in children
-.
Dear
with
PM
Sir:
We Mehta
are very interested in the paper by et al. (Am. J. Clin. Nutr. 28: 977, 1975), especially as we are working in this field. We should like to add to some of the data presented by the authors, by drawing an analogy with penicillin, and would care to bring up another point for discussion. We studied penicillin kinetics in a group of children with kwashiorkor and very marked edema (N. Buchanan, R. Robinson, and H. J. Koornhof, unpublished data). The purpose of the study was 2-fold: first, to assess the fate of intramuscular administration of pencillin in these children, when the injection had been given into an edematous thigh, and second, to study the excretion patterns of penicillin, by taking serial blood samples. The children each received 50,000 U of sodium benzyl penicillin on the day of admission, and thereafter on days 10 and 21 of hospitalization, i.e., once the edema had cleared. Two predominant observations arose from this study, which are pertinent to the present discussion: 1) It was shown that even in the presence of marked edema, the absorption of penicillin from an intramuscular injection was essentially normal; 2) With regard to excretion, we observed (Fig. 1) the same phenomenon as did Mehta et al., namely, that the excretion of a similar dose of penicillin was slower at the time of admission to hospital than on recovery. In these particular patients, we neither studied renal function, known to be abnormal in this clinical situation (1), nor did we study the unbound penicillin concentration. As stated by the authors, one would not expect, owing to the hypoalbuminemia of protein-calorie malnutrition and therefore the relatively high concentration of free drug present (2), that excretion would be delayed. For this reason, we suggest that inadequate The American
Journal
of Clinical
Nutrition
29: APRIL
renal function is probably the cause of this delay in excretion. This concept is supported by the urinary chloramphenicol levels described by Mehta et al. With regard to the subject of the protein binding of chloramphenicol, the authors make the unreferenced statement that “In normal adults, the drug is bound to plasma albumin to the extent of 60%.” There are, to our knowledge, which is open to correction, no data available as to the exact protein-binding site of chloramphenicol, and its manufacturers (Parke Davis Co.) have not been able to enlighten us further (personal communication). It is generally accepted that chloram-
-
Day
0
Day
10
Day 21
0
0 0
1
2
3
4 Tins
FIG. 1. The lar administration kor, on admission 10 and 21.
1976,
pp. 327-330.
5 wi
6
hours
excretion of penicillin after intramuscuof 50,000 U to children with kwashiorto hospital and subsequently on days
Printed
in U.S.A.
327
Downloaded from https://academic.oup.com/ajcn/article-abstract/29/4/327/4649856 by McMaster University Library, Collections - Serials Processing user on 31 January 2019
to the editor
letters
TO
phenicol is protein-bound, but to which fraction of the plasma proteins is uncertain. We have studied this problem in vitro using ‘4C-Chloramphenicol (Radiochemical Centre, Amersham). Using kwashiorkor serum (albumin 2.3 g/lOO ml) and normal serum (albumin 4.1 g/lOO ml), protein binding was studied by equilibrium dialysis (3). The results are shown in Table I. Although we have not as yet been able to localize the site of binding of chioramphenicol to the plasma proteins, those data showing no significant difference between the free (unbound) component in normal and kwashiorkor serum would suggest that chloramphenicol is not bound to serum albumin. N. Buchanan, M.D. J. D. L. Hansen Metabolic Baragwanath Johannesburg,
and
Nutrition Unit Hospital South Africa
THE
EDITOR
TABLE I The total, bound, of chloramphenicol Total concentration 1.94 3.97 6.0 7.99
Values
#{176}
Norma
and free concentrations in normal and kwashiorkor Kwashiorkor serum
I serum
Bound
Free
Bound
Free
0.66 (34.0) 1.26 (31.7) 1.87 (31.2) 2.43 (30.4)
1.28 (66.0) 2.71 (68.3) 4.13 (68.8) 5.56 (69.6)
0.66 (31.7) 1.31 (31) 1.88 (30.2) 2.69 (32.2)
1.42 (68.3) 2.92 (69) 4.12 (69.8) 5.66 (67.8)
given
are
Total concentration 2.08 4.23 6.23 8.35
(%).
g/ml
Steroids and Microbiology Departments School of Pathology South African Institute for Medical Research Johannesburg, South Africa References 1.
L. A. Van der R. Robinson H. J. Koornhof
Models for protein deficiency: Drs. Corey and Beaton
serum#{176}
Walt
rejoinder
Dear Sir: While Corey and Beaton (1) have withdrawn their original statement with respect to Sukhatme’s model, namely that it does not take into account the inter- and intravariability, they have replaced it by another which is even more objectionable: namely, that Sukhatme’s model fails to make full and proper use of the data. Evidently the authors are unaware of the stochastic models which have found their way into the heart of science and engineering and which we have adopted for explaining the generating mechanism of the time series on daily N balance. We would like to restate here that the linear model we have used (2) represents a data situation in which the population of individuals is partitioned into subpopulations with each individual constituting a subpopulation; observations over time in the same individual constitute a stochastic time series which makes up the
ALLEYNE, G. A. 0. The effect of severe protein malnutrition on the renal function of Jamaican dren. Pediatrics 39: 400, 1967. 2. BUCHANAN, N., AND C. EYBERG. Equilibrium ysis. S. African Med. J. 48: 1867, 1974.
calorie childial-
to
subpopulation
and
is of the y
=
Y
+
w
form: (I)
where Y stands for an individual’s true requirement and w stands for the deviation from Y at given point of time. As against this probabilistic model which respects the order of observations over time, the model used by Corey and Beaton is deterministic in time. They consider that day-to-day fluctuations are random in character arising from errors of measurement and as such can be eliminated by averaging the requirement over a sufficiently long period. That is to say, they assume that the intraindividual component w of the model represented by (1) is zero. However, these assumptions are not borne out by the data. Analysis of data on daily N balance shows that the successive values are correlated through the
Downloaded from https://academic.oup.com/ajcn/article-abstract/29/4/327/4649856 by McMaster University Library, Collections - Serials Processing user on 31 January 2019
LETTERS
328
-.
Dear
with
PM
Sir:
We Mehta
are very interested in the paper by et al. (Am. J. Clin. Nutr. 28: 977, 1975), especially as we are working in this field. We should like to add to some of the data presented by the authors, by drawing an analogy with penicillin, and would care to bring up another point for discussion. We studied penicillin kinetics in a group of children with kwashiorkor and very marked edema (N. Buchanan, R. Robinson, and H. J. Koornhof, unpublished data). The purpose of the study was 2-fold: first, to assess the fate of intramuscular administration of pencillin in these children, when the injection had been given into an edematous thigh, and second, to study the excretion patterns of penicillin, by taking serial blood samples. The children each received 50,000 U of sodium benzyl penicillin on the day of admission, and thereafter on days 10 and 21 of hospitalization, i.e., once the edema had cleared. Two predominant observations arose from this study, which are pertinent to the present discussion: 1) It was shown that even in the presence of marked edema, the absorption of penicillin from an intramuscular injection was essentially normal; 2) With regard to excretion, we observed (Fig. 1) the same phenomenon as did Mehta et al., namely, that the excretion of a similar dose of penicillin was slower at the time of admission to hospital than on recovery. In these particular patients, we neither studied renal function, known to be abnormal in this clinical situation (1), nor did we study the unbound penicillin concentration. As stated by the authors, one would not expect, owing to the hypoalbuminemia of protein-calorie malnutrition and therefore the relatively high concentration of free drug present (2), that excretion would be delayed. For this reason, we suggest that inadequate The American
Journal
of Clinical
Nutrition
29: APRIL
renal function is probably the cause of this delay in excretion. This concept is supported by the urinary chloramphenicol levels described by Mehta et al. With regard to the subject of the protein binding of chloramphenicol, the authors make the unreferenced statement that “In normal adults, the drug is bound to plasma albumin to the extent of 60%.” There are, to our knowledge, which is open to correction, no data available as to the exact protein-binding site of chloramphenicol, and its manufacturers (Parke Davis Co.) have not been able to enlighten us further (personal communication). It is generally accepted that chloram-
-
Day
0
Day
10
Day 21
0
0 0
1
2
3
4 Tins
FIG. 1. The lar administration kor, on admission 10 and 21.
1976,
pp. 327-330.
5 wi
6
hours
excretion of penicillin after intramuscuof 50,000 U to children with kwashiorto hospital and subsequently on days
Printed
in U.S.A.
327
Downloaded from https://academic.oup.com/ajcn/article-abstract/29/4/327/4649856 by McMaster University Library, Collections - Serials Processing user on 31 January 2019
to the editor
letters
TO
phenicol is protein-bound, but to which fraction of the plasma proteins is uncertain. We have studied this problem in vitro using ‘4C-Chloramphenicol (Radiochemical Centre, Amersham). Using kwashiorkor serum (albumin 2.3 g/lOO ml) and normal serum (albumin 4.1 g/lOO ml), protein binding was studied by equilibrium dialysis (3). The results are shown in Table I. Although we have not as yet been able to localize the site of binding of chioramphenicol to the plasma proteins, those data showing no significant difference between the free (unbound) component in normal and kwashiorkor serum would suggest that chloramphenicol is not bound to serum albumin. N. Buchanan, M.D. J. D. L. Hansen Metabolic Baragwanath Johannesburg,
and
Nutrition Unit Hospital South Africa
THE
EDITOR
TABLE I The total, bound, of chloramphenicol Total concentration 1.94 3.97 6.0 7.99
Values
#{176}
Norma
and free concentrations in normal and kwashiorkor Kwashiorkor serum
I serum
Bound
Free
Bound
Free
0.66 (34.0) 1.26 (31.7) 1.87 (31.2) 2.43 (30.4)
1.28 (66.0) 2.71 (68.3) 4.13 (68.8) 5.56 (69.6)
0.66 (31.7) 1.31 (31) 1.88 (30.2) 2.69 (32.2)
1.42 (68.3) 2.92 (69) 4.12 (69.8) 5.66 (67.8)
given
are
Total concentration 2.08 4.23 6.23 8.35
(%).
g/ml
Steroids and Microbiology Departments School of Pathology South African Institute for Medical Research Johannesburg, South Africa References 1.
L. A. Van der R. Robinson H. J. Koornhof
Models for protein deficiency: Drs. Corey and Beaton
serum#{176}
Walt
rejoinder
Dear Sir: While Corey and Beaton (1) have withdrawn their original statement with respect to Sukhatme’s model, namely that it does not take into account the inter- and intravariability, they have replaced it by another which is even more objectionable: namely, that Sukhatme’s model fails to make full and proper use of the data. Evidently the authors are unaware of the stochastic models which have found their way into the heart of science and engineering and which we have adopted for explaining the generating mechanism of the time series on daily N balance. We would like to restate here that the linear model we have used (2) represents a data situation in which the population of individuals is partitioned into subpopulations with each individual constituting a subpopulation; observations over time in the same individual constitute a stochastic time series which makes up the
ALLEYNE, G. A. 0. The effect of severe protein malnutrition on the renal function of Jamaican dren. Pediatrics 39: 400, 1967. 2. BUCHANAN, N., AND C. EYBERG. Equilibrium ysis. S. African Med. J. 48: 1867, 1974.
calorie childial-
to
subpopulation
and
is of the y
=
Y
+
w
form: (I)
where Y stands for an individual’s true requirement and w stands for the deviation from Y at given point of time. As against this probabilistic model which respects the order of observations over time, the model used by Corey and Beaton is deterministic in time. They consider that day-to-day fluctuations are random in character arising from errors of measurement and as such can be eliminated by averaging the requirement over a sufficiently long period. That is to say, they assume that the intraindividual component w of the model represented by (1) is zero. However, these assumptions are not borne out by the data. Analysis of data on daily N balance shows that the successive values are correlated through the
Downloaded from https://academic.oup.com/ajcn/article-abstract/29/4/327/4649856 by McMaster University Library, Collections - Serials Processing user on 31 January 2019
LETTERS
328
