Isokinetic Leg Muscle Strength in Older Americans and Its Relationship to a Standardized Walk Test: Data from the National Health and Nutrition Examination Survey 1999–2000

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Isokinetic Leg Muscle Strength in Older Americans and Its Relationship to a Standardized Walk Test: Data from the National Health and Nutrition Examination Survey 1999–2000
  Isokinetic Leg Muscle Strength in Older Americans and ItsRelationship to a Standardized Walk Test: Data from the NationalHealth and Nutrition Examination Survey 1999–2000 Yechiam Ostchega, PhD, RN,  Charles F. Dillon, MD, PhD,  Rosemary Lindle, PhD, w  Margaret Carroll,  and Ben F. Hurley, PhD w  OBJECTIVES:  Todescribeisokinetickneeextensormusclestrength in older U.S. men and women by age and race/ ethnicity and to ascertain its relationship to a standard,timed walking-speed test. SETTING:  The U.S. National Health and NutritionExamination Survey (NHANES) 1999–2000. DESIGN:  A cross-sectional nationally representativehealth examination survey. PARTICIPANTS:  All surveyed persons aged 50 and older(N 5 1,499) who performed muscle strength and timedwalk examinations in the NHANES mobile examinationcenter. MEASUREMENTS:  Concentric peak torque (strength) of thekneeextensorsat1.05rads/s  1 velocityanda6-mwalktimed in seconds. RESULTS:  Kneeextensorstrengthwasinverselyassociatedwith age ( P o .01), and women had less knee extensormuscle strength than men ( P o .01). After adjustment forstanding height, no significant difference in muscle strengthwas found across the three race/ethnicity groups (MexicanAmericans,non-Hispanicblacks,andnon-Hispanicwhites)for men or women.After adjustment for age, race/ethnicity,weight, and height, increasing knee extensor strength wasassociated with significant increases in meters walked persecond ( P o .01). CONCLUSION:  Knee extensor muscle strength is affectedby age and sex but not by race/ethnicity and it issignificantly associated with timed walk.  J Am Geriatr Soc52:977–982, 2004. Key words: isokinetic knee extensor strength; timed walk;NHANES R esearch shows that muscle mass and strength declinewith age. 1–3 The decline is associated with deteriora-tion in health status, increased limitation of activities of daily living, a loss of independence and quality of life, 4 andincreased use of healthcare services and their relatedcosts. 5,6 Muscle strength is an important determinant of physical functional capacity at older ages. 7 For example,age-related reductions in lower extremity strength areassociated with impairments in gait speed, balance, stairclimbing, and chair rising. 8–13 The description of age-specific muscle strength in theU.S. population may be a useful tool to measure theproportion of persons experiencing major activity limita-tions. 14 It is also essential to provide clinicians and publichealth officials with the basic information needed to targetpreventive care interventions. Indeed, a number of studiesshow that exercise, especially resistance exercise training,can increase muscle strength 15 and improve individualfunctional abilities and balance. 16,17 Because of the lack of population-based data in the United States, measurementsof muscle strength were incorporated into the NationalHealth and Nutrition Examination Survey (NHANES)beginning in 1999. The purpose of this study is to reportthe results of these measurements, comparing age, sex, andrace/ethnicity estimates of concentric peak torque in theknee extensors, the key muscle group for ambulation andbalance. Additionally, the relationship between musclestrength values and measured walk time, a measure of functional status, is described. 6 METHODSStudy Design and Population The NHANES survey uses a stratified multistage prob-ability design. The procedures followed to select the sampleand conduct the interview and examination were similar tothose for previous NHANES surveys. 18 NHANES consistsof a detailed home interview and a health examinationconducted in a mobile examination center (MEC). Begin-ning in 1999, NHANES became a continuous survey, witheach survey year a nationally representative sample of theU.S. civilian noninstitutionalized population. Two or more AddresscorrespondencetoYechiamOstchega,PhD,RN,NationalCenterforHealth Statistics, NHANES Program, 3311 Toledo Road, Rm. 4319,Hyattsville, MD 20782. E-mail: yxo1@cdc.govFrom the   Division of Health and Nutrition Examination Surveys, NationalCenter for Health Statistics, Centers for Disease Control and Prevention,Hyattsville, Maryland; and  w Department of Kinesiology, University of Maryland, College Park, Maryland.  JAGS 52:977–982, 2004 r 2004 by the American Geriatrics Society 0002-8614/04/$15.00  years of data are necessary to have adequate sample sizesfor subgroup analyses. This report is based on the first 2years of the continuous survey, NHANES 1999–2000.Informed consent was obtained from all participants, andthe institutional review board of the National Centerapproved the protocol for Health Statistics.Of the srcinal 3,278 potential participants aged 50and older who were screened as eligible for the NHANES1999–2000 survey, 2,420 (73.8%) completed an interview,and 2,156 (65.8%) attended the MEC for examination  , which included an assessment of their right isokineticquadriceps muscle strength and a 6-m timed walk test; 237persons were missing data because of equipment or datacapture failure, limited time available to do the examina-tion, refusal, or other reasons. This reduced subjectresponse to 58.5%. Those missing data in this latter groupwere found to be more likely to be aged 70 and older (44%vs 35% of those with data), non-Hispanic black (23% vs17% of those with data), and female (56% vs 49% of thosewith data). The data were reweighted to adjust for thesedifferences (see below).Additionally, for subject safety reasons, participantswere excluded from isokinetic muscle testing if they hadhad chest or abdominal surgery in the previous 3 weeks,myocardial infarction within the previous 6 weeks, ahistory of aneurysm/stroke, current severe right knee pain,right knee or hip replacement, or severe neck or back pain.Of participants aged 50 and older who attended the MECfor examination, an additional 420 were excluded for thesesafety and health reasons. Complete information on musclestrength and timed walk were therefore available on 758menand741women,foratotalof1,499individuals.Thosemissing data because of health reasons were more likely tobe aged 70 and older (52% vs 35% of those with data) andMexican American (32% vs 23% of those with sufficientdata).Theeffectofsubjectexclusionsbecauseofhealthandsafety considerations was to change the effective samplingframe for NHANES isokinetic muscle-strength testing; theresults of this study are properly generalized only to thespecific subset of older, relatively healthy noninstitution-alized U.S. adults who have no health-related contraindica-tion to isokinetic muscle-strength testing. Measurements Lower body peak torque (PT) was assessed using a KineticCommunicator isokinetic dynamometer (Kin Com MP,Chattecx Corp., Chattanooga, TN). Maximal voluntaryconcentric muscle force was measured in Newtons in theright knee extensor (quadriceps) muscles at an angularvelocity of 1.05 rads- s–1 (60 deg- s–1). Each study subjecthad a total of six trials during the strength test: threepractice trials for warm-up and three trials for maximalvoluntary effort. During the first three trials, subjects wereencouraged not to exert maximal effort, whereas maximaleffort was strongly encouraged during the last three trials.Only the right leg was tested, because previous studies havereported no significantdifference between torquegeneratedby left and right knee flexor and extensor muscle groups. 3 Also, for safety concerns, only concentric movements weretested. 13 PT was calculated according to the followingalgorithm: for each examinee who had more than fourtrials, the highest peak force (PF) value obtained after thethird trial was selected. If four or fewer trials werecompleted, the highest PF from the completed trials wasselected. Ninety-eight percent of the examinees had five ormore maneuvers. PT was calculated as (PF  mechanicalarm length in cm)/100. The mechanical arm lengthrepresented the distance from ankle to knee joint. Gravitycorrections to torque were based on a measured leg weightat 2.62 rads (150 1 ; 3.14 rads being equivalent to a straightlegposition).Theproceduremanual,whichcontainsdetailsabout calibration and the examination protocol, is avail-able on the NHANES Web site. 19 The timed walk test consisted of a 20-foot (6.15 meter)walk at the subject’s usual pace, measured using astopwatch and recorded to the nearest hundredth of asecond. 20,21 Walk time was measured from the time theexaminee’s foot first touched the floor across the start lineand stopped when the examinee’s foot touched the flooracrossthefinishline.Examineeswereexcludedfromtestingif they needed the assistance of another person to walk butwere allowed to use a walking device such as a walker orcane, if needed. Timed walk speed was converted to m/s toallow international standard for comparisons. Finally,height and weight measured using standard NHANESprotocols were included in the analysis. Statistical Analyses For all NHANES surveys, calculated sampling weights takeinto account the unequal probability of selection resultingfrom planned oversampling of certain subgroups. Samplingweightsarealsoadjustedfornonresponse.Allanalysestookinto account the sample weights and the complex multi-stage cluster design.ForthespecificNHANESsubsetofPTdata,thesampleweights were additionally adjusted for nonrandom non-response as follows. The eligible sample population forcomputing national weighted PT prevalence estimates isconsideredtobethesetofindividualsnotexcludedfromtheprocedureduetohealthreasons.Thisincludedpersonswithcollected data (n 5 1,499) and persons who were missingPT data because of equipment or data capture failure, timelimitations,refusals,orotherreasons(n 5 237).Withinthissubset, individuals with missing PT data were sufficientlydifferent by age, sex, and ethnicity from people withcompletedatatoprecludeconsideringthemtobemissingatrandom. Therefore, to compute weighted estimates propor-tional to the U.S. population, revised sampling weights thatadjustedfor the nonresponse due to these nonhealth relatedcauses in the PT data were constructed.When managing nonresponse that cannot be consid-ered missing at random, one study 22 recommends aprocedure to adjust the srcinal sampling weights of therespondent sample by a weight factor that accounts for thedifferences between respondents (the measured group) andnonrespondents (those eligible but not measured). Thecurrent analysis determined that the principal demographicvariables related to nonhealth-related nonresponse for thePT data set were age, sex, and race/ethnicity. Moreover, ageand sex are the principle characteristics affecting PT. 3 Sample persons were therefore classified into 24 subgroupsdefined by age (50–59, 60–69,   70), sex (M,F), and race/  978  OSTCHEGA ETAL.  JUNE 2004–VOL. 52, NO. 6 JAGS  ethnicity (non-Hispanic black (NHB), non-Hispanic white(NHW), Mexican American (MA), and other). It was thenassumed that the data were missing at random within eachof these 24 subgroups. Based on a previous report, 22 withineachofthe24subgroups,anadjustmentfactoriscalculatedasthesumoftheweightsforalleligiblepersonsinsubgroupA and B divided by the sum of the weights for thoseresponding in the subgroup (Group A). The samplingweight for each respondent in the subgroups is thenmultiplied by the subgroup weight factor to create revisedsampling weights. Except for slight rounding errors, thesum of the srcinal overall sample weights, including 1,736group A eligible but missing and group B respondents(58,459,257) was equal to the sum of revised samplingweights after adjustment for nonresponse (58,459,221).All descriptive statistics for PT were calculated usingthe adjusted weights. Race/ethnicity was categorized asNHB, NHW, and MA. Persons not classifiable into one of thesegroupswereincludedinalltotalpopulationestimates.Variance estimates were calculated using SUDAAN, afamily of statistical procedures for analysis of data from acomplex sample survey (Research Triangle Institute, Re-search Triangle Park, NC). 23 Following NHANES analyticguidelines, the Survey Data Analysis (SUDAAN) delete-one-jackknife method was used, partitioning the sampleinto52samplingunitsandforming52replicatesbydeletingone unit at a time. Ninety-five percent confidence intervals(CIs) were computed using the critical value for a  t   distri-bution with the appropriate number of degrees of freedomfor each subgroup. Age adjustment for comparing prev-alenceestimatesbetweenracial/ethnicgroupsandsexeswasperformed using the direct method, employing the stan-dard U.S. population estimates provided in the NHANESanalytic guidelines. 19 Coefficients of variation and designeffects were computed specifically for the study prevalenceestimates. All reported estimates met NHANES analyticguidelines for statistical reliability (coefficients of variationless than 30% and observed design of 3 or less). There wereinsufficient sample sizes to estimate PT prevalence simulta-neously by age, sex, and race/ethnicity.The reported univariate statistical significance testresults for comparisons between the three age and race/ ethnicity categories are the unadjusted values, but aunivariate statistical test was considered significant onlyat  P o .016 (Bonferroni correction method for multiplecomparisons;  a 5 0.05/3).A number of studies have shown skeletal muscle masstobepositivelycorrelatedwithbodyheight 24,25 andmusclestrength. 26–28 Theoverallcorrelation,inthisstudy,betweenPTand height was  r 5 0.67, and the current survey showedthat, in both sexes, MAs are on the average shorter thanNHWsandNHBs.MAmenareonaverage7.26cmshorterthan NHW men and 5.83 cm shorter than NHB men; MAwomen on average are 5.12 cm shorter than NHW womenand 5.56 cm shorter than NHB women (for all contrasts, P o .0001).Therewasnostatisticallysignificantinteractionbetween height and race/ethnicity ( F  5 1.01,  P 5 .4) whenPT was the dependent variable. PT estimates by race/ ethnicity are therefore reported as unadjusted values andheight-adjusted estimates. The SUDAAN Proc REGLSMEANS option was used to provide adjusted meansfor covariate and classification variables.Because recent studies suggest a possible thresholdeffect in the relationship between muscle strength andwalkingspeed, 13,29 apreliminaryanalysiswasperformedtoinvestigate thispossibility ineachsex.Specifically, residualswere obtained from a linear regression model for each sexthat included age, race/ethnicity (MA, NHB), standingheight, and weight as independent variables and time towalk 6 m as the dependent variable. These residuals can beinterpreted as performance results (timed walk) with thecovariatesinthemodelremoved.Theresidualswereplottedby sex against PT, the measure of isokinetic muscle strengthusing locally weighted regression smoothers (SAS 8.2PROC LOESS, SAS Institute, Inc., Cary, NC). 30 Nothreshold effect was identified for either sex (data notshown).Thereforeasex-specificweightedstepwisemultiplelinear regression analysis was used to determine whethermuscle strength as measured using PTwas an independentpredictorof6-mwalkperformanceasmeasuredinm/safteraccounting for age, race/ethnicity, height, and weight. RESULTS The overall average weighted PT for the total sample(N 5 1,499) was 111 Nm (95% CI 5 108–114). Table 1presents sample sizes for the study demographic subgroupsand mean PT values by sex and age groups and age-sexsubgroups. As expected, men had higher PT values thanwomen, and PT values decreased with age. These age andsex differences were statistically significant, allowing formultiple comparisons ( P o .01). In relative terms, thedecline in muscle strength with age is comparable inmagnitude between men and women. It decreases by 33%between ages 50 and 70 in men and by 29% from age 50 to70 in women. For all study subjects (N 5 1,499), the meanperformance time    standard deviation for the 6-m walkwas 6.63    1.92 seconds. There were significant overalldifferences between men and women in timed walkperformance ( P o .0002). Six-m walk times were somewhatlongerwithincreasingage;forexample,thoseaged50to59averaged 5.84  1.23 seconds, whereas those aged 70 andolderaveraged7.47  2.23seconds.Thesedifferenceswerestatistically significant ( P o .0001), and except for compar-ing women aged 50 to 59 with men aged 60 to 69, therewere significant trends by age and sex ( P o .01). MA andNHB mean walking time differed significantly ( P o .001)from NHW (means 6.8, 7.1, and 6.3 seconds, respectively).Figure 1 represents the linear relationship between PT andtimed walking for men and women.Table 2 presents the overall estimates for mean PTandtheir 95% CIs by race/ethnicity and for race/ethnicity bysex. In the unadjusted data, NHBs did not differ signifi-cantly from NHWs, but MAs were significantly differentfrom NHWs. This was true for both sexes. MA men andwomen had significantly lower PT values than NHW menand women. The absolute magnitude of PT values for MAwomen was the lowest among all women, but when themean PT values were adjusted for height, the mean PTvalues for MA men and women were not significantlydifferent from those of their NHWand NHB counterparts.Inessence, thisindicatesthat,afterheightadjustment, therewere no significant differences for mean PT values acrossthe race/ethnic categories. ISOKINETIC KNEE EXTENSOR STRENGTH  979 JAGS JUNE 2004–VOL. 52, NO. 6  Multivariate Analyses Results of the stepwise multivariate linear regressionanalysis by sex are shown in Table 3. After accounting forall other independent demographic and anthropometricvariables, muscular strength as measured using PT was asignificant independent linear predictor of timed walk. Itaccounted for 20% of the coefficient of determination ( R 2 )in men and women.The PT beta coefficient was positively associated withwalking the course, suggesting that an increase in musclestrength is associated with an increase in distance coveredper second. Indeed, transforming PT values into an ordinalscalebycategorizingthemintoquartilesshowsthatthefirstquartile (0–25%) of PT values was associated with anaverage walk of 0.8 m/s. whereas the fourth quartile (75–100%) was associated with an average walk of 1.1m/s. DISCUSSION ThisisthefirstU.S.nationalexaminationsurveytodescribeisokinetic knee extensor muscle strength by age, sex, andrace/ethnicity. Although theNHANES samplingdesign andstudy execution have many strengths, there are definitelimitations to the data. The overall examination responserate for muscle testing in older persons was 65.8%.Equipment and technical difficulties in the MEC unitreduced this figure to 58.5%. Although it was attemptedto address this examination-level nonresponse by reweight-ing the data, additional subject losses could not be adjustedfor because of health-related safety exclusions from testing(19% of subjects). It is important therefore to reiterate thattheseresults areproperly generalized onlyto thesegment of the older U.S. civilian, noninstitutionalized population freefrom these health-related exclusion criteria.Consistent with previous study results, these findingsshowmentobestrongerthanwomenandthatkneeextensormuscle strength decreases with age for both sexes. 26,31,32 Few studies have examined age-related differences in musclestrength with regard to race. The results of these studies areconflicting, but most have reported no racial difference inage-relateddecline in muscle strength. 33–35 Recently, a studyusing the Health, Aging and Body Composition (HealthABC) Study data reported that black women had higherabsolute strength than white women, and similar to thefindings of the current study, there were no racial differencesinstrengthinmen. 26 However,theHealthABCStudyresultscan not be directly compared with the results of the currentstudy because no disability restrictions were imposed ascriteria for selection in NHANES, whereas the Health ABCStudy excluded individuals who reported mobility disordersor activity of daily living disability. Although it was not Table 1. Mean Knee Extensor Peak Torque for Persons Aged 50 and Older: NHANES 1999 to 2000 Parameter SamplePopulation Estimate  (Thousands)Peak Torque(N/m)95% ConfidenceIntervalTotal 1,499 58,459 111 108–114SexMale w 758 27,317 140 134–146Female 741 31,142 86 z 84–88Age50–59 w 433 24,737 130 124–13560–69 541 18,063 115 z 110–12070 1  525 15,658 86 z 82–90Age and sex50–59 men w 207 11,987 163 154–17260–69 men 278 8,913 144 z 138–150  70 men 273 6,417 109 z 102–11750–59 women w 226 12,751 98 95–10260–69 women 263 9,150 87 z 83–91  70 women 252 9,242 70 z 67–72  Based on standard U.S. population estimates, National Health and Nutrition Examination Survey (NHANES) 1999 1 Analytic Guidelines. w Reference. z P o .01; all estimates have coefficients of variation of less than 30 and design effects of 3 or less. Figure 1.  Linearrelationshipbetweenisokineticmusclestrengthand timed walk in seconds, U.S. men and women, NHANES1999–2000. 980  OSTCHEGA ETAL.  JUNE 2004–VOL. 52, NO. 6 JAGS  possible to study Asians as a distinct category, the resultsshowed no significant difference between race/ethnic cate-gories when mean PT values were adjusted for height. Thefindings suggest that height is a significant covariate thatneeds to be included in future analyses of muscle strength byrace/ethnicity. Recent work 36 has suggested that lowermuscle mass and greater body fat may be importantdetermining factors in muscle strength. Data were notavailable to address this issue in the current study.Knee extensor strength was associated with timedwalk performance for men and women. The relationshipshowed a clear gradient and no threshold effect. A study 28 using timed chair stand as an outcome variable describeda similar relationship. Similarly, another study 13 describeda linear relationship between comfortable gait time of 50feet and PT. They also found that height and weightexplained some of the variability in comfortable timedwalk. Table 2. Mean Knee Extensor Peak Torque for Persons Ages 50 and Older: Unadjusted Estimates and Estimates Adjustedfor Standing Height, NHANES 1999–2000 PopulationSample(n)PopulationEstimate(Thousands)  Peak Torque(N/m) 95% CIStandingHeightAdjusted MeanStanding HeightAdjusted95% CIRace/ethnicity w Non-Hispanic white z 784 44,966 114 110–118 113 109–117Non-Hispanic black 250 4,835 113 107–119 117 111–123Mexican American 338 2,019 99 § 93–104 121 116–126Sex and race/ethnicity w MenNon-Hispanic white z 411 21,458 144 137–150 145 139–155Non-Hispanic black 124 2,225 137 126–147 144 138–152Mexican-American 166 88,1 126 § 120–132 147 134–154WomenNon-Hispanic white z 373 23,509 86 84–88 85 82–88Non-Hispanic black 126 2,610 93 87–99 94 87–101Mexican-American 172 1,138 77 § 71–82 88 84–92  Estimates based on standard U.S. population estimates, National Health and Nutrition Examination Survey (NHANES) 1999 1 Analytic Guidelines. w Age adjusted using 2000 U.S. population estimates. z Reference. § P o .01. All estimates have coefficients of variation of less than 30 and design effects of 3 or less.CI 5 confidence interval. Table 3. WeightedStepwiseRegressionofMeanKneeExtensorPeakTorqueonTimedWalkPerformance(m/s),Adjustingfor Demographic and Anthropometric Covariates, NHANES 1999 to 2000 Predictor  b  Standard Error  R  2 After StepTotal Contributionto R  2 Men w Intercept 0.4845 0.2764Mean peak torque, N/m 0.0016 § 0.0002 .2014 .2014Age   0.0059 § 0.0010 .0392 .2406Non-Hispanic black   0.1431 § 0.0204 .0373 .2779Standing height, cm 0.0053 k 0.0015 .0140 .2919Weight, kg   0.0024 k 0.0007 .0243 .3162Mexican American   0.0645 k 0.0215 .0031 .3194Women w Intercept 0.6224 0.2156Mean peak torque, N/m 0.0033 § 0.0005 .1992 .1992Age   0.0064 § 0.0011 .0545 .2537Weight, kg   0.0038 § 0.0006 .0591 .3128Non-Hispanic black   0.0925 § 0.0258 .0152 .3280Standing height, cm 0.0049 k 0.0014 .0142 .3422Mexican American z  0.047 0.02  Calculated using weighted multivariate regression analysis delete one jackknife method.Overall  R 2 5 w .32;  z .34.  R 2 5 coefficient of determination. § P o .001; k P o .01. z Variable did not meet the .15 significance level for entry into the final stepwise model. ISOKINETIC KNEE EXTENSOR STRENGTH  981 JAGS JUNE 2004–VOL. 52, NO. 6
Related Search
Similar documents
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!