Abstract
Criticisms of serological studies on Helicobacter pylori and ischaemic heart disease (IHD) include: undiagnosed heart disease in live controls; no assessment of severity or outcome of IHD; and qualitative not quantitative measurements of IgG to the bacteria. The aim was to assess quantitatively IgG levels specific for H. pylori (ng ml−1) among patients who survived a myocardial infarction (MI) with those who died of IHD. Sera were from four groups: (1) men who survived one MI; (2) men matched for age and socioeconomic background to group 1; (3) individuals who died suddenly of IHD; (4) accidental deaths matched for age and sex to group 3. Levels of IgG to H. pylori increased with age (P<0.005) but were not associated with smoking or socioeconomic groups. There was a correlation between IgG to the bacteria and decreasing socioeconomic levels only among group 1 (P<0.01). IgG levels were higher for subjects who died of heart disease (median=151 ng ml−1) compared with survivors (median=88 ng ml−1) (P=0.034) and higher for survivors compared with their controls (median=58 ng ml−1) (P=0.039). Future serological studies of H. pylori in relation to IHD should be quantitative and severity of disease considered in analyses.
1 Introduction
There is conflicting evidence for associations between chronic infection with Helicobacter pylori and ischaemic heart disease (IHD) [1–5]. Some studies indicated H. pylori is a marker for poor socioeconomic conditions in early life [6,7] or results are due to chance or preferential publication of positive results [8,9]. Major criticisms include undetected heart disease in controls and qualitative not quantitative tests for antibodies to H. pylori[8,9]. In patients with gastrointestinal disease, levels of IgG antibodies to H. pylori correlated with density of antral colonisation by the bacteria and with the degree of gastritis in the antrum [10]. Because levels of IgG to H. pylori were related to severity of gastrointestinal disease [11], our aim was to assess quantitatively IgG levels among patients who survived a myocardial infarction (MI), subjects who died of heart disease and their respective control groups.
2 Materials and methods
2.1 Subjects
Levels of IgG to H. pylori were determined for four groups: (1) men demonstrated by clinical parameters to have suffered a first MI (n=171); (2) men matched closely as possible for age, socioeconomic group and smoking habits to group 1 but with no clinical evidence of heart disease (n=175); (3) individuals who died suddenly and unexpectedly of IHD (n=125); (4) age- and sex-matched accident victims with no evidence of IHD at autopsy (n=50). Autopsies were performed by experienced pathologists by a standardised comprehensive protocol. All bodies were stored at 4°C after death, with intervals between death and post mortem examination varying from 4 to 130 h. Blood samples were collected from the iliac veins. IHD in the autopsy group was defined as: (1) evidence of old ischaemia in the left ventricular myocardium and/or severe coronary arterial atheroma; (2) changes of coronary arteries with stenosis but no evidence of acute occlusion or intraluminal thrombosis.
For groups 1 and 2, the number of cigarettes smoked per day was known but for autopsy subjects there was no information on smoking habits. Socioeconomic status was assigned as Carstairs index (CI) derived from the subject's postcode ranging from 1 (most affluent) to 7 (most deprived) [12].
2.2 Detection of IgG specific for H. pylori
IgG antibodies to H. pylori were detected as described previously [13]. The positive control was a serum sample obtained from a patient from whose antral biopsy H. pylori was isolated, and the negative control was a serum sample from a patient whose antral biopsy was negative on culture, Campylobacter-like organisms test for urease production and microscopic examination. All samples were diluted 1 in 100, duplicate samples tested and results expressed in ng ml−1 IgG derived from values of the standard human IgG (Sigma) run in each experiment [13].
2.3 Statistical methods
Results for patients and controls were compared by chi square with Yates' correction, Mann-Whitney U test and Kruskal-Wallis test. Levels of IgG to H. pylori and age, socioeconomic group and smoking habits were assessed by Spearman's correlation.
3 Results
3.1 Pilot study
In a pilot study of 70 autopsy subjects, IgG specific for H. pylori was higher in sera from IHD deaths (n=37) (mean=48 ng ml−1) than those from accidental deaths (n=33) (mean=31 ng ml−1), but the difference was not significant. Detailed examination of the autopsy findings revealed eight accident victims had significant evidence of heart disease, although this was not the cause of death. Reanalysis of the results including accident victims with evidence of heart disease in the IHD group found significantly higher levels of IgG to H. pylori among the IHD group (mean=49 ng ml−1) compared with the accident group (mean=25 ng ml−1) (P<0.01).
3.2 Characteristics of the study populations and levels of IgG to H. pylori
There were no significant differences between mean ages of the four groups or association between age and socioeconomic group (r=-0.03, 96% CI −0.11< R<0.06) or age and number of cigarettes smoked per day (r=0.01, 95% CI −0.09< R<0.11). Among men who survived a first MI, the mean number of cigarettes smoked per day was significantly higher than that of the control group. There was a significant association between numbers of cigarettes smoked per day and lower socioeconomic levels (r=0.18, 95% CI 0.06< R<0.28) (P<0.001) (Table 1).
Age, cigarette consumption and socioeconomic category for the groups examined
| Age (years) | Cigarettes (no. smoked per day) | Socioeconomic group (no.) | ||||
| mean | S.D. | mean | S.D. | 1–3 | 4–7 | |
| First MI (n=171) | 53.15 | 7.53 | 23.43 | 16.36 | 51 | 99 |
| Control group (n=175) | 53.43 | 7.23 | 13.57 | 13.87 | 118 | 78 |
| Died of IHD (n=125) | 54.74 | 10.48 | – | – | 45 | 21 |
| Accident victims (n=50) | 49.50 | 13.62 | – | – | 65 | 17 |
| Age (years) | Cigarettes (no. smoked per day) | Socioeconomic group (no.) | ||||
| mean | S.D. | mean | S.D. | 1–3 | 4–7 | |
| First MI (n=171) | 53.15 | 7.53 | 23.43 | 16.36 | 51 | 99 |
| Control group (n=175) | 53.43 | 7.23 | 13.57 | 13.87 | 118 | 78 |
| Died of IHD (n=125) | 54.74 | 10.48 | – | – | 45 | 21 |
| Accident victims (n=50) | 49.50 | 13.62 | – | – | 65 | 17 |
Age, cigarette consumption and socioeconomic category for the groups examined
| Age (years) | Cigarettes (no. smoked per day) | Socioeconomic group (no.) | ||||
| mean | S.D. | mean | S.D. | 1–3 | 4–7 | |
| First MI (n=171) | 53.15 | 7.53 | 23.43 | 16.36 | 51 | 99 |
| Control group (n=175) | 53.43 | 7.23 | 13.57 | 13.87 | 118 | 78 |
| Died of IHD (n=125) | 54.74 | 10.48 | – | – | 45 | 21 |
| Accident victims (n=50) | 49.50 | 13.62 | – | – | 65 | 17 |
| Age (years) | Cigarettes (no. smoked per day) | Socioeconomic group (no.) | ||||
| mean | S.D. | mean | S.D. | 1–3 | 4–7 | |
| First MI (n=171) | 53.15 | 7.53 | 23.43 | 16.36 | 51 | 99 |
| Control group (n=175) | 53.43 | 7.23 | 13.57 | 13.87 | 118 | 78 |
| Died of IHD (n=125) | 54.74 | 10.48 | – | – | 45 | 21 |
| Accident victims (n=50) | 49.50 | 13.62 | – | – | 65 | 17 |
The proportion of individuals in higher socioeconomic groups (1–3) (30%) was lower among patients who survived a first MI compared with their respective controls (57%) (χ2=23.80, df=1, P=0.000001). This was not observed for the autopsy groups.
In industrialised countries, the proportion of individuals seropositive for H. pylori increases with age. There was a significant increase in levels of IgG to H. pylori associated with age (r=0.14, 95% CI 0.05< R<0.22) (F=9.98, P<0.005). Analysis of individual groups found this was significant only for subjects who died of heart disease (r=0.25, 95% CI 0.08< R<0.41) (P<0.005).
Levels of IgG to H. pylori were not associated with numbers of cigarettes smoked per day for groups 1 and 2 combined (r=0.07, 95% CI −0.04< R<0.17), patients who survived a first MI (r=0.07, 95% CI −0.09< R<0.21) or their controls (r=0.08, 95% CI −0.07< R<0.23).
There was no association between IgG levels to H. pylori and CI for 490 individuals for whom this information was available (r=0.08, 95% CI −0.01< R<0.17). Among patients who survived a first MI (n=168), there was a significant correlation with lower socioeconomic levels and IgG to H. pylori (r=0.20, 95% CI 0.05< R<0.34) (P<0.01) but not for their control group. There was no significant association between levels of IgG to H. pylori and socioeconomic level in the autopsy groups.
3.3 Patient groups and IgG to H. pylori
Levels of IgG to H. pylori for patients who survived a first MI (n=171) (median=88 ng ml−1) were significantly lower than for those who died (n=125) (median=151 ng ml−1) (χ2=4.4485, P=0.034) (Fig. 1) but significantly higher than those of their controls (n=175) (median=58 ng ml−1) (χ2=4.247, P=0.039). Although the levels of IgG to H. pylori were higher among IHD deaths (n=125) compared with the accident victims (n=50) (median=50 ng ml−1), the difference was not significant (χ2=1.8, P=0.18).
Median levels of IgG to H. pylori (ng ml−1) for patients who survived a first HI (group 1), their age- and sex-matched controls (group 2), individuals who died of IHD (group 3) and their age- and sex-matched controls (group 4).
Median levels of IgG to H. pylori (ng ml−1) for patients who survived a first HI (group 1), their age- and sex-matched controls (group 2), individuals who died of IHD (group 3) and their age- and sex-matched controls (group 4).
To determine if higher levels of IgG for H. pylori were due to post mortem changes, we assessed total IgG levels in relation to time between death and autopsy. The autopsy specimens (groups 3 and 4) had significantly lower levels of total IgG than groups 1 and 2: group 1=3660 µg ml−1; group 2=4010 µg ml−1; group 3=1990 µg ml−1; group 4=2240 µg ml−1 (χ2=27.35, df=3, P<0.001). There was a significant decline in total IgG associated with interval between death and autopsy (r=0.324, 95% CI 0.85< R<0.105, P<0.025). Levels of IgG specific for H. pylori did not show this pattern (r=0.07, 95% CI −0.0102< R<0.0087, P<0.56).
4 Discussion
Studies of H. pylori in IHD are criticised for many reasons: small sample sizes; assessment of different factors in relation to infection with H. pylori; omission of systematic reviews of findings of other studies; selective reporting of results [8]. In addition, most studies on serological evidence of H. pylori infection were qualitative not quantitative, severity or outcome of IHD was not considered and undetected IHD in ‘healthy’ controls not recognised.
This study supports previous work on prevalence of antibodies to H. pylori among patients with IHD [1–5]. It differs in several ways: (1) it examined level of antibodies quantitatively; (2) it compared IgG levels to H. pylori among patients who survived a first clinical MI with sudden, unexpected IHD deaths; (3) direct evidence of heart disease was assessed for age- and sex-matched comparison group in the autopsy series.
Many serological studies on H. pylori examined risk factors for IHD in relation to presence or absence of antibodies to the bacteria [8]. Levels of IgG antibodies to H. pylori correlated with density of antral colonisation by the bacteria and with the degree of gastritis of the antrum [10]. Density of colonisation was associated with severity of gastrointestinal disease in another recent study [11]. There was a significant difference in levels of IgG to H. pylori between patients who survived a first MI and their respective control group, and significantly higher levels of IgG among those who died of IHD. The cases of IHD examined were unexpected deaths which might indicate these were more severe episodes. In a prospective study, severity of IHD and antibodies to H. pylori were assessed. The OR for seropositivity and MI was 1.77 (CI 1.06–2.95), but among fatal cases the OR was 2.41 (CI 1.113–5.12). Antibodies were not assessed quantitatively and the OR for both categories was not significant after adjustment for other risk factors [3].
We found no other studies on H. pylori infection and IHD with material from autopsies. Higher IgG levels for H. pylori observed in autopsy cases are probably underestimates of those at the time of death as there is a significant decline in levels of total IgG related to interval between death and collection of blood samples. The advantage of autopsy-derived specimens is that direct evidence of heart disease is available for both IHD and comparison groups. In the pilot study, reassessment of data with respect to presence of IHD in accident victims highlighted the possibility of unrecognised heart disease in controls. The findings agree with the report that H. pylori infection is associated with coronary atheroma when compared with patients without IHD [2].
Only one study assessed virulence factors of different strains of H. pylori present in patients with IHD. In gastrointestinal diseases, the cytotoxin-associated gene A (cag A) is present in a higher proportion of strains associated with symptomatic diseases [14]. H. pylori was isolated from 62% of patients with IHD and 40% without IHD, there was a higher proportion of cag A positive isolates from IHD patients (43%) than controls (17%) [1].
Infection-induced inflammatory responses are implicated in pathophysiological events associated with IHD [15]. We conclude that quantitative assessment of antibodies to H. pylori, careful assessment for evidence of IHD among controls and severity of IHD must be considered in analysis of the role of these infections in the aetiology of IHD.
Acknowledgements
This work was supported by Chest Heart and Stroke, Scotland. We are grateful to Dr. R. Riemersma for samples for groups 1 and 2.
