dicated an analgesic syndrome. The 7. ANGERVALL L, BENGTSSON U, ZETTERLUND CG, et al: Renal pelvic carexistence of renal disease was concinoma in a Swedish district with firmed by the patients' history, roabuse of a phenacetin-containing entgenographic findings and other drug. Br J Urol 41: 401, 1969 diagnostic studies, including tissue 8. BENGTSSON U, JoH.Nssor.i S, ANGERexaminations. VALL L: Malignancies of the urinary tract and their relation to analgesic Within the past 5 years two of abuse. Kidney mt 13: 107, 1978 the patients in this series have presented with hematuria and were found to have bladder cancer. In What is a stroke? fact, one of the patients had previously undergone left nephrectomy To the editor: It is unfortunate that for papillary transitional cell carci- in Dr. David L. Sackett's summary noma of the pelvis. The patient of the report of the hypertension subsequently underwent periodic task forces and study committees cystoscopy. Four years after ne- (Can Med Assoc 1 120: 1319, phrectomy, however, the hematuria 1979) the word "stroke" is used recurred and was found to be due to once again without qualification. transitional cell carcinoma of the Such use has now become custombladder. The other patient was sus- ary and almost acceptable. Howpected to have cancer of the renal ever, a stroke can be defined as pelvis when the bladder tumour a cerebrovascular accident of sudwas diagnosed, but during explora- den onset; no indication is given as tion only sloughed necrotic papilla to the nature of the underlying adherent to the right renal pelvis pathologic process. The type of was noted. The details of these stroke will have substantial relevance when one is considering the cases will be published elsewhere. In conclusion, patients who have cause, treatment and prognosis. I hematuria with analgesic nephro- believe it is somewhat of a wasted pathy should undergo thorough effort to obtain elaborate epideurologic evaluation for urothelial miologic data and draw conclusions tumours. Once the malignancy of when such fundamental details are the tissue is established, repeated not known. The rupture of a congenital berry histologic and cystoscopic studies aneurysm of the vessels in the circle are necessary. of Willis. is very different from an S. PAUL HANDA, MD, FRCP[C] embolic infarction occurring in a Division of nephrology patient with mitral stenosis or isSaint John Regional Hospital Saint John, NB chemic heart disease, yet the blanket term "stroke" will not take this difference into account. Similarly, References a massive hypertensive cerebral 1. LAKEY WH: Interstitial nephritis due hemorrhage should be considered to chronic phenacetin poisoning. Can different from a transient ischemic Med Assoc J 85: 477, 1961 attack due to carotid artery disease. 2. KOCH B, IRVINE AH, MCIVER JR. The nature of the conditions and et al: Renal papillary necrosis and their pathogenesis, treatment and abuse of analgesics. Can Med A ssoc outcome are completely different. J 98: 8, 1969 Furthermore, the continued use 3. GAULT MH, RUDWAL TC, REDMOND of the word stroke will make it NI: Analgesic habits of 500 veterans: incidence and complications of abuse. impossible to detect any changes in the patterns of the relative freIbid, p 619 4. Wnso. DR: Renal disease due to quency of diseases included under analgesics: II. Analgesic nephropathy that heading. For example, Yates1 in Canada: a retrospective study of has shown that, although the over351 cases. Can Med Assoc J 107: all incidence of stroke has probably 752, 1972 remained the same for many years, S. HULTENGREN N, LAGEROREN C, the frequency of hypertensive cereLJUNGQvIST A: Carcinoma of the bral hemorrhage has decreased, renal pelvis in renal papillary necrosis. Acta Chir Scand 130: 314, 1965 whereas that of cerebral infarction has increased. This change in fre6. BENGTSSON U, ANGERvALL L, EKMAN H, et al: Transitional cell tumors of quency would have remained hidthe renal pelvis in analgesic abusers. den if these two conditions, which Scand J Urol Nephrol 2: 145, 1968 are as different as chalk and cheese, 850 CMA JOURNAL/OCTOBER 6, 1979/VOL. 121
were not differentiated. Such differentiation can sometimes be made clinically, but it will certainly be made during necropsy. This is yet another reason for the continued performance of necropsy. F.M. COLE, MD
Chief, anatomical pathology Section of laboratory medicine McMaster University school of medicine Hamilton, Ont.
Reference 1. YATES P0: A change in the pattern of cerebrovascular disease. Lancet 1: 65, 1964
The logic of SI To the editor: Dr. J.M. Clark's letter (Can Med Assoc J 120: 1207, 1979) is certain to provoke a storm of criticism from the advocates of Syst.me international d'unit.s, the SI unit system. He dared to challenge the wisdom of replacing a familiar scientific system of measurement that works with a system that does not reflect methods of calibration or satisfy the following criteria: simplicity, familiarity, coherence and logic. I would guess that most practising physicians who must explain medical facts to their patients in plain language do not welcome the arbitrary changes. How important is it to retrain physicians in the use of unfamiliar and cumbersome units? The necessity of constant mental conversions will inevitably result in errors, possibly fatal ones. At least the metric system, which is based on a unit erroneously assumed to be 10. of one quarter of the world's meridian (law of the 19th frimaire, year VIII of the Revolution), works and has been universally accepted by the medical community. Now we are told that cells per cubic millimetre must be expressed as cells per litre, the dyne as the joule and the familiar millimetres of mercury as pascals. We are also told that the use of milliequivalents per litre is unscientific, and that millimoles should be used. Why? The new system is neither coherent nor logical. Since the earth and its satellite adamantly refuse to rotate according to the prescribed decimal system, the measurement of time (and circular measures) has remained