Antibiotic Treatment of Adults with Sore Throat by Community Primary Care Physicians a National Survey, 1989-1999

ORIGINAL CONTRIBUTION

Antibiotic Treatment of Adults With Sore Throat by Community Primary Care Physicians A National Survey, 1989-1999

Jeffrey A. Linder, MD Context Most sore throats are due to viral upper respiratory tract infections. Group Randall S. Stafford, MD, PhD A ␤-hemolytic streptococci (GABHS), the only common cause of sore throat warrant- ing antibiotics, is cultured in 5% to 17% of adults with sore throat. The frequency of ORE THROAT ACCOUNTS FOR antibiotic use for pharyngitis has greatly exceeded the prevalence of GABHS, but less 2.1% of ambulatory visits in the is known about specific classes of antibiotics used. Only penicillin and erythromycin United States and is the second are recommended as first-line antibiotics against GABHS. most common symptomatic Objectives To measure trends in antibiotic use for adults with sore throat and to Sreason for seeking care, after cough.1 determine predictors of antibiotic use and nonrecommended antibiotic use. Most sore throats are due to upper res- Design, Setting, and Subjects Retrospective analysis of 2244 visits to primary care piratory tract viruses such as rhinovi- physicians in office-based practices in the National Ambulatory Medical Care Survey, 2 rus, coronavirus, and adenovirus. The 1989-1999, by adults with a chief complaint of sore throat. main bacterial cause of sore throat is Main Outcome Measures Treatment with antibiotics and treatment with nonrec- ␤ group A -hemolytic streptococci ommended antibiotics, extrapolated to US annual national rates. (GABHS), which is cultured in 15% to 36% of children and 5% to 17% of Results There were an estimated 6.7 million annual visits in the United States by adults with sore throat between 1989 and 1999. Antibiotics were used in 73% (95% adults with sore throat.3-9 Group A confidence interval [CI], 70%-76%) of visits. Patients treated with antibiotics were ␤-hemolytic streptococci is the only given nonrecommended antibiotics in 68% (95% CI, 64%-72%) of visits. From 1989 common cause of sore throat warrant- to 1999, there was a significant decrease in use of penicillin and erythromycin and an ing antibiotic treatment.10 increase in use of nonrecommended antibiotics, especially extended-spectrum mac- Ideally, the proportion of patients rolides and extended-spectrum fluoroquinolones (PϽ.001 for all trends). In multivari- with sore throat receiving an antibi- able modeling, increasing patient age (odds ratio [OR], 0.86 per decade; 95% CI, 0.79- otic would approximate or perhaps 0.94) and general practice specialty (OR, 1.54 compared with family practice specialty; slightly exceed the prevalence of 95% CI, 1.10-2.14) were independent predictors of antibiotic use. Among patients GABHS in those with sore throat. How- receiving antibiotics, nonrecommended antibiotic use became more frequent over time (OR, 1.17 per year; 95% CI, 1.11-1.24). ever, 76% of adults and 71% of children diagnosed with pharyngitis in Conclusions More than half of adults are treated with antibiotics for sore throat by 1992 were treated with an antibi- community primary care physicians. Use of nonrecommended, more expensive, broader- spectrum antibiotics is frequent. otic.11,12 Therefore, pharyngitis repre- sents 5.6 million annual antibiotic pre- JAMA. 2001;286:1181-1186 www.jama.com scriptions and is responsible for 9% of America recommends using penicillin sicians—American Society of Internal all antibiotics used in adults, over half 14,15 11 or erythromycin (for those who are Medicine. Penicillin and erythro- of which are likely unnecessary. penicillin allergic) as the first-line agent mycin are recommended because of Compared with the frequency of an- for patients with sore throat due to tibiotic prescription, less is known 13 GABHS. These recommendations Author Affiliations: General Medicine Division, Mas- about the classes of antibiotics pre- were recently reiterated in the guide- sachusetts General Hospital, Boston (Dr Linder); and scribed for patients with sore throat. lines for the treatment of adults with Stanford Center for Research in Disease Prevention, The Infectious Diseases Society of Palo Alto, Calif (Dr Stafford). acute pharyngitis issued by the Cen- Corresponding Author and Reprints: Jeffrey A. Linder, ters for Disease Control and Preven- MD, General Medicine Division, Massachusetts Gen- See also Patient Page. eral Hospital, 50 Staniford St, 9th Floor, Boston, MA tion and the American College of Phy- 02114 (e-mail: [email protected]).

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ANTIBIOTIC TREATMENT OF ADULTS WITH SORE THROAT

proven efficacy, narrow spectrum, cluded in the NAMCS since 1990. Visit classification of “antimicrobial agents.” safety, and low cost. characteristics include up to 3 diag- We included only nontopical antibacte- We used the National Ambulatory noses (coded using the International rial agents in the analysis. Because the Medical Care Survey (NAMCS) to de- Classification of Diseases, Ninth Revi- National Drug Code Directory has only 8 termine if antibiotic prescribing prac- sion, Clinical Modification), 6 entries for categories for antimicrobial agents, we tices for adults with sore throat have medications, and the type of visit (eg, further subdivided antibiotics into 37 dif- changed over time. We tracked changes acute, chronic follow-up). ferent classes based on pharmacology in the use of different antibiotic classes and route of administration. for sore throat by community primary Data Extraction Recommended treatment was de- care physicians from 1989 to 1999. In We limited our analysis to adults with fined as the use of penicillin, including addition, we evaluated predictors of pa- a primary reason for their visit of “symp- intramuscular preparations, or erythro- tients receiving an antibiotic for sore tom referable to the throat” (reason for mycin for patients who received antibi- throat and evaluated predictors of re- visit code, 1455; over 97% complained otics. If more than 1 antibiotic was used ceiving nonrecommended antibiotics of “soreness” or “pain” in the throat). We in a single visit (2.5% of visits), each an- for sore throat. included only visits to primary care phy- tibiotic was considered prescribed in its sicians, including those who identified respective class. For all other calcula- METHODS themselves as belonging to the special- tions and statistical tests, use of mul- National Ambulatory ties of family practice, general practice, tiple antibiotics within a single visit was Medical Care Survey internal medicine, adolescent medi- counted as a single antibiotic use. If a The NAMCS has been administered by cine, geriatrics, and general preventive patient received both a recommended the National Center for Health Statis- medicine. In all analyses, adolescent and a nonrecommended antibiotic, they tics annually since 1989. The NAMCS medicine physicians were included with were considered to have received only collects information on patient visits to family practice physicians; geriatric and a recommended antibiotic. nonfederally funded, office-based phy- general preventive medicine physi- To ensure the validity of our main sician practices throughout the United cians were included with internal medi- analysis, all weights and analyses were States and has a multistage probability cine physicians. recalculated after excluding any diagno- design in which sampling is based on We excluded patients younger than 18 sis—primary, secondary, or tertiary— geographic location, physician spe- years and patients immunosuppressed that could account for an antibiotic cialty, and visits within individual phy- with diabetes mellitus, human immu- prescription other than pharyngitis. Ex- sician practices. Each visit is weighted by nodeficiency virus infection, or cancer. cluded diagnoses included acute bron- the National Center for Health Statis- We also excluded patients with a pri- chitis, acute exacerbation of chronic tics to account for location, specialty, and mary diagnosis of sinusitis, an alternate bronchitis, staphylococcal infection, bac- practice nonparticipation. These weights cause of sore throat that might prompt terial infection, gonococcal infection, allow extrapolation to national figures for antibiotic treatment (7% of the sample lymphadenitis, otitis media, otitis ex- all aspects of the survey. in a preliminary analysis). We excluded terna, pneumonia, urinary tract infec- The NAMCS has collected informa- visits described as nonacute; chronic, tions, and skin infections. Exclusion of tion on 355354 patient visits between flare-up; chronic, routine; preopera- these diagnoses reduced the sample size 1989 and 1999. The participation rate tive; postoperative; or visits that were the by 17.5% from 2244 to 1852 for this sec- of physician practices in the NAMCS result of an injury. This final sample con- ondary analysis. has declined over the last 11 years from sisted of 2244 patient visits. 74% in 1989 to 63% in 1999.16,17 Statistical Analysis At each visit, patient, physician, and Data Analysis Ninety-five percent confidence inter- clinical information is collected. Pa- We calculated annual national rates of vals (CIs) for percentage estimates were tient information includes demograph- antibiotic prescriptions for patients with obtained by calculating relative SEs as ics, insurance status, and up to 3 rea- sore throat and the rate of nonrecom- recommended by the National Center sons for the visit. Physician information mended treatment for patients with sore for Health Statistics.1 Coefficients for includes self-selected specialty, geo- throat. We assessed changes in the dis- these calculations are dependent on graphic location, and if the practice is tribution of different antibiotic classes physician specialty and year. Because in a standard metropolitan statistical over 11 years. For ease of interpretabil- our analysis spans 11 years and 3 spe- area (SMSA). An SMSA approximates ity, we also grouped changes in the use cialties, we used the single largest rela- an urban region and is defined as an of antibiotics and nonrecommended an- tive SE available. We used modified area that includes a city of at least 50000 tibiotics into the periods of 1989- sample weights, according to the people or that has a total area popula- 1992, 1993-1996, and 1997-1999. method of Potthoff et al,19 to derive ef- tion of at least 100000 people. Infor- Antibiotics were identified according fective sample sizes that account for mation about SMSAs has been in- to the National Drug Code Directory18 clustering by physician practice.

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ANTIBIOTIC TREATMENT OF ADULTS WITH SORE THROAT

We used linear trends to assess munity physicians by adult patients were prescribed in 49% (95% CI, 46%- changes in antibiotic use over time. with sore throat between 1989 and 1999 52%) of visits. Among those receiving Multivariable logistic regression was that met our inclusion and exclusion antibiotics, 68% (95% CI, 64%-72%) re- used to evaluate independent predic- criteria. There was no significant change ceived nonrecommended antibiotics. tors of antibiotic prescription and use in the number of estimated annual vis- of nonrecommended treatments. In the its over 11 years. The most frequent first model, the use of any antibiotic was physician-reported diagnoses were Table 1. Characteristics of Adult Visits for Sore Throat to Primary Care Physicians, the dependent variable. In the second acute pharyngitis (41%), upper respi- 1989-1999 (n = 2244)* model, use of a nonrecommended an- ratory infection (21%), acute tonsilli- Proportion of Sample tibiotic among those receiving an an- tis (8%), streptococcal infection (6%), Characteristic With Characteristic, % tibiotic was the dependent variable. Be- acute bronchitis (5%), and acute lar- Age, mean (SD), y 38 (13) Sex, % male 35 cause patient age had a roughly linear yngitis (2%). Race/ethnicity effect on these 2 outcomes, we mod- The sample had a mean age of 38 White, non-Hispanic 79 eled patient age as the odds ratio (OR) years (SD,13) and was 35% male and Black, non-Hispanic 7 Hispanic 8 associated with each increasing de- 79% white, non-Hispanic (TABLE 1). Asian 4 cade of age. We evaluated interaction Forty-one percent of patients had pri- Insurance† Private 41 terms, none of which were significant vate insurance, 29% made some type of Medicare 9 or included in the final models. self-payment at their visit, and 23% be- Medicaid 8 Self-pay 29 All analyses were done with SAS sta- longed to a health maintenance orga- Health maintenance 23 tistical software, version 8.01 (SAS In- nization. Family practice physicians, in- organization Specialty stitute Inc, Cary, NC). P values less than ternal medicine physicians, and general Family practice 45 .05 were considered significant, ex- practice physicians accounted for 45%, Internal medicine 29 cept for univariate trend tests by year, 29%, and 25% of visits, respectively. General practice 25 Region where, because of multiple compari- Twenty-three percent of visits oc- Northeast 20 sons, P values less than .001 were con- curred in non-SMSA areas. Midwest 29 South 30 sidered significant. Over the 11-year period studied, an- West 21 tibiotics were prescribed in 73% (95% Non-SMSA area‡ 23 RESULTS CI, 70%-76%) of visits (TABLE 2). Rec- *Data for some variables do not equal 100% due to rounding. There were an estimated 6.7 million ommended antibiotics were pre- †Insurance categories are not mutually exclusive. ‡SMSA indicates standard metropolitan statistical area. (range by year, 5.1 million to 8.7 mil- scribed in 23% (95% CI, 20%-26%) of Because SMSA was not available in 1989, n = 1950 for lion) annual visits to office-based, com- visits. Nonrecommended antibiotics this characteristic only.

Table 2. Distribution of Antibiotic Classes Prescribed Each Year by Primary Care Physicians to Adults With Sore Throat, 1989-1999* Antibiotic Class 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Overall Recommended antibiotics, % of visits† Penicillin, oral and intramuscular 17 14 15 13 12 9 7 5 10 9 6 12‡ Erythromycin 15 16 14 16 15 11 6847512‡ Subtotal 32 30 29 29 27 20 13 13 14 16 11 23‡ Nonrecommended antibiotics, % of visits Aminopenicillins 27 25 18 24 20 22 27 23 21 23 16 23 Cephalosporins, oral 12 10 23 14 8 18 16 9 16 10 19 14 Extended macrolides§ 00013691213124 4‡ Tetracyclines 23744322113 3 Amoxicillin/clavulanic acid 21224142443 2 Trimethoprim/sulfamethoxazole 00121112211 1 Fluoroquinolones࿣ 02010112212 1 Extended fluoroquinolones¶ 00000000131 0‡ Other antibiotic classes# 41441232522 3 Subtotal 45 40 53 49 42 51 61 53 61 56 46 49‡ Any antibiotic prescribed, % of visits 77 70 82 78 69 71 74 66 74 72 57 73 *Columns may not add up exactly due to rounding and individuals receiving multiple antibiotics. †Recommended first-line antibiotics for adults with sore throat are penicillin or erythromycin according to the Infectious Diseases Society of America. ‡P for trend, Ͻ.001. §Extended macrolides prescribed were azithromycin, clarithromycin, and dirithromycin. ࿣Fluoroquinolones prescribed were ciprofloxacin and ofloxacin. ¶Extended fluoroquinolones prescribed were levofloxacin and trovafloxacin. #Other antibiotic classes include intravenous penicillins, parenteral cephalosporins, lincomycin, trimethoprim, sulfonamides, clindamycin, aminoglycosides, nitrofurantoin, and entries labeled as “antibiotic agent.” Each of these antibiotic classes was used in less than 1% of visits overall.

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ANTIBIOTIC TREATMENT OF ADULTS WITH SORE THROAT

Recommended antibiotic use de- There was a significant decrease in the and erythromycin (15% of visits in 1989 creased from 32% of visits in 1989 to 11% use of both penicillin (17% of visits in to 5% in 1999; P for trend, Ͻ.001). of visits in 1999 (P for trend, Ͻ.001). 1989 to 6% in 1999; P for trend, Ͻ.001), Nonrecommended antibiotic use increased from 45% of visits in 1989 to 56% of visits in 1998, but decreased to Table 3. Proportion of Adults With Sore Throat Given Any Antibiotic and Nonrecommended Antibiotics Over an 11-Year Period, by Primary Care Specialty 46% of visits in 1999 (P for trend, P Value Ͻ.001). There was a significant in- 1989-1992 1993-1996 1997-1999 Overall for Trend crease in the use of extended- Any antibiotic, % 76 71 69 73 .01 spectrum macrolides (P for trend, Family practice 76 69 72 73 .20 Ͻ.001) and extended-spectrum fluo- Internal medicine 74 64 60 68 .01 roquinolones (P for trend, Ͻ.001). General practice 76 80 77 78 .41 Antibiotics were prescribed in 76% Nonrecommended antibiotic of visits in 1989-1992, 71% of visits in for those receiving antibiotics, % 60 74 80 68 Ͻ.001 1993-1996, and 69% of visits in 1997- Family practice 60 73 77 67 Ͻ.001 1999 (P for trend, .01; TABLE 3). The Internal medicine 66 73 88 72 .003 proportion of visits in which nonrec- General practice 56 78 72 65 Ͻ.001 ommended antibiotics were given to patients who received an antibiotic increased from 60% in 1989-1992, to 74% Table 4. Adjusted Predictors of Antibiotic Use and Nonrecommended Antibiotic Use in 1993-1996, to 80% in 1997-1999 for Adults With Sore Throat, 1990-1999* (P for trend, Ͻ.001). Nonrecommended Antibiotic Use Internal medicine physicians signifi- Among Patients Receiving cantly decreased the proportion of vis- Any Antibiotic Use Antibiotics† its in which an antibiotic was pre- Adjusted Nonrecommended Adjusted scribed from 74% of visits in 1989- Antibiotic Odds Ratio Antibiotic Odds Ratio Variable Prescribed, % (95% CI) Prescribed, % (95% CI) 1992, to 64% in 1993-1996, to 60% in Calendar year . . . 0.98 (0.93-1.02) . . . 1.17 (1.11-1.24) 1997-1999 (P for trend, .01). How- Patient age, mean, y‡ 37 (vs 41) 0.86 (0.79-0.94) 39 (vs 36) 1.11 (0.99-1.24) ever, when prescribing an antibiotic, Sex, % male 37 (vs 33) 1.21 (0.93-1.56) 37 (vs 38) 0.89 (0.67-1.20) internal medicine physicians signifi- Race/ethnicity cantly increased their use of nonrec- White, non-Hispanic 73 Referent 70 Referent ommended antibiotics: from 66% of vis- Black, non-Hispanic 74 1.30 (0.77-2.17) 65 0.71 (0.40-1.24) its 1989-1992, to 73% in 1993-1996, to Hispanic 66 0.75 (0.48-1.18) 72 1.10 (0.62-1.96) 88% in 1997-1999 (P for trend, .003). Asian 68 0.96 (0.53-1.74) 56 0.51 (0.25-1.01) Family practice and general practice Insurance§ physicians also increased their use of Private 72 1.22 (0.89-1.68) 73 0.90 (0.62-1.32) nonrecommended antibiotics, but with- Medicare 66 1.25 (0.73-2.13) 81 1.27 (0.62-2.60) out a significant change in their over- Medicaid 68 0.91 (0.56-1.48) 71 1.09 (0.60-1.99) Self-pay 79 1.43 (0.97-2.11) 64 0.78 (0.51-1.18) all use of antibiotics. Health maintenance 67 0.89 (0.63-1.24) 66 0.66 (0.44-1.00) In multivariable modeling, antibi- organization otic use was independently predicted by Specialty patient age (OR, 0.86 per increasing de- Family practice 72 Referent 71 Referent cade; 95% CI, 0.79-0.94) and general Internal medicine 67 0.92 (0.68-1.24) 71 1.16 (0.80-1.68) practice specialty (OR, 1.54 compared General practice 79 1.54 (1.10-2.14) 66 0.92 (0.66-1.30) with family practice specialty; 95% CI, Region ABLE Northeast 70 Referent 68 Referent 1.10-2.14; T 4). Among patients Midwest 72 1.00 (0.70-1.44) 72 1.30 (0.84-2.00) who received antibiotics, nonrecom- South 78 1.38 (0.95-2.00) 71 1.25 (0.82-1.90) mended antibiotic use was predicted by West 67 0.89 (0.60-1.32) 64 0.95 (0.59-1.53) calendar year (OR, 1.17 per year; 95% Non-SMSA area 76 1.21 (0.89-1.64) 68 0.86 (0.61-1.20) CI, 1.11-1.24). Nonrecommended an- *Analysis based on 1950 adult visits for sore throat to primary care physicians for the years 1990-1999 because stan- tibiotic use was less likely with health dard metropolitan statistical area (SMSA) was not available in 1989. Antibiotics were prescribed in 1427 (73%) of the maintenance organization coverage (OR, 1950 visits. Odds ratios are adjusted for calendar year, age, sex, race, insurance status, physician specialty, region, and SMSA status. Adjusted odds ratios are based on weightings provided by the National Center for Health Statis- 0.66; 95% CI, 0.44-1.00). tics and cannot be calculated from raw numbers in the table. CI indicates confidence interval. †Nonrecommended antibiotics are those other than penicillin or erythromycin. To ensure the validity of our find- ‡Odds ratio for patient age is per increasing decade. ings, we reanalyzed the data after re- §Insurance categories are not mutually exclusive. moving 392 visits with alternate diag-

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ANTIBIOTIC TREATMENT OF ADULTS WITH SORE THROAT

noses that could potentially account for sicians.20 In our analysis, general prac- toms or history of allergy is not avail- antibiotic use. The percentage of visits tice specialty may represent an older able. We also do not have access to de- in which antibiotics were used de- group of physicians accustomed to pre- tailed clinical information, such as vital creased from 73% to 71% (95% CI, scribing antibiotics for upper respira- signs, physical examination signs, or the 68%-74%). The trend in use of ex- tory tract infections. test results for GABHS. As such, we can- tended fluoroquinolones no longer met The only predictor for the use of non- not definitively assess the appropriate- our criteria for significance (P=.002). In recommended antibiotics was calen- ness of antibiotic use for any indi- the multivariable models, the results dar year. This effect persisted despite vidual visit. Despite these limitations, changed modestly from the main analy- controlling for patient demographics, the disparity between the known preva- sis to the reanalysis. In the first model, location, physician specialty, and af- lence of GABHS among adults with sore age (OR, 0.83 per decade; 95% CI, 0.75- ter eliminating potential competing di- throat and the proportion of visits in 0.91), general practice specialty (OR, agnoses in our reanalysis. We found a which an antibiotic was used is strik- 1.45 compared with family practice spe- significant increase in the use of the ex- ing. This disparity persisted in a re- cialty; 95% CI, 1.01-2.07), and self- tended-spectrum macrolides and a analysis of our data after removing payment (OR, 1.62; 95% CI, 1.07- small, but statistically significant in- 17.5% of visits that had a potential al- 2.47) were significant predictors of crease in the use of the extended- ternative reason for antibiotic use. antibiotic use. In the second model, cal- spectrum fluoroquinolones. Our analysis of the NAMCS shows endar year (OR, 1.18; 95% CI, 1.10- While many broad-spectrum antibi- that over half of adult patients with sore 1.25) was the sole significant predictor otics are effective in eradicating GABHS throat are treated with antibiotics by of nonrecommended antibiotic use. from the throat,21,22 there are 2 main community primary care physicians. concerns in the use of these agents. The Educational and policy efforts to re- COMMENT first is cost. A course of azithromycin duce unnecessary antibiotic use may be In a national community sample of is roughly 20 times more expensive having some effect, as evidenced by a adults presenting to primary care phy- than a course of penicillin ($39.32 vs decrease in the proportion of patients sicians with a chief complaint of sore $2.31), and a course of oral cephalo- receiving antibiotics in 1999. How- throat, 73% were treated with antibiot- sporins ranges in price from $4.41 to ever, the use of expensive, broad- ics. This greatly exceeds the 5% to 17% $80.05.23 spectrum antibiotics that can induce re- prevalence of GABHS in adults with sore The second concern with the use of sistance is still frequent. Efforts should throat.3-9 From 1989 to 1999, there has broad-spectrum agents is develop- be continued to encourage appropri- been a marginally significant decrease in ment of bacterial resistance. Extended- ate antibiotic use by both patients and the use of antibiotics overall, but an in- spectrum fluoroquinolones are ca- physicians. crease in the use of more expensive, pable of inducing resistance among 24 Author Contributions: Study concept and design: broader-spectrum antibiotics. Encour- GABHS isolates in vitro. In vivo, mac- Linder, Stafford. agingly, there was a decrease in the use rolide resistance among GABHS iso- Analysis and interpretation of data: Linder, Stafford. 25,26 Drafting of the manuscript: Linder, Stafford. of almost all antibiotic classes in 1999. lates varies between 2% and 17% and Critical revision of the manuscript for important in- Predictors of antibiotic use for sore is proportional to local macrolide use.27 tellectual content: Linder, Stafford. throat were younger patient age and Fortunately, interventions aimed at Statistical expertise: Linder, Stafford. Administrative, technical, or material support: Linder, general practice specialty. Prescribing lowering macrolide use decreases the Stafford. antibiotics more frequently for younger prevalence of macrolide-resistant Study supervision: Stafford. 28 Funding/Support: Dr Linder was supported by Na- patients makes biologic sense. Younger GABHS isolates. tional Research Service Award 5T32PE11001-12. Dr patients are more likely to be infected In contrast, GABHS has remained Stafford was supported by Mentored Clinical Scien- tist Development Award K08-HL03548 from the Na- with GABHS, and some treatment al- exquisitely sensitive to penicillin. Over tional Heart, Lung and Blood Institute, Bethesda, Md. gorithms for sore throat incorporate the past 80 years, there has never Previous Presentation: Presented in part at the 24th Annual Meeting of the Society of General Internal younger patient age as a risk factor for been a GABHS isolate resistant to peni- Medicine, May 4, 2001, San Diego, Calif. GABHS infection.4,13 cillin found in Europe25 or North 26,29,30 We also found in multivariable analy- America. Despite a small risk of REFERENCES sis that general practice physicians pre- serious allergic reaction, penicillin is 1. Woodwell DA. National Ambulatory Medical Care scribed antibiotics about 50% more fre- generally well tolerated, inexpensive, Survey: 1998 Summary. Hyattsville, Md: National Cen- quently than family practice physicians. and effective. Penicillin reduces the du- ter for Health Statistics; 2000. Advance Data From Vi- 31 tal and Health Statistics, No. 315. According to the masterfiles of the ration of symptoms and is the only an- 2. Gwaltney JM Jr, Bisno AL. Pharyngitis. In: Man- American Medical Association, those tibiotic proven to prevent rheumatic dell GL, Bennett JE, Dolin R, eds. Mandell, Douglas 32-34 and Bennett’s Principles and Practice of Infectious Dis- identifying themselves as general prac- fever in patients with sore throat. eases. 5th ed. Philadelphia, Pa: Churchill Livingtone; tice physicians tend to be older than Our analysis has limitations that de- 2000:656-662. 3. McIsaac WJ, White D, Tannenbaum D, Low DE. other physicians: 47% are older than 65 serve comment. Detailed historical A clinical score to reduce unnecessary antibiotic use years, compared with 17% of all phy- information such as duration of symp- in patients with sore throat. CMAJ. 1998;158:75-83.

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ANTIBIOTIC TREATMENT OF ADULTS WITH SORE THROAT

4. McIsaac WJ, Goel V, To T, Low DE. The validity Survey: 1989 summary. National Center for Health Sta- 26. Kaplan EL, Johnson DR, Del Rosario MC, Horn of a sore throat score in family practice. CMAJ. 2000; tistics. Vital Health Stat 13. 1992;(110):1-80. DL. Susceptibility of group A beta-hemolytic strepto- 163:811-815. 17. National Center for Health Statistics. Public use cocci to thirteen antibiotics: examination of 301 strains 5. Komaroff AL, Pass TM, Aronson MD, et al. The pre- micro-data file documentation, National Ambulatory isolated in the United States between 1994 and 1997. diction of streptococcal pharyngitis in adults. J Gen Medical Care Survey, 1999: machine readable data Pediatr Infect Dis J. 1999;18:1069-1072. Intern Med. 1986;1:1-7. and documentation produced by the National Cen- 27. Seppala H, Klaukka T, Lehtonen R, Nenonen E, 6. Walsh BT, Bookheim WW, Johnson RC, Tomp- ter for Health Statistics. Hyattsville, Md: National Tech- Huovinen P. Outpatient use of erythromycin: link kins RK. Recognition of streptococcal pharyngitis in nical Information Service; Springfield, Va: US Dept of to increased erythromycin resistance in group adults. Arch Intern Med. 1975;135:1493-1497. Commerce; 1999. A streptococci. Clin Infect Dis. 1995;21:1378- 7. Kljakovic M. Sore throat presentation and man- 18. Food and Drug Administration. National Drug 1385. agement in general practice. N Z Med J. 1993;106: Code Directory, 1995 Edition. Washington, DC: Pub- 28. Seppala H, Klaukka T, Vuopio-Varkila J, et al, for 381-383. lic Health Service; 1995. the Finnish Study Group for Antimicrobial Resis- 8. Kaplan EL, Top FH Jr, Dudding BA, Wannamaker 19. Potthoff RF, Woodbury MA, Manton KG. “Equiva- tance. The effect of changes in the consumption of LW. Diagnosis of streptococcal pharyngitis: differen- lent sample size” and “equivalent degrees of free- macrolide antibiotics on erythromycin resistance in tiation of active infection from the carrier state in the dom” refinements for inference using survey weights group A streptococci in Finland. N Engl J Med. 1997; symptomatic child. J Infect Dis. 1971;123:490-501. under superpopulation models. J Am Stat Assoc. 1992; 337:441-446. 9. Poses RM, Cebul RD, Collins M, Fager SS. The ac- 87:383-396. 29. York MK, Gibbs L, Perdreau-Remington F, Brooks curacy of experienced physicians’ probability esti- 20. Pasko T, Seidman B, Birkhead S. Physician Char- GF. Characterization of antimicrobial resistance in Strep- mates for patients with sore throats: implications for acteristics and Distribution in the US: 2000-2001 Edi- tococcus pyogenes isolates from the San Francisco Bay decision making. JAMA. 1985;254:925-929. tion. Chicago, Ill: American Medical Association, De- area of northern California. J Clin Microbiol. 1999; 10. Bisno AL. Acute pharyngitis. N Engl J Med. 2001; partment of Physician Practice and Communications 37:1727-1731. 344:205-211. Information, Division of Survey and Data Resources; 30. Macris MH, Hartman N, Murray B, et al. Studies 11. Gonzales R, Steiner JF, Sande MA. Antibiotic pre- 2000. of the continuing susceptibility of group A strepto- scribing for adults with colds, upper respiratory tract 21. Hooton TM. A comparison of azithromycin and coccal strains to penicillin during eight decades. Pe- infections, and bronchitis by ambulatory care physi- penicillin V for the treatment of streptococcal phar- diatr Infect Dis J. 1998;17:377-381. cians. JAMA. 1997;278:901-904. yngitis. Am J Med. 1991;91(3A):23S-26S. 31. Zwart S, Sachs AP, Ruijs GJ, Gubbels JW, Hoes 12. Nyquist AC, Gonzales R, Steiner JF, Sande MA. 22. Pichichero ME, Margolis PA. A comparison of AW, de Melker RA. Penicillin for acute sore throat: ran- Antibiotic prescribing for children with colds, upper cephalosporins and penicillins in the treatment of domised double blind trial of seven days versus three respiratory tract infections, and bronchitis. JAMA. 1998; group A beta-hemolytic streptococcal pharyngitis: a days treatment or placebo in adults. BMJ. 2000;320: 279:875-877. meta-analysis supporting the concept of microbial 150-154. 13. Bisno AL, Gerber MA, Gwaltney JM Jr, Kaplan EL, copathogenicity. Pediatr Infect Dis J. 1991;10:275- 32. Denny FW, Wannamaker LW, Brink WR, Ram- Schwartz RH, for the Infectious Diseases Society of 281. melkamp CH, Custer EA. Prevention of rheumatic fe- America. Diagnosis and management of group A strep- 23. Drug Topics Red Book. 103rd ed. Montvale, NJ: ver: treatment of the preceding streptococcal infec- tococcal pharyngitis: a practice guideline. Clin Infect Medical Economics Data; 1999. tion. JAMA. 1950;143:151-153. Dis. 1997;25:574-583. 24. Boos M, Mayer S, Fischer A, et al. In vitro devel- 33. Wannamaker LW, Rammelkamp CHJ, Denny FW, 14. Snow V, Mottur-Pilson C, Cooper RJ, Hoffman JR. opment of resistance to six quinolones in Streptococ- et al. Prophylaxis of acute rheumatic fever by treat- Principles of appropriate antibiotic use for acute phar- cus pneumoniae, Streptococcus pyogenes, and Staphy- ment of the preceding streptococcic infection with vari- yngitis in adults. Ann Intern Med. 2001;134:506-508. lococcus aureus. Antimicrob Agents Chemother. 2001; ous amounts of depot penicillin. Am J Med. 1951;10: 15. Cooper RJ, Hoffman JR, Bartlett JG, et al. Prin- 45:938-942. 673-695. ciples of appropriate antibiotic use for acute pharyn- 25. Betriu C, Casado MC, Gomez M, Sanchez A, Pa- 34. Del Mar CB, Glasziou PP, Spinks AB. Antibiotics gitis in adults: background. Ann Intern Med. 2001; lau ML, Picazo JJ. Incidence of erythromycin resis- for sore throat [Cochrane Review]. In: The Cochrane 134:509-517. tance in Streptococcus pyogenes: a 10-year study. Di- Library. Issue 1, 2001. Oxford, England: Update Soft- 16. Schappert SM. National Ambulatory Medical Care agn Microbiol Infect Dis. 1999;33:255-260. ware.

Downloaded From: https://jamanetwork.com/ on 09/27/2021