Total Body Potassium and Gross Body Composition in Relation to Age (Journal of Gerontology, Volume Contents, Vol. 15, No. 4)

Page 2

JOURNAL OF GERONTOLOGY TOTAL BODY eOTASSlUM AND GROSS BODY COMPOSITlON Denver, Colorado, and rhe L'niocrrity of Californit's Los .4lumos Scirntific Luborctorcr. Rubner (26) long ag3 introduced the mncept of an "active tissue mass," and postulated a relationship between this mass and body heat production. Because of its intracellular occurrence. potassium may offer a rather direct measure of this quantity. Potassium is usually the most abundant of the cations found inside living cells of vertebrates (20); u small quanrity also occurs in extracellular fluids. Potassium participates in propagation of nerve impulses, is involved with muscular contraction. and is recognized as necessary for biocatalysis. It is excreted chiefly by the kidneys and is replenished by ingestion of food. A detailed theory explaining accumulation OF potassium by living matter (especially in muscle cells) has been proposed by Conway (12). Although the importance of potassium in u numbeir of vital proca- is well established, it has been diPcult to determine the total quantity in the body. Direct analyses are available for only a limited number of cases. Results of chemical analysis of two cadavers b y Forbes and Lewis (17) indicate that about GO? of all body potassium is contained in muscle. The determination of body potassium by isotopic dilution using K" as a tracer ( 13) is nondestructive and can be applied in vivo but is time consuming, and the requirement of ingestion of radio-activity limits its applicability. This method determines exchangeable. not total, potassium. but the evidence is that if SURcient mixing time is allowed (48 hours) the two quantities are very similar. The first direct meaSurementS of total body potassium by means of its natural radioactivity Energy Commission and tha U. S. Army & a 1 Corp. were made by Sievert (28. 29) end h! Burch and Spiers [ 11). Sievert ( 2 9 ) and Sagild ( 2 3 established the change in body potassium content with age and indicated its esplanation. These early studies used ionization chambers to measure the 1.46-Mev gamma rays from the K". which wcur in constant proportion in all normal potassium. Development ot :he loa background sodium iodide crystal spec!rc:mcter (21. '11) reduced the time required !'or such measurements and improved the precision 1 1 1 body potassium determinations. Measurements can be made et'en more rapidly with a .In liquid scintillation counter (4. 3. '31 ), which can quantify total body potassium to a precision of 2% in less than 200 seconds of counting time. Using the liquid scintillation counter. Anderson and Langham ( 71 reported body potassium measurements in 1590 ;ubjects of both sexes. ranging in age from less than 1 year to 79 years. The average concentration of potassium per kilogram of gross body aeisht showed interestins correlations with #weid sex, one of the prominent features being steady decline with increasing nge lor .ldulrs. Confirmation of the general features of the curve and excellent quantitative agrrrment for two ndditional large population goups lone American. one German) have been bbteined with other liquid scintillation counters I 25. 33). Where large enough population priiups hare been measured to permit accurate comparisons. agreement among the various methods is excellent. For example, McNeill and Green (?'). using a NaI crystal spectrometer, have report& the average potassium concentration in a group of 30 Canadian male subjects 19 years of age (0 :Ion counter for a gro Gillivray and associatt Ron. found for 32 fern: 49 years an average va' [he LASL value of 1.6 of this age group. . Wayne and auociates p e d with the liquid the qtandard deviatior jets measured. More agreement of the pot. specific age groups in pies indicates that ax parameter is being rn' The purpose of thc lyze furtkr the age concentration and to effects of variation through the use of b Body water was chow iiwe it shows an u the body after remo tat (F), and water For isohted tissue, t t reted for the high c so that under the st This relationship ir for purposes of con!

Page 3

:, c o u n ~ ~ rfor ii group lit 21 American sub- , I . found ilir $2 fcmalcs (Scotland) aged 251) vears an i t verage value of I .60 '-t 0.04 versus .-.r'~.4~Lv;iliir of 1.63 z 0.02 for 176 subjects r 7 males I Scotland) aged 21-21 years. w m - :Ard with the liquid scintillator value of 2.12 0.04 for 37 American males aged 21-23 years. ill errors Siven above are estimates of the .:&rd error of the mean. calculated from -be standard deviation of individuals From the ::oup mean I 10-14%) and the number of wbmeasured. More detailed correlations with p data *ire discussed later. This excellent :jeement the potassium concentrations for pific age proups in different population sam- :k indicates that an nccurately reproducible The purpose of the present paper is to ana- :ze further the age dependence of potassium :,ncentratiun and to attempt to eliminnte the :tfecu of \.ariation in Fat and skeletal mass ? h u g h the use of body water determinations. Bady water was chosen as the other parameter. dnee it shows an unusually good correlation drawl defatted protoplasmic mass of the body, MI 12). M3is defined as the residual mass of ihc body after removal of bone mineral (m), irt (F), and water (W) from the gross body Fa isolated tissue, this residud mass has a uninitdonship holds even for bone, when corI mted for the high density of dry bone mineral a n OF age to : Independent of age and sex. the relationship t lorhat under the stated conditions PI to health md abstinence from food For 12 hours, This relationship is used in the present paper mahate M, from body water measurements ' h purpoxs of comparixrn with total body po- ; Wium determinations. In addition, a cornparbetween W and M3 is a constant, d (1,2): ison of total body potassium with basal metabolic rate (BNIR) is made. Measurcmcnts of the natural gamma radioactivity of the \ 590 su bjcvrs reported by Anderson and Langharn ( 7 ) were made with a large liquid scintillatioil counter demribcd elsewhere (4. 6). The coiitiwr prorided essentially loo? Kcometricol efficiency [-Insteradian solid angle) for a counting well I8 inches in diameter by 6 feet long in which the subjects were placed. By use OF cnergy discrimination. only the gamma rays from K'" were counted. Three minutes of counting rime per subject gave a statistical precision 1Jf t 3 Gm. of potassium (standard deviation 1. or about 2% For the aver- 'ige adulr mule. The counter was calibrated uith weighed quantitics of KCl mixed with quantities of suc~ose ranging from 80-200 pounds and ,~Is(Juirh known quantities of K" ingested by person5 r i f various weights. The precision of the counting method. as judged bv internal consistency and reproducibility. was essentially the same ns the counting statistical error (Le., about 0.08 equivalent OF potossium). The absolute accuracy of this method is harder to establish but seems to be better than 5%. as judged hy comparison with other methods. The population sample was predominantly New Mexican. but no significant geographical effects were apparent in the primary data. Excellent quantitative agreement has sine bcen obtained for a second U. S. popuiation sample (33) and for u German population (25). The original sample consisted of all visitors to the Health Research Laboratory who volunteered Potassium and body water by HTO dilution were determined on an additional 38 subjects 10 being women I ranging in age from 16-51 years. For 19 of these subjects, body water was independently estimated from HDO dilution (2). The estimated precision of the water determinations was u = 0.67 Kg. Unfortunately, it was not feasible to determine body water on the earlier group of 1590 subjects for whom total body potassium measurements were made. In using thcse data in the present study it is necessary. therefore, to make use of peviously reported average values lor body water as a Function of age. mined in 17 healthy male subjects 22-44 years

Page 4

I%. 1 . Correlrrion between total body potassium t t r n m K" wuntin8) cf males and females of all ages ind their W., mass derived from d and average total Fig. 2. Variation in the ratio of total body potas- .ium to M, as n function of age and sex. of ege within a few clays of the dcternrination of potassium. Three determinations nere of the average time required to Coiisuine I L. of pure oxygen while the subjecr i t ~ sstill in culate the BMR. a respiratory quotient of 0.d? was assumed. Kiving a caloric quiti1lent 4.825 kcal. to I L. of nxygen (8). From this. the BMR was calculated in the tinits kcal. suremenu on 1057 me iects in terms of the a\

Page 5

# BODY POTASSIUM. BODY COMPOSITION AND AGE Table 1 summarizes body potassium mea- ;urements on 1057 male and 532 Female subkts in terms OF the averages by age groups for' weight in kilogrrims. and svcrage total potassium in tquivalents are given. Shown also are the observed standard deviations of rhe individual values for ench age group from the group mean and the standard error of the mean, Urn. where n is the number OF subjects measured. Except for the extreme age prouh, the estimated precision OF the mean is from 1.5-5frr,. Table 2 summarizes data From the literature on body water as a function of age (2. 9. 16, 22). The values of u,,,are again estimated by dividing the standard deviation of each age group average by the square root of the number of subjects in the p u p . The average body water values for adult males is defined rather ues for children and females is much larger, ranging up to 8% because of the small number of subjects for whom data are nvailable. In the present analysis, thereFore. principal emphasis is placed on the adult male results. A plot of the data in table 2 as percentage body wnter versus age gives curves from which percentage h3dy water at intermediate ages may be obtained by interpolation. Smoothing of the data by this process probably decreases somewhat the standard error of the mean [or any given age. Table 3 gives the experimental data for the bth sexes. Number of s;bj&s. average p s s ments of b d y water and potagsium were made.

Page 6

, Figure 1 shows the average total body potassium for various age groups (based on the data in table I ) , plotted against the average M:!for the age group estimated by Eq. 3 from the measured average weight {table I) and the average body water graphically interpolated from the date in table 2. The number of subjects per point varied from 3-5 for very young to 50-250 for the adult groups. By plotting the ratio of K/M, in equivalents per kilogram as a function of age (fig. 2). deviations from the straight line of fig. 1 (corresponding to K/M, value of 0.332) can be emphasized. Ih figure 2, the standard error of each point is indicated. This error is assumed to be due entirely to the uncertainty in the value of percentage body water as a function of age, BS tabulated in table I. Because of the much larger population groups available for K40 counting as compared with body water measurements. this seems to be the case. While (for adults) the decrew of K/MIwith age and the sex difference are unmistakable, the scatter in the points is such that the slope is uncertain by 2 50%. This is particularly true for the female poinrs, because of the smell population sample for which body water data are availabIe. These curves may be compand with those showing age and sex variation in total body potassium given in figure 3 and with the previously published ones showing age and le\ dependence of potassium concentration expressed in Gm.!Kg. of gross body weight (7). [n the latter case. the results include changes 111 fat, connective tissue. and skeletal mass ~ i t hage. while figure 3 also includes effccu due to tarintions in total body mas. The correlations shown in figure 2 are b a d on statistical treatment of population at erages and on the assumption of the reality of certain correlations, specifically on the variation of average body water with age. One can. there fore, investigate whether the relationships indicated in figure 2 a h apply to individuals for whom simultaneous measurements of both p ~ - tassium and body water were made. Figure 4 shows the results of such e comparison. The curve is that deduced for males in figurc 2. The open circIes represent individual data on 7: adult males, 10 reported by McMurrev end associates (Z),28 cullected at the Los Alamm Scientific Laboratory, and 34 reported by rhe Naval Radiological Defense Laboratory (9). The solid circles show averagu for 5 age group of the 28 subjects measured at LASL. And the solid triangles represent similar average for the 34 subjects measured at the NRDL. In the latter subjects, the body potassium meisured by K'? isotope dilution after tvaiting 48 hours to ensure establishment of equilib-

Page 7

e2. The I Em 51 36.20 1.u 1o.m tassium concentratinn wirh nve nPr&tc tautan

Page 8

eral. The latter conclusion is supported further by the results in figure 4. showing individual values and age group averages for K , M: dcrived from body water and potassium measurements on the same individuals. These data were colIected in three different laboratories in which body potassium wm determined by two different methods. Although the individual points scatter widely, agreement between the age group averages and the statistically derived curve is strikingly good, despite the small number of determinations in each group. The average ratio of K,'M, valua by isotope dilution and by whole body counting is 0.967. The 3% discrepancy may be due to a consistent calibration error or m y represent the fraction of nonexchangeable potassium, or both. The age effect for both sexes seems to be due more to a slight loss of body potassium as noted by Sievert (29) and Sagild (27) than to an increase in M,. The rate of change of body potassium of adult males with age is reduced from 0.42%,'year. when total potassium is consium is related to M,. The scatter of individual points about the averaged curve is reduced from u = 12% to u = 5%. The standard deviation of the individual potassium values about the average for their age group corresponds to a n ape effect of 30 years in the Former case and 20 vears in the latter. The use of combined potassium and body water determinations, therefore, does not enhance the precision of the correlation suaiciently to make it a practical empirical method of estimating physiologic age of individuals. As a possible method of quantifying the aging process in population groups. the method seems to be worthy of further study. …