Article Text

Concise report
Ankylosing spondylitis: a risk factor for myocardial infarction?
  1. M J L Peters1,
  2. I Visman2,
  3. M M J Nielen3,
  4. N Van Dillen2,
  5. R A Verheij3,
  6. I E van der Horst-Bruinsma1,
  7. B A C Dijkmans1,2,
  8. M T Nurmohamed1,2,4
  1. 1Department of Rheumatology, VU University Medical Centre, Amsterdam, The Netherlands
  2. 2Department of Rheumatology, Jan van Breemen Institute Amsterdam, Amsterdam, The Netherlands
  3. 3NIVEL (Netherlands Institute for Health Services Research), Utrecht, The Netherlands
  4. 4Department of Internal Medicine, VU University Medical Centre, Amsterdam, The Netherlands
  1. Correspondence to Dr Michael T Nurmohamed, Departments of Internal Medicine and Rheumatology, VU University Medical Centre, PO Box 7057, 1007 MB Amsterdam, Netherlands; mt.nurmohamed{at}


Objective To ascertain the prevalence of myocardial infarction (MI) in ankylosing spondylitis (AS) relative to that in the general population.

Methods A questionnaire was sent to 593 patients with AS, aged between 50 and 75 years and registered at the Jan van Breemen Institute or VU University Medical Centre. A total of 383 (65%) patients with AS returned their questionnaire that covered the primary outcome, (non-fatal) MI. The prevalence of MI was calculated with data from the general population provided by Netherlands Information Network of General Practice databases as reference.

Results The overall prevalence for MI was 4.4% in patients with AS versus 1.2% in the general population, resulting in an age- and gender-adjusted odds ratio of 3.1 (95% CI 1.9 to 5.1) for patients with AS. When non-responders (35%) were considered as non-MI the odds ratio decreased to 1.9 (95% CI 1.2 to 3.2).

Conclusions These observations indicate that the prevalence of MI is increased in patients with AS.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


An excessive risk of cardiovascular (CV) mortality and morbidity has been observed in patients with inflammatory arthritis.1 2 The inflammatory process appears to have an important role in causing this excess CV risk, since it resembles similar processes that contribute to all stages of atherosclerosis, from early atheroma formation to plaque instability and thrombus development responsible for CV events.3 4 Accordingly, the CV risk in ankylosing spondylitis (AS) might also be increased. In an analysis of data from the integrated US health plan, an increased CV disease prevalence was seen in patients with AS in comparison with age- and gender-matched controls.5 Another study observed a higher prevalence of spondyloarthropathies among patients who underwent coronary artery bypass grafting (CABG), and additionally demonstrated that spondyloarthropathy was a stronger predictor of CABG than most traditional CV risk factors.6 However, despite these observations, publications on CV risk in AS are limited.7

Hence, the aim of this paper is to (a) ascertain the prevalence of myocardial infarction (MI) in patients with AS relative to the general population and (b) investigate differences in risk factors between patients with AS with and without a history of MI.


Patients with AS

Between 1 June and the end of September 2007, questionnaires covering the primary outcome MI were sent to all registered patients with AS aged between 50 and 75 years from the Jan van Breemen Institute and VU University Medical Centre in Amsterdam, The Netherlands. Patients not responding were sent a second and, if necessary, a third questionnaire with a 2-month interval.

General population

Data for comparison were used from the Netherlands Information Network of General Practice. Data include information on consultations, morbidity, prescriptions and referrals to other healthcare professionals from general practices. The patients as well as general practices are representative for the Dutch population.8 9 Only people between 50 and 75 years were included. Morbidity data were derived from consultations and all prescriptions issued by the participating practices were used. Diagnoses were recorded using the international classification of primary care (ICPC-1) coding system.10 When issuing a prescription, a diagnostic code was recorded, and the selected drug was automatically linked to the ATC coding system (, 2007 (accessed 27 December 2009)).This means that patients were classified as ‘having a MI event’ if they had a MI event and/or if they needed a prescription linked to a previous MI. Subjects were classified with MI when the diagnosis was recorded in 2006 or in previous years (up to 2004).

Definition of MI

In patients with AS, the ICD-9 code 410 was used to classify MI.11 Reported MIs were verified by chart review, by contacting the patient's general practitioner and/or local hospitals for documentation of the event, and were only used, when confirmatory medical information was available. In the general population, the presence of MI was determined by using the ICPC-1 code K75, which can be converted to the ICD-9 code 410.10

Nested case–control analysis: comparison of patients with AS with and without MI

Additional data on CV risk factors could not be obtained from four of the 17 patients with AS reporting MI as consent was withheld, leaving 13 patients with AS and MI for further investigation. Thirty-nine age- and gender-matched patients with AS were selected to serve as controls. The following AS-related variables were assessed: Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), Bath Ankylosing Spondylitis Functional Index (BASFI) and a history of orthopaedic surgery. A further (physical) examination was performed that included the AS mobility score (Bath Ankylosing Spondylitis Metrology Index (BASMI)), CV risk factors (smoking (pack-years), blood pressure, body mass index, waist to hip ratio, diabetes), and fasting blood samples (glucose, lipids, C-reactive protein) as described elsewhere.12 Hypertension was defined as a systolic blood pressure >140 mm Hg and/or a diastolic blood pressure >90 mm Hg and/or the current use of antihypertensive agents. Diabetes was defined as a documented medical history of diabetes, glucose-lowering agents usage or fasting glucose levels ≥7.0 mmol/l,13 and dyslipidaemia as a total cholesterol/high-density lipoprotein cholesterol ratio >4.0 and/or statin usage.

Statistical analyses

Logistic regression analyses were used to calculate odds ratios and 95% CI (a) for each stratum, adjusted for age, and (b) overall, adjusted for age and gender. First, analyses were done including only patients who returned the questionnaire. Second, to judge the effect of non-responders on the study results, best-case calculations were carried out—that is, including all non-responders and considering them as non-MI. To detect differences in risk factors between patients with AS and MI (case) and those without MI (control) and to deal with matched case–control pairs (for age and gender) conditional logistic regression analyses were performed.


Patients with AS

At conclusion of the protocol, 383 out of 593 (65%) patients with AS had responded. There was no difference in age between patients returning their questionnaire and those who did not, but male patients were less likely to respond than female patients (p<0.05).

MI prevalence in patients with AS versus controls

Table 1 summarises the prevalence rates for MI in AS and in the general population, in strata defined by age and gender. Odds ratios per stratum were higher in both male and female patients with AS. The overall prevalence was 4.4% in AS and 1.2% in the general population. Compared with controls, the odds ratio for a history of MI was approximately threefold higher in AS. Since the response rate of 65% might have resulted in bias, we performed additional analyses including all non-responders and considering them as non-MI. Then, in comparison with controls, the prevalence for MI decreased to 2.9%, resulting in an approximately twofold higher age- and gender-adjusted odds ratio.

Table 1

The prevalence of MI in patients with ankylosing spondylitis versus controls

Differences between patients with AS with and without prevalent MI

The prevalence of hypertension and dyslipidaemia was higher in patients with AS with MI (85% vs 54%, p=0.07 and 92% vs 38%, p=0.02). Patients with AS with MI had a less favourable body mass index and waist to hip ratio (31 kg/m2 vs 28 kg/m2, p=0.08 and 1.00 vs 0.95, p=0.05) and tended to have a higher prevalence of diabetes (23% vs 10%, p=0.24). With regard to smoking, no difference was found in the number of pack-years. Additionally, although most disease activity and severity markers were higher in patients reporting MI, these markers were not significantly different. Finally, we did not find a significant difference in the use of antirheumatic treatment—that is, disease-modifying antirheumatic drugs, non-steroidal anti-inflammatory drugs, or anti-tumour necrosis factor α treatment, between patients with AS with and without a history of MI. A table showing all results is available on the journal website as an online supplemental file.


Although the increased CV risk in patients with inflammatory arthritis is increasingly acknowledged, information about the magnitude of CV morbidity in AS is scarce. In this paper, we demonstrate that the prevalence rate for MI is approximately twofold to threefold increased in comparison with the general population. This observation indicates that AS, like rheumatoid arthritis and psoriatic arthritis, should be considered as an important CV risk factor.1 14 The key feature explaining the increased CV risk in patients with inflammatory arthritis is thought to be inflammation, as it is an intrinsic element of atherogenesis in all stages of its development.15 However, although a non-significant trend was observed which may indicate that patients with AS and a history of MI have a more severe disease, we observed no significant associations between disease activity markers or severity markers and a history of MI. This might be owing to the small number of cases, limiting the statistical power or the cross-sectional nature of this study. Additionally, an increased CV risk in patients with AS may be due to an unfavourable pattern of CV risk factors. Indeed, several risk factors—for example, hypertension, dyslipidaemia and obesity, were highly prevalent in AS and even more in patients with a history of MI. Untreated CV comorbidity is another potential explanation for our observations, and cannot be ruled out as a factor (partially) contributing to the increased CV risk in AS, since we observed a relatively high frequency of uncontrolled hypertension and dyslipidaemia (data not shown). Our findings have clinical implications and substantiate the importance of proper CV risk assessment in AS, in particular, CV risk screening and management.

Like all surveys, ours was subject to potential problems of response bias, since cases might be more interested in participating, which eventually might have led to an overestimation of actual MI rates. However, it should be emphasised that, when non-responders were considered as non-MI, the odds ratio for MI in AS decreased, but remained approximately twofold higher than in controls. Unfortunately, no information was available about CV risk factors, physical activity and social class from the general population, which prevents us from adjusting for these potential confounders. In conclusion, although our observations must be interpreted with caution, as the number of cases was smal, particularly among female subjects, they strengthen the need for larger well-controlled cohort studies to quantify the precise CV risk in patients with AS and to identify those at increased risk.


Supplementary materials

  • Web Only Data ard.2009.110593

    Files in this Data Supplement:


  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the local ethical commission board (Slotervaart Hospital Amsterdam and VU University Medical Centre, Amsterdam).

  • Provenance and peer review Not commissioned; externally peer reviewed.