When relative and absolute risks disagree (Part 2 of 3)
Click here to see part 1
When did your doctor last prescribe you a quinolone antibiotic? My last was in 2017, which was before the FDA warning about aortic aneurysm, but after warnings about severe nerve damage. Now I’m curious: Next time I get a urinary tract infection, what will my doctor prescribe…?
For the whole picture, make sure you consider both relative and absolute risks
Our journal club class recently reviewed a cohort study of the risk of abdominal aortic aneurysm associated with quinolone antibiotics.1 As we discussed the article, we observed a nice illustration of the phenomenon that absolute and relative risks often disagree. In this study, the overall finding was that quinolones increase the risk of aortic aneurysm by 66%. In subgroup analyses, the study authors reported the results separately for males and females, shown in the figure. Interestingly, the relative risk showed that the risk was more than doubled in women (a 114% increase), a finding that was statistically significant. In men, the relative risk was only increased 48% in patients who received quinolones.
However, when absolute risk was considered, we saw that the risk difference associated with quinolones was smaller for women than for men. For women, quinolones only contributed to an increase of 0.4 events per 1000 patient-years. For men, the increase was 0.6.
The entire discrepancy is driven by the fact that men had a higher baseline risk, and so the larger increase in absolute risk was really only a 48% increase over baseline. Women have a lower baseline risk, so the smaller increase in absolute risk actually constituted a doubling of her risk.
They did not do a hypothesis test on the risk differences — only on the relative risks. It was interesting that the larger absolute difference in men failed to reach significance in the hypothesis test of the relative difference, while the smaller absolute difference in women was significant in the hypothesis test of the relative difference.
- Pasternak B, Inghammer M, Svanström H. Fluoroquinolone use and risk of aortic aneurysm and dissection: nationwide cohort study. British Medical Journal. 2018;360:k678.