Fig. 2. Survival analysis for the unadjusted population. The in-hospital mortality up to 30 days is represented in 4 groups; (1) the population in whom there was a positive troponin (n = 119) is shown in the bottom line unbroken black line  (troponin I ≥ 0.7 μg/l). (2) The cohort in whom troponin I was < 0.07 μg/l and > 0.02 μg/l (n = 96) is shown in the dotted line . (3) The population in whom measured troponin I was not detected (n = 1,176) is shown in the thick dotted line . (4) The population in whom troponin was not measured is shown in the top thick solid line . This analysis shows that there is increased mortality in patients where troponin was ordered but not detected compared with patients in whom troponin was not  ordered (relative risk 2.4; 95% confidence interval 1.8–3.4).