THE NEUTRAL-SULFUR AND COLLOIDAL-NITROGEN TESTS IN THE DIAGNOSIS OF CANCER*

FREDERIC G. GOODRIDGE AND MAX KAHN (Biochemical Laboratories of Columbia University and the Beth Israel Hospital, New York City)

Introduction. During the past few years a number of urinary tests have been suggested for the early diagnosis of cancer. These tests have originated from German and Austrian laboratories; and, immediately after their publication, scientific workers in all parts of the world have endeavored to confirm or disprove the value of these tests, which, if specific, would aid greatly in the conquest of carcinoma. The reports of various observers have been either very favorable or totally discouraging. Accordingly, it is impossible to draw definite conclusions, at present, regarding the efficiency of these laboratory methods.

We have attempted to determine the relative values of the urinary colloidal-nitrogen and neutral-sulfur tests; to study the percentage of positive results obtained with these methods in known cases of malignancy; and to discover, if possible, whether the results of these tests run parallel in cancer and non-cancerous diseases.

COLLOIDAL-NITROGEN TEST. In 1892, Töpfer (1) found that the urine of patients suffering from cancer contained a very large amount of "extractive substance." This "extractive substance" was calculated by first determining the quantity of total nitrogen and then subtracting, from this amount, the sum of the nitrogen values for urea, uric acid, and ammonia, of the same urine. Bondzynski and Gottlieb (2), five years later, reported that the nitrogen in oxyproteic acid, in the urine, was 2 to 3 percent of the total urinary nitrogen. Salkowski (3), and Hess and Saxl (4), using different procedures, concluded that the oxyproteic acid portion of the alcohol-precipitable substances is increased in the urine of human beings suffering from carcinoma.

Salkowski and Kojo (5), in a preliminary communication, recently suggested several methods for the determination of colloidal nitrogen in the urine. A year later, Kojo (6) published the results of a comparative study of the various procedures suggested in this connection. Kahn and Rosenbloom (7) studied the zinc-sulfateprecipitable, colloidal, nitrogenous material from the urine of normal subjects, as well as of carcinomatous patients, and concluded that the amount of colloidal nitrogen was invariably increased in carcinoma. They also found that diseases like myocarditis, diabetes, leukemia, and anemia, likewise gave a high colloidal-nitrogen index. They concluded that this quantitative test was not specific for cancer. Kahn and Rosenbloom (8) studied the amount of colloidal nitrogen in the urine of a dog suffering from a malignant neoplasm. In this case they used dialysis as a part of the method, and found that the quantity of colloidal nitrogen was much greater in the urine of the diseased dog than the amount present in the urine of normal dogs.

Volpe (9) found that the colloidal-nitrogen index is of special value in cancer diagnosis. Mancini (10), using the Salkowski method, found that there were increased eliminations of colloidal nitrogen in the urines of patients afflicted with cancer, but this increase also occurred in pneumonia and pleurisy. Semionov (II) reported that the colloidal nitrogen output is low in normal individuals and is increased in cancer patients. He concluded that although the normal index excludes the possibility of a malignant growth, the increased amount of colloidal nitrogen in the urine is not specific for carcinoma. Konikov (12) found that the average amount of colloidal nitrogen in the urine, as determined by the Salkowski-Kojo method, was 1.68 percent of the total nitrogen in normal cases, and 2.47 percent in carcinomatous individuals. Of 73 cases of cancer investigated by him, only 9 showed a higher coefficient than 2.5 percent.

According to Marcel, Labbé, Dauphin (13) and others, on the other hand, increase in the urinary colloidal nitrogen is an index of a derangement of nitrogenous metabolism; and while it may serve to detect functional insufficiency in the liver, it is not at all specific for cancerous states. Carforio (14), also, concluded that the colloidal nitrogen index is not pathognomonic of cancer.

NEUTRAL-SULFUR TEST. Salomon and Saxl (15) have de

scribed a neutral-sulfur reaction in the urine. Like all the other tests in this connection, it has given excellent results in some hands but, in others, has proved valueless. The abnormal constituent in the urine of carcinomatous patients is a neutral-sulfur fraction, the sulfur of which can be split off by means of hydrogen peroxide, and can be determined as barium sulfate. Positive urines yield 0.010 to 0.018 gm. of barium sulfate from this fraction, for 100 cc. of urine. Of 41 carcinoma cases examined by Salomon and Saxl, 30 were positive, 4 faintly positive, I questionable, and 6 negative. Of 182 normal urines, 6 were positive, 3 faintly positive, I questionable and 172 negative.

Petersen (16) divided his cases into three classes. (A) Clinically non-cancerous suspects: of 26 patients examined, 25 gave a negative Salomon and Saxl neutral-sulfur reaction. (B) Clinically cancer suspects: of 20 cases examined, 5 were negative, 2 alternately positive and negative reactions, and 13 cases positive. (C) Manifest cancer: of 19 cases, 17 always gave a good positive reaction; the two negatives were icteric and cachectic. Dozzi (17) found that the test was invariably negative in all his patients free from cancer or tuberculosis, but the frequency of the positive responses in tuberculous patients detracted from its value as a sign of cancer, although cancer is rarely mistaken for tuberculosis. The only cancer cases that gave negative results were those in which the cancer had been excised. Murachi (18), also, found an increase in the neutral sulfur from cancer patients. The coefficient, according to him, may be 3.8 percent of the total sulfur.

In contrast to the foregoing, Pribram (19) found that only 60 percent of cancer patients gave a positive Salomon-Saxl test and that the test is, therefore, far from specific. Alekseev (20) came to a similar conclusion. Mazzitelli (21) has studied this test in 50 cases of cancer, with and without cachexia. Of 18 cases of the latter variety, the test was positive in 14; but also in 8 of 10 cases of tuberculous cachexia, and 16 of 23 cases of cachexia of various origins, including II with cancer and 4 with tuberculosis, Greenwald (22) concluded that this test has no value in the diagnosis of cancer. Stadtmüller and Rosenbloom (23) studied sulfur metabolism, in general, in carcinoma. They found that the lowest average totalsulfur excretion (0.88 gm. per day) occurred in a series of 13 cases of carcinoma. The same series also showed the lowest average neutral-sulfur excretion (not by the Salomon and Saxl method)-0.20 gm. per day. The proportion of neutral sulfur to total sulfur in the series was considerably higher than the normal proportion. They conclude, however, that "it is a precarious undertaking to diagnose a malignant tumor on the basis of the absolute or relative amount of neutral sulfur in the urine."

Experimental. The following methods were used by us for determinations of the colloidal nitrogen and the neutral sulfur in the urine.

COLLOIDAL-NITROGEN. The urine was first tested for coagulable protein, which, if found, was removed by means of heat coagulation, with addition to the boiling liquid of a few drops of dilute acetic acid sol. To 100 cc. of mixed, filtered, 24-hr. specimen of urine, zinc sulfate was added in sufficient quantity to effect saturation. The saturated liquid was allowed to stand for 24 hours, then was filtered through ashless paper, and the precipitate washed several times on the paper with saturated zinc sulfate solution, to remove nitrogenous substances adherent to the precipitate. The paper and precipitate were then placed in a Kjeldahl flask and the nitrogen content determined by the Kjeldahl method. The total nitrogen in 5 cc. of urine was also determined by the Kjeldahl method. The ratio of the nitrogen in the zinc sulfate precipitate to the total urinary nitrogen was computed.

NEUTRAL-SULFUR. The technic of the Salomon and Saxl neutral-sulfur test is the following: 150 cc. of urine, freed from coagulable protein by heat and acid, are diluted with 100 cc. of water. A mixture of 100 cc. of sat. aqueous sol. of barium hydroxid and 50 cc. of sat. aqueous sol. of barium chlorid is added, the liquid filtered and the filtrate tested with barium to see if precipitation is complete. In order to remove the ethereal sulfates, 300 cc. of the filtrate are treated with 30 cc. of conc. hydrochloric acid sol., and boiled for 15 min. in an Erlenmeyer flask, using a funnel condenser. The flask is then placed on a water-bath for 24 hr. Of the clear filtrate, 200 cc. are mixed with 3 cc. of hydrogen peroxide (perhydrolMerck), and boiled for 15 min. with a funnel condenser. After boiling, the liquid is transferred to a conical graduate, where, at the end of 6 hr., the amount of precipitate is observed. Antipyrin and creosote medications interfere, according to certain authors, with

The accompanying tables present our comparative results.

Table I shows the results obtained in normal subjects. The nitrogen values for the zinc-sulfate precipitate, as compared with those for total nitrogen, varied from 1.25 percent as a minimum, to 2.15 percent as a maximum, with an average of 1.67 percent. This agrees with the results obtained by Salkowski and Kojo, and Einhorn, Kahn and Rosenbloom, who obtained respectively averages of 1.75 percent and 1.9 percent. Of 22 urines examined, Io gave a precipitate by the Salomon and Saxl method that was so light as not to be weighable. The other 12 cases gave sulfate precipitates which varied between 1.22 percent of the total sulfur as a minimum, and 2.14 percent of total sulfur as a maximum. In general the Salomon and Saxl test was negative in all cases in which the neutral sulfur