- Email: [email protected]
DEATH ASSOCIATED WITH HYPERVENTILATION
635
including nicotinamide, nicotine, guanethidine, metyrapone, and pinacidil.lO,l1 Is the metabolic for
many
substances,
transformation of such substances also deficient in individuals with N-oxidation for TMA? Propositus A and her affected sister failed to form nicotine N-oxide on oral challenge with this alkaloid, suggesting co-segregation for the two defects. M. AL-WAIZ R. AYESH S. C. MITCHELL Department of Pharmacology, R. IDLE J. Medical St Mary’s Hospital School, R. L. SMITH London W2 IPG
impaired
JM. The chemistry and metabolism of nitrogenous extractives in fish. Biochem Soc Symp 1951; no 6: 28-48. 2. Tver DF, Russell P. Nutrition and health encyclopedia. New York: Van Nostrand Reinhold, 1980: 100. 3. De La Huerga J, Popper H. Urinary excretion of choline metabolism following administration in normal and patients with hepatobiliary diseases. J Clin Invest 1. Shewan
1951, 30: 463-70. Trimethylaminumbasen
4. Lmtzel W.
im menschlichen Ham. Biochem Z 1934; 273: 243-61. 5. Humbert JR, Hammond KB, Hathaway WE, Marcoux JG, O’Brien D. Trimethylaminuria: the fish-odour syndrome. Lancet 1970; ii: 770-71. 6. Todd WA Psychosocial problems as the major complication of an adolescent with trimethylaminuria. J Pediatr 1979; 94: 936-37. 7. Spellacy E, Watts RWE, Goolamali SK. Trimethylaminuria. J Inher Metab Dis 1979; 2: 85-88. 8. Brewster M, Schedewie H, MacDonald JA. Rapid identification of trimethylaminuria as a cause of body odour. Clin Chem 1980; 26: 1011. 9. Simenhoff ML, Dunn SR, Asatoor A, Milne MD. Determination of trimethylamine N-oxide in urine. Proc Am Chem Soc 1977 (meeting 173): 57. 10. Oelschlager H, Al Shaik M. Metabolic N-oxidation of alicyclic amines. In: Gorrod JW, Damani LA, eds. Biological oxidation of nitrogen in organic molecules. Chichester: Ellis Horwood, 1985: 66-75. 11. Damani LA. Oxidation of tertiary heteroaromatic amines. In: Gorrod JW, Damani LA, eds. Biological oxidation of nitrogen m organic molecules. Chichester: Ellis Horwood, 1985: 205-18.
ADULTERATION OF HOMOEOPATHIC REMEDIES
SIR,-Following publication of my Lancet letter on the adulteration of homoeopathic medicine with prednisolone and betarnethasone1 several doctors from the UK and elsewhere wrote to me recording similar experiences with apparently impressive responses to alleged homoeopathic treatments. I report here one further case of adulteration with a synthetic steroid. A patient of Dr L. S. Pal (Selly Oak, Birmingham) with severe bronchial asthma obtained homoeopathic capsules from Pakistan (’Deecap’; Shaham Homeo Pharmacy Ltd, Karachi). The patient obtained great relief from these capsules but after hearing of the use of steroids in such preparations wrote to the manufacturers and was assured that their product did not contain cortisone. A similar letter was sent to me by Mr Abdul Rafay Siddiqi, head of the parent company, Usman Enterprise, in which he claimed great concern at the possible hazards of adulterated medication and asked for his product to be exonerated. The capsules were red/yellow and contained a grey heterogeneous mixture. A 50 ul ethanolic extract was applied to Whatman LK6D silica gel chromatography plates and compared with standard steroid preparations. The chromatogram was run in 97 % ethyl acetate/3 % ethanol and was visualised under ultraviolet light (245 nm). A single discrete band was obtained which comigrated with prednisolone standard (RF value 0-7). Identity was confirmed by continued co-migration despite the addition of chloroform in varying amounts to the mobile phase. Analysis of suspected adulterated medication by this modification of the previously described method is simple and requires very little laboratory equipment. The practice of adding synthetic steroids to a "homoeopathic" medication is widespread, and such analysis should be undertaken before ascribing a clinical response to any homeopathic element in remedies obtained from abroad.
DEATH ASSOCIATED WITH HYPERVENTILATION
SIR,-Hyperventilation and hypopnoea have been attributed both to organic and psychiatric disorders.1 However, Bass and Gardner2found that some patients can have profound hypocapnia without organic or psychiatric disorder. We have seen a patient with no organic cause for his apnoea who died after prolonged overbreathing. A 25-year-old man was admitted to a psychiatric hospital because of his severely disturbed behaviour. His behavioural problems dated back to early childhood and had been ascribed to a general disturbance of personality in association with mild mental handicap (IQ 62) and learned maladaptive patterns of behaviour rather than a formal psychiatric disorder. He had threatened and sometimes assaulted people (including hospital staff and patients) and he had appeared in court for threatening behaviour, insulting language, and breach of the peace. He was in good physical health. During his last hospital admission there had been several episodes of disturbed behaviour but on most occasions he was able to calm down quickly. He was also observed to have frequent attacks of overbreathing during which he would arch his back and flex his arms. Some attacks were triggered by a specific emotional event such as a disagreement with a patient but others seem to have been spontaneous. Although an electroencephalogram was normal he was prescribed carbamazepine as well as the chlorpromazine he was already taking. After one typical attack of overbreathing, probably precipitated by an incident with a patient, he was given 10 mg diazepam intravenously, and he soon calmed down. However, later that day a further attack of overbreathing did not respond to diazepam or to rebreathing from a paper bag. 20 min later he had a cardiac arrest and died. Necropsy identified possible causes of death as myocardial infarction, coronary artery occlusion, and metabolic alkalosis. Although pCOz levels were not measured the clinical presentation left little doubt that the patient’s overbreathing did result in hyperventilation. He had had frequent overbreathing attacks and may have been chronically hypocapnic; short-term increases in ventilation would result not only in further decreases in pCO2 but also in a rise in blood pH, with consequent production of his symptoms.3,4 The patient was physically fit and had a normal EEG and we think that his hyperventilation had a psychological basis associated with frustration and anxiety. Hyperventilation affects the cardiovascular system in several ways: for example, it increases the heart rate and cardiac output, inhibits the carotid baroreflex, and shifts the acid-base balance towards alkalosis.5 Electrocardiographic abnormalities are often 6 seen during hyperventilation in the absence of coronary disease in with and patients coronary disease hyperventilation may impair myocardial oxygen supply, lower the threshold for anginal attacks, and induce coronary artery spasm.’ We know of only one convincing report of death after voluntary hyperventilation, and that was in a 44-year-old man with arteriosclerotic heart disease.8 Our patient had no clinical evidence of heart disease and it seems most unlikely that the intravenous diazepam contributed to his death. The important point is that prolonged hyperventilation is potentially dangerous not only in patients with heart disease but also in young, physically healthy patients with psychiatric problems. Division of Psychiatry, United Medical and Dental Schools,
Guy’s Hospital, London SE1 9RT
N. BOURAS
National Hospital for Nervous Diseases, London WC1
L. D. KARTSOUNIS
Division of Psychiatry,
UMDS, Guy’s Hospital, London 1. Pfeffer
JM.
The
P. K. BRIDGES
aetiology of the hyperventilation syndrome. Psychother Psychosom
1978; 30: 47-55. Clinical Pharmacology Unit, University of Cambridge Clinical School, Addenbrooke’s Hospital, Cambridge CB2 2QQ
C, Gardner WN. Respiratory and psychiatric abnormalities m chronic symptomatic hyperventilation. Br Med J 1985; 290: 1387-90 3. Okel BB, Hurst JW. Prolonged hyperventilation in man: associated electrolyte changes 2 Bass
ALYN MORICE
and subjective symptoms. Arch Intern Med 1961; 108: 757-62. HA, Heyman A, Sieker HO Correlation of clinical and physiologic manifestations of sustained hyperventilation N Engl J Med 1963; 268: 1431-36.
4. Saltzman 1 Monce AH. Adulterated
"homoeopathic" cure for asthma. Lancet 1986; i: 862-63.
636 5. Grossman P.
Respiration, stress and cardiovascular function. Psychophysiology 1983;
20: 284-300.
Kemp GL, Ellstead MH. The significance of hyperventilation and orthostatic changes in electrocardiograms. Arch Intern Med 1968; 121: 518-23. 7. Neill WA, Pantley GA, Nakomchai V. Respiratory alkalemia during exercise reduces angina threshold. Chest 1981; 80: 149-53. 8. Bates JH, Adamson JS, Pierce JA, Rock L. Death after voluntary hyperventilation. N Engl J Med 1966; 274: 1371-72.
6.
PREVENTION OF CORONARY HEART DISEASE
SiR,—The British Cardiac Society undervalues the benefits of by placing it fifth and sixth in its recommendations on coronary heart disease (CHD) prevention (Feb 14, p 377). The treatment of raised lipids and hypertension is undoubtedly important in certain individuals but is unlikely to have a dramatic effect on the incidence of CHD. By contrast, most people with normal values as well as those with risk factors, stand to gain from regular exercise. In terms of coronary mortality, intervention studies show definite benefit after drug and dietary treatment only for subjects with hypercholesterolaemia at the top 5 % of the scale.’ Similarly, treatment of mild hypertension may cause more problems than it preventsz and probably only subjects at about the
exercise
top 10% of the blood pressure range can expect to benefit. Exercise is valuable in maintaining good health. In addition to a direct effect on improving cardiovascular function, exercise reduces risk factors such as hypertension, lipid levels, and obesity. It also improves the ratio of high to low density lipoproteins, and by reducing platelet aggregability and increasing fibrinolysis, decreases clot formation.3 As little as 20 min aerobic exercise three times per week will produce these benefits. Siscovick and colleagues’ have demonstrated a dose-related effect in preventing sudden cardiac death, in which the incidence amongst those exercising at 6 kcal/min (ie, leisurely jogging) for 140 min per week was. one-third that of sedentary subjects. Paffenbarger5 has shown in his study of Harvard alumni that lack of exercise was the strongest predictor of coronary risk. It is reasonable to identify and treat those at especially high risk of CHD but, in such a common disease in which the majority of sufferers are in normal rather than high risk groups, population policies are also necessary. Low cost strategies such as exercise are particularly important to an underfinanced National Health Service. Thatched Cottage, Church Lane,
ERIC WATTS
Nursling, Hampshire SO1 97B
1. Oliver MF. Hypercholesterolaemia and CHD: an answer. Br Med J 1984; 288: 423-24. 2. Medical Research Council Working Party. MRC trial of treatment of mild hypertension. Principal results. Br Med J 1985; 29: 97-104. 3. Coronary Prevention Group. Exercise Heart Health conference held at Royal Society
of Medicine on Feb 3-4, 1987. 4. Siscovick DS, et al. Primary cardiac
arrest
during vigorous
exercise.
N
Engl J Med
1984; 311: 874-77. 5.
Paffenbarger RS. A natural history of athleticism and cardiovascular health. JAMA 1984; 252: 491-95.
the common subset of patients but it does indicate that recall bias in the Swedish-Norwegian study was a minor problem. However, we are left with the question why the conclusions of the SwedishNorwegian study differ from those of some other studies of oral contraception and the risk of breast cancer. Meirik and colleagues2 noted that the breast cancer/OC controversy had lately focused on the use of oral contraceptives before a first full-term pregnancy and that their results in that respect were "in principle not at variance with the findings of Pike et al and McPherson et al". "On the other hand," they continued, "our data indicate that the total duration of OC use might be just as much associated with an increased risk of breast cancer as OC use before the first full-term pregnancy." This statement does not, however, accord with their data on total duration of OC use (tables 11 and ill) and duration before first full-term pregnancy (tables vi and vn). In tables vi and vn women with all of their OC exposure after their first full-term pregnancy have not been excluded, which Meirik et al seem to believe, but have been transferred to the reference category and there amalgamated with never-users. This can be verified by looking at the totals of the cases. Had Meirik’s conclusion been valid this contamination of the reference category would have lowered the relative risk, but it did not. On the contrary, for Swedish women the relative risk was increased from 2-1to 4-4 and for the total material the duration needed to attain a doubled risk is halved. Meirik and colleagues’ conclusion about the effect of total duration is unsound and the results should be interpreted as fully supporting the Pike, McPherson, and Olsson studies.15 The lack of effect in the Norwegian material could indicate either recall bias in the Swedish material or insufficient latency time in the Norwegian one since OC use among young women is more recent in Norway than in Sweden. The latency time issue has been much discussed, and if it was the reason for the lack of effect in Norway then it would probably also be the key to understanding the heterogeneity of results in studies published so far. To evaluate the validity of the results in studies on OC use and breast cancer, information on the prevalence of OC use at different ages and calendar years is needed. Unfortunately, such information is seldom presented. We have questioned the results of the CASH study on these grounds’ but the necessary information was not given in reply. As shown by McPherson et al,8a latency time would bias the relative risks downwards, which means that the high relative risks presented in some studies could, in fact, be even higher when latency time is accounted for. Oral contraceptives are commonly used by young women. In Britain some 83% of women born 1955-68 have used OC.’ The current state of knowledge of adverse effects is unacceptable. We must try harder to improve this knowledge and an open attitude to debate, criticism, and cooperation is essential in this endeavour. Department of Community Health Sciences, Lund University, Malmo General Hospital, S-214 01 Malmö, Sweden
JONAS RANSTAM
Department of Oncology, University Hospital, Lund
HAKAN OLSSON
ORAL CONTRACEPTIVES AND BREAST CANCER
SiR,—The letters by Professor Schlesselman and Dr Stadel and Dr Meirik and his colleagues (Feb 14, p 383) on the CASHl and the Swedish-Norwegian2 studies on oral contraceptives (OC) and breast cancer discuss the possibility that the effect from long-term use found by Meirik et al2 is due to recall bias only. In Sweden, little publicity was given to the 1983 studies by Pike, McPherson, and their colleagues.3,4 However, the 1985 study by Olsson et al5 was reported widely in Swedish newspapers when participants in the Swedish-Norwegian study2 were being interviewed. This publicity could have had a substantial impact on the validity of the results, and it is essential to know if it did or not. The effect of recall bias among the cases could be estimated by looking at the subset of cases in the Swedish-Norwegian study already in an extension of the Olsson study** and so interviewed both before and after the publicity in Swedish newspapers. Notwithstanding the different conclusions, the results of the Olsson and SwedishNorwegian studies are similar; this is hardly surprising in view of
1. The Cancer and Steroid Hormone
Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. Oral contraceptive use and the nsk of breast cancer. N Engl J Med 1986; 315: 405-11. 2. Meirik O, Lund E, Adami H-O, Bergstrom R, Christoffersen T, Bergsjo P. Oral contraceptives and breast cancer m young women: a joint national case-control study m Sweden and Norway. Lancet 1986; ii: 650-54. 3. Pike MC, Henderson BE, Krailo MD, et al. Breast cancer in young women and use of oral contraceptives: possible modifying effect of formulation and age at first use Lancet 1983; ii: 926-30. 4. McPherson K, Neil A, Vessey MP, Doll R. Oral contraceptives and breast cancer. Lancet 1983; ii: 1414-15. 5. Olsson H, Landin Olsson M, Moller TR, Ranstam J, Holm P Oral contraceptive use and breast cancer in young women in Sweden. Lancet 1985; i: 748-49. 6. Olsson H, Moller TR, Ranstam J, Landin Olsson M, Holm P. Oral contraceptive use and premenopausal breast cancer in southern Sweden. Abstracts of IVth International Congress in Senology (Paris, 1986). 57 (abstr III: 21) 7. Ranstam J, Olsson H, Möller TR. Oral contraceptive use and the risk of breast cancer N Engl J Med 1986; 316: 162. 8. McPherson K, Coope PA, Vessey MP. Early oral contraceptive use and breast cancer theoretical effects of latency. J Epidemiol Commun Health 1986, 40: 289-94 9. RCGP’s Manchester Research Unit. New oral contraception study: pilot trial report J Roy Coll Gen Practic 1986, 36: 545-46.
including nicotinamide, nicotine, guanethidine, metyrapone, and pinacidil.lO,l1 Is the metabolic for
many
substances,
transformation of such substances also deficient in individuals with N-oxidation for TMA? Propositus A and her affected sister failed to form nicotine N-oxide on oral challenge with this alkaloid, suggesting co-segregation for the two defects. M. AL-WAIZ R. AYESH S. C. MITCHELL Department of Pharmacology, R. IDLE J. Medical St Mary’s Hospital School, R. L. SMITH London W2 IPG
impaired
JM. The chemistry and metabolism of nitrogenous extractives in fish. Biochem Soc Symp 1951; no 6: 28-48. 2. Tver DF, Russell P. Nutrition and health encyclopedia. New York: Van Nostrand Reinhold, 1980: 100. 3. De La Huerga J, Popper H. Urinary excretion of choline metabolism following administration in normal and patients with hepatobiliary diseases. J Clin Invest 1. Shewan
1951, 30: 463-70. Trimethylaminumbasen
4. Lmtzel W.
im menschlichen Ham. Biochem Z 1934; 273: 243-61. 5. Humbert JR, Hammond KB, Hathaway WE, Marcoux JG, O’Brien D. Trimethylaminuria: the fish-odour syndrome. Lancet 1970; ii: 770-71. 6. Todd WA Psychosocial problems as the major complication of an adolescent with trimethylaminuria. J Pediatr 1979; 94: 936-37. 7. Spellacy E, Watts RWE, Goolamali SK. Trimethylaminuria. J Inher Metab Dis 1979; 2: 85-88. 8. Brewster M, Schedewie H, MacDonald JA. Rapid identification of trimethylaminuria as a cause of body odour. Clin Chem 1980; 26: 1011. 9. Simenhoff ML, Dunn SR, Asatoor A, Milne MD. Determination of trimethylamine N-oxide in urine. Proc Am Chem Soc 1977 (meeting 173): 57. 10. Oelschlager H, Al Shaik M. Metabolic N-oxidation of alicyclic amines. In: Gorrod JW, Damani LA, eds. Biological oxidation of nitrogen in organic molecules. Chichester: Ellis Horwood, 1985: 66-75. 11. Damani LA. Oxidation of tertiary heteroaromatic amines. In: Gorrod JW, Damani LA, eds. Biological oxidation of nitrogen m organic molecules. Chichester: Ellis Horwood, 1985: 205-18.
ADULTERATION OF HOMOEOPATHIC REMEDIES
SIR,-Following publication of my Lancet letter on the adulteration of homoeopathic medicine with prednisolone and betarnethasone1 several doctors from the UK and elsewhere wrote to me recording similar experiences with apparently impressive responses to alleged homoeopathic treatments. I report here one further case of adulteration with a synthetic steroid. A patient of Dr L. S. Pal (Selly Oak, Birmingham) with severe bronchial asthma obtained homoeopathic capsules from Pakistan (’Deecap’; Shaham Homeo Pharmacy Ltd, Karachi). The patient obtained great relief from these capsules but after hearing of the use of steroids in such preparations wrote to the manufacturers and was assured that their product did not contain cortisone. A similar letter was sent to me by Mr Abdul Rafay Siddiqi, head of the parent company, Usman Enterprise, in which he claimed great concern at the possible hazards of adulterated medication and asked for his product to be exonerated. The capsules were red/yellow and contained a grey heterogeneous mixture. A 50 ul ethanolic extract was applied to Whatman LK6D silica gel chromatography plates and compared with standard steroid preparations. The chromatogram was run in 97 % ethyl acetate/3 % ethanol and was visualised under ultraviolet light (245 nm). A single discrete band was obtained which comigrated with prednisolone standard (RF value 0-7). Identity was confirmed by continued co-migration despite the addition of chloroform in varying amounts to the mobile phase. Analysis of suspected adulterated medication by this modification of the previously described method is simple and requires very little laboratory equipment. The practice of adding synthetic steroids to a "homoeopathic" medication is widespread, and such analysis should be undertaken before ascribing a clinical response to any homeopathic element in remedies obtained from abroad.
DEATH ASSOCIATED WITH HYPERVENTILATION
SIR,-Hyperventilation and hypopnoea have been attributed both to organic and psychiatric disorders.1 However, Bass and Gardner2found that some patients can have profound hypocapnia without organic or psychiatric disorder. We have seen a patient with no organic cause for his apnoea who died after prolonged overbreathing. A 25-year-old man was admitted to a psychiatric hospital because of his severely disturbed behaviour. His behavioural problems dated back to early childhood and had been ascribed to a general disturbance of personality in association with mild mental handicap (IQ 62) and learned maladaptive patterns of behaviour rather than a formal psychiatric disorder. He had threatened and sometimes assaulted people (including hospital staff and patients) and he had appeared in court for threatening behaviour, insulting language, and breach of the peace. He was in good physical health. During his last hospital admission there had been several episodes of disturbed behaviour but on most occasions he was able to calm down quickly. He was also observed to have frequent attacks of overbreathing during which he would arch his back and flex his arms. Some attacks were triggered by a specific emotional event such as a disagreement with a patient but others seem to have been spontaneous. Although an electroencephalogram was normal he was prescribed carbamazepine as well as the chlorpromazine he was already taking. After one typical attack of overbreathing, probably precipitated by an incident with a patient, he was given 10 mg diazepam intravenously, and he soon calmed down. However, later that day a further attack of overbreathing did not respond to diazepam or to rebreathing from a paper bag. 20 min later he had a cardiac arrest and died. Necropsy identified possible causes of death as myocardial infarction, coronary artery occlusion, and metabolic alkalosis. Although pCOz levels were not measured the clinical presentation left little doubt that the patient’s overbreathing did result in hyperventilation. He had had frequent overbreathing attacks and may have been chronically hypocapnic; short-term increases in ventilation would result not only in further decreases in pCO2 but also in a rise in blood pH, with consequent production of his symptoms.3,4 The patient was physically fit and had a normal EEG and we think that his hyperventilation had a psychological basis associated with frustration and anxiety. Hyperventilation affects the cardiovascular system in several ways: for example, it increases the heart rate and cardiac output, inhibits the carotid baroreflex, and shifts the acid-base balance towards alkalosis.5 Electrocardiographic abnormalities are often 6 seen during hyperventilation in the absence of coronary disease in with and patients coronary disease hyperventilation may impair myocardial oxygen supply, lower the threshold for anginal attacks, and induce coronary artery spasm.’ We know of only one convincing report of death after voluntary hyperventilation, and that was in a 44-year-old man with arteriosclerotic heart disease.8 Our patient had no clinical evidence of heart disease and it seems most unlikely that the intravenous diazepam contributed to his death. The important point is that prolonged hyperventilation is potentially dangerous not only in patients with heart disease but also in young, physically healthy patients with psychiatric problems. Division of Psychiatry, United Medical and Dental Schools,
Guy’s Hospital, London SE1 9RT
N. BOURAS
National Hospital for Nervous Diseases, London WC1
L. D. KARTSOUNIS
Division of Psychiatry,
UMDS, Guy’s Hospital, London 1. Pfeffer
JM.
The
P. K. BRIDGES
aetiology of the hyperventilation syndrome. Psychother Psychosom
1978; 30: 47-55. Clinical Pharmacology Unit, University of Cambridge Clinical School, Addenbrooke’s Hospital, Cambridge CB2 2QQ
C, Gardner WN. Respiratory and psychiatric abnormalities m chronic symptomatic hyperventilation. Br Med J 1985; 290: 1387-90 3. Okel BB, Hurst JW. Prolonged hyperventilation in man: associated electrolyte changes 2 Bass
ALYN MORICE
and subjective symptoms. Arch Intern Med 1961; 108: 757-62. HA, Heyman A, Sieker HO Correlation of clinical and physiologic manifestations of sustained hyperventilation N Engl J Med 1963; 268: 1431-36.
4. Saltzman 1 Monce AH. Adulterated
"homoeopathic" cure for asthma. Lancet 1986; i: 862-63.
636 5. Grossman P.
Respiration, stress and cardiovascular function. Psychophysiology 1983;
20: 284-300.
Kemp GL, Ellstead MH. The significance of hyperventilation and orthostatic changes in electrocardiograms. Arch Intern Med 1968; 121: 518-23. 7. Neill WA, Pantley GA, Nakomchai V. Respiratory alkalemia during exercise reduces angina threshold. Chest 1981; 80: 149-53. 8. Bates JH, Adamson JS, Pierce JA, Rock L. Death after voluntary hyperventilation. N Engl J Med 1966; 274: 1371-72.
6.
PREVENTION OF CORONARY HEART DISEASE
SiR,—The British Cardiac Society undervalues the benefits of by placing it fifth and sixth in its recommendations on coronary heart disease (CHD) prevention (Feb 14, p 377). The treatment of raised lipids and hypertension is undoubtedly important in certain individuals but is unlikely to have a dramatic effect on the incidence of CHD. By contrast, most people with normal values as well as those with risk factors, stand to gain from regular exercise. In terms of coronary mortality, intervention studies show definite benefit after drug and dietary treatment only for subjects with hypercholesterolaemia at the top 5 % of the scale.’ Similarly, treatment of mild hypertension may cause more problems than it preventsz and probably only subjects at about the
exercise
top 10% of the blood pressure range can expect to benefit. Exercise is valuable in maintaining good health. In addition to a direct effect on improving cardiovascular function, exercise reduces risk factors such as hypertension, lipid levels, and obesity. It also improves the ratio of high to low density lipoproteins, and by reducing platelet aggregability and increasing fibrinolysis, decreases clot formation.3 As little as 20 min aerobic exercise three times per week will produce these benefits. Siscovick and colleagues’ have demonstrated a dose-related effect in preventing sudden cardiac death, in which the incidence amongst those exercising at 6 kcal/min (ie, leisurely jogging) for 140 min per week was. one-third that of sedentary subjects. Paffenbarger5 has shown in his study of Harvard alumni that lack of exercise was the strongest predictor of coronary risk. It is reasonable to identify and treat those at especially high risk of CHD but, in such a common disease in which the majority of sufferers are in normal rather than high risk groups, population policies are also necessary. Low cost strategies such as exercise are particularly important to an underfinanced National Health Service. Thatched Cottage, Church Lane,
ERIC WATTS
Nursling, Hampshire SO1 97B
1. Oliver MF. Hypercholesterolaemia and CHD: an answer. Br Med J 1984; 288: 423-24. 2. Medical Research Council Working Party. MRC trial of treatment of mild hypertension. Principal results. Br Med J 1985; 29: 97-104. 3. Coronary Prevention Group. Exercise Heart Health conference held at Royal Society
of Medicine on Feb 3-4, 1987. 4. Siscovick DS, et al. Primary cardiac
arrest
during vigorous
exercise.
N
Engl J Med
1984; 311: 874-77. 5.
Paffenbarger RS. A natural history of athleticism and cardiovascular health. JAMA 1984; 252: 491-95.
the common subset of patients but it does indicate that recall bias in the Swedish-Norwegian study was a minor problem. However, we are left with the question why the conclusions of the SwedishNorwegian study differ from those of some other studies of oral contraception and the risk of breast cancer. Meirik and colleagues2 noted that the breast cancer/OC controversy had lately focused on the use of oral contraceptives before a first full-term pregnancy and that their results in that respect were "in principle not at variance with the findings of Pike et al and McPherson et al". "On the other hand," they continued, "our data indicate that the total duration of OC use might be just as much associated with an increased risk of breast cancer as OC use before the first full-term pregnancy." This statement does not, however, accord with their data on total duration of OC use (tables 11 and ill) and duration before first full-term pregnancy (tables vi and vn). In tables vi and vn women with all of their OC exposure after their first full-term pregnancy have not been excluded, which Meirik et al seem to believe, but have been transferred to the reference category and there amalgamated with never-users. This can be verified by looking at the totals of the cases. Had Meirik’s conclusion been valid this contamination of the reference category would have lowered the relative risk, but it did not. On the contrary, for Swedish women the relative risk was increased from 2-1to 4-4 and for the total material the duration needed to attain a doubled risk is halved. Meirik and colleagues’ conclusion about the effect of total duration is unsound and the results should be interpreted as fully supporting the Pike, McPherson, and Olsson studies.15 The lack of effect in the Norwegian material could indicate either recall bias in the Swedish material or insufficient latency time in the Norwegian one since OC use among young women is more recent in Norway than in Sweden. The latency time issue has been much discussed, and if it was the reason for the lack of effect in Norway then it would probably also be the key to understanding the heterogeneity of results in studies published so far. To evaluate the validity of the results in studies on OC use and breast cancer, information on the prevalence of OC use at different ages and calendar years is needed. Unfortunately, such information is seldom presented. We have questioned the results of the CASH study on these grounds’ but the necessary information was not given in reply. As shown by McPherson et al,8a latency time would bias the relative risks downwards, which means that the high relative risks presented in some studies could, in fact, be even higher when latency time is accounted for. Oral contraceptives are commonly used by young women. In Britain some 83% of women born 1955-68 have used OC.’ The current state of knowledge of adverse effects is unacceptable. We must try harder to improve this knowledge and an open attitude to debate, criticism, and cooperation is essential in this endeavour. Department of Community Health Sciences, Lund University, Malmo General Hospital, S-214 01 Malmö, Sweden
JONAS RANSTAM
Department of Oncology, University Hospital, Lund
HAKAN OLSSON
ORAL CONTRACEPTIVES AND BREAST CANCER
SiR,—The letters by Professor Schlesselman and Dr Stadel and Dr Meirik and his colleagues (Feb 14, p 383) on the CASHl and the Swedish-Norwegian2 studies on oral contraceptives (OC) and breast cancer discuss the possibility that the effect from long-term use found by Meirik et al2 is due to recall bias only. In Sweden, little publicity was given to the 1983 studies by Pike, McPherson, and their colleagues.3,4 However, the 1985 study by Olsson et al5 was reported widely in Swedish newspapers when participants in the Swedish-Norwegian study2 were being interviewed. This publicity could have had a substantial impact on the validity of the results, and it is essential to know if it did or not. The effect of recall bias among the cases could be estimated by looking at the subset of cases in the Swedish-Norwegian study already in an extension of the Olsson study** and so interviewed both before and after the publicity in Swedish newspapers. Notwithstanding the different conclusions, the results of the Olsson and SwedishNorwegian studies are similar; this is hardly surprising in view of
1. The Cancer and Steroid Hormone
Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. Oral contraceptive use and the nsk of breast cancer. N Engl J Med 1986; 315: 405-11. 2. Meirik O, Lund E, Adami H-O, Bergstrom R, Christoffersen T, Bergsjo P. Oral contraceptives and breast cancer m young women: a joint national case-control study m Sweden and Norway. Lancet 1986; ii: 650-54. 3. Pike MC, Henderson BE, Krailo MD, et al. Breast cancer in young women and use of oral contraceptives: possible modifying effect of formulation and age at first use Lancet 1983; ii: 926-30. 4. McPherson K, Neil A, Vessey MP, Doll R. Oral contraceptives and breast cancer. Lancet 1983; ii: 1414-15. 5. Olsson H, Landin Olsson M, Moller TR, Ranstam J, Holm P Oral contraceptive use and breast cancer in young women in Sweden. Lancet 1985; i: 748-49. 6. Olsson H, Moller TR, Ranstam J, Landin Olsson M, Holm P. Oral contraceptive use and premenopausal breast cancer in southern Sweden. Abstracts of IVth International Congress in Senology (Paris, 1986). 57 (abstr III: 21) 7. Ranstam J, Olsson H, Möller TR. Oral contraceptive use and the risk of breast cancer N Engl J Med 1986; 316: 162. 8. McPherson K, Coope PA, Vessey MP. Early oral contraceptive use and breast cancer theoretical effects of latency. J Epidemiol Commun Health 1986, 40: 289-94 9. RCGP’s Manchester Research Unit. New oral contraception study: pilot trial report J Roy Coll Gen Practic 1986, 36: 545-46.