CLINICAL PRACTICE GUIDELINE, PART 1

Position Paper

Principles of Appropriate Antibiotic Use for Acute Pharyngitis in Adults

Vincenza Snow, MD; Christel Mottur-Pilson, PhD; Richelle J. Cooper, MD, MSHS; and Jerome R. Hoffman, MD, MA, for the American
College of PhysiciansAmerican Society of Internal Medicine*

I n this guideline, we present the evidence and make

specific recommendations on how clinicians can dis-
tinguish and diagnose pharyngitis caused by group A -
rheumatic fever, valvular heart disease, immunosuppres-
sion, or recurrent or chronic pharyngitis (symptoms 7
days) or to patients whose sore throats are not due to
hemolytic streptococcus (GABHS). We also discuss acute pharyngitis. They are also not intended to apply
when antibiotic use is beneficial and which antibiotics during a known epidemic of acute rheumatic fever or
should be used. This guideline will not cover gonococcal streptococcal pharyngitis, or in nonindustrialized coun-
pharyngitis and diphtheria, for which the appropriate- tries in which the endemic rate of acute rheumatic fever
ness of immediate antibiotic treatment is well estab- is much higher than in the United States [1.2].
lished. The numbers in square brackets are cross-refer- Although it has been more than 50 years since treat-
ences to the numbered sections in the accompanying ment of streptococcal pharyngitis with penicillin was
background paper, Principles of Appropriate Antibiotic shown to prevent acute rheumatic fever, diagnosing
Use for Acute Pharyngitis in Adults: Background, GABHS infection remains a subject of controversy. This
which is part 2 of this guideline (see pages 509-517). is in part because the best criterion standard for the
diagnosis has not been definitively established, and test-
ing for a significant increase in antistreptolysin titers and
ACUTE PHARYNGITIS
use of throat swab cultures cannot provide real-time
Acute pharyngitis accounts for 1% to 2% of all visits
resultsthat is, results that are available when a decision
to outpatient departments, physician offices and emer-
regarding antibiotic use must be made. Because only
gency departments. A wide range of infectious agents
patients with GABHS (and a few other rare bacterial
produces acute pharyngitis, but viruses are the most
agents) benefit from antimicrobial therapy, the goal of
common cause. Approximately 5% to 15% of adult
the diagnostic evaluation should be to predict which
cases are caused by GABHS. In some patients, it can be
patients have a high likelihood of GABHS pharyngitis.
important to identify an infectious cause other than
GABHS (for example, gonococcal pharyngitis, Epstein
Barr virus, and acute HIV infection), but in the vast Diagnosis
majority of cases, acute pharyngitis in an otherwise Although recovery of GABHS from throat cultures
healthy adult is self-limited and rarely produces signifi- is reported in many clinical trials and may be the best
cant sequelae [1.1]. available predictor of treatment response, it has poor
Antimicrobial agents are prescribed to a substantial testretest agreement; does not always correlate with anti-
majority of patients with acute pharyngitis because of streptolysin titers; and produces results that vary de-
perceived patient expectations or physician desires to pending on technique, the site in which the sample is
avoid such potential complications as rheumatic fever obtained and plated, the culture medium, the condi-
and acute glomerulonephritis. Consequently, this dis- tions in which the culture is incubated, and whether
cussion will focus on the diagnosis and treatment of results are checked at 24 or 48 hours. Throat cultures
acute GABHS pharyngitis in adult patients. These also fail to distinguish acute infection from the carrier
guidelines do not apply to patients with a history of state. Furthermore, because culture results are not avail-

Ann Intern Med. 2001;134:506-508.

For author affiliations and current addresses, see end of text.

*This paper, written by Vincenza Snow, MD, Christel Mottur-Pilson, PhD, Richelle J. Cooper, MD, MSHS, and Jerome R. Hoffman, MD, MA, was developed for the Clinical Efficacy
Assessment Subcommittee: David C. Dale, MD (Chair); Patricia P. Barry, MD; William E. Golden, MD; Robert D. McCartney, MD; Keith W. Michl, MD; Allan R. Ronald, MD;
Sean R. Tunis, MD; Kevin B. Weiss, MD; and Preston L. Winters, MD. Approved by the Board of Regents on 16 July 2000.

Annals of Internal Medicine encourages readers to copy and distribute this paper, providing such distribution is not for profit. Commercial distribution is not permitted without the express
permission of the publisher.

506 2001 American College of PhysiciansAmerican Society of Internal Medicine

 Appropriate Antibiotic Use for Acute Pharyngitis, Part 1 Position Paper

able at the time of the index visit, and a delayed decision pared with a clinical decision alone, at the cost of po-
about use of antibiotics eliminates the primary benefit of tentially undertreating an additional small group of pa-
antimicrobial therapy in adults (symptom relief), we do tients with GABHS. Unfortunately, although many
not include use of throat culture for clinical decision physicians currently perform rapid antigen tests, there is
making in this set of management principles. In addi- evidence that they frequently prescribe antimicrobial
tion, there are several reasonable approaches to the di- agents even when test results are negative [4.2.2].
agnosis of GABHS in an otherwise healthy adult, such
as use of clinical criteria alone or use of rapid antigen Treatment
testing as an adjunct to clinical screening. Either of these Reasons to consider prescribing antimicrobials to
strategies is associated with reasonable diagnostic accu- treat streptococcal pharyngitis include a desire to pre-
racy (approximate sensitivity 70%, specificity 70%) vent rheumatic fever, prevent acute glomerulonephritis,
and allows treatment decisions to be made early in the prevent suppurative complications, decrease contagion,
course of illness, when patients can receive symptomatic and ameliorate symptoms. Although reported rates may
benefit [4.0 4.2]. underestimate the true incidence of acute rheumatic fe-
Although use of clinical screening (history and ver, repeatedly low and unchanging reported rates over
physical examination) alone would leave some patients more than a decade prompted the Centers for Disease
with GABHS untreated and would result in overtreat- Control and Prevention to drop acute rheumatic fever
ment of others, it would prompt treatment of most pa- from active national surveillance in 1994. It is still im-
tients with GABHS while dramatically decreasing excess portant to consider local epidemics and to be prepared
antibiotic use. The most reliable predictors include ton- to revise the approach to treatment if evidence of an
sillar exudates, tender anterior cervical lymphadenopa- outbreak exists, although the number of discrete out-
thy, absence of cough, and history of fever. Several stud- breaks of acute rheumatic fever in the United States is
ies examining these four criteria in a clinical decision very small and should not overly influence physician
rule indicate that in patients who have three or four of behavior. Similarly, poststreptococcal acute glomerulo-
the criteria, the sensitivity and specificity (compared nephritis does occur but is extremely rare, even in the
with those of throat culture) are approximately 75% and absence of antibiotic treatment. Furthermore, there is
75%, respectively [4.2.1]. no evidence that antimicrobial therapy decreases inci-
Rapid antigen tests for GABHS, when compared dence of this complication [3.0, 3.1].
with the criterion standard of throat culture, have The incidence of suppurative complications, regard-
widely variable reported sensitivity (58% to 96%) and less of treatment with antimicrobials, is also small. The
specificity (63% to 100%), depending on the type of most common complication today is peritonsillar ab-
test and practice setting of the trial. These rapid antigen scess (quinsy). Recent clinical trials provide some evi-
tests can be done at the bedside, and treatment decisions dence that targeting antimicrobials to a subset of pa-
can be made in real time. The potential advantage of the tients with higher clinical likelihood of GABHS may
rapid antigen tests compared with clinical models is that prevent quinsy. In another recent review of GABHS
they have approximately the same sensitivity and per- pharyngitis in practice, however, the authors reported
haps greater specificity for predicting results of throat that the risk for peritonsillar abscess was not reduced
culture. The disadvantage is that many patients would because many affected patients do not present for care
need to be tested to achieve the possible gain in speci- until after the complication has developed [3.2, 3.3].
ficity beyond that provided by clinical information Streptococcal infection often occurs in epidemics,
alone. This would shift economic costs from a few extra and contagion is a problem in areas of overcrowding or
prescriptions to many extra rapid antigen tests. Perform- close contacts. Antimicrobial agents lead to far greater
ing rapid antigen testing only in individual patients with microbiological eradication of streptococcus by 48 to 72
an intermediate clinical probability of GABHS (those hours. However, the presymptomatic incubation period
with three, or perhaps two, of the four clinical variables) for GABHS is 2 to 5 days, during which the infection
and withholding antimicrobial agents from those with can be unknowingly transmitted to others. The effect of
negative results would decrease antimicrobial use, com- treatment on spread of disease in a noninstitutionalized
www.annals.org 20 March 2001 Annals of Internal Medicine Volume 134 Number 6 507
Position Paper Appropriate Antibiotic Use for Acute Pharyngitis, Part 1

adult population is unknown. Nevertheless, it is not un- RECOMMENDATIONS

reasonable to consider whether an adult is living in close Recommendation 1. All patients with pharyngitis
quarters with small children when making clinical deci- should be offered appropriate doses of analgesics, antipyret-
sions about treatment [3.4]. ics, and other supportive care.
The relief of patient suffering is an appropriate con- The overwhelming majority of adults with acute
cern of physicians as well as patients. Antimicrobial pharyngitis have self-limited illness, which would do
agents, when instituted within 2 to 3 days of symptom well with supportive care only. Some suggested support-
onset, hasten symptomatic improvement among pa- ive care includes analgesics (both systemic and topical),
tients in whom throat culture ultimately grows GABHS antipyretics, and gargles.
or in populations in which a high likelihood of GABHS Recommendation 2. Physicians should limit antimicrobial
pharyngitis is identified clinically, but not in those with prescriptions to patients who are most likely to have GABHS.
a negative culture. Symptoms seem to resolve 1 to 2 days Group A -hemolytic streptococcus is the causal
sooner when antipyretics and other comfort measures agent in approximately 10% of adult cases of pharyngi-
are also instituted. One recent trial among unselected tis. The benefits of antibiotic treatment of adult pharyn-
patients with acute pharyngitis found that symptom du- gitis are limited to patients with GABHS infection. Rec-
ration was strongly related to patient satisfaction, and ommended strategies include a) empirical antibiotic
patient satisfaction was closely related to whether the treatment of adults with at least three of four clinical
physician addressed the patients concerns rather than to criteria (history of fever, tonsillar exudate, tender ante-
use of antibiotics. This further supports limiting antimi- rior cervical lymphadenopathy, and absence of cough)
crobials to the subset of patients most likely to benefit shown to be associated with GABHS pharyngitis and
and reemphasizes the importance of the quality of the nontreatment of all others; or b) empirical treatment of
physicianpatient interaction [3.5]. adults with all four clinical criteria, rapid antigen testing
If antimicrobial treatment is to be instituted, physi- of patients with three (or perhaps two) clinical criteria,
cians should choose an agent with the narrowest possible and treatment of those with positive test results and
spectrum of action that still covers GABHS. Thus, pen- nontreatment of all others.
icillin is the first choice for patients without a penicillin Recommendation 3. The preferred antimicrobial agent
allergy, and erythromycin is the first choice for penicil- for treatment of acute GABHS pharyngitis is penicillin, or
lin-allergic patients. To date, there is no evidence of erythromycin in penicillin-allergic patients.
GABHS resistance to or tolerance of penicillin, and eryth- Note: Clinical practice guidelines are guides only and may not apply to
romycin resistance rates are low in the United States [6.0]. all patients and all clinical situations. Thus, they are not intended to
override clinicians clinical judgment. All clinical practice guidelines from
Summary the American College of PhysiciansAmerican Society of Internal Med-
Antimicrobial treatment of GABHS pharyngitis icine are considered automatically withdrawn or invalid 5 years after
leads to a decreased risk for already rare complications publication or once an update has been issued.

and a decrease in the duration of some patient symp- Grant Support: Dr. Cooper is supported in part by a National Research
toms by 1 or 2 days. Symptomatic improvement re- Service Award (F32 HS00134-01) from the Agency for Healthcare Re-
search and Quality.
quires that treatment begin within 48 to 72 hours of
symptom onset. For otherwise healthy adult patients Requests for Single Reprints: Customer Service, American College of
with GABHS pharyngitis who are averse to antimicro- PhysiciansAmerican Society of Internal Medicine, 190 N. Indepen-
dence Mall West, Philadelphia, PA 19106.
bial agents or medication, it is reasonable to suggest a
policy of no antimicrobial treatment and expect few Current Author Addresses: Drs. Snow and Mottur-Pilson: American
College of PhysiciansAmerican Society of Internal Medicine, 190 N.
measurable adverse consequences. An appropriate strat-
Independence Mall West, Philadelphia, PA 19106.
egy in most adult patients is to limit antimicrobial ther- Drs. Cooper and Hoffman: Department of Emergency Medicine, Uni-
apy to the minority of adults with a high likelihood of versity of California, Los Angeles, 924 Westwood Boulevard, Los Ange-
GABHS pharyngitis, who are most likely to benefit. les, CA 90024.

508 20 March 2001 Annals of Internal Medicine Volume 134 Number 6 www.annals.org