ly unsound. I do not believe physicians should allow themselves to be placed in so untenable a position. Can Dr. MacLachlan be serious in his suggestion that no major problems exist in the processing of abortion applications because nearly all physicians quickly learn the "rules" of the "game"? Surely the issue is whether the female applicant can rely on an honourable, equitable, nonpunitive, and geographically and economically accessible means of obtaining an abortion. I fully agree with Dr. Thomas's conclusion in his recent editorial2 that these inequities would be ameliorated by eliminating the therapeutic abortion committee. M.E. KRAss, MD Greater Niagara General Hbspital Niagara Falls, Ont.
References 1. BADOLEY RE (chmn): Report of the Committee on the Operation of the Abortion Law, Ottawa, Supply and Services, 1977 2. THOMAS WDS: The Badgley report on the abortion law (E). Can Med Assoc / 116: 966, 1977
Probability of multiple myeloma in patients with low back pain To the editor: By equating the probability of abnormalities in the results of any of three screening tests with the probability of disease, Dr. J.C. Fernbach, in his letter "Multiple myeloma and low back pain" (Can Med Assoc J 116: 712, 1977), has committed the fallacy of affirming the consequent. The original thesis - if someone with back pain has a disease, then he will have an elevation of one of the screening test values - does not imply that if he has an elevated test value, then he has a disease. Decision analysis depends not only on the probability of abnormal results in disease, but also on the probability of abnormal results in "nondisease" This is best described by Bayes' theorem:1
measured may be derived as follows: Assuming that the probability of an abnormal ESR in a patient without multiple myeloma is 10% (likely low), that the probability of multiple myeloma in older patients with low back pain is 0.3% (likely high), and that the probability of an elevated ESR in multiple myeloma is 80%, then the probability of multiple myeloma after an elevated ESR has been found is 2.3%. Then, assuming that the probability of an elevated calcium concentration in a patient without multiple myeloma is 2.5% (random elevation), and that the probability of an elevated calcium concentration in multiple myeloma is 40%, then the probability of multiple myeloma after an elevated ESR and an elevated calcium value have been found is 27.4%. Therefore, measurement of the ESR and serum calcium concentration is inadequate as a screening procedure for multiple myeloma. Whether multiple myeloma should be screened for at all in patients with low back pain must be determined by a cost-effectiveness calculation.3 R.A. BLATrEL, PH D, MD 476 Holland Ave., Ste. 101 Ottawa, Ont.
References 1. KATZ MA: A probability graph describing the predictive value of a highly sensitive diagnostic test. N Engi J Med 291: 1115, 1974 2. FAGAN TJ: Nomogram for Bayes's theorem (C). N Engi J Med 203: 257, 1975 3. BEAR R, SCHNEIDERMAN J: Decision analysis in clinical medicine (E). Can Med Assoc J 115: 833, 1976
Newly identified constituents of renal calculi: 2,4-diketones
To the editor: The investigations of the late L.E. Francis, the late KI. Melville and myself1 into the chemical nature of the antiallergic principle in extracts of various human and mammalian tissues have led to the identification in such extracts of a group of homologous 2,4P(D) P(DIT) diketones with chain lengths ranging P(DjT) = P(D) P(T1D) + P(D) P(TjD) from 13 to 25 carbon atoms (a paper describing this work is in preparation). where P(DIT) is the probability of dis- In addition to their wide distribution in ease after the test, P(D) is the probabil- soft tissues, these compounds are presity of disease before the test, P(TID) ent in urine. Members of this lipid class is the probability of a positive_test in a are avid chelating agents, combining patient who has the disease, P(D) is the with di- and trivalent metal ions to probability of nondisease before the form neutral complexes with properties test and P(TID) is the probability of a of organic compounds. positive test in a patient who doesn't Since they occur in urine, we thought have the disease. This equation is most it would be of interest to determine if easily solved by a nomogram2 or by 2,4-diketones are present in renal cala pocket calculator. culi. A number of renal calculi of variWith Dr. W. Pruzanski's example ous types and one sample of multiple (ibid) of multiple myeloma, the prob- prostatic calculi submitted to our laboability of the disease after the erythro- ratory for analysis were examined for cyte sedimentation rate (ESR) and these compounds by a gas chromatoserum calcium concentration have been graphic method. In all cases in which 222 CMA JOURNAL/AUGUST 6, 1977/VOL. 117
a sufficiently large sample of calculus or calculi was available the 2,4-diketone fraction could be demonstrated, at a concentration of 3.6 mg per 100 g of calculus or less. It is likely that the diketones are present in the calculi as their chelates, and that they are associated with the organic matrix, which is regarded as an essential component in the formation of calculi.2 It is not known whether these 2,4diketones have some significant role in the formation and growth of calculi or whether their presence is merely fortuitous. We hope to continue this investigation as further clinical material becomes available. A more detailed account of our observations will be reported later. D.E. DOUGLAS, PH D, P CHEM Division of clinical chemistry Department of medicine Montreal General Hospital Montreal, PQ
References 1. Fs.i.cis LE, MELVILLE KI, DOUGLAS DE: Antiallergic activity and some chemical properties of an antihistamine principle in human and animal tissue extracts. Can J Biochem Physiol 41: 1961, 1963 2. Bovca WH: Organic matrix of human urinary concretions. Am I Med 45: 673, 1968
Canadian Critical Care Society To the editor: Over the past few months, interested individuals have been attempting to initiate the formation of a Canadian Critical Care Society. This society will be open initially to all medical practitioners interested in critical care medicine, including physicians, surgeons, pediatricians, anesthetists and others. Close liaison with the Royal College of Physicians and Surgeons of Canada is anticipated and, indeed, the inaugural meeting of this subsection of the Royal College will be at the annual meeting of the college in Vancouver in January 1978. A half-day program is anticipated. The scientific meeting will include an inaugural address by Dr. Bryan Kirk on the history and future of critical care medicine in Canada. It is hoped that there will be a response in the form of abstracts, which will be included in the Royal College call for abstracts; we will follow their format. Any person interested in participating in this new organization may contact me at the address below. Further information will be available as the society becomes more concrete in form and function. RONALD L. HOLLIDAY, MD Acting secretary Canadian Critical Care Society Department of surgery Victoria Hospital 391 South St. London, Ont. N6A 4G5
References 1. BADOLEY RE (chmn): Report of the Committee on the Operation of the Abortion Law, Ottawa, Supply and Services, 1977 2. THOMAS WDS: The Badgley report on the abortion law (E). Can Med Assoc / 116: 966, 1977
Probability of multiple myeloma in patients with low back pain To the editor: By equating the probability of abnormalities in the results of any of three screening tests with the probability of disease, Dr. J.C. Fernbach, in his letter "Multiple myeloma and low back pain" (Can Med Assoc J 116: 712, 1977), has committed the fallacy of affirming the consequent. The original thesis - if someone with back pain has a disease, then he will have an elevation of one of the screening test values - does not imply that if he has an elevated test value, then he has a disease. Decision analysis depends not only on the probability of abnormal results in disease, but also on the probability of abnormal results in "nondisease" This is best described by Bayes' theorem:1
measured may be derived as follows: Assuming that the probability of an abnormal ESR in a patient without multiple myeloma is 10% (likely low), that the probability of multiple myeloma in older patients with low back pain is 0.3% (likely high), and that the probability of an elevated ESR in multiple myeloma is 80%, then the probability of multiple myeloma after an elevated ESR has been found is 2.3%. Then, assuming that the probability of an elevated calcium concentration in a patient without multiple myeloma is 2.5% (random elevation), and that the probability of an elevated calcium concentration in multiple myeloma is 40%, then the probability of multiple myeloma after an elevated ESR and an elevated calcium value have been found is 27.4%. Therefore, measurement of the ESR and serum calcium concentration is inadequate as a screening procedure for multiple myeloma. Whether multiple myeloma should be screened for at all in patients with low back pain must be determined by a cost-effectiveness calculation.3 R.A. BLATrEL, PH D, MD 476 Holland Ave., Ste. 101 Ottawa, Ont.
References 1. KATZ MA: A probability graph describing the predictive value of a highly sensitive diagnostic test. N Engi J Med 291: 1115, 1974 2. FAGAN TJ: Nomogram for Bayes's theorem (C). N Engi J Med 203: 257, 1975 3. BEAR R, SCHNEIDERMAN J: Decision analysis in clinical medicine (E). Can Med Assoc J 115: 833, 1976
Newly identified constituents of renal calculi: 2,4-diketones
To the editor: The investigations of the late L.E. Francis, the late KI. Melville and myself1 into the chemical nature of the antiallergic principle in extracts of various human and mammalian tissues have led to the identification in such extracts of a group of homologous 2,4P(D) P(DIT) diketones with chain lengths ranging P(DjT) = P(D) P(T1D) + P(D) P(TjD) from 13 to 25 carbon atoms (a paper describing this work is in preparation). where P(DIT) is the probability of dis- In addition to their wide distribution in ease after the test, P(D) is the probabil- soft tissues, these compounds are presity of disease before the test, P(TID) ent in urine. Members of this lipid class is the probability of a positive_test in a are avid chelating agents, combining patient who has the disease, P(D) is the with di- and trivalent metal ions to probability of nondisease before the form neutral complexes with properties test and P(TID) is the probability of a of organic compounds. positive test in a patient who doesn't Since they occur in urine, we thought have the disease. This equation is most it would be of interest to determine if easily solved by a nomogram2 or by 2,4-diketones are present in renal cala pocket calculator. culi. A number of renal calculi of variWith Dr. W. Pruzanski's example ous types and one sample of multiple (ibid) of multiple myeloma, the prob- prostatic calculi submitted to our laboability of the disease after the erythro- ratory for analysis were examined for cyte sedimentation rate (ESR) and these compounds by a gas chromatoserum calcium concentration have been graphic method. In all cases in which 222 CMA JOURNAL/AUGUST 6, 1977/VOL. 117
a sufficiently large sample of calculus or calculi was available the 2,4-diketone fraction could be demonstrated, at a concentration of 3.6 mg per 100 g of calculus or less. It is likely that the diketones are present in the calculi as their chelates, and that they are associated with the organic matrix, which is regarded as an essential component in the formation of calculi.2 It is not known whether these 2,4diketones have some significant role in the formation and growth of calculi or whether their presence is merely fortuitous. We hope to continue this investigation as further clinical material becomes available. A more detailed account of our observations will be reported later. D.E. DOUGLAS, PH D, P CHEM Division of clinical chemistry Department of medicine Montreal General Hospital Montreal, PQ
References 1. Fs.i.cis LE, MELVILLE KI, DOUGLAS DE: Antiallergic activity and some chemical properties of an antihistamine principle in human and animal tissue extracts. Can J Biochem Physiol 41: 1961, 1963 2. Bovca WH: Organic matrix of human urinary concretions. Am I Med 45: 673, 1968
Canadian Critical Care Society To the editor: Over the past few months, interested individuals have been attempting to initiate the formation of a Canadian Critical Care Society. This society will be open initially to all medical practitioners interested in critical care medicine, including physicians, surgeons, pediatricians, anesthetists and others. Close liaison with the Royal College of Physicians and Surgeons of Canada is anticipated and, indeed, the inaugural meeting of this subsection of the Royal College will be at the annual meeting of the college in Vancouver in January 1978. A half-day program is anticipated. The scientific meeting will include an inaugural address by Dr. Bryan Kirk on the history and future of critical care medicine in Canada. It is hoped that there will be a response in the form of abstracts, which will be included in the Royal College call for abstracts; we will follow their format. Any person interested in participating in this new organization may contact me at the address below. Further information will be available as the society becomes more concrete in form and function. RONALD L. HOLLIDAY, MD Acting secretary Canadian Critical Care Society Department of surgery Victoria Hospital 391 South St. London, Ont. N6A 4G5
