1426
underlying the focal lymphocytic sialadenitis of chronic HCV liver disease thus seems different from that of primary Sjogren’s syndrome.
CORONARY EVENTS AND DEATHS FOR ELDERLY MEN AND WOMEN BY FIFTHS OF CHOLESTEROL
ANTONIO ACETI GLORIA TALIANI MAURIZIO SORICE MARIA A. AMENDOLEA
Institute for Tropical and Infectious Diseases, La Sapienza University, 00161 Rome, Italy
1. Fox RI, Robinson CA, Curd JG, Kozin F, Howell FV. Sjogren syndrome: criteria for classification. Arthritis Rheum 1986; 29: 577-85.
proposed
False-positive immunoassay for chlamydia in urine SIR,-Dr Wollenhaupt and colleagues report (May 2, p 1110) false-positive results in urine samples containing Staphylococcus aureus spp when tested for Chlamydia trachomatis with the IDEIA immunoassay (Novo Biolabs). Clinical samples containing high levels of S aureus or indeed any organism having protein A are well-known causes of false-positive EIA results and are acknowledged in the manufacturers’ instructions. In our extensive evaluations, clinical trials, and routine diagnostic procedures we have always confirmed reactive samplesl-4 in IDEIA by direct immunofluorescence with SYVA Microtrak monoclonal antibody. We note that Wollenhaupt et al recommend culturing all urine samples before testing for C trachomatis. We suggest an alternative approach in which all reactive EIA samples are confirmed by a method that uses monoclonal antibodies to a different C trachomatis epitope (MOMP) to that in EIAs-ie, direct immunofluorescence. This approach has distinct advantages because elementary bodies can be directly visualised, it does not require culture of all samples before testing, and it will still detect the rare sample containing C trachomatis and S aureus. Finally, in our wide experience of testing fresh first-catch urine samples it is exceptional to detect S aureus at levels that give rise to reactive EIAs in at-risk populations. Public Health
Laboratory,
Kingsdown, Bristol BS2 8EL, UK
E. O. CAUL I. D. PAUL
EO, Paul ID, Milne JD, Crowley T. Non-invasive sampling method for detecting Chlamydia trachomatis. Lancet 1988; ii: 1246-47. 2. Paul ID, Caul EO. Evaluation of three Chlamydia trachomatis immunoassays with an unbiased non-invasive clinical sample. J Clin Microbiol 1990; 28: 220-22. 3. Paul ID, Crowley T, Milne JD, Caul EO. A comparison of urine and urethral swabbing for the diagnosis of C trachomatis infections in males. Serodiag Immuno Infect Dis 1990; 4: 473-80. 4. Crowley T, Milne D, Arumainayagam JT, Paul ID, Caul EO. The laboratory diagnosis of male Chlamydia trachomatis infections: a time for a change? J Infect (in press). 1. Caul
this value is exceeded in over 40% of women. Indeed, the whole distribution is strikingly shifted to the right in elderly women, reflecting the sharp rise in cholesterol through the menopause. For coronary events and coronary deaths men had both a higher overall risk and a steeper risk gradient with cholesterol than did women. For elderly men, the gradient is clear and continuous with the top fifth (cholesterol 69 mmol/1) having 2-5 times the coronary risk of the bottom fifth (cholesterol < 5-0 mmol/1). For elderly women the risk gradient is less clear, and women with high cholesterol (78 mmol/1) have around 1-5 times the risk of low-cholesterol women. Furthermore, these high-cholesterol women have only about two-thirds the risk of the so-called average man with cholesterol of 5-7-6-0 mmol/l. The pronounced inverse association of malignant deaths with cholesterol concentration in both men and women is noteworthy. In support of previous opinion that this is a cholesterol-lowering effect of malignant diseasewe investigated the change in cholesterol concentrations with time in the 65 subjects who died from malignant disease for whom there were at least three previouis yearly cholesterol measurements. Within a year of deaths, mean cholesterol had fallen by 0-29 mmol/1 (SEM 0-09 mmol/1) compared with 2 years previously. This pronounced but causally inverted cholesterol/malignant-death association is sufficiently strong that it cancels out the cholesterol association with coronary death, leaving no observed association of total mortality with cholesterol in either men or women.
Cholesterol in
elderly women
SIR,—We agree with Dr Isles and colleagues (May 21, p 702) that there are few epidemiological data on cholesterol/disease-risk relations in women. There is also a need for further data for elderly subjects, since most studies relate to middle age. We present cholesterol findings for 1828 men and 2548 women aged 65-74 at entry to the Medical Research Council (MRC) trial of treatment of hypertension in older adults. We found no evidence that antihypertensive treatment affected cholesterol-risk associations, and only slight mean increases in serum cholesterol on diuretic and beta-blocker compared with placebo (+0-17 mmol/1 and +0-16 mmol/1, respectively, after 1 year of treatment). Hence we feel that this is a most valuable cohort for relating serum cholesterol at baseline to deaths and coronary events during a mean 5-8 years’
Although current and proposed lipid-intervention trials (eg, CRISP, the US lovastatin trial) will further elucidate the potential for coronary risk and mortality reduction in elderly patients, we feel that these observational data should have some influence on our In elderly women, there seems to be a much weaker case for cholesterol screening and intervention than in elderly men. In respect of absolute benefit, our efforts should be concentrated on men, with a view to seeing whether cholesterol-lowering in old age can match the observed gradient in coronary events and deaths, hopefully without any adverse effect on non-coronary deaths.
thinking.
Medical Statistics Unit, London School of Hygiene and Tropical Medicine, London EC1 E 7HT, UK
STUART J. POCOCK PAUL T. SEED
follow-up. The table shows the principal results. Classifying these elderly and women in roughly fifths of their cholesterol distributions emphasises the much higher cholesterol concentrations in women. Although 19% of the men have cholesterol of 6-9 mmol/1 or more, men
1. MRC Working party. Medical Research Council trial of treatment of hypertension in older adults: principal results. Br Med J 1992; 304: 405-12. 2. Law MR, Thompson SG. Low serum cholesterol and the risk of cancer: an analysis of the published prospective studies. Cancer Causes Control 1991; 2: 253-61.
underlying the focal lymphocytic sialadenitis of chronic HCV liver disease thus seems different from that of primary Sjogren’s syndrome.
CORONARY EVENTS AND DEATHS FOR ELDERLY MEN AND WOMEN BY FIFTHS OF CHOLESTEROL
ANTONIO ACETI GLORIA TALIANI MAURIZIO SORICE MARIA A. AMENDOLEA
Institute for Tropical and Infectious Diseases, La Sapienza University, 00161 Rome, Italy
1. Fox RI, Robinson CA, Curd JG, Kozin F, Howell FV. Sjogren syndrome: criteria for classification. Arthritis Rheum 1986; 29: 577-85.
proposed
False-positive immunoassay for chlamydia in urine SIR,-Dr Wollenhaupt and colleagues report (May 2, p 1110) false-positive results in urine samples containing Staphylococcus aureus spp when tested for Chlamydia trachomatis with the IDEIA immunoassay (Novo Biolabs). Clinical samples containing high levels of S aureus or indeed any organism having protein A are well-known causes of false-positive EIA results and are acknowledged in the manufacturers’ instructions. In our extensive evaluations, clinical trials, and routine diagnostic procedures we have always confirmed reactive samplesl-4 in IDEIA by direct immunofluorescence with SYVA Microtrak monoclonal antibody. We note that Wollenhaupt et al recommend culturing all urine samples before testing for C trachomatis. We suggest an alternative approach in which all reactive EIA samples are confirmed by a method that uses monoclonal antibodies to a different C trachomatis epitope (MOMP) to that in EIAs-ie, direct immunofluorescence. This approach has distinct advantages because elementary bodies can be directly visualised, it does not require culture of all samples before testing, and it will still detect the rare sample containing C trachomatis and S aureus. Finally, in our wide experience of testing fresh first-catch urine samples it is exceptional to detect S aureus at levels that give rise to reactive EIAs in at-risk populations. Public Health
Laboratory,
Kingsdown, Bristol BS2 8EL, UK
E. O. CAUL I. D. PAUL
EO, Paul ID, Milne JD, Crowley T. Non-invasive sampling method for detecting Chlamydia trachomatis. Lancet 1988; ii: 1246-47. 2. Paul ID, Caul EO. Evaluation of three Chlamydia trachomatis immunoassays with an unbiased non-invasive clinical sample. J Clin Microbiol 1990; 28: 220-22. 3. Paul ID, Crowley T, Milne JD, Caul EO. A comparison of urine and urethral swabbing for the diagnosis of C trachomatis infections in males. Serodiag Immuno Infect Dis 1990; 4: 473-80. 4. Crowley T, Milne D, Arumainayagam JT, Paul ID, Caul EO. The laboratory diagnosis of male Chlamydia trachomatis infections: a time for a change? J Infect (in press). 1. Caul
this value is exceeded in over 40% of women. Indeed, the whole distribution is strikingly shifted to the right in elderly women, reflecting the sharp rise in cholesterol through the menopause. For coronary events and coronary deaths men had both a higher overall risk and a steeper risk gradient with cholesterol than did women. For elderly men, the gradient is clear and continuous with the top fifth (cholesterol 69 mmol/1) having 2-5 times the coronary risk of the bottom fifth (cholesterol < 5-0 mmol/1). For elderly women the risk gradient is less clear, and women with high cholesterol (78 mmol/1) have around 1-5 times the risk of low-cholesterol women. Furthermore, these high-cholesterol women have only about two-thirds the risk of the so-called average man with cholesterol of 5-7-6-0 mmol/l. The pronounced inverse association of malignant deaths with cholesterol concentration in both men and women is noteworthy. In support of previous opinion that this is a cholesterol-lowering effect of malignant diseasewe investigated the change in cholesterol concentrations with time in the 65 subjects who died from malignant disease for whom there were at least three previouis yearly cholesterol measurements. Within a year of deaths, mean cholesterol had fallen by 0-29 mmol/1 (SEM 0-09 mmol/1) compared with 2 years previously. This pronounced but causally inverted cholesterol/malignant-death association is sufficiently strong that it cancels out the cholesterol association with coronary death, leaving no observed association of total mortality with cholesterol in either men or women.
Cholesterol in
elderly women
SIR,—We agree with Dr Isles and colleagues (May 21, p 702) that there are few epidemiological data on cholesterol/disease-risk relations in women. There is also a need for further data for elderly subjects, since most studies relate to middle age. We present cholesterol findings for 1828 men and 2548 women aged 65-74 at entry to the Medical Research Council (MRC) trial of treatment of hypertension in older adults. We found no evidence that antihypertensive treatment affected cholesterol-risk associations, and only slight mean increases in serum cholesterol on diuretic and beta-blocker compared with placebo (+0-17 mmol/1 and +0-16 mmol/1, respectively, after 1 year of treatment). Hence we feel that this is a most valuable cohort for relating serum cholesterol at baseline to deaths and coronary events during a mean 5-8 years’
Although current and proposed lipid-intervention trials (eg, CRISP, the US lovastatin trial) will further elucidate the potential for coronary risk and mortality reduction in elderly patients, we feel that these observational data should have some influence on our In elderly women, there seems to be a much weaker case for cholesterol screening and intervention than in elderly men. In respect of absolute benefit, our efforts should be concentrated on men, with a view to seeing whether cholesterol-lowering in old age can match the observed gradient in coronary events and deaths, hopefully without any adverse effect on non-coronary deaths.
thinking.
Medical Statistics Unit, London School of Hygiene and Tropical Medicine, London EC1 E 7HT, UK
STUART J. POCOCK PAUL T. SEED
follow-up. The table shows the principal results. Classifying these elderly and women in roughly fifths of their cholesterol distributions emphasises the much higher cholesterol concentrations in women. Although 19% of the men have cholesterol of 6-9 mmol/1 or more, men
1. MRC Working party. Medical Research Council trial of treatment of hypertension in older adults: principal results. Br Med J 1992; 304: 405-12. 2. Law MR, Thompson SG. Low serum cholesterol and the risk of cancer: an analysis of the published prospective studies. Cancer Causes Control 1991; 2: 253-61.
