| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Epidemiology, Demography and Biometry Program, National Institute on Aging, National Institutes of Health, Bethesda, Maryland
b The Diabetes Institutes and Center for Endocrine and Metabolic Disorders, Eastern Virginia Medical School, Norfolk
c Sytel, Inc., Bethesda, Maryland.
d Departments of Internal Medicine and Epidemiology, Johns Hopkins University, Baltimore, Maryland
Decision Editor: John E. Morley, MB, BCh
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. A total of 894 women aged 65+ years participating in the Women's Health and Aging Study received a baseline home interview and clinical examination during which PND was evaluated by the Vibratron II. Age and diabetes were examined in relation to the level of PND (normal, mild, moderate, or severe). Height, alcohol consumption, smoking, report of neurologic symptoms, and diabetes duration were examined as potential confounders.
Results. Eighteen percent of the sample reported diabetes, 42% had normal nerve function, and 23.9%, 14.5%, and 19.5% had mild, moderate, and severe PND, respectively. Women aged 85+ years had 6.5, 7.5, and 13.3 times the odds of mild, moderate, and severe PND relative to women aged 65–74 years, adjusted for diabetes and height. Women who reported diabetes had 1.8, 2.4, and 1.6 times the risk of mild, moderate, and severe PND relative to those who did not, adjusted for age and height. No interaction between age and diabetes was observed.
Conclusions. Age is strongly associated with decrements in large-fiber peripheral nerve function in disabled women aged 65+ years, with greatly accelerated risk among those aged 85+ years. Despite the overwhelmingly strong effects of advancing age on PND in this cohort, diabetes remains a significant correlate of PND. Future studies may determine whether prevention or control of diabetes is effective in reducing the occurrence of PND in old age and whether a reduction in PND will translate into reduced disability in this age group.
RECENT estimates indicate that 29 million Americans have diabetes mellitus, including more than 37% of Americans 
60 years of age (1). Diabetes, in turn, is associated with vascular and neuropathic complications, of which peripheral sensorimotor polyneuropathy may be the most common (2). Population estimates of the prevalence of neuropathy vary widely due in part to difficulty in diagnosis and also to the variety of tools used to evaluate neuropathy in epidemiologic studies (2)(3)(4)(5)(6)(7)(8)(9). Although there is considerable interpatient variability in the signs and symptoms of neuropathy, it is often characterized by loss of distal strength and vibration sensation, proprioceptive defect, and muscle atrophy, all features of damage to large, myelinated nerve fibers. Neuropathy with large-fiber features occurs with greater frequency in individuals with diabetes compared with those without, and large-fiber peripheral nerve dysfunction (PND) increases with advancing age (3)(6)(7)(10)(11)(12). Because diabetes occurs with increasing frequency as age increases and because PND increases with increasing duration of diabetes (5)(7)(11), it is challenging to isolate the contributions of age and diabetes when studying PND in older populations.
Data on PND are available from several epidemiologic studies that include middle-aged and older adults (6)(7)(8)(13)(14)(15), but many of these included only people with diabetes (6)(7), classified PND by the report of symptoms alone (13)(14)(15), or did not include the "oldest-old" individuals aged 85+ years (8)(14). There are few quantitative data on the individual and combined effects of age and diabetes on PND in older women with and without diabetes, or among the oldest-old, and virtually no data on these relationships among older adults with existing disability and a heavy burden of comorbidity.
The purpose of this report was to test several hypotheses related to large-fiber peripheral nerve function in a cohort of older women with existing disability. First, we hypothesized that significant age-associated decrements in peripheral nerve function would be observed in this sample of elderly, disabled women. We further hypothesized that the effects of diabetes mellitus on peripheral nerve function would be observed independently of age and that advancing age and diabetes would have synergistic effects on PND. We show that large-fiber PND occurs with advancing age and accelerates rapidly after age 85 years and that diabetes contributes to PND even in very old age.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
65 years. Participants were recruited from an age-stratified random sample of community-dwelling Medicare beneficiaries residing in 12 contiguous zip codes in the Baltimore, Maryland area. Eligibility was determined from an in-person screening interview conducted at participants' homes. All participants were self-respondents; no proxies were used. The WHAS received approval from the Institutional Review Board of the Johns Hopkins University, and participants provided informed consent.
Classification of Diabetes Status
WHAS participants received an extensive home interview that included questions about physician-diagnosed medical conditions and about symptoms characteristic of a number of conditions. Each participant was asked "Has a doctor ever told you that you had diabetes?" Women responding "yes" to this question were described as having "reported" diabetes. Participants with reported diabetes were also asked, "How old were you when you were first told that you had diabetes?" Duration of reported diabetes was calculated from the year the participant reported being diagnosed with diabetes and her age at the baseline exam. Age at diabetes diagnosis was used as a selection criterion: If the age at diagnosis was 
30 years, the subject was excluded due to the possibility that the diabetes she reported was type 1 rather than type 2.
Five questions related to peripheral neuropathy were asked, the first being a screening question: "Is your sensation or sense of feeling normal or abnormal?" If the subject reported that her sense of feeling was abnormal, four additional questions were asked: "Is your abnormal sensation found in your legs or feet?"; "Have you ever burned yourself without feeling pain?"; "Do you have a prickly-asleep-numbness feeling of the feet, like when your hand goes to sleep from lying on it?"; and "In your feet, do you have dead-asleep-numbness, like novocaine, without prickling?"
Examination and Evaluation of Peripheral Nerve Function
On a separate home visit, large-fiber peripheral nerve function was evaluated by measuring vibration perception threshold (VPT) with the Vibratron II (Physitemp Instruments, Inc, Clifton, NJ). The Vibratron II measures the sensitivity of the plantar aspect of the great toe in detecting small, vibratory stimuli, thereby providing quantitative information on large-fiber peripheral nerve function. WHAS employed a standardized protocol involving a two-alternative, forced-choice procedure in which the participant indicated which of two periods of supposed stimulation was accompanied by an actual vibration. The intensity of the stimulus was reduced by approximately 10% at each trial until the participant could no longer detect the vibration. This method was derived from a standardized protocol developed by Arezzo for Physitemp Instruments (18) and has been described in detail and validated (16)(19). VPT data are presented in vibration units, the actual reading displayed on the Vibratron II device. These units measure the amplitude of the stimulus and are related to "true" amplitude (measured in microns) by the following formula: A = KX2, where A = the peak amplitude in microns, K = 0.5, and X = vibration units (20). Higher values indicate that a stronger stimulus was needed to elicit a response, reflecting worse PND.
Methodological work with the Vibratron II has provided normal, age-specific values for peripheral nerve function. For adults >65 years of age, less than 3.43 vibration units is defined as the range of normal function; 3.43 to <4.87 units indicates evidence of mild dysfunction; 4.87 to <6.31 units is evidence of moderate dysfunction, and 6.31 vibration units indicates severe dysfunction (18). These values are specific to the forced-choice protocol of the Vibratron II.
Of the 1002 women in the WHAS, 104 were missing VPT data. Of these, 47 initiated but did not complete the protocol, 7 were amputees or could not perform the test because of other physical problems, 5 did not understand the instructions, 43 did not comply with the instructions, and 2 refused. Following exclusion of subjects with missing VPT, 4 additional women were excluded because they reported being diagnosed with diabetes at age 30 years. The level of PND could therefore be determined for 894 women (89.2% of the WHAS cohort).
Potential Confounders
Height is related to PND with taller individuals having worse measures (7)(11)(12). Standing height and knee height were both measured in WHAS according to standardized protocols (16) and were investigated separately in multivariate analyses. Because old age is associated with degenerative changes in stature, we used knee height to estimate corrected standing height with a validated prediction equation for women aged 60–90 years (21). Subtraction of measured standing height from corrected standing height (
standing height) yielded an estimate of change in stature. This variable was examined in regression analysis as an index of vertebral compression and kyphosis, both of which may be associated with PND.
Alcohol consumption has been associated with PND in older adults (13). Usual alcohol consumption was assessed during the interview with the following question: "Do you usually drink alcoholic beverages, including beer, wine, sherry, or liquor, at least once every week?" Alcohol was analyzed by contrasting women who reported any drinking to those who reported none. Current and former smokers were contrasted to never smokers.
Statistical Analysis
PND was analyzed as an ordinal outcome in descriptive and multivariate analyses, with the normal nerve function group as the reference. Women with mild, moderate, or severe PND were analyzed individually as the case groups. In descriptive analyses, the chi square test was used to study the distribution of PND (normal, mild, moderate, and severe) in relation to age and diabetes (reported diabetes vs no diabetes; age: 65–74, 75–84, and 85 years). PND categories were also examined in relation to potentially confounding variables.
In multivariate regression analyses using the CATMOD (Categorical Model) procedure in SAS (22), effects of age and diabetes were examined in relation to the odds of having a specific level of PND (mild, moderate, or severe) relative to normal function. The combined effects (interaction) of older age and diabetes were examined to determine if the presence of both factors resulted in a synergy that increased the odds of PND beyond the effects of each factor alone. Potentially confounding variables were entered into the models to determine if the effects of age and diabetes changed following adjustment for these factors. Odds ratios (OR) ratios and 95% confidence intervals (CI) show cross-sectional associations of having a specific level of PND relative to the reference group, adjusted for potential confounders. Associations were considered statistically significant at the p < .05 level, corresponding to 95% CI excluding unity.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
standing height was -0.20 cm, indicating a mean reduction in stature. Mean vibration threshold in this sample was 4.28 ± 2.23 units. The distributions of VPT and categories of PND are shown in Fig. 1. The distribution of VPT was skewed to the right. Forty-two percent of the sample had normal peripheral nerve function, and 23.9%, 14.5%, and 19.5% had mild, moderate, and severe PND, respectively.
|
85 years had normal nerve function; women aged 75–84 years were intermediate, with 43% having normal function. Conversely, only 13% of women aged 65–74 years had severe PND compared with 31% of women aged 85 years (p < .001). Reported diabetes status and level of PND are presented in Fig. 3. While 66% of women with reported diabetes had PND, 56% of women without reported diabetes had PND. The prevalence of severe PND was about 20% among women with and without reported diabetes. There was no statistically significant univariate association between diabetes category and level of PND.
|
|
85 years. The odds of severe PND in these two age groups increased dramatically to 2.6 and 13.3 compared with women aged 65–74 years. For mild and moderate PND, odds among women aged 85 years were more than three times that of women aged 75–84 years, and for severe PND, odds among women aged 85 years were more than five times that of women aged 75–84 years. These differences suggested an accelerated, nonlinear effect of age on the occurrence and severity of PND among women aged 85 years.
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
85 years. The odds of being at each level of PND among women aged 85 years were at least threefold greater than among women aged 75–84 years.
The accelerated effects of age on occurrence and severity of PND in women aged 85 years may be associated with functional impairments in the lower extremity, potentially placing this age group at high risk of neuropathy-related disability. Supporting an association between PND and decrements in lower extremity physical abilities are several small studies of middle-aged and older adults that showed relationships between PND and gait abnormalities, loss of strength, decrements in balance, and history of falls (24)(25)(26)(27)(28). These factors are important for maintenance of independence in old age and highlight the potential contribution of PND to disability in elderly individuals. Evaluation of large-fiber nerve function may therefore help to identify older individuals at increased risk of PND-associated disability, a critical first step in intervening on this potential disability risk factor. The strong association between women aged 85 years and risk of PND suggests that interventions to reduce PND-associated disability may have the greatest relative benefits among the oldest-old, a group expected to increase by more than 100% between 1993 and 2020 (23).
An important aspect of this study is the availability of large numbers of both diabetic and nondiabetic subjects. Previous studies of PND in older adults have focused solely on diabetic individuals (6)(7), preventing examination of the independent effects of diabetes and age on PND in the elderly. The availability of a nondiabetic comparison group is critical when studying peripheral nerve function in older adults because diabetes is highly prevalent in old age and because its contribution to PND relative to that of advancing age has not been examined explicitly. Consistent with a recent study in a similar age group (29), we showed that older women who report diabetes had a greater risk of PND compared with nondiabetic women, and our associations were independent of age and height. Compared with nondiabetic women, those with reported diabetes were 1.7 times more likely to have mild PND and 2.4 times more likely to have moderate dysfunction. The association between reported diabetes and severe PND was not statistically significant, due in part to the overwhelmingly strong effects of age in this PND category and in part to the small number of women aged 85 years reporting diabetes.
It is important to emphasize that, unlike advancing age, diabetes is a modifiable PND risk factor. Reducing the occurrence of diabetes or reducing the prevalence of uncontrolled hyperglycemia among diabetic individuals may reduce the risk and/or progression of PND that is attributable to diabetes among older adults. However, it has been proposed that diabetes represents an accelerated form of aging, with both phenomena sharing one or a number of similar metabolic abnormalities (30). Supporting this idea are findings from a study of diabetes and age in relation to peripheral blood flow, which showed that diabetes has an effect on precapillaries that resembles premature aging (31). Additional research is needed to clarify pathways potentially linking diabetes and chronological aging.
Our results did not indicate an interaction between age and diabetes, suggesting that reduced occurrence of diabetes in older adults would result in similar reductions in diabetes-associated PND across age. The oldest-old would therefore benefit as much as the young-old from a reduction in diabetes.
Long-term exposure to hyperglycemia has long been hypothesized to be causally associated with PND, but the mechanisms by which this occurs have not been fully elucidated. In this study, duration of diabetes (among the 158 women for whom these data were available) was not associated with worse PND, a finding observed in several previous studies (5)(7)(11). Lack of statistical power may have contributed to the absent association. An alternative explanation may be that the effects of age simply overpower the effect of diabetes duration in this elderly cohort.
Ascertainment of diabetes in the WHAS was not ideal. Because neither fasting glucose nor postchallenge glucose was measured, undiagnosed diabetes could not be categorized according to the criteria recommended by the American Diabetes Association or the World Health Organization (32)(33). This is important because the prevalence of undiagnosed diabetes is known to be approximately equal to that of reported diabetes (34). Absence of data on fasting and postchallenge glucose therefore raised obvious concerns about the misclassification of diabetes. If a 50% underestimation of diabetes is assumed in this study, there might have been as many as 165 women misclassified as nondiabetic (a number equal to those who reported diabetes), and this misclassification may have underestimated the magnitude of the diabetes-PND associations observed in this study. Misclassification of diabetes could also have contributed to the absence of interaction between age and diabetes.
One hypothesized mechanism through which hyperglycemia may contribute to or enhance the progression of PND is through the pathologic effects of advanced glycation endproducts (AGEs) on nerve tissues (30)(35). In addition to examination of the potential role of AGEs in development and progression of PND in old age, standardized methods to examine other mechanisms, such as microvascular insufficiency, growth-factor deficiencies, and immune activity should also be developed for use in population-based studies (36).
The WHAS baseline questionnaire contained questions related to symptoms of neuropathy, including items related to hot/cold sensation and tingling. These symptoms are consistent with small- rather than large-fiber neuropathy. However, the Vibratron II evaluates large-fiber PND; WHAS did not include quantitative sensory testing for small-fiber dysfunction, such as thermal discrimination. It was therefore not surprising that the neurologic symptoms reported by WHAS subjects were not associated with PND as measured by the Vibratron II. It should be noted that screening for small-fiber dysfunction would have likely resulted in identification of more cases of nerve dysfunction, and/or more neuropathic modalities (37). This is a critical distinction because dysfunction of large and small nerve fibers does not always occur simultaneously. Further, symptoms and objective tests of nerve function are central to neuropathy classification according to criteria proposed by the San Antonio Conference (38).
Alcohol consumption and smoking were not related to PND in this study. These findings are in contrast to a previous study of younger subjects (39) and also to a study of older individuals (13) and may be due to the strong effects of age among the oldest-old in the WHAS. Alternatively, a selection bias may have been present in this study: Individuals with behavioral characteristics such as smoking and alcohol consumption might have been less likely to have survived until age 65 years, the minimum age at which women were eligible for inclusion in the study. Supporting this possibility is the fact that at the time of the home interview, only 16% of the sample reported usual alcohol consumption, and only 12% were current smokers.
Kyphosis and vertebral compression present methodologic difficulties in studies of older adults in which accurate measurement of height is critical for interpreting quantitative data. We addressed this problem by calculating corrected height with a validated algorithm. While corrected height predicted worse PND in most analyses, standing height was not associated with PND, suggesting that large-fiber PND is length dependent. It is interesting to note that advancing age is associated with several potentially related phenomena: kyphosis, decreased bone mineral density, and dysfunction of peripheral nerves, the latter in both the presence and absence of diabetes. Kyphosis is characterized by gradual, usually painless vertebral deformities, accompanied by bone density in the lower range of age- and sex-adjusted norms. With age, diffuse loss of bone density is observed at numerous other sites, including the lower extremity. Marked bone loss predisposing to fracture also occurs in Charcot neuroarthropy, a chronic, progressive arthropathy related to sensory neuropathy. While mechanisms leading to development of Charcot foot are poorly understood, abnormalities of blood flow and increased bone resorption have been proposed (40)(41). The relationship between general, age-related loss of bone density and age-related decrements in peripheral nerve function has not been explored. These phenomena may be linked by mechanisms similar to those hypothesized to cause the focal bone loss observed in the Charcot foot or by pathways yet to be identified.
This study has several limitations, most of which are related to the design and selection of the cohort. First, the WHAS is limited to women, preventing examination of gender effects on associations between both diabetes and age on PND. While previous studies have suggested that PND is more common among men than among women (9), this association has not been consistent among persons with and without diabetes (11) and has been absent when height is accounted for in the analysis (5). From a public health standpoint, understanding modifiable risk factors related to PND is critical for older women because women have a longer life expectancy than men and, as we have shown, the likelihood of having PND increases dramatically at advanced old age. Our findings associated with the dramatic increase in PND among women aged 85 years is particularly important in light of current demographic trends in the United States: In the year 2050, it is estimated that there will be 4.7 million more women aged 85 years than men of the same age (23).
The WHAS cohort was selected on the basis of the presence of disability at the baseline examination, making it somewhat unrepresentative of the general population of women aged 65 years. However, the WHAS provides a unique opportunity to study peripheral nerve function in a large cohort of individuals whose existing disabilities could have masked the effects of diabetes and age on PND. Consistent with our hypotheses, however, we found that both age and diabetes independently affect PND in this cohort, indicating measurable effects of diabetes, even in the presence of considerable and comorbidity.
In summary, we have shown that both age and diabetes are independently associated with significant decrements in large-fiber peripheral nerve function in older women with existing disability and that the effects of age on PND accelerate rapidly after the age of 85 years. Despite the clear effects of age on PND and a growing body of evidence that PND may be associated with disability in old age, the effects of PND on functionally relevant outcomes in older adults have received surprisingly limited attention. Future epidemiologic studies of peripheral nerve function in old age should focus on clarifying the role of nerve dysfunction on functionally relevant outcomes.
|
Acknowledgments |
|---|
Received March 21, 2000
Accepted April 3, 2000
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. S. Strotmeyer, N. de Rekeneire, A. V. Schwartz, K. A. Faulkner, H. E. Resnick, B. H. Goodpaster, R. I. Shorr, A. I. Vinik, T. B. Harris, A. B. Newman, et al. The Relationship of Reduced Peripheral Nerve Function and Diabetes With Physical Performance in Older White and Black Adults: The Health, Aging, and Body Composition (Health ABC) Study Diabetes Care, September 1, 2008; 31(9): 1767 - 1772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Volpato, S. G. Leveille, C. Blaum, L. P. Fried, and J. M. Guralnik Risk Factors for Falls in Older Disabled Women With Diabetes: The Women's Health and Aging Study J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2005; 60(12): 1539 - 1545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Wray, M. B. Ofstedal, K. M. Langa, and C. S. Blaum The Effect of Diabetes on Disability in Middle-Aged and Older Adults J. Gerontol. A Biol. Sci. Med. Sci., September 1, 2005; 60(9): 1206 - 1211. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J. Brandon, D. A. Gaasch, L. W. Boyette, and A. M. Lloyd Effects of Long-Term Resistive Training on Mobility and Strength in Older Adults With Diabetes J. Gerontol. A Biol. Sci. Med. Sci., August 1, 2003; 58(8): M740 - 745. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Morley, H. M. Perry III, and D. K. Miller Editorial: Something About Frailty J. Gerontol. A Biol. Sci. Med. Sci., November 1, 2002; 57(11): M698 - 704. [Full Text] [PDF]
|
|
|
|
|
|
J. E. Morley Editorial: A Fall Is a Major Event in the Life of an Older Person J. Gerontol. A Biol. Sci. Med. Sci., August 1, 2002; 57(8): M492 - 495. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
