Group A streptococcal (GAS) tonsillopharyngitis is one of the most common diagnoses in children and adults for which an antibiotic is prescribed.1 The 10-day duration of antibiotic therapy for GAS tonsillopharyngitis is based on well-conducted studies with depot penicillin.2,3 Duration of therapy with oral penicillin preparations was extrapolated from comparative studies of depot penicillin and orally administered penicillin in adults.4,5 However, 10-day treatment courses are associated with poor patient compliance.6–12 If shorter duration therapy for GAS tonsillopharyngitis is as effective as 10-day treatment, reducing the duration of treatment could have an important impact on improved compliance, fewer antibiotic adverse events, a reduction in the development of bacterial resistance and a positive impact on the pharmacoeconomics of treatment. Shorter antibiotic treatment courses for GAS tonsillopharyngitis have been evaluated during the past 10 years as an alternative to traditional 10-day courses of therapy. The objective of this study was to use metaanalysis to compare the relative efficacy of short course antibiotic treatment with standard 10-day treatment courses for GAS tonsillopharyngitis. Additionally compliance and adverse event rates of short course antibiotic treatment and 10-day treatment were compared.
Randomized, controlled trials comparing antibiotic treatment for 4–5 days with 10 days for GAS tonsillopharyngitis in children and adults were identified from Medline (1966–2003) and Embase (1974–2003) searches. The searches had no language restriction; the search terms used were streptococcal pharyngitis/tonsillitis, cephalosporin, macrolide, amoxicillin and penicillin. Reference lists of relevant publications were reviewed to identify additional trials.
Trial Selection and Quality.
Trials comparing 4- to 5-day versus 10-day antibiotic treatment of GAS tonsillopharyngitis infections were independently reviewed for inclusion by the authors according to the following criteria: (1) bacteriologic confirmation of GAS tonsillopharyngitis; (2) random assignment to antibiotic therapy for a β-lactam or macrolide (other than azithromycin) antibiotic of a 4- to 5-day course versus a 10-day comparator; and (3) assessment of bacteriologic outcome with the use of a throat culture after therapy. The Jadad scale was used to assess the quality of the included trials. The scale assigned scores from 0 to 5 (best quality trial) based on the following criteria: (1) study participants were allocated randomly to treatment using an appropriate method such as a random numbers table; (2) the intervention was double blinded; and (3) accounting and description of study withdrawals were done.13
Data Abstraction and Definition of Terms.
The primary outcomes of interest were: (1) bacteriologic cure defined as failure to isolate GAS by throat culture after completion of the antibiotic course; and (2) clinical cure, defined as failure to isolate GAS by throat culture and resolution of, or improvement in, the presenting signs and symptoms of GAS infection after completion of antibiotic treatment and continuing throughout follow-up. To test the robustness of the analysis, sensitivity analyses were performed to assess the impact of detailed clinical illness descriptions, compliance monitoring, GAS serotyping or genotyping and timing of the test-of-cure culture. Secondary outcomes of interest were compliance rates and antibiotic adverse event rates of the treatment courses. We independently abstracted primary and secondary outcomes, and sensitivity analysis data from each trial using a data extraction form. Differences were settled by discussion, and consensus was reached.
Metaanalyses were performed on the trials grouped by shortened course antibiotic. Shortened course cephalosporin trials were further grouped by the 10-day comparator, penicillin or the same cephalosporin. The metaanalyses were performed with the Cochrane Collaboration's Revman 4.2 program (Cochrane Collaboration, Oxford, U.K.). The bacterial and clinical cure rates after 4–5 days of treatment compared with a 10-day comparator were calculated and expressed as an odds ratio (OR) with 95% confidence intervals (CIs). An OR >1 indicated a higher bacteriologic and clinical cure rate for the 4- to 5-day treatment, as compared with the 10-day comparator. ORs were calculated for individual trial outcomes with the Peto fixed effects model, when there was trial homogeneity,14 and the DerSimonian and Laird random effects model, when there was trial heterogeneity.15 A summary OR was determined for trials grouped by short course antibiotic. Statistical heterogeneity among trials was assessed by χ2 analysis.16,17
Sensitivity analyses were conducted to assess the robustness of the overall metaanalysis of the cephalosporin short course versus penicillin 10-day trials and to investigate potential clinical heterogeneity among the trials by comparing summary ORs among groups redefined by (1) excluding trials that were not investigator or double blinded, (2) excluding trials of a lower methodologic quality (Jadad score ≤ 2), (3) excluding trials that did not give specific details of the clinical status of the patients, (4) excluding trials that did not monitor compliance, (5) excluding trials that did not perform serotyping or genotyping of the GAS organism isolated on the initial and follow-up throat culture and (6) excluding trials that did not perform the test-of-cure follow-up culture 3–14 days after completion of the antibiotic treatment. Additional analyses included (1) studies grouped by site (ie, U.S. and European), (2) studies grouped by cephalosporin generation and (3) studies of individual cephalosporins grouped and analyzed separately.
A funnel graph of the standard effect versus the OR was plotted to determine whether publication bias existed. Compliance rates and antibiotic adverse event rates were secondary outcomes analyzed by the same metaanalytic methods. ORs were calculated for individual trial compliance, and adverse events rates and summary ORs were determined for trials that reported the data.
Literature Search and Trial Inclusion.
The Medline and Embase searches yielded 144 citations; 39 were randomized clinical trials comparing a shortened course (4–5 days) of treatment with another treatment course for GAS tonsillopharyngitis. Two trials not identified by Medline or Embase were retrieved from reference listings. The 41 citations were assessed further according to the inclusion criteria. Fifteen of these trials were excluded from the metaanalysis for the following reasons: (1) the trial did not evaluate a β-lactam antibiotic or a macrolide other than azithromycin (n = 12); (2) the data presented were a republication of previous data already included in the metaanalysis (n = 1); or (3) the treatment lengths were the same (n = 2). Twenty-two trials remained.18–41 Two trials had 3 treatment arms: cefpodoxime 5 days versus penicillin 10 days or cefpodoxime 10 days22; and clarithromycin 15 and 30 mg/kg versus penicillin 10 days.37 The treatment arms were analyzed separately in this metaanalysis.
Two trials comparing 6 days of treatment of amoxicillin versus penicillin for 10 days40,41 were analyzed separately and not included in the main metaanalysis.
The mean quality score for the trials was 2.5 (SD 1.5–3.5), of a maximum score of 5; 39% of the trials were of higher quality (Jadad score, >2) (Table 1). Four trials were double blind,25,35,36,39 and the investigators were blinded to treatment allocation in 4 additional trials.22,27,29,30 Eighteen trials adequately described the reasons for patient dropouts, with nearly all patients dropped from analysis because GAS was not isolated on the initial throat culture.
Description of Trials.
Twenty-two trials involving 7470 patients were analyzed (Table 1). Trials were grouped by short courses of cephalosporins (n = 14), macrolides (other than azithromycin) (n = 6), penicillin (n = 2) and amoxicillin (n = 2). Cephalosporin trials were further grouped by the comparator, penicillin (n = 12) or the same cephalosporin (n = 3). Five trials were conducted in the United States22,27,29,38,43; the remainder were conducted in Europe. Nineteen of the 22 trials were conducted in an outpatient private practice setting; 3 trials did not specify the setting.20,21,26 All trials required isolation of GAS on throat culture and all but 2 trials20, 38 performed a rapid antigen test at enrollment, but patients were dropped from analysis if GAS was not isolated from the throat culture. Twelve trials gave detailed descriptions of patient signs and symptoms at enrollment.21,22,24,30,33–38,40,41 Seven trials did not perform serotyping or genotyping of the infecting GAS organism.19,21,26,30,32,34,36 When serotyping or genotyping was performed, true bacterial failures could be distinguished from reinfection with another serotype of GAS. When available, true bacterial failure rates were used in the metaanalysis calculations. One trial specifically discussed identification and elimination of carriers from analysis.38
The timing of the test-of-cure follow-up culture varied among the trials. Most trials had an early and late follow-up culture and reported bacterial and clinical cure rates for each evaluation time point. Two trials had 2 follow-up visits but reported only 1 overall bacterial and clinical cure rate.25,38 Thirteen trials obtained test-of-cure cultures between 3 and 14 days after antibiotic completion.18,19,22,27,29,31,34–37,39–41 This timing is optimal because it occurs late enough to avoid antibiotic interference with culture results and early enough to avoid data contamination with reacquisitions or new infections. When available, the early follow-up test-of-cure data were used in the metaanalysis whenever possible.
Eighteen trials gave detailed compliance monitoring methods that included parent/patient questionnaires, tablet counts, record cards and urine testing.18,21,22,24,25,27,29–31,33–41 Six studies evaluated compliance as a study outcome and gave additional statistical data.18,21,23,25,33,41
Antibiotic-induced adverse events during therapy were assessed in 15 trials.23–27,29–31,33–37,40,41
Outcome of Bacterial and Clinical Cure Rates.
The primary outcome analyzed was the bacterial cure rate, comparing a 4- to 5-day treatment course with a 10-day treatment course. The summary OR for bacterial cure for the individual short course treatments gave differing results. Treatment with a cephalosporin for 4 or 5 days was superior for bacterial cure rate when compared with 10 days of penicillin treatment (12 trials, 4301 patients), OR = 1.47 (95% CI 1.06–2.03) (P = 0.002) (Fig. 1).
Sensitivity analyses were conducted with the trials comparing cephalosporin short course treatment to 10 days of penicillin treatment to evaluate the robustness of the overall summary OR for the outcomes of bacterial and clinical cure rates. (Tables 2 and 3) Sensitivity analyses were not performed on the other groups of trials as there were too few trials.
As a method to examine the validity of the overall analysis, we grouped the trials excluded from each of the sensitivity analyses and analyzed those for persistence of statistical significance. For bacteriologic cure, the OR favored short course cephalosporins in all groups except the trials that did not perform GAS typing: for the 8 trials18–21,23,24,26,28 that were not investigator or double blind, the OR was 1.32 (95% CI 1.04–1.68), P = 0.02; for the 8 trials18–21,23,26–28 with a low quality score (</= 2), the OR was 1.40 (95% CI 1.12–1.76), P = 0.003; for the 9 trials18–20,23,25–29 that did not define the clinical signs and symptoms of the studied patients, the OR was 1.59 (95% CI 1.30–1.95), P < 0.00001; for the 5 trials19,20,23,26,28 that did not monitor compliance, the OR was 1.34 (95% CI 1.04–1.73), P = 0.03; for the 7 trials20,21,23–26,28 with test-of-cure assessment outside of the 3- to 14-day period, the OR was 1.29 (95% CI 1.02–1.64), P = 0.03. In the 3 trials19,21,26 that did not perform GAS typing, the OR favored neither treatment arm: OR was 0.87 (95% CI 0.42–1.83), P = 0.72. For clinical cure, the same analyses were performed. In all but 1 analysis, the OR favored short course cephalosporin: OR was 2.10 (95% CI 1.44–3.07), P = 0.0001 for the 7 trials18,19,21,23,24,26,28 that were not investigator or double blind; OR was 1.98 (95% CI 1.43–2.73), P < 0.00001 for the 7 trials18,19,21,23,26–28 with a low quality score (</= 2); OR was 1.71 (95% CI 1.29–2.26), P = 0.0002 for the 8 trials18,19,23,25–29 with poorly defined clinical signs and symptoms of the included patients; OR = 2.34 (95% CI 1.55–3.54), P < 0.0001 for the 4 trials19,23,26,28 that did not monitor compliance; OR was 1.83 (95% CI 1.27–2.61), P = 0.001 for the 6 trials21,23–26,28 with the test-of-cure assessment outside of the 3- to 14-day period. The OR favored neither treatment of the 3 trials19,21,26 that did not perform GAS typing; OR 1.24 (95% CI 0.44–3.46), P = 0.68.
Tests for statistical heterogeneity were performed for all of the analyses. There was statistical heterogeneity for the primary outcome of bacterial cure for 4- to 5-day cephalosporin treatment compared with 10-day penicillin treatment, P = 0.03. However, we noted statistical heterogeneity was not present when the short course cephalosporin trials were grouped into U.S. and European trials (P = 0.20 and P = 0.41). The difference between the U.S. and European trials was caused by different bacterial cure rates from the 10-day penicillin therapy, 77% versus 86%, respectively; P < 0.0001. The overall short course cephalosporin bacterial cure rate was not different between the U.S. and European trials, 89% versus 90%, P = 0.579.
We also noted that statistical heterogeneity for the 4- to 5-day cephalosporin result was not present when grouped by generation, P = 0.50 and P = 0.06 for second and third generation cephalosporins, respectively. Heterogeneity was not present when the same cephalosporin was compared as a 4- to 5-day and 10-day therapy; P = 0.23.
Cefadroxil,20 the only first generation cephalosporin evaluated, had bacterial cure rates that favored neither cefadroxil nor the 10-day penicillin therapy; OR 0.72 (95% CI 0.33–1.58), P = 0.41. There were 3 trials evaluating the second generation cephalosporin cefuroxime.19,24,28 Cefuroxime achieved superior bacterial cure rates; OR 1.52 (95% CI 1.16–2.00), P = 0.003. There were 3 third generation cephalosporins evaluated in 7 trials; cefpodoxime,18,21,22 cefixime23,26 and cefdinir.27,29 The third generation cephalosporins produced superior bacterial cure rates when compared with 10-day penicillin therapy; OR 1.77 (95% CI 1.08–2.90), P = 0.02.
Individual cephalosporins were analyzed separately. Cefadroxil,20 and cefotiam25 were evaluated in one trial. Cefixime23,26 and cefdinir27,29 were evaluated by 2 trials, cefuroxime19,24,28 and cefpodoxime18,21,22 in 3 trials. Cefadroxil, cefotiam and cefixime did not demonstrate inferiority or superiority to 10 days of penicillin treatment: OR 0.72 (95% CI 0.33–1.58); OR .95 (95% CI 0.37–2.44); and OR 0.62 (95% CI 0.27–1.45), respectively. Cefdinir, cefpodoxime and cefuroxime were superior to 10 days of penicillin treatment: OR 2.39 (95% CI 1.67–3.41). P = 0.02; OR 2.21(95% CI 1.31–3.74), P = 0.007; and OR 1.52 (95% CI 1.16–2.00), P = 0.003, respectively.
Three trials (773 patients) compared 4- or 5-day treatment to 10-day treatment with the same cephalosporin. There was a trend toward superior bacterial cure rates with 10-day treatment: OR 0.70 (95% CI 0.42–1.14), P = 0.15 (Fig. 2). However, statistical power resulting from small sample size limited the strength of conclusion for this analysis. Heterogeneity was not significant; P = 0.98.
Results from trials comparing macrolide therapy for 5 days (6 trials, 1673 patients) with 10 days of penicillin treatment favored neither therapy; OR = 0.79 (95% CI 0.59–1.06), P = 0.11 (Fig. 3). Statistical power resulting from small sample size limited the strength of conclusion for this analysis. Heterogeneity was not significant; P = 0.14.
Two trials involving 309 patients, comparing 5 days of penicillin therapy with 10 days, showed short course penicillin therapy to be inferior in achieving bacterial cure; OR 0.29 (95% CI 0.13–0.63); P = 0.002 (Fig. 4). Heterogeneity was not significant; P = 0.93.
Six day amoxicillin therapy compared with 10-day penicillin therapy was evaluated in 2 trials with 511 patients.40,41 Results from the combined analysis favored neither therapy; OR 0.89 (95% CI 0.52–1.53), P = 0.7. Small sample size limited the statistical power and the strength of conclusion for this analysis.
Two trials20,38 did not report data on clinical cure. Therefore the primary outcome of clinical cure rate comparing a 4- to 5-day treatment course with a 10-day treatment course was assessed in 20 trials. Again the trials were grouped by short course of cephalosporins (13 trials, 4961 patients), macrolides (other than azithromycin) (6 trials, 1673 patients) and penicillin (1 trial, 137 patients). The cephalosporin trials were further grouped by the differing 10-day comparator. The summary ORs for clinical cure rate mirrored the bacterial cure rate results: 4 or 5 days of cephalosporin therapy was superior to 10 days of penicillin therapy [OR 1.50 (95% CI 1.16–1.93), P = 0.002], and 5 days of penicillin was inferior to 10 days [OR = 0.28 (95% CI 0.08–0.91), P = 0.03]. Small sample size limited the conclusions and statistical power of the short course macrolide trials and the trials of 4 or 5 days of cephalosporin therapy compared with 10 days. There was no significant statistical heterogeneity for any of the analyses of the clinical cure rates: cephalosporin trials versus 10 days of penicillin therapy P = 0.10 and versus 10 days of cephalosporin therapy P = 0.98; macrolide trials P = 0.86; and the sole penicillin trial had no determinable heterogeneity.
Specific compliance monitoring methods were used by 16 trials (Table 1),18,21,22,24,25,27,29–31,33–39 and 6 trials gave compliance rates for each of the treatment arms.21,22,25,28,33,39 The summary OR for compliance rate significantly favored the short course treatment [OR 3.10 (95% CI 2.22–4.32), P < 0.00001] (Fig. 5).
Adverse events associated with the antibiotic therapy were reported in 15 of 22 trials. The majority of events were gastrointestinal complaints. Eleven trials25–27,29,31,33–35,37,40,41 reported specific rates for both treatment arms. The rate of adverse events was equivalent between short course treatment and the 10-day treatment comparator [OR 1.01 (95% CI 0.80–1.27), P = 0.9]. One trial25 found significantly lower adverse events in the short course treatment group.
Statistical heterogeneity was assessed for the secondary outcomes of compliance and adverse event rates, and this approached significance at P = 0.06 for both.
Funnel plot analysis revealed no asymmetry (data not shown). Therefore publication bias, or bias introduced by smaller studies, was not found.
This metaanalysis shows that superior bacteriologic cure rates can be achieved with 4 or 5 days of treatment with an oral cephalosporin as compared with penicillin treatment of 10 days. Analysis of trials to date do not favor either short course macrolide therapy compared with 10-day penicillin therapy or short course treatment of a cephalosporin compared with 10-day cephalosporin treatment. With metaanalysis, 5 days of penicillin is confirmed to be inferior to 10 days of treatment. The 2 amoxicillin trials comparing 6 days of therapy were small, and there was insufficient power to conclude that there was a significant difference with 10 days of penicillin treatment.
There are important issues in interpreting the results of this metaanalysis. First, the trials were grouped according to the individual antibiotic (penicillin V) or the class of antibiotic (macrolides and cephalosporins). This grouping allowed trials evaluating the same antibiotic or antibiotics with similar pharmacokinetic/pharmacodynamic properties to be assessed together. Trials evaluating azithromycin were not included in this metaanalysis given that the action of azithromycin is not of short duration even though the antibiotic is administered for 5 days or fewer. We have analyzed and reported those results in another publication.42 There were 6 trials in the macrolide group, studying 4 different macrolides (erythromycin estolate, josamycin, clarithromycin and telithromycin), grouped together. The cephalosporin group included 7 different cephalosporins (1, first generation; 2, second generation; and 4, third generation).
The 22 trials included in this metaanalysis were randomized, prospective, trials; however, they were of varying quality according to the Jadad scale and used different methodologic designs. Sensitivity analysis was used to address this issue. The bacterial cure rates of short course cephalosporins as compared with 10-day penicillin therapy was remained superior in every analysis. When we performed a subset analysis on the 6 subsets of trials that were not in the sensitivity analysis of bacterial and clinical cure rate, 5 of 6 subsets maintained statistical significance, and 1 subset analysis was not significant. This approach enhances the statistical robustness of the overall analysis and strengthens the confidence about the appropriateness of combining all of the studies into a single metaanalysis.
In GAS tonsillopharyngitis, as in other community-acquired upper respiratory infections such as acute otitis media and sinusitis, one of the most common reasons for treatment failure is poor patient compliance.6–12 Therefore it is important that a compliance-enhancing antibiotic regimen be chosen. Poor compliance appears to be more common with longer treatment courses, because patients tend to stop their medication once symptoms have resolved.10 This typically occurs after 3–5 days in GAS tonsillopharyngitis. Given this reality of patient behavior, it is quite rational to prescribe short course therapy if it is as effective as the standard 10-day antibiotic treatment course. A metaanalysis comparing shortened course treatment to 10 days of treatment for acute otitis media in children found that the short course treatments were as effective as the 10-day treatment.43 The results of this metaanalysis show that short course treatment of GAS tonsillopharyngitis, another common upper respiratory infection, is as effective as, or even more effective than, the standard 10-day penicillin treatment if a cephalosporin is used.
Adverse events from antibiotic therapy are a common reason for poor patient compliance and can cause overall increased treatment costs due to additional physician visits, medications and monitoring. Shortened course treatment has been shown to reduce adverse events associated with antibiotic therapy.44,45 Although the overall the adverse event rates were equivalent in this metaanalysis between the short course treatment and the 10-day comparator, several of the trials included in this metaanalysis demonstrated fewer adverse events with the short course treatment arm.
Another important advantage of shortened therapy is the reduced impact on development of antibiotic resistance and nasopharyngeal colonization with resistant bacteria. Guillemot et al46 showed that inappropriate low dosages of aminopenicillins or third generation cephalosporins in children (as may occur in the latter half of a 10-day treatment course when patient compliance falls) is associated with an increased risk of penicillin-resistant Streptococcus pneumoniae nasopharyngeal carriage. Additionally their findings showed that longer duration of treatment was associated with increased risk of resistant S. pneumoniae carriage. Short course high dose therapy for GAS tonsillopharyngitis has the potential to have a positive impact on this important complication of antibiotic treatment.
Metaanalysis as a statistical tool has shortcomings. Specifically, a metaanalysis can incorporate existing biases from the included trials and introduce additional new biases.47–49 To minimize bias during trial selection, we used predetermined inclusion criteria. Publication bias was assessed,50 and no significant bias was evident. Statistical and clinical heterogeneity is a potential hazard of metaanalyses. Even though the trials included in this metaanalysis evaluated different antibiotics in both children and adults, significant heterogeneity was found only with the cephalosporin class and was eliminated when trials were grouped and analyzed by U.S. versus non-U.S. sites. Lastly statistical power was limited in several of the analyses because of small sample size; larger samples or additional studies would be of value in clarifying the comparator efficacy of short courses of macrolides and amoxicillin.
In conclusion, this metaanalysis shows that short course treatment of GAS tonsillopharyngitis can be more effective when prescribing 4 or 5 days of cefdinir, cefpodoxime or cefuroxime treatment than the standard 10-day treatment of penicillin. In the United States, 3 antibiotics have an indication for short course treatment of GAS tonsillopharyngitis, cefdinir, cefpodoxime and azithromycin, and meet the medicolegal standard of care. A 5-day treatment with cefuroxime was more effective than a 10-day penicillin treatment; however, it is not approved in the United States for a shortened course of treatment. The data to date are insufficient to recommend short course macrolide or amoxicillin treatment. Shortened course antibiotic therapy for GAS tonsillopharyngitis has the potential to improve the treatment efficacy by producing better compliance and fewer adverse events.
1. McCraig LF, Besser RE, Hughes JM. Trends in antimicrobial prescribing rates for children and adolescents. JAMA
2. Denny FW, Wannamaker LW, Brink WR, Rammelkamp CH Jr, Custer EA. Prevention of rheumatic fever. JAMA
3. Wannamaker LW, Rammelkamp CH Jr, Denny FW, et al. Prophylaxis of acute rheumatic fever by treatment of the preceding streptococcal infection with various amounts of depot penicillin. Am J Med
4. Wannamaker LW, Denny FW, Perry WD, et al. The effect of penicillin prophylaxis on streptococcal disease rates and the carrier state. N Engl J Med
5. Breese BB. Treatment of β-hemolytic streptococcic infections in the home. JAMA
6. Mohler DN, Wallin DG, Dreyfus EG. Studies in the home treatment of streptococcal disease, I: failure of patients to take penicillin by mouth as prescribed. N Engl J Med
7. Bergman AB, Werner RJ. Failure of children to receive penicillin by mouth. N Engl J Med
8. Leistyna JA, Macaulay JC. Therapy of streptococcal infections: do pediatric patients receive prescribed oral medication? Am J Dis Child
9. Charney E, Bynum R, Eldredge D, et al. How well do patients take oral penicillin? A collaborative study in private practice. Pediatrics
10. Green JL, Ray SP, Charney E. Recurrence rate of streptococcal pharyngitis related to oral penicillin. J Pediatr
11. Schwartz RH, Rodriquez WJ, Grundfast KM. Pharmacologic compliance with antibiotic therapy for acute otitis media: influence on subsequent middle ear effusion. Pediatrics
12. Finney JW, Friman PC, Rapoff MA, Christophersen ER. Improving compliance with antibiotic regimens for otitis media. Am J Dis Child
13. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials
14. Peto R, Collins R, Gray R. Large-scale randomized evidence: Large, simple trials and overviews of trials. J Clin Epidemiol
15. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
16. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ
17. Deeks JJ, Altman DG, Bradburn MJ. Fifteen statistical methods for examining heterogeneity and combining results from several studies in meta-analysis. In: Egger M, Smith G, Attman D, eds. Systematic Reviews in Health Care.
2nd ed. Oxford, United Kingdom: Blackwell BMJ Books; 2001:285–312.
18. Portier H, Chavanet P, Gouyan JB, Guetar F. Five day treatment of pharyngotonsillitis with cefpodoxime proxetil. J Antimicrob Chemother
. 1990;26(suppl E):79–85.
19. Gehanno P, Chiche E. Traitement des angines a streptocoque β hemolytique du groupe A par le cefuroxime axetil pendant 4 jours: etude comparative a la penicilline V pendant 10 jours. Med Mal Infect
20. Milatovic D. Evaluation of cefadroxil, penicillin, and erythromycin in the treatment of streptococcal tonsillopharyngitis. Pediatr Infect Dis J
21. Portier H, Chavanet P, Waldner-Combernoux A, et al. Five versus ten days treatment of streptococcal pharyngotonsillitis: a randomized controlled trial comparing cefpodoxime proxetil and phenoxymethyl penicillin. Scand J Infect Dis
22. Pichichero ME, Gooch WM, Rodriguez M, et al. Effective short-course treatment of acute group A β-hemolytic streptococcal tonsillopharyngitis. Arch Pediatr Adolesc Med
23. Peyramond D, Tigaud S, Bremard-Oury C, Scheimberg A. Multicenter comparative trial of cefixime and phenoxymethylpenicillin for group A β-hemolytic streptococcal tonsillitis. Curr Ther Res
24. Aujard Y, Boucot I, Brahimi N, Chiche D, Bingen E. Comparative efficacy and safety of four-day cefuroxime axetil and ten-day penicillin treatment of group A β-hemolytic streptococcal pharyngitis in children. Pediatr Infect Dis J
25. Carbon C, Chatelin A, Bingen E, et al. A double-blind randomized trial comparing the efficacy and safety of a 5-day course of cefotiam hexetil with that of a 10-day course of penicillin V in adult patients with pharyngitis caused by group A β-haemolytic streptococci. J Antimicrob Chemother
26. Adam D, Group Cefixime Study, Hostalk U, Troster K. Five-day cefixime therapy for bacterial pharyngitis and/or tonsillitis: comparison with 10-day penicillin V therapy. Infection
. 1995;23(suppl 2):S83–S86.
27. Tack KJ, Henry DC, Gooch WM, et al. Five-day cefdinir treatment of streptococcal pharyngitis. Antimicrob Agents Chemother
28. Adam D, Scholz H, Helmerking M. Comparison of short-course (5 day) cefuroxime axetil with a standard 10 day oral penicillin V regimen in the treatment of tonsillopharyngitis. J Antimicrob Chemother
29. Tack KJ, Hedrick JA, Rothstein E, et al. A study of 5-day cefdinir treatment for streptococcal pharyngitis in children. Arch Pediatr Adolesc Med
30. Esposito S, Noviello S, Iannielo F, D'Errico G. Short-course therapy with cefaclor for treatment of streptococcal pharyngotonsillitis. Int J Antimicrob Agents
31. Mehra S, VanMoerkerke M, Melck J, et al. Short course therapy with cefuroxime axetil for group A streptococcal tonsillopharyngitis in children. Pediatr Infect Dis J
32. Portier H, Lucht F, Lescale O, et al. Josamycine 5 jours versus penicilline V 10 jours dans le traitement des angines aigues a streptocoque β-hemolytique du groupe A. Med Mal Infect
33. Adam D, Scholz H, Pharyngitis Study Group. Five days of erythromycin estolate versus ten days of penicillin V in the treatment of group A streptococcal tonsillopharyngitis in children. Eur J Clin Microbiol Infect Dis
34. Portier H, Filipecki J, Weber F, et al. Five day clarithromycin modified release versus 10 day penicillin V for group A streptococcal pharyngitis: a multi-centre, open-label, randomized study. J Antimicrob Chemother
35. Norrby SR, Rabie WJ, Bacart P, et al. Efficacy of short-course therapy with the ketolide telithromycin compared with 10 days of penicillin V for the treatment of pharyngitis/tonsillitis. Scand J Infect Dis
36. Takker U, Dzyublyk O, Busman T, Notario G. Comparison of 5 days of extended-release clarithromycin versus 10 days of penicillin V for the treatment of streptococcal pharyngitis/tonsillitis: results of a multicenter, double-blind, randomized study in adolescent and adult patients. Curr Med Res Opin
37. Syrogiannopoulos GA, Bozdogan B, Grivea IN, et al. Two dosages of clarithromycin for five days, amoxicillin/clavulanate for five days or penicillin V for ten days in acute group A streptococcal tonsillopharyngitis. Pediatr Infect Dis J
38. Gerber MA, Randolph MF, Chanatry J, et al. Five vs. ten days of penicillin V therapy for streptococcal pharyngitis. Am J Dis Child
39. Stromberg A, Schwan A, Cars O. Five versus ten days treatment of group A streptococcal pharyngotonsillitis: a randomized controlled clinical trial with phenoxymethyl-penicillin and cefadroxil. Scand J Infect Dis
40. Clamorange French Study Group.Peyramond D, Portier H, Geslina P, Cohen R, Six-day amoxicillin versus 10-day penicillin V for group A β-haemolytic streptococcal acute tonsillitis in adults: a French multicentre, open-label, randomized study. Scand J Infect Dis
41. Cohen R, Levy C, Doit C, et al. Six-day amoxicillin vs. ten-day penicillin V therapy for group A streptococcal tonsillopharyngitis. Pediatr Infect Dis J
42. Casey JR, Pichichero ME. Higher dosages of azithromycin are more effective in group A streptococcal tonsillopharyngitis treatment. Clin Infect Dis
43. Kozyrskyj AL, Hildes-Ripstein E, Longstaffe SEA, et al. Treatment of acute otitis media with a shortened course of antibiotics: a meta-analysis. JAMA
44. Gooch WM, Blair E, Puopolo A, et al. Effectiveness of five days of therapy with cefuroxime axetil suspension for treatment of acute otitis media. Pediatr Infect Dis J
45. Gehanno P, Taillebe M, Denis P, al et. Short course cefotaxime compared with 5-day co-amoxyclav in acute otitis media in children. J Antimicrob Chemother.
46. Guillemot D, Carbon C, Balkau B, Geslin P, Lecoeur H, et al. Low dosage and long treatment duration of β-lactam: risk factors for carriage of penicillin resistant Streptococcus pneumoniae. JAMA
47. Moher D, Cook DJ, Eastwood S, et al. Improving the quality of reports from meta-analyses of randomized controlled trials: the QUOROM statement. Lancet
48. Moher D, Pham B, Jones A, et al. Does quality of reports of randomized trials affect estimates of intervention efficacy reported in meta-analyses? Lancet
49. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by simple, graphical test. BMJ
50. Schultz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA
51. Casey JR, Pichichero ME. In: Interscience Conference on Antimicrobial Agents and Chemotherapy, October 30–November 2, 2004, Washington, DC
. Washington, DC: American Society for Microbiology; 2004.