Fluoroquinolone use and risk of aortic aneurysm and dissection: nationwide cohort study
BMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k678 (Published 08 March 2018) Cite this as: BMJ 2018;360:k678
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
In some comments, more clinical data or more experimental data about Fluoroquinolones (FQ) and the aorta are asked for. However, the growing evidence is that FQ are deleterious for patients with Aortic Aneurysm (AA). In this setting, I would like to raise some issues which are directly connected with the care of those patients with an AA who are treated for whatever reason with FQ. The goal is to discuss practical clinical situations while waiting for more evidence or for binding recommendations.
1/ Should patients with known AA be treated with FQ in case of infection elsewhere?
The answer is not in the literature as no prospective study has been completed to date in such a population. But due to the high risk of a fatal issue in case of rupture or fissure, such a pre-existing condition should be searched for by anamnesis, physical exam or through EHR before filling out a prescription for FQ. Only severe sepsis caused by resistant bacteria to other antibiotics than FQ is a balanced risk indication. In all other cases, FQ should be avoided, especially in pneumonia and UTI without antibiogram. This hazard is not well known and a new regulation is waited (https://www.fda.gov/Drugs/DrugSafety/ucm511530.htm) to include this information in FQ label.
2/ What about patients treated with FQ and presenting with a growing AA?
In those patients, the first diagnosis suspected depends on the infectious context. If the patient is febrile with active infection and inflammation, a mycotic aneurysm is usually suspected either a true native AA if sacciform or a bacterial graft on a pre-existent AA. If the patient is apyretic and infection is healed, growth is usually considered to be a consequence of local inflammation in the arterial wall contemporary with the septic episode (https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.117.028433?u...). Indeed the role of FQ is rarely raised, and this bias of reasoning can impede fast and efficient treatment. In case one, considering the AA as mycotic induces a delay to reintroduce antibiotics before the operation and possibly changes the usual treatment for more aggressive and risky techniques. In case two, the increased risk of rupture which persists 60 days (as nicely shown by Pasternak) could be under-evaluated in the absence of FQ incrimination. This strategy based on non-FQ-aggravated AAs is not appropriate as repairing without delay an AA either by the endovascular or open technique is lifesaving if the threshold diameter of the AA is reached.
3/ Can FQ cause an AA in patients with a normal aorta?
It is not possible to answer this question in humans. However, experimental data suggest mechanistic evidence. Pre-existing conditions or peculiar genomics could define a high-risk group but it is too early for any advice except perhaps for patients with Marfan disease.
These three questions and others remain open, but the clinician cannot ignore accumulating data when choices must be made.
Eventually, the different prospective clinical studies on FQ are of great value for clinicians as sadly tendon rupture is still the main well-known side effect of FQ but not the worst.
Competing interests: No competing interests
Following this paper, the MHRA has advised that quinolones should only be used after careful benefit-risk assessment and consideration of other therapeutic options in patients at risk of aortic aneurysm or dissection. However, the considerations made by Dr Stephan Imfeld and Dr Tan Che-Kim are very strong and if and how quinolones may be eventually responsible for the increased risk of dissection remains elusive upon a pathophysiological point of view.
The same applies to their involvement in tendinopathy and tendon rupture, which we know to be far more evident with concurrent corticosteroid exposure.
It is therefore worthwhile considering more in depth those rare events to see what lessons may be learnt from each of them.
Competing interests: No competing interests
Dear Editor
We read with interest the recent study by Pasternak and colleagues on vascular complications from fluoroquinolone use (1). Despite the huge number of patients included, we have serious doubts about the validity of the study results and its implicit causation.
The difference in event rates is so small (0.05%) that even the slightest imbalances in the populations studied could bear major confounding effects. Although the authors meticulously included 47 covariates as predictors in their propensity score matching covering demographic information, medical history, prescription drugs, and healthcare use, they omitted major factors such as indication for the antibiotic prescription and more importantly the site of infection. Having a few percentage points higher abdominal infections (e.g. gastrointestinal or pyelonephritis) in one group could alter (a) the detection rate of abdominal aneurysms by a more careful clinically focused examination, (b) the vulnerability by adjacent inflammatory processes and (c) the chance of a slightly higher misdiagnosis of abdominal symptoms as infectious than an aortic process. As the primary differences were found primarily in the abdominal aorta, the above effects should be considered carefully
Even though the information on side effects of different antibiotics is of utmost interest, we have serious concerns that the results of the study, which could well influence future recommendations by the National Institute of Health on prescription procedures, are heavily affected by the above-mentioned biases, and that the implicit causation is not as strong (if existing at all) as suggested.
(1) BMJ 2018;360:k678
Competing interests: No competing interests
In an article in a recent issue of BMJ on the association between fluoroquinolone (FQ) use and the risk of aortic aneurysm and dissection, Pastermaket al1 reported that FQ use was associated with an increased risk of aortic aneurysm and dissection using amoxicillin as a comparator. However, we have several serious concerns.
When antibiotics are prescribed, there will be increased chances of a physician to prescribe an image study to survey for the infection source. Therefore there will probably be more aortic aneurysms that were incidentally discovered, and even more after usage of 2nd/3rd line antibiotics. Hence the authors need to clarify that the increased aortic aneurysms associated with the usage of FQ is not related to more utilization of image studying. Both this study1 and Daneman et al‘s study2 failed to demonstrate that amoxicillin is a negative tracer, ie. the usage of amoxicillin was associated with increase of aortic aneurysms too, though to a lesser extent than FQ.
In addition, the indication and the use of FQ should be largely different from amoxicillin. In contrast to amoxicillin, FQ is a broad-spectrum antibiotic, and ciprofloxacin – the most common FQ in this study can be indicated in complicated intra-abdominal infection, infectious diarrhea and complicated urinary tract infection. Therefore, it should be expected that FQ users are more ill than amoxicillin users, and FQ users may have more image studying, especially abdominal examinations than amoxicillin users. Moreover, the risk being highest during the first 10 days from the start of FQ use may be due to more examinations having been performed during the initial treatment period. All of these things suggest that more incidental aortic aneurysms can be found in FQ users than amoxicillin users, and help to explain why the most common type of aortic aneurysm is abdominal aneurysm among FQ users in this study.
Furthermore, it is supposed that after excluding the cases whose aortic aneurysm is incidentally diagnosed, the difference between FQ and amoxicillin users may disappear. It may help explain why the significant association is largely drive by aortic aneurysm in this study. This hypothesis can be demonstrated in the sensitivity analysis in this study. Regarding cases with dissection or rupture alone, which should cause severe discomfort, the difference did not reach statistical significance.
References
1. Pasternak B, Inghammar M, Svanström H. Fluoroquinolone use and risk of aortic aneurysm and dissection: nationwide cohort study. BMJ 2018;360:k678. doi: 10.1136/bmj.k678.
2. Daneman N, Lu H2, Redelmeier DA. Fluoroquinolones and collagen associated severe adverse events: a longitudinal cohort study. BMJ Open. 2015;5(11):e010077. doi: 10.1136/bmjopen-2015-010077.
Competing interests: No competing interests
Re: Fluoroquinolone use and risk of aortic aneurysm and dissection: nationwide cohort study
We appreciate the effort Drs Pasternak and others have put into carefully researching the important topic of the risk of aortic aneurysm associated with fluoroquinolone use. This is an important topic that is relevant to physicians in many specialties.
The results of this analysis have been widely quoted and so we feel that the authors should clarify their statistical methods. In the article they state that 64 cases of aneurysm were seen on the exposed group and 40 in the control group. Since both groups contained 360,088 participants. The excess rate is 24 cases or 67 per million not 82 as stated. The printed confidence interval is 15-181 which is not symmetric with 82.
For the cumulative incidence calculation for fluoroquinolone there were 64 cases per 360088 which is 1.7 x10-4 not 2.0 as written. The amoxicillin group had 40 per 360088 or 1.1 x10-4. This was written as 1.2.
We hope that the statistical calculations can be verified and corrected if needed. We would leave it to others to determine the clinical significance of an excess of 67 cases per million exposures to the fluoroquinolones.
Competing interests: No competing interests