Research Article The Impact of Body Mass Index …downloads.hindawi.com › journals › jar › 2013 › 905094.pdfmortality risks than those in the United States. However, when BMI

Hindawi Publishing CorporationJournal of Aging ResearchVolume 2013 Article ID 905094 11 pageshttpdxdoiorg1011552013905094

Research ArticleThe Impact of Body Mass Index and Weight Changes onDisability Transitions and Mortality in Brazilian Older Adults

Flaacutevia Cristina Drumond Andrade1 Ahmad Iqmer Nashriq Mohd Nazan2

Maria Luacutecia Lebratildeo3 and Yeda Aparecida de Oliveira Duarte4

1 Department of Kinesiology and Community Health University of Illinois at Urbana-Champaign 1206 South 4th StreetChampaign IL 61820 USA

2 Zilber School of Public Health University of Wisconsin-Milwaukee 1240 N 10th Street Milwaukee WI 53205 USA3 School of Public Health Departament of Epidemiology University of Sao Paulo Avenida Dr Arnaldo 715 Pinheiros01246-904 Sao Paulo SP Brazil

4 School of Nursing Department of Medical-Surgical Nursing University of Sao Paulo Avenida Dr Eneas de Carvalho Aguiar419 3o andar Sala 318 Cerqueira Cesar 05403-000 Sao Paulo SP Brazil

Correspondence should be addressed to Flavia Cristina Drumond Andrade fandradeillinoisedu

Received 2 November 2012 Revised 7 March 2013 Accepted 13 March 2013

Academic Editor Heather Keller

Copyright copy 2013 Flavia Cristina Drumond Andrade et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

The aimof this studywas to examine the association between bodymass index andweight changes on disability transitions andmor-tality among Brazilian older adults Longitudinal data from the HealthWell-Being and Aging in Latin America and the CaribbeanStudy conducted in Sao Paulo Brazil (2000 and 2006) were used to examine the impact of obesity on disability andmortality and ofweight changes on health transitions related to disability Logistic and multinomial regression models were used in the analysesIndividuals who were obese were more likely than those of normal weight to have limitations on activities of daily living (ADL)instrumental activity of daily living (IADL) and Nagirsquos limitations Obesity was associated with higher incidence of ADL and IADLlimitations and with lower recovery from Nagirsquos limitations Compared to those who maintained their weight those who gainedweight experienced higher incidence of ADL and Nagirsquos limitations even after controlling for initial body mass index Highermortality among overweight individuals was only found when the reference category was ldquoremaining free of Nagi limitationsrdquoThefindings of the study underline the importance of maintaining normal weight for preventing disability at older ages

1 Introduction

Brazil is among the 25 countries in the world with the fastestaging rates [1 2] In 1950 26 million (49) Brazilians wereolder than 60 years of age and this number has increased to206 million people (108) according to the 2010 census [3]Improvements inmedical care and living standards have beenshown to translate into higher life expectancy In 1950 forexample life expectancy at birth in Brazil was 509 years butthe figure increased to 722 years in 2010 [3] However thenumber of disabled people is expected to increase in the com-ing years given the rapid growth rate of the elderly popula-tion and the rise in the prevalence of obesity and chronic dis-eases [4]

Fast changes in the populationrsquos nutritional intake thathave occurred in Brazil in recent decades [5] have resulted inan increase in the prevalence of obesity [6 7] In the past threedecades obesity rates in Brazil tripled amongmen and almostdoubled among women [6] With the exception of higher-income urban women [8] the prevalence of obesity is higheramong women than among men

Few studies focus on the impact of BMI on mortality anddisability in the Latin American and the Caribbean (LAC)region Based on the baseline for the Health Well-Being andAging in Latin America and the Caribbean Study (SABE) AlSnih and colleagues [4] showed that obese individuals were16 times more likely to face difficulties performing activitiesof daily living (ADL) than thosewith normal BMI [4] Corona

2 Journal of Aging Research

and colleagues [9] found that older adults who were under-weight (BMIle 23 kgm2) and obese (BMIge 30 kgm2) weremore likely to develop limitations on instrumental activityof daily living (IADL) but their analysis did not exploreBMI associations with mortality or recovery from disabilityMonteverde and colleagues [10] found that based on relativecutoffs (quintiles) heavier older adults in Mexico face highermortality risks than those in the United States Howeverwhen BMI was categorized following the traditional WorldHealth Organization cutoffs no excess mortality was foundamong overweight and obese subjects [10] In fact coeffi-cients for overweight and obesity were not in the expecteddirection [10] However none of these studies jointly exam-ined the association between BMI disability transitions andmortality This paper addresses this gap in the literature Weexamine the association between bodymass index andweightchanges on disability transitions and mortality

11 Association between BMI andDisability Obesity has beenassociated with higher prevalence of disability in cross-sec-tional and longitudinal studies [4 11ndash13] This positive asso-ciation has been found among middle- and older-aged adults[13ndash15] and it appears that this association has not changedover time [16 17] Additionally the association between bodymass index (BMI) and disability is strongest among thosewho are underweight (BMIlt 185 kgm2) and among obesesubjects (BMIgt 30 kgm2) [4 11] Obese women face higherprevalence of mobility impairment than men [13] In theUnited States severe functional limitations are higher amongolder adults who gain or lose weight after age 50 compared tothosewith stable weight [18] Emerging evidence supports theproposition that BMI is an important predictor of the onsetof mobility limitations [13] Older adults in the United Stateswho gain weight over time have higher incidence of mobilitylimitations than those who maintain their weight [19]

12 Association between BMI andMortality Large systematicreviews have shown that the relationship between BMI andmortality seems to follow a J-shaped (sometimes U-shaped)curve [17 20ndash22] All-cause mortality appears to be lowest atBMI levels between 20 and 249 kgm2 [20 22] with obeseindividuals facing higher mortality risks than their normal-weight counterparts [20 22 23] Recent studies howevershow that overweight individuals have higher life expectancythan individuals of normal weight or those who are obese[12] whereas other studies have an L-shape (see [24] for ashort review) Higher mortality at low levels of BMI has beenassociated primarily with lung cancer and respiratory dis-eases [22] At older ages results from a systematic review andmeta-analysis indicated that BMI in the overweight range of25ndash299 kgm2 is not associatedwith increasedmortality [25]whereas other studies have shown that the burden of obesityon mortality seems to be reduced or eliminated among olderadults [11 12 26ndash30]

The use of BMI categories has been criticized for not re-producing well the complexities of the BMI and mortalityrelationship [31ndash33] These authors have suggested that theuse of alternativemodels to assess this association Gronniger

[33] used semiparametric models and found that men in themild-obese category (BMIs of 30ndash35 kgm2) had similarmor-tality than those of normal weight but among women BMIlevels above 27 kgm2 were associated with higher mortalityin the US Wong and colleagues [31] used multivariable frac-tional polynomials to explore the association between BMIandmortality in a sample of adults in the UAThey found thatthe best fitting model contained the powers minus1 and minus2 forBMI [31] Their results indicated that the nadir of the BMImortality curve was in the normal range for women but over-weight range for men [31] Zajacova and Burgard [32] usedgeneralized additive models and found that the nadir was atBMIs 23 to 26 which is also in the normal overweight rangeHowever they point out that there are important differencesdepending on the cause of death For example the associa-tion between BMI and diabetesmortality increasesmonoton-ically but for all-cause mortality it followed more a V-shape[32] However even though these alternative models oftenprovide better fit to the data the results of these studies areoften interpreted making references for traditional cutoffs asthey are more easily understood by the general audience andhealth practitionersThe use of BMI and BMI categories havealso been criticized because it can be related to underlyinghealth status [17] For example individuals may be under-weight based on their BMI because they have health condi-tions such as cancer thyroid problems infectious or digestivediseases that lead to low body weight One approach toaddress this limitation has been to take into account bodyweight changes [17]

Our study uses data a large cohort study conducted in SaoPaulo Brazil to examine the association between BMI andbody weight changes on disability transitions and mortalitywhile controlling for a series of demographic socioeconomicand health determinantsWe investigate these associations onthree types of disability (activities of daily living instrumentalactivities of daily living and Nagirsquos limitations) transitions

2 Materials and Methods

Data from the two waves (2000 and 2006) of the SABEcohort study conducted in Sao Paulo Brazil were used in thisstudy SABE is amulticenter surveywith respondents in sevenmajor cities throughout LAC countries that have been inves-tigating the health and well-being of older adults (age 60 andover) The study was approved by the Institutional ReviewBoards at the collaborating institutions [34 35] and the parti-cipants provided consent to have their data used for researchpurposes

The baseline sample was obtained using a two-stage strat-ified sampling based on the 1995 National Household Surveymaster sampling frame The data in the first wave were col-lected in two stagesThe first stage was a household interviewconducted by a single interviewer using a standardized ques-tionnaire that included several questions about the livingconditions and health status of the subjects The second stageof data collection consisted of a household visit by a pair ofinterviewers who completed anthropometric and physical-performance measurements At baseline the response rates

Journal of Aging Research 3

reached 846 in Sao Paulo In the first stage informationon 2143 individuals was collected Additional characteristicsof the baseline data collection process have been describedelsewhere [36ndash38]

In 2006 to reestablish contact trained interviewers vis-ited the addresses and neighborhoods of surviving partici-pants from the 2000 survey For those not found during thesevisits interviewers used the additional contact informationcollected at baseline (eg telephone numbers of children orother relatives) to obtain further information about the sub-jectsrsquo current location In 2006 researchers collected data viaface-to-face interviews using a standardized questionnaireThe 2006 questionnaire was very similar to the one used in2000 but included additional questions that complementedthe previous study Vital statistics records were used toidentify subjects who had died between 2000 and 2006 Thesearch was based on the names sex dates of birth andaddresses listed in the 2000 database

Of the 2143 participants in the first wave of SABE SaoPaulo 355 (166) had missing data on selected variablesMost of them (119899 = 347) had missing data on BMI measureThose with missing data were older (751 years) than thosewith complete data (729 years) (119875 = 00001) but there wereno sex differencesThe prevalence of all measures of disabilitywas higher among those with missing data (119875 lt 0001) Thefinal sample is composed of 1788 individuals with a subset of961 with weight change included in the analyses There were473 participants who died between the baseline and thefollowup in 2006

21 Measures Self-reported disability in six ADL measures(dressing bathing eating getting in and out of a bed toilet-ing and getting across a room) were used to measure dis-ability Individuals were given the following introductionldquoHere are a few everyday activities Please tell me if you haveany difficulty with these because of a health problem Excludeany difficulties you expect to last less than three monthsrdquoAfter this introduction they were asked ldquoDo you have dif-ficulty rdquo And the possible answers were ldquoyesrdquo ldquonordquo ldquodoesnot knowrdquo and ldquono responserdquo for each one of the six ADLmeasures Participants who answered ldquodoes not knowrdquo andldquono responserdquo were classified as missing IADL questions fol-lowed the ADL ones No additional introduction was madeIndividuals were asked ldquoDo you have difficulty rdquoThe IADLitems included were preparing a hot meal managing moneyshopping using of transportation within the communityability to use the telephone and responsibility for onersquos ownmedications The possible answers were ldquoyesrdquo ldquonordquo ldquocannotdo itrdquo ldquodoes not knowrdquo and ldquono responserdquo Those whoanswered ldquocannot do itrdquo were classified as having difficultyperforming the activity whereas those answering ldquodoes notknowrdquo and ldquono responserdquo were classified as missingTheNagiphysical performance measure included lifting or carryingobjects that weighed five kilograms or more lifting a coinpulling or pushing a large object such as a living-room chairstooping kneeling or crouching and reaching or extend-ing the arms above shoulder level Each of the three dis-ability measures was converted into binary form in which

respondents scored ldquo0rdquo if they did not indicate any limitationsand ldquo1rdquo if they reported having difficulty performing one ormore activity in the scale

Bodyweight andheightweremeasuredwithout shoes andwith light clothing by trained examiners BMI was calculatedas kgm2 Four BMI categories were defined according to thecriteria adopted by the Pan American Health Organizationfor the SABE study [38] underweight (BMIle 230) normal(BMIgt 23 and lt 28) overweight (BMIge 280 and lt 30) andobese (BMIge 30) Change in BMI was calculated as BMI in2006 minus the BMI at baseline This difference was dividedby the baseline BMI and then recoded into three categories(a) an increase of 5 or more (b) a decrease of 5 or moreand (c) changes within 5 of the baseline weight (referencecategory) [19 39]

The following sociodemographic characteristics wereincluded in the analysis age (in years) gender smoking status(never former or current smoker) schooling (in years offormal education) and household arrangement (living aloneor accompanied) All regression analysis also included amea-sure of number of chronic conditions at the baseline Healthstatus based on the number of self-reported chronic condi-tions included diabetes hypertension cardiovascular diseasestroke cancer arthritis and osteoporosis

STATA SE 121 forWindows (StataCorp College StationTX)was used for all the statistical analyses Descriptive statis-tics were conducted first Weighted logistic regressions werethen used to assess the influence of BMI on disability preva-lenceMultinomial logistic regressionswere used to assess theinfluence of BMI on disability transitions and mortality Forthose free of disability four outcomes were considered in themultinomial logistic regressions remained free of disability(reference category) became disabled (incidence) died orwere lost to followup For thosewhowere disabled at the base-line four outcomes were included in the multinomial logisticregressions remained disabled (reference category) recov-ered from disability died or were lost to followup Multi-nomial logistic regressions were used to analyze the roleof weight change on health transitions as discussed aboveexcludingmortality as we do not have information on weightchange prior to death in between waves

In the baseline there were 1420 individuals free of ADLand 368 individuals with ADL In 2006 among those free ofADL 606 individuals had remained free of ADL 226 haddeveloped ADL 329 had died and 259 were lost in the fol-lowup or had missing data on ADL status Among those whohadADL in the baseline 99 remainedwithADL 75 recoveredfrom ADL 144 died and 50 were lost in the followup or hadmissing data on ADL in 2006 For IADL limitations 1207were free of IADL and 581 had IADL in the baseline Amongthose who were free of IADL in the baseline 491 remainedfree of limitations 257 developed IADL 230 had died and229 were lost in the followup or had missing data on IADL inthe secondwave Among thosewith IADL in the baseline 220remained with IADL 36 recovered from IADL 243 had diedand 82 were lost in the followup or hadmissing data on IADLstatus in 2006 Regarding the Nagi 654 participants were freeof it in 2000 and 1134 had at least one Nagi limitationAmong those free of Nagi 192 remained free of it 210


developed Nagi 129 died and 123 were lost in the followup orhad missing data on the Nagi variable in 2006 Among thosewho had at least one Nagi limitation in 2000 539 remainedwith Nagirsquos limitations 70 recovered from Nagirsquos limitations344 died and 181were lost in the followup or hadmissing dataon Nagirsquos limitations in the second wave

3 Results

In the final sample 234 were underweight 433 had nor-mal weight 124 were overweight and 211 of the partici-pants were classified as obese Table 1 presents the prevalenceestimates of disability according to measures of ADL IADLand Nagirsquos limitations by sex and BMI category at baselinePrevalence of ADL and Nagirsquos limitations was highest amongobese individuals whereas prevalence of IADL was highestamong underweight older adults Weighted estimates indi-cated that 167 of the sample in Sao Paulo had difficulty per-forming at least oneADL Prevalence of IADL reached 244andmost (578) of the older Brazilian adults reported Nagirsquoslimitations In logistic regressions after adjusting for age andsex (not shown) individuals who were underweight did notdiffer from those of normal weight on their likelihood ofreporting having ADL IADL or Nagi in the baseline Obeseindividuals were more likely than normal weight participantsto report having at least one ADL (OR= 18 95 CI = 1226) andNagirsquos limitations (OR = 25 95CI = 18 36)Therewere no statistical differences between normal weight andobese participants regarding IADL prevalence at baselineThere were no statistical differences between normal andoverweight subjects regarding ADL and IADL prevalence atbaseline but overweight individuals were more likely thanthose of normal weight to have Nagirsquos limitations (OR = 1695 CI = 11 22) Women were more likely than men toreport having ADL IADL and Nagirsquos limitations at baseline(119875 lt 005)

Table 2 shows the multinomial logistic regression resultsof the disability transitions and mortality between 2000 and2006 among those who were free of disability in the baselineCompared to normal weight individuals obese individualswere more likely to develop ADL (RRR= 21) and IADL(RRR= 24) whereas individualswhowere underweightweremore likely to develop IADL (RRR= 19)Mortality risks werehigher among those who were overweight (RRR= 25) com-pared to those of normal weight in the Nagi model in whichthe reference category was remaining free of Nagirsquos limita-tions For all measures of disability the risk of becomingdisabled increased with age As expected older age wasassociated with higher mortality Women were more likely todevelop ADL and Nagirsquos limitations but not IADL betweenwaves Women had lower mortality in the ADL and IADLmodels Higher number of chronic conditions was associatedwith higher mortality and higher incidence of ADL andIADL

Table 3 shows the multinomial logistic regression resultsof the disability transitions and mortality between 2000 and2006 among those who had disability in the baseline Beingobese was also associated with lower recovery from Nagi

(RRR = 046) versus remaining with at least one Nagi limita-tionOlder age andhigher number of chronic conditionswerenegatively associated with recovery

In the last set of analyses we focus on the role of weightgain between waves on disability transitions (Table 4) Theanalyses are restricted to those who have survived betweenwaves The results presented in Table 4 indicate that thosewho gained weight between waves were more likely todevelop ADL (RRR= 23) and Nagirsquos limitations (RRR= 22)than those who maintained their weight even after control-ling for initial BMI categories and other covariates Higherage was also associated with higher incidence of disabilityWomen faced higher incidence of ADL (RRR= 18) and Nagi(RRR= 24) than men Obesity was associated with higherincidence of ADL and IADL Underweight individuals weremore likely to develop IADL Individuals with more chronicconditions also faced higher incidence of ADL and IADLlimitations When the analyses focused on those who haddisability in the baseline we found that weight gain wasassociatedwith lower recovery fromADL (RRR= 018)Olderage was negatively associated with recovery from disabilityA higher number of chronic conditions were associated withlower recovery of ADL and Nagi Obesity was negativelyassociated with recovery from Nagirsquos limitations

4 Discussion

Most previous studies have focused on the association bet-ween BMI and disability [4 9 13 15 16] or BMI andmortality[10 20ndash29 31ndash33] but few have analyzed the effect of BMIon both disability and mortality [11 12 14 30] Using threedisability measures and data from a large cohort study thisstudy contributes to the literature by exploring the impactof BMI and weight changes on disability status transitionsand on mortality This study confirmed the negative effectsof obesity on disability in Sao Paulo Brazil Higher levels ofNagirsquos limitations were also found among those who wereoverweight at baseline Most longitudinal studies have foundthat obese older adults are more likely to have experiencedincidence of disability in the followup than those of normalweight [13 14 40] and our study confirmed these findingsOlder adult Brazilians who were obese at baseline facedhigher risks of becoming disabled with ADL or IADL limita-tions compared to those of normal weight However beingoverweight was not associated with higher incidence ofdisability after controls were included in the analysis whichis consistent with previous findings [41] In terms of recoverywe also found that obese individualswere less likely to recover(versus remaining disabled in the followup) as other studieshave also found [40]

There is growing interest in the role of weight changeson health transitions [13 18 19 42] Studies have shown thatweight gain in older adults is associatedwith decreased physi-cal function and role limitations [18 19] We found similarfindings in which older adults who gained weight betweenwaves weremore likely to develop ADL andNagirsquos limitationsthan those who maintained their weight even after control-ling for initial BMI categories Al Snih and colleagues [19]


Table 1 Prevalence of ADL IADL and Nagirsquos limitations by sex and BMI categories Sao Paulo 2000 (weighted estimates)

Total Underweight Normal Overweight Obese 119875

Total 119899 = 1788 119899 = 419 119899 = 775 119899 = 217 119899 = 377

ADL 167 147 139 182 234 lowastlowast

IADL 244 289 210 250 267 lowastlowast

Nagi 578 527 507 615 749 lowast lowast lowast

Females 119899 = 1062 119899 = 212 119899 = 415 119899 = 130 119899 = 305


IADL 305 365 267 338 304Nagi 672 597 605 739 783 lowast lowast lowast

Males 119899 = 726 119899 = 207 119899 = 360 119899 = 87 119899 = 72

ADL 123 145 103 134 157IADL 156 199 146 139 127Nagi 441 444 396 459 618 lowastlowast

ADL activities of daily living IADL instrumental activities of daily livinglowastlowastlowastP lt 0001 lowastlowastP lt 005 lowastP lt 010

also reported higher ADL incidence among individuals whohad weight gain of more than 5 between waves Studieshave reported contradictory findings related to weight lossSome studies have indicated that weight loss is associatedwith improvements inmobility and functioning [13] whereasothers have reported increased ADL disability [19] Weight-loss therapy among obese older adults seems to be beneficialfor improving quality of life and physical functioning [43]Ritchie and colleagues [42] found that intentional weightloss was not associated with functional decline howeverthose who unintentionally lost weight faced higher rates offunctional decline regardless of the initial BMI In our studywe found no differences between those who lost weight andthose who maintained weight on disability transitions aftercontrolling for initial BMI Given the lack of data on intentrelated to weight changes further studies are necessary toexplore the impact of weight change on mortality and dis-ability in Latin American countries

In additional analyses (not shown and available uponrequest) we have explored additional models to test whetherBMI and weight changes influence changes in the number ofdisabilities over time We found that obesity was associatedwith the increases in the number of Nagirsquos limitationsWeightloss and weight gain were associated with an increase in thenumber of ADL and Nagirsquos limitations over time Changes inthe number of IADL limitations were not statistically asso-ciated with BMI categories or weight changes As expectedolder age was associated with the increases in the numberof ADL IADL and Nagirsquos limitations over time A highernumber of chronic conditions were also associated with anincrease in the number of ADL IADL and Nagirsquos limitationsover time Being female was also positively associated withincreases in the number of Nagi limitations We also testedfractional polynomial models following the approach sug-gested by Wong and colleagues [31] to examine the relation-ship between BMI disability and mortality (results availableupon request) and our main conclusions remain the samewhich indicates that findings are robust to different modelspecifications

The only mortality differential by BMI categories wasfound among overweight participants who were more likelyto die than to remain free of Nagirsquos limitations In furtheranalyses (not shown) results from a logistic regression thatcontrolled for the same covariates included in this studyrevealed no differences in mortality among underweightnormal weight overweight and obese participants This isconsistent with previous studies suggesting that the associ-ation between BMI and mortality becomes less U-shaped atolder ages [44] and others that suggest that higher BMI maynot be detrimental for mortality at older ages [12] Mon-teverde and colleagues [10] also did not find statistical differ-ences in mortality among older adults between higher BMIcategories (overweight and obese) and normal subjects whenusing traditional BMI cutoffs though they reported statisticaldifferenceswhenBMIwas categorized in relative termsHow-ever some studies allude to the fact that the associationbetween BMI and mortality is differential between individu-als who are healthy versus those with chronic conditions [45]The obesity paradox literature indicates that excess weight isactually protective among patients with chronic disease [45]In our sample there were no mortality differentials by BMIcategories among those with chronic conditions but over-weight individuals free of chronic conditions had lower mor-tality than those of normal weight (results available uponrequest)

Our findings also contribute to a growing debate in thefield about whether greater life expectancy implies betterhealth for the expanding surviving elderly female populationin Latin America [37 46ndash51]We found that Brazilian womenexperience higher levels of disability than men which is con-sistent with previous studies [52 53] Previous studies haveindicated that Brazilian women face lower mortality thantheirmale counterparts [37 54] and this study confirms thesefindings

Aging is related to the increase of fat mass and there isgrowing evidence of the detrimental impact of obesity on dis-ability at older ages There is evidence as well that changes inlifestyle such as walking have positive effects on preventing


Table 2 Relative risk ratios of the impact of body mass index categories on disability transitions and mortality among those who were freeof disability in the baseline Sao Paulo Brazilmdash2000ndash2006

Variables ADL119875

IADL119875

NAGI119875

RRRa 95 CI RRR 95 CI RRR 95 CIIncidence of disabilityAge 110 [107 113] lowast lowast lowast 111 [108 114] lowast lowast lowast 105 [101 109] lowastlowastFemale 170 [106 273] lowast 148 [086 254] 237 [140 402] lowastlowastSmoking status

Former smoker 094 [060 146] 115 [068 194] 136 [080 232]Current smoker 158 [088 285] 152 [073 316] 093 [052 165]

Number of chronic conditions 156 [133 183] lowast lowast lowast 133 [112 158] lowastlowast 123 [094 162]Schooling 095 [086 104] 091 [082 100] 088 [079 098] lowastLive alone 081 [050 131] 061 [035 106] 093 [039 224]BMI categories

Underweight 116 [066 201] 192 [121 303] lowastlowast 103 [056 189]Overweight 093 [049 175] 157 [084 296] 186 [089 387]Obese 207 [121 357] lowastlowast 242 [165 353] lowast lowast lowast 119 [063 226]

MortalityAge 111 [108 114] lowast lowast lowast 113 [110 116] lowast lowast lowast 111 [107 116] lowast lowast lowastFemale 060 [038 095] lowast 041 [025 065] lowast lowast lowast 068 [034 136]Smoking status

Former smoker 125 [082 190] 117 [067 203] 090 [046 177]Current smoker 289 [166 503] lowast lowast lowast 269 [136 533] lowastlowast 268 [113 637] lowast

Number of chronic conditions 140 [116 168] lowast lowast lowast 141 [117 169] lowast lowast lowast 182 [125 264] lowastlowastSchooling 095 [089 102] 094 [087 102] 093 [082 107]Live alone 101 [062 163] 116 [064 212] 136 [066 279]BMI categories

Underweight 121 [079 185] 122 [071 209] 187 [095 366]Overweight 130 [078 214] 141 [080 248] 250 [101 615] lowastObese 107 [065 175] 129 [077 218] 068 [022 207]

Lost to followup or missing disability status in 2006Age 103 [100 105] lowast 105 [102 108] lowast lowast lowast 103 [099 108]Female 138 [099 194] 124 [084 183] 219 [137 352] lowastlowastSmoking status

Former smoker 136 [090 206] 154 [100 238] 201 [110 370] lowastCurrent smoker 105 [060 184] 102 [053 198] 132 [061 289]

Number of chronic conditions 112 [096 131] 112 [094 134] 122 [094 158]Schooling 102 [092 113] 101 [091 112] 099 [085 116]Live alone 137 [084 222] 142 [084 240] 093 [038 230]BMI categories

Underweight 158 [097 259] 207 [124 345] lowastlowast 144 [070 298]Overweight 125 [069 226] 133 [078 226] 141 [059 335]Obese 136 [087 212] 155 [102 237] lowast 132 [056 314]119873 1420 1207 654BIC1015840 minus10455 minus82451 4773ADL activities of daily living IADL instrumental activities of daily living RRR relative risk ratio CI confidence interval BMI body mass indexaRemaining free of disability is the reference category Normal weight is the reference category for BMI living accompanied is the baseline category forhousehold arrangement and never smoked is the reference category for smoking statuslowastlowastlowast119875 lt 0001 lowastlowast119875 lt 005 lowast119875 lt 010


Table 3 Relative risk ratios of the impact of bodymass index categories on disability transitions andmortality among thosewho had disabilityin the baseline Sao Paulo Brazilmdash2000ndash2006

Variables ADL119875

IADL119875

NAGI119875

RRRa 95 CI RRR 95 CI RRR 95 CIRecovery from disabilityAge 093 [088 097] lowastlowast 092 [087 098] lowastlowast 096 [092 100] lowastFemale 065 [025 173] 039 [013 117] 056 [030 107]Smoking status


Number of chronic conditions 072 [052 099] lowast 093 [068 127] 059 [044 079] lowast lowast lowastSchooling 103 [083 127] 099 [076 128] 105 [089 125]Live alone 055 [017 174] 111 [031 392] 324 [159 657] lowastlowastBMI categories

Underweight 037 [010 134] 014 [001 145] 062 [025 158]Overweight 101 [035 297] 080 [022 288] 082 [029 228]Obese 048 [021 108] 077 [028 217] 046 [022 097] lowast

MortalityAge 112 [107 118] lowast lowast lowast 107 [104 110] lowast lowast lowast 110 [108 112] lowast lowast lowastFemale 040 [018 090] lowast 065 [030 138] 048 [030 077] lowastlowastSmoking status

Former smoker 133 [057 309] 175 [092 334] 175 [113 269] lowastCurrent smoker 335 [106 1064] lowast 309 [135 707] lowastlowast 230 [129 409] lowastlowast


Underweight 097 [044 212] 128 [074 221] 09 [057 144]Overweight 035 [011 114] 075 [041 139] 079 [045 138]Obese 054 [022 135] 066 [036 119] 078 [051 122]

Lost to followup or missing disability status in 2006Age 100 [094 106] 100 [096 104] 101 [097 104]Female 050 [018 139] 140 [048 408] 093 [054 160]Smoking status


Number of chronic conditions 073 [053 101] 088 [066 116] 088 [076 103]Schooling 116 [097 140] 107 [092 124] 103 [093 115]Live alone 173 [053 557] 143 [068 301] 238 [134 424] lowastlowastBMI categories

Underweight 167 [047 589] 142 [072 282] 185 [109 314] lowastOverweight 043 [012 154] 090 [036 227] 127 [063 253]Obese 092 [032 263] 114 [049 265] 124 [075 204]119873 368 581 1134BIC1015840 3033 6247 minus4291ADL activities of daily living IADL instrumental activities of daily living RRR relative risk ratio CI confidence interval BMI body mass indexaRemaining with disability is the reference category Normal weight is the reference category for BMI living accompanied is the baseline category for householdarrangement and never smoked is the reference category for smoking statuslowastlowastlowast119875 lt 0001 lowastlowast119875 lt 005 lowast119875 lt 010


Table 4 Relative risk ratios of the impact of body mass index categories and body mass index changes on disability transitions Sao PauloBrazilmdash2000ndash2006

Variables ADL119875

IADL119875

NAGI119875

RRRa 95 CI RRR 95 CI RRR 95 CIIncidence of disability (reference = remain free of disability)Age 110 [107 113] lowast lowast lowast 111 [108 115] lowast lowast lowast 105 [101 109] lowastFemale 175 [104 294] lowast 153 [088 265] 244 [144 412] lowastlowastSmoking status


Number of chronic conditions 156 [132 186] lowast lowast lowast 133 [112 159] lowastlowast 125 [092 168]Schooling 097 [088 107] 092 [082 103] 090 [080 101]Live alone 085 [050 144] 060 [034 108] 101 [042 241]BMI categories

Underweight 092 [048 177] 173 [105 284] lowast 093 [049 179]Overweight 072 [036 144] 139 [067 288] 187 [087 399]Obese 199 [110 357] lowast 238 [161 352] lowast lowast lowast 122 [066 227]

BMI changeLoss 123 [076 200] 099 [062 160] 085 [049 148]Gain 230 [103 512] lowast 197 [097 401] 215 [120 385] lowast119873 800 737 389BIC1015840 2746 940 2856Recovery from disability (reference = remain with disability)Age 092 [087 098] lowast 093 [088 099] lowast 095 [091 099] lowastFemale 043 [013 147] 041 [011 158] 053 [028 101]Smoking status




BMI changeLoss 052 [023 119] 053 [021 135] 107 [061 189]Gain 018 [005 068] lowast 064 [015 273] 053 [021 137]119873 161 224 572BIC1015840 3187 9145 8522ADL activities of daily living IADL instrumental activities of daily living RRR relative risk ratio CI confidence interval BMI body mass indexaRelative risk ratios were adjusted by smoking status Remaining free of disability is the reference category for those who started without disability andremaining with disability is the reference category for those who had disability in the baseline Normal weight is the reference category for BMI Stable weightis the baseline category for weight change Results for lost in the followup were omitted (available upon request)lowastlowastlowast119875 lt 0001 lowastlowast119875 lt 005 lowast119875 lt 010

mobility limitations [55] A large proportion of older adultshowever do not engage in physical activity In a study basedon an urban sample in Brazil for example 71 of older adultsreported living sedentary lives [56]When asked about neigh-borhood characteristics related to concerns of leaving hometo go out most (78) reported fear of being robbed whilealmost half (482) said that they were afraid of fallingbecause of sidewalk defects [56] Fear of falling due to poor

sidewalk conditions was associated with a 62 increase inthe expected number of ADL conditions [56] Thereforeinvestments aimed at improving urban infrastructure andsafety may be effective in addressing the health conditions ofolder adults in Brazil

This study advances the literature on the impact of bodyweight and body weight changes on disability and mor-tality This study however has some limitations First the


data used in the study on disability measures were self-re-ported Although this could be a possible source of biasmethodological studies have shown that self-reported dataon functional disability are consistent withmedical diagnoses[57] Second the use of BMI as a measure for body weightcomposition among older adults is very controversial as itdoes not take into account body fat distribution [17] In addi-tion BMI at baseline can be associated with health status [1617] Therefore it is important to control for weight changeswhich we accomplished in this study Some authors haveargued that waist circumference or waist-to-hip ratio couldbe better predictors of disability and mortality [10] howevermost studies to date have focused on the use of BMI and thecategories used here Other scholars have indicated that atleast for developed countries information on BMI waist cir-cumference or waist-to-hip ratio do not necessarily improveprediction ofmortality due to cardiovascular disease insteadthey suggest using information on systolic blood pressurediabetes status and lipids when those are available [58]And still others argue thatmdashin addition to BMImdashwaist cir-cumference and waist-to-hip ratio can be useful in betterunderstanding mortality risks [59 60] In Brazil as in otherdeveloping countries data on blood pressure and lipids areoften lacking so the use of anthropometric measures such aswaist circumference may improve our understanding of theimpact of body composition changes on mortality and dis-ability Third the first wave of SABE focuses on the civilianpopulation not residing in institutions As a result estimatesmay be biased if one expects institutionalized individualsparticularly those residing in nursing homes to be morelikely to have a higher prevalence of disability than the non-institutionalized population However because the institu-tionalized population in Brazil is relatively small [61] thispossible bias is likely not to be very significant

5 Conclusion

This study confirms previous studies that have found obesityto be associated with increased disability in Brazilian olderadults Historically Brazil has mainly been concerned withcurbing malnutrition however in recent years new policieshave targeted the marketing of highly processed and un-healthy foods [5] Owing to the fact that obesity rates in Brazilhave been increasing drastically for the past three decades [6]our findings have important implications for policymakers inBrazil with regard to curbing disability risk by promoting theuse of effective preventive measures to reduce body weightthereby making healthy aging a reality

Authorrsquos Contribution

F C D Andrade planned the study supervised the data anal-ysis andwrote the paper A I NMNazan contributed to thepaper writing M L Lebrao and Y A de O Duarte collectedthe data helped plan the study including instrumentationand revised the paper

Acknowledgments

This study was supported by grants from FAPESPBrazil theBrazilian Ministry of Health and the Lemann Institute forBrazilian Studies at the University of Illinois at Urbana-ChampaignThe authors gratefully acknowledge the feedbackreceived at the 2012 Population Association of AmericaAnnualMeetingThey also thank FernaoDias de Lima for thecareful management of the database

References

[1] D Cotlear Population Aging Is Latin America Ready WorldBank Washington DC USA 2011

[2] K Kinsella and W He An Aging World 2008 GovernmentPrinting Office Washington DC USA 2009

[3] Population Division of the Department of Economic and SocialAffairs of the United Nations SecretariatWorld Population Pro-spects The 2010 Revision United Nations New York NY USA2011

[4] S Al Snih J E Graham Y F Kuo J S Goodwin K SMarkidesand K J Ottenbacher ldquoObesity and disability relation amongolder adults living in latin america and the caribbeanrdquoAmericanJournal of Epidemiology vol 171 no 12 pp 1282ndash1288 2010

[5] BM Popkin L S Adair and SWNg ldquoGlobal nutrition transi-tion and the pandemic of obesity in developing countriesrdquoNutrition Reviews vol 70 no 1 pp 3ndash21 2012

[6] C AMonteiroW L Conde and BM Popkin ldquoIncome-speci-fic trends in obesity in Brazil 1975ndash2003rdquo American Journal ofPublic Health vol 97 no 10 pp 1808ndash1812 2007

[7] C A Monteiro W L Conde and B M Popkin ldquoThe burdenof disease from undernutrition and overnutrition in countriesundergoing rapid nutrition transition a view from BrazilrdquoAmerican Journal of Public Health vol 94 no 3 pp 433ndash4342004

[8] C A Monteiro W L Conde and B M Popkin ldquoIndependenteffects of income and education on the risk of obesity in theBrazilian adult populationrdquo Journal of Nutrition vol 131 no 3pp 881Sndash886S 2001

[9] L P Corona T R Pereira de Brito D P Nunes et al ldquoNutri-tional status and risk for disability in instrumental activities ofdaily living in older Braziliansrdquo Public Health Nutrition vol 13pp 1ndash6 2012

[10] MMonteverde K Noronha A Palloni and B Novak ldquoObesityand excess mortality among the elderly in the United States andMexicordquo Demography vol 47 no 1 pp 79ndash96 2010

[11] S Al Snih K J Ottenbacher K S Markides Y F KuoK Eschbach and J S Goodwin ldquoThe effect of obesity ondisability versusmortality in olderAmericansrdquoArchives of Inter-nal Medicine vol 167 no 8 pp 774ndash780 2007

[12] I M Majer W J Nusselder J P MacKenbach and A E KunstldquoLife expectancy and life expectancy with disability of normalweight overweight and obese smokers and nonsmokers inEuroperdquo Obesity vol 19 no 7 pp 1451ndash1459 2011

[13] H K Vincent K R Vincent and K M Lamb ldquoObesity andmobility disability in the older adultrdquo Obesity Reviews vol 11no 8 pp 568ndash579 2010

[14] I A Lang D J Llewellyn K Alexander and D MelzerldquoObesity physical function andmortality in older adultsrdquo Jour-nal of the American Geriatrics Society vol 56 no 8 pp 1474ndash1478 2008


[15] S Larrieu K Peres L Letenneur et al ldquoRelationship betweenbody mass index and different domains of disability in olderpersons the 3C studyrdquo International Journal of Obesity vol 28no 12 pp 1555ndash1560 2004

[16] D E Alley and VW Chang ldquoThe changing relationship of obe-sity and disability 1988ndash2004rdquo Journal of the American Medi-cal Association vol 298 no 17 pp 2020ndash2027 2007

[17] D E Alley L Ferrucci M Barbagallo S A Studenski and T BHarris ldquoA research agenda the changing relationship betweenbody weight and health in agingrdquo Journals of Gerontology SeriesA vol 63 no 11 pp 1257ndash1259 2008

[18] A P Carson D N Holmes and D L Howard ldquoWeight changeand functional limitations among older adults in North Caro-linardquo Journal of Community Health vol 35 no 6 pp 586ndash5912010

[19] S Al Snih M A Raji K S Markides K J Ottenbacher and JS Goodwin ldquoWeight change and lower body disability in olderMexican Americansrdquo Journal of the American Geriatrics Societyvol 53 no 10 pp 1730ndash1737 2005

[20] A B De Gonzalez P Hartge J R Cerhan et al ldquoBody-massindex and mortality among 146 million white adultsrdquo NewEngland Journal ofMedicine vol 363 no 23 pp 2211ndash2219 2010

[21] L C Hwang S C Chen J J Tjung H Y Chiou C J Chen andC H Tsai ldquoBodymass index as a predictor of mortality in olderpeople in Taiwanrdquo International Journal of Gerontology vol 3no 1 pp 39ndash46 2009

[22] Prospective Studies Collaboration ldquoBody-mass index andcause-specific mortality in 900 000 adults collaborative anal-yses of 57 prospective studiesrdquo The Lancet vol 373 no 9669pp 1083ndash1096 2009

[23] A Peeters J J Barendregt F Willekens et al ldquoObesity in adult-hood and its consequences for life expectancy a life-table anal-ysisrdquoAnnals of InternalMedicine vol 138 no 1 pp 24ndash32 2003

[24] A C Tsai and M Hsiao ldquoThe association of body mass index(BMI) with all-cause mortality in older Taiwanese results of anational cohort studyrdquo Archives of Gerontology and Geriatricsvol 55 no 2 pp 217ndash220 2012

[25] I Janssen and A E Mark ldquoElevated body mass index and mor-tality risk in the elderlyrdquoObesity Reviews vol 8 no 1 pp 41ndash592007

[26] R Bender K H Jockel C Trautner M Spraul and M BergerldquoEffect of age on excess mortality in obesityrdquo Journal of theAmerican Medical Association vol 281 no 16 pp 1498ndash15041999

[27] K M Flegal B I Graubard D F Williamson and M H GailldquoExcess deaths associated with underweight overweight andobesityrdquo Journal of the American Medical Association vol 293no 15 pp 1861ndash1867 2005

[28] RHuxley F Barzi andMWoodward ldquoExcess risk of fatal coro-nary heart disease associated with diabetes in men and womenmeta-analysis of 37 prospective cohort studiesrdquo British MedicalJournal vol 332 no 7533 pp 73ndash76 2006

[29] M Reuser L Bonneux and F Willekens ldquoThe burden of mor-tality of obesity at middle and old age is small A life table anal-ysis of the US Health and Retirement Surveyrdquo European Jour-nal of Epidemiology vol 23 no 9 pp 601ndash607 2008

[30] S L Reynolds Y Saito and E M Crimmins ldquoThe impact ofobesity on active life expectancy in older American men andwomenrdquo Gerontologist vol 45 no 4 pp 438ndash444 2005

[31] E Wong B Wang L Garrison et al ldquoExamining the BMI-mortality relationship using fractional polynomialsrdquoBMCMed-ical Research Methodology vol 11 no 1 article 175 2011

[32] A Zajacova and S A Burgard ldquoShape of the BMI-mortalityassociation by cause of death using generalized additive mod-els NHIS 1986ndash2006rdquo Journal of Aging and Health vol 24 no2 pp 191ndash211 2012

[33] J T Gronniger ldquoA semiparametric analysis of the relationshipof body mass index to mortalityrdquo American Journal of PublicHealth vol 96 no 1 pp 173ndash178 2006

[34] A Palloni G Pinto-Aguirre andM Pelaez ldquoDemographic andhealth conditions of ageing in Latin America and the Carib-beanrdquo International Journal of Epidemiology vol 31 no 4 pp762ndash771 2002

[35] R Wong M Pelaez A Palloni and K Markides ldquoSurvey datafor the study of aging in Latin America and the Caribbeanselected studiesrdquo Journal of Aging and Health vol 18 no 2 pp157ndash179 2006

[36] C Albala M L Lebrao E M L Diaz et al ldquoSalud Bienestar yEnvejecimiento (SABE) metodologia de la encuesta t perfil dela poblacion estudiadardquo Pan American Journal of Public Healthvol 17 no 5-6 pp 307ndash322 2005

[37] F C D Andrade P E Guevara M L Lebrao Y A de OliveiraDuarte and J L F Santos ldquoGender differences in life expect-ancy and disability-free life expectancy among older adults inSao Paulo Brazilrdquo Womenrsquos Health Issues vol 21 no 1 pp 64ndash70 2011

[38] M L Lebrao and R Laurenti ldquoSaude bem-estar e envelhec-imento o estudo SABE no municıpio de Sao Paulordquo RevistaBrasileira De Epidemiologia vol 8 no 2 pp 127ndash141 2005

[39] F C D Andrade I Vazquez-Vidal and T Flood ldquoOne-year fol-low-up changes in weight are associated with changes in bloodpressure in young Mexican adultsrdquo Public Health vol 126 no6 pp 535ndash540 2012

[40] S Walter A Kunst J MacKenbach A Hofman and HTiemeier ldquoMortality and disability the effect of overweight andobesityrdquo International Journal of Obesity vol 33 no 12 pp 1410ndash1418 2009

[41] H Dong M Unosson E Wressle and J Marcusson ldquoHealthconsequences associated with being overweight or obese aSwedish population-based study of 85-year-oldsrdquo Journal of theAmerican Geriatrics Society vol 60 no 2 pp 243ndash250 2012

[42] C S Ritchie J L Locher D L Roth T McVie P Sawyer andR Allman ldquoUnintentional weight loss predicts decline in activ-ities of daily living function and life-space mobility over 4 yearsamong community-dwelling older adultsrdquo Journals of Gerontol-ogy Series A vol 63 no 1 pp 67ndash75 2008

[43] D T Villareal CM Apovian R F Kushner and S Klein ldquoObe-sity in older adults technical review and position statement ofthe American Society for Nutrition and NAASO The ObesitySocietyrdquo American Journal of Clinical Nutrition vol 82 no 5pp 923ndash934 2005

[44] M Thinggaard R Jacobsen B Jeune T Martinussen and KChristensen ldquoIs the relationship between bmi and mortalityincreasingly u-shaped with advancing age A 10-year follow-upof persons aged 70ndash95 yearsrdquo Journals of Gerontology Series Avol 65 no 5 pp 526ndash531 2010

[45] J P Curtis J G Selter Y Wang et al ldquoThe obesity paradoxbody mass index and outcomes in patients with heart failurerdquoArchives of Internal Medicine vol 165 no 1 pp 55ndash61 2008

[46] B E Alvarado M V Zunzunegui F Beland M Sicotte andL Tellechea ldquoSocial and gender inequalities in depressivesymptoms among urban older adults of Latin America and theCaribbeanrdquo Journals of Gerontology Series B vol 62 no 4 ppS226ndashS237 2007


[47] F C Andrade ldquoMeasuring the impact of diabetes on life expec-tancy and disability-free life expectancy among older adults inMexicordquoThe Journals of Gerontology Series B vol 65 no 3 pp381ndash389 2010

[48] M C S Camargos C J Machado and R Do NascimentoRodrigues ldquoDisability life expectancy for the elderly city of SaoPaulo Brazil 2000 gender and educational differencesrdquo Journalof Biosocial Science vol 39 no 3 pp 455ndash463 2007

[49] M C S Camargos C J Machado and R N Rodrigues ldquoLifeexpectancy among elderly Brazilians in 2003 according to dif-ferent levels of functional disabilityrdquoCadernos de Saude Publicavol 24 no 4 pp 845ndash852 2008

[50] S Reyes-Beaman C Jagger C Garcia-Pena O Munoz P EBeaman and B Stafford ldquoActive life expectancy of older peoplein Mexicordquo Disability and Rehabilitation vol 27 no 5 pp 213ndash219 2005

[51] C A Reyes-Ortiz G V Ostir M Pelaez and K J OttenbacherldquoCross-national comparison of disability in LatinAmerican andCaribbean persons aged 75 and olderrdquo Archives of Gerontologyand Geriatrics vol 42 no 1 pp 21ndash33 2006

[52] R O Guerra B E Alvarado and M V Zunzunegui ldquoLifecourse gender and ethnic inequalities in functional disabilityin a Brazilian urban elderly populationrdquo Aging vol 20 no 1pp 53ndash61 2008

[53] J L F Santos M L Lebrao Y A O Duarte and F D De LimaldquoFunctional performance of the elderly in instrumental activi-ties of daily living an analysis in the municipality of Sao PauloBrazilrdquo Cadernos de Saude Publica vol 24 no 4 pp 879ndash8862008

[54] M F Lima-Costa S V Peixoto D L Matos J O A Firmoand E Uchoa ldquoPredictors of 10-year mortality in a populationof community-dwelling Brazilian elderly the Bambuı cohortstudy of agingrdquo Cadernos De Saude Publica vol 27 pp s360ndashs369 2011

[55] NM deVries C D van Ravensberg J S HobbelenMGOldeRikkert J B Staal and M W Nijhuis-van der Sanden ldquoEffectsof physical exercise therapy on mobility physical functioningphysical activity and quality of life in community-dwelling olderadults with impaired mobility physical disability andor multi-morbidity ameta-analysisrdquoAgeing Research Reviews vol 11 no1 pp 136ndash149 2012

[56] F R Ferreira C C Cesar V P CamargosM F Lima-Costa andF A Proietti ldquoAging and urbanization the neighborhood per-ception and functional performance of elderly persons in belohorizonte metropolitan area-Brazilrdquo Journal of Urban Healthvol 87 no 1 pp 54ndash66 2010

[57] M V Zunzunegui B E Alvarado F Beland and B VissandjeeldquoExplaining health differences betweenmen andwomen in laterlife a cross-city comparison in Latin America and the Carib-beanrdquo Social Science and Medicine vol 68 no 2 pp 235ndash2422009

[58] D Wormser S Kaptoge E Di Angelantonio et al ldquoSeparateand combined associations of body-mass index and abdominaladiposity with cardiovascular disease collaborative analysis of58 prospective studiesrdquoThe Lancet vol 377 no 9771 pp 1085ndash1095 2011

[59] T PischonH Boeing KHoffmann et al ldquoGeneral and abdom-inal adiposity and risk of death in Europerdquo New England Jour-nal of Medicine vol 359 no 20 pp 2105ndash2120 2008

[60] T Pischon ldquoCommentary use of the body mass index to assessthe risk of health outcomes time to say goodbyerdquo InternationalJournal of Epidemiology vol 39 no 2 pp 528ndash529 2010

[61] A A Camarano and S Kanso ldquoAs instituicoes de longa per-manencia para idosos no Brasilrdquo Revista Brasileira de Estudosde Populacao vol 27 no 1 pp 232ndash235 2010

Submit your manuscripts athttpwwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


and colleagues [9] found that older adults who were under-weight (BMIle 23 kgm2) and obese (BMIge 30 kgm2) weremore likely to develop limitations on instrumental activityof daily living (IADL) but their analysis did not exploreBMI associations with mortality or recovery from disabilityMonteverde and colleagues [10] found that based on relativecutoffs (quintiles) heavier older adults in Mexico face highermortality risks than those in the United States Howeverwhen BMI was categorized following the traditional WorldHealth Organization cutoffs no excess mortality was foundamong overweight and obese subjects [10] In fact coeffi-cients for overweight and obesity were not in the expecteddirection [10] However none of these studies jointly exam-ined the association between BMI disability transitions andmortality This paper addresses this gap in the literature Weexamine the association between bodymass index andweightchanges on disability transitions and mortality

11 Association between BMI andDisability Obesity has beenassociated with higher prevalence of disability in cross-sec-tional and longitudinal studies [4 11ndash13] This positive asso-ciation has been found among middle- and older-aged adults[13ndash15] and it appears that this association has not changedover time [16 17] Additionally the association between bodymass index (BMI) and disability is strongest among thosewho are underweight (BMIlt 185 kgm2) and among obesesubjects (BMIgt 30 kgm2) [4 11] Obese women face higherprevalence of mobility impairment than men [13] In theUnited States severe functional limitations are higher amongolder adults who gain or lose weight after age 50 compared tothosewith stable weight [18] Emerging evidence supports theproposition that BMI is an important predictor of the onsetof mobility limitations [13] Older adults in the United Stateswho gain weight over time have higher incidence of mobilitylimitations than those who maintain their weight [19]

12 Association between BMI andMortality Large systematicreviews have shown that the relationship between BMI andmortality seems to follow a J-shaped (sometimes U-shaped)curve [17 20ndash22] All-cause mortality appears to be lowest atBMI levels between 20 and 249 kgm2 [20 22] with obeseindividuals facing higher mortality risks than their normal-weight counterparts [20 22 23] Recent studies howevershow that overweight individuals have higher life expectancythan individuals of normal weight or those who are obese[12] whereas other studies have an L-shape (see [24] for ashort review) Higher mortality at low levels of BMI has beenassociated primarily with lung cancer and respiratory dis-eases [22] At older ages results from a systematic review andmeta-analysis indicated that BMI in the overweight range of25ndash299 kgm2 is not associatedwith increasedmortality [25]whereas other studies have shown that the burden of obesityon mortality seems to be reduced or eliminated among olderadults [11 12 26ndash30]

The use of BMI categories has been criticized for not re-producing well the complexities of the BMI and mortalityrelationship [31ndash33] These authors have suggested that theuse of alternativemodels to assess this association Gronniger

[33] used semiparametric models and found that men in themild-obese category (BMIs of 30ndash35 kgm2) had similarmor-tality than those of normal weight but among women BMIlevels above 27 kgm2 were associated with higher mortalityin the US Wong and colleagues [31] used multivariable frac-tional polynomials to explore the association between BMIandmortality in a sample of adults in the UAThey found thatthe best fitting model contained the powers minus1 and minus2 forBMI [31] Their results indicated that the nadir of the BMImortality curve was in the normal range for women but over-weight range for men [31] Zajacova and Burgard [32] usedgeneralized additive models and found that the nadir was atBMIs 23 to 26 which is also in the normal overweight rangeHowever they point out that there are important differencesdepending on the cause of death For example the associa-tion between BMI and diabetesmortality increasesmonoton-ically but for all-cause mortality it followed more a V-shape[32] However even though these alternative models oftenprovide better fit to the data the results of these studies areoften interpreted making references for traditional cutoffs asthey are more easily understood by the general audience andhealth practitionersThe use of BMI and BMI categories havealso been criticized because it can be related to underlyinghealth status [17] For example individuals may be under-weight based on their BMI because they have health condi-tions such as cancer thyroid problems infectious or digestivediseases that lead to low body weight One approach toaddress this limitation has been to take into account bodyweight changes [17]

Our study uses data a large cohort study conducted in SaoPaulo Brazil to examine the association between BMI andbody weight changes on disability transitions and mortalitywhile controlling for a series of demographic socioeconomicand health determinantsWe investigate these associations onthree types of disability (activities of daily living instrumentalactivities of daily living and Nagirsquos limitations) transitions

2 Materials and Methods

Data from the two waves (2000 and 2006) of the SABEcohort study conducted in Sao Paulo Brazil were used in thisstudy SABE is amulticenter surveywith respondents in sevenmajor cities throughout LAC countries that have been inves-tigating the health and well-being of older adults (age 60 andover) The study was approved by the Institutional ReviewBoards at the collaborating institutions [34 35] and the parti-cipants provided consent to have their data used for researchpurposes

The baseline sample was obtained using a two-stage strat-ified sampling based on the 1995 National Household Surveymaster sampling frame The data in the first wave were col-lected in two stagesThe first stage was a household interviewconducted by a single interviewer using a standardized ques-tionnaire that included several questions about the livingconditions and health status of the subjects The second stageof data collection consisted of a household visit by a pair ofinterviewers who completed anthropometric and physical-performance measurements At baseline the response rates













3 Results






4 Discussion






Total 119899 = 1788 119899 = 419 119899 = 775 119899 = 217 119899 = 377




Females 119899 = 1062 119899 = 212 119899 = 415 119899 = 130 119899 = 305



Males 119899 = 726 119899 = 207 119899 = 360 119899 = 87 119899 = 72










Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of




























3 Results






4 Discussion






Total 119899 = 1788 119899 = 419 119899 = 775 119899 = 217 119899 = 377




Females 119899 = 1062 119899 = 212 119899 = 415 119899 = 130 119899 = 305



Males 119899 = 726 119899 = 207 119899 = 360 119899 = 87 119899 = 72










Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















3 Results






4 Discussion






Total 119899 = 1788 119899 = 419 119899 = 775 119899 = 217 119899 = 377




Females 119899 = 1062 119899 = 212 119899 = 415 119899 = 130 119899 = 305



Males 119899 = 726 119899 = 207 119899 = 360 119899 = 87 119899 = 72










Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



















Total 119899 = 1788 119899 = 419 119899 = 775 119899 = 217 119899 = 377




Females 119899 = 1062 119899 = 212 119899 = 415 119899 = 130 119899 = 305



Males 119899 = 726 119899 = 207 119899 = 360 119899 = 87 119899 = 72










Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















Variables ADL119875

IADL119875

NAGI119875















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















Variables ADL119875

IADL119875

NAGI119875















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















5 Conclusion




Acknowledgments


References





































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of






































































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Documents

Research Article The Impact of Body Mass Index …downloads.hindawi.com › journals › jar › 2013 › 905094.pdfmortality risks than those in the United States. However, when BMI