= —_— 2 g 4 E 2 _ 8 q <q oO a an q x = # & 3 L* wa 2 a a ¥ 50 - G = RJ = g 1ok 4 gE 6 5 _ SE : 7 7 4F 7 @ 3f S et Ca COMPUTER OERIVED VALUES|_| z ° * EXPERIMENTAL DATA a St —— COMPUTER DERIVED VALUES o EXPERIMENTAL DATA @ 30}-—+—+—+—4+ + + ++ a x 20 . —~ 3 ~~ ‘ 40 |= Ca ——COMPUTER CERIVED VALUES * EXPERIMENTAL DATA St —— COMPUTER DERIVED VALUES © EXPERIMENTAL DATA o l 1 | 2 1 3 Lot 4 5 i 6 7 oe 1 8 od i 9, 4 GAYS Fic. 3. IBM-7094 computer-derived fit to experimental whole- & body counting and knee data. @, Ca experimental data point; o, Sr experimental data point; , computer-derived value ¥ for Ca; - --—-, computer-derived value for Sr. DISCUSSION 4 3 FECES dt 3 4 5 6 7 RESULTS The values of the transfer constants and compartment sizes of both Sr and Ca models obtained by the computer correspondence between the observed data and the curve of computer-derived values. Quantitative examina- tion of the variability of the observed data from cal- r enPrae From the experimental data and the postulated general ? relationships of the variables, the computer calculated a * set of values for the parameters which yielded a mean °. DAYS ‘3/8 6S least-squares fit for the observed data. The fit of the x model, thus obtained, to the experimental data for both = FIG. 2. IBM-7094 computer-derived fit to experimental plasma, Sr and Ca can be seen in Figs. 2 and 3. The “goodness ¢ feces, and urine data. @, Ca experimental data point; 0, Sr exof fit’? of the model is measured by the random scatter; perimental data point; , computer-derived curve for Ca; ——--, computer-derived curve for Sr. va of the data about the predicted values and the close f ight o ' 2 culated values indicates that the model represents both * the Ca and Sr data very well. The model can besaid to program are presented in Table 2. The standard deviations (o,) of each of these parameter values are also pre- approximate both Ca and Sr metabolism and to be consistent with the biological data collected. the standard error of the mean (¢,,) are also shown in the table. The computer-derived plasma and cumulative urinary and endogenousfecal excretion values for Ca and Sr are plotted along with the experimentally derived curves of a transfer constants (¢,) are functions of the individual * representative subject (subject 1) in Fig. 2. Whole-body counter data and knee data for Ca and Sr are plotted along with the computer-derived curves in Fig. 3. (The coinputer-derived values for both whole-body retention and knee retention represent the weighted summation of the tracers in all three compartments. ) The mean whole-body retention data for the seven subjects expressed as ratio of Sr/Ca over 30 days are shown in Fig, 4. The ratio falls rapidly due to the preferential excretion of Sr and reaches an approximate asyinptote of 0.660 by 18 days. The level of both Sr*® and Ca‘ in the knee can be expressed in the same manner as the whole-bodyretention data. The Sr/Ca ratio in the knees over 30 days, as shown in Fig. 4, also falls off rapidly and levels off at about 18 days but at a higher value than the whole-bodydata (0.730). The standard deviations of the compartmentsizes and | datum points as specifically related in the equationsof|‘ the model. The ability of the computer program toy assess the standard deviation or precision of the model on the basis of the data being considered permits the com- : computed for the seven subjects considered here. k parison of the fit of Ca and Sr data to the model. The t mean standard deviations (gn) (expressed as percentages) for the various calculated parameters of Ca and Sr were | Calculated Parameters Compartment1 Compartment 2 Compartment 1 + 2 Accretion (p31) Urine (pa) Feces (ps1) Flux (12.2) Cal on, % +21.9 +26.8 5.7 +13.0 +8.0 +9.2 +26.8 Sr8hon, % 19.4 21.5 | : 4.9 14.7 10.6 +11.0 21.5 The size of compartments 1 and 2 and the transfer rate é i OSE MgINSeeoO sented. The weighted mean of the seven subjects (%) and between them (p;2,91) are determined primarily from © data obtained in the ist day. Since only two plasma 3 samples were taken in this interval, the above three