BRITISH MEDICAL JOURNAL

218

but that the increase in the rate of remission in AML is small and not significant. A significant difference could be expected when larger numbers are analysed and/or when more effective induction regimens are discovered, but it seems to us that comparisons between expensive sophisticated systems and routine ward care are irrelevant. Certainly the purchase of expensive equipment of unproven value is to be avoided. What is needed in our several circumstances is the application of the principles of asepsis with economy and common sense. H E M KAY K ATKINSON H McD CLINK B JAMESON J LYNCH T J MCELWAIN Leukaemia Unit Royal Marsden Hospital, Sutton, Surrey 1 Levine, A S, Robinson, R A, and Hauser, Janet M,

European Journal of Cancer, 1975, 2, suppl. Gaya, H, Jameson, B, and Storring, R A, Chemotherapy Progress, 1975, 1, 93. 3Crowther, D, et al, British Medical Journal, 1973, 1, 131. 4Medical Research Council, British Journal of Haematology, 1974, 27, 373. 6Freeman, C B, et al, British Medical Journal, 1973, 4, 571. 6Levine, A S, et al, New England J7ournal of Medicine, 1973, 288, 477. 7Schimpff, S C, et al. Annals of Internal Medicine, 1975, 82, 351.

SIR,-While supporting the views of Dr C Bunch and his colleagues (3 January, p 40) that the place of plastic isolators in the management of bone marrow depression is not clearly established and that widespread use of such techniques is inappropriate at the present time we feel that it would be a pity if nothing were said in support of this isolator system. It is clearly demonstrable that the Trexler isolator effectively excludes exogenous microorganisms from the patient's environment, whereas other systems such as laminar air flow rooms appear less effective, as was brought out at a recent symposium on bone marrow transplantation held in New York. On the other hand, an isolator tent can be effective in reducing risk from infection only if endogenous sources are dealt with by adequate microbiological decontamination of the patient. This entails the vigorous use of a wide range of antimicrobial agents applied to the skin, oral cavity, gut, and body orifices. Clearly the indications for such a demanding and costly regimen must be restricted to profound, prolonged bone marrow depression, as in patients with aplastic anaemia in whom marrow transplantation is contemplated. Patients with severe combined immune deficiency disease also benefit since decontamination may significantly reduce the severity of graft-versus-host reactions following marrow transplantation. In our experience infection has been the major cause of death in bone marrow transplantation for aplasia and severe combined immune deficiency (three out of four patients who died in a series of eight recent grafts); and similarly the Seattle Transplant Group reported 10 deaths from bacterial sepsis in a series of 24 patients with aplastic anaemia who were given bone marrow grafts.' In marrow transplantation and in acute myeloblastic leukaemia under treatment with new chemotherapeutic regimens we feel it is desirable to use the maximum possible protec-

tion against microbial infection, and we now therefore use the Trexler isolator with full decontamination of the patient. A randomised trial comparing different forms of isolation is certainly required, but paradoxically this would entail the wider availability of these plastic isolators. We should like to add that our patients, who are mainly children, have tolerated the plastic isolator remarkably well. A J BARRETT S SELWYN Department of Haematology and Bacteriology, Westminster Medical School, London SWI 1

Storb, R, et al, Blood, 1974, 43, 157.

24 JANUARY 1976

dizygotic twinning rates4 and anencephalic rates5 by maternal age are different, the peak for twins being around 35-39 years and for anencephaly around 20-24 years. The most important piece of evidence not adequately explained, however, and commented on by the Liverpool group is the common occurrence of ASB in the first pregnancy. The accompanying table shows the distribution of ASB cases, all (live and still) births, and multiple births only by pregnancy order, defined as previous reproductive experience associated with livebirth, stillbirth, or abortion, in women resident in Belfast in 1964-8.6 7 Around 25% of ASB cases are born to women pregnant for the first time, which, although different, is of the same order as the proportions for all births (30%) or multiple births (20 %h) delivered to primiparae. The suggestion

Aetiology of anencephaly and spina bifida that mothers whose first pregnancies result in ASB offspring have previously been pregSIR,-The paper by Sir Cyril Clarke and nant but are unaware of this fact seems unlikely others (27 December, p 743) further elucidat- unless it is postulated that their recall of such ing the hypotheses formulated by Rogers' events differs markedly from primiparae and Knox2 postulating that anencephaly and having normal offspring. Some difference spina bifida (ASB) may result from an inter- with respect to genetic (for example, blood action between twin fetuses or between a fetus group) or environmental (for example, infecand residual trophoblastic material is of tions or diet) factors between women who have an affected first pregnancy and women considerable interest. On the basis of Knox's fetus-fetus interaction having later pregnancies affected seems worthy hypothesis a singleton ASB fetus at birth is the of investigation. Despite these inconsistencies with the outcome of a number of events (at least three) occurring in a certain time-related sequence early hypothesis as currently formulated, it has the in pregnancy. The conditions required are: firstly, merit of presenting a biologically plausible that the pregnancy at conception is of dizygotic mechanism for ASB occurrence which may be twin type; secondly, that these two fetuses differ tested by further epidemiological and genetic genetically in a sex-linked manner with respect research. to a dialletic gene system on the X chromosome; J H ELWOOD and thirdly, that some environmental factor (or factors) may behave as a trigger mechanism initiating the interaction. It follows therefore that the prevalence of anencephaly in a population, for example, is associated with the dizygotic twinning rate, the frequency of relevant genes, and the frequency of the relevant environmental factor. One further observation also may result-namely, that the anencephalic sex ratio and in particular the proportion of affected females is related to the frequency of the initiating environmental factor. Over relatively short periods of time gene frequencies may be assumed to be constant; hence any secular change in anencephalic rate should be associated with changes in dizygotic twinning rate, in the anencephalic sex ratio, or in both. Geographical variation in rates also should show similar patterns if examined cross-sectionally during similar time periods. From an epidemiological point of view it is too early to give a definite answer regarding the fit of observational data to this hypothesis. I agree with the view of the Liverpool workers and Knox's analysis of certain British data, indicating sufficient agreement to merit further research. At first glance findings from Irish data are similar because not only are ASB rates3 and twinning rates' high but also the anencephalic rate in Dublin and the twinning rate in the Irish Republic over the period 1959-68 have both declined. The latter association between these two rates, however, is not formally significant (Spearman's rank correlation coefficient=0-624; n= 10; for this sample size and at a 5% significance level r=0-648) probably because the number of years studied is too small. On the other hand the distributions of Irish

Number of previous pregnancies 0 1 2 3+ Total

All births

(live and still) 12 367 10 071 6 782 12 131 41 351

Department of Social and Preventive Medicine, Queen's University, Belfast

Rogers, S C, in press. Knox, E G, British Journal of Preventive and Social Medicine, 1974, 28, 73. 3Elwood, J H, Irish Journal of Medical Science, 1975, 144, 388. 4 Dean, G, and Kean, T, British Journal of Preventive and Social Medicine, 1972, 26. 186. Elwood, J H, unpublished. Elwood, J H, and Nevin, N C, British Journal of Preventive and Social Medicine, 1973, 27, 73. 7Elwood, J H, and Nevin, N C, Ulster Medical Journal, 1973, 42, 213. I 2

Paradoxical rise in blood pressure during propranolol treatment

SIR,-We 'have read with interest the short report by Dr I Blum and others (13 December, p 623) showing that in patients with various forms of psychotic illness treatment with the /-adrenoceptor blocking agent propranolol may increase blood pressure. Since propranolol may cause similar hypertensive episodes in patients with phaeochromocytoma the authors suggest that the hypertension is associated with high levels of circulating catecholarnines, which in the presence of fl-blockade produce unopposed a-adrenoceptor stimulation. We consider that the following experiments support this concept.

Multiple births only

Anencephaly

Spina bifida

Both defects

205 201 169 452

48 42 20 63

42 32 39 72

90 74 59 135

1 027

173

185

358

70

Noradrenaline

-

I

60

after

pro?ranolol -50E

LU 44

E40-

blockade n =14 after saline

C

n20

u

10

P, and Hartel, G, British Medical Yournal, 1973, 2, 178.

Effect of junior doctors' action on self-poisoning

70

Phenylephrine

SIR,-Admission rates to this department have not shown the recent dramatic changes reported by Dr G S Crockett (10 January, p 92). The number of female admissions due to self-poisoning in November and December 1975 was 93 compared with 85 for the same two months in 1974 and 72 in 1973. Junior doctors have been working normally in accident and emergency services in this city. This fait may explain the difference in admission rates between Bristol and Kettering, thus supporting Dr Crockett's conclusions.

4,o -

E50 O

-

T

40

after 30 infusednolol propra

-/

n=8

after 20 infused saline n =7

10

ROGER GABRIEL Renal Unit, Royal Infirmary, Hull 1 Karhunen,

8o c

pressor effect of propranolol in the circumstances reported from Israel would not have been seen if the same high dose level had been reached over a longer time. The experiment should be repeated over an extended period to show that the hypertension is an induced artefact and not a feature per se of the valuable drug propranolol.

X infused before

30

E

219

24 JANUARY 1976

BRITISH MEDICAL JOURNAL

-

A D BROOKMAN Accident and Emergency Department, Royal Infirmary, Bristol

8(XO) 100 3200 6400 200 4(X Rate of infusion ( nq/min) Dose-response curves for pressor agents infused Treatment of whooping cough intravenously into anaesthetised rats. SIR,-I was interested to see the letter from

Noradrenaline infused into pentobarbitoneanaesthetised rats at 310 and 1480 ng/min caused marked increases in arterial blood pressure (see fig). These responses were augmented' when the noradrenaline infusions were repeated after an intravenous injection of propranolol (1 mg/kg body weight), but they were not augmented by a control "blockade" with saline. Furthermore, pressor responses to a more specific a-adrenoceptor stimulant, phenyleplhrine, were not significantly enhanced by propranolol. F J IMMs MRC Environmental Physiology Unit, London School of Hygiene and Tropical Medicine, London WG1

Professor J A Davis (27 December, p 757) advocating the use of phenobarbitone as a very effective symptomatic treatment of whooping cough in young infants. Unfortunately I have never been impressed by this treatment and I suspect that this could be due to my failure to use the "relatively high dosage" recommended. I wonder if Professor Davis could be asked to detail the doses he has in mind. G MOSES London W12

The following example illustrates our principal concern. The relationship of serum calcium and albumin was studied in six subjects, inducing changes in serum albumin by altering the extracellular fluid volume and obtaining 10 measurements of calcium and albumin for each subject. The overall correlation of calcium on albumin was highly significant (P01). The error of the method was such that the results could easily be obtained by chance if the regression coefficients for the six subjects were in fact identical. Dr Pain and his co-workers reported individual regression coefficients between 0-018 and 0-080 mmol/g in 25 hospital inpatients and stated that the regression lines "were not always parallel," implying that they differed significantly. Unfortunately they had excluded data from nine patients in whom they could not fit a "satisfactory" regression line. When testing for significant non-parallelism of the regression coefficients there is no justification for excluding these nine patients, who contributed (probably considerably) to the error of the experiment. From the example we have given above it is quite possible that the range of coefficients observed by Dr Pain and his colleagues merely reflects the error of their experiment and not true intersubject variation. While uncertainty exists on this point the frightening calcium "corrections" shown in their table I are not really valid. In the preliminary information given in table II of the original paper it is clear that the regression coefficients for two of the three individuals studied using a tourniquet method did not differ significantly from the average population regression coefficient. The mean coefficient for the third subject was 33% higher than the population coefficient, a difference which, even if significant, is much

less worrying than the 700% variation highlighted earlier in their paper. Before abandoning the current method for calcium correction it needs to be shown that individuals deviate significantly from the population regression, and also that the magnitude of any deviation is sufficient to justify the trouble of obtaining individual regression coefficients. LAWRENCE E RAMSAY Gardiner Institute, Western Infirmary,

Glasgow

* * *We showed this letter to Professor Davis, G D Searle and Co, whose reply is printed below.-ED, BM7. High Wycombe, Bucks

JOHN R SHELTON

R L B NEAME D A Powis

SIR,-The dose of phenobarbitone that we have been using in babies with whooping cough is 5-10 mg twice daily depending on 1 Imms, F J, Neame, R L B, and Powis, D A, weight, with two-thirds of the dose given in 47P. journal of Physiology, 1974, 241, the evening and one-third in the morning. Babies will in fact tolerate larger doses without this interfering with feeding or affecting SIR,-The report by Dr I Blum and others respiration. (13 December, p 623) serves only to em- University Department of Child Health,JOHN A DAvIs phasise that one should not increase the dose St. Mary's Hospital, Manchester of a hypotensive drug too rapidly. The pressor effect they describe is not seen only when using propranolol: I have accidently "Corrected" calcium concept induced it using each of guanethidine, acebutolol, and sotalol. None of these patients SIR,-Dr R W Pain and his colleagues (13 had a phaeochromocytomna. In a case of December, p 617) suggest that the regression suicidal overdose practolol substantially in- coefficients for serum calcium on albumin creased blood pressure.' differ so widely between individuals that use It is not clear whether the pressor re- of an average population regression coefficient sponse is due to increased a-tone as sug- to correct serum calcium is invalid. They may gested. I have controlled my paradoxically prove correct, but the information presented hypertensive patients with parenteral diaz- in their paper does not allow this conclusion oxide or hydrallazine. It may be that the to be drawn. Department of PhysioLogy, St Bartholomew's Hospital, Medical College, London ECI

SIR,-Dr R W Pain and his colleagues (13 December, p 617) "question the validity of 'correcting' an individual's measured serum calcium concentration by an average correction factor." They propose individual determination of binding coefficient by a four-point cuffing study as an alternative. They believe this to be necessary because there is a "wide individual variation in the number of millimoles of calcium bound per gram of albumin in the serum." This belief cannot be soundly based on the two-point cuffing' or postural2 experiments that they cite nor on their own determinations of regression coefficients from 3 to 17 (mean 8) serial determinations of calcium and albumin in individual patients, for there are no analyses to show that the variations found are any greater than those to be expected from the potentially gross sampling errors alone. Their case must rely on their triple determinations of binding coefficient in three subjects using the four-

point cuffing method. Average binding

Letter: Paradoxical rise in blood pressure during propranolol treatment.

BRITISH MEDICAL JOURNAL 218 but that the increase in the rate of remission in AML is small and not significant. A significant difference could be ex...
535KB Sizes 0 Downloads 0 Views