222 Be aware of renal function when prescribing morphine Lancet 284-85. 8. Verbeeck RK Glucuronidation and disposition of drug glucuronides in patients with renal failure. A review Drug Metab Dis 1982; 10: 87-89. 9 Iwamoto K, Klassen CD First-pass effect of morphine in rats. Pharmacol Exp Ther 1977: 200: 236-44 10 Sawe J, Dahlstrom B, Paalzow L, Rane A. Morphine kinetics in cancer patients Clin Pharmacol Ther 1981; 30: 629-35.
7 McQuay H, Moore A 1984,
ii
11. Twycross RG The relief of pain in far advanced cancer Regional Anesth 1980, 5: 2-11. 12. Laidlaw J, Read AE, Sherlock S Morphine tolerance in hepatic cirrhosis Gastroenterology 1961; 40: 389-96. 13 Patwardhan RV, Johnson RF, Hoyumpa A, et al Normal metabolism of morphine in cirrhosis Gastroenterology 1981; 81: 1006-11.
IS THE TRH TEST NECESSARY?
SIR,-Conventional radioimmunoassays of thyroid-stimulating hormone (TSH) discriminate between primary hypothyroidism and euthyroidism but are not sensitive enough to distinguish hyperthyroid from euthyroid patients. Preliminary reports with an immunoradiometric assay (TSH-IRMA; ’Sucrosep’, BootsCelltech) suggest that it does discriminate between hyperthyroid and euthyroid patients and may make testing with thyrotropinreleasing hormone (TRH) unnecessary.l Most patients with socalled non-toxic goitre and a subnormal TSH-RIA response to TRH have basal serum TSH-IRMA levels below the detection limit,2but no information is available on the TSH-IRMA response to TRH in these patients. We have studied 15 patients with non-toxic goitre. Evaluated by a thyroid scintigraphy 10 goitres were multinodular, 1 was diffuse, and 4 were single adenomas. All patients had a normal freethyroxine index, free 3,5,3’-triiodothyronine index, and a TSHRIA response to intravenous TRH 200 g of 100 pU/ml or less. Basal TSH-IRMA values were below the detection limit (0’05 t/U/ml) in 10 patients whereas 5 had values from 0 - 05 to 0 - 18 pU/ml. 7 patients with undetectable basal serum TSH-IRMA did not respond to TRH, and may have been hyperthyroid. However, 3 patients with serum TSH-IRMA levels below 0-05 pU/ml had significant responses to TRH (0-09, 0-13, and 0-7MU/ml). Although the biological significance of such small TSH-IRMA responses to TRH remains to be established, they do suggest that these patients are not hyperthyroid and that the intravenous TRH test is still necessary in some patients with non-toxic goitre. Steno Memorial Hospital, DK-2820 Gentofte, Denmark
C. KIRKEGAARD
Frederiksberg Hospital, Copenhagen
C. BREGENGARD
DJ, Alexander WD. Is the TRH test usually unnecessary? Lancet 1984; ii: 1161. HA, Caldwell G, et al. A sensitive immunoradiometric assay for serum thyroid stimulating hormone: a replacement for the thyrotrophin releasing hormone test? Br Med J 1984, 289: 1334-36.
1 Kerr
2. Seth J, Kellett
have direct instantaneous digital read-out of cardiac output, and the new generation of relatively inexpensive Doppler monitors can measure vessel or valve diameter directly. A further advantage is that they may be used to measure cardiac output at other sites in the cardiovascular system such as at the pulmonary or mitral valves. This represents an important alternative if, for example, a patient has aortic stenosis, although the descending aorta may also offer a site for Doppler study since the high velocities found in the ascending aorta in aortic stenosis have probably returned to normal values once the blood has turned the aortic arch.1,2 Simplified monitoring of serial changes in cardiac output by the Doppler (or impedance) method has wide applications beyond cardiology, especially during and after surgery, in intensive care, and to monitor the cardiovascular effects of drugs.
to
Institute of Child
Health,
RICHARD WYSE
London WC1N 1EH 1. Tunstall Pedoe DS
Velocity distribution of blood flow in major arteries of animals and thesis, Oxford University, 1970 Wyse RKH, Robinson PJ, Deanfield JE, Tunstall Pedoe DS, Macartney FJ Use of continuous wave Doppler ultrasound velocimetry to assess the severity of man
2.
D Phil
of the
coarcation
aorta
by measurement of aortic flow velocities. Br Heart1984,
52: 278-83
SLEEP APNOEA AND HYPERTENSION
SIR,-In Professor Kales and colleagues’ study of sleep apnoea hypertension (Nov 3, p 1005) 48 of the 50 patients with hypertension were on medication. 9 were taking "centrally acting adrenergic blockers", 23 were on beta-blockers, and 6 patients were on both. 4 of the 6 patients with the most severe sleep apnoea events were taking minoxidil, but perhaps of more relevance is the fact that 5 of these 6 patients were taking either beta-adrenoceptor-blocker drugs or a centrally acting adrenergic blocker in combination with minoxidil. Both of these groups of drugs affect mental function;l Solomon et al2 demonstrated impaired verbal memory in patients treated with these drugs and Salem and McDevitt have shown that propranolol3 and atenolol4 reduce critical flicker frequency test scores in single oral dose studies in healthy volunteers, indicating a and
reduced level of arousal with these agents. Beta-blockers and centrally acting adrenergic blockers may have similar effects on sleep patterns, and until observer-blind studies with hypertensive patients on and off these drugs show otherwise we should be wary of suggesting a causal relation between sleep apnoea and hypertension. Department of Geriatric Medicine, St James’s Hospital, Dublin
JOHN P. Cox
of Clinical
Department Pharmacology, Royal College of Surgeons in Ireland,
KEVIN O’MALLEY
Dublin 2, Ireland
DOPPLER ULTRASOUND FOR MEASURING CARDIAC OUTPUT
SIR,-Dr Haites and colleagues (Nov 3, p 1025) bring to a wider audience the potential of Doppler ultrasound as a convenient noninvasive technique for monitoring in cardiac output. This technique has applications in many branches of medicine. I feel, however, that some technical points merit attention. be with the patient still and in a anatomical position (generally, supine with the head tilted to the left would be preferable). To derive cardiac output by Doppler ultrasound, a measure of vessel diameter is necessary. Haites et al assume aortic diameter to be a constant and derive serial measurements of stroke volume using the patient as his own control. This makes each study quicker and easier, it avoids the error inherent in an aortic diameter measurement and the need for more complex echocardiographic equipment. However, it may not be possible to assume constancy of aortic cross-sectional diameter during serial studies in some patients, especially when there are significant changes in blood volume or pressure. Haites and colleagues’ 1984 paper (ref 22), suggests that they are using a Doppler system that is very outdated: technological improvements in Doppler equipment have now greatly simplified this type of study yet added much more flexibility. It is now possible
Serial
measurements must
consistent
1 Editorial. Intellectual performance in 2 Solomon S, Hotchkiss E, Sarav SM, memory function
by
hypertensive patients Lancet 1984, i: 87 Bayer C, Ramsey P, Blum RS. Impairment of of antihypertensive medication. Arch Gen Psychiatry 1983, 40:
1109-12
SA, McDevitt DG. Central effects of single oral doses of propranolol in man Br J Clin Pharmacol 1984, 17: 31-36 Salem SA, McDevitt DG Central effects of beta adrenoreceptor antagonists in man. single oral doses of atenolol Clin Pharmacol Ther 1983; 33: 52-57.
3 Salem 4
THE HEART AND MENTAL STRESS, REAL AND IMAGINED
SIR,-The study by Deanfield et al,’ is an excellent portrayal of the effects of mental stress and myocardial ischaemia. Objective indices of myocardial ischaemia would include radioisotope
electrocardiographic displacements, increasing myocardial lactate production, wall motion abnormalities on ventriculography, and increase in left-ventricular end-diastohc filling pressures (LVEDP). We have seen a young man whose LVEDP rose sharply when he saw his own arteriogram. This increase was clearly the result of altered self-perception and imaginary stress. This 37-year-old man sustained a complicated anterior wall abnormalities,
infarction.
Risk factors
included
aggressive type A behaviour pattern.
cigarette smoking and an His blood pressure was 120/80
223
Hg and there was prominent S4 and S3 gallop. The electrocardiogram showed Q waves over the anterior precordium mm
and a nuclear cardiac scan demonstrated an ejection fraction of 35%. Coronary arteriography revealed 40% narrowing in the mid-portion of the right coronary vessel and a flushed total occlusion of the left anterior descending vessel at the origin of the first septal perforator. The circumflex anatomy was within normal limits. After the study was
completed, haemodynamic
assessments
were
once
again
obtained from the left ventricle with a multipurpose catheter. The patient was then asked to think of a previous stressful episode in his life and to perform mental arithmetic. LVEDP recordings were measured at this time. After about 3 min of each simulated mental task, the catheter was once again placed in the left ventricle and the patient was shown his own coronary ventriculogram and angiogram and the findings were discussed with him. The LVEDP was 12 at rest, 12 when a past imagined stressful situation was recalled, 15 during mental arithmetic, but 24 when the patient saw his own cardiac catheterisation data. This case suggests a powerful connection between emotion, mind, and body. The adrenergic response to stress probably resulted in a significant increase in LVEDP. During mental arithmetic, there was a small increase. However, when the patient viewed his own study he imagined the strong possibility of reconstructive surgery. This altered perception probably induced the fight/flight response with secondary increase in catecholamines and neurotransmitters, resulting in further deterioration of the LVEDP in an already impaired left ventricle. The case demonstrates.ahemodynamic cardiovascular response to stress. Increasing LVEDP may be another silent myocardial ischaemic factor as a result of psychological arousal and mental stress. Manchester Memorial Hospital, Manchester, Connecticut 06040, USA, and University School of Medicine, Mt Sinai
Department of Biomedical Statistics,
LEONARD A. FEITELL
Westphalian Wilhelms University Munster, D-4400 Munster, West Germany
and Institute of Human Genetics,
Lancet 1984,
ii:
et
al. Silent
myocardial
ischaemia due
to
1.
SiR,—The case of alleged ranitidine hepatotoxicity reported by Dr Lauritsen and colleagues (Dec 22/29, p 1471) could well have been a result of post-transfusion non-A, non-B viral hepatitis. The temporal relation to the blood transfusion, with an incubation period of six weeks, and the biochemical and histological data provided are all consistent with non-A, non-B hepatitis. Centrilobular and bridging necrosis, steatosis, bileduct proliferation, portal tract fibrosis, and infiltration with lymphocytes and plasma cells are well recognised histological features of this disease. 1-3 St Vincent’s Hospital, Dublin 4, Ireland
J. E. HEGARTY
1 Dienes
HP, Popper H, Arnold W, Lobeck H. Histological observations in human hepaptitis non-A, non-B Hepatology 1982, 2: 562-71. Gudat F, Eder G, Eder C, et al. Experimental non-A, non-B hepatitis in chimpanzees; light, electron and immune microscopical observations. Liver 1983, 3: 110-21. Rugge M. Vanstapel MJ, Ninfo V, et al. Comparative histology of acute hepatitis B and non-A, non-B
in
Leuven and Padova
Virchows Archiv Pathol Anat 1983; 401:
275-88.
SAMPLE SIZE NEEDED TO ASSESS RISK OF ABORTION AFTER CHORIONIC VILLUS SAMPLING agree with Dr Wilson and colleagues (Oct 20, p 920) that background rates of spontaneous abortion in ultrasonically normal pregnancies are an important requirement for evaluating the safety of chorionic villus sampling in the first trimester. For an unbiased assessment of the risk of spontaneous abortion with this new method of prenatal diagnosis, however, the rates of fetal losses should be compared with matched pregnancies without invasive 1 procedures in a prospective, randomised trial.’
SIR,-We
W. HOLZGREVE A. REISCH P. MINY F. K. BELLER
mental
1001-04.
RANITIDINE AND HEPATOTOXICITY
3
These calculations show that if chorionic villus biopsy increases the spontaneous abortion rate by 0’ 4%, which would be equivalent to the risk for second-trimester amniocentesis, about 69 000 pregnancies would be required in each group. The background rate of spontaneous abortion in the first trimester strongly influences the required numbers of patients-eg, a drop to about 2600 patients in the two groups if the difference in abortion rates is about 2%. Even though the numbers required to achieve statistical significance are large, a study with matched controls allows a more meaningful statement about the added risk of spontaneous abortion after chorionic villus biopsy than the mere comparison with fetal loss rates in ultrasonically normal pregnancies now available. Only a well-designed, statistically sound, multicentre (preferably international) study can answer the very important questions about the safety of chorionic villus sampling.
STEPHENT T. SINATRA
Hospital,
1 Deanfield JE. Shea M, Kensett M,
2
PI:
Women’s Clinic,
Hartford, Connecticut
stress
To be able to state with confidence that the fetal loss rate in a group of patients (P1) after,chorionic villus biopsy differs from that in a control group of ultrasonically normal pregnancies (P2) we have calculated the required sample size for the two populations, based on a probability of a type I error (a) of 1% and of a type 11 error (/3) of 10%. The most recent international survey2revealed a spontaneous abortion rate of about 4’ 4% after chorionic villus sampling, and this was the figure we used for the rate in P when calculating sample sizes by the Fleis formula, the arc-sine formula, and the formula of Casagrande, Pike, and Smith3for different assumed risk figures for
2. 3
Holzgreve W, Hogge WA, Golbus MS. Chorionic villi sampling (CVS) for prenatal diagnosis of genetic disorders. First results and future research. Europ J Obstet Gynecol Reprod Biol 1981; 17: 121-30. Jackson L. Chorionic Villi Sampling Newsletter (August, 1984) Fleis JL Statistical methods for rates and proportions. New York: Wiley, 1973.
PREVENTING POSTOPERATIVE PULMONARY COMPLICATIONS
SIR,-Your Nov 10 editoral prompted us to draw attention to our
study!
in which 14 cholecystectomy patients were randomly allocated to receive either subcutaneous salbutamol 0-5mg or placebo immediately before extubation and every 6 h for 48 h. Mean Pa02 was significantly higher in the salbutamol group 3 h after the operation (p =0 - 03), whereas 27 and 51 h after operation it was not significantly different from that in the placebo group. Mean
spirometric values showed insignificant differences. However, forced expiratory volume in 1 s in the salbutamol group 51 h after operation was 22% higher than the mean in the placebo group. According to Cordier et al2 continous positive airway pressure (CPAP) has beneficial effect on postoperative pulmonary complications, and thus CPAP with salbutamol might further reduce the risk of pulmonary complications. mean
Department of Anaesthesia, Herlev Hospital, University of Copenhagen, 2730 Herlev, Denmark
0. U. PETRING
Department of Anaesthesia, Gentofte Hospital
B. ADELHØJ
Adelhøj B, Petring OU, Larsen JC, Bigler DR, Andersen JB Effect of s.c. salbutamol on postoperative pulmonary function in patients undergoing elective cholecystectomy Acta Anaesth Scand (in press). 2. Cordier PH, Squifflet JP, Pirson Y, Carlier M, Alexandre GPJ. Postoperative continuous positive airway pressure helps to prevent pulmonary infection after human renal transplantation. Transplant Proc 1984; 16: 1337-39. 1.