Predicting max HR and the HR-VO, relationship for exercise prescription in obesity WAYNE C. MILLER, JANET P. WALLACE, and KAREN E. EGGERT Department of Kinesiology, Indiana University, Bloomington, IN 47405 ABSTRACT MILLER, W. C,, J, P. WALLACE, and K. E. EGGERT. Predicting max HR and the HR-VO; relationship for exercise prescription in obesity. Med. Sel. Sports Exerc, Vol. 25, No. 9, pp. 1077-1081, 1993, This research derived regression equations for predicting max- imal heart rate (MH) and examined the relationship between rela- tive oxygen consumption (VOz) and heart rate (HR) in obese (N = 86, body fat > 30%, hydrostatic weighing) compared with normal- weight (1V= $1, body fat = 30%) adults, Simultaneous measurements of HR and VOz were recorded at rest and every minute during a ‘maximal graded exercise test. When MHR was regressed on age, 1Wo distinct equations for the obese and normalweights were generated. ‘The relationship between MER and fomax VO; was similar be- tween groups (r = 0.83, obese; r = 0.87 normalweights). Likewise, when Sémax VOs was regressed on Sémax heart rate range similar ‘equations were derived for the obese (r = 0,81) and normalweights (¢ ='0.84). Correlation between Karvonen's predicted HR at a submax- imal VO; and the true HR at that VO; was 0.88, regardless of ‘These data indicate that when predicting MHR in normal- ‘weights the equation 220 ~ Age can be used, but for obese individuals the equation 200 ~ 0.5 x Age is more accurate; each having 12 as a standard error of estimate. Once MHR is determined, either the straight percentage technique or Karvonen's method would be appro- priate for prescribing exercise intensity for both populations. ‘TRAINING HEART RATE, EXERCISE AND WEIGHT LOSS, AEROBIC EXERCISE, BODY FAT, OVERWEIGHT egardless of the exact intensity selected for aero- Re exercise prescription, the exercise training intensity is invariably monitored by a predeter- mined training heart rate (THR). The two most com- mon methods used to compute the THR, without direct determination of the relationship between submaximal heart rate (HR) and oxygen consumption (VOz), are the straight percentage technique and the Karvonen formula (22). With the straight percentage technique, the THR is computed simply as a chosen percentage of the maximal HR (MHR; 1-3,20). In comparison, Kar- vonen’s formula computes THR where THR = [(MHR 0195913195/2500.107783.090, [MEDICINE AND SCIENCE IN SPORTS AND EXERCISE Copyright© 1993 bythe Amesean College of Spor Medicine Submited fr pubcaton September 1992. ‘Accepted for publication Ape 1993 1077 — resting HR) x 0.60 to 0.80] + resting HR. It is obvious that for either of these two methods to be effective in computing the proper THR required to elicit the desired aerobic exercise intensity for obese clients, one must have an accurate predictor of MHR and substantiation of the submaximal HR-VO, rela- tionship for the obese population. ‘The prediction equation for MHR that is used almost exclusively is 220 — Age. Although the actual derivation of this regression equation has never been published, ‘the American College of Sports Medicine (ACSM; 1,2) and several popular exercise physiology _ texts (8,13,15,26,27) still promote the use of it for predicting MHR. Furthermore, the relationship between MHR and age in obese men and women has never been examined, Consequently, the first objective of this study was to determine the relationship between MHR and age in an obese population, compared with a normal- weight population, and derive the appropriate predi tion equation(s) for MHR that would be useful in exercise prescription. Once a predictor for MHR has ‘been established, then either the straight percentage technique or Karvonen formula could be used to com- pute the THR that will elicit the desired metabolic Tesponse only if the HR-VO; relationship has been substantiated for obese subjects. Several papers have shown that this relationship between relative HR and ‘VOn is consistent across age, sex, coronary artery di ease, fitness level, training status, muscle groups exer- cised, and testing mode (17,19,25,30,35). However, none of these studies specifically examined’ the HR- VOz relationship in obesity. Thus, the second objective of this study was to compare the relationship between relative VOz and HR in obese adults with that found in normalweight adults, METHODS, Subjects, Subjects for this study were adult men and women entering either the university-administered1078 Official Journal ofthe American College of Sports Medicine Adult Fitness Program or Weight Loss Clinic. Each subject signed an informed consent approved by the University Committee for the Protection of Human Subjects. Routine medical history and health exami- nation forms were also completed before the study. Participants had no metabolic diseases and were not taking medications that may have affected heart rate or blood pressure. Test procedures. Each subject reported to the lab- oratory on two separate occasions after an overnight fast. On one occasion, body density was determined through hydrostatic weighing (33) and body fat was calculated by the Siri equation (29). On the other occasion, a graded exercise test (2) was administered according to a modified Balke protocol where speed ranged between 4.0 and 6.4 kph and grade was elevated by 2.5% every 2 min until maximal effort. During the exercise test, HR, VO2, blood pressure, and perceived exertion (7) were recorded simultaneously every min- ute, MHR was defined as the highest HR attained during the test and maximal VO2 (MVO,) was meas- ured as either the peak VO; achieved or the point at which VO> leveled with increasing workload. Maximal effort during the test was verified by a respiratory exchange ratio above 1.0 anda perceived exertion rating, of maximal, Prior to analysis, the data for each subject were assigned to either an obese or normalweight group according to the adiposity level of that subject. Obese was defined as >30% body fat whereas normalweight was defined as =30% body fat. Statistical analysis. The strength of relationship be- tween two variables was measured by the coefficient of linear correlation while regression analysis was em- ployed to generate prediction equations for any de- pendent variable within a specific analysis. The regres- sion slopes for different groups were tested for equality by the large sample Z test (21). Group mean compari- sons were made with a Student's ¢-test. Statistical sig- nificance was declared at P < 0.05. RESULTS ‘The demographic data for the subjects are reported in Table 1, The two groups were similar with respect to age and fitness level (as measured by MVOz), but significantly different with respect to body weight and adiposity level. Although obesity may be defined as percent body fat that increases disease risk, the absolute percent body fat at which disease risk increases is con- troversial (10,31). The level of 30% body fat that was used in this study to dichotimize the data was warranted in part by previous literature (17,27) as well as the definition of obesity proposed by Bray (9) where body mass index (BMI) >30. Accordingly, the mean values for percent body fat and BMI reported in Table 1 clearly MEDICINE AND SCIENCE IN SPORTS AND EXERCISE TABLE 1, Subjoct demography. ‘Normal weights or (west) Age 459218 420210 Weight a) 736424 94203" % Body fat 247 +06 388 +08" BMI 284 +08 80408 VO, (in) 2704 24204 ‘Values are means: SEM. Signifcantly diferent rom normal weights, P< 0.000%; ‘BM = body mas index, MUO = maximal oxygen consumption evel. sort the subjects into either an obese or normalweight group. Therefore, further comparisons between groups were justified. ‘When MHR was regressed on age, the equations listed in Table 2 were generated. Regressions for men and women within either the obese or normalweight group were remarkably similar and allowed for a com- bination of the sexes within each group. Furthermore, the regressions for the normalweights were not much different from the equation 220 — Age. However, be- ‘tween-group analysis revealed that equations generated for the obese were significantly different from those for the normalweights. To ensure the stability of the new regression equations, a split-half, cross validation was run for both groups. In each instance, there was no difference found for the generated regression equations or their respective r values. When %MVO; was regressed on %MHR for the normalweight and obese groups, the equations in Table 3 were formulated. It is apparent from the table that the relationship between relative submaximal oxygen consumption level and relative submaximal HR was similar for obese and normalweight adults. Next, %MVO, was regressed on Yomaximal HR re- serve (MHRR). The respective regression equations for this analysis are also reported in Table 3. Once again, the regressions for the obese were identical to those of the normalweights, which suggests that the ability’ of Karvonen’s formula to predict aerobic exercise inten- sity is the same for the obese and normalweight adults, This premise was further supported when the meas- ured HR for each individual exercising at 40, 50, 60, TABLE 2, Regression equations fr maximal heat rat. ‘Sample Equation tow SEE ermal weight men and WHR=217—085% Age 068 16% 12 women (W = 81) Normal weight men «MH = 2190.85 Age 0.70 18% 10 (v= 35) Normal weight women MHR=218—098% Age 0.70 27% 14 (= 18) (Obese men and women MHR = 200-048% Age’ 0.95 30% 12 MHA = 198-044 Ago “0.59 57% 10 MHR=200~049% Aget_0.34 359 19 “+ Sigicany diferent ram comparable nomal welght group; *P < 0.005, ¢ P< (0.04; MHA = maximal heart rate; r= colton coefclet; cv. = coeicent of variton; SEE = standard eror of estimateMAX HR AND THE HR-VO; RELATIONSHIP IN OBESITY TABLE 3. Regression equations for oxygen consumption fv Equation a SEE (11x %MHA— 19 087 4% 8 beso = 88) %MVOr= 1.15 S6MHA— 23063 3 8 = 085% %MHAA+ 27 084 4% 9 85x MHAR+27 O81 4% 9 {480 = pereent maximal oxygen consumplon lve; 96MHA = percent mia heart rat; %eMHRR = percont maximal hear rate range; r = corlaton coafient; x. = coset of variation; SEE = standard eror of estimate. 70, 80, and 90% of MVO> was regressed on the pre- dicted HR for each of those intensities according to Karvonen’s formula (see Table 4).’No difference was found between the regressions for the two groups and the correlations for both groups were similarly high while the coefficients of variation were low. DISCUSSION The purpose of this research was twofold: first, to derive equations for predicting MHR in an obese com- pared with a normalweight population and second, to verify whether the HR-VO, relationship during sub- maximal exercise was the same for the two groups. Accordingly, two distinct population-specific equations were derived for predicting MHR when age was an independent linear variable. The equation for the obese was MHR = 200 — 0.48 x Age, whereas that for the normalweights was dissimilar, being 217 — 0.85 x Age. However, when MHR was predicted for our normal- weight group using both our equation and 220 — Age, a paired t-test revealed that the mean difference was within a few beats. Although the 220 — Age equation had a tendency to overpredict MHR, it probably was not enough to make a difference in exercise pre~ scription. Early work from Robinson (28) suggested that MHR in adults peaked at about 200 and declined approxi- mately one beat per minute (bpm) with each progress- ing year. Data from Astrand (4,5) and Astrand et al. (6), published roughly 20 yr later, seemed to harmonize with that of Robinson (28). Others, over the next several years, reported widely differing MHR-age tables or regression equations, but none of these reports pro- duced the acclaimed 220 — Age equation. In fact, some of these researchers (1 1,18,23,24) even produced regres- TABLE 4, Regression equations for submaximal heart rate, amp Equation r Nemalweghts _ SHA=O71xKPHA+48 088 0% 9 W=51) ot 0) —SHR=O79xKPHR442 080 9% 9 ‘SAR = submaximal heart rate at 40, 60,60, 70,80, and 90% M03: KPHR kKervonenspredictedheart rat at 4,60, 60, 70, 80,2nd90% MHAR; r= corlaton cetfien; cx. = coefcient of vation; SEE = standard eror of estimate, Official Journal of the American College of Sports Medicine 1079 sion equations similar to what we found for obese subjects (Table 2). In 1971, Fox and coworkers (16) certainly suggested that MHR could be predicted by the equation 220 — Age, Nevertheless, the actual derivation was not published, Regardless, the possi regression equations for obese and normalweight pop- ulations has never been investigated. ‘The most thorough work on predicting MHR was performed when Londeree and Moeschberger (24) at- tempted to pool all published data on MHR and include in their regressions factors that may have been over- looked in previous work. These investigators analyzed data from 23,000 subjects from 5-81 yr old and used a multiple stepwise regression analysis to generate predic- tion equations for MR. In their analysis, MHR was the dependent variable and age, age’, age’, age’, sex, fitness level, type of ergometer, exercise protocol, con- tinent of residence, and race were independent vari- ables. In spite of the fact that obesity was still over- looked, their research produced several regression equa- tions for MHR. Since age accounted for about 70-75% of the variability in MHR in their work, the authors stated that the simple calculations involving only age as an independent variable could be used for predicting MRR without much loss in efficiency. Only one equa~ tion produced from their work used age as a linear factor (MHR = 206 - 0.711 x Age). When we com- pared this equation to ours for normalweights (217 — 0.85 x Age), we found it to be significantly different (P< 0.003). We cannot resolve the discrepancy at this point in time, but itis possible that the differences in regressions found among our study and any other previous study were due to failure to separate the normalweight sub- jects from the obese in the other studies. Our data clearly indicate that our population of obese adults was separate and distinct from our normalweight popula tion with respect to predicting MAR. The fact that the coefficient of variation for the obese regressions was greater than that of the normalweights can be explained in part by the smaller variance in the ages of the obese subjects (Table 1). However, the standard error of estimate for both groups was the same. Nonetheless, the regression equations for the two groups were significantly different and should be used for their respective populations. ‘One might suspect that the obese and normalweight groups in this study were not only distinguishable by adiposity but also by MVO;, and that the reported MHR-Age regressions merely reflected a difference in aerobic capacity (14,18,23,34). This is not a valid ar- gument because the MVO> for both groups was equiv- alent (Table 1). Furthermore, the coefficient of linear correlation between MVO; and %body fat for the entire data pool was rather low (rt = 0.27).1080 Official Journal ofthe American College of Sports Medicine In contrast to the differences found in regressions for ‘MER between the obese and normalweights, the rela- tionship between HR and VO, at submaximal work- loads was similar for the obese and normalweight groups (Table 3). These data are consistent with other papers that have shown that the relationship between relative HR and VO: is consistent across age, sex, coronary artery disease, fitness level, training status, muscle groups exercised, and testing mode (17,19,25,30,35). In their review, Franklin and associ- ates (17) reported that the value for the slope of the published regression equations for %MVO, on %MHR_ varied between 1,2 and 1.5 while the intercept varied between -24 and —52. These values are analogous to those generated by this research for obese and normal- weight adults (Table 3). This likeness in regression, equations among studies along with the high correla- tions and low coefficients of variation indicate that the telative oxygen cost of submaximal steady state exercise can be predicted from relative HR (17,20). Davis and Convertino (12), on the other hand are not that com- fortable with predicting %MVO, by simply calculating a particular fraction of the MHR. They contest that this straight percentage technique tends to overestimate exercise intensity whereas Karvonen’s method yields %MVO; values equivalent to measured values. Along this line of thinking, the regression of %MVOz on %MERR produced identical equations for the obese and the normalweights (Table 3). The high correlations and low coefficients of variation for both groups indi- cate that it is appropriate to use Karvonen's formula for prescribing exercise intensity for the obese as well as normalweight individuals. Even though the appro- priateness of Karvonen’s formula has already been validated by Davis and Convertino (12), Wier and Jackson (32) have recently refuted this validation, In their work, the regression of %MWVO; on %MHRR produced a regression equation (%MVOz = 0.47 x S6MHRR + 36.6) that obviously had a slope and inter- REFERENCES |, AMERICAN COLLEGE OF SPORTS MEDICINE. Position stand: the recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness in healthy adults. Med. Sci. Sports Exerc. 22:265-274, 1990, 2. AMERICAN COLLEGE OF SPORTS MEDICINE. Guidelines for Exer- cise Testing and Prescription, 4th Ed. Philadelphia: Lea & Febi- get, 1991, pp. 60-62, 96-100, 171-172, 195, 3. AMERICAN HEART ASSOCIATION—CoMMITTEE ON ExtRcIse, Ex- ercise Testing and Training of Apparently Healthy Individuals: A Handbook for Physicians. Dalias: American Heart Association, 1972, pp. 13-15. 4. AsrRAND, I. The physical work capacity of workers 50-64 years Old. Acta Physio. Scand. 4273-86, 1958. wND, I. Aerobic work capacity in men and women with spec eee to ape dea Postal Seared Sup 1) 92, 1960, 6. ASTRAND, [,, PO. AsTRAND, and K. RoDautL. Maximal heart tate during work in older men. J: ppl. Physiol. 1562-566, MEDICINE AND SCIENCE IN SPORTS AND EXERCISE cept different than 1.0 and 0, respectively, as Karvonen suggests (22). The large sample Z test revealed that our equation (%MVO; = 0.65 x %MHRR: + 27) was significantly (P < 0.0001) different from that of Weir and Jackson. The fact that the correlation between Karvonen’s predicted HR at a given submaximal VO2 with measured HR at that VO2 was high (0.88, Table 4) further suggests that Karvonen’s formula can be used for prescribing exercise intensity in both the obese and. normalweight populations. Summary and Recommendations Two distinct equations for predicting MHR were derived for obese and normalweight individuals. The equation for normalweights was very similar to the 220 — Age equation that has been used a priori. We ‘therefore propose that practitioners continue to use it for predicting MHR for normalweight clients and 200 = 0.5 x Age for their obese clientele, These equations can easily be remembered as well as obesity being defined as >30% body fat or >30 BMI. Once MHR has been determined, either the straight percentage technique or Karvonen’s formula could be used to calculate the THR that will be utilized for monitoring exercise intensity. It must be remembered that predic- tion equations are not absolute, but only guidelines to follow in prescribing exercise. The clinician must use judgment about how a given individual is responding to the exercise, and adjust the intensity accordingly to provide for the participant's comfort and safety while trying to achieve a training effect and a large energy expenditure (2). ‘This research was supported by the Obesity Foundation and the Indiana University Adult Fitness Program. The authors wish to thank David Kooeja for his technical assistance. ‘Address for correspondence: Wayne C. Miller, Ph.D., Department Of Kinesiology, HPER 296-H, Indiana Uniersty, Bloomington, IN 7. Bona, G. A. Psychophysical bases of perceived exertion. Med, Sci. Sports Exerc. 14:371-387, 1982. 8. Bowers, R, W. and E. L. Fox. Sports Physiology, 3rd Ed, Dubuque, IA: Wm. C, Brown, 1992, p. 240, 9. BRay, G. A. Pathophysiology of obesity. Am. J. Clin. Nutr '55:488S-4945, 1992. W. R. Fosres |. HinscH, and T. B. VAN ITALLIE. tions of obesity: an NIH consensus development conference. Int J. Obesity 9:155-169, 1985. 11. Cooper, K. H, The Aerobics Program for Total Well-Being Exercise Diet Emotional Balance. New York: M. Evans and Company, 1982, p. 125, 12, Davis, J. A. and V. A. ConverriNo. A comparison of heart rate ‘methods for predicting endurance training intensity. Med. Sci Sports. 7:295-298, 1975. 13. DEVRIES, H. A. Physiology of Exercise for Physical Education and hits, th Ed. Dubuaue, 1A: Wm, C, Brown, 1986, p. 298. 14, DrinkwareR, B. L,, , M, HovaTa, and C. L, WELLS. AerobicMAX HR AND THE HR-VO; RELATIONSHIP IN OBESITY 16. 7. 18, 20. 2 23. 24, power of females, ages 10 to 68. J. Gerontol. 30:385~394, 1975, Fox, E. L., R, W. Bowers, and M. L. Foss. The Physiological Basis of Physical Edueation and Athletics, 4th Ed. Philadelphia: Saunders College Publishing, 1988, p. 293. Fox, S. M. Ill, J. P. NaucHtToN, and W. L. Haskeut. Physical activity and the prevention of coronary heart disease. Ann. Clin Res, 3:404-432, 1971. FRANKLIN, B. A. J. Hopson, and E.R. Buskirk. Relationship between percent maximal O; uptake and percent maximal heart rate in women. Res. Q. 51:616-624, 1980. FROELICHER, V. F., JR. M, ALLEN, and M. C. LANCASTER. Maximal treadmill testing of normal’ USAF aircrewmen. Aero space Med, 45:310-315, 1974. ). GREEN, J. H., N. T. CABLE, and N. ELMS. Heart rate and oxygen consumption during walking on land and in deep water. J. Sports ‘Med. Phys. Fitness 30:49-52, 1990. HeLuersreis, H. K. and B. A. FRANKLIN. Exercise testing and prescription. in: Rehabilitation of the Coronary Patient, N. K. ‘Wenger and H. K. Hellerstein (Eds.). New York: John Wiley & Sons, 1978, p. 18i- Hopkins, K: D, and G. V. GLass. Basic Statistics for the Behav- ioral Sciences. Englewood Ciffs, NJ: Prentice-Hall, 1978, pp. 292-294, KAaRVONEN, M.J.,E, KENTALA, and O, MUSTALA, The effects of ‘training on heart rate a longitudinal study. Ann, Med. Exp. Biol. Fenn, 35:307-315, 1957. Lester, M,, L. T. SHEFFIELD, P, TRAMMELL, and T. J, REEVES. ‘The effect of age and athletic training on the maximal heart rate during muscular exercise. Am. Heart J. 76:370-376, 1968. LonDERee, B. R. and M. L. MOsscubeRcER. Effects of age and other factors on maximal heart rate. Res. Q. $3:297-304, 1982. Official Journal of the American College of Sports Medicine 25, 6. 2. 28 29, 30, 31. 32. 33. 34. 35, 1081 Lonpente, B. R. and S, A. AMES, Trend analysis of the %VO2 ‘max-HR regression. Med. Sci. Sports 8:22-125, 1978, Mcanpts, W. DF. I. Karen, and V. L. Karten. Exercise Physiology Energy, Nutrition, and Human Performance, 3rd Ed. Philadelphia: Lea & Febiger, 1991, p. 705. Powens, S. K. and E, T. HowLey. Exercise Physiology Theory ‘and Application to Fitness and Performance. Dubuque, 1A: Wi. C. Brown, 1990, p. 335. Roninson, S. Experimental studies of physical fitness in relation 10 age. Arbeitsphysiologie 10:251~323, 1938. Sint, W. E, Body composition from fluid spaces and density. In: Techniques for Measuring Body Composition, J. Brozek and A. Henschel (Eds.), Washington, DC: National Academy of Sci- ences, 1961, pp. 223-244, Sprincer, C,, T. J. BARstow, K, WASseRMAN, and D. M. Cooper. Oxygen uptake and heart rate responses during hypoxic exercise in children and adults. Med. Sei. Sports Exerc. 23:71~ 79, 1991 VAN ITALLB, T, B, Obesity: adverse effects on health and longev- ity. Am. J. Clin, Nutr, 32:2723-2733, 1979. Wier, L. T, and A. S, JACKSON. %VOsmus and 9SHRmax reserve are not equal methods of assessing exercise intensity. Med. Sci ‘Sports Exere. 24(Suppl.)S176, 1992. ‘WitMoré, J. H. A simplified method for determination of resid- uual lung volume. J. Appl. Physiol. 27:96-100, 1969. Winsore, J. H.,J, ROYCE, R. N. GIRANDOLA, F. I Katou, and V. L. KatcH. Physiological aiterations resulting from a 10-week program of jogging. Med. Sci. Sports 2:7-14, 1970. ‘Yasaut, K., M, GREENLEY, D. R. NoRTHEY, and R. L. Houon- SON. Oxygen uptake and heart rate responses to treadmill and ‘water running, Can. J. Sport Sci. 15:96-98, 1990.