Author + information
- Received March 9, 1999
- Revision received July 22, 1999
- Accepted September 21, 1999
- Published online January 1, 2000.
- Albrecht Hoffmeister, MD∗,* (, )
- Dietrich Rothenbacher, MD, MPH†,
- Peter Wanner, MS∗,
- Guenter Bode, PhD§,
- Kenneth Persson, MD, PhD‡,
- Hermann Brenner, MD, MPH†,
- Vinzenz Hombach, MD∗ and
- Wolfgang Koenig, MD, FACC∗
- ↵*Reprint requests and correspondence: Dr. A. Hoffmeister, Abteilung Innere Medizin II, Medizinische Universitätsklinik, Robert-Koch Str. 8, D-89081 Ulm, Germany.
We investigated the association between seropositivity to chlamydial lipopolysaccharide (cLPS) or Chlamydia pneumoniae (CP)and angiographically documented coronary artery disease (CAD), and we examined the relationship between serostatus and markers of systemic inflammation.
The potential contribution of CPto atherogenesis is still a matter of debate, and inflammation has been suggested to represent the link between infection and atherosclerotic disease.
Subjects age 40 to 68 years were recruited for this case-control study between October 1996 and November 1997: 312 patients with at least one coronary stenosis >50% and 479 age- and sex-matched blood donors without manifest CAD or history of angina. Antibodies against cLPS and CP, C-reactive protein (CRP), fibrinogen, plasma viscosity, leukocytes and neutrophils were determined. The study had a power of >80% to detect an odds ratio (OR) of 1.55 or above for the prevalence of immunoglobulin (IgG) antibodies against cLPS at a significance level of alpha = 0.05.
Prevalence of IgG antibodies against cLPS was not different between cases and controls (61% vs. 62%; p = 0.7). The adjusted OR for the presence of CAD given positive IgG serostatus against cLPS was 0.9 (95% CI; 0.6 to 1.3). Similarly, no difference in the prevalence of IgG antibodies against CPwas seen (88% vs. 87%; p = 0.6); the adjusted OR was 1.0 (95% CI; 0.6 to 1.6). Markers of inflammation did not show any statistically significant difference between cLPS seropositives and seronegatives.
Our results indicate no strong association between CPand CAD, and increased systemic inflammation in patients with CAD does not seem to be due to seropositivity to cLPS or CP.
Evidence of a potential role of infectious agents in the pathogenesis of atherosclerosis has been reported in several studies (see references [1–3]for recent reviews of the literature). In particular, an infection with Chlamydia pneumoniae (CP)has been claimed to be associated with atherosclerotic cardiovascular disease in seroepidemiological (4–6)and angiographical studies (7–10). Furthermore, CPhas been detected in atheromatous lesions of coronary arteries (11–14), yet the importance of these findings is still unclear. The organism may be actively involved in atherogenesis as a primary inducing agent or may exacerbate an existing inflammatory process. It cannot be ruled out, however, that CPis not involved in the process, but rather represents an innocent bystander passively brought to the site by monocytes.
Most of the previous seroepidemiological studies have reported a moderately strong association between CPand coronary artery disease (CAD). However, because of small sample size, results were statistically not significant in several of them, and often adequate control for potential confounders had not been carried out. Because the preferential publication of positive studies may have led to a publication bias, larger well-controlled studies are necessary to support or reject the hypothesis of an association between CPand CAD.
We therefore conducted a large case-control study to assess the association of angiographically documented CAD with seropositivity to chlamydial lipopolysaccharide (cLPS) and CP. Furthermore, because it has been suggested that systemic inflammation represents a link between infection and pathogenesis of atherosclerosis (1,2,15), we measured a variety of acute-phase reactants and related them to serostatus.
This case-control study was done in a university hospital setting in southern Germany. Subjects age 40 to 68 years and of German nationality were recruited between October 1996 and November 1997. Cases and controls were matched for age and sex; a sampling ratio of about 1:1.5 was intended. All subjects underwent blood sampling and a standardized interview. The study was approved by the ethics committee of the University of Ulm. Informed consent was obtained from all participants.
Cases were recruited in the University Hospital of Ulm. All subjects had a CAD with at least one coronary stenosis >50% at catheterization within the previous six months before entry into the study. Patients with acute coronary syndromes during the previous four weeks were excluded. The diagnosis “CAD” had been established no longer than two years earlier for the first time. Voluntary blood donors from the associated Red Cross blood bank served as controls. Potential controls with manifest heart disease or a history of angina pectoris according to the Rose angina questionnaire were excluded. Further exclusion criteria for cases and controls were acute infection, acute state of a chronic infectious or inflammatory disease, anticoagulant therapy (i.e., full-dose heparin or coumarins) during the preceding four weeks, severe liver or renal disease, neoplasm and hematological disorders.
Blood was taken under standardized conditions, and a complete blood cell count was done. The remaining blood was centrifuged at 3,000gfor 10 min within 30 min and aliquoted immediately. Plasma and serum specimens were frozen and stored at −70°C until analysis. All laboratory determinations were done in a blinded fashion. Antibodies (IgM, IgG, IgA) against Chlamydiaspecies (chlamydial lipopolysaccharide) were measured by a recombinant enzyme linked immunoassay (ELISA) (16)(Medac, Wedel, Germany). This ELISA used a cLPS antigen produced by recombinant DNA technology. This antigen contains a common epitope shared by all Chlamydiaspecies (17). An IgG titre ≥100 was defined as seropositive (≥50 for IgM and IgA). Determinations of specific antibodies against CPwere performed by microimmunofluorescence method as described elsewhere (18). An IgG titer ≥64 was considered to be seropositive (≥16 for IgA). C-reactive protein measurement was done by an immunoradiometric assay (range 0.05 to 10 mg/liter) calibrated with the WHO (World Health Organization) reference standard 85/506. Fibrinogen was measured by immunonephelometry (Behring, Marburg, Germany) and by the Clauss method. Measurement of plasma viscosity was done by a Coulter Harkness capillary viscometer (19). Leukocytes and neutrophils were counted automatically in a Coulter STKS chamber (Coulter, Krefeld, Germany). The intra-assay coefficient of variation was 4% for CRP, 3.7% for fibrinogen and 0.7% for plasma viscosity.
A standardized questionnaire was applied to all cases and controls by the same trained team of interviewers. It included a history of cardiovascular risk factors, history of heart disease and particular symptoms of angina pectoris according to the Rose questionnaire, current medication, previous antibiotic therapy, alcohol intake, physical activity and further socioeconomic variables.
Cases and controls were compared according to baseline characteristics in a descriptive way. The association of seropositivity to cLPS and CPwith presence of CAD was analyzed using a chi-square statistic. To assess differences in the distribution among various categories of each variable in seropositives, a chi-square statistic was also performed. We used the Wilcoxon rank-sum test to compare markers of inflammation between seropositive and seronegative subjects. Multivariable logistic regression was performed to assess the independent association of a seropositive status against cLPS and CPwith CAD; we controlled for age, gender, body mass index (BMI), pack years of smoking, history of hypertension, history of diabetes, years of school education, and alcohol consumption. Data were analyzed using SAS statistical software package (version 6.12).
Characteristics of the study population
In this study, 791 subjects (312 cases and 479 controls) were enrolled. Response rate among cases was 78% and among controls 84%. Table 1shows demographic characteristics and classical risk factors of the study population. Mean BMI, pack years of smoking, and prevalence of subjects with history of hypertension, dyslipidemia and diabetes as well as low school education were higher in cases than in controls (all p < 0.01). No significant differences were found for alcohol consumption, family status and history of chronic bronchitis.
Nearly two-thirds of the cases (62%) had a history of previous myocardial infarction within the previous two years. Based on the results of coronary angiography, 48% had a one-vessel disease, 34% a two-vessel disease and 18% a three-vessel disease.
IgG antibodies against cLPS were present in 62% of all subjects, and the prevalence of IgA antibodies against cLPS was 36%. The IgG antibodies against CPwere found in 87% of all participants; the overall prevalence of IgA antibodies against CPwas 36%. Table 2shows that the prevalence of IgG (IgA) antibodies against cLPS or CPwas not significantly different between cases and controls, independent of the serological cutoff point chosen (Fig. 1). The prevalence of IgG antibodies against cLPS in cases with previous myocardial infarction was 63%, and in cases without infarction 56% (p = 0.2). The prevalence of IgG antibodies against cLPS in cases with one-vessel disease was 64%, with two-vessel disease 56% and with three-vessel disease 61% (p = 0.4).
Seroprevalence, classical risk factors and socioeconomic variables
Most of the classical risk factors of CAD and of the socioeconomic variables were not related to the serostatus against cLPS (Table 3). A positive serostatus was less frequent in women than in men (p = 0.004). No significant positive association was detectable between age and serostatus in the age range 40 to 68 years (p = 0.2). Patients with history of diabetes revealed significantly fewer antibodies against cLPS compared to patients without such a history (p = 0.04).
Seropositivity and risk of cad
Table 4shows the association of cLPS and CPwith CAD after adjustment for age, gender and potential other confounders by multiple logistic regression. There was no increased risk of CAD associated with the presence of IgG or IgA antibodies against cLPS or CPin adjusted analyses. Furthermore, taking high levels of antibody titres into consideration did not change the results.
Inflammation and serostatus
In Table 5, mean values of different markers of systemic inflammation are compared with serostatus against cLPS. None of the variables showed a statistically significant difference between seropositives and seronegatives, neither in cases nor in controls. By contrast, a consistently increased inflammatory activity was seen in patients with CAD compared to controls.
The same results were found comparing markers of systemic inflammation by serostatus to CP(seronegative, IgG 64–511, and IgG ≥ 512).
In this study, the prevalence of IgG or IgA antibodies against cLPS or CPdid not differ between angiographically defined cases with CAD and carefully selected age- and gender-matched controls. This was true even when high antibody titres were considered as cutoff. Furthermore, the history of a previous myocardial infarction and severity of CAD was not associated with an increased seroprevalence. The results of bivariate analysis were also confirmed after controlling for various covariables by means of logistic regression, which revealed no strong and independent association between seropositivity to cLPS or CPand stable CAD. Thus, these findings do not confirm the results from several previous studies (4–10).
In 1988, Saikku et al. (4)reported for the first time an association of CPwith chronic CAD and acute myocardial infarction in a small study from Finland. Later, Saikku et al. (5)confirmed their findings using a nested case-control design in which subjects were identified from the Helsinki Heart Study; serum samples were obtained three to six months before a cardiac end point (n = 103) had occurred, and they were compared with matched controls. After adjustment for age, hypertension, and smoking, the odds ratio (OR) for developing a cardiac end point was 2.3 (95% CI, 0.9 to 6.2) for an elevated IgA titre (≥64) against CP, but only 1.5 (95% CI, 0.7 to 3.2) for elevated IgG titres (≥128) against CP, and 1.8 (95% CI, 0.9 to 3.6) for cLPS-containing immune complexes. These differences were not statistically significant, but they indicated a trend, that the presence of CPantibodies together with cLPS-containing immune complexes might be a risk factor for CAD. Miettinen et al. (20)found an independent positive association between seropositivity to CPand CAD in a cohort study from eastern Finland only in nondiabetics, but not in a population from western Finland.
In a cross-sectional study, Patel et al. (6)investigated a population-based random sample of 388 male London inhabitants. The OR for having electrocardiographic abnormalities suggestive of CAD or a history of angina or myocardial infarction, given seropositivity to CP(IgG ≥ 64), was 2.2 (95% CI, 1.1 to 4.6) after adjustment for several potential confounders.
No association between seropositivity to CPand CAD was reported in a large population-based case-control study from Israel for the first time in which potential confounders were carefully controlled for (21). This study included 302 patients with first acute myocardial infarction and 486 age- and gender-matched controls. The prevalence of anti-CP-IgG (≥32) was 81% in male patients (85% in male controls) and 65% in female patients (69% in female controls). The prevalence of anti-CP-IgA (≥20) was 47% in male patients (45% in male controls) and 20% in female patients (22% in female controls). Recent prospective data from the Physicians’ Health Study also showed no evidence of an association between positive anti-CP-IgG titre and future coronary risk (22).
Results from all these studies were based on clinical signs of CAD and did not use coronary angiography to characterize their cases.
Two case-control studies from Seattle, Washington, had included patients with angiographically documented CAD (7,8). In the first study, patients without angiographically detectable CAD (n = 95) served as controls and were compared to 461 cases (7). The investigators estimated a relative risk of CAD for subjects with positive IgG antibodies against CPof 1.6 (95% CI; 1.0 to 2.7). In the second study, population-based controls (n = 120) without CAD by history were included and compared with 171 cases (8). An adjusted OR of 4.2 (95% CI; 1.8 to 10.0) for a positive antibody titre against CPwas found. In both studies, adjustment was carried out only for age, gender, and month of blood drawing, but not for smoking behavior and other potential confounders.
Only high positive CP-IgG and IgA titres were related to angiographically defined CAD in a study from London (9)that included 170 subjects. A large study from northern Germany (10), with 412 cases and 431 unmatched controls, found a crude OR for angiographically defined CAD, given positive IgG antibodies against CPof 2.4 (95% CI, 1.6 to 3.7), but no adjustments for potential confounders were performed, although the controls were on average 20 years younger than the cases.
Systemic inflammation and infection with CP
Various potential pathophysiological mechanisms relating infectious agents to atherogenesis have been postulated (1–3)—that is, an increased production of cytokines (IL-6 or IL-1) and acute-phase reactants, a local or systemic disturbance of fibrinolysis and blood coagulation, a direct infection of the arterial wall via macrophages and alteration of vascular cell function and an immunological response (cross reaction) due to bacterial heat-shock protein. However, the potential importance and the mediating mechanisms of chlamydial infection in atherogenesis are only incompletely understood.
In the second part of the study we investigated the association between various markers of systemic inflammation and seropositivity to cLPS and CP. The CRP has been measured by an immunoradiometric assay that is able to detect values well below the clinically relevant range. However, seropositivity to cLPS or CPwas not associated with elevated CRP concentrations; neither was there an association between other inflammatory markers and antibodies against cLPS or CP. These results are in contrast to findings by Patel et al. (6), who reported elevated levels of fibrinogen in patients with a clinical or ECG diagnosis of CAD and a positive serostatus to CP. A recently published report by Anderson et al. (23)showed a borderline significant association between elevated CRP and combined seropositivity to Helicobacter pyloriand CP, but not to CPalone. Also, no relationship between CRP concentrations and levels of IgG antibody titre against CPwas found in the Physicians’ Health Study (22).
Study strengths and limitations
Our study included a relatively large number of patients and had a power of >80% (>84%) to detect an OR of 1.55 or above for an increased IgG (IgA) antibody titre against cLPS at a significance level of alpha = 0.05. Moreover, careful adjustment was made for potential confounders. It is widely recognized that higher age and male sex are positively associated with the prevalence of antibodies to CP(24,25); the prevalence rises from about 20% in teenagers to about 60% to 80% in subjects aged 40 to 80 years (24,25). Cigarette smoking also seems to be positively associated with chlamydial infection (6,26).
We included only cases with a diagnosis of CAD established during the two preceding years, documented either by myocardial infarction or coronary angiography. This was done to minimize potential selection bias by selective survival, and to focus on a potential association between the first appearance of CAD and serostatus. Because patients with an acute coronary syndrome have elevated markers of systemic inflammation, we enrolled only patients with stable CAD. The choice of blood donors as controls is generally suboptimal, because blood donors probably tend to be healthier than population-based controls. However, results as reported in the present study without statistically significant differences of prevalence against CPantibodies between both groups give further support to our conclusions. A limitation of the present study was the inability to perform an ECG or coronary angiography in controls, so that asymptomatic CAD in controls cannot be ruled out. However, prevalence of CAD in asymptomatic middle-aged subjects appears to be low (for review, see Parmley ), and selection of controls among subjects who undergo coronary angiography for different reasons would possibly introduce more severe bias if these conditions were related to CPinfection.
It is unclear in what way antibodies reflect previous or current chronic or latent infection of CP. Although serological criteria have been suggested for chronic CPinfection, they have not been validated against the presence of DNA in atherosclerotic plaques or mononuclear cells in the blood. However, we have used two different serological tests with almost entirely consistent results. First, the MIF method is able to detect specific antibodies against CP, which persist for a long time. This may explain a higher prevalence of specific CPantibodies as compared with antibodies against cLPS, which has also been reported by others (4,28). The MIF method may perform differently in different laboratories. Our method has been compared with that of several well-experienced laboratories and seems to produce specific antibody results. Second, we have used a genus-specific rELISA against cLPS (including Chlamydia trachomatisand C. psittaci), which may be particularly useful for acute (re-) infection (17). This rELISA test is commercially available and reliably standardized. This test is not specific for CP; however, the prevalence of IgG antibodies against C. trachomatisand C. psittaciis low (in the present study 7% and <1%, respectively).
Our data do not support an important role of CPin atherogenesis. The prevalence of antibodies against cLPS and CPwas almost identical in cases and controls, in cases with or without previous myocardial infarction and in cases with one-, two- or three-vessel disease. However, it cannot be excluded that, in patients with atherosclerosis, CPmodulates an inflammatory process after reaching the atherosclerotic plaques by macrophages attracted to the site of active inflammation and that this process might not be detectable by serology. Recently, viable CPhas been detected in human atherosclerotic lesions (29), and in an experimental cholesterol-feeding rabbit model, CPinduced progression of atherosclerosis (30). However, it may be a significant observation of our study that this suggested inflammatory effect of CPdoes not lead to an increase in systemic inflammatory activity. Moreover, prior or chronic infection with Chlamydiaspecies does not explain the elevation of inflammatory markers, which were consistently found in patients with CAD (for review, see Danesh et al. ) and which we were also impressed to see in the present study.
In conclusion, this case-control study did not confirm a strong association between CAD and antibodies to the common cLPS antigens or specific antibodies to CP. Furthermore, seropositivity to cLPS or CPwas not associated with an increased systemic inflammatory activity. Therefore, these findings, based on a serological approach, do not support an important role of CPin atherogenesis.
We are indebted to the staff of the blood bank for their help and to G. Trischler and S. Kuhn for excellent technical assistance. We would also like to thank Prof. Mark B. Pepys (F.R.S.), Hammersmith Hospital, London, U.K., for providing the monoclonal and polyclonal CRP antibodies and for his help in establishing the CRP-IRMA at our laboratory.
☆ Supported by grants from the Medical Faculty of the University of Ulm, from Astra Chemicals, Wedel, Germany, and from Medac Diagnostica, Wedel, Germany.
- body mass index
- coronary artery disease
- chlamydial lipopolysaccharide
- Chlamydia pneumoniae
- C-reactive protein
- enzyme linked immunoassay
- immunoglobulin G/A/M
- Received March 9, 1999.
- Revision received July 22, 1999.
- Accepted September 21, 1999.
- American College of Cardiology
- Libby P,
- Egan D,
- Skarlatos S
- Mehta J.L,
- Saldeen T.G.P,
- Rand K
- Patel P,
- Mendall M.A,
- Carrington D,
- et al.
- Thom D.H,
- Wang S.P,
- Grayston J.T,
- et al.
- Maass M,
- Gieffers J
- Kuo C,
- Shor A,
- Campbell L.A,
- et al.
- Muhlestein J.B,
- Hammond E.H,
- Carlquist J.F,
- et al.
- Brade L,
- Brunnemann H,
- Ernst M,
- et al.
- Brade L,
- Holst O,
- Kosma P,
- et al.
- Miettinen H,
- Lehto S,
- Saikku P,
- et al.
- Kark J.D,
- Leinonen M,
- Paltiel O,
- Saikku P
- Ridker P.M,
- Kundsin R.B,
- Stampfer M.J,
- Poulin S,
- Hennekens C.H
- Anderson J.L,
- Carlquist J.F,
- Muhlestein J.B,
- Horne B.J,
- Elmer S.P
- Leinonen M
- Hahn D.L,
- Golubjatnikov R
- Parmley W.W
- Verkooyen R.P,
- van Lent N.A,
- Mousavi Joulandan S.A,
- et al.
- Maass M,
- Bartels C,
- Engel P.M,
- Mamat U,
- Sievers H.H
- Muhlestein J.B,
- Anderson J.L,
- Hammond E.H,
- et al.