CHEMICAL METHODS FOR THE DETERMINATION OF CLINICAL VITAMIN C (ASCORBIC ACID) DEFICIENCY*

Transcription

1 CHEMICAL METHODS FOR THE DETERMINATION OF CLINICAL VITAMIN C (ASCORBIC ACID) DEFICIENCY* GEORGE J. KASTLIN, C. G. KING, CLARA R. SCHLESINGER AND J. WEST MITCHELL From the Medical Department of the Pittsburgh Skin and Cancer Foundation and the Department of Chemistry (Contribution 403), University of Pittsburgh, Pittsburgh, Pa. Chemical methods for studying vitamin C changes in blood, urine, and spinal fluid, and thereby evaluating the clinical state of nutrition, were developed following the isolation of the vitamin as a single substance in In the same year the dye, 2,6- dichlorophenolindophenol, prepared earlier by Mansfield Clark and associates 2 for oxidation-reduction studies, was introduced by Tillmans and his associates 3 for the direct titration of vitamin C. The method was modified by Bessey and King 4 and Harris 5 in 1933, and has since been modified in respect to details of procedure in a number of laboratories. The reduction is not rigidly specific for the vitamin, but is subject to interference by other substances that may be present in tissues having a reduction potential lower than that of ascorbic acid. The vitamin, however, has a much greater reaction velocity than is displayed by most of the other compounds at low ph ranges, and its relative concentration is generally sufficiently high that the possible interfering agents do not significantly alter analytical values. Suitable clinical methods based on titration with the dye have opened the field of vitamin C study for widespread clinical investigation. The daily requirements of the vitamin in health, and to some extent during disease, have been estimated by means * Read before the Nineteenth Annual Convention of the American Society of Clinical Pathologists, June 8th, 1940, New York. Received for publication July 20,

2 DETEKMINATION OF VITAMIN C DEFICIENCY 883 of the chemical methods with far greater certainty than would be possible by earlier types of study. To determine the state of vitamin C nutrition, however, chemical data should be considered in association with the clinical condition of the patient and the combined considerations must be based on the fundamental factors which influence vitamin C requirement and utilization by the body. Among these factors are: (a) Decreased intake of ascorbic acid. The food may be basically deficient in vitamin content or the vitamin may have been destroyed during storage, drying, or cooking of the food. (b) Increased physiological requirement. The high content of vitamin C in foetal tissue and in mother's milk indicates that during gestation and lactation there is an increased requirement of 50 to 100 per eent above normal. The claim that human foetal tissue can synthesize vitamin C has not been confirmed and there is much evidence that the claim is ill-founded. (c) Increased pathological requirement. Infections and toxic conditions may increase the metabolic requirement. Apparently hemochromogen formation in the blood stream may destroy the vitamin by oxidation. The activity of hemochromogens would be more destructive of the vitamin in the blood stream in the presence of a high oxygen concentration than in tissues where the oxygen concentration is lower. (d) Faulty absorption. Low gastric acidity and delayed assimilation may permit oxidative loss of the vitamin after ingestion. Decreased intestinal absorption may also subject the vitamin to destructive bacterial action. (e) Administration of drugs. Drugs may destroy the vitamin either through hemochromogen formation or by other means. This is a field that has been inadequately studied. (f) Age factor. There is reason to believe that the relative requirement per unit body weight is greater during infancy and youth. The growing organism is more sensitive to the tissue injury from borderline vitamin deficiency. (g) Idiosyncracy. Peculiar cases are found occasionally that do not give normal blood or excretion values under known conditions of diet or vitamin therapy.

3 884 KASTLIN, KING, SCHLESINGER AND MITCHELL We shall describe the chemical method which in our experience has been useful and practical in studying vitamin C blood plasma concentration, tissue saturation and urinary excretion in patients. From such observations, plus consideration of dietary history and clinical condition of the patient, we believe that it is possible to evaluate the state of nutrition with specific reference to vitamin C. THE DETERMINATION OF VITAMIN C IN BLOOD PLASMA The micro-technic described by Farmer and Abt 6 is in our experience reasonably simple and accurate. The method was devised to obviate obtaining blood by vein puncture. Sufficient test material was obtained from 0.3 ml. of capillary blood by skin puncture and placed in a special vial with potassium oxalate to prevent coagulation. To 0.1 ml. of blood plasma, distilled water 0.1 ml. and metaphosphoric acid solution 0.2 ml. were added to precipitate the protein. Deproteinized plasma 0.2 ml. was then titrated with a standard dye solution using a micro burette. The use of capillary blood for the test, although particularly useful with infants and for frequent examinations of adults, introduces certain errors. By skin puncture a variable arterial-venous mixture of blood is obtained with possible admixture with tissue fluids, and hemolysis is not infrequent. Blood obtained by vein puncture allows for greater constancy in results. A larger quantity of blood is obtained, which decreases the difficulties and attendant errors in handling minute quantities. Sufficient plasma is also available for checking results. For routine practice we use venous blood and have increased the individual quantities in the test technique, retaining the principles of the method. Technic using Venous Blood: Solutions: (a) 2.5 per cent and 5 per cent metaphosphoric acid. The solutions are stable for approximately two weeks if stored in the refrigerator (Bessey) 7. (b) Standardized solution of 2,6-dichlorophenolindophenol, made up with one tablet of dye or an equal amount of the weighed powder dissolved in 50 ml. of distilled water. Each ml. of the solution is equivalent to approximately 0.02 mg. of ascorbic acid, but the solution should be standardized against pure ascorbic acid or other suitable standard before use each day. The solution is stable for about one week if stored in a dark bottle in the refrigerator. The presence of a small amount of phosphate buffer ph 6.8, increases the stability of the dye solution. A standard thiosulfate solution provides a good means of standardizing the dye 8. Apparatus: Micro burette graduated to ml. (Supplied by E. H. Sargent and Co., Chicago, Illinois); white porcelain titration tile; Pipettes (0.2 and 1.0 ml.) graduated in 0.01 ml. divisions; small glass stirring rod.

4 DETERMINATION OF VITAMIN C DEFICIENCY 885 Collection of blood samples: Obtain 2 ml. of venous blood in a dry syringe, avoiding hemolysis by preventing unnecessary agitation, suction or tissue contact. Place in a tube containing 0.2 ml. of potassium oxalate, evaporated to dryness, to prevent coagulation. Centrifuge immediately. The serum can also be used from carefully protected clotted blood. Protein precipitation: Pipette 0.2 ml. of blood plasma into a 15 ml. centrifuge tube. With the same pipette add 0.2 ml. of distilled water. Add 0.4 ml. of 5 per cent metaphosphoric acid solution. Mix well to precipitate the protein and centrifuge. Plasma ascorbic acid will usually be retained in this state for several hours in the refrigerator. Titration: Pipette 0.2 ml. of deproteinized plasma into a porcelain tile depression. Fill the mercury controlled burette with standard dye solution. Dip the curved end of the burette into the fluid in the tile and while stirring with a fine glass rod, titrate until a pale pink color is obtained which lasts for 5 seconds. The actual titration should be completed in about ten seconds. Titrate a blank of 0.1 ml. of 2.5 per cent metaphosphoric acid to the same end point. Calculation: The number of ml. of dye used for the unknown, minus the number of ml. of dye used for the blank, times the mg. of ascorbic acid per ml. of dye, times 2000, equals the mg. of ascorbic acid per 100 ml. of the blood plasma. THE DETERMINATION OF VITAMIN C IN THE URINE MACRO METHOD Collection of sample: Single specimens are titrated immediately. Twentyfour hour collections are preserved by adding enough 12 per cent metaphosphoric acid to the receptacle to maintain a final concentration of 2 to 3 per cent. A less expensive but less certain method, is to add 10 ml. of glacial acetic acid for each 100 ml. of urine, kept in the refrigerator. 8-Hydroxyquinoline is also an excellent protective agent. Holmes 9 has added carbonate to acidified samples with good results. Titration: In titrating urines that are not excessively colored it is preferable to titrate directly into aliquot samples large enough (generally 10 to 100 ml.) to consume enough of the dye to give good readings. It is less convenient, but sometimes preferable, to titrate the urine from a burette into 5 ml. of standardized dye solution and 5 ml. of distilled water in an evaporating dish. When fresh samples are being titrated, acid should be added to give a ph of about 3.0 to 3.5, to minimize interference from reducing substances other than ascorbic acid. ADVANTAGES OF THE TESTS Metaphosphoric acid, first suggested by Fujita and Iwatake 10 and adapted to general use in the indophenol dye titration by Musulin and King 11 and Bessey 7, protects vitamin C in solution

5 886 KASTLIN, KING, SCHLESINGER AND MITCHELL against atmospheric oxidation more effectively than other acids such as acetic or sulfuric. Mercuric acetate, which has been used by a number of investigators as a precipitating agent for the removal of sulfur compounds, followed by H 2 S reduction, offers no advantage in our experience. The modification is very time-consuming, introduces a number of added hazards, such as removal of excess H 2 S without reoxidation, and involves an additional source of error through the formation of non-ascorbic acid reducing substances. The vitamin is fairly stable in blood plasma, serum (if there has been no hemolysis) and in most urines, if kept cold. The technic is applicable to the routine examination of venous blood samples, and with slight variations to capillary blood if care is taken in collecting and handling the samples. DISADVANTAGES OF THE TESTS The titration must be carried through quickly. The end point, which is not sharply specific for ascorbic acid, is not stable but fades at an indefinite rate after the ascorbic acid has reacted. Hemolysis in blood specimens destroys ascorbic acid rapidly by oxidation; hence careful collection and handling of the blood is necessary. The test solutions are stable for short periods only, so that frequent standardization is necessary. In the urine test other reducing agents, such as thiosulfate and sulfhydryls, constitute a potential source of error that may become appreciable when the excretion levels are low. The test requires scrupulous attention to details in taking samples and in performing the test. CRITIQUE OF OTHER METHODS If a photoelectric colorimeter is available its use is desirable as shown by Bessey 7, and Evelyn, Malloy and Rosen 12. It makes possible less interference from other reducing substances and from nominal amounts of colored or suspended material. Other dyes such as methylene blue, have been recommended from time to time but none has proved so satisfactory in general use as 2,6-dichlorophenolindophenol. Spinal fluid has been recommended for analysis but it apparently does not offer sufficient advantage to offset the greater difficulty in obtaining samples.

6 DETERMINATION OF VITAMIN C DEFICIENCY 887 Subcutaneous injection of the indophenol dye, followed by observing the time required for decolorization of the local area has not proved to be satisfactory. Capillary strength tests based upon either negative or positive pressure for the production of petechiae, commonly reveal cases of marked vitamin C deficiency, but such tests are neither as specific nor as quantitative as the chemical tests. Hence, when used alone the capillary tests are not suitable for the study of marginal deficiencies. NORMAL VALUES The normal daily requirement of vitamin C for the prevention of scurvy has been estimated to be from 10 to 25 mg. 13. It is evident from many investigations, including our own, that 100 mg. is a reasonable minimal standard for health in normal children and adults, and that this amount will normally maintain an approximate, though not full, saturation of the body tissues. From analysis of mother's milk it is clear that an infant normally received 50 to 100 mg. per day 14 ' " 16. It is reasonable to conclude, therefore, that formula-fed infants should receive such an intake and that mothers should be provided with approximately 200 mg. per day during the lactation period. The degree of tissue saturation will be reflected within certain limits, in the blood concentration of the vitamin if the tissue balance is stable. The normal blood levels, expressed in mg. per cent, reported by different workers have varied widely. This is apparently due in large part to variations in technic and in a basic concept of what constitutes the "normal" state of vitamin C nutrition 13 ' 17 ' 18 > 19 ' 20 ' 21. Most investigators believe that a desirable level is 1.0 to 2.0 mg. per cent, and that values above 0.7 mg. per cent are well above the scorbutic level. Estimations of the normal daily urinary excretion are also subject to great variation. The common range may be stated to be 15 to 50 mg., although no definite standard is available 12 ' 22 ' 23 ' 24 ' 25. Urinary excretion is not seriously influenced by diuresis (Johnson and Zilva) 26. A reasonably definite relationship exists between the blood level and urinary excretion, especially if the vitamin intake has

7 888 KASTLIN, KING, SCHLESINGER AND MITCHELL been sufficiently constant to allow a steady balance to develop. A depletion of the vitamin C body stores, with resultant decrease in the blood concentration to a range of 0.7 to 0.4 mg. per cent, may be considered to represent a state of mild deficiency. Blood levels below 0.4 mg. per cent, provide preliminary evidence of severe deficiency. There is an accompanying but irregular decrease in the urinary output of ascorbic acid with the fall in blood values. The statement is frequently made that blood values below 0.4 mg. per cent constitute the zone in which scurvy develops. Blood vitamin C concentrations in this range, particularly in adults, are not necessarily associated with clinical evidence of scurvy, however. A relatively short time is required to develop a deficiency that is evident by chemical tests, compared to the time required to develop tissue changes that are clinically recognizable. Vitamin C is not stored in the body tissues to a degree comparable to the storage of fat soluble vitamins. A relatively constant daily replenishment is necessary to maintain saturation, and a transient variation in vitamin intake may be reflected in the blood level. The latter does not necessarily reveal information in a strictly quantitative sense regarding saturation of the body tissues, but is probably the best single index available. Single determinations of blood level or urinary excretion are of relatively little value, with the possible exception of very high or very low levels. A series of observations correlated with a record of the dietary intake of ascorbic acid can be of distinct value. SATURATION TESTS A number of methods for determining the degree of vitamin C saturation have been reported, in which a test dose of ascorbic acid was given by mouth or parentally, followed by measurement of blood level response or of the excretion of ascorbic acid in the urine. Recently we reported 27 a method based on a test dose of 500 mg. of ascorbic acid intravenously, with observations on blood level and urinary excretion over a four hour period. In-

8 DETEBMINATION OF VITAMIN C DEFICIENCY 889 travenous administration insures maximum uniformity of utilization of the test dose. This test dose is large enough to produce unmistakable changes in the blood level of most patients, and induces sufficient acceleration of urinary excretion to be measured readily. Our observations and the reports from other laboratories indicate that such alterations in the blood concentration return to a stable level in about four hours, and that urinary excretion in this length of time is proportionate to the urinary excretion in the 24 hour period. The 4 hour test period is adaptable to ambulatory patients and eliminates the difficulty of longer periods of urine collection, particularly in children and in difficult patients. METHOD (a) Collect from the fasting (12 to 15 hours) patient a blood sample and specimen of urine to determine the content of ascorbic acid, (b) Give 500 mg. of ascorbic acid in 5 ml. of sterile distilled water intravenously in the arm. (c) Collect the first blood specimen 5 minutes after injection from the opposite arm. (d) Collect subsequent blood specimens and total urines at 1, 2, 3, and 4 hours. (e) Plot the curve of blood ascorbic acid (mg. per cent) and ascorbic acid excreted (mg.) as the ordinate, and time as the abscissa, (f) Compute the total urinary excretion in mg. and the percentage of the test dose excreted. The typical normal saturation curve has the following characteristics : (a) The fasting blood level is 0.7 mg. per cent of ascorbic acid or higher, (b) The five minute blood level is relatively high, ranging from 4.5 to 9 mg. per cent, (c) The rate of return of the blood concentration from the five minute peak to the four hour level is gradual, indicating no great avidity of the tissues for vitamin C. (d) The four hour blood level is distinctly above the fasting blood level, (e) Urinary excretion of the test dose is greatest in the first hour, and the total urinary excretion in four hours is 40 per cent or more of the test dose. The typical severe deficiency curve shows marked deviation from the normal: (a) The fasting blood level is below 0.4 mg. per cent of ascorbic acid, (b) There is only a slight variable rise of the blood level at five minutes, (c) The blood concentration falls rapidly to near the fasting level, (d) The total urinary excretion ranges from a few mg. to 20 per cent of the test dose.

9 890 KASTLIN, KING, SCHLESINGER AND MITCHELL The saturation curves of the normal and severe-depletion individuals are definite in type. The response indicated by both blood concentration and urinary excretion is usually associated with a corresponding dietary intake, though other factors should always be considered in an interpretation of the findings. Many tests which indicate a moderate depletion of ascorbic acid do not conform as nearly as one might expect to the severe depletion curves. The relation of the fasting blood level to the four hour blood level is fairly constant, but there are enough exceptions to the usual response to make this single measurement an unreliable guide to either saturation or depletion. The exceptions are greatest in the borderline or moderate deficiency cases. In our experience the information given by the blood concentration curve and the urinary excretion curve justifies the added work of testing both series of samples. The hourly excretion offers a means of detecting the occasional case of low or high kidney clearance and other types of irregularity. The rate of urinary excretion may be delayed occasionally even when the percentage excretion appears to approach normal. The difficulties of evaluating the saturation tests in patients in the borderline state of nutrition was shown by tests made during treatment of vitamin C deficiency. The first responses indicating an improved tissue storage of ascorbic acid are: (a) An increase in the five minute peak level, (b) a decrease in the rate of fall in blood concentration, and (c) an increase in the total urine excretion of ascorbic acid. These changes may be observed before there is an appreciable rise in the fasting blood level (i.e. above the fasting blood level of a previous test). During depletion there is an early change in urinary excretion, and in the blood concentration curve, before there is clear evidence of an alteration in the fasting blood level. In adults particularly, the saturation tests, as well as the fasting levels cited previously, may show evidence of severe ascorbic acid depletion when there is no clinical evidence of scurvy. This observation gives emphasis to the importance of both chemical and clinical considerations in relation to vitamin deficiency.

10 DETERMINATION OF VITAMIN C DEFICIENCY 891 In common with other required nutrients that have been identified chemically, vitamin C is now recognized as having a much wider scope of influence on nutrition and health than was attributed to it before the chemical tests came into use. An accurate history of the dietary intake of vitamins is difficult to obtain from patients, hence there is a distinct need for the use of direct chemical tests to supplement the other evidence available. The extreme instability of vitamin C in most foodstuffs may readily lead to deficiencies on dietary regimes that on paper appear to be adequate. In borderline cases, the nutritional state of patients is subject to rapid changes without recognition of a change in diet, particularly among those in the lower economic levels and among those who have been subjected to illness. Poor nutrition, in a non-caloric sense, is more common than is generally recognized. The functions of the vitamin in a chemical sense have not been established, nor is there yet a clear picture of the effects of chronic nutritional deficiency. It is becoming increasingly evident, however, that borderline deficiencies give rise to impaired health. Experiments with guinea pigs, for example, show clearly that in the borderline deficiency state, with no external evidence of malnutrition, sublethal quantities of diphtheria toxin produce marked injury to the growing tooth structure, 28 ' 29 though identical treatment of adequately nourished control animals does not induce such injury. SUMMARY Methods for the chemical determination of vitamin C in the blood plasma and urine are described. The limitations of single fasting blood determination as a measure of vitamin C nutrition are given. A clinical method of evaluating the state of nutrition with respect to vitamin C is described based upon (a) determining the fasting blood plasma level and urinary excretion, followed by (b) intravenous injection of 500 mg. of asorbic acid, and (c) subsequent determination of the blood plasma value at 5 minutes and both blood plasma and urinary excretion values after 1, 2, 3 and 4 hours.

11 892 KASTLIN, KING, SCHLESINGER AND MITCHELL The two types of curves (a) vitamin C saturation and (b) vitamin C depletion are complementary. Interpretation of the curves should be based on all factors of the test and should be considered jointly with the dietary history and clinical condition of the patient, particularly in borderline nutritional states. REFERENCES (1) KING, C. G.: Vitamin C, ascorbic acid. Physiol. Review., 16: 238 (1936). (2) COHEN, B., GIBBS, H. D., AND CLARK, W. M.: Studies in oxidation-reduction; Preliminary study of indophenols: (a) dibromo substitution products of phenol indophenol; (b) substituted indophenols of orthotype; (c) miscellaneous. Public Health Reports, U. S. P. H. S., 39: 804 (1924). (3) TlLLMANS, J., HlESCH, P., AND HIHSCH, W.: Das Reduktionsvermogen pflanzlicher Lebensmittel und seine Beziehung zum Vitamin C: des reduzierenda Stoff des Citronensaftes. Z. f. Untersuch d. Lebensmittel, 63: 1 (1932). (4) BESSEY, O. A., AND KING, C. G.: Distribution of vitamin C in plant and animal tissues, and its determination. J. Biol. Chem., 103:687 (1933). (5) HARRIS, L. J., AND RAT, S. N.: Vitamin C and suprarenal cortex: loss of'potency of guineapig suprarenals in scurvy, with notes on method for determining antiscorbutic activity (hexuronic acid) by chemical means. Biochem. J., 27: 303, 2006 (1933). (6) FARMER, C. J., AND ABT, A. F.: Determination of reduced ascorbic acid in small amounts of blood. Proc. Soc. Exper. Biol. Med., 34: 146 (1936). (7) BESSET, O. A.: Method for determination of small quantities of ascorbic acid and dehydroascorbic acid in turbid and colored solutions in presence of other reducing substances. J. Biol. Chem., 126: 771 (1938). (8) MENAKER.M. H., AND GTJERRANT, N. B.: Standardization of 2,6- dichlorphenolindophenol; an improved method. Indust. Engr. Chem., Anal. Ed., 10: 25 (1938). (9) HOLMES, H. N., AND CAMPBELL, K.: Determination of vitamin C in urine. J. Lab. Clin. Med., 24: 1293 (1939). (10) FUJITA, A., AND IWATAKE, D.: Uber die Bestimmung von Vitamin C mittels 2,6-dichlorophenolindophenol. Biochem. Z., 277: 293 (1935). (11) MTJSULIN, R. R., AND KING, C. G.: Metaphosphoric acid in extraction and titration of vitamin C. J. Biol. Chem., 116: 409 (1936). (12) EVELYN, K. A., MALLOT, H. T., AND ROSEN, C.: Determination of ascorbic acid in urine with photoelectric colorimeter. J. Biol. Chem., 126: 645 (1938). (13) SMITH, S. L.: Human requirements of vitamin C. J. Am. Med. Assn., Ill: 1753 (1938). (14) NETTWEILER, W.: Uber den Bedarf an Vitamin C Wahrend Graviditas und Lactation. Klin. Wochnschr., 14: 1793 (1935); Bemerkungen zur Frage des Vitamin C Bedarfes, 18: 769 (1939). (15) SELLEG, I., AND KING, C. G.: Vitamin C content of human milk and its variation with diet. J. Nutrition, 11: 599 (1936). (16) HARRIS, L. J., AND RAY, S. N.: Diagnosis of vitamin C subnutrition by urine analysis, with note on antiscorbutic value of human milk. Lancet, 1: 71, (1935). (17) ABT, A. F., AND FARMER, C. J.: The Vitamins, p. 411 (1939).

12 DETERMINATION OF VITAMIN C DEFICIENCY 893 (18) BKSSBT, 0. A.: Vitamin C; methods of assay and dietary sources. J. Am. Med. Assoc, 111: 1290 (1938). (19) SLOAN, R. A.: Comparison of methods for detecting and grading subclinical scurvy. J. Lab. Clin. Med., 23: 1015 (1938). (20) BBLSER, W. B., HATJCK, H. M., AND STORVICK, C. A.: Study of ascorbic acid intake required to maintain tissue saturation in normal adults. J. Nutrition, 17:513 (1939). (21) MINDLIN, R. L.: Variations in the concentration of ascorbic acid in the plasma of the newborn infant. J. Pediatrics, 16: 275 (1940). (22) SBNDROY, J., JR., AND MILLER, B. F.: Renal function as factor in urinary excretion of ascorbic acid. J. Clin. Investigation, 18:135 (1939). (23) BARRON, E. S. G., BRTJMM, H. J., AND DICK, G. F.: Ascorbic acid in blood and urine after intravenous injection of sodium ascorbate; clinical test for determining vitamin C deficiency. J. Lab. Clin. Med., 28: 1226 (1938). (24) TODHUNTER, E. N., AND ROBBINS, R. C.: Observations on the amount of ascorbic acid required to maintain tissue saturation in normal adults. J. Nutrition, 19:263 (1940). (25) FAULKNER, J. M., AND TAYLOR, F. H. L.: Observations on renal threshold for ascorbic acid in man. J. Clin. Investigation, 17:69 (1938). (26) LUDDEN, J. B., AND WRIGHT, I.: Effect of renal retention of vitamin C on saturation tests. Arch. Internal Med., 66: 151 (1940). (27) KASTLIN, G. J., AND SCHLESINGEB, C. R.: Saturation tests in the determination of vitamin C in the human. Abs. Am. Jr. Path., 16: 663 (1940). (28) SWANSON, W. F., SlGAL, A., AND KING, C. G.: Bacterial toxins and vitamin C in relation to tooth structure. J. Amer. Dental Assn., 23: 2089 (1936). (29) KING, C. G., MUSTTLIN, R., AND SWANSON, W. F.: Effects of vitamin C intake upon the degree of tooth injury produced by diphtheria toxin. Amer. J. Public Health, in press.