Prospective Study of Potentially Virulent Strains of Helicobacter pylori and Coronary Heart Disease in Middle-Aged Men
Background—Studies are needed to test claims that potentially virulent strains of Helicobacter pylori are more strongly related to coronary heart disease (CHD) than are other strains.
Methods and Results—We measured serum IgG antibodies to mixed H pylori antigens and separately to the virulence-associated H pylori antigen CagA (cytotoxin-associated gene product A) in 505 CHD cases and in 1025 age-matched controls “nested” in a prospective study of 7735 British men (mean duration of follow-up in controls, 16 years). Of the 505 cases, 401 (79%) were seropositive for H pylori antibodies compared with 740 (72%) of the 1025 controls, yielding an odds ratio for CHD of 1.55 (95% CI 1.19 to 2.03), which fell to 1.30 (95% CI 0.88 to 1.90) after adjustments were made for standard vascular risk factors and indicators of socioeconomic status. Of the CHD cases, 240 (48%) were seropositive for IgG antibodies to CagA compared with 450 (44%) of the controls. When CagA-seropositive individuals were compared with H pylori–seronegative individuals, the odds ratio for CHD was 1.42 (95% CI 1.06 to 1.91), which fell to 1.10 (95% CI 0.71 to 1.71) after adjustments. In an analysis restricted to the 1141 (75%) H pylori–seropositive participants, the odds ratio for CHD was 1.0 (95% CI 0.78 to 1.29) in CagA–seropositive men. No strong associations were observed between H pylori seropositivity and blood lipids, blood pressure, markers of systemic inflammation, or plasma homocysteine.
Conclusions—H pylori infection is not strongly related to the incidence of CHD in late middle-aged men, and CagA-positive strains appear to be no more strongly related to the disease than other strains. However, further studies are required to confirm or refute the existence of any moderate associations, particularly at younger ages.
Systematic reviews of epidemiological studies have suggested the existence of a weakly positive association between coronary heart disease (CHD) and chronic Helicobacter pylori infection, but this could be due to confounding and/or possible bias.1 2 To clarify the nature of this association, new studies that can confirm or refute the existence of modest associations are necessary. One approach is to study infection with bacterial subtypes that might yield stronger relative risks of disease, such as H pylori strains expressing cytotoxin-associated gene product A (CagA). Compared with other H pylori strains, these strains may be more strongly associated with upper gastrointestinal diseases, such as peptic ulceration and gastric adenocarcinoma.3 4 A retrospective study has reported a 4-fold increased risk of CHD in people with raised serum antibodies to the CagA protein, but this hypothesis-generating observation,5 based on only 88 CHD cases, needs to be tested in larger samples. Hence, we report a prospective study of H pylori serology, including measurements for CagA status, in 505 CHD cases and in 1025 age-matched controls.
Cases and Controls
From 1978 to 1980, 7735 men aged 40 to 59 years (response rate 78%) were randomly selected from general practice registers in each of 24 British towns and entered in the British Regional Heart Study.6 Nurses administered epidemiological questionnaires, performed physical measurements, and recorded an ECG. Nonfasting venous blood samples were collected in 5661 men in 18 of the towns and stored at −20°C for subsequent analysis. Additional questionnaires were posted at 5 years (98% response among survivors) and at 12 years (90% response among survivors) of follow-up; these questionnaires inquired about car ownership and childhood social circumstances (father’s social class and childhood household amenities), respectively. All men had been monitored since entry for all-cause mortality and for cardiovascular morbidity, with a follow-up loss of <1% to date.7 Cases in the present study were men with major CHD events, fatal and nonfatal, occurring between the beginning of follow-up and December 1995, who had a stored serum sample available for analysis. Of 642 CHD cases, 135 were included in a previous study of H pylori and CHD8 ; thus, 507 cases remained for the present study (223 CHD deaths and 284 nonfatal myocardial infarctions [MIs]). Fatal CHD cases were ascertained through National Health Service Central Registers on the basis of a death certificate with International Classification of Diseases, 9th Revision, codes 410 to 414, and the diagnosis of nonfatal MI was in accordance with World Health Organization criteria and was based on reports from general practitioners, supplemented by regular reviews of general practice records.8 One thousand twenty-six controls, “frequency-matched” to cases according to town of residence and age in 5-year bands, were randomly selected from men who had survived to the end of the study period free from incident MI. Two potential cases and 1 potential control had no available serum.
H pylori–specific IgG titers were measured by laboratory workers unaware of the case-control status of blood samples in December 1998. An ELISA kit (Premier, Meridian Diagnostics) was used, and the intra-assay and interassay coefficients of variation were 3% and 6%, respectively. Anti-CagA serum antibodies were measured by ELISA with the use of a recombinant CagA antigen (orv200, a gift from Orovax Inc, Cambridge, Mass), according to a method described previously.3 This assay had 92% (34 of 37 patients) sensitivity and 96% (24 of 25 patients) specificity in 62 patients in another study in whom CagA status was directly assessed by commercial Western blot in gastric biopsy specimens (Helicoblot 2.0, Genelabs Diagnostics). Measurements were made of serum lipids, albumin, leukocyte count, and hematocrit by use of standard assays soon after the baseline blood collection; measurements were made of C-reactive protein and serum amyloid A by use of sensitive enzyme immunoassays in stored samples.9
Case-control comparisons involved unmatched stratified logistic regression fitted by unconditional maximum likelihood (Stata Corp). Adjusted analyses included the following (as explanatory variables): age; cigarette smoking habit (never, former, or current); daily cigarette consumption; nonfasting blood concentrations of total cholesterol, HDL cholesterol, and triglycerides; systolic and diastolic blood pressures; current social class (according to the Registrar General’s 1980 classification and including a separate category for armed forces8 ); housing tenure (owner, private rent, or council rent); marital status; current car ownership; and childhood social circumstances (eg, father’s occupation, family car ownership, bathroom in house, hot water tap in house, and bedroom sharing). Concentrations of blood lipids and a number of other blood components were also investigated as possible correlates of H pylori seropositivity. Emphasis was mainly given to differences more extreme than 2.6 standard deviations (2P≈0.01) to make some allowance for multiple comparisons. Methods to identify studies for an updated meta-analysis of prospective studies published before 1999 involving H pylori IgG serology and CHD have been described previously.1 2 Combination of results involved inverse-variance–weighted averages of log odds ratios. Heterogeneity was assessed by standard χ2 tests. Odds ratios are given with 95% confidence limits, and 2-sided probability (2P) values are used.
The mean age at CHD event in cases was 62 years (mean duration of follow-up in cases, 9.5 years). As would be expected, there were highly significant differences between cases and controls with respect to various known vascular risk factors, such as smoking, obesity, blood pressure, and blood lipids (Table 1⇓). The well-known associations were observed of H pylori seropositivity with age and markers of socioeconomic status, but no important associations were observed of H pylori seropositivity with a number of other potential intermediates, including blood lipids, blood pressure, markers of systemic inflammation, plasma homocysteine, hematocrit, insulin, and markers of renal function (Table 2⇓). Hence, adjustment for these factors would not be expected to alter the association of H pylori infection with CHD, and it did not do so. Moreover, there were no substantial differences in the levels of these factors in H pylori–infected individuals who were CagA seropositive compared with those who were CagA seronegative. Only 2 H pylori–seronegative individuals tested CagA seropositive.
Of the 505 CHD cases, 401 (79%) were seropositive for H pylori IgG antibodies compared with 740 (72%) of the 1025 controls (Table 3⇓). This difference yielded an age- and town-adjusted odds ratio for CHD of 1.55 (95% CI 1.19 to 2.03; χ2=10.4, 2P=0.001), which fell to 1.30 (95% CI 0.88 to 1.90; χ2=1.5, 2P>0.1) after additional adjustments for smoking, blood lipids, blood pressure, and both adult and childhood indicators of socioeconomic status. Varying the cutoff titer around the manufacturer’s recommended level did not materially alter the estimates. Of the 505 CHD cases, 240 (48%) were seropositive for IgG antibodies to CagA compared with 450 (44%) of the controls (Table 3⇓). In a comparison of CagA-seropositive individuals with those who were H pylori seronegative, the odds ratio for CHD was 1.42 (95% CI 1.06 to 1.91; χ2=5.7, 2P=0.02), which fell to 1.10 (95% CI 0.71 to 1.71; χ2=0.1, 2P>0.1) after adjustments. When attention was restricted to only those 1141 (75%) participants who were H pylori seropositive, 240 (60%) of 401 such cases and 450 (60%) of 740 such controls were also CagA seropositive, yielding an odds ratio for CHD of 1.0 (95% CI 0.78 to 1.29). These results were not materially changed when the analyses were restricted to the 328 cases and 820 controls with no evidence of CHD at baseline (Table 3⇓) or when the analyses were restricted to the 227 cases and 779 controls who had complete information on all reported markers of childhood socioeconomic status. The data were too sparse to subdivide reliably by potentially relevant characteristics, such as age and nonfatal versus fatal CHD.
This prospective community-based study of 505 CHD cases and 1025 controls contributes to the previous evidence involving H pylori infection and CHD in 3 main ways. First, when taken together with previous prospective studies, it argues against the existence of >1.5-fold odds ratios for CHD in late middle-aged people seropositive for mixed strains of H pylori infection. Second, in contrast to a much smaller study, it observed no stronger association between CHD and serological evidence of infection with potentially virulent H pylori strains than with other strains. Third, it suggests that H pylori seropositivity is not importantly related to levels of blood lipids, blood pressure, plasma homocysteine, markers of systemic inflammation, and a number of other potential vascular intermediates.
CHD and H pylori Infection
Large prospective studies of H pylori and CHD should be more informative than small retrospective studies because they avoid selection biases, assess infection before the onset of clinical disease, and generally record more information on possible confounders and/or mediators.1 The present study involves 505 CHD cases and is larger than all but one10 of the 6 previously reported prospective studies involving H pylori and CHD death or MI.8 10 11 12 13 14 In aggregate, these 7 reports involve a total of 2286 CHD cases. All made adjustment for smoking and other standard vascular risk factors, but only 5 (including the present study) adjusted for markers of socioeconomic status.8 10 11 12 The weighted mean age at CHD event was 67 years (ie, weighted mean age at baseline of 54 years with a weighted mean follow-up of 13 years). There was no evidence of heterogeneity between the 7 studies (χ26=4.2, P>0.1), and a combined analysis of their results yielded an adjusted risk ratio of 1.16 (95% CI 0.94 to 1.43, Figure⇓). Hence, at least in older individuals with CHD, it seems unlikely that H pylori is strongly related to disease. Because relative risks for standard vascular risk factors tend to be greater in early middle age than at older ages, studies of early-onset CHD may be more sensitive to the existence of any modest association. Indeed, the odds of MI were about twice as great in people seropositive for antibodies to H pylori (adjusted odds ratio 1.87, 95% CI 1.52 to 2.32) in a study of 1122 early-onset cases (average age 44 years) and 1122 age- and sex-matched controls.15
CHD and Potentially Virulent Strains of H pylori
In the only previously published study, an adjusted odds ratio of 3.8 (95% CI 1.6 to 9.1) was reported for CHD in CagA-seropositive individuals.5 This result, however, was based on only 88 cases (34 MI cases and 54 cases of angina), involved very wide CIs, and was especially susceptible to the play of chance (because it was a hypothesis-generating finding). Selection biases were also possible because the controls were blood donors who may not have been truly comparable to the cases. By contrast, the present study involves >5 times as many CHD cases as the previous report. It avoids selection biases because cases and controls were selected within the same contemporaneous cohort of men and because blood samples were collected several years before the diagnosis of disease. We observed no strong association between CagA seropositivity and CHD (adjusted odds ratio 1.10, 95% CI 0.71 to 1.71), and the adjusted odds ratio was similarly null (odds ratio 1.0, 95% CI 0.78 to 1.29) when comparisons of CagA seropositivity were restricted to only the 1141 (75%) individuals who were H pylori seropositive. It seems unlikely, therefore, that CagA-seropositive strains of H pylori are much more strongly related to CHD than are other strains.
H pylori Seropositivity, Vascular Risk Factors, and Other Characteristics
It has been suggested that H pylori infection is importantly related to levels of various known risk factors, such as blood lipids or fibrinogen.16 However, a meta-analysis of 18 reports involving a total of 10 000 individuals found no strong correlations of H pylori seropositivity with blood total cholesterol, triglycerides, fibrinogen, blood pressure, C-reactive protein, or leukocyte count,17 nor did the present study in 1500 individuals. The present report extends these observations by suggesting no strong associations of H pylori seropositivity with additional suspected risk factors, such as plasma homocysteine, serum amyloid A protein, insulin, hematocrit, and markers of renal function.
H pylori infection is not strongly related to the incidence of CHD in late middle age, and CagA-positive strains appear no more strongly related to the disease than other strains. Further studies may be needed, however, to confirm or refute the existence of any modest associations, particularly at younger ages.
The British Regional Heart Study (initiated by Professor A.G. Shaper) is a British Heart Foundation Research Group, and also receives support from the Department of Health. Dr Danesh is supported by Merton College and the Frolich Trust. Dr Atherton is supported by a Medical Research Council Clinician Scientist Fellowship. The recombinant CagA antigen used in this study was a gift of Orovax Inc, Cambridge, Mass. H. Refsum and P. Ueland provided homocysteine assays, and J.R. Gallimore and M.B. Pepys provided C-reactive protein and serum amyloid A assays. J. John provided valuable assistance.
- Received July 21, 1999.
- Revision received October 28, 1999.
- Accepted November 10, 1999.
- Copyright © 2000 by American Heart Association
Danesh J, Appleby P. Persistent infection and vascular disease: a systematic review. Expert Opin Invest Drugs. 1998;7:691–713.
Blaser MJ, Perez-Perez GI, Kleanthous H. Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res. 1995;55:2111–2115.
Atherton JC. H pylori virulence factors. Br Med Bull. 1998;54:105–120.
Pasceri V, Cammarota G, Patti G, Cuoco L, Gasbarrini A, Grillo RL, Fedeli G, Gasbarrini G, Maseri A. Association of virulent Helicobacter pylori strains with ischemic heart disease. Circulation. 1998;97:1675–1679.
Shaper AG, Pocock SJ, Walker M, Cohen NM, Wale CJ, Thomson AG. British Regional Heart Study: cardiovascular risk factors in middle-aged men in 24 towns. BMJ. 1981;283:179–186.
Whincup PH, Mendall MA, Perry IJ, Strachan DP, Walker M. Prospective relations between Helicobacter pylori infection, coronary heart disease, and stroke in middle-aged men. Heart. 1996;75:568–572.
Danesh J, Muir J, Wong Y-K, Ward M, Gallimore R, Pepys MB. Risk factors for coronary heart disease and acute-phase proteins: population-based study. Eur Heart J. 1999;20:954–959.
Wald NJ, Law MR, Morris JK, Bagnall AM. Helicobacter pylori infection and mortality from ischaemic heart disease: negative result from a large, prospective study. BMJ. 1997;315:1199–1201.
Strachan DP, Mendall MA, Carrington D, Butland BK, Yarnell JWG, Sweetnam PM, Elwood PC. Relation of Helicobacter pylori infection to 13-year mortality and incident ischemic heart disease in the Caerphilly prospective heart disease study. Circulation. 1998;98:1286–1290.
Aromaa A, Knekt P, Reunanen A, Rautelin HI, Kosunen TU. Helicobacter pylori and the risk of myocardial infarction. Gut. 1996;39(suppl 2):A91. Abstract.
Folsom AR, Nieto FJ, Sorlie P, Chambless LE, Graham DY. Helicobacter pylori seropositivity and coronary heart disease incidence. Circulation. 1998;98:845–850.
Danesh J, Youngman L, Clark S, Parish S, Peto R, Collins R. Helicobacter pylori and early onset myocardial infarction: case-control and sibling pairs study. BMJ. 1999;319:1157–1162.
Patel P, Mendall MA, Carrington D, Strachan DP, Leatham E, Molineaux N, Levy J, Blakeston C, Seymour CA, Camm AJ, Northfield TC. Association of Helicobacter pylori and Chlamydia pneumoniae infections with coronary heart disease and cardiovascular risk factors. BMJ. 1995;311:711–714.
Danesh J, Peto R. Risk factors for coronary heart disease and infection with Helicobacter pylori: meta-analysis of 18 studies. BMJ. 1998;316:1130–1132.