Main Menu                  Health Library        Exercise and Cardiofitness  

 Cardiofitness and the Risk of Cardiovascular Disease and Death from All Causes  

Forrest H. Blanding,  CHE-CHE                                                                             

Draft to Internet at http://www.lifeahead.net                                                                                                       Jan 1,  2004

Reviews and Comments appreciated to forrestx@cox.net.                                                                  Updated  July 9, 2007       

Abstract:

The risks of Cardiovascular (CV) disease and death and death from all causes are related at high significance to an index  of cardiofitness called the CFR that can be measured using a treadmill. The CFR that formerly was named Heart Efficiency or HEF is an age and sex adjusted index of the usual measure of cardiofitness of VO2 Max in ml of oxygen/kg.min and that alternately can be identified as METS.   Cardiofitness that is alternately known as cardiovascular or cardiorespitory fitness varies from about 80 CFR for very unfit to 150 CFR for quite fit men and women.  A 47% reduction is risk of CV disease and CV death and a 37% reduction in death from all causes is associated with each improvement of 10 CFR in cardiofitness.  The cardiofitness of individuals is determined mostly by selected types and amounts of physical activity and exercise, but is in part determined by genetics.  Improvements in efficiency of the CV system and in potential coronary artery flow produced by the regular periodic increases in the blood flow from exercise provide a quantified explanation for most of the benefits of cardiofitness.  The benefits of the efficient CV exercise that produces improvements in cardiofitness appear to have been substantially underestimated by most research results on physical activity and exercise. Cardiofitness measured either as VO2 max adjusted for age and sex or by the CFR is a new independent health risk factor potentially larger in importance to US population risks than are those associated with serum cholesterol, blood pressure, or smoking.

Introduction

Many studies have shown significant relationships between various measures of physical activity and cardiovascular diseases and death, and death from all causes (3).  Most studies have attempted to relate estimates of physical calories of exercise to health risk.  An alternate theory long hypothesized is that cardiofitness provides a primary and more casual protection of health, and that physical activity benefits mostly to the extent that it improves this cardiofitness.  Most studies of physical activity or exercise have shown that those in populations that exercised usually obtained 30-50% lower risks of CV disease than did those that did not exercise (3).  But as will be shown, much larger differences in risks of disease and death than this are associated with differences in measured values of cardiofitness.

An accompanying paper introduced an index of physical fitness called the Cardiofitness Ratio (CFR) that is directly proportional to the usual measure of VO2 Max in ml of oxygen/(kg.min) but also provides a numerical scale of similar significance for men and women of all ages.(7)  The CFR of individuals can be determined by a simple treadmill test described.  It thus becomes of interest to explore the relationship between cardiofitness and the risk of cardiovascular disease and death and death from all causes determined in available population research. 

Problems Using Research on Physical Fitness and Disease.

Eleven studies relating risk of CV diseases or death from all causes were found that identified the physical fitness of groups studied by measured differences in cardiofitness that could be translated into a common basis. Although the studies used differing methods for measuring cardiofitness, results of the various cardiofitness measurements are translated approximately herein in Table 1 following to common values of VO2 Max and the CFR.

Each of the eleven studies was prospective, with cardiofitness of participants usually measured only at baseline, but with outcomes on included populations followed for periods ranging up to 17 years.  All studies showed the risks of disease were related at significance to differences in cardiofitness at study baselines.  But the studies conducted for shorter time periods measured much larger effects of risk of disease vs. baseline differences in cardiofitness than did the longer term studies.  The reason for this appears clear.  Cardiofitness is a physical state of the CV system that develops over a period of several months and that can be lost a few months after exercise is stopped. As for diet, individuals' habits of exercise are variable. Thus the actual cardiofitness of groups compared in these studies was different at the time of disease than that measured at baseline.  Risk of disease and mortality should be related most accurately to cardiofitness at actual time of event. (5,11)  Thus a key question needing answer is: “What differences in risk of cardiovascular disease and death from all causes is directly associated with differences in probable actual levels of cardiofitness?”

Prospective studies of cardiofitness and disease that could answer this question directly would monitor the actual cardiofitness of individuals at least each year, and preferably each 6 months. Such studies of the ten thousand men and women needed for good significance are unlikely to be achieved, and even if feasible results would be many years distant. Thus the only alternative for achieving a most probable answer to this important question is to obtain a best estimate of the cardiofitness of groups already studied at the time when events occurred.

Research data provide a useful if approximate answer to this problem. Leon showed that differences in leisure time physical activity (LPTA) of men in his study dropped from about 118 at baseline to about 40 after both 4 and 6 years, for a decline of 24% and 16% per year respectively.(15)  The exercise of those initially of lowest cardiofitness moved up, and that of those initially highest cardiofitness declined. Morris reported that about that about half of those doing vigorous exercise were still doing this about 7 years later. (22)

Movements in cardiofitness of population segments measured in the Cooper Institute studies showed similarly that less than half of the key unfit group remained unfit after 5 years into the study.(5)  Computations from these data suggest that a probable regression in cardiofitness differences of 15% per year occurred in this data set, with much of this due to those initially unfit becoming more fit.  And changes in risk of disease during this study were related to whether or not cardiofitness was maintained.  Similar changes in cardiofitness during the term of study were measured by Erikssen. (11)  Paffenbarger found that only about 10% of his population of Harvard graduates initially classed as “High activity” were still at this activity level 13 years later. (24)  Again a computation suggests an average regression in exercise differences of about 16% per year in this population group.

This usual decline of differences in cardiofitness of groups with age means that shorter-term studies would be expected to reveal larger effects of disease related to baseline cardiofitness than those measured in longer term studies. Smoothed plots of the cumulative survival curves from  coronary death reported by Hein for various study years and terminating at zero deaths at study start showed a clear fan of increasing deaths for each lower category of cardiofitness measured as VO2 Max.(12)  The high cardiofitness group had an indicated percentage risk of death from coronary disease vs. that of  the low cardiofitness group of 17% (or just 1/6th)  at year 3,  30% at year 7, but a much less favorable 63% over the full 17 years. Most actual disease events occurred during later years of the study when differences in cardiofitness probably were much lower than those at study baseline.  Sandvik noted that the CVD risk ratio for highest cardiofitness quartile vs. lowest was 21% at seven years, but dropped to 42% for full 16 years. (26)  These results from actual risks of populations also suggest that cardiofitness differences of groups usually declined at a rate of about 16% per year after baseline.

In summary, all of the above studies that reported results found that the amounts of exercise and cardiofitness differences of groups during most of their actual study periods probably were substantially lower than the values  measured at baseline.  It is of interest thus to compare risk of disease with both baseline differences in cardiofitness and with the more probable actual average differences in cardiofitness that existed during the times that  disease and death events were recorded.

Results of Research Relating Measured Cardiofitness to CV Disease and Death from all Causes.

Table 1 following includes the the results of all studies found wherein the measured cardiofitness of groups could be translated directly into common basis values of VO2 Max and CFR.  Estimated cardiofitness levels by group are shown for men (M) or women (W) for both VO2 Max and CFR at study baselines in columns 3 and 4.  A next column shows the probable average cardiofitness achieved within each group during the period of the study assuming that cardiofitness of groups regressed from that at baseline toward the mean at an average of 16% per year.  Other columns show the risk ratios of higher cardiofitness groups as a percentage of those of the lowest cardiofitness groups for cardiovascular diseases and all-cause death, certain characteristics about each study, and the method by which cardiofitness values were measured. 

                                                                             Table 1

   Risk Ratios of Cardiovascular Disease and Heart Attack and All-Cause Death vs. Measures of Cardiofitness

The only confounding factor with cardiofitness of usual significance was cigarette smoking that usually was somewhat lower for fit than for unfit people. Thus the study risk ratios included in Table 1 were adjusted for differences in smoking for the majority of studies wherein this information was provided. These adjustments were small.  Associated changes of cholesterol and blood pressure were of the modest level usually produced by the exercise that improves cardiofitness. Adjustments for these and other factors sometimes included in statistical assessments would not be valid for inclusion in this casually based analysis, but inclusion of such adjustments would have little effect on the conclusions that follow.  Most studies reported death from CV diseases, and others reported incidence of myocardial infarction. The effects of cardiofitness on these disease categories appeared to be similar and results of such are combined in the plots that follow.

The plots of the data using the CFR at baseline in Figure 1 following show via the open characters the consistent effect of cardiofitness on CVD for the 7-8 year studies, extending over a near 10 fold range in risk.  All shorter-term studies confirm the large effect on risk accompanying large differences in CFR.  But results for the 16-17 year studies via the solid characters show as anticipated much smaller and less consistent associations of CVD risk with baseline cardiofitness.

                                                                        Figure 1

Only study 7-M for the Norwegian men shows longer term results approaching those for the shorter-term studies.  Interestingly, Norwegian men maintained VO2 Max and cardiofitness better with age than did other populations.(8)  Some studies did not include consistent measures of 5-95% variability for individual risk ratios, but this variability appeared similar to that from most other population health studies.  For example, a risk ratio of 0.50 could have limits from 0.25 to 0.85.  Thus the deviations of individual risk ratios values from the correlation line in Figure 1 are expected. But risk ratios for highest vs. lowest cardiofitness groups were highly significant, and the correlations of the results from multiple studies produce an overall significance much higher than that obtainable from any individual study. Although results on women were limited, men and women appeared to experience the same effect of CFR on risk, and although not shown here, men of differing ages also appeared to obtain similar results.

The correlations of risk ratios vs. CFR differences at baseline are:

For 5-8 year studies, with coefficients followed by 5-95% ranges  

          (1) Log risk ratio  =  -0.0365 (-0.033 to -0.041)  *  CFR   

                t  of  coeff  = 25.9     s =  0.22    r  =  0.982

                reduction in risk for an increase of 10 in CFR:  31%

For 16-17 year studies,

          (2) Log risk ratio  =  -0.0138  (-0.010 to -0.018)  *  CFR  

                  t  of  coeff  = 5.1    s =  0.30    r  =  0.837

                  reduction in risk for an increase of 10 in CFR: 13%

But as discussed, the real differences in cardiofitness of all groups studied and especially those that continued for up to 17 years must have been significantly lower than those measured at baseline. The correlations of risk ratios vs. more the probable actual CFR differences during the time were:

For 5-8 year studies, regression of cardiofitness assumed at 16% per year,

    (3) Log risk ratio  =  - 0.0640 (-0.058 to  -0.070)  * D-CFR 

              t  of  coeff  =  21.1   s =  0.27  r = 0.974

              reduction in risk for an increase of 10 in CFR:  47%

For all studies, 5 to 17 years, regression of cardiofitness assumed at 16% per year, 

          (4) Log risk ratio =  - 0.0601 (-0.053 to  -0.067)  * D-CFR

                     t  of  coeff  =  16.5    s = 0.33     r =  0.893

                      reduction in risk for an increase of 10 in CFR: 45%

Recognition of the probable change in average cardiofitness over the terms of the studies brings results for the longer term studies in approximate agreement with results from the 5-8 year studies.

Formula (1) shows that accepting only the differences in CFR at baseline for 5-8 year studies (and which must be too high) produced a 31% reduction in risk per 10 unit in CFR increase in risk. This very conservative valuation still produces a very large effect of cardiofitness on risk for achievable differences of 30 CFR in cardiofitness.  The more probable effect of cardiofitness via (3) identifies a 47% reduction in risk for each 10 CFR improvement.  This more probable basis (3) is consistent with that from (4) but still does not recognize that a larger proportion of events occurred toward the latter part of the studies when group cardiofitness differences probably were lower and age related risk was much higher.  Although not shown here, a recognition of this further reality produces a probable effect of cardiofitness larger than the 47% reduction per 10 CFR improvement  estimated from formula (3). 

Figure 2 following plots risk of CVD vs. probable actual levels of cardiofitness in CFR for the more meaningful shorter-term studies. The data plot as a straight-line semi-logarithmic relation with no indication of a limit either overall, or within the studies.  A ten-fold reduction in CVD risk is associated with an improvement in CFR of 33 that can be obtained by most adults via a very good program of aerobic exercise. The full differences in CFR among groups tested and listed in Table 1 ranged from below 80 to more than 150, a twice-higher range than that now measured vs. probable risk.  This suggests that the full range of differences in risk of CVD that are associated with measured differences in cardiofitness probably will be much higher than the approximately 10 fold suggested from data now available. Figure 2 quantifies the results of 13 comparisons of 26 different groups of individuals, most of the major research on physical fitness and CVD published during about two decades.  Note here that similar plots could be obtained using VO2 Max values at age 50 which is the average age involved in the research.

                                                          Figure 2

Other studies confirm the above large advantages for cardiofitness on CVD, but are not included above because the cardiofitness measures used could not be translated directly to VO2 Max and CFR by this author.  A five year study by Sobolski showed a 7 fold difference for lowest vs. highest cardiofitness.(31)  Another of five years by Peters showed advantages of up to 3-5 fold for groups having a probable 20 higher CFR.(25)   A seven year study by Arraiz on 31,668 Canadians found risk ratios ranging from 0.13 to 0.21 or 4.8 to 7.8 fold for differences in a measure of cardiofitness, whereas measures of physical activity on the same population identified risk ratios only of above 0.50 or less than 2 fold. (1)  No study of cardiofitness has been found published that did not show an advantage for those at high vs. low cardiofitness, and all studies of 8 years or less found higher risks of CVD in the range of 3-9 times for those least vs those most cardiovascular fit. 

The risk of heart disease increases about 10% each year of age. A reduction in risk of the 50% obtained in most studies of physical activity thus implies that this benefit will provide an average individual about 8 more years to the time when risk again becomes similar, a useful but inadequate benefit for one approaching a coronary attack age 50. The geometric relationship of risk and  cardiofitness suggests that a reduction in risk via an improvement in CFR of the 30% that is achievable from a maintained good fitness program may provide a younger person 24 more years of life to equal risk. Although such a program will be much more protective than one providing only 8 more years, a high cardiofitness still may not prevent CVD if  risks from other factors are very high or atherosclerosis is advanced. 

The above results do not confirm cited opinions that moderate amounts of exercise produces most of the health benefit potential from exercise.  Moderate exercise that usually produces an improvement in cardiofitness of 10% or less does produce a useful risk factor of about 0.5 for CVD.  But higher cardiofitness levels will produce increasingly larger and much more useful reductions in risk.

Results of Research Relating Physical Fitness to Death from All Causes

The most important of all results of any health action is the reduction of the risk of death from all causes.  The analyses of the risks included in Table 1 above follow:Figure 1 shows a plot of the the risks of death vs. estimates of the difference in cardiofitness in CFR of the groups studied. All comparisons are based on a value of 100% for the lowest cardiofitness.  As for Figure 2 these comparisons are all from the studies of 8 years or less wherein probable differences in cardiofitness are most accurate.

                                                  Figure 3 

A proposed minimal improvement in CFR that most people should be able to accomplish from exercise reduces risk of death by nearly 40%.  A improvement of 20 in CFR that is obtainable by most people from good cardiofitness exercise reduces risk to only 37% of sedentary value of 100 for a reduction in risk of 63%.  If this cardiofitness is maintained, this could translate into several added years of healthy life.   The effect of cardiofitness on death for the various shorter and longer term studies follow:

The effects of baseline fitness in CFR on all-cause death where values in () represent 5% to 95% ranges of the coefficient are:

  For 7-8 year studies 

              (5) log risk ratio = - 0.0273 (-0.022 to -0.032) * D-CFR 

             t of  coeff  = 10.8     s  =  0.316     r  =  0.96

              reduction in risk for an increase of 10 in CFR:  24%

  For 16-17 year studies

              (6) log risk ratio = - 0.00795 (-0.0042 to -0.0112) * D-CFR

                t of coeff = 4.9     s = 0.181     r  =  0.82

                 reduction in risk for an increase of 10 in CFR:  8%

The most probable effect of actual fitness in CFR on all-cause death assuming 16% regression of fitness differences per year for the 7-8 year studies 

              (7) log risk ratio  -0.0458 (-0.0374 to –0.0541) * D-CFR 

                t of coeff = 11.0     s  =  0.310     r  =  0.96

                reduction in risk for an increase of 10 in CFR:  37%

The same pattern thus was found for death from all causes as for cardiovascular diseases and death. The 16-17 year results show risk ratios much lower than those of the 7-8 year study results probably because true differences in cardiofitness were much lower over the long term than they were at baseline.  The most probable result of formula (7) shows that each increase of 10 units of fitness in CFR is associated with a 37 % (5-95% limits 31-43%) lower risk of death.  This in turn reflects an increase in life expectancy of about 3 years for an improvement of 10 in CFR for the included populations that averaged about age 50.  Even the very conservative estimate from formula (5) shows a 24% (5-95% limits 20-27%) lower death risk for each 10 CFR increase in fitness. Again, it appears that the prospective studies longer than 7-8 years although confirming benefits for cardiofitness are not useful for quantifying its value. Only three data sets in Table 1 provided results for the more useful shorter-term results on death from all causes. But the large 7-year follow-up by Arraiz of 31,668 Canadians showed risk ratios for death from all causes of 33% to 45% for measured higher vs. lower fit populations, consistent with the result of formula  (7).  A significant part of the all-cause deaths were from CVD, but an increasing body of research now confirms that the exercise that improves cardiofitness also reduces risk of cancer.

Some of the researchers expressed their result as a lowered risk of all-cause death of about 0.82 or a 18% reduction per Met change in cardiofitness.  At age 50 This is a slightly higher reduction in risk than obtained from the above formula #7.   A Met value can have  quite differeing significance for individuals of differing ages and gender.  The CFR method that uses age and gender adjusted values corrects for this error.

Values of relative risk of heart disease and all-cause death are shown vs. CFR from formulas 3 and 7 in Table 2 following. The values in italics are extrapolations beyond the range of the actual data available in Figures 1 and 2 and thus may or may not be valid.

                                                   Table 2

                              Effect of Current Cardiofitness in CFR on Risk

                                   of CVD and All-Cause Death (ACD)

Risks related to Cardiofitness can be  Much Larger than those related to Differences in Calories of Physical Activity.

A usual answer to this is that separation of groups by measured differences in cardiofitness will provide sharper identification of the true effect of exercise than was obtained via separation of groups using user reports about their physical activity.  A cardiofitness measurement provides an objective measurement of a condition of the human body that is absent in any estimate of “Calories of physical activity.”

Most studies of physical activity and risk of disease or death provided no information on cardiofitness of participants.  The American Sports College Guide data showed that the measured cardiofitness differences between so-called ‘physically active’ and ‘sedentary’ men and women of most ages were about 10% or 10 CFR. (2)   It is difficult to separate individuals into groups that have differences in cardiofitness much larger than this by other than direct fitness testing.  Substantial differences in the estimated calories of physical activity from occupations do not have a much effect on cardiofitness. (6,8)   Leisure time physical activity calories from questionnaires usually have explained only about 33% of the variance in measured levels of fitness. (10,30)   One probable reason is that calories of physical activity do not value correctly the effective exercise that produces cardiofitness that in turn reduces risk of disease. A value of gross calories of physical activity  assumes that calories at low intensity levels have the same effect on risk as do equivalent calories at more intense levels.  cardiofitness is produced much more effectively from high intensity levels of exercise than from low intensity exercise. Further, only steady exercise for a continued 2-3 minutes produces the elevations in heart rate that produce cardiofitness effectively. Most physical activities of exercise are obtained from shorter periods of duration.

Another reason is that cardiofitness probably is in part genetically endowed.  Studies at the Cooper Institute showed that some individuals in their low fitness groups did useful exercise, and that some in the higher fitness groups did rather little exercise.  Some individuals have larger diameter coronary arteries and more robust cardiovascular systems than do others.  A measurement of cardiofitness identifies an overall efficiency of this system that includes benefits of genetics plus benefits from physical activity from occupation, from normal lifestyle and from leisure exercise. Estimates of physical activity calories do not recognize this genetic factor.

A review of all other published studies of physical activity and heart disease showed that results were generally consistent with those expected from the cardiofitness hypothesis.  The study of Morris on 9300 men in England showed that only reasonably vigorous aerobic type exercise that is current and non-seasonal provided substantial protection against coronary attacks.(22)   The few study comparisons of kinds of exercise that should produce differences in fitness larger than usual showed larger reductions in risk of disease.  Morris’s results on a “Most vigorously exercising” group found a risk ratio for this exercising group of only 18% of that of the inactive group.(21)  The large NY HIP study showed significant reductions of up to 10 times in sudden death from CHD for the group most active both in occupation and leisure.(28)   Lakka showed a 71% reduction in coronary heart disease for conditioning type exercise.(14)

The Mechanism by which Cardiofitness Reduces Risk of Disease and Death

Cardiofitness reduces blood pressure by an amount that from the studies cited in Table 1 and independent research might reduce heart disease by 5% for an increase of 10 in CFR.  It reduces total cholesterol and improves HDL by amount that should reduce risk by another 2% and 3% respectively for this change of 10 in CFR.  These mechanisms can explain only a 10% change in benefit, a small part of the actually measured benefit of 47% from a gain of 10 in CFR.  Only one known mechanism now can explain quantitatively how improvements in cardiofitness can accomplish the large reductions in risk of cardiovascular disease and death from all causes found. The pulsing blood flow from exercise that improves cardiofitness also enlarges coronary arteries. This can reduce very substantially the risk that they will close up completely from a blood clot or ruptured plaque.

Atherosclerosis usually blocks arteries gradually during life. Consider as a simplified example a coronary artery of 3 mm. diameter that becomes blocked 90% at age 50.  The remaining open space will be 10% of initial.  An increase in artery diameter of just 15% can increase the area of the remaining open artery cross section by 4.2 times and area blockage will become only 68%.  Assuming that atherosclerosis blocks the radius of the artery at a consistent rate, it will take 11 more years for the artery to become blocked to the same 90%. If an artery increases in diameter by 50%, a similar computation shows blockage at year will drop from 90% to 40%, and it will take 36 more years of average atherosclerosis progression for the artery to become 90% blocked.

Mitrani found that the arteries of physically active men – who probably had the usual advantage of 10 in CFR – had arteries 15% larger than those less active(20).  The arteries of African Masai - a group of which were measured at a high average fitness in CFR of 150 - were near twice-usual size(18,19).  The research of Kramsch provided pictures of the much larger coronary artery cross sections obtained by monkeys that exercised moderately(13).   Not only was artery size increased about 70%, but the atherosclerosis was pressed back against the artery wall.  Increased blood flow through arteries creates a force that pushes them open that via engineering should be proportional to a summation of the squares of the increases in blood flow produced by the exercising heart over time. This straightforward physical mechanism alone can explain a larger benefit of cardiofitness on risk of CHD than that estimated from formula (3).

A genetic component also may contribute differences of +/- 10 CFR or more to cardiofitness for a given level of physical activity^.(7)  An otherwise similar person with similar habits but with 2.5 mm. rather than a 3 mm. coronary artery sections could via the mechanism proposed above incur equal CVD risk 12 years earlier, or that computed from a 15% lower CFR.

Cardiofitness is an Important and Independent Health Risk Factor.

Some comparative risks of death from key coronary risk factors are:

  Cardiofitness        Serum Cholesterol       Blood Pressure           Smoking

  CFR        Risk             Total  Risk              Sys/Dia    Risk        Cig/Day Risk

  130         0.15               180    0.58              125/75      0.76              0         1.00

  120         0.38               190    0.67              135/85      1.00              5         1.42

  110         0.53               200    0.78              160/100    1.62            15        1.74 

  100         1.00               220    1.00              170/110    2.00            30        1.98

   90          1.90               250    1.46                                                     40+      2.36

   80          3.50               300    3.00  

In addition to the larger potential effect of cardiofitness on risk of death, perhaps 75-80% of today's US population could and should improve their cardiofitness.  Only 25% of our population now smokes cigarettes.  A still smaller fraction suffers substantially elevated cholesterol and modern drugs can substantially reduce both this risk and the levels and risk of very high blood pressures.  Only the proper exercise will improve cardiofitness.

Observations and Conclusions

Cardiofitness in CFR is a major health risk factor that is not now usually identified.  An increase in CFR from 90 of poor cardiofitness to the 120 achievable by individuals that did more exercise should reduce risk of death by more than three fold and risk of CV diseases by six-fold at age.

A separate paper shows that values of the CFR probably can be obtained with useful accuracy from a proposed convenient treadmill test.(7)  Individuals can determine their present level of cardiofitness and determine how this fitness and its associated reduction in risk can progress in response to their own exercise programs.  Efficient exercise programs should be designed to improve cardiofitness from a measured value to a desired goal.  The proposed measure of CFR and the above relationships between CFR and CVD disease and death provide substantial incentives for both understanding current levels of an individual's cardiofitness and for developing programs for developing and confirming its improvement.

References:

 1.  Arraiz, GA, Wigle, DT and Yang Mao (1992).  Risk assessment of physical activity and 
      physical fitness in the Canada Health Survey. J. Clin Epidmiol 45:419-28

 2.  American College of Sports Medicine (1990) ACSM’s Guidelines for Exercise Testing

      and Prescription, 4th Edition,  Lea and Febiger,  pp 148.

 3.  Berlin, JA and Colditz, GA (1990)  A meta-analysis of physical activity in the prevention of
      coronary heart disease.  Am J Epidemiol, 132:612-28.

 4.  Blair, SN, Kohl, HW, et al (1989)  Physical fitness and all cause mortality:  a prospective 

      study of  healthy men and women.  JAMA, 262:2395-

      2401.  Supplemented by data of time on treadmill by group supplied per request.

 5.  Blair, S., Kohl, HW III et al (1995): Changes in physical fitness and all-cause mortality.  

      JAMA 273:1093-98

 6.  Blanding, FH., Introduction and review by Fox, SM (1982), The Pulse Point Plan  Random

       House, New York, NY  pp 170

 7.  Blanding, FH. (2002)  A useful measure of physical fitness.   Draft to Internet at

      http://lifeahead.net

 8.  Cumming, GR (1967) Current levels of fitness. Can Med J 96:869 

 9.  Ekelund LG, Haskell,WL et al (1988)  Physical fitness as a predictor of
        cardiovascular mortality in asymptomatic North American men: The Lipid Clinics

        Mortality follow-up study.  N Eng J Med, 319:1379- 84

10.  Elosua, R. et al (2000) Validation of the Minnesota leisure time physical activity

       questionnaire in Spanish women.  Med Sci Sports Exerc  32(8):1431-37

11.  Erickssen, G. Liestol, k. et al (1998) Changes in physical fitness and changes in mortality. 

        Lancet 232(9130):759-62     

12. Hein HO,  Suadicani,P. & Gyntelberg F. (1992)  Physical fitness or physical
        activity as a predictor of ischaemic heart disease.  A 17 year follow-up in the
       Copenhagen male study. J Intern Med,  232:471-9

13. Kramsch, D.M. Aspen, A.J., et al. (1981) Reduction of coronary atherosclerosis by

        moderate conditioning exercise in monkeys on an atherogenic diet.  N Eng J Med 

       305:1483-89

14. Lakka TA, Venalainen, JM et al (1994)  Relation of leisure-time physical activity and

       cardiorespiratory fitness to the risk of acute myocardial infarction. N Engl J Med

       330(22):1549-54

15. Leon, AS, Connett, J. et al (1987)  Leisure-time physical activity levels and risk of

       coronary heart disease and death.  JAMA 258:2388-95

16. Lie, H, Mundal, R and Erikssen, J. (1985) Coronary risk factors and incidence of 
       coronary death in relation to physical fitness.  Eur Heart J, 6, 147-57
 

17. Makrides, L, Heigenhauser, JF and Jones, NL (1990) High-intensity endurance training in
       20-30 and 60-70 year old healthy men.  J Appl Physiol 69(5):1792-98.

18. Mann, GV, Shaffer, RD, Rich, AB (1965) Physical fitness and immunity to heart disease 

       in Masai.  Lancet II:14308-10.

19. Mann, GV, Spoerry, A. et al.(1972)    Atherosclerosis in the Masai.  Am J Epidemiol

       95: 26- 37

20. Mitrani, Y, Karplus, H. & Brunner, D. (1970) Coronary atherosclerosis in cases of traumatic

       death.  Medicine and Sport Vol. 4, Physical Activity and Aging: pp 241-48. Baltimore

       University Park Press.

21. Morris, JN, Chave, SPW et al (1973). Vigorous exercise in leisure-time and the incidence

       of coronary heart disease.  Lancet 1:333-39.

22. Morris, DG, Clayton, MG et al (1990)  Exercise in leisure-time: Coronary attack and death

        rates.  Brit Heart J 63:325-34

23. Paffenbarger, RS Jr, Wing, AL & Hyde, RT (1978) Physical activity as an index of heart

       attack risk in college alumni. Am J Epidemiol 108:161-75   (note: for sports example)

24. Paffenbarger, RS. Jr and Hsieh, C (1992)  Physical activity and risk of prostatic cancer

        among college alumni.  Am J Epidemiol 135:169-79 

25. Peters, RK, Cady, LD jr, et al (1983) Physical fitness and subsequent myocardial

        infarction in healthy workers.  JAMA 249:3052-56  

26. Sandvik, M., Erikssen, J et al (1993) Physical fitness as a predictor of mortality among

        healthy, middle-aged Norwegian men. N Engl J Med  328:553-7

27. Seals, DR Hegberg, JM et al (1984)  Endurance training in older men and women.  J Appl
       Physiol 57(4):1024-9

28. Shapiro, S, Weinblatt, E. et al (1969) Incidence of coronary heart disease in a population

        insured for medical care (HIP). Am J Publ Hlth 59: Suppl to June.

29. Sherman SE et al (1999)  Comparison of past versus recent physical activity in the

        prevention of premature death and coronary artery disease.  Am Heart J 138(5 Pt 1):900-7

30. Slattery, ML Jacobs, DR Jr. (1988)   Physical fitness and cardiovascular disease
        mortality; the US  railroad study.  Am J Epidemiol, 127:571-80

31. Sobolski, J,  Kornitzer, M. et al. (1987)  Protection against ischemic heart disease in the    
        Belgian physical fitness study: Physical fitness rather than physical activity?  Am J     
        Epidemiol 125:601-10

32.  Stensel, DJ, Brooke-Wavell, k,  et al (1994)  The influence of a 1-year program of brisk
        walking on endurance fitness and body composition in previously sedentary men aged      
        42-59 years.  Eur J Appl Physiol Occup Physiol 68(6):531

33. Talbot, LA et al (2000)  Leisure-time physical activities and their relationship to    
        cardiorespiratory fitness in healthy men and women 18-95 years old.  Med Sci Sports
        Exerc 32(2):417

34. Mora, S, Redberg, RF  et al  (2003)  Ability of Exercise Testing to Predict  Cardiovascular and All-Cause Death,

      JAMA 290:1600

35  Gulati, M, Pandey DK et al (2003)  Exercise Capacity and the Risk of Death in Women. 

      The St James Women Take Heart Project:  Circulation 108:1554-59

36  Myers, J, Kaykha, A, et al  (2004 )  Fitness versus Physical Activity Patterns in Pedicting 

      Mortality in Men.  Am J Med 117:912-18