The development of antibiotic-resistant bacteria seems to be an inevitable consequence of using antimicrobial drugs. Antimicrobial resistance has increased since the introduction of antibiotics. Penicillin was first used in the early 1940s and by 1950, most Staphylococcus aureus organisms were resistant to the drug by virtue of elaboration of a penicillinase enzyme.

The emergence of a resistant population of bacteria in a patient as a result of antibiotic use generally occurs through a process termed “selective pressure.” Studies using special culture techniques show that healthy persons normally harbor small numbers of bacteria that are intrinsically resistant to antibiotics.¹ The larger numbers of antibiotic-susceptible organisms usually keep resistant bacteria in check. The use of any antibiotic eliminates organisms susceptible to that drug at the site of infection and at all sites in the body into which the drug penetrates in adequate concentrations. Selective pressure occurs when the administration of an antibiotic decreases the numbers of normal flora, allowing resistant bacteria to proliferate.

When antibiotics are prescribed to large numbers of persons in a population, resistant bacteria may become the predominant organisms in that community. This situation is occurring with respiratory-tract pathogens that were once universally susceptible to antibiotics.² A national survey of clinically significant outpatient isolates of Streptococcus pneumoniae, the most common bacterial cause of most respiratory tract infections, gathered from 30 U.S. medical centers in 1994 and 1995, revealed that 23.6 percent of strains were not susceptible to penicillin. Intermediate resistance was detected in about 14.1 percent of isolates, while about 9.6 percent of strains demonstrated high-level resistance.³ A recent study in Iceland showed a direct correlation between the use of antibiotics and the prevalence of resistant S. pneumoniae strains.⁴

Appropriate prescribing of antibiotics may slow the rate at which resistance becomes widespread throughout the community. A study in Finland demonstrated that erythromycin resistance among group A streptococci decreased from 16.5 to 8.6 percent over a four-year period during a nationwide program relying on national guidelines to limit the use of erythromycin.⁵ Other studies in the United States show that decreased use of antibiotics for prophylaxis and treatment correlated with decreasing rates of colonization with resistant organisms.^6,7

Inappropriate Antibiotic Use

Despite the association between the increased use of antibiotics and the spread of resistant organisms in the community, antibiotic use in this country appears to be increasing. A national office-based study suggested an increase of 48 percent in antibiotic prescribing for children between 1980 and 1992.⁸ The use of antimicrobials is especially prevalent in the very young and the elderly. In one study, 37 and 70 percent of children, by three and six months of age respectively, had received at least one antibiotic prescription.⁹ The emergence of in vitro resistance in previously susceptible pathogens, such as S. pneumoniae in respiratory tract infections and Escherichia coli in urinary tract infections, has temporally coincided with the increase in antibiotic use.

A good portion of antibiotic use appears to be for viral or spontaneously resolving bacterial infections. The Centers for Disease Control and Prevention (CDC) estimates that about 100 million courses of antibiotics are prescribed by office-based physicians each year, and that approximately one half of those prescriptions are unnecessary.¹⁰ Most antibiotic prescriptions in the ambulatory setting are for respiratory infections.² Studies evaluating physicians' prescribing patterns have found that almost 50 percent of office visits for colds and upper respiratory tract infections (URIs), and 80 percent of visits for acute bronchitis are treated with antibacterial agents.⁸ This prescribing pattern persists despite the fact that antibacterial agents have no effect against a viral disease such as the common cold. Data from numerous studies show that antibacterial agents do not significantly shorten the duration of illness in acute bronchitis.¹¹

Physicians continue to prescribe antibiotics despite acknowledging the association of antimicrobial use and bacterial resistance. In one study,¹² 97 percent of surveyed physicians agreed that overuse of antibiotics is a major factor contributing to antimicrobial resistance. However, an evaluation of the practice patterns of these same physicians revealed continued prescribing of antimicrobials for viral illnesses. Other potential harms of indiscriminate antibiotic prescribing include allergic reactions, adverse reactions and drug-drug interactions.²

Reasons for Prescribing

Patients want antibiotics, and physicians continue to prescribe them in situations where antibiotics may be withheld for many reasons. The act of prescribing an antibiotic has social and medical implications. From the patient's point of view, the prescribing of an antibiotic validates that the patient does have an illness, that a diagnosis has been made and that the illness is amenable to treatment. The fact that there is a “cure” for their problem reassures them that the illness is not serious.¹³

Patients are accustomed to receiving antibiotics for benign URIs and have come to believe, sometimes very strongly, that antibiotics demonstrate efficacy and are necessary for proper treatment of these illnesses. So convinced are patients of the efficacy of antibiotics that in one clinical trial of acute bronchitis, 60 percent of patients screened for the study refused to participate because they did not wish to be randomized to the placebo arm of the trial.¹⁴

The most common clinical factor causing patients to seek medical help for common infections is relief of bothersome symptoms. In one study,¹⁵ 90 percent of patients with URIs came to the physician's office because of a persistent cough. Another two thirds of patients complained of disturbed sleep, while nearly one half made an appointment at the urging of family and friends. Antibiotics have little effect on the rate of resolution of cough or disturbed sleep.¹¹

Certain clinical features of disease may inappropriately influence physicians' treatment decisions. The presence of purulent nasal discharge is more likely to result in a physician's prescribing an antibiotic for a patient with acute bronchitis.¹⁶ However, the purulent nasal discharge does not adequately differentiate viral from bacterial etiologies and is not associated with a worse outcome.^17–19

The presence of cough for more than three days may also trigger an antibiotic prescription. The natural history of acute bronchitis is such that many patients cough for two to three weeks as a result of bronchial hyperresponsiveness and not as a result of persistent infection.²⁰ In the absence of chronic obstructive lung disease, fever and a smoking history are not indications for antibiotic therapy in patients with acute bronchitis.¹⁷ Physicians may prescribe antibiotics in the hope of preventing more serious superinfections, but a meta-analysis of studies on antibiotic use in URIs did not show an effect of antibiotics in decreasing the incidence or severity of subsequent disease.²¹

Physicians' prescribing patterns appear to be affected by nonclinical factors in up to one half of cases.²² In today's managed-care environment, physicians are pressed to see more patients in less time. They have insufficient time to educate patients about the ineffectiveness of antibiotics.¹⁹ Prescribing an antibiotic may be an effective way to conclude an office visit.¹³ Although some physicians believe that prescribing an antibiotic will decrease return office visits, up to one fourth of patients with URIs make a return visit within one month of the initial encounter regardless of the prescription of antibiotic agents.¹¹

Although physicians often feel compelled to prescribe an antibiotic to satisfy patient demands, patient satisfaction surveys indicate that patients do not acknowledge putting such pressure on their physicians. One survey indicated that while 65 percent of patients expected to receive an antibiotic for treatment of a URI, there was no correlation between patient satisfaction and receipt of an antibiotic prescription.²³ Instead, patient satisfaction correlated highest with the quality of the physician-patient interaction. Results from focus groups indicate that patients would be satisfied if an antibiotic was not prescribed as long as the physician explained the reasons for the decision to withhold antibiotics.²⁴

Approach to the Patient

The physician should take a well-reasoned, step-wise approach to a patient with an infection before prescribing an antibiotic (Table 1). The first step is to make a tentative diagnosis based on data from the history and physical examination. Often, it is difficult to distinguish a bacterial from a viral illness by clinical features alone. Clinical and laboratory factors can be helpful in selecting subsets of patients more likely to have bacterial disease. More important than the actual microbiologic cause of the infection is knowledge of whether antibiotics will affect the course of the illness. For example, antibiotic therapy has little effect on the course of acute bronchitis caused by the bacterial pathogens Mycoplasma pneumoniae or Chlamydia pneumoniae.²⁵

The second step is to decide if antibiotics are warranted for that particular infection. Antibiotics are almost never necessary for treatment of the common cold or acute bronchitis in otherwise healthy persons. Table 2¹⁰ lists the indications for appropriate antimicrobial treatment in common infections.

In the third step, the physician should evaluate the potential for side effects or drug-drug interactions. Finally, with all else being equal, the physician should choose the least expensive alternative among equally effective and safe drugs.

The physician should also address the social aspects of the illness and explain to the patient that antibacterial drugs are not effective in viral infections and emphasize that unnecessary antibiotics can be harmful in several ways (Table 3). Several studies show that recent antibiotic use is a risk factor for colonization and infection with resistant organisms.⁴ Unwarranted prescription of an antibiotic exposes patients to potential adverse effects of the drug. Many antimicrobial agents are costly.

Perhaps most important to patients, an antibacterial agent will not shorten the duration of bothersome symptoms. The physician should not dismiss the importance of a patient's illness and the impact of those symptoms on the daily routine. Instead, the physician should explain the natural course of the illness and what the patient can expect in terms of time to symptom resolution. Realistic patient expectations may help decrease future office visits for a resolving illness.

Finally, the physician should actively manage symptoms with analgesics, antipyretics and decongestants, where appropriate. Therapy with beta agonists often is effective in shortening the duration of cough in acute bronchitis when clinical evidence of bronchial hyperresponsiveness (such as wheezing or bothersome cough) is present.^26,27 The CDC Web site provides a preprinted prescription form for patients with URIs to help physicians educate patients.²⁸ The site also includes educational pamphlets for physicians to provide to patients and parents as part of the office visit.

Physicians may take a preemptive approach to educating patients about appropriate use of antibiotics. The American Academy of Pediatrics recommends discussing antibiotic use as part of well-child care visits. Educational materials in the office waiting room also may help disseminate the principles of judicious antibiotic use to patients. Several sites on the Internet offer posters, pamphlets, videotapes and other materials for office use.^10,29 Office personnel may be helpful in changing patient attitudes and decreasing patient demand for antibiotic therapy.