Principles of Judicious Antibiotic Prescribing for Upper Respiratory Tract Infections in Pediatrics

More than 1 in 5 pediatric ambulatory visits to a physician result in an antibiotic prescription, which accounts for nearly 50 million antibiotic prescriptions annually in the United States.¹ It is widely documented that inappropriate antibiotic prescribing, especially for upper respiratory tract infections (URIs) of viral origin, is common in ambulatory care.^1,–3 As many as 10 million antibiotic prescriptions per year are directed toward respiratory conditions for which they are unlikely to provide benefit.¹ Recent evidence shows that broad-spectrum antibiotic prescribing has increased and frequently occurs when either no therapy is necessary or when narrower-spectrum alternatives are appropriate.^1,2 Such overuse of antibiotics causes avoidable drug-related adverse events,^4,–6 contributes to antibiotic resistance,^7,8 and adds unnecessary medical costs. This is compounded by the fact that few new antibiotics to treat antibiotic-resistant infections are under development.⁹ The growing health and economic threats of antibiotic resistance make promoting judicious antibiotic prescribing, which encompasses both reducing overuse and ensuring that appropriate agents are prescribed, an urgent public health and patient safety priority (http://www.cdc.gov/drugresistance/threat-report-2013).

Clinical decision-making about whether to prescribe antibiotics for a patient with URI symptoms is a daily occurrence for ambulatory-care physicians and other health care professionals who provide care for children. Although antibiotic prescribing is a routine part of clinical care, judicious antibiotic prescribing is challenging because it is difficult to distinguish between viral and bacterial URIs. A major objective of this clinical report is to provide a framework for clinical decision-making regarding antibiotic use for pediatric URIs. A point of emphasis is the importance of using stringent and validated clinical criteria when diagnosing acute otitis media (AOM), acute bacterial sinusitis, and pharyngitis caused by group A Streptococcus (GAS), as established through clinical guidelines. Additionally, this document emphasizes situations in which the use of antibiotics is not indicated, in particular for viral respiratory infections. Considering the frequency of URIs and the large proportion of antibiotic prescribing attributable to URI visits, these conditions represent a high-impact target for guidelines and other interventions designed to optimize antibiotic prescribing. The careful application of these criteria has the potential to mitigate overuse of antibiotics for pediatric URIs.

The first “Principles of Judicious Use of Antimicrobial Agents for Pediatric Upper Respiratory Tract Infections” were published in 1998 in response to concerns over the emergence and spread of antibiotic-resistant organisms.¹⁰ The Centers for Disease Control and Prevention, in collaboration with the American Academy of Pediatrics (AAP), sought to update these principles in a current context. Antibiotic resistance remains a major public health concern, and appropriate antibiotic use is an important health care quality goal. Although the introduction of a 7-valent pneumococcal polysaccharide-protein conjugate vaccine (PCV7) in 2000 led to large declines in the incidence of invasive pneumococcal infections,¹¹ an increase in the prevalence of nonvaccine serotypes, most notably serotype 19A, a commonly antibiotic-resistant serotype,^12,13 prompted the 2010 introduction of a 13-valent pneumococcal polysaccharide-protein conjugate vaccine (PCV13). Provider concerns about antibiotic resistance may be 1 factor leading to increasing use of broad-spectrum antibiotics. In recent years, several high-quality randomized controlled trials, meta-analyses, and new and updated clinical guidelines have been published that better define the effectiveness of antibiotic use for selected URIs, including AOM and acute bacterial sinusitis.^14,–23 At the same time, new evidence highlighting the extent to which antibiotics lead to adverse events requiring medical attention^4,–6 or potentially life-threatening events^24,25 has emerged.

This clinical report focuses on antibiotic prescribing for key pediatric URIs that, in certain instances, may benefit from antibiotic therapy: AOM, acute bacterial sinusitis, and pharyngitis. The specific recommendations are applicable to healthy children who do not have underlying medical conditions (eg, immunosuppression) placing them at increased risk of developing serious complications. The purpose of this report is to provide practitioners specific context using the most current recommendations and guidelines while applying 3 principles of judicious antibiotic use: (1) determination of the likelihood of a bacterial infection, (2) weighing the benefits and harms of antibiotics, and (3) implementing judicious prescribing strategies (Table 1).

Many aspects of the clinical history, symptoms, and signs of bacterial URIs overlap with or mirror those of viral infections or noninfectious conditions. To make a judicious decision about antibiotic use, it is essential first to determine the likelihood of a bacterial infection. When a practitioner has made the diagnosis of viral infection and has reasonably excluded the presence of concurrent bacterial infection, antibiotics should not be used because the potential for harm outweighs the potential benefit. In the specific cases of AOM, acute bacterial sinusitis, and pharyngitis, there are well-established stringent criteria that aid in distinguishing bacterial from nonbacterial causes.

The AAP and American Academy of Family Physicians released updated clinical practice guidelines for the diagnosis and treatment of AOM in 2013.²² AOM may be defined as “the rapid onset of signs and symptoms of inflammation in the middle ear.” The signs include bulging with or without erythema of the tympanic membrane (TM), and the symptoms may include otalgia, irritability, otorrhea, and fever. The diagnosis of AOM always requires a careful otoscopic examination to confirm the presence of inflammatory changes in the TM. The AAP guideline recommends that physicians diagnose AOM definitively under either of 2 conditions: (1) evidence of middle-ear effusion, as demonstrated by moderate to severe bulging of the TM, or (2) new onset of otorrhea that is not attributable to otitis externa. AOM may also be diagnosed when a child presents with only mild bulging of the TM but with additional symptoms of recent onset of ear pain or with intense erythema of the TM. Although clear visualization of the TM at times is difficult and because AOM is typically a self-limiting disease, a high degree of diagnostic certainty is essential to minimize antibiotic overuse. After AOM is diagnosed, judicious antibiotic use can be enhanced by further categorizing patients on the basis of illness severity (severe otalgia, otalgia lasting >48 hours, or temperature ≥39°C), laterality of infection (bilateral versus unilateral), and age (≤23 months vs ≥24 months). Patients with more severe symptoms, bilateral involvement, and younger age are more likely to benefit from antibiotics. Watchful waiting is reasonable for patients who are older and have nonsevere, unilateral disease.

The AAP²³ and the Infectious Diseases Society of America²¹ recently developed evidence-based clinical guidelines for the diagnosis and treatment of acute bacterial sinusitis. These guidelines support use of strict diagnostic criteria to distinguish bacterial from viral URIs. In particular, acute bacterial sinusitis is diagnosed on the basis of symptoms that are (1) persistent and not improving, (2) worsening, or (3) severe. Persistent symptoms are most common and include nasal discharge (of any quality) or daytime cough not improving by 10 days. Worsening symptoms include a worsening or new onset of fever, daytime cough, or nasal discharge after improvement of a typical viral URI. Severe symptoms include persistent fever (temperature ≥39°C) and purulent nasal discharge for at least 3 days. These clinical criteria are the basis for the diagnosis of acute bacterial sinusitis. Because many children with viral URI will have radiographic abnormalities, imaging should not be performed routinely.

Pharyngitis, or sore throat, may be accompanied by other nonspecific symptoms including cough, congestion, and fever. The most important diagnostic consideration is whether β-hemolytic GAS is the cause. Unlike AOM and acute bacterial sinusitis, the diagnosis of GAS infection can be confirmed with laboratory testing (either a rapid-antigen detection test or culture).^26,27 Scoring systems (Modified Centor or McIsaac Scores²⁸) can assist in identifying candidates for testing. Patients with 2 or more of the following features should undergo testing: (1) absence of cough, (2) presence of tonsillar exudates or swelling, (3) history of fever, (4) presence of swollen and tender anterior cervical lymph nodes, and (5) age younger than 15 years. Children with URI signs and symptoms, including cough, nasal congestion, conjunctivitis, hoarseness, diarrhea, or oropharyngeal lesions (ulcers, vesicles) more likely have viral illnesses and not GAS infection and should not be tested for GAS. Testing should generally not be performed in children younger than 3 years in whom GAS rarely causes pharyngitis and in whom rheumatic fever is uncommon. GAS should not be diagnosed in the absence of testing, even among patients with all of the aforementioned clinical criteria, with rare exceptions (eg, symptomatic and household contact with confirmed GAS pharyngitis). The importance of limiting testing to children with appropriate clinical criteria is further supported by the fact that colonization rates can reach 15% to 20% even among asymptomatic children.

Symptoms of the common cold, nonspecific URI, and bronchitis may overlap with or mirror those of bacterial URIs and can include cough, congestion, and sore throat. Collectively, these viral conditions account for millions of office visits per year. Acute bronchitis, in particular, is a cough illness that is diagnosed during more than 2 million pediatric office visits annually, and antibiotics are prescribed more than 70% of the time.¹ Application of diagnostic clinical criteria for AOM, sinusitis, and pharyngitis should aid clinicians in excluding these conditions. Management of the common cold, nonspecific URI, acute cough illness, and acute bronchitis should focus on symptomatic relief. Antibiotics should not be prescribed for these conditions.

If a bacterial infection is determined to be likely, the next step is to compare the evidence about the benefits of antibiotic therapy for each condition to the potential for harms. Relevant outcomes to consider for benefits include the cure rate, symptom reduction, prevention of complications, and secondary cases. Outcomes for harms include antibiotic-related adverse events (eg, abdominal pain, diarrhea, rash), Clostridium difficile colitis, development of resistance, and cost.

Several high-quality randomized controlled trials and meta-analyses have been published since the publication of the first principles of judicious use of antibiotics.^{18,–20,29,–33} Collectively, these have emphasized the following: (1) at least half of patients with AOM will recover without antibiotic therapy; (2) recovery is more likely and is hastened for children who receive antibiotic therapy compared with placebo; and (3) recovery without antibiotic therapy is less likely for younger children, those with bilateral versus unilateral disease, and those with more severe signs and symptoms. These observations underlie the rationale for treatment recommendations for AOM.

Multiple meta-analyses indicate that children receiving antibiotic therapy are more likely to achieve clinical success in terms of symptom resolution compared with placebo with a number needed to treat (NNT) of 7 or 8 patients.^18,33 Two recent randomized controlled trials among younger children that used even more stringent diagnostic criteria demonstrated that children who received antibiotics had more favorable symptom scores than those who received placebo, achieved faster symptom recovery, and had significantly lower rates of clinical failure as measured by otoscopic examination and persistence of symptoms, with an NNT closer to 4.^19,20 Nonetheless, it is important to note that in numerous studies of antibiotic efficacy for AOM, the majority of patients have symptoms that ultimately resolve spontaneously regardless of therapy and without complications. The potential for preventing complications, such as mastoiditis, may contribute, in part, to the clinical decision to use antibiotics for AOM. However, across the aforementioned controlled studies and meta-analyses, antibiotics have not demonstrated significant benefit in preventing these rare but serious complications. Observational data from the United Kingdom including more than 1 million AOM episodes indicates that when mastoiditis occurs, it typically is present at time of initial clinical presentation to care.³⁴ The estimated NNT to prevent 1 episode of mastoiditis is nearly 5000.³⁴ 

The AAP recommends antibiotic therapy for children diagnosed with AOM on the basis of presence of established clinical criteria. Observation can be considered for selected children, particularly children older than 2 years with nonsevere symptoms and unilateral disease.

The evidence base evaluating the effectiveness of antibiotics for treatment of acute bacterial sinusitis in children is limited and mixed. Three randomized controlled trials have assessed the effectiveness of antibiotics versus placebo for clinically diagnosed acute bacterial sinusitis in children, 2 of which have been published since the 1998 principles of judicious use of antibiotics.^14,17,35 Two trials concluded that antibiotics significantly improved the likelihood of symptom resolution after both 3 and 14 days,^14,35 but 1 study revealed no benefit of antibiotics over placebo.¹⁷ Key differences in the study design between these studies likely contributed to the differences in outcomes; the trials showing benefit included patients with more severe symptoms and applied more strict diagnostic criteria. This emphasizes the importance of careful attention to clinical diagnosis because antibiotics confer no clinical benefit for patients without diagnostic criteria suggesting acute bacterial sinusitis.

The benefit of antibiotic therapy in preventing suppurative complications, such as orbital cellulitis or intracranial abscess, is unproven. Individual efficacy trials lack the statistical power to demonstrate effectiveness against these rare complications, and a meta-analysis of randomized controlled trials in children and adults found no significant association between antibiotic use and the rate of complications.³⁶ 

The AAP recommends antibiotic therapy for children with clinical features of acute bacterial sinusitis, especially those with symptoms that are worsening or severe. Observation with close follow-up or antibiotic therapy can be considered for those with persistent symptoms (>10 days).

Antibiotic treatment of acute pharyngitis has been studied with respect to the effects on symptom resolution, transmission, and prevention of complications, including rheumatic fever. Five randomized controlled studies and 1 meta-analysis have examined the effect of immediate antibiotics on resolution of symptoms, 1 of which was completed since publication of the first principles of judicious use of antibiotics.^37,–41 These studies provide strong evidence that antibiotic therapy for children with pharyngitis and confirmation of GAS shortens the duration of symptoms, including sore throat and headache, by approximately 1 day. These benefits are apparent within as few as 3 days. However, the benefits of antibiotic therapy on shortening duration of fever are uncertain. Although data are somewhat limited, antibiotic therapy for index cases of GAS may reduce horizontal transmission and thereby prevent secondary cases.^40,42 These benefits are especially relevant in large households, child care settings, schools, and military settings.

Historically, the primary motivation for prescribing antibiotics for GAS pharyngitis was prevention of rheumatic fever. Randomized controlled trials in children before 1975 showed a fourfold benefit in preventing the onset of rheumatic fever, which occurred in approximately 3% of untreated patients.⁴³ Although localized outbreaks have occurred in recent decades, the incidence of rheumatic fever in most developed countries has declined dramatically.⁴⁴ Some of this decline might be attributable to better recognition and antibiotic treatment,⁴⁵ but more likely this relates to a decline in the prevalence of rheumatogenic strains of GAS.⁴⁶ 

Antibiotics may also have a role in preventing suppurative complications associated with GAS pharyngitis, such as peritonsillar abscess (PTA), AOM, and acute sinusitis. One meta-analysis suggested that antibiotic treatment prevents PTA; however, the majority of cases were derived from a single study conducted in 1951.⁴³ Data from a large observational cohort conducted in the United Kingdom suggest that antibiotic treatment may prevent development of PTA, but with an NNT >4000.⁴⁷ 

The AAP recommends antibiotic therapy for children with pharyngitis confirmed to be caused by GAS.

Because the predominant etiologies for these conditions are viruses, antibiotic therapy is not indicated. Because of uncertainty about the relevance of the diagnosis of acute bronchitis for children, data are limited. Nonetheless, a large meta-analysis concluded that there was no benefit to antibiotic therapy (including for delayed prescriptions) for patients with nonspecific cough and cold.⁴⁸ 

It is crucial to account for the potential for antibiotics to cause harm when used for treatment of URIs. The significance of potential harms should be directly balanced against the potential for benefit on a case-by-case basis. The importance of harms associated with antibiotic use is directly related to (1) an assessment of the magnitude of potential benefit (eg, greater benefit achieved for young children with bilateral AOM than unilateral) and (2) the extent to which uncertainty remains in the diagnosis. The preponderance of evidence for benefits of antibiotic therapy in treatment of bacterial URIs relates to attenuation of symptoms. When it is unclear whether the URI represents an acute bacterial infection, in general, the harms of antibiotic use have the potential to outweigh benefits. The importance of applying stringent clinical criteria to establish the diagnosis of a bacterial infection aids in differentiating children with nonspecific URI and common cold. Prescribing antibiotics for nonspecific URI and colds generally does not provide benefit and only exposes these children to potential harm.

Antibiotics are responsible for the largest number of unplanned medical visits for medication-related adverse events among children, which exceeds 150 000 per year and incurs substantial potential morbidity and cost.⁴ Antibiotic-associated adverse events can range from mild (diarrhea and rash), to more severe (Stevens-Johnson syndrome), to life-threatening (anaphylaxis or sudden cardiac death) reactions. Most clinical trials conducted to assess the treatment of AOM, sinusitis, and pharyngitis have used amoxicillin or amoxicillin-clavulanate, and these remain the first-line recommended agents for antibiotic therapy for these conditions. Studies comparing antibiotic treatment to placebo for AOM suggest a modestly increased rate of adverse events among treated patients, particularly diarrhea and rash. Two meta-analyses estimated rate differences of approximately 5% for adverse events.^18,32 Not included in these are the results from 2 recent trials using amoxicillin-clavulanate (older studies frequently used amoxicillin), which demonstrated even higher rates of diarrhea and dermatitis among patients receiving antibiotic therapy.^19,20 Among studies of sinusitis, in the most recent trial that demonstrated a benefit of antibiotic therapy, adverse events (defined as rash, diarrhea, vomiting, and abdominal pain) occurred in 44% of patients treated with high-dose amoxicillin-clavulanate compared with 14% in the placebo group.¹⁴ 

The adverse events described previously occur relatively frequently, although are relatively mild in most cases. Antibiotics can produce serious allergic reactions such as Stevens-Johnson syndrome.²⁵ There is rapidly growing evidence that antibiotic exposures early in life may disrupt the microbial balance of the intestines and other parts of the body in such a way as to contribute to long-term adverse health effects, such as inflammatory bowel disease, obesity, eczema, and asthma.^49,–51 A recent study highlighted risk of sudden death in adults treated with azithromycin, likely related to drug-associated prolongation of the QT interval.²⁴ Azithromycin is not a first-line antibiotic for any pediatric URI and is the antibiotic most likely to be used inappropriately (inadequate coverage for the most common pathogens causing AOM and sinusitis).¹ The incidence of C difficile colitis in hospitalized children has increased substantially during the past decade.⁵² Although children with comorbid conditions are at greatest risk, community-onset infections occur,⁵³ with recent antibiotic exposure as an important risk factor.

The relationship between antibiotic exposure and development of antibiotic resistance at the level of the individual patient and at the level of the community is well established.^7,8 Because of limited therapeutic options, antibiotic-resistant infections are difficult to treat and, in some cases, are associated with poor clinical outcomes.⁵⁴ Application of stringent diagnostic criteria and use of therapy only when the diagnosis and potential benefits are well established is essential to minimizing the impact of antibiotic overuse on resistance in individuals and within communities.

When evidence suggests that antibiotics may provide benefit, several aspects of judicious prescribing should be considered. These include selecting an appropriate antibiotic agent that treats the most likely pathogens (including accounting for local resistance patterns), selecting the appropriate dose, and treating for the shortest duration required. Additionally, physicians may consider the role of observation and use of delayed prescribing strategies.

The treatment of AOM and acute bacterial sinusitis illustrates several key aspects of judicious antibiotic use. Amoxicillin has traditionally been the recommended first-line agent for these conditions because Streptococcus pneumoniae is the most important cause. However, in some communities, the prevalence of amoxicillin-resistant β-lactamase-producing Haemophilus influenzae among bacterial URIs has increased significantly.⁵⁵ This underlies (in part) the recommendation to consider amoxicillin-clavulanate in certain instances (eg, severe symptoms, recent [<6 weeks] antibiotic exposure, known high local prevalence of amoxicillin-resistant H influenzae). It is important to note, however, that the benefits of antibiotic therapy appear to be greatest for patients with S pneumoniae infection, compared with other bacterial causes of URI, including H influenzae and Moraxella species, which may have higher rates of spontaneous resolution.¹⁶ In recognition of the possibility of a higher rate of adverse events caused by amoxicillin-clavulanate compared with amoxicillin, some physicians may choose to use amoxicillin as the first-line agent in most instances.

An understanding of local epidemiology and resistance patterns is especially important for understanding appropriate antibiotic selection. The rates of pneumococcal resistance to macrolides⁵⁶ and oral third-generation cephalosporins^57,58 make these agents poor choices for treating most children with suspected bacterial URIs. Emergence of macrolide resistance to GAS is also an important problem, although susceptibility testing is not routinely performed.

The role of observation (also termed “wait and see” or “delayed prescribing”) instead of immediate antibiotic therapy is an important consideration for children with AOM and acute bacterial sinusitis. Studies among patients with AOM have shown that this approach reduces antibiotic use, is well accepted by families, and, when supported by close follow-up, does not result in worse clinical outcomes.²² Observation therapy may be considered as an alternative strategy to immediate therapy for AOM and sinusitis for older patients without severe symptoms.^22,23 The use of this approach is an opportunity to engage in shared decision-making with patients and families to include a discussion about the potential benefits and risks associated with immediate antibiotic therapy.

Another important consideration for judicious antibiotic use is overall magnitude of exposure. Relatively short courses of therapy may achieve the same clinical benefits as longer courses while minimizing the risks of adverse events and development of resistance and lead to better compliance. Important examples are the use of once-daily amoxicillin for GAS pharyngitis²⁶ (vs 2 or 3 times daily dosing but the same daily dose of 50 mg/kg) and short-course therapy (eg, 7 days vs 10 days) for older children with AOM.²² 

This clinical report discusses principles of judicious antibiotic use for pediatric URIs. There is a strong emphasis on appropriate diagnosis, which is the foundation for making judicious decisions about prescribing antibiotics. Although focused on specific URIs, the main message has broader application for antibiotic use in general. These principles can be used to promote educational efforts for physicians, amplify the messages from recent clinical guidelines, assist with communication about appropriate antibiotic use to patients and families, and support local guideline development for judicious antibiotic use.

Michael T. Brady, MD, Chairperson, Red Book Associate Editor

Carrie L. Byington, MD

H. Dele Davies, MD

Kathryn M. Edwards, MD

Mary Anne Jackson, MD, Red Book Associate Editor

Yvonne A. Maldonado, MD

Dennis L. Murray, MD

Walter A. Orenstein, MD

Mobeen Rathore, MD

Mark Sawyer, MD

Gordon E. Schutze, MD

Rodney E. Willoughby, MD

Theoklis E. Zaoutis, MD

Marc A. Fischer, MD – Centers for Disease Control and Prevention

Bruce Gellin, MD – National Vaccine Program Office

Richard L. Gorman, MD – National Institutes of Health

Lucia Lee, MD – Food and Drug Administration

R. Douglas Pratt, MD – Food and Drug Administration

Jennifer S. Read, MD – National Vaccine Program Office

Joan Robinson, MD – Canadian Pediatric Society

Marco Aurelio Palazzi Safadi, MD – Sociedad Latinoamericana de Infectologia Pediatrica (SLIPE)

Jane Seward, MBBS, MPH – Centers for Disease Control and Prevention

Jeffrey R. Starke, MD – American Thoracic Society

Geoffrey Simon, MD – Committee on Practice Ambulatory Medicine

Tina Q. Tan, MD – Pediatric Infectious Diseases Society

Henry H. Bernstein, DO, Red Book Online Associate Editor

David W. Kimberlin, MD, Red Book Editor

Sarah S. Long, MD, Red Book Associate Editor

H. Cody Meissner, MD, Visual Red Book Associate Editor

Adam L. Hersh, MD, PhD

Lauri A. Hicks, DO

Jennifer Frantz, MPH

AAP

American Academy of Pediatrics

AOM

acute otitis media

GAS

group A Streptococcus

NNT

number needed to treat

PTA

peritonsillar abscess

TM

tympanic membrane

URI

upper respiratory tract infection

The authors acknowledge the contributions of Daniel Shapiro and Jeffrey Gerber for assistance in systematic review and critical review of early versions of this report.