certainly when there is dissent about a standard volume. But if we would all accept the recommendations of the International Union of Biochemistry, then for enzymes in solution " per ml." would be understood, and no-one could object to its pedantic use. most
Group Pathology Department, North Lonsdale Hospital,
TRANSPLACENTAL PASSAGE OF IRON-DEXTRAN
SIR,łIron-deficiency anaemia is not uncommon in pregnancy. Although most women respond satisfactorily to oral iron, about 10% require parenteral iron, which is usually given from the second trimester onwards. A method of giving parenteral iron therapy which is becoming more widely accepted is the total dose infusion (T.D.I.) of iron-dextran complex (’ Imferon ’).1-5 During the first few days after infusion high concentrations of iron-dextran are circulating in the plasma. Values of the order of 5000-10,000 g. iron per 100 ml. have been reported 5 days after infusion. These fall to roughly 300 g. iron per 100 ml. at 10 days, and are within physiological limits by 15 days .6It is desirable to ascertain whether or not iron in excessive quantities, either as iron-dextran or in some other form, crosses the placenta and enters the foetal circulation. The placental transfer of a high-molecular-weight ferric carbohydrate has been studied in rats by Nylander.18 After intravenous injection, the preparation was metabolised in maternal tissues and did not cross the placenta in an unchanged form. At term there was no evidence of iron overload in the foetuses. More recent experiments have been carried out in rats and rabbits using 5’Fe-labelled iron sorbitol citrate, iron saccharate, and iron-dextrin. After parenteral administration there was no significant placental transfer of the materials studied.9 10 Uptake of iron by the foetus from iron-dextrin was 1.85% of the dose at 72 hours, and from iron-sorbitol 1,:"7% at 24 hours. Roughly a third was in the foetal liver. Transfer of iron was a function of the weight of the foetus.ll Clinical studies with iron sorbitol have shown that iron transport across the placenta is enhanced, but could be accounted for by physiological mechanisms. It has been estimated that 0-15% of the maternal dose crosses the placenta.12 The hxmoendothelial placentx of the rat, mouse, and rabbit are very different from the hxmochorial placentae of man and of the macaque. For this reason transplacental transfer of iron in the rodent and the rabbit cannot be assumed to take place by the same mechanism as that which occurs in man, but in the rhesus monkey (Macaca mulatta) the transport mechanism is more likely to be comparable with that of man. Cotes et ap3 have studied the fate of iron administered to pregnant rhesus monkeys as a single intravenous infusion of 59Fe-labelled irondextran during the last third of gestation. They found that 0-25-4-5% of the maternal dose of iron derived from the iron-dextran was recovered in the foetuses, and this amount appeared unrelated to the maternal dose of iron. There was no microscopic evidence of excessive iron storage in the foetuses, although there was some siderosis in foetal tissues. Light-microscopic studies on experimental rhesus placentae and on human placentae after infusion show prussian-bluepositive granules in the trophoblast.14 Iron-containing granules are also seen on the foetal side of human placentar 15; similar staining granules are, however, present in control human material. 1. 2. 3. 4.
Basu, S. K. J. Obstet. Gynæc. Br. Commonw. 1965, 72, 253. Bonnar, J. Br. med. J. 1965, ii, 1030. Clay, B., Rosenberg, B., Sampson, N., Samuels, S. I. ibid. 1965, i, 29. Dawson, D. W., Goldthorp, W. O., Spencer, D. J. Obstet. Gynæc. Br. Commonw. 1965, 72, 89. 5. Varde, K. N. ibid. 1964, 71, 919. 6. Marchasin, S., Wallerstein, R. O. Blood, 1964, 23, 354. 7. Nylander, G. Acta Soc. Med. upsal. 1954, 59, 372. 8. Nylander, G. Nature, Lond. 1954, 174, 310. 9. Wohler, F. Curr. ther. Res. 1964, 6, 464. 10. Högberg, K. G., Lindvall, S. Br. J. Pharmac. Chemother. 1964, 22, 275. 11. Lindvall, S. Jectofer Proceedings Conference (edited by H. E. D’Amato); p. 53. Södertälje, Sweden, 1964. 12. Evers, J. E. M., Van Kessel, H. Ned. Tijdschr. Geneesk. 1964, 108, 1476. 13. Cotes, P. M., Moss, G. F., Muir, A. R., Scheuer, P. J. Br. J. Pharmac. Chemother. (in the press). 14. Muir, A. R. Personal communication. 15. Walters, G. Personal communication.
Examination of the infused human and rhesus trophoblast by electron microscopy shows small particles with an inherent electron density contained within membranous vesicles in the trophoblast. These particles are similar in size and structure to pure iron-dextran; they do not have the ultrastructure of ferritin or hasmosiderin, and it is surprising that these substances cannot be identified in the specimens obtained after
iron-dextran infusion.14 These histological studies suggest that some of the infused material reaches and is retained by the trophoblast. This is confirmed by chemical analysis of human placenta: after irondextran infusion; these show a high initial level of nonhxm iron, which drops rapidly in the first 14 days after infusion. The control levels of placental iron do not return until about 5 weeks after infusion. Preliminary studies on the serum-iron and total iron-binding capacity (T.LB.C.) in babies born of mothers given iron-dextran infusions in the later stages of pregnancy suggest that there is no passage of iron-dextran or excess iron across the placenta once
maternal serum-iron values have returned
levels. In two patients, when infusions were given within a fortnight of delivery, foetal serum-iron values at term exceeded physiological levels. The values observed in cord-blood were 360 g. and 1060 g. iron per 100 ml. The excess serum-iron in the foetus is thought to be present as iron-dextran.15 The evidence suggests that, after infusion of iron-dextran in pregnancy, a small proportion of the infused iron crosses the placenta, possibly as unchanged iron-dextran. The iron stores of the offspring of infused mothers do not contain excessive amounts
Research and Development Department, Fisons Pharmaceuticals Limited, Holmes Chapel, Cheshire.
J. S. G. Cox P. MILLS G. F. Moss.
ADENOSINE AND PLATELET ADHESIVENESS SIR,-I WISII to dispute rrolessor (.,lUICk’Sstatement (jan. " Even though Dr. Misirlioglu claims that A.T.P. (adenosine triphosphate) is an activator of the fibrinolytic mechanism, it must be recognised that these findings are obtained by testtube experiments." My own work and that done with Professor Shafiroff, which I cited in my letter,l has been carried out in man and the results confirmed in the test-tube. Application of this principle has resulted in decalcification of arteries and reversal of essential hypertension in about 200 patients.
Y. I. MISIRLIOGLU. DIABETES MELLITUS AND NEOPLASIA SiR,łThe possibility of an association between diabetes mellitus and neoplasia has been considered many times, with remarkably varying results. As the establishment of either a positive or a negative association could be of value, the question has been investigated using data now available for 24 populations from five continents. The data consist of death-rates for diabetes, and for neoplasms in 21 sites. They were adjusted for age to a standard population, and the rates for diabetes were compared with those for the neoplasms. A positive correlation was found between the rates for diabetes and those for malignant neoplasms of the prostate, and a negative one between diabetes and malignant neoplasms of the stomach (both sexes). The probability of the result being due to chance was less than 1 in 100 for prostate neoplasms, and less than 1 in 1000 for stomach neoplasms. Positive correlations which may be of importance were also found between diabetes and malignant neoplasms of the female breast and of the uterine cervix. There are undoubtedly many sources of error in these data, the more obvious of which are as follows: 1. These data
death-rates-for most malignant neoplasms reflection of the incidence, but this is not true for diabetes. 2. It is known that there is a large number of undetected diabetics, and most of these will not be reflected in the death-rates.
Lancet, 1965, ii, 850.
822 3. It is very improbable that in all the countries concerned medical will agree on what constitutes a death from diabetes. The commonest causes of death in this disease are cardio-renalvascular lesions, and such deaths will not always be certified as due to diabetes.
It is therefore clear that the data must be regarded with considerable suspicion, and that nothing could be regarded as firmly established on the basis of such evidence. The best way to determine whether these associations are real is by means of a prospective survey, involving the follow-’iip of a large population of diabetics over a period of years. Such an investigation is now being carried out by this fund with the cooperation of the British Diabetic Association and the Registrar General’s Office. Should these associations be confirmed, then the reasons for them and the possibility of their use in the treatment or prevention of the neoplasms concerned would be considered. It is felt that speculations on such lines would be premature until the completion of the investigation outlined. I shall be pleased to supply copies of the data used to anyone interested. Statistics Unit,
Imperial Cancer Research Fund, Lincoln’s Inn Fields,
A. J. LEA.
intravenous liquid should be given fast, but when the circulation is restored as much time as possible must be allowed for equilibration. If this is done, fits are uncommon, and there is usually a urea-induced diuresis. Acidosis, if severe as judged by the respiration, may need correction with bicarbonate or lactate, but often settles as the circulation improves. If the similar picture in diabetes is due to dehydration from glucose diuresis in a patient drinking too little water (or too much sugary " tonic fizz "), the use of 5% dextrose solution initially, as suggested by Dr. Halmos and his colleagues, may be unwise.
T. H. HUGHES-DAVIES.
INSULIN SENSITIVITY OF HUMAN BRAIN
SIR,-Professor Butterfield and his colleagues (March 12)
the well-known clinical observation that in f insulin-dependent diabetic patients, hypoglycxmic symptomsI bear little relation to blood-sugar levels. This feature surely ! underlines the fact that disturbances of consciousness are more closely related to the intracellular metabolism of the cerebrum than to levels of circulating metabolites. An in-vivo assessment of such metabolic activity may be made by the measurement of cerebral oxygen consumption,I which according to Kety1 has a mean value in the region of 3-3 c.cm. of oxygen per 100 g. of brain tissue per minute (i.e., 40 c.cm. per minute for a brain weighing 1200 g.) in conscious " and alert individuals. Under normal conditions the fuel for this activity is exclusively carbohydrate, since values of cerebral respiratory quotient (R.Q.) in normal man approximate to units. 1Since each mole of glucose (180 g.) requires 6 moles of oxygen (134-1 litres) for its complete oxidation, a simple calculation reveals that an expected cerebral oxygeni consumption of 40 c.cm. per minute should involve the utilisation of roughly 53 mg. per minute of glucose. This is within the range of values of glucose uptake measured by Professor Butter- í field and his colleagues, representing in terms of energy con- I sumption about 20% of the expected total base metabolism. Traditionally, the brain is said not to depend on insulin for its metabolism. In support of this view are the observations that normal cerebral glucose uptakes and R.Q.s close to unity are found even in dogs rendered ketotic by fat-feeding,3 or diabetic by pancreatectomy.4 Professor Butterfield and his colleagues demonstrate that the intravenous infusion of insulin can actually i decrease the rate of cerebral glucose-uptake in man, while [ simultaneously increasing the rate of peripheral glucose-uptake. The interpretation they offer is that the cerebral threshold for glucose falls more slowly than does the peripheral threshold in response to insulin. Perhaps an alternative explanation for their unexpected observation is that a rising level of extracellular insulin may actually tend to block the entry of glucose into cerebral tissue, while facilitating its passage across the cell membranes of peripheral tissues. Conversely, the diabetogenic growth hormone or fat-mobilising substancewhich is released from the pituitary gland during starvation, fat-feeding, or diabetic ketosis may block the entry of glucose into periperal tissues while permitting its entry into the brain. Since consciousness appears to depend on the organised and uninterrupted flow of electrons from hydrogen to oxygen within the cerebrum, any interference with this final common (aerobic) pathway is liable to lead to disturbance of consciousness pari passu with increasing concentration of hydrogen ion within the neurone. Interference with this orderly transport may result from hypoxia (acute or chronic), lack of substrate (in hyperinsulinism) or of coenzymes (in vitamin-B-deficiency state), or the presence of depressant agents of endogenous or exogenous origin. In diabetic precoma these agents include excess ketone bodies and hydrogen ions produced (anaerobically) by peripheral
HYPEROSMOLAR COMA IN DIABETES
SiR,łIread with interest the
of eight cases of hyperosmolar non-ketoacidotic coma in diabetes reported by Dr. Halmos and his colleagues (March 26). I was surprised that they did not mention possible upsets in potassium metabolism in this condition. The prolonged osmotic diuresis produced by the severe glycosuria in these patients, by embarrassing potassium reabsorption in the renal tubules, may lead to hypokalxmia, especially if the process continues over several weeks. The accompanying severe dehydration might mask the hypokalsmia, so that the serum-potassium could appear normal. Treatment of these patients by rehydration and large doses of insulin will promote the movement of potassium ions from the extracellular fluid into the cells, and this will further aggravate the hypokalxmia. Hypokalsemia might well have been the cause of the sudden death of two of the patients reported, and might also have contributed to the episodes of hvuotension. Edgware General Hospital, Middlesex. J. O. HUNTER. account
SIR,-The hyperosmolar coma in diabetic patients described by Dr. Halmos and his colleagues is very like the hypernatrasmic form of gastroenteritis in infants. These patients have usually received milk during their illness. This milk, if digested but not absorbed, fills the gut with a very hyperosmotic solution which draws water from the circulation, producing a rapid fall in blood-volume with a rise in hoemoglobin and electrolyte concentration. (Some infants have a very high blood-sugar, too-perhaps because of a failure of peripheral utilisation.) Sodium levels of 180 mEq. per litre, and urea levels of 300 mg. per 100 ml. or more, are not uncommon. The traditional response to such biochemical abnormalities was to give the
child intravenous glucose in water. Unfortunately this often leads to death, for, after preliminary improvement, the sunken fontanelle bulges as water passes into the cerebrospinal space and cells more quickly than urea and electrolytes can escape, as in the postdialysis syndrome; and sudden collapse from " coning " or fits may occur. If the child survives, the kidneys may fail temporarily-perhaps because of similar passage of water into recently anoxic renal cells-and treatment becomes difficult. Later, thrombosis is common. In such patients, normal saline is hypotonic enough at first, and nothing weaker than half-normal physiological saline in glucose solution should be given intravenously. Initially
Kety, S. S., Schmidt, C. F. J. clin. Invest. 1948, 27, 476. Gibbs, F. L., Lennox, W. G., Nims, L. F., Gibbs, F. A. J. biol. Chem. 1942, 144, 325. 3. Milder, A. G., Crandall, L. A. Am. J. Physiol. 1942, 137, 437. 4. Himwich, H. E., Nahum, L. H. Prov. Soc. exp. Biol. Med. 1929, 26, 496. 5. Chalmers, T. M., Pawan, G. L. S., Kekwick, A. Lancet, 1960, ii, 6. 1. 2.