UBC Faculty Research and Publications

Frailty, fitness and late-life mortality in relation to chronological and biological age Mitnitski, Arnold B; Graham, Janice E; Mogilner, Alexander J; Rockwood, Kenneth Feb 27, 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12877_2001_Article_6.pdf [ 301.8kB ]
JSON: 52383-1.0221617.json
JSON-LD: 52383-1.0221617-ld.json
RDF/XML (Pretty): 52383-1.0221617-rdf.xml
RDF/JSON: 52383-1.0221617-rdf.json
Turtle: 52383-1.0221617-turtle.txt
N-Triples: 52383-1.0221617-rdf-ntriples.txt
Original Record: 52383-1.0221617-source.json
Full Text

Full Text

ralBioMed CentBMC GeriatricsBMC Geriatrics 2002, 2Research articleFrailty, fitness and late-life mortality in relation to chronological and biological ageArnold B Mitnitski1, Janice E Graham2, Alexander J Mogilner3 and Kenneth Rockwood*4Address: 1École Polytechnique, Montreal QB, Canada, 2Department of Anthropology and Sociology, University of British Columbia, Vancouver BC, Canada, 3Montreal, QB, Canada and 4Division of Geriatric Medicine, Dalhousie University, Halifax, NS, CanadaE-mail: Arnold B Mitnitski - arnold@grbb.polymtl.ca; Janice J Graham - igraham@interchange.ubc.ca;Alexander E Mogilner - alex.mog@rocketmail.com; Kenneth Rockwood* - rockwood@is.dal.ca*Corresponding authorAbstractBackground: People age at remarkably different rates, but how to estimate trajectories ofsenescence is controversial.Methods: In a secondary analysis of a representative cohort of Canadians aged 65 and over (n =2914) we estimated a frailty index based on the proportion of 20 deficits observed in a structuredclinical examination. The construct validity of the index was examined through its relationship tochronological age (CA). The criterion validity was examined in its ability to predict mortality, andin relation to other predictions about aging. From the frailty index, relative (to CA) fitness andfrailty were estimated, as was an individual's biological age.Results: The average value of the frailty index increased with age in a log-linear relationship (r =0.91; p < 0.001). In a Cox regression analysis, biological age was significantly more highly associatedwith death than chronological age. The average increase in the frailty index (i.e. the averageaccumulation of deficits) amongst those with no cognitive impairment was 3 per cent per year.Conclusions: The frailty index is a sensitive predictor of survival. As the index includes items nottraditionally related to adverse health outcomes, the finding is compatible with a view of frailty asthe failure to integrate the complex responses required to maintain function.BackgroundAlthough the prevalence of both illness and functionalimpairment rise with age, individuals with same chrono-logical age vary widely in health and function [1]. Howbest to summarize this variability in impairments is notclear. While functional disabilities tend to follow a hierar-chical pattern, [2–4] summarizing disabilities without ref-oretical base, such as frailty [1] and allostatic load [5],have been proposed as a better means of assessing the het-erogeneity of health status amongst elderly people, butnone has yet proved entirely satisfactory [6–9].We recently proposed that the concept of functional agecan be derived from a representative database which in-Published: 27 February 2002BMC Geriatrics 2002, 2:1Received: 14 December 2001Accepted: 27 February 2002This article is available from: http://www.biomedcentral.com/1471-2318/2/1© 2002 Mitnitski et al; licensee BioMed Central Ltd. Verbatim copying and redistribution of this article are permitted in any medium for any purpose, provided this notice is preserved along with the article's original URL.Page 1 of 8(page number not for citation purposes)erence to the associated illnesses omits importantinformation. Broader concepts, with a more rigorous the-cludes information on a range of variables [10]. Earlier,we identified a constellation of signs and symptoms fromBMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1this database which show synergy (close inter-relatednessin complex patterns) in health and loss of synergy in neu-rodegenerative disease [11]. Recalling that both advancedage and the frail state are associated with loss of complex-ity and resiliency and, ultimately, with death [7,12–15]we now extend these observations to propose that, as afirst approximation, an individual's health status can bequantified as the proportion of ageing associated deficitswhich they have incurred. Further, we suggest that bothfitness and frailty can be estimated by comparing thenumber of symptoms and signs (jointly referred to hereinas deficits) which are present in an individual with themean number of deficits present in others of the samechronological age.Materials and methodsSampleThese data come from the inception cohort of the Canadi-an Study of Health and Aging (CSHA) [16], a representa-tive survey of people aged 65 and over. The chief goals ofthe CSHA were to identify the prevalence, incidence out-comes of and risks for dementia. In the first phase (CSHA-1) data collection took place between February 1991 andMay 1992. In Figure 1, the sample population is presentedas a flow chart. Initially, 10,267 people were interviewed– 9,008 in the community, and the remainder in long-term care institutions. The response rate in the communi-ty was 72 % and was 82% for those in long-term care. Par-ticipants in the clinical component were selected from arandom sample of elderly Canadians, based on their scoreon the Modified Mini-Mental Status Examination (3 MS)[17]. Those who screened positive, and a sample whoscreened negative were invited to a clinical examinationdesigned to detect cognitive impairment and diagnose itscause (n = 2914). Demographic and diagnostic detailshave appeared elsewhere [16,18–20]. Briefly, of the 2914who came for a clinical examination 64.4% were women.Their mean age at baseline was 82.0 years; SD 7.43, range65–106. Dementia, and its subtypes, were diagnosed andstaged by standard protocols, [16,19,20] based on the ex-amination and an informant interview, in 1132 people.Of the remainder, 861 were classified as having CognitiveImpairment No Dementia (CIND) [19] and 921 with nocognitive impairment. All 1338 surviving clinical partici-pants received a follow-up examination approximatelyfive years after the baseline assessment [20]. Date of deathwas recorded for 1465 of the 1521 subjects who did notsurvive. The median time to death was 33 months.Selection of deficitsTo select the signs and symptoms for analysis, we havehad to compromise between the desire to use as manyitems as possible from the database (more than 400 itemsthis early stage of our investigations, we have restrictedourselves to deficits that represent a variety of functions,and at the same time, occurred with reasonable frequency(on average, in 26.7% of subjects) had minimum (on av-erage 12%) missing data, and maximum variability acrossthe database. Our earliest studies demonstrated that tak-ing into account from 19 to 26 signs and symptoms (de-pending on the specific applications) we were able toaddress different aspects of cognitive aging from diagnos-tics [21,22] to functional decline [23]. The most impor-tant finding of the initial studies was the demonstrationthat the deficits are not independent. In other words, thevariables are closely interrelated and, in a sense, any defi-cit contains some information about many of the others[11]. This implies that even a restricted set of deficits,which show the above properties can represent a wide va-riety of impairments. In a prior analysis, we identified aset of 20 symptoms, signs, impairments and disabilities(referred to collectively as deficits) that represent loss offunctional activities, sensory impairment, and generalmedical, health and behavioural problems [10,11]. Thevariables represent deficits, which are more common withage and include informant-based (e.g. data on functionfrom CAMDEX H) [16] observer-assessed (e.g. clinical re-ports of hypertension, diabetes) and test (e.g. 3 MS) data.The list of deficits is presented below. (The numbers in pa-rentheses refers to their position in Figure 6 and corre-spond to those used in [27]). Vision loss (22), hearing loss(23), impaired mobility (3), vascular problem (36), gaitabnormality, impaired vibration sense (56), difficulty toi-leting (11), difficulty cooking (7), difficulty bathing (10),difficulty going out (6), difficulty grooming (9), skinproblems (40), resting tremor (47), changes in sleep (2),difficulty dressing (8), urinary complaints (29), gastro-in-testinal problem (28), diabetes (31), hypertension (24),limb tone abnormality (46).AnalysisWe considered binary deficits (represented by either 1 or0), depending on their presence or absence in a given in-dividual. For simplicity, we considered the proportion ofthe deficits in the i-th individual (0 < qi < 1) as a state var-iable, measuring an individual degree of impairment andfrailty. We analyzed the proportion q(t) averaged acrossall subjects at age t, for those with no cognitive impair-ment and checked the hypothesis of the linear increase ofln(q) with age, as a test of the construct validity of ourmodel. Regression techniques were applied to estimatethe parameters of the linear model. Cox proportional haz-ards regression with covariates and Gompertz's modelwere used in analysis of the survival data to test the crite-rion validity of the index in relation to death. The signifi-Page 2 of 8(page number not for citation purposes)were available) and the need to minimize noise due tomeasurement error, including missing data. Therefore, atcance level was set at p < 0.05.BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1ResultsThe impairment index, q ranged from 0 to 1. At any givenage, the impairment index is distributed about the meanvalue for individuals of that age. Figure 2 illustrates thisfor subjects aged 77 years old who did not demonstratecognitive impairment. An individual's health status, fimay be defined as a ratio of that person's impairment in-dex to the mean index value, averaged across individualswithout cognitive impairment, but of the same chrono-logical age: fi = qi/m; f > 1 if the individual is frail and f <1 if the individual is fit. The logarithm of the ratio ln(f)may also be considered as an appropriate index of relativefrailty/fitness; a positive value of the logarithm indicatesfrailty, whereas a negative value indicates fitness. (Whenrelative fitness/frailty is to be measured as a dichotomousvariable, the mean case is taken as fit; i.e. f < 1.) Figure 3presents the proportion of fit and frail people who sur-vive, averaged by 5-years intervals. As expected, this de-creases with age for both groups, but note that a higherproportion of fit individuals than frail survive in each agegroup.Figure 4 presents the log of the mean value of the frailtyindex. Consistent with our hypothesis, the frailty indexvalue increases with age. Because age and the frailty indexsonable precision, across the age spectrum. The regressionline corresponds to the equation:ln(q) = - 4.23 + 0.03 t  (1)For example, according to equation (1), after elementarycalculations, 67, 77 and 87 year old individuals expect tohave the following proportions of deficits, m = 0.11, 0.15and 0.2, respectively. Now consider a 67 individual with3 deficits from the list of 20. The proportion of deficits q= 0.15, and therefore such individual may be consideredas frail since q > m = 0.11. However, a 87 years old indi-vidual with the same q would be considered as fit, since q< m = 0.2. In other words, the frailty index represents rela-tive fitness and frailty, so that a given degree of fitness at age87 may represent frailty at age 67.A separate, but related way of quantifying relative frailtyand fitness is to use equation (1) as a calibration tool,which again can be done because it holds with a satisfac-tory degree of certainty (r = 0.91 corresponding to 83 % ofexplained variance). By inverting equation (1) and input-ting the proportion of deficits, q we arrive at a value of aget as an output. Calculated in this way, the biological age ofan individual can be understood as the age at which theFigure 1Population sample flow chart.CSHA-1 CORE SAMPLE (n = 10,267)Consented, n = 9008 CommunityConsented, n = 1259 InstitutionsCLINICAL EXAMINATION (n = 2914)64.4% womenmean age: 82; SD 7.43; range 65 -106yrs.DEMENTIAand its subtypes(n = 1132)CIND(n = 861)NCI(n = 921)Page 3 of 8(page number not for citation purposes)are well correlated (r = 0.91, p < 0.001) we can use this in-formation to estimate relative fitness and frailty, with rea-average individual has q deficits. Similarly to equation (1)a precise estimate of relative fitness and frailty can be de-BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1rived: if the calculated age is less than that individual'schronological age, he/she is considered fit, if it is greaterthan the average age at which people have such deficits,otherwise he/she considered frail. Such a definition offrailty therefore depends on two inputs: chronological ageand the proportion of deficits seen at any given age. Thedifference between the individual's chronological age andthe expected age given the proportion of deficits can beused to assess a given individual's relative frailty or fitness.Recalling the log-linear relationship between the impair-ment index and chronological age (Figure 4), we suggestcalculation of personal biological age (PBA) by applyingan inverse regression, (regression of age on ln(q)), so that:Figure 2The distribution of the frailty index at chronological age 77suggests varying levels of fitness and frailty, even in thosewith no cognitive impairment.Figure 3The proportion of survivals for frail (solid rectangles) and fitindividuals (total rectangles) decreases with the chronologicalage. However, frail individuals show lower survival at all agegroups than do fit individuals.Proportion of deficits, qNo of cases048121620240 0.1 0.2 0.3 0.4 0.5Chronological age, yearsProportionofsurvived00. 70 75 80 85 90 95 100Figure 4Mean proportion of deficits at any given age for subjects withno cognitive impairment. Solid circles represent the propor-tion of the deficits averaged across all subjects at age CA: In(q) = - 4.23 + 0.03 CA. Correlation coefficient, r = 0.91, p <0.0001Figure 5Time to death by cognitive diagnostic groups as a function ofchronological and biological age. Solid circles correspond toaverage values of time to death (across all subjects of theparticular diagnostic group) with respect to averaged BA dis-tributed along a straight line (r = - 0.98, p < 0.001). Empty cir-cles, (o) correspond to time to death with respect to CA anddid not show such a pattern. The following abbreviations areused for the diagnostic groups: NCI (No Cognitive Impair-ment), CIND (Cognitive Impairment No Dementia), AD(Alzheimer's disease), VD (vascular dementia).Chronological age (years)Ln(q)-2.6-2.2-1.8-1.4-1-0.660 70 80 90 100 Chronological and Biological age, yearsMonths to deathDiagnoses262830323470 80 90 100 110 120NCICIND,moderate ADsevere ADmoderate VDsevere VDooooooo mild ADPage 4 of 8(page number not for citation purposes)that this value can be used as a first approximation in the PBA = 126.65 + 26.09 ln(q)  (2)BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1An individual's PBA is defined as the value at which thosedeficits are present, on average, in the successfully ageinggroup. Chronological (CA) and biological age coincidewhen an individual's number of deficits corresponds tothe age given by the equation above. Otherwise, an indi-vidual may be biologically younger (more fit) or older(more frail) than their chronological age, and the degreeof relative fitness or frailty can be estimated as the differ-ence between PBA and CA. For example, a person with 4deficits from 20 (q = 0.20), according to formula (2) hasan estimated PBA of 83 years. Comparing this value withhis/her chronological age one may consider him/her fit orfrail depending on the difference between PBA and chron-ological age.Figure 5 presents the mean months to death, plottedagainst mean chronological and biological age, for themain CSHA cognitive diagnostic groups. As hypothesised,biological age predicts death in a dose-response relation-ship. While chronological age has only a weak relation-ship to time to death, given diagnosis, biological age isstrongly inversely correlated with the time to death (r = -0.98, p < 0.01).PBA and CA were compared in relation to survival usingCox regression, with each as covariates (Table 1). Notethat while the regression coefficients of CA and PBA arethe same, the standard errors are quite different, so thatPBA shows a much stronger relationship with mortality.Figure 6 serves as a further test of the ability of PBA andCA to predict mortality. A least squares estimates of thesurvival functions using a Gompertz model was fitted tothe survival data for two groups (younger than 80 yearsand older than 80 years) for each of CA (panel A) and PBA(panel B). (The same relationships hold when other agesare used to dichotomize the sample.) The survival func-tions clearly show greater differences with PBA comparedwith CA, indicating a more refined ability to discriminateadverse outcomes with the former.DiscussionWe have extended an earlier conceptualization of func-tional age [10] to present a method of estimating personalbiological age, and from that, to estimate relative fitnessand frailty. In particular, we claim that the proportion ofdeficits accumulated by an individual at a given chrono-logical age allows an operational definition of relative fit-ness and frailty. To validate this definition of relativefitness and frailty we compared survival both as functionof PBA in the cohort, and by varying levels of fitness andfrailty within cognitive diagnostic groups.Calibration equations (1) and (2) comprise our knowl-discovery in databases (KDD). KDD is a set of techniquesbeing developed through a number of disciplines to takeadvantage of existing databases as means of discoveringnew knowledge. [24,25] KDD, defined as the identifica-tion of meaningful and useful patterns in databases, in-cludes both the discovery of previously unseen groupingswithin existing databases [11,22] and predictive model-ling, as also developed in this inquiry.We are aware of important limitations in our study. First,in order to initially describe a simple model, we have lim-ited our analysis to a set of 20 deficits. While these provedsufficient in earlier analyses, [21–23] alternate formula-tions of the impairment index may be more efficient. Themore essential the deficits that are taken into account, themore precise the estimation of frailty that is obtained [26].However, it is not yet clear which are the essential proper-ties of signs, symptoms and functional deficits that needto be selected. This is an ongoing area of investigation, butit appears that the number of deficits, rather than theirprecise nature, might be the most important determinantFigure 6Cumulative proportion of surviving as a function of time todeath for two groups with chronological age (A) and biologi-cal age (B) less and greater than 80 years. Circles correspondto experimental data for individuals below 80 years old andtriangles for individuals above 80 years. Curves correspondto the least square Gompertz's functions (solid lines fit datafor the individuals below 80 years and dashed lines for thosewho was older than 80 years).Time to death (months)Cumul ative Proportion Survi ving0. 10 20 30 40 50 60 70 800. 10 20 30 40 50 60 70 80ABCA < 80CA > 80BA < 80BA > 80Page 5 of 8(page number not for citation purposes)edge about the accumulation of deficits in older adults,and thus can be considered as an example of knowledge[26,27]. We interpret this to mean that frailty might be in-terpreted as a loss of redundancy in a complex system.BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1In addition, we were limited in our databases to individu-als aged 65 years and older, and drew from the screenedclinical sample, so that we cannot make a claim about therepresentativeness of the data. The incorporation of dataon middle-aged and representative samples should allowmore general claims about PBA to be examined. Third, tosimplify the calculations we have suggested, as a first ap-proximation, that a state variable can be estimated as theproportion of deficits. This may seem naive, as if has theeffect of equalizing all the deficits. Evidently, at an indi-vidual level, not all deficits are equally important: heartproblems or diabetes likely may cause death sooner thanfor example, difficulties in getting dressed or skin prob-lems per se. The finding that the proportion of deficits in agiven individual can include seemingly arbitrary or eventrivial ones requires further investigation. For now, we un-derstand this finding to mean that accumulating severaldeficits results overall in impaired adaptive ability. This islikely to be the case if the signs are redundant, i.e. if a giv-en deficit represents a set of others, and if the items of theindex are related. The latter appears to be the case, as illus-trated in Figure 7, which shows that the deficits are not in-dependent. The nodes of the graph correspond to thedeficits (numbered in the Materials and methods) and theedges represent statistically significant relationships be-tween the deficits (defined as the difference between theunconditional probability of the occurrence of deficit Xand conditional probability of deficit X given deficit Y)[11]. This is not surprising when we consider that syner-getic relationships are typical for age-associated deficits. Inother words, roughly speaking, everything is dependenton everything else in complex organisms so that changesin one subsystem affects many others. For example, visionimpairment may be caused by the numerous reasons.Since vision loss, by itself, is not readily regarded as a life-threatening factor, it may indicate a more serious problem(e.g. diabetes, stroke). The more deficits that are used inderiving the frailty index, the greater the chance that suchsecondary signs are linked to serious illnesses. As arguedelsewhere [1,26,28] this is a central aspect of many char-acterizations of frailty. Moreover, whether this holds forany combination of deficits (and not just age-associatedWe considered data together for men and women. The dif-ferences in mortality of men and women are well knownand we recognise that in the next approximation theyhave to be treated separately. However, here we were lim-ited by the power of data representing, in average, 100 in-dividuals at each age and division the sample by sexwould make statistics worse. We intend to address the is-sue of gender differences in accumulation of deficits in theother paper when dealing with the representative largesample of data.Have we offered any special insight beyond the common-sense observation that as people age they are more likelyto become ill, or that ill people are more likely to die? WeTable 1: Parameters of Cox regression model for time to death as a function of CA and BA tag for table legends and titlesVariable Beta STD t-value Wald statist p-valueCA 0.0081 0.0038 2.15 4.63 0.0313BA 0.0081 0.0014 5.70 32.49 <0.000001Figure 7Inter signs synergy graph. Nodes indicate the deficits (codescorrespond to the deficits from Methods and [29]) and edgesindicate the statistically significant relationships between defi-cits, i.e. when the conditional probability of one deficit, givenanother is statistically different (p < 0.05; t-test) from theunconditional probability of the first deficit.234678910112223242829313640464756Page 6 of 8(page number not for citation purposes)ones) additionally requires further study, although we rec-ognize that such summarization does not allow the influ-ence of individual disease states to be tested. [22,29]believe that we have. In the first instance, we empiricallyderived an index to distinguish biological from chrono-BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/1logical ageing with a result that seems to offer reasonableprecision. For example, equation 2 predicts a maximumlife span 126 years, compared with the maximum-record-ed life span of about 122 years. This is of interest, but wehave to be careful with its interpretation. We are notclaiming that 126 years is the absolute limit of the humanlifespan. Note that this is an average characteristic of thesample, in which older adults with deficits have beenover-represented. Moreover, as has been well argued,there is no evidence for such a limit, and considerable ev-idence against it [30]. The accumulation of deficits in ourequations suggests a process whereby damage is initiallycompensated by redundancy of systems, but when the re-dundancy is exhausted (i.e., too many deficits are accumu-lated) any new insult leads to death [31]. Similarly,equation (1) corresponds to the differential equation re-lating the proportion of the deficits q to the instantaneousrate of increase in deficits; i.e. dq/dt: dq/dt = kq (k = 0.03).In other words, the average annual accumulation of defi-cits in successful ageing (in this case for individuals withno cognitive impairment) is 3% per year.Here we used death as the outcome. However, other ad-verse outcomes, such as institutionalisation can also beconsidered and will be the subject of further inquiry. Thisapproach also incorporates an important feature of frailty,which has otherwise received little attention in its opera-tionalization, namely its relationship to age. While ageingis readily accepted as being associated with frailty, as ar-gued elsewhere, [1,26] the notion of frailty finds its rootsin the imprecision of chronological age as an explanatorymodel in predicting adverse outcomes in individual cases.As presented here, PBA represents the individual case, butits relationship to age is inherent in the definition. As not-ed (Figure 2), individual values of the impairment indexcan vary widely at any given CA. As a statistical analogy,PBA for a subject represents their individual value; CA rep-resents the mean. In other words, by taking into accountboth the proportion of deficits accumulated by an indi-vidual at a given age, and the average proportion of defi-cits estimated from the successful ageing group, we areable to describe respectively, an individual or a character-istic of a population. While this is of interest, and servedas the basis of further inquiry [27] we are not making aclaim to have definitively calculated the PBA. Rather, wehave presented one method of so doing, and note that thisapproach has properties which encourage us to pursue theanalysis further in other databases.ConclusionsA frailty index can be estimated as the proportion of defi-cits older adults accumulate over time. By knowing theproportion of deficits a given individual has accumulatedestimated from their frailty index value, as the age atwhich the person has accumulated that frailty index value.Personal biological age is a stronger correlate of mortalitythan chronological age. Relative frailty and fitness cantherefore be estimated as the difference between chrono-logical and biological age. The frailty index is a sensitivepredictor of survival. As the index includes items not tra-ditionally related to adverse health outcomes, the findingis compatible with a view of frailty as the failure to inte-grate the complex responses required to maintain func-tion.Competing interestsNone declared.AcknowledgementsThe data were collected as part of the Canadian Study of Health and Aging, funded by the Seniors Independence Research Program, administered by the National Health Research and Development Program (NHRDP) of Health and Welfare Canada (Project No. 6606-3954-MC(S)). This analysis was supported by grants from the National Health Research Development Program of Health Canada (grant no. 6603-1471-55) and the Queen Eliza-beth II Research Foundation. The data reported in this article were collect-ed as part of the Canadian Study of Health and Aging. The core study was funded by the Seniors' Independence Research Program, through the Na-tional Health Research and Development Program (project no. 6606-3954-MC(S)). Additional funding was provided by Pfizer Canada Incorporated through the Medical Research Council/Pharmaceutical Manufacturers As-sociation of Canada Health Activity Program, the National Health Research and Development Program (project no. 6603-1417-302(R)). The study was coordinated through the University of Ottawa and the Division of Aging and Seniors, Health Canada. Additional funds for analysis came from Hoechst Marion Roussel Canada, through a grant administered by the Med-ical Research Council of Canada (grant MRC PA-13560). Kenneth Rock-wood and Janice Graham are supported by Investigation Awards from The Canadian Institute of Health Research. We are grateful to two anonymous referees for their constructive comments.References1. Rockwood K, Fox RA, Stolee P, Robertson D, Beattie BL: Frailty inelderly people; an evolving concept. CMAJ 1994, 150:489-4952. Reisberg B, Ferris SH, de-Leon MJ, Crook T: The Global Deterio-ration Scale for assessment of primary degenerative demen-tia. Am J of Psych 1982, 139:1136-11393. Hébert R, Brayne C, Spiegelhalter D: Factors associated withfunctional decline and improvement in a very elderly com-munity-dwelling population. AMJE 1999, 150:501-5104. Thomas VS, Rockwood K, McDowell I: Multidimensionality in in-strumental and basic activities of daily living. J Clin Epidemiol1997, 51:315-3215. Seeman TE, Berkman LF, Charpentier PA, Blazer DG, Albert MS, Ti-netti ME: Behavioural and psychosocial predictors of physicalperformance: MacArthur studies of successful aging. J GerontolA Biol Sci Med Sci 1995, 50:M177-M1836. Campbell AJ, Buchner DM: Unstable disability and the fluctua-tions of frailty. Age & Ageing 1997, 26:315-3187. Van Boxtel MP, Buntinx F, Houx PJ, Metsemakers JF, Knottnerus A,Jolles J: The relation between morbidity and cognitive per-formance in a normal aging population. J Gerontol A Biol Sci MedSci 1998, 53:M147-M1548. Finch CE: Variations in senescence and longevity include thepossibility of negligible senescence. J. Gerontol. Biol. Sci. 1998,53A:B235-B2399. Rockwood K, Stadnyk K, MacKnight C, et al: A brief clinical instru-ment to classify frailty in elderly people. Lancet 1999, 353:205-20610. Graham JE, Mitnitski AB, Mogilner AJ, Rockwood K: The dynamicsPage 7 of 8(page number not for citation purposes)at a particular age, personal biological age can be estimat-ed. Personal biological age for a given individual can beof cognitive aging: distinguishing functional age and diseasefrom chronological age in a population. Am J Epidumiol 1999,150:1045-1054BMC Geriatrics 2002, 2 http://www.biomedcentral.com/1471-2318/2/111. Graham JE, Mitnitski AB, Mogilner AJ, Gauvreau D, Rockwood K:Symptoms and signs in dementia: synergy and antagonism.Dementia 1996, 7:331-33512. Yates FE, Benton LA: Loss of integration and resiliency withage: a dissipative destruction. In: Handbook of physiology: aging1995591-61013. Lipsitz LA, Goldberger AL: Loss of 'complexity' and aging. JAMA1992, 2671:806-80914. Berg L, Danziger WL, Storandt M, Coben LA, Gado M, Hughes CP,Knesevich JW, Botwinick J: Predictive features in mild senile de-mentia of the Alzheimer type. Neurology 1984, 34:563-56915. Small BJ, Backman L: Cognitive correlates of mortality: evi-dence from a population-based sample of very old adults. Psy-chol Aging 1997, 12:309-31316. Canadian Study of Health and Aging Working Group: The CanadianStudy of Health and Aging Study methods and prevalence ofdementia. CMAJ 1994, 150:899-91317. Teng EL, Chui HC: The Modified Mini-Mental State (3 MS) Ex-amination. J Clin Psychiatry 1987, 48:314-91318. Graham JE, Rockwood K, Beattie BL, McDowell I, Eastwood R,Gauthier S: Standardization of the diagnosis of dementia inthe Canadian Study of Health and Aging. Neuroepidemiology1996, 15:246-25619. Graham JE, Rockwood K, Beattie BL, et al: Prevalence and severityof cognitive impairment with and without dementia. Lancet1997, 349:1793-179620. Canadian Study of Health and Aging. Working group: The incidenceof dementia in Canada. Neurology 2000, 55:66-7321. Graham JE, Mitnitski AB, Mogilner AJ, Gauvreau D, Rockwood K: Analgorithmic approach to the differential diagnostics of de-mentua dementia. Dementia 1996, 7:324-33022. Mitnitski AB, Graham JE, Mogilner AJ, Rockwood K: Vector diag-nostics in dementia derived from Bayes' theorem. Am J Epide-miol 1997, 146:665-67123. Mitnitski AB, Graham JE, Mogilner AJ, Rockwood K: The rate of de-cline in functions in Alzheimer's disease and other demen-tias. J Geront Med Sci 1999, 54A:M65-M6924. Fayyad UM, Piatetsky-Shapiro G, Smyth P, Uthurusamy R, eds: Ad-vances in Knowledge Discovery and Data Mining. Menlo Park,CA: AAAI Press 199625. Tamayo P, Slonim D: Molecular classification of cancer: classdiscovery and class prediction by gene expression with self-organizing maps: methods and applications to hematopoiet-ic differentiation. Proc Nat Acad Sci USA 1999, 96:2907-291226. Rockwood K, MacKnight C, Hogan D: Conceptualization and in-strumentation of frailty. Drugs Aging 2000, 17:295-30227. Mitnitski AB, Mogilner AJ, Rockwood K: Accumulation of deficitsas a proxy measure of aging. The Scientific World 2001, 1:323-33628. Brown M, Sinacore DR, Binder EF, Kohrt WM: Physical and per-formance measures for the identification of mild to moder-ate frailty. Journal of Gerontology: Medical Sciences 2000, 6:M35-M35329. Ferraro KF, Wilmoth JM: Measuring morbidity: disease counts,binary variables, and statistical power. Journal of Gerontology: So-cial Sciences 2000, 3:S173-S18930. Gavrilov LA, Gavrilova NS: The Biology of Life Span. A QuantitativeApproach 199131. Gavrilov LA, Gavrilova NS: The reliability theory of aging andlongevity. J. Theor. Biol. 2001, 213:527-545Publish with BioMed Central  and  every scientist can read your work free of charge"BioMedcentral will be the most significant development for disseminating the results of biomedical research in our lifetime."Paul Nurse, Director-General, Imperial Cancer Research FundPublish with BMC and your research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central yours - you keep the copyrightBioMedcentral.comPage 8 of 8(page number not for citation purposes)editorial@biomedcentral.comSubmit your manuscript here:http://www.biomedcentral.com/manuscript/


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items