Scribd
Upload
82 views7 pages

Blair, Steven N., Et Al. Physical Fitness and All-Cause Mortality A Prospective Study of Healthy Men and Women

Uploaded by

12trader
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
82 views7 pages

Blair, Steven N., Et Al. Physical Fitness and All-Cause Mortality A Prospective Study of Healthy Men and Women

Uploaded by

12trader
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 7
Original CONtribUtiONS ttt Physical Fitness and All-Cause Mortality A Prospective Study of Healthy Men and Women ‘Steven N. Blait, PED; Harold W. Koh! Ill, MSPH; Ralph S, Paffenbarger, Jr, MO. DrPH; Debra G. Clark, MS; Kenneth H. Cooper, MD, MPH; Larry W. Gibbons, MD, MPH We studied physical fitness and risk of all-cause and cause-specttic mortality in 10 224men and 3120 women who were given a preventive medical examination. Physical fitness was measured by a maximal treadmill exercise test. Average {follow-up was slightly more than 8 years, for a total of 110.482 person-years of observation. There were 240 deaths in men and 43 deaths in women. Age- adjusted all-cause mortality rates declined across physical fitness quintiles from 64.0 per 10 000 person-years in the least-fitmen to 18.6 per 10 000 person-years in the most-fit men (slope, ~ 4.5). Corresponding values for women were 39.5 er 10.000 person-years to 8.5 per 10000 person-years (slope, — 5.5). These trends remained after statistical adjustment for age, smoking habit, cholesterol level, systolic blood pressure, fasting blood glucose level, parental history of coronary heart disease, and follow-up interval. Lower mortality rates in higher fitness categories also were seen for cardiovascular disease and cancer of combined sites. Attributable risk estimates for all-cause mortality indicated that low physical fitness was an important risk factor in both men and women. Higher levels of physical fitness appear to delay all-cause mortality primarily due to lowered rates of cardiovascular disease and cancer. PHYSICAL activity is inversely associ- ated with morbidity und mortality from several chronie diseases. The apparent- ly protective effect of a more active life is seen for occupational activity and death from cardiovascular disease! and colon cancer,’ and for leisure-time phys- ieal activity and cardiovaseular dis ease." Higher levels of leisure-time physical activity are associated with in creased longevity in eollege alumni. ‘These associations of sedentary habits to health appear to be independent of confounding by other well-established For editorial comment see p 2437. risk factors.’ Furthermore, the rela- tionship of physical fitness (an attri- bute) to physical activity (a behavior) and disease rates is controversial," and it is uncertain whether physical zetivity sufficient to increase physical fitness is required for health benefits. In contrast to physical activity, pub- lished studies on physical fitness and mortality are few, typically with fewer than 20000 person-years of follow-up, and usually limited to men, Physieal ac~ tivity is an important determinant of ~Fromine inser rabies Resench Dales Tex Raping eaests 9 the nse es ooo Fe search, 12830 ester Re Oss, 76230 0 fia) JAMA, November 3, 1989—Vo! 262, No. 17 (LAMA. umsranecenn-2 ness is an objective marker for habitual physical activity. Physical fitness ean be measured more objectively than physi= ‘al activity, and thus may be more u ful clinically. Research studies that in- clude the measurement of physical fitness may provide additional insight into the contribution of a physically ac- tive way of life to decreased risk of mor- bidity and mortality. Here, we report all-cause and cause- specific’ mortality by physical fitness categories in men and women followed up for 110 482 person-years, or an aver- age of more than 8 years, SUBJECTS AND METHODS. Subjects ‘The 13344 study participants com- prised 10224 men and 3120 women who received a preventive medical examina- tion at the Cooper Clinie in Dallas, Tex, during 1970 to 1981. Patients were in cluded in the study if they were resi- dents of the United States at their frst clinie visit, had a complete examination, and achieved at least 85% of their age: predicted maximal heart rate ona ‘treadmill exercise test at the baseline linie visit. Patients not achieving this ‘maximal heart rate standard were pre- ‘sumed to be more likely to have preex- isting disease or be receiving medica- tion with B-blockers, These conditions Would be associated with poorer tread. rill est performance and higher risk of death during follow-up. Thus, excluding patients with these characteristics is a conservative decision that reduces the chance of finding a spurious inverse re- lationship hetween fitness and mortal- wy. At baseline, all patients had no personal history of heart attack, hyper tension, stroke, or diabetes: no resting electrocardiographie (ECG) abnormali- ties; and no abnormal responses on the exercise BCG. Clinical Examination ‘The baseline examination was given alter an overnight fast of at least 12 hours and after patients gave their in formed consent, The examination was 2 complete preventive medical evaluation that. ineluded a personal and family health history, a physical examination, a questionnaire on demographie charac. teristiesand healthhabits, anthropome- try, resting ECG, blood chemistry tusts, blood pressures, and a maximal treadmill exercise test. Examination ‘methods and procedures followed 2 standard manual of operations and have been described further in earlier re- ports." All patients were free of known chronie disease as determined by the following criteria: no personal his- tory of heart attack, hypertension, stroke, or diabetes; no resting ECG ab- xnormalities; and no abnormal responses onthe exercise ECG. Physical fitness was measured by a maximal treadmill exercise test.” ‘Treadmill speed was set initially at ‘88mv/min. The grade was 0% for the first minute, 29 the second minute, and in- ereased 1% each minute until 25 min- utes. After 25 mimutes, the grade did not change and speed’ was increased 5.4 m/min each minute until test termi- nation, Patients were given encourage- ment to give maximal effort. ‘Total treadmill test time in seconds was the variable used in analysis. ‘Treadmill ‘time from this protocol is highly corre- lated with measured maximal oxygen Physica Fitness and Moralty—Blairetal 2308 uptake in men‘ (7.92) and women* (r=.94), which is the most widely accepted index of eardiorespiratory fit- ness. Patients were assigned to physical fitness categories based on their age, sex, and maximal time on the treadmill test, Tresdmill-time quintiles were de- termined for each age and sex group. Individuals with a treadmill time in the first quintile were assigned to the low- fit group. Those with scores in the sec- ond through the fifth quintiles consti- tuted fitness groups 2 through 5, respectively. This, assignment to a ness ealegory was based on age and sex norms of treadmill performance rather than by an absolute fitness standard. (Treadmill-time quintile cutoff points for each age group for men and women may be obtained from us.) Cigarette-smoking status was deter- mined from the medical questionnaire. Patients who reported smoking at pre~ sent or within the 2 years preceding the baseline examination were designated as current smokers. This conservative definition for smoking was adopted be- cause many smokers may have quit tem- poratily in preparation for their preve tive medical examination, and mortality risk for recent quitters is similar to eon- tinuing smokers." Results from the smoking analyses were essentially un- changed when current smoking was de- fined as cigarette smoking at baseline or during the year preceding the exami- nation Height and weight were measured on a standard physician’ seale, and body ‘mass index was calculated (kilograms per meter squared). Blood pressure was measured by the auscultatory method with a mereary sphygmomanometer, diastolic pressure being recorded as the disappearance of sound, Serum samples were analyzed for cholesterol and glu- cose by automated techniques. Mortality Surveillance Study subjects were followed up for mortality from their first clinie visit through 1985. The average length of fol. low-up was slightly more than 8 years, and the total follow-up experience for the cohort was 110482 person-years, Several follow-up methods were used. Decedents were identified by reports from family, friends, and business asso- ciates; responses to appointment re- minders; and other mailings from the clinic. The entire cohort was sent acase- finding and disease-identifying ques- tionnaire in 1982." Nonrespondents were followed-up via the Social Security Administration files, the Department of Motor Vehicles in the subject’ state of residence, and a nationwide eredit bu- 2306 JAMA, November, 1989 vol 262, No.17 ‘Table 1. Baseline Characteristics of Suvving and Deceased Male and Female Patents, Aerobics Center Longtudna Soy. 1970 10 1981 Surviving (oro5e) 7 z sp ag, ae as 5h contdence ks) (1.4.7) Wig i wie at {259% coience hts) (317.821) Figs, on Was 63 (0694 corience Has) (1787. 709) Bea, mass inden 256 38 (Gist eontaerce is) (255, 257) Feagemaane To er (ss% conieence tts) (hor, 022) Fotomun y aa 28 {5h coiionce nis) (23.85) Taalcroksteol lee mmol S450 {95% confidence is) 66.54 Bysesc Hood pressure, mimtg 1204197 "ose conerce ts) 2b, 2019) Dac tingg esr, wag 7878 {95% contigence hers) | 79.5, 799) ‘rere smokes 285 Deceased Deceased — ae i 0 zs oe as 99 Si as oasis “era. 560) m2 133 Sop 07607 9a wise “basco “ere.639) Wo; gi ies 66 iar 86 928) “Treas. 1550) “0624, 1658) 28 2235 m2 a7 (54.25) “een.ees)ees.249) ar ot O aa 50 (726.56) (ee2,680) “tava. s60) Bs ao a2 28 es ae 60.70) re “Gar6) aT (80.62 iSt7.52)_“er,ea) agra 4d 198 tiesov03) “innizan__‘Gibs, wos) Cra aS 702-94 (10.000 742.748) “764,420, ws Ba Re ————— eau network, The National Death In- dex has been used since it was estab- lished in 1979 to search for possible matches inthis cohort. Finally, individ- uals with unknown vital status and with a Dallas-area address were checked in local telephone directories, Follow-up has been difficult. since patients come from all 50 states and are mobile and since a significant portion of the follow- up occurred prior to the establishment. of the National Death Index. Despite these limitations, vital status has been ascertained for 95% ofthe cohort. ‘There have been 283 deaths in the study group, Official death certificates were obtained from the states. ‘The un- derlying cause and up to four eontribut- ing eauses of death were coded by a nosologist according to the Internation- al Classification of Diseases, Ninth Edition, Revised. Data Analysis A total of 288 deaths were identified in the cohort over the average of ap- proximately 8 years of follow-up. Mor- tality rates per 10000 person-years of {follow-up were computed for each of the five fitness categories and age-adjusted by the direct method, using the total experiencein the populiationas the stan- dard, Age differences were adjusted by the following groupings: 20 to 39, 40 to 49, 50 to 59, and 60 or more years. These rates were then used to compute rela- tive risks (RR) of death for each fitness ‘quintile as well as for examination of the role other variables played in confound- ing the relationship between fitness and mortality. Attributable risk percent ages (etiologic fractions) for those {groups exposed to adverse characteris- ties were ealeulated 2s were population- based estimates of attributable risks.” Multiple logistic regression was used to estimate RRs of death among the fitness quintiles after control for associ- ated confounding risk factors." Interval estimation was used to ealeulate confi- dence intervals (Cs) around point esti- mates of risk. RESULTS Patients n this study are from middle to upper sociveconomie strata; approxi ately 70% are college graduates. Most are employed in professional, exeeu- tive, or white-collar position’. More than 9 are white. Baseline character- isties on selected demographic and clini cal variables are shown in Table I. Dece- dents were somewhat older, less physically fit, and had less favorable sk profiles, Table 2 shows the age-adjusted all- cause death rates by physical fitness categories in men and women, Relative risks of death with the 95% Cls are shown with the mostft quintile as the reference category. Less-ft individuals hada higher risk ofdeath than the more- fit men and women. Increased RR for all-cause mortality’ was significantly higher for the lesst-fit quintile in men, and for the two least-fit quintiles in Women. The 95% Cls for the test for linear trend across Gtness categories did not include 1.0 in either men or women, Univariate age-adjusted RR for all- ‘cause mortality for several important clinieal and life-style variables for men and women are presented in Table 3. Physical ness anc Mortality—Blairetal ‘Table 2.—Age-Adustes AB-Cauce Death Rates par 10000 Person Years of Follow-up (1970 to 1985) by Physical Fines Groups n Men and Women inthe Aerobics Center Longhunal Study SS =e = vot nARrpaSo rahe || Seeeerwoee | ee 1 0 = 5 = 7 = mr = = : xs az2%8 i Bs fois @ a eens 10 ‘ a ey] 7 s FF 10) Se eee zt ‘fest for linear Wend, slope ~ 5.5; 95% contigence lis. -$.2, - 1.9. gs +10} “Table 3.—Petatve Fisk for ALCause Motay to: Selected Cirical and Lte-syle Venables, Men and Women the Aerobics Centr Longitudinal Study Guenter or aut npaslay_ Ener parent hd of coronary heart daease ‘Serum glos9 98 7 Serum etoeseottvel 6.20 meno. 150,494 Boy macs ex 269 ‘Curent ereker or aut n paca Ear parent 90d of coronary nev osoaso 00,4235 ‘The findings show an increased risk, as expected, for all variables except body mass index, which shows a trend in the expected direction only in women. Multiple logistie analyses were done to estimate RR of death in the fitness categories while adjusting for potential confounding. The dependent measure ‘was all-cause mortality and the model included physical fitness and all vari- ables in Table 3. All variables were in- cluded, although overweight for height in both men and women and parental history of coronary heart disease (CHD) in women were not statistically signifi- cantly associated with mortality in uni- variate analyses. The RRs (95% Cl) of, Jow physical fitness for all-cause mortal- ity foreach quintile (to Q4) compared with the most-fit quintile were as fol- lows: Q1=1.58 (1.23 to 1.89), Q2= 1.08 (0.81 to 1.80), Q8=1.12 (0.89 to 1.40), and Q4 = 1.08 (0.81 to 1.28) for men; and QL= 1.98 (1.18 to 3.47), Q2=1.45 (0.80 to 2.62), Q3=1.07 (0.55 to 2.08), and 07 (0.55 to 2.23) for women, A SAMA, Novertoer 3, 1989— Vo! 262, No. 17 ‘more pronounced dose-response gradi- tent was seen when length of follow-up (as a continuous variable) was added to the model, Relative risks (95% CI) for the four less-ft quintiles relative to the most-fit quintile were as follows: QL = 1.82.88 to2.40), Q2=1.38.(.0t0 1.78), Q3=1.29 (0.97 to 1-70), and Qk=1.06 (0,78 to 1.44) for men and Q1~3.02 4.89 to 11.04), Q2=3.01 4.05 to 8.65), Q8=2.06 (0.66 to 6.22), and Q1= 1.55 (0.49 to 4.91) for women, Sev- eral interaetion terms among the inde- pendent variables were tested, and the assumption of no interaction on a multi- plicative scale was not violated, ‘Subclinical disease could eause poor performance on the treadmill and also lead to elevated death rates in patients presumed to be healthy at baseline. Mortality rates in both short- and long term follow-up were examined to test the hypothesis that preexisting disease was confounding the relationship be- tween fitness and mortality. Logistie regression analyses were done for two {f0p) and 9120 women (boron) it Aerob Center Longtusinal Std, bysysica! finess que ‘es a5 delomined by maximal Teas exerise teste subgroups.as follows: the first 3 years of follow-up and for extended follow-ap af- ter 3 years. The dependent measure was all-cause mortality. Low again was defined as the first quintile of the fitness distribution. Other indepen- dent variables in the analyses were those in Table 8, to control for possible confounding. Adjusted RRs for all cause mortality in low-fit men were as follows: follow-up less than or equal to years, 1,60 (95% CI, 1.18 to 2.16); and follow-up greater than 3 years, 1.45 (95% CI, 1.08 to 1.96). Corresponding values for women were as follows: less than or equal to 3 years, 1.47 (95% Cl, 0.14 to 2.94); and greater than 3 years, 3.00 (95% CI, 1.06 08.61), The elevated RR in later follow-up suggests that the relationship between fitness and mor- tality isnot likely to be due entirely to confounding by subclinical disease. ‘Age-specific, all-cause mortality rales actoss ‘fitness categories are shown in Fig 1. ‘The upper panel pre- sents data for men, and the lower, for women, Inboth analyses, the decline in death rates with higher levels of fitness ‘is more pronouneed in the older individ- uals. The small number of deaths in the younger women leads to unstable esti- mates of the death rate in this group. ‘Table 4 shows cause-specific death rates by fitness categories in men and women, The fitness quintiles were col- Japsed into three groups for these ana- lyse due to smaller numbers of deaths Physical Fitness and Mortaliy—Saretal 2397 “eble 4.—Age-hajisted Cause-Spectic Ova Rates per 10000 Person-Years of Follow-up (1970 10 1965) by Physical Finess Groups n Men and Women in the ‘Aerobics Cones Longtudnal Stay a onthe at = td et nhingcasn tonite teas st eter me a ew! esas sa -s2 Se raeas 3 aaa Pa a rms fet as n02.-07 or a we Pil Scanian ee 1 wes “ese 7 a ae Ee, 008) = ese “8-33 abe tS) 5 ue ee oe we eee 7 is $a aw sas Seton 7 ts ee rot SS on oan ere renee Cao Sores Tone eT TOOT w a we oo! name & ne, aaa santos gb a Nore § om : - ve, ey EY A 1 le, — i } 7 ey eA g2 z ge an ‘ o samen [pie wae econ 7 : . t ro tones é ty aoe i aoae . 1 : [zn — i ; i : i ie vol Ps j y a 7 J coun / eee; “Toes Te om 0 ashes Pi ran ws ay Fanos Cnooo) Firesscatgory Fig 2—Relative risks of al-cause mentaly in 3120 women inthe Aerobics Center Longitudinal Stay, by ‘hysical fness catorae ad ood preeeue (A, eum chotaetr! level (8), serum ghcose iva (). ‘Smoking habits (0) Body mass index (E) and parental history of coronary heart dscese (F). Each Dar ‘epracese the rete isk based on oge-edjustus, all cause death rates per 10.00 person-years ct olow= LUnwthtn elaiverskoftne ont ight callaet at 1.0. Numbers entopotthe bars aretrealrcase death rloe ‘er 10000 person‘ears of folow-up for each call. The number of deaths fh each cel is shown Inthe paraleograms| in the specific causes. There are few strong gradient across fitness groups in presented in Figs 2 and 8. In these ste- deaths for the specific causes in women, _bothmen and women, while none isseen _reograms, the back-left cell shows the which leads to unstable estimates of for other causes of death. RR for ‘the presumed highest-risk rates; these results should be inter- The RRs for all-cause mortality by group (eg, ow fit and high systolic blood preted cautiously. Death rates for ear- cross-tabulations of fitness groups and pressure). The lowest risk group (refer- diovascular disease and cancer show a other elinical and life-style variables are ent) is in the front-right cell of the fig- 2398 JAMA, Novernber 3, 1988—Vol 262, No.17 Pysicel Finest and Mortlty —lai eta) g 0” Ne otbeats a 4° dea & 2% | Cee ie resrgnise 3 [is bobo, aches) one Freeaconmey ° e r s of Sfenre Be No olDetns ro.stoeere 4 o.tbeste a ase i Seay 3 a 3) AERA 5 * ay fe : ae ou 7 = Panty Hy Deh com eo ee ret AE ttn ARTE Fines Category — fusca Fig 3.—Ralatin ric of allcauso mertaly in 10.224 man in tha Aarobics Canter Longtudieal Study. by [Physical tiness categores and bcos pressure (A), serum cholesterol ove (B), sam glucce lave! (C), Smoking nabts (0), Booy mass index represents afd patent history of coronary heart csease (F). Each Bat wen the rete iskof the tront-ght call set at 0, Numbers on top ote bar are the slave death ates per 10000 porzonyoare of fellow-up for each oe. The numéer of athe in each cal i shown in the paraleiograms, ures. Cutoff points for the eliniea! and behavioral risk factors in these analyses were established somewhat arbitrarily, 80 as to provide an adequate number of person-years in each cell for analysis. Increased risk of death in low-fit men andwomenis clearly illustrated in these stereograms, and this pattern generally holds across risk strata for the other variables. In several cases, notably ste- reograms for men onblood pressure and

You might also like