Objective. To evaluate the efficacy of oral penicillin V and intramuscular benzathine penicillin G (BPG) in eradicating group A streptococci from the upper respiratory tract.
Methodology. Two randomized, single-blind, multicenter antibiotic efficacy trials in children using recommended doses of either oral penicillin V or intramuscular BPG for treatment of acute-onset pharyngitis associated with isolation of group A streptococci were conducted. Throat examinations and cultures were obtained at enrollment and on days 5 to 8, 10 to 14, and 29 to 31.
Results. Thirty-five percent of 284 evaluable patients treated with oral penicillin V and 37% of BPG-treated patients were microbiologic treatment failures at either 10 to 14 or 29 to 31 days.
Conclusions. Although these findings do not provide sufficient evidence to change current treatment recommendations or public health policy, important questions are raised about currently recommended penicillin doses, about the role of the carrier state, and possibly about adequate bioavailability of intramuscular BPG. These findings require confirmation.
Although occasionally questioned, the efficacy of penicillin in eradicating group A streptococci from the upper respiratory tract and in preventing nonsuppurative sequelae has not been convincingly challenged. Some have reported equal microbiologic efficacy of other antibiotics, including cephalosporins and macrolides (several even in shortened courses of oral therapy). Yet, penicillin has remained the recommended antibiotic of choice for treating these infections according to guidelines published by respected national and international advisory bodies.1–4
Nevertheless, questions about the efficacy of penicillin for this common infection continue to be raised. Criticisms primarily have been based either on eradication rates of the organism from the upper respiratory tract or on recent reports of reduced serum penicillin levels after administration of intramuscular repository benzathine penicillin G (BPG).5–7 Other reports mention compliance problems associated with oral penicillin.8Intramuscular BPG eliminates the problem of compliance and theoretically provides therapeutic serum penicillin levels for longer than the conventional 10-day course of oral penicillin V. However, virtually all recent antibiotic efficacy studies have included only oral penicillin for comparison, although the original supporting studies used a parenteral formulation to demonstrate that penicillin therapy of streptococcal throat infection effectively prevents rheumatic fever.9
The opportunity to carefully further examine this issue was presented during 2 large, multicenter studies of children with acute-onset symptomatic pharyngitis and recovery of group A streptococci by throat culture. The original intended objective was to address the patient compliance problem by examining the efficacy of a single injection of 500 mg of intramuscular ceftriaxone. Two separate trials were conducted: the first compared intramuscular ceftriaxone with oral penicillin V given for 10 days in doses recommended by the American Academy of Pediatrics and the American Heart Association.1,,2 The second concurrently conducted study, using an essentially identical protocol, compared the same dose of intramuscular ceftriaxone with intramuscular BPG, also in recommended doses.1,,2
Although ceftriaxone was examined during the clinical trials, this report focuses specifically on the microbiologic efficacy (ie, eradication of group A streptococci from the throat) of the 2 penicillin preparations. The data reveal lower than expected streptococcal eradication rates from the upper respiratory tract after recommended doses of either oral penicillin V or intramuscular BPG.
The studies were conducted in 1994 to 1995 according to recognized published joint Infectious Diseases Society of America (IDSA) and US Food and Drug Administration (FDA) clinical guidelines for enrollment of participants and evaluation of anti-infective drugs for the treatment of respiratory tract infections.10 That document states: “Signs and symptoms of pharyngitis or tonsillitis of acute onset include sore throat and evidence on physical examination of inflammation of the uvula and pharynx or tonsils, including erythema, often with edema of the tissues with or without exudate. Fever may or may not be present.”10 The mean onset of symptoms in the participants reported here was only 1.7 days before enrollment, reducing the likelihood that these participants represented long-standing or chronic upper respiratory tract illness.
The protocol for these prospective, randomized, and investigator-blinded studies is shown in Fig 1. The study protocols were approved by institutional review boards at each of the clinical sites. Written informed consent was obtained; a total of 1259 patients (2–12 years) were enrolled at 42 geographically separate clinical sites in the United States during a 15-month period (January 1994–March 1995). After initial detection of group A streptococci by rapid antigen detection test (Test Pak; Abbott Laboratories, Abbott Park, IL) at the enrollment visit (baseline; V1), patients in the 2 concurrently conducted studies were randomized into antibiotic treatment groups using a table of random numbers. Those in the oral penicillin study received either a single intramuscular injection of ceftriaxone (500 mg) or recommended doses of oral penicillin V (250 mg 3 times daily for 10 days). Participants in the intramuscular penicillin study received either the single 500-mg intramuscular ceftriaxone injection or a single intramuscular injection of BPG also in recommended doses (600 000 units if body weight <60 pounds; 1.2 million units if body weight >60 pounds).1,,2 At each of 3 follow-up visits (V2: 5–8 days after the initial visit; V3: 10–14 days; V4: 29–31 days), a throat culture was obtained and an examination of the throat was conducted. At V1 and again at V4 sera were obtained for acute and convalescent antistreptolysin O (ASO) and anti-deoxyribonuclease B (anti-DNase B) titers. Antibiotic susceptibility minimum inhibitory concentrations (MICs) were determined by the broth microdilution method following the National Committee for Laboratory Standards.11
Throat swabs were placed in transport medium (Culturette, Becton-Dickinson, Baltimore, MD) and sent by overnight express courier to a central reference laboratory (World Health Organization Collaborating Center for Reference and Research on Streptococci, Department of Pediatrics, University of Minnesota Medical School). Specimens were immediately plated on to 5% sheep blood agar plates, incubated at 37°C, and examined at 24, 48, and 72 hours for the presence of β-hemolytic streptococci. Single colonies were subcultured, incubated, and then frozen (−20°C) in blood broth for later characterization. Blood samples were spun down at each clinical site and the serum separated before freezing at −20°C. Frozen serum samples were shipped on dry ice on a regular basis to the reference laboratory where they were kept frozen until analyzed. Acute and convalescent sera from each participant were analyzed at the same time for the 2 streptococcal antibodies. As in previous reports from our laboratory, significant rise in antibody titer was defined as an increase of ≥ 0.2 log10 units (2-tube dilution).12,,13 Conventional laboratory methods were used for processing throat isolates, for serologic grouping, T-agglutination pattern, and M/OF serotyping, and also for determining ASO and anti-DNase B titers.13
The primary intended measure of antibiotic efficacy in these analyses, microbiologic eradication, was defined as eradication of the concordant serotype of group A β-hemolytic streptococci at follow-up visits. Microbiologic failure was defined for the ceftriaxone treated patients as isolation of the baseline (concordant) serotype at any follow-up visit after completion of therapy (V2, V3, or V4) and for the 2 penicillin-treated groups as isolation of the concordant serotype at V3, V4, or both. (Fig 1)
Subjects from all clinical sites were included in these analyses only if they met the published IDSA/FDA recommended clinical criteria for acute pharyngitis, were compliant, and completed all follow-up visits and laboratory studies. This is in agreement with suggested criteria for proper analysis of pharyngitis studies.14
Statistical analyses were performed using Fisher's exact test (InStat for Macintosh version 2.0. GraphPad Software, San Diego, CA).
There were no significant differences among the 3 treatment groups in the distribution of age, sex, or mean body weight. Analysis of participant ages shows a normal distribution with a median age of 7 years (14% of participants) and a range from 2 years to 12 years (Fig 2A) The months of highest enrollment were in winter and early spring, as one would expect for group A streptococcal (GAS) pharyngitis in temperate climates (Fig 2B).
Among 1259 participants enrolled in the 2 trials, 1123 met the criteria for evaluation of the microbiologic efficacy (136 patients were eliminated from analysis either because they did not meet the strict IDSA published guidelines for enrollment or because they were noncompliant with the oral drug, missed a convalescent visit, or were missing culture or serum specimens). Thirty-seven percent of the 284 evaluable participants treated with oral penicillin V had a positive throat culture (concordant serotype) at the time of at least 1 follow-up visit. However, if only posttreatment visits (ie, after completion of 10 days of oral penicillin V) are considered, 100 (35%) of the 284 participants still harbored the concordant streptococcal serotype and therefore were classified as microbiologic failures (Fig 3A).
Although 42% of 271 participants treated with intramuscular BPG still had a concordant GAS isolate present in the throat during at least 1 of the 3 follow-up visits, it is more important to look specifically at the 10- to 14-day (V3) and the 29- to 31-day (V4) follow-up visits because these intervals extend beyond the 9-day period during which the organism should be eradicated to most effectively prevent rheumatic fever15 (Fig 3B). Thirty-seven percent of these 271 participants had a positive throat culture for the concordant serotype at either or both of these latter 2 convalescent visits. In fact, even at the 29 to 31 day visit, 25% still harbored the concordant GAS serotype.
Because of the possible influence of body weight on the dose and efficacy of antibiotic, we carefully analyzed this variable. There was no correlation between the antibiotic dose in mg (or units) per kg of patient body weight and the bacteriologic eradication rate for any of the 3 antibiotics (data not shown).
To attempt to determine the possible influence of streptococcal carriers, ASO, and anti-DNase B titers from participants who were microbiologic failures and those from whom the organism was successfully eradicated were also examined and compared. There was no significant difference in geometric mean baseline (V1) antibody titers between those demonstrating successful eradication and those who were microbiologic treatment failures. Similarly, analysis of rises in titers among the treatment groups failed to provide a credible explanation for the high failure rates (eg, differentiate carriers) although there were some differences noted. Fifty-eight percent (59/101) of BPG treatment failure patients and 52% (88/170) of bacteriologic treatment success patients showed a significant antibody response (≥0.2 log10increase) to either streptolysin O or DNase B. This difference was not statistically significant. Seventy percent (70/100) of oral penicillin treatment failure patients developed antibody to either of these streptococcal antigens compared with 52% (95/184) of treatment success patients. This higher antibody response by treatment failure patients was statistically significant (P = .0037) and was the opposite of what one might expect if carriers were the primary reason for the high percent of treatment failures.
Presentation with fever at visit 1 in patients who were bacteriologic treatment failures was compared with patients whose antibiotic treatment was successful. In patients treated with BPG, 40% of treatment failure patients had fever at visit 1 compared with 38% of bacteriologic treatment successes. In patients treated with oral penicillin, the percentages were 51% and 48%, respectively. The presence or absence of fever did not seem helpful in distinguishing streptococcal carriers.
MIC determinations were done on GAS isolates from 202 randomly selected treatment failure patients. The penicillin MIC90for all of these isolates was 0.016 μg/mL, and the MIC range was from 0.0008 to 0.032 μg/mL.
Of the 568 evaluable participants who received the single 500-mg intramuscular injection of ceftriaxone, 54% had at least 1 positive throat culture (concordant serotype) at a follow-up visit. All strains tested from this study were susceptible (in vitro) to ceftriaxone.
No participants developed suppurative or nonsuppurative sequelae of GAS infection.
Although the focus of this report is on the efficacy of penicillin for treatment of GAS pharyngitis, the results obtained with ceftriaxone are also of interest. Although all strains tested from this study were susceptible (in vitro) to ceftriaxone based on an MIC90 of 0.047 μg/mL reported during a recently published study of >300 group A isolates,16 this high failure rate, in retrospect, likely was attributable to too small a dose of ceftriaxone. The selection of a 500-mg ceftriaxone dose was based on limited pharmacokinetic data, and it was not known whether this dose would result in and maintain “adequate” therapeutic ceftriaxone levels. A recent report by Blumer and colleagues17 has indicated that two 500-mg ceftriaxone doses given 18 hours apart might be necessary to adequately eradicate group A streptococci. Thus, suboptimal doses likely explains the observed unacceptable eradication rate (46%) of group A streptococci from the pharynx after 500-mg intramuscular ceftriaxone therapy.
The documented degree of reduction of the microbiologic efficacy of oral penicillin V and especially of intramuscular BPG was surprising.18 We are unaware of any recent, large, carefully conducted and thoroughly analyzed prospective study directly evaluating recommended doses of intramuscular BPG for the treatment of GAS pharyngitis or tonsillitis. The microbiologic failure rates for both penicillin formulations were essentially the same, probably reducing the influence of lack of compliance with oral penicillin V as a possible explanation. The administered doses were recommended ones, but recently published data raise the question whether larger penicillin doses might have been more effective.19
The most important acknowledged goal of antibiotic therapy for GAS pharyngitis is prevention of rheumatic fever; eradication of the pathogen from the throat within 9 days has been documented as necessary to achieve this objective.15 The relatively high microbiologic failure rates observed in these studies cannot be explained on the basis of in vitro antibiotic resistance. There have been no reports of either penicillin or ceftriaxone resistance among group A streptococci,16 and all GAS isolates tested from this study proved to be susceptible to both antibiotics. The penicillin MIC90 of 0.016 μg/mL observed in 200 group A isolates recovered from patients during this investigation is comparable to that reported for GAS isolates collected during the past 80 years as well as to other recent reports.16,,20 Despite this unchanged in vitro penicillin susceptibility, the penicillin eradication rates described in this report are lower than previous reports from this laboratory.18 However, there are no available data to link this failure rate with the increase in sequelae of GAS infections either in this study population or elsewhere.
It also seems unlikely that the high penicillin failure rates are resulting from any single streptococcal clone or group of clones. The isolates were obtained from 42 clinical sites located in 23 states and the District of Columbia. No single state or region provided a disproportionate number of isolates. The treatment failure isolates included at least 21 different serotypes. Twelve distinct recognized M protein serotypes were identified, the most common being M-types 1, 3, 12, 4, 28, 75, and 6 (in order of recovery). In addition, at least 9 distinct M nontypeable strains could be distinguished based on T-pattern serotyping and production of serum opacity factor. Thus, the isolates obtained from the penicillin treatment failure patients were from widely dispersed geographic sites in the United States and were serotypically diverse.
The, as yet, unproven hypothesis of adverse clinical and microbiologic influences of β lactamases produced by normal pharyngeal flora on penicillin as well as the possibility of an detrimental influence attributable to GAS tolerance to penicillin remain as possibilities. However, they seem unlikely to fully explain our findings in view of a number of published studies with conflicting conclusions about their influence. Additionally, one must always consider the possibility of initial eradication followed by recolonization from family contacts or schoolmates to explain the positive convalescent cultures. Although we do not believe that this possibility significantly influenced the results, the extensive sampling of family and school contacts required to address this issue was not included in the protocols.
To attempt to determine whether contamination of the study population with streptococcal carriers could have, as has been previously reported,21 contributed to the unexpected frequency of penicillin treatment failure rates, analyses of clinical status were also conducted. The percentage of patients presenting with specific clinical signs or symptoms could not be associated in any statistically meaningful way with microbiologic failure. All of the patients evaluated had met the published joint IDSA and FDA clinical recommendations for enrollment in such a study. Although recent studies have reported a beneficial influence of antibiotic therapy on resolution of clinical symptoms in patients with acute GAS pharyngitis,22,,23 evaluation of “clinical cure” in the patients reported here is particularly difficult because it is well documented that clinical symptoms associated with pharyngitis normally markedly resolve within 72 hours in a majority of patients, even without antibiotic administration.24
A previous report from this laboratory described penicillin as less effective in eradicating group A streptococci from the upper respiratory tract of “carriers” than from acutely infected individuals.21 It has been suggested that this may be because organisms present in the upper respiratory tracts of GAS carriers are not rapidly dividing when compared with those in the throat of acutely infected individuals.25 There have also been recent suggestions that the presence of intracellular streptococci in carriers may make eradication with penicillin more difficult because it does not effectively penetrate into cells.26,,27However, there are no data from the present study that allow us to comment about either of these hypotheses. If this study population was contaminated with “carriers” (despite the efforts to clinically and by using the laboratory identify this group), this could provide yet another possibility for the reduced microbiologic efficacy of penicillin. Such a reduced efficacy of penicillin to eradicate group A streptococci from the throats of carriers recently has been commented on in more detail by Gerber and colleagues.28
We also previously have reported that streptococcal upper respiratory tract “carriers” tend to have higher initial anti-DNase B and ASO titers than patients with bona fide streptococcal infection.12,,29 In the hope of evaluating contamination of the study population with streptococcal carriers despite published strict clinical criteria for enrollment, the acute ASO and anti-DNase B titers from these participants were carefully analyzed. In this study, there was no trend toward higher initial titers among those who proved to be microbiologic failures. In fact, geometric mean baseline (visit 1) ASO and anti-DNase B titers differed by <0.1 log (<1 tube dilution) between the microbiologic treatment success and treatment failure groups; the standard deviations were virtually identical (data not shown). Thus, by analysis of acute titers, there is no direct evidence to indicate significant contamination of the study population by carriers.
We have also previously suggested that GAS carriers are less likely to experience a significant rise (≥0.2 log) in either ASO or anti-DNase B titers than are patients with bona fide streptococcal infections.12 The analyzed date regarding individuals with a significant rise in titer from this study do not allow discrimination between suspected carriers and those with infection. Thus, the antibody data unfortunately do not allow us to definitively exclude contamination with carriers as at least partially responsible for the microbiologic failures in the present studies of penicillin efficacy.
We have presented a paucity of detailed information about convalescent clinical status in this report. Because it is well known that clinical diagnosis of GAS pharyngitis is fraught with many pitfalls even at presentation,30,,31 and because a high percentage of patients continued to harbor the organism despite penicillin therapy, changes in clinical status would not be reliable measurements and, in fact, could be very misleading. For that reason we have concentrated on the more important endpoint, eradication of group A streptococci. Reviewing the varied and often controversial published data about convalescent clinical findings reveals no adequately reliable guidelines to assist the clinician. Analysis of the clinical data from this study led us to conclude that the clinical findings did not influence the bacteriologic eradication data in any meaningful way. Interpretation of the clinical findings proved difficult, especially in light of the accepted observations that patients tend to spontaneously improve from streptococcal pharyngitis with or without antibiotic therapy.24,,32
Many experienced clinicians and investigators believe, as we do, that the primary goal of antibiotic therapy for patients with streptococcal pharyngitis is eradication of the organism. It has been known for many years that prevention of rheumatic fever was predicated on eradication of the group A streptococcus.9 Although it is true that the FDA does require pharmaceutical companies to determine clinical findings in their applications associated with new drug applications, for this specific infection convalescent clinical findings do not represent a primary goal. In patients with pharyngitis, in contrast to several other pediatric bacterial infections, clinical improvement from pharyngitis does not have to be a surrogate criterion for microbiologic effectiveness. One can measure the desired effect (eradication) in patients with streptococcal pharyngitis.
Another important consideration in arriving at this decision to only summarize convalescent clinical information is based on review of the individual clinical findings reported during the convalescent visits in these studies. Many participants who were categorized by the examining physician as clinical failures were noted to have pharyngeal erythema alone at the time of a convalescent visit. This somewhat subjective assessment of erythema alone probably was too frequently used, resulting in a larger number of clinical “failures.” Again, we are aware of no association of the presence or absence of erythema as a characteristic of “improvement.”
Finally, we believe that there are important questions about available injectable BPG preparations. There is increasing published evidence in children6,,7,33 and in adults34 of reduced duration of adequate serum penicillin levels after injection of recommended doses of repository BPG when compared, for example, with the early studies by Stollerman and Rusoff.35 Is it possible that either the potency of or the manufacturing process for BPG has been modified during the past half century resulting in reduced bioavailability of penicillin? Although there is no direct evidence to firmly support this possibility in the present study, the recent report by Zaher and colleagues7 indicates that variations in manufacture of the product can adversely influence both peak levels and duration of adequate serum levels. Differences in the viscosity and even of crystal size of BPG have been documented. (S. R. Zaher, unpublished observations). In this present study, all BPG came from the same manufacturer (Wyeth-Ayerst Laboratories, Philadelphia, PA), although several different manufactured lots were used by the different clinical sites. Attempts to correlate failure rates at individual clinical sites with specific lots of BPG were not revealing. Although admittedly speculative, we believe our findings also may be compatible with reduced bioavailability of penicillin as a possible explanation. To our knowledge, there has been no published detailed description(s) of the manufacturing process by the numerous pharmaceutical companies that produce this formulation of penicillin around the world. Furthermore, we are aware that a pharmaceutical company can market BPG using penicillin prepared in several different countries.
The reason(s) for the high microbiologic failure rates after either oral or parenteral penicillin in these carefully designed and rigorously conducted clinical trials in patients with acute pharyngitis remains unclear. However, before prematurely concluding that there is no place for penicillin (or perhaps for ceftriaxone in adequate doses) in the therapy of GAS upper respiratory tract infections, additional studies are required; larger antibiotic doses should be one important part of any proposed protocol. Although not providing sufficient evidence to change current recommendations for therapy, the present study raises important questions that require confirmation, as confirmatory data could significantly impact currently accepted medical regimens and public health policy worldwide.
This study was supported by a grant from Roche Laboratories, Nutley, New Jersey.
We thank Constance D. Rothermel, PhD, whose dedication to carrying out the studies, proper analysis of the data, and review of the manuscript were essential. We are also grateful to Melanie Bishop, Louise Cockey, Dr Steve Lee, and Michael McGuire of Roche Laboratories for their cooperation and assistance with data analysis. We also express our appreciation to the physicians and study nurses at the 42 clinical sites where participants were enrolled. Finally, we thank Cheryl Kunde, JoAnn Nelson, and Carey Rehder for their technical assistance.
- Received February 28, 2001.
- Accepted June 14, 2001.
Reprint requests to (E.L.K.) Mayo Mail Code 296, Department of Pediatrics, University of Minnesota Medical School, 420 Delaware St SE, Minneapolis, MN 55455. E-mail:
Presented in part at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy; September 26–29, 1999; San Francisco, CA.
- BPG =
- benzathine penicillin G •
- IDSA =
- Infectious Diseases Society of America •
- FDA =
- Food and Drug Administration •
- ASO =
- antistreptolysin O •
- DNase B: deoxyribonuclease B; MIC =
- minimum inhibitory concentration •
- GAS =
- group A streptococcal
- ↵American Academy of Pediatrics, Committee on Infectious Diseases.Red Book: Report of the Committee on Infectious Diseases. 24th ed. Elk Grove Village, IL: American Academy of Pediatrics; 1994
- Dajani AS,
- Bisno AL,
- Chung KJ,
- et al.
- ↵World Health Organization. Rheumatic Fever and Rheumatic Heart Disease. Report of a WHO Study Group. Geneva, Switzerland: World Health Organization; 1988. Technical Report Series No. 764
- Bisno AL,
- Gerber MA,
- Gwaltney JMJ,
- Kaplan EL,
- Schwartz RH
- Zaher S,
- Kassem A,
- About-Shlieb H,
- Kholy AE,
- Kaplan E
- Dajani A
- Chow AW,
- Hall CB,
- Klein JO,
- Kammer RB,
- Meyer RB,
- Remington JS
- ↵National Committee for Clinical Laboratory Standards. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically. 3rd ed. Approved Standard. Villanova, PA: NCCLS; 1993. NCCLS Document No. M7-A3
- Kaplan E,
- Top F Jr.,
- Dudding B,
- Wannamaker L
- ↵Johnson DR, Kaplan EL, Sramek J, et al. Laboratory Diagnosis of Group A Streptococcal Infections. Geneva, Switzerland: World Health Organization; 1996
- Blumer JL,
- Reed MD,
- Kaplan EL,
- Drusano GL
- Currie B
- Nelson J
- ↵Kaplan EL. Group A streptococcal carriers and contacts: (when) is retreatment necessary? In: Shulman S, ed. Management of Pharyngitis in an Era of Declining Rheumatic Fever. Columbus, OH: Ross Conference of Pediatric Research; 1984:92
- LaPenta D,
- Rubens C,
- Chi E,
- Cleary P
- ↵Facinelli B, Magi G, Giovanetti E, Capretti R, Varaldo P. Persistence of Streptococcus pyogenes within cultured human respiratory cells. In: 39th Interscience Conference on Antimicrobial Agents and Chemotherapy; September 1999; San Francisco, CA; 1999:73
- Gerber M,
- Tanz R,
- Kabat W,
- et al.
- ↵Wannamaker LW. Diagnosis of pharyngitis. Clinical and epidemiologic features. In: Shulman S, ed. Pharyngitis: Management in an Era of Declining Rheumatic Fever. New York, NY: Praeger Publishers; 1984:33–46
- Denny F,
- Wannamaker L,
- Hahn E
- Ginsburg C,
- McCracken G,
- Zweighaft T
- Bass J,
- Longfield J,
- Jones R,
- Hartmann R
- Stollerman G,
- Rusoff J
- Copyright © 2001 American Academy of Pediatrics