Article Text

Download PDFPDF

How long should we treat this infection for?
  1. Carroline Kerrison,
  2. F Andrew I Riordan
  1. Department of Paediatric Infectious Diseases, Alder Hey Children's NHS Foundation Trust, Liverpool, UK
  1. Correspondence to Dr F Andrew I Riordan, Department of Paediatric Infectious Diseases, Alder Hey Children's NHS Foundation Trust, Eaton Road, Liverpool L12 2AP, UK; andrew.riordan{at}


Infection accounts for approximately half of all paediatric admissions to hospital and an even greater proportion of primary care. Guidelines on duration of antibiotic therapy exist, but antibiotic therapy for children needs to be individualised. If a child is not improving the clinical condition and treatment should be reviewed and/or discussed with an expert. However, slavishly completing the recommended course of antibiotics in a child who is well, may not be appropriate. Recent studies on treatment duration advocate shortened courses with certain caveats, but guidelines and clinical practice do not always follow the evidence from the few randomised trials of treatment duration of infection.

  • Infectious Diseases
  • Pharmacology
View Full Text

Statistics from


Infection accounts for approximately half of all paediatric admissions to hospital and an even greater proportion of primary care prescriptions. One-third to half of the children admitted to hospital are given antibiotics.1 Yet, how long to treat with antibiotics remains a subject of great debate. Despite the existence of multiple guidelines for most common childhood infections, many of the recommendations on duration of treatment are based more on expert opinion than strong scientific evidence. There are few infections for which trials demonstrating optimal duration of antimicrobial therapy exist; those that do exist are subject to design flaws and contradictory conclusions.2 Well-designed randomised controlled trials (RCTs) of treatment duration for paediatric infections are rare.3 ,4

Hospital doctors are now encouraged to review the need for antibiotics after 48 h of treatment. Decision options are: stop antibiotics, switch IV to oral, change antibiotic, continue the current treatment and send home on outpatient parenteral antibiotic therapy (see online supplementary e1).

There are a number of advantages of shorter courses of antibiotics including: better compliance, fewer adverse effects, reduced cost and a decrease in antibiotic resistance. This article will discuss the data for antibiotic duration of common childhood infections and new evidence for different approaches to defining duration of treatment.

The choice of an appropriate antibacterial agent and the optimal duration of therapy depend on a number of variables including: site of infection, host factors, bioavailability and palatability of medication, local resistance patterns, and importantly, knowledge of the infectious agent. In paediatric practice, diagnoses of infectious syndromes are often based solely upon symptoms and signs. Clinical specimens to identify causative agents are not, or cannot, be practicably obtained for many respiratory tract infections, which account for the majority of consultations and prescriptions. Clinicians diagnose and treat syndromes (such as acute otitis media, tonsillitis or lower respiratory tract infection), often without knowledge that the causative organism is bacterial. This makes the decision of how long to treat (if treatment is required at all) challenging and possibly contributes to the heterogeneity of results from trials and guidelines.

Why should we use short courses of antibiotics?

Short-course antibiotic therapy has many advantages. Antibacterial resistance is an international problem resulting in adverse clinical outcomes and increasing healthcare costs. There are good data to suggest that antibiotic exposure increases carriage of resistant organisms, especially prolonged courses at low dose.5 ,6 Evidence from RCTs, observational studies and modelling exists to show that both judicious use of antibiotics and shorter courses with higher doses would reduce the burden of resistant organisms in the community and hospital setting.6

Compliance with antibiotic prescriptions is variable and influenced by a number of factors including dose frequency and course duration; reducing both has been shown to improve compliance.7 ,8 Shorter courses of antibiotics are also associated with fewer adverse effects, and so, reduced costs in terms of additional healthcare visits or changes in antibiotic.9

Short-course antibiotic therapy can only be adopted if it is shown to be equivalent to established standard of care. A number of studies have evaluated shortened courses of antibiotics for many infections including those of the upper respiratory tract, pneumonia, urinary tract infections (UTIs), bone and joint and meningitis.

Respiratory tract infections


Acute pharyngitis is common in children, but only 20%–30% are due to a bacterial cause (mostly Group A Streptococcus).10 Even in proven streptococcal infection, antibiotics shorten the duration of symptoms by about 16 h overall.11 Some guidelines suggest that antimicrobial therapy should only be given to children with pharyngitis and diagnosed group A streptococcal infection.12

The main reason to treat streptococcal pharyngitis with antibiotics is to prevent rheumatic fever and other sequelae. However, the incidence of rheumatic fever in the UK is extremely low, and current guidelines suggest that there is no support for the routine treatment of sore throat with antibiotics to prevent rheumatic fever or glomerulonephritis.12

If children with acute streptococcal pharyngitis are given antibiotics, then they should be treated with an antibiotic long enough to eradicate the organism from the pharynx.10 For the commonly recommended drug (penicillin V), this is 10 days;10 however, few patients complete the 10-day course, making this ‘Gold Standard’ treatment of dubious value.7

Shorter courses (3–6 days) of some other antibiotics (amoxicillin, macrolides and cephalosporins) have been shown to be equivalent to 10 days of penicillin V for streptococcal pharyngitis.13 Thus, the duration of antibiotic treatment for streptococcal pharyngitis can vary depending on the antibiotic prescribed. However, these shorter courses are not recommended in recent guidelines because of concerns about study design, their broader antimicrobial spectrum and increased cost.10 Guidelines still recommend a 10-day course of penicillin V, despite the poor adherence to this regimen (see table 1).

Table 1

Suggested antibiotic treatment duration in days as recommended by various organisations


Community-acquired pneumonia (CAP) may be caused by a large number of organisms: bacterial, viral or both. Although likely pathogens vary with age, it is usually not possible to reliably differentiate between causative organisms clinically or radiographically. Lack of microbiological diagnosis makes studies of optimal treatment duration difficult. Despite a number of well-conducted large trials demonstrating good outcomes with shorter courses of oral antibiotics, guidance on treatment duration is varied and vague. It may be difficult to adopt the recommendations from many of the studies as they were conducted in resource-poor settings where immunisation coverage may be limited and does not include the 13-valent pneumococcal vaccine. However, a recent Cochrane review suggests that a 3-day course may be appropriate in children under 5 years old with pneumonia.14 Although it may not yet be possible in the UK to adopt the 3-day course, suggested safe by a variety of studies, there is clear evidence that courses shorter than the traditional 7–14 days are appropriate for uncomplicated CAP.14 Current guidelines acknowledging shorter courses may be appropriate, but continue to recommend longer courses.15 ,16 The recent British Thoracic Society guidelines give no clear statement on duration although they discuss the evidence for shorter courses and reasons for not adopting the 3-day courses suggested in studies based in the developing world.16

Urinary tract infections

Infections of the urinary tract have traditionally been treated for longer in children than in adults; the British National Formulary suggests a 1-day (two high doses) course of oral amoxicillin for uncomplicated UTI  in adults. The terminology of UTI classification varies depending upon the author, as do the criteria for these classifications. The National Institute for Health and Clinical Excellence (NICE) guidance refers to ‘upper tract infections’/pyelonephritis as infants or children with fever >38°C and bacteriuria or bacteriuria with loin/flank pain, while all others with bacteriuria but no other features should be considered to have cystitis/‘lower tract infection’.17 Upper tract infections include pyelonephritis. The American Academy of Pediatrics (AAP) guidance on the management of UTIs in febrile infants 2–24 months of age suggest 7–14 days of treatment.18 NICE recommend oral antibiotics for 7–10 days, if oral treatment is not possible initially, then the recommendation is 2–4 days intravenous therapy before switching to an oral antibiotic to complete a total of 10 days treatment. There are a number of studies that evaluate management of complicated/upper tract infections. Meta-analysis of management of pyelonephritis concludes that oral therapy is appropriate when tolerated, but suggest treatment duration of 7–14 days.19

Lower UTIs or ‘cystitis’ involve the bladder only and so produce fewer systemic symptoms. An updated Cochrane review20 assessed 10 studies involving 652 children and concluded that short courses (2–4 days) are as effective as longer courses (7–14 days). The inclusion criteria for the studies assessed were proven bacteriuria with or without other features such as fever. NICE recommend a 3-day treatment course for lower tract infections (see table 1).17

Bone and joint infections

Osteomyelitis and septic arthritis have traditionally been treated with extended courses of antibiotics (6 weeks) with a tendency to prolonged intravenous therapy, and then, an oral antibiotic to complete the course.

Septic arthritis

A benchmark for the management of septic arthritis in children was set 40 years ago by Nelson et al21 who gave at least 3 weeks of intravenous therapy. Since then, new antibiotics have been trialled using the same regimen. Moves to shorten treatment regimes have been proposed in a recent study.22 Peltola et al. published a prospective RCT comparing 10-day antibiotic treatment (2–4 days IV) with the traditional 30-day course. They reported no difference in short-term outcomes or recurrence at 1 year.22 The ideal duration of antibiotic therapy for septic arthritis is not yet fully defined,23 so a UK Multi-Centre observational, case-controlled study of Bone and Joint infection in children is now underway (DINOSAUR (see online supplementary e2).


Osteomyelitis with or without adjacent joint involvement is usually of haematogenous origin in children. Some authorities differentiate between acute osteomyelitis, diagnosed within 14 days of onset of symptoms and chronic osteomyelitis, the presence of symptoms or signs for longer than 14 days prior to diagnosis. This differentiation then impacts upon treatment duration, the former being treated for 4–6 weeks and the latter considerably longer. A recent Finnish study has evaluated shorter courses for the treatment of acute haematogenous osteomyelitis.4 Peltola et al randomised children with a clinical and microbiologically confirmed diagnosis of osteomyelitis (+/− adjacent joint involvement) to receive a minimum of 30 days antibiotics or a shorter course defined by resolution of most but not necessarily all clinical symptoms and reduction in C reactive protein (CRP) to <20 mg/l. In both groups, initial therapy was intravenous for a median of 3 or 4 days. In both groups, if inflammatory markers remained high or went backup, or if clinical symptoms or signs persisted, the antibiotics were continued until two CRP values <20 were obtained. The range of antibiotic duration in the short-course group was 10–21 days. Both groups extended therapy when indicated by parameters described above. Outcomes were full recovery and recurrence of symptoms up to 1 year. Their study noted no difference in outcomes between children receiving shorter therapy compared with 30 days.

Critics of this study quote difference in population between acute osteomyelitis in children in Finland compared with other settings such as the USA or UK.24 This study adds to the evidence that early transition to oral therapy is safe and effective. These studies demonstrate that it is ethical to conduct further trials of treatment duration in bone and joint infection. A UK national survey of the management of Bone and Joint infection in children (DINOSAUR) (see online supplementary e2) is now underway to see if a randomised trial of short antibiotic versus long antibiotic therapy is feasible in the UK.


Bacterial meningitis is an important cause of morbidity and mortality worldwide. Because of the potential for poor outcomes, if treatment was inadequate, protracted treatment durations were used initially (three antibiotics for 3 weeks).25 Meningitis is one of the few infections in children where a microbiological diagnosis is likely, and as a consequence, guidelines for organism-specific treatment durations exist.26 However, a meta-analysis of the available data could not show the differences between short-course and long-course antibiotic treatment for bacterial meningitis in children.27

A recent groundbreaking study has published data to challenge traditional treatment durations for all types of bacterial meningitis.3 Molyneux and colleagues suggest that 5 days intravenous therapy is sufficient for the majority of bacterial causes of meningitis, but critics question if we are ‘Fine with five?’. Molyneux's study was a large international RCT involving over 1000 children over 2 months of age. It was set in low-income and middle-income countries where vaccine coverage and uptake may be considerably different to our own. NICE recently updated its guidance on bacterial meningitis and has elected not to change its pre-existing recommendations,26 which differ slightly from US recommendations28 (see table 1).

New approaches to considering treatment duration

Antibiotic therapy for infections in children needs to be individualised for each specific child. If a child is not improving, the clinical condition and treatment should be reviewed and/or discussed with an expert but certainly not discontinued, even if the ‘course’ is due to end. However, slavishly completing the recommended course of antibiotics in a child who is well may not be appropriate. Many of the newer studies on treatment duration advocate shortened courses with certain caveats, namely, clinical and/or laboratory markers of improvement. Some authors suggest policies that actively guide treatment duration by clinical judgement and laboratory markers such as CRP or procalcitonin.4 ,29 In adult practice, reviews of this strategy to guide antibiotic therapy of septic patients in an intensive care unit setting appear to reduce antibiotic consumption without impacting upon outcome.30 A multicentre prospective RCT using procalcitonin as a marker to guide treatment duration in possible early-onset neonatal sepsis is ongoing (see online supplementary e3).

Guidelines and clinical practice do not always follow the evidence from the few randomised trials of treatment duration of infection (see table 1). Given the heterogeneity in treatment recommendations for most infections, clinicians should make decisions based on guidelines, but also take into account the individual circumstances of each case. It may be safe and appropriate to reduce or extend treatment depending upon the clinical and/or laboratory findings as treatment progresses.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors CK and FAIR devised, wrote and edited the review.

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

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.