SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Background  Coeliac disease is more prevalent than was previously thought. The association between coeliac disease and cardiovascular outcome is not clear.

Aim  To investigate whether coeliac disease patients have an increased risk of cardiovascular events.

Methods  A community-based cohort study using a record-linkage database. Three hundred and sixty-seven coeliac patients identified by a positive antiendomysial antibody test or a diagnosis with small bowel biopsy, and 5537 subjects who were tested and had a negative coeliac immunology, were included in the study.

Results  The crude rates of cardiovascular events were 9.5 per 1000 person-years (95% CI: 4.4–14.6) in the coeliac cohort and 8.9 per 1000 person-years (95% CI: 7.6–10.3) in the antiendomysial antibody-negative cohort. Compared with the antiendomysial antibody-negative cohort, the adjusted relative risk of cardiovascular events for coeliac cohort was 1.9 (95% CI: 1.00–3.60). When we excluded patients who had previous hospitalization for cardiovascular disease, the adjusted relative risk was 2.5 (95% CI: 1.22–5.01). The use of any cardiovascular drugs prior to and after entry to the study were 36% and 29% for the coeliac cohort (P = 0.05), and 34% and 26% for the antiendomysial antibody-negative cohort (P < 0.01).

Conclusion  Our findings suggest that coeliac disease seems to be associated with an increased risk of cardiovascular outcome.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

It is now accepted that coeliac disease is more prevalent than was previously thought being between 0.5% and 1% of the general population.1 We previously reported a case of a patient with both hypertension and coeliac disease who had hyperhomocysteinaemia.2 This was attributed to malabsorption of essential co-factors such as folate and vitamin B6. Treatment with a gluten-free diet and co-factor supplements improved blood pressure control. Recently, an analysis from the US Nurses’ Health Study reported that a high folate intake was associated with lower risk of hypertension in women.3 However, it has not been clear whether there is an association between coeliac disease and cardiovascular disease. Some studies have reported that there was a reduced risk of cardiac disease in coeliac disease patients, while other studies have found that coeliac disease was associated with increased risk of cardiac disease.4–8 More recently, case reports have suggested an association between coeliac disease and stroke.9, 10 We hypothesized that patients with coeliac disease in whom malabsorption is common may be at increased risk of hypertension and therefore cardiovascular disease.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study design

We have done a community-based cohort study using the Medicines Monitoring Unit’s record-linkage database in Tayside Scotland from 1993 to 2003. The database contains several data sets including all dispensed community prescriptions, hospital discharge data and other data that are linked by a unique patient identifier, the community health index number.11 These data were made anonymous for the purposes of research as approved by the Tayside Caldicott Guardians. The project was also approved by the Tayside Committee on Research Medical Ethics.

Study population

Subjects resident in Tayside and registered with a general practitioner between January 1993 and December 2003 formed the study population. These subjects were residents of Tayside throughout the study period, or died during the study period.

Study subjects

Coeliac cohort.  All incident subjects with the diagnosis of coeliac disease with either a positive antiendomysial antibody (AEA-IgA) test or positive small bowel biopsy between July 1993 and December 2003 formed the study group. Patients entered the study at the date of the positive test result.

Comparator group.  Subjects were patients who had a negative AEA-IgA test in the same laboratory between July 1993 and December 2003. Patients entered the study at the date of the negative test result.

Definition of cardiovascular disease

A diagnosis of ischaemic heart disease, heart failure, cerebrovascular disease or cardiovascular death was ascertained from the hospital discharge diagnosis data (SMR1) or from death certification from General Register Office in Scotland coded by primary International Classification of Diseases (ICD)-9 code and ICD-10 code. The accuracy of diagnosis for these hospital discharge diagnosis data was about 88%.12

Outcome variables

The outcome of the study was cardiovascular events defined as the composite end point of hospitalization with a primary diagnosis of myocardial infarction, angina, stroke, transient ischaemic attack, congestive cardiac failure or cardiovascular death during the follow-up period.

Covariates

Concomitant drug prescription.  Each dispensed prescription has details of the date of prescription, the strength of the tablet, the number of tablets dispensed and the instructions on how these should be taken. Cardiovascular drugs were examined in the study. Other drugs of interest were gluten-free foodstuff prescriptions, folic acid, allopurinol, hormone replace therapy (HRT), nonsteroidal anti-inflammatory drugs (NSAIDs) and oral glucocorticoids.

Other covariates.  Other covariates included age, sex and the Carstair’s social deprivation code (derived from patients’ postcode and 1991 census data comprising social class, overcrowding, male unemployment and no-car ownership), prior hospitalization for cardiovascular disease and diabetes mellitus.

Statistical analysis

Data were summarized as number of subjects (percentage) for categorical variables. Chi-square was performed to determine significant differences. Events that occurred during the study period were counted and event rates were compared between cohorts. A Cox regression model was used to assess whether there was a risk difference in cardiovascular events between cohorts. Analyses were carried out univariately and multivariately. The multivariable analyses were adjusted for age, gender, social deprivation, diabetes mellitus, any cardiovascular drug use, allopurinol, HRT, NSAIDs, oral glucocorticoids, folic acid and gluten-free foodstuff prescriptions. Diabetes mellitus history was defined as previous hospitalization or taking diabetic medication. Cardiovascular drugs included angiotension-converting enzyme (ACE) inhibitors, lipid-lowering drugs, beta-blockers, antiplatelet agents, cardiac glycosides, diuretics, nitrates, antihypertensive drugs, anticoagulants and calcium channel blockers. All statistical analyses were carried out using sas (version 8).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Baseline characteristics

There were 463 incident patients diagnosed with coeliac disease in Tayside between 1993 and 2003. Of these, 138 had both AEA-IgA test and small bowel biopsy, 217 had AEA-IgA test only and 108 had small bowel biopsy only. Ninety-six coeliac patients were excluded from the analysis because they did not live in Tayside throughout the study period. Thus, 367 coeliac patients (117 men and 250 women) were analysed. The number of new cases rose significantly from 13 in 1993 to 61 in 2003 (trend test, P < 0.01). About one-third of subjects were men and two-thirds were women. The average age of coeliac patients was 46 years (standard deviation, 18.3). There were 5537 patients who had negative test results for AEA in the same laboratory between July 1993 and December 2003. They formed comparator group. Table 1 shows the characteristics of the two cohort patients. Coeliac patients were significantly older than patients in the AEA-negative comparator cohort (P < 0.01). There were more female patients and more previous cardiovascular hospitalizations in the coeliac cohort than in the AEA-negative comparator cohort. However, the differences were not statistically significant.

Table 1.   Characteristics of patients in the three cohorts
 Coeliac patients, n (%)AEA-negative comparators, n (%)
  1. * P < 0.01.

  2. † Excluding missing data.

  3. AEA, antiendomysial antibody.

Total number of patients367 (100)5537 (100)
Gender
 Male117 (32)1917 (35)
 Female250 (68)3620 (65)
Age band*
 <1016 (4)546 (10)
 10–1917 (5)537 (10)
 20–2928 (8)623 (11)
 30–3965 (18)849 (15)
 40–4982 (22)916 (17)
 50–5976 (21)815 (15)
 60–6941 (11)622 (11)
 70–7935 (10)463 (8)
 80+7 (2)166 (3)
Social deprivation category†
 131 (8)375 (7)
 266 (18)1005 (18)
 3107 (29)1436 (26)
 433 (9)495 (9)
 537 (10)674 (12)
 657 (16)994 (18)
 734 (9)526 (10)
Previous disease history
 CVD18 (5)256 (5)
 Diabetes mellitus 22 (6)393 (7)

Cardiovascular drug use in the two cohorts

Table 2 shows the use of cardiovascular drugs in the two cohorts. In the coeliac cohort, the proportions of patients who used cardiovascular drugs prior to entry of the study ranged from 1% for cardiac glycosides to 19% for beta-blockers. The corresponding figures for comparator cohort were 1% for cardiac glycosides to 18% for diuretics for the AEA-negative cohort. The use of cardiovascular drugs after entry of the study ranged from 1% for cardiac glycosides to 14% for antiplatelets for the coeliac cohort and 1% for cardiac glycosides to 12% for diuretics for the AEA-negative cohort. The uses of any cardiovascular drugs prior and after entry of the study were 36% and 29% for the coeliac cohort (P = 0.05), and 34% and 26% for the AEA-negative cohort (P < 0.01).

Table 2.   Cardiovascular drug use in the two cohorts
 Coeliac patients, n (%)AEA-negative comparators, n (%)
  1. AEA, antiendomysial antibody.

Drug use prior entry of the study
 ACE inhibitors15 (4)255 (5)
 Lipid-lowering drug19 (5)349 (6)
 Beta-blockers71 (19)902 (16)
 Antiplatelets43 (12)570 (10)
 Cardiac glycosides4 (1)70 (1)
 Diuretics62 (17)966 (17)
 Nitrates34 (9)469 (8)
 Antihypertensive drugs8 (2)101 (2)
 Anticoagulants5 (1)104 (2)
 Calcium channel blockers46 (13)618 (11)
 Any cardiovascular drugs131 (36)1864 (34)
Drug use post entry of the study
 ACE inhibitors14 (4)286 (5)
 Lipid-lowering drug30 (8)514 (9)
 Beta-blockers33 (9)460 (8)
 Antiplatelets50 (14)604 (11)
 Cardiac glycosides3 (1)48 (1)
 Diuretics37 (10)654 (12)
 Nitrates22 (6)322 (6)
 Antihypertensive drugs5 (1)80 (1)
 Anticoagulants7 (2)171 (3)
 Calcium channel blockers31 (8)438 (8)
 Any cardiovascular drugs106 (29)1465 (26)

Cardiovascular events during the follow-up period

The average follow-up was 3.7 years for the coeliac cohort. Thirteen events were observed in the coeliac cohort in 1371 person-years (PY) of follow-up, giving a rate of 9.5 per 1000 PY (95% CI: 4.4–14.6). In the AEA-negative comparator cohort, 171 events were observed in 19 128 PY of follow-up, giving a rate of 8.9 per 1000 PY (95% CI: 7.6–10.3). Table 3 shows the results of univariate and one covariate adjusted relative risks (RR). The first column shows all study subjects and we found that gluten-free prescription was the main confounding for the results. The second column of Table 3 shows the results excluding the users of gluten-free prescriptions (164 from the coeliac cohort and 100 from the AEA-negative comparator cohort). Among 264 users of gluten-free prescriptions, the adjusted RR of cardiovascular events for the coeliac cohort was 1.4 (95% CI: 0.12–15.16) when compared with the AEA-negative cohort. Compared with the AEA-negative cohort, the adjusted RR of cardiovascular events for the coeliac cohort was 1.9 (95% CI: 1.00–3.60). When we excluded patients who had a previous hospitalization for cardiovascular disease, the figure was 2.5 (95% CI: 1.22–5.01; Table 4).

Table 3.   Univariate and one covariate adjusted hazard ratios for cardiovascular disease outcome
 Coeliac patients vs. AEA-negative comparatorsCoeliac patients vs. AEA-negative comparators (excluding 264 patients who were on gluten-free prescriptions)
  1. AEA, antiendomysial antibody; HRT, hormone replace therapy; NSAID, nonsteroidal anti-inflammatory drug.

Unadjusted
 HR (95% CI)1.1 (0.62–1.92)2.5 (1.36–4.59)
Adjusted for age
 HR (95% CI)1.1 (0.60–1.87)2.4 (1.29–4.37)
Adjusted for sex
 HR (95% CI)1.1 (0.63–1.94)2.5 (1.33–4.50)
Adjusted for deprivation
 HR (95% CI)1.0 (0.57–1.86)2.3 (1.22–4.39)
Adjusted for diabetes mellitus
 HR (95% CI)1.1 (0.63–1.96)2.6 (1.41–4.77)
Adjusted for previous cardiovascular hospitalization
 HR (95% CI)1.1 (0.65–2.00)2.6 (1.42–4.81)
Adjusted for cardiovascular drugs
 HR (95% CI)0.95 (0.54–1.66)2.4 (1.32–4.46)
Adjusted for allopurinol
 HR (95% CI)1.1 (0.62–1.91)2.4 (1.28–4.36)
Adjusted for oral glucocorticoids
 HR (95% CI)1.1 (0.61–1.88)2.5 (1.34–4.54)
Adjusted for HRT
 HR (95% CI)1.1 (0.65–2.00)2.5 (1.36–4.60)
Adjusted for NSAIDs
 HR (95% CI)1.0 (0.59–1.82)2.6 (1.42–4.81)
Adjusted for folic acid
 HR (95% CI)1.1 (0.59–1.85)2.4 (1.32–4.48)
Adjusted for gluten-free prescriptions
 HR (95% CI)2.4 (1.32–4.42)
Table 4.   Univariate and multivariate hazard ratios for cardiovascular disease outcome
 Coeliac patients vs. AEA-negative comparators
  1. * Covariates included age, gender, social deprivation, diabetes mellitus, any cardiovascular drug use, folic acid, gluten-free prescriptions, allopurinol, HRT, NSAIDs and oral glucocorticoids.

  2. † Excluding patients who had previous hospitalization for cardiovascular disease.

  3. AEA, antiendomysial antibody; HRT, hormone replace therapy; NSAID, nonsteroidal anti-inflammatory drug.

Unadjusted HR (95% CI)1.1 (0.62–1.92)
Adjusted HR (95% CI)*1.9 (1.00–3.60)
Adjusted HR (95% CI)*†2.5 (1.22–5.01)

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Coeliac disease is a lifelong condition in which early diagnosis and strict adherence to a gluten-free diet improves the outcome of the disease. In this population-based cohort study, we found evidence that there was an association between coeliac disease and cardiovascular outcomes especially in subjects free from cardiovascular disease.

No previous studies have reported the details of cardiovascular drug use in coeliac disease patients. Our study has shown that coeliac patients used slightly more cardiovascular drugs than comparator group. West et al.8 reported that coeliac patients had a lower prevalence of hypertension and hypercholesterolaemia (a composite measure defined as a diagnosis of hypertension or hypercholesterolaemia and a prescription of antihypertensive medication or lipid-lowering drug) compared with the general population between 1987 and 2002 using the General Practice Research Database, a UK-based longitudinal primary care database. This is in direct contrast to the findings in our study. However, the methodology of the study was different from our study. Nevertheless, they found that the risk of stroke was higher in the coeliac patients than in the population controls (1.6, 95% CI: 0.99–2.59 for the incident subject and 1.2, 95% CI: 0.85–1.68 for the prevalent subjects).

We found that subjects who were prescribed gluten-free products had lower rates of cardiovascular disease. This throws up the tantalizing possibility that a gluten-free diet, which abolishes disease activity, may return cardiovascular risk to normal. An alternative hypothesis is that such subjects are compliant with medical therapy and exhibit good health behaviour. In support of this concept, it has been shown that compliance with placebo appears to be associated with reduced mortality.13

The underlying biological mechanism as to why coeliac subjects might be at increased cardiovascular risk might be that they may have higher levels of plasma homocysteine because of malabsorption of folic acid and vitamin B.2 However, two recent large randomized studies have failed to find any impact of supplementation with these vitamins in subjects with cardiovascular disease.14, 15 The issue of whether or not folate is associated with cardiovascular disease prevention remains controversial, with some arguing that current evidence from both observational studies and randomized-clinical trials is consistent with a short-term protective effect of 12% for ischaemic heart disease and 22% for stroke.16

We have observed an increased number of coeliac patients diagnosed during the study period. Given the stable population in Tayside, it is likely that the increase in newly diagnosed cases of coeliac disease is due to increased awareness of the disease by both health professionals and patients and the availability of reliable and sensitive antibody tests, thus more patients were suspected and tested with time.

Our study has some limitations. We ended up with smaller sample size than expected. Forty-seven per cent of coeliac patients were diagnosed by a positive AEA-IgA test. These may include some subclinical cases of coeliac disease. There may also be some false-positive results. However, a previous study has reported that the test has a sensitivity of 94% and a specificity of 99%.17 We were limited by the number of covariates on which we had data and consequently were not able to adjust for factors of possible influence such as smoking, obesity, exercise and disease severity, all of which are important risk factors in patients with heart disease. However, we did use the Carstairs socio-economic deprivation score as a surrogate, which provides adjustment for at least some of these factors.18 However, this adjustment can never fully account for residual confounding. The strengths of the present study are the population-based cohort design and complete follow-up over the study period. This approach allows a real-world population to be studied representing all socio-economic groups and within a universal healthcare coverage scheme.19

In conclusion, our findings suggest that coeliac disease seemed to be associated with an increased risk of cardiovascular outcome.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Declaration of personal interests: T. M. MacDonald has served as a speaker, a consultant and an advisory board member in the last year for Pfizer, Novartis, Servier and Kaiser Permanante. MEMO has had research grants from GSK, Aventis, Novartis, AstraZeneca, BMS, Boehringer Ingelheim, Pfizer and Novartis. The remaining authors have no conflict of interest. Declaration of funding interests: L. Wei holds a Special Training Fellowship in Health Services and Health of the Public Research award from the MRC, UK. This study was funded by Chief Scientist Office of the Scottish Executive (CZG/2/211).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References