Nov.

10, 1951 DIATHERMY TREATMENT OF ANGIOMATA m=H 1125

the' untoward sequelae from diathermy, such as inter- remaining secondary varicosities were treated with the
ference with the epiphysial growth, damage to the eyes, diathermy needle electrode on four occasions. The
or epilation, that may result from radiotherapy. The tumour disappeared entirely, and when the patient was
superficial nature and delicate endothelium of the large examined two years later it had not recurred.
capillary angioma tempted me to treat a few with a flat- I have had no experience of the treatment of the
button electrode of 7 mm. diameter. Cosmetic results angiomata involving a whole limb, and .1 very much
were excellent, and the method was less tiring than doubt if this method of treatment could be applied with
treatment from beneath by the insulated needle, but a success. Fortunately these extensive- lesions are rare.
low-intensity current must be used to avoid overheating
the 'tissues, with subsequent sloughing and scar forma- Summary
tion. The latter method is undergoing further trial with A series of 82 patients with angiomata have been treated
electrodes of various diameters. with diatherny, and these have been followed up for one
Cavernous Angiomata.-These are generally situated year after treatment. All forms of angiomata, except the
deeply and are covered by healthy skin or mucosa, little port-wine stain and angiomata involving a whole limb, may
or no blemish being left after treatment. Some of my be treated with a specially insulated needle. This method is
best results were obtained in these cases (see illustration), simple and safe, and gives a cosmetic result which is com-
parable with that of other methods of treating capillary
angiomata, and is superior to them in the treatment of
cavernous angiomata. The preparation of the special
needle and the way to use it are described.
...

I wish to thank Mr. Wilfred Hynes, our plastic surgeon, for

his great help and criticisms, and for allowing me to treat patients
under his care. I am also grateful to Mr. H. Blacow Yates and
Dr. E. R. A. Cooper for the help they have given me.
REFEPRNCES
Aiken, D. (1944). Lancet, 2, 212.
Ribbert, V. A. (1898). Virchows Arch., 151, 381.

USE AND INTERPRETATION OF THE

FANTUS ESTIMATION OF
Photographs showing cavernous angioma of tower'lip before and
URINARY CHLORIDE
after treatment with the insulated needle. BY

the response often being dramatic. With care the tissues W. H. TAYLOR, B.M., B.Ch., M.R.C.P.
and configuration of the area-apart from the deep (From the Department of Biochemistry, Radcliffe
scarring in the tumour-may he returned to normal. Infirmary, Oxford)
Some of the cases had been treated by irradiation pre-
viously, and in these diathermy was used carefully Since its advocacy by Marriott (1947), the ward test of
because of the danger of sloughing of the irradiated Fantus (1936) for estimating urinary chloride ion con-
tissue. centration has been widely used in clinical practice. This
Port-wine Stains.-These present a problem best left communication describes experiments which have been
to the plastic surgeon. In the one case treated there made to determine its reliability, and records circum-
were obvious capillary overgrowths in parts of the stain. stances' in which interpretation of the test in the normal
The recorded case responded well and the patient was fashion would lead apparently to mistaken conclusions
delighted with the result, but the treatment was an about the electrolyte requirements of certain patients.
arduous and difficult undertaking. The lesion involved Methods
the right side of the face and both eyelids, and great
tempogralh
care artery.gto TheT
was needed arer and lwasresultant
animai
avoid scarring ligatedorandth
contrac- Fantus Test.-The method described by Fantus (1936)
tures. The port-wine portion of the stain remained was used. It is essentially a modification of the Mohr
almost entirely unaffected by the treatment. (1856) titration. To 10 drops of urine a drop of 20%
T'he two capillary lymphangiomata occurred in the potassium chromate solution is added. With the same
tongue. They were considerably reduced in sizke and dropper, silver nitrate solution (29 g. per litre) is added
showed marked improvement after four treatments with drop by drop until a permanent and distinct colour
the needle electrode. Neither case_has required further change to red-brown occurs. The number of drops thus
attention 'for the past two ybars. required is taken as a measure of the content of chloride
The cystic hygroma was a small one situated in the in grammes of NaCl per litre.
left supraclavicular fossa. It disappeared completely Volhard Titration.-The method described by Peters
after a single treatment by the needle electrode four and Van Slyke (1932) was used.
years ago, and there has been no recurrence. Because
of the remarkable power of growth which these tumours Results: (a) Anions Estimated by the Test
possess, treatment at the onset is advocated. It is usual to estimate chloride ions in urine by a
In the present series there was only one case of method involving their precipitation as silver chloride
in acid solution, as in neutral or alkaline solution other
silver compounds may be precipitated and render the
 1126 Nov. 10, 1951 FANTUS ESTIMATION OF URINARY CHLORIDE BRmsss
MEDICAL JOURNAL

estimation inaccurate. The Fantus test is, of course, TABLE I

carried out in neutral solution, or nearly so, depending No. of Drops No. of
upon the pH of the urine. The extent to which the
Correct Urinary Chloride
Concentration (g./litre) of AgNO, Used Readings
formation of other compounds renders the test unreli- 10
able must therefore be determined. This was done by 19 .2 3 ~~~~2
the addition of various anions in the form of the solid 3.
sodium or ammonium salts to 10 drops of urine and
8-5 ... . .{
7
85~~~~~~~~~~~
4
9

then performing the test. 4-0 .. .. . ...8 8

Of the anions commonly present in urine, sulphate,
phosphate, and bicarbonate were without effect upon 4.5 45 1
the result of the test. Urates, however, did affect the 2
result, and the magnitude of this effect must be assessed. 4-5 ... . . . {45 6
6 1
The normal daily excretion of urates in urine is 0.1-2 g.
(Harrison, 1947) measured as uric acid, and' the normal
urinary volume is 1-1.5 litres. The normal concentra- TABLE II
tion of uric acid in urine could therefore vary from
0.007 g./100 ml. to 0.2 g./100 ml. Uric acid is known Correct Urinary Chloride No. of Drops No. of
to form monobasic and dibasic salts with cations, and, Concentration (g./litre) of AgNO, Used Readings
assuming that one molecule of silver nitrate reacts with
one of uric acid, then one drop of the 2.9 g./100 ml.
85.I1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
17 2
silver nitrate solution used in the Fantus test is equivalent 6 2
to one drop of 2.8 g./litre uric acid solution, or to 10 8-5 8~~~~{
LO 21
drops of a 0.28 g./litre solution. .At the most, therefore, .5.7 ... . . .< 6 1
the amount of urate present in 10 drops of urine used 48 2
in the Fantus test would utilize 0.7 drop of the silver
nitrate solution. If the di-silver urate were wholly 6-2 7. . . . { 6 143
formed, 1.4 drops would be utilized.
Carbonate ion was also found to interfere with the
test. The amounts normally present in urine are small, 8 13
but in urine which has stood for some time, especially
if infected, urea is converted into ammonium carbonate,
and this may cause an appreciable error. It was found
that the Fantus test result may be increased by up to
2 g./litre if the urine is allowed to stand for 24 hours at These results afford an illustration of the inability of
room temperature. The presence of silver carbonate the test to give reproducible results even in the hands of
in the reaction mixture at the end-point of the Fantus the more experienced manipulators, and the decreased
test can often be demonstrated by the addition of a reliability with those less experienced. The urine con-
mineral acid to the filtered precipitate. There is an taining 4 g. NaCl/litre was markedly aLkaline and
effervescence of carbon dioxide, which can be detected contained much carbonate ion, giving high Fantus test
by the usual tests. It is important, therefore, always to values which illustrate well the points made in the
perform the test upon a fresh urine specimen. preceding section.
There are other factors which could potentially affect It should be noted, however, that the smaller the
the result of the Fantus test. The chromate ion is only amount of chloride ion in the urine the less will be the
stable in the pH range 6.5-9 (Vogel, 1939). In more inaccuracy caused by drops of unequal size, as fewer
acid solutions the reaction -drops will be needed.
2CrO4-- + 2H+ 2HCrO4- = Cr2O7- + H2O
=

occurs. The tendency of this would be to reduce the (c) Effect of Errors in Practical Use of Test
chromate ion concentration and delay the appearance In 62 urine specimens the results obtained by the
of the end-point in an acid urine. A further factor Fantus test were checked by the Volhard titration. The
which may be of significance in a urine containing large Fantus test was in each case performed by the same
amounts of ammonium ion,-that is, when infected or observer on urines which had stood for not longer than
in acidosis-is the ability of silver ions to form the four hours, so that errors due to the formation of ammo-
argentammonium ion (Ag(NH3)2)+. This factor would nium carbonate and to inexperience with a dropping
again tend to increase the number of drops of silver pipette were minimal. The results are given in Table III.
nitrate solution used.
TABLE III
(b) Error in Performance of the Test
No. of Fantus Test Results
The accuracy of the test depends also upon producing Volhard Titration No. of Giving Error-f
drops of equal volume from a dropping pipette with g./litre Urines Tested
1-2 g.litre 2-3 g.llitre
rubber teat. The ability to do this depends in turn on 0-1 .. .. 26 8 2
the experience of the observer. Thus, two biochemists 1-2 .. .. 5 0 0
and three technicians who were accustomed to using 3-4 2-3 .. .. 3 1 0
.. .. 8 4 0
dropping pipettes made many estimations on the same 4-55-6
..
..
..
..
3
5
0
2 0
urine specimen, giving results shown in Table I. 6-7 .. .. 5 3 1
.. .. 1 0
Three house officers and two medical students, 7-8 8-9 .. ..
2
1 0 0
9-10
unaccustomed to using the method, gave a rather wider 10-11 .. .. 3 1 1
.. .. 0
scatter (Table II).
 Nov. 1951 FANTUS ESTIMATION OF URINARY CHLORIDE BRMLqH 1127
Nov. 10,
10,es 1951r FATU ESTMAIO OFlireURIAR CHLRID occurred
MMICAL JOURNAL
112
The largest error is 2-3 g. / litre, and this occurred in 5 Thereafter for five days an average oral daily intake of
out of 62 estimations (8%). An error of g./litre 2,850 ml. (5 pints) of water or 50% milk in water was given.
Aspirations averaged 1,140 ml. (40 oz.) daily, and collect-
or

over occurred in 26 out of 62 estimations (42%).

able urine 910 ml. (32 oz.) daily. During the last three of
From a statistical point of view these specimens show
these days a total of 3,410 ml. (6 pints) of 5% glucose
a preponderance of urines containing little chloride ion.
solution and 1,140 (2 pints) of 0.9 g./100 ml. saline was
The natural accuracy of the test is greater in these urines, given.
so that the results expressed show the test in as favour- At this stage, on the seventh day after 'operation, the
able a light as possible. plasma chloride ion concentration was 700 mg. NaCl/100
ml. (or 120 mEq/litre) and blood urea 176 mg./100 ml.
(d) Interpretation of the Results of the Fantus Test
The urine, of specific gravity 1015, was completely free of
the chloride ion. Treatment was designed to reduce this
Marriott (1947) considers that in concentrated "
urine,
high chloride level. Over the next four days an average
with a specific gravity of more than 1020, less than oral intake of 2,840 ml. of milk, water, and glucose was
3 g./litre (of chloride measured as NaCl) suggests (salt) given with 1,700 ml. of 5% glucose solution intravenously
depletion." Black (1950) considers that a urinary salt daily. The only sodium chloride received was in the milk.
concentration of less than 5 g./litre indicates the need
The aspirations averaged 1,360 ml. (48 oz.) daily and the
for increasing salt intake. He makes the reservations measurable urine 1,140 ml. (40 oz.). The plasma chloride
that the urine must be reasonably concentrated, levels fluctuated as follows: 690, 700, 790, and 650 mg./
there must be no disease of the suprarenal glands,
that
100 ml. (I 18, 120, 135, and 111 mEq/litre) over the next four
and
days. The serum sodium level on the last of these days
that the Fantus test may fail to show salt deficiency in
was 162 mEq/litre (372 mg./100 ml.) and the plasma bi-
the presence of acidosis. Marriott also mentions that
carbonate 15 mEq/litre (35 vols./100 ml.). The blood urea,
during intravenous saline therapy there may be an after rising to 417 mg./100 ml., fell to 271 mg./100 ml.
adequate concentration of chloride ion in the urine Throughout this time the urinary chloride concentration
although the body is still salt-deficient. varied between 0 and 0.4 g./litre, as measured by the
There are patients, however, in whom such an inter-
Volhard method. The urinary urea concentration on two
pretation of the Fantus test may be misleading occasions was 2.6 and 3 g./litre. Fantus test estimations
: first,
made by the nursing staff usually gave urinary chloride
those in whom there is hyperchloraemia and hyper-
concentrations of 0-1 g./litre and 1-2 g./litre, although
natraemia but a urinary chloride ion concentration of
on two occasions misleadingly high values were obtained-
less than g./litre in a concentrated specimen; and, 2-3 g./litre and 3-4 g./litre.
secondly, those with hypochloraemia and hyponatraemia At this point, 11 days after the operation and five before
yet with a high urinary chloride concentration and symp- death, chloride ion was passed for the first time and the
toms of salt deficiency but with no evidence of Addison's urinary concentration rose to 4.5 g./litre. However, chest
disease. This second group may be subdivided into those infection developed and the patient became febrile and
(mentioned by Marriott) in whom saline transfusion is
generally more ill. It proved impossible to give more than
being carried out and those in whom there is no such
1,700 ml. (3 pints) of fluid by mouth daily. Blood, glucose
therapy. solution, hypotonic saline, and sodium sulphate were given
intravenously in addition. The patient's condition deterior-
During the past year 12 patients of the hyperchlor- ated steadily and he died 16 days after the operation.
aemia group have been noted in the Radcliffe Infirmary. At necropsy there was thrombosis of several mesenteric
They consist of 6 fatal cases of head injury, non-fatal veins, causing gangrene and perforation of the colon at a
case of perforated peptic ulcer (post-operative), fatal point near the recto-sigmoid junction. Uraemic ulceration
case of carcinoma of the oesophagus (post-oesophag- of the large and small bowel and bilateral pulmonary con-
ectomy), fatal case of fat embolism (post-traumatic), gestion and oedema were also found. The coronary arteries
fatal case of diabetic coma, fatal case of meningitis
were atheromatous, with narrowing of the lumina. There
and lung abscess, and fatal case of post-operative was no significant renal disease, macroscopically or micro-
par-

scopically.
tial gastrectomy. The cases with head injury have been

reported in some detail in a separate communication Comment

with other workers (Higgins et al., 1951). Had this patient been given intravenous saline therapy
The biochemical findings in the following patient are
during the period of hyperchloraemia and hyper-
fairly typical. natraemia, as the urinary chloride tests would have
indicated, it seems likely that the plasma sodium and
Case Report chloride levels would have risen even higher.
A aged 57 with a 20-year history of
man epigastric pain During the same period 10 patients falling into the
ch&racteristic of peptic ulceration developed the symptoms second, or hypochloraemic, group were noted. Six were
of pyloric stenosis. Seven months previously he had
rec.i'vIng intravenous saline transfusions post-opera-
an attack of renal colic, but a subsequent intravenous
tively, and four not. Three of the six were children.
pyelogram had shown good concentration of the dye

One, a boy of 11 years subjected to investigative lapar-

no calculi in the kidneys. The blood-urea concentration

was 33 mg./100 ml. After a period of preparatory

otomy, passed 7.5 g. NaCl/litre in a urine of specific
cal treaiment, a Billroth I partial gastrectomy was
gravity 1012 when his plasma chloride level was 87mEq/
formed. Post-operative progress was stormy, with litre (510 mg./
100 ml.) and his serum sodium level 128
rent haematemesis and a persistently low blood pressure. mEq/litre (297 mg./100 ml.). Another child of 8 weeks,
Two weeks after the operation a hypostatic chest infection
following Ramstedt's operation for pyloric stenosis,
with rise of temperature developed, and the patient
passed 5.4 g. NaCl/litre in his urine on a day when his
For the first two days after operation fluid balance was plasma chloride level was 82 mEq/litre (480 mg./100
maintained mainly by intravenous transfusion; a
ml.) and he appeared clinically to be salt-deficient. Of
1,140 ml. (2 pints) of blood, 1,140 ml. (2 pints) of 0.9
g./
the three adults, two were post-abdominal operation
100. ml. saline, and 1,700 ml. (3 pints) of Ringer-Locke
cases and one was in diabetic hyperglycaemic coma.
solution
given. The was given.
measurable By mouth, 1,140
urinary output
ml. of water
totalled 1,730 ml. The remaining four all suffered headinjuries severe
(61 oz.) and the stomach-tube aspirations 1,250 ml. (44 oz.). enough to keep them unconscious for several days.
 1128 Nov. 10, 1951 FANTUS ESTIMATION OF URINARY CHLORIDE M
Despite the maintenance of an adequate milk, water, and disturbance is inability of the kidneys to excrete the
glucose intake, and an adequate urinary output, they chloride ion. An intravenous saline transfusion or oral
developed unmistakable signs of salt deficiency. One salt intake will then produce high plasma chloride and
man aged 84, for example, was thought to have had a sodium levels. Use of the urinary chloride concentra-
cerebral thrombosis causing him to fall and sustain a tion as a guide to therapy would lead to precisely this
head injury. Five days later the normal skin turgor state of affairs and run the risk of oedema, particularly
was lost, th; intraocular tension was diminished, the of the lung bases.
cheeks were hollowed and sunken, and the tongue was The hypochloraemic group of patients, who pass a
dry and wrinkled. His urine was found to contain urine of relatively high chloride concentration while
7.8 g. NaCl/litre and his serum sodium ion concentration receiving saline transfusion, are most interesting and
was 130 mEq/litre (300 mg./ 100 ml.) and plasma chloride need characterizing further. Again, only the minority
ion concentration 82 mEq/litre (480 mg./100 ml.). No of patients so treated behave in this.way. I

intravenous fluids had been given. These four patients The remaining group of patients lose both sodium and
responded slowly to Ringer-Locke solution by mouth. chloride ions spontaneously and become progressively
Had the Fantus test been performed and interpreted in salt-deficient. They behave as though they had Addi-
the usual way, this fluid would have been denied them, son's disease, and appear to differ essentially from the
as they would not have been considered salt-deficient. transfusion group, who may not pass chloride ions in the
urine prior to transfusion. Again, it is apparent that use
Discussion of the urinary chloride concentration would lead to treat-
The foregoing results pose two separate problems ment quite the reverse of that required.
frst, the accuracy of the Fantus test as a measure of Other workers have recently criticized the value of
urinary chloride ion concentration; and, secondly, the urinary chloride tests. Wilkinson et al. (1950), after
value of that concentration as a guide to sodium and investigating sodium and chloride balance post-opera-
chloride ion requirements in disease. tively, remark: " It is a corollary of the hypothesis we
Assuming for the time being that the urinary chloride have advanced that determinations of urinary chloride
concentration is of some value, it is important that the afford no reliable guide to the requirements of salt and
water during the immediate post-operative period, excel-
Fantus test should give a fairly accurate result for values ,lent though they are in other circumstances." Spencer
from 0 to 5 g. NaCl/litre, over which range it is going (1950) simPnarly remarks : " Chloride determinations on
to provide a basis for Na and Cl therapy. Although, the urine are sometimes valuable guides to diagnosis and
for reasons given above, the natural accuracy of the test treatment, but they do not necessarily reflect the balance
is greater at lower urinary chloride concentrations, the of chloride and sodium in the body." He then goes on
values obtained under conditions approximating to the to quote examples from the literature.
best still show an undue amount of error. A urine The conclusions seem to be (a) not to use the Fantus
containing less than 1 g./litre ought not to be estimated
as containing 2-3 g./litre as it was in 2 out of 26 cases. test if a more accurate method is available, and (b) never
In the hands of less practised persons, such as house to rely on a urinary chloride estimation alone as a basis
officers and nursing staff, an error of this magnitude for salt therapy, but to take into account also the clini-
might occur more frequently. In the case described. cal signs and the plasma chloride (or, better, sodium)
for example, such urines were twice estimated as con- concentration.
taining 2-3 g./litre and once as containing 3-4 g./litre. Summary
It is surely indefensible that such results be used as a The errors and reliability of the Fantus test have been
basis of treatment, especially when the Volhard esti- investigated.
mation is in itself so easy and accurate and involves the Its use may lead to serious error in urinary chloride
use of only two solutions, a pipette, a burette, and a estimations, especially in the hands of inexperienced persons.
conical flask. The urinary chloride concentration may not be an accu-
The usefulness of urinary chloride estimations may be rate guide to salt therapy. Examples where this is the case
questioned on a number of grounds. In the first place, are given.
sodium is probably the most important ion maintaining
extracellular fluid volumes, and urinary chloride ion Elliot I wish to thank Dr. V. Smallpeice, Professor L. J. Witts, Mr. A.
and Mr. J. C. Scott for permission to use the case
variations are of use only in so far as they reflect changes recordsSmith,
of patients under their care, and the staff of the Bio-
in urinary sodium ion concentration. This they may chemistry Department for performing the blood analyses in the
fail to do in acid (Black) or alkaline urines, for example. case described. My thanks are also due to the technical staff,
Secondly, their usefulness is considerably diminished house officers, and medical students who performed the experi-
of the body for sodium ments mentioned in Section B, and to Mr. J. R. P. O'Brien for
if they do not reflect the need advice and encouragement.
ions. The schemes of treatment given by Marriott and
by Black assume that they do reflect this need, with the REFERENCES
exceptions stated. There is considerable evidence to Black, D. A. K. (1950). British Medical Journal, 1, 893'
support this, particularly in the salt-depleted states pro- Fantus, B.G.-(1936). J. Amer. med. Ass., 107, 14.
A. (1947). Chemical Methods in Clinical Medicine,
duced in the Tropics (Marriott) or by forced sweating Harrison, 3rd ed. Churchill, London.
(McCance, 1936) in otherwise normal individuals. In Higgins, G., Lewin, W. S., O'Brien, J. R. P., and Taylor, W. H.
(1951). Lancet, 1, 1295.
most patients in whom problems of fluid balance arise in Marriott, H. L. (1947). British Medical Journal, 1, 245, 285, 328.
hospital this relationship between urinary chloride con- McCance, R. A. (1936). Lancet, 1, 825.
centration and the body's requirement of sodium ions Peters, J. P., and Van Slyke, D. D. (1932). Quantitative Clinical
but arise. Those quoted in Chemistry, vol. II. Bailliere, Tindall and Cox, London.
still holds good, exceptions Spencer, A. G. (1950). Lancet, 2, 623.
this paper have all come to light in a hospital of 842 beds Vogel, A. I. (1939). Quantitative Inorganic Analysis. Longmans,
Green and Co.. London.
in one year. The hyperchloraemic group is discussed Wilkinson, A. W., Billing, B. H., Nagy, 0G., and Stewart, C. P.
more fully elsewhere, where it is argued that the primary (1950). Lancet, 2, 135.