Regarding ‘Lack of an effective drug therapy for abdominal aortic aneurysm’

We read with great interest a review by Golledge et al. 1 entitled ‘Lack of an effective drug therapy for abdominal aortic aneurysm‘, which was published in Journal of Internal Medicine on July 6, 2019. The review summarized past randomized placebo-controlled drug trials in patients with abdominal aortic aneurysm (AAA). A meta-analysis conducted by authors suggested that antibiotics do not affect the expansion of AAA, and the forest plot was presented in Fig 1 of the paper. Here we would like to make several comments on the methodology of the review. Figure 1 in the article combined the effects of several antibiotics on the growth of AAA. Of included antibiotics, azithromycin and roxithromycin are macrolide antibiotics and doxycycline is tetracycline antibiotic. The potential mechanism how macrolide antibiotics influence the growth of AAA is still debatable. However, experimental studies indicated that tetracycline antibiotic has pleiotropic effects to reduce aortic matrix metalloproteinase activity and limit inflammation of aorta 1, 2. The potential mechanisms of the effects of two antibiotics on AAA could be totally different, which induces a significant clinical heterogeneity in the meta-analysis. Therefore, the final result of the meta-analysis may not be valid and should be interpreted carefully. A subgroup analysis on two antibiotics is needed. Otherwise, the forest plot in Fig. 1 showed a significant and substantial statistical heterogeneity (I2 = 85.6%, P = 0.000). This was caused by clinical heterogeneity and methodologic heterogeneity of included studies. To interpret the results of the meta-analysis and find the heterogeneity source, a sensitivity analysis is needed. A meta-analysis with rigorous methodology could provide high-level evidence on clinical research and clinical practice, and we thank authors to provide a detailed review on drug therapy for AAA. We believe our suggestions and observations will help readers interpret the result of the review and give inspirations to future scientific research. No conflicts of interest were declared.