ation of voltage as a source of error in coulometric titration. Am. J. Clin. Path., 34: 195, 1960.
187. Woodard, H. Q.: The clinical significance of serum acid phosphatase. Am. J. Med., 27: 902-910, 1959.
188. Wootton, I. D. P.: Standardization in clinical chemistry. Clin. Chem., 3: 401405, 1957.
189. Wr6blewski, F.: The clinical significance of transaminase activities of serum. Am. J. Med., 27: 911-923, 1959.
190. Wr6blewski, F., and LaDue, J. S.: Lactic dehydrogenase activity in blood. Proc. Soc. Exper. Biol. & Med., 90: 210-213, 1955.
K.: Isoenzymes and myocardial infarction. New England J. Med., 263 : 531-536, 1960.
192. Yonan, V. L., and Reinhold, J. G.: Effects of ambient temperature on thymol, phenol and zinc turbidity tests. Am. J. Clin. Path., 24: 232-238, 1954.
193. Yonan, V. L., and Reinhold, J. G.: Effects of delayed examination on the results of certain hepatic tests. Clin. Chem., 3: 685-690, 1957.
194. Yti, T. F., and Gutman, A. B.: Quantitative analysis of uric acid in blood and urine. Methods and interpretation. Bull. Rheu-mat. Dis., 7 (Suppl.): S17-S20, 1957.
195. Zauchau-Christiansen, B.: The variation in serum uric acid during 24 hours and from day to day. Scandinav. J. Clin, k Lab. Invest., 9: 244-248, 1957.
new method for the direct determination of serum cholesterol. J. Lab. & Clin. Med., 41: 486-492, 1953.
197. Zucker, M. B., and Borrelli, J.: Platelets as a source of serum acid nitrophenyl-phosphatase. J. Clin. Invest., 38: 148-154, 1959.
COMMENTARY TO Predictive Value of a Single Diagnostic Test in Unselected Populations.
New England Journal of Medicine 274(21): 1171 -1173.
Tolbutamide (Orinase®) was the first successful oral hypoglycemic agent introduced into clinical practice in the 1950s (1). Tolbutamide, a sulfonylurea derivative, stimulates the beta cells in the pancreas to produce insulin. Unger and Madison in 1958 developed a tolbutamide challenge test to screen for diabetes (2). Patients with a glucose level of 80-84% of fasting levels after injection of a standard dose of tolbutamide were claimed to have a 50% probability of having diabetes (3). Thomas J. Vecchio and co-workers at Upjohn pharmaceutical, the manufacturer of tolbutamide, compared the tolbutamide test to the standard oral glucose tolerance test in 102 controls and 40 diabetics (4,5). The results were less than exciting. There was some crossover between the patients classified as diabetic by the glucose tolerance test and those termed diabetic by the new tolbutamide challenge test. Today, this challenge test is considered to have no value in the detection of diabetes (6).
These papers on tolbutaminde by Vecchio and co-workers and many others published during this period are of minor historical note. They demonstrate however how limited the statistical analysis of new tests were at the time. Vecchio in 1966 introduced the concept of the predictive value calculation and provided a statistical tool that accurately assessed the efficacy of a laboratory test. His 1966 paper presented here provided for the first time a single value on a % scale that directly stated the probability that a laboratory test would correctly predict a disease condition.
Vecchio first provides standard definitions for the sensitivity and specificity of a laboratory test. Sensitivity is the ability of a test to detect the disease. Specificity is the ability of a test to give a negative result in a patient who does not have the disease. The predictive value takes these analyses a step further. It calculates the probability that a positive or negative test result is correct. The positive predicative value of a test is obtained from, total positive test results 100
true positive results + false positive results
The first application of Vecchio's predictive value model to published clinical laboratory data may have been by Robert S. Galen in 1974. In that same year a paper appeared in the literature that claimed that the presence of serum antibodies to either cow's milk protein or egg whites was a strong indicator of increased mortality following myocardial infarction (7). From the published data in the paper on the presence of antibodies to cow milk proteins Galen calculated the sensitivity of the test as 74.4%, the specificity as 54.0% and the predictive value as 26.6% (8). This means that almost three-fourths of all results will be false positives. Galen then applied the statistic to the results reported in another paper on a screening method for hepatocellular carcinoma using the alpha-fetoprotein test. He calculated a predictive value of 70% for this test which indicated a 30% false positive rate (9). In 1975 Galen, Reiffel and Gambino examined the efficiency of four cardiac marker enzymes and two isoenzyme markers in the diagnosis of acute myocardial infarction (10). In 100 patients studied only two indicators had a predictive value for myocardial infarction that was above 60%. The presence of an elevated CK-MB (Creatine kinase, EC 188.8.131.52) level or the presence of CK-MB on electrophoresis with a flipped LDH isoenzyme (Lactate dehydrogenase, EC 184.108.40.206) pattern both had a predictive value of 100%.
Galen and Gambino in 1975 published a book on the predictive value method (11). They applied the calculations to the data in numerous published articles and helped establish this parameter as one of the most important measures of a test's efficacy.
(1) Miller, W.L. and Dulin, W.E. (1956) Orinase, a new oral hypoglycemic compound. Science. 123(3197):584-585.
(2) Unger, R.H. and Madison, L.L. (1958) Comparison of response to intravenously administered sodium tolbutamide in mild diabetic and nondiabetic subjects. Journal of Clinical Investigation. 37(5):627-630.
(3) Carawy, W.T. (1970) Carbohydrates, in Fundamentals of Clinical Chemistry. Tietz, N.W. (ed), W.B. Saunders, Philadelphia, Chapter IV, pg 166.
(4) Vecchio, T.J., Smith, D.L., Oster, H.L., and Brill, R. (1964) Oral sodium tolbutamide in the diagnosis of diabetes mellitus. Diabetes. 13(1):30-36.
(5) Vecchio, T.J., Oster, H.L., and Smith, D.L. (1965) Oral sodium tolbutaminde and glucose tolerance tests. Archives of Internal Medicine. 115(2):161-166.
(6) Sacks, D.B. (1999) Carbohydrates, in Tietz Textbook of Clinical Chemistry, 3rd Edition, Burtis, C.A. and Ashwood, E.R. (eds), W.B. Saunders, Philadelphia, Chapter XXIV, pg 776.
(7) Davies, D.F., Johnson, A.P., Rees, B.W.G., Elwood, P.C., and Abernathy, M. (1974) Food antibodies and myocardial infarction. The Lancet. 303(7865):1012-1014.
(8) Galen, R.S. (1974) Food antibodies and myocardial infarction. The Lancet. 304(7884):832.
(9) Galen, R.S. (1974) False-positives. The Lancet. 304(7888):1081.
(10) Galen, R.S., Reiffel, J.A., and Gambino, S.R. (1975) Diagnosis of acute myocardial infarction. Relative efficiency of serum enzyme and isoenzyme measurements. Journal of the American Medical Association. 232(2):145-147.
(11) Galen, R.S. and Gambino, S.R. (1975) Beyond Normality: The Predictive Value and Efficiency of Medical Diagnosis. Wiley, New York.
New Engl. J. Med 1966; 1171-1173
Copyright ©1966 Massachusetts Medical Society. All rights reserved.
Reproduced with permission. 435
Was this article helpful?