All results in micromole equivalentsof norleucineper 24 h. Av of four runs. ...
positive alkaline hydrolyzable materi- al in urines of three boys with muscu-.
Dermatomyositis:Urinary ChromatographicFindingsby Alkaline Autoanalysisfor Peptides. (One Case) To the Editor: In a previous study of the ninhydrinpositive alkaline hydrolyzable material in urines of three boys with muscular dystrophy and eight normal controls we demonstrated significant differences between the two groups for certain of the fractions separated by cation-exchange chromatography with the Model PNC-1 Technicon Peptide Analyzer (1). Resultsof laterwork on other patients were similar. We have now done a series of analyses for ninhydrin-positive and alkalisensitive peaks, by the same technique, on urine of a patient with dermatomyositis. The patient (an 18-yearold girl weighing 41 kg) received no
i
ii
patient with dermatomyositis that of a normal girl and a boy Duchenne muscular dystrophy indicatesthe possibility that in
I
diseases the Fig.
1. A
typical
ion-exchange
chro-
matogram of a normal urine, indicating the ninhydrin-positive peaks in the unhydrolyzed sample (U) and the hydrolyzable components in the al-
kali-hydrolyzed
sample
(H)
for the patient on prednisone alone and on simultaneous prednisone gram
and
months, continuously with prednisone and periodically with methotrexate. When first seen at this institution in 1969, the patient had a far-advanced loss of skeletal muscle power and a loss of skeletal muscle mass estimated at 90%. She had skin granulomas of the face and arms, which improved after medication during the first year, but her motor disability continued. During the subsequent three years of treatment with decreasing doses of prednisone alone, the patient improved remarkably. Figure 1 shows a chromatogram of a normal urine, before and after hydrolysis. Table 1 gives data for cer-
comparable data for a normal untreated girl of the same age and a boy of the same age with muscular dystrophy. The drugs themselves give no peaks on such a chromatogram.
tain components
in this chromato-
methotrexate
Urinary
treatment,
and
excretion of peptide-like
pathological changes
from normal are reflected in an elevated peak V (which is mainly creatine), a decreased peak F, and an especially marked decrease in peak Z (which is mainly creatinine). In many (but not all)patients with muscular
dystrophy, we also see elevation of peaks G and X, but not of peak W. In this study, peaks W and X were elevated in the patient’s urinary chromatogram,
special diet, but was treated for six
with with only both
and
this
may
be
specific
for dermatomyositis. Although both these two diseases involve disorders of muscle metabolism, the comparison is open to criticism because the excretion patterns were not monitored before the patient was treated with prednisone. (The patient was first seen in our clinic
substances varies greatly between individuals, although it is fairly consistent for an individual. Therefore, inter-individual comparisons are of limited value. In this study the decrease of all substances included in the Table-and especially those comprising peak V (which is mainly cre-
shortly after the treatment with pred-
atine)-appears
really related to the beneficialinfluence of this drug in this particular
to
result
from
the
treatment with methotrexate and prednisone. Comparison of excretion data of the
nisone ported
was begun.) The changes here have occurred during four years of treatment.
rethe
Methotrexate, an antimetabolite in
the folic acid pathway, processes. How decrease
disease
in
slows cellular the observed
much
excretion
would
patients with
of
require
studies
creatine
on more
this illness.Further
Table 1. Data Obtained with the Technicon Peptide Analyzer for Urine from Two Girls (One with Dermatomyositis, the Other Normal) and One Boy with Duchenne Type of Muscular Dystrophy Peak (see Figure Subject
Age, yr
Weight, 41
Prednisone(1969) Methotrexateand prednisone (1970)
2 yr of treatment;
V
W
X
Z
3456
2967
436
1126
2137
3160
2232
358
946
1803
2560
825 594
125
590
1800
149
331
2305
90
285
5286
180
440
670
patient
strikingly improved (1971) 4 yr of treatment (1973) Normalc dystrophy
(Duchenne)d
1053 17.5
56
3748
17
51
2000
418 2700
All results in micromole equivalentsof norleucineper 24 h. Av of four runs. “Av of 15 runs. Av of two runs, except the figureforpeak W, which isfrom one run only. dOne singleanalysis(whichcorrelateswellwithaverages foryounger patients with this disease
1412
CLINICAL
1)
F
18
Dermatomyositis
Muscular
kg
CHEMISTRY,
Vol. 19, No. 12, 1973
(1)1.
is
identification of peaks X and W might be particularly interesting, especially if patients treated with methotrexate and separately with prednisone could be compared with similarly treatedcontrols. Quantitation and identification of peak G is more difficult because of overlapping of several components, hippuric acid being the major one.
Table 1. Comparison of Standards Prepared in Different Sera’ Blank DIgoxin, ng/ml
3
this
study,
and
Miss
Juanita
Lyons,
0 1
for
3
hertechnical assistance.
Reference 1. Maskaleris,M. L.,Gross,S.,and Milhorat,A. T., Urinary amino acidand pep-
No.
New std.
%
0 1
We especially thank Dr. J. Marchand, under whose care the patient was during
subtracted
Old std.
blank
Old std.
New std.
Binding
36.8 27.8 19.0
36.2 28.0 19.2
39.4 30.2 21.4
39.1 30.6 21.2
36.1 28.2 18.9
35.5 27.4 18.8
38.7 30.6 21.3
38.1 29.4 21.3
Percent binding has been calculated both with, and without subtracting a blank. Values are average of duplicates.
tideexcretionpatternsin patients with muscular dystrophy (Duchenne). Clin. Chem. 15,545(1969).
Milton
S. Gross L. Maskaleris
Institute for Muscle Disease, 515 East 71st St. New York, N. Y. 10021
Inc.
More Accurate Radioassayof Digoxin To the Editor:
A recent report (1) brings into question the accuracy of routine digoxin assays.A given amount of digoxinis shown to give different values for percent binding in the presence of different normal plasmas. These differences as reported would cause a coefficient of variation (CV) of about 40% for digoxin values of about 1 ng/ml in plasma in the assay using [3H]digoxin.We findthat thesedifferences can be easilyavoided without extraction, and the assay can be routinelyperformed with a CV of 6% or less. Burnett et al. (I) do not detail the method they are using, but indicate it is similar to that of Smith et al. (2). In our laboratory, this procedure is used as described (2) with the addition that blanks (serum without antibody) are run with standards and with each sample. Counts per minute in the blank supernate are subtracted from counts per minute in the sample
standard deviationof 1.05forthe percent binding. The 95% confidence limits are 42.7 ± 2.3% binding. Burnettet a!.assayed 20 normal plasmas (plasma containing no digoxin) with a
says reported by Burnett et a!. Our standards are made up in serum, and digoxin concentrations are usually determined in serum from patients.Oc-
standard deviation for percent binding of 2, and 95% confidence limit of 52.7 ± 4.2 percent binding. Both re-
Table 2. Examples of High Serum Blanks for Five Patients’
sultsdescribewithin-assayvariation. We ran dilution curves on several
Blankstandard blank
patients’ sera. Figure 1 shows results
obtained forone patient.Others were similar.The percent binding curve for the patient is superimposable on the standard curve, indicating identical
reactivities
for
patient
and
Digoxin, ng/ml
stan-
dard. Severalreasons may be considered to account for the difference between these results and those obtained by Burnett et al.: (a) They do not mention running blanks, i.e.,samples containing no antibody, with each plasma. However, when standard curves are calculatedwithout blanks,
Serum -
CPM
5.25 0.75 3.02 3.36
906 2681 581 405
61 2449 56 8
2.00
1111
45
‘Serum
radioactivity
was determined
by
assaying samples without the addition of (8Hidigoxin. Second patient had had a brain scan with use of technetium 99.
casionally plasma
is
and serum
reportedly
results
4).
(2,
as plasma give the same
run,
Recently
we
assayed
plasma (heparin) and serum from blood drawn at the same time from 10 patients. Both plasma and serum were run in duplicate,and the standard deviation and CV calculated for the paired samples. The standard deviation for the serum samples was 1.08 percent binding (CV, 4%); standard deviation for the plasmas was 1.64% (CV, 5.8%). When serum and plasma from the same patient were
r
supernate.
standard solutions are prein 95% ethanol, and working standards are made by adding 250 zl of 1 g/mg stock standard to 50 ml of digoxin-free serum. Working standards are prepared every two months, and a standard curve with the new standard is compared to the previous standard. Table 1 shows two typical comparisons. When the same [3H]digoxin, antibody, and charcoal-dextran are used, the curves for percent binding are identical. Zero standard values for 12 pairs of normal sera (serum containing no digoxin) gave a
the reproducibility of the curves is not altered (Table 1). This would be expected, as blanks with normal serum usually have about the same value. Patients’ serum blanks, however, are occasionally observed to be high compared to standard blanks (Table 2). Sometimes, but not always, this can be accounted for by radioactivity in the serum. These blank values may be twice the standard blank or more, and would significantly alter the percent binding observed ifthey were not subtracted.We have not been able to correlate high blanks with specificdrugs or diseasestates, but they are generallyfound forsera from very sick patients. The normal plasmas described by Burnett et al. would not fall into this category. (b) Plasma was used in all the as-
Stock
pared
compared, the standard deviation was 4.64 percent binding (CV, 16%). The CV for the data of Burnett et al. for
_________________________ ,‘
,
‘
,‘j’
20 control plasma samples is 4%, indicating the variationseen is probably a function of the precision of the assay.
Fig. 1. Standard curve compared withdilution ofpatientserum The percent binding for the highest serum sample is fitted to the curve and the other values are plotted versus amount of serum used. X, patient serum; #{149}, standard; values are av. of duplicates
CLINICAL
The difference in standard curves cannot be so readily explained. In spite of our lack of precision with plasma samples,we found good correlationbetween our standard curve in serum and standard curves prepared CHEMISTRY.
Vol. 19. No. 12. 1973
1413
