Opportunities to Improve Fluoroquinolone Prescribing in the United States for Adult Ambulatory Care Visits

Abstract

The Food and Drug Administration warned against fluoroquinolone use for conditions with effective alternative agents. An estimated 5.1% of adult ambulatory fluoroquinolone prescriptions were for conditions that did not require antibiotics, and 19.9% were for conditions where fluoroquinolones are not recommended first-line therapy. Unnecessary fluoroquinolone use should be reduced.

Fluoroquinolones (FQ) are the third most commonly prescribed outpatient antibiotic class in the United States in adults, with an estimated 115 prescriptions per 1000 persons annually [1]. In 2016, the US Food and Drug Administration (FDA) updated the 2008 boxed warning to highlight serious side effects associated with systemic FQ use, including damage to tendons, muscles, joints, nerves, and the central nervous system [2]. The warning advises healthcare providers to not use FQ when the potential risks outweigh the benefits, specifically in conditions such as acute bronchitis where antibiotics are not typically required, and acute sinusitis and uncomplicated urinary tract infections for which other effective antibiotic treatment options exist [2]. In light of these recent warnings, we aimed to quantify and characterize the distribution and frequency of FQ prescribing for ambulatory care visits in US adults and to quantify and identify specific target conditions to focus public health and antibiotic stewardship efforts to reduce inappropriate FQ use.

METHODS

Data from 3 data sources were combined to report the number of outpatient FQ prescriptions and to characterize the distribution of diagnoses associated with FQ prescriptions. The QuintilesIMS™ Xponent database was used to report the number of outpatient FQ prescriptions dispensed in 2014. This database represents 100% of outpatient prescription activity for retail pharmacies projected from data collected from over 90% of retail pharmacies [1].The distribution of diagnoses leading to FQ prescriptions were estimated using the US Centers for Disease Control and Prevention’s (CDC) National Center for Health Statistics’ National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS) data from 2013 to 2014 [3]. These cross-sectional surveys systematically sample office-based physicians (NAMCS) and emergency departments (NHAMCS) in nonfederal hospitals throughout the United States. Visits among persons ≤19 years of age were excluded because FQ are infrequently prescribed to children. Diagnoses were coded by International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) codes. Diagnostic categories for common outpatient infectious diseases were created as described previously [3].

Specific diagnoses for which no antibiotic therapy should be prescribed, that is, viral upper respiratory tract infections and bronchitis, were identified [4]. Diagnoses for which FQ are not recommended as first-line agents were highlighted. These conditions included sinusitis and uncomplicated cystitis in women, for which the guideline-recommended antibiotics are amoxicillin with/without clavulanate and nitrofurantoin or trimethoprim-sulfamethoxazole, respectively [5, 6]. NAMCS/NHAMCS provide national estimates on antibiotic prescribing in physician offices and emergency departments, and account for approximately 64% of total FQ dispensing as measured by Xponent from all outpatient pharmacies. Thus national estimates of overall outpatient prescriptions were calculated by multiplying the percentage of NAMCS/NHAMCS visits for each diagnostic category with total number of outpatient FQ transactions in 2014 reported by QuintilesIMS™ and the CDC. The rank and proportion of FQ out of total antibiotics prescribed for each diagnostic category were also calculated. All statistical analyses were performed in STATA (StataCorp 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP) and accounted for components of the complex survey design (patient visit weights, strata, and primary sampling unit design variables). Estimates not meeting standards of reliability or precision (if based on fewer than 30 sampled visits or if the relative standard error < 0.3) were not reported [3].

RESULTS

During 2014, 31.5 million FQ courses were dispensed to adults from U.S. community pharmacies (QuintilesIMS™ Xponent data). The distribution of visits by condition associated with FQ prescribing (NAMCS/NHAMCS data) is shown in Table 1. The largest proportion (24.5%) of FQ prescriptions were prescribed for genitourinary conditions. Uncomplicated urinary tract infections represented the largest subgroup of genitourinary conditions treated with FQ (15.0%). FQ were prescribed for 40.3% of uncomplicated urinary tract infections, more than any other antibiotic class. Respiratory conditions accounted for 21.6% of all FQ prescriptions. Of antibiotics prescribed for all respiratory conditions, FQ were the third most common antibiotic class prescribed. Skin and gastrointestinal conditions accounted for 14.6% and 11.6% of all FQ prescriptions respectively.

Table 1.

Fluoroquinolone prescribing, according to diagnosis, for ambulatory care visits among adults >19 years of age in the United States

Diagnosisa	Estimated number of fluoroquinolone prescriptions from QuintilesIMS™	Proportion of visits with a fluoroquinolone prescription from (NAMCS/NHAMCS)b	Percentage of fluoroquinolone prescriptions of all antibiotics per conditionc
Genitourinary conditions	7721082	24.5%	40.4%
Urinary tract infections	5185676	16.4%	42.0%
Complicated	452881	1.4%	74.3%
Uncomplicated	4732795	15.0%	40.3%
Other genitourinary conditions	2535406	8.0%	37.5%
Respiratory conditions	6797720	21.6%	15.0%
Sinusitis	1529060	4.8%	11.1%
Pneumonia	1339111	4.2%	48.5%
Bronchitis/bronchiolitisd	852376	2.7%	12.9%
Viral upper respiratory tract infections	757477	2.4%	10.8%
Suppurative otitis media	199859	0.6%	9.9%
Other respiratory conditions	1008109	3.2%	24.6%
Skin conditions	4599751	14.6%	26.5%
Gastrointestinal conditions	3642547	11.6%	29.2%
Remaining Codese	8768900	27.8%	NA
All conditions	31530000	100%	NA

Abbreviations: NAMCS/NHAMCS, National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey.

^aDiagnostic categories were created based on the most likely indication for an antibiotic prescription in a tiered fashion as done previously [3]. One or more of the following ICD-9-CM codes were used to identify complicated urinary tract infections: 590.1 (acute pyelonephritis), 590.2 (renal and perinephric abscess), and 590.8 (other pyelonephritis or pylonephrosis not specified as acute or chronic).

^bWeighted annual proportion of sampled visits with fluoroquinolone prescribed out of total visits where an antibiotic was prescribed per year from 2013 to 2014, N = 20.1 million

^cAntibiotic classes included were penicillins, cephalosporins, macrolides, quinolones, lincomycin derivatives, tetracyclines, sulfonamides, urinary anti-infectives, metronidazole, and linezolid. Percentage of total antibiotics.

^dBronchitis or bronchiolitis includes visits with bronchitis, not specified as acute or chronic, and acute bronchitis and bronchiolitis but excludes visits in which the second or third diagnosis was chronic bronchitis, emphysema, or chronic obstructive pulmonary disease, included in other respiratory conditions.

^eIncludes miscellaneous conditions and remaining codes not listed elsewhere.

Viral upper respiratory tract infections and bronchitis, for which no antibiotics should be prescribed, led to an estimated 1.6 million FQ prescriptions (5.1% of total). Sinusitis and uncomplicated urinary tract infections, for which FQ are not first-line recommended therapy, accounted for an estimated 6.3 million FQ prescriptions (19.9% of total) (Table 1).

DISCUSSION

In 2014, 31.5 million FQ prescriptions were dispensed; visits for genitourinary, respiratory, skin, and gastrointestinal conditions accounted for most FQ prescriptions. An estimated 7.9 million FQ prescriptions were for conditions where no antibiotics should be prescribed or for which FQ are not first-line recommended therapy, which include acute bronchitis, acute sinusitis, and uncomplicated urinary tract infections, conditions recently highlighted in the FDA warning [2]. With the documentation of serious adverse events and rising rates of antibiotic resistance, prescribing FQ only when indicated and recommended will improve patient outcomes. Although all antibiotics can cause potential adverse events, the perceived safety and risk-benefit ratio when prescribing FQ needs to be reconsidered. FQ are also associated with a high risk for subsequent development of Clostridium difficile infection; a recent ecologic analysis of data from England suggested that restricting FQ prescribing played the most important role in decreasing the incidence of C. difficile infections at the country level [7].

FQ have well-documented efficacy, a broad spectrum of activity covering many common pathogens, favorable pharmacokinetics, and perceived safety, which may have led to overprescribing [8]. However, in this analysis, an estimated 1.6 million FQ prescriptions were for acute respiratory conditions for which antibiotics are not recommended. Although antibiotics are not always indicated in cases with acute sinusitis, even when antibiotics are needed, quinolones are not first-line agents [4]. A recent study reported that physicians prescribe appropriate first-line antibiotic therapy in only 37% (95% confidence interval 32%–43%) of cases of sinusitis and pharyngitis in adults [9], indicating that improvements in antibiotic selection for specific conditions are needed. FQ prescribing may be appropriate in some cases of acute bacterial sinusitis and uncomplicated urinary tract infection, such as in patients who have a beta-lactam allergy, risk for an antibiotic- resistant infection, or failed initial therapy. Also, FQ may be appropriate for specific clinical situations such as the treatment of urinary tract infection in males. However, FQ were the most commonly prescribed antibiotic for uncomplicated urinary tract infections, where the Infectious Diseases Society of America (IDSA) treatment guidelines emphasize the avoidance of FQ for this condition due to the possibility of “collateral damage” with their use. IDSA cited that FQ should be reserved for more serious infections, where the broad spectrum of activity of FQ is important [6]. In addition, resistance to ciprofloxacin has been increasing among urinary pathogens, specifically Escherichia coli, in adult female outpatients [10]. Prescribing FQ for empiric treatment of uncomplicated cystitis in adult women should be discouraged. Pediatric providers have long considered FQ to be unsafe in children, and treated acute respiratory conditions and urinary tract infections without their use. If recommended first-line therapies were used for sinusitis and uncomplicated urinary tract infections, a -proportion of 6.3 million FQ prescriptions per year reported in this study would be avoided. Although not specifically mentioned in the recent FDA warning [2], in this analysis gastroenterological and skin conditions accounted for over 8 million prescriptions in 2014. The practice of treatment of gastrointestinal infections with FQ is changing, with rising rates of resistance, particularly among Campylobacter species in certain destinations in Asia [11]. FQ are not recommended for treating Staphylococcus aureus, the most common cause of skin infections, as resistance may emerge during therapy [12]. However, in our study, 26.1% of FQ prescriptions were given for gastroenterological and skin conditions, where FQ are not considered first-line agents.

One limitation of this analysis is that total outpatient FQ prescriptions tin QuintilesIMS™ Xponent FQ prescribing data may not be distributed similarly to the NAMCS/NHAMCS data across clinical diagnoses. Also urgent care and retail clinics are not included in NAMCS/NHAMCS, so these results may not be generalizable to all outpatient settings. ICD-9-CM codes are used in NAMCS/NHAMCS data, which may not differentiate all the diagnoses of interest, and may not fully reflect the clinical diagnosis. Also “diagnosis shifting,” where clinicians select a more antibiotic-appropriate diagnosis, may also occur. Assessing appropriateness of prescribing in this analysis was difficult, since information on antibiotic allergies, previous treatment history and patient comorbidities was not collected, which could have justified FQ prescribing, even when not recommended by treatment guidelines. The most recently available data from NAMCS/NHAMCS are several years old; however, there is no indication that there have been significant changes in antibiotic prescribing rates in adults (https://gis.cdc.gov/grasp/PSA/AUMapView.html).

This analysis suggests that a quarter of FQ prescriptions were given for conditions for which no antibiotics are indicated, or for which FQ are not recommended first-line therapy. With the threats to patient safety and rising rates of antibiotic resistance, FQ should not be prescribed for conditions where alternative effective therapies are recommended. Antibiotic stewardship efforts should target inappropriate FQ prescribing in adults, specifically for acute respiratory tract infections for which no antibiotics are needed, and for ambulatory infections for which FQ are not recommended first-line therapy. This will optimize clinical outcomes by reducing unnecessary antibiotic therapy and preventing FQ adverse drug reactions.

Notes

Financial support. This work was supported by the Centers for Disease Control and Prevention. A. P. is supported by funding from the National Institutes of Health (1R01AI125642-01 and R21HD090955). A. H. is supported by funding from the Agency for Healthcare Research and Quality (K08 HS23320-03). A. P., A. H., and D. S. are supported by the Centers for Disease Control and Prevention (CDC/NCEZID 17IPA1708453). The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Potential conflicts of interest. All authors report no conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.