Association of Infections and Use of Fluoroquinolones With the Risk of Aortic Aneurysm or Aortic Dissection

Yaa-Hui Dong; Chia-Hsuin Chang; Jiun-Ling Wang; Li-Chiu Wu; Jou-Wei Lin; Sengwee Toh

doi:10.1001/jamainternmed.2020.4192

. 2020 Sep 8;180(12):1587–1595. doi: 10.1001/jamainternmed.2020.4192

Association of Infections and Use of Fluoroquinolones With the Risk of Aortic Aneurysm or Aortic Dissection

Yaa-Hui Dong ^1,^2,^✉, Chia-Hsuin Chang ^3,^4,^5,^✉, Jiun-Ling Wang ^6,⁷, Li-Chiu Wu ³, Jou-Wei Lin ^4,^8,⁹, Sengwee Toh ¹⁰

¹Faculty of Pharmacy, National Yang-Ming University School of Pharmaceutical Science, Taipei, Taiwan

²Institute of Public Health, National Yang-Ming University School of Medicine, Taipei, Taiwan

³Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan

⁴Department of Medicine, National Taiwan University College of Medicine, Taipei, Taiwan

⁵Institute of Epidemiology and Preventive Medicine, National Taiwan University College of Public Health, Taipei, Taiwan

⁶Department of Internal Medicine, National Cheng Kung University Hospital, Tainan, Taiwan

⁷Department of Medicine, National Cheng Kung University Medical College, Tainan, Taiwan

⁸Department of Internal Medicine, National Taiwan University Hospital Yunlin Branch, Douliou City, Yunlin County, Taiwan

⁹Cardiovascular Center, National Taiwan University Hospital Yunlin Branch, Douliou City, Yunlin County, Taiwan

¹⁰Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts

Accepted for Publication: April 27, 2020.

^✉

Corresponding Authors: Chia-Hsuin Chang, MD, ScD, Department of Internal Medicine, National Taiwan University Hospital, Seven Chung-Shan S Rd, Taipei City, 100 Taiwan (chiahsuin@yahoo.com); Yaa-Hui Dong, PhD, Faculty of Pharmacy, National Yang-Ming University School of Pharmaceutical Science, 155 Sec 2 Linong St, Taipei City, 112 Taiwan (yaahuidong@gmail.com).

Published Online: September 8, 2020. doi:10.1001/jamainternmed.2020.4192

Author Contributions: Drs Dong and Chang had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Dong, Chang, Wang, Lin, Toh.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Dong, Chang, Wu, Lin, Toh.

Critical revision of the manuscript for important intellectual content: Dong, Chang, Wang, Lin, Toh.

Statistical analysis: Dong, Chang, Wu, Lin, Toh.

Obtained funding: Dong, Chang, Lin.

Administrative, technical, or material support: Chang, Wang, Lin.

Supervision: Chang, Lin, Toh.

Conflict of Interest Disclosures: Dr Chang reported receiving grants from the Ministry of Science and Technology, Taiwan, during the conduct of the study and outside the submitted work. No other disclosures were reported.

Funding/Support: The study was partly supported by research grant 107F016-3 from National Yang-Ming University, grant YLH106.A001 from National Taiwan University Hospital Yunlin Branch, and grant CI-109-28 from Yen Tjing Ling Medical Foundation.

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

^✉

Corresponding author.

PMCID: PMC7489369 PMID: 32897358

This case-control study estimates the risk of aortic aneurysm or aortic dissection associated with infections and the comparative risk associated with fluoroquinolones vs other antibiotics with similar indication profiles among patients with the same types of infections.

Key Points

Question

Is the risk of aortic aneurysm or aortic dissection independently associated with infections and fluoroquinolone use (vs other antibiotics with similar indication profiles)?

Findings

In this nationwide, nested case-control study of 28 948 cases and 289 480 matched controls identified from 21 651 176 adult patients, the odds ratio (OR) of aortic aneurysm or aortic dissection for any indicated infections, adjusted for baseline confounders and concomitant antibiotic use, was 1.73 (95% CI, 1.66-1.81). Fluoroquinolones were not associated with an increased risk of aortic aneurysm or aortic dissection when compared with amoxicillin-clavulanate or ampicillin-sulbactam (OR, 1.01; 95% CI, 0.82-1.24) or extended-spectrum cephalosporins (OR, 0.88; 95% CI, 0.70-1.11) among patients with indicated infections.

Meaning

These results highlight the importance of accounting for coexisting infections while examining the safety of antibiotics using real-world data; the concern about aortic aneurysm or aortic dissection should not deter fluoroquinolone use for patients with indicated infections.

Abstract

Importance

Prior observational studies have suggested that fluoroquinolone use may be associated with more than 2-fold increased risk of aortic aneurysm or aortic dissection (AA/AD). These studies, however, did not fully consider the role of coexisting infections and the risk of fluoroquinolones relative to other antibiotics.

Objective

To estimate the risk of AA/AD associated with infections and to assess the comparative risk of AA/AD associated with fluoroquinolones vs other antibiotics with similar indication profiles among patients with the same types of infections.

Designs, Settings, and Participants

This nested case-control study identified 21 651 176 adult patients from a nationwide population-based health insurance claims database from January 1, 2009, to November 30, 2015. Each incident case of AA/AD was matched with 10 control individuals by age, sex, and follow-up duration in the database using risk-set sampling. Analysis of the data was conducted from April 2019 to March 2020.

Exposures

Infections and antibiotic use within a 60-day risk window before the occurrence of AA/AD.

Main Outcomes and Measures

Conditional logistic regression was used to estimate the odds ratios (ORs) and 95% CIs comparing infections for which fluoroquinolones are commonly used with no infection within a 60-day risk window before outcome occurrence, adjusting for baseline confounders and concomitant antibiotic use. The adjusted ORs comparing fluoroquinolones with antibiotics with similar indication profiles within patients with indicated infections were also estimated.

Results

A total of 28 948 cases and 289 480 matched controls were included (71.37% male; mean [SD] age, 67.41 [15.03] years). Among these, the adjusted OR of AA/AD for any indicated infections was 1.73 (95% CI, 1.66-1.81). Septicemia (OR, 3.16; 95% CI, 2.63-3.78) and intra-abdominal infection (OR, 2.99; 95% CI, 2.45-3.65) had the highest increased risk. Fluoroquinolones were not associated with an increased AA/AD risk when compared with combined amoxicillin-clavulanate or combined ampicillin-sulbactam (OR, 1.01; 95% CI, 0.82-1.24) or with extended-spectrum cephalosporins (OR, 0.88; 95% CI, 0.70-1.11) among patients with indicated infections. The null findings for fluoroquinolone use remained robust in different subgroup and sensitivity analyses.

Conclusions and Relevance

These results highlight the importance of accounting for coexisting infections while examining the safety of antibiotics using real-world data; the findings suggest that concerns about AA/AD risk should not deter fluoroquinolone use for patients with indicated infections.

Introduction

Aortic aneurysm (AA) and aortic dissection (AD) are potentially fatal conditions. Population-based studies in the United States, European countries, and Taiwan estimated the annual incidence to be 2.4 to 14.8 per 100 000 persons for AA^1,2,3,4 and 3.8 to 8.8 per 100 000 persons for AD.^3,5,6,7 Although the incidence varied across countries, the number has universally increased over time.^1,2,3,4,5,7 Without appropriate treatment of ruptured AA/AD, mortality can increase to 90%.⁸

The known risk factors for AA/AD include congenital connective tissue disorders, older age, male sex, atherosclerotic cardiovascular disease, and tobacco smoking.^8,9,10,11 Cumulative case reports and case series^{12,13,14,15,16,17} also suggested that endocarditis, septicemia, intra-abdominal infections, bone-related infections, genitourinary tract infections (GUTIs), and lower respiratory tract infections (LRTIs) may be associated with AA/AD.

Past epidemiological studies^18,19,20,21 observed more than 2-fold increased risk of AA/AD with oral fluoroquinolones that are commonly used in treating LRTIs and GUTIs.²² These findings prompted the US Food and Drug Administration²³ and the European Medicines Agency²⁴ to issue safety warnings about fluoroquinolones. However, these studies did not fully consider the role of infections on the risk of AA/AD.^18,19,20,21 Most studies compared use vs nonuse of oral fluoroquinolones,^18,19,20 which could be susceptible to confounding by indication and overestimate the risk with fluoroquinolones because patients receiving fluoroquinolones may have different infection types or severity compared with those not receiving fluoroquinolones. Using a nested case-control study design in a nationwide population-based database, we estimated the risk of AA/AD with infections and the comparative risk of AA/AD with fluoroquinolones vs other antibiotics with similar indication profiles.

Methods

Data Source and Source Population

We used data from the Taiwan National Health Insurance Research Database, which included deidentified data of 23 million individuals covered by a national health insurance system.^25,26 Our source population consisted of patients 20 years or older who entered the cohort from January 1, 2009, to November 30, 2015. The cohort entry date was the date patients reached 20 years of age. We excluded patients with ambiguous sex information, history of AA/AD (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM], code 441), or any congenital disorders that potentially predisposed them to AA/AD.^8,9,10,11 We followed up patients from cohort entry to the earliest of AA/AD occurrence (defined below), death, or November 30, 2015 (see eMethods and eFigure 1 in the Supplement for detail). The National Yang-Ming University Research Ethics Committee approved the study and waived the need for informed consent for the use of deidentified data. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Selection of Cases and Controls

We defined the index date as the date of the first hospital admission or emergency department visit for AA/AD based on the ICD-9-CM code 441 in any diagnosis position for the cases (see eTable 1 in the Supplement for codes). The algorithm had positive predictive values of 89% to 100% for AA and 78% to 92% for AD based on prior studies.^27,28 We determined how the cases were clinically managed by identifying imaging examinations and treatments for AA/AD within 30 days before and after the index date (see eTable 2 in the Supplement for codes). We also determined whether the diagnosis was in the primary diagnosis position, which indicated that AA/AD was the main reason for the encounter.

We used risk-set sampling to identify controls among patients who remained free of the outcome at the time a case occurred. For each case, we randomly sampled as many as 10 controls matched on birth year, sex, and follow-up duration from cohort entry to the index date. The index date for the controls was the same as the index date for the case to which they were matched.

Ascertainment of Infections and Antibiotic Use

As in prior studies,^19,20 we defined the risk window as 1 to 60 days before the index date. For the cases and matched controls, we identified inpatient and outpatient infection episodes that were potential indications for fluoroquinolones (termed indicated infections), including LRTIs only; GUTIs only; skin, soft tissue, and bone infections only; intra-abdominal infections only; or any combinations of aforementioned infections. We further included septicemia, which is also an indication for fluoroquinolone treatment in Taiwan and included infections without an identified source of infection. We classified the remaining patients as having no indicated infection. The positive predictive values of the algorithms were 70% to 97% for LRTIs; 73% to 100% for GUTIs; 74% to 92% for skin, soft tissue, and bone infections; 77% to 84% for intra-abdominal infections; and 80% to 100% for septicemia based on previous studies (see eTable 3 in the Supplement for codes).^{29,30,31,32,33,34,35,36,37,38,39,40,41,42,43}

We used inpatient and outpatient pharmacy dispensing claims to identify oral or injectable antibiotic use within the risk window. As in prior studies,^19,20 we defined patients as being exposed to antibiotics if at least a 3-day supply of antibiotics was available within the window. We classified antibiotics into fluoroquinolones, combined amoxicillin-clavulanate or combined ampicillin-sulbactam, extended-spectrum cephalosporins (second-, third-, or fourth-generation cephalosporins), and miscellaneous antibiotics (see eTable 4 in the Supplement for codes). We selected amoxicillin-clavulanate or ampicillin-sulbactam and extended-spectrum cephalosporins as comparison antibiotics because their indication profiles are similar to that of fluoroquinolones based on the treatment guidelines in Taiwan (see eMethods in the Supplement for detail).^44,45,46 Because patients could receive more than 1 antibiotic class, we further classified antibiotic exposure into mutually exclusive groups of fluoroquinolone monotherapy, comparison antibiotic monotherapy, other antibiotic regimens (any combination therapies of fluoroquinolones and comparison antibiotics or use of miscellaneous antibiotics), or no use of any antibiotics.

Ascertainment of Baseline Covariates

We identified potential baseline confounders between cohort entry and 60 days before the index date. These included individual comorbidities, Charlson comorbidity score (range, 0-33, with higher scores indicating greater number of comorbidities),⁴⁷ tobacco smoking,⁴⁸ a claims-based frailty index,⁴⁹ any infection episodes defined above, and any use of fluoroquinolones or nonfluoroquinolone antibiotics for at least 1 day (see eTable 5 in the Supplement for codes).

Statistical Analysis

Analysis of the data was conducted from April 2019 to March 2020. We used conditional logistic regression to estimate the odds ratios (ORs) and 95% CIs for the association between infections and AA/AD. The first model only adjusted for the matching factors. The second model additionally adjusted for the potential baseline confounders. The third model further included any use of fluoroquinolones, comparison antibiotics, and other antibiotics for at least 3 days during the risk window to adjust for concomitant antibiotic use and to simultaneously estimate the effects of infections and antibiotics.

To further reduce confounding by infection and infection severity on the association between fluoroquinolone use and risk of AA/AD, we performed another set of prespecified analyses restricted to patients with an indicated infection and using an active comparison approach. Specifically, we reidentified and rematched cases and controls on age, sex, follow-up duration, and infection type among patients with an indicated infection. We used 2 conditional logistic regression models to estimate the ORs and 95% CIs associated with fluoroquinolone monotherapy compared with amoxicillin-clavulanate or ampicillin-sulbactam monotherapy and extended-spectrum cephalosporin monotherapy; one model adjusted for the matching factors only, and the other adjusted for both matching factors and potential baseline confounders.

Subgroup and Sensitivity Analysis

We conducted additional analyses to examine the robustness of our findings for fluoroquinolones vs comparison antibiotics among patients with an indicated infection. Specifically, we (1) evaluated a possible duration-response association for fluoroquinolones; (2) mitigated potential spillover effects of fluoroquinolones initiated before the risk window; (3) evaluated the effects of the choice of risk window and minimum antibiotic treatment duration; (4) reduced potential reverse causation between fluoroquinolone use and AA/AD occurrence (see eFigure 2 in the Supplement for conceptual temporality of events); (5) mitigated potential confounding due to infection severity by classifying antibiotic use by treatment setting, dosage form, and cephalosporin generation; (6) examined potential effect measure modification by infection type and patient characteristic; (7) examined whether the risk varied by AA/AD subtype; (8) examined the effect of the operational definition of AA/AD; and (9) mitigated the concern about misdiagnosis of AA as LRTIs or GUTIs. eTable 6 in the Supplement gives additional details of each analysis.

Examination of Positive Control Outcomes

Previous studies showed an association between fluoroquinolones and tendon disorders, especially for Achilles tendon rupture or in older patients.^50,51 Therefore, we selected Achilles tendon rupture and any type of tendon rupture as positive control outcomes, restricted the analyses to older patients if the sample size was sufficient, and examined whether our active comparison approach could identify an increased risk with fluoroquinolones. Details are given in the eMethods and eTable 7 in the Supplement.

Results

Eligible Cases and Controls

Among 21 651 176 eligible patients, 28 948 cases and 289 480 matched controls were included in the analysis (71.37% male; 28.63% female; mean [SD] age, 67.41 [15.03] years; mean [SD] follow-up duration, 1303.82 [723.12] days). Among the cases, 88.89% received an imaging examination, 74.51% received treatment intervention, and 52.59% had a primary diagnosis (eTable 8 in the Supplement). The cases had more comorbidities (mean [SD] Charlson comorbidity score, 2.27 [2.41] vs 1.24 [2.00]), tended to be smokers (1.79% vs 0.87%), had a higher claims-based frailty index (mean [SD], 0.17 [0.05] vs 0.14 [0.05]), and had more prior infections (52.79% vs 34.22%) and antibiotic use (fluoroquinolones, 19.50% vs 10.24%; nonfluoroquinolone antibiotics: 75.35% vs 53.63%) compared with matched controls (Table 1).

Table 1. Distribution of Baseline Covariates Between Cases and Matched Controls of AA or AD.

Characteristic	Study group^a		Adjusted for matching factors, OR (95% CI)
Characteristic	Cases (n = 28 948)	Matched controls (n = 289 480)	Adjusted for matching factors, OR (95% CI)
Demographic characteristics
Age at cohort entry, mean (SD), y	67.41 (15.03)	67.41 (15.03)	NA
Male	20 661 (71.37)	206 610 (71.37)	NA
Duration of follow-up, mean (SD), d	1303.82 (723.13)	1303.82 (723.12)	NA
Comorbidities and prior infections^b
Hypertension	19 979 (69.02)	105 055 (36.29)	4.98 (4.84-5.13)
Ischemic heart disease	9793 (33.83)	42 215 (14.58)	3.26 (3.17-3.36)
Valve disorder	3406 (11.77)	10 742 (3.71)	3.55 (3.41-3.70)
Ischemic stroke	4287 (14.81)	18 477 (6.38)	2.67 (2.57-2.77)
Disorders of lipid metabolism	9559 (33.02)	62 917 (21.73)	1.89 (1.84-1.95)
COPD	6520 (22.52)	31 126 (10.75)	2.56 (2.49-2.65)
Chronic kidney disease	3546 (12.25)	12 529 (4.33)	3.22 (3.10-3.36)
Charlson comorbidity score, mean (SD)^c	2.27 (2.41)	1.24 (2.00)	1.23 (1.22-1.24)
Tobacco smoking	519 (1.79)	2513 (0.87)	2.11 (1.92-2.32)
Claims-based frailty index, mean (SD)^d	0.17 (0.05)	0.14 (0.05)	3.79 (3.70-3.88)^e
Any episodes of infections^f	15 282 (52.79)	99 059 (34.22)	2.43 (2.36-2.49)
Antibiotic use^b
Fluoroquinolones	5646 (19.50)	29 635 (10.24)	2.22 (2.15-2.29)
Nonfluoroquinolone antibiotics	21 812 (75.35)	155 251 (53.63)	3.48 (3.37-3.60)

Open in a new tab

Abbreviations: AA, aortic aneurysm; AD, aortic dissection; COPD, chronic obstructive pulmonary disease; NA, not applicable; OR, odds ratio.

^{^a}

Cases and controls were matched on age, sex, and follow-up duration in the database. Unless otherwise indicated, data are expressed as number (percentage) of participants.

^{^b}

Measured between cohort entry and 60 days before the index date.

^{^c}

Scores range from 0 to 21, with higher scores indicating more comorbidities.

^{^d}

Scores range from 0.05 to 0.46, with higher scores indicating a more vulnerable state.

^{^e}

The OR was presented as per 0.1 increase in the claims-based frailty index.

^{^f}

Includes lower respiratory tract infection; genitourinary tract infection; skin, soft tissue, and bone infection; intra-abdominal infection; and septicemia.

Risk of AA/AD Associated With Infections

A total of 5391 cases (18.62%) and 17 084 matched controls (5.90%) had an episode of an indicated infection during the risk window. Lower respiratory tract infections (1511 [5.22%] in cases and 3891 [1.34%] in controls) and GUTIs (1665 [5.75%] in cases and 5663 [1.96%] in controls) accounted for most of the episodes. The OR of AA/AD comparing indicated infections with no indicated infection was 3.69 (95% CI, 3.57-3.81) when we only adjusted for the matching factors (Table 2). The OR lowered to 2.27 (95% CI, 2.19-2.36) after further adjustment for all baseline covariates.

Table 2. Risk of AA or AD Associated With Indicated vs No Indicated Infections.

Infection type	No. (%) of participants		OR (95% CI)
Infection type	Cases (n = 28 948)^a	Matched controls (n = 289 480)^a	Adjusted for matching factors^a	Adjusted for matching factors and baseline covariates^a^,^b	Adjusted for matching factors, baseline covariates, and concomitant antibiotic use^a^,^b^,^c
Any indicated infections	5391 (18.62)	17 084 (5.90)	3.69 (3.57-3.81)	2.27 (2.19-2.36)	1.73 (1.66-1.81)
Specific indicated infections
LRTI	1511 (5.22)	3891 (1.34)	4.54 (4.27-4.83)	2.78 (2.60-2.97)	2.11 (1.96-2.27)
GUTI	1665 (5.75)	5663 (1.96)	3.46 (3.27-3.66)	2.20 (2.07-2.34)	1.77 (1.66-1.89)
Skin, soft tissue, or bone infections	1049 (3.62)	5301 (1.83)	2.29 (2.14-2.46)	1.51 (1.41-1.62)	1.27 (1.18-1.36)
Intra-abdominal infections	174 (0.60)	346 (0.12)	5.92 (4.93-7.11)	3.99 (3.29-4.85)	2.99 (2.45-3.65)
Mixed infections^d	759 (2.62)	1516 (0.52)	5.88 (5.38-6.43)	2.93 (2.66-3.22)	1.75 (1.57-1.95)
Septicemia	233 (0.80)	367 (0.13)	7.41 (6.28-8.74)	4.29 (3.59-5.12)	3.16 (2.63-3.78)
No indicated infection	23 557 (81.38)	272 396 (94.10)	1 [Reference]	1 [Reference]	1 [Reference]

Open in a new tab

Abbreviations: AA, aortic aneurysm; AD, aortic dissection; GUTI, genitourinary tract infection; LRTI, lower respiratory tract infection; OR, odds ratio.

^{^a}

Matched on age, sex, and follow-up duration in the database.

^{^b}

Baseline covariates in the model included hypertension, ischemic heart disease, valve disorder, ischemic stroke, disorders of lipid metabolism, chronic obstructive pulmonary disease, chronic kidney disease, Charlson comorbidity score, tobacco smoking, claims-based frailty index, any episodes of infections, any use of fluoroquinolones, and any use of nonfluoroquinolone antibiotics measured between cohort entry and 60 days before the index date.

^{^c}

Concomitant antibiotic use in the model included any use of fluoroquinolones, amoxicillin-clavulanate or ampicillin-sulbactam, extended-spectrum cephalosporins, and other antibiotics in the risk window.

^{^d}

Include any combinations of LRTIs; GUTIs; skin, soft tissue, and bone infections; and intra-abdominal infections.

The OR further attenuated but remained elevated after additional adjustment for concomitant antibiotic use (1.73; 95% CI, 1.66-1.81). Septicemia was associated with the highest risk of AA/AD (OR, 3.16; 95% CI, 2.63-3.78), followed by intra-abdominal infections (OR, 2.99; 95% CI, 2.45-3.65), LRTIs (OR, 2.11; 95% CI, 1.96-2.27), GUTIs (OR, 1.77; 95% CI, 1.66-1.89), mixed infections (OR, 1.75; 95% CI, 1.57-1.95), and skin, soft tissue, or bone infections (OR, 1.27; 95% CI, 1.18-1.36). The adjusted ORs ranged from 1.32 (95% CI, 1.27-1.37) to 1.43 (95% CI, 1.32-1.54) in the same model comparing the use of a specific antibiotic vs nonuse of that antibiotic (eTable 9 in the Supplement).

Risk of AA/AD Associated With Fluoroquinolone Use Among Patients With Indicated Infections

We identified 5391 cases and 53 880 matched controls with the same age, sex, follow-up duration, and infection type. Compared with all the cases and matched controls (Table 1), the distributions of baseline covariates between cases and matched controls with indicated infections were more similar, although the cases still tended to have more disease burden than the controls (eg, the mean [SD] Charlson comorbidity score was 3.17 [2.65] for the cases and 2.98 [2.70] for the controls) (eTable 10 in the Supplement).

Approximately 60% of the cases and controls received more than 2 classes of antibiotics during the risk window; only 200 cases (3.71%) and 1679 controls (3.12%) received fluoroquinolone monotherapy; 221 cases (4.10%) and 1823 controls (3.38%) received amoxicillin-clavulanate or ampicillin-sulbactam monotherapy; and 145 cases (2.69%) and 1071 controls (1.99%) received extended-spectrum cephalosporin monotherapy. The OR adjusting for matching factors and baseline covariates was 1.01 (95% CI, 0.82-1.24) comparing fluoroquinolone monotherapy with amoxicillin-clavulanate or ampicillin-sulbactam monotherapy and 0.88 (95% CI, 0.70-1.11) comparing fluoroquinolone monotherapy with extended-spectrum cephalosporin monotherapy (Table 3).

Table 3. Risk of AA or AD Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections^a.

Variable by duration of use	No. (%) of participants		OR (95% CI)
Variable by duration of use	Cases (n = 5391)^b	Matched controls (n = 53 880)^b	Adjusted for matching factors^b	Adjusted for matching factors and baseline covariates^b^,^c
Fluoroquinolones (≥3 d)	200 (3.71)	1679 (3.12)	0.98 (0.80-1.20)	1.01 (0.82-1.24)
3-7 d	111 (2.06)	843 (1.56)	1.08 (0.85-1.38)	1.13 (0.88-1.45)
8-14 d	39 (0.72)	440 (0.82)	0.73 (0.51-1.05)	0.74 (0.52-1.06)
>14 d	50 (0.93)	396 (0.73)	1.04 (0.75-1.45)	1.04 (0.75-1.45)
Amoxicillin-clavulanate or ampicillin-sulbactam (≥3 d)	221 (4.10)	1823 (3.38)	1 [Reference]	1 [Reference]
Fluoroquinolones (≥3 d)	200 (3.71)	1679 (3.12)	0.88 (0.70-1.10)	0.88 (0.70-1.11)
3-7 d	111 (2.06)	843 (1.56)	0.97 (0.74-1.26)	0.99 (0.76-1.29)
8-14 d	39 (0.72)	440 (0.82)	0.65 (0.45-0.95)	0.65 (0.44-0.94)
>14 d	50 (0.93)	396 (0.73)	0.93 (0.66-1.31)	0.91 (0.64-1.28)
Extended-spectrum cephalosporins (≥3 d)	145 (2.69)	1071 (1.99)	1 [Reference]	1 [Reference]

Open in a new tab

Abbreviations: AA, aortic aneurysm; AD, aortic dissection; OR, odds ratio.

^{^a}

Data for other antibiotics (3265 cases and 30 153 matched controls) and nonuse of any prespecified antibiotics (1560 cases and 19 154 matched controls) are not shown but were considered in the regression model for proper estimation of ORs of fluoroquinolones vs comparison antibiotics.

^{^b}

Cases and controls were matched on age, sex, follow-up duration in the database, and type of infection.

^{^c}

Baseline covariates in the model included hypertension, ischemic heart disease, valve disorder, ischemic stroke, disorders of lipid metabolism, chronic obstructive pulmonary disease, chronic kidney disease, Charlson comorbidity score, tobacco smoking, claims-based frailty index, any episodes of infections, any use of fluoroquinolones, and any use of nonfluoroquinolone antibiotics measured from cohort entry to 60 days before the index date.

Subgroup and Sensitivity Analysis

We did not observe a duration-response association comparing fluoroquinolones with comparison antibiotics (Table 3). The analysis restricted to patients without any fluoroquinolone use at baseline did not show an increased risk with fluoroquinolones (eFigure 3 and eTable 11 in the Supplement). Use of different risk windows (1-60 or 1-30 days) and minimum antibiotic treatment durations (3 or 1 day) yielded similar results (eFigure 4 and eTable 12 in the Supplement). The ORs moved toward the null when we ignored antibiotic use within 1 to 3 days before the index date (eFigure 4 and eTable 12 in the Supplement).

We did not observe an increased risk with fluoroquinolones by treatment setting (eFigure 5 and eTable 13 in the Supplement), dosage form (eFigure 6 and eTable 14 in the Supplement), or cephalosporin generation (eFigure 7 and eTable 15 in the Supplement). The results did not change materially when we stratified by infection type (eFigure 8 and eTable 16 in the Supplement) or patient characteristics (Figure and eTable 17 in the Supplement). We also did not observe an increased risk with fluoroquinolones in different AA/AD subtypes (eFigure 9 and eTable 18 in the Supplement) or with more specific outcome definitions (eFigure 10 and eTable 19 in the Supplement). The analysis comparing injectable fluoroquinolones vs injectable third- or fourth-generation cephalosporins also showed a null association (eFigure 7 and eTable 15 in the Supplement).

Findings of Positive Control Outcomes

Fluoroquinolone use was associated with a numerically increased risk of Achilles tendon rupture vs amoxicillin-clavulanate or ampicillin-sulbactam (OR, 1.56; 95% CI, 0.56-4.36) or vs extended-spectrum cephalosporins (OR, 2.33; 95% CI, 0.60-9.07) in adult patients. The limited sample size precluded an analysis restricted to older patients. Fluoroquinolone use was also associated with a higher risk of any type of tendon rupture compared with either amoxicillin-clavulanate or ampicillin-sulbactam (OR, 1.13; 95% CI, 0.72-1.77) or extended-spectrum cephalosporins (OR, 2.04; 95% CI, 1.08-3.84) in elderly patients (eFigure 11 and eTable 20 in the Supplement).

Discussion

This nationwide, nested case-control study examined the risk of AA/AD associated with infections and antibiotics. Infection was found to be a risk factor for AA/AD after we adjusted for baseline covariates and concomitant antibiotic use. In contrast, fluoroquinolones were not associated with an elevated risk of AA/AD vs antibiotics with similar indication profiles after accounting for coexisting infections. The null findings did not change materially in different subgroup and sensitivity analyses.

Association Between Infections and AA/AD

Infection has been suspected as a risk factor for AA for decades. Salmonella, Staphylococcus, and Streptococcus species are common microorganisms.^{13,14,15,16,17,52,53,54,55,56} Septic emboli, bacteremic seeding into the arterial wall, and infections from adjacent surroundings of the aorta are potential causes.^57,58,59 During an infection episode, patients may have hemodynamic instability, impaired immunity, and systemic inflammation.^57,58,59 Bacteremia may also produce collagenases that may break down aorta integrity.^57,58,59 These common mechanisms may explain why we observed increased risk of AA associated with several indicated infections, although the magnitudes vary with infection type and severity. To our knowledge, our study is the first to quantify the magnitude of risk of AA/AD with various infections.

Association Between Fluoroquinolones and AA/AD

Several observational studies reported an elevated risk of AA/AD with oral fluoroquinolone use. Daneman et al¹⁸ found that fluoroquinolone use (vs nonuse) was associated with an increased AA risk in a cohort study of elderly Canadian patients (hazard ratio, 2.24; 95% CI, 2.02-2.49). Lee et al^19,20 analyzed a random sample of 1 million individuals from the same database used in the current study using various designs and observed an elevated risk of AA/AD for fluoroquinolone use vs nonuse; the OR was 2.28 (95% CI, 1.67-3.13) in the nested case-control design, 2.71 (95% CI, 1.14-6.46) in the case-crossover design, and 3.61 (95% CI, 3.56-3.63) in the case-time-control design. In a Swedish cohort study with an active comparator design, Pasternak et al²¹ reported a greater risk of AA/AD when comparing fluoroquinolones with amoxicillin (hazard ratio, 1.66; 95% CI, 1.22-2.46). These studies might not have adequately adjusted for the effect of coexisting infections. The nonuser comparison approach is susceptible to confounding by indication^60,61,62 because patients treated with fluoroquinolones may have different infection types or severity vs those not treated with fluoroquinolones.

The present study analyzed a nationwide database and meticulously controlled for infections and infection severity. In the model that simultaneously included infections and concomitant antibiotic use, we also observed an increased risk of AA/AD with fluoroquinolone use (vs nonuse) as in the prior studies. However, the observed association was probably biased, as in those studies. Specifically, we observed an elevated risk of AA/AD with indicated infections and multiple other antibiotics in the model. All these antibiotics may have been associated with a greater risk of AA/AD, but a more plausible explanation would be that the associations were biased owing to residual confounding.

Specifically, there could be residual confounding by risk factors for AA/AD, such as tobacco smoking, frailty, infections, and infection severity. Besides adjusting for multiple measured risk factors, we performed a series of analyses restricted to patients with indicated infections that matched cases and controls on infection type and that used an active comparison approach to further reduce confounding bias. These approaches have been shown to reduce measured and unmeasured confounding in observational studies.^{60,61,62,63,64,65} Fluoroquinolone use was not associated with a greater risk of AA/AD vs comparison antibiotics. On the other hand, our approach replicated the known association between fluoroquinolone use and tendon rupture disorder,^50,51 which lends support to our choice of active comparators.

Reverse causation may be one of the possibilities leading to fluoroquinolones being the apparent cause of AA. Specifically, patients may have infections followed by the onset of AA and fluoroquinolone use within a short period, and the diagnosis of AA was confirmed only after clinical workup (eFigure 2 in the Supplement). The hypothesis was partly supported by the attenuated risk estimates for fluoroquinolones when we ignored antibiotic use within 1 to 3 days before the index date.

Despite these concerns, we could not fully rule out an actual causal relation between fluoroquinolones and AA/AD in certain patients. Ciprofloxacin hydrochloride, a commonly used fluoroquinolone, may increase matrix metalloproteinase levels and degrade collagen.^66,67 One animal model showed that oral administration of ciprofloxacin hydrochloride (100 mg/kg/d for 4 weeks) did not increase the risk of aortic destruction or enlargement or of AA/AD in mice receiving a normal diet and saline infusion, but conferred a higher adverse consequence in mice receiving a high-fat diet and angiotensin infusion mimicking high exogenous stress.⁶⁸ However, we did not observe a higher risk of AA/AD with fluoroquinolone use in several high-risk subgroups.

Limitations

Our study has limitations. First, previous studies generally examined the risk for any use of fluoroquinolones without accounting for concomitant use of other antibiotics.^18,19,20,21 To reduce drug exposure misclassification, we examined the risk with monotherapy of fluoroquinolones and comparison antibiotics suggested by the treatment guidelines in Taiwan. However, this approach yielded a limited sample size in some subgroup analyses. Second, although we used a validated outcome algorithm, we could not rule out the possibility of outcome misclassification because coding practice may vary by country. However, our results were robust under various outcome definitions.

Third, although claims databases offer a sufficient sample size to examine rare adverse events, they lack important clinical information, such as image findings, microbiology testing results, and laboratory data. This may lead to missing or misclassified AA/AD, infections, infection severity, and other confounders. However, we applied validated claims-based algorithms to identify indicated infections. Among our cohort patients, approximately 4% of antibiotics were dispensed without an accompanying infection diagnosis. This indicated that undercoding of infection diagnoses may be minimal in the present study. We could not rule out potential misdiagnosis of AA as intra-abdominal infections resulting in the observed positive association. However, contiguous infections, such as intra-abdominal infections, have been reported as a possible cause leading to AA. Our findings are consistent with biological plausibility.^57,58,59 Similarly, potential misdiagnosis of AA as LRTIs or GUTIs may raise concerns about more frequent use of comparison antibiotics and underestimation of the risk with fluoroquinolones. However, the analysis comparing injectable forms of fluoroquinolones and third- or fourth-generation cephalosporins, which are usually used in the presence of strong evidence of infections, yielded null findings. In terms of potential confounding due to infection severity, the analyses stratified by proxy of infection severity, such as treatment setting, dosage form, or cephalosporin generation, produced consistent null associations.

Finally, although we applied rigorous designs to disentangle the roles of infections and antibiotics on AA/AD, we recognize that this is challenging to do in real-world settings (see eFigure 12 in the Supplement for a plausible directed acyclic graph). Specifically, if there was still residual confounding between antibiotic use and AA/AD due to tobacco smoking, frailty, or other unmeasured confounders, the observed association between infections and AA/AD in the model that adjusted for concomitant antibiotic use could be biased owing to improper adjustment for an intermediate variable.⁶⁹ However, we would expect the bias to lead to a greater (or spurious) effect of infections and not an attenuation of effect as observed in our analysis.

Conclusions

This study’s results emphasize the importance of considering coexisting infections while examining the safety of antibiotics using real-world data. Concern about AA/AD should not preclude patients with indicated infections from necessary treatment with fluoroquinolones.

Supplement.

eMethods. Data Source and Source Population, Selection of Comparison Antibiotics, and Analysis of Association Between Fluoroquinolones and Tendon Rupture Disorder

eFigure 1. Study Design

eFigure 2. Conceptual Time Sequence for Infections, Antibiotic Use, and Aortic Aneurysm or Aortic Dissection

eFigure 3. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections, Restricting to Those Without Any Use of Fluoroquinolones at Baseline

eFigure 4. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections, by Risk Window and Minimum Antibiotic Treatment Duration

eFigure 5. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections, by Treatment Setting

eFigure 6. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections, by Dosage Form

eFigure 7. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Extended-Spectrum Cephalosporin Monotherapy Among Patients With Indicated Infections, by Cephalosporin Generation and by Dosage Form

eFigure 8. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections, by Type of Infection

eFigure 9. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections Stratified by Subtype of Aortic Aneurysm or Aortic Dissection

eFigure 10. Risk of Aortic Aneurysm or Aortic Dissection Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections Stratified by Outcome Definition

eFigure 11. Risk of Tendon Rupture Outcomes Associated With Fluoroquinolone Monotherapy vs Comparison Antibiotic Monotherapy Among Patients With Indicated Infections

eFigure 12. Potential Directed Acyclic Graph for Infections, Antibiotic Use, and Aortic Aneurysm or Aortic Dissection

eTable 1. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) Diagnosis Codes Used to Identify the Outcomes of Interest

eTable 2. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) Procedure Codes, Taiwan Health Insurance Service Claims Codes, or Anatomical Therapeutic Chemical (ATC) Classification System Codes Used to Describe Clinical Management for Patients With Aortic Aneurysm or Aortic Dissection

eTable 3. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) Diagnosis Codes Used to Identify Episodes of Indicated Infections

eTable 4. Anatomical Therapeutic Chemical (ATC) Classification System Codes Used to Identify Study Antibiotics

eTable 5. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) Diagnosis Codes or Procedure Codes, Taiwan Health Insurance Service Claims Codes, or Anatomical Therapeutic Chemical (ATC) Classification System Codes Used to Identify Baseline Comorbidities, Indicated Infections, or Antibiotics, and to Define Patients With Cardiovascular Disease

eTable 6. Summary of Subgroup and Sensitivity Analyses

eTable 7. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) Diagnosis or Procedure Codes or Taiwan Health Insurance Service Claims Codes Used to Identify the Positive Control Outcome of Achilles Tendon Rupture and Any Type of Tendon Rupture

eTable 8. Study Cohort Assembly and Case Description

eTable 9. Risk of Aortic Aneurysm or Aortic Dissection Associated With Any Use of a Specific Antibiotic vs No Use of That Antibiotic

eTable 10. Distribution of Baseline Covariates Between Cases and Matched Controls of Aortic Aneurysm or Aortic Dissection Among Patients With Indicated Infections

eTable 11. Cases and Matched Controls Among Patients With Indicated Infections, Restricting to Those Without Any Use of Fluoroquinolones at Baseline

eTable 12. Cases and Matched Controls Among Patients With Indicated Infections by Risk Window and Minimum Antibiotic Treatment Duration

eTable 13. Cases and Matched Controls Among Patients With Indicated Infections, by Treatment Setting

eTable 14. Cases and Matched Controls Among Patients With Indicated Infections, by Dosage Form

eTable 15. Cases and Matched Controls Among Patients With Indicated Infections, by Cephalosporin Generation and by Dosage Form

eTable 16. Cases and Matched Controls Among Patients With Indicated Infections, by Type of Infection

eTable 17. Cases and Matched Controls Among Patients With Indicated Infections, Restricting to Those With Potentially High-Risk Characteristics

eTable 18. Cases and Matched Controls Among Patients With Indicated Infections, by Subtype of Aortic Aneurysm or Aortic Dissection

eTable 19. Cases and Matched Controls Among Patients With Indicated Infections, by Outcome Definition

eTable 20. Cases of Tendon Disorder and Matched Controls Among Patients With Indicated Infections

eReferences

Click here for additional data file.^{(829KB, pdf)}

References

1.Clouse WD, Hallett JW Jr, Schaff HV, Gayari MM, Ilstrup DM, Melton LJ III. Improved prognosis of thoracic aortic aneurysms: a population-based study. JAMA. 1998;280(22):1926-1929. doi: 10.1001/jama.280.22.1926 [DOI] [PubMed] [Google Scholar]
2.Acosta S, Ogren M, Bengtsson H, Bergqvist D, Lindblad B, Zdanowski Z. Increasing incidence of ruptured abdominal aortic aneurysm: a population-based study. J Vasc Surg. 2006;44(2):237-243. doi: 10.1016/j.jvs.2006.04.037 [DOI] [PubMed] [Google Scholar]
3.von Allmen RS, Anjum A, Powell JT. Incidence of descending aortic pathology and evaluation of the impact of thoracic endovascular aortic repair: a population-based study in England and Wales from 1999 to 2010. Eur J Vasc Endovasc Surg. 2013;45(2):154-159. doi: 10.1016/j.ejvs.2012.12.007 [DOI] [PubMed] [Google Scholar]
4.Wang SW, Huang YB, Huang JW, Chiu CC, Lai WT, Chen CY. Epidemiology, clinical features, and prescribing patterns of aortic aneurysm in Asian population from 2005 to 2011. Medicine (Baltimore). 2015;94(41):e1716. doi: 10.1097/MD.0000000000001716 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Pacini D, Di Marco L, Fortuna D, et al. Acute aortic dissection: epidemiology and outcomes. Int J Cardiol. 2013;167(6):2806-2812. doi: 10.1016/j.ijcard.2012.07.008 [DOI] [PubMed] [Google Scholar]
6.Howard DP, Banerjee A, Fairhead JF, Perkins J, Silver LE, Rothwell PM; Oxford Vascular Study . Population-based study of incidence and outcome of acute aortic dissection and premorbid risk factor control: 10-year results from the Oxford Vascular Study. Circulation. 2013;127(20):2031-2037. doi: 10.1161/CIRCULATIONAHA.112.000483 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Yeh TY, Chen CY, Huang JW, Chiu CC, Lai WT, Huang YB. Epidemiology and medication utilization pattern of aortic dissection in Taiwan: a population-based study. Medicine (Baltimore). 2015;94(36):e1522. doi: 10.1097/MD.0000000000001522 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Kent KC. Clinical practice: abdominal aortic aneurysms. N Engl J Med. 2014;371(22):2101-2108. doi: 10.1056/NEJMcp1401430 [DOI] [PubMed] [Google Scholar]
9.Chung J. Epidemiology, risk factors, pathogenesis, and natural history of abdominal aortic aneurysm. Updated March 17, 2020. Accessed October 6, 2019. https://www.uptodate.com/contents/epidemiology-risk-factors-pathogenesis-and-natural-history-of-abdominal-aortic-aneurysm
10.Black JH III, Burke CR. Epidemiology, risk factors, pathogenesis, and natural history of thoracic aortic aneurysm. Updated January 6, 2020. Accessed October 6, 2019. https://www.uptodate.com/contents/epidemiology-risk-factors-pathogenesis-and-natural-history-of-thoracic-aortic-aneurysm
11.Black JH III, Manning WJ. Overview of acute aortic dissection and other acute aortic syndromes. Updated July 22, 2020. Accessed October 6, 2019. https://www.uptodate.com/contents/overview-of-acute-aortic-dissection-and-other-acute-aortic-syndromes
12.Johansen K, Devin J. Mycotic aortic aneurysms: a reappraisal. Arch Surg. 1983;118(5):583-588. doi: 10.1001/archsurg.1983.01390050059011 [DOI] [PubMed] [Google Scholar]
13.Brown SL, Busuttil RW, Baker JD, Machleder HI, Moore WS, Barker WF. Bacteriologic and surgical determinants of survival in patients with mycotic aneurysms. J Vasc Surg. 1984;1(4):541-547. doi: 10.1016/0741-5214(84)90040-5 [DOI] [PubMed] [Google Scholar]
14.Lane GP, Cochrane AD, Fone DR. Salmonellal mycotic abdominal-aortic aneurysm. Med J Aust. 1988;149(2):95-97. doi: 10.5694/j.1326-5377.1988.tb120513.x [DOI] [PubMed] [Google Scholar]
15.Moneta GL, Taylor LM Jr, Yeager RA, et al. Surgical treatment of infected aortic aneurysm. Am J Surg. 1998;175(5):396-399. doi: 10.1016/S0002-9610(98)00056-7 [DOI] [PubMed] [Google Scholar]
16.Müller BT, Wegener OR, Grabitz K, Pillny M, Thomas L, Sandmann W. Mycotic aneurysms of the thoracic and abdominal aorta and iliac arteries: experience with anatomic and extra-anatomic repair in 33 cases. J Vasc Surg. 2001;33(1):106-113. doi: 10.1067/mva.2001.110356 [DOI] [PubMed] [Google Scholar]
17.Chen SH, Lin WC, Lee CH, Chou WY. Spontaneous infective spondylitis and mycotic aneurysm: incidence, risk factors, outcome and management experience. Eur Spine J. 2008;17(3):439-444. doi: 10.1007/s00586-007-0551-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Daneman N, Lu H, 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 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Lee CC, Lee MT, Chen YS, et al. Risk of aortic dissection and aortic aneurysm in patients taking oral fluoroquinolones. JAMA Intern Med. 2015;175(11):1839-1847. doi: 10.1001/jamainternmed.2015.5389 [DOI] [PubMed] [Google Scholar]
20.Lee CC, Lee MG, Hsieh R, et al. Oral fluoroquinolone and the risk of aortic dissection. J Am Coll Cardiol. 2018;72(12):1369-1378. doi: 10.1016/j.jacc.2018.06.067 [DOI] [PubMed] [Google Scholar]
21.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 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Emmerson AM, Jones AM. The quinolones: decades of development and use. J Antimicrob Chemother. 2003;51(suppl 1):13-20. doi: 10.1093/jac/dkg208 [DOI] [PubMed] [Google Scholar]
23.US Food and Drug Administration . FDA Drug Safety Communication: FDA warns about increased risk of ruptures or tears in the aorta blood vessel with fluoroquinolone antibiotics in certain patients. Published December 20, 2018. Accessed October 6, 2019. https://www.fda.gov/drugs/drug-safety-and-availability/fda-warns-about-increased-risk-ruptures-or-tears-aorta-blood-vessel-fluoroquinolone-antibiotics
24.European Medicines Agency . Pharmacovigilance Risk Assessment Committee (PRAC). Minutes of PRAC meeting on 10-13 May 2016. Published June 9, 2016. Accessed October 6, 2019. https://www.ema.europa.eu/docs/en_GB/document_library/Minutes/2016/07/WC500209623.pdf
25.National Health Insurance Administration . 2017-2018 National Health Insurance Annual Report. Updated June 17, 2020. Accessed October 6, 2019. https://www.nhi.gov.tw/English/Content_List.aspx?n=8FC0974BBFEFA56D&topn=ED4A30E51A609E49
26.Lin LY, Warren-Gash C, Smeeth L, Chen PC. Data resource profile: the National Health Insurance Research Database (NHIRD). Epidemiol Health. 2018;40:e2018062. doi: 10.4178/epih.e2018062 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Landenhed M, Engström G, Gottsäter A, et al. Risk profiles for aortic dissection and ruptured or surgically treated aneurysms: a prospective cohort study. J Am Heart Assoc. 2015;4(1):e001513. doi: 10.1161/JAHA.114.001513 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Sundbøll J, Adelborg K, Munch T, et al. Positive predictive value of cardiovascular diagnoses in the Danish National Patient Registry: a validation study. BMJ Open. 2016;6(11):e012832. doi: 10.1136/bmjopen-2016-012832 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Aronsky D, Haug PJ, Lagor C, Dean NC. Accuracy of administrative data for identifying patients with pneumonia. Am J Med Qual. 2005;20(6):319-328. doi: 10.1177/1062860605280358 [DOI] [PubMed] [Google Scholar]
30.Drahos J, Vanwormer JJ, Greenlee RT, Landgren O, Koshiol J. Accuracy of ICD-9-CM codes in identifying infections of pneumonia and herpes simplex virus in administrative data. Ann Epidemiol. 2013;23(5):291-293. doi: 10.1016/j.annepidem.2013.02.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Kern DM, Davis J, Williams SA, et al. Validation of an administrative claims-based diagnostic code for pneumonia in a US-based commercially insured COPD population. Int J Chron Obstruct Pulmon Dis. 2015;10:1417-1425. doi: 10.2147/COPD.S83135 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Grijalva CG, Chung CP, Stein CM, et al. Computerized definitions showed high positive predictive values for identifying hospitalizations for congestive heart failure and selected infections in Medicaid enrollees with rheumatoid arthritis. Pharmacoepidemiol Drug Saf. 2008;17(9):890-895. doi: 10.1002/pds.1625 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Schneeweiss S, Robicsek A, Scranton R, Zuckerman D, Solomon DH. Veteran’s Affairs hospital discharge databases coded serious bacterial infections accurately. J Clin Epidemiol. 2007;60(4):397-409. doi: 10.1016/j.jclinepi.2006.07.011 [DOI] [PubMed] [Google Scholar]
34.Patkar NM, Curtis JR, Teng GG, et al. Administrative codes combined with medical records based criteria accurately identified bacterial infections among rheumatoid arthritis patients. J Clin Epidemiol. 2009;62(3):321-327, 327.e1-327.e7. doi: 10.1016/j.jclinepi.2008.06.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Wiese AD, Griffin MR, Stein CM, et al. Validation of discharge diagnosis codes to identify serious infections among middle age and older adults. BMJ Open. 2018;8(6):e020857. doi: 10.1136/bmjopen-2017-020857 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Levine PJ, Elman MR, Kullar R, et al. Use of electronic health record data to identify skin and soft tissue infections in primary care settings: a validation study. BMC Infect Dis. 2013;13:171. doi: 10.1186/1471-2334-13-171 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Flum DR, Koepsell TD. Evaluating diagnostic accuracy in appendicitis using administrative data. J Surg Res. 2005;123(2):257-261. doi: 10.1016/j.jss.2004.08.020 [DOI] [PubMed] [Google Scholar]
38.Kleif J, Thygesen LC, Gögenur I. Validity of the diagnosis of appendicitis in the Danish National Patient Register. Scand J Public Health. 2020;48(1):38-42. doi: 10.1177/1403494818761765 [DOI] [PubMed] [Google Scholar]
39.Lo Re V III, Lim JK, Goetz MB, et al. Validity of diagnostic codes and liver-related laboratory abnormalities to identify hepatic decompensation events in the Veterans Aging Cohort Study. Pharmacoepidemiol Drug Saf. 2011;20(7):689-699. doi: 10.1002/pds.2148 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Iwashyna TJ, Odden A, Rohde J, et al. Identifying patients with severe sepsis using administrative claims: patient-level validation of the Angus implementation of the international consensus conference definition of severe sepsis. Med Care. 2014;52(6):e39-e43. doi: 10.1097/MLR.0b013e318268ac86 [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Carnahan RM, Herman RA, Moores KG. A systematic review of validated methods for identifying transfusion-related sepsis using administrative and claims data. Pharmacoepidemiol Drug Saf. 2012;21(suppl 1):222-229. doi: 10.1002/pds.2322 [DOI] [PubMed] [Google Scholar]
42.Su VY, Liu CJ, Wang HK, et al. Sleep apnea and risk of pneumonia: a nationwide population-based study. CMAJ. 2014;186(6):415-421. doi: 10.1503/cmaj.131547 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Hsiao LC, Muo CH, Chou CY, Tseng CH, Chen MF, Chang KC. Chronic osteomyelitis is associated with increased risk of new-onset atrial fibrillation: evidence from a nationwide cohort of 23 million people. Can J Cardiol. 2016;32(12):1388-1395. doi: 10.1016/j.cjca.2016.04.006 [DOI] [PubMed] [Google Scholar]
44.Chou CC, Shen CF, Chen SJ, et al. ; Infectious Diseases Society of Taiwan; Taiwan Society of Pulmonary and Critical Care Medicine; Medical Foundation in Memory of Dr. Deh-Lin Cheng; Foundation of Professor Wei-Chuan Hsieh for Infectious Diseases Research and Education; CY Lee’s Research Foundation for Pediatric Infectious Diseases and Vaccines; 4th Guidelines Recommendations for Evidence-Based Antimicrobial Agents Use in Taiwan (GREAT) Working Group . Recommendations and guidelines for the treatment of pneumonia in Taiwan. J Microbiol Immunol Infect. 2019;52(1):172-199. doi: 10.1016/j.jmii.2018.11.004 [DOI] [PubMed] [Google Scholar]
45.Hsueh PR, Lau YJ, Ko WC, et al. Consensus statement on the role of fluoroquinolones in the management of urinary tract infections. J Microbiol Immunol Infect. 2011;44(2):79-82. doi: 10.1016/j.jmii.2011.01.015 [DOI] [PubMed] [Google Scholar]
46.Infectious Diseases Society of the Republic of China; Medical Foundation in Memory of Dr Deh-Lin Cheng; Foundation of Professor Wei-Chuan Hsieh for Infectious Diseases Research and Education; Lee CY’s Research Foundation for Pediatric Infectious Diseases and Vaccine . Guidelines for antimicrobial therapy of urinary tract infections in Taiwan. J Microbiol Immunol Infect. 2000;33(4):271-272. [PubMed] [Google Scholar]
47.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. doi: 10.1016/0895-4356(92)90133-8 [DOI] [PubMed] [Google Scholar]
48.Desai RJ, Solomon DH, Shadick N, Iannaccone C, Kim SC. Identification of smoking using Medicare data—a validation study of claims-based algorithms. Pharmacoepidemiol Drug Saf. 2016;25(4):472-475. doi: 10.1002/pds.3953 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Kim DH, Schneeweiss S, Glynn RJ, Lipsitz LA, Rockwood K, Avorn J. Measuring frailty in Medicare data: development and validation of a claims-based frailty index. J Gerontol A Biol Sci Med Sci. 2018;73(7):980-987. doi: 10.1093/gerona/glx229 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.van der Linden PD, Sturkenboom MC, Herings RM, Leufkens HG, Stricker BH. Fluoroquinolones and risk of Achilles tendon disorders: case-control study. BMJ. 2002;324(7349):1306-1307. doi: 10.1136/bmj.324.7349.1306 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.van der Linden PD, Sturkenboom MC, Herings RM, Leufkens HM, Rowlands S, Stricker BH. Increased risk of Achilles tendon rupture with quinolone antibacterial use, especially in elderly patients taking oral corticosteroids. Arch Intern Med. 2003;163(15):1801-1807. doi: 10.1001/archinte.163.15.1801 [DOI] [PubMed] [Google Scholar]
52.Chan P, Tsai CW, Huang JJ, Chuang YC, Hung JS. Salmonellosis and mycotic aneurysm of the aorta: a report of 10 cases. J Infect. 1995;30(2):129-133. doi: 10.1016/S0163-4453(95)80007-7 [DOI] [PubMed] [Google Scholar]
53.Wang JH, Liu YC, Yen MY, et al. Mycotic aneurysm due to non-typhi salmonella: report of 16 cases. Clin Infect Dis. 1996;23(4):743-747. doi: 10.1093/clinids/23.4.743 [DOI] [PubMed] [Google Scholar]
54.Hsu RB, Tsay YG, Wang SS, Chu SH. Surgical treatment for primary infected aneurysm of the descending thoracic aorta, abdominal aorta, and iliac arteries. J Vasc Surg. 2002;36(4):746-750. doi: 10.1067/mva.2002.126557 [DOI] [PubMed] [Google Scholar]
55.Luo CY, Ko WC, Kan CD, Lin PY, Yang YJ. In situ reconstruction of septic aortic pseudoaneurysm due to Salmonella or Streptococcus microbial aortitis: long-term follow-up. J Vasc Surg. 2003;38(5):975-982. doi: 10.1016/S0741-5214(03)00549-4 [DOI] [PubMed] [Google Scholar]
56.Maeda H, Umezawa H, Goshima M, et al. Primary infected abdominal aortic aneurysm: surgical procedures, early mortality rates, and a survey of the prevalence of infectious organisms over a 30-year period. Surg Today. 2011;41(3):346-351. doi: 10.1007/s00595-010-4279-z [DOI] [PubMed] [Google Scholar]
57.Spelman D. Overview of infected arterial aneurysm. Updated July 8, 2020. Accessed October 6, 2019. https://www.uptodate.com/contents/overview-of-infected-mycotic-arterial-aneurysm
58.Valentine RJ, Chung J. Primary vascular infection. Curr Probl Surg. 2012;49(3):128-182. doi: 10.1067/j.cpsurg.2011.11.004 [DOI] [PubMed] [Google Scholar]
59.Sekar N. Primary aortic infections and infected aneurysms. Ann Vasc Dis. 2010;3(1):24-27. doi: 10.3400/avd.ctiia09000 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Schneeweiss S, Patrick AR, Stürmer T, et al. Increasing levels of restriction in pharmacoepidemiologic database studies of elderly and comparison with randomized trial results. Med Care. 2007;45(10)(suppl 2):S131-S142. doi: 10.1097/MLR.0b013e318070c08e [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Lund JL, Richardson DB, Stürmer T. The active comparator, new user study design in pharmacoepidemiology: historical foundations and contemporary application. Curr Epidemiol Rep. 2015;2(4):221-228. doi: 10.1007/s40471-015-0053-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
62.D’Arcy M, Stürmer T, Lund JL. The importance and implications of comparator selection in pharmacoepidemiologic research. Curr Epidemiol Rep. 2018;5(3):272-283. doi: 10.1007/s40471-018-0155-y [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Psaty BM, Siscovick DS. Minimizing bias due to confounding by indication in comparative effectiveness research: the importance of restriction. JAMA. 2010;304(8):897-898. doi: 10.1001/jama.2010.1205 [DOI] [PubMed] [Google Scholar]
64.Pearce N. Analysis of matched case-control studies. BMJ. 2016;352:i969. doi: 10.1136/bmj.i969 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.de Graaf MA, Jager KJ, Zoccali C, Dekker FW. Matching, an appealing method to avoid confounding? Nephron Clin Pract. 2011;118(4):c315-c318. doi: 10.1159/000323136 [DOI] [PubMed] [Google Scholar]
66.Tsai WC, Hsu CC, Chen CP, et al. Ciprofloxacin up-regulates tendon cells to express matrix metalloproteinase-2 with degradation of type I collagen. J Orthop Res. 2011;29(1):67-73. doi: 10.1002/jor.21196 [DOI] [PubMed] [Google Scholar]
67.Corps AN, Harrall RL, Curry VA, Fenwick SA, Hazleman BL, Riley GP. Ciprofloxacin enhances the stimulation of matrix metalloproteinase 3 expression by interleukin-1beta in human tendon-derived cells: a potential mechanism of fluoroquinolone-induced tendinopathy. Arthritis Rheum. 2002;46(11):3034-3040. doi: 10.1002/art.10617 [DOI] [PubMed] [Google Scholar]
68.LeMaire SA, Zhang L, Luo W, et al. Effect of ciprofloxacin on susceptibility to aortic dissection and rupture in mice. JAMA Surg. 2018;153(9):e181804. doi: 10.1001/jamasurg.2018.1804 [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Cole SR, Platt RW, Schisterman EF, et al. Illustrating bias due to conditioning on a collider. Int J Epidemiol. 2010;39(2):417-420. doi: 10.1093/ije/dyp334 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials