided by the number of person-months of follow-up for each group. The measure of the effect of neuromuscular and gait function on the rate of first falls was the rate ratio or incidence density ratio. Testbased interval estimation procedures were used to determine 95 percent confidence intervals for the rate ratios (12). Potential confounders were adjusted for using the Cox proportional hazards regression model (13).

RESULTS

The individual follow-up time varied from 2 to 9 months. Over the average 5.6 months of follow-up, 57 of 169 persons (34 percent) fell at least once. The rate of first falls was 57 per 758.5 person-months of follow-up, or 7.5 people fell at least once per 100 personmonths. Fourteen people fell more than once; the total number of falls was 77. The rate of falls was 77 per 946 total personmonths of follow-up, or 8.1 falls per 100 person-months. The distributions of selected demographic variables of the study cohort are presented in table 1. The mean age of the subjects was 79.8 years (standard deviation = 7.3). Table 2 presents the rates of first falls for selected variables measured at baseline and thought to be potential confounders in this study. No clear trend of increasing rates of falls was seen with increasing age. Males had a slightly higher rate of falls than did females. The average Mini-Mental State examination score was 26 (standard deviation = 3.7, range = 14-30). Subjects with a MiniMental State examination score of less than 24, adjusted for years of education, did not have an elevated rate of falls (rate ratio =

TABLE 1. Distribution of selected demographic characteristics of the study cohort, tenants of senior citizen housing, New Jersey, 1986-1987 Characteristic Age (years) 61-74 75-84 85-94 Race/ethniclty White Black Hispanic Asian Sex Mate Female Marital status Married Widowed Separated/divorced Never married Not stated

NO. (n-169) 45 75 49

26.6 44.4 29.0

151 12 4 2

89.3 7.1 2.4 1.2

35 134

20.7 79.3

20 116 20 8 5

11.8 68.6 11.8 4.7 3.0

Neuropathy and Falls in the Elderly

587

TABLE 2. Rates of first (alls (or selected baseline variables, tenants of senior citizen housing, New Jersey, 1986-1987 Variable

Age (years) 61-74 75-84 85-94 Sex Male Female Stroke No Yes Heart failure No Yes Emphysema No Yes Use of walker or cane No Yes Blocks walked per week 0-4 5-14 15-90 Benzodiazepine exposure None As needed Daily

95% confidence interval

No.*

%

Persons whofel

Personmonths

Ratet

Rate ratio

45 75 49

26.6 44.4 29.0

19 20 18

243.5 349.0 166.0

7.8 5.7 10.8

0.7 1.4

35

20.7 79.3

13



44

141.5 617.0

9.2

134

7.1

0.8

150 19

88.8 11.2

48 9

683.0 75.5

7.0 11.9

1.7

0.8-3.4

157 11

93.5 6.5

51 6

711.5 44.0

7.2 13.6

— 1.9

0.8-4.4

162 6

96.4 3.6

54 3

730.0 20.5

7.4 14.6

— 2.0

0.6-6.2

139 27

83.7 16.3

41

643.5 95.5

6.4 15.7

— 2.5

1.4-4.4

59 51 58

35.1 30.4 34.5

22

245.0 214.5 298.5

9.0 7.9 5.7

— 0.9 0.6

0.5-1.7 0.3-1.2

125 21 23

74.0 12.4 13.6

39 7

590.5 80.5 87.5

6.6 8.7 12.6

— 1.3 1.9

0.6-2.9 1.0-3.7

15

17

17

11

-t

0.4-1.4

0.7-2.6

0.4-1.4



* Total subjects does not always add to 169 because of Incomplete Information for some subjects, t Rates per 100 person-months. X Reference category.

1.1, 95 percent confidence interval 0.5-2.3). Stronger associations were found for stroke, heart failure, emphysema, daily benzodiazepine use, and use of a cane or a walker. A trend toward reduced rates of falls with increased distances walked outside was found. The rate ratios for first falls were below unity for history of diabetes (rate ratio (RR) = 0.9) and for use of bifocals (RR = 0.8). No association (RR = 1.0) was seen for having cataracts at baseline or for any hospitalization in the previous year. A weak positive association was seen for heart attack (RR = 1.2), previous fracture (RR = 1.2), falling in the previous year (RR = 1.3), and history of arthritis (RR = 1.5). All 95 percent confidence intervals for these estimates included the null value.

The prevalence of reduced sensory, motor, and gait functions is presented in table 3. Numbers of subjects with reduced unilateral and reduced bilateral function are also noted. Vibration sense at the ankles was most often reduced, and the Babinski sign was least often found. Vibration sense was reduced most often bilaterally (79 percent, 77/97); Babinski sign was found most often unilaterally (95 percent, 21/22). Table 4 shows the relations between neuromuscular and gait function test results and the rates of first falls. Separate analyses for unilateral compared with bilateral reduced function showed similar results and were combined. Weak associations were found between first fall and reduced vibratory sense (RR = 1.4), touch sense (RR = 1.3),

588

Sorock and Labiner

TABLE 3. Prevalence of reduced sensory, motor, and gatt test results, tenants of senior citizen housing, New Jersey, 1986-1987 Test

Vibration sense Touch sense AchiQes reflex Patellar reflex Sharp-dull discrimination Stocking sense changes Romberg Position sense Ankle strength Steps to walk 12 feet Steps to turn around Babinski sign

No. with unflateral/ bflateral reduced test results

% reduced

No. tested

No. reduced*

167 159 166 166

97 86 84 71

20/77 44/42 26/58 26/45

58.1 54.1 50.6 42.8

158

66

40/26

41.8

157 163 166 164 167 167 160

60

35/25

38.2 25.2 24.7 23.8 16.8 14.4 13.8

41 41 39 28 24 22

25/16 22/17

21/1

* Defined as unilateral or btateral reduced function.

TABLE 4. Reduced sensory, motor, and gait test results and rates of first falls, tenants of senior citizen housing, New Jersey, 1986-1987 Test

Position sense Normal Reduced Sharp-dull discrimination Normal Reduced Ankle strength Normal Reduced Steps to turn around Normal Reduced

Rate ratio

95% confidence interval

5.9 15.0

2.6

1.5-4.3

463.5 242.0

5.4 11.6

2.1

1.3-3.6

589.5 148.0

6.3 12.2

1.9

1.1-3.4

655.0 88.5

12.4

1.8

0.9-3.4

Persons who fen

Personmonths

Rate'

36 20

614.0 133.0

25 28 37 18 46

11

7.0

' Rate per 100 person-months of follow-up.

stocking sense changes (RR = 1.6), reduced ankle (RR = 1.1) or knee (RR = 1.1) reflexes, 15 or more steps to walk 12 feet (368.4 cm) (RR = 1.0), positive Romberg test (RR = 1.2), and Babinski sign (RR = 1.3). Three results showed substantially elevated rate ratios (RR of close to 2.0): reduced toe joint position sense, reduced sharp-dull discrimination, and reduced ankle strength. Adjustment of these rate ratios for stroke, emphysema, heart failure, or use of cane or walker changed the rate ratio for reduced ankle strength by 21 percent (table 5).

Toe position sense, sharp-dull discrimination, and ankle strength were intercorrelated as determined in a cross-sectional analysis. Odds ratios from this analysis ranged from 1.7 for reduced ankle strength and sharp-dull discrimination to 5.0 for reduced toe position sense and sharp-dull discrimination. Combinations of these three deficits may be indicative of peripheral neuromuscular dysfunction (PND) associated with an elevated rate of falls. Subjects with at least one of these three sensory or motor deficits were designated as having PND.

Neuropathy and Falls in the Elderly

589

TABLE 5. Selected sensory and motor test results: unadjusted and adjusted rate ratios for first falls using Cox proportional hazards regression, tenants of senior citizen housing, New Jersey, 1986-1987 Test result*

Unadjusted rate ratio

95% conflctencs interval

Adjusted rateratiot

95% confidence Interval

Position sense Sharp-dun discrimination Ankle strength

2.5 2.0 1.9

1.4-4.3 1.2-3.5 1.1-3.3

2.2 2.0 1.5

1.2-3.9 1.2-3.5 0.8-2.9

• Reduced versus normal test result; unilateral and bilateral reduction combined. f Adjusted In separate models for stroke, heart faBure, emphysema, and use of cane or walker in own apartment.

TABLE 6. Combination of selected sensory and motor function deficits in position sense, sharp-dull discrimination, and ankle strength, and rate of first falls, tenants of senior citizen housing, New .Jersey, 1986-1987 No.

Persons falling

Personmonths

Rate

72 59 38

17 19 21

400.5 230.5 127.5

4.2 8.2 16.5

169

57

758.5

7.5

Combination of deficits

None of three One of three (possible PND*) 2 or 3 (probable PND*) Total

Rate ratio

95% confidence Interval

1.9 3.9

1.1-3.7 2.1-7.0

' PND, peripheral neuromuscular dysfunction.

Ninety-seven (57.4 percent) of 169 study subjects had PND (table 6). Subjects were further classified as follows: having one of these three deficits was defined as possible PND, while having two or three deficits was defined as probable PND. Compared with persons without PND, persons with possible PND nearly doubled the rate of first falls; probable PND nearly quadrupled the rate (RR = 3.9). The rate ratio for any degree of PND adjusted for stroke; emphysema; heart failure; use of a cane or walker, use of benzodiazepine sedatives; blocks walked per week; and history of arthritis, heart attack, fracture, and falling in the previous year was 2.4 (95 percent CI 1.3-4.5). Cross-sectional analysis of baseline data indicated that a history of at least one fall in the previous year (44 percent of the subjects) and PND were associated (odds ratio = 1.9, 95 percent CI 1.0-3.6). There was a modest interaction effect between PND and any use of benzodiazepine sedatives in stratified analysis. The crude rate ratio for PND and first fall was 2.6 (95 percent CI 1.5-4.5). For nonusers of benzodiazepine sedatives, the rate ratio for PND was 2.1 (95 percent CI 1.1-4.1); for benzo-

diazepine users, the rate ratio for PND was 4.6 (95 percent CI 1.7-12.5). DISCUSSION

The association between reduced proprioception in the toes and falls was reported previously (3) in persons aged 75-84 years who reported falling in the previous year. However, reduced position sense measured at the ankles was not associated with recurrent falls in another prospective study (8). This difference in findings may be due to the more distal location of measurement in this study (toe vs. ankle). Persons with reduced toe proprioception (and with PND) have difficulty knowing where their feet are positioned relative to the floor and thus may have greater postural instability (14) and a higher rate of falls. In our study, reduced sharp-dull discrimination was associated with the rate of first falls. Perhaps reduced sensory information from the soles of the feet (more distal than where the testing was done) makes detection of irregular floor or ground surfaces difficult. Alternatively, reduced sharp-dull discrimination may be associated with first falls be-

590

Sorock and Labiner

cause position sense and sharp-dull discrimination are intercorrelated, or the association may be due to chance. This finding suggests that these modalities should not be considered separately, but rather as part of a clinical picture of PND. Reduced ankle strength was also marginally associated with first falls in this cohort. The association was explained in part by stroke history and use of a cane or walker, with the latter being a marker of unstable gait. Lower leg weakness has been associated with falls in prospective studies (6-8) of community-dwelling elderly and among nursing home residents (5). Ankle weakness makes it difficult for persons to respond to alterations in their center of gravity. Furthermore, gait can be altered by motor deficits, making tripping more likely because of foot dragging or shuffling. The measurements of sensory and motor functions made in this study were qualitative in nature. However, these tests are routinely performed by physicians to assess neurological and motor function in the lower extremities and are clinically important. Quantitative evidence (15) of deficits in sensory nerve conduction in those who fall compared with those who do not supports our qualitative findings. Moreover, quantitative tests indicate that reduced proprioception (of the large toe) contributes more to postural instability than does either reduced vision or vestibular sense (14). The low participation rate in this study (15.4 percent) may have introduced a selection bias that influenced the results (16). Selection bias is most likely to affect point estimates when selection fractions differ according to specific combinations of exposure and disease (17). Possibly, the selection fraction for persons with both PND and a first fall were higher than the other three combinations of exposure and fall outcomes and thus distorted the rate ratios in this study. However, this would mean that a future event influenced the selection process as well as knowledge of PND status, which potential study subjects were unlikely to know at the time of recruitment. Selection bias due to loss to follow-up was not likely, two (1.2

percent) of 169 subjects moved and were lost to follow-up. Information bias was an unlikely explanation for the association between PND and falls. At baseline, sensory and motor functions were assessed before history of falls, and follow-up interviews were conducted without knowledge of sensory and motor function test results. A strength of this study was its prospective design. History of falls is often an inadequate assessment of risk of falls, as people tend to forget falls which occurred in the previous year (18); thus, prospective designs with frequent follow-up are probably the best method of studying risk factors for falls in the community-dwelling elderly. PND probably reduces gait stability by reducing sensory input from the lower limbs while walking. Benzodiazepine sedatives may further increase the rate of first falls in subjects with PND. Benzodiazepines bind selectively to neurons that affect motor coordination (19) and thus reduce the coordinated protective motor responses to loss of balance (20, 21). This possible interaction should be examined in other studies of falls in the elderly. Falls in the elderly are an example of a recurrent health problem for which the ideal outcome may be the rate of total falls over the follow-up period (22) rather than the rate of first falls as in this study. This is because the probability of fracture increases with the number of falls. However, in this study, perhaps because of its small size or short follow-up period, the unadjusted rate ratio for PND and total falls (RR = 2.2; 95 percent CI 1.4-3.6) was similar to the unadjusted rate ratio for PND and first falls (RR = 2.6; 95 percent CI 1.5-4.5). This study suggests that impaired neuromuscular function of the lower legs is related to falls in the elderly. If these results are confirmed by other studies, health providers may be able to identify older persons at risk for falls by using three tests of neuromuscular function of the lower extremities. It is less certain, however, whether interventions to improve sensation and muscle strength of the lower extremities can reduce the rate of falls in the elderly.

Neuropathy and Falls in the Elderly

REFERENCES 1. Wolfson LI, Whipple R, Ameran P, et al. Gait and balance in the elderly. Two functional capacities that link sensory and motor abilities to falls. Clin GeriatrMed 1985; 1:649-59. 2. Stelmach GE, Worringham CJ. Sensorimotor deficits related to postural stability: implications for falling in the elderly. Clin Geriatr Med 1985; 1:679-94. 3. Brocklehurst JC, Robertson D, James-Groom P. Clinical correlates of sway in old age sensory modalities. Age Ageing 1982; 11:1 -10. 4. MacLennan WJ, Timothy Jl, Hall MRP. Vibration sense, proprioception and ankle reflexes in old age. J Clin ExpGerontol 1980;2:159-71. 5. Whipple RH, Wolfson LI, Amerman PM. The relationship of knee and ankle weakness to falls in nursing home residents: an isokinetic study. J Am Geriatr Soc 1987;35:13-20. 6. Campbell AJ, Borrie MJ, Spears GF. Risk factors for falls in a community-based prospective study of people 70 years and older. J Gerontol: Med Sci 1989;44:M 112-17. 7. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988;319:1701-7. 8. Nevitt MC, Cummings SR, Kidd S, et al. Risk factors for recurrent nonsyncopal falls. A prospective study. JAMA 1989,261:2663-8. 9. DeMyer W. Technique of the neurologic examination: a programmed text. 3rd ed. New York, NY: McGraw-Hill Book Co., 1980. 10. Bates B. A guide to physical examination. Philadelphia, PA: J. B. Lippincott Co., 1983. 11. Folstein MF, Folstein SE, McHugh PR. Minimental state a practical method for grading the cognitive state of patients for the clinician. J Psy-

591

chiatrRes 1975; 12:189-98. 12. Kleinbaum DG, Kupper LL, Morgenstem H. Epidemiologic research: principles and quantitative methods. Belmont CA: Lifetime Learning Publications, 1982:301-2. 13. Cox DR. Regression models and life tables. J R StatSoc(B) 1972;34:187-220. 14. Lord SR, Clark RD, Webster IW. Postural stability and associated physiological factors in a population of aged persons. J Gerontol: Med Sci 1991;46: M69-76. 15. Pack DR, Wolfson LI, Amerman P, et al. Peripheral nerve abnormalities and falling in the elderly. (Abstract). Neurology 1985;35:79. 16. Sorock G. Re: "Issues in carrying out epidemiologic research in the elderly." (Letter). Am J Epidemiol 1991;133:317-18. 17. Kelsey JL, Thompson WD, Evans AS. Methods in observational epidemiology. New York, NY: Oxford University Press, 1986:96-100. 18. Cummings SR, Nevitt MC, Kidd S. Forgetting falls. The limited accuracy of recall of falls in the elderly. J Am Geriatr Soc 1988;36:613-16. 19. Haefely W. Tranquilizers. In: Grahme-Smith DG, Cowen PJ, eds. Preclinical psychopharmacology 2, part 1: preclinical psychopharmacology. Amsterdam, The Netherlands: Elsevier, 1985:92-105. 20. Cummings SR, Nevitt MC. A hypothesis: the causes of hip fracture. J Gerontol: Med Sci 1989; 44:M107-ll. 21. Sorock GS, Shimkin EE. Benzodiazepine sedatives and the risk of falling in a community-dwelling elderly cohort. Arch Intern Med 1988;148: 2441-4. 22. Cumming RG, Kelsey JL, Nevitt MC. Methodologic issues in the study of frequent and recurrent health problems: falls in the elderly. Ann Epidemiol 1990; 1:49-56.

Peripheral neuromuscular dysfunction and falls in an elderly cohort.

In a prospective study of 169 tenants of senior citizen housing in New Jersey in 1986-1987, the relations between tests of peripheral sensory and moto...
518KB Sizes 0 Downloads 0 Views