Chlamydia pneumoniae IgA titres and coronary heart disease

Transcription

1 European Heart Journal (2002) 23, doi: /euhj , available online at on Chlamydia pneumoniae IgA titres and coronary heart disease Prospective study and meta-analysis J. Danesh 1, P. Whincup 3, S. Lewington 2, M. Walker 4, L. Lennon 4, A. Thomson 4, Y.-K. Wong 5, X. Zhou 5 and M. Ward 5 1 Department of Public Health and Primary Care, University of Cambridge, Cambridge, U.K.; 2 Clinical Trial Service Unit and Epidemiological Studies Unit, University of Oxford, Oxford, U.K.; 3 Department of Public Health Sciences, St George s Hospital Medical School, London, U.K.; 4 Department of Population Sciences and Primary Care, Royal Free University College London Medical School, London, U.K.; 5 Departments of Cardiology and Molecular Microbiology, University of Southampton, Southampton, U.K. Aims To examine associations between Chlamydia pneumoniae IgA titres and incident coronary heart disease, and to compare them with associations previously reported between C. pneumoniae IgG titres and coronary heart disease. Methods and Results We measured serum concentrations of C. pneumoniae IgA antibodies in 502 coronary heart disease cases and in 1005 age- and town-matched controls nested in a community-based prospective study of 5661 British men (mean follow-up in controls, 16 years), and conducted a meta-analysis of published prospective studies to place our findings in context. Two hundred and twentyone (44%) of the cases were in the top third of C. pneumoniae IgA titres compared with 336 (33%) of the controls, yielding an odds ratio for coronary heart disease of 1 84 (95% confidence interval ) which was largely unchanged after adjustment. In aggregate, the present study and nine previously reported prospective studies of C. pneumoniae IgA titres involved 2283 cases, yielding a combined odds ratio for coronary heart disease of 1 25 ( ), with no significant heterogeneity among the ten studies (χ 2 9=7 8; P>0 1). This combined odds ratio is compatible with that previously reported for C. pneumoniae IgG titres and coronary heart disease (1 15, ). Conclusion Neither C. pneumoniae IgA titres nor IgG titres are strongly predictive of coronary heart disease in the general population. (Eur Heart J, 2002; 23: , doi: /euhj ) 2001 The European Society of Cardiology Key Words: Coronary heart disease, infection, Chlamydia pneumoniae, epidemiology. See page 343, doi: /euhj for the Editorial comment on this article Introduction Studies of coronary heart disease and serum markers of persistent infection with Chlamydia pneumoniae (also known as Chlamydophila pneumoniae) have reported apparently inconsistent results [1]. It has been proposed that such differences may be partly due to the use of different antibody subtypes as markers of the infection across studies [2]. In particular, it has been suggested Revision submitted 10 May 2001, accepted 16 May 2001, and published online 19 September Correspondence: J. Danesh, CTS Department of Public Health and Primary Care, Cambridge CB2 2SR, U.K X/02/ $35.00/0 that serum immunoglobulin A antibody titres to C. pneumoniae may be more predictive of coronary heart disease than are immunoglobulin G titres because IgA antibodies have different biological characteristics (such as a shorter half-life) and may be better indicators of persistent chlamydial infection in certain chronic diseases [2]. A previous meta-analysis of 15 prospective studies of C. pneumoniae IgG titres and coronary heart disease, involving over 3000 cases and over 9000 controls, yielded a combined odds ratio of 1 15 (95% confidence interval, ) [1]. Although a few small studies have reported stronger associations of C. pneumoniae IgA titres with myocardial infarction [3] or stroke [2], the evidence on IgA markers has not yet been 2001 The European Society of Cardiology

2 372 J. Danesh et al. Table 1 Baseline characteristics of men with coronary heart disease and of age- and town-matched male controls. Values are means SD or numbers (%) Characteristic Cases (n=502) Controls (n=1005) P value Questionnaire Age (years) matched Current smoker 266 (53%) 427 (43%) < Evidence of coronary disease* 176 (35%) 197 (20%) < Treated diabetic 12 (2%) 14 (1%) ns >2 drinks alcohol/day 109 (22%) 227 (23%) ns Occupation in social classes I II 111 (22%) 276 (27%) 0 03 Home owner** 272 (64%) 654 (69%) 0 03 Physical measurements Body mass index (kg. m 2 ) Systolic blood pressure (mmhg) < Diastolic blood pressure (mmhg) < FEV 1 (l) < Blood sample Total cholesterol (mmol. l 1 ) < HDL cholesterol (mmol. l 1 ) Triglyceride (mmol. l 1 ) < Chlamydia pneumoniae IgA titres (FC 10 6 ) < Top third 221 (44%) 336 (33%) Middle third 160 (32%) 335 (33%) Bottom third 121 (24%) 334 (33%) *Evidence of ischaemia on baseline electrocardiogram or reported history of angina or myocardial infarction. **Information on home ownership was available for only 428 cases and 943 controls. + Geometric mean approximate SD. ++ FC=fluorescent count; FEV 1 =forced vital capacity in 1 s. systematically assessed. We therefore measured C. pneumoniae IgA titres in stored baseline samples of 502 coronary heart disease cases and 1005 controls in a prospective epidemiological study (in which we had previously studied C. pneumoniae IgG titres [1] ), and conducted a meta-analysis of other long-term prospective studies of C. pneumoniae IgA titres and coronary heart disease to place our findings in context. Methods Methods used to establish and analyse the British Regional Heart Study have been described previously [4]. During nurses administered epidemiological questionnaires, made physical measurements, and recorded ECGs in 7735 males aged (response rate 78%) randomly selected from general practice registers in each of 24 British towns. Non-fasting venous blood samples were collected in 5661 men in 18 of the towns and stored at 20 C for subsequent analysis. Further questionnaires were posted after 5 years (98% response among survivors) and 12 years of follow-up (90% response) that enquired about car ownership and childhood social circumstances (father s social class and childhood household amenities), respectively. All men have been monitored since entry for vascular mortality (ascertained through National Health Service Central Registers on the basis of a death certificate with ICD-9 codes ) and for cardiovascular morbidity (based on reports from general practitioners, supplemented by evidence from general practice records meeting World Health Organisation criteria), with a follow-up loss of <1% to date [4]. As in our previous studies of infective agents and coronary heart disease [1,4], cases in the present study were men with fatal coronary heart disease or non-fatal myocardial infarction occurring between 1978 and 1996 and with a stored serum sample available for analysis. Five hundred and two men had available C. pneumoniae IgA measurements. One thousand and five controls frequency matched to cases on town of residence and age in 5-year bands were randomly selected from those men who had survived to the end of the study period free from incident myocardial infarction. Laboratory workers, unaware of the disease status of the participants blood samples, made C. pneumoniae IgA measurements using a validated assay involving whole organism antigen and time-resolved fluorimetry [1,5]. Case-control comparisons involved unmatched stratified logistic regression fitted by unconditional maximum likelihood (SAS Institute, North Carolina, U.S.A.), as previously described [1,6]. For associations between C. pneumoniae IgA titres and a variety of known and suspected risk factors, emphasis was mainly given to differences more extreme than 2 6 standard deviations (2P 0 01) to make some allowance for multiple comparisons. To avoid selective reporting, we pre-specified case-control analyses of C. pneumoniae

3 Chlamydia and coronary heart disease 373 Table 2 Comparisons of the levels of risk factors and other characteristics in controls by thirds of Chlamydia pneumoniae IgA titres. Values are means SD or numbers (%) Top (n=336) Middle (n=335) Bottom (n=334) t + t ++ Age (years) Current smoker 154 (46%) 139 (42%) 134 (40%) >2 alcohol drinks/day 82 (24%) 79 (24%) 66 (20%) Evidence of CHD at baseline 68 (20%) 73 (22%) 56 (17%) Physical measurements Body mass index (kg. m 2 ) Height (cm) Weight (kg) Systolic blood pressure (mmhg) Diastolic blood pressure (mmhg) FEV 1 (l) Blood sample Chlamydia pneumoniae IgG (FC 10 6 ) **** Helicobacter pylori seropositivity 254 (76%) 243 (73%) 229 (69%) C-reactive protein (mg. l 1 ) Serum amyloid A protein (mg. l 1 ) Albumin (g. l 1 ) White cell count ( 10 9.l 1 ) Total cholesterol (mmol. l 1 ) HDL cholesterol (mmol. l 1 ) Triglyceride (mmol. l 1 ) Homocysteine (mmol. l 1 ) Haematocrit (%) Socioeconomic factors Non-manual occupation 79 (24%) 101 (30%) 96 (29%) Home owner 201 (64%) 222 (71%) 231 (64%) Married 319 (95%) 316 (94%) 311 (93%) Car owner 234 (74%) 251 (80%) 243 (77%) Father with non-manual job 60 (24%) 54 (22%) 75 (29%) Family owned a car 37 (14%) 40 (15%) 38 (14%) Bathroom in house 120 (45%) 137 (52%) 134 (50%) Hot water tap in house 123 (46%) 143 (54%) 137 (51%) Bedroom shared 182 (68%) 174 (66%) 167 (62%) t tests derived from regression of C. pneumoniae IgA values on each characteristic separately adjusting for age and town only. ++ t tests derived from regression of C. pneumoniae IgA values adjusting for age, town, smoking, body mass index, and markers of socioeconomic status (including height). Adjustments for social class were omitted in the regressions involving markers of socioeconomic status. FC=fluorescent count; CHD=coronary heart disease. Table 3 Odds of coronary heart disease and 95% confidence intervals in men who were in the top third (fluorescent count > ) compared with those in the bottom third of Chlamydia pneumoniae IgA titres (< ) Age and town only Age, town, and smoking other risk factors* adult SES** lifetime SES*** All cases and all controls 1 84 ( ) 1 78 ( ) 1 85 ( ) 1 82 ( ) 1 76 ( ) Only participants without evidence of CHD at baseline 2 19 ( ) 2 11 ( ) 2 21 ( ) 2 19 ( ) 2 14 ( ) *risk factors=total cholesterol, HDL cholesterol, triglycerides, body mass index, blood pressure. **adult SES=occupation, housing tenure, marital status, car ownership. ***lifetime SES=adult SES plus father s social class, family car ownership, bathroom in house, hot water tap in house, bedroom sharing, height. CHD=coronary heart disease.

4 374 J. Danesh et al. Study Wald, 2000 Present study Strachan, 1999 Tavendale, 1999 Miettinen, 1996 Saikku, 1992 Glader, 2000 von Hertzen, 1999 Ossewarde, 1998 Roivanen, 2000 Total No. of cases Degree of adjustment ++ Odds ratio and confidence intervals 1.25 (95% CI 1 03 to 1 53) 99% or 95% limits Figure 1 Odds ratios compare seropositives and seronegatives at the baseline measurement. Black squares indicate the odds ratio in each study, with the square size proportional to the number of cases and the horizontal lines representing confidence intervals. The combined odds ratio and its 95% confidence interval are indicated by a black diamond. + =adjustment for age and sex only; + + =for these plus smoking; =forthese plus some other classical vascular risk factors; =forthese plus markers of socioeconomic status. IgA antibody titres by thirds of values in controls. We also pre-specified that odds ratios would be reported both with and without adjustments for various sets of possible confounders. We conducted a meta-analysis of prospective studies of coronary heart disease and C. pneumoniae IgA titres published before 2001 using methods that have been described previously [6]. Cases were compared only with controls within the same studies to avoid potential biases [1]. Results Present study There were highly significant differences between cases and controls with respect to various known vascular risk factors (Table 1). The correlation coefficient between C. pneumoniae IgA titres and C. pneumoniae IgG titres was 0 34 (P< ). C. pneumoniae IgA titres were not significantly associated with blood lipids, blood pressure, serum homocysteine, C-reactive protein, serum amyloid A protein, albumin, or haematocrit (Table 2). Two hundred and twenty-one (44%) of the 502 cases were in the top third of C. pneumoniae serum IgA titres compared with 336 (33%) of the 1005 controls (Table 1). This difference yielded an age- and town-adjusted odds ratio for coronary heart disease of 1 84 (95% confidence interval ) for men in the top third of baseline C. pneumoniae IgA titres compared with men in the bottom third. The odds ratio was 1 78 ( ) after additional adjustment for smoking, 1 85 ( ) after additional adjustment for serum lipids, blood pressure, and body mass index, and 1 76 ( ) after additional adjustment for indicators of socioeconomic status. The results were not materially changed when the analyses were restricted to the 325 cases and 806 controls with no evidence of coronary heart disease at baseline (Table 3). Meta-analysis Including the present study, we identified ten prospective studies of C. pneumoniae IgA titres and coronary heart disease up to 2001 [3,7 14] involving a total of 2283 cases of non-fatal myocardial infarction or coronary heart disease death and 7143 controls. The weighted mean age of participants at baseline was 54 years with a weighted mean follow-up of 11 years. All but one of the studies [3] reported adjustment for smoking and some other classical risk factors, but only four [7 9] reported adjustment for markers of adult socioeconomic status (and only two for markers of childhood social class [8] ). Six used microimmunofluorescence assays (two [8,11] of which reported 1:64 as a cut-off titre for seropositivity, and others reported cut-off titres of 1:40 [10], 1:32 [13], 1:16 [12], or the top quarter of titres in controls [3] ), and four used other methods (two used enzyme-linked immunoassays [9,14] and two, including the present study, used time-resolved fluorimetry [7] ). Despite these differences, there was no significant heterogeneity among the ten studies (χ 2 9=7 8; P>0 1), and a combined analysis yielded an odds ratio of 1 25 ( : Fig. 1), which is not significantly different from the combined odds ratio of 1 15 ( ) previously reported for C. pneumoniae IgG titres (χ 2 1=0 3; P>0 1). To explore any effects of adjustment, we grouped reports on C. pneumoniae IgA titres by degree of adjustment for social class. The combined odds ratio was 1 23 ( ) in the six studies (623 cases) without any reported adjustment for indicators of adult or childhood socioeconomic status [3,10 14] and 1 16 ( ) in the eight studies (1503 cases) without adjustment for indicators of childhood social class [3,7,9 14]. In the six studies (863 cases) that used microimmunofluorescence assays [3,8,10 13], the combined odds ratio was 1 16 ( ) There was no significant heterogeneity in any of these subsidiary analyses.

5 Chlamydia and coronary heart disease 375 Discussion Although some small-scale studies have suggested that C. pneumoniae IgA titres are much more strongly related to coronary heart disease risk than are C. pneumoniae IgG titres [2,3], this may have been due to chance, selection biases and/or selective emphasis on particular reports. To minimize such potential biases, we measured C. pneumoniae IgA titres in a prospective, communitybased study with more coronary heart disease cases than all but one [7] previous study. Moreover, to avoid placing undue emphasis on particular reports, we placed our findings in the context of a meta-analysis of other long-term prospective studies. Hence, although data from our study were somewhat more positive than results from most previous prospective reports, the available studies are statistically compatible with one another, yielding a combined odds ratio for coronary heart disease of 1 25 ( : Fig. 1) in people with raised C. pneumoniae IgA titres. The relevance of this combined odds ratio is, however, uncertain because it is only weakly positive and of marginal statistical significance. Existing studies may have somewhat exaggerated the observed coronary heart disease risk due to failure to control for residual biases (such as residual confounding by cigarette smoking and/or publication bias ). Conversely, existing studies may have underestimated coronary heart disease risk due to failure to correct for within-person fluctuations in serum antibody levels (since C. pneumoniae antibodies may disappear and recur [1] ). Consequently, further studies are needed to investigate whether any modest association exists, particularly at younger ages where associations may be stronger than at older ages [15]. Despite these uncertainties, the present report reliably refutes previous suggestions that C. pneumoniae IgA titres strongly predict coronary heart disease risk. Conclusion Neither C. pneumoniae IgA nor IgG titres are strongly predictive of coronary heart disease in the general population. Professor A. G. Shaper established the British Regional Heart Study. We thank J. R. Gallimore and Prof. M. B. Pepys for the C-reactive protein and serum amyloid A protein assays; H. Refsum and P. Ueland for the homocysteine assays; J. Atherton and Prof. C. Hawkey for H. pylori assays; and J. John for valuable help. Paul Appleby plotted the Figure. The British Regional Heart Study is a British Heart Foundation research group, and also receives support from the Department of Health. J. D. was supported by a Merton College fellowship, a Frohlich award, and the Raymond and Beverly Sackler Award in the Medical Sciences. S. L. was supported by a British Heart Foundation research fellowship. References [1] Danesh J, Whincup P, Walker M et al. Chlamydia pneumoniae IgG titres and coronary heart disease: prospective study and meta-analysis. BMJ 2000; 321: [2] Elkind MSV, Lin I-F, Grayston JT et al. Chlamydia pneumoniae and the risk of first ischemic stroke: the Northern Manhattan Stroke Study. Stroke 2000; 31: [3] Roivainen M, Viik-Kajander M, Palosuo T et al. Infections, inflammation and the risk of coronary heart disease. Circulation 2000; 101: [4] Whincup P., Danesh J, Walker M. et al. Prospective study of virulent strains of Helicobacter pylori and coronary heart disease. Circulation 2000; 101: [5] Wong Y-K, Sueur JM, Fall CHD et al. The species specificity of the microimmunofluorescence antibody test and comparisons with a time resolved fluoroscopic immunoassay for measuring antibodies against Chlamydia pneumoniae. J Clin Pathol 1999; 2: [6] Danesh J, Whincup P, Walker M et al. Low grade inflammation and coronary heart disease: prospective study and meta-analysis. BMJ 2000; 321: [7] Wald NJ, Law MR, Morris JK et al. Chlamydia pneumoniae infection and mortality from ischaemic heart disease: results from a large prospective study. BMJ 2000; 321: [8] Strachan DP, Carrington D, Mendall MA et al. Relation of Chlamydia pneumoniae serology to mortality and incidence of ischaemic heart disease over 13 years in the Caerphilly prospective heart disease study. BMJ 1999; 318: [9] Tavendale R, Parratt D, Brook RA et al. Antibodies to Chlamydia pneumoniae antibodies do not predict subsequent CHD in the Scottish Heart Health and MONICA studies. Eur Soc Cardiol 1999; P2221. [10] Miettinen H, Lehto S, Saikku P et al. Association of Chlamydia pneumoniae and acute coronary heart disease events in non-insulin dependent diabetic and non-diabetic subjects in Finland. Eur Heart J 1996; 17: [11] Saikku P, Leinonen M, Tenkanen L et al. Chronic Chlamydia pneumoniae infection as a risk factor for coronary heart disease in the Helsinki Heart Study. Ann Intern Med 1992; 116: [12] Glader CA, Boman J, Saikku P et al. The proatherogenic properties of lipoprotein(a) may be enhanced through the formation of circulating immune complexes containing Chlamydia pneumoniae-specific IgG antibodies. Eur Heart J 2000; 21: [13] Von Hertzen L, Isoaho R, Kivela SL et al. Relation of C pneumoniae antibodies to ischaemic heart disease. BMJ 1999; [14] Ossewaarde JM, Feskens EJ, De Vries A et al. Chlamydia pneumoniae is a risk factor for coronary heart disease in symptom-free elderly men, but Helicobacter pylori and cytomegalovirus are not. Epidemiol Infect 1998; 120: [15] Danesh J, Youngman L, Clark S et al. Helicobacter pylori infection and early onset myocardial infarction: case-control and sibling pairs study. BMJ 1999; 319: