(De)troubling transparency: artificial intelligence (AI) for clinical applications

Artificial intelligence (AI) and machine learning (ML) techniques occupy a prominent role in medical research in terms of the innovation and development of new technologies. However, while many perceive AI as a technology of promise and hope-one that is allowing for more early and accurate diagnosis-the acceptance of AI and ML technologies in hospitals remains low. A major reason for this is the lack of transparency associated with these technologies, in particular epistemic transparency, which results in AI disturbing or troubling established knowledge practices in clinical contexts. In this article, we describe the development process of one AI application for a clinical setting. We show how epistemic transparency is negotiated and co-produced in close collaboration between AI developers and clinicians and biomedical scientists, forming the context in which AI is accepted as an epistemic operator. Drawing on qualitative research with collaborative researchers developing an AI technology for the early diagnosis of a rare respiratory disease (pulmonary hypertension/PH), this paper examines how including clinicians and clinical scientists in the collaborative practices of AI developers de-troubles transparency. Our research shows how de-troubling transparency occurs in three dimensions of AI development relating to PH: querying of data sets, building software and training the model The close collaboration results in an AI application that is at once social and technological: it integrates and inscribes into the technology the knowledge processes of the different participants in its development. We suggest that it is a misnomer to call these applications 'artificial' intelligence, and that they would be better developed and implemented if they were reframed as forms of sociotechnical intelligence.

Effect of naratriptan on myocardial blood flow and coronary vasodilator reserve in migraineurs

BACKGROUND

Migraine drugs can produce adverse cardiac effects. The authors have demonstrated previously that ergotamine can lead to a significant reduction of hyperemic myocardial blood flow, but little is known about the effect of the newer serotonin analogues. Coronary artery constriction caused by serotonin or its analogues is mediated mainly by 5HT2 receptors. The selective 5HT1B/1D agonist naratriptan has no significant activity at 5HT2 receptors; however, like all 5HT1B/1D agonists developed for the acute treatment of migraine, naratriptan could potentially constrict coronary arteries by activation of 5HT1B receptors.

METHODS

The effects on myocardial blood flow of subcutaneous naratriptan 1.5 mg compared with placebo were assessed under resting and hyperemic conditions with PET using oxygen-15 labeled water during two separate visits. This study was a randomized, double-blind, placebo-controlled crossover trial in 34 migraine subjects with no evidence of ischemic heart disease, studied outside a migraine attack.

RESULTS

Naratriptan did not differ significantly from placebo in its effects on resting myocardial blood flow, but did evoke a small, significant fall in hyperemic myocardial blood flow (-13% versus placebo) and an increase in hyperemic coronary resistance (+19% versus placebo) without any signs or symptoms suggestive of myocardial ischemia. Naratriptan did not significantly affect the coronary vasodilator reserve (hyperemic/resting blood flow) compared with placebo.

CONCLUSIONS

These results show that at therapeutic doses, naratriptan exerts only a minor effect on myocardial blood flow, coronary vasodilator reserve, or coronary resistance among subjects with no evidence of ischemic heart disease. These results should not be extrapolated to patients with coronary artery disease, in whom all 5HT1 agonists for migraine are contraindicated.

Mortality from Parkinson's disease and other causes in men who were prisoners of war in the Far East

BACKGROUND

During World War II, more than 140000 Allied prisoners of war (POWs) were held captive by the Japanese in conditions of extreme privation. There have been concerns that the survivors are at increased risk of degenerative neurological disorders, especially Parkinson's disease. We assembled a cohort of British ex-POWs and analysed their mortality in a 46-year follow-up study.

METHODS

Using records held by the War Pensions Agency, we abstracted data on 11915 British former POWs. 11134 men were traced, and observed numbers of deaths between 1952 and 1997 were compared with those expected from national rates for the male population of England and Wales. Standardised mortality ratios (SMR) were calculated.

FINDINGS

Overall, mortality was lower than expected (7474 deaths vs 8796.2 expected; SMR 0.85 [95% CI 0.83-0.87]). Death rates from Parkinson's disease among the former POWs were slightly below the national average, though this difference was not statistically significant (35 deaths vs 43.2 expected; SMR 0.81 [0.56-1.13]). A similar pattern was seen for other degenerative neurological disorders (motorneuron disease 0.62 [0.31-1.11], multiple sclerosis 0.88 [0.42-1.61], and dementia 0.88 [0.68-1.11]). The former POWs had significantly lower than expected mortality from all major causes of death (ischaemic heart disease 0.81 [0.78-0.85], cerebrovascular disease 0.88 [0.81-0.95], all malignant neoplasms 0.92 [0.88-0.95], and respiratory disease 0.79 [0.74-0.85]). They also had below average rates of death from tuberculosis (0.44 [0.26-0.71]) and suicide (0.77 [0.57-1.02]), though the latter relation was not statistically significant. Mortality from diseases of the liver was increased (chronic liver disease and cirrhosis 1.68 [1.28-2.17], primary carcinoma of the liver 2.42 [1.75-3.26]).

INTERPRETATION

There is little evidence that men who were POWs in the Far East have higher rates of death than the male population generally. The only exception is diseases of the liver, which may be due to infection with hepatitis B or C virus during captivity. Death-certification data cannot provide a complete picture of physical and mental health, but the period of severe malnutrition, frequent infections, exhaustion, and intense psychological stress seems not to have increased susceptibility to neurodegenerative disease.

Gender differences in the healthy worker effect among synthetic vitreous fiber workers

The aim of this study was to determine whether the healthy worker effect and its component parts operate similarly for women and men. A cohort of workers from 14 synthetic vitreous fiber factories in seven countries, employed for at least 1 year between 1933 and 1977 and followed up to the early 1990s, included 375 deaths and 53,608 person-years among females and 2,568 deaths and 210,073 person-years among males. Standardized mortality ratios for all-cause and circulatory diseases were adjusted for country, age, calendar time, and gender. In addition, internal comparisons were adjusted for time since hire and employment status. The analyses addressed the following: 1) the healthy hire effect, 2) the time since hire effect, and 3) the healthy worker survivor effect. In this cohort, an overall healthy worker effect was not present in either gender. The healthy hire effect, based on standardized mortality ratios for years 1-4 since hire, was observed in males (standardized mortality ratio (SMR) = 0.8; 95% confidence interval (CI): 0.7, 1.0) but was less in females (SMR = 0.9; 95% CI: 0.5, 1.6). The relative risks increased slightly with time since hire in males but not in females. Higher mortality ratios were seen among those leaving employment than among those who remained actively employed; however, this effect was substantially greater for women (relative risk (RR) = 3.4; 95% CI: 1.8, 6.3) than men (RR = 1.8; 95% CI: 1.5, 2.1). The gender difference for active versus inactive status was stronger up to age 60 (men: RR = 1.7; 95% CI: 1.4, 2.0; women: RR = 3.6; 95% CI: 1.8, 7.1) than above that age. In conclusion, it appears that there is a stronger selection of healthy men than women into the workforce, while health-related selection out of the workforce is stronger for women than men.

Outcomes of undernutrition in patients in the community with cancer or cardiovascular disease

Public health concern has tended to focus on the dangers of obesity, but there is evidence that undernutrition may also pose a risk to physical and mental well-being, particularly in those who are already ill. Using the General Practice Research Database (see office for Population Censuses and Surveys, 1995), we followed up 10,128 men and women aged 18 years and over who had been diagnosed with cancer or cardiovascular disease to examine whether nutritional status, as indicated by BMI, affected rates of use of health care resources and mortality. In both diagnostic groups, patients with a BMI below 20 kg/m2 had higher rates of consultation with GP, higher rates of prescription and higher death rates during the follow-up period compared with those with a BMI of 20-< 25 kg/m2. In men and women with cardiovascular disease, poor nutritional status was associated with a sharply increased risk of hospital admission. Patients whose BMI was 30-< 40 kg/m2 also tended to have increased rates of GP consultation and prescription, and if they were under the age of 65 years, they had an increased risk of death. The results of the present study suggest that in men and women with cancer or cardiovascular disease, even minor degrees of undernutrition are associated with a marked increase in morbidity and mortality.

Catch-up growth in childhood and death from coronary heart disease: longitudinal study

OBJECTIVE

To examine whether catch-up growth during childhood modifies the increased risk of death from coronary heart disease that is associated with reduced intrauterine growth.

DESIGN

Follow up study of men whose body size at birth was recorded and who had an average of 10 measurements taken of their height and weight through childhood.

SETTING

Helsinki, Finland.

SUBJECTS

3641 men who were born in Helsinki University Central Hospital during 1924-33 and who went to school in Helsinki.

MAIN OUTCOME MEASURES

Hazard ratios for death from coronary heart disease.

RESULTS

Death from coronary heart disease was associated with low birth weight and, more strongly, with a low ponderal index at birth. Men who died from coronary heart disease had an above average body mass index at all ages from 7 to 15 years. In a simultaneous regression the hazard ratio for death from the disease increased by 14% (95% confidence interval 8% to 19%; P<0.0001) for each unit (kg/m3) decrease in ponderal index at birth and by 22% (10% to 36%; P=0.0001) for each unit (kg/m2) increase in body mass index at 11 years of age. Body mass index in childhood was strongly related to maternal body mass index, which in turn was related to coronary heart disease. The extent of crowding in the home during childhood, although related to body mass index in childhood, was not related to later coronary heart disease.

CONCLUSION

The highest death rates from coronary heart disease occurred in boys who were thin at birth but whose weight caught up so that they had an average or above average body mass from the age of 7 years. Death from coronary heart disease may be a consequence of poor prenatal nutrition followed by improved postnatal nutrition.

Effect of nutritional status on use of health care resources by patients with chronic disease living in the community

The prevalence of malnutrition in patients with chronic disease living in the community in the UK is around 8%. Whether such patients experience greater morbidity and mortality or make increased use of health care resources is unknown. The aim of this study was to investigate how the use of health care resources by patients with chronic disorders of the respiratory, gastrointestinal and neurological systems varied by nutritional status. We used longitudinal data, collected since 1987, which formed part of the General Practice Research Database in the UK. Subjects were 11 357 men and women aged 18 years or over. Main outcomes were consultation rates in general practice, prescription rates, hospital referral rates, hospital admission rates and mortality. Consultation and prescription rates were lowest amongst patients whose body mass index (BMI) was between 20 and 25. Rates were higher in patients whose BMI was below 20, or 25 and above. There was no statistically significant relation between rate of hospital outpatient referral and nutritional status, but both hospital admission rate and mortality were greatest in those people whose BMIs were below 20 and declined as BMIs increased. In patients with differential use of health care resources in both primary care and hospital practice, and with differences in mortality.

Effects of ergotamine on myocardial blood flow in migraineurs without evidence of atherosclerotic coronary artery disease

The effects of intravenous ergotamine (0.25 mg) on basal and hyperemic (dipyridamole) myocardial blood flow (MBF), measured with positron emission tomography and H2(15)O, were assessed in 15 migraineurs in a double-blind, randomized, placebo controlled, crossover study. Ergotamine produced a 27% reduction in hyperemic MBF (2.62 +/- 0.11 vs 3.72 +/- 1.05 ml x min(-1) x g(-1); p <0.05), a 31% reduction in the coronary vasodilator reserve (1.81 +/- 0.50 vs 2.71 +/- 1.15; p <0.01), and a 55% increase in minimal coronary resistance (42.2 +/- 15 vs 26.7 +/- 8 mm Hg x min x ml(-1) x g(-1); p <0.001), suggesting vasoconstriction of the coronary microcirculation.

Cancer mortality among man-made vitreous fiber production workers

We have updated the follow-up of cancer mortality for a cohort study of man-made vitreous fiber production workers from Denmark, Finland, Norway, Sweden, United Kingdom, Germany, and Italy, from 1982 to 1990. In the mortality analysis, 22,002 production workers contributed 489,551 person-years, during which there were 4,521 deaths. Workers with less than 1 year of employment had an increased mortality [standardized mortality ratio (SMR) = 1.45; 95% confidence interval (CI) = 1.37-1.53]. Workers with 1 year or more of employment, contributing 65% of person-years, had an SMR of 1.05 (95% CI = 1.02-1.09). The SMR for lung cancer was 1.34 (95% CI = 1.08-1.63, 97 deaths) among rock/slag wool workers and 1.27 (95% CI = 1.07-1.50, 140 deaths) among glass wool workers. In the latter group, no increase was present when local mortality rates were used. Among rock/slag wool workers, the risk of lung cancer increased with time-since-first-employment and duration of employment. The trend in lung cancer mortality according to technologic phase at first employment was less marked than in the previous follow-up. We obtained similar results from a Poisson regression analysis limited to rock/slag wool workers. Five deaths from pleural mesothelioma were reported, which may not represent an excess. There was no apparent excess for other categories of neoplasm. Tobacco smoking and other factors linked to social class, as well as exposures in other industries, appear unlikely to explain the whole increase in lung cancer mortality among rock/slag wool workers. Limited data on other agents do not indicate an important role of asbestos, slag, or bitumen. These results are not sufficient to conclude that the increased lung cancer risk is the result of exposure to rock/slag wool; however, insofar as respirable fibers were an important component of the ambient pollution of the working environment, they may have contributed to the increased risk.

Cancer mortality among local authority pest control officers in England and Wales

OBJECTIVE

To examine cancer mortality by tumour site among local authority pest control officers.

METHODS

Prospective mortality study, and follow up to the end of 1994, of 1485 male pest control officers aged between 17 and 69 and employed in 296 local authorities in England and Wales for at least six months between January 1980 and April 1984. Observed numbers of deaths were compared with those expected on the basis of the rates for relevant calendar year, cause, sex, and age specific groups for England and Wales.

RESULTS

200 deaths occurred during the follow up period of which 65 were certified as due to malignant neoplasms. No tumour type showed significantly more deaths than expected. Total all cause, lung cancer, and respiratory disease mortality were significantly lower than expected.

CONCLUSIONS

15 year follow up of a group of men handling a wide range of pesticides did not show any significant risk of cancer. This may be partially explained by the healthy worker effect and also the limited power of the study to detect significant increases in the less common tumours. Further long term follow up of this cohort will continue. Chemical control of pests that can cause human disease and can contaminate food and water has been, and will continue to be, a major public health measure. It is important to ensure that the health of those applying pesticides is not at excess risk. Negative results are important.

Vitamin C and risk of death from stroke and coronary heart disease in cohort of elderly people

OBJECTIVES

To determine whether vitamin C status, as measured by dietary intake and plasma ascorbic acid concentration, is related to mortality from stroke and coronary heart disease in people aged 65 and over.

DESIGN

A 20 year follow up study of a cohort of randomly selected elderly people living in the community who had taken part in the 1973-4 Department of Health and Social Security nutritional survey and for whom dietary and other data had been recorded.

SETTING

Eight areas in Britain (five in England, two in Scotland, and one in Wales).

SUBJECTS

730 men and women who had completed a seven day dietary record and who had no history or symptoms of stroke, cerebral arteriosclerosis, or coronary heart disease when examined by a geriatrician in 1973-4.

RESULTS

Mortality from stroke was highest in those with the lowest vitamin C status. Those in the highest third of the distribution of vitamin C intake had a relative risk of 0.5 (95% confidence interval 0.3 to 0.8) compared with those in the lowest third, after adjustment for age, sex, and established cardiovascular risk factors. The relation between vitamin C intake and stroke was independent of social class and other dietary variables. A similar gradient in risk was present for plasma ascorbic acid concentrations. No association was found between vitamin C status and risk of death from coronary heart disease.

CONCLUSION

In elderly people vitamin C concentration, whether measured by dietary intake or plasma concentration of ascorbic acid, is strongly related to subsequent risk of death from stroke but not from coronary heart disease.

Pooled reanalysis of cancer mortality among five cohorts of workers in wood-related industries

OBJECTIVES

To provide more information regarding the risk of cancer associated with wood dust, a pooled reanalysis of data from five cohort studies was performed.

METHODS

The combined cohort consisted of 28,704 persons from five studies: British furniture workers, members of the union representing furniture workers in the United States, two cohorts of plywood workers, and one of wood model makers, among whom 7665 deaths occurred. Pooled analyses were carried out for all of the cohorts combined, the two furniture worker cohorts combined, and the two plywood workers cohorts combined.

RESULTS

Significant excesses of nasal [observed 11, standardized mortality ratio (SMR) 3.1, 95% confidence interval (95% CI) 1.6-5.6] and nasopharyngeal (observed 9, SMR 2.4, 95% CI 1.1-4.5) cancer were observed. That for nasal cancer appeared to be associated with exposure to wood dust but was based solely on cases from the British furniture worker cohort, while that of nasopharyngeal cancer was observed for furniture and plywood workers and was associated with both high and low probability of wood dust exposure. Some support for an excess risk of multiple myeloma was also observed but was less clearly associated with wood dust exposure. No excesses of lung, larynx, stomach, or colon cancer were found to be associated with any surrogate indicators of wood dust exposure.

CONCLUSIONS

Workers exposed to wood dust may have an excess risk of nasopharyngeal cancer and multiple myeloma in addition to sinonasal cancer. The limitations of this study would tend to obscure relationships, rather than create false positive findings.

Weight gain in infancy and cancer of the ovary

The geographic association between ovarian cancer and tall stature suggests a link with rapid growth in early childhood. Among 5585 women born in Hertfordshire, UK, 41 who died from ovarian cancer had had a high rate of weight gain in infancy. Whereas their mean birthweight was the same as that of the other women, their mean weight at 1 year was higher (22.3 pounds [10.1 kg] vs 21.4 pounds [9.7 kg], p = 0.01). These observations are consistent with the hypothesis that ovarian cancer is linked to altered patterns of gonadotropin release established in utero when the fetal hypothalamus is imprinted.

Early growth and death from cardiovascular disease in women

OBJECTIVE

To determine whether the link suggested between growth in utero and during infancy and death from cardiovascular disease in men is also present in women.

DESIGN

Follow up study of women and men whose birth weight and weight at 1 year of age had been recorded.

SETTING

Hertfordshire, England.

SUBJECTS

5585 women and 10,141 men born during 1911-30.

MAIN OUTCOME MEASURES

Standardised mortality ratios for cardiovascular disease.

RESULTS

Among women and men death rates from cardiovascular disease fell progressively between the low and high birth weights groups (chi 2 = 4.3, p = 0.04 for women, chi 2 = 8.5, p < 0.005 for men). Cardiovascular deaths in men but not women were also strongly related to weight at 1 year, falling progressively between the low and high weight groups (chi 2 = 27.5, p < 0.0001). The highest cardiovascular death rates in women were among those with below average birth weight but above average weight at 1 year. In men the highest rates were among those with below average birth weight and below average weight at 1 year.

CONCLUSION

Relations between cardiovascular disease and birth weight are similar in men and women. In men cardiovascular disease is also related to weight gain in infancy.

A cohort study of workers exposed to formaldehyde in the British chemical industry: an update

A cohort study of workers exposed to formaldehyde in the British chemical industry in any one of six factories has been extended after the earlier published report in 1984. A further eight years of follow up to the end of 1989 have been included for the originally reported 7660 workers first employed before 1965, and a first follow up to the same date has been carried out for 6357 workers first employed since 1964. Extensive checking of the database has taken place including records at the factories, the MRC Environmental Epidemiology Unit, and the National Health Service Central Register. The updated findings include one death from nasal cancer compared with 1.7 expected in this number of men during the follow up period--which gives no support to the original hypothesis based on animal experimental data that formaldehyde may be a nasal carcinogen in humans. There have been no cases of nasopharyngeal cancer in the cohort compared with an estimated 1.3 expected--which gives no support to the findings in a similarly designed study in the United States of an excess of cancers of the nasopharynx associated with exposure to formaldehyde. There has been a slight excess of about 12% for lung cancer with 402 deaths compared with about 359 expected. This is similar to that found in the United States study, but higher than we reported earlier before the checking procedures and extended follow up. Further analysis gives no definitive indication of this excess of lung cancer being clearly related to formaldehyde exposure, and the increase is within that generally thought consistent with possible confounding effects of cigarette smoking (although no data are available on this point).

Relation of infant feeding to adult serum cholesterol concentration and death from ischaemic heart disease

OBJECTIVE

To examine whether method of infant feeding is associated with adult serum lipid concentrations and mortality from ischaemic heart disease.

DESIGN

Follow up study of men born during 1911-30.

SETTING

Hertfordshire, England.

SUBJECTS

5718 men, for 5471 of whom information on infant feeding had been recorded by health visitors and 1314 of whom had died. 485 of the men born during 1920-30 and still living in Hertfordshire who had blood lipid measurements.

MAIN OUTCOME MEASURES

Death from ischaemic heart disease; serum cholesterol and apolipoprotein concentrations.

RESULTS

474 men had died from ischaemic heart disease. Standardised mortality ratios were 97 (95% confidence interval 81 to 115) in men who had been breast fed and had not been weaned at 1 year, 79 (69 to 90) in breast fed men who had been weaned at 1 year, and 73 (59 to 89) in men who had been breast and bottle fed. Compared with men weaned before one year men not weaned had higher mean serum concentrations of total cholesterol (6.9 (not weaned) v 6.6 (weaned) mmol/l), low density lipoprotein cholesterol (5.0 v 4.6 mmol/l) and apolipoprotein B (1.14 v 1.08 g/l). Men who had been bottle fed also had a high standardised mortality ratio for ischaemic heart disease (95; 68 to 130) and high mean serum concentrations of total cholesterol (7.0 mmol/l), low density lipoprotein cholesterol (5.1 mmol/l), and apolipoprotein B (1.14 g/l). In all feeding groups serum apolipoprotein B concentrations were lower in men with higher birth weight and weight at 1 year.

CONCLUSIONS

Age of weaning and method of infant feeding may influence adult serum low density lipoprotein cholesterol concentrations and mortality from ischaemic heart disease. Adult serum apolipoprotein B concentrations are related to growth in fetal life and infancy.

Fetal and infant growth and impaired glucose tolerance at age 64

OBJECTIVE

To discover whether reduced fetal and infant growth is associated with non-insulin dependent diabetes and impaired glucose tolerance in adult life.

DESIGN

Follow up study of men born during 1920-30 whose birth weights and weights at 1 year were known.

SETTING

Hertfordshire, England.

SUBJECTS

468 men born in east Hertfordshire and still living there.

MAIN OUTCOME MEASURES

Fasting plasma glucose, insulin, proinsulin, and 32-33 split pro-insulin concentrations and plasma glucose and insulin concentrations 30 and 120 minutes after a 75 g glucose drink.

RESULTS

93 men had impaired glucose tolerance or hitherto undiagnosed diabetes. They had had a lower mean birth weight and a lower weight at 1 year. The proportion of men with impaired glucose tolerance fell progressively from 26% (6/23) among those who had weighted 18 lb (8.16 kg) or less at 1 year to 13% (3/24) among those who had weighed 27 lb (12.25 kg) or more. Corresponding figures for diabetes were 17% (4/23) and nil (0/24). Plasma glucose concentrations at 30 and 120 minutes fell with increasing birth weight and weight at 1 year. Plasma 32-33 split proinsulin concentration fell with increasing weight at 1 year. All these trends were significant and independent of current body mass. Blood pressure was inversely related to birth weight and strongly related to plasma glucose and 32-33 split proinsulin concentrations.

CONCLUSIONS

Reduced growth in early life is strongly linked with impaired glucose tolerance and non-insulin dependent diabetes. Reduced early growth is also related to a raised plasma concentration of 32-33 split proinsulin, which is interpreted as a sign of beta cell dysfunction. Reduced intrauterine growth is linked with high blood pressure, which may explain the association between hypertension and impaired glucose tolerance.

Relation of birth weight and childhood respiratory infection to adult lung function and death from chronic obstructive airways disease

OBJECTIVE

To examine whether birth weight, infant weight, and childhood respiratory infection are associated with adult lung function and death from chronic obstructive airways disease.

DESIGN

Follow up study of men born during 1911-30 whose birth weights, weights at 1 year, and childhood illnesses were recorded at the time by health visitors.

SETTING

Hertfordshire, England.

SUBJECTS

5718 men born in the county during 1911-30 and a subgroup of 825 men born in the county during 1920-30 and still living there.

MAIN OUTCOME MEASURES

Death from chronic obstructive airways disease, mean forced expiratory volume in one second (FEV1) and forced vital capacity (FVC), and respiratory symptoms.

RESULTS

55 men died of chronic obstructive airways disease. Death rates fell with increasing birth weight and weight at 1 year. Mean FEV1 at age 59 to 70 years, adjusted for height and age, rose by 0.06 litre (95% confidence interval 0.02 to 0.09) with each pound (450 g) increase in birth weight, independently of smoking habit and social class. Bronchitis or pneumonia in infancy was associated with a 0.17 litre (0.02 to 0.32) reduction in adult FEV1 and with an increased odds ratio of wheezing and persistent sputum production in adult life independently of birth weight, smoking habit, and social class. Whooping cough in infancy was associated with a 0.22 litre (0.02 to 0.42) reduction in adult FEV1.

CONCLUSIONS

Lower birth weight was associated with worse adult lung function. Intrauterine influences which retard fetal weight gain may irrecoverably constrain the growth of the airways. Bronchitis, pneumonia, or whooping cough in infancy further reduced adult lung function. They also retarded infant weight gain. Consistent with this, death from chronic obstructive airways disease in adult life was associated with lower birth weight and weight at 1 year. Promoting lung growth in fetuses and infants and reducing the incidence of lower respiratory tract infection in infancy may reduce the incidence of chronic obstructive airways disease in the next generation.

Weight in infancy and death from ischaemic heart disease

Environmental influences that impair growth and development in early life may be risk factors for ischaemic heart disease. To test this hypothesis, 5654 men born during 1911-30 were traced. They were born in six districts of Hertfordshire, England, and their weights in infancy were recorded. 92.4% were breast fed. Men with the lowest weights at birth and at one year had the highest death rates from ischaemic heart disease. The standardised mortality ratios fell from 111 in men who weighed 18 pounds (8.2 kg) or less at one year to 42 in those who weighed 27 pounds (12.3 kg) or more. Measures that promote prenatal and postnatal growth may reduce deaths from ischaemic heart disease. Promotion of postnatal growth may be especially important in boys who weigh below 7.5 pounds (3.4 kg) at birth.

Lung cancer in the meat industry

Routine statistics of occupational mortality and incidence of cancer have consistently shown high rates of lung cancer in butchers. Possible explanations include infection by carcinogenic papilloma viruses, exposure to polycyclic aromatic hydrocarbons and nitrites in the preservation of meat, or a confounding effect of tobacco. To explore these possibilities, we have examined the mortality of 1610 men employed at three British companies processing pork, beef, lamb, bacon, and other meat products. The overall death rate was less than in the national population (271 deaths observed, 310 expected) but there was an excess of deaths from cancer (87 observed, 80 expected), and in particular from lung cancer (42 observed, 32 expected). The risk of lung cancer was concentrated in subjects exposed to recently slaughtered meat, especially after an interval of 10 or more years. These findings increase suspicions of a risk of lung cancer in butchers, although further information is needed about smoking habits in the meat industry. If there is a hazard infection by a papilloma virus would seem the most likely cause.

Lung cancer among glass fibre production workers: a case-control study

A cohort study among 4734 employees at an English glass fibre plant previously reported no excess of lung cancer mortality either overall or when examined in broad occupational groups. To investigate occupation in more detail, and to test the hypothesis that processes producing or using finer (respirable) fibres may be related to a higher risk of lung cancer, a nested case-control study has now been carried out. Included are 73 cases of lung cancer and 506 matched controls, for whom jobs held and processes worked on have been blindly recorded in more detail than for the cohort study. Workers known to have been employed on processes containing respirable fibres had a relative risk of lung cancer of 1.2 (95% confidence interval 0.7-2.0) compared with other workers. There was no evidence of a relationship of lung cancer to fibre diameter, duration of exposure, or time since first exposure. The results by broad occupational group were similar to those of the cohort study, and although some of the many detailed occupational categories examined had significantly raised relative risks, these did not appear to be related to exposure to respirable glass fibre. Although the study has not indicated a differential risk of lung cancer among workers exposed to finer diameter glass fibres, the exposure levels were low and the number of cases small.

Mortality of workers exposed to styrene in the manufacture of glass-reinforced plastics

Epidemiologic studies have suggested an increased risk of leukemia and lymphoma among workers exposed to styrene. In a further exploration of this possible hazard, an analysis was conducted of the mortality among 7,949 men and women employed during 1947-1984 in eight British companies manufacturing glass-reinforced plastics. The subjects were identified from company files and traced to the end of 1984 through National Health Service and National Insurance records. The overall mortality in the cohort was less than in the national population (693 deaths observed, 830.1 expected) as was mortality from cancer (181 deaths observed, 223.7 expected). In particular, there was a deficit of deaths from lymphoid and hemopoietic cancer (6 observed, 14.9 expected). The small excess of lung cancer (89 deaths observed, 80.1 expected) was not statistically significant and can probably be attributed to chance. Among 3,494 hand laminators (the job with the highest exposure to styrene) there was one death from lymphoma and none from leukemia. The findings do not exclude the possibility that styrene is a human carcinogen, but give no support to the hypothesis that it causes leukemia and lymphoma.

Follow up study of workers manufacturing chrysotile asbestos cement products

A cohort study has been carried out of 2167 subjects employed between 1941 and 1983 at an asbestos cement factory in England. The production process incorporated the use of chrysotile asbestos fibre only, except for a small amount of amosite during four months in 1976. Measured airborne fibre concentrations available since 1970 from personal samplers showed mean levels below 1 fibre/ml, although higher levels had probably occurred previously in certain areas of the factory. No excess of lung cancer was observed in the mortality follow up by comparison with either national or local death rates, and analyses of subgroups of the workforce by job, exposure level, duration of employment, duration since entry, or calendar years of employment gave no real suggestion of an asbestos related excess for this cause of death. There was one death from pleural mesothelioma and one with asbestosis mentioned as an associated cause on the death certificate, but neither is thought to be linked to asbestos exposure at this factory. Other suggested asbestos related cancers, such as laryngeal and gastrointestinal, did not show raised risks. Although the durations of exposure were short in this study, the findings are consistent with two other studies of workers exposed to low concentrations of chrysotile fibre in the manufacture of asbestos cement products which reported no excess mortality.

Mortality of workers exposed to 2 methyl-4 chlorophenoxyacetic acid

The authors examined the mortality and cancer incidence of employees at a company which has manufactured, formulated, and sprayed 2 methyl-4 chlorophenoxyacetic acid (MCPA) and other phenoxy acid herbicides. Ninety-eight percent of the 5,784 men employed by the company during 1947-1975 was traced to the end of 1983. The overall mortality of the cohort was less than that of the national population, as was mortality from cancer. When allowance was made for rural residence, the deficit of cancer deaths became a slight excess, but not statistically significantly so. Among workers whose jobs entailed potential exposure to MCPA, there was one death from soft tissue sarcoma (0.6 expected). No further cases of soft tissue sarcoma were registered among living members of the cohort. Three potentially exposed workers died from nasal carcinoma, but this tumor has not previously been associated with phenoxy herbicides and the cluster of cases may have occurred by chance. The findings do not exclude the possibility that MCPA is a human carcinogen, but they suggest that any risk of soft tissue sarcoma is less than that indicated by earlier studies of 2,4,5-T (2,4,5-trichlorophenoxyacetic acid) and 2,4,5-trichlorophenol and is small in absolute terms.

Mortality of tanners

The mortality of 833 male tannery workers known to have been employed in the industry in 1939 and who were followed up to the end of 1982 was studied. A total of 573 men had been employed in making leather tanned by vegetable extracts for soles and heels, and 260 men had used chrome tanning to make leather for the upper parts of shoes. No significant excesses of deaths were found for any of the common sites of cancer in either group of workers. One death from nasal cancer (0.21 expected) was reported among the men who worked with sole and heel leather.

Mortality of hepatitis B positive blood donors in England and Wales

Of 2880 men and 1054 women who were found to be hepatitis B surface antigen positive by the Blood Transfusion Service in England and Wales between 1971 and 1981, more than 92% were traced to the end of 1983. 5 deaths from hepatocellular carcinoma had been reported in men, giving a relative risk compared with the male population of England and Wales as a whole of 42. There had been no deaths from liver cancer in women. In both men and women there was a greater than ten-fold increase in the risk of death from chronic liver disease. Secondary preventive action may be indicated.

Diet and renal stones: a case-control study

The diets of 51 men with renal stones were compared with those of 94 age-matched controls, using a measured record of all food and drink consumed during 4 days. The mean intakes of three nutrients suspected in the aetiology of renal stones, animal protein, refined carbohydrate (sugar) and dietary fibre, were similar for both cases and controls. However, the cases had a higher intake of ascorbic acid. A greater proportion of cases had low fluid intakes or took no exercise in their leisure time.

Mortality of English furniture makers

Men (5,108) who worked in the Buckinghamshire furniture industry before 1968 have been followed to the end of 1982; 1,638 (32.1%) had died. With the exception of nasal cancer, there was no significant increase in mortality, nor any trend towards increasing mortality with increasing dustiness of the work, for cancer of any site.

Extensions to a technique for relating mortality and environment--exemplified by nasal cancer and industry

The use of geographic data to interrelate mortality rates and environmental variables is an important approach to determining reasons for the variations in disease frequency. An alternative methodology to standard regression analysis was recently suggested, which is particularly useful when there are a large number of geographic areas and relatively few deaths. An extension of this technique is described, which includes a modified regression approach using areas aggregated according to their level of the environmental factor. An example is given relating nasal cancer mortality to the distribution of industry in England and Wales. The known hazards in the furniture and leather industries are indicated and other potential risk industries, including tailoring, are suggested.

Dietary fibre and acute appendicitis: a case-control study

In order to investigate the hypothesis that low dietary-fibre intake is a cause of acute appendicitis the diets of 30 children who had had appendicitis were compared with those of 55 age-sex matched controls, using a 7-day weighed food record. Although there were no statistically significant differences the average daily intake of cereal fibre was lower in the cases than the controls, giving limited support to reduced cereal fibre intake acting as a determinant of appendicitis. Low water intake may also be a causative influence. There is some evidence that infection and a familial predisposition may increase susceptibility to the disease.

Cancer in a factory using amosite asbestos

The paper describes the mortality experience of 5969 men employed in a factory where insulation board was manufactured using amosite asbestos from 1947 to 1979. 422 (7%) of the men were known to have died by the end of 1980. Among the 4820 men engaged in the manufacture of insulation board a doubling of the risk of lung cancer has occurred (57 deaths observed; 29 deaths expected). An excess is present both in men who entered the factory before and after 1960. Among the 2461 for whom smoking information is available a detectable excess risk is limited to current smokers exposed to higher levels of asbestos. Apart from five deaths from mesothelioma no statistically large or significant excesses of mortality from cancers of other sites have occurred, but further follow-up of the cohort is in progress. Nine deaths from asbestosis have been recorded. The results are discussed in the light of other studies of the effects of exposure to amosite asbestos.

Mapping small area cancer mortality: a residential coding story

During the compilation of an atlas of cancer mortality in England and Wales on a small area scale, one rural district was found to have raised death rates for most sites of cancer. On investigation this proved to be related to the opening in that district of a home for patients terminally ill with cancer. Deaths occurring in the home to patients from outside areas were not consistently transferred back to the deceased's area of residence. The effect of the transferability status of the home on the district's cancer rates is described. The Office of Population Censuses and Surveys is now reviewing the procedure for coding the area of residence of deaths in long stay institutions.

Chronic effects of beta 2-adrenergic agonists on body composition and protein synthesis in the rat

Chronic treatment of rats with the beta 2-adrenergic agonists clenbuterol and fenoterol over 16-19 d raised energy intake, expenditure, and body weight gain but did not affect fat or energy deposition, and body protein gain was increased by 50 and 18%, respectively. Both drugs increased the protein content and mitochondrial GDP-binding capacity of brown adipose tissue. Clenbuterol did not affect plasma insulin, growth hormone, or triiodothyronine levels, although insulin levels were reduced by fenoterol. Both drugs caused hypertrophy of skeletal muscle (gastrocnemius), and muscle protein synthesis in vivo (fractional rate) was elevated by 34 and 26% in clenbuterol and fenoterol-treated rats, respectively.

Mortality from thyrotoxicosis in England and Wales and its association with the previous prevalence of endemic goitre

The distribution of mortality from thyrotoxicosis among women in England and Wales during 1968-78 correlates with the previous prevalence of endemic goitre. Mortality from the disease rose to a peak in the decade 1931-40 and then declined. This peak affected all age groups and can be attributed to the high fatality from surgery before the introduction of preoperative iodine to prevent thyroid crisis. An apparent cohort effect, whereby cohorts born from 1871 to 1886 experienced the highest mortality, may be explained by generations with a high prevalence of endemic goitre becoming exposed to increasing dietary iodine intake in later life.

Diurnal variations in circulating hormone levels and brown adipose tissue activity in "cafeteria"-fed rats

Circulating hormone levels and interscapular brown adipose tissue (BAT) mitochondrial GDP-binding capacity were measured at 4-hr intervals over 24 hr in rats fed a stock diet or a palatable "cafeteria" diet. Plasma thyroxine, insulin and glucagon levels were unaffected by diet, but plasma triiodothyronine, BAT mass, mitochondrial protein content and GDP-binding were higher in cafeteria-fed rats. Blood glucose levels and BAT mitochondrial GDP-binding were higher at night in the cafeteria rats. The diurnal variations in these parameters correlated with changes in diet-induced thermogenesis previously observed in cafeteria-fed rats.

Nasal cancer in the Northamptonshire boot and shoe industry: is it declining?

This paper reports a survey of nasal cancer in Northamptonshire during the period 1950-79. An increased risk of various histological types of nasal tumour has been observed within the footwear manufacturing industry, which seems to be limited to the minority of men and women exposed to the dust of leather soles and heels. In Northamptonshire this exposure has usually occurred in the preparation, press and finishing rooms of factories making boots and shoes by the welted process. This type of leather is tanned by treatment with vegetable extracts, not chrome salts. Although the population of workers involved has diminished over the period of the study there has been no evidence of a decline in incidence of these tumours within it.

Mortality from mesothelioma of the pleura during 1968-78 in England and Wales

The geographical distribution of mortality from mesothelioma of the pleura during the years 1968-78 in England and Wales has been studied using extracts from the death records held by the Office of Population Censuses and Surveys. Using the national death rate as standard, Local Authority areas with raised mortality are identified. The patterns are somewhat different for each sex. In men the high-mortality areas are mainly the major ports where shipbuilding and repairing have been concentrated, whereas in women areas where gas masks are manufactured are predominant. In both sexes there are also high death rates on the eastern side of London. Nearly all the areas of high mortality are known to have had a major asbestos-using industry in the past. Over the 11-year period the annual number of deaths from pleural mesothelioma rose by approximately 75%. This marked increase was virtually confined to men, in whom the number of deaths had reached almost 200 per annum by 1978. The indications are that the effect of past high exposures, in particular to amphibole asbestos, have not yet reached a peak in terms of mortality. On the other hand imports and usage of amphiboles, particularly crocidolite, have decreased rapidly since the mid-1960s, and dust levels in the working environment have improved even more radically.

Variations in cancer mortality among local authority areas in England and Wales: relations with environmental factors and search for causes

Geographical variations in specific causes of mortality among the 1366 local authority areas of England and Wales as defined at 1971 were studied by examining extracts from death certificates held on computer tape. Five items of information on each death--year of death, age at death, sex, local authority area of residence, and the underlying cause of death, during the 11 years 1968-78--permitted a more detailed investigation than had been possible before. Analysis of some early results of the study--including maps of mortality for pleural mesothelioma, nasal cancer and bladder cancer--suggested that, despite the known limitations of death certification, systematic study of the mortality of small areas may give clues to aetiological factors in the environment. Analyses relating mortality to the distribution of environmental factors and examining disease profiles of each area may also provide clues. These will be followed up by other methods of study, such as case-control techniques.