brand logo

Am Fam Physician. 1998;58(6):1335-1342

This is Part II of a two-part article on the principles of judicious use of antimicrobial agents in the treatment of upper respiratory tract infections in children. Part I, on otitis media and sinusitis, appeared in the last issue (Am Fam Physician 1998;58:1113–23).

See related patient information handout on runny nose in children, adapted from a CDC handout.

This article summarizes the principles of judicious antimicrobial therapy for three of the five conditions—cough, pharyngitis, the common cold—that account for most of the outpatient use of these drugs in the United States. The principles governing the other two conditions, otitis media and acute sinusitis, were presented in the previous issue. This article summarizes evidence against the use of antibiotic treatment for illness with cough or bronchitis in children, unless the cough is prolonged. Although empiric treatment may be started in patients with pharyngitis when streptococcal infection is suspected, the authors recommend withholding antibiotic treatment until antigen testing or culture is positive. There is never any indication for antibiotic treatment of the common cold; it is important to understand the natural history of colds, because symptoms such as mucopurulent rhinitis or cough, even when they persist for up to two weeks, do not necessarily indicate bacterial infection.

Data from the National Center for Health Statistics indicate that, in recent years, approximately 75 percent of all outpatient antimicrobial prescriptions have been issued for five conditions: otitis media, sinusitis, bronchitis, pharyngitis and nonspecific upper respiratory tract infection.1 Antimicrobial drug use rates are highest among children1; therefore, the pediatric age group is the appropriate initial focus for these guidelines. The principles presented here focus on areas in which antimicrobial therapy could be curtailed without compromising patient care. Part I of this article reviews appropriate use of antimicrobial medications in the treatment of otitis media and sinusitis; Part II reviews appropriate therapy for bronchitis, pharyngitis and nonspecific upper respiratory tract infections (common colds). The complete set of principles with supporting evidence and references, has been published separately.27

Cough Illness or Bronchitis

Principles of Antimicrobial Therapy

  1. Regardless of duration, nonspecific cough illness or bronchitis in children (or both) rarely warrants antimicrobial treatment.

  2. Antimicrobial treatment for prolonged cough (longer than 10 days) may be indicated occasionally. Pertussis should be treated according to established recommendations. Mycoplasma pneumoniae or Chlamydia pneumoniae infection may cause pneumonia and prolonged cough (usually in children older than five years of age); a macrolide agent (or tetracycline in children eight years of age or older) may be used for treatment. Children with underlying chronic pulmonary disease (not including asthma) may benefit occasionally from antimicrobial therapy for acute exacerbations.


Seven randomized, placebo-controlled antimicrobial trials for bronchitis among adult patients have been published in the English-language, peer-reviewed medical literature. A meta-analysis8 that included six of these studies led to the conclusion that there is no evidence to support the use of antimicrobial therapy in acute bronchitis. Three trials911 that used erythromycin, doxycycline or trimethoprim-sulfamethoxazole demonstrated minimal improvement in duration of cough and time missed from work in the group treated with antimicrobial medications. The remaining four trials,1215 including two that the authors concluded best fulfilled the criteria for methodologic soundness, showed no difference in outcomes between subjects who received placebo and subjects treated with erythromycin, doxycycline or tetracycline.

No randomized, placebo-controlled antimicrobial trials have evaluated children with cough illness or bronchitis (or both) defined strictly by sputum production; however, several pediatric studies1619 have evaluated the use of antimicrobial therapy in cough illnesses, which in common practice are termed “bronchitis” and are treated with antimicrobial medications. None of these studies showed any benefit from antimicrobial therapy for cough.

Although most practitioners recognize that most cough illness results from viral infections, some believe that lower respiratory bacterial superinfections might be averted by prophylactic use of antimicrobial agents. At least nine trials have evaluated the role of antimicrobial treatment in preventing bacterial complications of viral respiratory illnesses. A meta-analysis20 of these trials led to the conclusion that antimicrobial agents do not prevent or decrease the severity of such bacterial complications.

The lack of benefit from antimicrobial therapy is consistent with community- and hospital-based studies in the United States and other areas of the world that implicate non-bacterial organisms as the etiologic agents of cough illness and bronchitis. These studies2124 demonstrate that viral pathogens such as parainfluenza virus, respiratory syncytial virus and influenza virus account for most of the agents identified among children with cough illness or bronchitis. Among children older than five years, Mycoplasma pneumoniae was also recognized to cause cough illness and bronchitis.21 Recently, Chlamydia pneumoniae has been isolated from children with nonspecific cough illness.25 Taken together, the evidence is ample that cough illness or bronchitis in children is principally caused by viral pathogens or, in the case of older children, sometimes by M. pneumoniae or C. pneumoniae. Little if any microbiologic evidence indicates an important role for other pathogenic bacteria in the etiology of cough illness and bronchitis.


Principles of Antimicrobial Therapy

  1. Diagnosis of group A streptococcal pharyngitis should be based on results of appropriate laboratory tests in conjunction with clinical and epidemiologic findings.

  2. Antimicrobial therapy should not be given to a child with pharyngitis in the absence of diagnosed group A streptococcal or other bacterial infection.

  3. A penicillin remains the drug of choice for treating group A streptococcal pharyngitis.


Symptoms of “classic” streptococcal pharyngitis include acute onset of pharyngeal pain, dysphagia and fever. Malaise, headache, abdominal pain and vomiting commonly occur. Rhinorrhea, cough, hoarseness, conjunctivitis and diarrhea are uncommon and strongly suggest a viral cause. On examination, the pharynx is erythematous, a patchy exudate is often present on the posterior pharynx and tonsils, and palatal petechiae may be present. The anterior cervical lymph nodes often are enlarged and tender.

Unfortunately, these clinical findings are neither sensitive nor specific for group A streptococcal infection. When a diagnosis is based on clinical impression alone, physicians generally overestimate the probability that patients have streptococcal infection.26 Several schema have been developed to improve the ability to predict which patients have group A streptococcal pharyngitis by scoring clinical and epidemiologic findings.2729 None of them, however, accurately identifies all children who need treatment or avoids therapy for those who do not. Although the negative predictive value of a low score is good and may help a physician decide when a diagnostic test is not needed, the positive predictive value of even the highest score is limited. An evaluation of one such system among adults found that only 54 percent of patients in the most predictive group—those with a history of fever, tonsillar exudate, anterior cervical lymphadenopathy and an absence of cough—had group A streptococci identified by culture.29

Since the clinical presentation of pharyngitis does not reliably predict the etiologic agent, when group A streptococcal infection is suspected, the diagnosis should be based on results of a throat culture or an antigen detection test with culture confirmation of negative results. Culture of a specimen obtained by throat swab is recommended as the “gold standard” for diagnosis.30 Some studies report the sensitivity of antigen detection tests to be 90 percent or greater in carefully controlled clinical settings,3134 but such tests often have proved less sensitive in routine clinical practice.3540 Survey results show that many physicians initiate antimicrobial therapy for pharyngitis while awaiting results of throat culture and that antimicrobial therapy often is continued despite a negative culture result.41 This approach results in substantial antimicrobial overuse and fails to capture the full benefits of obtaining a culture. If antimicrobial medications are prescribed pending results of culture, parents should be contacted if the cultures are negative and told to stop therapy and discard any remaining antimicrobial medication.

Because early antimicrobial therapy may limit transmission of group A streptococci and facilitate a child's return to school or day care, appropriate therapy should be initiated as soon as the diagnosis is supported by a laboratory test. Antimicrobial treatment while a culture is pending may offer the advantages of symptom relief and, possibly, decreased transmission, but relief may also be obtained with symptomatic treatment.4244 Furthermore, no evidence suggests that early antimicrobial therapy decreases the rate of recurrent infection45 or is necessary to prevent acute rheumatic fever.27 Negative consequences of empiric antimicrobial therapy include cost, selection of resistant bacterial pathogens and risk of hypersensitivity or other adverse reactions. Use of a rapid antigen detection test can help clinicians resist the pressure for immediate therapy, since a negative result may facilitate return to school or day care.

Viral agents cause most episodes of pharyngitis. Even in patients with pharyngeal exudate and fever, group A streptococci account for a minority of infections. In one study, diagnostic tests for bacterial and viral pathogens were performed in 110 children with exudative pharyngitis and fever who had not been treated previously with antimicrobial agents. Group A streptococci were isolated from only 12 percent of the children, while viral infection was documented in 31 percent. In addition, viral agents for which diagnostic testing was not available, including rhinovirus and coronavirus, may have accounted for infection in some of the children in whom no etiologic agent was identified.46 The predominance of viral infection was even higher among children who were younger than three years of age—a group in which classic group A streptococcal pharyngitis occurs less often.

Since most episodes of pharyngitis are not caused by group A streptococci, empiric antimicrobial therapy would result in substantial overtreatment. The widespread availability of accurate, inexpensive diagnostic tests for group A streptococcal infections makes a diagnostic strategy of culture or antigen detection testing, or both, for children with suspected streptococcal pharyngitis effective and cost-effective,47 and offers a cogent approach to avoiding the overuse of antimicrobials.48

Common Cold

Principles of Antimicrobial Therapy

  1. Antimicrobial agents should not be given for the common cold.

  2. Mucopurulent rhinitis (thick, opaque or discolored nasal discharge) frequently accompanies the common cold. It is not an indication for antimicrobial treatment unless it persists for longer than 10 to 14 days.


Controlled trials of antimicrobial treatment of the common cold have consistently failed to show that it changes the course or outcome of the illness (Table 1). For example, a 1962 prospective, double-blind study17 of 781 children with colds demonstrated that 3.5 percent of those treated with antimicrobial agents developed purulent upper respiratory tract infections, compared with 2.6 percent of those treated symptomatically.17 A more recent study49 of 261 children randomly treated with penicillin, tetracycline or placebo showed similar results: 5 percent of the placebo group either did not improve or presented with evidence of a complication (e.g., pneumonia) by eight days, compared with 5 percent of the antimicrobial-treated group.

Study (year)Number of study subjectsComparison groupsOutcomeConclusion
Cronk, et al.66 (1954)2,177Penicillin G and/or symptomatic treatmentRequired return outpatient visit(s):No difference between groups
Penicillin G, 26%
Symptomatic, 20%
Hardy, et al.67 (1956)217Antimicrobial* or placeboRate of all infectious complications:No difference between groups
Antimicrobial, 15%
Placebo, 15%
Townsend16 (1960)845Antimicrobial or symptomatic treatmentRate of all infectious complications:No difference between groups
Antimicrobial, 14%
Symptomatic, 9%
Townsend17 (1962)781Antimicrobial or symptomatic treatmentRate of complications (e.g., acute otitis media):No difference between groups
Antimicrobial, 3.5%
Symptomatic, 2.6%
Lexomboon, et al.49 (1971)261Penicillin V or tetracycline or placeboNot improved or complicated:No difference between groups
Antimicrobial, 5%
Placebo, 5%
Gordon, et al.18 (1974)89Antimicrobial or placeboImproved symptoms or signs:Antimicrobials do not change short-term course of upper respiratory tract infection
Data not provided in publication
Stott and West12 (1976)212Doxycycline or placebo§Runny nose at day 5:Doxycycline beneficial at day 5; benefit no longer apparent by day 10
Doxycycline, 14%
Placebo, 30%
Taylor, et al.19 (1977)197Amoxicillin, co-trimoxazole or placeboPurulent rhinitis at day 8:Marginal benefit from antimicrobial therapy
Amoxicillin, 6%
Co-trimoxazole, 4%
Placebo, 15%
Normal activity:
Amoxicillin, 89%
Co-trimoxazole, 95%
Placebo, 97%
Kaiser, et al.50 (1996)314Co-amoxiclav or placeboAt day 5, in patients with positive cultures, persistent/worse symptoms:Antimicrobial therapy may be indicated in a subset of adult patients with sinusitis
Co-amoxiclav, 73%
Placebo, 96%

Two studies12,50 in adults showed modest benefits in those treated with antimicrobial agents and should be considered in evaluating the potential for benefit to children. The first,12 involving 212 adults with coughs or colds who were randomized to undergo treatment with doxycycline or placebo, showed a reduction in the proportion of study subjects with rhinorrhea at day 5 that was no longer apparent by day 10. More recently, a trial50 of treatment with amoxicillin-clavulanate versus placebo among 314 adults with cold symptoms showed that antimicrobial treatment, although not beneficial overall, was helpful at day 5 in a subgroup of patients with identified Streptococcus pneumoniae, Moraxella catarrhalis or Haemophilus influenzae infection. Given the modest benefit of antimicrobial agents even in these subgroups and the lack of benefit shown in most other studies (Table 1), antimicrobial agents are not indicated for treatment of viral rhinosinusitis.

Although a great majority of physicians realize that antimicrobial therapy will not hasten resolution of a cold, antimicrobial agents are often prescribed in an attempt to “prevent bacterial complications.” Some data indicate that this is not an effective strategy. A recent meta-analysis21 of five randomized clinical trials of the efficacy of antimicrobial treatment for colds to prevent lower respiratory infections offered no evidence for a protective effect. While some studies51,52 suggest that intermittent antimicrobial therapy begun at the onset of respiratory symptoms can help prevent acute otitis media in children at highest risk of recurrent disease, other studies53 do not suggest this. Furthermore, this approach is less effective than standard continuous antimicrobial prophylaxis in some high-risk children who meet stringent criteria3 when the two approaches are directly compared.54

Mucopurulent rhinitis is a common feature of uncomplicated viral rhinosinusitis and is not a separate indication for antimicrobial therapy.55,56 Unnecessary antimicrobial therapy can be avoided by recognizing the signs and symptoms that are part of the usual course of this disease and thus are not suggestive of a secondary bacterial infection. Viral rhinosinusitis begins with the inoculation of virus onto the nasal, oral or conjunctival mucosa, followed by infection of the local respiratory epithelium. The initial symptoms, which are caused both by cellular damage and the inflammatory response, include nasal stuffiness and throat irritation. Within a few hours, sneezing and watery nasal discharge may occur, often accompanied by systemic complaints such as low-grade fever, malaise, headache, anorexia and myalgias. Cough occurs in 60 to 80 percent of cases of viral rhinosinusitis5759 and does not necessarily suggest a bacterial infection.

One to three days after the onset of illness, nasal secretions typically become thicker and more mucopurulent because they contain desquamated epithelial cells, polymorphonuclear cells and bacteria that normally colonize the upper respiratory tract.59 The duration of illness usually ranges from two to seven days. Although patients generally improve by the 10th day, lingering symptoms, including cough (in up to 31 percent of patients) and nasal discharge (in 35 percent), can persist in children and adolescents for more than two weeks.59,60 With an average of six respiratory tract infections per year—;more in children who attend day care6165—;many children will have sequential episodes of viral rhinosinusitis with little time for improvement between episodes.

In 1984, one study56 randomized 142 children with mucopurulent nasopharyngitis into groups that received cephalexin, symptomatic therapy or placebo. Although bacteria often were identified in nasal secretions, antimicrobial treatment did not reduce the number of potentially pathogenic organisms obtained from nasopharyngeal cultures. Furthermore, no differences in clinical outcomes between antimicrobial- and placebo-treated groups were found at five to six days, based on both physicians' and parents' assessments. Of the children treated with antimicrobial medications, 76 percent had continued nasal discharge compared with 63 percent of those treated with placebo; 7 percent had evidence of complications compared with 8 percent of those treated with placebo, and clinical improvement was reported for 35 percent, compared with 31 percent of those treated with placebo.

Final Comment

By using these principles along with those outlined in Part I of this article as a starting point for individualized patient treatment, concerned clinicians can reduce unnecessary use of antimicrobial agents and help protect their patients from infections with resistant bacteria. Patient education is an essential part of any effort to promote judicious antimicrobial use. To assist clinicians in educating their patients, a variety of materials, including pamphlets, posters and fact sheets, are available from the Centers for Disease Control and Prevention (CDC). These materials may be obtained free of charge by calling 404-639-4702 for a fax order sheet.

If these efforts to reduce unnecessary antimicrobial use are successful, the reduction in antimicrobial resistance in this country may be substantial. The CDC has estimated that following the above principles would avoid approximately 50 million unnecessary antimicrobial prescriptions annually in the United States. This summary may help to reassure physicians and patients that these practices are in line with both scientific evidence and expert opinion.

Continue Reading

More in AFP

More in PubMed

Copyright © 1998 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See permissions for copyright questions and/or permission requests.