Antibiotic risks are greater than the benefits? In this 2008 study from Brigham and Woman’s Hospital and Harvard, the researchers clearly state that “For an individual patient, the risks are greater than the benefits and the discussion should stop there.”
Despite little microbiological reason to think they would be helpful, antibiotic prescribing for predominantly viral acute respiratory infections remains nearly as popular as ever. Acute bronchitis in adults is an instructive example. According to guidelines and performance measures, the right antibiotic prescribing rate is close to zero percent.1, 2 However, in the United States, physicians prescribed antibiotics to 77% of adults with acute bronchitis in 1995, 59% of adults in 2000, and 67% of adults in 2005 (unpublished data from the National Ambulatory Medical Care Survey and National Hospital Medical Care Survey [NAMCS/NHAMCS]). Together, acute respiratory infections other than pneumonia account for 50% of antibiotic prescribing to adults and 75% of antibiotic prescribing to children.3
Microbiology aside, clinicians may prescribe and patients may use antibiotics with the thought of reducing symptoms or preventing complications. Research does not bear out these rationales. Systematic reviews from the Cochrane collaboration have found marginal to no benefit of antibiotics for the common cold, acute otitis media in children, maxillary sinusitis, sore throat, and acute bronchitis.4–8 Recent randomized controlled trials that carefully measured patient symptoms found no symptomatic benefit of antibiotics for acute bronchitis or sinusitis.9, 10
So, for most acute respiratory infections, there appears to be little benefit to antibiotics. However, if the risks to individual patients are minimal, then it might make sense to use antibiotics for acute respiratory infections. However, evidence of the risks of antibiotics continues to increase. Depending on the antibiotic, 5% to 25% of patients will develop antibiotic-associated diarrhea.14 Clostridium difficile is an increasingly recognized adverse event following antibiotic treatment. Quantifying the risk per antibiotic exposure is challenging, but the antibiotics most commonly associated with C. difficile are clindamycin, cephalosporins, and fluoroquinolones.15 About 2% of patients who take an antibiotic will develop a skin
reaction.16 About 1 in 5000 patients who receive an antibiotic will have an anaphylactic reaction.17
In addition, in studies examining adverse drug events, antibiotics keep emerging as a common culprit. Penicillins are one of the most common causes of adverse drug events in ambulatory practice.18 Antibiotics are the second most common cause of adverse drug events in the elderly.19 Antibiotics are one of the most common causes of adverse drug events following hospital discharge20 and adverse events from antibiotics often cause hospital admissions.21 Antibiotics are one of the most common classes of medication associated with malpractice claims.22
In this issue of Clinical Infectious Diseases, using the NAMCS/NHAMCS and National Electronic Injury Surveillance System–Cooperative Adverse Drug Event Surveillance (NEISS-CADES), Shebab and colleagues add the risks of an emergency department visit for an adverse drug event from antibiotics. They found that antibiotics are responsible for nearly 20% of emergency department visits for adverse drug events and caused 10.5 emergency department visits for adverse drug events for every 10,000 outpatient visits at which an antibiotic was prescribed. Because the NAMCS/NHAMCS only captures antibiotic prescribing that occurs in the context of a visit, Shebab are likely undercounting antibiotic prescribing. However, the authors are also undercounting adverse drug event emergency department visits as well: the sensitivity of the NEISS-CADES for detecting adverse events is around 45%.23 What is most striking is that the risk of antibiotics – especially sulfonamides and clindamycin – is comparable to insulin, warfarin, and digoxin, medications that are perceived as relatively dangerous.
Of course, in looking at only emergency department visits, Shebab and colleagues are detecting the tip of the iceberg. Many more patients are having mild adverse drug events that led them to seek non-emergency ambulatory care or to simply stop the antibiotic. Strom has noted that the present US drug safety system focuses on rare effects of new drugs instead of common adverse effects of older drugs.24 Indeed, according to the US Food and Drug Administration Adverse Event Reporting System (AERS) – a collection of voluntary reports – antibiotics do not emerge as a top cause of serious adverse drug events.25 However, depending on several factors, estimates of the sensitivity of the AERS system are as low as 0.3% and there has been widespread agreement that it takes too long to identify problem drugs and quantify risks of older drugs.
Shebab and colleagues rightly point out that antibiotic-associated adverse events are happening when antibiotics are taken as prescribed and as intended; the only error for many antibiotic prescriptions is that they are prescribed at all. If as many as half of antibiotic prescriptions are unnecessary, what interventions work to reduce inappropriate antibiotic prescribing? Clearly finger wagging in the medical literature is not working. Potential interventions to decrease inappropriate antimicrobial prescribing for acute respiratory infections include physician education, physician audit and feedback, patient education, multidimensional interventions, delayed antimicrobial prescriptions, health information technology solutions, and financial or regulatory incentives. Unfortunately, most interventions result in only modest absolute reductions in inappropriate antimicrobial prescribing, on the order of 10%. A recent systematic review found that multidimensional interventions involving physicians and patients appear more effective than clinician educational interventions, which, in turn, were more effective than interventions that used audit and feedback.27 However, for acute bronchitis, even successful interventions in the US have only reduced the antimicrobial prescribing rate to about 40% to 50%.13, 28, 29
When considering an antibiotic prescription, physicians may feel a need to balance individual benefit with the societal risks of increasing antimicrobial resistance.30 This is not necessary. First, “treating” a viral illness with antibiotics does not make microbiological sense. Second, the benefits of antibiotics for most acute respiratory infections appear to be shrinking in the literature. Third, the real risks of antibiotic prescribing are becoming clearer. The article by Shebab and colleagues is part of growing evidence that antibiotics have greater risks than previously appreciated. For everything we do, we have an obligation to weigh and clearly spell out benefits and risks for our patients. The decision and discussion about antibiotic prescribing should focus on benefits and risks for the individual patient. Physicians should be comfortable telling the following to their patients with most acute respiratory infections: “For your infection, there is about a 1 in 4000 chance an antibiotic will prevent a serious complication, a 5% to 25% chance it will cause diarrhea, and about a 1 in 1000 chance you will wind up in an emergency room from a bad reaction to the antibiotic.”
For most acute respiratory infections antimicrobial resistance is irrelevant. For an individual patient, the risks are greater than the benefits and the discussion should stop there.