Reduced Peripheral Nerve Function Is Related to Lower Hip BMD and Calcaneal QUS in Older White and Black Adults: The Health, Aging, and Body Composition Study

Authors

  • Elsa S Strotmeyer PhD, MPH,

    Corresponding author
    1. Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    • Department of Epidemiology, University of Pittsburgh, 130 North Bellefield Avenue, Room 515, Pittsburgh, PA 15213, USA
    Search for more papers by this author
  • Jane A Cauley,

    1. Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • Ann V Schwartz,

    1. Department of Epidemiology and Biostatistics, The University of California, San Francisco, California, USA
    Search for more papers by this author
  • Nathalie de Rekeneire,

    1. Centers for Disease Control, Atlanta, Georgia, USA
    Search for more papers by this author
  • Helaine E Resnick,

    1. Department of Epidemiology and Statistics, MedStar Research Institute, Hyattsville, Maryland, USA
    Search for more papers by this author
  • Joseph M Zmuda,

    1. Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • Ronald I Shorr,

    1. Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, Tennessee, USA
    Search for more papers by this author
  • Frances A Tylavsky,

    1. Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, Tennessee, USA
    Search for more papers by this author
  • Aaron I Vinik,

    1. Eastern Virginia Medical School, Norfolk, Virginia, USA
    Search for more papers by this author
    • Dr Vinik serves as a consultant for Neurometrix, R.W. Johnson Pharmaceutical Research Institute, Takeda, TEVA Pharmaceutical Industries, Astrazeneca, Sanofi-Aventis, DLife, Health ABC, Mitsubishi Pharma America, Inc., and Paramount Biosciences. All other authors state that they have no conflicts of interest.

  • Tamara B Harris,

    1. Laboratory of Epidemiology, Demography and Biometry, Intramural Research Program, National Institute on Aging, Bethesda, Maryland, USA
    Search for more papers by this author
  • Anne B Newman,

    1. Department of Epidemiology and Division of Geriatric Medicine, University of Pittsburgh Graduate School of Public Health and School of Medicine, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • for the Health ABC Study


Abstract

Bone tissue is innervated, and peripheral nerve function may impact BMD. Older black and white men and women (N = 2200) in the Health, Aging, and Body Composition Study with worse sensory and motor peripheral nerve function had lower hip BMD and calcaneal BUA independent of lean mass, strength, physical ability, and diabetes. Poor peripheral nerve function may directly affect bone.

Introduction: Bone tissue is innervated, yet little is known about the impact of nerve function on BMD. Poor peripheral nerve function may contribute to lower BMD and higher fracture risk, particularly in those with diabetic neuropathy.

Materials and Methods: The Health, Aging, and Body Composition (Health ABC) Study included annual exams in white and black men and women 70–79 years of age recruited from Pittsburgh and Memphis. Nerve function in legs/feet was assessed by 1.4- and 10-g monofilament detection, vibration threshold, and peroneal motor nerve conduction velocity (NCV) and amplitude (CMAP). Total hip BMD, heel broadband ultrasound attenuation (BUA), and total fat and lean mass were measured 1 year later (QDR 4500A, Sahara QUS; Hologic).

Results: Participants (N = 2200) were 48% men and 37% black. Poor nerve function (lower monofilament detection, higher vibration threshold, lower CMAP, lower NCV) was associated with 1.4–5.7% lower BUA and significant for all but NCV, adjusted for demographics, diabetes, body composition, and physical ability. Results were similar for adjusted hip BMD, with 1.0–2.9% lower BMD, significant for monofilament and CMAP testing. When considering the components of BMD, total hip area was 1.8–4.9% higher in those with the worst nerve function, although BMC showed little difference. Lower monofilament detection and CMAP were independently associated with lower heel BUA (p < 0.01), and monofilament detection was associated with lower hip BMD (p < 0.05) in regression additionally adjusted for lifestyle factors, bone-active medications, and diabetes-related complications.

Conclusions: Poor peripheral nerve function may directly related to lower BMD, likely through an increase in bone area in older adults, independent of lean mass, strength, physical ability, and diabetes. Whether those with impaired nerve function are at higher risk for fracture independent of falls needs to be studied.

INTRODUCTION

Bone tissue is innervated, and neurotransmitters directly affect bone remodeling.(1) However, very little is known about the impact of peripheral nerve function on BMD. The prevalence of poor peripheral nerve function is higher in older adults, even among those without diabetes.(2–11) In the United States for 1999–2000, 28% of adults 70–79 years of age and 35% of adults ≥80 years of age had peripheral neuropathy based on a simple screen for reduced sensation at the foot.(11) Peripheral nerve function decline is a well-recognized complication of type 1 and type 2 diabetes.(12) To our knowledge, the relationship of peripheral neuropathy and bone mass has not been studied in community-dwelling type 2 diabetic adults or nondiabetic adults. However, in type 1 diabetes, peripheral neuropathy is associated with lower BMD and lower calcaneal quantitative ultrasound (QUS).(13–17) Microvascular complications such as peripheral neuropathy may also be a marker for tissue ischemia that influences bone,(18,19) particularly at peripheral sites. Conversely, an increase in blood supply to the bone caused by nerve damage may increase bone resorption and cause weakening of bone,(20) as is theorized for Charcot neuroarthropathy.

Poor peripheral nerve function may increase the risk of fractures indirectly because of loss of balance and greater risk of falling.(21–25) Both type 1(26,27) and type 2 diabetic adults(26–31) are at a greater risk for fracture. Loss of protective sensation was related to increased risk of nonvertebral fracture in older white women in addition to diabetes.(28) We recently reported that older type 2 diabetic adults enrolled in the Health, Aging and Body Composition Study (Health ABC) had a higher rate of bone loss(32) and fractures.(31) For these analyses, we tested the hypothesis that sensory and motor peripheral nerve function is related to BMD and calcaneal QUS independent of diabetes status in an older community-based population.

MATERIALS AND METHODS

Study participants

Participants (N = 2200) were from 3075 well-functioning older white and black adults (48.4% men and 41.6% black), 70–79 years of age, in the Health ABC, with a baseline examination in 1997–1998 and follow-up exams in 2000–2001 and 2001–2002. Health ABC is an ongoing prospective cohort study investigating changes in body composition as a common pathway by which multiple diseases contribute to disability. Potential participants were recruited from a mailing to (1) a random sample of white Medicare beneficiaries and (2) all age-eligible black community residents in Pittsburgh, PA, and Memphis, TN. The mailed invitation was followed by a telephone screening interview to determine eligibility. Eligible participants reported no difficulty walking a quarter of a mile (400 m), climbing 10 steps, or performing activities of daily living; were free of life-threatening cancers with no active treatment within the past 3 years; and planned to remain within the study area for at least 3 years. Eligible participants were scheduled for a home interview, where eligibility was confirmed and a comprehensive interview was conducted followed by a clinic examination. Participants provided informed consent before exams, approved by the institutional review boards at the University of Pittsburgh and the University of Tennessee Health Science Center. Of 3075 participants at baseline, 2396 had an in-person exam at the clinic or home in 2001–2002, and of these, 2384 (99.5%) had data on either calcaneal BUA or hip BMD. The remaining portion of the original cohort from 1997–1998 had telephone follow-up only (N = 337), were deceased (N = 263), withdrew (N = 9), or missed the exam (N = 70) in 2001–2002. We excluded clinic or home exam participants with missing diabetes or impaired fasting glucose (IFG) status based on a fasting blood glucose test in 2000–2001 (n = 27) or diabetes onset in childhood (≤20 years old; n = 5). An additional 152 participants were missing data on peripheral nerve function from the 2000–2001 clinic exam. We therefore included 2200 participants (723 white men, 334 black men, 664 white women, and 479 black women), representing 71.5% of 3075 baseline participants in 1997–1998 and 91.8% of 2396 participants with a clinic or home exam in 2001–2002.

Peripheral sensory and motor nerve function

Peripheral nerve function measures at a clinic exam in 2000–2001 included monofilament testing (reduced sensation defined as inability to feel three of four touches at the great toe for both 1.4- and 10-g monofilament), average vibration threshold in μ (VSA-3000 Vibratory Sensory Analyzer; Medoc), and peroneal motor nerve conduction amplitude in millivolts (CMAP) and velocity in meters per second (NCV) from the popliteal fossa to ankle (NeuroMax 8; XLTEK). The right leg was tested on all participants unless contraindicated.

BMD, QUS, and body composition

Height was measured using a stadiometer, and weight was measured with a calibrated balance beam scale. Body mass index (BMI) was calculated as weight divided by square height (kg/m2). Total hip BMD (g/cm2 from BMC in g/total area in cm2) and whole body bone mineral–free lean mass (LM) and fat mass (FM) were assessed by DXA (Hologic 4500A, software version 9.03; Hologic) at a clinic exam in 2001–2002. Change in total LM and FM in 1 year, from the time of peripheral nerve assessment to the time of BMD measurement, was calculated. Calcaneal QUS (Sahara; Hologic) at either a home or clinic exam in 2001–2002 assessed broadband ultrasonic attenuation (BUA). QUS was performed in duplicate with repositioning of the heel in all participants, with a third measurement was taken if the first two differed by 10 dB/MHz or more, and the measures were averaged. DXA and QUS quality assurance measurements performed at both study sites ensured scanner reliability and identical scan protocols. The right hip and heel were tested on all participants unless contraindicated.

Other measures

Health and medical histories included smoking in 1999–2000, alcohol consumption at baseline, osteoporosis at baseline, and diabetes-related complications at baseline (peripheral arterial disease; cerebrovascular disease—transient ischemic attack or stroke; cardiovascular disease—bypass/coronary artery bypass graft (CABG), carotid endarterectomy, myocardial infarction (MI), angina or congestive heart failure); and eye diseases in 1999–2000—retinopathy/retinal disease, cataracts, macular degeneration, and glaucoma). Other clinical and questionnaire data were collected in 2000–2001 unless otherwise noted. Weekly physical activity from walking and stair climbing (kcal/kg/week) and falls in the prior 12 months were determined by an interviewer-administered questionnaire. Medications from the prior week were brought to the clinic in 1999–2000. Thiazide diuretic, statin, estrogen, oral steroid, thyroid, and osteoporosis medications (bisphosphonates, calcitonin, raloxifene, fluoride), and calcium and vitamin D supplement use were coded using the Iowa Drug Information System (IDIS) ingredient codes.(33) Serum creatinine levels ≥1.5 mg/dl for men and ≥1.3 mg/dl for women defined renal insufficiency at baseline.(34) Fasting blood draw was used to measure total cholesterol(35) and glucose. Diabetes was defined as self-report of physician diagnosis, hypoglycemic medication use, or fasting glucose ≥126 mg/dl (≥7.0 mM).(36,37) Hypertension was defined through self-report, medication use, and/or blood pressure measured at clinic exam. Ankle-brachial index <0.9 measured subclinical cardiovascular disease as previously described.(38) The Health ABC performance battery was a supplemented version of the lower-extremity performance test used in the Established Populations for the Epidemiologic Studies of the Elderly (EPESE; chair stands, standing balance, 6-m walk for gait speed)(39) with increased test duration, a single foot stand, and a narrow walk test of balance as previously described (score range, 0–12).(40) Knee extension strength and ankle dorsiflexion strength in 2001–2002 were measured concentrically at 60°/s on an isokinetic dynamometer (Kin-Com dynamometer, 125 AP) in three to six trials. The right leg was tested on all participants unless contraindicated. Quadriceps strength was calculated as the mean maximal torque produced (Nm) between 90° and 30° of knee extension. Ankle dorsiflexion strength was calculated as the mean maximal torque produced (Nm) between 72° and 30° of ankle extension.

Statistical analyses

Differences in prevalence and univariate associations were tested separately by race and sex using Pearson χ2 methods and Fisher exact test when appropriate. For continuous variables, nonparametric one-way Mann-Whitney tests were performed because of non-normal distributions. Partial correlations coefficients were computed for nerve function measures, adjusting for age, race, and sex. Nonparametric Kruskal Wallis test of means were performed to test trends across monofilament categories (no detection, detected only 10-g, detected 1.4-g) or quartiles of nerve function measures. The quartiles (from worst to best) were ≥73.30, 42.00–73.28, 22.98–41.98, and ≤22.93 μ for vibration threshold; ≤1.6, 1.7–3.0, 3.1–4.5, and ≥4.6 mV for CMAP; and ≤39.2, 39.3–43.0, 43.1–46.8, and ≥46.9 m/s for NCV. Data are presented for each nerve measure separately, and each is entered into the model as a separate dependent variable, because each represents a distinct component of nerve function with modest correlation (r = 0.10–0.24, adjusted for age, sex, and race) between measures. Vibration threshold was analyzed by quartiles for regression analyses because of its skewed distribution to very low or very high threshold values.

Total hip BMD/BMC/area, femoral neck BMD/BMC/area, and BUA means across categories or quartiles of nerve function measures were calculated with ANCOVA, adjusted for age, race, sex, clinic site (Memphis or Pittsburgh), diabetes, LM, FM, changes in LM and FM, strength (quadriceps for BMD/BMC/area and ankle for QUS), and Health ABC performance battery categorical score (0–12). Stepwise multiple linear regression was performed with BMD, BMC, total area, or BUA as dependent variables and nerve function measures (monofilament detection, quartiles of vibration threshold, CMAP, and NCV) as the independent variables of interest, while adjusting for covariates included in ANCOVA analyses and additionally for height, falls, weekly physical activity from walking and stair climbing, smoking status, drinking frequency, previously diagnosed osteoporosis, bone-active medication use (osteoporosis medications, oral estrogens, oral steroids, calcium and vitamin D supplementation, thiazide diuretics, statins), thyroid medication use, and diabetes-related conditions or complications (cholesterol, blood pressure, hypertension, low ankle-arm index, history of cardiovascular disease: coronary heart disease, peripheral arterial disease, transient ischemic attack (TIA)/stroke, kidney disease: creatinine > 1.5 men and ≥ 1.3 women; eye disease: glaucoma, retinopathy/retinal disease, cataracts). Interactions (race × sex, each nerve measure × sex, each nerve measure × race, each nerve measure × diabetes) entered into the models were not significant, and therefore, models were not stratified by any of these variables. Models met underlying assumptions and were built progressively by entering variables stepwise in the following order: demographic factors, diabetes, body composition factors, nerve function measures, physical performance, other risk factors for BMD, and diabetes-related comorbidities. Sex, race, age, study site, and diabetes were included in all models, and the remaining variables were removed in a stepwise manner at p > 0.10. Models were also run that excluded participants with osteoporosis, diabetes, or oral steroid users. Multicollinearity for independent variables was assessed using the variance inflation factor (VIF), the inverse of the proportion of variance not accounted for by other independent variables; no VIF was >10, and the mean VIF for each regression model was ≤2.(41) Data were analyzed using SPSS statistical software.

Percentage difference in BMD or BUA caused by nerve function in the final linear regression models was calculated using the following formula: [(unstandardized β for nerve function)(unit change in nerve function)/unadjusted BMD or BUA mean for entire sample] × 100. Confidence intervals (95%) for percentage change in BMD or BUA caused by nerve function were calculated using the following formula: {[(unstandardized β for nerve function)(unit change in nerve function) ± (SE of β for nerve function)(1.96)]/unadjusted BMD or BUA mean for entire sample} × 100.

RESULTS

Descriptive characteristics for each race and sex group are shown in Table 1. Participants were 48% men and 37% black. Men and black participants were more likely to have diabetes then women or white participants. White women were more likely to take bone-active medications and had lower BMI, LM, FM, and strength than black women. White men and women were more likely to drink and less likely to smoke than black men and women, respectively. White men and women were more likely to have physical activity from walking and stair climbing and higher Health ABC performance scores than black men and women, respectively. More women than men reported falling in the past year.

Table Table 1.. Descriptive Characteristics by Sex and Race*
original image

Men had worse sensory and motor nerve function than women for all measures. Men were less likely to detect a 10-g monofilament (11.8% versus 5.8%; p < 0.001), had higher average vibration threshold (58.3 versus 43.5 μ; p < 0.001), and lower CMAP (2.9 versus 3.6 mV; p < 0.001) and NCV (41.6 versus 44.7 m/s; p < 0.001) than women. White women had worse CMAP than black women. Men were more likely to have a history of vascular disease, and black men and women had more CVD risk factors than white men and women, respectively. High creatinine was more common in men and blacks. History of glaucoma was more often reported in blacks, but retinal disease and cataracts were more common in white men than black men.

The tests of trend in unadjusted hip BMD and femoral neck BMD means were statistically significant across categories or quartiles for all nerve function measures. After adjustments for covariates, the worst nerve function categories were associated with 1.0–2.9% lower adjusted total hip BMD. The tests of trend in unadjusted calcaneal BUA means were statistically significant across quartiles for average vibration threshold and CMAP. After adjustments for covariates, the worst nerve function categories associated with 1.4–6.1% lower adjusted BUA. Table 2 shows the adjusted means for total hip BMD and calcaneal BUA by groups or quartiles of each nerve function measure. Significant differences in total hip BMD and BUA were evident across the monofilament categories and quartiles of nerve conduction testing, with the lowest hip BMD and BUA in the worst sensory and motor nerve function groups. Hip BMD was 2.9% lower across monofilament categories (0.19 SD of BMD) and 1.5% lower across CMAP quartiles (0.11 SD of BMD). BUA was 3.8% lower across monofilament categories (0.17 SD of BUA), 6.1% lower across CMAP quartiles (0.28 SD of BUA), and 5.1% lower across NCV quartiles (0.23 SD of BUA). No differences in adjusted BMD or BUA means existed for average vibration threshold quartiles. Relationships of adjusted femoral neck BMD means to each nerve function measure were similar to those for total hip BMD (data not shown).

Table Table 2.. Adjusted Means (±SD) for Total Hip BMD (g/cm2), BMC, and Area and Calcaneal BUA (dB/MHz) in Nerve Function Test Groups/Quartiles
original image

The adjusted means for total hip BMC and area are also shown in Table 2. Test for trend in unadjusted hip BMC and area means were statistically significant across categories or quartiles for all nerve function measures. After adjustments for covariates, the worst nerve function categories were associated with a 1.8–2.3% higher total hip area. Significant differences in total hip area were evident across all nerve function measures, with the highest hip area in the worst sensory and motor nerve function groups. Hip area was 1.8% higher across monofilament categories (0.21 SD of area), 2.3% higher across vibration threshold quartiles (0.28 SD of area), 2.3% higher across CMAP quartiles (0.28 SD of area), and 2.8% higher across NCV quartiles (0.34 SD of area). The only difference across nerve function in adjusted BMC means was in the monofilament detection groups, with 0.6% lower BMC in the worst group (0.04 SD of BMC). Relationships of adjusted femoral neck BMC and area means to each nerve function measure were similar to those for total hip BMC and area (data not shown).

Table 3 shows the fully adjusted multivariate linear regression models for total hip BMD, total hip area, and calcaneal BUA. Lack of monofilament detection (either 10- or 1.4-g) was associated with 1.1% (95% CI: 0.02–2.3%) lower total hip BMD and 2.5% (95% CI: 0.7–4.2%) lower calcaneal BUA, respectively, and a 1 mV lower CMAP was associated with 1.1% (95% CI: 0.5–1.6%) lower calcaneal BUA, when all nerve function measures were entered into the models simultaneously. A 1 mV lower CMAP was associated with 0.29% (95% CI: 0.19–0.48%) higher total hip area. Total hip BMC was not related to any nerve function measure in fully adjusted models (results not shown). Average vibration threshold and NCV were removed from all models at p > 0.10. If diabetic participants were excluded from analyses, all relationships of BUA or total hip area and nerve function measures remained the same. However, in the model with only nondiabetic participants, total hip BMD was related marginally (p < 0.10) to CMAP but not monofilament detection. If oral steroid or osteoporosis medication users were excluded from analyses, the relationships with nerve function did not change.

Table Table 3.. Multivariate Linear Regression Models* for Total Hip BMD (g/cm2) and Calcaneal BUA (dB/MHz)
original image

DISCUSSION

In this population of community-dwelling older adults, both poor sensory and motor peripheral nerve function were independently associated with lower BMD 1 year later. We show an association of poor peripheral nerve function and larger total hip area, although there was no relationship with BMC. These data suggest that, although the mineral and periosteum of bone are both innervated,(1) poor peripheral nerve function may affect separate components of bone differently. Our results imply that the age-related worsening of peripheral nerve function(2–11) may be associated with lower BMD through an increase in bone area. Our adjustments for lean mass, strength, and physical performance did not eliminate this relationship, suggesting that peripheral nerve function directly impacts bone. The BMD differences observed with poor nerve function are significant and, although fairly small in absolute magnitude and SD of BMD, may possibly have an impact on fracture. A 1 SD decrease in BMD is associated with a 2- to 3-fold hip fracture increase.(42,43) Calcaneal QUS, a peripheral bone site that could be more influenced by peripheral nerve function and peripheral vascular disease, was related to sensory and motor nerve function in our study. Potentially, vascular alterations related to peripheral nerve problems(44) may influence BMD.(18–20)

Sensory nerve testing with the monofilament showed the strongest relationship to BMD and BUA after complete multivariate adjustment. Lack of 10-g monofilament detection, generally associated with clinical disease that is predictive of future foot ulcers,(45,46) was perhaps an indicator of more severe reduction in peripheral sensation than the vibration testing, our other sensory nerve assessment. However lack of 1.4-g monofilament detection, a more sensitive measure than the 10-g monofilament, also showed a relationship with BMD and BUA.

The motor nerve measures of CMAP were related to BUA and total hip area after multivariate adjustment. We are uncertain as to why NCV showed a weaker association with bone than CMAP. Low CMAP is related to axonal damage of the nerve, and low NCV is related to demyelination of the nerve.(47) Possibly certain types of nerve degeneration, such as motor axon loss, may have a greater impact on BMD. Interestingly, the InCHIANTI Study recently found that CMAP, but not NCV, was independently related to calf muscle density in older adults.(48) These data suggest that the relationship we observed between poor nerve function and BMD may be similar to, although independent of, the relationship with muscle, because we adjusted for lean mass and strength. In diabetes, severe peripheral neuropathy is quite clearly related to muscle atrophy.(49) Diabetic neuropathy in type 1 diabetes is associated with a 50% reduction in muscle volume, with a high correlation between neuropathy score and muscle volume (r = −0.75, p < 0.001).(50)

In type 1 diabetes, peripheral neuropathy is related to lower BMD and QUS measures,(13–17) although there is a lack of information for type 2 diabetic adults and nondiabetic adults. We found no significant interaction of type 2 diabetes and any nerve function measure, indicating that the relationship with BMD was consistent for diabetic and nondiabetic participants. The relationship with peripheral nerve function remained after diabetic participants were excluded from analyses, particularly for the peripheral bone site (calcaneal QUS).

Poor peripheral nerve function increases the risk of falling in older adults.(21–25) The loss of protective sensation was associated with risk of nonvertebral fracture in older white women independent of diabetes.(28) Several other recent studies, including the Health ABC Study, have also shown that older type 2 diabetic adults have higher fracture rates(26–31) than expected given their generally higher BMD.(51) Our previous work in this population found that type 2 diabetic participants with fractures were less likely to detect a standard monofilament and had more frequent falls than the diabetic participants without fractures,(31) suggesting that poor nerve function contributes to fracture risk. Older diabetic adults in the Health ABC cohort had greater bone loss than nondiabetic older adults.(32) One mechanism for this greater bone loss among diabetic adults could be greater declines in peripheral nerve function. Although peripheral nerve function contributes to falls, our study indicates that there may also be a direct impact on BMD. We are currently collecting longitudinal fracture data that will allow us to determine if poor nerve function is prospectively related to increased fracture rates.

The relationship between peripheral nerve function and BMD has not been previously examined in an older cohort that included nondiabetic adults. Our study included state-of-the-art comprehensive assessments of both sensory and motor peripheral nerve function and BMD, including a peripheral site. We were able to control for a large number of covariates that may be related to both our outcome and exposure of interest. Reduced nerve function is associated with decreased physical functioning,(52–56) potentially through body composition changes, and we adjusted for performance battery scores, strength, body composition, and body composition changes to address this. Our study allows only suggestions of causal relationships and more longitudinal data on peripheral nerve function changes and BMD alterations are needed. Peripheral nerve function was likely poorer in individuals that did not participate in the Health ABC Study because of greater physical disability and in Health ABC participants that were not able to return for follow-up clinic exams; therefore, these results may apply only to ambulatory community-dwelling older adults.

In summary, this unique multiethnic study of men and women showed a consistent association of poor sensory and motor nerve function and lower BMD 1 year later, which was not accounted for by diabetes status or risk factors for low BMD and independent of lean mass, strength, and physical performance. Therefore, poor peripheral nerve function may be directly related to lower BMD, likely through an increase in bone area. Given the high prevalence (>30%) of lack of 10-g monofilament detection among older adults in the 1999–2000 NHANES population,(11) peripheral nerve deficits are an unappreciated problem in the elderly. Future research may determine if a simple screening for loss of touch sensation with a monofilament can identify older diabetic and nondiabetic adults at risk for low BMD and fractures. Whether the relationship of poor peripheral nerve function and lower BMD will impact clinical osteoporosis or fracture rates needs to be examined, and longitudinal fracture data are currently being collected for our cohort.

Acknowledgements

This work was funded by the National Institutes on Aging (NIA) contracts N01-AG-6-2101, N01-AG-6-2103, and N01-AG-6-2106. This research was supported in part by the Intramural Research program of the NIH, National Institute on Aging. Dr Strotmeyer is supported by a Junior Faculty Award from the American Diabetes Association.

Ancillary