Table 2

Risk of MI compared between sDMARD and TNFi cohorts

sDMARD; n=3058TNFi; n=11 200
Median duration of follow-up per patient, years (IQR)3.5 (1.8, 4.9)5.3 (3.6, 6.4)
Total person-years of exposure, pyrs10 33755 636
Primary drug exposure model: on-TNFi+90 days
 Number of verified first MIs58194
 Crude incidence rate of verified first MI per 10 000 pyrs (95% CI)56 (43 to 73)35 (30 to 40)
 Unadjusted HR (95% CI)Referent0.78 (0.58 to 1.05)
 HR adjusted for age and gender (95% CI)1.19 (0.89 to 1.59)
 HR after adjusting for PD* (95% CI)0.61 (0.41 to 0.89)
Sensitivity analyses
 In subjects ever exposed to TNFi; PD-adjusted HR (95% CI)0.67 (0.46 to 0.96)
 Trimming the PD at 5%; PD-adjusted HR (95% CI)0.56 (0.34 to 0.93)
  • *Deciles of propensity score (PD). The PD included age, gender, DAS28, disease duration, health assessment questionnaire score, whether the patients used four or more sDMARDs prior to study registration (yes/no), whether the patients were recruited to the register before or after 30 June 2004, hypertension, diabetes, chronic lung disease, smoking (ever/never), antiplatelet therapy, NSAID/COX-2 inhibitor use, glucocorticoid use and statin use.

  • COX, cyclooxygenase inhibitor; DAS28, disease activity in 28 joints; MI, myocardial infarction; NSAID, non-steroidal anti-inflammatory drugs; sDMARD, synthetic disease-modifying antirheumatic drug; TNFi, tumour necrosis factor α inhibitor.