Article Text

Download PDFPDF

Once-daily amoxicillin versus twice-daily penicillin V in group A β-haemolytic streptococcal pharyngitis
  1. D R Lennon1,
  2. E Farrell2,
  3. D R Martin3,
  4. J M Stewart1
  1. 1
    University of Auckland, Auckland, New Zealand
  2. 2
    Counties Manukau District Health Board, Manukau City, New Zealand
  3. 3
    Institute of Environmental Science and Research Ltd, Porirua, New Zealand
  1. Professor Diana R Lennon, Department of Paediatrics, University of Auckland, Private Bag 92019, Auckland, New Zealand; d.lennon{at}


Background: Rheumatic fever is a preventable chronic disease preceded by group A β-haemolytic streptococcal (GABHS) pharyngitis.

Objective: To test the non-inferiority of once-daily (QD) oral amoxicillin to the recommended twice-daily (BID) oral penicillin V in GABHS pharyngitis.

Methods: This was a randomised non-inferiority trial carried out in a school-based clinic in New Zealand. Children presenting with GABHS pharyngitis were randomised to oral amoxicillin 1500 mg QD (or 750 mg if bodyweight was ⩽30 kg) or to oral penicillin V 500 mg BID (or 250 mg if bodyweight was ⩽20 kg) for 10 days. Observed medication and weekend diary cards were used to monitor adherence.

Outcome: Eradication of GABHS, determined with follow-up throat cultures on days 3–6, 12–16 and 26–36. GABHS isolates were serotyped to distinguish bacteriological treatment failures (and relapses) from new acquisitions. Non-inferiority was defined as an upper 95% confidence limit (CL) for the difference in success of eradication in the amoxicillin and penicillin V treatment groups of ⩽10%.

Results: 353 children with positive throat swabs for GABHS were randomised to amoxicillin (n = 177) or penicillin V (n = 176). The upper 95% CL for the differences in positive cultures between the antibiotics was 4.9% at days 3–6, 6.5% at days 12–16 and 8.5% at days 26–36. Treatment failures (including relapses) occurred at each visit in 5.8%, 12.7% and 10.7% of amoxicillin recipients and 6.2%, 11.9% and 11.3% of penicillin V recipients, respectively. No significant differences in resolution of symptoms were noted between treatment groups. One case of unsubstantiated acute rheumatic fever occurred after 7 days of amoxicillin.

Conclusion: In this adequately powered study, once-daily oral amoxicillin is not inferior to twice-daily penicillin V for the treatment and eradication of GABHS in children with pharyngitis.

Statistics from

Rheumatic fever is a preventable chronic disease preceded by a group A β-haemolytic streptococcal (GABHS; Streptococcus pyogenes) pharyngitis.1 Its sequela, rheumatic heart disease, is the commonest cause of cardiac morbidity in children globally.2

The use of antibiotics in individuals with probable GABHS pharyngitis for the primary prevention of acute rheumatic fever was first tested in a randomised controlled trial by Denny and colleagues in 1950 in United States military personnel.3 When subjects were treated with a single intramuscular dose of long-acting penicillin, a protective effect for the use of penicillin against acute rheumatic fever was shown (the relative risk of developing acute rheumatic fever was 0.12; 95% confidence interval (CI) 0.03 to 0.50). Recently, the overall protective effects of antibiotics in reducing the incidence of acute rheumatic fever following a suspected episode of GABHS pharyngitis have been validated in a meta-analysis of 10 randomised or quasi-randomised studies predominantly conducted in the army setting.4 Analysis of the nine trials evaluating penicillin showed a protective effect of 80% against acute rheumatic fever. Similarly, a Cochrane review found benefit from antibiotics for acute rheumatic fever prevention, reducing the incidence to less than one third of that in the placebo group (odds ratio 0.30; 95% CI 0.20 to 0.45).5 Based on the results of these studies, primary prevention of acute rheumatic fever by depot penicillin treatment of pharyngitis (using the surrogate microbiological endpoint of eradication of GABHS) is considered the benchmark against which oral antibiotic formulations are measured. Lack of eradication has been shown to be an important risk factor for the development of rheumatic fever.6

While antibiotics may not be justified for treating a sore throat in populations at low risk of acute rheumatic fever,7 a different approach is warranted in high-risk populations.8 American clinical practice guidelines recommend oral antibiotics for the treatment of GABHS in children.911 In particular, oral penicillin V administered twice or three times daily for 10 days is indicated as first-line therapy. The universal sensitivity of GABHS isolates to penicillin makes this an effective therapeutic option when the entire 10-day course of treatment is given9; however, the arduous dosing regimen is often a barrier to compliance, particularly in children.

Amoxicillin is an inexpensive congener of penicillin, with a relatively narrow spectrum of antimicrobial activity. It has a longer half-life than penicillin V, thus allowing less frequent administration. A systematic review was undertaken to seek out published studies on once-a-day amoxicillin as we developed a guideline for rheumatic fever control, including sore throat management ( Small studies have indicated that once-daily (QD) administration of amoxicillin produces a similar rate of GABHS eradication as penicillin V administered twice daily (BID) or three times daily (TID).12 13 A well-powered study conducted in 2001–2003 has recently been published.14 This study, however, did not compare QD amoxicillin to the current standard of care, penicillin V BID or TID. The purpose of our randomised study conducted in 1996–1998 was to test the non-inferiority of oral amoxicillin QD to oral penicillin V BID in children presenting to a school-based sore throat clinic with GABHS pharyngitis in an area with a high incidence of rheumatic fever (∼60/100 000) in school-aged children.



From May 1996 to November 1998, children aged 5–12 years presenting to a sore throat clinic at a primary school in Auckland, New Zealand with signs and symptoms indicative of acute pharyngitis or tonsillitis were enrolled in this study following informed consent. Signs and symptoms included a sore throat and at least one of the following: core temperature ⩾38°C; headache, nausea or abdominal pain; difficulty in swallowing; inflamed or infected throat; or tender glands in the neck. Children were excluded from study participation if they had hypersensitivity to penicillin, were likely to require treatment with other antimicrobials during the study period or had received antimicrobial therapy within 72 h prior to study entry, had a previous history of acute rheumatic fever, cardiac disease or kidney disease, had a rash suggestive of scarlet fever or mononucleosis, were immunocompromised, had a neoplastic disease, a terminal illness or neutropenia (absolute neutrophil count <1.5×10/9 cells/l) or had previously been included in this study within the current school term (approximately 12 weeks in duration).


The study was embedded in a pilot project in the planning phases of a large school programme to study rheumatic fever prevention (D Lennon, personnal communication). All consented children were encouraged to present with sore throat symptoms. Children/guardians with cultures positive for S pyogenes were approached for this study.

Study design

A rayon-tipped swab of the posterior pharynx and tonsils (or tonsillar fossae) was obtained from each clinically symptomatic child by trained research assistants and transported by courier overnight to the Institute of Environmental Science Streptococcal Laboratory (Porirua, New Zealand) where it was streaked on crystal violet blood agar. All plates were incubated in CO2 at 37°C for 48 h with the first reading at 24 h. Pure colonies from the primary plate were serogrouped. β-Haemolytic colonies recultured on a secondary plate (Columbia blood agar) were identified by latex agglutination (Streptex, Murex Biotech, Dartford, UK). Bacitracin disc testing was used only on secondary plates. For patients with positive follow-up throat cultures, serotyping (M type and/or emm sequence typing) was performed on pre- and post-treatment GABHS isolates.15

Children who were positive for GABHS were randomised on study day 1 in a 1:1 ratio to amoxicillin 1500 mg (or 750 mg for children with bodyweight ⩽30 kg) administered orally QD for 10 days or to penicillin V 500 mg (or 250 mg for children with bodyweight ⩽20 kg) administered orally BID for 10 days. The randomisation schedule was computer generated, and allocated by telephone implemented by a third party. Medications were provided as tablets or elixir. A second placebo dose for the amoxicillin recipients was not used. To aid compliance, children were supervised taking the study drugs at school (9 am for those randomised to amoxicillin and 9 am/3 pm for those randomised to penicillin V). Weekend compliance was monitored by completion of drug diary cards and return of labelled bottles to the school each Monday.

This study was conducted in accordance with Good Clinical Research Practice guidelines and the Declaration of Helsinki. The study protocol, patient information sheet and consent form were approved by the Auckland Regional Ethics Committee. Parental consent was obtained for all children participating in this study.


Eradication of GABHS was determined with follow-up throat cultures on days 3–6 (visit 2), days 12–16 (visit 3) and days 26–36 (visit 4). For children with positive follow-up throat cultures, GABHS isolates characterised by serotyping distinguished between bacteriological treatment failures (same serotype as initial isolate), relapses (same serotype as initial isolate, but previous negative swab) and new acquisitions (different serotype to initial isotype).16 The impact of therapy on the clinical course of infection was determined by evaluation of symptoms at randomisation (visit 1; day 1), days 3–6 (visit 2) and days 12–16 (visit 3) by trained research assistants. Patients from either treatment group with signs and/or symptoms of pharyngitis and a positive culture at visit 3 or 4 were retreated with a 10-day course of penicillin V BID (ie, the conventional therapy).

Statistical analysis

With no difference in treatment effect in the two arms of the trial and assuming 85% eradication, 155 evaluable subjects per treatment group would have 80% power to demonstrate non-inferiority (defined as the upper 95% confidence limit (CL) of the treatment effect difference of <10%). To allow for non-completion, it was planned to enrol 185 patients per arm. To determine if the treatments differed in their effect upon clinical symptoms, a logistic regression model was fitted for each symptom, with the presence or absence of the symptom at visit 2 as the binary outcome and treatment group and the presence or absence of the symptom at visit 1 as explanatory variables. χ2 Tests for independence were used to investigate if symptoms at completion of treatment differed depending on whether or not GABHS was still present and also to ensure there was no evidence of baseline symptoms differing depending on completion of the trial or persistence of the original GABHS serotype. Analyses were conducted using SAS statistical software.

The primary outcome of the study was bacteriological failure (including relapse) at visit 3. This included bacteriological persistence (the presence of the same M or emm type of GABHS as for visit 1) and relapse (ie, the reappearance of the visit 1 M or emm type). Symptoms were evaluated separately. Streptococcal antibody titres were not measured.


Patient disposition and characteristics

A total of 353 patients who had positive throat swabs for GABHS were randomised to treatment with amoxicillin QD (n = 177) or penicillin V BID (n = 176) following provision of parental informed consent. The demographic characteristics and symptomology of the two treatment groups were similar at baseline (table 1). Almost half of all patients were indigenous Maori, with another third of Pacific Island origin, reflecting the population of the school where the study was conducted. All patients had a sore throat at presentation, with tonsillar exudate and tender cervical lymphadenitis present in 61% and 69% of all patients, respectively.

Table 1 Baseline characteristics of children attending a school clinic with a positive group A streptococcal culture

Seven children in the penicillin V group and 11 children in the amoxicillin group did not complete medication (table 2). No difference could be demonstrated in occurrence of fever, tonsillar exudate or tender lymph nodes at visit 1 in those who did and did not complete. Some other children were not available for assessment on all evaluation days.

Table 2 Number (%) of randomised patients who discontinued treatment and reasons for discontinuation

Both drugs were well tolerated. Adherence was high in both treatment groups, with 97% of amoxicillin recipients and 98% of penicillin V recipients receiving ⩾80% of their scheduled doses.

Effects of treatment on eradication of GABHS

Bacteriological responses to amoxicillin QD or penicillin V BID are shown in table 3. At visit 2 (days 3–6), the between-treatment difference in the incidence of positive cultures was 0.3% (upper 95% CL of 4.9%), with a bacteriological failure of 5.8% for amoxicillin and 6.2% for penicillin. At visit 3 (days 12–16), bacteriological failure was similar between groups (12.7% and 11.9% for amoxicillin and penicillin, respectively), with an upper 95% CL of 6.5% for the between-treatment difference in positive cultures. The higher rate of positive cultures at visit 3 compared with visit 2 was mainly attributable to relapse. At visit 4 (days 26–36), the incidence of positive cultures had increased with a between-treatment difference of 1.9% (upper 95% CL of 8.5%), but bacteriological failure decreased slightly (10.7% for amoxicillin and 11.3% for penicillin V). New acquisitions had increased in both groups. There was no evidence of inferiority of amoxicillin to penicillin V at any time period, with inferiority defined as an upper 95% CL of the difference in effect of ⩾10%.

Table 3 Bacteriological response to antibiotic therapy (number (%) of patients with throat swab available)

Effects of treatment on symptoms

The effects of antibiotic treatment on clinical manifestations (sore throat, tonsillar exudates and tender lymph nodes) were evaluated after 3–6 days of medication (visit 2). There was no evidence of difference in the degree of change in any symptoms from premedication to day 3–6 between the amoxicillin and penicillin V groups. At 12–16 days, symptoms were more prevalent in children who had persistent or recurrent GABHS compared with those who had eradication of GABHS. Of those with the original infecting GABHS (n = 39), 54% had a sore throat compared with 32% of those eradicated (p = 0.009), 59% had tonsillar exudates compared with 27% of those eradicated (p<0.001) and 58% had tender lymph nodes compared with 30% of those eradicated (p<0.001).

One child after 1 week of amoxicillin (see table 2) developed symptoms of possible acute rheumatic fever as evidenced by monoarthritis in the absence of non-steroidal anti-inflammatory drugs and an erythrocyte sedimentation rate of 120 mm/h and falling rapidly, with a PR interval on ECG of 0.18 s and mild mitral regurgitation on echocardiogram in the absence of an audible heart murmur. His temperature was 38°C and streptococcal titres were 265 IU/ml, antiDNAse B 366 IU/ml and antihyaluronidase antibody 1024 IU/ml. The patient received a further 10 days of penicillin. Repeat throat swabs at the time of acute rheumatic fever symptoms were negative. In line with the standard of care in our environment (, the patient was commenced on benzathine penicillin prophylaxis. No other diagnosis was elicited; however, his symptoms and signs did not fulfil the Jones criteria.17


The results of this randomised study provide a notable addition to the small body of evidence1214 that 10 days of once-daily amoxicillin is as effective as 10 days of twice-daily penicillin V in children with GABHS pharyngitis. There was no evidence of inferiority of amoxicillin to penicillin at any time period during or following treatment. Amoxicillin is a generically available, relatively low cost and narrow spectrum antibiotic that can be taken with or without food18 and presents a useful therapeutic alternative to penicillin V. In this study, non-inferiority (defined as being 95% confident that the effect of amoxicillin was not more than 10% less than that of penicillin V) was established in terms of microbiological effect (a surrogate marker for primary prevention of acute rheumatic fever). No difference in clinical symptom resolution was demonstrated between treatments. Over 90% of children participating in the study had eradication of GABHS within 3–6 days of commencing therapy. Relapse and/or treatment failure occurred at a similar rate in each treatment group (12.7% and 11.9% of amoxicillin and penicillin recipients, respectively, by day 12–16), which could be distinguished from new acquisitions by serotyping. New acquisitions occurred in 5.0% of amoxicillin recipients and 2.5% of penicillin recipients by day 26–36. There was no evidence of inferiority of amoxicillin to penicillin V at any time period. As would be expected, there was a higher rate of persistent clinical symptoms in children who failed treatment.

Our bacteriological failure rates at the end of treatment visits (days 12–16) were comparable with most other amoxicillin and penicillin studies, although higher than the two small published studies on amoxicillin once a day (0% and 5%, respectively).12 13 In other studies with more frequent dosing, amoxicillin bacteriological failure rates of 5%, 16% and 24% were reported with twice-daily treatment1921 and between 3% and 15% with three-times daily dosing.19 2226 In a careful review of bacteriological failure in oral penicillin trials, including assessment of comparability of protocols and use of serotyping, rates of 11–17% were found (10.5–12.5% at days 1–14 and 14–16.9% at days 1–60).27

The children admitted to this study had high rates of exudative tonsillitis and cervical adenitis, suggesting (in the absence of streptococcal serology) that the majority at least had true streptococcal pharyngitis.28 There was no discernable difference in the symptom profiles of the small number of children who had retention of their original GABHS strain compared to the remainder. Throat carriage of GABHS has not been measured in this predominantly Maori and Pacific Island population, although rheumatic fever is common.29

Acute pharyngitis is one of the most common reasons for childhood visits to primary care practitioners.30 31 Practitioner-based studies observe higher rates of GABHS positivity (10–30%) than our population-based study in a school (∼6%), presumably because of more severe symptoms leading to consultation and other factors. While the viral form of the infection can only be managed symptomatically, antibiotic treatment is indicated for GAHBS infection, particularly to prevent the development to acute rheumatic fever and accompanying rheumatic heart disease. This is of particular importance in high-risk populations, predominantly in developing nations, where the incidence of paediatric acute rheumatic fever may exceed 50 per 100 000 school-aged population,32 and in some indigenous populations within developed nations.29 33 For example, in New Zealand, less than 10 cases of acute rheumatic fever per 100 000 school-aged children of European descent were reported per annum between 1982 and 1997; however, among indigenous Maori and Pacific Islanders in New Zealand (more likely to be socioeconomically disadvantaged), 80–100 cases per 100 000 were reported each year.29 The serious sequelae to GABHS infection of acute rheumatic fever and associated rheumatic heart disease (including chronic heart disease, heart failure and death) provide a strong imperative to treat GABHS pharyngitis in an appropriate and timely manner.

While antibiotic availability is not the sole determining factor for preventing rheumatic fever, it certainly contributes to reducing the burden of disease.34 The use of intramuscular benzathine penicillin for secondary prophylaxis has the greatest and most cost-effective impact on rheumatic fever recurrence.35 Treatment of the GABHS pharyngitis trigger for acute rheumatic fever has good support in high-risk settings.4 5 Given global efforts to prevent widespread antimicrobial resistance, there is a concerted effort to ensure that patients prescribed antibiotics receive the appropriate therapy. The recommended regimen for penicillin V treatment of paediatric GABHS pharyngitis (BID or TID for 10 days) is potentially a barrier to treatment adherence. A shorter course of treatment or a once-daily dosing regimen may help improve compliance. However, a number of studies have shown that penicillin V is not as effective at GABHS eradication when administered once daily or for less than 10 days.3638

The once broad spectrum of amoxicillin is being slowly eroded with the increasing resistance, for example, of Escherichia coli, Haemophilus influenzae type b and more recently Streptococcus pneumoniae in many geographical areas, thus justifying its use in this arena.

Several once-daily formulations of non-penicillin-based antibiotics are available for GABHS pharyngitis in children, including azithromycin, cefadroxil, cefibuten, cefdinir, cefpodoxime and cefixime. Furthermore, 5-day courses of some such antibiotics have shown comparable clinical efficacy to a 10-day course of penicillin V.3945 However, the utility of these agents is limited by a broad spectrum of antimicrobial activity, cost or higher bacterial eradication failure rates than oral penicillin. Compliance was stringently monitored in our study by observed administration of medications; therefore, the potential additive benefits of a simplified dosing regimen for amoxicillin could not be observed. It would be expected, however, that improved compliance with once-daily amoxicillin in a non-clinical trial setting would provide superior effectiveness to twice-daily penicillin V.

The setting of this study provides further insight into the utility of amoxicillin QD for primary prevention of acute rheumatic fever. The study was conducted in a population with a high risk of rheumatic fever (predominantly Maori and Pacific Island children)29 in a school-based setting in New Zealand. This allowed targeted and effective health care access for those who are most vulnerable to GABHS infection. Translation of similar programmes to other high-risk populations may be a useful strategy for reducing the incidence of acute rheumatic fever. We also delivered the antibiotic doses at approximately 9 am and 3 pm to fit in with the school day. Furthermore, the generic availability and low cost of amoxicillin widen its access for some developing nations with high rates of rheumatic fever.

Shorter courses (6 days) of amoxicillin (administered BID) have been shown to be comparable to 10-day courses of penicillin V (administered TID) in small studies.20 38 46 Whether a similar effect could be gained with shorter courses of once-daily amoxicillin is yet to be determined.

What is already known on this topic

  • Rheumatic fever is a preventable chronic disease causing chronic rheumatic heart disease preventable by treating GABHS pharyngitis with oral penicillin V twice or three times daily for 10 days.

  • Amoxicillin is inexpensive, with a narrow spectrum and a longer half-life than penicillin V, allowing less frequent administration.

What this study adds

  • This randomised intervention demonstrates that once-daily oral amoxicillin is not inferior to twice-daily penicillin V for the treatment and eradication of GABHS pharyngitis in children in a population at high risk for acute rheumatic fever.

This study has at least two potential limitations. We have compared the new potential treatment, amoxicillin QD to the current standard of care, penicillin V BID, but are unable to comment on any fine tuning of dosage of amoxicillin or length of treatment. Our study was conducted in a single school with unique demographic features which might limit the generalisability of our findings. We have demonstrated that once-daily oral amoxicillin is not inferior, with respect to a prespecified margin of 10%, to standard of care twice-daily penicillin V for the treatment and eradication of GABHS in children with pharyngitis, adding to the body of knowledge on this subject.12 13 A more recent study14 has contributed to the area by comparing QD to BID amoxicillin. The once-daily dosing schedule, cost, availability and reasonably narrow antibacterial spectrum of amoxicillin make it a favourable treatment option for this disease.


We gratefully acknowledge the assistance of Shirley Maihi, school principal and the parents and children involved in this study. Maninoa Tasi’s help was also invaluable.



  • Funding: This study was funded by the New Zealand Heart Foundation, who had no role in the study design, the collection, analysis and interpretation of data, the writing of the report or the decision to submit the paper for publication.

  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.