Main Menu                  Health Library              Exercise and Cardiofitness  

 The CFR:  A USEFUL MEASURE of PHYSICAL FITNESS  

 

Forrest H. Blanding,  CHE                             Draft to Internet at http://lifeahead.net

                                                                      Jan 2, 2004.  Reviews and  Comments appreciated to forrestx@cox.net.

                                                       Updated July 9, 2007 by author.

Abstract:

 

Cardiorespiratory or Cardiovascular (CV) physical fitness abbreviated as Cardiofitness and measured by exercise tests has been related importantly to risk of cardiovascular disease and premature death, and is a health risk factor of substantial significance.  Yet no measure of cardiofitness suitable for use by both researchers and the public is in use.  A proposed index of CV physical fitness called Cardiofitness ratio or CFR  is easily interpretable and can be obtained with useful accuracy from either a measurement of VO2 Max in ml of oxygen/kg.min or from a described sub-maximal treadmill test of moderate intensity.  The CFR is shown to classify properly widely varying cardiofitness  levels of groups of men and women of different ages.  The variation of CFR within populations is shown, and data are summarized showing how the CFR develops from lifestyle physical activity and programmed exercise.  The CFR provides a useful basis for identifying how risks of disease and death relate to levels of cardiofitness ratios.  The proposed index provides a simple and practical basis for individuals to identify their present fitness and general health risk, to set objectives for exercise programs to reduce that risk, and to learn how their cardiofitness improves with exercise.  Other methods for measuring the CFR that should be widely useful for adult populations are proposed.

 

Introduction:  Physical fitness can refer to the physical state of a wide variety of body muscles. But the key physical fitness component most often related to risk of health is cardiovascular or cardiorespiratory fitness abbreviated herein as cardiofitness that identifies the fitness of the heart and its related system of arteries or cardiorespitory fitness that also appraises the respiratory system.  Cardiofitness has been related more consistently and more importantly to risk of cardiovascular diseases and death from all causes than has any other measure of physical activity or exercise (Arraiz et al 1992, Berlin and Colditz; 1990; Blair et al 1989; Ekelund, et al 1988;  Hein et al 1992; Lie et al 1985;  Slattery and Jacobs, 1988).  But measures of cardiofitness differed in each of these studies, and no consistent measure of cardiofitness useful to both the lay public and the scientific community that is verified as health related is in general use. An important need is for an index of cardiofitness that can be related to the risks of major disease and death identified in differing population research studies. Without such an index people have had no practical way of learning what their cardio fitness is now, how it improves as they exercise, or what improvement in this cardiofitness might mean to their risk of major disease and death.

 

The popular scientific measure for identifying cardiofitness is called VO2 max, the ml/kg.min of oxygen consumed at an individual’s maximum possible exercise. A value called Mets that refers to the number of energy equivalents of the energy required at rest also is used. VO2 Max often is taken as 3.5 times Mets.   Unfortunately, the VO2 Max and Met  is not a method potentially useful to the public. Although it does reflect the effect of exercise, its results also are substantially age and gender dependent.  Table 1 summarizes data (American College of Sports Medicine 1990) showing the VO2 max values of ‘sedentary’ and ‘physically active’ men and women.  Because of the dominating effect of age, a sedentary man of age 40 can have a higher VO2 max than a so-called active man of age 50, and values differ for men and women.  The use of such a cardiofitness measure that requires comparison of each value to extensive tables of targets for age and gender will have minimal value to the public.  A VO2 max measurement also requires an expensive test that involves some if small risk of a heart attack or death.

                                                       

                                                         Table 1. 

     Average VO2 max Values in mg-kg.min for Healthy Active

                    and Sedentary Men and Women vs. Age

 

Age

Men

          Women

 

Active

Sedentary

Active

Sedentary

20

57.5

48.9

36.7

35.2

30

51.3

44.5

33.5

31.6

40

45.2

40.0

30.4

28.3

50

39.1

35.3

27.3

24.5

60

33.0

31.1

24.2

20.9

Some Alternates as Measures of cardiofitness:  VO2 max is an approximate measure of the total energy the heart can produce.  It derives largely from a combination of a stroke volume of the heart per beat (or more accurately, the oxygen delivered per beat) and the maximum rate that the heart will beat.  Exercise and particularly CV exercise improves the efficiency of the heart by increasing stroke volume and oxygen transported per beat.  Maximum heart rate depends mostly on age, and is little affected by exercise.  Thus a useful measure of cardiofitness should identify selectively the efficiency factor that is improved by exercise.  The contribution of age that affects every aspect of health and differences by gender then can be adjusted for separately then when needed.  A most useful measure would be an easily understandable index that identifies similar meaningful values for the cardiofitness of men and women of all ages, and that perform the similar amounts of CV exercise needed to obtain this efficiency.

 

There are alternates for solving this problem.  A straightforward method is (1) to simply take a ratio in percentage of actual VO2 Max to values for sedentary men and women as from Table 1.  A problem here is that the average amounts of exercise people perform decline with age, and thus a common index of 100 will identify differing amounts of exercise for individuals of different ages.  Method (2) is the use of heart stroke volume or VO2 per pulse as a basis, as this is the factor improved by exercise, and stroke volume varies little with age (Shepard, RJ 1994).   But the use of a computed delivery per beat involves need for a resting heart rate that introduces potential complexity and error.  A method (3) proposed earlier (Blanding, FH 1982) called the HPI was an empirical adjustment to VO2 max that produced approximations of heart stroke volume.  A method (4) and that is used herein involves obtaining the percentage of VO2 max to fixed values for typical sedentary men and women of age 50, and inserting a correction that removes the contribution of the age-related maximum heart rate from VO2 max.  This method called Cardiofitness ratio or CFR (it formerly was called the HPI or HEF) also produces values that are approximately proportional to heart stroke volume and VO2 per pulse but has a definable basis for age. 

 

Values of 33 VO2 Max and 25.3 VO2 Max are selected herein as average values for sedentary men and women of age 50 based on a review of multiple studies of VO2 Max values of sedentary populations.  Using these arbitrary values with each of the above methods will produce the same simple result at age 050 as:  A cardiofitness of 100 becomes simply VO2 Max times 3.03 for men and VO2 Max times 3.95 for women.  Because most research considered population groups of ages near 50, the exact method selected for a cardiofitness index or the alternative of using VO2 Max directly will have little effect on the relationships between cardiofitness and disease or death from population studies.

 

Importantly, the CFR from method (4) above produces numerical values for cardiofitness that have a similar significance for men and women, and remain as useful values for persons most ages. Younger people usually will have somewhat higher average values of CFR than those older in part because they usually exercise more.  And as will follow, CFR values are easily interpretable as to meaning and can be measured usefully from simple and convenient sub-maximal exercise tests.

 

Values of the CFR for adults of age 20 and older can be obtained from any measurement of VO2 max from the following formulas:

 

For Men

            (1)        CFR = VO2 max * 100 / (33.0 + (50 - age) * 0.24)

For Women

           (2)         CFR = VO2 max * 100 / (25.3 + (50 – age) * 0.18)

 

With this simple straightforward method a value of 100 designates arbitrarily by definition the cardiofitness of average sedentary men or women of age 50 having the above 33 and 25.3 VO2 max levels respectively. Values of CFR are proportional to VO2 max, and thus maintain the integrity of the VO2 max measure.  The correction for age modifies the average sedentary value for VO2 max for an effect of age and the CFR thus values cardiofitness approximately as a percentage of usual ‘sedentary’.  A useful benefit of the CFR is that values below 100 identify deficiencies in cardiofitness, and values above 100 show increasingly favorable levels of cardiofitness for all adults.  An CFR value as a simple percentage number immediately identifies a level of cardiofitness that is directly related both to amounts of cardio effective exercise and that can be related to risks of major disease.  The use of 100 as a base for the CFR essentially converts values of cardiofitness as\\to percent changes that are far easier for people to understand than are the fractional numbers needed when using the direct scientific values.

 

Applicability of the CFR to a Large Population:   A unique set of data on CV or CR fitness (Blair, SN et al 1998) on about 18,000 men and 5700 women provides a basis for verifying the usefulness of the CFR measure. Values were developed from the popular Balke exercise test on each individual, and average values of populations were separated by sex and cardiofitness levels for each age. (Balke B, 1969) cardiofitness level groups for each sex and age included the lowest 20%, the 20%-60% group, and the highest 40% of each population. VO2 max values and CFR values for the average cardiofitness in each age group for both men and women are computed from this data set in Table 2. The values of VO2 max decline about 30% with age in each group, overlap from group to group, and are much different for men and women.  But the CFR values provide a clearer classification of the cardiofitness groups for all ages of both men and women. The modest declines of about 10% in CFR with age are about those expected from usual lesser physical activity with age. Also note that the sub-maximal heart rates in beats per minute at the 5% grade in the Balke test that provide a direct measure of CV efficiency are quite consistent within each cardiofitness group for both men and women, are sharply different for successive cardiofitness groups, and vary modestly and inversely with age in accord with the modest declines in CFR with age.  The risk ratios for cardiovascular death and death from all causes from the primary study on this population (Blair, SN et al 1989) shown in Table 2 decline similarly and substantially for both men and women with increases in cardiofitness in CFR.

             

                                                                  Table 2.

 

     Values of cardiofitness as VO2 max in ml/kg.min and CFR Estimated from Treadmill Tests and

            Heart Rates in Beats per Minute, Exercise Calories,  and risks of CVD and Death.

                 

     Results for MEN

                         Low cardiofitness               Moderate cardiofitness              High cardiofitness               

 

    Age

VO2

Max

 

CFR

5 Min

Heart

Rate

 

VO2

Max

 

CFR

5 Min

Heart

Rate

 

VO2

Max

 

CFR

5 Min

Heart

Rate

   20-39

 33.6

 91

 131

 

 40.6

110

116

 

 49.7

138

103

   40-49

 31.5

 92

 130

 

 38.5

113

116

 

 47.2

140

103

   50-59

 28.0

 87

 133

 

 35.0

109

118

 

 44.1

137

104

    60+

 23.5

 78

 137

 

 30.5

102

121

 

 39.9

133

107

 

    Risk

    Ratios

All Cause Death      1.0

Cardiovasc Death   1.0

 

All Cause Death    0.41

Cardiovasc Death 0.32

 

All Cause Death   0.32*

Cardiovasc Death 0.13

 

 

Exercise Calories/Week

              260 

 

Exercise Calories/Week

            525

 

Exercise Calories/Week

               1650

 

     Results for WOMEN

                      Low cardiofitness               Moderate cardiofitness              High cardiofitness

   20-39

 25.2

 89

151

 

 32.9

114

136

 

 43.4

153

119

   40-49

 22.8

 87

151

 

 29.8

114

138

 

 39.9

153

120

   50-59

 19.6

 80

152

 

 25.9

105

139

 

 35.0

142

124

     60+ 

 

 17.9

 77

ns

 

 23.1

100

140

 

 31.5

137

126

Risk

ratios

 All-Cause Death     1.0

 Cardiovasc Death  1.0

 

 All-Cause Death     0.41

 Cardiovasc Death  0.39

 

All-Cause Death   0.18*

Cardiovasc Death 0.11

 

 

Exercise Calories/Week

              200

 

           

Exercise Calories/Week

                420

 

          

Exercise Calories/Week

            1260

 

 

 

 

 

 

 

 

 

*    5-95% confidence limits about 0.17 – 0.49 for men; 0.10 - 0.45 for women.

      Exercise Calories probably selectively from aerobic type exercise

 

Regressions of the values in Table 2 show that the CFR can be estimated with very high accuracy from the average group sub-maximal heart rate in beats per minute at the 5-minute and 5% grade (HR5) at the speed of 88 m/min or 3.3 miles per hour in the Balke test for both men and women: 

 

 (3)   For Men:         CFR = 318.2 – 1.753 * HR5. (5-95% values of coefficient 1.620-1.886),

                                  s = 2.4%,  r = 0.99,  t of coefficient = 26.3

 

 (4)   For Women:  CFR = 406.7 – 2.139 * HR5  (5-95% values of coefficient 2.040-2.238), 

                                  s = 3.1%,  r = 0.99,  t of coefficient = 25.1                                              

Note that these regressions are of highest significance with correlation coefficient of 0.99 and very high values of the 't' statistic. The associated effect of age was of nil significance (t = 0.01). This suggests that the age correction used in the CFR produces equivalent CV efficiency values for ages from 30 to 65+ for both men and women.  A qualification is that these relationships are verified only for individuals to about age 65, and that results for older individuals will require further verification.

 

Similar but no better correlation coefficients and standard errors were obtained using the higher heart rates associated with higher treadmill inclines of 10%, 15% and 20%.   Thus no advantage is indicated for use of such higher risk exercise loads and heart rates.  An attempt to classify cardiofitness levels using VO2 max values directly rather than CFR values produced no better correlation accuracy and required differing correlations of VO2 vs. cardiofitness level for both sex and each age group, as the added effect of age was very significant  (t  = 10). 

 

Formulas relating VO2 Max to 5 minute heart rates are:

 

 7)  For men      VO2 Max = 113.4 - 0.5638 * HR5 - 0.201 * age

                           s = 0.89; r=0.99;    t of HR5 coefficient = 25..3

 

(8) For women:  VO2 Max = 113.9 - 0.5438 * HR5 - 0.212 * age

                          s = 0.96;  r=0.99;     t of HR5 coefficient = 23.9

 

Note that for predicting VO2 Max a quite substantial and significant (t = 9.5) effect of age is needed.  But with this, formula is quite robust.

 

There long has been uncertainty regarding the accuracy of estimating VO2 max from sub-maximal heart rates.  One problem has been the fact that VO2 Max values vary far more with age than do sub-maximal heart rates, and thus no simple direct relation between a sub-maximal heart rate that applies to individuals of various ages or gender can exist.  The near perfect above correlations of the CFR with sub-maximal heart rates are accomplished because the CFR is an age-adjusted VO2 max that largely eliminates this problem.  But because the CFR also by definition relates to VO2 max, it becomes possible via this approach to first estimate an CFR from the sub-maximal heart rates from (3) or (4) and then compute the age-dependent VO2 max from the mathematical  translation of formulas (1) and (2) as:

 

  (5)    For Men:               VO2 max = CFR * .01 * (33.0 +  (50 - age) * 0.24)

 (6)    For Women:          VO2 max = CFR * .01 * (25.3 +  (50 - age) * 0.18)

 

This combination approach provides a method for reconciling the VO2 Max values with sub-maximal heart rates that is verified at high accuracy for the averaged values from this data set.  Note also that this data set is far larger and more extensive than is any prior data set relating submaxmimal exercise test values to maximal values.

 

The fact that the CFR (or VO2 Max) can be estimated usefully from a simple sub-maximal exercise treadmill test means that the expense and risk involved in determining maximum cardiofitness values is not required for an estimation of cardiofitness of a healthy person.  The above 5-minute heart rate in beats per minute on the Balke test that correlated with both CFR and VO2 max at r = 0.99 requires a 5% treadmill grade at 3.3 mph or 5.3 kph speed.  Although this test uses an increasing incline of 1% each minute, heart rates reach near equilibrium levels after 3 minutes at most loads. Thus similar heart rates and results from the above formulas probably could be obtained from a simple 3 minute treadmill test walking at 3.3 mph (5.3 kph) and 5% grade without the stepwise increases in grade, perhaps with some early warm-up.  This test can be performed easily in any exercise facility or even by most individuals themselves on a quality home treadmill with precautions to be noted later.

 

Results then can be read from a simple table relating CFR values to exercise heart rates as per example Table 3 which can be expanded using formulas (3) and (4).  A suggestion for a potentially more widely useful test method that should produce equivalent results for individuals of nearly all cardiofitness levels more acceptably and with lesser risk is provided later.

 

                Table 3

 

CFR from Heart Rates in beats per

minute on treadmill  at 3.3 mph or

    5.3 kph and 5% incline

 after walking for 3 minutes

 

Heart Rate

At Grade

 Men

 CFR

Women

  CFR

150

-

  86

145

-

  96

140

  73

107

135

  81

117

130

  90

129

125

  99

139

120

107

150

115

117

-

110

125

-

105

134

-

100

143

-

95

151

-

  

For perspective, the values obtained from this analysis are based on a comprehensive database that is unlikely to be duplicated.  This base of 23,700 tests included cardiofitness ranges in CFR from about 80% to 140-150% of sedentary for both men and women of ages ranging from below 30 above 65.  This will include the cardiofitness levels of most involved adults in our population.  Although the above regressions (3) and (4) were based on averaged data values at each from each of the cells in Table 2,  each of these values were averages of several hundred actual VO2 max measurements that should in turn provide a high level of precision.  This explains the unusually high correlation coefficient obtained.  The expected deviations of individual CFR and VO2 Max values from exercise tests on individuals will be higher than these averaged values.  

 

An assumption in exercise testing often has been that the VO2 max value is of key interest, and that a sub-maximal exercise test provides only a way for approximating this value. A maximum test often is used and felt needed as a medical procedure. Yet it is possible that a sub-maximal exercise test may measure pertinent cardiofitness and its related health risk more accurately for most healthy people than will a maximum test. The sub-maximal test identifies more directly the relative efficiency of the CV system and its capability of transporting blood and oxygen at heart rates more consistent with those that are involved in both usual lifestyle and in most actual heart attacks. Also, the result of a maximum test can be in error if a user does not exercise to maximum.

 

A Proposed Exercise Test for cardiofitness:  An exercise test indicated from the above analysis is:

 

Obtain an individual’s heart rate after 3 minutes of walking on a treadmill operated at a speed of 3.3 mph (5.3 kph) at 5% grade.  Convert this heart rate via expanded tables in format of Table 3 to a value of CFR for men or women. This test is herein identified as Test C.   

 

The relationships in Table 3 are essentially duplicates of the relationships found from the same data in Table 2.  This provides a clear bridge of relationships at highest significance between the CFR measure of cardiofitness from sub-maximal test, values of VO2 Max, and actual measured risks of heart disease and cancer. It seems unlikely that such a series of such direct relationships on such a large data set will be duplicated in the foreseeable future.

 

More than 20,000 men and women from ages 20 to beyond age 60 took an approximate equivalent of this Test C successfully in the Cooper Institute.  This suggests that most people can take this test. But these measurements all were made in a monitored facility.  And Text C is demanding for many women and for older men having low cardiofitness.  It is not an appropriate test for a individual having any diagnosis of heart disease.  Thus adults that take this test should have prior approval from a doctor and desirably should take the test in a monitored exercise facility. 

 

A Proposed New Test Format:  Although the above Test C should be provide a useful measurement of cardiofitness, it would be desirable to establish a single standard test protocol for cardiofitness that is as convenient as possible, and that can be taken safely by nearly everyone. Although any test method that provides a result as VO2 max could be used for measuring the CFR, the use of differing methods could produce somewhat differing results and cause confusion.  A study of many different exercise options was made to explore how a more convenient test protocol could be constructed.  The conclusions from this study follow

 

A test that will produce minimum risk will use lowest feasible testing heart rates. A data set relating heart rates to Mets of energy (1 Met equals 3.5 ml/kg.min) on 362 men showed that at heart rates above 30-35 beats per minute or 40-50 % above resting heart rates the hearts of most men produced a consistent increase in energy output per pulse (Spangler, RD et al 1970).  This identifies the minimum increase in heart rate needed, and there may be little advantage for testing at higher heart rates than this.  The hearts of most people fully fill and pump at near constant volumes per pulse at rates above this level. Results discussed above showed no advantage for using higher heart rates than those from the 5% grade and 3.3 mph (5.3kph).  Exercise heart rates of only 103 to 137 beats per minute at the 5% grade were obtained by many in the groups of men included in Table 2.

 

Thus a useful cardiofitness test should require exercised heart rates of at least 30-35 beats per minute above resting heart rate.  And a heart rate of no more than 50 beats per minute above resting heart rate would be a useful target to keep health risk from the test to a minimum.  A study of probable exercise heart rates in beats per minute at various treadmill loads for men and women of widely differing cardiofitness levels using basic relationships of VO2 vs. speed and incline showed that no single exercise load could provide a test basis for men and women of all levels of cardiofitness that would meet these guides. (Formulas in Exercise Guide of the American College of Sports Medicine, published by Lea and Febiger, 1975).  

 

Using a desired heart rate range of from 35 to 50 beats per minute above resting rate, a protocol of two tests was computed as optimum.  A ‘Test A’ for most women would involve a heart rate in beats per minute obtained walking for 3 minutes on a treadmill at 3.5 mph (4.82 kph) at level grade.  A ‘Test B’ for most men would require a heart rate after walking for 3 minutes on the treadmill at 3.5 mph (5.65 kph) and a 2% grade. Reasonably fit women could use test B, and men of poor cardiofitness could use test A.   Where applicable, the two tests should produce similar results.

 

CFR values would be read directly from simple tables of CFR vs. heart rates obtained in these tests expanded from the illustration in Table 4 following.  Tests such as one on a calibrated bicycle could be added to the protocol for those with partial physical disabilities, a test at higher incline might be useful for very fit men and athletes, and one of lower level might be needed for those few older and of very poor cardiofitness that cannot walk at 3.5 mph. An objective for such a test protocol would be one convenient for use in research studies of exercise and as accompaniments to a doctor's usual physical examination.  

 

Complete tables of probable values of CFR from heart rates from these proposed tests have been estimated and formulas and further suggestions about developing and verifying best values for these tests can be provided.  But further research is needed to verify these values of heart rate vs. those monitored from Test C above, or from VO2 Max via the above formulas (3) and (4).  It is felt that using results from Test C above rather than complete VO2 Max tests as a monitoring basis for developing these more convenient  tests for cardiofitness in the CFR should be adequate and involve much less expense.  Such testing must verify that unexpected problems do not occur in these low level tests.

 

Any cardiofitness test must be conducted at least 3 hours after any meal or food intake, consumption of alcohol, tobacco, or caffeine. There should be no major exercise preceding the test during the day.  Medications that affect heart rates may confuse test results and this should be recognized. Treadmill grade must set accurately. Exercised heart rates must be measured accurately because a difference of 2 in CFR can be computed from a difference of only 1 beat per minute in heart rate. With such precautions cardiofitness measurements with a usual accuracy of  +/- 6 to 7% might be achievable. Although the conventional medical precautions should be taken for those with low cardiofitness or possible disease, the above proposed tests use heart rates that will be obtained commonly by most people during brisk walking and thus should nearly always be safe.  For individuals that exercise regularly, are reasonably fit, and are medically approved the use of Text C should be preferred. 

 

A Simple Method for Identifying Cardiofitness as the CFR should be Widely Available.   cardiofitness has been a subject of publicity since before 1970. It was shown from an extensive study of research to be a factor relating to disease and death of high significance as early as 1982 (Blanding, FH 1982).   Several studies cited above have now confirmed that poor vs. good measured cardiofitness is associated with differences in risk ratios of cardiovascular diseases ranging up to 4 to 9 times. There is a need for individuals in the public to be able to identify their cardiofitness in terms of test that provides as result a simple, easily measurable and understandable single number. The CFR provides a method for accomplishing this that maintains integrity with established scientific measures of VO2 Max in ml/kg.min.  There is a serious need for researchers to develop and confirm a suitable standard protocol for measuring cardiofitness that can be used by everyone. The CFR provides a suitable measure for this needed method.

 

                                                       Table 4

 

                Example of CFR Values and Heart Rates (beats per minute)

                       Estimated from Proposed Standard cardiofitness Test

                           These estimates require further confirmation

 

                       Values for Men                             Values for Women           

               Test A                    Test B                  Test A                   Test B

Heart

 Rate

CFR

 

Heart

Rate

CFR

 

Heart 

Rate

CFR

 

Heart

Rate

CFR

 115

  70

 

130

  71

 

130

  70

 

140

  81

 110

  78

 

125

  78

 

125

  77

 

135

  87

 105

  86

 

120

  85

 

120

  84

 

130

  95

 100

  97

 

115

  93

 

115

  92

 

125

101

   95

107

 

110

102

 

110

102

 

120

109

   90

119

 

105

112

 

105

111

 

115

118

   85

132

 

100

122

 

100

122

 

110

128

   80

145

 

  95

134

 

  95

133

 

105

138

   75

160

 

  90

145

 

  90

145

 

100

150

 

 

References:

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

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

      Arraiz, GA,  Wigle, DT, and Mao,Y. e(1992)  Risk assessment of physical activity and

      physical fitness in the Canadian Health Survey followup study.  J  Clin  Epidemiol,

      45:419-28

      Balke B and Ware RW.  An experimental study of physical fitness in Air Force personnel. 

      US armed Forces Med J 1959:10:675-688

      Berlin, J.A. and Colditz, GA. (1990)  A meta-analysis of physical activity in the

      prevention of coronary heart disease.  Am J Epidemiol, 132:612-28.

      Blair, S.N., Kohl H.W. et al (1989)  Physical fitness and all cause mortality:  a 
      prospective study of healthy men and women.  Journal of the American Medical 
      Association  262:2395-2401.  Supplemented by data on time on treadmill by group

      supplied per  request.

      Blair, S.N., Ming, W., and Chong, D.L. (1998).  Cardiorespiratory fitness determined by

     exercise heart rate as a predictor of mortality in the Aerobics Center Longitudinal Study.      

     J Sports Sci 16, S47-S55.  

     Blanding, F.H., intro by Fox, MD. (1982) The Pulse Point Plan, Random House,

     New York, NY

     Blanding, FH Physical fitness and the risk of cardiovascular disease and death.    

     Draft to Internet at http://lifeahead.net             

     Ekelund L.G., 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

     Hein H.O.,  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

     Kohrt W.M., Malley, MT et al.(1991)  Effects of gender, age and fitness level on

     response of VO2 max to training in 60-71 year olds. J Appl Physiol ,71(5):2004-11 

     Calculation of the relative energy expenditure for various

     activities in Mets.  Supplement to Guidelines for Graded Exercise Testing, 

     American College of Sports Medicine. Publisher  Lea and Febiger, (1975)  

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

      Peters, R.K.and Cady, L.D. (1983)  Physical fitness and subsequent myocardial

      infarction in healthy workers.  JAMA, 249:3052-56

      Shepard, R.J.  (1994) Aerobic Fitness and Health, Human Kinetics, Champaign, 

      IL  pp. 236-7

      Shepard, R.J. (1997)  Aging, Physical Activity, and Health.  Human Kinetics,

      Champaign, IL., pp. 89-90, 157

      Slattery, M.L. Jacobs, D.R. Jr. (1988)   Physical fitness and cardiovascular disease
      mortality; the US  railroad study.  Am J Epidemiol, 127:571-80

      Sobolski, J. et al. (1987)  Protections against ischemic heart disease in the 

      Belgian physical fitness study  Am J Epidemiol, 125:601-10

      Spangler, R.D. et al (1970)   A submaximal exercise electrocardiographic test as

      method of determining occult ischemic heart disease. Am Heart J 80:752-58

      Stofan, M.S.,  et al. (1998)   Physical activity patterns associated with

      cardiorespiratory fitness and reduced mortality:  The Aerobics Center longitudinal

      study.  American Journal of Public Health, 88:1807-13